RWIN J. HERSKOWITZ L*n ■ |||"< a— — ■ *-^A Second Edition fJCCX^t^t GENETICS //v IRWIN H. HERSKOWITZ Hunter College, The City University of New York Second Edition mawn£ L\B RARV woods hoiU^-1. LITTLE, BROWN AND COMPANY Boston and Toronto COPYRIGH1 j 1962, 1965, Bl LITTLE, BROWN AND COMPANY Mm.) \1 I RK.IIIs Kl si RVFD. NO I'ART OF THIS BOOK \1 \1 HI REPRODUCED IN ANY FORM WITHOUT I'l RM1SSION IN WRITING FROM THI PUBLISHER. LIBRARY OF CONGRESS CATALOG CARD NO. 65-17335 FIRS I I'RIN LING Published simultaneously in Canada by Little, Brown & Company {Canada) Limited PRINTED IN THE UNITED STATES OF AMERICA PREFACE S Iince the beginning of this cen- tury the science of genetics has had a spectacular growth. The discovery of basic principles and the application of these principles are occurring at an ever-increasing rate. It is generally agreed that a knowledge of genetics is essen- tial for an understanding of present and fu- ture biology. The impact of genetics is not restricted, however, to professional teachers and research workers in pure and applied biology, nor to physicians and dentists. More and more students whose major in- terest is in psychology, biochemistry, chem- istry, biophysics, physics, or mathematics find that the study of genetics offers new and challenging opportunities in these various fields. How can a text for an introductory course in genetics be best organized to serve stu- dents with such varied interests? An intro- ductory text must provide the reader with an understanding of the nature of the genetic material, for this knowledge is a prerequisite for a fruitful genetic approach to the solution of problems in biology and all the other fields mentioned. Accordingly, insofar as possible, the subject matter of this book is arranged so that principles dealing with the nature of the genetic material are separated from the applications of these principles. The nature of the genetic material is studied through the use of the operations or methods of recombination, mutation, chemistry and physics, replication, and function. The pres- entation is designed to encourage the reader to use his powers of inductive reasoning to arrive at the primary generalizations of genetics on the basis of experimental evi- dence. Whenever feasible, genetic principles are derived scientifically — by recognizing and stating a problem, designing appropriate experiments to test hypotheses, analyzing the experimental results, and drawing conclu- sions. The aim is to present genetics as a rational, organized body of knowledge. Because of its importance the introductory genetics course is being offered more fre- quently in the earlier rather than the later years of college study. Since such a course is elected more and more frequently by students who do not wish to specialize in biology, simple biological examples and ter- minology are used whenever possible, and certain biological phenomena generally un- derstood by students specializing in biology are explained in some detail. Because many students in a first course in genetics may not have an adequate background, certain as- pects of chemistry and physics important for understanding genetics are described in greater detail than in other texts. No single text can include the ways each principle of genetics apply to every plant and animal studied, or give examples of the application of each of these principles to all the different kinds of organisms. Accord- ingly, only one or a few experimentally favorable or historically important organisms are usually employed in this text to establish a principle or to illustrate an application. Additional proofs, applications, or examples are left to the instructor who, depending upon his students' training and interest, can supply other illustrations by means of lec- tures and laboratory sessions or by means of assignments to detailed accounts in other texts and in the original literature. It is hoped that the text will stimulate readers to utilize the books and journals in their libraries. The reading of genetic works in the original after studying appropriate sections of the text can be a very rewarding experience. Accordingly, references requir- VI PREFACE Ing different degrees of sophistication are given at the ends of chapters. Since the later chapters deal with recent advances in genet- ics, whose discussion may be absent from already published textbooks, more references are given to particular workers in the later than in the earlier chapters. Part of a letter by G. Mendel and the Nobel Prize Lectures presented by geneticists are included in the book as supplements. These supplements should be completely un- derstandable, or nearly so. if appropriate chapters have been read beforehand, and can serve as a review and overview of genetic principles and their applications. The cita- tions to the literature included in the Nobel Prize talks should prove especially valuable to those who wish to do additional reading on key topics. The supplements can also function to bridge the gap between the text- book and the research worker, giving the reader some idea of the history of the subject and the personalities of the people involved. The Second Edition The subject matter is presented in forty-two chapters, each ending, as before, with a sum- mary, questions for discussion, and refer- ences. An appendix. "Elementary Biomet- rical Inferences," has been added, and the supplements now include three additional Nobel Prize lectures. A description of rep- resentative life cycles and genetic maps of several higher organisms have also been added. Recent advances — for example, in human, corn, and Drosophila genetics and in our understanding of the genetic code and the regulation of gene synthesis and gene action — have required that several chapters be combined or rearranged in sequence and that new chapters be written. Additional fig- ures, photographs, problems, and references are also included. Suggestions for Use of the Book The text contains more information than is usually covered in a one-semester, introduc- tory course for undergraduates. The chap- ters or chapter sections that are marked by an asterisk do not contain principles or terminology needed to understand unmarked parts and are, therefore, optional. A one-semester lecture course (meeting about 30 to 45 periods) can be based upon ( 1 ) thirty chapters — those whose chapter numbers are unstarred, or (2) twenty-eight chapters — numbers 1 through 4 and 19 through 42. A two-semester lecture course (meeting a total of about 60 to 90 periods) can be based upon (1) the first eighteen chapters for the first semester and the last twenty-four chapters for the second semester, or (2) all thirty unstarred chapters for the first semes- ter and all starred chapters and sections for the second semester. Acknowledgments I wish to thank my wife, Reida Postrel Herskowitz, for preparing the typescript, and my sons, Ira and Joel, and my present and former students for numerous suggestions. CONTENTS 1 Genetic Material and Mitosis 1 2 Meiosis and Chromosomal Segregation 75 3 Segregation of Alleles 31 4 Independent Recombination by Nonalleles 42 5 Multiple Alleles; Multigenic Traits 57 *6 Phenotypic Effects of Gene Action 69 7 Sex Chromosomes and Sex-Linked Genes 90 *8 Sex Determination 102 9 Linkage and Crossing Over Between Genes 116 10 Gene Arrangement; Crossover Maps 131 11 Changes Involving Unbroken Chromosomes 149 12 Structural Changes in Chromosomes 164 *13 Radiation-Induced Structural Chromosome Changes 179 *14 Point Mutations 189 15 The Gene Pool; Equilibrium Factors 201 *16 Genetic Loads and Their Population Effects 216 *17 Chromosomal Rearrangements in Nature 228 *18 Races and the Origin of Species 241 19 Chemical Nature of Genes 252 20 Organization and Replication of DNA in Vivo 265 21 Replication of DNA in Vitro 279 22 Clones; Transformation; Strand Recombination in Vitro 292 23 Bacterial Mutation and Conjugation 306 24 The Episome F 317 25 Transduction 330 26 Bacteriophage; Recombination and Genetic Maps 339 27 Bacterial Episomes and Genetic Recombination 355 28 RNA as Genetic Material 363 %iii CONTENTS 29 Extranuclear Genes 369 30 The Genetic Control of Mutation 383 31 The Molecular Basis of Mutation 391 32 Gene Action and Polypeptides 404 33 Polypeptide Synthesis and RNA 42J 34 Genetic Amino Aeid Coding 436 35 Regulation of Gene Synthesis 449 36 Regulation of Gene Action — Operons 457 37 Regulation of Gene Action — Gene Control Systems in Maize 465 38 Regulation of Gene Action — Position Effect in Drosophila 473 39 Regulation of Gene Action — Dosage Compensation 484 40 Regulation of Gene Action — Its Molecular Basis in Higher Organisms 492 41 Regulation of Gene Action — Growth, Differentiation, and Development 501 42 The Origin and Evolution of Genetic Material 509 Appendix — Elementary Biometrical Inferences Supplements 5-7 519 I Part of a Letter (1867 II Nobel Prize Lecture ( 1 III Nobel Prize Lecture ( 1 IV Nobel Prize Lecture ( 1 V Nobel Prize Lecture ( 1 VI Nobel Prize Lecture ( 1 VII Nobel Prize Lecture ( 1 VIII Nobel Prize Lecture ( 1 IX Nobel Prize Lecture ( 1 X Nobel Prize Lecture ( 1 Author Index 541 Subject Index 546 ) from Gregor Mendel to C. Nageli s-9 934) of Thomas Hunt Morgan s-15 946) of Hermann Joseph Muller 5-79 962) of Maurice H. F. Wilkins s-31 959) of Arthur Romberg s-60 958) of George Wells Beadle s-75 958) of Edward Lawrie Tatum s-88 958) of Joshua Lederberg s-98 962) of James Dewey Watson 5-77/ 962) of Francis H. C. Crick s-135 The essential feature of the operational viewpoint is that an object or phenomenon under experimental investigation cannot usefully be defined in terms of assumed properties beyond experimental determi- nation but rather must be defined in terms of the actual operations that may be applied in dealing with it. . . . What is a gene in operational terms? L. J. Stadler, "The Gene," Science, 120:81 1-819, 1954 Must we geneticists become bacteriologists, physiological chemists and physicists, simultaneously with being zoologists and botanists? Let us hope so. H. J. Muller, "Variation Due to Change in the Individual Gene," American Naturalist, 56:32-50, 1922 GENETICS Chapter 1 GENETIC MATERIAL AND MITOSIS S urely each of us has observed that we are the same kind of creatures as our parents. They gave rise to us, other humans — not to a plant, or a fish, or a bird. Let us start, therefore, by assuming the existence of some intrinsic factor which determines that hu- mans shall beget humans, and let us call this inborn factor for the genesis of like from like the genetic factor. Since each plant and animal produces offspring of its own kind, or species, we can generalize and hypothesize that every species of organism has such a built-in genetic factor. But it must also be admitted that the genetic factors for dog, for apple tree, and for man all differ in some way in order to produce such different or- ganisms as end products. In addition to basic likenesses within the species, each person is similar to and differ- ent from his parents in respect to certain details. What is the basis for this? If par- ents and offspring have similar caloric in- takes, all will weigh more nearly alike at a comparable age than if their caloric intakes were different. Apparently, then, environ- ment in which parents and children live can sometimes be the cause of their similarities and differences. But are all similarities and differences among human beings produced by environment, or does the genetic factor presumed responsible for like begetting like play a role in their production? 1 In trying to formulate the answer to this question, it may be helpful to consider the results of certain studies with bean plants.1 The particular kind of bean plant concerned reproduces sexually, a single plant perform- ing the functions both of male and female parent. For the present, assume that the genetic ) actor is transmitted from the parent to the offspring, and that the transmitted factor must be the same as that of the par- ent. Assume also that the genetic factor has a natural rather than a supernatural or spiritual basis. If the genetic factor has a natural basis then it ought to have a material basis and have chemical and or physical properties, as have other material things. One is led, therefore, to postulate the exist- ence of genetic material. Genetic Material Consider a particular bean. When the plant grown from this seed produces offspring beans (Figure 1-1 A), these are found to vary in size, some being very small, some small, and some medium. According to the assumptions made, these beans all have the same type of genetic material or genetic con- stitution— or genotype. The simplest ex- planation one can offer for the size differ- ences is that they were caused by environ- mental differences occurring during seed for- mation. This view can be tested by growing each of the beans and scoring the size of seeds each produces. When this is done, each bean is also found to produce offspring beans of very small, small, and medium sizes, regardless of the size of the parent seed itself. This test can be repeated gen- eration after generation with the same result. Such a line of descent, whose members carry the same genotype, can be called a pure line. The manifestation of the genotype in traits or characteristics (size, in our example) is 1 Based upon W. Johannsen's experiments. See reference on p. 12. CHAPTER 1 Typical Offspring Typical Offspring /l\ /l\ /l\ PURE LINE I B /l\ /l\ PURE LINE II /l\ /l\ /i\ * • • • ^ NEW MUTANT PURE LINE OLD PURE LINE figure 1-1. Relative sizes of seeds obtained from self-fertilized bean plants. Genetic Material and Mitosis called the phenotype. Environmental differ- ences can cause the same genotype to pro- duce a variety of phenotypes, and one can conclude that the differences between the beans of a pure line are environmentally pro- duced and are not due to differences in genotype. Now consider another bean, of the same species, which gives rise to offspring beans (Figure 1-1 B) that are very large, large, and medium sized. Since each of these pro- duces offspring beans which again show the same range of phenotypes. another and dif- ferent pure line is clearly involved, within which phenotypic variability is attributable to environmental fluctuation. How can one explain the differences be- tween these two different pure lines, one producing some very small and small beans and the other producing some very large and large ones? Since all the beans are grown under the same environmental conditions, these phenotypic differences cannot be due to environmental differences; instead they must be due to genotypic differences. It must be concluded, then, that the genetic material in these two pure lines is different. How can one explain that some of the seeds in both of these genotypically different pure lines are similar — medium sized? Apparently, different genotypes have produced the same phenotype due to the influence of the en- vironment. As already mentioned, under similar en- vironmental conditions the average size of the beans produced within a pure line re- mains the same regardless of the size of the specific beans planted. That is, in the pure line first described the offspring beans have the same average size whether the very small or the medium seed is used as parent. Sim- ilarly, the average size of seed produced in the second pure line is the same whether the medium or the very large seed is the parent. In other words, selection within pure lines is futile, as expected in view of the hypoth- esis that all members of a pure line are ge- netically identical. Throughout the bean experiments de- scribed, every effort was made to keep the environment the same. This does not mean that the environment did not vary, but that it varied approximately in the same ways and to the same extent for all the groups in the study. In this particular work it happens that phenotypic variability due to the fluc- tuations of environment is not so great as to mask completely the phenotypic effect of a genetic difference. In any randomly chosen case, however, one cannot predict offhand to what degree any particular phenotype will be influenced by the genotype and by the en- vironment. Hypothetically, then, two indi- viduals of the same species can have both phenotypic similarities and phenotypic dif- ferences resulting from each of the following four combinations, as the examples indicate: 1. Identical genotypes in near-identical environments Phenotypic difference — one small and one medium sized bean from the same pure line. Phenotypic similarity — two small sized beans from the same pure line. 2. Different genotypes IN NEAR-IDENTICAL ENVIRONMENTS Phenotypic difference — one small and one large bean from genetically different pure lines. Phenotypic similarity — two medium sized beans from genetically different pure lines. 3. Identical genotypes in different environments Phenotypic difference — one bean plant grown in the light is green, while an- other grown in the dark is white, though both are from the same pure line. ( II \l' I I K 1 FIGURE 1-2. Male Siamese cat, grown under temperate conditions, showing the same pig- mentation pattern as the Himalayan rabbit. {After C. E. Keeler and V. Cobb.) Phenotypic similarity — two rabbits from a certain pure line (genetically black rabbits) both have black coats even though one individual grew at high temperatures and the other grew at low temperatures. 4. Different genotypes in different environments Phenotypic difference — a rabbit from a genetically black line, grown in a cold climate, has black fur, while a rabbit from a Himalayan line, grown under temperate conditions, is Himalayan, i.e., white except for the extremities (paws, tail, snout, and ears), which are black (see Figure 1-2). Phenotypic similarity — a rabbit from a genetically black line grown at a mod- erate temperature and a rabbit from a genetically Himalayan line grown at a cold temperature both have black fur. The final example illustrates that genotypi- cally different individuals which are pheno- typically different in one environment may become phenotypically similar when placed in different environments. The all-black Himalayan rabbit is termed a phenocopy of the genetically black rabbit. Persons who arc genetically diabetic and take insulin arc phenocopies of genetically normal persons who do not take insulin. Genetically normal embryos whose mothers are exposed to the drug thalidomide develop into phenocopies o\' genetically abnormal, phocomelic persons lacking most or all of the lour limbs. So both normal and abnormal phenotypes can be phenocopied. The case of coat color in rabbits is instruc- tive in another respect. The rabbit that is genetically black will always produce a black coat no matter what the temperature is, pro- vided the temperature is not lethal. In the case of this genotype there seems to be no range of phenotypic expression with respect to temperature variations. In the Himalayan strain, however, the situation is different, as already described in part. If grown at very high temperatures such rabbits have entirely white coats. In this case the phenotypic range of reaction, or norm of reaction, of the genotype is relatively great, varying with increasing temperature from completely black through the Himalayan pattern to completely white. We are now in a position to answer the question concerned with the basis of simi- larities and differences among offspring or between them and their parents. Extending the principles just described for beans and rabbits to all other kinds of organisms, in- cluding man, it is concluded that not only is the genetic material different in different species of organisms, but that it can also differ from one organism to another in the same species. Phenotypic similarities be- tween individuals may occur when they are carrying the same or different genotypes, and phenotypic differences between individuals may or may not be accompanied by geno- typic differences. Having agreed that genetic variation exists within as well as between species, one may now ask: How does genetic varia- Genetic Material and Mitosis tion arise? If a pure line of large beans is bred for many generations, one finds, on rare occasions, a very small bean which gives rise to offspring beans ranging from tiny to small, and which clearly make up a new. different, pure line (Figure 1-1C). What has apparently happened is that the genetic material in the pure line of large beans somehow changed to another transmissible form which henceforth caused the produc- tion of beans which are, on the average, very small. Such a change in the genotype that is transmitted to progeny may be attributed to a process called mutation. The result of mutation is a mutant, a term which is ap- plicable to either or both the genotype and the phenotype of the new kind of individual. Just as it is easy to ascribe differences be- tween dogs and cats to genetic differences, so it is often simple to tell that certain dif- ferences between lines of the same species have a genetic basis. There are many strains or breeds of pigeons, dogs, cattle, and of other domesticated animals, each of which differs from the other in phenotype. That many of these differences are due to genetic differences is proved by finding that the phenotypic differences are retained even after the different breeds are raised together gen- eration after generation in essentially iden- tical environments. Revealed in this way, the genotypes within a species are of im- mense variety. This already-present genetic variation should be kept in mind in seeking to learn more about the nature of the genetic material. In order to learn more about the genetic material, the material things comprising or- ganisms can be examined more closely, par- ticularly those substances transmitted from parent to offspring. Most types of organ- isms are composed of (usually microscopic) building blocks, or cells, plus substances that have been manufactured by cells. Such an organism begins life either as a single cell, or by the fusion of two cells into one, or as a group of nonfusing cells derived from the parents. The cell serves as the link or bridge between generations. In those cases where the new individual begins life as one cell or as a group of nonfusing cells derived from a single parent, reproduction is asexual, whereas in cases where two parents con- tribute cells, reproduction is sexual. In sex- ual reproduction two mature sex cells, or gametes, fuse in the process of fertilization into one new cell, the zygote, which is the start of a new individual. In higher animals the gametes are called egg (female) and sperm (male), and the zygote the fertilized egg. In the bean plant, as already men- tioned, male and female gametes are pro- duced in the same individual and self-fertili- zation normally occurs; in human beings the two kinds of sex cells are produced in sepa- rate individuals of different sex, so that cross- fertilization always occurs. When might the hypothesized genetic ma- terial be transferred from parent to offspring? Consider certain organisms, composed of only a single cell, which reproduce asexually by dividing into two cells. In this process the parent becomes extinct, so to speak, its individuality being replaced by two daughter cells of the same kind. Once formed, the two daughters often separate, never to meet again. In such a case, the genetic material must have been transmitted before the com- pletion of cell division. Accordingly, the cell and this process of cell division should be studied in some detail for clues concern- ing the physical basis and transmissive char- acteristics of the genetic factors. Mitosis Attention has already been called to the cellular bridge between generations. It is only via this bridge that genetic transmission may take place, at least in single-celled or- ganisms for which cell division is equivalent 6 CHAPTER 1 to reproduction. All cellular organisms are remarkably similar in the way they accom- plish cell division. Accordingly, let us ex- amine briefly certain general features of cell structure and the appearance, under the microscope, of cells undergoing division, in initiating our search for the material basis of the genotype. There are two major parts of the cell (Figure 1-3): a peripheral portion compris- ing the cytosome, containing substances mak- ing up cytoplasm, and a more central portion called the nucleus, containing nucleoplasm. In the final stages of cell division in higher plants, the cytoplasm is divided by a cell plate, whose growth starts internally and proceeds toward the periphery until the sep- aration into two daughter cells is complete. In the case of animal cells, a furrow starts at the periphery of the cell and deepens until the parent cell is cleaved into two. The degree to which the two daughter cells are identical with respect to cytoplasmic com- ponents depends upon the position of the cell plate or furrow in the parent cell. In some cases these occur in the middle of the cell, but in many other cases they are located off-center, producing daughter cells which contain very different amounts of cytoplasm. Although the cytoplasmic components of a parent cell are often distributed unequally between daughter cells, this is not true for the nuclear contents. Ordinarily, nuclear division directly precedes cytosomal division. But the nucleus does not simply separate into two parts by the formation of a furrow or cell plate. Instead, the nucleus under- goes a remarkable series of activities in order to divide; this whole process of indirect nu- clear division is called mitosis. During the time that a nucleus shows no visible evidence of mitosis, it is nevertheless very active biochemically. In appearance (Figure 1-4A), it is bounded by a nuclear membrane and is filled by a more or less homogeneous-appearing ground substance or matrix in which are located one or more small bodies, called nucleoli. The first indication that the nucleus is pre- paring to divide is the appearance in its ground substance of a mass of separate fibers (Figure 1-4B), some of which seem to be associated with the nucleoli. These fibers are called chromosomes, and their appear- ance marks the start of the first phase of mitosis, or prophase. Careful cytological observation reveals that each chromosome is in turn composed of two major delicate threads irregularly coiled about each other. Each of the paired threads within each chro- mosome is called a chromatid. As prophase continues, the chromatids within each chro- mosome become shorter and thicker and untwist from each other (Figure 1-4C). Some of the material incorporated to thicken the chromatids may be derived from the nucleoli, which are seen to become smaller. By the end of prophase (Figure 1-4D), the nucleoli and nuclear membrane have disap- peared and the chromatids have formed thick rods which begin to move actively for the first time. Active motility is not the property of the entire chromosome, however, but is restricted to a particular region of it called the centromere or kinetochore (see p. 379). The centromeres move in a particular di- rection relative to a fibrillar structure called the spindle which has been forming through- out prophase. The completed spindle has a shape similar to what is produced when one extends and separates the fingers and touches corresponding fingertips together. The wrists represent the poles of the spindle and the fingers, the spindle fibers. The chro- mosomes migrate from whatever position in the spindle region they may have, until each centromere comes to lie in a single plane perpendicular to the axis between the poles, that is, at the equatorial plane or equator Genetic Material and Mitosis n>T©17967DA'v7"""ff "0"'SeC"0n "' " "'L ^0"™« »'<"< amission. Copy, right © 1961 by Scientific American, Inc. All rights reserved.) s CHAPTER 1 of the spindle, which is represented by the plane formed where the fingertips touch. The rest of eaeli chromosome, being pas- six e. can be in an\ position in the spindle. When all the centromeres have arrived at the equatorial plane of the spindle, mitosis has reached the middle phase, or metaphase (Figure 1-4E). Until this point the chromatids of a chro- mosome are still attached to each other at or near the centromere, although elsewhere they are largely free. Next they also separate at the centromere and the two daughter centromeres suddenly move apart, one going toward one pole of the spindle, the other toward the other pole, with the rest of each chromatid, which is now recognized as a chromosome, being passively dragged along. This stage, in which the chromatids separate, move toward, and arrive at the poles as chro- mosomes, is called anaphase (Figure 1-4F). When the chromosomes have reached the poles, the last stage, or telophase, occurs (Figure 1-4G), in which the events appear to be the reverse of those that happened in prophase. Specifically, the spindle disinte- grates, a new nuclear membrane is formed around the chromosomes, and nucleoli re- appear. The chromosomes become thinner and longer and then can be seen to consist of two delicate threads (chromatids) wound one about the other. Finally, as the chromo- somes lose their visible identity, the nucleus enters the interphase, inter mitotic , or meta- bolic stage (see again Figure 1-4A). The impression may have been gained that, in one respect, the preceding general- ized account of the mitotic phases was either incomplete or misleading. It was stated that the prophase chromosome is composed of two chromatids or threads, that metaphase puts these into position for separation at anaphase, and that after separation their newly attained individuality is recognized by calling them chromosomes. But chromo- somes were defined as containing two visible threads! The question rightly asked is: does the anaphase chromosome contain the two threads that are later seen at telophase? This would be true if each chromatid some- how visibly reproduced itself between the time it was seen relatively uncoiled at pro- phase and the next time it was seen rela- tively uncoiled, at telophase. Remember that we have been discussing the replication of chromatids as detected by microscopic observation. Chromosome and chromatid replication can also be studied by other means. Let us consider some evidence re- garding chromosome replication at the chem- ical level, which may help us understand its replication at the visible level. Chromosomes ("colored bodies") are unique since they are the only objects in the cell that are made purple by the Feulgen staining technique. It is possible to measure the amount of chromosomal material by the amount of purple stain held by the chromo- somes. The amount of chromatin — Feul- gen-stainable chromosomal material — does not change between prophase and telo- phase, but doubles over a period of hours during the intermitotic stage. By the be- ginning of prophase, therefore, each chro- mosome, as revealed by its stainability, has already replicated chemically. At the visible level, however, this is not yet ap- parent, so that each of the two visible chromatids in a chromosome also contains the chemical materials for an identical chro- matid which is still not resolved as a sepa- rate thread under the microscope. This new material is unresolved either because it has not yet assumed a proper chromatid form or has done so but is so tightly paired with its sister chromatid that together they ap- pear as one strand. Before the next occa- sion when unwound threads can be seen — that is, at the telophase of the same mitosis — this replication at the visible level has al- Genetic Material and Mitosis 10 CHAPTER 1 read) been accomplished. Thus, the chem- ical replication that takes place in a given interphase is not visible in chromatid form until the succeeding telophase. What are the consequences of mitosis'.' Speaking in terms of visible structures, the chromosomal content of the parent nucleus has become repeated in each daughter nu- cleus, so that the subsequent division of the cytoplasm produces daughter cells whose chromosomal composition is identical to each other's and to that of the parent cell from which they were derived. Mitosis merely provides the cellular machinery for the exact partitioning of previously replicated chromo- somal material. The cells of different species are different in that they have different num- bers of chromosomes per nucleus, or the chromosomes differ in appearance, or both. One chromosome may differ from another in size, in stainability with various dyes, and in the position of the centromere. Most chromosomes have a single centromere which is not located terminally, i.e., at an end, and therefore separates the two arms of a chro- mosome; all chromosomes and chromatids are unbranched fibers. Examination of the kinds of chromosomes present at metaphase in sexually reproducing organisms typically reveals that for each chromosome which arrives at the equatorial plane, there is another chromosome very similar or identical in appearance which also takes a position independently in this plane. Chromosomes thus occur as pairs; the mem- bers of a pair are called homologous chro- mosomes, or homologs, whereas chromo- somes of different pairs are nonhomologous, or nonhomologs. It should be repeated that the members of a pair of homologs take their positions at mitotic metaphase inde- pendently of each other. The number of chromosomes seen in typical mitosis of the garden pea is 7 pairs; in Indian corn (maize) there are 10 pairs, in the domesticated silkworm 28, and in human beings 23; thus the chromosomes of a species are characteristic in number - as well as in form. Whatever the number o\ chromosomes present in the zygote, then, other things being equal, the same number of chromosomes will be found in every cell of a multicellular organism descended from the zygote by cell divisions in which mitosis has occurred. Chromosomes as Genetic Material The chromosomes are one of the character- istic components transmitted by all cells to daughter cells. Chromosomes reproduce themselves and are transmitted in mitosis equally to the daughter cells so that these are identical, in this respect, to each other and to their parent cell. Let us make the additional reasonable assumptions that ge- netic material arises only by the replication of pre-existing genetic material, and also that different genotypes arise only from each other by mutation, that is, by a genotype's changing to an alternative mutant form which in turn is involved in reproducing the alter- native form until it undergoes mutation. A chromosome may occasionally become visi- bly altered in certain ways; in these cases all chromosomes ordinarily derived from such a modified chromosome via mitosis have ex- actly the same alteration. Therefore, both genetic material and chromosomes are con- sidered capable of mutation and are subse- quently involved in replicating their new form. On this basis, then, we can hypoth- esize that the chromosome is, or carries, the genetic material. It has been implied that the genetic ma- terial routinely retains its individuality or integrity regardless of the nature of the en- vironment. One indirect piece of evidence has already been cited for believing this is -'See S. Makino (1951) and C. D. Darlington and E. K. Janaki-Ammal (1945). Genetic Material and Mitosis 11 true for the chromosomes — namely the in- dependence with which each chromosome arrives at metaphase. It might be thought, when the chromosomes "disappear" during interphase, that their individuality is lost and even that their contents are dispersed. That the nuclear material is not dispersed into the cytoplasm between successive mi- toses is indicated by the retention of the full amount of chromosomal material within the nucleus during interphase, insofar as re- vealed by the Feulgen staining technique. Even so, it is still possible that those com- ponents of chromosomes which remain in the nucleus become scrambled during inter- phase and later resynthesize their proper form during the next prophase. Four lines of evidence bearing on this matter can be mentioned. The first three come from study- ing the appearance of successive mitoses. It is possible to observe the relative positions of the chromosomes at late anaphase or telo- phase and also their relative positions as they enter the next prophase. When this is done, the chromosomes are found to have held the same relative positions, as expected had they retained their integrity during the intervening interphase. Second, since the nucleolus does not fragment during inter- phase, those parts of the chromosomes, called nucleolus organizers, with which the nucleolus is associated probably remain as- sociated during that interval.3 Third, it sometimes happens that two originally iden- tical homologs are modified by mutation so that each is changed in a different respect. The finding, mitosis after mitosis, that both homologs retain their separate differences is evidence that each homolog has retained its individuality cell generation after cell gen- eration. Finally, more direct evidence on the retention of chromosomal individuality during interphase is available from cells of 3 See also F. H. Ruddle (1962). larval salivary glands of certain flies. These giant cells have interphase nuclei that con- tain giant chromosomes which, though rela- tively uncoiled, are clearly equivalent to the more contracted chromosomes seen during mitosis. The number of points of similarity be- tween genetic material and chromosomes is already impressive. However, if all nuclei divide by mitosis, a gamete should contain the same number of chromosomes as the other cells derived from the original zygote; and since the zygote of any generation com- bines two gametes, the number of chromo- somes should increase in the zygotes of suc- cessive generations. One would therefore expect an increase in the amount of genetic material in successive sexual generations. This expectation is not realized, however, since one finds that all individuals of a species have a characteristic, typically stable, chromosomal content. In fact, as expected, human gametes do not contain the paired, diploid, chromosome number, that is, 23 pairs of nonhomologous chromosomes. In- stead each usually contains 23 chromosomes, one of each nonhomologous type, comprising a complete, unpaired, haploid or monoploid, set of chromosomes. The zygote, therefore, has the diploid chromosome constitution re- stored because each gamete furnishes a haploid set of chromosomes, one set con- tributed by the sperm from the father, and another set by the egg from the mother. In this way chromosomes remain as pairs, sexual generation after sexual generation, and the number of chromosomes in zygotes remains unchanged. Clearly, then, the cell divisions preceding gamete formation cannot be invariably mitotic, but must involve at some point a special mechanism for reducing the number of chromosomes from diploid to haploid. The nature of this special kind of nuclear behavior is considered in the next chapter. 12 CHAPTER 1 SUMMARY AND CONCLUSIONS Organisms are assumed to contain an intrinsic genetic factor which is responsible for like reproducing like. Ibis genetic factor is presumed to have a physical basis in genetic material. The genetic material must be different in dilferent species of organisms, and may be different in different lines or breeds of the same species. Variations in phenotype m.i\ be due to genetic or environmental differences, or both. The contribution made to phenotypic variability bj one of these two factors may be evaluated by holding the other factor as constant as possible. Genotypic differences arise by the process of mutation. The genetic material is presumably transmitted from parents to offspring by means of the cellular bridge be- tween generations, and is assumed to be self-replicating and to arise only from pre- existing genetic material. Studies of cell division in which nuclei divide mitotically reveal that, of all cellular components, the chromosome is the structure most likely to serve as the genetic material or as its carrier. This hypothesis receives support from several of the properties of chromosomes which parallel established or assumed properties of genetic material. Chromosomes come only from pre-existing chromosomes; different species have differ- ent chromosomal compositions; the chromosome content is identical both quantitatively and qualitatively in each cell of a line produced by asexual reproduction; each chromo- some retains its individuality, mitotic cell generation after mitotic cell generation, re- gardless of the nature of the other chromosomes present; chromosomes can occasionally mutate, the mutant chromosome then replicating the mutant form. REFERENCES Darlington, C. D., and Janaki-Ammal, E. K., Chromosome Atlas of Cultivated Plants, London: Allen and Unwin, 1945. Flemming. W., 1 879. "Contributions to the Knowledge of the Cell and its Life Phe- nomena," as abridged and translated in Great Experiments in Biology, Gabriel, M. L., and Fogel, S. (Eds.), Englewood Cliffs, N.J.: Prentice-Hall, 1955, pp. 240-245. Johannsen, W., 1909. Elemente der exakten Erblichkeitslehre. Jena. See also a trans- lation of the summary and conclusions of his 1903 paper, "Heredity in Populations and Pure Lines," in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, 1959, pp. 20-26. Makino, S., An Atlas of Chromosome Numbers in Animals, Ames, Iowa: Iowa State College Press, 1951. Mazia, D., "Mitosis and the Physiology of Cell Division," in The Cell, Vol. 3, Meiosis and Mitosis, pp. 77-412, Brachet, J., and Mirsky, A. E. (Eds.), New York: Aca- demic Press, 1961 . Ruddle, F. H., "Nuclear Bleb: A Stable Interphase Marker in Established Lines of Cells in Vitro," J. Nat. Cancer Inst., 28:1247-1251, 1962. Schrader, F., Mitosis: the Movement of Chromosomes in Cell Division, New York: Columbia University Press, 1953. Scientific American, Sept. 1961, Vol. 205, No. 3, "The Living Cell," articles by J. Brachet and D. Mazia. Genetic Material and Mitosis 13 WlLHELM LUDWIG JOHANNSEN (1861-1926). (From Genetics, vol. 8, p. 1, 1923.) Spector, W. S. (Ed.), "Chromosome Numbers," in Handbook of Biological Data, Phila- delphia: Saunders, 1956, pp. 92-96. Swanson. C. P., Cytology and Cytogenetics, Englewood Cliffs, N.J.: Prentice-Hall, 1957. Swanson, C. P.. The Cell, 2nd Ed., Englewood Cliffs, N.J.: Prentice-Hall, 1964. QUESTIONS FOR DISCUSSION 1.1. Does the phenotype of one generation have any effect upon the genotype of the next? Explain. 1.2. Evaluate the thesis that the genotype is more important to organisms than is the environment. 1.3. Is the environment for two organisms ever identical? Explain. 1.4. What is meant by an operational definition? 1.5. Define the genetic factor. Have you given an operational or a nonoperational definition? Explain. 1.6. When the same similarities or differences in phenotype can be produced by either the environment or the genotype, can one ever be sure which is the determining factor? Explain. 1.7. What evidence can you give to support the view that genetic material is trans- mitted from parent to offspring? Do you think this evidence constitutes conclu- sive proof of transmission? Explain. 1.8. What conclusions can you reach regarding the genetic factor in Himalayan rabbits and in Siamese cats? 14 CHAPTER 1 1.9. Assume the genetic factor has a supernatural basis. Could we learn anything about it b> the scientific method of investigation? Explain. l.K). Do you think human beings provide good material for the study of the genetic factor? Explain. 1. 11. What si/e limitations can you give to the genetic material? 1.12. Is the existence o[ genetic material presumed or proved? Explain. LI 3. What do the bean experiments reveal about genetic material? 1.14. August Weismann ( 1834-1914) cut off the tails of mice for a series of genera- tions and found that tail length remained normal each new generation. Why are these experiments significant? 1.15. What are the consequences of mitosis? 1.16. For each of the properties of chromosomes listed in the Summary and Conclu- sions, state the corresponding property of the genetic material and identify it as one that is either proved or assumed. 1.17. [f the chromosomes serve as the genetic material, each cell of the body derived by mitosis should carry the same genotype. Describe how you would test this idea, using a multicellular plant. 1.18. What are the advantages or disadvantages of chromosome coiling? 1.19. Can you imagine a spindle which is too small for normal cell division? Explain. 1.20. Suppose certain nuclei normally do not divide with the aid of a spindle. How would this affect your ideas about genetic material? 1.21. Discuss the statement that all cell divisions are normally mitotic. 1.22. Differentiate between replication of chromatids and of chromosomal material. 1.23. List the events that presumably take place before a given telophase chromosome can give rise to a chromosome made entirely of chromosomal material not yet synthesized. 1 .24. Why should the peas in a pod be similar? Different? 1.25. What do each of the following observations mean with regard to the origin and/or integrity of chromosomal material? (a) Nonhomologous chromosomes retain their characteristic morphological dif- ferences mitosis after mitosis. (b) A loss or gain of one entire chromosome occurs occasionally, with all mitotic descendants having the same aberration. (c) T. Boveri noted in Ascaris cleavage that sister cells entering the next mitosis often have a mirror-image arrangement of their chromosomes. 1.26. What conclusions can you draw from the fact that there are three genotypically different kinds of Indian corn: one always has red kernels, one always has yellow kernels, and one has kernels which are yellow but become red if exposed to sunlight? Chapter 2 MEIOSIS AND CHROMOSOMAL SEGREGATION H "ow do both male and female gametes come to contain only one set of chromo- somes, composed of one member of each pair of chromosomes found in the nucleus of an ordinary body, or somatic, cell? If gametes were produced by regular mitotic division, they would be diploid. The reduc- tion from two sets to one is brought about by another type of indirect nuclear process, called meiosis, which actually requires two successive nuclear divisions to accomplish its result. Meiosis To render the cytological description of the meiotic process more meaningful, several as- sumptions will be made. Suppose that the processes directing the division of the nucleus act especially early in the case of meiosis, before the chromosomes have attained the degree of coiling first seen in mitotic pro- phase. Suppose further that a relatively more uncoiled state of the chromosome is, under these conditions, associated with an especially strong attraction between homo- logs of like chromosome parts for like parts and that this attractive force extends over considerable, though still microscopic, dis- tances. Then, with one additional novelty yet to be described, the meiotic process will occur in the following predictable way when the chromosomes, without further replica- 15 tion, undergo two successive mitotic divi- sions. In prophase of the first meiotic division, just as in mitotic prophase, each chromo- some contains two chromatids plus an equal amount of chromosomal material not yet visible as chromatids (see p. 8). But now, because of the early onset of nuclear division, homologous chromosomes pair point for corresponding point (making a bundle of four chromatids plus an equal amount of future chromatid material). Ac- cordingly, the chromosomes proceed as pairs to the equator of the spindle for the meta- phase. (Recall that in mitosis, on the other hand, each chromosome of the two sets pres- ent goes to the equator of the spindle in- dependently of its homologous chromo- some.) At anaphase the members of a pair separate and go to opposite poles, each ana- phase chromosome still containing two chro- matids plus an equivalent amount of future chromatid material. In the interphase after the first telophase, no synthesis of future chromatid material takes place since what was made in the previous interphase had not been used to make visible chromatids in the first meiotic division. The second meiotic division may start at any time and proceed as a typical mitosis. In the second meiotic prophase each chromosome contains two chromatids and the material for two future chromatids. Each chromosome pro- ceeds to metaphase independently; at ana- phase the two chromatids separate and go to opposite poles of the spindle (after sep- aration the chromatids may be called chro- mosomes). By telophase the future chro- matid becomes visible; thus each telophase chromosome contains two chromatids. Although mitosis always involves chromo- some duplication and separation alternately, one duplication is followed by two separa- tions in meiosis. The result is the mainte- nance of the diploid chromosome condition IT) CHAPTER 2 in mitosis, but a reduction from the diploid to the haploid condition upon the completion of meiosis. I et us examine in some detail the actual meiotic process as seen under the micro- scope (Figure 2-1 ). Prophase of the first meiotic division (prophase I ) is of long dura- tion, as compared to mitotic prophase, and is divided into several substages, each with its own distinguishing characteristics. 1 . As they emerge from the interdivision phase the chromosomes are long and thin, more so than in the earliest prophase of mitosis. This is the leptonema (thin thread) stage of prophase I. 2. Next the thin threads pair with each other in a process called synapsis. This pairing is very exact, being not merely between homologous chromosomes, but be- tween exactly corresponding individual points of the homologs.1 Synapsis proceeds zipperwise until the two homologs are com- pletely apposed. This is the zygonema (join- ing thread) stage. 3. The apposition of homologs becomes so tight that it is difficult to identify two sep- arate chromosomes in the pachynema (thick thread) stage (Figure 2-2A). 4. Next, the tight pairing of the pachy- nema is relaxed, whereupon it can be clearly seen in the diplonema (double thread) stage that each pair of synapsed chromosomes contains four threads, two visible chromatids for each chromosome (Figure 2-2B, C). A pair of synapsed chromosomes is called a bivalent (composed of two univalents) when referring to chromosomes, but is called a tetrad (composed of two dyads or four monads) when referring to cytologically de- tectable chromatids. Although the chromatids in a tetrad sepa- rate from each other in pairs here and there, they are all still in close contact with each other elsewhere. Each place where the four 1 See H. Jehle (1963) for a discussion of the physical basis for the attraction of like for like. figure 2-1 . Meiosis in the lily (Courtesy of R. E. Cleland.) -general view. chromatids are still held together is called a chiasma (Greek, cross; plural, chiasmata) (Figure 2-3 A). In a chiasma the two chro- matids that synapse to make a pair on one side of the point of contact, separate at that point and synapse with other partners on the other side of the contact point; i.e., the partners making up two synapsed pairs of chromatids are different on the two sides of the place of contact (Figure 2-3B). A tetrad typically has at least one chiasma. The occurrence of a chiasma assures that the univalents are held together. When sev- eral chiasmata occur per bivalent, loops are figure 2-2 (opposite) . Meiosis in the lily. The leptonema and zygonema stages of pro- phase I have been omitted. (Courtesy of R. E. Cleland.) (By permission of McGraw-Hill Book Co., Inc., from Study Guide and Work- book for Genetics, by I. H. Herskowit: :. copy- right I960.) Meiosis and Chromosomal Segregation 17 A. PACHYNEMA DIPLONEMA D. DIAKINESIS PROPHASE I TH 4ifi&Jv I , L. ${ 'te^ E. METAPHASE I F. ANAPHASE I G. INTERPHASE I (Equatorial View) (Side View) ■Mwr H. METAPHASE II (Side View) I. TELOPHASE II is CHAPTER 2 formed, adjacent ones at right angles to each other. As diplonema continues, the chromosomes become shorter and thicker, more compacted than they ever become in mitosis. 5. In some animals, during the formation oi female gametes especially, a diffuse or growth stage follows diplonema, in which the chromosomes ami nucleus revert to the appearance kmnd in a nondividing cell. During this stage a great amount of cyto- plasmic growth takes place. In human be- ings this stage may last for decades, after which the rest of meiosis occurs and mature eggs ready for ovulation are produced. 6. Diakinesis (Figure 2-2D) is charac- terized by the maximal contraction of diplo- nema chromosomes, or by maximal recon- traction of the chromosomes which had entered a diffuse stage. By the end of this stage nucleoli and nuclear membrane have disappeared, the spindle has formed, and prophase I is completed. Metapha.se I (Figure 2-2E) is attained by the movement of chromosomes to the midspindle, as in mitosis, except that they move as bivalents, made up of a tetrad of chromatids still held together by chiasmata. Between diplonema and metaphase I the chiasmata move toward the end of the chro- mosome arms, that is. away from the centro- mere, especially if the bivalent is short. As a consequence of this chiasma terminaliza- tion the number of chiasmata present at metaphase I may be less than it was at diplo- nema. During anaphase I (Figure 2-2F) the uni- valents in each bivalent separate from each other at the region of the centromere and proceed to opposite poles of the spindle. This movement completely terminalizes all remaining chiasmata. The dyad nature of each univalent is readily seen in the figure. In telophase I the two daughter nuclei are formed, and interphase I (Figure 2-2G) follows. The length of interphase I varies in different organisms. Each daughter nucleus undergoes the sec- ond meiotic division, which proceeds as ex- pected from mitosis. In prophase II, each univalent (equivalent to a chromosome with its two visible chromatids) contracts; at metaphase II (Figure 2-2H) each lines up at the equator of the spindle independently; at anaphase II the members of a dyad sepa- rate and go to opposite poles as monads (each equivalent to a single chromosome, since now each contains two visible chro- matids). Because two nuclei undergo this second division, four nuclei are formed at telophase II (Figure 2-21). Photographs of the meiotic process in corn can be seen in Figure 2-4 (pp. 20-21 ). Chromosomal Segregation Consider next the consequences of meiosis. The organism undergoing meiosis starts its existence as a zygote produced by fertiliza- tion involving the union of two haploid sets of chromosomes, one maternal and one paternal. When meiosis is completed the diploid, paired, chromosome number is re- duced to the haploid, unpaired, chromosome number. Since any postmeiotic nucleus nor- mally contains only one representative of any given pair of chromosomes present in a pre- meiotic nucleus, chromosome segregation has occurred. Two questions come to mind at this point. First, is the haploid set of chromosomes, or genome, in a gamete com- posed of replicas of all the chromosomes contributed by the female parent or of all those contributed by the male parent? For typical meiosis, the answer depends upon two events. The first of these is the manner in which the centromeres of the bivalents arrange themselves at the equator of the spindle at metaphase I. Relative to the poles of the spindle, each bivalent ar- ranges itself at the equator independently of Meiosis and Chromosomal Segregation 19 other bivalents, so that it is purely a matter of chance whether the copy of the ma- ternally-derived chromosome will go to one specified pole and the copy of the paternally- derived chromosome to the other, or vice versa. Consider the distribution of two bi- valents, for example. Since there are many cells undergoing meiosis in any sex organ, or gonad, at metaphase I, approximately half of these will have the two paternal univalents going to one pole and the two maternal uni- valents going to the other pole at anaphase I, and approximately half will have one ma- ternal and one paternal going to one pole and one paternal and one maternal to the other. As a result, the chromosomal con- tent of a pool of all the haploid nuclei pres- ent at the completion of meiosis will be 25% paternal + paternal; 25% maternal + ma- ternal; 25% paternal -f- maternal; 25% ma- ternal + paternal. Because the centromeres of each bivalent line up at metaphase I in one direction with a frequency equal to that in the other and because each bivalent does so independently of all other bivalents, we see that the segregation which follows occurs independently for different pairs of chromo- somes.2 Note also, from the fate of two bi- valents, that 50% of haploid products have the same combinations of nonhomologous chromosomes as entered the individual in the parental gametes, therefore retaining the old or parental combinations, whereas 50% of haploid products carry new, nonparental combinations or recombinations . Let us defer considering the genetic im- plications of these conclusions until we have considered the second question, which also bears upon the maternal-paternal chromo- some content of gametes: our answer may modify the conclusions just reached. Is a chromosome in a gamete, in fact, a com- pletely uniparental replica, or has it a bi- 2 As shown by E. E. Carothers (1921). B figure 2—3. Lily diplonema showing chro- matids (1-4) with different synaptic partners on different sides of a chiasma. (Courtesy of R. E. Cleland.) parental derivation? The latter situation would obtain if one segment of a gametic chromosome were a copy of a portion of one homolog and another segment a copy of a portion of the other homolog. Considerable evidence exists that some time between the onset of meiosis and diplo- nema a cytologically undetected event occurs which results in two of the four chromatids in a tetrad having segments which are bi- parental copies, exactly reciprocal in con- tent. Thus, if one biparental segment of a chromatid has a linear sequence that is ma- ternal-paternal, the other is paternal-mater- nal in composition. The other two chro- 20 CHAPTER 2 2 ^ X 4 5rX. t ^N» -a -a < LU z >- X u < a. <7 * Ok ■ Meiosis and Chromosomal Segregation 21 Q. O | S r ^ ' » « > !<■ ^> C a J <3 =3 -2 3 - (parents of the second generation) and reproduce by self-fertilization to yield F2 progeny. When this is done, and sufficient numbers of F2 are obtained from each P2 plant, one finds among the offspring of every P2 that some are colored and some are white. In terms of genetic material, these FL> must carry, respectively, C or c. It is no surprise that some FL. contain C, but where did the c come from? In these cases, one could at first suppose either that c arose spontaneously from some non-genetic origin or that C mutated to c. The first explanation can be bypassed in view of the previous assump- tions (p. 10) that genetic material can arise only from pre-existing genetic material, and that this material is self-reproducing (self- replicating). The second explanation can be eliminated by the observation for the pure line containing C that mutations to c are thousands of times rarer than the occur- rence of c among the FL>. If the P-(FX) are genotypically like pure-line C individuals, as assumed, mutation cannot be the explana- tion for the difference in breeding behavior between P,C and PLC. The results of the bean experiments in Chapter 1 are consistent with the view that the genetic material in any individual is a Segregation of Alleles 33 single indivisible unit. In the absence of a simpler explanation for the present findings with peas, it seems necessary to postulate that the genetic material is not always com- posed of a single indivisible unit. The ap- pearance of c in F2 can be explained by making the more complex assumption that each PL.(Fi) individual contains not only C but c as well; in other words, that in some individuals the genetic material contains two units. Let us use the word gene to refer to a unit or restricted portion of the genetic material. But, if it is assumed that there is a pair of genes in each PL., then all other individuals in our experiment must be as- sumed to have a pair of genes, too. For, in science, we adhere to the law of parsimony {Occam's rule), which states that one must use the minimal number of hypotheses or assumptions to explain a given set of obser- vations. Instead of having some individuals with paired genes and others without pairs, then, all are assumed to have a pair of genes in their genetic material. Accordingly, the two pure lines and the Pi must have been CC and cc, and all Ft must have been Cc. Those F2 which are colorless must be cc. Attention is called to the individuals in F2 that are cc. These have colorless flowers phenotypically identical with those of the original colorless pure line used in the P^ In fact, crosses of F2 colorless individuals either with themselves or with any other colorless individual (F2, or pure line) pro- duce all colorless progeny. In other words, F- cc individuals are genotypically just as pure with respect to the trait under consid- eration as are pure-line individuals. This is true despite the fact that both c's in the F2 had been carried in F] individuals in which C was the other member of the pair of genes. We conclude, therefore, that when c is transmitted to the F2, it is uncontam- inated, or untainted, by having been in the presence of C in the Ft, even though c had not been expressed in any noticeable way in the phenotype of the F, individuals. We can generalize this conclusion and state that the nature and transmission of any gene is uninfluenced by whatever its partner gene may be. The members of a gene pair are said to be alleles (partner genes), a term also applied to alternative forms of a given gene. Since each P2 produced colored and color- less F2 offspring, each P2 had the genotype Cc composed necessarily of C from the CC P, and c from the cc P^ This specifies that one and only one member of a pair of genes in a parent is transmitted to each offspring, so that in the transmission process the members of a parental pair of genes must become separated, or segregated, from each other. The paired, or diploid, status of the genes becomes unpaired, or haploid, during transmission; but diploidy is restored in the offspring because a haploid genotype is con- tributed to it by each parent. Accepting the hypothesis that paired genes are segregated by the time they are trans- mitted to progeny, are the two alleles in a parent equally likely to be transmitted to offspring? The F2 produced by self-fertiliza- tion of Fx Cc demonstrate that both genes of a given individual are transmissible. Let us test the hypothesis that both members of this pair of alleles are equally transmissible. If so, the Fx male parent (or part) would contribute C one half the time and c the other half; similarly the Fi female parent (or part) would contribute C half the time and c the other half. Finally, assume di- ploidy is restored at random; that is, the haploid gene contributed to the offspring by one parent is uninfluenced by the haploid gene contributed by the other parent. Ac- cordingly, an offspring that receives C from the male (50% of offspring) will have an equal chance of receiving C or c from the female, so that of all offspring 25% will be CC and 25%- Cc. Those offspring receiv- ing c from the male (50% of offspring) will 34 ( II \l'l p.r 3 P CC x cc (Cross - fertilization) I J O. all C all c (Gametes) all Cc Cc x Cc (Self-fertilization of F, A A Vi C, Vt c Vi C, Vic Male gametes Vi C 7* c »/« cc '/« Cc '/« Cc '/« cc Female '7 C gametes ,/ 3 Vi c or ' 4 CC »/j Cc V, cc when * t P3 self- breeds breeds breeds fertilized like like like P, CC P Cc P cc figure 3—1. Genotypic model proposed to explain the phenotypic results of certain crosses involving colored and colorless flowers in pea plants. also have an equal chance of receiving C or c from the female, so that the contribution to all the offspring genotypes will be 25% Cc and 25% cc from this source. On this basis the FL. would be predicted to contain 25% of individuals that arc CC. 50% that are Cc, and 25% cc. This expectation can be expressed as relative frequencies in several ways: % CC: '{> Cc % cc, or 1 CC: 2 Cc : 1 cc, or .25 CC : .50 Cc : .25 cc. As al- ready reasoned CC and Cc are phenotyp- ically indistinguishable, having colored flow- ers, so that phenotypically 75% of the F2 would be colored and 25% would be color- less. What is their relative frequency in the F2 actually observed? Although a penny has in theory a 5()r i chance of falling head up and a 50% chance oi falling tail up, obviously a sufficiently large Dumber of tosses is required to obtain approximately 5095 heads, 509? tails. Sim- ilarly, an accurate test of the theoretical ex- pectation of 75% colored and 25% color- less will be obtained only if a sufficiently large sample of offspring is scored. Instead of scoring the offspring of just one P2, then, the results for the offspring of all P2 should be totalled. When this is done, the actual F2 results (among 929 plants. 75.9% are colored and 24. 1 % colorless ) are very close to the expectation. It should be emphasized that the concept of paired, untaintable, segregating genes has not depended upon obtaining any particular phenotypic ratio for the F2. Granting these characteristics of the genetic material, ob- taining or not obtaining the phenotypic ratio :;, colored to % colorless merely tests the suppositions ( 1 ) that there is an equal chance for offspring to receive either haploid product of gene segregation from a parent, and (2) that the haploid products from dif- ferent parents come together at random to restore the diploid condition. If all the assumptions so far made about genetic material are correct, the 75% of F2 that are colored should have one of two genotypes: Vs of them should be CC, breed- ing like pure-line CC individuals, and % should be Cc. breeding like the F, Cc indi- viduals. Accordingly, each F2 colored plant is permitted to self-fertilize and, in fact, very nearly K produce only colored F:i. whereas -;. produce F;! of both colored and colorless types. The theoretical genotypic ratio ex- pected in the FL., % CC:'1^ ^c lA cc> is, in this way, fully confirmed in experience. The gene model proposed to explain these pheno- typic results is summarized in Figure 3—1. It is convenient to introduce two additional terms at this time. A homozygote is an individual that is pure with respect to the Segregation of Alleles 35 paired genes in question, like CC or cc, whereas a heterozygote, or hybrid, is im- pure in this respect, like Cc. An independent test of all the genetic hypotheses presented so far can be made in the following way. F, colored plants are crossed with colorless plants, this cross being symbolized genetically: F, Cc X cc. As the result of segregation half of the offspring should receive C and half c from the Cc parent, and all should receive c from the cc parent. The genotypes of the offspring from this cross should be, theoretically, Cc 50% of the time and cc 50% of the time, and the expected phenotypic ratio should be, there- fore, 1 {» colored: l/2 colorless. This expecta- tion is fulfilled experimentally (85 colored: 81 colorless). Are the principles just established gen- erally applicable? Thus far they apply strictly only to the genetic determination of flower color in garden peas. All these ideas can be tested six additional times, using six other traits in garden peas, each of which occurs in two clearcut alternatives and ful- fills the prerequisites for suitability already described. In each case, when two appro- priate pure lines are crossed, the Fi hybrids produced are phenotypically uniform, as be- fore. Moreover, self-fertilization of the F, produces Fj in approximately the expected 1:2:1 genotypic ratio. Recall that the Cc phenotype is indistin- guishable from CC. In Cc individuals the phenotypic expression of c is masked by the expression of C. The ability of a gene to express itself phenotypically in the presence of a different allele is described in terms of dominance. In the case of flower color, C produces a dominant effect when present with c, whose effect is, accordingly, recessive. It should be emphasized that our concept of the gene does not depend upon the occur- rence or nonoccurrence of dominance. In- deed, testing our genetic postulates has been made more complicated by the fact that the effect of C is, for all intents and purposes, completely dominant to that of c. The Fi Cc expressed only C, and the presence of c was detected only by breeding F, individuals, and observing cc progeny. Only by breeding the colored FL> were we able to determine that ' :{ were CC and % Cc. Dominance, then, re- fers to the phenotypic expression of genes in heterozygous condition and has no rela- tion to their integrity, replication, or mech- anism of transmission. For convenience, dominant and recessive will be used here- after to refer to genes, but the precise mean- ing of these terms should always be kept in mind. As mentioned, six other traits have been used to test the general applicability of the gene concept. In each case it happened that one allele was dominant to the alternative one in the hybrid. It is tempting to con- clude that dominance is a universal phe- nomenon since it was found to hold for each of seven different traits based on genes in the garden pea. Before making this de- cision, however, examine the results with re- gard to feather color of breeding certain chickens. Here black X white produces blue-gray F]. Mating two blue-gray Fx pro- duces in F_. ]4 black, 1/> blue-gray, and % white. In this case complete dominance does not occur, so that complete dominance is not a rule for the phenotypic expression of alleles in heterozygotes. Whenever domi- nance is incomplete or absent, genotypes can be stated with certainty from a knowledge of phenotypes. Cross-fertilization made it possible to show that genes occur as pairs, which be- come unpaired after segregation, then re- combine to form pairs in the offspring. In other words, the view that the genetic ma- terial contains separable paired units is based upon the recombination which these units undergo in cross-fertilizing species. The meaning of the term genetic recombination ought to be considered at this point. The 36 CHAPTER 3 genetic units themselves are not required to undergo novel changes (mutations) when undergoing recombination. That is. the types of genes present in a genetically re- combinant individual existed before recom- bination. Given an individual whose gene pair is AA', segregation followed by self- fertilization may produce A A' again. This genotype is not considered to be a genetic recombination, but rather a reconstitution of the original arrangement of the units. The self-fertilization under discussion may also produce A A or A' A'. These represent two new genetic combinations relative to the parental combination, and are considered to be genetic recombinations. Accordingly, when events lead to the production of "old" combinations and "new" combinations of genes, only the latter type of grouping is called genetic recombination. This usage is reasonable in view of the importance that new combinations have for our understand- ing of genetic material (it was possible to derive the principle of gene segregation only because new combinations of genes were produced via sexual reproduction). Ac- O NORMAL FEMALE V I I NORMAL MALE r>f \/ UNKNOWN 9 AFFECTED W I AFFECTED ^ f") — i— I Marriage Line QrO sn Offspring Line Offspring, in order of birth | I, to r. ) Dizygotic Twins Monozygotic Twins FIGURE 3-2. Symbols used in human pedigrees. cordingly, genetic recombination should be identified with any reassortment or regroup- ing of genes which results in a new arrange- ment of them. Any process, like segregation or fertilization, that has the potential of pro- ducing new arrangements of genetic units is. therefore, a mechanism for genetic recom- bination. The phenotypic results of the experiments discussed in Chapter 1 led us to hypothesize the existence of genetic material which is self-replicating, mutable, and transmissible. The pea plant experiments reveal that the genetic material can be divided into a pair of units by means of the operation or tech- nique of genetic recombination. Techniques or operations other than recombination can be employed to study the nature of the ge- netic material. Should other operations also divide the total genetic material into units, this would not necessarily mean that the units revealed by different operations are equivalent. Thus, to use a nongenetic anal- ogy, a book (equivalent to the total genetic material ) can be described operationally in terms of words, letters, numbers, pages, il- lustrations, and so forth. Each operation reveals something about the book, but the different units by which it is described are necessarily neither identical nor mutually inclusive. What bearing has the discovery of segre- gating alleles upon the hypothesis that the chromosomes represent genetic material? Both genes and genomes are unpaired in gametes and paired in zygotes. Can a genome be the physical basis of a gene? Though the gametes contain a single genome, this is usually constituted (ignoring the ex- changes leading to chiasmata, for the mo- ment) of replicas of some maternal and some paternal chromosomes. Since segre- gated genes are uncontaminated by their former alleles, being just as pure in the gametes as they were when they entered the organism at fertilization, one can reject the Segregation of Alleles 37 CM^^ O o ^t^6]u66t666to6 i6W55c57) c56i 65 figure 3-3. A pedigree of albinism in man. 65 hypothesis that a genome is the material equivalent of a gene. Nor can an entire chromosome be identified with a gene, since any given gametic chromosome is typically constituted of some maternally-derived and some paternally-derived parts, as a conse- quence of the exchanges leading to chias- mata. However, the possibility still remains that the gene is associated with a particular segment of a chromosome which is so short that it cannot undergo an exchange leading to a chiasma with a corresponding segment on a homologous chromosome. Such a seg- ment would always retain its pure maternal or pure paternal constitution after meiosis. Accordingly, the maximal size of such an uncontaminable segment would equal the maximal size of the gene. * Segregation of Alleles in Man The genetic principles discovered in garden peas should also hold equally well for any other sexually reproducing species, including man. Because we naturally have great in- terest in ourselves, let us consider some human traits which may be based upon the action of a single pair of genes. The study of human genetics is complicated by the fact that, unlike other species of plants and ani- * Throughout this book, the asterisk indicates an optional chapter or section. mals. our species is not bred experimentally. Because of this scientific difficulty special methods of investigation have to be em- ployed. These include the pedigree, family, population, and twin methods. The present discussion deals primarily with the first two of these methods. The pedigree method uses phenotypic records of families (family trees or gene- alogies) extending over several generations. In recording pedigrees certain symbols are used by convention (Figure 3-2). In a pedigree chart a square or { represents a male, a circle or 9 represents a female: filled-in symbols represent persons affected by the anomaly under discussion. In con- trast, the family method utilizes the pheno- types only of parents and their offspring: that is. it uses data that span only one gen- eration. Albinism, or lack of melanin pigment, is a rare disease which occurs approximately once per 20.000 births. Studies of families, and of pedigrees like the one in Figure 3-3. yield the following facts, each of which is discussed relative to the hypothesis that albinism occurs in homozygotes for a reces- sive gene. a. Alternative hypotheses that albinism is due to a completely or a partiallv dominant gene soon prove untenable: 1. Both of the parents of albinos mav be 18 ( MAI'I'ER 3 nonalbino, i.e., normally pigmented. This may be explained bv the occurrence of .1 homozygous albino child (aa) from the marriage of two nonalbino heterozygotes (Aa ■ Aa). 2. The trait appears most frequentlv in progenj sharing a common ancestor. In Sweden and Japan, although the percentage of all marriages between cousins is less than 595 in the general population, it is 20 to 509? among the parents of albino children. Since albinos are rare, so is the a gene. Ac- cordingly, the chance of obtaining homozy- gous, aa. individuals is relatively slight if the parents are unrelated, because even when the first parent is A a or aa, the second parent is most likely to be A A. On the other hand. if once again the first parent is A a or aa, marriage with a related individual makes it much more likely that the second parent carries an a received from the ancestor held in common with the first parent. 3. What relative frequencies of nonalbino and albino children are expected when one tallies all the children in two-child families in which albinism appears in the progeny even though both parents are nonalbino? Based on the hypothesis under consideration, the parents must be Aa X A a. From such a marriage, the chance that a given child is nonalbino is ::4 and that it is albino 14- Each child produced from such a marriage has these same chances for nonalbinism and albinism, chances which are independent of the genotypes (or phenotypes) of the chil- dren preceding or following it in the family. Accordingly, of all two-child families whose parents are Aa, % have the first child non- albino. and of these % also have the second child nonalbino. Thus, %6 of all two-child families from heterozygous parents are ex- cluded from our sample, since both children are normally pigmented. Our sample in- cludes the following, however: families whose first child is normal (% ) and second child is albino ( ', ). making up %6 ( % of ;; , ) o{ all tWO-child families; families where the reverse is true ( :; , o\' ',). comprising another ;,,, oi all two-child families; and families in which both children are albino ( ' 1 of ' , ). which make up ' , 0 o\ all two- child families. On the average, then, everv seven albino-containing families scored should contain six normal children (three from each of the two kinds of families con- taining one albino) and eight albinos (three from each of the two kinds of families con- taining one albino, and two from each family containing two albinos), so that the ratio expected is 3:4 as nonalbino: albino. The ratio actually observed closely approximates the expected one. The observed proportions of nonalbino and albino children in families of three, or of four, or of more children from normal parents also fit the expected proportions cal- culated in a similar manner. 4. Marriages between two albinos pro- duce only albino children, as expected ge- netically from aa X aa. 5. Twins arising from the same zygote (monozygotic or identical twins) are either both albino or nonalbino. Since ordinarily such twins are genetically identical, both arc expected to be normal, AA or Aa, or albino. aa. Twins arising from different zygotes (dizygotic, nonidentical, or fraternal twins) are no more likely to be the same with re- spect to albinism than any two children of the same parents. These evidences offer clear proof that an albino person is usually homozygous reces- sive for a single pair of segregating genes. The anomaly of woolly hair is a rare trait in Norwegians. After studying pedigrees, woolly hair can be attributed to the presence of a rare dominant gene, represented by W. When woolly-haired individuals (Ww) marry normal-haired individuals (tin), it is expected and found that approximate!} 50% of children have woolly hair and 50% have normal hair. Note that the affected Segregation of Alleles 39 parent is represented as a heterozygote. Be- cause the trait is so rare, and because, bar- ring mutation, both parents of a homozygous WW child would have to have woolly hair, WW probably does not occur. Finally, consider the genetic basis for cer- tain kinds of anemia. Two special kinds occur among native or emigrated Italians. One type, severe and usually fatal in child- hood, is called thalassemia major or Cooley's anemia; the other type, a more moderate anemia, is called thalassemia minor or micro- cytemia. Pedigree and family studies show that both parents of t. major children have t. minor, and all the data support the hypoth- esis that individuals with t. major are homo- zygotes (tt) for a pair of genes, and that persons with t. minor are heterozygotes (Tt ) for this gene. More than 100,000 people in Italy have been classified as TT. Tt, or tt. Notice that in the case of thalassemia neither T nor / is completely dominant (nor com- pletely recessive). Although with respect to phenotypic ex- pression the relation between the alleles in the heterozygote may involve complete, partial, or no dominance, it should be re- called that gene action ordinarily has no effect upon either the integrity of the genes or their segregation and recombination. One can study a large number of other human traits by the pedigree and family methods and apply the principles known about genes to explain the data genetically, using the simplest suitable explanations in much the same way as was illustrated here for albinism and other traits. Sometimes, unfortunately, the data are insufficient and the investigator is left with several equally probable genetic explanations. SUMMARY AND CONCLUSIONS The gene is a unit or restricted portion of the total genetic material as discovered via any operational procedure. The genes discovered in the present chapter were revealed by recombination. Genes occur in pairs. When they are transmitted in sexual reproduction the mem- bers of a pair segregate so that any offspring receives only one member of a pair from each parent. The gene is uncontaminated by the type of gene that is its partner prior to segregation, and enters the new individual uninfluenced by the allele being con- tributed from the other parent. The hypothesis that the chromosomes comprise or carry the genetic material can be made more specific — a recombinationally detected gene may be represented by a short chromosome segment within which an exchange leading to a chiasma cannot occur. Data furnished in pedigree and family studies provide evidence that a number of human traits are based upon the action of a pair of segregating genes. REFERENCES Gates, R. R., Human Genetics, 2 Vols., New York: Macmillan, 1946. Mendel, G., 1866. "'Experiments in Plant Hybridization,*' translated in Sinnott, E. W., Dunn, L. C, and Dobzhansky, Th... Principles of Genetics, 5th Ed.. New York: McGraw-Hill, 1958. pp. 419-443; also in Dodson, E. O.. Genetics, the Modern Science of Heredity, Philadelphia: Saunders. 1956. pp. 285-311: also in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall. 1959, pp. 1-20. 40 CHAPTER 3 Mohr, O. 1 .. "Woolly Hair a Dominanl Mutanl c haractei in Man," J. Herod.. 23:345- 352, 1932. Neel, J. V., and Schull, W. J., Human Heredity, University of Chicago Press, 1954, pp. 83 86, 89 91, 240 241. Stern. (.'.. Principles oj Human Genetics, 2nd Ed., San Francisco: Freeman, I960. QUESTIONS FOR DISCUSSION 3.1. How would you recognize a line of garden peas that had become genotypically pure for a given trait'.' 3.2. Criticize the assumption that genes come only from pre-existing genes and do not arise de novo. 3.3. Differentiate between phenocopies and phenotypic overlaps. 3.4. Does a parent lose its own genetic material when it is transmitted to progeny? Defend your answer. 3.5. Is it necessary to assume that genes are able to reproduce themselves? Explain. 3.6. List all the assumptions required to explain a 3:1 ratio in F2 on a genetic basis. 3.7. A mating of a black-coated with a white-coated guinea pig produces all black offspring. Two such offspring when mated produce mostly black but some white progeny. Explain these results genetically. 3.8. A cross of two pink-flowered plants produces offspring whose flowers are red, pink, or white. Defining your genetic symbols, give all the different kinds of genotypes involved and the phenotypes they represent. 3.9. What operation was employed in studying the gene in the present chapter? Define a gene in terms of size. 3.10. Discuss the role of dominance in the study of genes. 3.11. Do organisms that reproduce asexually have genes? Explain your answer. 3.12. What relation has a gene to the phenotypic effect with which it is associated? 3.13. Do you agree with the statement on p. 33 that a cross between two colorless pea plants results in "all colorless progeny"? Why? 3.14. Throughout this book the use of the word "heredity" and its derivatives has been avoided. Why do you think this is, or is not, justified? 3.15. What is the difference between the pedigree and family methods of investigation? 3.16. What evidence is there that pigmentation (albinism vs. nonalbinism) is due to genes that are segregating? 3.17. Two nonalbinos marry and have an albino child. What is the chance that the next child is albino? Nonalbino? That of the next two children, both are albinos? Nonalbinos? One albino and one nonalbino? 3.18. What proportion of three-child families, whose parents are both heterozygous for albinism, have no albino children? All albino children? At least one albino child? 3.19. Would you conclude that the gene for woolly hair is completely dominant to nonwoolly hair? Explain. Segregation of Alleles 41 3.20. What are the similarities and differences regarding the segregation of genes and of chromosomes? 3.21. If the genetic material is in the chromosome, is it necessary to assume that the members of a gene pair occupy exactly corresponding positions in the two homo- logs? Why? 3.22. If a mitotic chromosome normally contains two identical chromatids, can we decide whether each chromosome contains one gene or an identical pair of genes? Justify your answer. 3.23. In what respect do you suppose that a sample of two-child families may be biased? How would you attempt to avoid this error? 3.24. The electron microscope shows that the sperm heads of some organisms contain a mass of uniformly thin threads. Do such cases offer evidence against the retention of the integrity of chromosomes or of genes? Why? 3.25. Differentiate between the genetic recombinations that occur at the time of fertiliza- tion and at the time of meiosis. 3.26. What is the difference between the study of genetics at the cellular and at the organismal levels? 3.27. Describe the environment of a single gene. 3.28. A lack of neuromuscular coordination, ataxia, occurs in the pedigrees of certain families in Sweden. How can you explain that one form of this rare anomaly occurs in certain families where the parents are apparently unrelated, and an- other form occurs in other families where the parents are first cousins? 3.29. What bearing have the following facts relative to the generality of the phenomenon of dominance? When pure lines of smooth-seeded plants and shrunken-seeded plants are crossed, the Fj seeds are all smooth. Microscopic examination reveals that the margins of the starch granules in the seeds are smooth in the smooth P,, highly serrated or nicked in the shrunken Pl9 and slightly serrated in all the Fv Chapter 4 INDEPENDENT RECOMBINATION BY NONALLELES I n i hi: preceding chapter we dis- cussed the transmission genetics of alternatives tor a single trait and found that a single pair of genes could explain the data in each case. But what is the genetic unit of transmission when two or more different traits are followed simul- taneously in breeding experiments? The answer may be found in the results of some additional experiments performed with the garden pea.1 Other studies show that seed shape and seed color, like the flower color trait described in Chapter 3, are each due to a single pair of genes. That is, a P] of pure-line round X pure-line wrinkled seeds gives round F,, round being dominant. Self- fertilizing the F, round produces Fj in the proportion of 3 round: 1 wrinkled. Simi- larly, a P, of pure-line yellow x pure-line green seeds gives yellow F,, yellow being dominant, and self-fertilization of the yellow Fi gives 3 yellow: 1 green in F2. What actually happens in a crossing of individuals that differ simultaneously with regard to both seed shape and seed color? A round yellow strain is crossed with a wrinkled green strain, these strains being available as pure lines. In F, only round yellow seeds are obtained (Figure 4-1). This result is what we would expect had we been studying shape and color of seeds sep- arately. In this case, there is no phenotypic Based upon experiments of G. Mendel. 42 etleet ol the dominance of one trait upon the phenotypic expression o\ the other trait. Self-fertilization of the round yellow F, gives offspring which, when counted in sulli- ciently large numbers, occur approximately in the relative frequencies of 9 round \ el- low: 3 round green : 3 wrinkled yellow:! wrinkled green. Notice that segregation and recombination are involved for each trait, as revealed in F2 by approximately 12 round: 4 wrinkled and by about 12 yel- low^ green. In this generation also there is no effect of one trait upon the recombina- tional behavior of the genetic material for a different trait. From these results, what else can we de- cide regarding the gene? Until now, we have been able to explain all the experi- mental data on the basis of only a single pair of genes, as if the total genetic material of a diploid cell is divisible into only two genes, any one gene having numerous dif- ferent alleles, each one having effects on many different traits. To continue to con- sider that each Pi individual in the present Round Yellow x Wrinkled Green ALL Round Yellow F, Round Yellow x F, Round Yellow PHENOTYPE NUMBER RATIO Round Yellow 315 9.06 Round Green 101 2.9 Wrinkled Yellow 108 3.1 Wrinkled Green 32 0.9 FIGURE 4-1. Phenotypic results from studying two traits simultaneously. Independent Recombination by Nonalleles 43 case carries but a single pair of genes, each gene must have two simultaneous effects, one on seed shape and the other on seed color. The results obtained are consistent with this requirement in the following re- spect: the F, are round yellow, and the FL. give a 3 : 1 ratio for yellow vs. green and also for round vs. wrinkled. According to this hypothesis, the FL. should be of only two types — 3 round yellow: 1 wrinkled green. But in FL. not only are these grandparental (Pi) combinations found, but two new, re- combinational classes of offspring appear. namely, round green and wrinkled yellow! Apparently, then, what is genetically trans- mitted is not composed of a single pair of indivisible units, but is composed of pairs of units, or genes, with each gene pair ca- pable of undergoing segregation separately . Let us assume, therefore, that each sex- ually reproducing organism contains more than one pair of genes. In the present case, let/? (round) and r (wrinkled) be the alleles of one pair of genes and Y (yellow) and y (green) be the alleles of the second pair. The Pj, then, would be RR YY (round yel- RR YY RR Yy A n ^ Round Rr YY W Yellow Rr Yy /2 Y RR yy r yy Round Green rr YY r r Y y Wrinkled Yellow r yy » Wrinkled figure 4-2. Expected genotypes and pheno- types in F2 following segregation. '/a R — %RY — Va R y V2 r y2 y Va r Y y2 y Va r y i IGl rh 4-3. Genotypes of gametes formed by a dihybrid, Rr Yy, undergoing independent segregation. low) and rr yy (wrinkled green). Each pair of genes would undergo segregation so that a gamete would contain only one mem- ber of each pair. In this manner the former parent would produce only RY gametes and the latter only ry, and all F, would be Rr Yy (round yellow), as observed. Based on the current hypothesis, the gam- etes formed by the Fj would contain either R or r. and. moreover, would contain either Y or y. Since the F2 show that R and Y do not always go together into a gamete, nor do r and y, there must be four genotypes possible in gametes. RY. Ry. rY. ry. Since these possible haploid genotypes will be found both in male and in female gametes, it is expected that the ¥_. would contain the diploid genotypes and their corresponding phenotypes indicated in Figure 4—2. Note that nine different genotypes are possible in F2, four giving the round yellow phenotype, two giving round green, two wrinkled yellow, .--•••:• - C = = 5-R SG , . • • • • v w - • • - I / - I rY : ::0 ■ r: • . ■ RR Yv Rr YY RxYy r r • r: - Rr Yy 1 Rt YY RrYy - YY - Yy RR YY RR Yy Rr YY Rr yy rr YY rr Yy "■ yy ~ ; - - ' . r ; 9 Round Yellovs . * . . - - j •'•--•, e c Yellow 1 Wrinkled C- r r ~ - " • RrYy Rryy rr Yy "ryy J .■*.""." " :' ." :' :' " : - - _"- ±i: -. F. - Tks qmsssk s of ame pmr of gams r_r= --"- R, of - . - - ._ ' ■J-.c - - - Y. Rm. rY, and rr. Independent leles The branching track in Figure 4—4 can be read from the top: 14 of female gametes are /?y and are fertilized :4 of the time by /?y mal; (producing %$ of all off- spring a Ul? > > | of the time fertiliza- tion is by .R > male gamete all offspring are RR > :rom this origin), etc. Summing up like classes, the kinds and relative numbers of gene n i of phenotypes are obtained as shown in the figure. The observed ratio (Figure 4—1 . client agreement with the expected one. The branching track may be used to ob- tain the 9:3:3:1 pbenotypic ratio more simp. . know that crossing together two identical monohybrids (beterozygotes for only one of all the gene pairs under con- sideration ) yields a 3:1 phenotypic ratio of dominant : rec . . trait If the recombina- tional activity of two pairs of genes is inde- pendent- both pairs being heterozygous and showing dominance, then the two indepen- dently produced 3 : 1 ratios may be combined in the progeny as shown in Figure This may be read: among the offspring, the "ich are round (because of segregation, random fertilization, and the dominance of R in the cross Rr ■ Rr will also be yellow 24 of the time and gre. d die time cause of segregation, random fertilization, and the dominance of J' in the cross ?. of all progerr. 11 be round yellow and s1<5 round greet see, men, mat independent segrega- tion by two pairs of genes results in the ' ' ' - ■ '-' - - - i - ; ' ■-.:.'. :' ■-.'.'. In the present case, the F3 dihybrid (hetero- zygote for two of all the gene pairs under consideration ) received bom R and Y from .:..-.-'-. ■'._■ HV_ • ' - ? -'-... . . ':- parent and rY from the other, the gametic recombinations would have been RY and . . - - - Regardless of how die genes enter the in- dividual, then, the dihybrid forms four, equally frequent, genetically different gam- The types and frequencies of gametes : "' -- '. "-'- '■ : --'-. ..- - . ..- • . . , ~ - . . - .-----.--. • . . -. - - double recessive individual, Le.. an individ- ual homozygous recessive for both pairs of genes concerned. In the cross of Rr Yy by ----- . :-.-:--. - - - . -. - . - . ::'--/-. - _ pected to produce four different and equally :: "-'-.'- ■- -' R '':' \ -.::'-:- r -r - ~ '- *--"-; : f s r r. - 1 ± cross (Figure 4—6) that very nearly 1 ----- " " :-- If -_-: r . . f r 25 . _ "1 ' " ' - ---..:. ---- .:--•; ;- PARENTS Rr Yy x Rr Yy .; Round <- I ellow OFFSPRING < I '/« Wrinklec <^ Greei • - - Gfeei lound elow Rc-= G-ee- •Vrinlcled Yellow •'•-■ e r 3 '£ £ - FlGU? I — S Phenotypic results of a cross between identical dihybrids. 16 CHAPTER A GAMETES Va r y 1 r y GENOTYPES Va rr yy PHENOTYPES 9 o* Va R Y 1 ry '/< Rr Yy Va Round Yellow ■ARy 1 ry k V4 Rr yy k Va Round Green '4 r Y 1 ry r 'A rr Yy f Va Wrinkled Yellow 'A Wrinkled Green figure 4-6. Test cross of the F3 dihybrid ( Rr Yy) with the double recessive indi- vidual ( rr yy). of segregation by the members of a single pair of genes and of independent segrega- tion by different pairs of genes. Whenever one is dealing with complete dominance, a cross to an individual reces- sive for the pairs of genes involved will al- ways serve to identify the genotype of the other parent, since the phenotypic types and frequencies of the offspring will correspond to the genotypic types and frequencies oc- curring in the gametes of the latter. This kind of cross is, therefore, called a test cross, or a backcross when one of the parents in the series of crosses is homozygous recessive for the genes under study. We are now in a position to return to a consideration of the material basis for genes. If one gene pair is to be associated physically with the corresponding short regions in a pair of homologous chromosomes, within which an exchange leading to a chiasma cannot occur, the question is, where, in rela- tion to one pair of genes, is a second pair located? Two possibilities occur — either both pairs are on the same chromosome pair or they are on different, nonhomologous chromosome pairs. Consider the latter as- sumption— that different pairs of genes are located on different pairs of chromosomes. If this is true, then there are several differ- ent arrangements that the parts of different pairs of chromosomes may take relative to each other at metaphase I of meiosis (Fig- ure 4-7). It has been established that different pairs of chromosomes arrive at metaphase I in- dependently of each other. Moreover, it is entirely reasonable that the orientation to- ward the poles, of the centromeres in tetrads at metaphase I and in dyads at metaphase II, is not influenced by the presence or ab- sence of chiasmata or exchanges. If, as in Case A (Figure 4-7), no exchange occurs — and, hence, no chiasma is formed — between the centromere and gene pair A a or between the centromere and gene pair Bb, alignments I and II, being equally frequent, will result in four different, equally frequent types of gametes at the end of meiosis. The same result is also obtained either when there is a chiasma between the centromere and the gene in question in one tetrad but not the other (Case B), or when a chiasma occurs in each of the tetrads (Case C). Both in Case CI and CII the dyads can orient to the poles at metaphase II in four equally likely arrangements, with the same net re- sult, four equally frequent types of gametes. Therefore, independent segregation of dif- ferent pairs of chromosomes can serve as Independent Recombination by Nonalleles 47 Pole—— Metaphase I— ■- Pol< CASE A No chiasma II CASE B After one chiasma "^ in one pair or II Haploid Meiotic Products at Telophase II AB, AB, ab, ab Ab, Ab, aB, aB AB, Ab, aB, ab Ab, AB, ab, aB CASE C AB, ab, AB, ab After one J { \\ orAb, aB' AB, ab chiasma fl I or AB, ab, Ab, aB in each pair or Bl Ub Bl Ub orAb, aB, Ab, aB Ab, aB, Ab, aB orAB, ab, Ab, aB orAb, aB, AB, ab orAB, ab, AB, ab figure 4-7. Meiotic fate of gene pairs presumably located in nonhomologous pairs of chromosomes. Note that when all alternatives in Case CI (and 11) are considered AB - ab Ab = aB. 48 CHAPTER 4 Pole-* Metaphase I Pole Haploid Meiotic Products at Telophase II CASE A No chiasma between gene pairs n A a B b AB, AB, ab, ab CASE B After one chiasma between gene pairs It A a b B I AB, Ab, aB, ab figure 4-8. Meiotic fate of gene pairs presumably located in the same pair of chromosomes. the physical basis for independent segrega- tion of different pairs of genes, regardless of chiasma formation. Let us examine next the consequences of assuming that A and B are on the same chro- mosome, and a and b are on the homologous chromosome of the pair (Figure 4-8). When no chiasma occurs between the two different pairs of genes, Case A, only the old (maternal and paternal) combinations are found in the gametes. When such a chiasma occurs. Case B, four gametic classes are produced with equal frequency (two old and two new combinational types ) . But, un- less every tetrad has a chiasma in the region between the nonalleles, the number of old gene combinations found among the gametes will exceed the new combinations. Al- though a tetrad usually contains one or several chiasmata, there are numerous points along the chromosome where an exchange leading to a chiasma might occur. An addi- tional hypothesis would be needed if each tetrad were required to form a chiasma within a given interval, such as between A and B. Moreover, we have no knowledge as to the genie interval, or the distance be- tween nonalleles presumed to be in the same chromosome. Accordingly, we shall neg- lect, for the time being, the possibility that nonalleles in the same chromosome pair can form old and new combinations with equal frequency — that is, we shall assume that two pairs of genes which do so, and are therefore segregating independently of each other, are located in different pairs of chro- mosomes. Evidence consistent with this as- sumption is obtained from studies with gar- den peas. From the breeding behavior of hybrids for two or more gene pairs, it is possible to establish the existence of seven different pairs of genes (each happening to Independent Recombination by Nonalleles 49 AA' x AA' BB' x BB' AA' BB' x AA' BB' x V» Va AA Vi AA' figure 4-9. Genotypic re- combination frequencies. y4 a'a' Va BB 1 AA BB (1) Vi BB' \ 2 AA BB' (2) Va B'B' / 1 AA B'B' (3) Va BB 2 AA' BB (4) Vi BB' \ 4 AA' BB' / (5) Va B'B' 2 AA' B'B' (6) Va BB 1 A'A' BB (7) y2 bb' \ 2 A'A' BB' (8) V4 B'B' f 1 A'A' B'B' (9) show dominance in the hybrid condition), each pair seeming to segregate independently of all the others. Since the garden pea pos- sesses a diploid number of seven pairs of chromosomes, there are enough chromo- some pairs for each pair of genes to be lo- cated on a different pair of chromosomes. * Different Phenotypic Ratios - A monohybrid may show the phenotypic effects of only one allele, some of the effects of both alleles, or the complete effects of both alleles. These phenotypic conse- quences have already been designated as complete, partial, and no dominance, re- spectively. In the garden pea experiments already discussed, complete dominance pro- duced the 3 : 1 phenotypic ratio obtained from a cross between identical monohybrids. This necessitated breeding the offspring with the dominant phenotype in order to identify the 1:2:1 genotypic ratio predicted from such crosses. Had no dominance or partial -See W. Bateson (1909). dominance occurred, the phenotypic and genotypic ratios would have been identical. Nevertheless, in all cases the recombining genes retained their individuality, and the specific ratios observed depended only upon the dominance relation within the gene pair — that is, the relation between the expres- sion of one allele and that of its partner. Complete dominance also has no influence upon the individuality or segregation of non- allelic pairs of genes. Although the geno- typic ratio expected from crossing two par- ticular dihybrids has already been derived (Figure 4-4), let us rederive this ratio, em- ploying more general symbols for the genes, using the branching-track method in a slightly different way. Let A and A' be one pair of alleles and B and B' another. Mating AA' BB' by AA' BB' gives the genotypic ratio seen in Figure 4-9. Notice here that among every 16 off- spring, on the average, there would be 9 different genotypes in the ratio of 1:2:1:2: 4:2:1:2:1. How did this genotypic ratio 50 CHAPTER 4 give rise to the 9:3:3: I phenotypic ratio in crosses between identically dihybrid garden peas'.' I'wo factors were responsible. One was the occurrence of dominance for each pair oi alleles; this converted the 1:2:1 genotypic ratio for each gene pair to a 3:1 phenotypic ratio. The other factor was that the action of the two gene pairs was inde- pendent and resulted in detectable effects on different traits. This permitted both 3:1 ratios to be recognized separately even when these ratios were distributed at random in the progeny (Figure 4-5). It becomes ap- parent, therefore, that the particular pheno- typic ratios obtained, following crosses in- volving more than one gene pair, depend both upon the dominance relationships be- tween alleles and the gene interaction rela- tionships between nonalleles. If neither gene pair shows dominance, and if each pair acts both independently and on different traits, two 1:2:1 phenotypic ratios will be produced, and these, when distributed at random, will result in the 1:2:1:2:4:2: 1:2:1 phenotypic ratio. Here, because no genotype is masked phenotypically by any other, the phenotypic and genotypic ratios are the same. (This would also be true of the following crosses: A A' BB' X AA BB, AA' BB' X A' A' B'B', AA' BB X AA BB'. ) This kind o\ result is illustrated in the prog- eny of parents both of whom have thalas- semia minor (77) and MN (MM') "blood type." (MM is phenotypically M, M'M' is phenotypically N, as described on p. 58.) When, however, the aforementioned con- ditions are changed so that one of the two pairs of genes shows dominance, two dif- ferent genotypes will produce the same phenotype, and fewer than 9 phenotypes are expected. Thus, if B is dominant to B\ genotypes 1 and 2 (in Figure 4-9) are ex- pressed as one phenotype, genotypes 4 and 5 as another, and 7 and 8 as another, so that the phenotypic ratio becomes 3:1:6:2:3:1. This is the phenotypic ratio expected in the progeny of parents both heterozygous for albinism (Aa) and having MN blood type (MM'). If both gene pairs show domi- nance, one phenotype is expressed by geno- FIGURE 4-10. Drosophila melanogaster itui- tants showing the no-wing (left) and the curled wim; ( right ) phenotypes. ( Drawn by E. M. Wallace.) Independent Recombination by Nonalleles 51 types 1, 2, 4, 5, another by genotypes 3 and 6, another by 7 and 8. and another by geno- type 9, producing the 9:3:3:1 ratio already discussed. Dominance causes a reduction in the number of phenotypic classes. What phenotypic ratios are expected when two different, independently active, pairs of genes affect the same trait in the same man- ner or direction? If one or more allelic combinations for one gene pair produce the same phenotype as one or more allelic com- binations for the other gene pair, the num- ber of phenotypes will also be reduced from the maximum (9 when identical dihybrids are crossed). Of course, the number of different phenotypes detected will be further reduced if the absence of dominance in both gene pairs is changed to dominance in one gene pair, and still further reduced if both gene pairs show dominance. Thus, if A and B produce equal amounts of melanin pigment in human skin, the amount of pig- ment being cumulative, and A' and B' pro- duce none, dominance being absent, a cross between identical dihybrids yields the ratio 1 "black" (type 1 ) :4 "dark" (types 2, 4) : 6 "mulatto" (types 3, 5, 7) :4 "light" (types 6, 8):1 "white" (type 9), instead of 9 dif- ferent phenotypes. Moreover, if both A and B show complete dominance, for exam- ple either gene producing full flower color, the phenotypic ratio becomes 15 colored (types 1-8) :1 colorless (type 9). Note that when different pairs of genes act on the same trait in the same direction or way, they have a common phenotypic background on which their effects superpose, and the effect of one gene pair interferes with the detection of the effect of the other pair. Sometimes different gene pairs act inde- pendently on the same trait in different — antagonistic or cooperative — ways. For ex- ample, in Drosophila (Figure 4-10), A' is a recessive allele which reduces the wing to a stump, whereas B' is a recessive allele which causes the wing to be curled, the dominant allele A making for normal sized wings and the dominant allele B straight wings. A cross between two identical dihy- brids does not produce the customary 9:3: 3:1 ratio. In the present case, the ratio becomes 9 flies with long, straight wings: 3 with long, curled wings: 4 whose wings are mere stumps (of which one quarter would have had curled wings if the full wing had formed). Here, then, the phenotypic ex- pression of one gene pair can prevent de- tection of the phenotypic expression of an- other gene pair. In another case, either of two pairs of genes may prevent a given phenotype from occurring. Suppose the dominant alleles A and B each independently contribute some- thing different but essential for the produc- tion of red pigment, whereas their corre- sponding recessive alleles A' and B' fail to make the respective independent contribu- tions to red pigment production. Then crosses between two identical dihybrids will produce 9 red: 7 nonred (composed of 3 homozygotes for A' only, 3 homozygotes for B' only, and 1 homozygote for both A' and B ') . Notice that if the recessive alleles are considered, we have just dealt with examples of unilateral and mutual opposition to pheno- typic expression, respectively, but if the dom- inant alleles are considered these become cases of unilateral and mutual cooperation in phenotypic expression. In all cases where two pairs of genes af- fecting the same trait interact phenotypically by superposition, antagonism or cooperation, one pair of genes has had an influence upon distinguishing the effects of the other. The general term epistasis may be used in these cases to describe the interference with — suppression or masking of — the phenotypic expression of one pair of genes by the mem- bers of a different pair. Genes whose de- tection is hampered by nonallelic genes are 52 CHAPTER 4 said to be hypostatic, or to exhibit hypo- stasis. As dominance implies recessiveness, so epistasis implies hypostasis. There need be no relationship between the dominance o\ a gene to its allele and the ability of the gene to be epistatic to nonalleles. In theory. then, epistatic action may depend upon the presence of A. A\ or A A'; moreover, hypo- static reactions may depend upon the pres- ence of B, B\ or BB'. Consequently, it should be noted that in crosses between identical dihybrids. epistasis-hypostasis can produce phenotypic ratios still different from those already described. Consider another example of a dihybrid in which both pairs of genes show domi- nance but no epistasis. In Drosophila, the dull-red eye color of flies found in nature is due to the presence of both red and brown pigments. Let A be the allele which pro- duces the red pigment and A' its recessive allele which produces no red pigment; let B be the nonallele producing the brown pig- ment whose allele B' makes no brown pig- ment. A mating between two dull-red dihy- brid flies (from a cross of pure red, A A B'B' by pure brown. A' A' BB) produces offspring in the proportion 9 dull-red (containing A — B — ):3 red (containing A — B'B'): 3 brown (containing A' A' B — ) : 1 white (A' A' B'B'). The last phenotypic class, resulting from the absence of both eye pigments, is new in this series of crosses. This case illus- trates that the interaction of nonallelic genes may result in apparently novel phenotypes. Such interactions change not the number but the kind of phenotypes obtained. The preceding discussion suggests that any given phenotypic trait may be the result of the interaction of several gene pairs. One is even led to conclude that the total pheno- type is the product of the total genotype acting together with the environment. The difference between phenotypic and geno- typic ratios is often due to products of gene action — by alleles and nonalleles — which superpose, cooperate, or conflict at the phys- iological or biochemical level. It is also pos- sible that there is sometimes a direct in- fluence of one gene upon the ability of an allele or nonallele to act. Although the na- ture of gene interactions can be predicted partially, in a general way. from the kind of modified ratio obtained, an understanding of the mechanisms involved must ultimately be based upon a knowledge of how genes act and the nature and fate of gene products. In no case has a phenotypic ratio that differs from the expected genotypic one served to disprove either segregation or independent segregation. In fact, segregation and inde- pendent segregation were first proved despite the misleading phenotypic simplifications of genotypic ratios wrought by the occurrence of dominance; moreover, the principle of independent segregation could also have been first proved from crosses involving epistasis or apparently novel phenotypes. SUMMARY AND CONCLUSIONS When two different traits were studied separately, the phenotypic alternatives were found to be due to the presence of a single pair of genes in each case. Studies were then made of the distribution of phenotypes in successive generations when these two pairs of traits were followed simultaneously in the same individuals. The data ob- tained showed that each trait is due to the presence of a different pair of genes, proving that the genetic material is made up not of a single segregating pair of genes but of a number of segregating gene pairs. Moreover, the results are best explained by the principle that the segregation of one pair of alleles is at random with respect to the segregation of all the other nonalleles tested. The simplest hypothesis for the physical Independent Recombination by Nonalleles 53 William Bateson (1861-1926). (From Genetics, vol. 12, p. 1, 1927.) basis for the independent recombination of such nonalleles is that different pairs of genes reside in nonhomologous pairs of chromosomes. The phenotypic expression of genes depends upon their alleles, insofar as dominance is involved, and upon nonalleles, insofar as epistasis (including superposition, coopera- tion, and antagonism) and the production of apparently novel phcnotypes are involved. The absence both of dominance and of epistasis will always produce phenotypic ratios which directly represent genotypic ratios, whereas the occurrence of one, the other, or both reduces the number of phenotypic classes. In any case, segregation and inde- pendent segregation, being genetic properties, are totally uninfluenced by the manner whereby genes do or do not come to phenotypic expression. REFERENCES Bateson, W., Mendel's Principles of Heredity, Cambridge, England: Cambridge Uni- versity Press, 1909. Mendel, G. See reference at the end of Chapter 3. Supplement I (at the end of this book). QUESTIONS FOR DISCUSSION 4.1. Make genetic diagrams for the crosses and progeny discussed in the second and third paragraphs on p. 42. Be sure to define your symbols. 4.2. Is a test cross or backcross used to determine genotypes from phenotypes in cases of no dominance? Explain. ~>l CHAPTER 4 4.3. What types and frequencies of gametes are formed by the following genotypes, all gene pans segregating independently? (a) An lib C< (C) A a lib Cc (b) AABBCcDD (d) MmNnOoPp 4.4. How mam different diploid genotypes are possible in offspring from crosses in which both parents are undergoing independent segregation for the following numbers o\ pairs of heterozygous genes — 1, 2, 3, 4. n? 4.5. What conclusions could you reach about the parents if the offspring had pheno- types in the following proportions? (a) 3 : 1 (c) 9 : 3 : 3 : 1 ( b ) I : I ( d ) 1 : 1 : 1 : 1 4.6. Would you be justified in concluding that a pair of chromosomes can contain only a single pair of genes? Explain. 4.7. Suppose a particular garden pea plant is a septahybrid. What proportion of its gametes will carry all seven recessive nonalleles? All seven dominant nonalleles? Some dominant and some recessive nonalleles? 4.8. What proportion of the offspring of the following crosses, involving independent segregation, will be completely homozygous? (a) AaBbXAaBb (c) Aa BB Cc X A A Bb cc (b) AA BBCCX AA bbec (d) AA' X A" A'" 4.9. Following independent segregation, why would you expect that gametes fertilize at random with respect to their genotypes? 4.10. Discuss the particulate nature of the genetic material. 4.11. Does the discovery of independent segregation of nonalleles affect your concept of gene size? Explain. 4.12. Discuss your current understanding of the term "genetic recombination." 4.13. Discuss the factors that can modify the expected phenotypic ratio. 4.14. In snapdragons, red flowers ( R ) are incompletely dominant to white (/), the hybrid being pink; narrow leaves (N) are incompletely dominant to broad leaves (/;), the hybrid being intermediate in width ("medium"). Assuming the gene pairs recombine independently, give the genotypic and phenotypic ratios ex- pected among the progeny of a cross between the following: (a) a red medium and a pink medium plant (b) a pink medium and a white narrow (c) two identical dihybrids 4.15. Suppose an albino child also suffers from thalassemia minor. Give the most likely genotypes of the parents. 4.16. How can you explain that a certain kind of baldness, due to a single gene pair, is dominant in men and recessive in women? 4.17. Though blue-eyed couples ordinarily have only blue-eyed progeny, brown-eyed couples may also have blue-eyed children. Select and define gene symbols so you can give the complete genotypes of all individuals mentioned in each of the families listed: (a) One member of a pair of twin boys has brown eyes, the other has blue eyes. (b) A blue-eyed, woolly-haired child resembles his father in one of these respects and his mother in the other respect. (c) A brown-eyed, nonthalassemic child is like the grandmother but unlike the mother in both of these respects. Independent Recombination by Nonalleles 55 4.18. Differentiate between dominance and epistasis. 4.19. What is the maximum number of genotypes possible in the progeny if the parents are monohybrids? 4.20. Two green corn plants are crossed and produce offspring of which approximately '■'\,; are green and 7ic are white. How can you explain these results? 4.21. Does gene interaction occur only when identical monohybrids (or identical di- hybrids) are crossed? Explain. .MA. WALNUT COMB SINGLE COMB 4.22. A chicken from a pure line of "rose" combs is mated with another individual from a pure line of "pea" combs (see the accompanying illustration). All the Fj show "walnut" combs. Crosses of two F, "walnut" type individuals provide F2 in the ratio 9 "walnut": 3 "rose": 3 "pea":l "single." Choose and define gene symbols to provide a genetic explanation for these results. 4.23. Three walnut-combed chickens were crossed to single-combed individuals. In one case the progeny were all walnut-combed. In another case one of the progeny- was single-combed. In the third case the progeny were either walnut-combed or pea-combed. Give the genotypes of all parents and offspring mentioned. 4.24. Matings between walnut-combed and rose-combed chickens gave 4 single, 5 pea. 13 rose, and 12 walnut progeny in Fv What are the most probable genotypes of the parents? 4.25. A mating of two walnut-combed chickens produced the following Fx with respect to combs: 1 walnut, 1 rose, 1 single. Give the genotypes of the parents. 4.26. The hornless, or polled, condition in cattle is due to a completely dominant gene, P, normally horned cattle being pp. The gene for red color (R) shows no domi- nance to that for white (/?'), the hybrid (RR') being roan color. Assuming in- dependent segregation, give the genotypic and phenotypic expectations from the following matings: (a) Pp RRX pp RR' (b) Pp RR' X pp RR (c) Pp RR' X Pp RR' (d) hornless roan (whose mother was horned) X horned white 4.27. When dogs from a brown pure line were mated to dogs from a white pure line all the numerous F] were white. When the progeny of numerous matings be- tween Fj whites were scored there were 118 white. 32 black, and 10 brown. How can you explain these results genetically? 4.28. Using your answer to the preceding question, give the phenotypic and genotypic expectations from a mating between the following: (a) a black dog (one of whose parents was brown) and a brown dog (b) a black dog (one parent was brown, the other was black) and a white dog (one parent was brown, the other was from a pure white strain) 56 CHAPTER 4 4.2l>. When one crosses pure White 1 eghorn poultry with pure White Silkies, all the I are while. In the 1 ■'■_.. however, large numbers of progeny occur in a ratio approaching 13 white:3 colored. Choosing and defining your own gene symbols, explain these results genetically. 4.30. (a) In the yellow daisy the Bowers typically have purple centers. A yellow- centered mutant was discovered which when crossed to the purple-centered type gave all purple-centered I,, and among the F2 47 purple and 13 yellow. Explain these results genetically . (b) later, another yellow-centered mutant occurred which also gave all purple F, from crosses with purple-centered daisies. When these F, were crossed to- gether, however, there were 97 purples and 68 yellows. Explain these results genetically. (c) How can you explain that a cross between the two yellow-centered mutants produced all purple-centered F,? 4.31. Give a single genetic explanation that applies to all the following facts regard- ing human beings: (a) One particular deaf couple has only normal progeny. ( b ) One particular deaf couple has only deaf progeny. (c) One particular normal couple has many children, about % are normal and Y4 deaf. (d) One particular normal couple has all normal children. (e) Normal, identical twins marry normal, identical twins and have a total of 9 normal and 9 deaf children. 4.32. How can you explain the observations with regard to lint color of cotton that brown X green gives green F,, which when mated together produce F-, which contain mostly brown, some greens, and a few whites? 4.33. Suppose two unrelated albinos married and had 8 children, 4 albino and 4 non- albino. How could you explain these results? 4.34. When, during the life cycle, can dominance and/or epistasis occur or not occur in maize? Neurospora? 4.35. When two plants are crossed it is found that °%4 of the progeny are phenotypically like the parents, and 1) wings of Drosophila melanogaster. (Courtesy of C. Stern; from Genetics, vol. 28, p. 443, 1943.) an effect of this gene pair rather than a modifying effect of some other gene pair. Apparently, then, the ci+ allele in popula- tion 3 is different from that in populations 1 and 2. We are dealing, consequently. with two isoalleles in a multiple allelic series. (Note that a slightly different system of symbolizing genes was used here.) 4. Eye Color in Drosophila. Another series of multiple alleles in Drosophila involves eye color. In this case the different alleles can be arranged in a series that shows different grades of effect on eye color, ranging from dull-red to white: dull-red (w+), blood (u'-'). coral (W"), apricot (u"'). buff (w6/), and white (w). The w * allele is dominant to the others listed and is the allele com- monly found in wild-type flies. One can think of all the different alleles as producing the same kind of phenotypic effect, but less of it in proceeding from w+ to w, the white allele being completely inefficient in this respect. We have already described isoalleles for genes normally expressed in individuals liv- ing in the wild (wild-type isoalleles). Iso- alleles for mutant genes (mutant isoalleles) also occur. For instance, it has been proved that the gene producing white eye color in different strains of Drosophila is actually composed of a series of multiple isoalleles (vv1, w'-, w:!, etc.). 5. Self-sterility in Nicotiana. Among sex- ually reproducing plants it is not uncommon to find that self-fertilization does not occur even though the male and female gametes are produced at the same time on a given plant. The reason for this has been studied in the tobacco plant, Nicotiana, where it was found that if pollen grains fall on the stigma of the same plant, they fail to grow down the style. When this happens self-fertiliza- tion is impossible. A clue to an explanation for this phenomenon comes from the ob- servation that different percentages of pollen from a completely self-sterile plant may grow down the style of other plants. The results of certain crosses are shown in Figure 5-4. Genetically identical pistils are exposed to pollen from the same plant (A), from a second one (B), and from a third (C). No pollen, approximately half, and approximately all, respectively, are able to grow down the style of the host. Note, in B, that although all the pollen used came from one diploid individual, half of it can Multiple Alleles; Mucigenic Traits 61 and half of it cannot grow on its host. Re- call that the stigma and style are diploid tissues, whereas pollen grains are haploid. These results suggest that most important in determining whether or not a pollen grain can grow down a style is not the diploid genotype of its parent but the haploid geno- type contained in itself. Let us assume that self- or cross-sterility is due to a single pair of genes. Call s3 the allele contained in the pollen which permits pollen to grow in case B. The pollen grains from the host plant furnishing the pistil can- not contain s3, or the pollen would be able to grow on their own parent; and they can- not (Case A). So, the host pistil tissue in this experiment cannot contain s3, and one of its alleles can be called si. Then, half of the pollen from the host individual will carry si (Case A); but since these fail to grow, we must assume that any pollen grain carrying an s allele also present in the host pistil will fail to grow. Excluding the pos- sibility of a mutation, the other allele in the host pistil cannot also be 57, since one si would have had to be received from a pa- ternal pollen grain growing down a maternal style that carried si as one of its two alleles. Since the second allele in the pistils illus- trated cannot be either si or s3, let us call it s2. The other half of pollen from the pistil parent thus will contain s2, and also fail to grow in self-pollination (Case A). In B the pollen grains that fail to grow are either s/ or s2 (adhering to the law of par- simony) ; however, their precise identity can- not be determined without additional tests. In C, since all the pollen grew, one pollen allele must be a different one — call it s4. The other pollen allele may be s3 or a still different one, s5. Here again more tests are needed to determine the precise identity. In these cases the phenotypic alternatives for pollen are to grow or not to grow. Whenever the pollen grains from any one plant are placed on a given stigma and both alternatives occur, the phenotypes are in a 1 : 1 ratio. All these results and others are Stigma VA *'% PISTIL < Style figure 5-4. Multiple alleles for cross- or self -sterility. Ovary S1S2 STS2 S1S7 62 CHAPTER 5 consistent with the assumptions made, that self- 01 cross-sterility is regulated by a single pair of genes which form a multiple allelic series. Some species prose to have fifty or more multiple alleles forming a series re- sponsible for self-sterility, group sterility, or group incompatibility. ^Multigenic Traits Up to now. the traits chosen to study genes occur in clear-cut. qualitatively different al- ternatives like flower color in garden peas, or albinism vs. pigmentation, and various blood types in human beings. These are called discontinuous or qualitative traits be- cause in each case an individual belongs clearly to one phenotypic class or another. Although the interaction of many or all genes may ultimately be involved in the appearance of a given phenotype, the phenotypic alter- natives previously considered have been ef- fected primarily by only one or a few pairs of genes. Moreover, in these cases the non- genetic environment had much less or no effect upon the phenotypic differences in- volved. For practical and theoretical reasons one may also be interested in the genetic basis for certain continuous traits like height of corn or intelligence in man, for which there are so many grades that individuals are not separable into discrete types or classes. Such traits are also called quantitative traits because the continuous range of phenotypes observed requires that an individual be meas- ured in some way in order to be classified. Are quantitative traits also determined ge- netically? Let us make the simplest assump- tion that quantitative traits differ from qual- itative ones only in degree, the former being due to the combined effects of many gene pairs. In the case of multigenic (polygenic) traits, although many phenotypic classes would be made possible by the action of multiple gene pairs, the effect of any single pair would be difficult to distinguish. Con- sequently, since each pair of genes would contribute only slightly toward the expres- sion of the quantitative trait, one would expect the effect of environment to be rela- tively larger than that of any single gene or gene pair. The large effect of fertilizer upon corn ear si/e and of diet upon height in human beings illustrate the importance of environment in multigenic traits. A given trait may be determined qualita- tively in certain respects and quantitatively in other respects. For example, in garden peas one pair of genes may determine whether the plant will be normal or dwarf, the actual size of a normal plant being de- termined by multigenic interaction with the environment playing a significant role. Sim- ilarly, a single pair of genes can determine whether a human being has a serious mental deficiency or normal mentality, though nearly all individuals have a degree of mental ability which varies in a continuous way due to environment and polygenes. If quantitative traits are determined multi- genically, it ought to be possible to derive other characteristics of them which are con- sistent with actual observations by consider- ing the same trait, first as a qualitative trait (i.e., determined by one or two or three gene pairs), and then as a quantitative trait (i.e., determined by many gene pairs). Let the trait be color, and the alternatives in P, be black and white. Assume first that there is no dominance or epistasis (see p. 51); then, whether one, two, three, or many gene pairs are involved, the F, will be uniform and phenotypically intermediate (medium gray) between the two P,. Examine, in Figure 5-5, results of matings between F] (by cross- or self-fertilization) in each case. As the number of gene pairs increases, the number of classes of Fj offspring increases. As the number of classes becomes large, one would expect environmental action to cause Multiple Alleles; Multigenic Traits, 63 individuals to fall out of their phcnotypic class, so to speak, and into the space be- tween classes or into an adjacent phenotypic class. And so, as gene pair number in- creases, classes become more numerous, then indiscrete, resulting finally in a continuum of phenotypes. Note also that as the number of gene pairs determining the trait increases, the fraction of all F2 resembling either P, becomes smaller. Thus, with one pair of genes y2 of FL> are black or white, with two pairs %, with three pairs YS2, etc. Consequently, as the number of genes increases from 1 0 to 20 and more, the continuous distribution of phenotypic types gives rise to an F2 curve which becomes narrower and narrower. In other words, the chance of recovering in F2 any phenotype a given distance off the mean decreases as gene pair number increases. Although it may be relatively easy to identify whether one, two, or three gene pairs cause a given characteristic, it is much more diffi- cult to determine exactly how many pairs are involved whenever more than three are involved. In multigenic cases, measurement of how the population varies relative to the average phenotype can give information as to the approximate number of polygenes involved. The variability of a trait can be measured statistically as follows: the mean, m (the simple arithmetic average), is found. The variance, v (the measure of variability from the mean), for a group of measurements is determined by finding the difference between each measurement and the mean, squaring each such difference, adding all the values obtained, and dividing the total by 1 less than the number of measurements involved. With a given sample size, all other things being equal, the greater the variance the smaller the number of gene pairs involved, as would be expected from Figure 5-5. One may find detailed statistical procedures for P x P, 1 F- >< F, F, < involving one gene pair involving two gene pairs -■ill.. involving three gene pairs involving many gene pairs figure 5-5. Dependence of number of pheno- typic classes upon number of gene pairs. Horizontal axis shows classes, vertical axis indicates relative frequencies. 64 CHAPTER 5 3/4 A- % B Va bb 9/i6 A- B- (1 unit) V,6 A- bb (2 units) V4 a a % B- y4 bb _^ 3/i6 aa B- (0 units) .^ Vi6 aa bb (1 unit) figure 5-6. Results of cross- ing together the dihybrids described in the text. using variance this way in any standard text on elementary statistical methods. Consider next the effect of dominance upon the expression of quantitative traits. When a qualitative trait is determined by one. two. or three pairs of heterozygous genes not showing dominance, there are (as in Figure 5-5 ) three, five, or seven possible phenotypic classes, respectively. As a result of dominance, however, the number of classes is reduced (cf. Chapter 4, p. 51). Since the estimated number of gene pairs responsible for a phenotype is directly re- lated to the number of phenotypic classes, the number of gene pairs involved in a quan- titative trait is underestimated whenever dominance occurs. This effect is important because many genes show complete or par- tial dominance. One can construct a hypothetical case in which two pairs of genes both showing domi- nance can give much the same phenotypic result as one pair with no dominance. Sup- pose gene A (as AA or Aa) adds 2 units of effect and its recessive allele a (as aa) adds only 1 unit; suppose B (as BB or Bb) subtracts 1 unit of effect and its recessive allele b (as bb) has no effect at all. Then a 2-unit individual {A A bb) mated with a 0-unit one (aa BB) will give all intermediate 1 -unit F| (AaBh). The F_. from the mat- ing of the F] can be derived by a branching track as shown in Figure 5-6. The pheno- typic ratio obtained in F2 of 3:10:3 might be, in practice, difficult to distinguish from the 1:2:1 ratio obtained from crossing monohybrids showing no dominance.3 Dominance has a second effect with re- gard to quantitative traits; this can be illus- trated by means of two crosses involving the genes just described. In the first, two 0- unit individuals are crossed, aa Bb X aa Bb, yielding % aa B- (0 unit) and % aa bb (1 unit). In this case the parents, which are at one phenotypic extreme (0 unit), produce offspring which are, on the average, less extreme (0.25 unit). In the second case, two 2-unit individuals are crossed, Aa bb X Aa bb, yielding % A- bb (2 units) and % aa bb (1 unit). Here the parents are at the other phenotypic extreme (2 units) but produce offspring which are, on the average, less extreme (1.75 units). These results demonstrate regression, the conse- quence of dominance which causes individ- uals phenotypically extreme in either direc- tion to have progeny less extreme. Figure 5-7 illustrates the principle of re- 3 See J. H. Edwards ( 1960). Multiple Alleles; Multigenic Traits 65 gression in polygenic situations. When no dominance occurs, the average offspring from parents at A, B, and C will be at the corresponding points A'. B\ C\ respectively, in the offspring curve. (The environment will cause some fluctuation around these phenotypic mean points in the offspring curve.) In the case of dominance, how- ever, the offspring of A will be, on the aver- age, to the right of A, as shown by arrows, whereas the offspring of C will generally be to the left of C. Contrary to what one might expect, the loss of extreme individuals gen- eration after generation will not make the entire population more and more homo- geneous phenotypically; there will be a closely counterbalancing tendency for the average members, B, of the population to produce offspring more extreme than them- selves in either direction. The result is that, as in cases of no dominance, the distribution curve for the offspring will be the same as for the parent population. PARENT GENERATION Group Selected as Parents Mean Mean of selected group OFFSPRING GENERATION FIGURE 5- character. Mean 8. Selection for a quantitative PARENTS 7^TS<7\ OFFSPRING figure 5-7. The principle of regression. To obtain a line of phenotypically extreme individuals from a population showing a quantitative trait, one would choose the ex- treme individuals as parents (Figure 5-8). If dominance were absent, the very first off- spring generation would have the same mean as the group selected as parents. Some de- gree of dominance usually occurs, and hence regression will usually occur; and the mean size of the first generation offspring will be somewhat less extreme than that of the se- lected parents, but somewhat more extreme than the original mean. As one continues to select appropriately extreme individuals as parents, the offspring in successive gen- erations will, on the average, approach more and more closely the extreme phenotype desired. 66 CHAPTER 5 SUMMARY AND CONCLUSIONS \ gene can cxisl in any one of two or more genetic states. These alternatives comprise a multiple allelic (sometimes isoallelic) series. Some alleles ol a given gene may produce apparently different qualitative phenotypic effects (and therefore show no dominance when hybrid); other alleles may produce different degrees of a quantitative phenotypic effect (in which case they may or may not show dominance when hybrid). Genes are the basis for both continuous and discontinuous traits. Continuous traits are usually determined by many gene pairs, each of which has a phenotypic effect that is small and often matched or exceeded by the action of the environment. The variability of a quantitative trait is such that the larger the number of heterozy- gous polygenes determining it, the narrower is the distribution curve and the smaller the chance of recovering either of the extreme phenotypes in the offspring. When polygenes are heterozygous, dominance has the effect of reducing the number of pheno- typic classes and of placing proportionally more offspring in extreme classes. Conse- quently, dominance usually causes one to underestimate the number of genes determin- ing a quantitative trait. Dominance also causes regression, so that selection must be continued for a number of generations to obtain a line which approaches the desired phenotype. REFERENCES Crow, J. F., Genetics Notes, 5th Ed., Minneapolis: Burgess, 1962. Edwards. J. H.. "The Simulation of Mendelism," Acta Genet., Basel, 10:63-70. 1960. Falconer, D. S.. Introduction to Quantitative Genetics, New York: Ronald Press, 1961. Race. R. R.. and Sanger. R.. Blood Groups in Man, 4th Ed., Philadelphia: F. A. Davis, 1962. Wiener, A. S., and Wexler. I. B.. Heredity of the Blood Groups, New York: Grune & Stratton. 1958. QUESTIONS FOR DISCUSSION 5.1. Discuss the occurrence of dominance with respect to blood group types. 5.2. Why was it necessary to assume that a gene may have more than two allelic forms? 5.3. A baby has blood type AB. What can you tell about the genotypes of its parents? What would you predict about the blood types of children it will later produce? 5.4. If one parent is A blood type and the other is B, give their respective genotypes if they produced a large number of children whose blood types were: (a) All AB (c) Half AB, half A (b) Half AB, half B (d) h AB. Va A. *» B. U O Multiple Alleles; Multigenic Traits 67 5.5. Give examples of complete dominance and of no dominance as found in human beings. 5.6. Is the occurrence of complete dominance helpful in determining the genetic basis of alternatives for a given trait? Explain. 5.7. A father with blood group types M and O has a child with MN and B blood types. What genotypes are possible for the mother? 5.8. Criticize the statement: "Genes can be explained on the basis of a presence-or- absence hypothesis." 5.9. A woman belonging to blood group B has a child with blood group O. Give their genotypes and those which, barring mutation, the father could not have. 5.10. What do you think of the view that all the different genes that exist can be described as being different multiples of a single basic unit which is capable of retaining its integrity and is able to self-replicate? 5.11. How many different genotypes are possible when there are four different alleles of a single gene? 5.12. Does the discussion of multiple alleles in the text imply that: (a) There is an infinite variety of isoalleles? (b) No two genes are ever identical? Explain. 5.13. Describe how you would test whether the genes for white eye color in two different populations of Drosophila were alleles, isoalleles, or nonalleles. 5.14. In rabbits the following alleles produce a gradation effect from full pigmentation to white: agouti (C), chinchilla (crh) and albino (c). Another allele, ch, pro- duces the Himalayan coat-color pattern. C is completely dominant to all these alleles, ch is completely dominant to c, whereas cclt shows no dominance to ch or c. (a) How many different diploid genotypes are possible with the alleles men- tioned? (b) A light chinchilla mated to an agouti produced an albino in Fx. Give the genotypes of parents and F1. (c) An agouti mated to a light chinchilla produced in Fx one agouti and two Himalayan. Give the genotypes possible for parents and Fj. (d) An agouti rabbit crossed to a chinchilla rabbit produced an agouti offspring. What genotypic and phenotypic results would you expect from crossing the F1 agouti with an albino? 5.15. For each of the following matings involving Nicotiana give the percentage of aborted pollen tubes and the genotypes of the offspring. $ 9 6 9 (a) sls2Xsls3 (c) si s4 X si s4 (b) sls3Xs2s4 (d) s3 s4 X s2 s3 5.16. Could you prove the existence of multiple allelism in an organism that only re- produces asexually? Explain. 5.17. Do the genes for quantitative traits show epistasis? Explain. 5.18. Does the environment have a more important role in determining the phenotype in cases of quantitative than in cases of qualitative traits? Explain. 5.19. Under what circumstances are only seven phenotypes possible when three pairs of genes determine a quantitative trait? 5.20. Discuss the statement: No new principles of genetics have originated from the study of polygenic traits. 68 CHAPTER 5 5.21. Suppose each gone represented by a capital letter causes a plant to grow an additional inch in height, aabbceddee plants being 12 inches tall. Assume independent segregation occurs tor all gene pairs in the following mating: A a 1111 cc /)r/ 1.1. ! N 1 1 i 1 1.3 1.2 1.1 1.0 .9 1 .8 1 1 1 1 1 1 II 7 .6 .5 .4 .3 .2 .1 0 RELATIVE VIABILITY FIGURE 6-2. Classification of c fleets that mu- tants have on viability. merit, or at any stage in between. Some- times a lethal effect is produced not by one gene or a pair, but by the combined effect of several nonallelic genes. In such a case, some of the nonalleles are contributed by each parent, and the offspring dies because the nonalleles, viable when separate, are lethal when together. Different alleles, recessive or dominant, have been shown to affect viability in dif- ferent degrees. These effects cover the en- tire spectrum — ranging from those which are lethal, through those which are greatly or slightly detrimental, to those which are ap- parently neutral or even beneficial (Figure 6-2). When different combinations of alleles or nonalleles have different viabilities, the phenotypic ratios observed may differ significantly from those expected. The im- portance of the precautions to be taken, relative to the viability and fertility of the individuals bred in experiments designed to establish principles of transmission genetics, has already been discussed in Chapter 3. Pleiotropism Does each gene affect only one trait or can it have multiple, manifold, or pleiotropic effects? An investigation ' can be undertaken to answer this question, using two strains of Drosophila that are practically identical ge- netically (isogenic), except that one is pure 1 See Th. Dobzhansky and A. M. Holtz ( 1943 ). tor the gene for dull-red eye color (vv+) and the other is pure for its allele white (w). Another trait is then chosen for examination in these two strains, one apparently uncon- nected with eye color. The trait selected is the shape of the spermatheca, an organ found in females which is used to store the sperm received. The ratio of the diameter to the height of this organ is determined for each of the two strains. This index of shape is found to be significantly different in the dull- red as compared to the white strain. From this result it can be concluded that the eye- color gene studied is pleiotropic. The re- sults of other studies also show many differ- ent genes to be pleiotropic for morpholog- ical traits. In Drosophila a recessive lethal gene called lethal-translucida causes pupae to be- # $ # # A- A fi # J? * ./ £ tr g & i dene Action 87 Richard Benedict Goldschmidt (1878-1958). (From Genetics, vol. 45, p. I, 1960.) tf aH* * J) L .- ?sji ^^^J|& Wi k9 V? Hfr^sS^j QUESTIONS FOR DISCUSSION 6.1. Can two genetically different individuals ever have identical viabilities? Explain. 6.2. Why can you not conclude, from the evidence presented, that the genes for MN blood type in man, or for auria phenotypes in the snapdragon, are pleiotropic? 6.3. How can genes be lethal to a genotype without producing a corpse? 6.4. Two curly-winged, stubble-bristled Drosophila are mated. Among a large num- ber of adult progeny scored the ratio obtained is: 4 curly stubble: 2 curly only: 2 stubble only:l neither curly nor stubble (therefore normal, wild-type). Explain these results genetically. 6.5. In Drosophila, a mating of <5 A X 9 B or of S CX $ D produces F,, Vi of which turn brown and die in the egg stage. If, however, the matings are t! A X 9 D or <3 C X 2 B, none of the F, eggs turn brown and die. How can you explain these results genetically? 6.6. In what respects are the terms penetrance and dominance similar and in what respects are they different? 6.7. Is it the gene for dull red eye color that is pleiotropic in Drosophila, or is it the allele for white eye color? Explain. 6.8. Most of the genes studied in Drosophila affect the exoskeleton of the fly. Do you suppose these genes also have effects on the internal organs? Why? 6.9. Would you expect to find individuals who are homozygous for Polydactyly? Explain. What phenotype would you expect them to have? Why? 6.10. Why are genes whose penetrance is 100% and expressivity is uniform particularly valuable in a study of gene properties? 6.11. Two normal people marry and have a single child who is polydactylous on one hand only. How can you explain this? SS CHAPTER 6 6.12. A certain type ol baldness is due to a gene thai is dominant in men and recessive in women. A nonhald man marries a bald woman and they have a bald son. Give the genotypes ol all individuals and discuss the penetrance of the genes involved. 6.13. A man has one brown eye and one blue eye. Explain. 6.14. How could you distinguish whether a given phenotype is due to a rare dominant gene with complete penetrance or a rare recessive gene of low penetrance? 6.15. In determining whether or not twins are dizygotic, why must one study traits tor which one or both parents arc hctcrozygotes? 6.16. Are mistakes ever made in classifying twins as dizygotic in origin? Why? 6.17. Can the gene P for Polydactyly be considered as being partially dominant? As having pleiotropic effects? Explain. 6.18. When nonidentical twins are discordant for ABO blood type, why must one or both parents have been heterozygous for /' or //;? 6.19. Invent a particular situation that would result in greater discordance for identical than for nonidentical twins. 6.20. What would be the probability of twins being dizygotic in origin if both have the genotype aa Bb CC Dd Ee Ff, each pair of alleles segregating independently, if the parents are genotypically Aa Bb CC DD Ee Ff and Aa BB CC dd ee FF? 6.21. How would you test whether, in women, there is a genetic basis for the matura- tion of more than one egg at a time? 6.22. In what way can you imagine that the paternal genotype could influence the frequency of twinning? 6.23. Is tuberculosis "inherited"? Explain. 6.24. What can twin studies by themselves tell you about genes? About genetic re- combination? 6.25. Is it valid to apply the conclusions from twin studies to nontwin members of the population? Explain. 6.26. Does this chapter present any new information about genetic properties? Explain. 6.27. In a genetically black strain of the house mouse, W. L. Russell found a mouse with a splotchy phenotype — having white spotting on the belly and occasionally on the back. Splotchy X black gives both splotchy and black types of progeny. Splotchy X splotchy also produces the same types, but a number of embryos die in utero at 14 days of age and are characterized by a kinky tail and spina bifida. Discuss the genetic basis for and the dominance relationships involved in these results. 6.28. It has been found that mouse ovaries transplanted from embryos to adult females can develop to maturity and produce offspring. Describe how you would proceed to determine the genotype of the abnormal embryos described in 6.27. 6.29. Do you agree with J. H. Sang that penetrance (P) and expressivity (E) ". . . are descriptive terms which cloak our ignorance of the underlying reactions which determine particular values of P and E in any situation"? Explain. 6.30. In the Japanese quail (Coturnix coturnix) matings between normal-appearing individuals of certain strains produce some micromelic embryos, having a short broad head with bulging eyes, which die between 11 and 16 days of incubation. How would you proceed to determine whether these abnormal embryos are homozygotes for a single recessive lethal gene? Phenotypic Effects of Gene Action 89 6.31. What conclusions can you draw from the data of B. Harvald and M. Hauge (J. Amer. Med. Assoc, 186:749-753, 1963) obtained from an unbiased sample of Danish twins? One twin Both twins cancerous Twin pairs cancerous At same site At different sites Identical 1528 143 8 13 Nonidentical 2609 292 9 39 6.32. In what way does the study of genes help us understand normal embryonic development? 6.33. If most somatic cells have the same genetic content, why do different cells not differentiate in the same way? 6.34. In what ways can genes regulate embryonic development? 6.35. Do the studies of Creeper, of Brachy, or of pituitary dwarfism in mice offer any support for the view that most, if not all, genes have a single, primary effect? 6.36. What is the relationship between phenogenetics, developmental genetics, physio- logical genetics, and biochemical genetics? 6.37. Discuss the comparative importance of genes that act earlier, as compared with those which act later, in development. 6.38. Do you suppose that all genes act at all times in all cells of the body? Why? 6.39. "This chapter tells more about development than it does about genes." Do you agree? Why? 6.40. What can be learned about gene action if the gene studied ( 1 ) has only two alternatives, (2) has many alternatives? Chapter 7 SEX CHROMOSOMES AND SEX-LINKED GENES S |i\( i sex has phenotypic alter- natives (maleness and female- ness ) . the genetie basis for sex can be investigated. This basis cannot be determined by studying garden pea plants because they are bisexual; that is. both sex- ual alternatives occur in one individual, and no phenotypic differences will be produced by genetic recombination. The typical Drosophila individual, however, being either male or female (Figure 2-6, p. 23), can be used to study the genetic basis for sex. When normal males and females mate to- gether, the male: female ratio of their prog- eny is approximately 1:1. This suggests the simplest hypothesis — that sex in Drosophila is determined by a single gene pair, one of the sexes being a homozygote and the other a heterozygote. For the moment, however. which sex corresponds to which genotype cannot be designated. In accordance with the view that chromo- somes contain the genes, one pair of chro- mosomes should be concerned with sex. Let us call the presumed homologous pair of chromosomes carried by the homozygote lor the sex genes the XX pair and those carried by the heterozygote, the XY pair. Segregation and random cross fertilization then would produce equal numbers of XX and XY progeny. The X and the Y chro- mosomes presumed to carry the genes for sex can be called sex chromosomes; each of the other chromosomes which an individual carries can be called an autosome (A). Since Drosophila melanogaster has a diploid chromosome number of four pairs, each individual can be represented as either XX + 3AA or XY + 3AA. Sex-Linked Genes Consider the results of certain crosses in- volving the recessives cubitus interruptus (ci) and ebony body color (e) and their dominant alleles ci+ (normal wing venation) and e+ (gray body color). One cross, ci+ci e + ehy ci ci e e — a dihybrid parent and a double recessive parent (Figure 7-1 ) — produces offspring in a 1:1:1:1 ratio thus + + ci ci e e ci ci e e + + + + 'Aci e, V4ci e, V-ici e, 'Aci e ci e + + Aci ci e e Aci ci e e 'Aci ci e e 'Aci ci e e FIGURE 7-1 . Results of hack- crossing a dihybrid. 90 ,S>.\ Chromosomes and Sex-Linked denes \)\ demonstrating that the two pairs of genes are segregating independently. The same result and conclusions hold for the cross of ci ci e? by ci ci e * e. Consider, next, crosses in which the sex and wing venation traits are studied simultaneously in recip- rocal matings — ci+ci XX by ci ci XY and ci ci XY by ci ci XX. In both cases the result is a 1:1:1:1 ratio of cubitus male, cubitus female, normal male, normal female. Here, then, the sex genes segregate inde- pendently of the genes for cubitus. There- fore, according to our hypothesis, the ci alleles are located autosomally. Similarly, each of the reciprocal crosses — e fXXbyee XY and e • e X Y by e e XX — also gives a 1:1:1:1 ratio, indicating that the gene for ebony is located autosomally. Since the genes for ebony and cubitus segre- gate independently of each other, they must be located on different pairs of autosomes. Even though we cannot yet specify which sex is XX or XY, the last two types of crosses can be described as reciprocally made backcrosses of a monohybrid; that is, one time the hybrid parent was the male, the other time the hybrid parent was the female. In both cases the two traits appear in a 1:1 ratio among the sons, and in a 1:1 ratio among the daughters. At this point, an earlier statement (p. 32) — that all crosses give the same results when made reciprocally — can be understood to mean that the observed phenotypes and their proportions are the same for sons and daughters even though the crosses were made reciprocally. So, for example, in a cross of the dihybrids ci+ci e + e by ci+ci e + e there would be a 9:3:3:1 ratio among the sons and a 9 : 3 : 3 : 1 ratio among the daughters because the parents' sex genes were located in the sex chromosomes, where- as the other gene pairs were in nonhomolo- gous pairs of autosomes. Previously, all the crosses we dealt with involved autosomal recombination. Because autosomal genes A B P -'9* whi^ P "hi,e CJ*^ lei'? 9 /hite dull red 0*0* ?? figure 7-2. Phenotypic results of reciprocal matings involving eye color. & 6 nudes. 9 9 = females. always segregate independently of the sex genes, sex did not influence the results; that is, the phenotypic results of autosomal re- combination crosses are the same for sons and daughters even though reciprocal mat- ings are made. But consider the results of crosses involv- ing the dull-red (w+) and white (w) eye color alleles. Using pure lines, dull-red 9 by white i (Figure 7-2 A) produces all dull-red sons and daughters in F,, as ex- pected, since w+ is dominant. However, the reciprocal cross (Figure 7-2B), white 9 by dull-red 6 , produces only white sons and dull-red daughters. Although the first cross gives the same result for sons as for daughters, the second (reciprocal) cross gives different results: sons resemble their mothers; daughters resemble their fathers. Because such different results are never ob- tained from reciprocal matings involving autosomal genes, we can conclude that w+ and its alleles are not located autosomally. Let us assume that the gene for white eye is located in the sex chromosomes and, therefore, is sex-linked and investigate the consequences of this on the gene's transmis- sion relative to the sex phenotype.1 If we assume that females are XY and males XX, the first cross then is dull-red female X""Y"" by white male X"X"' (Fig- ure 7-3, A-l ), and the F, expected are 1 See T. H. Morgan ( 1910). 92 CHAPTER 7 A-l B-l 9W W ^fl W W ("*) WW X xxx(j p, x y qp x X X (J W fl^^ WW fl ^ x OO F x x OCT +99 W99 + + W W ( 1 WW P, X Y x X X w w w X Y A-2 + + P, X X X : B-2 Q w w + + ^ w wO w w Q P,XXqpxXY a* + w w X Y w w X Y + w w X X A-3 99 B-3 + + O WW ^_^ W W ("") WW ^ p,xx YxXYU p, x x ; x x y Q w >99 w w X Y 0*0* ' x->99 figure 7-3. Three attempts (A-l and B-l, A-2 and B-2, A-3 and B-3) to represent matings A and B in Fig. 7-2 genotypically. Shaded genotypes must he incorrect. X"X" sons and X"Y"' daughters, all dull-red-eyed, as found. The reciprocal cross (Figure 7-3, B-l), therefore, is white 9 X"Y" by dull-red $ XW+XM'\ The Fi daughters (X"'*YW) are expected to be dull-red-eyed, as found. The Fi sons (X"*X" '), also expected to be dull-red-eyed, are, however, actually white-eyed. There- fore, we must reject this particular hypoth- esis for correlating sex chromosomes and eye color genes. So let us assume the reverse situation — that females are XX and males XY. The same crosses are represented now as dull-red Sex Chromosomes and Sex-Linked Genes 93 9 Xw+Xw+ by white 6 X"Y'r producing X"+X"' (dull-red) daughters and X"+Y'" (dull-red) sons (Figure 7-3, A-2); recip- rocally, white 9 X"'X'r by dull-red $ Xw+Yw+ produces Xw+Xw (dull-red) daugh- ters and XWYW* (dull-red) sons (Figure 7-3, B-2). Again an expected phenotype is contrary to fact — the phenotype of the Fi sons being white, not dull-red. Since we cannot explain these observa- tions by identifying maleness with XX or XY alone, we must increase the number of assumptions. Let us then test two hypoth- eses simultaneously: (1) Assuming that Drosophila males are XY and (2) the Y chromosome carries w but no other allele, then the genotypes and results of the first cross described in the preceding paragraph remain unchanged (Figure 7-3, A-3); the reciprocal cross (Figure 7-3, B-3) becomes white 9 X"X"" by dull-red <$ XM'+Y\ producing X"X" (dull-red) daughters and XI(Y"" (white) sons. Since these hypoth- eses fit the observations, we may accept them. The recombination genetics of several traits in Drosophila other than sex and eye color also proves to be based upon a pair of genes in the sex chromosomes; that is. each case gives different results in Fi when lines pure for different alternatives are crossed reciprocally. Moreover, each case can be explained by assuming that females are XX, males XY, and that the Y carries the most recessive and least effective allele known for the gene pair under test, as is w in the case of eye color. In such cases, the absence from the Y of a partially or completely domi- nant allele must mean that such alleles can- not be produced by mutation of the most recessive allele simply because this recessive allele does not exist on the Y. Accordingly, the Y ordinarily lacks an allele of a gene located on the X; therefore, in Figure 7-3, A-3 and B-3 a Y should be substituted for each Y" . Whenever the Y carries no allele of a gene on the X, sons will express phenotypically whatever allele is contained in the single X each son receives from his mother. With regard to these genes, therefore, a Drosoph- ila female is being test crossed whenever and to whomever she mates, since her X chromosome genotype can be determined di- rectly from the phenotypes of her sons. An otherwise diploid individual carrying one or more unpaired genes is said to be hemi- zygous in this respect. For example, a gene in the X chromosome with no allele in the Y is hemizygous in the Drosophila male; half of the zygotes he produces will receive this allele in the X he contributes, whereas the half receiving the Y will not get one. The X of a Drosophila male is obtained from his mother and transmitted to each of his daughters; the Y is transmitted from father to son. In poultry a mating of a female with non- barred feathers to a male with barred feathers produces offspring which are all barred — barred (B) being dominant to nonbarred (b) ( Figure 7-4 A ) . In the reciprocal cross (Figure 7-4), barred 9 by nonbarred $, all sons are barred and all daughters non- barred. Here also the results of reciprocal matings differ, so that we are dealing again with sex-linkage. In the reciprocal cross, note that the exceptional Fx are nonbarred, showing the recessive trait as in the case of Drosophila. But, in poultry the sex is op- posite, since the exceptional Fi are females. (The exceptional F! Drosophila were white- eyed males.) To explain these results we must assume that in poultry, as in Drosoph- ila, sex is determined by XX vs. XY, and that the X chromosome does and the Y chromosome does not contain a gene for barred or nonbarred feathers. But, contrary to Drosophila, poultry males are XX and females, XY. The genotypes of the bird crosses are, on these hypotheses, X6Y (nonbarred 9 ) by 94 < HAPTER 7 \ V (barred >, producing XBY (barred 9 ) and X'.V (barred ) in F,; the re- ciprocal mating of .VY (barred V ) by \ \ (nonbarred ; ) produces VY (non- barred 5 ) and V V (barred ) in F, ( Figure 7-4. A-l and B-l ). Support for die existence ol sex chromo- somes may be sought from etiological ob- servations. If the gene content of the X and > is different as in Drosophila and poultry. the cytological appearance of the two kinds oi sex chromosomes might also be different, i Note, however, that the preceding explana- tion of sex-linkage was made independently of any cytological expectation.) In Drosophila it is found cytologically (Figure 7-5) that three of the four pairs of chromosomes seen at mitotic metaphase cor- respond in the male and the female, homo- logs being \er\ similar morphologically. In the female the homologs of the fourth pair are also morphologically similar. In the male, however, only one member of this pair looks like its homologs in the female; its partner's morphology is distinctly differ- ent. Thus, the distinctive cytological ap- pearance of this last chromosome is con- sistent with the genetic expectation for a Y chromosome being present once in the male and not at all in the female. The other homolog in the male is then called the X and is present twice in the female. The reverse cytological picture is observed in poultry; here the homologs are similar for each pair of chromosomes in the male, whereas the female has one heteromorphic pair; that is, A Nonbarred ^ Barred f~Y?~) Barred QO x Barred B /-V^ P, Barred V/ x Nonbarred r^ a"a" 99 Barred Nonbarred b X Y A- V x X X (J B X Y 2 B-l b b * X x O B b X X x%99 B b X X X Y do' 99 FIGURE 7-4. Phenotypic (A and li) and genotypic (A-J and B-l) results of reciprocal nuttings involving barred and nonbarred feathers in chickens. Sex Chromosomes and Sex-Linked denes 95 FEMALE MALE A tf xx * figure 7-5. Silhouettes of chromosomes of Drosophila melanogaster as seen at mitotic metaphase. one pair whose members are morphologically different, one being similar to, one different from, the corresponding pair in the male. As in birds, the male of moths, butter- flies, and some amphibians and reptiles is XX and the females, XY. In human beings, genetic and cytological evidence shows XY to be male and XX to be female, just as in Drosophila. So in different species different sexes make two kinds of gametes; that is, different sexes are heterogametic with re- spect to sex chromosomes. In man, a certain kind of red-green color- blindness is sex-linked due to a recessive allele, c, present on the X and absent on the Y. Accordingly, color-blind women (X'XC) who marry normal men (XrY) have normal daughters (XrXr) and color-blind figure 7-6. Pedigree showing a woman homo- zygous for the gene for hemophilia. m^5~i r 6 b sons (X'Y). The classical bleeder's disease in human beings, hemophilia type A, is also due to an X-linked recessive gene, h, absent from the Y. This rare disease usually occurs in males. Recently, however, a few hemo- philic women have been discovered in Eng- land. These homozygotes are extremely rare because, barring mutation, they must have a hemophilic father (X'Y) and a heterozy- gous mother (X"X*) — (Figure 7-6). A. PHENOTYPES P, White O x DuM w-* B. GENOTYPES WW w X X x X Y F, TYPICAL EXCEPTIONAL X Y w w X X H [w wl X xj figure 7-7. Nonmutant exceptions in crosses involving eye color in Drosophila. Nondisjunction Certain additional experiments have been performed with the sex-linked gene for white eye in Drosophila.- When white females (X"X") are crossed with dull-red males (X" Y) and large numbers of progeny are scored, not all F( are white sons (X"Y) and dull-red daughters (Xw+Xw) as ex- pected according to sex-linkage. One or two flies per thousand Ft are exceptional dull-red-eyed sons or white-eyed daughters (Figure 7-7). These exceptional flies can- not be explained away as the result of care- less scoring of phenotypes or contamination. Moreover, they cannot be explained as being due to mutation, since the mutation frc- - Based upon work of C. B. Bridges. 96 CHAPTER 7 quencj from w to w or the reverse is several orders of magnitude lower than the observed frequency of exceptional Hies. Since the exceptional Fi females are white- eyed, each must cam \ V | Figure 7-7B). The onlj source of X's containing w is the mother. Accordingly, the father must fail to contribute his X"' chromosome to an exceptional daughter. Each exceptional dull-red-eyed son must carry X"", which could be contributed only by the father. In order to understand how this excep- tional situation may come about, let us examine the normal consequence of meiosis in the Drosophila female as regards the sex chromosomes. Normally, the two X's syn- apse and form a tetrad, and due to segrega- tion four nuclei are produced at the end of meiosis, each containing one X (Figure 7-8 A ) . One of the four nuclei becomes the gametic (egg) nucleus; the other three are discarded (in polar bodies). Suppose, however, segregation of the four strands in the X chromosome tetrad occa- sionally occurs improperly in either of two ways: 1. At anaphase I, instead of one dyad going to each pole, both dyads go to the same pole (Figure 7-8B). The daughter nucleus containing no X dyad undergoes the second meiotic division to produce two nuclei, neither one having an X. The other daughter nucleus, containing two dyads, pro- ceeds through the second division, during which the two members of each dyad sepa- rate and go to opposite poles at anaphase II. The result is two daughter nuclei each con- taining two X's, one from each dyad. Therefore, at the end of meiosis, the failure of dyads to disjoin at anaphase I will result ultimately in four nuclei, two with no X and two with two X's. As a consequence the nucleus which becomes the gametic nucleus has a 50% chance of carrying no X and a 50% chance of carrying two X's. 2. Alternatively (see Figure 7-8C), METAPHASE TELOPHASE figure 7-8. Consequences of normal segrega- tion of X chromosomes (A) and of its failure to occur (B and C) . anaphase I is normal, and at telophase I two daughter nuclei are formed each con- taining one X dyad. The second meiotic division occurs normally in one of the daugh- ter nuclei, producing two telophase II nuclei, each of which contains one X. In the other daughter nucleus, however, the members of the X dyad fail to separate at anaphase II and go instead into the same telophase II nucleus. This failure of monads to disjoin at anaphase II produces two nuclei, one con- taining no X and the other containing two X's. Consequently, the gametic nucleus has a 25% chance of carrying no X, a 25% chance of carrying two X's, and a 50% chance of carrying one X. The occasional failure of normal separa- tion of chromatids at either the first or the second meiotic division would result in the occasional production of eggs containing either no X or two X's. Such a failure of the members of a pair of chromosomes to segregate is also referred to as nondisjunc- tion of chromosomes. According to the hy- pothesis that the X chromosome carries an allele for w, chromosomal nondisjunction can provide the mechanism by which a pair of genes fails to segregate, with the result that after meiosis, eggs are sometimes pro- duced containing two members or no mem- Sex Chromosomes and Sex-Linked Genes 97 bers of the gene pair. Any egg produced following nondisjunction will usually be fer- tilized by a sperm carrying either an X or a Y in addition to a haploid set of auto- somes. (Nondisjunction can also occur dur- ing meiosis in the male. We can ignore this complication here, because it is an in- frequent event and the probability that an egg produced after nondisjunction would be fertilized by a sperm produced after nondis- junction is negligible.) If the hypothesis of chromosomal nondis- junction is valid, it should be consistent with the genetic results. After nondisjunction the exceptional eggs produced by a white (X"X,r) female would be either XWXW or 0 (zero designating the absence of the homolog normally expected to be present). The normal sperm produced by a dull-red (X't+Y) male would carry either Xw+ or Y. The expected genotypes of Fi following random fertilizations between these gametes are given in Figure 7-9. Let us momentarily ignore the sex of these exceptional offspring and classify them only for eye color. Type 1 would be dull-red- eyed, type 2 white-eyed, type 3 dull-red- eyed, and type 4's eye color undetermined. The genetic observations would be explained if types 1 and 4 were lethal; type 2, female; type 3, male. (On the hypothesis that XX is female and XY is male, it is reasonable to assume that types 1 and 4 would be neither, and therefore might be lethal. ) Even more specific requirements must be fulfilled before accepting these hypotheses, namely, that each exceptional white female must prove to be XXY cytologically; that is, such females must have, in addition to the normal diploid chromosomes of a female, an extra chromosome which is Y. More- over, each exceptional male must have, in addition to the normal autosomes, one X but no Y. When the somatic cells of excep- tional females and males are examined cyto- logically, these chromosomal prescriptions are found to be filled completely. It is also possible to show that YO zygotes are lethal, and that X"*X"X" individuals— these usu- ally die before adulthood — are dull-red-eyed. Although XY individuals are fertile males, XO flies are invariably sterile males. This, therefore, implies that the Y chromosome is necessary for male fertility, the trait being attributable to a gene on the Y which has no allele on the X. Moreover, our sex chro- mosome formula for maleness must be mod- ified to include XY and XO individuals and similarly the femaleness formula also modi- fied to include XX and XXY individuals. Chromosomes as Genetic Material We should now re-evaluate the hypothesis that the chromosomes serve as the material basis for genes. In preceding chapters, the following parallels were found between the properties and behavior of genes and of chromosomes: both come only from pre- existing counterparts; both are self-replicat- ing; both occur as pairs in all cells of the diploid stage of sexually reproducing organ- isms except gametes; both are replicated in each mitotic division; both maintain their individuality from one mitotic division to the EGGS SPERM OFFSPRING w w X X + w X (1) + www XXX w w X X Y (2) w w XXY 0 + w X (3) + w X 0 0 Y (4) Y 0 figure 7-9. Genotypic expectation after fer- tilization of nondisjunctionally produced eggs by normal sperm. 98 CHAPTER 7 next; both arc capable of imitation and sub- sequent replication of the new form; both segregate during gametogenesis so that they occur unpaired in the gamete; both show independent segregation for different pairs; both are combined randomly at fertilization. It was also hypothesized that the chromo- some is larger than a rccombinational gene (the smallest rccombinational unit of the genetic material ). since a gene was described as the largest distance along the length of a chromosome within which an exchange lead- ing to a chiasma cannot form. These parallels still might be considered merely coincidental. The present chapter provides the following additional tests of the idea that chromosomes function as the ma- terial basis for genes: Se.x-linka^e. detected by the nonrandom association between sex and the genes for certain traits, was found to be an exception to the mode of transmission of the auto- somal genes studied previously. This phe- nomenon was explained only by the assump- tion that certain genes have an allele in the homologous chromosome of a pair in one sex but no allele in the homologous chromo- some in the other sex. Hemizygosity was a necessary assumption in the case of the Drosophila male and the poultry female. This genetic aberration was exactly paral- leled by the occurrence of a pair of hetero- morphic homologs in the Drosophila male and poultry female, one homologous mem- ber being present as a pair in the female Drosophila and in the male chicken. Finally, in Drosophila, an exception to the exception of sex-linkage was found and ex- plained genetically as resulting from the failure of the members of a single pair of sex-linked genes to segregate. This failure produces gametes containing two or no alleles of a given sex-linked gene. This genetic nondisjunction was shown to result from chromosomal nondisjunction; that is, Calvin Blackman Bridgi s (1889-1938). (From Genetics, vol. 25. p. 1. 1940.) the members of a pair of X chromosomes failed to segregate during meiosis. From chromosomal nondisjunction, it was pre- dicted that the different genetically excep- tional individuals would have different spe- cific and unique sex chromosomal composi- tions, and further tests proved this was the case. In the light of these results, the view that the chromosomes are the material basis for all the genes so jar studied can no longer be considered merely a hypothesis based upon limited — therefore possibly circumstantial — evidence, but must now be accepted as a theory supported both by all the typical and all the atypical recombinational properties of genes and of chromosomes. Ordinarily, no further comment will be made in this book about new tests that sub- stantiate the theory. Henceforth, assume that all tests do so unless note is made to the contrary. Sex Chromosomes and Sex-Linked Genes SUMMARY AND CONCLUSIONS 99 Up to now we have studied genes located in autosomes. We have found that the recombination of autosomal genes is such that reciprocal crosses between different pure lines produce F, which are genotypically and phenotypically uniform; that is. there is no dependency between the traits which appear and the sex of the offspring because autosomal genes segregate independently of the genetic material in sex chromosomes. In Drosophila, sex is not the only trait directly associated with genetic material located in the sex chromosomes. Several other traits in Drosophila yield results which differ in reciprocal matings made between lines that are pure for different alternatives: the difference appearing in the phenotypes shown by one of the sexes. Genes behaving in such a sex-linked way are not located autosomally. The Y sex chromosome carries no allele of these genes, and the X does. In human beings and Drosophila, the XY sex chromosomal constitution is male and the XX female; in birds and moths, it is the female which is heteromorphic, and, therefore, heterogametic with reference to sex chromosomes. Occasionally, as a consequence of nondisjunction of sex chromosomes at meiosis, chromosome segregation fails, and gametes are formed containing two or, comple- mentarily, no sex chromosomes. When this nondisjunction occurs in a Drosophila female homozygous for an X-linked recessive gene and such a female is mated to a male carrying the dominant allele, some offspring appear that are simultaneously ex- ceptions to sex-linkage and to sex chromosome content; the exceptional feature of the one accurately predicts the exceptional feature of the other, and vice versa. Sex-linkage and nondisjunction offer additional tests of the hypothesis that the ma- terial basis of all the genes studied thus far is in the chromosomes. This view is sup- ported by so many and diverse lines of evidence, and contradicted by none, that it must be accepted as theory. REFERENCES Bridges, C. B., '"Non-Disjunction as Proof of the Chromosome Theory of Heredity " Genetics, 1:1-52, 107-163, 1916. Morgan. T. H., "Sex Limited Inheritance in Drosophila," Science, 32:120-122, 1910. Reprinted in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood' Cliffs N.J.: Prentice-Hall, 1959, pp. 63-66. QUESTIONS FOR DISCUSSION 7.1. Under what circumstances would sons fail to receive a Y chromosome from their father? 7.2. In the cross X""Y by X"X" what would you expect to be the genotypes of the zygotes produced, after sex chromosome nondisjunction during meiosis in both male and female Drosophila? What is the phenotypic outcome in each case? 7.3. If a trait is found to be due to a gene unlinked to any autosome, does this prove that the gene is linked to a sex chromosome? Explain. 7.4. A husband and wife both have normal vision, although both their fathers are red-green color-blind. What is the chance that their first child will be: (a) a normal son? (c) a red-green color-blind son? (b) a normal daughter? (d) a red-green color-blind daughter? 100 CHAPTER 7 7.5. One child is hemophilic, his twin hrother is not. (a) What is the probable sex of the hemophilic twin? (h) Are the twins monozygotic? Explain. (c) Give the genotypes of both twins and of their mother. 7.6. A hemophilic father has a hemophilic son. Give the most probable genotypes oi the parents and child. 7.7. A Drosophila male with cubitus interruptus wing venation, ebony body color, and white eye color is mated to a pure wild-type female (normal wing venation, gray body color, and dull-red eyes); then the F, females are crossed to males like their father. Give the kinds and relative frequencies of genotypes and of phenotypes expected among the offspring of the last cross. 7.S. Are you convinced that all genes have their material basis in the chromosomes? Explain. 7.9. What reason can you give for believing that in Drosophila the Y chromosome is lacking a gene present in the X chromosome? That the X is lacking a gene present in the Y? 7.10. List evidence in support of the theory that chromosomes contain the material basis for genes. 7.1 1. Has any evidence been presented that a chromosome carries more than one gene? Explain. 7.12. What proportion of all genes causing hemophilia type A is found in human males? Justify your answer. 7.13. Does a gene have to be hemizygous in one sex to be sex-linked? Explain. 7.14. Two phenotypically wild-type Drosophila were mated in a vial. By accident all but one Fx was lost. The survivor was a male with white eyes, ebony body color, and cubitus interruptus venation. Give the most probable genotypes of the parents. 7.15. Using pure stocks of Drosophila, yellow-bodied male by gray-bodied (wild-type) female produced 1241 gray-bodied daughters, 1150 gray-bodied sons, and 2 yellow-bodied sons. The reciprocal mating produced 1315 gray daughters, 924 yellow sons, and 1 yellow daughter. Give the genetic and chromosomal makeup of each type of individual mentioned. Discuss the relative viability and fertility of the different chromosomal types. 7.16. Females of Drosophila having a notch in their wing margins mated to wild-type males gave the following F, results: 550 wild-type 9 9, 472 notch 9 9, 515 wild-type 6 6 . Explain these results genetically. 7.17. A line of Drosophila pure for the sex-linked gene, coral (W") was maintained in the laboratory for many generations. To demonstrate sex-linkage to a class, a coral male was mated to a wild-type female, and all the F^ were as expected. The reciprocal cross, between a coral female and a wild-type male, gave 62 coral females and 59 wild-type males. Present a hypothesis to explain this unusual result. How would you test your hypothesis? 7.18. The wild-type eye shape in Drosophila is ovoid. A certain mutant, X, narrows the eye. Using pure lines, and ignoring rare exceptions, mutant 9 X wild-type $ produces mutant sons and daughters in ¥x; wild-type 9 X mutant S produces wild-type sons and mutant daughters in Fj. Another mutant. Y, also narrows the eye. Using pure lines of Y and wild-type, mutant 6 or 9 X wild-type 9 or 2 produces 2 mutant 6 $ and 9 9:1 wild-type £ $ and 9 9. Discuss the genetics of mutants X and Y. Sex Chromosomes and Sex-Linked Genes 101 7.19. Reciprocal matings using pure lines of Drosophila produce all wild-type F, from wild-type by vestigial wings. What are the genotypic and phenotypic expecta- tions if the Fa $ of this mating is crossed with a white-eyed vestigial-winged 9 w being sex-linked? s ' 7.20. Give two ways in which knowledge of sex-linked genes could be put to practical use. r 7.21. A normal man of blood type AB marries a normal woman of O blood type whose father was hemophilic. What phenotypes should this couple expect in their children and in what relative frequencies? 7.22. The diagram below is a partial pedigree of the descendants of Queen Victoria of England (II) which contains information regarding hemophilia only for gen- eration IV. In this generation, the entire symbol is filled in if the person has hemophilia. A heterozygote for hemophilia would have been represented by a half filled-in symbol. Fill in the symbols of previous generations using this system. & 111 = Princess Alice 112 = Leopold, Duke of Albany 1111 = Irene 1112 = Alexandra 1113 = Alice III5 = Victoria Eugenie I VI = Prince Waldemar of Prussia IV3 = Prince Henry of Prussia IV8 = Tsarevitch Alexis of Russia IV10 = Viscount Trematon IV12 = Alfonso I VI 7 = Gonzalo II IV HI ( )1 -n. oooo 12 3 4 5 6 7 8 OBO 9 10 11 12 13 14 15 16 17 (After J. B. S. Haldane.) ( 'hapter *8 SEX DETERMINATION Drosophila In Chapter 7. it was mentioned that the ordinary Drosophila melanogaster female is 3AA + XX and the male. 3AA + X + Y. One cannot decide the chromosomal basis for sex determination from these facts alone, however, since two variables are involved, the X's and the Y. Is the male a male be- cause he has a Y, because he has only one X, or because he has both one X and one Y? By knowing the sex of flies that carry — besides two sets of autosomes — either XXY (female). XX YY (female), or XO (male), we can see that the Y is not sex determining in this organism. (As has been indicated in Chapter 7, the Y is necessary for fertility. XO males having nonmotile sperm. ) Knowing that sex in Drosophila is cor- related with the chromosomal alternatives of XX versus X, one can ask: What is the detailed genetic basis for sex in terms of genes located in the X chromosome? The data so far presented can be interpreted to mean that only a single pair of genes (in the case of XX) or a single gene (in the case of X ) is the total genetic basis for sex de- termination. There are several implications in this interpretation. The X-linked sex gene need not have an alternative allele if the presence of one such gene produces one sex, and the presence of two, the other sex — dominance not being involved. It can also be claimed that the Y carries no allele for this sex gene. Two additional assumptions must be made, however, in Drosophila and in other species having heteromorphic sex chromosomes, to correlate the genetics with the cytology of sex. 1. That the sex gene must be located in a region of the X which distinguishes X from Y cytologically. 2. No chiasma may occur between X and Y within this cytologically different segment. These postulates are necessary to preserve the exact correspondence between the mor- phology of the X and its sex gene content. Consequently, even though a chiasma occurs between the X and Y in a segment which they share (for example, both carry an allele of bobbed), the resultant strands that appear cytologically as X will carry the sex gene, whereas those that appear as Y will not. These requirements are reasonable A. PHENOTYPiC RESULTS B. GENOTYPIC EXPLANATION Female x Male P, XX tra tra x XY tra tra G, 'AX trat 'AX tra VjX tra, 'AY tra F, V , 25% XY tra tra d* / 25% XY tra tra a* \ \ > 25% XX tra tra (_} (transformed ^p) O r n Pii.iiibIiii • 25% XX tra tra ? figure 8-1. Abnormal sex ratio in Drosophila. /d o Females 102 Sex Determination 103 since synapsis does not occur between non- corresponding regions of homologous chro- mosomes, and, in the absence of pairing, exchanges leading to chiasmata cannot occur. To further establish the cytogenetic basis for sex, we shall consider the results of crosses between certain laboratory strains of D. melanogaster.1 One strain produces about 75% males and 25% females (Figure 8-1 A), instead of the normal sex ratio of approximately 50% males and 50% females. Since just as many eggs become adult in this unusual strain as in a normal one, the abnormal result cannot be due to a gene that affects the viability of one sex. In this exceptional case it can be hy- pothesized that an autosomal gene is affect- ing the determination of sex. This gene is called transformer and is postulated to have two alleles, tra + , and tra. Homozygotes for tra are presumed always to form males re- gardless of the X genes present ( tra tra is epistatic and the X genes hypostatic), - whereas heterozygotes or homozygotes for tra + have their sex determined by the pres- ence of the sex gene on the X (in this case the X sex gene is epistatic). Accordingly, XX individuals that are also tra tra will ap- 1 Based upon work of A. H. Sturtevant. -These terms are defined in Chapter 4. pp. 51-52. figure 8-2. Some abnormal sex types in Drosophila: A = superfemale; B = supermale; C = intersex. {Drawn by E. M. Wallace.) Compare with normal male and female in Fig. 2-6. p. 23. pear as males {transformed females), ex- plaining the excess number of males in the progeny. Thus, a cross of XY tra tra (male ) by XXtra+ tra (female) (Figure 8-1B) produces one-fourth each XY tra tra (males), XY tra+ tra (males), XX tra tra (males, transformed females), XXtra+ tra (fe- males)— accounting for the numerical re- sults. All these assumptions have been tested in additional crosses and are con- firmed, proving that autosomal genes are also concerned with sex determination. Note, however, that the tra allele is very 104 CHAPTER 8 rare: almosl all Drosophila found in nature arc homozygous tra So far we have described only two sex types in Drosophila. Occasionally, how- ever, individuals occur which have, overall, an intermediate sexual appearance: that is. they arc both male and female in certain respects. Such sexual intermediates, called intersexes (see Figure 8-2). are sterile. Intersexes arc relatively frequent among the progeny oi' triploid (3N) females (whose chromosomes at mitotic mctaphase are dia- gramed in Figure 8-3; X chromosomes arc represented by filled-in blocks, autosomes h\ blanks, and the Y by a broken line). Some o\ the gametes of triploid females are haploid and some diploid; still others contain one. two, or three nonhomologs with or without a haploid set. Whereas haploid eggs produce normal males and females when fertilized by sperm from a normal male, diploid eggs produce triploid females when fertilized bj X-bcaring sperm. Diploid eggs produce XX Y individuals with three sets o\ autosomes, however, when fertilized by Y-bearing sperm. Some intersexes have this chromosomal constitution; other inter- sexes carry three autosomal sets and XX — one X derived from an egg containing two autosomal sets and the other from an X- bearing sperm. Close observation reveals two additional sex types among the progeny of triploid Drosophila (Figures 8-2, 8-3). These do not appear as intersexes but as sterile "super- sexes" — one type, called a superfemale, shows characteristic female traits even more strongly than does the normal female, the SUPERFEMALE i^^fl !&'??■>■ SUPERMALE FEMALE INTERSEX MALE FIGURE 8-3. Chromosomal complements of the sexual types found among the progeny of triploid females of D. melanogaster. Sex Determination 105 PHENOTYPES figure 8-4. Sex index and sexual type in D. mclano- gaster. other type, supermale, shows characteristic male traits even more strongly than does the normal male. Chromosomally, the superfemale contains two sets of autosomes and three X's; the X's are derived from an egg which carries one set of autosomes plus XX and is fertilized by an X-carrying sperm. The superfemale usually dies before adult- hood (see p. 97). The supermale con- tains three sets of autosomes plus XY; the chromosomes are derived from an egg which carries two sets of autosomes plus X and is fertilized by a Y-bearing sperm. What conclusions can we draw about sex determination from a knowledge of the chro- mosomal composition of different sex types in Drosophila? 3 Since we know that genes in the X and in the autosomes are sex-deter- mining, let us refer to Figure 8-4, which tabulates the number of X's and sets of auto- somes present for each sex type and also the ratio of X's to sets of autosomes — a numer- ical sex index. This index ranges from 0.33 for supermales to 1.5 for superfemales. Note that an index of 0.50 makes for male and that adding a set of autosomes can be 3 The following is based upon work of C. B. Bridges. Superfemale 3 " tetraploid 4 Normal . triploid 3 Female diploid 2 ^ haploid 1 Intersex 2 Normal male 1 Superma e 1 NO. X NO. SETS OF CHROMOSOMES AUTOSOMES (A sets) SEX INDEX No. X's No. A sets 1.5 1.0 1.0 1.0 1.0 0.67 0.50 0.33 interpreted as creating more maleness, pro- ducing the supermale. When the sex index is 1 .0, essentially normal females are pro- duced, indicating that the female tendency of one X overpowers the male tendency of one set of autosomes. But if the index is between 0.50 and 1.00, intersexes are pro- duced, indicating, by the same line of rea- soning, that the effect of two X's is partially overpowered by the extra autosomal set present. Finally, when the sex index is 1.5, the female tendency of the X's becomes so strong that superfemales result. These results strongly suggest that sex de- termination is due to the balance of genes located in the X on the one hand, in the autosomes on the other. According to this view, only the balance of the genes involved is important, so that a sex index of 1 .0 should (and does) produce a typical female, whether the individual is diploid (2X + 2 sets of A), triploid (3X -f 3 sets of A), or tetraploid (4X -f 4 sets of A). Individuals that contain haploid (IX -f- 1A set) sections have been found and, as expected from their sex index of 1.0, these parts were female. Since all known facts support the exact cor- respondence between chromosomal constitu- tion and sexual types, we can accept chromo- 106 CHAPTER 8 some balance as the typical basis of sex de- termination in Drosophila. What is the relationship between \-auto- some balance and tra, the sex-transforming gene'.' Sex is determined by X-autOSome balance when the individuals carry tra + , which they normally do. When tra is homo- zygous, however, the balance view does not apply and 2X + 2A sets produces a male. Gynandromorphs On relatively rare occasions, abnormal Dro- sophila appear with some of their parts typically male and the remainder, typically female. Such individuals are said to be mosaic for sex traits; sex mosaics are also called gynandromorphs or gynanders (Fig- ure 8-5). The male and female parts are clearly demarcated in such flies, sometimes the front and hind halves, at other times the right and left sides are of different sex. The sharp borderline between male and female parts in an insect gynander is due to the relatively small role that hormones play in insect differentiation, so that each figure 8-5. D. melanogaster gynandromorph whose left side is female and right side is male. (Drawn by E. M. Wallace.) body part is formed according to the geno- type it contains. In view of the preceding discussion, one would predict that the dip- loid cells in the female part of a gynander contain XX and those in the male part X, the chromosome number being otherwise normal. If this prediction is correct, then approximately half-and-half gynanders could originate as follows: the individual starts as a zygote containing 3AA -+- XX — that is, as a female. The first mitotic division of the zygotic nucleus is abnormal — one daughter nucleus contains 3AA + XX and is nor- mal, the other daughter nucleus contains 3AA + X and is defective, because one of the X's failed to be included in this nucleus, degenerated, and was lost. However, sub- sequent nuclear divisions are normal — cells produced following mitosis of the XX nucleus and its descendants giving rise to female tissue, and cells derived from the X nucleus giving rise to male parts. In this case the gynander has about half its body male and half female. If, however, the X is lost at some later mitosis, a correspondingly smaller portion of the body will be male, explaining gynanders one quarter or less male. We can test whether this explanation is sometimes correct by making use of an X- linked gene which produces a phenotypic effect over a large portion of the body sur- face; that is, a gene that affects the size and shape of the bristles and hairs. Such a gene is forked, two of its mutant alleles being /346 and /. In homozygotes (females) and hemizygotes (males), /::"' produces bristles and hairs of normal length and shape; / causes them to be shortened, split, and gnarled. The f'ih f heterozygotes have bristles and hairs slightly abnormal in these respects, showing a "weak forked" pheno- type. If a cross is made to produce female offspring that are f"ib/f heterozygotes, the following predictions can be made regarding the phenotype of the gynanders occasionally Sex Determination 107 present among the siblings: All gynanders, originating as postulated, will be weakly forked in their female parts; their male parts will have either normal or strongly forked bristles and hairs, depending upon whether the lost X carried / or f:1\ respectively. Experimental results obtained confirm ex- actly these expectations. Gynanders also occur in moths. Whereas male moths usually have large, beautifully colored wings and females, small stumps of wings, gynanders have been found with wings like the male on one side and those like the female on the other side. The ex- planation for these exceptions is similar to that given for Drosophila. In the case of the moth, however, the gynander usually starts as a male zygote (XX). Although most gynanders in Drosophila and other insects in which the male has the heteromorphic sex chromosomes, can be ex- plained in this manner, some gynanders originate another way. In extremely rare cases, an abnormal egg is produced after meiosis which contains not one but two hap- loid gametic nuclei. Because polyspermy sometimes occurs in insects — that is, more than the one sperm normally involved in fertilization enters an egg — one of the two haploid egg nuclei may be fertilized by an X-carrying sperm, the other by a Y-carrying one. The resultant individual is approxi- mately a half-and-half gynander. This type of gynander can be identified if the two paternal (or the two maternal) haploid gametic nuclei are marked differently for a pair of autosomal genes. Man and Mouse In human beings sexual type is determined at fertilization, XY zygotes becoming males; XX zygotes, females. In early development, all sex organs or gonads are neutral; that is, they give no macroscopic indication whether they will later form testes or ovaries. The early gonad has two regions, an outer one, the cortex, and an inner one, the medulla. As development proceeds, the cortex degenerates in those individuals that carry a Y (male), and the medulla forms a testis; in individuals genetically determined to be females, the medulla degenerates, and the cortex forms an ovary. Once the testis and ovary are formed, they take over the regulation of further sexual differentiation by means of the hormones they produce. The hormones direct the de- velopment or degeneration of various sexual ducts, the formation of genitalia, and other secondary sexual characteristics. Since sex- ual differentiation is largely controlled by the sex hormones, it is not surprising that genetically normal individuals are morpho- logically variable with regard to sex. Any change in the environment that can upset the production of, or tissue response to, sex hormones can produce effects which modify the sex phenotype. So, the phenotypes nor- mally considered male and female show some variability — providing some of the spice of life. Genetically normal persons exposed to abnormal environmental conditions can dif- ferentiate phenotypes that lie between the two normal ranges of sex type, and, there- fore, are intersexual in appearance. Though it is sometimes easy to classify an individual as being an intersex because the person is clearly between the two sex norms, other individuals at the extremes of normality can- not readily be labeled normal, or intersex, or supersex. Intersexual phenotypes due to environmental factors can result either from genotypic males who have developed par- tially in the direction of female, or from genotypic females partially differentiated in the direction of male. Otherwise-diploid individuals are known who have various numbers of sex chromo- somes. Only one type has a single sex chro- mosome; this is the X0 individual, who is female. The typical phenotypic effect of this condition is called Turner's syndrome 108 CHAPTER 8 (after its discoverer) and is characterized by the failure to mature as a woman. Turner type females usually do not develop breasts. ovulate, or menstruate. Because of vari- ability in the genotypic details and in the environment (including medical treatment), considerable variation occurs in the pheno- ls pic consequences of the X0 condition. In fact, one woman of this constitution is known to have given birth to a normal (XY) son. The X0 mouse is apparently less variable phenotypically since it always seems to pro- duce a fertile female. The other single sex chromosome type, YO. presumably lethal in man, is known to be lethal in mouse. Otherwise-diploid individuals having three sex chromosomes are of three types: XXX is female (sometimes mentally defective); XYY is male; XXY is male. The XXY in- dividual who is characteristically sterile, may have undersized sex organs, and may de- velop various secondary sexual characteris- tics of females, possesses Klinefelter's syn- drome (named after its discoverer). Along with the X0 female, he is phenotypically variable; for instance, some Klinefelter males are mentally retarded, others are not; al- though all those presently known are sterile, some show normal sexual drive and be- havior. In the mouse, XXY is a sterile male. Otherwise-diploid persons of the follow- ing additional types are also known: XXXX ( 9 ) ; XXXY ( $ ) ; XXYY ( $ ) ; XXXXX ( 9 ) ; XXXXY ( $ ); XXXYY ( $ ). Con- trary to the situation in Drosophila, it is clear from all these results that the Y chro- mosome is the primary sex-determining chromosome in man and mouse. Presence of a single Y determines the sex as male; absence of a Y produces the female. All individuals require an X in order to be viable. The Y versus no Y sex-determining mech- anism in human beings and mice implies that the Y must carry one or more genes for maleness in that portion which makes it cytologicaUy unique, the X having no cor- responding allele(s). Admitting that the presence of gene(s) for maleness on the Y makes for male, what is genetically responsi- ble for the femaleness produced in the ab- sence of the Y? Clearly other genetic factors are present — not limited in location to the Y chromosome — which affect sex and, there- fore, femaleness. The female tendency often shown by the human XXY suggests that the X contains genes affecting normal sexual differentiation which, when present in excess, cause a shift toward femaleness. Presum- ably, the X also has this capacity when Y is absent. All cases in which the entire body seems to contain an abnormal number of sex chro- mosomes can be explained as the result of nondisjunction leading to chromosome loss or gain which occurs either during meiosis or at an early cleavage division — probably the first — of the fertilized egg. Such nondis- junctions are correlated in human beings with the mother's advanced age at the time of pregnancy. By following the distribution of X-l inked mutants, it has been shown, however, that the nondisjunction which produces an ab- normal sex-chromosome number sometimes involves the paternally contributed sex-chro- mosome material. This origin is exemplified by a red-green colorblind father having an X0 daughter of normal vision. Since cer- tain aged Drosophila eggs cause the loss of paternal chromosomes after fertilization, it is important to recognize the possibility that the loss of a paternal chromosome in man can occur post- as well as pre-meiotically. Due to a premeiotic paternal nondisjunction colorblind women can, of course, have XXY Klinefelter sons of normal vision. A considerable number of persons having different chromosomal compositions in dif- ferent body parts are mosaic for sex chro- mosomes. These include the following Sex Determination 109 mixed constitutions: XXX/XO; XX/XO; XY/XO, XXY/XX; XXXY/XY. Such cases are usually due to one or more errors in chromosome distribution among the daughter nuclei produced after fertilization. Although such individuals are sex-chromo- some mosaics, and some may even have one ovary-like and one testis-like gonad, they are not gynanders in superficial char- acteristics because of their whole-body dis- tribution of sex hormones. Although the XXY male is often clearly an intersex, the X0, XXX, and so on females that show in- complete maturity are best considered infra- females, being underdeveloped sexually. It should now be clear that some specific phenotypic sexual abnormalities may be based primarily either on an abnormal en- vironment or on an abnormal chromosomal composition (recognizing also the possibility that mutants other than those involving an abnormal number of sex chromosomes can affect sex). Accordingly, chromosomal counts are often desirable in order to deter- mine the cause — and, hence, the treatment — of sexual abnormality. Human Sex Ratio Consider how the genotype is related to the sex ratio, that is, to the relative numbers of males and females born. On the average, 106 boys are born for each 100 girls. This statistic might be surprising at first, since half the sperm are expected to carry X, half Y, and all eggs, an X, the ratio of boy to girl expected at conception is one to one. Even if the four meiotic products of a given cell in spermatogenesis usually carry X, X, Y, Y, there is the possibility that during or after spermiogenesis (conversion of the telo- phase II cell into a sperm) some X-bearing sperm are lost. This possibility is supported by a report 4 that human ejaculates contain sperm heads of two sizes and shapes (Fig- ure 8-6); the smaller type sufficiently in 4 By L. B. Shettles (1960). figure 8-6. Head shapes in human sperm. Round-headed sperm are reported to be smaller and more numerous than oval-headed sperm, suggesting these carry the Y and X chromo- somes, respectively. {Courtesy of L. B. Shettles.) excess to explain an excess of males at fer- tilization provided the smaller sperm con- tains the small Y chromosome, and the larger sperm carries the larger X chromosome. Other evidence suggests that at conception males are much more numerous than fe- males; since more male fetuses normally abort than female, the numbers of boys and girls are more nearly equal at the time of birth than they were at conception. A study of the sex ratio at birth shows that the ratio 1.067:1.000 is found only among young parents, and that it decreases steadily until it is about 1.036: 1.000 among the children of older parents. How may this significant decrease be explained? Per- haps in older mothers there is a greater chance for chromosomally normal male 10 CHAPTER 8 babies to abort, or tor chromosome loss in the earliest mitotic divisions o( the fertilized egg. If the chromosome lost is an X and the zygote is \Y. the loss is expected to be lethal, so that a potential boy is aborted. If the zygote losing an X is XX. a girl can still be born. Moreover, if the chromosome lost in the XY individual is a Y, a girl can be born instead of a boy. Part of the effect must be due to the increase in meiotic non- disjunction with maternal age (zygotes of XXX type form viable females, whereas zygotes of Y0 type are expected to abort). We must include the possibility that the fathers may also contribute to this shift in sex ratio. Postmciotic selection against Y- carrying sperm may increase with paternal age. Or. as fathers become older, the XY tetrad may be more likely to undergo non- disjunction to produce sperm containing re- spectively, X, X, YY, 0. The first two can produce normal daughters; the last one can produce an underdeveloped X0 daughter; and only the YY is capable of producing males. Even though the XYY individual is male, it may frequently abort. Other genetic and nongenetic explanations for the shift in sex ratio with age are also possible. This discussion merely demonstrates how the basic facts of sex determination, chro- mosome loss, and nondisjunction may be used to formulate various hypotheses whose validity is subject to test. When many pedigrees are examined for sex ratio, several consecutive births of the same sex occasionally occur. This phe- nomenon could, of course, happen purely as a matter of chance when enough pedigrees are scored. One family, however, is re- ported to have only boys in 47 births and, in another well-substantiated case, out of 72 births in one family, all were girls. In both these cases the results are too improb- able to be attributed to chance. We do not know the basis for such results in man, but two different cases of almost exclusive female progeny production in Dro- sophila might suggest an explanation for those human pedigrees in which only one sex occurs in the progeny. In the first case, an XY male carrying a gene called sex ratio is responsible. Because of this gene, the X and Y fail to synapse, and the X replicates an extra time to form a tetrad; since almost all Y chromosomes degenerate during meio- sis, almost all sperm carry an X. In the second case, a female transmitting a spiro- chaete microorganism to her offspring through the egg is responsible. Such a fe- male mated to a normal male produces zygotes which begin development; soon thereafter the XY individuals are killed by the spirochaete, leaving almost all female survivors. The sex ratio can be controlled if the genotypes of the zygotes formed can be con- trolled. Since X- and Y-bearing sperm of men apparently differ in cytological appear- ance (Figure 8-8), it should be possible to separate them and thereby control the sex of progeny. Using various animal forms, such experiments have been performed with some success by Russian, American, and Swedish workers, using electric currents or centrifugation. Although these experiments have been encouraging, the results are not yet consistent, and the techniques not yet suitable for practical use. Hymenoptera In Hymenoptera (for example, bees, ants, wasps, and saw flies) unfertilized eggs de- velop as males (haploids) and fertilized eggs, usually, as females (diploids). Hap- loid males produce haploid sperm via suit- able modifications of the meiotic process, and all gametes of males and females have morphologically identical chromosomal com- positions. In the parasitic wasp, Habrobracon jug- landis, when the parents are closely related, some of the sons are haploid, but others are Sex Determination 111 diploid having ten pairs of chromosomes like their sisters. Genetic study shows that such diploid males have a biparental origin. Not only are diploid males relatively in- viable, but the hatchability of sibling eggs is very poor. A study of intrastrain and interstrain breeding supports the interpreta- tion 5 that a multiple allelic series determines sex in this form. With respect to this sex- determining locus or chromosome region, haploids are males, diploid heterozygotes are females, and diploid homozygotes are semisterile males. Role of Genotype in Sex Determination In certain organisms, male and female gam- etes are produced in the same individual. Animals of this type are said to be hermaph- roditic (after Hermes and Aphrodite), and plants, monoecious. The hermaphrodite snail, Helix, has a gonad which produces both eggs and sperm from cells which some- times lie very close together. In the earth- worm, eggs and sperm are produced in sepa- rate gonads located in different segments of the body. In certain mosses, egg and sperm- like gametes are also produced in separate sex organs (located on the same haploid gametophyte). In all these cases, the two types of gametes are produced by an organism that has but a single genotype; that is, one that is not genetically mosaic. Nevertheless, it might be supposed, at first, that the haploid geno- type carried by eggs and by the sperm is different and causes the difference in pheno- type and behavior. In the case of the gametophyte of mosses, however, the indi- vidual is haploid and so are both types of gametes it forms. Accordingly, in such or- ganisms we cannot expect differences in gene content to be the basis either for the formation of gametes or for the different types of gametes produced. 5 See P. W. Whiting (1943). Gamete formation in hermaphroditic and monoecious organisms, therefore, must de- pend primarily upon environmental differ- ences. Such differences must exist even between cells which lie close together, as is the case in Helix. It is reasonable to sup- pose that the same kinds of environmental factors which can direct one group of cells to form muscle cells and an adjacent group to form bone cells, can direct the differen- tiation of still other cells to make gonadal tissue in which adjacent cells can further differentiate as sperm and egg. Note, however, that sex involves another kind of differentiation, which, at least in organisms like the mosses, is separate from the type of gamete formed. This problem (which will not be discussed in detail here) concerns the genetic and environmental fac- tors responsible for the onset of meiosis, which is, of course, the feature most funda- mental to the success of the sexual process as it presently occurs in many species. In the examples already mentioned, the type of gamete differentiated depends upon the different positions which cells have within a single organism; consequently, they are subject to differences in internal and external environments. In the marine an- nelid, Ophryotrocha, the two sexes are in separate individuals, and the sex type formed is determined by the size of the organism. When the animal is small, because of youth or because it was obtained by amputation from a larger organism, it manufactures sperm; when larger, the same individual shifts to the manufacture of eggs. In this case the environment of the gonad is changed by the growth of the organism. Finally, consider sex determination in the marine worm, Bonellia, in which the sepa- rate sexes are radically different in appear- ance and activity — females being walnut- sized and having a long proboscis, males being microscopic ciliated forms that live as parasites in the body of the female. Fer- 112 CHAPTER 8 tilized eggs, grown in the absence of adult females develop as females; they develop as males in the presence either of adult fe- males or simply an extract of the female's proboscis. In this case, then, differentiation as a whole including sexual differentiation, is regulated by the presence or absence of a chemical messenger manufactured by fe- males. Nothing has been stated about the specific genetic basis for the determination or dif- ferentiation of sex in any of the examples given in this section because different sexes or gametes are determined not by genetic differences between cells, organs, or indi- viduals, but by environmental differences acting upon a uniform genotype. The genes, nevertheless, must play a role in all these cases by making possible different sexual responses to variations in the environment. Importance of Sexuality Even if reproduction occurred only by asex- ual means, the earth would now be popu- lated by genetically different kinds of or- ganisms, each variant having arisen by muta- tion in a pre-existing individual who was, in turn, produced from an unbroken line of descent. This method of direct descent is inefficient, however, since biologically fit individuals must wait for the rare occurrence of mutation to make them more fit. The biological innovation of sexuality has a tremendous genetic advantage over ascx- uality by providing genetic recombination which speeds up the process of the evolu- tion of more adaptive organisms. A more adaptive genotype may be produced in one individual by the combination of allelic and nonallelic genes originally located in two parents who, individually, may have been less well or even poorly adapted. Since genetic recombination normally occurs each generation for each gene pair, adaptive com- binations of genes originate much more rap- idly by recombination than by the relatively rare event of mutation. It should be clear, therefore, that sexuality, which produces a greater variety of adaptive genotypes in a given period of time than asexuality, is pri- marily responsible for the great variety of adapted kinds of individuals that have ap- peared on the Earth in recent times. SUMMARY AND CONCLUSIONS An understanding of the basis of sex requires the answer to two questions: What is responsible for the onset of meiosis? What is the basis for the formation of different kinds of gametes? Only the latter question is discussed in significant detail. In some cases the environment and in other cases the genotype is primarily responsible for sex determination. In the latter cases, sexual differences can often be correlated with cytogenetic differences. Genes responsible for sex determination are located not only in the sex chromosomes but in the autosomes as well. Although sex type may be changed through the action of a single pair of genes, a given sex is usually the result of the interaction of several, and probahly many, pairs of genes. Sex is, therefore, a polygenic trait (Chapter 5). Chromosomal differences found among zygotes serve as visible manifestations of differences in the balance of genes concerned with sex. Whenever, as in female Drosophila, genie balance is unaffected by the addition or subtraction of whole sets of chromosomes, sex also is unaffected. However, changes in chromosome number which produce intermediate genie balances also produce intermediate sex types — intersexes; those which make the balance more extreme than normal produce extreme sex types — supersexes. Sex Determination 113 These principles of sex determination apply also to human beings. In man and many other organisms, a large part of sexual differentiation is controlled by sex hor- mones produced by the gonads. This type of control rarely, if ever, permits the occur- rence of individuals who are typically male in one part and typically female in another part; it may also contribute to the formation of abnormal sex types for nongenetic reasons. REFERENCES Bangham, A. D., "Electrophoretic Characteristics of Ram and Rabbit Spermatozoa," Proc. Roy. Soc, Ser. B, 155:292-305, 1961. Bridges, C. B., "Sex in Relation to Chromosomes and Genes," Amer. Nat., 59:127-137, 1925. Reprinted in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, 1959, pp. 117-123. Goldschmidt, R. B., Theoretical Genetics, Berkeley and Los Angeles: University of California Press, 1955. Hannah-Alava. A.. '"Genetic Mosaics," Scient. Amer., 202:118-130, 1960. Lancet, No. 7075, Vol. 1, 1959, pp. 709-716. McKusick, V. A., Human Genetics, Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1964. Shettles, L. B., "Nuclear Morphology of Human Spermatozoa," Nature, London, 186: 648-649, 1960. Shettles, L. B., "Nuclear Structure of Human Spermatozoa," Nature, London, 188- 918-919, 1960. Sturtevant, A. H., "A Gene in Drosophila melanogaster that Transforms Females into Males," Genetics, 30:297-299, 1945. Whiting, P. W., "Multiple Alleles in Complementary Sex Determination in Habro- bracon," Genetics, 28:365-382, 1943. QUESTIONS FOR DISCUSSION 8.1. If sexual reproduction is as advantageous as discussed, why do so many organ- isms still reproduce asexually? 8.2. Does the study of sex determination offer any test of the theory that chromo- somes furnish the physical basis for genes? Explain. 8.3. Is it possible to consider the factors responsible for the meiotic process separately from the factors responsible for gamete formation? Explain. 8.4. Why is meiosis the most fundamental feature in the success of sexuality? 8.5. Give the genotypes and phenotypes of the unexceptional, the nondisjunctional, and the gynandromorphic offspring expected from a mating of f34b/f with / Drosophila. 8.6. Are there isoalleles for the genes determining the size and shape of the bristles and hairs of Drosophila? Explain. 8.7. Using first the autosomal alleles e and e+ and then the X-linked alleles y and y + , devise crosses by which you could identify gynanders in Drosophila resulting from two fertilizations of a single egg. Ill ( HAPTER 8 S.S. Compare the genotypes and phenotypes of sex chromosome mosaics of flies, mollis, and men. 8.9. All human beings have the same number o\ chromosomes in each somatic cell. Discuss this statement giving evidence in support of your view. 8.10. The following types oi mosaics are known in human beings: XXX/XO XXY/XX XX/XO XXXY/XY XY/XO Give a reasonable explanation lor the probable origination of each. 8.11. In human beings, can the members of a pair of monozygotic twins ever be of different sexes? Explain. S.I 2. Does a gene have to have an alternative allele before it can be discovered? Explain. 8.13. Assuming each homolog carried a different allele, a1, a2, a'\ o\ the same gene, make a schematic representation of a trivalent as it might appear during synapsis. Show diagrammatically the chromosomal and genetic content of the four meiotic products that could be obtained from your trivalent diagram. 5.14. Ignoring chiasma formation, how many chromosomally-different kinds of eggs can a triploid Drosophila female produce? How many of these eggs have more than a 5% chance of occurring? 8.15. (a) Scurfy, sf, is an X-linked recessive gene that kills male mice before they repro- duce. How is a stock containing this gene maintained normally? (b) Occasionally, the stock containing this gene produces scurfy females which also die before reproductive age. Suggest a genetic explanation for these female exceptions. Describe how you would test your hypothesis genetically by trans- planting ovaries and obtaining progeny from them. 8.16. What explanations can you offer, other than those already presented, for the shift in sex ratio with age of human parents? 8.17. No Y0 human beings are known. Why is this chromosomal constitution con- sidered to be lethal? 8.18. List the types of human zygotes formed after maternal nondisjunction of the X chromosome. What phenotypes would be expected for each of the zygotes that these, in turn, may produce? 8.19. List specific causes for the production of abnormal sex types in human beings. 8.20. How can you explain that only one X0 individual is known to have had a suc- cessful pregnancy, whereas other XO's are sterile? 8.21. Discuss the general applicability of the chromosomal balance theory of sex determination. 8.22. In Drosophila, why are gynanders not intersexes? Is this true in man also? Explain. 8.23. What chromosomal constitution can you give for a triploid human embryo that is "male"? ■"Female"? 8.24. A non-hemophilic man and woman have a hemophilic son with Klinefelter's syndrome. Describe the chromosomal content and genotypes of all three indi- viduals mentioned. Sex Determination 115 8.25. A white cat reported by H. C. Thurline, had one yellow and one blue eye, a phallus and one testis, one uterine horn, and one ovary. Although the animal had 38 chromosomes (the normal diploid number), some nuclei had an XX and others, an XY content. Suggest hypotheses to explain the chromosomal content of this individual. 8.26. Give a possible chromosomal formula for human individuals who are: (a) Triploid males (b) Klinefelter's type of male 8.27. A 26-year-old somewhat mentally retarded man is known to be XYY but other- wise diploid. To what do you attribute the apparent rarity of this type of chromosomal constitution? 8.28. Klinefelter-type males occur who are XXXYY. Give a possible origin of this chromosomal constitution. 8.29. In the insect Protenor and certain short-horned grasshoppers, all eggs have the same number of chromosomes, and half of the sperm are different, in that they carry one less chromosome. What is the cytogenetic basis for sex determination in such cases? 8.30. In the plant genus, Melandrium, one observes individuals of the following types: Diploid: XX + 11AA-5 XY + 1 1 AA = 6 Triploid: XXX + 1 1 AAA = 9 XXY + 1 1 AAA = <$ Tetraploid: XXXX + 1 1 A AAA = $ XX YY + 1 1 AAAA = $ or XXXY + 1 1 AAAA = $ Discuss the cytogenetic basis for sex determination in Melandrium. 8.31. Discuss the sex ratio expected in the honey bee from unmated and mated females. 8.32. Does the chromosomal balance hypothesis of sex determination apply in the case of parasitic wasps? Explain. 8.33. Compare the self-sterility alleles in Nicotiana (see p. 60) with the sex-determination alleles in Habrobracon. Chapter 9 LINKAGE AND CROSSING OVER BETWEEN GENES T: Jhe alleles of a gene pair af- fecting the seed coat of the garden pea were symbolized in Chapter 4 as round (/?) and wrinkled (r). This symbology follows the conven- tion that uses upper and lower case of the first letter (or so) of the phenotype pro- duced by the dominant allele — the one usu- ally found in nature — to represent the domi- nant and recessive alleles, respectively. In other conventions (see Figure 9-1), the first letter (or so) of the recessive trait (wrinkled) is used in lower case for the recessive allele (w), and the normally dom- inant allele (round) is given as one of the following: the same symbol in upper case (W); a + symbol as a superscript or base to the lower case symbol (vv+ or +"); or -+- alone. Henceforth in this book we will usually use one form of the -f- system for symbolizing genes. In this system, a mutant gene — Beadex, for example — which is dom- inant to the normal wild-type allele is rep- resented by one (or more) letters of which the first is capitalized (Bx or + "■'') and its wild-type allele is + (or Bx+). The hy- brid + vv can be represented as = or or r WW -\-/w to show that these alleles are on dif- ferent members of a pair of homologous chromosomes. Each of the first seven pairs of genes studied in the garden pea (see Chapter 4, p. 48) appeared to segregate independently. If this kind of segregation is attributed to 116 each gene pair being located in a dilferent one of the seven pairs of chromosomes car- ried by this organism, what result will be obtained when an eighth pair of genes, showing dominance and affecting an unre- lated trait, is included in such a study? When a dihybrid is made of one of the seven gene pairs and the eighth pair men- tioned above, a 9:3:3:1 phenotypic ratio is obtained when the dihybrid is self-ferti- lized, and a 1:1:1:1 phenotypic ratio is obtained when the same dihybrid is test crossed to the double recessive. These two independent tests demonstrate that the two pairs of genes involved are segregating in- dependently. The phenotypic ratios are radically different, however, when a dihy- brid is made with still another of the seven gene pairs — the one affecting seed coat — W w + w w + w w +_ w — +/ w w figure 9-1. Various ways of representing the round-wrinkled hybrid by gene symbols. and the same eighth. The other pair of genes involved (the eighth) determines the presence and absence of tendrils — the threadlike structures serving as a means for attachment as the plant climbs. The ten- drilless allele (t) is recessive. When a dou- ble recessive pea plant — wrinkled, tendril- less (w w 1 1) is crossed to a pure double dominant — round, tendrils ( + + + + ), all Fi are round with tendrils (+w +/), as expected. When the F, are self-fertilized (dihybrid by dihybrid), the following re- sults are obtained in F>: Phenotype No. Individuals round, tendrils 319 round, tendrilless 4 wrinkled, tendrils 3 wrinkled, tendrilless 123 Linkage and Crossing Over Between Genes 117 Note that each gene pair shows segrega- tion in the F2 since the ratio of round to wrinkled is 323:126 (a 3:1 ratio), and the ratio of tendrils to no tendrils is 322:127 (a 3:1 ratio). Had these gene pairs been segregating independently, the resultant ra- tio would have been 9:3:3:1. Instead, the FL» has relatively too many plants pheno- typically like the Px parents (wrinkled, no tendrils; round, tendrils) and relatively too few new recombinational types (round, no tendrils; wrinkled, tendrils). Examine also the phenotypic results ob- tained from test crossing the dihybrid in question ( + w -\-t by w w t t) : Phenotypes No. Individuals round, tendrils 516 round, tendrilless 9 wrinkled, tendrils 7 wrinkled, tendrilless 492 Independent segregation would have given a 1:1:1:1 ratio for each of the types. But again the dihybrids produced relatively excessive numbers of gametes containing the old (parental) combinations (-\ — h and w t) and relatively too few new combina- tional or recombinational types. Based on the results of both crosses, we conclude that independent segregation does not occur in this dihybrid. The very existence of re- combinational types proves — what had pre- viously been an assumption — that we are dealing with two separate pairs of genes. Let us assume now that the two pairs of nonalleles involved are located in the same pair of homologous chromosomes, a possi- bility already mentioned in Chapter 4 (p. 48). In this situation the nonalleles in the same chromosome are linked to each other. Recall that sex-linkage involves the linking of a single gene (such as the one for white eye in Drosophila) to a particular chromo- some (the X chromosome). Our concern here is with intergenic linkage, which in- volves all the nonallelic genes presumed to be located in the same chromosome. We can obtain evidence for this only by study- ing the transmission genetics for at least two traits simultaneously. Since no genetic re- combination was detected between the ge- netic material for sex and for a sex-inde- pendent trait (like eye color) in the X chro- mosome, the linkage between the two traits, sex-linkage (or, more precisely in this case, X-linkage) was complete and presented no evidence that this chromosome contained two or more separable nonalleles. Because the present experiments with peas involved two separable pairs of genes, we were able to propose the hypothesis that a chromo- some contains more than one gene. Let us reexamine the results of the two kinds of pea crosses described. In Figures 9-2 and 9-3 a horizontal line is used to represent a chromosome and to indicate the presence of one member of each pair of alleles in each chromosome. Where the genes could be either the dominant or the recessive allele, a question mark is placed in the appropriate position. Down through the genotypes of the P2 the results in Fig- ure 9-2 are consistent with the view that linkage is complete; that is, the chromo- somes carrying w t or + -f are unchange- able (except by mutation). However, the occurrence of seven recombinational indi- viduals in F2 shows that linkage is not com- plete— that these recombinants have a chro- mosome which has kept one allele and re- ceived the nonallele present in the homolog. Moreover, reciprocal types of recombinants are approximately equal in frequency, sug- gesting that a given pair of genes switched positions in the homologs; that is, they had reciprocally crossed over. For this reason, such recombinational individuals are said to carry a crossover chromosome produced by a process called crossing over. Therefore, complete linkage between genes is prevented IIS CHAPTER 9 Wrinkled, no tendrils x Round, tendrils wt + + wt F] Round, tendrils ^z^ wt P, F, Round, tendrils (self-fertilized) w t w t + + F Round, tendrils Round, no tendrils Wrinkled, tendrils Wrinkled, no tendrils mgurf. 9-2 (above). Linkage between non- allelic genes in the garden pea. figure 9-3 (below). Linkage between non- allelic genes in the garden pea. The dihybrid parent is the same as the F, in Fig. 9-2. P^ F| Round, tendrils x Wrinkled, no tendrils + + w t 319 ? ? + t 4 ? t w + 3 w? wt 123 wt TOTAL 449 w t Round, tendrils Round, no tendrils Wrinkled, tendrils Wrinkled, no tendrils O 10 w t + t 9 w t w • 7 w t w t 492 w t TOTAL 1024 In a crossing-over process that produces genetic recombinations called crossovers. What other characteristics can we estab- lish lor the crossing-over process and the crossovers it produces? Among the prog- eny obtained from backcrossing the dihy- brid ( Figure 9-3 ) . 16 received crossovers in the gametes contributed by the dihybrid. 1008 did not. Again, the reciprocal cross- over classes are about equal in frequency. So approximately one crossover was produced for each 63 noncrossovers. A simple cal- culation will show that the F_. results in Fig- ure 9-2 are consistent with this proportion. These genes can also make a dihybrid which receives one mutant (recessive) and one normal (dominant) gene from one par- ent (w+) and one normal and one mu- tant gene from the other parent (-M). When such a dihybrid is test crossed, the crossovers (w(or -j- +) and noncrossovers (w -\- or -f O also occur in the proportion 1:63. Crossing over, apparently, occurs with the same frequency whether the two mutant genes enter the dihybrid from the same parent or from different parents. Crossovers, therefore, occur in the gametes of an individual with a frequency that is constant and independent of the specific combination in which the nonalleles were received. If this is the normal behavior, it must follow that even in -| — \-/-\ — |- and w t/w t individuals, one gamete in each 64 produced is a crossover for these genes but undetected because it carries no new com- bination of nonalleles. Notice that the crossover progeny are fewer than the non- crossover progeny. This must mean that when two linked mutants enter a dihybrid in the same gamete, the mutants tend to be transmitted together to the gametes made by this dihybrid (coupling); if, on the other hand, the mutants enter the dihybrid sepa- rately, they tend to be transmitted separately to the next generation (repulsion). Linkage and Crossing Over Between Genes 110 In another species, the sweet pea, the trait purple flowers is due to a single gene ( + ) whose recessive allele (r) produces red flowers. Long pollen ( + ) is dominant to round pollen (ro). Assume two pairs of genes are involved in a cross between a pure line of purple long (-)- -+-/+ +) and red round (r ro rro). The F] produces all purple long (-| — \-/rro) and self-fertiliza- tion of the F, produces in F2 too many P, phenotypes and too few new recombina- tional types (purple round and red long) for independent segregation. Therefore, these genes must be linked. But, as before, linkage is incomplete. In this case, the crossovers obtained can be accounted for if the Pj(Fj) dihybrid forms gametes in the relative proportions 10 + -f:10rro:l + ro:l r + . This fre- quency of crossovers is obtained no matter how the genes enter the dihybrid. Notice, however, that the constant frequency (1/11) in the sweet pea differs from the frequency (1/64) observed previously in the garden pea. Consider also, the following cases: 1 . In Drosophila you recall, the mutant gene (w) for white eye is X-linked. So also is another (presumably nonallelic) mu- tant gene which produces miniature wings (m). Using pure lines, a white-eyed long- winged fly is crossed to a dull-red-eyed min- iature-winged fly. The Fi female carries two X's and is, presumably, w +/+ m- This female is then mated, and the sons are scored phenotypically. (Any male can be used as parent since it will usually trans- mit to each son a Y chromosome lacking alleles of the genes under consideration. In fact, the Y is found to lack alleles of almost all of the genes known to be present on the X except the gene for bobbed bristles, bb. Moreover, the Y contains several genes for male fertility that have no alleles on the X.) Since sons normally receive their single X from their mother, their phenotypes directly indicate which hemizygous X-linked alleles each has received. Among the sons of this mating, about one crossover type appears for every two that are noncrossovers. 2. In man. color-blindness (c) and he- mophilia type A (h) are recessive X-linked mutant genes absent on the Y chromosome. Though rare, some women have the geno- type + h/c H with one of these mutants on each X. Available data indicate that crossover (c h or + +) and noncrossover ( + h or c + ) sons occur in the approxi- mate ratio of 1:9. These examples show that when linkage between nonalleles is incomplete, the per- centage of progeny carrying crossovers is constant for a given case but can be quite different in different organisms. The possibility that the strength of link- age varies in the same organism can be tested using two mutants, b (black body color) and vg (vestigial wings), of Drosoph- ila melanogaster. A P, cross between vg -f- vg -f- (vestig- ial)1 females and + b /+ b (black) males produces all normal F, (vg + /+ b). As shown in Figure 9-4A, a test cross of the Fj female {vg + + b 9 by vg b/vg b $ ) pro- duces in Fj only 20% with recombinant chromosomes. (All Fj carry vg b from the father, their maternal chromosome 40% of the time is vg +; 40%, + b; 10%, + +; 10%, vgb.) Since these results are inde- pendent of sex, we conclude that b and vg are linked autosomally. Recall that the X- 1 inked genes m and w showed 33% cross- overs; therefore, the linkages between differ- ent pairs of nonalleles on different pairs of homologs can have different strengths. When the reciprocal cross ( vg -f-/+ b 6 1 The convention used here, and usually hereafter, is to describe the phenotypes of individuals only with respect to the appearance of mutant traits, all traits not mentioned being of the normal type. 120 CHAPTER 9 r\ vg + vg \ 9* Ho* dW °"d ?? vg + b A _+Fo x nip^ vg b vg b vg b B ? vg + _ x F. + b ' a* F2 40% 40% 10% 10% vg + vg b + b vg b + + vg b vg b vg b F2 50% 50% vg + vg b + b vg b figure 9-4. Results of reciprocal crosses involving black body color (b) and vestigial (vg) wings. by vg b/vg b $ ) is made with the Fi, 50% of offspring are vg-\-/vgb (vestigial), and 50% are + b/vg b (black) (Figure 9-4B). This cross produces no offspring with cross- overs, so that linkage is complete for these genes in the male Drosophila. (Had link- age been complete in the female also, we should not have had any evidence that vg and b are separable and, therefore, two genes instead of one. ) One finds, more- over, that in Drosophila any genes showing incomplete linkage in the female are com- pletely linked in the male; the male, there- fore, does not undergo the process of cross- ing over to produce crossovers.2 It may be noted that in animals in general, crossing over is reduced or absent in the hetero- gametic sex. For example, no crossing over occurs in the females of birds. 2 On rare occasions a special kind of "crossing over" does occur in the male Drosophila but is not of the kind that typically occurs in females. Linkage and Crossing Over Between Genes 121 What is the strength of linkage between a given gene and several nonalleles located in the same chromosome? This problem can be readily studied for certain X-linked genes in Drosophila. In Figure 9-5, one column shows genotypes of females; the other column shows the frequencies of crossover combinations as detected in their sons. The recombination frequencies given are those found between the gene for yel- low body color (y) and for each of the fol- lowing: white eye (w); crossveinless wings (cv); cut wings (ct); miniature wings (m); forked bristles (/). For example, 13 of each 100 eggs produced by the female di- hybrid for v and cv carry crossovers (-| — f- or ycv). What does this value, and the other still different linkage values, mean in terms of meiosis? So far, no commitment has been made as to where or when crossing over takes place. Since we have been concerned with complete and incomplete linkage as studied in successive generations of individuals, let us consider only crossing over that occurs in the cell line that gives rise directly to the FEMALES % CROSSOVER CHROMOSOMES AMONG SONS y + / + w 1.5 y + / + cv 13 y + / + ct 20 y + / + m 34 Y + / + I 48 figure 9-5. Crossover percentages between one gene and others linked to it. figure 9-6. The genetic consequences ex- pected after a crossing over between linked genes. gametes (the germ line), ignoring the pos- sibility that crossing over occurs in somatic (nongerminal) cells. Although crossing over might be premeiotic, meiotic, or post- meiotic in occurrence, we shall assume that all crossovers are produced during meiosis. The genetic consequences of an exchange (which we now call crossing over) between two pairs of linked genes during meiosis were discussed earlier (Chapter 4, p. 48) and the assumption was made (pp. 19-22) that a chiasma represents physical cytolog- ical evidence that a crossing over has oc- curred. These cytogenetic events are diagramed in Figure 9-6 in somewhat more detail than those originally shown (Figure 4-8). In stage I, one member of a pair of homolo- gous chromosomes (hollow bar) is carrying the recessives a and b and the other (solid bar) is carrying their normal alleles. The black dots represent centromeres. The homologs synapse and form a tetrad (each ( II \I'I ER 9 univalent is now represented b\ two sister •strands). After crossing over, the tetrad seems to appear at diplonema as depicted in stage 11. which shows a chiasma between the a and b loci (the places in a chromo- some containing the genes). Note that when the univalents are initially identical in appearance, a chiasma shows there was a physical exchange of apparently exactly equivalent segments between two nonsister strands oi a tetrad, the strands being just as long after as before the exchange. Stage 111 shows the dyads present after the first meiotic division is completed. The upper cell or nucleus contains one -f + noncross- over strand and one + b crossover strand, whereas the lower one contains the recip- rocal crossover strand a + and the non- crossover strand a b. Stage IV shows the four haploid products (cells or nuclei) pro- duced after the dyads form monads, and the second meiotic division is completed. Ac- cording to this hypothesis, if one chiasma (representing a crossing over) occurs in any position between the loci of a and b, two of the four haploid nuclei produced con- tain noncrossover parental combinations, and the other two contain crossover non- parental recombinations. Evidence that the crossovers found in gametes originate in this way is ordinarily difficult to obtain because, in females, only one of the four haploid products from each nucleus entering the meiotic divisions is usually retained as the nucleus of a func- tional gamete, the others being lost (as polar body nuclei or cells). Even when each of the four haploid products becomes or gives rise to a gamete, as in sperm or pollen formation, the four gametes — pro- duced from a cell containing a given chi- asma— mix with gametes produced from other meiotic cells which may or may not have had a similar chiasma. For these rea- sons, only one of the four meiotic products is normally observed or identified at a time. If each chiasma results from a prior cross- ing over in the four-strand stage, approxi- mate^ equal numbers o\' the two reciprocal kinds o\ crossovers would be expected, as seen in the crossover data already pre- sented. However, crossing over during the two-stranded stage 1 is also expected to pro- duce this result. The occurrence of non- crossover types, which are equally frequent and more numerous than the crossovers, can be explained if crossing over between the loci of a and b occurs less than 509? of the time at the two-strand stage or less than 100% of the time at the four-strand stage. The morphology of a chiasma, how- ever, supports the view that crossing over takes place sometimes, if not always, at the four-strand stage. Genetic evidence as to whether crossing over occurs at the two-strand or the four- strand stage might be obtained from gam- etes that retain not one but two or more strands of a tetrad. Finding a gamete that carries one strand which is a noncrossover and another homologous one which is a crossover, would support only the four- strand hypothesis. A suitable system for this test is found in Drosophila females whose two X's are not free to segregate since they are joined and have a single cen- tromere. One type of such attached-X's is V-shaped at anaphase. During meiosis this attached-X replicates once, and the four arms synapse to form a tetrad, yielding two meiotic products each of which carries at- tached-X's and two products devoid of X chromosomes. Using females whose at- tached-X's are dihybrid and scoring their female progeny, one finds attached-X's hav- ing one arm a crossover and one that is not (Figure 9-7). Though this evidence also supports the four-strand hypothesis, it does not eliminate crossing over at the two-strand stage. Linkage and Crossing Over Between Genes 123 METAPHASE I w f w -» \ % f wa f34b)| wa f 34b / — / B w f w Vi '' wa/\f34b; wa f34b^ w f wa f34bi w f \ TELOPHASE II light apricot weak forked light apricot weak forked 34b white weak forked apricot weak forked figure 9-7. Genotypic and phenotypic consequences of no crossing over {A ) and of one type of crossing over (B) between marker genes in an attached-X female of Drosophila. 124 CHAPTER 9 The red bread mold Neurospora ma) provide critical evidence as to the time of crossing over. Recall that in the sexual process, so-called "fruiting" bodies arc formed composed of cells containing two haploid nuclei, each of which was derived originall) from a different parent (Figure 9 8). Two such haploid nuclei fuse to form a diploid nucleus containing seven pairs of chromosomes, and the cell elon- gates to form a sac. The diploid nucleus immediately undergoes meiosis, as shown in the figure, so that the four haploid products are arranged in tandem; that is. the two top nuclei come from one first-division nu- cleus, the bottom two from the other first- Q b X DIPLOID NUCLEUS DIPLONEMA D AFTER FIRST DIVISION FOUR MEIOTIC PRODUCTS JO I TWO HAPLOID NUCLEI DIPLOID NUCLEUS DIPLOID NUCLEUS FIRST MEIOTIC DIVISION. SECOND MEIOTIC DIVISION MITOTIC DIVISION AND SPORE FORMATION EIGHT SPORES figure 9-9. Chiasma and crossing over in Neurospora. figure 9-8. Meiosis in Neurospora. division nucleus. Since each haploid nu- cleus subsequently divides mitotically once, each meiotic product is present in duplicate within the ascus. Each haploid ascospore can be removed from the ascus, grown in- dividually, and its genotype determined di- rectly. In this organism, then, all of the meiotic products derived from a single dip- loid nucleus can be obtained and identified. Using the symbols in Figure 9-6, let us Linkage and Crossing Over Between Genes 125 follow in Figure 9-9 the genetic conse- quences of a single crossing over between the loci under study. Since only one of the seven pairs of chromosomes present is being traced, the others were omitted from the figure. As shown, a single crossing over in the four-strand stage produces two cross- over and two noncrossover meiotic prod- ucts. On the other hand, a crossing over in the two-strand stage (in the topmost nu- cleus) would produce only crossover meio- tic products. When numerous asci of a particular di- hybrid for linked genes were tested, 90% had all eight spores noncrossovers for the two loci; in the remaining 10%, exactly four of their eight ascospores were cross- overs. In other words, never were all eight spores from a single sac crossovers. This fact demonstrates conclusively that crossing over occurs only in the four-strand stage, as depicted in Figure 9-9 and Figure 9-10. It has already been implied that chiasma formation is a normal part of meiosis (p. 16). The chiasma prevents the premature separation of dyads by holding them to- gether as a tetrad until anaphase I. (Usu- ally at least one, and as many as six chi- asmata occur per tetrad.) Therefore, the crossing over that subsequently leads to the useful chiasma is also a normal part of mei- osis. Since chiasmata are found at numerous positions along a chromosome, it seems rea- sonable to suggest that the greater the dis- tance between two loci, the greater will be the chance for a crossing over to occur be- tween them, and the greater will be the frequency of crossovers for them. Con- versely, the closer two loci are, the smaller will be the chance that crossing over oc- curs between them, and the smaller will be the frequency of crossovers for them. Ac- cording to this view, the frequency of cross- overs can be used as an indication of the relative distances between loci. (The re- sults presented in Figure 9-5 should now have additional meaning for us.) In the particular Neurospora test men- tioned, no crossing over occurred in 90% of spore sacs in the genetically marked re- gion (a-b). These sacs produced 90% of the total number of spores and carried only parental, noncrossover genotypes. From the 10% of spore sacs which did undergo TETRAD FIRST MEIOTIC DIVISION ) SECOND MEIOTIC "\ DIVISION v m + th ^ th + -•- th (-•- th 8 Spores 8 Spores + th + th MITOTIC th DIVISION th + th th th th figure 9-10. Arrangement of spores in the Neurospora ascus when segregation occurs at the first meiotic division {left) and at the second meiotic division {right), as determined by the absence and presence, respectively, of a chiasma between the segregating genes and the centromere. L26 CHAPTER 9 crossing over, half oi the spores were of the parental types and half were crossovers. So. equating the chiasma with the crossing over, a chiasma Frequency of 109? resulted in 5r't of all spores having crossovers. We can express the distance between the loci of a and /' as being five crossover units long. a crossover unit representing that distance between linked nonalleles which results in one crossover per hundred postmciotic prod- ucts (spores, in the present case). Gener- ally, when the genes are sufficiently close together (as in the present example), cross- over frequency (crossover distance) is just one half the chiasma frequency, supporting our expectation of one chiasma for each preceding crossing over. Crossover frequency can be measured in several ways in Neurospora: 1. Spores are tested from each sac (two to five per sac are sufficient) to determine whether or not the sac carries a crossover in the region under investigation. In the a-b example above, 10% of the sacs would have crossovers, 90% would not. Since each sac in the 10% group contains four spores that are crossovers and four that are not, crossover frequency would be 5%. TETRAD WITH ONE CHIASMA S MEIOTIC PRODUCTS A - B A • b .— n a • B a . '■' — Noncrossover Crossover Crossover Noncrossover figurl 9—11. Correlation between genetical and cytological crossovers. 2. All the spores from many sacs are mixed, then a random sample of spores is taken and tested. This method would also give 595 recombination with a-b and is sim- ilar to the sampling procedure involved in determining crossover frequency in animal sperm. 3. One randomly chosen spore from each sac is tested; the others are discarded. Again. 5% crossovers are obtained. This procedure resembles the situation in many females (including Drosophila and human beings) in which one random product of meiosis normally enters the egg and the others are lost. In the discussion above, no direct correla- tion was made between a genetically de- tected crossover and a cytologically detect- able event involving a particular chromo- some region. Such a connection cannot be made if both members of a pair of homol- ogous chromosomes are identical in cyto- logical appearance (as is assumed in Figure 9-6) because a crossover strand, having ex- changed a cytologically identical segment with its homolog, appears the same as a non- crossover strand. A dihybrid for linked genes can be constructed, however, in which one homolog differs cytologically from its partner on both sides of the loci being tested. Such a genetic dihybrid is also cytologically dihybrid as specified in Figure 9-11. In this case it is possible to collect noncross- over progeny and show cytologically that they invariably retain the original chromo- somal arrangement; crossovers on the other hand always show cytologically a new chro- mosome arrangement explained by a mutual exchange of specific chromosome regions between the homologs.:{ :! Using this method, genetic crossovers were cor- related exactly with cytological crossovers by C. Stern (1931) using Drosophila and by H. S. Creighton and H. McClintock (1931) using maize. Linkage and Crossing Over Between Genes 127 Maize workers (/. to r.) at Cornell University in 1929: C. R. Burnham, M. M. Rhoades, G. W. Beadle {crouched), R. A. Emerson, and Barbara McClintock. SUMMARY AND CONCLUSIONS The nonallelic genes in a given chromosome are linked and tend to be transmitted together to the next generation. Just as intergenic linkage produces an exception to independent segregation, crossing over produces an exception to linkage between non- alleles and causes linkage to be incomplete. Incomplete linkage proves that a chromo- some contains more than one gene. In any given case, the degree to which linkage is incomplete — as measured by crossover frequency — is constant and independent both of the specific alleles which are present at the two different loci and of the gene com- binations that enter the individual forming the gametes. Moreover, reciprocal cross- over types are equally frequent. The crossover frequency in different cases is found to vary considerably. A crossover chromosome is derived from a tetrad in which a crossing over between the linked genes showing recombination involves only two of the four strands. For closely linked genes, the crossover frequency is one half the frequency with which a chiasma or a crossing over occurs between their loci. It is hypothesized that crossover frequency is directly related to the distance between genes in a chromosome. One unit of crossover distance between genes is defined as one crossover per one hundred postmeiotic cells (spores or gametes). Since different genes linked to the same gene show different percentages of crossing over with this gene, they are presumably different distances from it. 128 CHAPTER 9 Curt Stern, in the early 1930's. REFERENCES Committee on Standardized Genetic Nomenclature for Mice, "A Revision of the Standardized Genetic Nomenclature for Mice," J. Heredity, 54:159-162, 1963. Creighton, H. S., and McClintock. B., "A Correlation of Cytological and Genetical Crossing-over in Zea Mays," Proc. Nat. Acad. Sci., U.S., 17:492-497, 1931. Reprinted in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, 1959, pp. 155-160; also in Great Experiments in Biology, Gabriel, M. L., and Fogel, S. (Eds.), Englewood Cliffs, N.J.: Prentice-Hall, 1955, pp. 267-272. Morgan, T. H.. "Random Segregation Versus Coupling in Mendelian Inheritance," Science, 34:384, 1911. Reprinted in Great Experiments in Biology, Gabriel, M. L., and Fogel, S. (Eds.), Englewood Cliffs, N.J.: Prentice-Hall, 1955, pp. 257-259. Stern, C, "Zytologisch-genetische Untersuchungen als Beweise fur die Morgansche Theorie des Faktorenaustauschs," Biol. Zbl., 51:547-587, 1931. QUESTIONS FOR DISCUSSION 9.1. Distinguish between sex-linkage and the linkage of nonalleles. 9.2. Does the linkage of two genes prove they are located on the same chromosome? Explain. 9.3. Discuss the advantages and disadvantages of linkage and crossing over with respect to the fitness of individuals carrying certain genotypes. 9.4. In Drosophila, y and spl are X-linked. A female genotypically + +/y spl pro- duces sons. If 3% carry either y + or + spl, what are the genotypes and rela- tive frequencies of gametes produced by the mother? Is the father's genotype important? Explain. Linkage and Crossing Over Between Genes 129 9.5. Name all the processes so far discussed which lead to genetic recombination. 9.6. Do you think that one of the main principles in this chapter is that chromosomes contain more than one gene? Explain. 9.7. In light of your present knowledge how would you proceed to state a "law of independent segregation"? 9.8. What evidence do you have that crossing over does not involve the unilateral movement of one gene from its position in one chromosome to a position in the homologous chromosme? 9.9. Does crossing over always result in genetic recombination? Explain. 9.10. In what respect do you think the development of the principles of genetics in this text would have been affected had the first two pairs of genes (Rr and Yy), simultaneously studied in crosses, been linked? 9.11. Assume that the gene for woolly hair (Chapter 3, p. 38) is located autosomally. A nonwoolly-haired, nonhemophilic man marries a woolly-haired, nonhemophilic woman. They have a woolly-haired, hemophilic son. Give the genotypes of all three individuals, and the genotypes and frequencies of the gametes usually pro- duced by the son. 9.12. How would you defend the conclusion that the point of crossing over is located at exactly equivalent positions on the two homologs? 9.13. What are the relative frequencies of the phenofypes and genotypes expected from a mating between two Drosophila: dihybrid vg+/+b? 9.14. Two dihybrids, a +/+ b, for autosomally linked mutants a and b in Drosophila, are crossed. If 2p equals the frequency of noncrossover eggs and 1— 2p equals the frequency of crossover eggs, and if p < 0.5, give the relative frequencies of the phenotypes expected among the F, of this cross. 9.15. What result would you expect from 9.14 if the cross were between dihybrids a b/+ +? 9.16. A wild-type Drosophila female, whose father had crossveinless wings and mother had yellow body color, is mated to a yellow male. Give the relative frequencies of genotypes and phenotypes expected in the Fv 9.17. A mating in Drosophila produces the results shown below. Give the genotypes of the parents, and determine which genes are linked and which are not. Sons Daughters 75 wild-type 92 wild-type 70 yellow body color, white-eyes 75 white 21 yellow, white, vestigial wings 28 vestigial 27 vestigial 20 white, vestigial 2 yellow 1 white, vestigial 9.18. What is the relationship in Neurospora between crossing over and first and second meiotic division segregation? 9.19. In tomatoes, the gene for tall ( + ) is dominant to short (s), and the gene for smooth epidermis ( + ) is dominant to rough (/-). A cross between two plants produces 208 tall smooth, 9 tall rough, 6 short smooth, 195 short rough. Give the genotypes of the parents. 9.20. What is the percentage of crossovers for two loci, in a species in which both sexes undergo crossing over with equal frequency, if a mating between identical dihybrids (Ab/aB) gives four equally viable classes of offspring, the smallest class comprising 1% of all offspring? 130 < HAPTER 9 9.21. How would you prove genetically that the last division in a spore sac of Ncuro- spora is a mitotic one.' 9.22. In the absence ot crossing over, could you determine whether the alternatives for two different traits were due to a single pair of genes or to two pairs of linked genes? Explain. 9.23. Draw an attached-X chromosome of Drosophila heterozygous both tor y and for in. Show the kinds of gametes which could be obtained after: (a) No chiasma (b) One chiasma between the nonallelic genes (c) One chiasma not between the genes mentioned 9.24. Suppose one member of a long pair of chromosomes in a plant has a large knob at one of its ends, and the other has a small knob at the opposite end. Suppose, moreover, that there is also a shorter pair of homologs. one member terminating with a large knob, and the other at the opposite end with a small knob. What combinations and configurations would you expect to readily find in the gametes of this individual? 9.25. What reasons can you present for believing that germ-line crossing over is based neither upon premeiotic nor upon postmeiotic events? 9.26. Calculate the number of crossover units between black body (b) and dumpy wings (dp) in the following Drosophila crosses: (a) P! pure black X pure dumpy P_, Fj 9 9 X black dumpy $ $ F-, wild-type 272 " black' 774 dumpy 801 black dumpy 239 (b) Pj black dumpy X pure wild-type P2 Fj 9 9 X black dumpy $ $ F., wild-type 360 black 103 dumpy 97 black dumpy 314 9.27. What phenotypic results would you expect in 9.26(a) and (b) if the reciprocal mating had occurred in P_,? 9.28. Test statistically the F.> results of 9.26(a) with those expected from independent segregation. 9.29. Test statistically whether the FL, results in 9.26(a) and (b) differ significantly. 9.30. A trihybrid Aa Bb Cc is test crossed to aa bb cc. The Fj show that the trihybrid produced the following gametes: 29 ABC 21 abc 235 ABc 215 abC 210 Abc 239 aBC 27 AbC 23 aBc (a) Which loci are linked and which are segregating independently? (b) Write the genotypes of both parents in view of your answer to (a). (c) Give the percentage of crossovers wherever applicable. Chapter 10 GENE ARRANGEMENT; CROSSOVER MAPS I n the preceding chapter, the fre- quency of crossing over was presumed to be dependent upon the distance between genes, the interval being measured in crossover units. Differ- ent genes linked to a given gene were found to give different, essentially constant, cross- over frequencies or crossover distances. Let us now investigate how these different genes are arranged spatially. Crossover distances can be used to study whether linked genes are arranged in some regular three-dimensional configuration such as a sphere, cube, prism, or some two- dimensional one such as a line, circle, or triangle. To map the genes on the basis of crossover data, that is. make a crossover {or linkage) map, it is necessary to determine all the crossover distances for a minimum of three linked loci, since two points (such as those defined by the crossover distance between two genes) are not enough to de- termine a specific geometrical arrangement. Gene Arrangement The arrangement of linked loci can be in- vestigated with Drosophila. Using the three X-linked genes, y (yellow body color), \v (white eyes), and spl (split bristles), di- hybrid females of the following types are obtained: yw/-\ — (-; y spl -\ — |-; and wspl/-\ — |-, and each type is test crossed with the appropriate double recessive male. The corresponding crossover distances are: 131 y to w, 1.5; >• to spl, 3.0; and u> to spl, 1.5. Since the crossover distance between y and spl equals the sum of the crossover distances from y to w and from w to spl, a linear ar- rangement for these three genes is described, namely, y w spl or spl w y. In other words, the genetic map based on crossovers is linear. If the reasonable assumption is made that crossing over is a function of physical dis- tance between genes, the genes are also linearly arranged in the chromosomes. When the positions of a fourth X-linked gene and all other X-linked genes are mapped relative to the three studied above, all are found to be arranged in a linear order (Figure 10-1, and page s-16). In such a crossover map, y is arbitrarily assigned the position, or locus, zero. On a standard crossover map for the Dro- sophila X, the genes y, w, spl, cv. ct, m, and / line up respectively at positions 0, 1.5, 3.0, 13.7, 20, 36.1, 56.7, and one can see that ct and spl are 17 map units apart (20 — 3). Since one crossover map unit equals one crossover per hundred gametes, the dihybrid for spl and ct (Figure 10-2) should produce 17% crossovers (8.5% + + and 8.5% splct). However, such a result is obtained only under special condi- tions. The crossover frequency actually observed will depend upon several factors. One of these is the number of individuals making up the sample. In small samples it is very likely that, by chance, the observed values will deviate considerably in both directions from the standard map distance. As the size of the sample increases, the observed value will more closely approach the stand- ard one. Standard distances, therefore, are determined only after large numbers of prog- eny have been scored. The relative viability (see p. 69) of dif- ferent phenotypic classes is another factor influencing observed crossover frequency. 132 CHAPTER 10 Xorl centromere 0.0 0.1 Hw svr S^. \IHnl llw .SIT pn w spl ec hi rh cv rux cm ct sn 0.8 3.0 6.9 7.5 15.0 18.7 pn spl bi rb rux cm 10 unit i of Map Distance 21 0 27.7 32.8 38.7 40.7 59.5 sn Iz ras fw wy fu I II I I I 1 1 I I I figurh 10—1. Crossover map of the X chromosome of D. melanogaster. Name yellow body color Hairy-wing — extra bristles on wing veins, head, and thorax scute — absence of certain bristles, espe- cially scutellars silver body color prune eye color white compound eyes and ocelli split bristles echinus- — large and rough textured eyes bifid — proximal fusion of longitudinal wing veins ruby eye color crossveinless — crossveins of wings absent roughex — eyes small and rough carmine eye color cut — scalloped wing edges singed — bristles and hairs curled and twisted Kii ro Symbols Symboi OC I Iz ra.s v m fw wy s 8 sd f B fu car bb Name ocelliless — ocelli absent; female sterile tan body color lozenge — eyes narrow and glossy raspberry eye color vermilion eye color miniature wings furrowed eyes wavy wings sable body color garnet eye color scalloped wing margins forked — bristles curled and twisted Bar — narrow eyes fused longitudinal wing veins; female sterile carnation eye color bobbed — short bristles The phenotypic expression of a + allele is usually more viable than that of its mutant forms. For example, in Figure 10-2 the phenotypically split, cut sons are not as viable as the normal (wild-type) sons; al- though both types are equally frequent as zygotes, the former fail to complete their development more often than the latter and, therefore, are relatively less frequent when the adults are scored. Zygotes destined to become either split or cut males are also less viable than zygotes destined to produce wild-type males. Whenever phenotypes are scored after some long developmental pe- riod, much of the error due to differential viability may be avoided by providing op- timal culture conditions. Another way to avoid most of this kind of error is to delay the scoring of crossovers for one generation. The cross is arranged in such a way that individuals with the chromosomes to be scored have a homologous chromosome con- taining the normal alleles of all genes under crossover test. Since the progeny of such a cross are phenotypically normal, their viability will be approximately the same, and they can be scored for chromosome type from the offspring each produces when individually test crossed. For example, the female in Figure 10-2 is crossed with wild- type males, and the Fi daughters (all pheno- typically normal) are individually mated to any male. Daughters carrying an X of one of the following types: spl +, + ct, -\ — \-, spl ct — in addition to a H — \- homolog ob- tained from the father — will produce sons of Gene Arrangement; Crossover Maps 133 the following type respectively: some split, but none cut; some cut, but none split; all normal; some both split and cut. In this way the generation being tested for cross- over frequency is largely protected from differential viability, its genotypes being scored in the next generation. For map- ping and other purposes, the extra labor entailed by this method is often justified. Variability in crossover frequency may be due also to factors — such as temperature, nutrition, age of the female, and presence of specific genes — which influence the very process of crossing over. To better understand the relationships be- tween crossover maps and chiasmata, con- sider the properties of an over-simplified model (Figure 10-3). Assume that a chro- mosome (ignoring the centromere) is com- posed of five equally long regions, the ends of each marked by a gene; that each tetrad of this type contains one, and only one, chiasma following a crossing over; and that this chiasma can occur in a random posi- tion among these segments. For the hexa- hybrid shown in Figure 10-3, the chance the chiasma will occur in the a-b region is 20%; out of each 25 tetrads (producing 100 haploid meiotic products), five or 20% will have the chiasma in the a-b region. These five will produce 10 crossover and 10 noncrossover strands. Adding the latter 10 and the 80 noncrossover strands from the remaining 20 tetrads, gives 90 noncrossover strands. For this region, therefore, 20% of tetrads have a chiasma and 10% of haploid meiotic products are crossovers as explained in Chapter 9. Similarly, in the b-c region 10% crossovers would be noted. The chiasma would occur in the a-c region 40% of the time, and 20% of all haploid meiotic figure 10-3. Crossover consequences of a single chiasma. pi *+l? 9 x ANY O* F Sons: spl + / Y 41.5 % + ct / Y 41.5 % + + / Y 8.5 % spl ct / Y 8.5 % figure 10-2. Crossover frequency for two X- linked genes in Drosophila. MEIOTIC 40 PRODUCTS 40 45 45 PERCENT OF ALL PRODUCTS 134 CHAPTER 10 I FIG1 R] 10-4. Chromatid recombinations pos- sible in a double chiasmata. {See text for details. ) products would be crossovers relative to the markers a and c. Since the sum of the dis- tances from a-b and b-c equals the distance between a and c measured directly, the genes of the model would be aligned linearly, just as was observed in the experiment described earlier in this chapter. In the model proposed above, the pres- ence of a chiasma in one region automat- ically excludes it from being in some other region. Consequently, the chance that a chiasma will be found in the a-c region is equal to the sum of the separate chances that a chiasma will be found in the a-b and b-c regions. It is a general rule that the overall probability jor the occurrence of any one of a series of mutually exclusive events is equal to the sum of their separate probabili- ties of occurrence. Therefore, the chance of a chiasma occurring between a and / is (20 + 20 + 20 + 20 + 20)%, or 100%. As a result, 100% of the tetrads have one crossing over (that is, one chiasma) which produces 50% crossovers, and the model chromosome has 50 map units. The oversimplification of this model can be appreciated by remembering that a given tetrad usually contains more than one chi- asma. This prompts us to ask: When a tetrad contains two or more chiasmata, which strands are involved in the exchanges? To answer this question, let us specify the strands in a tetrad as 1,2, 3, 4, where 1 and 2 are the sister strands carrying the normal alleles and 3 and 4 the sister strands carry- ing the recessive alleles (Figure 10-4). If one chiasma involves an exchange between nonsister strands 2 and 3 in the a-b region, a second chiasma, involving nonsister strands in the b-c region, can result from any one of four exchanges: 2 with 3; 2 with 4; 1 with 3; 1 with 4. The positions of these chiasmata are indicated in Figure 10-4. The four types of single chiasma in the b-c region together with the single chiasma in the a-b region form double chiasmata of three types: 2-strand (the same two strands exchange in both chiasmata); 3 -strand (one of the two strands of the first chiasma exchanges in the second, there being two ways this double chiasmata can occur); and 4-strand (those strands which do not exchange in the first chiasma, exchange in the second). Let us examine the genetic consequences of these four nonsister types of double chi- asmata (shown separately at the left of Fig- ure 10-5). The middle column shows the meiotic products of each, and the right column indicates whether these products are noncrossovers, single crossovers, or double crossovers for the a-b-c region. From 2- strand double chiasmata, two of the four meiotic products are genetic noncrossovers (+ + + and a be), and two are double crossovers (+ b + and a + c). The dou- ble crossovers, or "doubles" as they are called, are characterized by a change in the position of the middle gene relative to the end genes. A 3-strand double chiasmata produces one double crossover, two single crossovers (in each, the position of one end Gene Arrangement; Crossover Maps 135 gene is changed relative to the other two genes), and one noncrossover. The 4- strand double chiasmata yields four single crossover strands. Note that each type of double chiasmata produces some strands with a new genetic combination, that is, crossover strands; each of the three differ- ent types also produces a characteristic pat- tern of noncrossover and crossover types. Moreover, the genetic products obtained from each type of double chiasmata differ from those obtained from a single chiasma (which produces two noncrossovers and two "singles"). In view of the preceding discussion, it should be possible to learn, from the geno- CHIASMATA MEIOTIC PRODUCTS CROSSOVERS 2-STRAND 2 Doubles £ J a / x _ b _\/_ _ c_ _ »\ 2 b_/ \ __»■_. 3-STRAND c 1 Double c 2 Singles + 1 Noncrossover Ia_ _b_ + b + + 1 Double a + 2 Singles a b c 1 Noncrossover 3-STRAND FIGURE 10-5. Double nonsister chiasmata types and their genetic consequences. + + y b !\ 4 Singles 4-STRAND i:m ( II \ I'M K Id types of the meiotic products, the relative frequency with which the tour types oi double chiasmata occur. It all four types occur with equal frequency, the strands forming one chiasma would he unaffected by those which form an adjacent chiasma. Indeed, experiments with Neurospora reveal that all four types c\o occur — in some, the four types occur with equal frequency. For our purposes, we can accept the view that there is usuallj no chromatid interference in chiasma formation; in other words, the particular nonsister chromatids forming a chiasma arc not influenced by nonsister strands which may or may not form a chi- asma in an adjacent region. Thus, nonsister strands crossing over in two different re- gions of the same tetrad are independent. Does the occurrence of one chiasma in- fluence the probability that a second chiasma will occur in the same tetrad, even though when both chiasmata occur there is no chro- matid interference? Suppose that in the genetic system of Fig- ure 10-4, each of the two regions under observation has a 20% chance of forming a single chiasma. If the occurrence of a chiasma in the a-b region is independent of a chiasma in the b-c region, then, of all tetrads, 20% of the 20% with an a-b chi- asma will simultaneously have a b-c chiasma; that is, 4% will contain double chiasmata. (According to the previous discussion, this 4% will be composed of the four nonsister types in equal frequency.) It is a general rule that the overall probability for the si- multaneous or consecutive occurrence of two or more events of independent origin is equal to the product of their separate probabilities of occurrence. If 4% double chiasmata were actually ob- served in the a-c region, one would conclude there was no chiasma interference (or better still, no crossing-over interference); that is, the formation of one chiasma would not affect the formation o\' another in an adja- cent region. It. on the other hand, only 29? double chiasmata were observed, this would mean that some chiasma interference had occurred. The degree o\' chiasma interference can be written as double chiasmata observed 0.02 double chiasmata expected 0.04 = 0.50 This fraction, called the coefficient of coin- cidence, expresses the frequency with which the coincidence of two chiasmata is actually obtained. Consequently, a coefficient of co- incidence equal to zero would mean that one chiasma completely prevented the other from occurring; whereas a value of one would mean that the one chiasma had no effect at all on the occurrence of the other. In practice, however, because of the errors involved — particularly those stemming from chiasma terminalization (see p. 22) — one does not usually determine the frequencies and positions of double chiasmata cytolog- ically. Can we use the frequency of ge- netically-detected double crossovers as an alternative for measuring chiasma or cross- ing-over interference? We can be sure that each double cross- over observed has resulted from multiple crossing over or chiasmata. The expected frequency of double crossovers in the a-b-c region of the example can be calculated in the following way: since each region {a-b and b-c) has a 0.2 chance for one crossing over, the chance for a double crossing over is 0.2 times 0.2, or 0.04. (If, as before, the coefficient of coincidence were 0.5, one would expect 0.02 tetrads to have double crossing over.) Recall that the double crossing over can occur in four ways and can involve 2, 3, or 4 strands of a tetrad. If these alternatives occur with equal fre- quency, only one quarter ( % 6 ) of all mei- otic products from double crossing over will Gene Arrangement; Crossover Maps 137 appear as double crossovers (Figure 10-5). Since the remaining three quarters (1tig) of the meiotic products are noncrossovers or single crossovers, they are not useful in identifying the occurrence of double cross- ing-over events, because they could have been produced in tetrads of other types, for example, those having single or no crossing over. Accordingly, a frequency of 0.04 double crossing over would lead us to ex- pect a frequency of .01 double crossovers; and a frequency of only 0.005 would actually be detected were the coefficient of coinci- dence 0.5. In this way, the coefficient of coincidence can be determined from double crossovers observed divided by the double crossovers expected. There is another, perhaps simpler, way to calculate the expected frequency of double crossovers. In our example above, the chance a crossing over will occur is 0.2, and the chance that a given strand will be a crossover, 0.5. The chance that both will occur once is 0.1, and that both will occur twice is 0.1 times 0.1 or 0.01. That is, the expected chance that a given strand will be a double crossover is one percent. Accord- ingly, the frequency of observed single cross- overs in the a-b region multiplied by the frequency of observed single crossovers in adjacent b-c region equals the expected fre- quency of double crossovers (one in each region). In practice, therefore, one may readily determine the coefficient of coinci- dence from double crossovers. Generally the coefficient of coincidence is negligible — equal to zero for all practical purposes — for short map distances and be- comes larger with increased distance. This relation suggests that a tetrad in which one crossing over occurs is somehow precluded from having a second one occur close by, with this restriction diminishing as the dis- tance to the second region increases. In Drosophila, for example, the coefficient of coincidence is zero for distances up to 10-15 map units and, consequently, no double chi- asmata (or no double crossovers) occur within such distances. As the distance in- creases beyond \5 map units, however, the coefficient gradually increases to 1, at which point nothing interferes with the formation of double chiasmata. In two equal-armed chromosomes there does not seem to be chiasma interference across the centromere. If each tetrad has only a single chiasma, the maximum frequency with which the end genes recombine relative to each other is 0.5. What happens to the frequency of recom- bination for the end genes when the chro- mosome has double chiasmata? If each tetrad has two chiasmata, one might think that the end genes would form new combinations with a frequency greater than 0.5. Examination of Figure 10-5 re- veals, however (each type of double chias- mata being equally probable), that on the average eight products (single crossovers) will carry a new combination with respect to one end gene, and eight products will not. Of the latter, four will be noncrossovers and four, double crossovers in which the middle gene has changed position relative to the end genes. Therefore, even if every tetrad has double chiasmata, the maximum fre- quency of recombination for the end genes is 0.5. When four loci are studied and three chi- asmata occur in each tetrad — one in each region — one finds that for every 64 meiotic products, 32 are recombinational for the end genes and 32 are not. For cases where four or more chiasmata lie between end genes, the frequency of meiotic products bearing odd numbers of crossover regions is easily calculated to be 0.5. In each of these cases the gene at one end is shifted relative to that at the other. However, the remaining strands contain either even num- bers of crossover regions (which do not 138 CHAPTER 10 + + + 0.31 a b c 0.31 + b c 0.14 a + + 0.14 + + c 0.01 a b + 0.01 + b + 0.04 a + c 0.04 + + + 0.31 a c b 0.31 + c b 0.14 a + + 0.14 + c + 0.01 a + b 0.01 + + b 0.04 a c + 0.04 1.00 1.00 figure 10-6. Determination of gene order from a test crossed trihybrid. cause the genes at the two ends to shift relative to each other) or are noncrossovers. Accordingly, the maximum frequency of re- combination of 0.5 holds for the endmost genes (and, therefore, of course, for any genes between them). If two genes in a chromosome are suffi- ciently far apart, the frequency with which they undergo recombination will be near 0.5. Since a recombination frequency of 0.5 means that nonalleles are independent in their segregation, one cannot conclude from such a recombination frequency that non- alleles are on the same chromosome. Ac- cordingly, two pairs of genes that show re- combination frequencies near 0.5 can be either far apart in the same pair of homologs or located in different pairs of homologs. However, if two nonalleles segregate inde- pendently but are both linked to a third nonallele, all three are linked to one an- other. Whenever the number of gene pairs in- vestigated is considerably larger than the Dumber of chromosome pairs, the number of groups of linked genes equals or ap- proaches the number of chromosome pairs. The result is a limitation in the number of linkage groups, the maximum number equalling the haploid chromosome number. (Examination of the linkage groups of the garden pea now reveals that two of the first seven gene pairs studied ' arc in the same linkage group although a considerable dis- tance apart. The initial recombination data were sufficiently meager for acceptance of the hypothesis that the genes were segregat- ing independently.) The sequence of three linked genes can be determined from the results of a single cross. Suppose the trihybrid -+- -+- + /a b c is test crossed, and the frequencies of the various phenotypes in the progeny are those shown at the left in Figure 10-6. These values, we remember, represent the fre- quencies of the corresponding genotypes in the gametes of the trihybrid. The middle gene in the actual sequence is the one which switches least often from the original gene combinations (+ -\ — h and a be), because only the middle one requires two chiasmata for its switch. Consequently, this gene is identified as c, and the actual gene order is acb (or be a). This reasoning may be easier to follow if the data are examined with the genes listed in their correct order, as shown at the right in Figure 10-6. The frequency of observed crossovers be- tween the a and c loci is 0.30; between c and b it is 0.10. Between a and b the fre- quency of single crossovers is 0.36. Cross- over frequency between a and b, however, also includes double crossovers. Since each double crossover represents two single cross- overs between the end genes, the frequency 1 By G. Mendel. Gene Arrangement; Crossover Maps 139 of double crossovers, .02, must be doubled and added to the frequency of single cross- overs to obtain the total crossover frequency between a and b. The genetic map based on crossover frequency becomes linear (0.30 + 0.10 = 0.36 + 0.04), therefore, when double crossovers are taken into ac- count. The expected frequency of double crossovers is 0.3 times 0.1 or 0.03, so that the coefficient of coincidence in this case is 0.02 0.03, or 0.66. ( In the y w spl example discussed earlier in this chapter, the longest region, y-spl, was too short for double cross- ing over.) Would it be satisfactory to use the data in Figure 10-6 to construct a standard link- age map for the distances between these genes, assuming that large numbers of prog- eny had been scored and standard experi- mental conditions had been used? For this purpose, the observed distance from c to b is acceptable since only a single chiasma can occur in such a short interval. The situa- tion is otherwise for the a-c region, however, which is 30 map units long in the present experiment. Double chiasmata are expected to occur under these circumstances, yet the absence of genetic markers between a and c prevents their identification. Therefore, the standard map distance for a-c must be longer than 30 map units (and a-b longer than 40) . Note that the identical error foreshortens the a-c and the a-b distances; therefore, for the distances observed, {a-c) plus (c-b) is equal to (a-b). Whether or not the chro- mosome is genetically marked so that all multiple crossover strands are detected, the correct order of three linked genes can al- ways be determined, provided that two are not 50 map units away from the third. It should now be clear why the crossover frequencies observed for large distances are less than the standard map distances, and why the standard map distances are always obtained by the summation of the short dis- tances in which only a single chiasma can occur. Although end genes can show at most 50% recombination, the length of the cross- over map may exceed 50 units. For exam- ple, if a given pair of homologs contains an average of two chiasmata in each tetrad (see Figure 1 0-5 ) , a total of 1 00 crossovers will occur among 100 meiotic products, and the map length will be 100 units even though the end genes will have recombined 50% of the time. In fact, it can be predicted that the length of the standard map is equal to fifty times the mean number of crossing-over events (or chiasmata) per tetrad. Crossover Maps Utilizing crossover frequency, genetic maps cf chromosomes have been made for a num- ber of multicellular organisms. Figures 10-7 through 10-10 give the linkage maps for a considerable number of genes in man, mouse, maize, and Neurospora. 29 G6PD Deutan Hemophilia A ► e5->|«— 12 • I 3 38- -41 V figure 10-7. Tentative linkage map of a segment of the human X chromosome. The numbers given are the values for the map distance found in five separate studies. The loci mapped are the Xg {blood group) locus, the G6PD (glucose-6-phosphate dehydrogen- ase) deficiency locus, the deutan (green) color- blindness locus, the classic hemophilia locus. (Courtesy of V. A. McKusick. From Human Genetics, 1964, Prentice-Hall, Inc., Englewood Cliffs, N.J.) I III H CD II I I _ m ( II \l> I I l< 10 <0 <^ Q ft £° « £? ■H E £5^ Ml 11 2 M •H M I II I * - T +— f ro in m c«j tt» W »H 00 ZO -c^t S N in. i mil _n lO lO CJ— a. (/) N ■& B* £jE* +-H *£ 4-^ & 2 " r- U g I io ixSo* £• io. 6- -8 2. I I n II I I I I ■ I ■ III Gene Arrangement; Crossover Maps 141 £- -* z. j 0 V 3 5, '— B 'A DC y z: J : &TJ - - > 0 CS. k ■r to co c.3 5 8. E & I "3 £ 1 £ £ 1 8 &j- 8.3 8,1 ? S I -S 1 o s-. *.— ^ >-> *k **■ ^- -* 2" ^■w'^v k x x = = s E a 2 3 » £ « I J2 g 1 >* a 3 « s >, .s 5 - § J5 *g w i J- _ — 60 <" |ll|l| ^l|llll||||l||||slf|1-Sgp||||«1 V3< < O 0 O ft. ftft.ft,ft,ftft.ft.ft.ft.*.,GSi;^T;.~C^£c/5£^2)^;^::;.o^; 3. c — Tt "3 4) 4> >>>>>>>> & « e3 331313 .. « S £ E. c o 3 3 3 3 c 3 ^ u _ H. !3 a & c a. c u •>, _e as*' 3 ^3 60 o <« <" '3 OOOO u 3 cb t> f2 ^ "S o I •- •§ I 1 3 4> 3 ft .-6 E o M 60 c3 C E 2§-?s» ^Rl •? >> P- 3 i >>3^ 2 13 > u :2^ 3 Q.— b « « 2 |I^|lss||.g ||sa|EtS llllllfl* li^lllll 5 « 5rC7^ .3 30l>C; Oi-60-3JoFr:— _5.3w3-zC-j3^. ~3uOne-raj Cr>"ca<; JC U o O O CJ) -O M- E OK 0 O •— "^ CN (N n " -H- 00 OJ ^3 O *> c 10 II M( |« O 'f 00 cm ^ 4 OM cn cn co -*r Tf D -O-Q > JJ 4-44^ O O K O cm -H co ^ "f 10 O — O o O i 111(1 — HH- >o -O |\ (s. 00 o »o o o 5 * OS S « ^ S ** v; 8 .R < ? J SB I I S3 CO 2 "o -? CM 44 en _* E ~ o a: o> o. J? c H 1 — Oil II n ", 3 01 e r - I 1 )!(( — H hK 1 h4 >o >o K CO o r— (N CS (N n iO o o K rs 00 H Gene Arrangement; Crossover Maps 143 « «» 8 S- S S S* S U e o E x: 9- o c « re A.^!c«-^ E «) Hi ., - a i u h. ~ C Jr w o — *2 r^ o - yi CD * — CO *ri S s. s s. - c/5 X c s re ■— s. IE re 13 >> SUQ N ^ ^ ^ ;-; X -c N 0 o £ ■* Cloi<4_ -tj O C n m " p •= , > ~ 2 » * S © &«><*- T3 O c-OcjCC "E !->*"* « u j-5 o a • — Co * « --3 * g. Si, o nftn = flEj3g &« e o c S •£ - c. — u s 3 re c g >. E 5 re £ x: — « S§ § o ca. ^H: I §. o. eifti gfcfftjlj O c c « ,.;•- co •= *-. 5=. 6c ?r C3 "t> Sd -5 :=T -S -~. -« -£? O c« O <*> " o <•- H c « «; =«^2 g'c a° c- '^ u a re re re s th 26 ag+ ag+ + th +th 23 + th +th ag+ ag+ Discuss the positions of these loci with respect to each other and their centro- mere(s). How would you determine the genotypes present in these spores? 10.21. Map mutant a relative to its centromere when Neurospora heterozygous for this mutant produces asci having the following spore orders: Spores asci 1 +2 3 + 4 5 + 6 7 + 8 44 a a + + 48 + + a a 2 a + a + 3 a + 1 a 2 + a + a 1 + a a 4- 148 CHAPTER 10 10.22. Suppose asci from a given Neurospora cross had spores in the following relative order: Spores % 1 I + 2 3 + 4 5 + 6 7 ! 8 asci or i 7 + 8 5 + 6 3 + 4 1 +2 92 xy XV + + + + 2 xy ++ xy + + 2 xy ++ ++ xy 1 xy x+ +y ++ 1 xy x+ ++ +y 1 x+ xy +y ++ 1 X+ xy ++ +v (a) Are x and y linked? (h) If not linked, give the crossover distance from its centromere for each, [f linked, construct a crossover map for x and y and their centromere. (c) Are any double crossovers involved? Why? 10.23. The accompanying photograph (courtesy of R. G. Isaacson) shows asci of the fungus Sordaria fimicola. The asci are in various stages of maturity, the most mature containing dark ascospores. What genetic conclusions can you draw knowing that all the asci shown are products of the same parental genotype? Chapter 11 CHANGES INVOLVING UNBROKEN CHROMOSOMES E xcept for Chapters 6 and 8 the preceding chapters sought to determine the characteris- tics of the genetic material through the oper- ation of genetic recombination. This opera- tion revealed the existence of different re- combinational units of the genetic material, which in order of size, include the genome, the chromosome, and the genes in a chro- mosome— the smallest recombinational unit being the recombinational gene. This chapter begins a study of the genetic material through the operation of mutation. We shall be especially interested in learn- ing the extent to which the genetic material can be divided into mutational units, always remembering that the recombinational and mutational units may or may not be mate- rially identical. We have been able to learn the recom- binational properties of the gene only be- cause it exerts a detectable phenotypic effect, and because it exists in an alternative state. One can readily see that if a gene were pres- ent in the same form in all organisms, it would not be detectable, since all individuals would have the same genotype and, there- fore, the same range of phenotypic expres- sion. In other words, the genes detected thus far in this book were only those that occur either in different numbers in different individuals, or have an alternate allele, or both, provided that such a genetic difference produced a detectable phenotypic change. 149 A great deal of genetic variation of this kind exists among living organisms. We have seen that some of the phenotypic varia- tion attributed to genes is actually due to sexuality which by segregation, independent segregation, crossing over, and fertilization produces new combinations of already-pres- ent genes. These mechanisms of recombina- tion shuffle the genes, just as shuffling a deck of playing cards produces the great variety of card combinations. Detecting Mutations We would like to learn two things concern- ing genetic differences; namely, what they are, and how they are produced. To do this we must first have some way to distinguish between a mutant — a really new genetic form produced by the process of mutation — and a recombinant for already-existing genes. We can use an example in Drosophila to illustrate how this distinction may be made. Suppose (as was the case at one time) none of the flies in laboratory strains has an ap- pendage on the anterior-dorsal part of the thorax. Then a single fly occurs with an appendage in this region (Figure 11-1) and, when crossed with the wild type of a dif- ferent strain (outcrossed), this trait appears in approximately one half of the progeny. How is the new phenotypic variant (He.x- aptera) to be explained? If the culture conditions had not changed. Hexaptera could not be due to environmental factors alone. Could Hexaptera result from a new combination of already-existing ge- netic units? It could not be due to the inter- action between two particular alleles al- ready present in the population which hap- pened to combine in the same zygote at fertilization, for such a combination would have to be rare and, following segregation, this phenotype would not be expected to appear in any appreciable number of the progeny of the outcross. Moreover, it could 150 ( II M'TI R 1 1 not be due to the rare combination of two already-existing unlinked nonalleles since, at most, onlj one quarter of the progenj would have the novel phenotype. Consequently, neither segregation nor independent segrega- tion could be associated with ttie appearance or nonappearance o\ Hexaptera. The new phenotype might have appeared alter a rare crossing over between two \er\ close loci brought two previously separated nonalleles into the same chromosome. Once produced, this new combination of linked genes would almost always remain intact and be trans- mitted to one half of the progeny. How- ex er. suppose also that the parents' chromo- somes were suitably marked witli genes, and it was found that the chromosome region. essential for the production of the new phe- I IGl Rl 11-1. melanogaster. 1949.) The Hexaptera phenotype in D. (from Genetics, vol. 34, p. 13, notype, was o\' a noncrossovcr type. In such a case, crossing over would not explain the results either. The only reasonable remaining explana- tion would be that a novel change, a muta- tion, occurred in the genetic material. We sec. therefore, that when the mutant pro- duces a dominant phenotypic effect it is not too difficult to determine whether a novel phenotype is due to mutation rather than to genetic recombination. In the case of a dominant mutant, only one parent needs to have a specific genotype to produce a dominant mutant trait in the progeny; no particular genetic recombina- tion is a prerequisite. In other cases, the novel phenotype appears in the progeny only when both parents have specific genotypes, and genetic recombination is required for its appearance. Note that the detection of a completely recessive autosomal mutant gene is postponed for the number of genera- tions required for two heterozygotes to mate and produce a mutant homozygote. Before a recessive mutant becomes homozygous, many generations may elapse, during which time the mutant allele may become relatively widespread in the population in heterozygous condition. When the genotype of the pop- ulation is uniform or is known, it may be possible to trace back to the origin of a new recessive mutant. If, however, the popula- tion genotype is not known, it is impossible to determine when a recessive mutant first arose, and it may be considered — correctly or not — one of the genes normally present in the population. Obviously, the detection of mutants, both recessive and dominant, would be made rela- tively easy by using pure lines. As men- tioned in Chapter 1. suddenly-appearing phenotypic variants which are due to muta- tion and not to environmental fluctuation are occasionally found in pure lines of self- fertilizing bean plants. When completely Changes Involving Unbroken Chromosomes pure lines cannot be obtained because self- fertilization does not occur, detection of mutations is facilitated by knowing the pre- existing genotypes. Although we have seen how a phenotypc is proved to be the result of a mutation, we have not determined the basis for the genetic change involved. The change could con- ceivably encompass as much as an entire genome, or as little as the genetic material in a single gene locus. The latter type of change may not be detectable cytologically. Although a cytological study has not been made, genetic studies indicate that the chromosomal change associated with the occurrence of Hexaptera is submicroscopic. Let us now look at mutants known to be as- sociated with a gross visible change in chro- mosome composition, as detected by either genetical or cytological methods, or both, and leave for later the consideration of muta- tions involving submicroscopic changes in chromosomes. Heteroploidy In the evening primrose, Oenothera, a giant type called gigas is found to be a mutant. Other Oenothera, like most sexually repro- ducing species, are diploid, having two sets of chromosomes — one genome contributed by each of the gametes. In the gigas type, cytological examination shows that there are three genomes; that is, the individuals are triploid. Studies of other groups of diploid plants reveal related types which prove to have four genomes (tetraploids), others may have six sets {hexaploids) or eight (octa- ploids). A chromosomal composition made up of an abnormal number of normal chro- mosome sets is said to be heteroploid. The occurrence of extra whole genomes is called polyploidy, a term which is applicable for multiples of the haploid number when mono- ploidy is the normal condition. Note that changes in genome number preserve the 151 figure 11-2. Ploidy in Datura (N = 12) (silhouettes). same ratios that chromosomes (and genes) have to each other under normal conditions. Such changes are said to be euploid (right- fold). 1. Autopolyploidy Different forms of the Jimson weed. Datura, carry different numbers of whole genomes, or ploidies.1 Some are haploid, others diploid, triploid, or tetraploid. The flowers that each of these types produces are shown in Figure 11-2, line B (the respective seed capsules are shown in line A). Note that flower size increases with ploidy. The seed capsules illustrated are those which might have been obtained had the individual under test been fertilized by pollen from a diploid — the differences in size being due partly to the number of seeds that have set or developed. Triploid and tetraploid embryos are found 1 The following is based upon work of A. F. Blakeslee and J. Bellinc 152 CHAPTER 11 iiwri: 11-3. Normal {left) and triploid (right) D. melanogaster females. The body of the 3N female is slightly larger than the 2N female and also has slightly larger cells. (Drawn by E. M. Wallace.) in a variety of mammals, and even races of some animals are polyploid. For example, tetraploids of the following are known: the water shrimp, Artemia; the sea urchin. Echinus; the roundworm, Ascaris; and the moth, Solenobia. Polyploid larvae of sala- manders and of frogs also have been ob- tained, although races are not formed. Poly- ploidy is also found in Drosophila. Female Drosophila have been found that are triploid (3X + 3 sets of A) (Figure 11-3) and tetraploid (4X + 4 sets of A). Somatic parts of Drosophila individuals have been found to be haploid (IX + 1A set). One way that ploidy can increase is by the addition of genomes of the same kind as those present — by allopolyploidy — as was the case with Datura. Autopolyploidy can arise several different ways: 1 . Mitotic anaphase may be abnormal, so that the doubled number of chromosomes becomes incorporated into a single nucleus which thereafter divides normally to produce daughter polyploid nuclei and eventually — by asexual reproduction — polyploid progeny. 2. Sometimes two of the haploid nuclei produced by meiosis fuse to form a diploid gamete which, after fertilization with a hap- loid gamete, forms a triploid zygote. (Com- plementarily, fertilization of a gamete formed without a nucleus may initiate development of a haploid.) 3. Haploid individuals may undergo meio- sis and, although this usually results in gam- etes containing only part of a genome, a complete haploid gamete can sometimes be produced which, upon fertilization with an- other haploid gamete, forms a diploid zygote. By interfering with mitosis and meiosis, autopolyploidy can be artificially induced by: drugs like colchicine or its synthetic substi- tute, colcemide (which destroys the spindle, thereby preventing the anaphase movement of chromosomes); environmental stresses like starvation and cold; or energetic radia- tions. Some females of Solenobia produce hap- loid eggs; others produce diploid eggs. Both types of eggs start development without fertilization; that is, they begin developing parthenogenetically. During development, however, nuclei of the respective individuals fuse in pairs to establish the diploid and tetraploid conditions. In this case, normal parthenogenesis leads to normal diploidy and tetraploidy. In many other organisms, arti- ficially induced parthenogenesis may begin haploid development. In the case of an ordinarily diploid indi- vidual, development as a haploid usually pro- duces abnormalities. These must sometimes be due to the expression of detrimental genes which are not expressed in a diploid because their normal alleles are present in homol- Changes Involving Unbroken Chromosomes 153 ogous chromosomes. However, this is not always the case. If chromosome doubling — naturally or artificially induced — occurs at an early stage, a normal diploid (and homo- zygous) embryo may be produced; for ex- ample, chromosome doubling has produced parthenogenetic salamanders and (female) rabbits. In these instances, at least, abnor- mal development as a haploid must have its basis in quite a different factor — probably one involving the surface-volume relation- ships within the nucleus and between the nucleus and the cytosome. These relation- ships are changed when cells that are adapted to be diploid are haploid. A sim- ilar explanation can be offered for the ob- servation that development of triploid and tetraploid mouse zygotes ceases after a few days, even though initially they have a nor- mal mitotic rate. Ploidy changes also occur during gameto- genesis and fertilization. These and certain other examples of ploidy change already dis- cussed are normal in various organisms. (A ploidy change should be considered muta- tional only when it is novel.) Autopoly- ploidy can occur as a normal process in a portion of a multicellular organism; for ex- ample, it occurs normally in certain somatic tissues in man such as liver cells. Many of the examples of autopolyploidy mentioned involve an increase in ploidy which is accom- plished by endoreplication; that is, the ge- nomic contents replicate and remain in one nucleus. In these cases, the daughter chro- mosome strands separate to produce an in- creased number of separate chromosomes, each chromosome in the nucleus proceed- ing independently to mitotic metaphase. In another consequence of endoreplication, all daughter chromosome strands remain syn- apsed, so the number of separate chromo- somes is not increased. Let us consider an example of this condition as found in the giant salivary gland cells of Drosophila larvae. 2. Poly ne my Recall that the metaphase chromosome in the usual cell of Drosophila is rod-shaped (see Figure 7-5) and contains chromatids each of which is coiled tightly in a series of spirals like those in a lamp filament, and that during interphase the chromatids un- wind. The chromatids in the chromosomes of the salivary gland cell nucleus are also in an unwound state, perhaps even more so than in ordinary interphase, and undergo three special changes: 1. Each chromosome present endorepli- cates synchronously a number of times in succession, so that one chromosome pro- duces two, two produce four, four produce eight, and so on. Endoreplication can occur at least nine times, so each chromosome can produce 512 daughters. 2. All daughter strands, instead of sep- arating, remain in contact with the homol- ogous loci apposed, giving the appearance of a many-threaded — polynemic or polytenic — cable. 3. The original members of a pair of ho- mologous chromosomes are paired at homol- ogous loci, demonstrating what is called somatic synapsis. Accordingly, a double cable is formed which can contain as many as 1024 chromosomes. When seen under the microscope (Figures 1 1-4 through 6), these double cables have a cross-banded appearance due to differ- ences in density along the length of the un- wound chromosomes. A band is formed by the synapsis of the same dense regions in all the strands; in this case, an interband region is also formed by the synapsis of correspond- ing regions of lesser density (Figure 11-5). The pattern of bands is so constant and characteristic that it is possible to identify not only each chromosome but different re- gions within a chromosome (Figure 11-6). The giant size of salivary chromosomes, very long because they are unwound and thick I. 14 ( IIAI'I I K II V\ ' "* X \ 3L '. v 2L « '*» Vx % kttt* } 2R '-'% ♦ 3R % - v* *3 figure 1 1 -4. Salivary gland chromosomes of a female larva of D. melano- gaster. (Courtesy of B. P. Kaufmann; by permission of The American Genetic Association. Journal of Heredity, Frontispiece, vol. 30, No. 5, Max, 1939.) I K.i RE 1 1-5. A hand (at lop) and interhand (below) region of a stretched Drosophila salivary gland chromosome. Photographed with the electron microscope at a magni- fication of approximately J2.200X- Present enlargement is about 13.000X. (By permission of The American Genetic Association. Jour- nal of Heredity, vol. 43, p. 231. 1952.) figure 1 1-6. The pair of fourth chromosomes as seen in salivary gland nuclei ( each homolog is highly polynemic) and at mitotic metaphase (arrow), drawn to the same scale. I By permission of The American Genetic Association. C. B. Bridges, "Salivary Chromosome Maps," Jour- nal of Heredity, vol. 26, p. 62, 1935.) '•/ SCALE 5 A*- &%£} #l%tviin ■ .-^•'*-n»v.U-Vt6' «iii Changes Involving Unbroken Chromosomes 155 because of synapsed polynemes, offers a unique opportunity to correlate genetical and cytological events. ( It should be noted that, as a rule, giant polyncmic chromosomes are produced in cells which will never divide again. ) At any given stage of a cell cycle, most of the chromosomal material reacts similarly to certain staining procedures and, therefore, is called euchromatic (truly or correctly colored). Other portions of the chromo- somes stain either darker or lighter and are said to be heterochromatic. Although het- erochromatin may be located at various places along a chromosome arm, it is nor- mally found adjacent to the centromere and, to a lesser extent, near the ends. Hetero- chromatin also has the characteristic of being less specific in synapsis than is euchromatin, different heterochromatic regions located in the same chromosome, its homolog, or in nonhomologous chromosomes often being found synapsed. In the giant salivary gland nuclei of Drosophila larvae, the heterochro- matic regions nearest the centromeres of all chromosomes synapse to form one mass, called the chromocenter. This is the center from which the double cables radiate in Figure 1 1-4 and at the left of Figure 1 1-6. Also, the heterochromatic regions nearest the ends are sometimes found synapsed with other heterochromatic regions, especially the chromocenter. In squashing the nuclei to separate and flatten the salivary chromo- somes, two synapsed heterochromatic re- gions may be pulled apart, but show evi- dence of synapsis because they are still connected by strands of apparently sticky material. The right end of the fourth chromosome polynemes in Figure 1 1-6 shows such glutinous matter, probably indicating synapsis with the chromocenter. Heterochromatin is chromatic and is not to be identified with the regions between bands; interband regions do not seem to contain the Fculgcn-stainable material - and apparently are achromatic, as is the spindle. 3. Allopolyploidy Ploidy can increase another way besides allopolyploidy. Two species can each con- tribute two or more genomes to form a third species which is called an allopolyploid. Cultivated wheat is an allopolyploid. As expected, allopolyploids often show a com- bination of characteristics of their different parent species. This type of polyploidy is discussed in more detail in Chapter 18. Changes in genome number represent the class of normal and mutational events in- volving the largest unit of genetic material. Although many plants are polyploid and one plant has 512 chromosomes, polyploidy will produce a chromosome number that is un- wieldy in nuclear division if it occurs many times in succession. It should also be noted that certain other classes of mutation, like those involving a single locus, have greater difficulty expressing themselves in autopoly- ploids than they have in haploids or diploids, in which no other, or just one other, homol- ogous locus is able to mask the mutant effect. Aneusomy The next category of mutations to be dis- cussed involves the addition or subtraction of part of a chromosome set. Such muta- tions upset the normal chromosomal and gene balance and produce aneuploid ("not right-fold") chromosomal (genetic) consti- tutions by having the incorrect number of particular chromosomes (aneusomy). By what mechanisms can single whole (un- broken) chromosomes be added to or sub- tracted from a genome? 1 . In Drosophila Recall that nondisjunction in the germ line of Drosophila can produce offspring, - See D. M. Steffensen (1963). 156 CHAPTER 1 1 otherwise diploid, that arc XO, XXX, and XXY. Nondisjunction of the small fourth chromosome can load to the production of individuals with one fourth chromosome (huplo-IY individuals) or three (triplo-IY individuals) (Figure 11-7) being in this respect monosomic and trisomic, respec- tivelj : instead of disomic as is normal. Even though addition or subtraction o\' a chromo- some IV makes visible phenotypic changes from the disomic condition as can be seen from the phenotypes, both aneusomic changes are viable. On the other hand, individuals monosomic or trisomic for cither of the two large autosomes die before com- pleting the egg stage. When triploid Drosophila females — with all chromosomes trisomic — undergo meiosis, bundles of three homologous chromosomes (trivalents) may be formed at synapsis. This is because, at one place along the length of a chromosome, the pairing is between two homologs, and at another place it is between one of these two and the third homolog. In this way, although pairing is two-by-two at all levels, all three homologs are held to- gether as a divalent. At the first mciotic division the two homologs that are synapscd at their ccntromeric regions, separate, and go u> opposite poles, while the third homolog goes to either one of the poles. At the end o\' the second mciotic division, two nuclei each have one homolog of the trivalent. and two nuclei each have two homologs. The same result is obtained when synapsis is entirely between two homologs and excludes the third. Since each of the four trisomies present at metaphase I segregates independ- ently, eggs are produced which have one of the following: 1 . Each chromosome type singly and, therefore, contain one complete ge- nome (being haploid) 2. Two chromosomes of each type and, therefore, contain two genomes (be- ing diploid) 3. Any combination in which some chro- mosomes are represented once and others twice (being aneusomic). figure 11-7. Haplo-IV (left) and triplo-IV (right) females of D. melano- gaster. The haplo-IV is smaller than the wild-type female shown in Fig. 2-6. (Drawn by E. M. Wallace.) Changes Involving Unbroken Chromosomes 157 We see, therefore, that meiosis produces many aneusomic gametes when the number of homologs is odd, as it is in triploids, pentaploids, etc. In tetraploids, since each chromosome can have a partner at meiosis, the four homologs often segregate two and two. Sometimes, however, the four homo- logs form a trivalent and segregate three and one, so that some aneusomic gametes are produced by polyploids with even numbers of homologs. Because the phenotypic effect of any gene depends directly or indirectly upon the phenotypic effects of most, if not all, of the other genes present, it is expected that a diploid individual contains, in its two sets of chromosomes, a proper balance of genes for the production of a successful phenotype. It is not surprising, then, that a haploid in- dividual mated to a diploid produces very few progeny, since after fertilization most zygotes are chromosomally unbalanced by the absence of one or more chromosomes needed to make two complete genomes. Mated to a diploid the triploid individual also produces zygotes that are imbalanced but in the opposite direction, having one or more chromosomes in excess of two ge- nomes. In matings with diploids, however, the triploid individual usually produces more offspring than the haploid. This observa- tion can be explained as the result of the lesser imbalance brought about by the addi- tion of chromosomes to the diploid condi- tion than by the subtraction of chromosomes from it. This effect can be seen by com- paring how far from normality (diploidy) each of the two abnormal conditions is. When one chromosome is in excess, the ab- normal chromosome number of three is one and a half times larger than the normal num- ber of two; when one chromosome is miss- ing, the abnormal chromosome number of one is two times smaller than the normal number. Thus, the addition of a chromo- some makes for a less drastic change in bal- ance than the subtraction. Accordingly, knowing that the triple dose of a large auto- some is lethal in Drosophila, we can cor- rectly predict that the single dose is lethal also. In these cases, death is attributable to genetic imbalance due to an excess of the genes present in a long autosome in trisomic individuals and to a deficiency of these genes in monosomic individuals. 2. In Datura Chromosome addition and subtraction can also be studied in Datura 3 whose haploid chromosome number is twelve. It is pos- sible to obtain twelve different kinds of in- dividuals, each having a different one of the twelve chromosomes in addition to the dip- loid number. Each of these trisomies is given a different name such as "Globe." It is also possible to obtain viable plants that are diploid but missing one chromosome of a pair; these are monosomies or haplosomics. Individuals with two extra chromosomes of the same type (tetrasomics) or with two extra chromosomes of different types (dou- ble trisomies) are also found. Datura enables us to test the phenotypic consequences of disturbing the normal bal- ance among chromosomes. Compare, in Figure 11-8, the seed capsules of the normal diploid (2N) with those of diploids having either one extra chromosome (2N + 1 ) of the type producing Globe or two of these (2N + 2). The latter two polysomics can be called trisomic diploid and tetrasomic diploid, respectively. Although the tetra- somic is more stable chromosomally (each chromosome can have a partner at meiosis) than is the trisomic, the tetrasomic pheno- type is too abnormal to establish a race, since it has a still greater genetic imbalance than the trisomic and produces a still greater deviation from the normal diploid phenotype. 3 Based upon work of A. F. Blakeslee and J Belling. 15S ( ii \r i i i' II t TETRAPLOID AN #• ##? 2N I 2N 2 IGlobel 4N I 4N 2 4N 3 FIGUR] 11-8. /.//t'<7 upon f/ie capsule of Datura of the presence of one or more extra "Globe" chromosomes. In comparison, the tetraploid (4N) indi- vidual is phenotypically almost like the dip- loid, since chromosomal balance is undis- turbed. The tetraploid which has one extra Globe chromosome (4N -+- 1, making it a pentasomic tetraploid) deviates from the tetraploid in the same direction as the 2N -f- 1 deviates from 2N, but does so less extremely. Hexasomic tetraploids (4N + 2) de\ iate from 4N just about as much as 2N + 1 deviates from 2N. It is clear, therefore, that adding a single chromosome to a tetraploid has less phenotypic effect than its addition to a diploid, since the shift in balance between chromosomes is relatively smaller in the former than in the latter. Thus, polyploids can stand whole chromo- some additions and subtractions better than diploids can. Since crosses between tetraploid Datura produce fertile seed in amounts sufficient to maintain a tetraploid race, the question arises, can a tetraploid race of Drosophila be produced? As mentioned, the gametes of a tetraploid Drosophila female contain complete genomes more often than do those of triploids. Since it produces many diploid eggs, the tetraploid female presents no diffi- culty for the continuity of a tetraploid race. To be o\' normal sex, a tetraploid male has to carry 2X + 2Y + 4 sets of A (Chapter 8). But the X's (and Y's) in such a male usually synapse with each other during meio- sis so that after meiosis each sperm carries IX and I Y in addition to the 2A sets. In fertilizing eggs (from tetraploid females) containing 2X + 2 sets of A. sperm of this type produce zygotes with 3X -f- 1Y + 4 sets of A which develop as sterile inter- sexes. Thus, a self-maintaining tetraploid race of Drosophila cannot be established. In fact, we can conclude that any species containing a heteromorphic pair of sex chro- mosomes (such as X and Y) cannot form polyploid races, since the correct balance between sex chromosomes and autosomes is upset by the meiotic divisions. This factor probably explains why polyploid races and species are rarer among animals than among plants whose sexuality (as in monoecious forms) is not associated with heteromorphic homologs. 3. /// Man Down's syndrome, or mongolism, in hu- man beings is sometimes the result of a trisomic diploid chromosomal constitution. In this case, the trisomic is number 21 the third smallest of all human chromosomes (the smallest being the Y) (Figures 11-9, 1 1-10). Trisomies for several other of the smaller autosomes are also known, each pro- ducing its own characteristic set of congenital abnormalities. Trisomy for the largest auto- somes is apparently lethal before birth, prob- ably due to the imbalance of too many genes. The very severe phenotypic defects observed among the least affected auto- somals trisomic individuals makes it a rea- sonable expectation that the monosomic condition of any autosome is lethal before birth — in accordance with the view that chromosome subtraction is even more detri- mental than chromosome addition. Changes Involving Unbroken Chromosomes 159 figure 1 1-9. The chromosomal complement of a normal human female. Cell was in mitotic metaphase {hence chromosomes appear double except at the centromere) when squashed and photographed. (Courtesy of T. C. Hsu.) The frequency among live births of Down's syndrome due to trisomy has been determined as approximately 0.2%. Most cases of Down's syndrome occur among the children of older mothers and are due pri- marily to nondisjunction during oogenesis. If other chromosomes have a similar fre- quency of nondisjunction, there might be a minimum of 4.4% (22 X 0.2% ) of zygotes autosomally trisomic at conception. There might also be another 4.4% of zygotes that are autosomal monosomies, due to the equal chance that the haploid meiotic product complementary to the one which is disomic — the nullosomic one — becomes the egg. In fact, more nullosomic than disomic gametes are expected, since a chromosome left out of one daughter nucleus need not be included in the sister nucleus. Supporting a normally high frequency of aneusomy is the observa- tion that about one quarter of aborted hu- man fetuses show a chromosomal derange- ment. It is expected, moreover, that main conceptions involving aneusomy. especially monosomy, are lost so early in pregnanes that they go unnoticed. Nondisjunction leading to aneusomy can also occur in the paternal germ line of man, 160 CHAPTF.R I I 1) 11 (1 1 2 3 I) H n 4 5 6 II X 1) if II 7 8 9 11 If 10 11 II 12 ft it tt 13 14 15 16 17 li 18 ** I » id 20 A 4 A 21 II 22 figure 11-10. Chromosomal constitution found in a female showing Down's syndrome. (By permission of M. A. Ferguson-Smith and A. W. Johnston, and The Annals of Internal Medicine, vol. 53, p. 361, I960.) After photographing a squash preparation like that in Figure 11-10, the chromosomes are cut out and "paired" as shown here. although this does not seem to contribute very significantly to the total observed fre- quency. In man, almost all nondisjunction is associated with the aging of oocytes. The reverse is true in the mouse, however, even though mouse females — like human fe- males— are born with all their germ cells in the oocyte stage. Thus, in the mouse, marked chromosomes show that spontaneous aneusomy almost always has a paternal origin. It should be noted that, in the mouse, viable aneusomy also occurs for the sex chromosomes and for certain small auto- somes when trisomic.' 4 See A. B. Griffen and M- C Bunker (1964). The incidence of nondisjunction can be increased by high energy radiations. Carbon dioxide, other chemical substances, and cer- tain diploid genotypes can increase the non- disjunction rate in Drosophila. In human beings, the evidence that older women are more apt to have trisomic children suggests that some metabolic defect associated with increased age increases the chance for non- disjunction. Although chromosome loss may result from spontaneous mciotic and nonmeiotic nondisjunction in diploids as well as from the meiotic process normally taking place in polyploids, it should not be inferred that Changes Involving Unbroken Chromosomes 161 these are the only ways entire chromosomes can be lost. Mosaic Heteroploidy and Aneusomy Mutations leading to heteroploidy need not involve germ cells or the entire organism, as mentioned with respect to asexually re- producing species (p. 152). Sexually repro- ducing species of plants and animals may also show mosaicism for ploidy involving reproductive or nonreproductive tissues, or both. In man, for example, a baby boy has been studied who is diploid in some tissues and triploid in other, normally dip- loid, tissues. About 3% of cells in certain human tissue cultures show such changes in ploidy. Aneusomy can also originate at any mitotic, as well as meiotic, nuclear division. Thus, nondisjunction at the first nuclear di- vision of a normal human zygote might pro- duce one nucleus that is monosomic and one trisomic for chromosome 21. In this case the former nucleus is expected to die, and the latter nucleus, to produce a completely mongoloid individual. Some of the aneusomics born of older mothers may have originated in such a post- zygotic nondisjunction, as is the case in mice. If nondisjunction occurs later in development, it produces complementary monosomic and trisomic mutant patches in a diploid background, which — in the case of autosomes in man and mouse — are usually expected to be lethal to the individual. That such nondisjunctions or chromosome losses do occur with appreciable frequency is sug- gested by the frequent occurrence in human adults of a few cells per hundred which are scored as having one or two chromosomes too few or too much. It is extremely un- likely that all, or even most, of these abnor- mal counts are due to experimental errors in preparing or in scoring the cells. Under normal circumstances one would expect the aneusomic cells produced after birth to be functionally inferior to their neighboring euploid cells and, therefore, at a selective disadvantage. Because of the large genetic unbalance it produces, addition and subtraction of whole chromosomes is a class of mutation which involves a phenotypic change too drastic to play a very significant role in evolution. SUMMARY AND CONCLUSIONS The mutational events involving the largest recombinational unit of genetic material are euploid changes in the number of whole sets of chromosomes — heteroploidy. Ploidy can increase by allopolyploidy, autopolyploidy, and polynemy. The modes of origin and the breeding behavior of autopolyploids, and the origin and structure of the giant polynemic chromosomes in the salivary gland of Drosophila larvae are considered in detail. Loss or gain of part of a genome — aneuploidy — can result from nondisjunction and the segregation of chromosomes in polyploids, especially those possessing an odd num- ber of genomes. Not only do such mutations occur in the germ and somatic lines spontaneously, but they may be initiated or have their frequency enhanced by physical and chemical factors. The addition or subtraction of single chromosomes results in aneusomy. The ab- sence of a chromosome is more detrimental to survival than an excess. Aneusomy produces too drastic a phenotypic change to be as inportant in evolution as heteroploidy. Ki2 CHAPTER 1] REFERENCES ^uerbach, C, Mutation. An Introduction to Research on Mutagenesis. Part I. Meth- ods, Edinburgh: Oliver and Boyd, 1962. Blakeslee, \. I .. "New Jimson Weeds from Old Chromosomes," J. Hered., 25:80 108, 1934. Blakeslee, \. I . and Belling, J., "Chromosomal Mutations in the Jimson Weed. Datura Stramonium," J. Hered., 15:194 206, ll>24. Bridges, C. G., and Brehme, K. S., The Miliums oj Drosophila Melanogaster, Wash- ington. D.C.: Carnegie Institution of Washington, Publ. 552. 1944. Burdette, W. J. (Ed.). Methodology in Mammalian Genetics, San Francisco: Holden- Day, Inc.. 1963. Dobzhansky, lh.. Genetics and the Origin of Species, 2nd Ed.. New York: Columbia University Press. Chap. 7. pp. 223-253. 1941. Griffen, A. B., and Bunker. M. C. "Three Cases of Trisomy in the Mouse." Proc. Nat. Acad. Sci.. U.S.. 52:1194-1198. 1964. Heitz. E.. and Bauer. H., "Beweise fiir die Chromosomennatur der Kernschleifen in den Knauelkernen von Bibio hortulanus L. (Cytologische Untersuchungen an Dipteren, 1)." Z. Zellforsch.. 17:67-82. 1933. Painter. T. S.. "A New Method for the Study of Chromosome Rearrangements and Plotting of Chromosome Maps," Science, 78:585-586, 1933. Reprinted in Classic Papers in Genetics, Peters. J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, pp. 161-163. 1959. Patau, K.. Smith. D. W.. Therman, E., Inhorn. S. L., and Wagner, H. P., "Multiple Congenital Anomaly Caused by an Extra Autosome." Lancet, 1:790-793, 1960. Russell. L. B.. "Chromosome Aberrations in Experimental Mammals," Progress in Medical Genetics. 2:230-294. 1962. Steffensen, D. M.. "Evidence for the Apparent Absence of DNA in the Interbands of Drosophila Salivary Chromosomes." Genetics, 48:1289-1301, 1963. Suomalainen, E.. "Significance of Parthenogenesis in the Evolution of Insects," Ann. Rev. Ent., 7:349-366, 1962. White. M. J. D.. Animal Cytology and Evolution, 2nd Ed.. Cambridge: Cambridge University Press. 1954. QUESTIONS FOR DISCUSSION 11.1. How do we know that the genetic differences in a population today were not always present? 11.2. What have you learned in this chapter about the characteristics of mutation? 11.3. What is the relation between mutants and genes? Mutants and recombination? 11.4. From your present knowledge, how would you modify the statements on page 1 1 relative to the ploidy of gametes? 11.5. Describe at least two different ways that the trisomy causing Down's syndrome may originate. 1 1.6. The only presently known case of trisomy for a chromosome of the 19-20 group occurred mosaically in a six-year-old boy. To what do you attribute this? Changes Involving Unbroken Chromosomes 163 1 1.7. Discuss the statement: All somatic cells from diploid zygotes are chromosom- ally identical. I 1.8. Do you suppose that the human species will henefit from a discovery that certain of its members are trisomic? Explain. 1 1 .9. What are the advantages of allopolyploidy? Of allopolyploidy? 11.10. What genetic explanation can you offer for the fact, demonstrated in Figure 1 1-2. that the seed capsule of the Datura haploid is smaller than that of the triploid? 1.11. What do you consider to be the advantages and disadvantages of polynemy? 1.12. Unfertilized mammalian eggs can contain ploidies of IN. 2N. 3N, or 4N. Explain how each of these could be produced. 1.13. How can you explain the fact that persons with Down's syndrome are more susceptible to leukemia than normal diploids? 1.14. Explain why individuals with Down's syndrome show a wide variety of pheno- typic differences as well as similarities in their abnormalities. 1.15. Some babies classified as normal at birth are clearly mongoloid when a year old. What would you do to assure an early diagnosis of Down's syndrome? 1.16. Would you expect a correlation between producing a child with Down's syn- drome and the frequency with which the mother has abortions? Subsequent children with Down's syndrome? Explain. 1.17. Should a woman with a trisomic mongoloid sibling be more than ordinarily con- cerned about having a child of this type? Explain. 1.18. After examining Figures 11-8 and 11-9, discuss the precision with which a given human chromosome can be identified. 1.19. How would you proceed to determine the somatic chromosome composition of a given human individual? 1.20. Discuss the phenotypic effects of adding an N-l genome to individuals that are normally N. 2N, 3N. or 4N. 1.21. R. A. Turpin reported two cases of monozygotic twins. One set contains an XY male and an XO female; the other set is composed of a disomic-21 male and a trisomic-21 male. Discuss the mechanisms probably involved in pro- ducing such twins. Include in your hypothesis the additional fact that one XO cell is also found in the first XY individual mentioned. Chapter 12 STRUCTURAL CHANGES IN CHROMOSOMES T |he two classes of mutation dealt with in Chapter 11 in- volved changes in chromo- somal content of unbroken individual or sets o( chromosomes. In some instances, mu- tants are based upon the gain, loss, or shift of a part of one or more chromosomes. All such structural changes are preceded by chromosome breakage, which — ignoring chromatids for the present — results in two new. "sticky" ends. When several breaks are produced, the new ends can join together but only in pairs, any new end capable of joining any other new end. Moreover, an end produced by breakage cannot join the normal (unbroken) end of a chromosome. Thus, originally free ends of chromosomes are not sticky because they have genes, called telomeres, which serve to seal them off, making it impossible for a normal end to join to any other. The two ends produced by one break usually join together in what is called resti- tutional union even when ends produced by other breaks coexist in the same nucleus. This indicates that proximity favors the union of sticky ends. Although restitu- tional union usually occurs and thereby restores the original linear order of the chro- mosome, the ends uniting may sometimes come from different breaks, so that a new chromosomal (gene) arrangement is pro- duced. The latter union is, therefore, a non- restitutioncd, or exchange, or cross-union type. Let us see how nonrestitutional unions 164 produce various structural changes in chro- mosomes. Consequences of a Single Chromosome Break Consider first the consequences of a single chromosome break; that is. a break through both chromatids (Figure 12-1). Diagram 1 represents a norma] chromosome (its chro- matids are not shown), whose centromere is indicated by a black dot. In diagram 2 this chromosome is broken. If the new chromosome ends join together, that is. restitute, no chromosomal rearrangement is produced. Although restitution usually oc- curs, it may sometimes fail because the new ends spring apart or are moved apart by Brownian movement or protoplasmic cur- rents. In nonrestitution. chromosome repli- cation produces a daughter chromosome just like the parent — with a break in the same position — as shown in diagram 3 where the two broken sister chromosomes are indi- cated. The union of the piece containing no centromere ( a ) to the centromere-bear- ing piece of the other sister chromosome ( b' ) would, in effect, be restitution as would the joining of a' to b. (Sometimes, only one of these restitutional unions occurs.) If restitution does not occur before or after the chromosome replicates, the ends closest together usually join together, these being the corresponding ends of the sister chromosomes (a with a' and b with b'). As shown in diagram 4. the results of such nonrestitutional unions are one chromosome with no centromere (an acentric chromo- some), and one with two (a dicentric chro- mosome). Note that both the acentric and the dicentric chromosomes are composed of identical halves lengthwise, each, therefore, being termed an isochromosome. (This dia- gram shows the chromosomes contracting preparatory to metaphase.) In diagram 5 we can see that in mitotic anaphase the acentric isochromosome is not pulled toward either pole, whereas the dicen- Structural Changes in Chromosomes 165 trie one is pulled toward both poles at once. The acentric isochromosome is, therefore, not included in either daughter nucleus and so is lost to both. (The acentric pieces in diagram 3 will be lost in any subsequent division, whether they are joined to each other or do not join at all.) The dicentric isochromosome. being pulled to both poles at once, forms a bridge. A bridge can pre- vent any part of the chromosome from enter- ing either daughter nucleus, so that the dicentric is lost. Alternatively, the centric regions of the dicentric piece can enter the daughter nuclei, and the bridge can either snap at any one of a number of places be- tween the centromeres and free the daugh- ter nuclei from each other, or it can persist, joining the daughter nuclei together. The amount of phenotypic detriment that a single nonrestituting chromosome break will produce in the daughter cells and their progeny cells depends upon the particular chromosome involved, the place of breakage, and the fate of the dicentric piece. Suppose. for example, that chromosome IV of Dro- sophila (often viable as a haplo-IV indi- vidual) is the chromosome involved. The break can occur at any position in IV, and the loss of the genes in the acentric piece, though detrimental, does not usually cause death; neither does the loss of the entire dicentric fragment if excluded from both daughter nuclei, nor, probably, does a snap in the bridge between the daughter nuclei. (In the last case, each daughter nucleus is deficient, at least for the genes in the acen- tric piece.) Note what happens when a bridge, involv- ing a dicentric isochromosome linearly dif- ferentiated as a.bcddcb.a (the centromere is between a and b). does not snap between the d's. If it breaks between b and c, one fragment is even more deficient (yet viable in this example), whereas the other con- tains an extra dose of the genes in the cd region (and is most probably viable). Re- gardless of where the bridge snaps, both daughter nuclei carry a centric fragment which, after replicating, usually forms a new dicentric isochromosome and can again form a bridge at the next mitotic division. It is possible, therefore, to have bridge-breakage- fusion-bridge cycles in successive nuclear generations. When a bridge fails to break leaving the two daughter nuclei tied together, the en- tanglement of the nuclei may interfere with subsequent attempts at nuclear division. In our example, this interference may be of much greater importance than the presence figure 12-1. Conse- quences of a single nonrestituting chromosome break. w > < !(>(• ( ii \i- ni< I 2 or absence of all of the genes located in the bridge. Suppose, however, that the broken chro- mosome is one of the largo autosomes of Drosophila. Detriment or death to one or both daughter cells may occur because of the genes lost when either the acentric or the dicentric fragment is left out of one or both daughter nuclei. In addition, suc- cessive bridge-breakage-fusion-bridge cycles ma) harm future cell generations via the abnormal quantities o\' chromosomal regions; that is. the ancuploidy. resulting from the off-center breakage of dicentric isochromo- somes. Other things being equal, shorter dicentrics are expected to break more often than longer ones. Of course, any inter- nuclear bridge that does not break may frus- trate future nuclear division. Single chromosome breaks can occur in either the somatic or the germ line. In the latter case, aneuploid gametes may be pro- duced. Since the genes are found to be physiologically inactive in the gametes of ani- mals, aneuploid genomes can enter the egg and sperm without impairing their function- ing (as implied on p. 104). Accordingly, in animals, aneuploid genomes can be carried by unaffected gametes into the zygote, which may subsequently suffer dominant harmful or lethal effects. In many plants, however, the products of meiosis form a gametophyte generation which performs physiological functions requiring gene action, in which case, ancuploidy is usually more lethal or detrimental before fertilization than after. Chromatid Breaks A break can be produced in one and not the other chromatid of a chromosome. Such chromatid breaks are more likely to restitute than chromosome breaks, since the unbroken strand serves as a splint to hold the newly- produced ends close to each other. What appears under the microscope as a break involving only one chromatid may initially have been a chromosome (or isochromatid) break that was followed by restitution of one but not (yet) the other chromatid. Nonrestituted chromatid fragments be- come nonrestituted chromosome fragments if they persist long enough to replicate. To be seen cytologically, a conjoined chromatid or chromosome break produced during inter- phase usually has to persist until nuclear division occurs. Some chromatid and. per- haps, chromosome breaks induced in con- tracted ( metaphasc ) chromosomes may not be visible, the pieces being held together without joining by the nongenetic auxiliary material in a chromosome. To detect such unjoined breaks one would have to wait until the next division. Essentially all ends pro- duced by breaks are not sticky when the chromosome is contracted as during nuclear division; joinings are restricted largely, if not completely, to the period between late telo- phase and early prophase. Accordingly, the later in this period a break is produced, the less likely it is that the ends will join; broken ends produced between early prophase and late telophase have the maximum time for joining but probably also the maximum op- portunity to cross-unite. For simplicity, the discussion which fol- lows is restricted to isochromatid breaks that fail to restitute. The reader is given the task of working out the consequences of aneuploidy resulting from single nonresti- tuted broken chromatids. The lack of fur- ther discussion on this type of mutation does not reflect on the relative frequency or im- portance of chromosome versus chromatid breaks. Agents capable of producing chro- mosome breaks can also produce chromatid breaks; moreover, certain agents may prefer- entially produce chromatid breaks. Consequences of Two Breaks in One Chromosome When a chromosome is broken twice, the two breaking points may be paracentric, that Structural Changes in Chromosomes 167 PARACENTRIC BREAKS ABCDEFGHIJ r A G H I J Deficiency C D B C D E F or B C J E lost F (a) AFEDCBG HIJ Inversion (b) PERICENTRIC BREAKS ABCDEFGHIJ A B C D I J lost J Deficiency G F ABCDHGFEIJ Inversion (c) (d) FIGURE 12—2. Some consequences of two breaks in the same chromosome. is, to one side of the centromere, or peri- centric, that is, with the centromere between them (Figure 12-2). 1. Deficiency Consider a chromosome linearly differen- tiated as ABCDEFG.HIJ, the centromere being between G and H. When the breaks are paracentric in position (for example, between A and B, and between F and G). the fragments can unite to produce a centric chromosome (AG. HIJ, Figure 12-2a) de- ficient for the acentric interstitial (nonter- minal) piece (BCDEF). The ends of the latter fragment may join to produce a ring chromosome, or they may not. In either event, the acentric fragment is usually lost before the next nuclear division. When the breaks are pericentric (for example, between D and E, and between H and I ), the acentric end pieces are lost, even if they join together (Figure 12-2c). The middle centric piece can survive if its ends join to form a ring and if the deficient sections are not exten- sive. Even if a ring survives because it is not too hypoploid (the aneuploid condition in which genes or chromosomal regions are missing), it is still at a disadvantage because a single crossing over either with a nonring (rod) homolog or with another ring results L68 CHAPTER 12 in a dicentric rod or ring, respectively, as can be seen by drawing the appropriate con- figurations. Of course, a nondividing nucleus, in which breakage or another structural change oc- curs, is still euploid. The first occurrence of hypoploid)/ or hyperploidy (aneuploidy due to an excess of genes or chromosome parts) is in the daughter nuclei formed by such a nucleus. This delay in producing an aneu- ploid nucleus should be remembered when we state that chromosomes with small de- ficiencies can be lethal when homozygous, and detrimental when heterozygous; chro- mosomes with large deficiencies usually act as dominant lethals in the next cell genera- tion. Remember also that we have ignored — and shall continue to do so for the rest of the chapter — the usual consequence of two breaks, that is. restitution for all ends pro- duced by breakage. 2. Inversion Another structural consequence of two breaks in the same chromosome is repre- sented in Figure 1 2-2 b and d. In this case, the middle piece is inverted with respect to the end pieces and undergoes exchange unions with them. The result which is either a paracentric or a pericentric inversion (Fig- ure 12-2 b and d), is due to the middle seg- ment moving while the ends are relatively stationary, or the reverse. Note that inver- sion is a euploid rearrangement. Structural rearrangements in chromo- somes can occur in either the somatic or the germ line. An inversion which occurs in the germ line may be retained in the pop- ulation long enough to become homozygous in some individuals. Meiotic behavior is normal in such inversion homozygotes whether or not the tetrad undergoes cross- ing over, since all the strands in the tetrad are identically inverted. Other individuals in the population, however, may possess one inverted and one noninverted homolog, be- ing inversion heterozygotes. Provided the inversion is wry small, these homologs will pair properly everywhere but in the inverted region. Because the homologs cannot twist enough to make homologous loci meet in so short a region, they will fail to synapse and no crossing over will occur. Insofar as crossing over can lead to more adaptive re- combinants, such inversion heterozygotes are at a disadvantage compared to noninversion or inversion homozygotes because of the ab- sence of recombination among genes within the inverted region. Nevertheless, very small inversions do survive in many species. Consider the meiotic process in heterozy- gotes for larger paracentric inversions. In this case (Figure 12-3A), synapsis between homologs occurs for all regions except those adjacent to the points of breakage. This synapsis requires one homolog twisting in the inverted region while the other does not. The figure happens to show the inverted and not the noninverted chromosome twisting, but the reverse is equally likely to occur. If crossing over occurs anywhere outside the inverted region, each of the four meiotic products will be eucentric (having one cen- tromere), as usual. If, however, a single crossing over occurs anywhere within the region inverted — as shown between C and D — the two noncrossover strands of the tetrad will be eucentric (one with and one without the inversion), and the two cross- overs will be aneucentric (having more than one centromere or none). One of the aneu- centrics will be acentric (duplicated for A and deficient for G.HIJ); the other will be dicentric (deficient and duplicated for these respective regions). If the inversion is only moderately long, only one crossing over can occur within it; if sufficiently long, double crossing over is possible. When such double crossing over is of the two-strand type, both crossover strands are eucentric. In animals, gametes function regardless of the ploidy of the meiotic products they Structural Changes in Chromosomes 169 A. PARACENTRIC INVERSION iA .A«B"C — D B. PERICENTRIC INVERSION figure 12-3. A single crossing over in an inversion heterozygote. {See text for explana- tion.) contain. If a gamete contains an aneucentric produced by crossing over in a paracentric inversion heterozygote, this chromosome will usually have a dominant lethal effect after fertilization; that is, individuals heterozygous for moderate to large paracentric inversions are at reproductive disadvantage, which often leads to the elimination of the inversion from the population soon after it arises by muta- tion. This disadvantage is partly avoided in those species having no crossing over in one sex. For example, in the Drosophila male each homolog, inverted or not, is a noncross- over and has the same chance of being in- cluded in the gamete. A special factor oper- ates during meiosis in some species in which the female undergoes crossing over, occur- ring only if the two meiotic divisions occur in tandem, as they do in female Drosophila. In the Drosophila oocyte heterozygous for a paracentric inversion, a single crossing over within the inverted region produces the usual dicentric at anaphase I. But this di- centric serves to hold the dyads at metaphase II, so that the two eucentric monads proceed to the two outermost of the four poles. Therefore, at the end of telophase II, the cen- tric meiotic products are arranged in a row: eucentric; part of dicentric; remainder of di- centric; eucentric. It is one of the two end eucentric-containing nuclei which becomes the egg nucleus, the others degenerate. In this way the dicentric strand is prevented from entering the nucleus that becomes gametic; the gamete, therefore, receives one of the two eucentric, noncrossover strands. That is why in Drosophila, paracentric in- versions of any size rarely cause aneuploid gametes in either sex and can become estab- lished in nature. What products result from a crossing over within the inverted region in a heterozygote for a larger pericentric inversion? As seen in Figure 12-3B, a single crossing over, such as between F and G, produces four eucentric strands: two noncrossovers (one with and one without the inversion) ; one with a dupli- cation (for ABCD) and a deficiency (for IJ) ; the last with a deficiency and a duplica- tion of the respective regions. All strands enter the gametes of males if crossing over occurs in the male. Each strand also has an equal chance of being present in the gametes of females capable of crossing over. This is true even in Drosophila where shunt- ing of euploid strands into the egg nucleus does not occur because all the meiotic prod- ucts are eucentric. Consequently, aneuploidy which results from crossing over within a 170 ( II \PTHR 12 BREAKAGE ABCDEFG HIJ REPLICATION A B C D E F G HIJ AB CDE FGHIJ i u.i ri I 2 4 ( right ) . Duplication ( tandem type ) . CROSS-UNION ABCDECDE FG HIJ A B FGHIJ FIGURE 12-5 (below). Reciprocal transloca- tion between nonhomologous chromosomes. K L M N O P Q R S T U ANEUCENTRIC TYPE: K L M N O Q R S T U EUCENTRIC TYPE: K L Q R S T U P M N O pericentric inversion always puts the hetero- zygote at a reproductive disadvantage. For this reason, only the smallest pericentric in- versions— those which do not synapse when heterozygous — are usually able to survive in the wild. 3. Duplication If, following two breaks in the same chro- mosome, joining is delayed until after the broken chromosome reproduces ( Figure 12-4). the two interstitial pieces and the appropriate end pieces can join to produce a eutelomcric chromosome with the inter- stitial region repeated. This rearrangement is called a duplication. Neither, either, or both of the regions involved in the duplica- tion can be inverted with respect to the orig- inal arrangement. The two remaining end pieces may or may not join to form a de- ficient chromosome. (A deficiency can also be produced without a duplication when the end pieces join before chromosome replica- tion.) Provided that the duplicated region is small enough and acentric, it can survive in nature. Structural Changes in Chromosomes 171 Consequences of Two Breaks in Two Chromosomes What happens when two breaks occur, one in each of two different chromosomes? In the first such case, the two broken chromo- somes are nonhomologous (Figure 12-5). If the two centric pieces unite, a dicentric is formed and the two acentric pieces are lost in the next division, whether they join each other or do not join at all. If all pieces join as indicated, then there is a mutual ex- change of segments between nonhomologous chromosomes, which is called a segmental interchange, or more often, a reciprocal translocation. This is the aneucentric type of reciprocal translocation and often acts as a dominant lethal in a subsequent division, particularly when the dicentric is pulled to- ward both poles at once. The reverse is often just as likely, how- ever; union occurs between the centric piece o\' one chromosome and the acentric piece of the nonhomolog, with the centric piece of the second joining the acentric piece of the first. This reciprocal translocation is of the dicentric type. In individuals heterozygous for such an exchange (Figure 12-6), hav- ing two nonhomologs translocated and two nontranslocated, gametes are formed with deficiencies and duplications if, by segrega- tion, they receive one but not both members of the reciprocal translocation. When the chromosomes in nuclei are com- pressed in a relatively small volume, no broken end is far from any other; usually, FIGURE 12-6. Diagrammatic representation of segregation in dicentric reciprocal translocation heterozygotes. (Chromatids not shown; the spindles — also not shown — have their poles oriented vertically.) <><» SYNAPSIS F Zigzag Circle / \ Open Circle DIAKINESIS AND METAPHASE I TELOPHASE I GAMETES ViV All Euploid vv vv All Aneuploid (Half-Translocationall 172 CHAPTER 12 if one of the two unions deeded tor recip- rocal translocation occurs, so does the other. Such is the case in the nucleus of the Dro- sophila sperm just alter fertilization. In oocytes and probably in other cells that have a relatively large nuclear volume, the dis- tance between the broken ends of nonhomo- logs is so great that reciprocal translocations are comparatively rare and. even if one cross union occurs, the two other broken ends usually tail to join to each other, so that onlj half of a reciprocal translocation — a half-translocation — is produced. The loss or behavior of the unjoined fragments usually causes descendent cells to die or to be ab- normal, as would be expected. Half-trans- locations can also result when heterozygotes for a eucentric reciprocal translocation un- dergo segregation (see Figure 12-6), and only one of the two reciprocals is present in a gamete. Some children with Down's syndrome have 46 chromosomes. These chromosomes include — in addition to two normal num- ber 21*s — an autosomal pair (from group 13-15 or from group 16-18) which is het- eromorphic. one member being longer than usual. The extra piece is probably the long arm of 21, so that the individual is hyper- ploid for 21, being almost trisomic 21. In some cases the mother is phenotypically nor- mal although she is heterozygous for a eucentric reciprocal translocation between 21 and, for example, 15. Her chromosome constitution can be represented by 15, 15.21 (centromere of 15), 21.15 (centromere of 21 ). 21. An egg containing 21 and 15.21 (the half-translocation) fertilized by a nor- mal sperm (containing 21 and 15) produces the almost-trisomic-21 mongoloid under dis- cussion. (The break in 15 must have been so close to the end that the hypoploid seg- ment in the half-translocation mongoloid in- dividual was not lethal.) In other cases such half-translocational mongoloids have half-translocational nonmongoloid mothers with 45 chromosomes. These mothers have only one normal 21, one normal 15, for example, and the half-translocation 15.21. The hypoploidy for both 21 and 15 must be small enough to be viable in the mother, who can produce the aneuploid gamete that makes her child mongoloid. (Note in the cases cited above no relation exists between mother's age and the occurrence of half- translocational mongoloid children. ) In the second case in which two chromo- somes are broken once, the chromosomes are homologs (ABCDEFG.HIJ). The breaks are usually at different places, for example, between A and B in one chromo- some, and between D and E in the other. Here, also, reciprocal translocation can occur two ways. The aneucentric type produces a dicentric and an acentric chromosome whose fate can be readily predicted. The eucentric type produces two eucentric chro- mosomes, the BCD region being deficient in one and duplicated in the other. From the preceding discussion, one would expect eucentric reciprocal translocations to tend to be eliminated from the population soon after arising by mutation, since they are usually heterozygous and cause about 50% of gametes to be half-translocational aneuploids. Certain eucentric reciprocal translocations, however, seem to be excep- tions. In these cases, almost a whole arm of each chromosome is mutually exchanged. Such whole-arm reciprocal translocations — when heterozygous in Drosophila and prob- ably in most other species — tend to synapse and disjoin in the following way: at synapsis the heterozygous reciprocal translocation forms an X configuration composed of two tetrads (Figure 12-6). Later, when homol- ogous centromeres repel each other, alter- nate centromeres move toward the same pole, so that as the chiasmata move towards the ends, a zigzag arrangement of four dyads results (Figure 12-6). Because of this al- ternate centromeric orientation, anaphase I Structural Changes in Chromosomes 173 produces one nucleus without the transloca- tion and the other with the full transloca- tion. Since euploid gametes are usually formed, such translocation heterozygotes are not at an appreciable reproductive disad- vantage. Increasing Gene Number Both here and in Chapter 1 I , it has been pointed out that a change in ploidy can sur- vive in nature when it involves either no shift in chromosome balance (because it deals with whole genomes) or eucentric aneuploidy due to small segments of chro- mosomes which are hypo- or hyperploid. In the latter cases, the number of deficient or duplicated genes is small enough to pro- duce a tolerable phenotypic effect. It is reasonable to assume that the greater the amount of chromosomal material, the greater the complexity possible in an organism and. consequently, the greater the diversity pos- sible in its phenotype and adaptiveness. Ac- cordingly, viable changes in ploidy must be particularly important in organic evolution. It is desirable, therefore, to specify some of the different ways that small numbers of genes can be added to a genome after break- age. Two methods of increasing gene number after breakage have already been described. One requires two breaks in the same chro- mosome; the entire chromosome then repli- cates, after which the broken ends join to form a chromosome with the interstitial piece duplicated (p. 170); the other involves each member of a pair of homologs break- ing once in a different region before eucen- tric cross union (p. 172). A third mechanism involves three breaks in one chromosome. The two interstitial pieces exchange positions, producing what v "> figure 12-7. Inversion het- erozygotes in corn (pachy- nema) (courtesy of D. T. Morgan, Jr.) and in Drosoph- ila (salivary gland) (courtesy of M. D enter ec). 171 CHAPTER 12 i m ^ v. # % v *„ t I [GURE I 2 8. Salivary gland chromosomes heterozygous for a shift within the right arm of chro- mosome 3 of Drosophila melano- gaster. A piece from map region "98" is inserted into map region "91." I lie rightmost buckle is clue to the absence of the shifted segment; the leftmost buckle is due to its presence. ( Courtesy of B. P. Kaufmann. ) is called a shift. If, in the heterozygote for a shift, the homologs pair up and a crossing over occurs in the region of the shift, a sec- tion of one of the crossovers will be in duplicate, as can be seen by tracing the re- sultant strands. Two breaks in one chromosome and one in a nonhomolog can result in the interstitial piece of the first chromosome being inserted into the second. This result is called trans- position. A transposition-containing chro- mosome can occur in subsequent generations not with the nonhomologous, deficient chro- mosome from which the piece was trans- posed, but with two normal chromosomes of that type. In this way an individual is produced containing a pair of normal hom- ologs and a part of the normal homolog present in hyperploid condition in a non- homolog. The preceding indicates how the same type of structural change — duplication — can result from different types of breakage events. For this reason, one cannot always specify the particular number of nonrestitut- ing breaks originally involved by observing the resultant rearrangement and, therefore, the explanation proposed is always the sim- plest one. Note also that loss of an entire chromosome can occur after breakage; thus, not all such losses come from nondisjunc- tion. Contrary to nondisjunction, however, breakage events cannot produce trisomies. Cytogenetic Detection of Structural Changes The question of how structural changes in chromosomes are detected may have arisen during the preceding discussions. Such mu- tants may be detected initially by cytological examination, or they may be noted first by their effects on the phenotype when genetic tests are made. Thus, detection and identi- fication of structural changes can be made cytologically, or genetically, or by a com- bination of both methods. When heterozygous, deficiencies can some- times be recognized genetically since they permit the expression of all genes which are hemizygous in the nondeficient chromosome. Inversions and translocations can be sus- pected when mutant heterozygotes show a marked reduction in offspring carrying cross- overs. Using appropriate genetic markers, inversion homozygotes show some genes in the reverse of normal order, whereas in heterozygotes or homozygotes for transloca- Struct iircd Changes in Chromosomes 175 tions genes normally not linked are found linked. Sometimes a cytological study is preceded by genetic studies indicating the class of structural change involved and the particular chromosome (s ) affected. Of course, detailed knowledge of the cytological appearance of the normal genome is a pre- requisite for such work. The prophase of meiosis of some organ- isms and the giant salivary gland chromo- somes of Diptera are particularly suited for cytological studies, because in both cases synapsis between homologs helps locate the presence, absence, or relocation of chromo- some parts. For example, inversion hetero- zygotes show either a reversed segment which does not pair with its nonreversed homologous segment (if the inversion is small), or (if the inversion is larger) show one homolog twisting in order to synapse ( Figure 1 2-7 ) . A deficiency-heterozygote will buckle in the region of the deficiency. Since a chromosome with a duplication may also buckle when heterozygous, careful cyto- logical study is needed to distinguish this case from deficiency (see Figure 12-8). Heterozygotes for reciprocal translocations (Figure 12-9) show two pairs of nonho- mologous chromosomes associated together in s) napsis. The present discussion should suffice as an introduction to the origin, nature, and consequences of the more common types of structural changes in chromosomes and to the methods used in identifying such mu- tants. i IGURE 12-9. Heterozygous reciprocal trans- location in corn (pachynema) (courtesy of M. M. Rhoades) and Drosophila (salivary gland) ( courtesy of B. P. Kaufmann ) . \ ) *^t*|^* £ I7(> « HAP1 EB 12 SUMMARY AND CONCLUSIONS Structural change in chromosomes is a type o! mutation involving the gain, loss, or relocation of chromosome parts. All such mutations require chromosome or chromatid breakage. Since proximitj favors union, most of the ends produced hy breakage restitute. Unions occur mainly during interphase. Noniestitutional unions produce structural changes ill chromosomes. The occurrence of one. two. or three nonrestituting hreaks in one or two chromosomes is discussed in relation to the production of whole- chromosome losses, deficiencies, duplications, inversions, translocations, shifts, and transpositions. Chromosomes that have undergone structural change may be euploid or aneuploid. The cells in which these mutations arise are euploid but can become aneuploid follow- ing mitosis, segregation, or crossing over. The structural changes most likely to be retained in the population are the smallest ones; those changes which directly or in- directly cause an increase in gene number are most likely to be important in evolution. REFERENCES Beam. A. G.. and German III, J. L., "Chromosomes and Disease," Scient. Amer., 205. No. 5:66-76, 1961. Muller. H. J.. 'The Nature of the Genetic Effects Produced by Radiation," in Radiation Biology, Hollaender, A. (Ed.), New York: McGraw-Hill, 1954, Chap. 7, pp. 351-473. Patau. K., "The Origin of Chromosomal Abnormalities," Pathologie-Biologie, 11:1163- 1 170, 1963. Russell, L. B.. "Chromosome Aberrations in Experimental Mammals," Progress in Medical Genetics, 2:230-294, 1962. QUESTIONS FOR DISCUSSION 12.1. The terms euploid and aneuploid (hypo- or hyperploid) have been applied both to individual chromosomes and to whole nuclei. Give an example of: (a) A hypoploid chromosome in a euploid nucleus (b) A hyperploid chromosome in a hyperploid nucleus (c) An aneuploid nucleus containing all structurally normal chromosomes 12.2. As used on p. 170, what does the term eutelomeric mean? Name two types of aneutelomeric chromosomes. 12.3. Given the chromosome AB/CDE/F.GHI/J, where the period indicates the centromere and the slanted lines the positions of three simultaneously produced breaks, draw as many different outcomes as possible. Indicate which one is most likely to occur. 12.4. In Drosophila, the loss of a given chromosome results in monosomy; this situa- tion is approximately three to five times as frequent as its gain, resulting in trisomy. Explain. 12.5. Discuss the origin of monosomies among human zygotes. Structural Changes in Chromosomes 111 12.6. In human chromosomes at mitotic metaphase, discuss the detectability of the following: (a) Paracentric inversion (c) Deficiency (b) Pericentric inversion (d) Duplication (e) Half-translocation 12.7. What advantages may inversion provide? 12.8. What characteristics of cells undergoing oogenesis favor the production and viable transmission of half-translocations? 12.9. In Drosophila, a male, dihybrid for the mutants bw and ,s7, when back-crossed to bw bw st st, normally produces offspring whose phenotypes are in a 1:1:1:1 ratio. On exceptional occasions, this cross produces offspring having only two of the four phenotypes normally obtained. How can you explain such an exception? 12.10. Is the telomere a gene? Why? 12.11. Explain how you could cytologically determine the position of the locus for white on the X chromosome of Drosophila by each of the following: (a) deficiencies of various sizes (b) inversions of various sizes (c) various reciprocal translocations 12.12. Suppose you had a self-maintaining strain of Drosophila in which all females were yellow-bodied and males, grey-bodied. How would you explain this con- sistency if the egg mortality were always 50%? Low, as it is normally? How would you test your hypothesis cytologically? 12.13. (a) Several X-linked mutants in Drosophila cause notched wings. One of these mutants is lethal in the male and also in the mutant homozygote female. How do you suppose such a homozygote is produced? (b) A female heterozygous for this mutant (N/+) is mated to a fa/Y male. In Fx all sons are normal, half the daughters are normal, and half are both notched and faceted. Explain this result showing how you might test your hypothesis. 12.14. Make a diagram of the different eucentric reciprocal translocations between autosomes 2 and 3 in Drosophila which you would expect to be lethal in the following cases: (a) when either half-translocation is present (b) when one half-translocation but not the other is present (c) under no circumstances 12.15. Does the absence of crossing over in male Drosophila facilitate the detection of heterozygous reciprocal translocations? Explain. 12.16. Given a Drosophila heterozygous for a eucentric reciprocal translocation be- tween chromosomes 2 and 3 and assume both half-translocations are lethal when present separately. Discuss the nature of the linkage maps one would obtain from mating (a) genetically marked females of this type with appropriately marked non- translocation males (b) genetically marked males of this type mated to appropriately marked non- translocation females 12.17. A chromosome A.BCDEEDCFG has a reverse repeat, or duplication, for CDE. Compare the stability of this chromosome with A.BCDECDEFG, which carries a tandem repeat, or duplication, for the same region. ITS ( II AIM IK 12 I 2. is. In Drosophila each ol the genes for curly wings (Cy), plum eye color (/'///). hairless ( // ) and dichaete wings ( /> ) are lethal when homozygous. A curly, han less male mated to a plum, dichaete female produces 16 equally frequent types ol sons and daughters. One curly, plum, hairless, dichaete F, son is irradiated \uth \-ravs ami then crossed to a plum, dichaete female. Three F2 sons phenotypically like the father, collected and mated separately with wild- type females, produce the following males and females in the F3 progeny. Phenotype Son I Son 2 Son 3 C\ H 140 120 76 ( \ D 120 81 Pm H 135 84 Pm D 154 117 79 Explain these results, using cytogenetic diagrams for all individuals mentioned. I2.1l>. (a) Discuss the frequency of abortions in normal mothers who produce half- translocational children with Down's syndrome. (b) Would you sometimes expect the occurrence of children with Down's s\ndrome to he correlated with the father but not with his age? Explain. 12.20. The Y chromosome is of different size in different phenotypically normal men. On the other hand, a woman with a small X chromosome is phenotypically defective. How can you explain the origin of such different Y and X chromo- somes and the difference in the way they affect the two sexes? Chapter *13 RADIATION-INDUCED STRUCTURAL CHROMOSOME CHANGES I n the preceding chapter struc- tural changes in chromosomes were discussed with respect to types and consequences, but little was said about the events responsible for their pro- duction, namely, breakage and cross-union. Chromosomes break spontaneously; that is. they occasionally break in cells exposed to normal conditions. Because spontaneous breakage is relatively rare, agents that are able to produce great numbers of breaks are very useful in studies of chromosome break- age and its consequences. Our discussion in this chapter is restricted to one of these agents, radiation. The process of breaking a chromosome is a chemical reaction requiring energy. The biochemical effect of radiation depends upon the type and amount of energy left in tissue. Less energetic radiations (such as visible light) leave energy in the form of heat; more energetic radiations (such as ultraviolet light) leave energy in the form of heat and activation; the latter type of energy makes an electron move from an inner to an outer orbit of an atom. The more ener- getic the radiation, the greater the likelihood that the energy absorbed will lead to chem- ical change. For example, ultraviolet light produces more breaks in chromosomes than does visible light. Radiations of energy higher than ultraviolet light (X rays and gamma rays; alpha and beta rays; electrons, neutrons, protons, and other fast-moving 179 particles) are even more capable of causing breaks. Although such high-energy radia- tions also heat and activate, most of the energy left in the cells is in the form of ioni- zation, and this leads to most of the chromo- some breaks. Before discussing how ioniza- tion energy leads to breakage, we should first understand what ionization is and what its consequences are. Like visible and ultraviolet light, X and gamma rays are electromagnetic waves; how- ever, they have relatively shorter wave lengths and can penetrate tissue more deeply than visible or ultraviolet light. When a highly energetic wave is stopped (or a fast- moving particle is captured or slowed down ) , energy is absorbed by the atoms of the medium. This energy can cause an atom to lose an orbital electron, creating a charged particle, or ion, by the process of ioniza- tion. Such an electron, torn free of the atom, goes off at great speed and can, in turn, cause other atoms to lose orbital electrons — to be ionized. All atoms losing an electron, of course, become positively charged ions, and atoms that capture free electrons become negatively charged ions. Since each electron lost from one atom is eventually gained by another atom, ions occur as pairs. In this way a track of ion pairs, or an ion track, is produced which often has smaller side branches. The length of the main or primary ion track and its side branches and the density of ion pairs differ with the type and energy of the radiation involved. Fast neutrons make a relatively long, rather uniformly thick ion track; fast beta rays or electrons make a relatively long, uniformly thin or interrupted track of ions; ordinary X rays make a relatively short track sparse in ions at its origin becoming only moderately dense at its end. It is sufficient to say that all known ionizing radiations pro- duce clusters of ion pairs within microscopic distances. In other words, no amount or 180 CHAPTER 13 kind of high-energy radiation presently known can produce only single ions, or single pairs of ions evenly spaced over microscopic (hence, relatively large) dis- tances. Since one ion or a pair sufficiently separated from the next does not exist, the genetic effects of ionization must be deter- mined from the activity of clusters of nega- tively and positively charged ions Ions undergo chemical reactions to neutralize their charge to reach a more stable con- figuration. It is during this process that ion clusters are able to produce chromosome and chromatid breaks (Figure 13-1). The amount of ionization produced by radiation is measured in terms of an ioniza- tion unit called the roentgen, or /• unit, one r being equal to about 1.8 X 10° ion pairs per cubic centimeter of air. A sufficiently 1 V .*' . 4 - ft %' V + *-*\ B * V. • ^ c'- figure 13—1. Structural changes X-ray-induced (75-150 r) in normal human male fibro- blast-like cells in vitro. Arrows show: (A) broken chromosomes, (B) translocation {cen- ter) and dicentric (lower left), (C) ring chromosomes. A, B are in metaphase (see Fig. 11-9); C is late prophase. (Courtesy of T. T. Puck, Proc. Nat. Acad. 5c/., U.S., 44: 776-778, 1958.) Radiation-Induced Structural Chromosome Changes 181 penetrating radiation (such as fast elec- trons), producing this 1.8 X 10!) ion pairs in a given cm:{ of air, can also produce this amount in successive cm'5 of air because only a very small fraction of the incident radiation is absorbed at successive depths. If not very energetic X rays are used ("soft" X rays of relatively long wavelength — also called Grenz rays), all radiation may be absorbed near the surface of the medium, keeping the deeper regions free from ionization. The amount of energy left at any level depends not only upon the energy of the incident radiation, but also upon the density of the medium through which the radiation passes. Thus, in tissue, which is approximately ten times as dense as air, a penetrating high- energy radiation produces about one thou- sand times the number of ion pairs per cm3 as it does in air. Knowing this, it can be calculated that one r (always measured in air) produces about 1.5 ion pairs per cubic micron (/a3) of tissue. Since the volume of the Drosophila sperm head is about 0.5 ju3, one r is able to produce, on the average, less than one ion pair in it. Since ions occur in clusters, one r may place dozens of ion pairs in one sperm head and none in dozens of other sperm heads. The r unit measures only the absorbed energy which produces ions; another unit, the rad, measures the total amount of radiant energy absorbed by the medium. In the case of X rays, about 90% of the energy left in the tissue is used to produce ions; the rest produces heat and excitation. Since ultraviolet radiation is non- ionizing, its dosage is measured in rads and not r units. The number of chromosome breaks pro- duced by X rays increases linearly with the radiation dose (r) (Figure 13-2). This re- lationship means that X rays always produce at least some ion clusters large enough to cause a break. Moreover, clusters of ions from different tracks of ions do not combine their effects to cause a break. (If there were co7 or 0Q5 - t- o (T2 10 20 30 40 50 60 70 80 90 100 110 120 130 DOSAGE IN R UNITS figure 13-2. The relation between X-ray dos- age and the frequency of breaks induced in grasshopper chromosomes. (Courtesy of J. G. Carlson, Proc. Nat. Acad. Sci., U.S., 27:46, 1941.) such cooperation between clusters, the break frequency at low doses would be lower than what has been found because of the waste of clusters too small to break; the frequency at higher doses would be higher because of the cooperation among such clusters.) Cer- tain radiations, like fast neutrons, produce fewer breaks per r than X rays because one r of these radiations produces larger — and, hence, fewer — clusters of ions than do X rays. These larger clusters more often ex- ceed the size needed to produce a break, and therefore, are relatively less efficient in this respect. Ion clusters can produce breaks either di- rectly by attacking the chromosome itself, or indirectly by attacking oxygen-carrying molecules (which, in turn, react with the chromosomes) or other chemical substances (which, in turn, affect the chromosome or oxygen-carrying molecules). In any case, this indirect pathway must be of nearly sub- microscopic dimensions; otherwise, differ- ent ion clusters would be able to cooperate in causing breakage. Thus, only ion clus- ters in or very close to the chromosome can produce breaks in it, as has been visibly demonstrated by using beams of penetrating L82 CHAPTER 13 radiation of microscopic diameter. Such a beam passing through a metapliase chromo- some can break it. but fails to do SO when directed at the protoplasm adjoining the chromosome. From what has been stated, it is reason- able to assume that the number of breaks produced by a given dose of a certain radia- tion depends upon the volume which a chro- mosome occupies. This volume is different at different times in the nuclear cycle (for example, it changes during chromosome replication). Because of variations in poly- nemy or gene activity, the same chromo- some can occupy different volumes in dif- ferent tissues of an individual and the vol- ume of the same sex chromosome can be different in the two sexes. Because break- age requires energy, it is also reasonable to assume that the number of breaks indirectly produced increases if, during irradiation, either the amount of oxygen is increased, or the cell's reducing substances are poisoned. And conversely, replacement of oxygen by nitrogen during irradiation reduces the num- ber of breaks produced. After this preliminary discussion of some of the factors that influence the production of radiation-induced breaks, we are ready to consider the factors that influence the fate of the ends produced by breakage. Just as breakage involves a chemical reaction, so does the union between two sticky ends. The joining of break-produced ends appar- ently involves adenosine triphosphate and protein synthesis.1 Joining is enhanced by the oxygen (and inhibited by the nitrogen) present after irradiation. Accordingly, resti- tution is prevented if nitrogen replaces oxy- gen after irradiation, thus increasing the time that ends from the same break stay open, and. therefore, the chance for cross- union when the supply of oxygen is later resumed. (Note that the presence of oxygen 1 See J. G. Brewen (1963). has two contrary effects on rearrangement frequency — during irradiation it increases the number iA' breaks, whereas after irradia- tion it increases restitution. ) Since, under given conditions, the num- ber of breaks increases linearly with an ion- izing dose — each part of the dose independ- ently producing its proportional number of breaks — clearly, the number of breaks pro- duced is also independent of the rate at which a given total dose is administered. It also follows that all structural changes in chromosomes resulting from single breakages are also independent of the radiation dose rate. Radiations such as fast neutrons which produce long and dense ion tracks can fre- quently induce two chromosome breaks with the same track. In this case, if the same chromosome — having folded or coiled tightly — is broken twice by being twice in the path of the track, then large and small structural changes of inversion, deficiency, and dupli- cation types can be produced. The fre- quency of these rearrangements increases linearly with fast neutron dose and is inde- pendent of the dose rate. A single fast neutron-induced track of ions can also break two different chromosomes when chromosomes are closely packed to- gether, as they are in the sperm head. The linear increase with dose in the frequency of reciprocal translocations obtained after sperm are treated with fast neutrons pro- vides evidence for concluding — as was done in Chapter 1 2 — that proximity of sticky ends favors their union. Such a linear dose-effect can be obtained only if both breaks are pro- duced by the same track and if the broken ends capable of exchange union are located near each other — broken ends produced by different tracks being too far apart. When ordinary X rays are employed, however, the clusters are smaller, and the track of ions is shorter than fast neutron tracks. Accordingly, two breaks in the same chromosome are produced by the same Radiation-Induced Structural Chromosome Changes 183 X-ray ion track less frequently, and if they do occur, they are usually quite close to- gether. Note, however, that two breaks oc- curring within submicroscopic distances in successive gyres of a coiled chromosome produce structural changes whose size ranges only from minute to small. Nevertheless, a small proportion of single X-ray tracks — in the treatment of sperm, for example — do cause two breaks, each in a different chro- mosome. Therefore, for X-ray doses that produce fewer than two tracks per sperm, gross chromosomal rearrangement frequency increases linearly with dose. So, there is actually no dose of X rays which does not have some chance of producing a gross re- arrangement. In other words, no matter how small a dose of ionizing radiation is received, the possibility of a chromosomal break and a gross chromosomal mutation always exists. In the case of X rays or fast electrons, two breaks that occur in the same nucleus usually result from the action of two ion clusters, each derived from a different, inde- pendently arising track, so that each break is induced independently. Fast electron or X-ray-induced, two-break gross rearrange- ments of this origin are dose dependent, for when a small enough dose is given, a nucleus is traversed by only one track, and only one-track — not two-track — gross rearrange- ments can result. But when the dose is large enough for a nucleus to be traversed by two separate tracks, the two breakages required for two-break gross rearrangements can be produced independently. Therefore, the higher the dose of X rays used, the greater the efficiency in producing multi-break gross rearrangements caused by breaks independ- ently induced by separate tracks. Accord- ingly, for doses causing some cells to experi- ence two such independently produced breaks and higher doses, the frequency of these mutations increases more than in di- rect proportion to the amount of dose. One example is the exponential rise in the fre- quency of reciprocal translocations obtained after treating sperm in inseminated Dro- sophila females with increasing dosages of fast electrons (Figure 13-3, curve T). X-ray-induced rearrangements involving two (or more) breaks induced by separate tracks also depend upon the rate at which a given dose is administered. When a suit- ably large dose is given over a short interval, i i i i — i — i — r 10 14 18 22 26 30 34 38 DOSE IN RADS(XIOO) figure 13-3. Percentage of mutations, ±2X standard error, recovered from Drosophila sperm exposed to different dosages of 18 mev electrons. The sex-linked recessive lethal fre- quencies (L) are joined by solid lines and are adjusted for the control rale; sex chromosome loss frequencies (5) are connected by broken lines and are corrected for the control rate; reciprocal translocation frequencies (T) be- tween chromosomes II and III are connected by dot-dash lines. (From I. H. Herskowitz, H. J. Midler, J. S. Laugh lin. Genetics, 44:326, 1959.) 184 ( HAPTER 13 the ends produced h\ separate breaks exist simultaneously and are able to cross-unite. But when the same dose is given more slowly, the pieces of the fust break ma\ restitute before those o\ the second arc pro- duced, thus eliminating the opportunity for cross-union. In this event, the same dose produces fewer gross rearrangements when given in a protracted manner than when given in a concentrated manner. Although this dose-rate dependence for X rays is true for most cells — at least during part of the interphase stage — it does not apply to ma- ture sperm of animals, probably including man. In these gametes and during most of nuclear division in other cells, the broken pieces cannot join each other (sec p. 166) and. therefore, accumulate. For this rea- son, it makes no difference how quickly or slowly the dose is given to the chromo- somes in such a sperm head, since the breaks remain unjoined at least until the sperm head swells after fertilization. As already mentioned, the spatial arrange- ment of chromosomes with respect to each other influences the number of breaks and the kinds of structural changes they produce. It should be noted that the possibilities for multiple breakages and for joinings are quite different for chromosomes packed into the tiny head of a sperm than they are for chro- mosomes located in a large nucleus. But even within a given type of cell, a number of other factors can influence breakage or rejoining, such as the presence or absence of a nuclear membrane, the degree of spiral- ization of the chromosomes, the stress or tension under which the parts of a chromo- some are held, the degree of hydration, the amount of matrix in which the genes are embedded, protoplasmic viscosity and the amount of fluid and particulate movement around the chromosomes, gravity, centrif- ugal force, and vibration. In cells whose chromosomes have just replicated and in somatic or meiotic cells where homologS arc synapscd, a special re- striction on the movements of the pieces is produced when only some of the apposed strands are broken (see p. 166). In this situation, the forces which keep parts of one strand adjacent to the corresponding parts of its sister or homolog may prevent the broken pieces from moving apart freely, so that the unbroken strand or strands serve as a splint for the broken one(s) and reduce the opportunities for cross-union. Many factors exist, therefore, which determine to what degree chromosome and chromatid fragments can move or spring apart; those affecting the distances between different chromosomes or the parts within a chromo- some also affect chromosome and chromatid breakability. The frequencies and types of structural changes depend also upon the total amount of chromosomal material present in the nu- cleus and the number and size of the chro- mosomes into which this material is divided. The rearrangements that occur in different cells of a single individual depend upon whether the cell is haploid, diploid, or poly- ploid, and whether or not the chromosomes are polynemic, are in the process of replica- tion, or are otherwise metabolically active. Radiation can produce important non- mutating effects upon the chromosomes by damaging nonchromosomal cellular compo- nents which, in turn, affect chromosomal be- havior and function. If the cells are capable of repairing such nonchromosomal, struc- tural or functional damage, they will have a longer time in which to repair when a radia- tion dose is given slowly than when given quickly. The most obvious example is the effect of radiation upon mitosis (and prob- ably meiosis). Cells at about midprophase or a later stage in nuclear division usually complete the process even though irradiated. Cells no farther advanced than about mid- prophase often return to interphase when irradiated. For this reason, ionizing radia- Radiation-Induced Structural Chromosome Changes 185 tion causes a greater degree of synchromy in division than occurs in the absence of the radiation. Accordingly, starting with a pop- ulation of cells in various stages of nuclear division, the chromosomal targets for muta- tion are different in the later stages of re- ceiving a protracted dose and of receiving a concentrated dose. The capacity to produce recoverable struc- tural changes is not the same in euchro- matic and heterochromatic chromosomal regions. Recovered radiation-induced struc- tural changes involve heterochromatic regions more frequently than they do euchromatic ones. It has not been determined whether this excess is due to heterochromatin hav- ing a greater breakability, a lesser resti- tutability (which might be associated with the general ability of different heterochro- matic regions to synapse with each other), or both. Nevertheless, in many rearrange- ments, at least one of the points of breakage is located in the heterochromatin nearest the centromere. This is one reason why whole- arm reciprocal translocation is the type most frequently observed. The present discussion has been motivated by the ability of energetic radiations to in- duce many breaks and, subsequently, many structural changes. The great supply of re- arrangements readily available via radiation treatment has made it possible to discover many of the factors influencing breakage and joining. Many other important discoveries were made possible by the study of structural changes, including 1 . The genetic basis of the centromere 2. The reduced incidence of crossing over near the centromere 3. The genetic basis of the telomere 4. The existence in some species of ge- netic elements (collochores) near the centromere of special importance to synapsis. - ^ See K. W. Cooper (1964). X CHROMOSOME cv ct y sc br pn wrb Centromere H CHROMOSOME 0 / I0\20 30 40^50 17? ^^ \ \ ,60-^70^80 90 100 107 al dp 7 v" b Bl/\cn vg c It tk P* sp Centromere IE CHROMOSOME Q__jg_Aig__3Q_l40?t ,5Q^60 70 80 90 IQO D th cu sr e Centromere figure 1 3-4. Comparison of chromosome (hollow bar) and crossover (solid bar) maps in D. melanogaster. Perhaps the most fundamental contribution was the finding, via structural changes, that the genes have the same linear order in the chromosome, that is, in chromosome maps, as they have in crossover maps. The spac- ing of these, however, is different in the two cases (Figure 13-4). Thus, for example, because of the reduction in crossing over near the centromere, the genes nearest the centromere — spaced far apart in the meta- phase chromosome map — are found to be close together in the crossover map. Although our subject matter has been re- stricted to the factors influencing the origin and joining of breaks produced by ionizing radiation, these factors are expected to oper- ate on breaks produced by any other spon- taneously occurring or induced mechanism. For. in general, no matter how broken chro- mosomes are produced, all possess the same properties. 186 CHAPTER 13 SUMMARY AND CONCLUSIONS The components ol structural chromosome change, breakage and cross-union, are readilj studied through the use ol ionizing radiations. These radiations induce break- aye in chromosome strands primarily by the clusters of ion pairs they produce. These clusters form tracks ol ions whose thickness and length determine the number and loca- tion ol the breaks, hacks ol ions must occur very close to. or within, the chromosome lh.it the) break. The number of breaks increases linearly with radiation dose. Whether they result from one or from two breaks, all chromosomal rearrangements induced by a single ionizing track increase linearly with radiation dose, have no threshold dose, and show no effect from protracting or concentrating the dose. Accordingly, there is no dose of ionizing radiation which does not produce breaks and at least single-track-induced rearrangements. Two-or-more-break structural changes produced by ion clusters in separate, inde- pendently-occurring tracks increase in frequency faster than the amount of dose and do have a threshold dose. If joining of chromosome ends produced by breakage can take place during the course of irradiation, such rearrangements are reduced in fre- quency by protracting the delivery of the total dose. Since both the breakage and joining processes involve chemical changes, their fre- quencies can be modified by the metabolic state of the cell. All types of rearrange- ments are expected to be affected by: the physical and chemical state of the chromo- some and the amount and distribution of its euchromatin and heterochromatin; by the number and arrangement of the chromosomes present; by the presence or absence of a nuclear membrane; and by the movements of broken ends as influenced by cellular particles, fluids, and extracellular factors. REFERENCES Bacq, Z. M., and Alexander. P., Fundamentals of Radiobiology , 2nd Ed., New York: Pergamon Press, 1961. Bender, M. A., and Gooch, P. C, "Types and Rates of X-ray-Induced Chromosome Aberrations in Human Blood Irradiated in Vitro," Proc. Nat. Acad. Sci., U.S., 48:522-532, 1962. Brewen, J. G.. "Dependence of Frequency of X-ray-Induced Chromosome Aberra- tions on Dose Rate in the Chinese Hamster," Proc. Nat. Acad. Sci., U.S., 50:322- 329, 1963. Chu, E. H. Y., Giles, N. H., and Passano, K., "Types and Frequencies of Human Chromosome Aberrations Induced by X-rays," Proc. Nat. Acad. Sci., U.S.. 47: 830-839. 1961. Cooper, K. W., "Meiotic Conjunctive Elements Not Involving Chiasmata.' Proc. Nat. Acad. Sci.. U.S.. 52:1248-1255. 1964. "Ionizing Radiation," Scicnt. Amer., 201 :No. 3 (Sept.), 1959. Muller. H. J.. "General Survey of Mutational Effects of Radiation," in Radiation Biolo^x and Medicine, Claus, W. D. (Ed.), Reading, Mass.: Addison-Wesley, Chap. 6, pp. 145-177. 1958. Puck, T. T., "Radiation and the Human Cell," Scient. Amer., 202, No. 4:142-153, 1960. Sobels, F. H.. Repair from Genetic Radiation, New York: Pergamon Press, 1963. Sparrow, A. H.. Binnington. J. P., and Pond, V., Bibliography on the Effects of Ionizing Radiations on Plants, 1896-1955, Brookhaven Nat. Lab. Publ. 504 (L-103), 1958. Radiation-Induced Structural Chromosome Changes 187 Lewis John Stadler (1896-1954) is noted for his studies on the nature of mutation and of the gene (see p. ix). He and H. J. Muller discov- ered independently the mutagenic effect of X rays. {From Genetics, vol. 41, p. 1, 1956.) QUESTIONS FOR DISCUSSION 13.1. After both are exposed to the same amount of radiation why should tissue, which is only about ten times as dense as air, contain about one thousand times more ions than air? 13.2. What evidence can you give to support the view that the ions causing breakage need not always attack the chromosome directly? 13.3. Does the observation that the volume of a chromosome is variable under dif- ferent conditions mean that it has an inconstant gene content? Explain. 13.4. Do you suppose that chromosomes exposed to X rays are more likely to under- go structural change when they are densely spiralized than when relatively un- coiled? Why? 13.5. Discuss the role of heterochromatin in changes involving chromosome number and chromosome shape. 13.6. Do you suppose that the oxygen content of a space capsule can affect the mutability of Drosophila passengers? Explain. 13.7. Discuss the relative efficiency, per r, of small doses of X rays and of fast neu- trons in producing structural changes in chromosomes. 13.8. Do you suppose that the mutability of ultraviolet light threatens man's survival? Explain. 1SS CHAPTER 13 ] 3.9. Compare the number and fate of breakages induced b) the same dose of X rays administered to: (a) A polyploid and a diploid liver cell m man (h) A diploid neuron in man and Drosophila (e) A sperm and a spermatogonium in man 13.10. Discuss the importance of the nonmutant effects ol a given dose of radiation upon the mutation frequency induced b\ a subsequent radiation dose. 13.11. Using Figure 13-4. discuss the likelihood of crossing over in different regions of the X chromosome of D. melanogaster, 13.12. Compare the roentgen unit with the rad unit. 13.13. What specific aspects of our present environment tend to reduce the number of mutations induced by penetrating radiations from the number induced when man first evolved? Chapter *14 POINT MUTATIONS W: e have already found that the mutational unit of the genotype may be a whole genome, a single chromosome, or a part of a chromosome. Perhaps a study of these units will reveal more about the mutational characteristics of a single gene; perhaps the recombinational properties of individual genes will illuminate this area of investiga- tion. Let us consider what we already know about the mutation of single genes — the class of mutation that is probably the most impor- tant in evolution because it causes the small- est shift in gene balance. All chromosomes are linear and un- branched whether or not they have under- gone segmental rearrangement by crossing over or breakage. This linear arrangement could be due to the linkage of gene to gene directly, or indirectly by a nongenetic ma- terial which connects adjacent genes. In either case, the fact remains that a chromo- some is invariably either a rod or a ring, providing almost conclusive evidence that a gene cannot be joined to other genes at more than two places, and that a mutation which permits a gene to be joined to more than two others cannot occur spontaneously or be in- duced. That this type of mutation is never observed regardless of the organism studied can be interpreted to mean that genes never had this property or that all existing genes have lost this property. We are led to con- clude, therefore, that all interstitial genes are bipolar, and that mutation is incapable of 189 causing the gene to be more than bipolar. After chromosome breakage, the "stick- iness" of the new ends is evidence that al- most all mutations retain the bipolarity of genes. In some relatively rare cases, how- ever, break-produced ends (broken ends) are known to become permanently unsticky or healed, so that mutation from bipolarity to unipolarity does occur. That mutation can change genes from a bipolar to a unipolar type, or vice versa, is evidenced also by the presence of telomeres — unipolar genes which seal off the normal ends of chromosomes. The chromosomal change from bipolarity to unipolarity occurs regularly in the life history of certain animals. In particular species of the roundworm Ascaris, for ex- ample, nuclei which remain in the germ line have a single pair of chromosomes. When the nuclei first enter the somatic line, how- ever, these chromosomes break up into a number of small linear fragments whose ends are sealed off and behavior during mitosis is normal — normal mitotic behavior being possible because a germ line chromosome has numerous centromeres along its length (each surviving fragment of the chromosome in a somatic cell has at least one). In the germ-line polycentric chromosome all centro- meres but one are suppressed. Because chromosome fragmentation in Ascaris takes place only in somatic cells, these polarity changes can be attributed to some physiolog- ical difference between cells entering the so- matic line and cells remaining in the germ line. These polarity changes should be con- sidered recombinational rather than muta- tional events because the changes from bipolarity to unipolarity are numerous, si- multaneous, and normal — therefore lacking the novelty of mutations. Although mutations which change polarity from bipolarity to unipolarity apparently oc- cur, no unambiguous case has ever been reported of the reverse, that is, of a muta- 190 CHAPTER 14 tion from unipolarity to bipolarity. Since the chance of detecting and proving a change from uni- to bipolarity is verj small indeed, the occurrence of such a change cannot, at present, be denied with certainty. Do muta- tions to nonpolaritv occur'.' It is evident that a unipolar or bipolar gene that mutates to a Donpolar alternative must necessarily drop out of the chromosomal line-up. If this hap- pens, the freed, not-at-all-sticky gene will not he linked to any chromosome. Since no evidence has yet been presented for the existence of genetic material liberated from its chromosomal locus in this way, we cannot give an affirmative answer at this point. The gene was first identified in sexually reproducing individuals whose chromosomes synapse during meiosis. Synapsis results from the attraction between different seg- ments of one or more chromosomes. That different degrees of specific attraction exist between genes is illustrated by the fact that genes located in heterochromatin synapse much less specifically than those found in euchromatin. Specific genes (such as one in maize called asynaptic) are known which not only lack synaptic attraction for their alleles but also destroy this attraction be- tween pairs of genes at other loci, or cause general desynapsis. The occurrence of col- lochores — genes which assist in pairing — has already been mentioned in Chapter 13. Corresponding euchromatic loci located in homologous chromosomes synapse with each other whether or not the particular alleles contained are identical or different. Yet euchromatic genes in nonhomologs do not usually synapse with each other, although it is presumed that some presently nonallelic genes were previously allelic. Consequently, mutation must be capable of changing the synaptic specificity of a gene; and it must follow, at least in a general way, that iden- tical genes attract each other more than non- identical ones. Since at least some genes have multiple alleles, it is clear that different forms of a gene do exist, and mutations of such genes are not explicable merely in terms of their complete loss or inactivation. Since some mutations produce no visible change in the handing pattern of salivary gland chromo- somes of Drosophila, mutations involving but a single gene, that is, gene mutations, can be submicroscopic. At present, we can only detect gene mutations by the pheno- typic changes they produce. Consequently, the characteristics of gene mutation must be determined from the phenotypic changes produced by recombinationally detected genes. Accordingly, we are unable to de- termine from such phenotypic changes whether gene mutation involves the recom- binational gene in toto, a one portion or site within it. or many different sites within it. If gene mutation involves a change in the entire gene, then the material composition of the genes detected by recombination and by mutation would be identical. If, on the other hand, the recombinationally detected gene contains one or more sites at which mutation can occur, the basic recombina- tional unit of genetic material would be larger than the basic mutational unit. Until such time as critical evidence is obtained to the contrary, we have no choice other than con- tinuing to accept the mutational and recom- binational genes as materially equivalent, an assumption (Chapter 3, p. 36) which is in accord with the law of parsimony. As mentioned in the first chapter, any given gene is rather stable, having been faith- fully replicated many thousands of times be- fore a detectable mutation occurs. The greater the sensitivity of our tests for de- tecting mutations, however, the larger is the frequency of mutation observed (recall the detection of isoallelism, p. 59). It is rea- sonable to assume therefore that transmis- sible modifications of single genes do occur which escape our present means of detection. Nevertheless, within the limits of our present Point Mutations 191 methods of analysis, the gene appears to be a very stable entity. Consider the following method for obtain- ing information with regard to gene muta- tion. All mutants involving the one or more genes being investigated are collected and then analyzed. Some mutants involving a given locus prove to be based upon changes in the number of whole chromosomes; others prove to be associated with gross or small chromosomal rearrangements. All these mutants are eliminated from further consid- eration even though gene mutation may also have occurred. Insofar as feasible, all ge- netic and cytological tests known are applied to eliminate mutants involving the minutest chromosomal rearrangements including, for example, tiny duplications or deficiencies. All, or a considerable number, of the mu- tants remaining can then be assumed — for lack of evidence to the contrary — to have resulted from mutations involving either a single gene (gene mutations) or at most only a few genes (intergenic mutations). Each of the remaining mutants behaves as though it resulted from a change at a single point in the crossover and cytological maps and is, therefore, called a point mutant. Since at this point we have no criterion for differentiating between a mutant involving only one gene (including its complete loss) and one involving a few genes, the entire heterogeneous class of point mutants will have to be studied in the hope of revealing some of the characteristics of gene mutation. Consider some of the characteristics of spontaneous and induced point mutations. Since point mutation of a vast number of different genes occurs, this process is not restricted to a very limited type of gene. Although the conditions causing point muta- tion might be of such a nature that, in the diploid cell, both members of a pair of alleles tend to mutate at the same time, actually the evidence is that only one gene of the pair is affected. Because only one member of a pair of genes in a nucleus mutates, point mutation must be a very localized, submicro- scopic event. If point mutation usually involved either a series of stable gene changes or an in- stability of the gene extending over more than one cell generation, the resultant mu- tants would usually occur in clusters and within a cluster the same gene might not always mutate to the same allele. But many point mutants occur singly. Moreover, those which do appear in a cluster often seem to be identical. Such a cluster can usually be explained by assuming a single cell has undergone mutation, having divided a number of times before the tests to detect the mutants were performed. Although such data do not prove that point mutation is instantaneous, they indicate that it is usu- ally completed within one cell generation and the change in this respect is quick more than it is gradual. The number of point mutations obtained from X-ray or ultra- violet ray treatments is reduced, however, if posttreatment with certain types of visible light or chemical substances is given im- mediately (but not if such treatment is post- poned for some hours). Such immediate posttreatments produce photo- or chemo- recovery from point mutation and prove that the point mutation process often does not occur or is not completed for some minutes. Certain chemical changes, which themselves may or may not be mutational, can lead to other, genetic changes such as breakage. If the first changes are repaired before they can induce the second, an apparent recovery from mutation is observed. Only after the point mutation process is completed is the new genetic alternative just about as stable as the old. Because point mutants are just about as stable as their parent genes or other genes in the genotype, it does not necessarily mean that all allelic and nonallelic genes have the same spontaneous mutation frequency. 192 CHAPTER 14 Study o\ a representative sample ol specific loci in Drosophila reveals an average ol one point mutation at a given locus in each 200,- 000 germ cells tested. In mice the per locus frequency is about twice this, or one in 100,000. In man. by scoring the mutants detected in heterozygous condition, the per locus rate is found to be one per 50.000 to 100,000 germ cells per generation. Within a species, different loci have about the same order of mutability. Even though some genes are definitely more mutable than others, the average spontaneous point muta- tion rate per genome per generation can be estimated for Drosophila, mouse, and man. In one Drosophila generation, one gamete in twenty (or one zygote in ten) contains a new detectable point mutant. In mice, this frequency is about one in ten gametes, whereas in man it is about one in five gam- etes (or two in five zygotes). The point mutations which occur spon- taneously— that is, under natural conditions — bear no obvious relation to the environ- ment, either with respect to the locus af- fected or the type of alternative produced. Modifications in the environment do, how- ever, influence point mutation frequency. For example, in the range of temperatures to which individuals are usually exposed, each rise of 10 C produces about a fivefold increase in point mutation frequency. The magnitude of this increase is similar to, al- though somewhat greater than, that obtained with an increase in temperature in ordinary chemical reactions. Violent temperature changes in either direction produce an even greater effect upon point mutation fre- quency. Actually, detrimental environmen- tal conditions of almost any kind increase point mutation frequency. Physical and chemical agents which raise the mutation frequency enormously are called mutagens. All high-energy radia- tions (see Chapter 13) are mutagenic (see Figure 13-3) as are many highly reactive chemical substances including: mustard gas and its derivatives; peroxides; epoxides; and carbamates. The point mutation frequencies obtained with radiation and certain chemical mutagens can be 150 times the spontaneous frequency. One speaks of a "spectrum of spontaneous point mutations in that, as men- tioned, certain loci are normally somewhat more mutable than others. The loci affected and the types of mutant alternatives pro- duced by ionizing radiation are not radically different from those involved in spontaneous mutation. That these radiations produce a mutational spectrum much like the sponta- neous one is expected, since radiant energy is more or less randomly distributed in the nucleus and generally enhances many differ- ent kinds of chemical reactions. The point- mutational spectra for different chemicals are somewhat different from each other as well as from the spectra induced by radia- tion mutagens or by spontaneous factors. These differences can be attributed to the nonrandom penetration of these chemical substances into the nucleus, or to their spe- cific capacities for combining with different nuclear chemicals, or both. Nevertheless, the frequency of point mutation, which in- creases linearly with the dose of ionizing radiation (although the frequency is in- fluenced by the amount of oxygen present), probably also increases linearly with the nu- clear dose of many different chemical muta- gens. So point mutation probably has no threshold dose with chemical mutagens, and the number of point mutations produced by a given total dose is constant, other things being equal, regardless of the rate of de- livery. For ultraviolet light — which is not a highly energetic radiation — the situation is differ- ent. Here the probability for the individual unit or quantum of energy inducing point mutation is considerably less than 100 per cent. Moreover, because several quanta can cooperate to produce mutation, ultraviolet Point Mutations 193 induced point-mutation frequency increases taster than linearly with dose — at least for low doses — and an attenuated dose is less mutagenic than a concentrated one. Point mutation is not restricted to the genes of any particular kind of cell, occur- ring in males and females, in somatic tissues of all kinds, and in the diploid and haploid cells of the germ line. Later stages in gam- etogenesis and very early developmental stages — perijertilization stages — are found to be relatively rich in spontaneous point mutations. Despite very great differences in life span, one does not find correspondingly great differences in the spontaneous germ line mutation frequencies of flies, mice, and men. This similarity in mutation frequency is not surprising if most of these mutations occur in the perifertilization stages, since each of these organisms spends a comparable length of time in these stages. Still another similarity among these species is the com- parable number of cell divisions required for each to progress from a gamete of one gen- eration to a gamete of the next. In fact, the differences in mutation frequency for these organisms are approximately proportional to the differences in the number of germ cell divisions per generation. When during the history of the gene does mutation occur? The finding that the point mutation frequencies in Drosophila, mouse, and man are proportional to the number of cell divisions they undergo suggests that some of these mutations occur at synthesis of the new gene, although the experimental results do not specify whether it is the old or the new gene that mutates. Aging of spermatids and sperm of Drosophila is known to in- crease the point mutation frequency. Since the viability of these cells is not impaired when aneuploid, the increase in point muta- tions may be due to an effect upon the old, physiologically quiescent gene, implying that point mutational changes can occur while a gene is linearly attached to its neighbors. The larger number of mutations obtained from aged cells may also be explained as resulting from a mutagen accumulated over a period of time which acts on the old or the new gene once gene replication is re- sumed. The possibility also remains that changes can occur in the steps leading to gene synthesis — before the new gene is com- pleted and attached to its linear neighbors; such changes could be scored later as point mutants. Phenotypic Effects of Point Mutants The biological fitness of a mutant gene — pure or hybrid — is best described in terms of its effect upon the organism's ability to produce surviving offspring, that is, upon re- productive potential. This potential includes the mutant-carrying individual's capacity to reach the reproductive stage and its fertility and fecundity during this period, as well as the viability of its offspring until sexual ma- turity. Although each mutant has many phenotypic effects, point mutants with small phenotypic effects occur much more fre- quently than those with large effects. For instance, pure (homo- or hemizygous) mu- tants which lower the viability of males with- out being lethal are at least three to five times more frequent than those which are recessively lethal (Figure 13-3). The vast majority of point mutants have a detrimental effect on the reproductive potential; beneficial point mutants are extremely rare. In terms of the past evolutionary history of a species, it is under- standable that in the great majority of cases, mutants affecting a trait or organ cause its degeneration. All the genotypes in a species have been subjected to selection for many generations, those producing the greatest re- productive potential having been retained. Although point mutation at any locus is a rare event, many of the possible alternatives for each gene must have occurred at least several times in past history. Of these 194 CHAPTER 14 alternatives, only the more advantageous alleles were retained, and these are the ones found in present populations. So. a point mutation today is likely to produce one of the genetic alternatives which occurred also in the past but had been eliminated beeause ol its lower biological fitness, that is, its lower reproductive potential. It should be realized, moreover, that reproductive poten- tial is the result of coordinated action of the whole genotype. The genotype may be likened to the machinery that makes modern automobiles — the automobile representing the phenotype — with the environment fur- nishing the necessary raw materials. Pres- ent genotypes, like the machines that manu- facture automobiles, are complex and have had a long evolutionary development. The chance that a newly-occurring point muta- tion will increase reproductive potential is just as small as the chance that a random local change in the present machinery will result in a better automobile. The differences between the phenotypic effect of a point mutant and its normal alter- native can be studied by adding more repre- sentatives of the mutant allele to the geno- type and examining the effect. In Drosoph- ila, for example, the normal fly has long bristles when the normal, dominant gene bb + is present. A mutant strain has shorter, thinner bristles because of the recessive allele bb (bobbed bristles), which — it should be recalled — has a locus both in the X and the Y chromosomes. We might suppose that the male, or female, homozygous for bb has bobbed bristles because this allele results in thinning and shortening the normal bristle. Since otherwise-diploid XYY males and XXY females can be obtained which carry three bb alleles, according to this view, one would expect the bristles formed to be even thinner and shorter than they are in ordinarv mutant homozygotes. But, on the contrary,1 3M 1 • DOSAGE OF GENES figure 14-1. The relationship between dos- age of normal and mutant genes and their phenotypic effect. in the presence of three representatives of bb — that is, three doses of bb — the bristles are almost normal in size and shape. This finding demonstrates that bb functions in the same way as bb+ does, but to a lesser degree. Mutants whose effect is similar but less than the normal gene's effect are called hypomorphs. Many point mutants are hy- pomorphs, since, in the absence of the nor- mal gene, additional doses cause the pheno- type to become more normal. Of the remaining point mutants, most are amorphs; these produce no phenotypic effect even when present in extra dose. One ex- ample is the gene for white eye (w) in Dro- sophila. Some mutants, neomorphs, produce a new effect — adding more doses of a neomorphic mutant causes more departure from normal, whereas adding more doses of the normal alternative has no effect. The relationship between the normal, wild-type gene and its hypomorphic mutants is indicated diagrammatically in Figure 14-1.1' The vertical axis represents pheno- typic effect; the normal, wild-type effect is 1 As shown by C. Stern. -Adapted from H. J. Muller. Point Mutations 195 indicated by -f . The horizontal axis refers to the dosage of either the normal gene or a hypomorphic mutant. Notice that a sin- gle + gene itself produces almost the full normal phenotypic effect (and often the dif- ference between its effect and the normal effect is not readily detected) . Two -f genes reach the wild-type phenotypic level. In the case of the hypomorphic mutant, however, even three doses may not reach the pheno- typic level produced by one -(- gene (recall the discussion oibb). Note also that genetic modifiers or environmental factors, which can shift the position of the genes on the horizontal axis and thereby shift the pheno- typic effect, have a decreasing influence as one proceeds from individuals carrying only one dose of mutant toward individuals carry- ing two + genes. Natural selection would clearly favor alleles that result in phenotypic effects close to wild-type — that is, near the curve's plateau — for such alleles assure phenotypic stability. Any mutant which produced such a phenotypic effect would, in the course of time, become the normal gene in the population and would automat- ically be dominant when heterozygous with a hypomorphic gene alternative. This model illustrates how the heterozygote with one + and one mutant gene has practically the same effect as the normal homozygote, and it seems to best explain most cases of com- plete or almost complete dominance. Since the normal gene alternative already pro- duces a near-optimum phenotypic effect, this scheme also illustrates why, other things be- ing equal, so few mutants are beneficial. Although it is understandable from the preceding discussion that hypomorphic and amorphic mutants are usually detrimental when pure, one may still wonder what effects these mutants have when heterozygous with the normal gene. If the mutant is an amorph, the mutant heterozygote can fall short of producing the wild-type phenotypic effect and, therefore, such mutants are ex- pected to be sJightly detrimental when het- erozygous. Hypomorphs are expected to be less or not at all detrimental when hetero- zygous, at least with respect to the trait for which they are classified as hypomorphic. But since each gene affects many different biochemical processes, a mutant hypo- morphic in respect to one trait may be amorphic in respect to another. In Dro- sophila, for example, the normal allele apr+ which results in dull-red eye color also pig- ments the Malpighian tubules. One of its alleles, apr, causes a lighter eye color (being hypomorphic in this respect) but no color in the Malpighian tubules (being amorphic in this respect). Experience confirms the expectation that most "recessive" lethal point mutants — these are lethal when homozygous — also have some detrimental effect on reproductive po- tential when heterozygous. Such mutants are not completely recessive, therefore, and when heterozygous in Drosophila cause death before adulthood in about two per cent of individuals. Usually mutants which are detrimental but not lethal when pure also show a detrimental effect when heterozy- gous; this effect is somewhat less than that produced by heterozygous recessive lethal point mutants. The principles of phenotypic action discussed here are expected to apply both to spontaneous and to induced point mutants. Detection of Point Mutants in Drosophila We have already mentioned the existence of genetic methods for collecting point mutants. Let us now consider in some detail one ele- gant procedure ;; employed in Drosophila melanogaster for this and other purposes. The commonly-used technique for detect- ing recessive lethals is called "Base" (see Figure 14-2) and was designed 4 to discover such mutants arising in the male germ line, :i Invented by H. J. Muller. 4 To replace the old "C1B" method. L96 ( II M'TER 14 son, however, carries this mu- tant in hemizygous condition and usually dies before adulthood, so that no wild-type sons appear in F>. It becomes clear, then, since an Fi female is formed by fertilization with a wild-type X-carrying sperm, the ab- sence of wild-type sons among her progeny is proof that the particular Py sperm carried a recessive lethal, X-linked mutant. Such a lethal mutant must have occurred in the germ line after the fertilization that produced the Pi male; he would not have survived had it been present at fertilization. It is unlikely that many of the X-linked lethals detected in sperm originate very early in development, for in this case a large por- tion of the somatic tissue would also carry the lethal and usually cause death before adulthood. Even when a few hundred sperm from one male are tested, only one is usually found to carry an X-linked recessive lethal mutant. This indicates that most X-linked lethals present in sperm involve only a very small portion of the germ line. Occasion- ally, however, the mutation occurs early enough in the germ line so that several sperm tested from the same male carry what proves to be the same recessive lethal. When a thousand sperm from normal, un- treated males are tested for X-linked reces- sive lethals by means of a thousand separate matings of Fj females, approximately two of these matings are found to yield no wild- type sons. This X-linked recessive lethal mutation frequency of 0.2% is fairly typical in D. melanogaster. For every 1000 r of X rays to which the adult male is exposed, approximately 3.1% more sperm are found to carry X-linked recessive lethals (see Fig- ure 13-3, for the similar frequency obtained after exposure to fast electrons). When used as described, the Base tech- nique detects only those recessive lethals which kill before adulthood. Other reces- sive lethals that produce wild-type adult males which are sterile or die before they can mate are not detected. No recessive lethals are detected unless they are hemi- zygous in the F2 male, as mentioned. Since a considerable number of X-linked mutants whose hemizygous lethality is prevented by genes normally present in the Y chromo- some is known to occur, this group is missed because each F2 male is normally provided with a Y chromosome. Suitable modifications of the Base procedure can be made to detect this special kind of Y-sup- pressed recessive lethal. On the other hand, the advantages and applications of the Base technique as described are numerous. For example, the presence or absence of wild-type males in F2 is easily and objec- tively determined. Since the recessive lethal detected in F2 is also carried by the hetero- 198 CHAPTER 14 zygous-Bar F2 females, further study of the sive lethal is possible in FL. and subse- quent generations. Such studies reveal that certain lethals are associated with intergenic changes; Lethals not associated with inter- genic changes are designated as recessive lethal point mutants. The Base technique can also be used to deteet recessive lethals that occur In a P, Base chromosome, the absence of Base males among the F2 prog- eny indicating such a mutation. Moreover, if the environmental conditions are standard- ized, it becomes possible to detect hemi- zygous mutants which either lower the via- bility of the F2 males without being lethal or raise their viability above normal. The opportunity for studying the viability effects of recessive lethals in heterozygous condition is also provided by this technique. Although the Base technique can also be used to detect X-linked mutants producing a visible morphological change when hemi- zygous, all those "visibles" which are also hemizygous lethals are missed. The "Maxy" technique 5 overcomes this difficulty. In this method, the tested female has fifteen X- linked recessive point mutants on one homo- log and their normal alleles on the other. -See H. J. Muller (1954). Suitable paracentric inversions maintain the identity oi these chromosomes in successive generations. Mutants are detected when such females show one or more of the re- cessive traits. Maxy detects, therefore, any mutation involving the normal alleles of the fifteen reccssives, provided that the mutant does not produce the normal phenotype when heterozygous with the recessive allele and is not a dominant lethal. Once such mutants are obtained, they can be screened for point mutants. The study of recessive lethals in the X chromosome and in the autosomes shows that there are hundreds of loci whose point mutations may be recessively lethal. It should be noted that the recessive lethals detected by Base and the visibles detected by Maxy are not mutually exclusive types of mutants, for some Maxy-detected visibles are lethal when hemizygous, and about ten per cent of Base-detected hemizygous lethals show some morphological effect when hetero- zygous. It can be stated, in general, that any mutant in homo- or hemizygous condi- tion which is a "visible" will produce some change in viability, and, conversely, that any mutant which affects viability will produce a "visible" effect, "visible" at least at the bio- chemical level. SUMMARY AND CONCLUSIONS The mutational units of a genotype are, in order of size: the genome; the chromosome; chromosomal segments involving more than one gene; and the gene. Since a recom- hinational gene can have multiple alleles, gene mutation may involve the entire recom- hinational unit or one or more mutational sites within it. Although the genes delimited operationally by recombination and by mutation may not be materially equivalent, we shall continue to assume that this is so until we have evidence to the contrary. The occurrence of gene mutation is not limited by any ploidy, type of cell or gene, or effect it can have on synapsis. It is limited with respect to the effect it can have on gene polarity. Tripolar genes are excluded, bipolarity being the usual and unipolarity the less usual alternative. Point mutations are the remainder of all mutations not identifiable as intergenic changes. Since point mutants include gene mutants, the former can be studied to reveal the mutational characteristics of the gene. The frequency of point mutants increases linearly with the dose of high-energy radiations; there is no effect from dose Point Mutations 199 H. J. Muller, at Cold Spring Har- bor, N.Y., 1941. protraction and no threshold dose below which the genetic material is safe from change. Point mutations also indicate that a given gene is relatively stable over many cell gen- erations— changes in genes resulting from very localized physico-chemical events last- ing a matter of minutes, after which the new gene is stable. Point mutations are en- hanced or induced by temperature changes, aging, gene replication, and physical and chemical mutagens. It is possible that changes resulting in point mutants take place in the old gene, in the new gene, or during the formation of the new gene. Genetic schemes for the detection of X-linked recessive lethal and recessive visible mutants in Drosophila are described. A single representative of most normal genes fails to produce the full normal phenotypic effect, and most point mutants act on the phenotype in a hypomorphic or amorphic manner. The study of point mutants of these and other types reveals that almost all are detrimental to the reproductive poten- tial of individuals when pure (not hybrid), and to a lesser extent when hybrid. Ac- cordingly, most point mutants are not completely recessive to their normal genetic alternatives. REFERENCES Alexander, P., "Radiation-Imitating Chemicals," Scient. Amer., 202. No. 1:99-108, 1960. Crow, J. F., and Temin, R. G., "Evidence for Partial Dominance of Recessive Lethal Genes in Natural Populations of Drosophila," Amer. Nat., 98:21-33, 1964. Muller, H. J., "Variation Due to Change in the Individual Gene," Amer. Nat., 56:32- 50, 1922. Reprinted in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, 1959, pp. 104-116. Muller, H. J., "Artificial Transmutation of the Gene," Science, 66:84-87, 1927. Re- printed in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, 1959, pp. 149-155, and also in Great Experiments in Biology, Gabriel, M. L., and S. Fogel (Eds.), Englewood Cliffs, N.J.: Prentice-Hall, 1955, pp. 260-266. Muller, H. J., "A Semi-automatic Breeding System ('Maxy') for Finding Sex-linked Mutations at Specific 'Visible' Loci," Drosophila Info. Serv., 28:140-141, 1954. 200 ( HAPTER 14 \l tiller. H. J.. "The Nature oi the Genetic Effects Produced by Radiation," in Radiation Biology, Hollaender, A. (Ed.), Vol. I. Chap. 7:351 473. New York: McGraw- Hill, ll>54. Mul lor. H. J., and Oster, I. I.. "Some Mutational Techniques in Drosophila," pp. 249- 278. in Methodology in Basic Genetics, Burdette, W. J. (Ed.). San Francisco: Holden-Day, 1963. Schalet, A.. "A Study of Spontaneous Visible Mutations in Drosophila Melanogaster," Proc. X Intern. Congr., Genetics. Montreal. 2:252 (Abstr.), 1958. See Supplement 111. QUESTIONS FOR DISCUSSION 14.1. Is there a safe dose of X rays and/or ultraviolet radiation; that is, a dose that cannot produce some point mutations? Explain. 14.2. Can we he sure that any given mutation involves a single gene change rather than intergenic one? Explain. 14.3. Would we know of the existence of genes if all genes had identical mutational capacity? Explain. 14.4. Would you expect the mutation rate to Polydactyly, P, from normal, /?, to be greater among normal individuals in a pedigree for Polydactyly than it is among normals in general? Explain. How might you test your hypothesis? 14.5. Do the mutational properties discussed suggest any limitations with respect to the chemical composition of genes? Explain. 14.6. When a chromosome is broken, is the breaking point within a gene, between genes, or both? Justify your answer. 14.7. Point mutations are sometimes called gene mutations. Do you think this is permissible? Why? 14.8. In what way is the study of mutation dependent upon genes? In what way is the reverse true? 14.9. What is your opinion regarding the validity of applying principles of point mutation directly to gene mutation? 14.10. Are all of the mutants detected by the Base or Maxy techniques point mutants? Explain. 14.11. Suppose, in the Base technique, an F2 culture produced both of the expected t\pes of daughters but no sons at all. To what would you attribute this result? 14.12. How can you determine whether a recessive lethal detected in the FL> by the Base technique is associated with an inversion or a reciprocal translocation? 14.13. A wild-type female produces 110 daughters but only 51 sons. How can you test whether this result is due to the presence, in heterozygous condition, of a recessive X-linked lethal? 14.14. How can you explain the phenotype of a rare female in the Maxy stock that produces only unexceptional progeny but has compound eyes distinctly lighter than normal? 14.15. Compare the relative suitability of man and Drosophila for the determination of mutation frequencies. 14.16. The genes in the X chromosomes are incompletely linked in the females of the Base stock. Do you agree with this statement? Why? Chapter 15 THE GENE POOL; EQUILIBRIUM FACTORS T! |he recombinational and muta- tional properties of the genetic material have been studied in cross-fertilizing individuals and the nature and phenotypic consequences of various ge- netic units have been described in terms of the traits found in such individuals and their relatives. Cross-fertilizing individuals are members of a general population. In a general population, each individual usually has an opportunity to choose a mate from a large number of the other members. The gametes of all mating individuals furnish a pool of genes, or gene pool, from which the genes of the next generation are drawn. Over successive generations what happens to the frequency of a particular gene in the gene pool? Let us construct a gene pool and investigate this question. Suppose that Mars is colonized by human beings, that the population sent there is suffi- ciently large, and that — with respect to eye color genes — only the B (brown) allele and the completely recessive b (blue) allele are present in the gene pool in the frequencies .2 B and .8 b. Presuming that marriages are not influenced by eye color phenotype, what genotypes and phenotypes will the F, have? The answer can be seen in Figure 15-1. As the result of the random union of gam- etes, 4% of these children are BB; 32% are Bb; and 64%, bb. Phenotypically, the Fi population is composed of 36% brown- and 64% blue-eyed people. In the absence of mutation, what is the 201 gene pool in the gametes of the F,? The 4% of F, BB individuals furnish 4% of all gametes, and these carry B. The 32% of F, Bb individuals furnish 32% of all gametes of which half (16% ) carry B and the other half b. Therefore, the total gene pool con- tains 20% gametes with B. The b gametes comprise 80% of the gene pool (16% from the 32% of Bb heterozygotes and 64% from the 64% of bb individuals). Note that the gene pool of the F, is identical to that of the P,. Therefore, in the F2 and all subsequent generations, the same genotypic and pheno- typic ratios are found, because the fre- quencies of B and b in the gene pool remain constant. What would be the consequence if, in- stead of starting the Martian colony with a gene pool of 20% B and 80% b, some other proportion were used? We can generalize FEMALE GAMETES .2 B .8 b .04 BB .16 Bb .2 B Brown Brown MALE Eyes Eyes GAMETES 16 Bb .64 bb .8 b Brown Blue Eyes Eyes The F, Population .04 Brown (BB) + .32 Brown (Bb) + .64 Blue (bb) The F, Gene Pool B = .04 + .16 = .2 b = .16 + .64 = .8 figure 15-1. F, genotypes and the gene pool these produce. 202 CHAPTER 15 EGGS p B (Brown) q b (Blue) SPERMS PB (Brown) qb (Blue) p BB (Brown Eyes) p q Bb (Brown Eyes) p q Bb (Brown Eyes) q2 bb (Blue Eyes) FIG1 ki 15-2. The types and frequencies of genotypes produced by a gene pool composed oj p B nnd q b. the analysis by letting p equal the fraction of male and female gametes in the popula- tion which carries B. and q equal the frac- tion which carries /). Naturally, for eggs, p -+- q = 1, as is also true for sperm. These sex cells combine at random to produce the result shown in Figure 15-2. The offspring population, then, is p- BB + 2 pq Bb + q2 bb The fraction of brown-eyed individuals is p- + 2 pq, whereas q- is the blue-eyed frac- tion. The frequency of B and b among the gametes produced by the offspring popula- tion is: B = p- + pq = p(p + q) = p b = q2 + pq = q(q + p) = q Thus the gene frequencies have remained the same as they were in the gametes of the previous generation, and all future genera- tions will have the same gene pool and the same relative frequencies of diploid geno- types. The formula p- BB + 2 pq Bb + q- bb describes the genotypic equilibrium produced by a static gene pool.1 1 This is called the Hardy-Weinberg equilibrium principle. It should be noted that this equilibrium principle is independent oi the occurrence of dominance. Moreover, the B and b in the formula can represent any two alleles whose frequency in the gene pool is known, even if the sum of their frequencies is less than one, as in cases of multiple allelism. If this equilibrium principle applied in- definitely, gene frequencies would remain unchanged, and the evolution of different genotypes and their resultant new pheno- types would not occur. In the Martian model described, certain conditions had to be fulfilled in order to maintain a genetic equilibrium. One condition was met by barring mutation, for if it were permitted, obviously the frequency of the two alleles, B and b, in the population would have been reduced, and the equilibrium upset. The frequency of any allele would also have been changed if the mutation rates to and from it were different. In either or both types of events, the genetic equilibrium is shifted until a new one is attained. There- after, the new equilibrium is maintained until some new factor acts on mutation rate in a directional way. Our model also assumed that the repro- ductive potential ( biological fitness, or adap- tive value) was the same regardless of the genotype for eye color. But it is possible, under certain conditions, that persons with blue (or with brown) eyes are preferred as mates, in which case the reproductive po- tential of an individual is not independent of the alleles under consideration. Accord- ingly, if individuals with a certain genetic endowment produce more surviving off- spring than those produced by a different genetic endowment, the genes which transfer this higher biological fitness tend to increase their frequency in the population, whereas those genes with lower fitness tend to de- crease it. In this way selection, by operat- ing on genotypes of different adaptive value. The Gene Pool; Equilibrium Factors 203 causes changes in gene frequencies and shifts in the genotypic frequencies found at equi- librium. The Martian population was also pre- sumed to be large. Suppose, however, that the Martian population (whose gene pool is 2B and .8b) ran short of food, and only one couple, determined by chance, could have children. The chance that this hus- band and wife would be blue-eyed is .64 X .64, or about .41. Accordingly, there is a 41% chance that the gene pool will drift at random in this particular manner, producing the new gene frequencies of 1 .0 for b and 0 for B. This random genetic drift can also be illustrated in a less extreme situation: If a population is very large, and a certain family happens to produce a rela- tively large number of children for several generations, then the proportion of all indi- viduals in the population with this family name is still very small. But if the popula- tion decreases while this family's reproduc- tive rate is unchanged, the proportion of the population with this surname increases. Ac- cordingly, when populations are very large, oscillations in the number of children pro- duced by different genotypes occurring by chance are unimportant, for they do not change the gene pool. In small populations, however, such chance oscillations can change gene frequencies via random genetic drift. In our Martian model, the possibility that the colony would have emigrants or immi- grants was not considered. If the emigrants' gene frequencies are different from those re- maining in the population gene pool, then the gene frequencies in the remaining popu- lation will be changed. If the immigrants' gene frequency is different from the natives', and they interbreed, the gene pool will again be changed. In this way migration can shift the genetic equilibrium. We see then that a cross-fertilizing pop- ulation remains static — in genetic equilib- rium— in the absence of mutation, selection, random genetic drift, and differential migra- tion. The occurrence of one or another or all of these factors changes the frequencies of genes in the gene pool and thereby shifts the frequencies of genotypes at equilibrium. Different species possess different gene pools, and it is natural to presume that they are different species because of their different gene pools. Accordingly, the factors which change gene frequencies are considered to be the main causes of species formation. Insofar as the formation of higher taxonomic categories is, like speciation, based upon change in gene pools, the principal causes of biological evolution are: 1. Mutation (which supplies the raw ma- terials) 2. Selection (which shapes these raw ma- terials into the biologically fit geno- types of races and species) 3. Random genetic drift (which can pro- duce rapid changes in gene frequency in small populations) 4. Differential migration (which can shift gene frequencies via interchange of in- dividuals between populations). * Selection of Genotypes The disequilibrating effect of selection upon the gene pool has already been noted. Selec- tion acts at the phenotypic level to conserve in the population those genotypes which pro- vide the greatest reproductive potential. Se- lection takes place at all stages in the life cycle of an individual. Since it acts to pre- serve whole phenotypes and not single traits, selection conserves genotypes and not sing'e genes. Sometimes selection acts upon the phenotypes produced by single genomes in haploid species or stages; at other times — in sexually reproducing organisms — it acts upon the combined phenotypic effect of two genomes. It should be noted that what is a relatively adaptive genotype at one stage of 204 CHAPTER 15 the life cycle may be relativelj ill-adaptive at another whether or not these stages have the same or different ploidies. It is. of course, the total adaptiveness of all these separate features which determines the over- all reproductive potential oi an individual. Finally, it should he noted that in cross- fertilizing populations, selection favors geno- types which produce maximal fitness of the population as a whole. Because selection acts this way, it is possible that some por- tion of the population receives genotypes which are decidedly not advantageous. If this is so. the same genetic components are expected to be advantageous when present in en her. more probable, combinations. * Selection Against Mutants Since the human being is primarily a diploid, it is upon the diploid-produced phenotype that selection principally operates. If one asks, "What is the fate of mutants in the gene pool?*' the answer must include knowl- edge of the frequency with which the mu- tants arise as well as their effects upon re- productive potential in a diploid genotype. Remember that the phenotypic effect of a mutant gene depends not only upon the na- ture of its allele but also upon its relationship to the rest of the genotype. Let us consider, in turn, the fate in the human gene pool of mutants whose overall phenotypic effect is: dominant lethal; domi- nant detrimental; recessive lethal; or reces- sive detrimental, as influenced by selection and mutation. Dominant lethal mutants are lethal when heterozygous and are eliminated from the gene pool the same generation they arise. Accordingly, the biological fitness of such mutants is zero. If the normal homozygote (AiA]) is considered to have a selective disadvantage of zero, then the dominant lethal is at a complete selective disadvantage, and the selection coefficient, s, is one. We can readily see that the imitation frequency, ii, of this dominant lethal condition must equal one half the frequency of affected in- dividuals (A1A2), since each affected indi- vidual has one mutant and one normal gene. In the absence of special medical treatment. retinoblastoma, a type of cancer of the eye, is an example of such a dominant lethal in man. A chondroplastic ( or chondrodystrophy ) dwarfism is characterized by disproportion — normal head and trunk size but shortened arms and legs. This rare, fully penetrant (see p. 73) disease is attributed to the pres- ence of a gene in heterozygous condition which therefore acts as a dominant detri- mental mutant. Such dwarfs (A^A2) are known to produce only 20 per cent as many children as normal people. Because of this lower reproductive potential the A\A-> geno- type selection coefficient is .8. In one study the frequency of AxA-2 in the population was found to be 10 dwarf babies in 94,075 births. The dwarf children in this sample must have resulted from nor- mal parents who carried new mutations to A ■> or from one normal and one dwarf par- ent. The gene frequency, p, of A-< in the population, therefore, must be 10 per 2(94,075) or .000053. From the incidence of dwarfs known to have normal parents the mutation frequency, u, to A2 is .000042. If the value s = .8 is correct, then p = u/s, or .000042 .8, or .0000525, which is in excel- lent agreement with the gene frequency (p) value observed. Gene frequency for a dominant lethal equals mutation frequency (p = u) because s = 1; however, in the present case s is less than one, so the gene frequency is greater than the mutation fre- quency. Actually the gene frequency for dwarfism is not very much larger than the mutation frequency, demonstrating the effi- ciency of natural selection in eliminating such mutants from the gene pool. The gene for juvenile amaurotic idiocy {a-,) has no apparent effect when hetero- The Gene Pool; Equilibrium Factors 203 zygous (A^a-j)', since homozygous children die, a-2 is a recessive lethal mutant. Affected individuals are found with a frequency of one per 100,000, or .00001. What is the frequency of a2 in the gene pool? As shown in Figure 15-3, the frequency of a^a% indi- viduals at equilibrium is equal to q-. Ac- cordingly, the frequency of a-2 (q) must be equal to Vq12, or V .00001, or about .003, whereas the frequency of A\ must be one minus .003. or .997. Note that heterozy- gotes (carriers) are 600 times more fre- quent than afflicted homozygotes. What is the mutation frequency from Ax to a-P. As- sume that the gene pool is at equilibrium; in other words, the frequency with which a> enters the population by mutation equals the frequency with which it leaves the pop- ulation in flotfo homozygotes. Accordingly, the mutation frequency to a2 must be .00001. The selection coefficient for normal indi- viduals (AXAX and A^a^) is zero, and for a2a-2 it is one. We see, therefore, that at equilibrium the frequency of a recessive GENOTYPE PHENOTYPE FREQUENCY AT EQUILIBRIUM A, A, A,a: Normal Normal Dies P5 2pq q* u = Mutation rate from A, — a, q ="U u/s Here s = 1, hence q ="\| 0 u = 10~5 = 0.000,01 Hence q =M 0.000,01 = 0.003 ACTUAL FREQUENCY AT (0.997)' (2pq) 2I0.997M0.003) tq2) (.003I2 EQUILIBRIUM 0.994 0.006 0.000,01 OtQ DtO (WSSd lolorioo figure 15-3. Juvenile amaurotic idiocy. {See text for explanation.) figure 15—4. Pedigree showing the occur- rence of phenylketonuria among the offspring of cousin marriages {denoted by thick marriage lines). mutant in the gene pool can be expressed as q = Vii s, where s = 1 for a recessive lethal. When the homozygous recessive mutant is detrimental without being lethal, s becomes less than one (but more than zero) and the frequency of the mutant in the gene pool increases. Thus, if s were Y4 instead of one, q would be twice as large. * Nonrandom Matings In deriving the types and frequencies of genotypes in a population at equilibrium, we assumed that marriages were random with respect to the genotypically-determined trait under consideration. Such a randomly- mating population is said to be panmictic or to undergo panmixis. What happens if the different genotypes do not marry at ran- dom? Consider the disease phenylketonuria (Figure 15-4) which involves a type of feeblemindedness in individuals who are homozygous for a recessive gene, and who metabolize the amino acid phenylalanine to phenylpyruvic acid, which is toxic. The frequency in the gene pool of the normal gene (A) is .99 and of the abnormal gene 206 CHAPTER 15 (a) is .01. In the population at equilib- rium, therefore. AA:Aa:aa individuals have frequencies oi 9801 10,000: 198 10,000: I 10,000, respectively. Notice that Aa in- dividuals are 1C)S times more frequent than aa, so that even if every aa did not repro- duce, only one per cent of the a genes pre- sent in the gene pool would be eliminated each generation. This fact illustrates the inefficiency of selection against homozygotes for rare recessive genes, at least insofar as lowering the frequency of such genes is con- cerned. A A and Aa individuals apparently marry at random but feebleminded people do not. So panmixis does not occur with respect to this trait, and persons with dif- ferent genotypes tend to be restricted in their marriages — all the available marriage part- ners making up a person's reproductive isolate. The occurrence of different repro- ductive isolates for normals and phenyl- ketonurics has little effect on the relative frequencies of different genotypes in succes- sive generations, because aa people have so few of all the a genes present in the popula- tion. Clearly, only marriages between two A a individuals are of consequence, since those are the major source of aa offspring. The example of phenylketonuria shows that when a gene is rare and apparently com- pletely recessive, nonrandom marriage has little influence either upon gene frequency or the diploid (heterozygous or homozygous) genotypes in which it is found in the popula- tion. When the mutant is relatively fre- quent in the population, however, it is ob- vious that nonrandom marriages raise the frequencies of certain diploid genotypes and lower others. Moreover, if there are adap- tive differences for the different genotypes, the composition of the gene pool can be changed in a different direction or at a dif- ferent frequency than would be predicted for a population mating at random. Consider two ways in which mating can be nonrandom. The first invokes the tend- ency of phenotypically similar individuals (except for sex) to mate and is referred to as assortive mating. This kind of breeding pattern is generally true in animals including human beings. The genetic result is the production of more homozygotes than would occur by randomly-chosen matings. The second departure from random mat- ing involves inbreeding, the tendency for mates to be more closely related in descent than randomly chosen mates. What is the effect of inbreeding carried out for a single generation? This can be determined by studying what happens to genes that are heterozygous in the parent generation. There are various degrees of inbreeding, the closest form being self-fertilization. In self- fertilization the heterozygote for a given pair of genes, A a, produces progeny of which one half are homozygous. In general, the de- crease in heterozygosity because of self- fertilization can be expressed as follows: the chance that an offspring receives a given gene in the male gamete is y2, and the chance that it receives the same allele in the female gamete is H; the chance that the offspring is a homozygote for that allele, therefore, is %. But there is an equal chance that the offspring becomes homozy- gous for the other allele, so that the total chance for homozygosis from this type of inbreeding is 50%. If all members of the population are heterozygotes and self-fer- tilize, then in each successive generation, half of the genes that were heterozygous become homozygous. Suppose, on the other hand, that a portion of a population mating at random has X% homozygous individuals. These could come from matings between two heterozygotes, two homozygotes, or a heterozygote and a homozygote. If the gene pool is at equi- librium, the random matings that tend to increase homozygosis are counterbalanced The Gene Pool; Equilibrium Factors 207 by others which decrease it, so that X% homozygosis remains constant generation after generation. Consider what happens in another portion of this population which happens to practice self-fertilization for one generation. Since this segment of the pop- ulation already shows X% homozygosis, its offspring will also have X% homozygosis. But, if this segment is Z% heterozygous, after self-fertilization the offspring will have only V-2 Z% heterozygosis, and, therefore, will show a total homozygosis of X% + % Z% . In other words, each generation of self- fertilization makes half of all heterozygous genes homozygous, and, in a normally ran- dom-mating population, the effect of self- fertilization is to increase the random-mating frequency of homozygosis by y2 the fre- quency of heterozygosis. How much is homozygosity increased in brother-sister (sib) matings? The chance that a particular gene in the father is present in the male sib is y2, and the chance that the male sib's child receives this is similarly ]{>; the chance for the occurrence of both events is %. The chance that the female sib receives and transmits this same gene to her child is also %. Therefore, the chance that the child of the sib mating receives two representatives of this same allele is l/4 times % j or it has ] \ 6 chance of being homozygous for this gene. Since the child has an equal chance to become a homozygote for the other allele in his grandfather and for each of the two alleles in his grandmother, this gives him 4 times y16 or a 25% chance of homozygosis. In other words, sib matings cause % of the heterozygous genes to be- come homozygous. This chance of homo- zygosis from sib mating is in addition to the chance of homozygosis from mating at random. Matings between individuals who have one parent in common are called half-sib matings. In this case, the frequency with which a given allele in the common parent passes to the male half-sib is y2, and the frequency with which an offspring of this sib receives this allele is l/2', the chance of both events occurring is, therefore, %. The chance is also r4 for these events to occur through the female half-sib, so that the chance of a given allele becoming homo- zygous from a half-sib mating is y4 times y4, or y16. Since the other allele in the common parent could, in this way, also be- come homozygous y16 of the time, the com- bined additional chance of homozygosity for half-sib matings is 1£, or, in other words, y8 of the heterozygous genes become homo- zygous because of this type of inbreeding. The amount by which heterozygosity is reduced because of inbreeding is called the inbreeding coefficient, F. In a similar man- ner we can determine that in the case of cousin marriage, F is Vlr>. The values of F for more complicated pedigrees can be worked out accordingly. All forms of inbreeding increase homozy- gosity. Let us calculate the consequence of cousin marriage upon the frequency of phenylketonuria. Its frequency of hetero- zygotes per 10,000 people is 198 (see p. 206). Cousin marriage reduces heterozy- gosity by y16, or by twelve individuals, of which half of them are expected to be nor- mal (A A) and half affected {ad). Since random mating produces one affected indi- vidual per 10,000, cousin marriages bring the total number of affected homozygotes in this population to seven (six from in- breeding, one from random breeding). Ac- cordingly, there is a sevenfold greater chance for phenylketonuric children from cousin marriages than from marriages between un- related parents. Another example of how cousin marriages increase the risk of defect comes from a study which found that in a Japanese pop- ulation (Figure 15-5) congenital malforma- 208 CHAPTER 15 nons. stillbirths, and infant deaths were 24 to 48 per cent higher when cousins married than when parents were unrelated. Since, in sonic cases, delects such as these are known to be due to recessive genes in homozygous condition, these results support the view that homozygosis resulting from inbreeding can produce detrimental effects. Although in- breeding produces homozygosis and homo- zygosis can lead to the appearance of de- tects, it must not be inferred that inbreeding is disadvantageous under all circumstances. Many individuals do become homozygous for detrimental genes as a result of inbreed- ing, but just as many become homozygous for the normal alleles. The success of self- fertilizing species is testimony to the advan- tage of homozygosity at least for some types of organisms. * Heterosis In normally cross-fertilizing species, how- ever, inbreeding usually results in a loss of vigor which is directly linked to homozygosis. What is the functional basis for the adaptive superiority of heterozygotes. usually known as heterosis or hybrid vigor? Consider the three genotypic alternatives, A A, A A', A' A' relative to their phenotypic effects. Suppose A' A' is less vigorous than A A. Whether A is completelj or incompletely dominant to A' or shows no dominance to it. the A A' hetcr- OZygOte will be superior to one of the homozygotes. It is also possible that the heterozygote has a greater adaptive value than either type of homozygote. To illus- trate this possibility, imagine that A is pleiotropic. having a relatively great adaptive effect with respect to trait X but a relatively less adaptive effect with respect to trait Y. whereas the reverse is true of A', namely, relatively less adaptive for X and relatively more adaptive for Y. In the event of no dominance, the heterozygote is superior to either homozygote. Heterosis can be pro- duced, therefore, when the heterozygote is superior to either one or both homozygotes. The first type of heterotic effect can be demonstrated by crossing two pure lines, homozygous for different detrimental reces- sives (A A bb CC dd by aa bbCC DD). The F, (Aa bb CC Dd) is uniform yet more vigorous (having normal alleles at three loci) than either parent (each of which had nor- mal alleles at two loci) because the domi- nant alleles hide the detrimental effects of the recessive ones. In this case the hetero- zygous F2 progeny carrying Aa bb CC Dd are no more adaptive than the homozygotes, AA bb CC DD. Frequency from Increase in Frequency Per cent Unrelated Parents with Cousin Marriage Increase CONGENITAL MALFORMATION .011 .005 48 STILLBIRTHS .025 .006 24 INFANT DEATHS .023 .008 figure 15-5. Increased risk of genetic defect with cousin marriages. Hiroshima and Nagasaki. ) 34 ( Data from The Gene Pool; Equilibrium Factors 209 The second type of heterotic effect can be illustrated in human beings. As men- tioned on p. 71, homozygotes for the gene for sickle cell anemia (/3B[Ja) usually die from anemia before adolescence. f}A{lA in- dividuals have normal blood type, whereas ffAfja individuals are either normal or have a slight anemia. In certain countries the frequency of /3s in the gene pool follows the expectation for a recessive lethal gene. In other countries, however, fjs is more fre- quent than expected. This difference is at- tributable to the (3Af3s heterozygote being more resistant to certain kinds of malaria than the f3Af3A homozygote. Of course, in nonmalarial countries, (3s confers no anti- malarial advantage, and so the fitness of the heterozygote (1 — s) is lower than that of the normal homozygote ( 1 ) , whereas the ,3's7?8 individual has a fitness of zero. As expected, therefore, sickle cell anemia is rare or absent in most of the world where certain forms of malaria are absent. On the other hand, in certain malarial countries, even though heterozygotes may be slightly anemic, the advantage of being re- sistant to malaria produces a greater overall fitness than does the (3A/3A genotype. Here the fitness of the heterozygote, f3A(3s, is maximal and therefore must be assigned the value one, whereas that of the normal homo- zygote, f3ApA, is one minus Si. Mutant homozygotes, /?>s'/?'s', have a fitness of one minus s2, where s2 equals one, since all f3s(3s die (even if extremely resistant to malaria). In this situation natural selection maintains both fiA and /3s in the gene pool, (3s having a frequency equal to This fraction Si + s2 can be read as "the advantage of f3s (as shown by the advantage of flA(3s over /3Af3A ) divided by the total disadvantage of f3A and /3s." Thus, when the heterozygote, being more adaptive than either homozygote, shows heterosis in this way, natural selection maintains a gene such as (1* in the gene pool even though it is lethal when homozygous. Although we have discussed heterosis in terms of the phenotypic effects of the mem- bers of a pair of alleles, it should not be in- ferred that the unit of heterotic action is always a single pair of genes. Since we know that different pairs of genes interact in various ways to produce phenotypes, it would not be surprising to find that heterosis results from the effects of combinations of nonalleles and alleles. Natural populations of Drosophila pseudo- obscura contain various paracentric inver- sions. Laboratory populations can be started with some individuals carrying the normal chromosome arrangement and others, a particular one of these inversions. After a number of generations has passed, in some cases the population comes to contain only normal chromosomes, because the inversion chromosome behaves like a detrimental gene which provides no advantage when hetero- zygous and is eliminated from the gene pool. When other particular inversions are tested this way, however, an equilibrium is reached — both the normal and inverted chromo- somes are retained in the gene pool. In these cases, the inversion heterozygote is adaptively superior to either homozygote, showing heterosis just as the gene for sickling in malarial countries. It is difficult to decide the genetic basis for heterosis in such cases, however, since the hybrid vigor could be due to: the genes gained or lost at the time the inversion was initially produced; or the new arrangement of the inverted genes; or the types of genes or groups of genes con- tained within the inversion. Recall that in- dividuals with paracentric inversions are not at a reproductive disadvantage in Drosophila and suppose a heterotic system exists or de- velops in Drosophila heterozygous for a paracentric inversion. If the heterosis is due to the action of several specific nonalleles 210 CHAPTER 15 figure 15-6. The variability of normal corn is pointed out by James F. Crow. {Photographed in 1959 by The Calvin Company.) within the inverted region, this adaptively favorable gene content tends to remain in- tact in the heterozygote because of the fail- ure of single crossovers within the inverted region to enter the haploid egg nucleus. Breeding procedures that result in hybrid vigor have been widely applied to econom- ically important plants and animals. For example, it has been estimated that the use of hybrid corn has enriched society by more than a billion dollars. We might ask: What is wrong with normal corn? The answer is that it is too variable in quality and vigor (Figure 15-6). Inbreeding decreases vari- ability, but unfortunately inbreeding also re- sults in loss of vigor or other desirable traits. The way to overcome this problem is to obtain inbred lines which are uniform (because they are homozygous) and carry different favorable dominant genes (yet are also homozygous for various undesirable re- cessive genes ) , and cross the different inbred lines to each other. Their Fi will be multi- ply-heterozygous, uniform, and more vigor- ous than cither parental inbred line. Consequently, hybrids are made from two selected inbred lines — of corn in this case. Although the Fi plants are vigorous and uniform, they come from kernels grown on one of the less vigorous inbred lines. For this reason, hybrid seeds are not sufficiently numerous, and consequently, commercially unfeasible. In practice this difficulty is over- come (Figure 15-7) by crossing four se- The Gene Pool; Equilibrium Factors 211 INBRED A INBRED B INBRED C INBRED D SINGLE CROSS CxD SINGLE CROSS AxB ffll J DOUBLE CRO yjri (AxB)x(CxD CROSS ) figure 15-7. The production of commercial hybrid corn by the "double cross" breeding procedure. 212 CHAPTER IS ■ inbred lines img sold inexpensively. Hel practical importance; a fuller understanding single cross fu brids arc then this phenomenon requii other Since I rous the biochemical, molecular I c % c I . rbrid plant, seeds produced b) - . and w touble Ct plentiful .\n<.\ can be SUMMARY AND CONCLUSIONS I he whose functi produce the I popi* pool The gene pool and the l •n> will remain forever unchanged if: the population rtk drift does not occur, mutation does not - direction preferentially . no differential selection is made I •IK |iist like the • lr however. conditions is not Bed. a shift in the composition of the gene pool will occur, in other words, frequencies will change and so will the frequencies of different genotvpes until a new equilibrium is attained. It 't onlv species formation hut all oH hiological evolution is b upon changes in the gene pool I he roles th..r mutation and selection have in establishk mc equilibrium is discussed tor those rare mutants which lower reprodlK lethal, dominant detriment.il. recessive lethal, or l lemmental lorn breeding resulting the freq' I he per generation rate oi reduction in betel due to inbreeding is ': tor self-fertilization, '» for sib • »r half-sib and ¥u for cousin matings. H<" Lilizing individuals leads i h\ heterosis, or h\bnd '• Hel • i phenotvpic result ol in because the heiero- idaptive! ft to one or to both types o\ homo' peal tm nee economically. REFERENCES \lli». klc Cells and Evolution." Scical tin >5:87 M If. polls H' and the Origin of Sp )rd EdL, New Yofi Cob ;' Doha netics and M in WHej Gowcn, I. W. (Ed), A wa: low . S Mixed Populatioi • Reprinted a < Peters, J. A. (Ed.). Eng -wood , s i Pi H trimenta in 8 ,,! m i od ( tiffs N I Prentice-Hall, ■ •.v-Hiil. I in- (icnr Pool, Equilibrium ii\ •i ; Rasmuson m Genetics on iii< Population Level, Stockholm, Sweden Svenska ii"i f( n I a gel Bonnici i; i ondon Heincmann 1 961 Spi.s, I H (Ed.) Papers on \nimal Population Genetics, Boston l ittle, Brown 196 ■ ■. ii i i . .Lin i V., Kessingei . M . A and Harris, w Differential Rates ol Di vclopmenl ol I leteroth .mil Nonheterotii V< g Mai/c Seedlings I < orrelation ol Dill ii.il Morphological Development with Physiological i>iii In Qermlnating s,. ,i. I'm.. Nal A. id. Sci U.S 11:212 118 1964 Spragui ' > I 1 1 ii ) Corn and Corn Improvement, New York \cademii Press \'> • ■ Weinberg, W I bei d des Vererbung belm Menschen lahresh Vercin i vaterl Naturl In WUrttemberg 64 168 182 1908 rranslated In part, In Stern ( i he Hardy Weinberg I • I len >7: 137—138, 194 I 'jiir.flONS FOR DISCUSSION i I i An- the ' iiises "i evolution the lame In populations reproducing onl) asexuall) .is in those reprodu< ing lexually? i (plain i . • Suppose in a population obeying the Hard) Weinberg rul< mutation o i.ii mil • 'i iration and changes the compositii i the gene pool H ■ man) additional generations are required befori a new gjenotli equilibrium is . itabli ihi 'i ' i iplain I Ml OD( - III ■ I lOR/.HAI SK ■ iii 1937 SBWAL] WatOHl U noted fOf his n\r,inli in physiological genetics and In the mathematics o) population genetics Photograph was taken in 193 I -It CHAPTER 15 15.3. Discuss the statement: "The Hardy-Weinberg Law is the cornerstone of evolu- tionary genetics." 15.4. Assuming that the Hard) Weinberg principle applies, what is the frequency o\ the gene R h us onlj allele R' is homozygous in the following percentages of the population: 49%? 49? ? 25%? 36%? 15.5. In the United Stales about 709? of the population gets a hitter taste from the drug phenyl thiocarbamide (PTC). I hese people are called "tasters" and the remaining 309? who get no hitter taste from PTC are called "nontasters." All marriages between nontasters produce all nontaster offspring. Every experi- mental result supports the view that: a single pair of nonsex-linked genes de- termines the difference between tasters and nontasters; dominance is complete between the only two kinds of alleles that occur; penetrance o\ the dominant allele is complete. ( a ) Which of the two alleles is the dominant one? (b) What proportion of all marriages hetween tasters and nontasters have no chance (barring mutation) of producing a nontaster child? (c) What proportion of all marriages occurs between two nontasters? Two tasters? 15.6. The proportion of A A individuals in a large crossbreeding population is .09. Assuming all genotypes with respect to this locus have the same reproductive potential, what proportion of the population should be heterozygous for A) 15.7. What do you suppose would happen to a population whose gene pool obeyed the Hardy-Weinberg rule for a very large number of generations? Why? 15.8. Can a population obey the Hardy-Weinberg rule for one gene pair but not for another? Explain. 15.9. Explain whether the mutation frequency to a particular allele is of primary im- portance in shifting its frequency in the population, when this gene is: (a) a dominant lethal in early developmental stages (b) a recessive lethal (c) phenotypically expressed only after the reproductive period of the individual (d) very rare (e) present in small cross-fertilizing populations 15.10. Can the adaptive value of the same gene (15.9) differ in: (a) haploids and diploids? (b) males and females? (c) two diploid cells of the same organism? Explain your answer in each case. 15.1 1. Other things being equal, what will happen to the frequency in the gene pool of a dominant mutant whose selection coefficient changes from one to V? If the mutant is completely recessive? 15.12. If persons carrying detrimental mutants never marry, these particular genes are removed from the gene pool. Under what conditions is the failure to marry likely to appreciably reduce the frequency of detrimental mutants in the gene pool? 15.13. Are inbreeding and assortive mating mutually exclusive departures from random mating (panmixis)? Explain. 15.14. Explain why the inbreeding coefficient, F, is Vie for cousin marriages. The Gene Pool; Equilibrium Factors 215 15.15. Suppose the frequencies of A and a are .3 and .7, respectively, in a population obeying the Hardy-Weinberg rule and mating at random: (a) What per cent of the population is composed of homozygotes with respect to these genes? (b) What would be your answer to (a) after one generation of mating hybrids only with hybrids? (c) How would the conditions in (b) affect the composition of the gene pool? 15.16. Discuss, from a genetic standpoint, the advantages and disadvantages of cousin marriages in man. 15.17. In Thailand, heterozygotes for a mutant gene that results in the formation of hemoglobin E are more frequent in the population than would be expected from the Hardy-Weinberg rule. How can you explain this? 15.18. Two inbred strains of mice and their Vl hybrids are tested for locomotor activity (measured for each subject in each group during three consecutive five-minute periods) and for oxygen consumption. In both these respects the F, hybrid is less variable than the parental strains. Propose a genetic hypothesis to explain these results. 15.19. Compare the reproductive isolates of people who were marrying in 1900 with those marrying today. Which factors are the same and which are different? Is the change desirable from a biological standpoint? Explain. Chapter *16 GENETIC LOADS AND THEIR POPULATION EFFECTS Genetic Loads in Drosophila The fruit fly, Drosophila pseudoobscura, is commonly found in northern Mexico and the western United States. When collected in the wild, almost all its individuals are phenotypically alike, except for the sex differences, appearing wild-type or normal. We cannot accept this phenotypic uniform- ity as evidence of genotypic uniformity, how- ever, since a Drosophila population appear- ing wild-type can conceal considerable genetic variability in the form of isoalleles, recessive point mutants, reciprocal transloca- tions, paracentric inversions, and so on. We would like to estimate the genetic load — the total amount of this genetic variability pres- ent in a natural population of D. pseudo- obscura.1 D. pseudoobscura has five pairs of chro- mosomes— the usual X and Y sex chromo- somes, three pairs of large rod-shaped auto- somes (II, III, IV), and a dotlike pair of autosomes (V) (Figure 16-1). Numerous laboratory strains of this species are avail- able whose autosomes are marked by various point and rearrangement mutants. We can, therefore, make a suitable series of crosses between laboratory strains and flies collected in the wild which will yield information on the presence of autosomal mutants in the wild-type flies. In practice, autosomes II, III, and IV of individual wild-type flies are made homozygous to detect the presence of the following recessive mutants (see Figure 6-2): 1. Lethal (causing death to all individuals before adulthood) or semilethal (caus- ing more than ninety and less than one hundred per cent mortality before adulthood) 2. Subvital (causing significantly less than normal but greater than ten per cent survival to adulthood ) 3. Female sterile (sterile to females) 4. Male sterile (sterile to males). The results of this study are summarized in Figure 16-2. About 25% of all auto- somes tested this way carry a recessive lethal or semilethal mutant. Recessive subvital mutants are found in about 40% of III chro- mosomes tested and in more than 90% of II's and IV's tested; mutants causing sterility are present in 4 to 14% of tested chromo- somes. Obviously the natural population carries a tremendous load of detrimental mutants. How is this load of mutants distributed in the fly population? Consider first one pair of the autosomes tested. Each member has a 25% chance of carrying a lethal or semilethal and a 75% chance of being free of such mutants. The chance that both members of a pair of chromosomes will carry a lethal or semilethal is (0.25 )- or 6.25% . From the data presented we cannot tell whether all the lethals and semilethals found in a particular pair of autosomes are 1 The following is based upon Dobzhansky and collaborators. 216 work of Th. FIGURE 16-1. Chromosomal complement of D. pseudoobscura. Genetic Loads and Their Population Effects 217 MUTANT TYPE PER CENT OF CHROMOSOMES II III IV Lethal or Semilethal 25 25 26 Sobvital 93 41 95 Female Sterile 11 14 4 Male Sterile 8 11 12 figure 16-2. Genetic load in natural popula- tions of D. pseudoobscura. {After Th. Dob- zhansky.) allelic (in which case up to 6.25% of zy- gotes in nature would be mutant homozy- gotes and fail to become adults), or whether all the mutants involve different loci (in which case 6.25% of zygotes would be hy- brid for linked mutants of this kind), or whether some combination of these alterna- tives is obtained. In any case, the chance that both members of a given chromosome pair are free of lethals or semilethals is (0.75)- or 56%. What portion of individuals in the pop- ulation carry no lethal or semilethal on any member of autosomes II, III, and IV? This percentage is calculated as (0.75)2 times (0.75)- times (0.75)-' or about 17%. However, if one considers the X and V chromosomes which can also carry such mutants, the frequency of lethal-semilethal- free individuals in nature is still lower. Moreover, when the subvital mutants (which comprise the most frequent mutant class de- tected) and the sterility mutants are also considered, it becomes clear that very few, if any, flies in natural populations are free of a detrimental mutant load. Genetic Loads in Man What is the genetic load in man? The vast majority of mutants are detrimental in homo- zygous condition (as already noted in Chap- ter 15). Since inbreeding increases the fre- quency of homozygosis, a comparison of the detriment produced in an inbreeding segment with that in a noninbreeding segment of a human population may provide us with an estimate of the genetic load present in het- erozygous condition. From the population records of a rural French population during the last century listing fetal deaths and all childhood and very early adult deaths we can compare the frequency of death to off- spring of unrelated parents with that of cousin marriages.-' The frequency of death to progeny from unrelated parents was .12, whereas it was .25 from cousin marriages. We are not concerned here with establishing the genetic or nongenetic cause of death in the normal outcrossed human population; however, it can be assumed that the extra mortality of .13 (.25 minus .12) has a ge- netic basis in the extra homozygosity result- ing from cousin marriage. This assumption is reasonable in the absence of any known nongenetic factor that tends to cause death to more or fewer offspring from marriages between cousins than from marriages be- tween unrelated parents. (These data would have a nongenetic bias if, for example, it were the custom — which it was not — that all children from cousin marriages are pur- posely starved.) Apparently, then, 13% more offspring died because their parents were cousins. The total amount of recessive lethal effect present in the population in heterozygous condition can be calculated as follows: recall (Chapter 15) that of all heterozygous genes, an extra l/16 become homozygous in off- spring of cousin marriages. In the model half of the l/16, or y32, must have become homozygous for the normal genes and half of Y1G, or y32, for their abnormal alleles. Therefore, to estimate the total heterozy- gous content of mutants which would have been lethal if homozygous, it is necessary to - Based upon an analysis of N. E. Morton, J. F. Crow, and H. J. Muller. 218 CHAPTER 16 multiply .13 by 32. The resultant value of about 4 represents a 40095 chance that the ordinar) individual carried in heterozygous condition a genetic load of detrimental mu- tants which would have been lethal if homo- zygous. In other words, on the average, each person carried four lethal equivalents in heterozygous condition, or, lour times the number of detrimentals required to kill an individual if the genes involved somehow became homozygous. The preceding analysis did not reveal the number of genes involved in the production of the tour lethal equivalents. These lethal equivalents might have been due to the pres- ence in heterozygous condition of four re- cessive lethals, or eight mutants producing 50% viability, or sixteen mutants with 25% viability, or any combination of detrimental mutants whose total was four lethal equiv- alents. Because of environmental improve- ments (better housing, nutrition, and med- ical care) since the last century, it is likely that the effect of the same mutants in present- day society would be expressed by some- what less than four lethal equivalents. For the same reason, the detrimental effects of these mutants in heterozygous condition are expected to be somewhat less at present than they were a century ago. For example, in the last century a particular hypothetical homozygous combination having variable penetrance and expressivity would have pro- duced no detectable effect 25% of the time; a detrimental effect — but not death before maturity — 15% of the time; and death be- fore maturity 60% of the time; today, the respective values would be 50%; 10%; 40%. A century ago this combination would have produced .6 of a lethal equiva- lent; at present, the portion is .4. Notice also that the detriment not lethal before maturity would also have been reduced dur- ing this period from 15% to 10% or, speak- ing in terms of detrimental equivalents, what had been .15 would now be .10. Appar- ently the genes responsible for lethal equiva- lents and for detrimental equivalents must be the same, at least in part. It is also apparent that present-day man carries a genetic load. Some of those mu- tants transmitted to him arose in his parents (probably two of each five zygotes carry a newly arisen mutant, as mentioned on p. 192), and others arose in his more remote ancestry. It has been calculated :i that, on the average, each of us is heterozygous for what is probably a minimum of about eight such mutant genes. This genetic load does not include the mutants carried in homozy- gous condition. What happens to this load of mutants in successive generations? Balanced vs. Mutational Loads To predict, in a general way, the fate of the "usual" mutant in the population, it is neces- sary to determine its "usual" phenotypic effect.4 Since the typical mutant is detri- mental when homozygous — at least to some degree — the homozygous condition tends to eliminate it from the gene pool. But two opposite effects are possible for mutants when heterozyg3us (see Chapter 15): either the heterozygote is superior to both homo- zygotes (as is found for the sickling-causing gene in malarial countries), or the hetero- zygote is somewhat inferior to the nonmu- tant homozygote (as is true for most point recessive lethal heterozygotes). In the for- mer case the heterotic effect tends to in- crease the frequency of the mutant, and both the normal and mutant genetic variants are retained in the population gene pool at equi- librium. A population which normally re- tains more than one genetic (or chromo- somal) alternative in its gene pool, there- fore, exhibits balanced polymorphism in its phenotypes. This component of the genetic load is balanced, and is, therefore, a balanced ■■ By H. J. Muller and by H. Slatis. •» See B. Wallace (1963), J. F. Crow and R. G. Temin (1964), and Th. Dobzhansky (1964). Genetic Loads and Their Population Effects 219 load. When the heterozygote is inferior to one homozygote, the heterozygous condition increases the rate at which the mutant is eliminated from the gene pool, and the pop- ulation shows unbalanced polymorphism and tends to become genotypically and pheno- typically monomorphic. This component of the genetic load, called the mutational load, is maintained in the population chiefly by recurrent mutation. Experimental evidence in Drosophila 5 and a statistical analysis of data for man ,; support the view that the great majority of point mutants are detri- mental when heterozygous. We shall, there- fore, consider most of the genetic load to be a mutational load. Genetic Death How is a mutant gene eliminated from the population? It need not be eliminated by the death of an individual, although some- times it is. A more general way to express the removal of a mutant gene from the gene pool is by genetic death — the failure of a mutant-carrying individual to produce de- scendants carrying the mutant. Thus, all an individual's genes, whether normal or mu- tant, suffer genetic death if that individual fails to produce children. Since mutants are stable, they are usually removed from the gene pool by genetic death and only occasionally by mutation. A person carrying a dominant lethal like retinoblastoma suffers genetic death (as well as physical death). In this case the mutant gene is eliminated from the population the generation in which it arises; it has, there- fore, only one generation of persistence. A dominant detrimental mutant with a selec- tion coefficient of .2 and, therefore, an adap- tive value of .8 as compared to normal, will n Based upon works of H. J. Muller and co-work- ers, C. Stern and co-workers, J. F. Crow and co- workers, I. H. Herskowitz and R. C. Baumiller, and others. 11 Based upon an analysis by N. E. Morton. persist for five generations, on the average, before suffering genetic death; that is, given a population approximately the same in size for successive generations, in each genera- tion the mutant-containing individual has a 20% chance of not transmitting the mutant. After this mutant arises, it sometimes fails to be transmitted the very first generation; it may suffer genetic death at the fifth gen- eration or at the tenth, but. on the average, the mutant persists five generations. The principle of persistence holds even though genetic drift, migration, or other factors cause fluctuations in the frequency of the mutant. Consider the fate in the population of a rare recessive lethal gene like the one pro- ducing juvenile amaurotic idiocy. Each time homozygosis for this gene occurs, it results in genetic death, and two mutant genes are eliminated from the gene pool. But consider the fate of heterozygotes which are 600 times more frequent (Chapter 15) and carry 300 times as many of these genes as do homozygotes. Since it is generally true that heterozygotes for a recessive lethal suffer genetic death about two per cent of the time (see p. 195), approximately .02 times 600, or twelve, heterozygous people would suffer genetic death, thus involving the removal of 24 genes, twelve of them being the recessive lethal alleles. Accord- ingly, six times as many of these particular recessive lethal genes suffer genetic death in the heterozygote than in the mutant homo- zygote, even though the reduction in repro- ductive potential in the former type is only '.-o of that in the latter. It is apparent that the rarer a mutant is, the smaller will be the proportion of all genetic deaths it causes in homozygotes and the larger the proportion in heterozygotes. For rare mutants, then, natural selection re- moves mutant genes primarily via the ge- netic death of heterozygotes, the small amount of detriment being more important •220 ( IIAI'TER 16 when heterozygous— from the population 01 gene pool standpoint — than the greater det- rimental effect when homozygous. How- ever, each rare mutant, in terms of its etleet on reproductive potential, is equally harmful to a constant-sized population in that each eventually causes a genetic death. Thus, hy- poploid}1 which aets as a dominant lethal per- sists onlv one generation before it causes a genetic death; a rare point mutant whose reproductive disadvantage is only UU)% will persist, on the average, one thousand gen- erations before causing genetic death. Consider, on one hand, the gross chromo- somal abnormality which kills in utero, de- stroying a life early. Neither the individual involved nor its parents suffer very long, since such deaths may occur as abortions which pass unnoticed. On the other hand, the heterozygous point mutants in individuals who are past the reproductive age — and, therefore, already have or have not suffered genetic death — will continue to subject these people to the previously and newly pro- duced, small phenotypic detriment of hetero- zygosity which adds to their aches, pains, and disease susceptibility. In this respect, then, the mutant with a small effect on re- productive potential can cause more suffer- ing than one with a large effect, for the longer the persistence, the more the damage in postreproductive life. In general, speak- ing not in terms of biological fitness but in terms of the total amount of suffering to which a human population is subject, point mutants with the smallest heterozygous detriment are the most harmful type of mutant. One might at first suppose that the amount of gene-caused human suffering can be reduced by medical science. This pos- sibility exists, particularly for an individual such as the diabetic who takes insulin; no doubt he is better off than he would be without medicine. But remember that this medicine does not cure the genetic defect. Moreover, by increasing the diabetic's repro- ductive potential, the medicine serves to increase the persistence of the mutants in- voked, and the genetic death which must eventually occur is only postponed to a later generation — each intervening generation re- quiring the same medication. The total amount of human sutlering would be re- duced only if medicine could correct the gene-produced defect. To correct all of the multiple effects of the mutant, the medicine would have to replace the primary product of the defective gene with normal product. Insofar as most, if not all, currently known medicines act later than this earliest stage in gene action (Chapter 6), they serve to alleviate only some detrimental effects, thus causing an increase in human suffering by increasing persistence. Unfortunately, this situation will continue until medical science is much further advanced. In view of the preceding discussion, we can assume that it is primarily the euploid or nearly euploid mutants which persist in the gene pool and are mainly responsible for changes in its composition during the course of evolution. By far the most common and most important class of such mutants is the point mutant. Mutation and Evolution In Chapter 1 5 we only suggested that muta- tions provide the raw materials for biolog- ical evolution. The reason for our hesitance in specifying evolution as the natural out- come of changes in gene pools was that the great majority of mutants, including point mutants, are harmful in homozygous or hemi- zygous condition. In this chapter and in Chapter 15, we indicated that most mutants are also detrimental when heterozygous. Under these circumstances, how can muta- tion provide the more adaptive genotypes postulated as necessary for evolution? It is true that for a given genotype under a given set of environmental conditions the great Genetic Loads and Their Population Effects 221 majority of point mutants are detrimental, and that, perhaps, only one point mutant in a thousand minutely increases the reproduc- tive potential of its carrier. Yet, provided the mutation rate is not too large and there is sufficient genetic recombination, these rare beneficial mutants offer the population the opportunity to become better adapted. Moreover, mutants which lower biological fitness under one set of environmental con- ditions may be more advantageous than the normal genes under different environmental circumstances.7 For example, a mutant producing vestigial wings in Drosophila is clearly inferior to its normal genetic alterna- tive in an environment where flight ability is advantageous; but this mutant might be advantageous for Drosophila living on a small island where flight is not only un- necessary but harmful because insects that fly can be blown out to sea and lost. Con- sider a second example of this type. Several decades ago the environment was DDT-free. and mutants which confer immunity to DDT were undoubtedly less adaptive than the normal genetic alternatives present. But once DDT was introduced into the insect environment, such mutants — even if detri- mental in other respects — provided such a tremendous reproductive advantage over their alternatives that they became estab- lished in the population as the new wild- type genes. Still other examples can be cited involving antibiotic-resistant mutants in microorganisms, which in an antibiotic- free environment are less adaptive than the genes normally present. It becomes clear, then, that mutation pro- vides the opportunity for a population to become better adapted to its existing en- vironment. It also provides the raw mate- rials needed to extend the population's range to different environments, either those al- ready existing elsewhere or those that will •See Th. Dobzhansky (1964). arise through changes. A population that is already very well adapted to its present environment is appreciably harmed by the occurrence of mutation. But environments differ, and any given environment will even- tually change, so that a nonmutating popu- lation though successful at one time will, in the normal course of events, eventually face extinction. Mutation, therefore, is the price paid by a population for future adap- tiveness to the same or different environ- ments. We can now appreciate that muta- tion and selection, together with genetic drift and migration, are primarily responsible for the origin of more adaptive genotypes. We can also better appreciate the advantage of genetic recombination in speeding up the production of adaptive genotypes and the importance of the genetic mechanisms which regulate mutation frequency. Somatic Mutations In view of the preceding discussion, it is not at all difficult to predict the consequences of increasing the mutation frequency in human beings, an increase that doubtlessly is oc- curring as a result of our exposure to man- made penetrating radiations and certain re- active chemical substances. Man-made as well as spontaneous mutations can occur in either the somatic line or the germ line. Somatic mutants are, of course, restricted to the person in which they occur. The earlier the mutation occurs in a person's life, the larger will be the sector of somatic tissue to which the mutant cell gives rise. When an adult is exposed to an agent which causes a mutation to occur in a cer- tain percentage of all cells, the cells carrying induced mutants will usually be surrounded by nonmutant ones of the same tissue whose overall action produces a near-normal phe- notypic effect. When an embryo is exposed, a proportionally smaller number of its cells will mutate. Mutant embryonic cells can, however, give rise later to whole tissues or CHAPTER 16 organs which arc defective; in such cases there is no compensator) action of normal tissue. Furthermore, since many mutants affect the rate of cell division, the earlier in development they occur, the more ab- normal the si/e o\' the resulting structure will be. It is understandable, then — assum- ing that cells at all stages are equally mutable — that the earlier somatic mutations occur in the development of an individual, the more damaging they will be to him. Newly arisen mutants produce almost all their somatic damage when heterozygous, since mutation involves loci which are usu- ally nonmutant in the other genome. Al- though somatic mutants cannot be trans- mitted to the next generation, they can lower the reproductive potential of their carriers, thus affecting the gene pool of the next generation. The damage which new mutants produce in a somatic cell depends upon whether or not the cell subsequently divides. Certain highly differentiated cells in the human body, like nerve cells or the cells of the inner lining of the small intestine, do not divide. In such cases, it is ordinarily difficult to detect mutations since the cells have no progeny classifiable as mutant or nonmutant. Non- dividing cells may be more or less mutable than those retaining the ability to divide. In any event, a variety of mutations can occur in nondividing cells, including point mutations which inactivate or change the type of allele present, as well as structural rearrangements of all sizes. Nevertheless, the nondividing cell remains euploid or nearly euploid, and the phenotypic detriment produced must be due almost entirely to point mutants in heterozygous condition and to shifts in gene position. Although this may considerably impair the functioning of nondividing cells and give the impression that they are aging prematurely, their sud- den and immediate death due to mutation is probably very rare. Although the same kinds of mutations occur in somatic cells that subsequently di- vide and in those that do not, nuclear divi- sion can result in gross aneuploidy (Chapters I I and 12). Accordingly, most of the phe- notypic damage of induced mutants in divid- ing cells is the result of aneuploidy — mostly the consequence of single breakages that fail to restitute. It should be noted that all known agents causing point mutation also break chromosomes. Germinal Mutations Since somatic cells comprise a population produced by asexual reproduction (cell divi- sion), the preceding discussion of the effects of somatic mutation is appropriate to this chapter. Consider next, in a general way, the consequences of increasing the frequency of mutations in the human germ line. The earlier that mutation occurs in the germ line, the greater the portion of all germ cells carrying the new mutant will be. Of course, the upper limit of gametes carrying a par- ticular induced mutant is usually fifty per cent. Consider the effect of exposing the gonads of each generation to an additional constant amount of high-energy radiation (Figure 16-3). The load of mutants pro- duced spontaneously is presumably at equi- librium— the rate of mutant origin equals the rate of mutant loss via genetic death. Beginning with the first generation to receive the additional radiation exposure, the mu- tant load increases with each generation until a new equilibrium is reached; at this point the higher number of genetic deaths per generation equals the higher number of new mutants produced each generation. If the additional radiation exposure ceases at some still later generation, the mutational load will decrease gradually (because of variations in persistence) via genetic deaths, until the old spontaneous equilibrium is reached again. Genetic Loads and Their Population Effects 223 GENETIC LOAD Spontaneous Radiation Exposure GENERATIONS figure 16-3. Genetic load and exposure to radiation. It is clearly important to learn in detail the genetic effects of high-energy radiation to which human populations are being sub- jected either purposely or circumstantially. In order to make the best evaluation, we will need to know much more about: the dis- tribution of the energy of various radiations in tissue; the exact amount of gonadal ex- posure to radiations of different types; the detriment of the induced mutants in hetero- and homozygous conditions; the persistence of mutants; and the different types and the frequencies of mutations that each kind of radiation produces in different stages of male and female gametogenesis. In the last respect, it is necessary to de- termine for various types of mutations, the relative mutagenicity of a concentrated dose and one given in a protracted manner. It is also necessary to learn as accurately as possible the mutability of spermatogonia and oocytes, because these are the stages in which the human germ cells producing the next generation remain for the longest pe- riod of time. It is suggested that the largest number of germ-line mutations occurs in oocytes. Because spermatogonia are con- stantly dividing, mutants producing a detri- mental effect may be selected against so that they are reduced in frequency by the time gametes are formed; the human female, how- ever, is born with all, or almost all, her future gametes already in the oocyte stage so that there is no parallel mitotic selection in this germ line. Not only do oocytes fail to undergo mitosis, but they remain rela- tively inactive for decades before becoming ova; as oocytes age during this period, they become disproportionately sensitive to spon- taneous mutation (at least to factors leading to aneusomy). Although at present we do not have as much information about any one of these factors as we would like, available informa- tion along these lines already gives us ap- proximate answers (see the references at the end of this chapter). It should be noted, therefore, that all values given in the dis- cussion below may be in error by as much as a factor of two or more. It has been a practice to discuss the germ- line effect of radiation in terms of the amount of increase any particular exposure would produce in our spontaneous mutation fre- quency. The general impression is held that, as a species, man is fairly well adapted to his spontaneous mutation rate, and that if this rate is doubled it will not threaten his survival. Accordingly, the question be- comes, how much man-made radiation would produce as many mutations as occur nor- mally? A United Nations report calculates that about 30 rads (roughly equal to 30 r) is sufficient to double the human sponta- neous mutation rate — the frequency per gen- eration. This amount is called the doubling dose. In a population of one million peo- ple, one rad delivered to the gonads, or sex organs, of each person is calculated to pro- duce between 100 and 4,000 mutants which could be transmitted to future generations. Thus, one rad of gonadal exposure for one generation will result in the birth of 100 to 4,000 people with new heterozygous mu- tants. Affected descendants will occur for many generations, since only a small por- tion of the genetic deaths from these mutants will occur in the first generation. These 224 CHAPTER 16 will not be evident when added to the number of genetic deaths resulting from spontaneous mutation. It' the one rad gon- adal exposure were repeated every genera- tion, an equilibrium would eventually be established in which, for every generation, 100 to 4.000 people per million would show the effects of radiation-induced mutants in the form of genetie death. However, since the kinds of phenotypic effects produced by the radiation-induced mutants would be the same as those from mutants which occur normally, we would not be able to recognize specifically those people hurt by the radia- tion. What part of our normal load of mutants comes from naturally-occurring penetrating radiation? Since human beings receive about five rads in the course of a reproduc- tive generation — that is, in 30 years — it is possible that as little as %0, or y6, of our mutations normally are radiation-induced. How much additional radiation are we exposed to during medical treatment? If medical use of radiation were to continue at its present level, it has been estimated that each person in the United States would re- ceive a total dose to the sex cells of about three r per generation. Of course, some people do not receive this amount of radia- tion, while others get considerably more. But this average radiation dose to the germ cells from medical uses alone is 60% of the amount received spontaneously and is rais- ing the mutation rate about 10% above the spontaneous rate. In the years to come, with increased use of radiation for diagnosis and therapy, the average dose from medical radiation might increase greatly. Already radioactive materials have been used in one million medical treatments in a single year. Many governments as well as private dental and medical groups are investigating such radiation exposure, and many ways of re- ducing radiation exposure without hinder- ing its usefulness are being instituted, It is difficult to determine the number of germ-line mutations resulting from the radia- tion associated with fallout following atomic explosions, because some radiation reach- ing the gonads could come from fallout on the ground, breathed in, or ingested with food. In the latter case, the distribution of particular radioactive substances in the body makes a large difference in the amount of radiation reaching the sex cells. With re- spect to sex cells, the three most important radioactive substances in fallout are cesium- 137, strontium-90, and carbon- 14. Because cesium is distributed through the tissues — including the gonads — more or less evenly whereas strontium is preferentially localized in bone, we expect cesium- 137 to produce more gonadal radiation damage from in- gested fallout than strontium-90 produces. The period of time over which radioac- tive substances produce new mutations also varies. The induction of mutations by rela- tively short-lived radioactive substances, like strontium-90 and cesium- 137, is restricted almost entirely to a few generations. On the other hand, carbon- 14 — C-14 — is long- lived, having a half-life of 6,000 years. So, if the exposure to C-14 in the environment does not change, there will be about half as many new mutations induced after 200 gen- erations as there are in the first generation. Because of its abundance and long half-life, carbon- 14's potential for delivering radia- tion to the gonads has been calculated as being 4 to 17 times more than radioactive cesium's and strontium's combined and, therefore, carbon- 14 is capable of producing proportionally more point mutations. In the United States National Academy of Sciences — National Research Council report of 1956 (see References), the gonadal dose expected from fallout — if weapons of the same type continued to be tested at the same rate — was given as about 0.1 rad in the next thirty years. On the basis of the United Nations report, we could expect approxi- Genetic Loads and Their Population Effects 225 mately 10 to 400 mutations per million peo- ple. How much modification does this figure now need in order to bring it up to date? Before accurate estimates of germ-line mu- tational damage due to fallout can be ob- tained, many factors need be taken into ac- count, among them: 1. Carbon- 14, whose long half-life was not considered in this report 2. The changes in rate of testing (accord- ing to the United States Atomic En- ergy Commission, in 1958 alone the amount of fallout-producing radioac- tive material in the stratosphere was doubled by the numerous test explo- sions of nuclear weapons conducted by the United States and the U.S.S.R.) 3. The unequal distribution of fallout in different parts of the world 4. Reduction in decay taking place in the stratosphere since fallout is descending faster than expected 5. Changes in the nature of bombs tested and in the location of the test sites 6. The decrease in exposure as a result of the test ban treaty. Each month brings more of the data re- quired to estimate the fallout risk to the germ cells. Apparently, the possible dam- age has been underestimated. In 1953 the International Commission on Radiological Protection recommended — and various U.S. Government agencies adopted — 80 units as the maximum permissible concentration of strontium-90 in food. In 1958 the Com- mission recommended this maximum be lowered to 33 units, and the new value has subsequently been employed as a guideline by the U.S. Government. In principle, exposure to man-made radia- tion undoubtedly produces point mutants in the somatic and germ lines of man, but this possibility is not easy to demonstrate in prac- tice principally for two reasons: The first is that the expected point mutants are not qualitatively different from those which oc- cur spontaneously; the second is that the quantitative effect, although large in total, is small enough in any one generation to be masked by the general variability of hu- man genotypes and environment. Through the use of improved statistical methods, however, the evidence that radiation has pro- duced such genetic effects is becoming in- creasingly strong. On the other hand, clear proof that radiation can cause structural changes in human chromosomes does exist. With the recent perfection of cytological methods for studying human chromosomes and the evidence that aneusomy is a rela- tively frequent event in oocytes, it is likely that additional data will be forthcoming about the numbers and kinds of gross chro- mosomal mutations which different types and doses of radiation can induce in man. In discussing the genetic effects of low radiation doses, we recognized a danger which is not likely to be calamitous to the human gene pool; however, the very high radiation doses from a nuclear war could be disastrous, for if the whole body receives 500 r in a short period of time, the chance is 50% that the affected person will die in a few months. If the person survives this period, his life expectancy is reduced by some years, probably because of somatic mutations, and children conceived after ex- posure will be handicapped by many detri- mental mutants. It is even possible, but not probable, that in a nuclear war enough radiation would be released to destroy the human species. Finally, it should be realized that we are being constantly exposed to man-made mu- tagenic chemical substances. Although it is very probable that we are getting fewer germ-line mutations from chemical sub- stances than from radiation, more somatic mutants may be produced by chemical sub- stances than by our present exposure to radiation. 22(5 CHAPTER 16 SUMMARY AND CONCLUSIONS Cross-fertilizing species carrj a large load oi mutants in heterozygous condition. The vasl majorit) ol them are detrimental when homozygous and to a lesser extent — when heterozygous, although some heterozygotes are superior to either homozygote. Other things being equal, almost all mutants are harmful to the same degree in that each eventually causes genetic death. Mutants producing the smallest detriment to repro- ductive potential cause the greatest total amount of suffering. More detriment and more genetic deaths occur in heterozygotes than in homozygotes for rare mutants. Persistence of a mutant in the population is inversely related to its selection coefficient. Mutation is the current price paid by a population for the possibility of having a greater reproductive potential in the same or a different environment in the future. So, despite the rarity of mutants which increase reproductive potential in a given en- vironment, mutation provides the raw materials for evolution. Natural and man-made penetrating radiations are undoubtedly causing mutations in our somatic and germ cells, increasing our load of detrimental mutants. This exposure, though harmful, is most likely no threat to man's survival as a species at present, although it might be in the future should the exposure become large enough. Further research is needed to accurately assess the effects of high-energy radiations and chemical substances upon man's mutation rate and well-being. REFERENCES Auerbach, C, Genetics in the Atomic Age, Fair Lawn, N.J.: Essential Books, 1956. Background Material for the Development of Radiation Protection Standards, Report No. 1, Federal Radiation Council, Washington, D.C.: U.S. Government Printing Office, 1960. Chu, E. H. Y., Giles, N. H., and Passano, K., "Types and Frequencies of Human Chro- mosome Aberrations Induced by X-rays," Proc. Nat. Acad. Sci., U.S., 47:830-839, 1961. Crow. J. F., "ionizing Radiation and Evolution," Scient. Amer., 201:138-160, 1959. Crow, J. F., and Temin, R. G., "Evidence for Partial Dominance of Recessive Lethal Genes in Natural Populations of Drosophila," Amer. Nat., 98:21-33, 1964. Dobzhansky, Th., Evolution, Genetics, and Man, New York: John Wiley & Sons, 1955. Dobzhansky. Th., "How Do the Genetic Loads Affect the Fitness of Their Carriers in Drosophila Populations?" Amer. Nat., 98:151-166, 1964. Herskowitz, I. H.. "Birth Defects and Chromosome Changes," Nuclear Information, 3 (No. 2): 1-2, 4, 1960. Krieger, H., and Freire-Maia, N., "Estimate of the Load of Mutations in Homogeneous Populations from Data on Mixed Samples," Genetics, 46:1565-1566, 1961. Morton, N. E.. 'The Mutational Load Due to Detrimental Genes in Man," Amer. J. Human Genet., 12:348-364, 1960. Muller, H. J., "Mutational Prophylaxis," Bull. N.Y. Acad. Med., 2nd Ser., 24:447-469, 1948. Muller, H. L. "Radiation Damage to the Genetic Material," Amer. Scientist, 38:33-59, 126, 399-425, 1950. Miintzing, A., "A Case of Preserved Heterozygosity in Rye in Spite of Long-Continued Inbreeding," Hereditas, 50:377-413, 1963. Genetic Loads and Their Population Effects 227 Report of the United Nations Scientific Committee on the Effects of Atomic Radiation, New York: General Assembly Official Records: 13th Session, Suppl. 17 (A/3838), Chaps. 5-6, Annexes G-I, 1958. Selected Materials on Radiation Protection Criteria and Standards: Their Basis and Use, Joint Committee on Atomic Energy, Congress of the United States, Washington, D.C.: U.S. Government Printing Office, 1960. The Biological Effects of Atomic Radiation, Summary Reports, Washington, D.C.: National Academy of Sciences — National Research Council, 1956 and 1960. (See Reports of the Genetics Committee.) Wallace, B., "A Comparison of the Viability Effects of Chromosomes in Heterozygous and Homozygous Condition," Proc. Nat. Acad. Sci., U.S., 49:801-806, 1963. QUESTIONS FOR DISCUSSION 16.1. Do you suppose that the mutations which occur in man serve a useful function? Why? 16.2. Compare the fate of a mutational load in asexually reproducing populations that are haploid, diploid, and autotetraploid. 16.3. Discuss the effect upon the gene pool of mutants restricted to the somatic line. 16.4. Can the gene that comprises part of a detrimental equivalent also comprise part of a lethal equivalent? Explain. 16.5. Give examples of balanced and unbalanced polymorphism in the genetics of man. 16.6. What is the relation between phenotypic detriment, genetic death, and genetic persistence? 16.7. Discuss the relative importance of point mutants and gross structural changes in chromosomes to the individual and to the population. 16.8. What is the difference, in terms of mutation, between a maximum permissible dose and a doubling dose of ionizing radiation? Is any dose of any radiation safe from a mutational standpoint? Explain. 16.9. Compare the genetic composition of the mutant load caused by fallout, by med- ical uses of radiation, and by atomic reactor accidents. 16.10. Do you believe it is essential for the general public to become acquainted with the genetic effects of radiation? Why? 16.11. What are some of the beneficial uses of radiation? Are any of these based upon the genetic effects of the radiation? If so, give one or more examples. 16.12. One of the components of fallout is radioactive iodine, 1-131, which has a half- life of about a week. Discuss the genetic effects expected in the somatic and germ lines of persons exposed to fallout. 16.13. Susceptibility to leprosy may be due to a single irregularly dominant gene. S. G. Spickett notes that leprosy is increasing in some human populations that have been free of it for many generations. List some factors which may be responsible for this finding. 16.14. Is a genotype adaptive in man today, one which would have been adaptive 2,000 or 20,000 years ago? Explain. 16.15. "The danger of mutation lies primarily in the rate with which it occurs." Crit- icize this statement. 16.16. How can you explain the finding that in the genus Drosophila apparently the heaviest genetic loads occur in common and in ecologically most versatile species, whereas the lightest loads are found in rare and in specialized species and in marginal colonies of common species? Chapter *17 CHROMOSOMAL REARRANGEMENTS IN NATURE Oenothera ' The evening primrose, Oenothera (Figure 17-1 ). is a common weed found along road- sides, railway embankments, and in aban- doned fields. It exists in nature in a number of pure breeding, self-fertilizing strains — each with a characteristic phenotype. These strains can be cross-fertilized in the labora- tory. If the two strains crossed are La- marckiana and biennis the outcome in Fx is surprising. First, the F, are not all uniform in phenotype as one would expect from pre vious experience with crossbreeding two pure lines, but three distinct types (which we will call A, B, C). Second, upon self-fertiliza- tion each of these three F, types is thereafter pure breeding. If the Fi were hybrid, we would expect self-fertilization to produce recombinants and. therefore, more than one phenotype in its progeny. These two pe- culiarities are summarized in Figure 17-2 where the results obtained from similar crosses with garden peas are shown side by side. We must conclude from the Oenothera results that self-fertilizing strains cannot au- tomatically be considered pure homozygous lines, despite any contrary impression gained in Chapter 1. In order to obtain three dif- ferent genotypes in F,. either Lamarckiana, 1 Based upon work of H. DeVries. O. Renner. R. E. Cleland. F. Oehlkers, A. F. Blakeslee, J. Belling, S. Emerson, and A. H. Sturtevant. 228 or biennis or both must be heterozygous. Assume that Lamarckiana is heterozygous for a single pair of genes. If so, how can this strain produce only Lamarckiana upon self- fertilization'.' To do this would require that the heterozygote produces only heterozy- gote progeny. But suppose that self-fertili- zation does, as expected, produce the two homozygotcs, both of which are lethal. ( Recall that for yellow mice — p. 69 — only one homozygote is lethal; the other is viable. In the present case the two different alleles would both have to act as recessive lethals. ) This hypothesis which predicts that one half of the zygotes die before becoming mature Lamarckiana, is supported by the finding that approximately one half of the ovules regularly fail to produce seed upon self- figure 17-1. Oenothera. (Courtesy of R. E. Cleland.) Chromosomal Rearrangements in Nature 229 fertilization — evidence that in nature La- marckiana is a permanent heterozygote in this respect, with a balanced lethal system. In this case both lethals kill the individual sometime before seed formation, in fact, the lethal kills at the time of fertilization or very soon thereafter, being in effect a zygotic lethal (Figure 17-3). Recall that some plants, including Oeno- thera, have a haploid gametophyte genera- tion. Permanent heterozygosis could be maintained also, if one allele were lethal to the male gametophyte and the other to the female (Figure 17-3). Consequently, game- tophytic lethals can also provide a balanced lethal system which prevents half of the ovules from producing seed. We have al- ready seen an example of this kind of lethal in the self-sterility gene in Nicotiana (p. 60). In general, all strains of Oenothera found in nature, including biennis, have en- forced heterozygosity due to the zygotic and gametophytic lethals which produce bal- anced lethal systems. Does a balanced lethal system explain why the phenotype of Lamar ckiana, for ex- ample, is the only one produced in the prog- eny after self-fertilization? Since all sexual organisms so far studied have many pairs of genes, it would not seem reasonable that Oenothera has only a single pair of recessive PEAS Tall Dwarf I t Tall Dwarf I * Tall Dwarf \ / F Tall / \ F 3 Tall; 1 Dwarf OENOTHERA Lamarckiana Biennis t I Lamarckiana Biennis ♦ t Lamarckiana Biennis F A B C 'III F A B C ZYGOTIC LETHAL # © figure 17-2. Comparative breeding results from garden peas and Oenothera. GAMETOPHYTIC LETHAL figure 17-3. Balanced lethal systems that enforce heterozygosity. lethal genes whose manifold (pleiotropic) effects produce the entire phenotype. It is more likely that many gene pairs exist which form a single linkage group, so that the dip- loid individual has one genome whose genes are all linked to one recessive lethal and another genome whose genes are all linked to the allelic lethal. In other words, Lamarckiana behaves as though it contains two complexes of linked genes. Within a strain these genes are com- pletely linked by some mechanism that pre- vents recombination, leaving the gametes with only two kinds of genotypes. The two gene complexes are so constant in natural populations of a strain that in the case of Lamarckiana they are given the names gau- dens and velans, and identified as gaudens. velans; the biennis strain as described by its 230 ( II M'TER 17 BIENNIS LAMARCKIANA Albicans IAI Gaudens IGI Rubens IRI Velans (VI A R 1 @) GAMETES ((3) ) G V GAMETES ((o)) G GAMETES (O)) G / \ FIGURE 17-4 {above). Balanced lethal gene complexes in O. biennis and O. Lamarckiana. figure 17-5 (below). Linkage groups in hy- brids from interracial crosses. m P B Sp Cu FLAVENS-CURVANS FLAVENS-PERCURVANS FLAVENS-FLECTENS FLAVENS-VELANS RUBENS-FLAVENS RUBENS-CURVANS CURVANS-VELANS RUBENS-VELANS gene complexes is albicans. rubens. Figure 17-4 shows how these balanced lethal gene complexes are distributed generation after generation in biennis and Lamarckiana. All the recessive lethal alleles in either of the two strains cannot be identical to those in the other; if they were the F, from crossing them would consist of only two different phenotypes, whereas three types are actually obtained. We can conclude, therefore, that the balanced lethal system generally found in Oenothera involves either a multiple allelic series or several pairs of genes or both. Each of the three different Fi hybrids obtained from crossing Lamarckiana with biennis breeds true upon self-fertilization, showing that each hybrid contains two com- pletely linked gene complexes. This con- clusion may or may not be true, however, of the breeding behavior of other hybrids obtained from interracial crosses. This ambivalence is illustrated in Figure 1 7-5 with the gene complexes present in the dif- ferent hybrids shown at the left. The dis- tribution of the various genetic markers (top of Figure 1 7-5 ) in the gametes of these hybrids was determined from breeding tests. For example, the curvans. velans hybrid pro- duced only two kinds of gametes though heterozygous for all these marker genes, the markers behaving as if they were all com- pletely linked. On the other hand, the Havens. velans hybrid produced four kinds of gametes. The genes R, m, and P (all still linked to each other) segregated independ- ently of the genes B and Sp (both still linked to each other), so that half of the gametes contained one of the two parental combina- tions, the other half carried one of the two recombinations. In this case, therefore, genes which belonged to a single linkage group in the parent races behaved as two linkage groups during the gametogenesis of their hybrid. Since 50% recombination oc- curred in gametogenesis of the hybrid, these results are not really explained by postulat- Chromosomal Rearrangements in Nature 231 ing that flavens (or velans) is actually a single linkage group which cannot undergo crossing over with its partner gene complex in the parent race, but which can do so when its partner is velans (or flavens). Tests of the hybrid containing flavens. curvans showed m and P still completely linked but segregating independently of 5, which was, in turn, segregating independ- ently of Sp and Cu, so that in this case three linkage groups existed. Perhaps more link- age groups would have been found with additional genetic markers. In all cases, however, a given hybrid combination always showed the same linkage groups in its game- togenesis. Because at least three linkage groups can be identified in certain interstrain hybrids (even though these act as one in the self- fertilizing parental strain), it is expected that the diploid Oenothera has at least three pairs of chromosomes and cytological examination confirms this genetic expectation — all of the Oenothera strains discussed in this chapter having seven pairs of chromosomes. (Oeno- thera gigas, the triploid mentioned on p. 151, has 2 1 chromosomes. ) If the balanced lethal system is based upon a single pair of genes located on a single pair of homologs, this pair of chromosomes must be hetero- zygous in viable progeny. But this hetero- zygosity would not be expected for the other six pairs of chromosomes, if they segregated independently. Consequently, all gametes of O. biennis, for example, which carry the albicans complex recessive lethal should sim- ilarly be expected to carry the rubens or the albicans homolog in each of the other six cases of independent segregation. However, this distribution is not found. We could then suppose that each of the seven chro- mosome pairs is heterozygous for a different recessive lethal. Upon self-fertilization, such a genotype would produce only viable Fi like the parent. Since this explanation predicts that only about (U)7 of all ovules should develop as seeds, it cannot be the correct one for Oenothera in which, as men- tioned, about 50% of all ovuies mature into seeds. A clue to the orderly segregation oi com- plete gene complexes in Oenothera may be found by cytological study of meiosis. The typical self-fertilizing Oenothera in nature does not form seven separate bivalents as expected, but, as seen clearly at metaphase I, forms a closed circle of 14 chromosomes synapsed end to end (Figure 17-6). At ana- figure 17-6. Circle of 14 chromosomes in Oenothera. Chromosome number is clear in upper cell where the circle has broken open. (Courtesy of R. E. Cleland.) 232 CHAP II R 17 figure 17-7. Ralph E. Cleland points to zig-zag chromosomal arrangement at the start of anaphase I of an Oenothera having a circle of 14 chromosomes at metaphase I. {Photograph, 1959, courtesy of The Calvin Company.) figure 17-8. thera. Manner of chromosome segregation during meiosis of Oeno- Chromosomal Rearrangements in Nature 233 phase I, moreover, adjacent chromosomes in the circle go to opposite poles of the spindle, so that at the start of the separation the chromosomes assume a zigzag arrange- ment (Figure 17-7). If we assume that paternal and maternal chromosomes alter- nate in the circle, then all paternal chromo- somes would go to one pole and all ma- ternal chromosomes to the other. The com- plete linkage of all genes in a complex would be explained by such chromosome segrega- tion (if crossing over is rare), and the gam- etes produced by an individual would be identical to those which united to form it (Figure 17-8). If in an alternate segregation procedure maternal and paternal genomes separate, a circle should always contain an even number of chromosomes. Moreover, we could pre- dict that when one gene complex no longer behaves as a single linkage group, it will also no longer form a single circle of four- teen chromosomes with the other gene com- plex. Fourteen chromosomes can be ar- ranged fifteen different ways in circles (com- posed of even numbers of chromosomes) and pairs as shown in Figure 17-9. Indeed, when various race hybrids are made, all fif- teen types and no others are found at meta- phase I — any particular hybrid always form- ing the same meiotic configuration. (The top cell in Figure 17-7 shows an inner circle of four and an outer circle of ten chromosomes.) If what has been supposed about alternate segregation is true, it should also follow that even though alternate chro- mosomes within a circle show complete link- age with each other, such linkage groups should segregate independently of other link- age groups consisting of chromosomes either in separate circles or in separate pairs. This expectation can be tested by comparing the number of genetically determined linkage groups in the different hybrids of Figure 17-5 with the chromosome arrangements Q 14 Q 10, 2 Po.rs O10' O 4 O 6- O4- 2Pairi (7) 8, (JJ) 6 Q 8, 3 Pairs O •• O <• O 4 O 4- O 4- 3 *- Q\ 12, 1 Pair Q 6, 4 Pairs 0 8, Q4, 1 Pair O 4' 5 Pairs 0 6, (~) 6, 1 Pair 7 Pairs O "' O 4' O 4' } Pair Q=CIRCIE figure 17-9. Circle and pair arrangements possible for Oenothera chromosomes. seen cytologically during their meiosis. Such a comparison reveals that the number of separate groups of chromosomes observed in meiosis is always equal to, or greater than, the number of linkage groups detected ge- netically. In fact, whenever a sufficient number of genetic markers are used, the number of linkage groups always equals the number of chromosome groups. Although the preceding discussion indi- cates that a rather unique segregation of alternate chromosomes in a circle and the presence of balanced lethal systems can ex- plain most of the unusual genetic behavior of Oenothera, other matters still need ex- planation. What causes these chromosomes to form circles in the first place? A clue to this, contained in the observation made on p. 172 (see also Figure 12-6), is that two pairs of nonhomologs will be associated as a double tetrad during synapsis if a recip- rocal translocation involving them is present in heterozygous condition. Figure 17-10 illustrates this situation in Oenothera. All Oenothera chromosomes are small, are roughly the same size, and have median centromeres. To help us identify homol- 234 CHAPTER 17 1 2 13 3 4 r? 3 3 I 5 5 FIGURE 17-10. Heterozygous reciprocal trans- locations and circle formation. Chromatids arc not shown. OgOUS chromosomes, the ends of dilTerent chromosomes in a genome are given dif- ferent numbers. Suppose, at some time in the past, a dicentric reciprocal translocation occurred between the tips marked 2 and 3 (Figure 17-10A, B). This rearrangement in heterozygous condition ( B ) would pro- duce an X-shapcd configuration at the time of synapsis in prophase I (C) and a circular appearance at metaphase I-early anaphase 1 (D). In this way a circle of four chro- mosomes would be produced. If a second reciprocal translocation oc- curs between any chromosome arm in a cir- cle of four and an arm of some other pair of chromosomes, a circle of six chromosomes will form in the individual heterozygous for both reciprocal translocations. This type of formation is illustrated in Figure 17-10D, E; D shows the configuration before arms 4 and 5 have exchanged, E shows the circle of six produced in meiosis after this ex- change. Still larger circles can be formed by successive interchanges of this type; six such interchanges are required to form a circle of 14 chromosomes. The presence of reciprocal translocations in heterozygous condition could explain how various sized circles containing even numbers of chromo- somes are produced in Oenothera. Although the cytogenetic analysis of Oeno- thera is known in some detail, the picture is, however, not yet complete. One of the questions remaining is: What is the mech- anism whereby alternate chromosomes in a circle proceed to the same pole during meio- sis? No fully acceptable answer to this ques- tion has yet been given. A second question stems from the fact that almost all the dilTer- ent strains or races of Oenothera found in nature form a circle of 14. Are the six translocations involved the same in all races? No — for if they were, viable hybrids be- tween races would form either circles of 14 or seven separate chromosome pairs at meio- Chromosomal Rearrangements in Nature 235 sis. That all the configurations in Figure 17-9 are found in meiosis of such hybrids must mean that different gene complexes differ from each other in the specific ways that their chromosome arms have become translocated. Many thousands of ways are possible for 14 ends to be arranged in seven groups of two. How can we determine the number of these different arrangements oc- curring in nature? We can start by choosing a particular gene complex — calling it the "standard'" — and considering its chromosome ends to be 1-2, 3_4, 5_6, 7-8, 9-10, 11-12, and 13-14. Normally, that is, in nature, this complex would form a circle of 14 with the other gene complex, which would therefore have no chromosome with the same pair of num- bered ends as any chromosome in the stand- ard complex. Proceeding further, we form a series of interracial hybrids with the stand- ard as one of the complexes and score the meiotic chromosome arrangements of the hybrids. Suppose in one case that the hy- brid forms five pairs and a circle of four. This result must mean that the ends of 5 chromosomes are in the same order in the complex under test as in the standard, but that they are in a different order in the re- maining two chromosomes. Although there was previously no reason to assign ends 1-2 and 3-4 of the standard complex to any par- ticular chromosomes, we can presently assign these ends arbitrarily to the two standard chromosomes in the circle of four. The chromosomes in the circle from the complex under test can then be called 2-3 and 4-1 (or 2-4, 1-3). In this way the composition of ends of two chromosome pairs is specified permanently. The top of Figure 17-11 shows the standard and tested complexes (of our example) synapsed according to iden- tical numbers. Call the complex just tested A. Suppose another complex, B, is made hybrid both A. COMPLEXES DIFFERING BY ONE RECIPROCAL TRANSLOCATION 1.2 3.4 5.6 7.8 9.10 11.12 13.14 \ / \ II II I I I I I I 23 41 56 7.8 9.10 11 12 13.14 B COMPLEXES DIFFERING BY SIX RECIPROCAL TRANSLOCATIONS 1.2 3.4 5.6 7.8 9.10 11.12 13.14 2.3 4.5 6.7 8.9 10.11 12.13 14.1 C. MURICATA RACE'S ACTUAL COMPOSITION OF Q 14 1.2 3.4 6.5 13.12 7.11 10.9 8.14 \ / \ / \ / I / I / \ / 1 2.3 4.6 5.13 12.7 11.10 9.8 14.1 A and B are theoretical figure 17-11. Arrangement of chromosome ends in different Oenothera complexes. with the standard and with A. The meiotic configuration of the hybrid may specify other ends of A, B, and the standard com- plexes. Such procedures can be carried out until all of the standard's chromosomes are specified and the complete order of all four- teen ends determined for any other complex. In this manner, we can verify that a circle of fourteen is produced in many different ways in nature; a hypothetical and an actual example is shown in the central and lower parts of Figure 17-11. In fact, of 350 com- plexes analyzed, more than 160 different segmental arrangements have been found. All these results are consistent with the hy- pothesis that during the course of evolution, the ends of Oenothera chromosomes have been shuffled many times in different ways by reciprocal translocation. A most convinc- ing test of the reciprocal translocation inter- pretation would be the ability to predict the meiotic chromosomal arrangement to be found in a hybrid not yet formed. This type of predicting has been done many times and all such expectations have been verified. At various points in this discussion Oeno- thera's behavior has seemed exceptional, ap- 236 ( II MTI.R 17 parently violating our concepts of pure lines and independent segregation. More com- plete analysis has shown, however, that ( Oenothera's failure to behave as expected was due to the operation of other, already known, genetic events. Oenothera is an ex- ception which should be treasured; for in the exact correspondence between its atypical genetics and its atypical cytology, it is an outstanding example of the validity of the chromosome theory of transmission genetics. Three aspects of the cytogenetic behavior of Oenothera are disadvantageous under many circumstances: reciprocal transloca- tions; recessive lethals; and self-fertilization. By combining all three of these disadvan- tages in one plant, however, Oenothera's sur- vival value is probably greater than it would be without them. The self-fertilization mechanism involves bringing the stigma down to the level of the anther, so that a much heavier pollination is attained than would be likely were the plant pollinated by insects. This self-fertilization mechanism offsets the 50% mortality due to balanced lethals. These lethals, together with the re- ciprocal translocations and alternate segre- gation, prevent the homozygosity usually consequent to self-fertilization, enforce het- erozygosity, and produce maximum hybrid vigor. The great survival value of Oenothera is demonstrated by the distribution of this genus: it can be found from the southern tip of South America to the far reaches of Northern Canada and from the Atlantic Ocean to the Pacific. It is interesting to note that the most numerous sections of the genus and those which have ranged the farthest are the ones with large circles, bal- anced lethals and self-pollination. Drosophila Although reciprocal translocations have played an important role in the evolution of Oenothera, it might be claimed that this X >x >,< >,< >r '(* T T V s >> ^i4 figure 17-12. Chromosome configurations in several Drosophila species. genus is an unrepresentative example of the importance of chromosomal rearrangements in evolution because its cytogenetic behavior is so unorthodox. Hundreds of different species of Drosophila occur in nature. These species can be compared ecologically, mor- phologically, physiologically, and biochem- ically. For those species able to interbreed, recombinational genetic properties can also be compared; banding patterns of salivary gland chromosomes and the appearance of chromosomes at metaphase of different spe- cies are additional areas of comparison. After all available information of this kind is gathered, it is possible to arrange the chro- mosomes of various species on a chart so that those closest together are more nearly related in descent — evolution — than are those farther apart.- This arrangement is illus- trated in Figure 17-12 which shows the karyotype — the haploid set of chromosomes - Based upon work of C. W. Metz and others. Chromosomal Rearrangements in Nature 237 at metaphase — including the X but not the Y chromosome for different Drosophila spe- cies or groups of species. The karyotype of the melanogaster species group, for ex- ample, is shown in row 2, column 1; the bottom chromosome is the rod-shaped X, the two V's are the two large autosomes (II and III), and the dot represents the tiny chromosome IV. In the other karyotypes, whole chromosomes or chromosome arms judged to be homologous are placed in the same relative positions. What can be learned from a comparison of these karyotypes? Since the amount of detail in a metaphase chromosome is limited basically to size and shape, one cannot expect to discern any small-sized rearrangements at this stage. Accordingly, regardless of their importance, small rearrangements involving duplication, deficiency, shift, transposition, inversion, and translocation cannot be detected on the chart. Even a large paracentric inversion is undetected at metaphase, since it does not change the shape of the chromosome. Other gross structural changes, however, can be detected. In row 4 the chromosome pat- terns in columns 2 and 3 seem identical, except that a pericentric inversion has changed a rod to a V, or vice versa. (Peri- centric inversions always change the rela- tive lengths of the arms when the two breaks are different distances from the centromere.) Compare the karyotype for melanogaster (row 2, column 1 ) with the one to its right (row 2, column 2). A V-shaped autosome in melanogaster appears as two rods in its evolutionary relative. (Note also that the dot chromosome is missing.) In the next karyotype to the right (row 2, column 3), two rods have combined to form a V that is different from either of the two V's in melanogaster. Other examples in this chart indicate that two rod-shaped chromosomes have formed a V-shaped chromosome, or a V has formed two rods. Consider first how a V can origi- nate from two rods (Figure 17-13). Re- call that a rod-shaped chromosome typically has two arms, though one is very short. The short arm may not be noticeable at meta- phase or anaphase; however, its presence may be demonstrated either cytologically at an earlier or later stage of the nuclear cycle, or genetically by studying genetic recom- bination. Suppose two rods are broken near their centromeres, one in the long arm of one chromosome, the other in the short arm of the other chromosome. If the long acen- tric arm of the first chromosome becomes joined to the long centric piece of the sec- ond, a V is formed. Notice that this union involves the joining of two whole or almost- whole arms in a eucentric half-translocation. The remaining pieces may join together to form a short eucentric chromosome, thereby completing a reciprocal translocation; or they may not join. In either instance, if the short pieces are lost in a subsequent nuclear /* \ HALF (OR RECIPROCAL) TRANSLOCATION /\ figure 17-13. Formation of a V-shaped chro- mosome from two rod-shaped chromosomes. RECIPROCAL TRANSLOCATION PARACENTRIC DELETIONS figure 17-14. Formation of two rod-shaped chromosomes from a V-shaped chromosome and a Y chromosome. division and the number of genes lost is small enough, the absence of these parts may be tolerated physiologically by the or- ganism. The reverse process, the formation of two rods from a V, necessitates the contribu- tion of a centromere from some other chro- mosome. In Drosophila, this second chro- mosome may be the Y (Figure 17-14). Suppose the V is broken near its centromere and the Y is broken anywhere. Should a eucentric reciprocal translocation follow, two chromosomes would be produced, each hav- ing one arm derived predominantly from the Y. If subsequent paracentric deletions oc- cur in these Y-containing arms, rod shapes will result, thereby completing the change from a V to two rods. Note that almost every part but the centromere of the Y chro- CHAPTER I 7 mosome is eventually lost in this process. But this loss may have little or do disadvan- tage to the Drosophila. since the Y carries relatively few loci and is primarily concerned with sperm motility. For example, this series of mutations may be initiated in the male germ line, producing two chromosomes — each containing part of the Y. Deletion of Y parts can occur without detriment if these chromosomes happen to enter the fe- male germ line; they may stay in the male germ line provided that a regular Y chro- mosome is included in the genotype in due time. The small IV chromosome in melano- gaster, whose monosomy is tolerated in either sex, may also contribute a centromere in the process of changing a V to two rods by an identical or similar series of mutational events. Karyotype comparisons of Drosophila confirm the expectation (Chapter 12), that whole arm translocations are able to survive in natural populations. Such rearrange- ments and pericentric inversions are ex- tremely useful in helping us establish evolu- tionary relationships among different species. But it should be emphasized that this kind of information by itself does not prove that: 1 . The formation of different species is a primary consequence of the occurrence of these rearrangements 2. These rearrangements are of secondary importance in species formation 3. These mutational events occur after species formation is complete. As exemplified by Oenothera and Dro- sophila, we have seen that gross chromo- somal rearrangements of various types have persisted in the evolutionary course of differ- ent groups of organisms. For this reason it would perhaps be wise at this point to retrain from predicting — except generally as in Chapter 12 — which, if any, structural changes might be associated with the evolu- tion of other particular groups of organisms. Chromosomal Rearrangements in Nature 239 SUMMARY AND CONCLUSIONS Although the cytogenetics of Oenothera has several unusual aspects, present knowledge renders these differences quite understandable; consequently, Oenothera provides an outstanding confirmation o\ the validity of the chromosomal basis for genetic material. Evolution in this genus is intimately associated with self-fertilization, balanced lethals, and numerous reciprocal translocations. Pericentric inversions (which change chromosome shape) and whole arm reciprocal translocations (which lead to changes in chromosome number) have been frequent in the past evolutionary history of Drosophila. REFERENCES Cleland, R. E., "A Case History of Evolution," Proc. Indiana Acad. Sci. (1959), 69: 51-64, 1960. Cleland, R. E., "The Cytogenetics of Oenothera," Adv. in Genet., 11:147-237, 1962. Patterson, J. T., and Stone, W. S., Evolution in the Genus Drosophila, New York: Mac- millan. 1952. White. M. J. D., "Cytogenetics of the Grasshopper Moraba scurra, VIII." Chromosoma, 14:140-145, 1963. Hugo De Vries (1848-1935), pio- neer in the study of mutation and Oenothera genetics. (From Genet- ics, vol. 4, p. 1, 1919.) 240 CHAPTER 17 QUESTIONS FOR DISCUSSION 17.1. Whal evidence can you present for saying that the genes which make up the balanced lethal system in Lamarckiana arc different from those in biennisl 17.2. Discuss the following statement: "All evening primroses found in nature are constant hyhrids." 17.3. With respect to chromosomes, how does the origin of a circle differ from the origin of a ring? 17.4. Can a circle contain an odd number o\ chromosomes? Explain. 17.5. What new investigations regarding the genetics and or cytology of Oenothera has this chapter suggested to you? 17.6. List the genetic principles you could have deduced had Oenothera been the only organism studied so far. 17.7. If this chapter contains no new principles of genetics, why do you suppose it was written? 17.8. Curly-winged Drosophila mated together always produce some non-curly off- spring. Plum eye-colored flies mated together always produce some non-plum offspring. But, when flies that are both curly and plum are mated together, only flies of this type occur among the offspring. Explain all three kinds of results and define your symbols. 17.9. (a) Draw a diagram representing a heterozygous whole-arm translocation in Drosophila at the time of synapsis. Number all chromosome arms involved, (b) What would be required for a mating between two flies with this constitu- tion to produce offspring flies only of this type? 17.10. Do you suppose that the preservation of heterozygosity has an adaptive ad- vantage in Oenothera? In other organisms? 17.1 1. Discuss the evolutionary flexibility of the genus Oenothera and Drosophila. 17.12. Is the balanced lethal system in Oenothera part of its genetic load? Explain. If so, are the lethals components of a balanced load or a mutational load? Explain. 17.13. Compare the genetic effects of ionizing radiation on populations of Oenothera and Drosophila. 17.14. Explain how a Drosophila zygote formed with a sperm carrying a centric, grossly-deleted Y chromosome can develop into a fertile male. Chapter *18 RACES AND THE ORIGIN OF SPECIES I 'n cross-fertilizing species, differ- ent individuals in a population are heterozygous for different genes (see Chapter 16), even though the gene pool is at equilibrium with the factors that cause shifts in gene frequency — namely, mutation, selection, drift, and migration. In other words, in reaching genetic equilibrium, all the members of cross-fertilizing popula- tions do not eventually become homozygotes, nor do they all become heterozygotes. Such populations, therefore, do not become either genetically pure or uniform with the passage of time. Although any given population is poly- morphic for some genes, it is not necessarily polymorphic with regard to a particular gene. For example, Indians in South America are almost all of O blood type, being homozy- gous (//) in this respect, but have a poly- morphic pool with respect to other genes. Moreover, an allele, like IB, may be rare or absent in one population, as in certain North American Indians, and relatively fre- quent in the gene pool of another population, as in central Asia. Thus, populations lo- cated in different parts of the world may differ both in the types and frequencies of alleles carried in their gene pools. For many purposes it is desirable to identify a popula- tion with certain gene pool characteristics as a race. Races An investigator may choose to define races only according to the distribution of the //; 241 gene for ABO blood type. He might define populations that do or do not contain I': in their gene pool as different races. On this basis there would be only two races of man, the South American Indian (without I11) and all the other people (with lB in their gene pool). On the other hand, an investigator may decide to define races on the basis of the relative frequency of i and //; in the popula- tion. The frequency of these alleles in the gene pool has been determined for many populations all over the world. The results show that in western Europe, Iceland, Ire- land, and parts of Spain, three-fourths of the gene pool is i, but this frequency begins to decrease as one proceeds eastwardly from these regions. On the other hand, IB is most frequent in central Asia and some popula- tions of India but becomes gradually less and less frequent as one gets farther away from this center. Since the change in frequency of these alleles is gradual, any attempt to sharply separate people into races having different gene frequencies would be arbitrary. In practice, therefore, the number of races recognized is a matter of convenience. For some purposes separating mankind into only two races is adequate; for other reasons, as many as two hundred have been recognized. As a rule, most anthropologists recognize about half a dozen basic races but may in- crease the number to about thirty when con- sidering finer population details. Regardless of the number of races defined, however, each is best characterized according to the genes it contains. Since the people in a population are either A, B, AB, or O in blood type and intermediates do not occur, no average genotype exists for the ABO blood group, nor is there an average geno- type for any other polymorphic gene. Be- cause a population has no average genotype. a race should be defined according to the relative frequency of alleles contained in its gene pool. Without an average genotype, 242 CHAPTER 18 a race cannot have an average phenotype; accordingly, it is futile to try to picture a typical (average) member of any race. Other blood traits, besides ABO blood group, whose genetic basis is understood. are also useful in characterizing races. In fact, it is valid to utilize any phenotypic difference due to a genetic difference. For example, in delimiting races one can employ certain genetic differences in color of hair. eyes, and skin, and differences in stature and head shape; one should avoid using pheno- ls pic ditTerences whose genetic basis is un- proved, for the environment itself can cause phenotypic differences (Chapter I ). Re- member also that the same phenotypic result may be produced by different genotypes be- cause of gene interaction in dominance and epistasis (Chapter 4). Knowledge of the distribution of genes for ABO blood types in different populations provides important information to geneti- cists, anthropologists, and other scientists. To what can the different distributions be attributed? Since people do not choose their marriage partners on the basis of their ABO blood type, and since there does not seem to be any pleiotropic effect making persons of one blood type sexually more attractive than those of another, it is very likely that mating is at random with respect to ABO genotype. However, in other respects some evidence indicates that different ABO geno- types do not have the same biological fitness. Differential mutation frequencies can also explain part of the differences in gene dis- tribution. During the past few thousand years the greatest shift in ABO gene fre- quencies of different populations has prob- ably been the result of genetic drift and migration. In fact, the paths of past migra- tions can be traced by utilizing — along with other information — the gradual changes in the frequencies of ABO and other blood group genes in neighboring populations. It has already been mentioned (p. 209) that different paracentric inversions are found in natural populations of D. pseudo- obscura. All oi these flies are very similar phenot\ pically, even though their chromo- somal arrangements are different. Sample populations of this fly in the southwestern part of the United States (Figure 1S-1) have been studied to determine the relative frequency of these inversions." California populations proved to be rich in the inver- sion types called Standard and Arrowhead. Eastward, in nearby Arizona and New Mex- ico, the populations contain relatively few Standard and Pikes Peak chromosomes, most chromosomes having the Arrowhead arrangement. Finally, in still more easterly Texas, one finds almost no Standard and some Arrowhead with most chromosomes being of the Pikes Peak type. The shift in the frequency and type of inversions in the three different geographic regions cannot be explained as the result of differential mutation, since the spontaneous mutation rate for inversions is extremely low. Moreover, since there is no indication that the gene flow among these populations has changed appreciably in the recent past, migration rates have probably had a rela- tively small influence upon genotypic fre- quencies; there is also no indication that genetic drift has had a major role in causing the differences in inversion frequency in the three areas. These observations lead us to suppose that the primary basis for these pop- ulation differences lies in the different adap- tive values which different inversion types confer on individuals in different territories. Despite the absence of any obvious morpho- logical effects, these inversions prove to have different physiological effects in laboratory tests; different inversion types survive best in different experimental environments. Since 1 Bused upon work of Th. Dobzhansky and col- laborators. Races and the Origin of Species 243 figure 18—1. Distribution of inversion types in D. pseudoobscura collected in the South- western United States. (After Th. Dohzhan- sky and C. Epling.) these inversion types show different adaptive values in the laboratory, it is reasonably cer- tain that they do so in nature too. Accord- ingly, natural selection is primarily responsi- ble for the inversion differences among the three geographic populations, which can be defined as three different races. Similar results have been obtained with three California races of the cinquefoil plant species, Potentilla glandulosa, which live at sea level, mid-elevation, and the alpine zone. The sea level race is killed when grown in the alpine environment, whereas the alpine race grown at lower elevations proves less resistant to rust fungi than the lower-eleva- tion races. Such experiments show that dif- ferent races are adapted to their own habitats but not to others. The inorganic and organic environment — including its organisms — is different in different parts of the territory occupied by a species. Clearly, then, no single genotype will be equally well adapted to all the different environments encountered within a particular territory. One way in which a cross-fertilizing species can attain maximal biological fitness as a whole is to remain genetically polymorphic and sepa- rate into geographical populations or races which differ genetically. Whenever, as in all of the examples dis- cussed so far, different races of a cross- fertilizing species occupy geographically sep- arate territories, they are said to be allo- patric; different races occupying the same territory are said to be sympatric. In the absence of geographical separation, what factors operate to keep sympatric races from hybridizing to become one race? One may find the answer by considering the fate of races — originally allcpatric — which have be- come sympatric, a kind of change which has occurred in man. Several thousand years ago, mankind was differentiated into a num- ber of allopatric races. With the develop- ment of civilization and improved methods of travel, many of these races have become sympatric. Gene exchange in the now- sympatric races, however, is sometimes in- hibited by social and economic forces, so that some of these races continue to main- tain their identity. Domesticated plants and animals provide another example of what can happen when allopatric races become sympatric. Many different breeds, or races, of dogs originally allopatric are now found living in the same locality. Yet these now- sympatric races do not exchange genes with sufficient frequency to form a single mongrel breed, or race, because their reproduction is controlled by man. It should be realized that, under other circumstances, such allo- patric races which become sympatric can form a single polymorphic race via cross- breeding. 244 CHAPTER 18 Speciation Involving One Species A species of cross-fertilizing organisms usu- ally consists of a Dumber of races adapted to the different environments of the terri- tories thej occupy. All these races are kept in genetic continuity by interracial breeding and hybrid race types, so that the species, as a whole, has a single gene pool containing no portion completely isolated from any other. On the other hand, different cross- fertilizing species are genetically discontinu- ous from each other. Thus, the gene pool of one species is so isolated from the gene pools of all other species that none can lose its identity via crossbreeding, or hackcross- ing subsequent to crossbreeding. Moreover, the gene pools of different species are iso- lated from each other for genetic — not merely environmental — reasons. The formation of new species, speciation, has occurred frequently in past evolution; since evolution is continuing, new species are still being formed. The speciation mech- anism considered most common for cross- fertilizing individuals involves the production of two or more species from a single one. How can this come about? Hypothetically, one can start with a single panmictic, genetically-polymorphic species. Since environments vary we will assume that different populations occupy different por- tions of a territory and, although enough interpopulation breeding takes place to form one gene pool, most of the breeding is intra- population. If, in the course of time, two (or more) of these populations diverge ge- netically— each one uniquely adapted to its own territory — these populations become different races of the same species. The dif- ferences in the gene pools of these two races may increase more and more because of mu- tation, natural selection, and genetic drift. As this differentiation process continues, the genes which make each of the races adaptive in their own territories may, by their mani- fold phenotypic effects, make matings be- tween the two races still less likely to occur or may cause the hybrids of such matings to be less adaptive than the members of either parent race. Accordingly, partial re- productive isolation may be initially an acci- dental or an incidental byproduct of the adaptability of genotypes to a given en- vironment. The greater this effect, however, the greater we would expect the selective advantage to be of genes which increase the reproductive isolation between two diverging races further still. If races continued to di- verge genetically in this way, they would eventually form separate and different gene pools, and instead of being two races of the same species would become two different species. Note that speciation is an irrevers- ible process; once a gene pool has reached the species level, it can never lose its identity via cross breeding with another species. In this generalized account of how specia- tion usually occurs races have acted as in- cipient species. But remember that under other circumstances two races can also cross- breed to become a single race. For exam- ple, although several thousand years ago dif- ferent allopatric populations of human be- ings were definitely different races which might have formed different species had the same conditions of life continued, some of these races subsequently merged into one race because civilization and migration facili- tated crossbreeding. Gene exchange between races can be hin- dered in several ways. Those barriers lead- ing to complete reproductive isolation in- clude the following: 1. Geographical. Water, ice, mountains, wind, earthquakes, and volcanic activity may separate races. 2. Ecological. Changes in temperature, hu- midity, sunlight, food, predators, and parasites may alter or completely change a race's habitat. Races and the Origin of Species 245 3. Seasonal. Seasonal changes may cause different races to become fertile at differ- ent times even if their territories overlap, or if they are sympatric. 4. Sexual or ethological. Intrarace mating, due to preference or domestication ef- fected by man. 5. Morphological. Incompatibility of the sex organs between some races. 6. Physiological. Failure of a race's sex cells to fertilize those of another, so that the hybrid zygote is formed infrequently, or not at all. 7. Hybrid inviability. Even when formed, the development of hybrid zygotes may be so abnormal that it cannot be com- pleted. 8. Hybrid sterility. A possibility even if hybrids complete development and are hardy. Although geographical, ecological, and sea- sonal differences do not automatically initi- ate genotypic differences, they furnish the environmental variations which select from the available genotypes those which are adaptive; that is, those with the greatest re- productive potential under the given condi- tions. Of course, mutation must provide the raw materials for natural selection; since no single genotype is equally well adapted to all conditions, different races come to contain different genotypes. The remaining barriers listed can complete reproductive isolation. The many genes by which two incipient species differ may produce seasonal, sexual, morphological, and physiological barriers. Hybrid inviability may result from develop- mental disharmony caused by the presence of two genetically different genomes in each cell. Although hybrid sterility can be caused by such genetic action, it also results when two races become quite different with respect to gene arrangement — because of structural changes within and between chromosomes — so that during meiosis, synapsis between the two different genomes in the hybrid is irreg- ular. Improper pairing causes abnormal segregation, which results in aneuploid meio- tic products. Recall that aneuploidy in pol- len is lethal, and that aneuploid gametes in animals usually result in dominant lethality of the zygotes they form. Consequently, reproductive isolation can be based upon either genetic activity or chromosomal be- havior, or both. It seems reasonable that the more morpho- logically divergent two forms are, the more likely it is that they will differ physiologically and that these differences will have orig- inated in very different and isolated gene pools. Simply by comparing horse and mouse morphologies, one certainly expects them to be different species; thus the occur- rence of morphological differences is some- times a good index of a species difference. However, when the groups being compared are closely related in descent, one finds that morphology is not well correlated with re- productive isolation. For example, Euro- pean cattle and the Tibetan yak are quite different in appearance and usually are placed in different genera, but these two species can be crossed. Moreover, in Tibet, many cattle have yak-like traits, so that widely different phenotypes do not neces- sarily result in complete reproductive isola- tion between closely related species. On the other hand, consider D. persimilis and D. pseudoobscura. These two species — for- merly considered races of the same species — are so similar morphologically that they can be differentiated by their genitalia only if very careful measurements are made. Nevertheless, these two species have com- pletely isolated gene pools in nature, even where their territories overlap. Such mor- phologically similar species are called sibling species. They originated from different 246 CHAPTER 18 i aces of a single species. Sibling species are found in mosquitoes and Other insects as well as in Drosophila; they are also found in plants — among the tarweeds of the aster family and in the blue wild rye. The study of D. pseudoobscura and D. persimilis illustrates two other principles re- lating to species formation. First, any par- ticular reproductive barrier usually has a multigenic and or a multichromosomal basis; second, any two species are separated not by one but by a number of reproductive bar- riers. Although each of the barriers involved is incomplete, together they result in com- plete reproductive isolation — there being no stream of genes between the two gene pools in nature. The known differences between these two particular sibling species include: 1 . Pseudoobscura lives in drier and warm- er habitats than persimilis 2. Females accept the mating advances of males of their own species more often than they do male advance of the other 3. Pseudoobscura usually mates in the evening, persimilis in the morning 4. Interspecific hybrids are relatively in- viable and when viable, they are mostly sterile. The nature and origin of the reproductive isolation mechanisms involved in forming new species from races shows that valid species originate not by a single or simple mutation, but as the result of many different, independently occurring genetic changes. Moreover, as already noted, speciation is accomplished not merely by an accumulation of mutants which distinguish races, but also by those which contribute to reproductive isolation. Usually populations are physically separated while reproductive barriers are being built up; otherwise, hybridization would break down these barriers. Experi- mental evidence also supports our expecta- tion that natural selection acts to further the accumulation of the genetic factors promot- ing reproductive isolation between races. The preceding discussion illustrates how one species can give rise to two or more species via races which serve as incipient species.- It was stated earlier that a species has an isolated gene pool, that is, a gene pool closed to individuals of some other al- ternative condition (species). A species is expected to undergo numerous changes in its gene pool during the course of many generations. At the end of this time, is it the same or a new species? Here is an example of one type of species formation which would not be recognized by the cri- terion above because the alternative state would no longer exist. Suppose some mem- bers of the original population had been (miraculously) preserved, then we might find that they were reproductively isolated from the members of the new population. In such an event we could admit the forma- tion of a new species whose origin is de- pendent upon the "extinction" of the parent species. This type of speciation will be- come a valid subject of study once man learns how to preserve sample genotypes indefinitely. One species can give rise to another via allopolyploidy — an increase in the number of genomes present in a normally cross- fertilizing species. Mechanisms for the pro- duction of autopolyploid cells, tissues, and organisms have already been described on pp. 151-153. In the genus Chrysanthemum, species occur with 2n chromosome numbers of 18, 36, 54, 72, and 90. Thus, it appears that nine is the basic n number. In the genus Solanum (the nightshades, including the potato ) the basic n number seems to be twelve, since species of this genus are known '-' See Th. Dobzhansky, L. Ehrman, O. Pavlovsky, and B. Spassky (1964). Races and the Origin of Species 247 having 24, 36, 48, 60, 72, 96, 108, and 144 chromosomes. These two examples sug- gest that autopolyploidy has played a role in the speciation of these two genera. Auto- polyploidy, however, is not considered an important mechanism of speciation in forms reproducing primarily by sexual means, since autopolyploids having more than 2n chromosomes form multivalents at meiosis and, therefore, numerous aneuploid gametes. Autopolyploids can succeed, though, if they are propagated asexually, by budding or grafting, as in the case of the triploid apples — Gravenstein and Baldwin. Triploid tulips are also propagated asexually. Speciation Involving Two or More Species Many new cross-fertilizing species originate not only from a single species or its races, but — in relatively recent times — from hy- bridization between two or more different species, that is, via interspecific hybridiza- tion. Although interspecific hybrids pose no threat to the isolation of the gene pools of their parental species, they may form a suc- cessful, sexually-reproducing population that has its own closed gene pool. Interspecific hybrids, particularly of plants, can be con- verted into stable, intermediate types iso- lated from their parental species by three methods. The first method involves amphiploidy (allopolyploidy, see p. 155). If one species has 2n = 4 and another has 2n = 6, the Fi hybrid between them will have five chromo- somes (Figure 18-2). If the hybrid sur- vives, it may be sterile because each chro- mosome has no homolog and, therefore, no partner at meiosis. As a result, meiosis pro- ceeds as if the organism were a haploid and produces mostly aneuploid gametes. If, however, the chromosome number of the Fi hybrid is doubled — either artificially (via colchicine) or spontaneously — the individual or sector will be 2n = 10; each chromosome 2n 4 2n 6 Aneuploid -*■ Meiotic Products AMPHIPLOID Euploid(n) "*" Meiotic Products figure 18-2. Interspecific hybridization lead- ing to new species formation via amphiploidy (allopolyploidy). will have a meiotic partner; and euploid gametes of n = 5 will be formed. Upon uniting, such gametes produce 2n = 10 progeny, which are fertile and more-or-less phenotypically intermediate to and isolated from both parental species. It has been estimated that twenty to twenty-five per cent of the present flowering plant species originated as interspecific hy- brids whose chromosomes doubled in num- ber (therefore being "doubled hybrids" or amphiploids). Moreover, in the past many (or more) species originated in this way, then diverged to form different genera. Nat- urally-occurring amphiploidy was involved in the origin of cotton in the New World and in the appearance of new species of goatsbeard during the present century. In the early 1800's, the American marsh grass, S parti na alterni flora (2n = 70), was 2»s CHAPTER 18 DIPLOID HYBRID AMPHIPLOID figure 18-3 {above). Seed pods of cabbage and radish, of their hybrid and amphiploid. {After G. D. Karpechenko.) figure 18-4 (below). Distribution of Del- phinium species in California. Each species has a unique habitat. D. GYPSOPHILUM D. RECURVATUM D. HESPERIUM accidentally transported by ship to France and England and became established along- side of the European marsh grass, 5". Stricta (2n 56). By the early 1900's a new marsh grass, S. townsendii (2n = 126), ap- peared and largely erowded out the two older species. Since S. townsendii has a chromosome number equal to the sum of the diploid numbers of the older species, is fertile, breeds true, and has an appearance intermediate between the two older forms, this species is undoubtedly an amphiploid of S. alterniflora and S. stricta. S. town- sendii is so hardy that it has been purposely introduced into Holland (to support the dikes) and to other localities. Amphiploidy also can be produced arti- ficially. For example, in the greenhouse it is possible to cross radish (2n = 18) with cabbage (2n = 18) (Figure 18-3), thus producing an F] hybrid with 18 unpaired chromosomes at meiosis. If, however, the chromosome number of the hybrid doubles early enough in development, it can produce amphiploid progeny with 2n = 36 chromo- somes (containing nine pairs each from radish and cabbage). Since the amphiploid is fertile and genetically isolated from both radish and cabbage, it constitutes a new species. If each chromosome contributed to an in- terspecific hybrid is different and the hy- brid's chromosome number doubles, then each chromosome would have just one part- ner at meiosis, and segregation would be normal. Therefore, the breeding success of the amphiploid is enhanced by greater dif- ferences between the chromosomes of the two species that contributed haploid ge- nomes to the interspecific hybrid. It is not surprising, then, in hybridizing two chromo- somally similar species, that at meiosis their amphiploid produces trivalents and quad- rivalents leading to abnormal segregation and sterility. Although amphiploidy is not successful Races and the Origin of Species 249 for hybrids between similar species, there is a second way interspecific hybrids can be- come stabilized as a new species, provided that the two hybridizing species are very similar chromosomally. If the two species have the same haploid number, their Fi hybrid may have all chromosomes synapsed in pairs at meiosis. Segregation, independ- ent segregation, and crossing over may yield progeny of the hybrid whose recombinations can become stabilized in nature yet are iso- lated from either parental species. Con- sider certain species in the larkspur genus, Delphinium: D. gypsophilum is morpholog- ically intermediate between D. recurvatum and D. hesperium; all three species have 2n = 16; and the "parent" species, recurva- tum and hesperium, can be crossed to pro- duce an ¥\ hybrid. When the F! hybrid is crossed to gypsophilum, the offspring are more regular and more fertile than those produced by backcrossing the F1 hybrid with either parent species. Similarly, the progeny from crosses between gypsophilum and either of its parent species are not as regular or as fertile as are those from the cross between gypsophilum and the hybrid of the parent species. These results provide good evi- dence that gypsophilum arose as the hybrid between recurvatum and hesperium. Fig- ure 1 8-4 shows the distribution of these species in California. The third way that interspecific hybrids can become stabilized as new species is by introgression. In this process a new type arises after the interspecific hybrid back- crosses with one of the parental types. The backcross recombinant types favored by nat- ural selection may contain some genetic com- ponents from both species, may be true- breeding and, eventually, may become a new species. SUMMARY AND CONCLUSIONS A race of a cross-fertilizing species is characterized by the content of its gene pool. Each race is adapted to the territory in which it lives. Different races can be sympatric or allopatric. Races can become species by accumulating genetic differences whose end effect is genetic discontinuity — that is, the formation of isolated gene pools. Sep- aration of two gene pools is usually accomplished by a combination of different repro- ductive barriers each of which is incomplete by itself and has a polygenic and/or a polychromosomal basis not necessarily correlated with morphological differences. It is generally recognized that most cross-fertilizing species arose from the further differentiation of races. Occasionally a new species can arise via autopolyploidy, and it is possible that a new species can also arise by the gradual change of one species as a whole into another species. Two (or more) species can give rise to a new one after interspecific hybridization. An interspecific hybrid can form a new species via amphiploidy, by selection of re- combinants among its progeny, or by selection of individuals produced after intro- gression. 2"0 CHAPTER 18 REFERENCES Dobzhansky, I h.. Genetics and the Origin of Species, 3rd Ed.. New York: Columbia University Press, 1951. Dobzhansky, Th., Evolution, Genetics, and Man, New York: John Wiley & Sons, 1955. Dobzhansky, In.. Ehrman, 1... Pavlovsky, O., and Spassky, B., "The Supcrspecics Dro- sophila paulistorum," Proc. Nat. Acad. Sci., U.S., 51:3-9, 1964. Dodson. E. O.. Evolution: Process and Product (Rev. Ed.), New York: Rinehart, 1960. Dunn. L. C. and Dobzhansky, Th., Heredity, Race, and Society, 3rd Ed., New York: New Amer. Libr. of World Lit.. 1957. Ehrlich. P. R.. and Holm. R. W.. The Process of Evolution, New York: McGraw-Hill Book Co., Inc.. 1963. Mayr, E., Animal Species and Evolution, Cambridge: Harvard University Press, 1963. Merrill. D. J.. Evolution and Genetics, New York: Holt, Rinehart and Winston, 1962. Stebbins, G. L., Variation and Evolution in Plants, New York: Columbia University Press, 1950. QUESTIONS FOR DISCUSSION 18.1. Discuss the validity of the concept of a pure race. 18.2. To use the frequencies of ABO blood types in tracing the course of past migra- tion, what assumptions must you make? 18.3. Under what future circumstances would you expect the number of races of human beings to decrease? To increase? 18.4. Can the definition we have used for a species be applied to forms that reproduce only asexually? Why? 18.5. Differentiate between genetic sterility and chromosomal sterility. Invent an example of each type. 18.6. Discuss the hypothesis that a new species can result from the occurrence of a single mutational event. 18.7. Is geographical isolation a prerequisite for the formation of a new species? Explain. 18.8. What is the relative importance of mutation and genetic recombination in species formation? 18.9. Is a species a natural biological entity, or is it — like a race — defined to suit man's convenience? 18.10. Does the statement, "We are all members of the human race," make biological sense? Why? 18.11. Suppose intelligent beings, phenotypically indistinguishable from man, arrived on Earth from another planet. Would intermarriage with Earth people be likely to produce fertile offspring? Why? 18.12. Invent circumstances under which the present single species of man could evolve into two or more species. 18.13. The cells of triploid and tetraploid autopolyploids are usually larger than those of the diploid. What importance has this fact for fruit growers? Races and the Origin of Species 251 18.14. H. Kihara and co-workers have produced triploid (33 chromosomes) water- melons with no seeds, and tetraploid (44 chromosomes) watermelons with seeds but larger than the diploid. How do you suppose this was accomplished? How do you suppose these types are maintained? 18.15. In each of the following cases, an interspecific hybrid can be formed experi- mentally. State whether or not you would expect each of the hybrids produced from the parents described below to become established in nature. (a) In California, the Monterey cypress grows along the coast on the rocks. whereas the allopatric Gowen cypress grows two miles inland in the sand barrens. (b) Two sympatric species of pine occur in California. One of these, the Monterey pine, sheds its pollen before March; the other, the bishop pine, some time later. (c) The hybrid between Crepis neglecta (n = 4) and C. fuliginosa (n = 3) shows unpaired and paired chromosomes as well as multivalents during meiosis. 18.16. Drosophila pseudoobscura crossed with D. persimilis produces sterile males, but partially fertile females. Using marked chromosomes, the interspecific hybrid female can be backcrossed to pseudoobscura, and their progeny can have various combinations of the chromosomes of the two species. When male progeny of the backcross are examined for the length of the testis, it is found that the testis is essentially normal when the X chromosome is from pseudoobscura; when the X is from persimilis, the testis is shorter — increasing in abnormality as the num- ber of autosomes coming from pseudoobscura increases. What can you con- clude about reproductive barriers from these results? 18.17. The cotton species Gossypium hirsutum and G. barbadense are tetraploids (2n = 52) and are phenotypically intermediate between the diploid species G. herbaceum and G. raimondii. (Each is 2n = 26.) If cytological examination is made of meiosis in various hybrids, what would the following results reveal about the origin of these species? (a) barbadense X raimondii shows 13 pairs and 13 singles. (b) barbadense X herbaceum shows 13 pairs and 13 singles. (c) raimondii X herbaceum shows 26 singles. 18.18. What do the following two cases have in common? (a) Through artificial selection the evolution of corn, Zea mays, was aided by genes incorporated from teosinte, Zea mexicana. (b) Commercial wheat contains genes for rust resistance obtained from goat grass chromosomes. Chapter 19 CHEMICAL NATURE OF GENES I n the preceding chapters the pri- mary concern was with the defi- nition of the genetic material on the basis of its capacity to recombine and mutate; our concern here will be with the chemical nature of the genetic material as revealed through chemical analyses. Let us try to determine which of the cell's chemical components are and which are not suitable to serve as genetic material. Since the nu- cleus contains genetic material in its chro- mosomes, any chemical substances located exclusively in the cytoplasm can, of course, be eliminated from consideration as the basis for nuclear genetic material. Because the genetic material seems to possess complex properties, one would expect that its chemical properties were also complex. On this basis, we can eliminate from consideration all in- organic compounds (compounds not con- taining carbon), since no class of inorganic compound enters into a sufficient variety of chemical reactions. One unique feature of protoplasm is the speed and orderliness of its chemical activi- ties. These two characteristics are due to the presence of proteins in the form of en- zymes and cellular structures. Different kinds of proteins contain different numbers of amino acids. Since twenty or so different kinds of amino acids are found in the pro- tein of organisms, the total number of dif- ferent combinations is, for all practical pur- poses, infinite. Protein clearly possesses adequate complexity, so it is not unreason- 252 able to hypothesize that the genetic material is composed of protein. It' the gene were protein in nature, one would expect to find protein in the chromo- somes but not, perhaps, of a type usually found in the cytoplasm. Chemical analyses of nuclei and chromosomes confirm both ex- pectations by revealing the existence of histories — complex proteins which act as bases and are found primarily in chromo- somes. Although the chromosomes of many cells contain histones, they are not found in the chromosomes of all cells. For ex- ample, histones are usually present in the somatic nuclei of fish; however, the sperm of trout, salmon, sturgeon, and herring in- stead contains protamine, a basic protein of simpler composition. The protamine in fish sperm is in turn replaced by histone in the somatic cells produced mitotically after fer- tilization. If genetic material is protein, the genetic specifications or information must be transferred from protamine to histone to protamine. At least in some organisms, then, the same genetic specifications would have to be carried in two chemical forms, protamine and histone. Present knowledge does not prevent us from entertaining the view that alternative chemical compositions are possible for the genetic material, but any alternative must be capable of performing a number of activi- ties in accordance with the principles already established. Nevertheless, the hypothesis that protamine and histone are both genetic material in the same organism is rather com- plicated, at least when one considers that there would be two chemical formulae for a single genotype. For the sake of sim- plicity, it would be more satisfactory if a single nuclear chemical substance were the genetic material. Other proteins are found in chromosomes. Their quantity changes, however, according to the type and rate of metabolic activity Chemical Nature of Genes 253 performed by the cell. There is, therefore, no simple one-to-one relationship between their quantity and gene quantity. Conse- quently, additional hypotheses are required to explain genetic behavior. Despite the initial attractiveness of the hypothesis that the genetic material is proteinaceous, one can conclude that the types and amounts of nuclear protein actually found do not ade- quately support this view. Chromosomes contain another chemical substance which seems to be absent in the cytoplasm (Figure 19-1). This chemical is a type of nucleic acid 1 called deoxyribo- nucleic acid, or DNA, a substance usually found combined with basic proteins like protamine and histone (by means of a chem- ical linkage not completely understood) to form deoxyribonucleoproteins. Before in- vestigating the possibility that chromosomal DNA is genetic material, let us consider first the chemical composition of its molecule. Chemical Composition of DNA Organic bases. Chromosomal DNA con- tains organic ring compounds of which nitro- gen is an integral part. The fundamental N-containing structure is a six-membered ring, as found in benzene, C(;H(i. Figure 19-2a shows the complete structural ar- rangement of benzene; Figure 19-2a' is an abbreviated version with the carbon atoms of the ring omitted; Figure 19-2a" is the same model condensed further by eliminat- ing the hydrogen atoms attached to ring carbon atoms. The basic N-containing ring in DNA is a pyrimidine . This molecule (Figure 19-2b) has N substituted for the CH group at position 1 as well as at position 3 in the benzene ring. Figure 19-2b' and Figure 19-2b" show successive abbrevia- tions of this formula corresponding to those used for benzene. 1 Discovered by F. Miescher (1869). The nitrogen found in DNA is also found in a derivative of the basic pyrimidine ring, called a purine. This molecule consists of a pyrimidine ring — minus the H atoms at positions 4 and 5 — to which an imidazole ring (5-membered) is joined, so that the carbons at these positions are shared by both rings (Figures 19-2c, c', and c"). Hence- forth, the most abbreviated structural repre- sentation will be used for pyrimidines and purines. Since all pyrimidines and purines act chemically as bases, they are termed organic bases. *r ••'• • *«• x»\ 9 IXx "•*y.i u * figure 19-1. Whole mount of a larval salivary "land of Drosophila. DNA stain is restricted to the nuclei. {Courtesy of J. Schultz.) 2.")4 CHAPTER 19 Figure 19—3 shows the structural formulae for various types of pyrimidine. The names underlined are found in DNA. All the dc- rivatives shown o\' the basic pyrimidine ring have an oxygen at position 2 replacing the H which is now at position 3. This oxygen R is shown in the keto form (O = C<^ , with R R representing an atom or group other than H). Two pyrimidines are commonly found in DNA: cytosine and thymine. Cytosine differs from the basic pyrimidine ring by having an amino (NHL.) group attached to the C at position 6 instead of the H. Conse- quently, cytosine can also be called 6-amino- 2-oxypyrimidine. Replacement of the H at position 5 in cytosine by a methyl (CH.{) group produces 5-methyl cytosine; this DNA pyrimidine is found in appreciable amounts H •is H— Ci sC— H I II H— Q , 4C— H H a a' BENZENE H Ni sC— H I II H— Q , 4C— H b' PYRIMIDINE b" H Ni C H— O C H *C— H H PURINE [KiLRi 19-2. Relationship between certain ring compounds. Chemical Nature of Genes 255 Pyrimidine I 5 N NH, (6-amino-2-oxy pyrimidine) NH, NH. CH, CH2OH 5-METHYL CYTOSINE (6-amino-2-oxy-5- methylpyrimidine) 5-HYDROXYMETHYL CYTOSINE (6-amino-2-oxy-5-hydroxy- methyl pyrimidine) 0 O H— Nf^ji H-Nf^jpCH3 1 1 H 1 H URACIL THYMINE (2,6-oxypyrimidine) (2,6- -oxy-5-methyl pyrimidine (5-methyl uracil) figure 19-3. Pyrimi- dines. Names of pyrimi- dines occurring in DNA are underlined. in wheat germ and in trace amounts in mam- mals, fish, and insects. Another pyrimidine, found only in the DNA of certain viruses attacking bacteria, has a hydroxymethyl (CH.OH) group replacing the H at position 5 of cytosine. and is therefore called 5 -hy- droxy methyl cytosine. The other pyrimidine commonly found in DNA is thymine. Thymine is unique in having a keto group replace the H attached to the C at position 6; in addition a methyl group replaces the H at position 5. So thymine can also be called 2,6-oxy-5-methyl- pyrimidine. Note that all the pyrimidines shown differ primarily in the groups present at the 5 and 6 positions in the ring. Figure 19-4 shows the structural formulae for various purines; those found in DNA are underlined. Two purines are commonly found in DNA: adenine and guanine. Ade- nine differs from the basic formula of purine by having an NH. group in place of H at position 6; therefore this compound can also be identified as 6-amino-purine. A deriva- tive of adenine has a CH;! substitute for an H in the NH2 group at position 6 with 6- methylaminopurine resulting; limited amounts of this purine have been found in DNA. The other purine most frequent in DNA is guanine (Figure 19-4). Since guanine has an NH2 group at position 2 and an O in keto form at position 6, it can also be 256 CHAPTER 19 figure 19-4. Purines. Names oj purines occurring in DNA arc un- derlined. Purine NH CH 3 N- H N^^ 0 :> N H 6-METHYLAMINOPURINE NH (6-aminopurine) Cm NT N 2-METHYL ADENINE (2-methyl-6-aminopurine) CH, N— CH3 ^N" "N I H 6-DIMETHYLAMINOPURINE GUANINE [2-amino-6-oxy purine) N I H H 2-METHYLAMINO GUANINE 1 -METHYL GUANINE Chemical Nature of Genes 257 OH OH OH or OH D-RIBOSE or OH 2-DEOXY-D-RIBOSE figure 19-5. Pentoses found in nucleic acids. called 2-amino-6-oxypurine. The purines differ largely in the groups attached at the 2 and 6 positions of the double ring. Pentoses. D-ribose is a sugar (Figure 19-5 a) containing five carbons (being, therefore, a pentose), four of which are joined with an O to form a five-membered ring. Figure 19-5 a' employs the conven- tion, used hereafter, of not showing the car- bons of the ring. The carbons in pentose are given primed numbers to indicate their positions. DNA contains a pentose modified from the D-ribose structure by the absence of an oxygen at position 2', so that this sugar is named 2' -deoxy -D-ribose and is often called 2-deoxyribose or simply, deoxyribose (Figures 19-5b and b')- Deoxyribosides. Each purine or pyrimi- dine base in DNA is joined to a deoxyribose sugar to form the combination called a de- oxyribonucleoside or deoxy riboside. The four main deoxyribosides in DNA are: de- oxy cytidine (for cytosine); (deoxy) thy- midine (for thymine); deoxy adenosine (for adenine); and deoxyguanosine (for gua- nine). The structure for these is shown in Figure 19-6. Note that the deoxyribose always joins to the organic base at its Y position. The linkage involved occurs at position 3 in pyrimidines and at position 9 in purines. Deoxyribotides. In DNA a phosphate (PO^ group is always joined to a deoxy- riboside, forming a deoxyribonucleotide or deoxyribotide. The phosphate is attached either at the 3' or 5' position of the sugar :.-»s CHAPTER 19 I li.IRE 19-6. Common deoxy 'ribosides. OH NH NP 5il .A O N OH CH OH H Deoxycytidine OH H Thymidine PYRIMIDINE DEOXYRIBOSIDES NH, OH H Deoxyadenosine OH CH OH H Deoxyguanosine PURINE DEOXYRIBOSIDES figure 19-7. Deoxyrihotides. OH 5-CH Purine or Pyrimidine Deoxyriboside 3 -monophosphate OH Deoxyriboside 5 -monophosphate Chemical Nature of Genes 259 (a generalized form appears in Figure 19-7 ). This combination is shown specifically for the deoxyribotides containing the pyrimidine cytosine and the purine adenine in Figure 19-8. The deoxyriboside 5'-monophos- phates of cytosine. thymine, adenine, and guanine are called, respectively, deoxycyti- dylic acid, thymidylic acid, deoxyadenylic acid, and deoxyguanylic acid. In summary, then, the basic unit of DNA is the deoxy- ribotide which is composed of a phosphate joined to a deoxyriboside; this, in turn, is composed of a deoxyribose joined to an or- ganic base. These bases are either pyrimi- dines (most commonly cytosine and thy- mine) or purines (most commonly adenine and guanine). Poly deoxyribotides. Chromosomal DNA occurs not as single deoxyribotides but as polydeoxyribonucleotides or polydeoxyribo- NHo Deoxycytidine 3 -monophosphate OH H Deoxycytidine 5'-monophosphate or Deoxycytidylic acid NH. NH Deoxyadenosine 3 -monophosphate figure 19-8. Specific deoxyribotides OH Deoxyadenosine 5 -monophosphate or Deoxyadenylic acid 260 CHAPTER 19 tides. These molecules are actually chains in which the individual deoxyribotides com- prise the links. The way these links are joined can be understood by examining the two separate deoxyriboside 5'-monophos- phates shown at the right of Figure 19-8. I hese two compounds can become linked if the topmost () of the lower compound replaces the OH at position 3' of the sugar * Pyrimidine or purine base of appropriate type (usually cyto- sine. thymine, adenine or gua- nine). figure 19-9. Polydeoxyribotide. in the upper compound. (The same reac- tion occurs when a phosphate is added to position 3' of a deoxyriboside to produce a deoxyriboside 3'-monophosphate as illus- trated in the two molecules shown at the left of Figure 19—8.) Since deoxyriboside ^'-monophosphates are capable of joining together by means of phosphate linkage at 3', polydeoxyribotide chains of great length are produced. Figure 19-9 shows a portion of such a chain. Note that the polydeoxy- ribotide is a linear — that is, unbranched — molecule, whose backbone consists of sugar- phosphate linkages and whose linearity is in- dependent of the particular bases present at any point. This independence means that the structure of the chain is uninfluenced by the sequences of bases which can be in any array. Notice, moreover, that this polymer (a molecule composed of a number of iden- tical units) of deoxyribotides does not read the same in both directions. As indicated by the arrows, the sugar linkages to phos- phate read 3'5', 3'5', and so on; whereas in the opposite direction they read 5'3', 5'3', et cetera. Because of this difference, the polymerized DNA molecule is said to be polarized. Measuring DNA Quantity Two main methods are commonly used in determining the amount of DNA present in the nucleus: the histochemical and the cyto- chemical. The histochemical method em- ploys whole tissues for the chemical extrac- tion and measurement of DNA. Sometimes chemical analysis is made of masses of nuclei, from which most of the adhering cytoplasm has been removed by special treat- ment, to determine the average amount of DNA per nucleus. In the second, cytochem- ical, approach the DNA content of single nuclei, chromosomes, or chromosomal parts is determined. This method is based upon the finding that DNA is the only substance Chemical Nature of Genes 261 in the cell which stains under certain condi- tions. The Feulgen technique stains DNA purple (see p. 8), whereas the methyl green method stains DNA green. When properly applied, not only are these stains specific for DNA, but the amount of stain retained is directly proportional to the amount of DNA present. A given amount of dye bound by DNA will make a quantitative change in the amount of light of differ- ent wavelengths it transmits. These meas- urements can then be used to calculate the amount of DNA present. For example: a stained nucleus is placed under the micro- scope; different appropriate wavelengths in the visible spectrum are sent through the nucleus, and a series of photographs is taken; its DNA content is measured by density changes of the nucleus. From the italicized portions of words comes the name of this procedure, microspectrophotometry . A different application of microspectro- photometry utilizes another property of the purines and pyrimidines in DNA. These bases absorb ultraviolet light of wavelengths near 2600 A {Angstrom units). When other substances absorbing ultraviolet of these wavelengths are removed by enzymatic or other treatments, the quantity of DNA can be measured by its absorbence. As one test of the validity of the absorbency, one can remove the DNA from the chromosome by the use of enzymes — deoxyribonucleases , DNA uses, DNAses, or DNases. These or- ganic catalysts break the long DNA chains into short pieces which then can be washed out of the chromosomes and the nucleus. Such treatment produces the expected loss of absorbency. DNA as Genetic Material Having described the chemical nature and quantitative measurement of chromosomal DNA, we are in a position to consider some results bearing upon the relationship be- tween chromosomal DNA and the genetic material of the nucleus: 1 . The quantity of DNA increases during the metabolic stage until it is exactly double (within the limits of experimental error) the amount present at the begin- ning of this stage. Mitosis apparently distributes equal amounts of DNA to the two telophase nuclei. Therefore, when first formed, all the diploid nuclei of an individual have approximately the same DNA content. 2. The amount of DNA in a haploid gamete is roughly half that found in a newly formed diploid metabolic nucleus of the same individual. Fertilization, which re- stores the diploid chromosome condition, also restores the DNA content character- istic of the diploid cell. 3. Polyploid cells increase proportionally in DNA content. 4. Different cells of a tissue such as the salivary gland of larval Drosophila show different amounts of polynemy in their chromosomes. Since the DNA content of these different nuclei is found to be proportional to their volume, it is as- sumed to be a direct reflection of the degree of polynemy. 5. The capacity of different wavelengths of ultraviolet light to induce mutations in fungi, corn, Drosophila, and other organ- isms is paralleled by the capacity of DNA to absorb these wavelengths. 6. By tagging or labeling atoms (those that are radioactive or have an abnormal weight), it is found that many cellular components are being replaced continu- ously during metabolism. Despite this "atomic turnover," however, the total amount of cellular material does not in- crease. DNA is unusual because it shows little, if any, turnover; in other words, DNA maintains its integrity at the molec- ular level. 262 CHAPTER 19 NUCLEIC COMMON PENTOSE NUCLEOSIDE (MONO-) NUCLEOTIDE ACID PYRIMIDINE (PY) with PO.at 5' 4 or PURINE (PU) BASE 2 -deoxy-D-ribose deoxyriboside deoxyribotide Cytosinc PY Deoxycytidine Deoxycytidylic acid Thymine PY Thymidine Thymidylic acid DNA Adenine PU Deoxyadenosine Deoxyadenylic acid Guanine PU Deoxyguanosine Deoxyguanylic acid D - ribose riboside ribotide Cytosine PY Cytidine 5 Cytidylic acid Uracil PY Uridine 5 Uridylic acid RNA Adenine PU Adenosine 5 Adenylic acid Guanine PU Guanosine 5 Guanylic acid figure 19-10. Terminology for nucleic acids and their components. 7. DNA is a linear, unbranched, polymer — a reasonable finding if one expected DNA to represent a sequence of genes. Just as interstitial genes are bipolar (see p. 1 89 ) , so are interstitial segments of DNA, since each deoxyriboside joins only to two other deoxyribosides via its 3' and 5' sugar linkages to phosphate. In its cellular location and in all of the re- spects mentioned above, the observations are consistent with the view that DNA either is, or is intimately associated with, the ge- netic material. Chemical Composition of RNA In addition to DNA another type of nucleic- acid is found in the chromosome. This is ribonucleic acid or RNA. Normally chro- mosomal RNA is found in combination with protein in the form of rihonucleoprotein. Because the RNA content of chromosomes varies within a cell and among diploid cells of the same organism according to meta- bolic activity, RNA is unlikely to be the chemical basis of genes in typical (DNA- containing) chromosomes. Nevertheless, let us discuss the chemical composition of RNA, noting in particular its resemblance to DNA. Chromosomal RNA, like DNA, is a long, unbranched polymer with the basic unit being a ribonucleotide or ribotide. The ribotide is like the deoxyribotide, for it, too, is a combination of organic base plus pentose plus phosphate; one way in which it differs is that the pentose is D-ribose (Figure 19-5) rather than 2'-deoxy-D-ribose. Another dif- ference is found in RNA's pyrimidines. The two pyrimidines commonly found in RNA are cytosine (also common in DNA) and uracil (2,6-oxypyrimidine — -not found in Chemical Nature of Genes 263 typical DNA). Uracil's structure is shown in Figure 19-3. The two purines commonly found in DNA, adenine and guanine, are also common in ribotides. In RNA the base plus sugar combination is called a ribonu- cleoside or riboside. Ribosides are joined together by phosphates joined both at the 3' and 5' positions of the sugar just as in DNA; consequently Figure 19-9 can repre- sent a polyribotide if an O is added at each 2' position (making each sugar D-ribose), and if among the bases usually present uracil is substituted for thymine. It should be noted that RNA also absorbs ultraviolet light of 2600 A but can be removed from the chromosome by treatment with ribo- nucleases or RNases. In summary, typical chromosomes con- tain two nucleic acids, DNA and RNA. These normally occur in combination with protein to form nucleoproteins (deoxyribo- nucleoprotein and ribonucleoprotein) in which DNA and RNA occur as polynucleo- tides (polydeoxyribotides and polyribotides). Each polynucleotide is built of (mono-) nucleotides (deoxy- and ribotides, respec- tively), which in turn are composed of phos- phates joined at 5' of nucleosides (deoxy- ribo- and ribosides ) . These nucleosides are made up of a pentose (2'-deoxy-D-ribose and D-ribose) joined to a pyrimidine (usu- ally cytosine or thymine and cytosine or uracil) or to a purine (usually adenine or guanine). A portion of this terminology is summarized in Figure 19-10. Although the RNA in chromosomes pos- sesses neither the proper quantitative varia- tion nor the constancy expected of ordinary chromosomal genes, it does have the same linear organization as DNA. Moreover, some viruses composed primarily of ribo- nucleoprotein (influenza, poliomyelitis, and other encephalitic viruses; plant-attacking viruses such as the tobacco mosaic virus; and certain bacteria-attacking viruses) possess genetic properties but do not contain DNA. Since DNA rather than protein is favored as being the genetic chemical under typical chromosomal conditions, it is reasonable to consider RNA rather than the protein to be the chemical basis of genetic specification in these particular viruses. SUMMARY AND CONCLUSIONS This chapter is an initial attempt to throw some light on the chemical nature of the genetic material. The search for chemical substances with properties of the genetic material has led to a consideration of the protein found in chromosomes, but the avail- able evidence does not actively support such a primary role for protein. It is hypothesized that DNA either is or, at least, is intimately associated with the genetic material in chromosomes in view of the following: the localization of DNA; its quantity and distribution in mitosis, meiosis, and fertilization; its quantity in poly- ploid and polynemic situations; the parallel between DNA absorption and the muta- genicity of ultraviolet light; the maintenance of molecular integrity; and its long, linear, unbranched arrangement. It is also hypothesized that RNA may assume the genetic role in certain DNA-free viruses. Some details of the chemical nature of DNA and RNA are presented. Subsequent chapters will aim to further test the hypothesis that DNA (and RNA in special cases) is typically either the genetic material or intimately associated with it. Our ultimate objective is to determine the chemical units of the genetic material — chemical units corresponding to the genetic units of replication, mutation, recombina- tion, and function. 264 CHAPTER 19 REFERENCES Chargaff, E., and Davidson, J. N. (Eds.), The Nucleic Acids, 2 Vols., New York: Academic Press, 1955; Vol. 3, New York: Academic Press, 1960. Davidson. J. N.. and C'ohn, W. E. (Eds.), Progress in Nucleic Acid Research, 2 Vols., New York: Academic Press. 1963. Miescher. F.. "On the Chemical Composition of Pus Cells," translated in Great Experi- ments in Biology, Gabriel, M. L., and S. Fogel (Eds.), Englewood Cliffs, N.J.: Prentice-Hall. 1955. pp. 233-239. Potter. V. R.. Nucleic Acid Outlines, Vol. I. Minneapolis: Burgess Publ. Co., 1960. Steiner, R. F.. and Beers, R. F., Jr., Polynucleotides. Natural and Synthetic Nucleic- Acids, Amsterdam: Elsevier Publ. Co., 1961. QUESTIONS FOR DISCUSSION 19.1. Is it simpler to postulate that DNA rather than protein constitutes the genetic material? Why? 19.2. What is the chemical distinction between: (a) a mononucleotide and a polynucleotide? (b) a nucleotide and a nucleoside? (c) a pyrimidine and a purine? (d) a ribose and a deoxyribose sugar? 19.3. Draw the complete structural formula of a polyribotide having the base sequence adenine, uracil, guanine, cytosine. 19.4. Express thymine as a derivative of uracil. What part of the term deoxythymidine is superfluous? Why? 19.5. What evidence can you provide to support the view that viruses possess genetic properties? 19.6. How would you proceed to measure the absorbency of ultraviolet light by chro- mosomal DNA? Chromosomal RNA? 19.7. Do you believe that the evidence so far presented provides conclusive proof that DNA is genetic material in chromosomes? Why? 19.8. What is your opinion of the hypothesis that DNA is the chemical basis for re- combination, but that protein is the chemical basis for gene function? 19.9. Is DNA complex enough to serve as the chemical basis of gene action? Explain. 19.10. Do you think the term chemon could be defined usefully? Justify your opinion. 19.11. Discuss the similarities and the differences between DNA and RNA. Chapter 20 ORGANIZATION AND REPLICATION OF DNA IN VIVO T: |hat DNA serves as the chem- ical basis of chromosomal ge- netic material was supported by the indirect evidence presented in the last chapter. The primary structure of DNA was described as a single, long, unbranched, polarized chain of nucleotides. If the DNA polymer were genetic material, one would expect it to be linearly differentiated so that successive portions could represent different genes. This differentiation cannot be based upon either the deoxyribose sugar or the phosphate, since one of each is present in every nucleotide. Therefore, all differences in genetic information along the length of the DNA strand would have to be due to the organic bases present. Since species differ genetically, one might expect them to differ in DNA quantity and/or base content. Figure 20-1 gives the per genome DNA content of various types of organisms. It is generally true that the higher an organism is on the evolutionary scale, the larger is its genomic DNA content. Perhaps it would be more meaningful to say that the DNA content per genome increases as the number of functions controlled by genes increases. Histochemical analyses reveal the organic base content in DNA extracted from various species. Considering the total amount of the bases in an extract as 100%, we see in Figure 20-2 the portions found as adenine (A), thymine (T), guanine (G), and cyto- sine (C). There is considerable variation in the relative frequency of bases, ranging 265 from organisms relatively rich in A and T and poor in C and G (sea urchin) to those in which A and T are much less abundant than C and G (tubercle bacillus). The DNA samples taken from radically different species contained relatively different amounts of the four bases. Do these data suggest that a shift in the sequence of bases can produce genetic dif- ferences? The assumption that different orders of the same bases might be involved in specifying different genetic units is con- sistent with the fact that the chicken, salmon, and locust — certainly all very different ge- netically— have very similar base ratios. An alternative explanation would be that these species are molecular polyploids which differ only in the multiples of a basic set of DNA molecules they contain. This possibility can be eliminated from serious consideration in light of our knowledge that chromosomal polyploidy has made a limited contribution to evolution, at least in the animal kingdom (Chapters 11, 18). As long as relatively crude histochemical analyses are made of the total amount of DNA in cells with a large DNA content, we should expect to find roughly the same base ratios among the different members of a single species. This expectation has proved true. Moreover, the same base ratios are found in different normal and neoplastic Man, Mouse, Maize 5-7 x 109 Drosophila 8x 107 Aspergillus 4x 107 Escherichia 1 x 107 Bacteriophage T4 2 x 105 Bacteriophage XI 74 4.5 x 103 (Unpaired) figure 20-1. DNA nucleotide pairs per ge- nome in various organisms. 266 ( II A I'll R 20 ADENINE THYMINE GUANINE CYTOSINE Man (sperm! 1 31.0 31.5 19.1 18.4 Chicken 28.8 29.2 20.5 21.5 Salmon 29.7 29.1 20.8 20.4 Locust 29.3 29.3 20.5 20.7 Sea urchin 32.8 32.1 17.7 17.7 Yeast 31.7 32.6 18.8 17.4 Tuberculosis bacillus 15.1 14.6 34.9 35.4 Escherichia coli 26.1 23.9 24.9 25.1 Vaccinia virus 29.5 29.9 20.6 20.3 E. coli bacteriophage T2 32.6 32.6 18.2 16.6 figure 20-2. Base composition of DNA from various organisms. (*5-/;v- droxymethyl cytosine. ) tissues in the same and among different human beings. Nevertheless, a genome is expected to contain many DNA molecules which differ in base content and sequence. A _l_ t The variation found in — — in differ- G + C ent species — the ratio is about 0.4 for the tubercle bacillus and about 1 .8 for the sea urchin — is consistent with our chemical knowledge, since the DNA strand imposes no limitation upon either the types or the frequencies of the bases present along the length of the fiber. However, the amount of A and the amount of T in the DNA of a given species are remarkably equal as are the amounts of G and C (Figure 20-2). Since in each species A = T and G = C, it is also true that A + G = T + C;in other words, the total number of DNA purines always equals the total number of DNA pyrimidines. Although this regularity is common to all the chromosomal DNA's listed, there is nothing in the primary struc- ture of DNA which helps to explain this significant fact. That the primary structure of DNA is the same in all these organisms suggests, however, that these regularities may be connected with some additional, general characteristic of chromosomal DNA structure. An understanding of the basis for the A = T and G = C relationships may come from studies of an entirely different kind. It has been known for some time that a beam of X rays is bent or refracted when it passes through material. If the material through which the rays pass is completely heterogeneous in structure and orientation. Organization and Replication of DNA in Vivo 267 figure 20-3. X ray diffraction photographs of suitably hydrated fibers of DNA, showing the so-called B configuration. A. Pattern obtained using the sodium salt of DNA. B. Pattern obtained using the lithium salt of DNA. This pattern permits a most thorough analysis of DNA . ( Courtesy of Biophysics Re- search Unit. Medical Research Coun- cil. King's College. London.) ♦ * ♦ < figure 20-4. The Watson-Crick dou- ble-stranded helix configuration of DNA. 2(»S CHAPTER 20 the emergent beam shows qo regularity when it is refracted. But it' the material is com- posed oi macromolecular units and or mo- lecular subunits spatially arranged in a reg- ular manner, then the emergent beam will form an X-ray diffraction pattern. This particular X-ray pattern can be used to identify units and subunits that arc repeated at regular intervals. It is known that each nucleotide in a DNA chain occupies a length of 3.4 A along the chain; this repetition is detectable by the characteristic X-ray dif- fraction pattern it produces — the black spots located symmetrically near the upper and lower edges of both photographs in Figure 20-3. X-ray diffraction patterns have been ob- tained from the DNA of numerous species. In some cases the DNA was not removed from the nucleus; in other cases it was re- moved and also separated from the nucleo- protein protein. The spacings between DNA parts, and hence the X-ray diffraction pat- terns, depend upon the degree to which the DNA is hydrated. In all cases, provided that the DNA is similarly hydrated, essen- tially the same patterns attributable to DNA are found (Figure 20-3). In addition to the 3.4 A repetition, a study of these com- mon patterns shows other repeated units which can be explained only if DNA does not usually occur as a single strand. (On the other hand, X-ray diffraction studies show that chromosomal RNA usually is single-stranded. ) Here then is clear evi- dence for the existence of a secondary structure to DNA normally found in all chromosomes. (We might infer some sort of secondary organization for DNA from the independent observation of the equivalences A = T and G = C.) The simplest explana- tion consistent with the diffraction results of M. H. F. Wilkins and co-workers was proposed by J. D. Watson and F. H. C. Crick (1953a). They hypothesized that DNA is normally two-stranded (Figure 20-4 ) ; each strand being a polynucleotide, and the two strands coiled around each other in such a manner that they cannot be sepa- rated unless the ends arc permitted to re- volve. This kind of coiling is called plecto- nemic (coiled like the strands of a rope) in contrast with paranemic coiling, which per- mits the separation of two coils without revolving their ends (just as two bedsprings pushed together can be separated). The Watson-Crick model for the second- ary organization of DNA macromolccules involves a double helix in which each strand is coiled right-handedly (clockwise). This coil direction is the same as that found in the secondary structure of amino acid chains. polypeptides. The model shows the pentose- phosphate backbone of each strand on the outside of the spiral (comprising the rib- bon), whereas the relatively flat organic bases projecting into the center (as bars) lie perpendicular to the long axis of the fiber (indicated by a vertical interrupted line). The backbone completes a turn each 34 A. Since each nucleotide occupies 3.4 A along the length of a strand, 10 nucleotides occur per complete turn, and each successive nu- cleotide turns 36° in the horizontal plane (so that 10 nucleotides complete the 360° required for a complete turn). The two helices are held together by chemical bonds between bases on different strands. The two strands can form a reg- ular double helix with diameter uniformly about 20 A only if the bases on different strands join in pairs, each pair composed of one pyrimidine and one purine. A pair of pyrimidines (being single rings) would be too short to bridge the gap between back- bones, whereas two purines (being double rings) would take up too much space. Moreover, it is found that the pyrimidine- purine pairing must be either between C and G or between T and A, for only in this way is the maximum number of stabilizing bond- ages between them produced. The type of Organization and Replication of DNA in Vivo 269 stabilizing bond holding the members of a base pair together is called a hydrogen bond or H bond. The base pairs, their H bonds indicated by interrupted lines, are shown in Figure 20-5; the hydrogens that are re- moved when the base pairs join the back- bones are included in the diagrams. The top half of the illustration shows the C-G (and G-C) arrangements. Note in the C-G pair that cytosine has been turned over (from left to right) relative to the way it was shown in Figure 19-3. Three H bonds are formed. Two occur between NFL. and O (the 6— NH, of C with the 6— O of G; the 2— O of C with the 2— NFF of G). One occurs between the 1 — N of C and the 1 — NH of G. The GjC pair is identical to C;G shown except that, in this case, the base turned over is guanine. The bottom half of Figure 20-5 shows the other type of base pair (T:A or A:T, in which T and A have been turned over relative to the way they were shown in Fig- ures 19-3 and 19-4). In this pair only two H bonds are formed, one between the 6 — O of T and the 6— NH,. of A; the other between the 1 — NH of T and the 1 — N of A. Although the hydrogen bond is a weak chemical bond as compared to the C — C bond, there are so many H bonds along a long double helix that the entire structure is fairly rigid and paracrystalline even when moderately hydrated. Note that the region surrounding two base-paired nucleosides can be separated into two portions relative to the pentoses. The smaller portion is called the minor groove (the region sur- rounding the lower parts of the base pairs shown in Figure 20-5), and the larger por- tion, the major groove (the region surround- ing the upper parts). Recall that the double helix configuration of DNA does not dictate the sequence of bases along the length of a chain. But re- member also that the sizes of the pyrimidines and purines and their H bonds do dictate Cytosine Guanine H --H-N / Guanine Cytosine CH- O N-H \ i>— - u // \\ H7 > Thymine Adenine O H Thymine FIGURE 20-5. Base pairs formed between single DNA strands. 270 ( HAP I BR 20 i k,i ki 20—6. ///<• opposite direction of the sugar-phosphate linkages in the two strands of a DNA double helix. 5'CH H2C5' that A in one chain can pair only with T in the other chain — similarly C with G — to form a double helix of constant diameter whose strands are held together by the maxi- mum number of H bonds. Since A and T always go together (as do C and G), the equivalences A = T and C = G, derived from chemical analysis of DNA, become meaningful as the direct consequence of the secondary structure of DNA. In fact, these chemical equivalences provide the first in- dependent test of the Watson-Crick model constructed initially on the basis of the X-ray diffraction diagrams among other considera- tions. To form the maximum number of H bonds between a purine and a pyrimidine, it is necessary to represent one of the two as Organization and Replication of DNA in Vivo 271 being turned over, so that the number 1 atoms of both face each other. This ar- rangement has an important consequence for the orientation of the two chains rela- tive to each other, as represented in the two- dimensional diagram, Figure 20-6. The bases in the chain at the right all face the accustomed way; those in the left chain are all turned over. For each base to join to its sugar in the same three-dimensional way, the sugars must be arranged as shown. No- tice, in proceeding downward from the top of the right chain, the P04~ linkages to sugar read 3'5', 3'5', and so on; reading down in the same way, however, the left chain is 5'3', 5'3', et cetera, so that the member chains in a double helix run in opposite directions, as indicated by the arrows. The X-ray diffraction results, which led to the double helix hypothesis, do not tell us that all DNA in chromosomes is two- stranded, or that a double strand is never single-stranded at certain places or at cer- tain times. Such data prove only that, in the wide variety of organisms studied, a very appreciable part of the chromosomal DNA is not single-stranded. The base content and organization of DNA in viruses attack- ing bacteria have also been studied by chem- ical analysis and by X-ray diffraction. In the varieties T2 and T7, for example, the data are entirely consistent with DNA's being present in the Watson-Crick double- helix configuration. In the mature bacterial virus particles of two other smaller varieties (called 0X174 and 0S13), however, the DNA is definitely single-stranded. This is reflected in the nonequivalence of A and T and C and G and the absence of those pat- terns indicating a secondary structure in the X-ray diffraction photographs. Whenever the DNA is in the double-helix configuration, we can consider one strand is the complement of the other, so that if the sequence of bases in one strand is known, the composition of the other strand can be determined. Thus, if one strand has the base sequence ATTCGAC, the other strand would have to contain TAAGCTG in the corresponding region. If DNA is genetic material, we expect DNA to be replicated just as accurately as genetic material. Since the base sequence in one strand is complementary to the se- quence in the other, we immediately see a simple way in which the double helix might be replicated: ] the two strands separate, and then each strand builds its complement. In this explanation, called the strand separa- tion hypothesis of DNA replication, each strand is visualized as a mold or template. We know that complex surfaces (like statues) can be copied exactly by making a mold which, in turn, can be used to make a second mold which is an exact copy of the original configuration. In the present case, the two complementary strands of DNA can be viewed as molds, or templates, for each other. One strand or both strands act as a mold on which the complementary strand is synthesized. Figure 20-7 shows one pos- sible sequence of events. At the top of this figure, the two strands are coming apart due to rupture of the H bonds. At the center the two single chains exist in the presence of single nucleotides or their pre- cursors. When the complementary free nu- cleotide approaches the single strand, its base is H-bonded. Then, after two or more nu- cleotides have bonded to the single strand, they are linked — perhaps by an enzyme — to start the new complementary strand. The bottom diagrams show sections of the com- plementary strands whose synthesis is al- ready completed. Experiments can be designed - to simul- 1 Based upon the hypothesis of J. D. Watson and F. H. C. Crick (1953b, c). - Based upon those of M. Meselson and F W Stahl. 272 ( II M'TF.R 20 taneously test the hypotheses for both the double-helix structure of DNA and its repli- cation after strand separation. Remember that every pyrimidine or purine base nor- mally found in DNA contains two or four N atoms, respectively. Ordinarily, these are atoms of N-14, or light nitrogen. It should be possible to grow bacteria in a culture medium whose only nitrogen source is in the form o\' a heavier isotope, N-15, or heavy nitrogen. If so, after a number of genera- tions have passed, almost all of the DNA present will have been synthesized using heavy nitrogen. Suppose also that one can synchronize the multiplication of the bacteria containing heavy DNA. What will we ex- figure 20-7. Diagrammatic representation of the hypothesis of DNA replication after strand separation. Organization and Replication of DNA in Vivo 273 pect to happen if these bacteria are quickly washed, placed in a culture medium con- taining only light nitrogen, and permitted to continue their synchronous multiplication? The DNA should replicate each time the bacteria undergo cell division. During the first replication of DNA, the two strands containing heavy nitrogen should separate, and each should synthesize a complementary strand containing only light nitrogen. Thus, after one DNA replication, the density of the DNA molecules should be exactly mid- way between completely light and com- pletely heavy DNA. To test whether or not this expectation is actually observed, DNA is extracted from "all-heavy" bacteria and also from "all- light" bacteria. These extracts, serving as controls, first are ultracentrifuged separately and then together in a fluid medium con- taining cesium chloride. When a solution of cesium chloride is ultracentrifuged for about twenty hours, a gradient of densities is established because the concentration of cesium chloride is greatest at the bottom of the ultracentrifuge tube and least at the top. In the ultracentrifuge tube DNA assumes the position corresponding to its own density. In the density gradient the position of the DNA can be detected by its absorption of ultraviolet light at 2600 A. Two separate bands of DNA are found in the medium, one containing the all-heavy and the other the all-light DNA. When DNA is extracted at various time intervals after the originally all- heavy bacteria have been placed in the all- light nitrogen culture medium, the DNA band in the ultracentrifuge tube is observed to move from the all-heavy DNA position to a position exactly intermediate between the all-heavy and all-light positions (Figure 20-8). This result is exactly what is ex- pected if after one replication the DNA is "hybrid'' in density. What would one expect to find after an additional DNA replication? In this case, the two strands of the hybrid DNA should separate, and light complementary strands should be made by both the light and heavy single strands. So, after a second replica- tion, half of the double-stranded DNA mole- cules should be all-light, and half should be intermediate between all-light and all-heavy (that is, they should be hybrid). In fact, the samples of DNA taken at later intervals show the single band at the intermediate position in the ultracentrifuge tube has be- come two bands, one at the hybrid position and one at the all-light position. It should be noted, moreover, that the time required for the change from all-heavy to all-hybrid molecules, or for the change from all-hybrid to half all-light and half hybrid molecules, is approximately the interval occupied by a bacterial generation. Although these results are consistent with the hypothesis of replication of double- stranded DNA following chain separation, they do not automatically exclude other pos- sible explanations. It might be claimed, for instance, that the double helix grows not by separation of strands followed by the synthesis of complementary ones, but by the addition of new double strand material to the ends of the original double strand. This alternative explanation can be tested in two ways. If the all-heavy molecules present initially grew by adding light material to their ends, they should be composed linearly of double strands that are successively heavy and light. It should then be possible for sonic vibra- tions to fragment the macromolecules into smaller segments, some all-heavy and others all-light. This result should be detectable in the ultracentrifuge tube by some of the sonicated hybrid DNA assuming the all- light and some the all-heavy positions. However, nothing happens; the DNA re- mains in essentially the same hybrid position whether or not it is sonically fragmented. A second test of the view that DNA syn- 274 CHAPTER 20 a b EXP. NO. GENERATIONS 0.3 0.7 1.0 I.I 1.5 1.9 2.5 3.0 figure 20-8. Test of the "replication after chain sep- aration" hypothesis, using the technique of density gradient centrif ligation. DNA was extracted from all-heavy (N-l 5-labeled) bacteria grown for different generation times on all-light (N-14-containing) medium. The extracts were subjected to ultra- centrifugation to position the DNA in the centrifuge tube according to its density. (Density increases to the right of the figure. ) DNA absorption of ultraviolet light is indicated by the bands in different photographs under a and the height of the peaks in the corresponding densi- tometer tracings under h. The rightmost band in the bottom two frames and the bund in the top frame represent all-heavy DNA. The leftmost band, seen clearly in all generation times after 1.5 generations, represents all-light DNA. The only other clear band is between the all- heavy and all-light ones. I his is the only band present after 1.0 generations, and represents DNA which is hybrid in density. Note that at 1.9 generations, half the DNA is all-light and half is hybrid in density (see row showing 0 and 1 .9 mixed). {Courtesy of M. Mesel- son and F. W. Stalil, Troc. Nat. Acad. Sci., U.S., 44:675, 1958.) Oandl.9 mixed Oand 4.1 mixed Organization and Replication of DNA in Vivo 275 thesis is at the ends of double strands in- volves some of the following facts: When any sample of natural or native, double- stranded DNA is heated to an appropriate temperature (near 98° C), the H bonds are broken and the complementary strands sepa- rate. Double-stranded DNA's with high A 4- T — ratios become single-stranded at a G + C B lower temperature than do those with low ratios. This result is expected since high- ratio DNA is richer in A-T than low-ratio DNA, each pair of which has one less H bond than a C-G pair, so that less energy is needed to break the smaller total of H bonds. If the appropriately heated mixture is cooled quickly, the chains remain single, producing denatured DNA . That heat de- naturation followed by quick cooling pro- duces single strands from double helices can be confirmed by the loss of that part of the DNA X-ray diffraction pattern which de- notes polystrandedness. The change to single-strandedness is also accompanied by an increase of as much as 40% in the ab- sorption of ultraviolet light of 2600 A, so that single-stranded DNA is relatively hy- perchromic. It also is slightly denser than double-stranded DNA. If the hot mixture, containing denatured DNA, is cooled slowly, base pairing occurs and renatured DNA is obtained which shows a hypochromic effect and, from X-ray diffraction studies, evidence of double helices. The second test of endwise DNA synthesis involves converting double-stranded, all-light and all-heavy DNA to the single-stranded condition and locating the positions of the two types of single strands in the ultra- centrifuge tube. The "hybrid" double- stranded DNA is then made single-stranded and is ultracentrifuged. This preparation shows only two major components, one lo- cated at the all-light single-stranded position and the other at the all-heavy single-strand position. This result also is inconsistent with the hypothesis being tested. Not only do the two tests eliminate the view that ap- preciable endwise synthesis of DNA occurs in bacterial DNA, but they offer additional support for the hypothesis of replication after strand separation. Similar experiments yielding similar re- sults have been performed using the uni- cellular plant, Chlamydomonas, and higher- organisms, including man. The general agreement in the results of all these experi- ments apparently furnishes conclusive proof of the correctness of the Watson-Crick hy- potheses for the double-helix configuration of chromosomal DNA and for its replication after strand separation. Although the nuclear DNA of most or- ganisms is present as nucleoprotein, being combined with histones or protamines, the DNA in bacteria and in the viruses attack- ing them seems to exist uncombined with basic protein. • In the latter case, the DNA- containing fibers have a diameter of about 25 A. Ordinary chromosomes are probably polynemic with respect to DNA double helices,4 although the exact number of double helices per chromatid is not yet known with certainty. The basic fibril in protamine- containing sperm seems to be about 40 A in diameter, containing one DNA double helix plus basic protein.3 In cells contain- ing histones, two DNA double helices bound side by side, plus the histone. form a fibril which is about 100 A thick.'1 So far, the evidence presents no clue as to how either end of a DNA polymer termi- nates. The possibility exists that DNA is a circular molecule, although this explana- tion still leaves the problem of how chain separation occurs if there are no free ends to revolve. When DNA is extracted from human sperm, about 0.1% of the "purified" material is reported to be composed of amino 3 See H. Ris and B. L. Chandler (1964). 4 See W. J. Peacock (1963). 276 ( II M'TER 20 acids. Digestion of the DNA at those points where it is joined to amino acid pro- duces extensive depolymerization. Such re- sults indicate that the amino acid sequence is short and. therefore, must sometimes be located internally rather than always being at the end of a DNA strand. On the aver- age, there is a sequenee of three amino acids per thousand nucleotides. The amino acid sequences appear bound to the phosphate of the DNA. not as a side ehain. but as an integral part of the molecule. Thus, the backbone of the DNA strand appears to be interrupted by short amino add sequences. About 33 c'( of the amino acids in DNA are the hydroxyl amino acids, serine and threonine, and 10 to 15% are glutamic acid. As expected, available evidence indicates that the amino acid attached to the phos- phate of DNA is often serine. The occur- rence of amino acids in the DNA of human leucocytes, calf thymus, and the vaccinia virus has also been reported. The DNA ■"■See A. Bendich and H. S. Rosencranz (1963). in the larval salivary gland chromosomes of Drosophila also seems to be interrupted." If confirmed, these results are significant since they involve sequences of amino acids which arc so short that, when interstitial, they may be immune to the action of certain digestive enzymes. Such amino acid groups may be involved in: 1. The bending of DNA double-helices (which are rather rigid) especially where DNA is much coiled on itself 2. The mechanism by which strand sep- aration leading to replication occurs 3. The separation of functional DNA units 4. The functioning of individual DNA units 5. The mechanism of crossing over 6. Mutagenesis by agents capable of af- fecting amino acids.7 ,; See reference to D. M. Steffensen (1963) on p. 162. 7 See 1. A. Rapoport and R. G. Kostyanovskii (1959). SUMMARY AND CONCLUSIONS DNA /'// vivo usually exists in the Watson-Crick double helix configuration and usually replicates, after the strands separate, by the formation oi complementary strands. In certain viruses, <2>X174 and <2>S13 for example, the DNA is single-stranded. In bacteria and bacterial viruses, the DNA exists uncombined with protein. In ordinary chromosomes, the following are found: a basic 40 A thick fibril con- taining protamine and one DNA double helix; a basic 100 A thick fibril composed of histone and two DNA double helices. A chromatid probably contains a number of 100 A fibrils. The DNA molecule seems to be interrupted periodically by short amino acid sequences. Organization and Replication of DNA in Vivo 277 REFERENCES Bendich, A., and Rosencranz, H. S., "Some Thoughts on the Double-Stranded Model of Deoxyribonucleic Acid," Progr. Nucleic Acid Res., 1 :219-230, 1963. Crick, F. H. C, "Nucleic Acids," Scient. Amer., 197:188-200. 1957. Jehle, H.n Ingerman, M. L., Shirven, R. M., Parke, W. C, and Salyers, A. A., "Replica- tion of Nucleic Acids," Proc. Nat. Acad. Sci., U.S., 50:738-746, 1963. Luzzati, V., "The Structure of DNA as Determined by X-ray Scattering Techniques," Progr. Nucleic Acid Res., 1:347-368, 1963. Meselson, M., and Stahl, F. W., "The Replication of DNA in Escherichia coli," Proc. Nat. Acad. Sci., U.S., 44:671-682, 1958. Peacock, W. J., "Chromosome Duplication and Structure as Determined by Auto- radiography," Proc. Nat. Acad. Sci., U.S., 49:793-801, 1963. Rapoport, I. A., and Kostyanovskii, R. G., "The Mutagenic Activity of Several Inhibitors of Cholinesterase," Doklady Akad. Nauk SSSR, 131:191-194, 1959 (in Russian). Ris, H., and Chandler, B. L., "The Ultrastructure of Genetic Systems in Prokaryotes and Eukaryotes," Cold Spring Harb. Sympos. Quant. Biol., 28:1-8, 1964. Sueoka, N., "Mitotic Replication of Deoxyribonucleic Acid in Chlamydomonas rein- hardi," Proc. Nat. Acad. Sci., U.S., 46:83-90, 1960. Watson, J. D., and Crick, F. H. C, "Molecular Structure of Nucleic Acids. A Structure for Deoxyribose Nucleic Acid," Nature, London, 171:737-738, 1953a. Reprinted in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs, N.J.: Prentice- Hall, 1959, pp. 241-243. Watson, J. D., and Crick, F. H. C, "Genetical Implications of the Structure of Deoxy- ribonucleic Acid," Nature, London, 171:964-969, 1953b. Reprinted in Papers on Bacterial Genetics, Adelberg, E. A. (Ed.), Boston: Little, Brown, 1960, pp. 125- 130. Watson, J. D., and Crick, F. H. C, "The Structure of DNA," Cold Spring Harb. Sympos. Quant. Biol., 18:123-131, 1953c. Reprinted in Papers on Bacterial Viruses, Stent, G. S. (Ed.), Boston: Little, Brown, 1960, pp. 193-208. See bibliography and all but last portion of Supplement IV. QUESTIONS FOR DISCUSSION 20.1. Can you draw any conclusions from the observation that most of the multi- cellular organisms studied are richer in A + T than C -f- G? 20.2. Among the DNA molecules contained in a genome, why is it expected that many would differ in base sequence and content? 20.3. How many different base pairs normally occur in a double helix of DNA? What are they? 20.4. If a coil is right-handed when looked at from one end, is it also right-handed when seen from the other end? 20.5. What would you have expected to see in the ultracentrifuge tube following sonic treatment of DNA or following the conversion of DNA to its single-stranded condition, if synthesis had occurred at the ends of the double DNA helix? 20.6. What evidence can you give that heating double-helix DNA causes the strands to separate? 278 CHAPTER 20 20.7. When is DNA single-stranded? 2o. s. A double helix oi DNA lias a base sequence \l I AG< A on one strand. Can you complete an inversion after breaking the backbone at two places on this single strand? Explain. Can you complete an inversion if the backbone of the complementary chain is also broken at exactly the same two levels? Explain. 20.9. Given two double helices whose backbones are broken at the places indicated by periods: ATCG.GCAT AT.TAG TAGC.CGTA TA.ATC draw the base sequences which can occur following reciprocal translocation between double helices. 20.10. In what respects is Figure 20-7 incorrect? 20.11. M. Green and M. Piha (1963) report that the G + C ratio of a number of human and animal carcinogenic viruses is similar and lower than that of com- parable nontumorigenic viruses. Discuss the possible implications of this anomaly for the origin and action of tumor-inducing viruses. 20.12. Locate on Figure 20-6 the minor and major grooves of the two pairs of nucleo- tides shown. Discuss the accuracy of this figure. NH, Chapter 21 REPLICATION OF DNA IN VITRO o o o O— P— O— P— o— p— o- o- o- o- E Iarly in this book (p. 10) we assumed self-replication to be a characteristic of the genetic material. In light of the indirect evidence that chromosomal DNA is genetic material, it is of great interest to learn as much as possible about the replication of the DNA double helix. Although the evidence (Chap- ter 20) is fairly conclusive that comple- mentary chains are synthesized after chain separation, no evidence has yet been pre- sented about how this replication is accom- plished. Figure 20-7 and the discussion on page 271 only postulate a mechanism which includes an enzyme that joins the nucleotides forming a new complementary strand. Since the linear combination of nucleo- tides undoubtedly requires energy, consider the possible source of this energy. Consid- erable chemical energy is contained in the ribotide, adenosine triphosphate (ATP), a riboside 5 '-triphosphate (Figure 21-1). The energy hitherto needed to bond two phosphates to adenosine monophosphate be- comes available when ATP reacts with other nucleotides or acids and loses its two ter- minal phosphates as inorganic pyrophos- phate. Because ATP is known to supply the energy for many chemical reactions in the cell, it is reasonable to suppose that it may also supply the energy needed to join individual deoxyribotides to a DNA strand during replication. Since DNA removed from the nucleus and separated from protein still retains what ap- 279 OH OH FIGURE 21-1. Adenosine 5' -triphosphate (ATP) (APPP). pear to be its main characteristics in situ (in the living cell), we may well be able to study DNA synthesis under nonliving condi- tions. What should we extract from cells in order to study DNA synthesis in vitro? Basically, we ought to use all the apparatus the cell normally utilizes for this function. From the strand separation viewpoint, DNA is needed to serve as a template for DNA synthesis, so the extract should contain the cell's DNA. ATP is added to the extract as the source of energy required for the synthesis. MgCl. can also be added; since the magnesium ion, Mg++, is known to activate many enzymes, perhaps it will also act on the one required for DNA strand formation. How can we tell whether DNA is syn- thesized in the extract? Any crude cellular extract would be expected to contain DNases. These enzymes might depolymerize or other- wise degrade DNA as fast as — or faster than — any process synthesizing DNA. The problem of identifying DNA synthesis in the absence of a net increase in DNA quantity can be solved by preparing the deoxyribo- side thymidine with radioactive Cu incor- porated in its pyrimidine and adding this •'hof chemical to the extract. If any radio- actively-labeled thymidine is incorporated into DNA, it would happen as part of the synthetic reaction, since incorporation into DNA only occurs during synthesis. 280 CHAPTER 21 Finally, we ought to obtain the extract From cells that are growing and dividing rapidly, for these cells are likely to contain the greatest amount of functional apparatus for DNA synthesis. In line with this rea- soning, an experiment is performed with an extract of the bacterium Escherichia coli? ATP, Mg " ions, and radioactive thymidine are added and the pH is adjusted to suit experimental conditions. After an incuba- tion interval (about 30 minutes), the pH is made suitably acidic for precipitating a DNA polymer; single deoxyribosides — that is, monomers — remain soluble. The acid precipitate is washed many times until it is certain that the DNA precipitate is not contaminated by adsorbed deoxyribosides. When the DNA is examined, it is found to be only slightly radioactive (50 counts per unit time as compared with 5 million counts in the thymidine substrate added). In fact, so little thymidine is incorporated in the DNA that it is 10,000 times too small to be detected by ordinary chemical analysis. Nevertheless, the radioactivity is without doubt due to thymidine incorporated into DNA and can be released from the precipi- tated DNA by treatment with DNase. Although this result is not quantitatively impressive, the process furnishes C,4-thymi- dine-labeled, acid-precipitable, DNase-sensi- tive DNA as the end product. The amount of this labeled material formed can be used to determine the effect of changes in the ex- perimental procedure. This fact has already led to a change in the procedure and to a better understanding of the nature of the reaction. Reactions that produce derivatives of adenosine commonly start with ATP as one of the reactants. Similarly, derivatives of uridine, cytidine, and guanosine involve their respective triphosphates and the liberation 1 The preceding and following account is based primarily upon work by A. Kornberg and his associates. oi inorganic pyrophosphate. Such facts lead to the conclusion that the fundamental unit in the formation of diribotides or polyribo- tides is the riboside 5'-phosphate, activated in the form of riboside 5 '-triphosphate. It is reasonable, therefore, to assume that the active building block of polydeoxyribotides is the deoxyriboside 5'-triphosphate. If this molecule is the building block, the ATP added in the in vitro experiments may be converting various deoxyribosides — al- ready present or added to the extract — to the 5 '-triphosphate condition (making, for instance, C1 '-thymidine 5'-triphosphate). This view is supported since DNA synthesis occurs in vitro when labeled thymidine 5'- triphosphate (T*PPP) is used instead of labeled thymidine (T*)+ATP (APPP). To learn more about the ingredients es- sential to DNA synthesis, the initial extract, obtained from the sonic treatment of bac- teria, is fractionated and its protein, concen- trated. This procedure results in a nearly 4,000-fold increase in synthetic activity. From this and other evidence, it becomes clear that the presence of a protein catalyst — the enzyme E. coli DNA polymerase (or DNA duplicate) — is essential for the syn- thetic reaction to take place. Once E. coli DNA polymerase is concen- trated, it is possible to obtain a large net increase in DNA (final amount minus initial amount). Such a net increase, however, is obtained only if the 5'-triphosphates of all four deoxyribosides commonly found in DNA are added to the incubation mixture. Deoxyriboside 5'<7/phosphates are not active, nor are riboside 5'-triphosphates. The other requirements for net increase in DNA amount are: 1. The presence of already-formed DNA molecules of high molecular weight 2. Mg+ f ions 3. DNA polymerase. The already-formed, high molecular weight Replication of DNA in Vitro 281 P-P--P * >DNA >DNA figure 21-2. Growth of a primer DNA strand at its nucleoside end (left) and nucleo- tide end (right). Arrows show 5' positions of subsequent degradation by micrococcal DNase plus splenic phosphodiesterase. DNA may come from a plant, animal, bac- terium, or virus. Similar DNA-synthesizing extracts can be prepared from other bacteria, calf thymus, and various animal tissues. Extended and Limited Syntheses Using E. coli preparations, we can obtain an extended synthesis of DNA which pro- duces twenty or more times as much DNA as was initially present. In this case, there- fore, 95% or more of the DNA present at the end must have been synthesized from the triphosphates added as substrate, the ex- tended synthetic reaction proceeding until the supply of one of the four triphosphates is exhausted. One inorganic pyrophosphate is released for each deoxyribotide incorpo- rated into DNA. Although extensive synthesis of DNA does not occur if only one of the deoxyriboside 5'-triphosphates is added as substrate, some incorporation of this nucleotide into the DNA strand occurs in what is called a lim- ited reaction. By what mechanism does the nucleotide add on to the pre-existing DNA strand? In this case, the already-present DNA must provide a point of linear attach- ment for newly-forming DNA, thereby func- tioning as a primer. Suppose that the only triphosphate added to the substrate is de- oxycytidine 5'-triphosphate whose innermost phosphate carries radioactive PA- (dCP*PP). The two possible ways in which the DNA strand might lengthen are shown at the left and right of Figure 21-2; the DNA strand, present as primer, is shown enclosed in brackets. The primer strand can be con- sidered to have a nucleotide end (top of the figure) — to which pyrophosphate, P-P, is added in the diagram to the right — and a free 3'-OH nucleoside end (bottom of illus- tration). (The removal of a sugar and base by a single break at the 5' position involves the removal of a nucleoside at the nucleoside 282 CHAPTF.R 21 end and the removal oi a oucleotide at the nucleotide end. i The diagram at the left of the figure shows the X174, is excellent for this purpose, and heat-treated DNA is better than unheated DNA. Moreover, the prep- arations containing the most active DNA polymerase do not work well with double- stranded DNA unless it is first heated or treated with DNase. The DNA produced in extended syntheses behaves as though it is primarily two-stranded, but differs from native double-stranded DNA by appearing to be markedly branched in electron micro- graphs and by being readily renatured after heat or alkaline denaturation. These results suggest that strand separation is usually in- complete in the in vitro system. In view of these results, we can conclude that the physical characteristics of the DNA synthesized in vitro and in vivo are extremely similar, though not identical. Synthesis clearly involves single strands which pro- duce double strands probably held together by H bonds. One can also study the detailed chemical and physico-chemical characteristics of the in vitro synthesis of DNA. If single strands produce double strands by forming comple- mentary structures, then the capacity to form a complementary strand should depend upon the presence in the substrate of purine and pyrimidine bases which can form appropriate H bonds with the bases in the added DNA. In other words, in an extensive synthesis, pre-existing DNA should serve as a template for the synthesis of complementary DNA strands. Figure 21-4 shows some pyrimi- dine and purine bases which do not naturally or frequently occur in DNA as well as the four principal types which do. The un- natural or infrequent bases include: uracil and 5-bromo uracil (both of which are ex- pected to have the same H-bonding capaci- ties as thymine); 5-methyl cytosine and 5- bromo cytosine (both of which are expected to have the same H-bonding capacities as cytosine); and hypoxanthine (which has two of the three H-bonding sites found in gua- nine). If A in the single-strand preformed DNA dictates its complement — by specify- ing that the complementary base is a pvri mi- dine that provides the proper sites for H- bonding A — then one would expect thai uracil or 5-bromo (or 5-fluoro) uracil can substitute for the thymine in thymidine 5'- triphosphate. When the substrate used contains do- 284 CHAPTER 21 CH; Thymine Uracil 5-Bromo uracil NH NH, CH; CT N' H 5-Methyl cytosine NH. ^HN NH^ NT N H Guanine HN Hypoxanthine figure 21-4. Various bases utilized in in vitro synthesis of DNA. Ar- rows point to groups capable of typical H-bonding. oxyuridinc 5'-triphosphate (or 5-bromo or 5-fluoro deoxyuridine 5'-triphosphate), \ which is not sufficient to penetrate. Only high-molecular-weight DNA penetrates. These tacts should be considered with re- gard to the circumstances under which DNA uptake occurs in mammalian tissue culture. In this case, DNA enters by phagocytosis which occurs only when the DNA adheres to a suitably large non-DNA particle. Pino- cytosis, similar to phagocytosis, is another process by which materials can enter ordi- nary cells. Although pure nucleic acids are not pinocytosed, protein is. However, if pure nucleic acid is mixed with protein, pinocytosis is stimulated, and the nucleic acid is carried into the cell with the protein. Perhaps the penetration of high-molecular- weight DNA into bacteria is dependent upon the presence of sufficient contaminating ma- terial capable of stimulating pinocytosis or some other mechanism for DNA penetra- tion. Whatever the precise method by which transforming DNA penetrates, it is found that relatively short sequences of DNA will enter microbial cells if sufficient protein is also present. The bacterial surface contains a finite number of sites which act as receptors for DNA. Since non-transforming DNA (such as DNA from a widely separated genus) can also penetrate readily, receptor sites can be saturated by nontransforming DNA, thereby preventing the penetration of transforming DNA. Synapsis. Alternatives of the same trait — for example, resistance and sensitivity to streptomycin, or auxotrophy and prototrophy for a particular nutrient — can be found in different species of bacteria. Since it is a reasonable assumption that the same type i)| gene (and its alternatives) performs the same or similar functions in different species, interspecific transformations ought to be possible. Although this result has been achieved, in any given case the interspecific transformation is usually less frequent than the intraspecific one. Moreover, the trans- formation frequency is actually lower and not due to a delay in phenotypic expression which occurs in interspecific (but not in intraspecific) transformation. That inter- specific transformation does take place favors the idea that the transformed locus is nor- mally part of the genotype of both species. The relative infrequency of interspecific transformations is, therefore, not due to in- competence of the recipient cell or a failure of the foreign DNA to bind to or penetrate the recipient. The transforming capacity of already- penetrated DNA may depend not only upon the homology of the loci transformed but upon the nature of the genes adjacent to those undergoing transformation. These neigh- boring genes might influence transformation by their effect upon the synapsis of the trans- forming DNA with the corresponding region of the host's genetic material. In intraspe- cific transformation, the loci adjacent to those transformed are very probably homol- ogous in transformer and host, so that syn- apsis between the two segments can occur properly; in interspecific transformation, these loci are likely to be nonhomologous and, therefore, may often fail to synapse or act to prevent synapsis. Integration. Even if the hypothesized synapsis occurs properly between host and transforming DNA, some process has yet to occur by which the host gene — whose transformation is being followed — is lost from the chromosome, and the donor's locus becomes an integral part of it. Some under- standing of the mechanism of this final stage in transformation may be gained from a study of transformation frequency. First of all. different loci transform intraspecifically with different frequencies. Using genes that transform with suitably high frequencies, we are able to study the frequency of double transformations, that is, the frequency with Clones; Transformation; Strand Recombination in Vitro 299 which bacteria are transformed with respect to two markers present in the donor DNA. In several cases (for example, penicillin- and streptomycin-resistance), the frequency of doubly-transformed bacteria is approxi- mately equal to (actually somewhat less than) the product of the frequencies for the single transformations. Such results prob- ably mean that the transforming DNA car- ries the two loci either on separate particles or in widely separated positions on the same particle. On the other hand, the markers for streptomycin-resistance and mannitol- fermentation are transformed together with a frequency (.1%) which is about 17 times that expected from the product of the fre- quencies of the single transformations (.006%). This result implies that these two genetic markers are located on the same transforming particle; that is, they seem to be reasonably close together in the same bacterial chromosome. If two loci are closely linked, how can we explain the occurrence of single and double transformations for them? Because of frag- mentation during extraction, a given pene- trating DNA particle may not always have the same composition relative to the two markers; it may sometimes carry only one and, at other times, may carry both of these markers. The effect of reducing the particle size of penetrating DNA upon the fre- quencies of single and double transforma- tions can be tested. When particle size is reduced by DNase or sonic treatment, one expects — according to the present hypothe- sis— the particles sometimes to be broken between the two markers, reducing the rela- tive frequency of the double transformation and increasing the relative frequencies of the single transformations. When the particle size is reduced, the overall rate of trans- formation is lower, as expected. No change is found, however, in the ratio of double to single transformations, implying that the two markers are so closely linked, they are rarely separated when particles are fragmented. Accordingly, it seems that the penetrating particles must usually carry both markers, or neither, and the failure to obtain 100% double-transformations from the former type must be because only a small portion of a penetrating, synapsing particle is integrated. Integration of a portion of a synapsed particle can occur in two possible ways (Fig- ure 22-5): One involves copy-choice (Fig- ure 22-5B) in which a daughter chromo- some is formed by the alternate use of the host chromosome and the donor DNA as a template. When completed, the daughter chromosome is exactly like the original chro- mosome except for the daughter segment formed with transforming DNA as the tem- plate. One expects the recombinant chro- mosome produced by the copy-choice method to contain all newly-synthesized DNA. The second method involves breakage and exchange of the kind that takes place in chromosomal rearrangement or in crossing over. In this case (Figure 22-5A), "breaks" have to occur on each side of the marker being integrated, so that a "double crossover" (p. 134) is produced. Although double crossovers within a short distance are expected to be extremely rare between two homologous chromosomes of higher organisms, this kind of exchange can occur under special circumstances and may be pos- sible between the chemically less complex chromosome of bacteria and the shorter, synapsed segment of transforming DNA. Linkage of transforming DNA to host mark- ers does not require DNA synthesis in the region involved, ,! although the integrated segment — which must be at least 900 nu- cleotide pairs long — appears to replicate in synchrony with the host DNA. Experiments with labeled DNA show that in transforma- tion single-stranded donor DNA is inserted «See M. S. Fox (1962). 300 Cl b c c d' d e' e P f g h n 11 CHROMOSOME TRANSFORMING SEGMENT CHAPTER 22 a a b b c' c c' c d' d d' d' e' e e' e' P f P f g g h h * INTEGRATION COMPLETED + a a a b b b c c c' c c d' d d' d' d e' e e e' e P !: f P f f ■ g g g ■ ■ h ' h h REPLIC DAUGHTER CI HROM OSOME figure 22—5. Postulated mechanisms for the incorporation of a segment of genetic information into a host chromosome. A. Breakage method; B. Copy-choice method. into the DNA of the host, and that probably either strand of the donor DNA could be used this way.7 Thus, the evidence favors the breakage hypothesis of integration. It is important to note that the portion ■ See O. H. Siddiqi (1963), and M. S. Fox and M. K. Allen (1964). of penetrating donor DNA which is not in- tegrated, obviously, is also not retained or conserved as chromosomal genetic material. If integration occurs by copy-choice, the transforming segment is not conserved either; if integration occurs by "breakage- exchange." the replaced DNA segment is not conserved. Clones; Transformation; Strand Recombination in Vitro 301 Strand Recombination in Vitro Heating chromosomal DNA denatures it by causing strand separation. After quick cool- ing, the resultant single strands of denatured DNA can fold forming a considerable num- ber of complementary base pairs between bases at different levels of the single strand. It should be noted that, under certain con- ditions, homopolymers of RNA containing A, U, C, or inosinic acid are capable of base- pairing after folding, thus forming regular double-helical structures. To understand de novo synthesis and limited reaction in vitro it may be important to learn the ex- tent to which intrastrand base-pairing oc- curs in DNA homopolymers containing C or G. After pneumococcal DNA is heated for ten minutes at 100° C, all strands are essentially single and all H bonds, broken; 8 this denaturation, called melting, occurs sharply at 71° C for dAT and at 83° C for dGdC. When DNA denatured by heat is cooled slowly, only about 70% renaturation occurs for native DNA, although, as ex- pected, dAT and dGdC apparently show 100% renaturation. Renatured and native DNA differ from denatured DNA in several properties: 1. Under the electron microscope, rena- tured DNA looks very much like na- tive DNA, whereas denatured DNA is irregularly coiled with clustered re- gions 2. Renatured and native DNA have sim- ilar and lighter densities than dena- tured DNA 3. Renatured and native DNA have about twice the molecular weight of dena- tured DNA 4. Although all DNA has the same ab- sorption spectrum, renatured and na- 8 The following account is based primarily on work reported by P. Doty, J. Marmur, J. Eigner, and G. Schildkraut (1960), and J. Marmur and D. Lane (1960). tive DNA absorb less ultraviolet than denatured DNA. Several factors affect renaturation: 1 . The concentration of DNA in a slowly cooling mixture. When the concentration of single strands is high, so is the amount of renaturation; when the concentration is low, slow cooling does not produce any substan- tial recombination of strands. 2. Salt concentration. The negatively- charged phosphate groups of single strands tend to prevent union with other strands. This inhibition can be overcome by adding KC1 to the solution to act as a shield against the repulsion between phosphates. Conse- quently, within a certain range, the more KC1 present, the greater the amount of re- naturation obtained by slow cooling of heated DNA. 3. The source of DNA. Assuming the molecular weight of native DNA to be ap- proximately the same in all organisms, a mammalian cell, which has about a thousand times as much nuclear DNA as a bacterial cell, also has about a thousand times as many DNA molecules. Assuming that all the DNA molecules within a genome differ in base sequence, then, in a given concentra- tion of denatured DNA, on the order of one thousand times fewer complementary strands are present in a sample from calf thymus than in one from Pneumococcus. When equal concentrations of denatured DNA are heated to 80° C, double strands are formed by a large fraction of the bacterial DNA, but by no detectable fraction of the calf thymus DNA. The concentration of com- plementary strands, therefore, is important in renaturation. DNA can be denatured in vitro by a large number of organic chemical substances in- cluding urea, aromatic compounds, and a variety of alcohols. This finding, however, does not necessarily mean that such com- pounds have this function in vivo, or that 302 ( II M'lI.K 22 thej reveal what is responsible for holding the strands in a DNA double helix together under in vivo or the usual in vitro conditions. Vnother physical-chemical change can occur when DNA is heated in vitro. As noted, native pneumococcal DNA has a molecular weight of about six million. When certain preparations of this native DNA are heated, the single strands obtained have a molecular weight o\' less than half this value. This reduction in molecular weight can be explained by the presence o\ DNase as a contaminant. Even though single strands in a double helix are enzy- matically severed by DNase, the whole com- plex can still retain the double-helix con- figuration. Once these complementary strands are separated by heat denaturation, however, the fragments of each single strand separate. As already mentioned, DNA from differ- ent sources and DNA particles of different sizes behave differently in various parts of the sequence leading to transformation. When DNA in vitro is exposed to dilute con- centrations of DNase, the results 9 indicate that single strands of the double helix are attacked first, and only later — when both strands have been attacked at reasonably nearby positions — is the molecule severed. This scission produces smaller DNA mole- cules which, it should be recalled, penetrate a host cell poorly. Even if only one strand of the double helix has been attacked, how- ever, transformation capacity declines. This effect is attributed partly to the failure of penetrant molecules to transform because the transforming locus or because a locus neeessary for synapsis or integration has been inactivated. In Pneumococcus, denatured DNA has a small amount of transforming ability. The molecular basis for this is still undetermined. On the other hand, the transforming ability 1 Of L. S. Lerrrum and I.. J. Tolmach. of renaturcd DNA can be as much as 509? o[ that shown by an equivalent con- centration o\' native DNA. An increased concentration of DNA plus a high ionic strength increase both renaturation and transforming ability. Hybrid molecules can be made by rena- turing a mixture of N-14 and N-15 DNA from E. coli. ( Recall that these synthetic molecules can be identified by the inter- mediate position they assume in the ultra- centrifuge tube.) Hybrid molecules can also be formed between single DNA strands from different speeies, but only if the species are closely related genetically (as would be sug- gested if they showed interspecific transfor- mation) and, therefore, have similar base sequences. Molecular hybrids are useful for comparing base sequences in closely- related organisms even when genetic recom- bination between them cannot take place. Several additional observations should be made: 1 . Strand separation is accomplished by heat in a matter of a few minutes or less. One wonders if this kind of extensive strand separation occurs in vivo. It has been sug- gested that chain separation normally is pro- duced enzymatically through the activity of ravelase, or better, unravelase. 2. The now-routine ability to separate and combine single strands should lead to a better understanding of transformation, in particu- lar, the mechanism of integration. 3. The smallest recombinational unit of the genetic material in bacteria can he iden- tified as the smallest unit of DNA capable of being integrated or replaced in a host genotype in a genetic transformation. Although the physical and chemical prop- erties of the DNA product of an extensive synthesis in vitro closely resemble those of the natural DNA used as primer-template, and although the synthesis is considered to be a biological process, it has not been dem- onstrated that the DNA product has biolog- Clones; Transformation; Strand Recombination in Vitro 303 ical properties, that is, functions genetically in vivo. When transforming DNA is used in an in vitro synthesis, the total transform- ing capacity of the incubating mixture de- creases with time as the synthesis continues. It is very likely that the trace amounts of DNases present in the polymerase prepara- tion cause the loss of overall transforming activity by interrupting the continuity of the DNA strands, so that transforming capacity is lost by DNase action faster than it is gained by means of DNA synthesis. The biological activity of newly-synthesized DNA can be detected, however, by differentially labeling the old and the new DNA and sep- arately testing each for transforming capac- ity. Consequently,10 nonradioactive DNA con- taining 5-bromo uracil is used as a primer- template to synthesize radioactive DNA with no 5-bromo uracil. After synthesis, density gradient centrifugation of the DNA provides a double-stranded fraction containing essen- tially all newly-synthesized DNA (which is radioactive and less dense than the DNA with 5-bromo uracil). Since other explana- tions seem to be ruled out, and this new DNA is found capable of transforming a variety of gene loci, it apparently is proved that biologically-active genetic material, that is, functionally-active genes, can be syn- thesized in vitro. 10 See R. M. Litman and W. Szybalski (1963). SUMMARY AND CONCLUSIONS Genetic recombination occurs in cells of bacteria and of other organisms by means of genetic transformation. In bacteria, transformation involves a sequence of events in which competent cells transiently and then permanently bind DNA. Once bound DNA has penetrated, it apparently undergoes a synapsis-like process with a corresponding segment of the bacterial genome. Transformation is completed when a small segment of donor DNA becomes integrated and replaces a similar segment of the host genome. Transformation provides direct and conclusive evidence that chromosomal DNA is genetic material. In bacteria, the smallest recombinational unit of the genetic ma- terial is the smallest unit of DNA integrated or replaced in transformation. Strand separation and recombination /'/; vitro produce denatured and renatured DNA. respectively. Strong evidence has been obtained that functionally-active genes, introduced into bacteria via transformation, can be synthesized in vitro. REFERENCES Akinrimisi, E. O., Sander, C, and Ts'o, P. O. P., "Properties of Helical Polycytidvlic Acid," Biochemistry, 2:340-344, 1963. Avery, O. T., MacLeod, C. M., and McCarty, M.. "Studies on the Chemical Nature of the Substance Inducing Transformation of Pneumococcal Types," J. Exp. Med., 79:137-158, 1944. Reprinted in Papers on Bacterial Genetics, Adelberg. E. A. (Ed.). Boston: Little, Brown, 1960, pp. 147-168, and in Classic Papers in Genetics, Peters. J. A. (Ed.). Englewood Cliffs, N.J.: Prentice-Hall. 1959. pp. 173-192. 304 CHAPTER 22 Doty, P.. Marmur, J.. Eigner, J., and Schildkraut, C. 'Strand Separation and Specific Recombination in Deoxyribonucleic Acids: Physical Chemical Studies." Proc. Nat. Acad. Sci.. U.S.. 46:461 476. 1960. Ephrussi-Taylor. H.. "Genetic Aspects of Transformations of Pneumococci," Cold Spring Harb. Sympos. Quant. Biol.. 16:445-456, 1951. Fox, M. S., "The Fate of Transforming Deoxyribonucleate Following Fixation by Transformable Bacteria. III." Proc. Nat. Acad. Sci., U.S., 48:1043-1048, 1962. Fox, M. S., and Allen, M. K., "On the Mechanism of Deoxyribonucleate Integration in Pneumococcal Transformation," Proc. Nat. Acad. Sci., U.S., 52:412-419, 1964. Gellert, M.. Lipsett. M. N., and Davies, D. R., "Helix Formation by Guanylic Acid," Proc. Nat. Acad. Sci., U.S., 48:2013-2018, 1962. Herriot. R. M., "Formation of Heterozygotes by Annealing a Mixture of Transforming DNAs," Proc. Nat. Acad. Sci., U.S., 47:146-153, 1961. Hotchkiss. R. D.. "Transfer of Penicillin Resistance in Pneumococci by the Desoxyribo- nucleate Derived from Resistant Cultures," Cold Spring Harb. Sympos. Quant. Biol., 16:457-461, 1951. Reprinted in Papers on Bacterial Genetics, Adelberg, E. A. (Ed.), Boston: Little, Brown, 1960, pp. 169-176. Hoyer, B. H., McCarthy, B. J., and Bolton, E. T., "A Molecular Approach in the Systematics of Higher Organisms," Science, 144:959-967, 1964. Lerman, L. S., and Tolmach. L. J., "Genetic Transformation. I. Cellular Incorpora- tion of DNA Accompanying Transformation in Pneumococcus," Biochim. Bio- phys. Acta, 26:68-82, 1957. Reprinted in Papers on Bacterial Genetics, Adelberg, E. A. (Ed.), Boston: Little, Brown, 1960, pp. 177-191. Levine, L., Gordon. J. A., and Jenks, W. P., "The Relationship of Structure to the Effectiveness of Denaturing Agents for Deoxyribonucleic Acid," Biochemistry, 2:168-175, 1963. Litman, R. M., and Szybalski, W., "Enzymatic Synthesis of Transforming DNA," Biochem. Biophys. Res. Commun, 10:473-481, 1963. Marmur, J., Falkow, S., and Mandel, M., "New Approaches to Bacterial Taxonomy," Ann. Rev. Microbiol., 17:329-372, 1963. Marmur, J., and Lane, D., "Strand Separation and Specific Recombination in De- oxyribonucleic Acids: Biological Studies," Proc. Nat. Acad. Sci., U.S., 46:453- 461, 1960. Marmur, J., Rownd, R., and Schildkraut, C. L., "Denaturation and Renaturation of Deoxyribonucleic Acid," Progr. Nucleic Acid Res., 1:231-300, 1963. Ravin, A. W., "Experimental Approaches to the Study of Bacterial Phylogeny," Amer. Nat., 97:307-318, 1963. Roger, M., "Fractionation of Pneumococcal DNA Following Selective Heat Denatura- tion: Enrichment of Transforming Activity for Aminopterium Resistance," Proc. Nat. Acad. Sci., U.S., 51:189-195, 1964. Siddiqi, O. H., "Incorporation of Parental DNA into Genetic Recombinants of E. coli," Proc. Nat. Acad. Sci., U.S., 49:589-592, and 50:581, 1963. Szybalska, E. H.. and Szybalski, W., "Genetics of Human Cell Lines, IV. DNA- Mediated Heritable Transformation of a Biochemical Trait," Proc. Nat. Acad. Sci., U.S., 48:2026-2034, 1962. Clones; Transformation; Strand Recombination in Vitro 305 QUESTIONS FOR DISCUSSION 22.1. Which single characteristic of bacteria provides the greatest advantage for ge- netic studies? Why? 22.2. Assuming that all members of a clone are genetically identical, could sexual processes have influenced the results of any of the experiments described in this chapter? Explain. 22.3. Distinguish between auxotrophic and prototrophic bacteria. 22.4. Design an experiment to test whether the dye, acriflavin, is mutagenic in E. coli. 22.5. Compare the "chromosome" of bacteria with that of man. 22.6. Is the use of the phrase "bacterial chromosomal DNA" justified even though bacteria do not contain typical chromosomes? Explain. 22.7. What is meant by integration in genetics? Without using diagrams, describe the mechanisms by which it may occur. 22.8. List those features of crossing over which are difficult to explain on a copy- choice basis. 22.9. Discuss the genetic control of gene synthesis and gene degradation. 22.10. Criticize the statement (p. 5) that genetic transmission can occur between gen- erations only by means of a cellular bridge. 22.11. On what basis is transformation classified as a type of genetic recombination rather than as a mutation? Do you agree with this interpretation? Why? 22.12. Devise an experiment to detect whether chain separation occurs during exten- sive in vitro synthesis of DNA. 22.13. Do the studies on transformation offer any clues as to the ploidy of Pneumococ- cus? Explain. 22.14. What kinds of problems would you investigate if you had a feasible method of studying the fate of individual cells exposed to transforming DNA? 22.15. What do studies of genetic transformation reveal regarding the genetic nature of conserved and nonconserved chromosomal DNA? 22.16. Redraw Figure 22-5 showing hypothetical base sequences in double-stranded DNA. Has your drawing any bearing on your answer to question 22.8? Explain. 22.17. How can you explain the finding (p. 299) that the frequency of double trans- formations for certain markers is sometimes somewhat less than the product of the frequencies of the single transformations? 22.18. Interspecific transformation in bacteria is rare or absent when the relative G + C contents of host and donor differ. When the G + C contents are the same, donor-host hybrid DNA's can form even when interspecific transformation is rare. Discuss the relative values of G + C content, hybrid DNA formation, and interspecific transformation in taxonomic studies of bacteria. Chapter 23 BACTERIAL MUTATION AND CONJUGATION Mutation Practically all of the mutants produced after treatment with a mutagenic agent have non- adaptive or detrimental phenotypic effects (Chapter 16). The detriment produced by these mutants clearly is not dependent upon the mutagen's continued presence in the en- vironment. Similarly many of the rare mu- tants that increase adaptability continue to be beneficial in the absence of the mutagen which induced them. On rare occasions, however, a mutagen (like X rays) produces a mutant with an adaptive advantage in the presence of the mutagen (for example, re- sistance to the genetic or nongenetic detri- mental effects of X rays). Is such an adap- tive mutant produced by chance, or is it a special genetic response elicited by the mu- tagen? The same question can be raised about adaptive mutants that occur "spon- taneously." Are these mutants produced as an adaptive genetic response to unidentified factors in the environment? This general problem can be illustrated with a particular strain of E. coli which ap- parently has never been exposed to the drug streptomycin. If such a strain is plated onto an agar medium containing this drug, almost all the individuals will not grow and, there lore, will not form colonies. These in- dividuals are streptomyt insensitive. How- ever, about one bacterium in ten million does grow on this medium and forms a colony composed o\' streptomycin-resistant individuals, the basis for this resistance 306 clearly being transmissible. Is the adaptive, resistant mutant produced in response to the streptomycin exposure, with the streptomy- cin acting as a directive mutagenic agent? Or, do streptomycin-resistant mutants occur in the absence of streptomycin, sponta- neously, with the streptomycin acting only as a selective agent to reveal the prior oc- currence (or nonoccurrence) of resistant mutants? Or, are both explanations true? Restating the problem more generally, we ask whether mutants adapted to a treatment are postadapted (having arisen after treat- ment), pre adapted (having already been present before treatment), or of both types. Clearly, an ambiguous decision results so long as it is necessary to treat the individuals scored with what is being tested — strepto- mycin, in this example — for, under these conditions, one cannot decide whether the resistant mutant had a post- or preadaptive origin. This difficulty can be resolved. If streptomycin-resistant mutants are preadap- tive, they should occur in the absence of the drug and give rise to clones all of whose members are resistant. It should be noted again that the mutation to streptomycin- resistance is a very rare event however it originates. Consequently, one must grow about ten million clones on streptomycin- free agar medium and test each clone for streptomycin resistance by placing a sample of each on a streptomycin-containing me- dium. After this transfer, part or all of one sample is expected to be resistant to the drug. If resistance is due to a preadapted mutant, one can return to the appropriate original clone — which has never been ex- posed to streptomycin — and readily obtain other samples which prove to be resistant. If, on the other hand, the mutant is post- adaptive, additional samples of the original clone will have no greater chance o\' furnish- ing resistants than additional samples taken from different clones. Three clone-sampling procedures are Bacterial Mutation and Conjugation 307 available for testing the preadaptive or post- adaptive origin of mutants. The first method starts with growing a single (presumably) streptomycin-sensitive clone in liquid me- dium and then plating it on an agar medium to produce a large number of separate col- onies. A sample of each colony is then streaked across an agar medium perpendicu- lar to a strip in the agar containing strepto- mycin. Where there is no streptomycin, each streak of bacteria will grow on the agar and, if enough clones are tested, at least one streak will also grow in the streptomycin region (Figure 23-1). If streptomycin- resistant mutants are postadaptive in origin, the growth on the streptomycin-containing area will be sharply discontinuous, since the members of the clone streaked across the streptomycin were originally sensitives, and only rarely will more than one of these bacteria respond to streptomycin by post- adaptive mutation. Moreover, other sam- ples from the original clone will succeed in growing on streptomycin only to the same limited degree as did the first sample. If, on the other hand, the mutation is preadap- tive, the growth across the streptomycin will be practically as continuous as after streak- BAND OF STREPTOMYCIN RESISTANT CLONE SENSITIVE CLONE figure 23-1. Streptomycin-sensitive and streptomycin-resistant E. coli as determined by streaking individual clones. ing the drug with a pure clone of resistant bacteria or with a mixture of bacteria rich in resistant individuals. The proof that the parental clone contained a spontaneous, pre- adaptive, streptomycin-resistant mutant will be complete if other samples of that clone also grow readily when streaked across this drug. Considering the rarity of mutation from streptomycin-sensitivity to -resistance in this strain (one per 107 cells), the labor involved in testing the preadaptive or postadaptive nature of streptomycin-resistant mutants by this clone-sampling technique is prohibitive. (Nevertheless, the method has numerous uses in other genetic studies of bacteria.) The second method which can be used to sample clones involves replica plating.1 As before, this procedure starts by spreading the members of a single clone on streptomy- cin-free agar in a petri dish. As many as a thousand separate colonies can form after the plate is incubated and by pressing this master plate on the top of a sheet of velvet, a sample of almost every colony can be ob- tained simultaneously. The velvet — whose fibers pick up 10 to 30% of each colony — is then used to plant a corresponding pattern of growth on a series of additional replica plates (Figure 23-2). Preliminary control tests show that the velvet makes several excellent replicas of the master plate, and that both streptomycin-resistant and -sensi- tive clones can be replicated this way. The first replica is made on drug-free medium, whereas the second and later ones are made with streptomycin-containing plates on which, obviously, only streptomycin-resist- ant bacteria can grow into colonies. If the postadaptive view is correct, the chance that cells from one colony will grow on two replicas in the presence of streptomycin is the same as it is for two colonies to grow on the same replica. In other words, the 1 Based upon work of J. Lederberg and E. M. Lederberg. 30S CHAPTER 23 positions of the resistant colonies on differ- ent replicas arc random. II' the mutants are preadaptive, however, all replicas probably will be resistant in the same position (al- though some exceptions can occur if the velvet tails to place a portion o\' the same colony on every replica plate). Of course, one also readily finds resistant bacteria in the corresponding position on the master plate. Although this clone-sampling tech- nique is advantageous for many other pur- poses, it is still too laborious for testing the preadaptation or postadaptation hypothesis. since replicas of about ten thousand master plates are required to be reasonably sure of rinding one clonal streptomycin-resistant mutant. This difficulty can be avoided by using a third method for clone-sampling that in- volves replica-plating contiguous colonies. A billion or so bacteria (from a streptomy- cin-sensitive clone) plated on drug-free agar will produce small clones so closely spaced that they grow together and form a bacterial lawn (Figure 23-3A). Nevertheless, rep- licas of this growth can be made on strepto- mycin-containing agar and will show growth wherever drug-resistant mutants occur (Fig- ure 23-3B, CD). One can then turn to the corresponding regions on the master plate to obtain samples to be tested for re- sistance to the drug. If such samples are no richer in resistant mutants than samples from randomly-chosen sites on the master plate corresponding to those which are not mutant on any replica, the postadaptive view is proved. When the experiment is actually performed, the master plate is found to be much richer in mutants at replica sites that are mutant than at those that are nonmutant. Moreover, replicas tend to have mutant clones at corresponding positions on all replica plates (Figure 23-3B-D). Accord- ingly, most mutants are clearly preadaptive. Other experiments show conclusively, in the case of streptomycin, that almost all, if not all, mutants resistant to the drug are pre- adaptive— that is, streptomycin does not in- duce a detectable number of resistant mu- tants. Since the same results are obtained with the drug chloramphenicol, one can extrapolate and conclude that, in general, the figure 23-2. Separate colonies replica-plated (right) from a master plate (left). (Courtesy of N. E. Melechen.) Bacterial Mutation and Conjugation 309 resistant mutants on drug plates arise spon- taneously, prior to exposure to the drugs and, there jore, are preadaptive in origin. Large numbers of bacteria can be easily tested for mutations. For example, a billion drug-sensitive individuals can be plated on agar containing the drug, and the number of resistant mutant clones detected by count- ing the colonies formed; or, similarly, the number of mutants to prototrophy can be scored by plating auxotrophs on agar which lacks the nutrient required for their growth and counting the number of colonies formed. To give information in terms of a rate of mutation, however, it is necessary to state the number of mutants occurring per unit event. In multicellular organisms, mutation rate is usually expressed in terms of muta- tions per cell, per individual, or per genera- tion. This definition can be applied to bac- teria also. Thus, the mutation rate from streptomycin-sensitivity to -resistance in one particular strain of E. coli (a different strain from that used previously) is one per billion bacteria — one of the lowest mutation rates so far measured in any organism. It is sometimes desirable to express muta- tion rate in terms of mutations per unit time; for example, in describing the increase in mutations obtained by aging Drosophila spermatids or sperm (Chapter 14). In bac- teria, one can considerably vary the length of time required to complete a generation. For generation times between 37 minutes and two hours, the shorter the generation time, the larger the mutation rate per hour. When generation time is lengthened from two to twelve hours, the rate of mutations per hour is constant — each hour of delay increasing the number of mutants by the same amount. (Thus, in the two- to twelve- hour range, the number of mutations in- creases linearly per generation. ) Even when the generation time is extended from twelve hours to infinity (the nondividing cells kept alive in a medium which provides a source figure 23-3. Replica-plating a bacterial lawn for the detection of mutants to streptomycin- resistance. (After J. Lederberg and E. M. Lederberg. ) of energy), some mutations are found to take place. It becomes apparent, therefore, that mu- tation rate is best defined as the chance of a mutation per cell {or individual) per unit time. When, however, each of the division cycles or generations requires the same length of time (as would be true for bacteria under optimal environmental conditions), mutation rate is usually measured with one generation as the unit of time. Conjugation Genetic recombination in bacteria may occur as a result of genetic transformation. The transformation process has two features hitherto unencountered in discussions of ge- netic recombination in multicellular organ- isms: 1. The donor DNA enters the host bac- terium without intervention of any other or- ganism, as is shown by the infectivity of pure DNA. (Although transformation involves 310 ( HAPTER 23 MASTER PLATE T L B Contains -r -r -r ♦ TLB, REPLICAS ^ o o o - + + T L B, + - + + + - TLB, TLB, B Pa C ^ o o o Medium Contains t _ -r -i | B Pa C B Pa C + - + B Pa C B Pa C figure 23—4. Use of replica-plating {shown diagrammatically) to detect spontaneous mutations in E. coli. Replica I detects one mutant to T . replica 3 detects one mutant to B, . and replica 2' detects one mutant to Pa+. the genetic material of two different cells, it is not a typical sexual process, since trans- formation does not depend upon contact between donor and recipient cells.) 2. The integration process leading to ge- netic recombination requires the presence of only a portion of the entire genome of the donor cell. (Integration results in a small segment of the penetrant donor DNA replac- ing a small homologous segment of the host genome.) At this point it does not seem unreason- able to hypothesize that any homologous DNA penetrating a bacterial cell can inte- grate by the same mechanism involved in transformation. Of course there may be other means of introducing DNA into a cell and, consequently, experiments are now designed to test whether or not DNA passes from one bacterium to another when two are in contact. One such experiment 2 starts with a proto- trophic strain (K12) of E. coli treated with -' Based upon work of J. Lederberg and E. L. latum ( 1946). a mutagen (like X rays or ultraviolet light) to obtain single auxotrophic mutants which require different nutritional supplements in order to grow. The mutagenic treatment is repeated — first on the single and then on the double mutant auxotrophs — eventually obtaining two lines which differ from each other by three nutritional mutants, all six mutants having arisen independently. One triple mutant strain is auxotrophic for threo- nine (T~), leucine (L ). and thiamin ( #, ) ; the other triple mutant is auxotrophic for biotin (B~), phenylalanine (Pa), and cystine (C~). The genotypes of these two lines can be given, respectively, as and T-L-B1B + Pa+C+ T + L + B, + BPaC-. Of course, the given gene sequence may be different in the linkage map. The pure lines are grown separately on complete liquid culture medium, that is, one which contains all nutrients required for growth and reproduction. To form a bac- Bacterial Mutation and Conjugation 311 terial lawn, about 10s bacteria from one line are plated onto agar containing complete medium. Then three replica plates are made for the TLBX- line (Figure 23-4), each plate contains complete medium deficient in a different single nutrient (T, L, and Bt, respectively). Occasionally, a replica shows a clone that is able to grow because a pre- adaptive mutant produces prototrophy for the nutrient missing from the medium. However, such clonal growth is not found in the corresponding position on all three replicas (or even on two) with greater than chance frequency. The same results are ob- tained when an equal number of bacteria of the B~Pa~C~ line are plated on complete medium and tested on appropriate replicas. We may conclude, therefore, that on rela- tively rare occasions mutants to prototrophy for one nutrient do occur singly, but double or triple mutants do not occur with detect- able frequency. In another test, the preceding experiment is repeated exactly, with the exception that the same numbers of the two triply-mutant strains are mixed in the liquid medium before being plated on agar containing complete medium. In this case (Figure 23-5), six replicas are made with medium which is complete except that three lack B, Pa, and C in addition to lacking T, L, or Bx; the other three lack T, L, and B, and also B, Pa, or C. Individuals of the T~L~BX~ strain cannot grow on the first three replicas mentioned because a single required nutrient is missing; they cannot grow in the last three because all three required nutrients are miss- ing. Individuals of the B~Pa~C~ strain cannot grow on the first three replicas be- cause all three required nutrients are miss- ing; they cannot grow on the last three be- cause one of the three is absent. If the master plate contains a mutant preadaptive to nutritional independence for one of the nutritionally dependent loci, in only one of the six replicas will the mutant form a colony. For example, if a T+ mutant occurs among the individuals of the T~L-Bi~ strain on the master plate, a colony will grow only on the replica lacking B, Pa, C, and T. Actually, about a hundred different positions on the master plate show growth on the replicas. This number is very much larger than that found in the two groups of three replicas made after plating the two lines separately. In the present case, some posi- tions show growth only on one of the six +++ ++ + ___ TLBBPaC T L B, B Pa C \ / Mixed MASTER PLATE ++++++ ^ T L B,B Pa C •Medium Contains SIX REPLICAS 2 3 .9 . (phi) is used to denote phage. 330 whereas the part of the phage that remains attached to the bacterial surface is unneces- sary. These observations suggest a new way by which homologous DNA may en- ter a phage-infected cell. The virus might carry a segment of DNA derived from a previous bacterial host. This piece could penetrate the new host at the same time as part of the phage docs, the phage's entry providing the opening for the bacterial DNA. With this possibility in mind, consider a series of experiments 2 involving the mouse typhoid organism. Salmonella typhimurium. This bacterium, like its close relative E. coli, can also be cultured on a simple nutrient medium. A large number of auxotrophic strains of Salmonella are available, includ- ing one that requires methionine (M~T+ ) and another that requires threonine (M + T~). When these two strains are mixed and plated on a culture medium lack- ing both methionine and threonine, proto- trophic colonies appear in such large num- bers that they cannot be explained entirely as the result of mutation. Prototrophs are also obtained when a liquid culture of the M+T~ strain is centrifuged (to remove most of the bacteria), the supernatant liquid heated for 20 to 30 minutes (to kill any remaining bacteria), and this liquid added to the M~T+ strain. This procedure dem- onstrates that living M+T donor cells are not required to furnish the M+ factor needed to establish prototrophy. So here the production of prototrophs clearly docs not result from conjugation. Moreover, the filtrate retains its full M+ capacity after treatment with DNase. Accordingly, this is not a case of genetic recombination via transformation. Since the M+ factor can pass through filters that hold back bacteria but not viruses, the factor is a "filterable agent." The reverse experiment — using fil- '-' The following discussion is based upon the work, of N. D. Zinder and J. Lederberg (1952). Transduction 331 trates of the M~T+ strain on M+T~ cells — does not produce recombinants. The two strains differ, therefore, in donor ca- pacity. The M + T~ donor strain (but not the M~T+ strain) is found to harbor a phage. This virus, P22, is said to be nonvirulent or temperate. Nevertheless, about one in a thousand times this phage replicates and lyses or bursts the host cell, liberating up to several hundred progeny phage. Ac- cordingly, a culture of bacteria harboring temperate phage does not show a conspicu- ous amount of lysis. Because each cell of the M+T~ strain carrying P22 is poten- tially subject to lysis, the strain is said to be lysogenic. (The lysogenic bacterium, or lysogen, is immune to new infection — that is, to superinfection — by identical or ho- mologous phage.) On the other hand, the M~T+ strain normally lacks P22 and is a nonlysogenic or sensitive strain. When a sensitive strain is exposed to temperate phage, a relatively large fraction of the newly-infected cells lyse and liberate phage. But a small fraction is able to survive, be- come lysogenic, and give rise to lysogenic progeny. If lysogens are lysed artificially and tested for phage, none are detected. Apparently, the phage in a lysogen is con- verted to a new form, called prophage, which reproduces at the same rate as the host chromosome. Usually, to lyse a lyso- gen, prophage must first rapidly replicate a number of times to produce the infective phage liberated at the time of lysis. What is the relationship between the fil- terable M+ factor and the phage P22? 1. Both are unaffected by RNase and DNase. 2. Both show the same inactivation pat- tern with temperature changes. 3. Both have the same susceptibility to an antiserum that blocks the attachment of phage to the bacterium. 4. Both become attached to susceptible cells simultaneously. 5. Both have the same size and mass as determined by filtration and sedimentation tests. 6. Both appear in the medium at the same time and in a constant ratio. 7. Both retain this ratio even though var- ious purification and concentration proce- dures are applied. From these results, it is evident that M + is associated with the phage. Since the ge- netic material of Salmonella is known to be composed of DNA, it is likely that the ge- netic factor M+ is also composed of DNA. Moreover, because the M+ genetic factor cannot be located on the outer surface of the phage particle, the M+ gene must be located in the interior of the virus. Genetic transduction is the process of ge- netic recombination made possible by a vi- rus particle introducing homologous DNA into a recipient cell. Are there any restrictions on the genetic material of Salmonella which can be trans- duced by P22? This virus can be grown on sensitive bacteria genetically marked M + T+X+Y~Z~\ the crop of phage pro- duced after this infection can be harvested, and a portion tested on sensitive indicator strains (M~, T~ , X~, Y~, Z~) one at a time. The results of such tests show trans- duction of M + , of T+, and of X+ — but not of Y + or Z + . Another portion of the har- vested phage is grown on another genetically- marked, sensitive strain — M+T~X+Y + Z-, for example. When the new phage crop is harvested and then tested on the indicator strains already mentioned, it is found now that the new crop of phage has lost T+ but has gained Y+ transducing ability. These results demonstrate that a phage filtrate has a range of transduceable markers exactly equal to that of the markers present in the bacteria on which the phage was last grown. 332 ( II \I»TER 25 In other words, the phage is passive with respect to the content of genes it transduces and retains no transducing memoiy oi an\ hosts previous to the last. Since additional tests demonstrate that ever) locus in Sal- monella is transduceable by P22, we can call this a case of unrestricted or general- ized transduction. In generalized transduc- tion one cell is transduced for a given marker for about each 10''' infecting phage particles. Any chromosomal marker is transduce- able by P22. but is it possible to transduce more than one at a time? P22 can be grown on M + T+X + , harvested, and then grown on M~T~X-. The latter bacteria are replica-plated on three different media one selecting only for M+ recombinants (it contains T and X), another only for T+, and the third only for X l . When the M+ clones are further typed, they are still T-X~. Similarly. r+ clones are still M-X~, and X+ clones are still M~T~. These results show that only a single bac- terial marker or a relatively short DNA seg- ment is transduced at one time. In this re- spect, transduction is similar to transforma- tion but different from conjugation, in which — especially in Hfr strains — large sequences of genes can be transmitted and integrated. In Salmonella, however, examples are known of several genetic markers trans- duced together in what is called linked trans- duction or cotransduction. Other work has established that the biological synthesis of the amino acid, tryptophan, is part of a se- quence of genetically-determined reactions that proceed from anthranilic acid through indole to tryptophan. Different genes con- trolling different steps of this biosynthetic sequence are cotransduced ■; this finding suggests such genes are closely linked to each other. The biosynthesis of histidine in Salmonella is known to involve at least eight loci, four of which produce identifiable ef- • As shown by M. Demerec and coworkers. fects on the sequence of chemical reactions involved. Linked transductions have been found between two or more of these loci.4 In fact, using the relative frequencies of dif- ferent cotransductions and other evidence. all eight loci are found to be continuous with each other and to lie arranged linearly (see Figure 25-1). Using cotransduction, one can build up a complete and detailed ge- netic map of Salmonella which proves to be a single circle. Cotransduction of closely- linked markers is also known to occur " in E. coli by phage P 1 . In a generalized transduction experiment, when a prototroph is obtained by transduc- ing an auxotroph, the new prototroph is usually stable and produces clones pheno- typically identical to typical prototrophs. This process is called complete transduc- tion. In this case, the prototrophic gene introduced must have integrated into the Salmonella chromosome in place of the re- cipient's auxotrophic allele. However, in addition to the large prototrophic colonies formed on selective agar (each of these clones represents a complete transduction), on occasion about ten times as many minute colonies are present (see Figure 25-2). These minute colonies do not appear in platings of auxotrophic mutants on minimal medium and, also, are not the result of an interaction between auxotrophs and colonies of normal or transduced prototrophs located elsewhere on the plate. Minute colony for- mation is explained as follows: Through phage infection the cell initiating the minute colony receives the segment of DNA con- taining the gene for prototrophy under test. This gene, however, fails to be integrated and fails to replicate but retains its func- tional ability to produce a phenotypic effect. Consequently, a hybrid merogenote or het- erogenote is produced in which the domi- nant injected gene for prototrophy is func- 4 By M. Demerec. P. E. Hartman. and coworkers. ■ From the work of E. Lennox. 1 c ' 1 1 o CT> E ■J (T o o E o c -C Op O O •— " a: »3 333 ( ii \i> 1 1 u 25 i K.i ri 25-2. Large and minute (period-sized) colonics of Salmonella, representing complete ami abortive transductions, respectively. (Cour- tesy of P. E. Hartman. ) tional. Because the prototrophic gene prod- uct is made, the cell is able to grow and divide. Only one of the first two daughter cells, however, receives the added chromo- somal fragment, the exogenote. The daugh- ter cell without the exogenote is able to grow and divide only until the prototrophic gene product received from the parent becomes too scarce; on the other hand, the hetero- genotic daughter cell can continue to grow and divide, in turn producing only one het- erogenotic daughter cell. In this way a mi- nute colony is produced which contains a single genetically-prototrophic cell. This has been proved in a variety of cases and by various methods.' This consequence of the failure of complete transduction is called abortive transduction. Hypothetically, the exogenote in an abor- tive transduction has two possible fates: the exogenote might eventually be lost; or it might be integrated, resulting in a complete transduction. Regardless of its ultimate fate, the exogenote is considered to be ge- netic in nature, even though it does not self- replicate. Remember, however, that sell- replication is an assumed characteristic of the total genetic material: this capacity was not required when a ^nc was first defined (p. 33). In most transduction studies, an excess of phage is used: that is, each cell is infected with more than one phage particle. In such experiments, it is always found that trans- duced cells become lysogenic simultaneously. Thus, the cell transduced receives not only the exogenote but an apparently-complete genome of a phage as well — the former re- sulting in genetic recombination for one or more host markers; the latter in lysogeny and immunity. The phage particle whose contents make the host cell lysogenic need not be the same particle which introduces the exogenote, since high concentrations or multiplicities of infecting phage are used. and, on the average, each host cell is pene- trated by the contents of two or more phage particles. Consequently, one particle might furnish the exogenote, and another might cause lysogeny and immunity. Using low concentrations of phage to obtain low mul- tiplicities so that almost no bacterium can be infected by more than one phage, it is possible to prove that, at least in some cases, only one phage particle is needed per trans- duction. It is found.7 moreover, that when a single phage attacks a susceptible bac- terium, the virus can usually produce only one of three mutually-exclusive effects on its host — namely, lysis, lysogeny, or trans- duction. E. coli strain K12 is normally lysogenic for the temperate phage lambda (A). An- other strain of /:. coli is nonlysogenic, that ,; By B. A. I). Stocker, J. Lederberg, and N. D. Zinder. and by H. Ozeki (1956). By J. N. Adams and S. E. Luria. Transduction 335 is, sensitive to lambda. This phage can be collected from a culture of lysogenic bac- teria in great quantities a few hours after a brief exposure of these cells to ultraviolet light. Such UV -induction causes the pro- phage to replicate, and the progeny phage lyse the cell. Using A phage collected from lysogenic cultures, one finds that only a very limited number of different bacterial mark- ers can be transduced. They are restricted to a region that controls galactose fermen- tation, the Gal locus, whose markers are known from conjugation studies to be very closely linked. Lambda is therefore capable only of restricted or specialized transduction. As mentioned, lysis of a lysogen and the consequent liberation of infective phage can be induced by ultraviolet light. When lyso- genic Hfr conjugate with sensitive F~, a number of zygotes are induced to lyse and liberate infective phage. Initiated by con- jugation this method of inducing prophage to produce infective phage and lysis, is called zygotic induction. Moreover, zygotic induction by A occurs, with a given Hfr strain, at a specific time after the start of mating. This precise timing suggests that the chromosome has a locus with which lambda prophage is physically associated during lysogeny. Jn a nonlysogenic cell, no prophage is attached to or associated with this site. When the site with the pro- phage enters a sensitive F~ cell, zygotic in- duction occurs. When crosses are made between nonlysogenic Hfr (without pro- phage) and lysogenic F^ (with prophage), however, zygotic induction does not occur, and the nonlysogenic locus is transferred and segregates in the heterogenotes just as any other genetic marker. From these results and others, the locus for lambda prophage maintenance is found to be closely linked to the Gal locus which lambda can transduce (see Figure 24-5, p. 324, in which the at- tachment locus is given as A). The original lambda-containing lysogenic K12 bacterium is stable and haploid with aspect to the Gal locus and produces only about one <7 See W. Arber. G. Kellenberger. and J. Weigle (1957). and A. Campbell (1964). 10 See A. D. Kaiser and D. S. Hogness (1960). pathways of action open to them upon in- fecting a sensitive bacterium: the phage either lyses or lysogenizes its bacterial host. As clearlj shown in the case of lambda, the infecting phage either remains in the cyto- plasm where it replicates taster than the chromosome and eventually lyses and liber- ates progeny phage, or it integrates in the chromosome where it resides as prophage and is replicated as a regular chromosomal marker. Accordingly, lambda and most other temperate phages are episomes. What is the basis for the difference be- tween the temperate phages capable of gen- eralized and those capable of restricted transduction? A restrictive-transducing phage usually has a specific chromosomal locus for attachment to the host chromo- some, a generalized-transducing phage has not. Assuming — correctly — that the phage genome is nucleic acid, it can be suggested that the nucleotide sequence held in com- mon between prophage and chromosome is shorter for the generalized-transducing phage than it is for the specialized trans- ducer. In this connection it is noteworthy that evidence has been obtained 1! that a portion of the lambda genome is homolo- gous to the E. coli chromosome, as revealed by the ability of their denatured DNAs to base pair with each other. Several experi- ments suggest that a prophage makes the host cell immune to further infection by homologous phage, by preventing not the penetration of the DNA but its replication. This action parallels the suppression of free- F replication by integrated F. Transduction by temperate phages has been found to occur also in Pseudomonas. Vibrio. Staphylococcus, and Proteus, and it would not be surprising to find transduc- tion occurring in a wide variety of other types of cells, including human. 11 By D. B. Cowie and B. J. McCarthy, and by \1. H. Green. Transduction 337 SUMMARY AND CONCLUSIONS Genetic recombination of loci of the bacterial chromosome can be mediated by tem- perate bacteriophages in the process of genetic transduction. The transduced segment can be derived from any region of the bacterial chromosome (as in generalized or unrestricted transduction) or from a narrowly limited region (as in specialized or restricted transduction). The DNA segment transduced may, by integration, replace a chromosomal marker of the host (as in complete transduction), or it may produce a merozygote, in which case the exogenote can still function, whether it can replicate (as Gal exogenotes in E. coli) or not (as the exogenote in abortive transduction in Salmonella). A transducing <£ lambda is defective in its own genome. The deficient portion is replaced by a small segment of bacterial DNA acquired at the time the prophage was induced in its last host. Most temperate phages are episomes which, when attached to the chromosome, have some characteristics resembling those of integrated F. Norton D. Zinder. about 1954. REFERENCES Arber, W., Kellenberger, G., and Weigle, J., "The Defectiveness of Lambda-Transduc- ing Phage" (in French), Schweiz. Zeitschr. Allgemeine Path, und Bact., 20:659- 665, 1957. Translated and reprinted in Papers on Bacterial Genetics, Adelberg, E. A. (Ed.), Boston: Little, Brown, 1960, pp. 224-229. Campbell, A., "Transduction," pp. 49-89, in The Bacteria, Vol. 5, Heredity, Gunsalus, I. C., and Stanier, R. Y. (Eds.), New York: Academic Press, 1964. Jacob, F., and Wollman, E. L., "Spontaneous Induction of the Development of Bac- teriophage A in Genetic Recombination of Escherichia Coli K 12" (in French). C. R. Acad. Sci. (Paris), 239:317-319, 1954. Translated and reprinted in Papers on Bacterial Viruses, Stent, G. S. (Ed.), Boston: Little, Brown, 1960, pp. 336-338. Jacob, F., and Wollman, E. L., "Genetic Aspects of Lysogeny," pp. 468-500, in A Symposium on the Chemical Basis of Heredity, McElroy, W. D., and Glass, B. (Eds.), Baltimore: The Johns Hopkins Press, 1957. 338 < ii\ri ii< 25 Kaiser, A. D., and Hogness, I) S., "The rransformation of Escherichia Coli with Deoxyribonucleic \cid Isolated from Bacteriophage \dg," J. Mol. Biol., 2:392- 415. 1960. Morse, M. I . 1 ederberg, I M., and I ederberg, J., "Transduction in Escherichia Coli k-12." Genetics, 41:142-156, 1956. Reprinted in Papers on Bacterial Genetics, Adelberg, E. \. il d.), Huston: Little, Brown. I960, pp. 209-223. Ozeki. H.. "Abortive fransduction in Purine-Requiring Mutants ol Salmonella Ty- phiniurium." Carnegie Inst. Wash. Publ. 612, Genetic Studies with Bacteria, 97- 106, 1956. Reprinted in Papers on Bacterial Genetics, Adelberg. E. A. (Ed.), Boston: Little, Brown, I960, pp. 230-238. Stent. G., Molecular Biology of Bacterial Viruses, San Francisco: Freeman & Co.. 1963. Wollman, F. F.. and Jacob. F., "Lysogeny and Genetic Recombination in Escherichia Coli K 12" (in French). C. R. Acad. Sci. (Paris). 239:455-456, 1954. Translated and reprinted in Papers on Bacterial Viruses, Stent. G. S. (Ed.). Boston: Little. Brown. I960, pp. 334 335. Zinder. N. D.. "'Transduction' in Bacteria." Scient. Amer.. 199:38-43. 1958. Zinder. N. D.. and Lederberg, J., "Genetic Exchange in Salmonella," J. Bact., 64:679- 699. 1952. QUESTIONS FOR DISCUSSION 25.1. How would you deline the term provirus? How do the terms merozygote and heterogenote differ? How would you define a homo°enote? 25.2. What characteristics are conferred upon a host cell infected by a nontransducing temperate phage which becomes a prophage? Does not become a prophage? 25.3. How would \ou prove that only one exogenote exists in a microcolony of Salmonella produced by an abortive transduction? 25.4. Discuss the statement: "Temperate phage has chromosomal memory, and the chromosome has temperate phage memory." 25.5. F particles are known which carry the prophage of A as "memory." How could you prove the existence of such a particle? 25.6. Describe the procedure and genotypes you would use in demonstrating that E. coli can undergo genetic transformation with respect to Gal. 25.7. Is there any reason to believe that the close linkage of genes with related effects might be more advantageous in microorganisms than in higher organisms? Explain. 25.8. List the different ways that the Gal locus in the E. coli chromosome can undergo recombination. 25.9. Are temperate phages good or bad for bacteria? Explain. 25.10. Is a cell which has presumably stopped undergoing mutation, genetic recom- bination, and self-replication of its DNA still considered to contain genetic material? Explain. 25.11. Discuss the origin and relative numbers oi Adg present among the phages in l.FT and HFT lysates. Chapter 26 BACTERIOPHAGE: RECOMBINATION AND GENETIC MAPS Tl I he morphology of the T-even group (T2, T4, T6) of phages that attack E. coli has been studied in some detail.1 Its members are tadpole shaped, 0.1 to 0.2 p. long — roughly a tenth the bacterial diameter (Figure 26- 1 ) . The surface of the head has a hexa- gonal outline and facets like a crystal. The head membrane is composed of numerous subunits each having a molecular weight of about 80,000. The tail is cylindrical and is used by the phage for attachment to the host cell. The outer sheath of the tail, com- posed of about 200 spirally-arranged sub- units each having a molecular weight of approximately 50,000, forms a hollow cylin- der. The sheath can contract, shortening its length while increasing its diameter with- out changing its volume appreciably. Be- neath the sheath is the core, a hollow cyl- inder with a central hole about 25A in diameter. At the distal end of the core is a hexagonal plate to which six tail fibers are attached; each is bent in the middle and seems to contain subunits with molecular weight of not less than 100,000 a piece. The subunits of the head membrane, sheath, and tail fibers are composed of protein. When digested with trypsin, each of these subunits produces a unique set of peptides indicating that each is different. The core is also protein. A serologically distinct pro- tein, 4 to 6% of the total phage protein, is found in the interior of the phage par- 1 See S. Brenner et al. (1959). 339 tide; polyamines, putrescine, spermadine, lysozymc. and a minor polypeptide are also reported in the phage interior. In addition to these components, the T- even phage interior contains DNA whose volume is approximately the same as that of the total protein. This DNA is com- posed of a single double helix about 200,- 000 nucleotides long. Since such a poly- nucleotide would be about 68 p. long, the DNA inside the phage must be highly coiled.- No RNA has been reported in DNA-containing phages. Not all phages contain DNA; several bac- teriophage contain RNA and no DNA. Moreover, the physical and chemical com- plexity of the T-even phages is not typical -See R. Kilkson and M. F. Maestre (1962). 700 A HEAD MAIL figure 26-1. Diagrammatic representation of the structures observed in intact and triggered T-even phages of E. coli. 340 CHAPTER 26 oi all viruses. Certain plant viruses, such is the tobacco mosaic and turnip yellow mosaic, arc relative!) simple helical or spherical structures. Although X174 seems to have a simple spherical structure. 4>R, a closely related single-stranded DNA phage, shows a small knob which may func- tion as a tail. Identification of the genetic material in DNA-containing phages is made somewhat easier because DNA contains no sulphur and T2 phage protein contains no phospho- rous. The DNA in one sample of phage can be labeled by feeding the E. coli host cells radioactive P:-, while the protein in another phage sample is labeled by feeding the host cells radioactive S35. Each sample of radioactive phage is then permitted to infect nonlabeled cells. '■ The following re- sults are obtained: in one sample all of the P32 (hence all of the DNA) enters the bac- terium; in the other all but about 3% of the S35 (hence almost all the protein) re- mains outside and is removed by blendor treatment. As implied earlier (p. 330), when most of the protein of an attached phage is removed from the host cell by blendor treatment the normal outcome of infection remains unaffected. These results are consistent with the view that DNA and not protein is the carrier of phage genetic information. ; This account follows the work of A. D. Hershey and M. Chase ( 1952). Recall (p. 336) that pure DNA does not penetrate normal E. coli unassisted. The cell wall of /.. coli can be removed by suit- able culture conditions, leaving a protoplast that can be penetrated by purified DNA. After phenol, CaCT, or other treatments, the entire protein coat of phage can be re- moved, leaving pure DNA. When proto- plasts of /•-'. coli are mixed with such pure, single-stranded DNA of phage XI 74 (which has only about 4,500 dcoxyribotides per par- ticle), typical <£X174 progeny, including the characteristic protein envelope arc pro- duced.1 Consequently, the only genetic ma- terial in DNA-containing phage is DNA. The course of events leading to lysis of a phage-infected bacterium can be sum- marized as follows (Figure 26-2): The phage becomes attached tail-first to specific receptors on the bacterial surface. All the DNA and a small amount of protein are injected into the host; probably, the injec- tion is assisted by the contraction of the spiral sheath protein. An eclipse period follows (Figure 26-2, B-D), during which no infective phage can be recovered if the host cell is artificially lysed. During this eclipse period, the infected cell is said to carry immature phage, and the phage DNA is replicating to produce a pool of phage DNA units. Starting at the end of the eclipse period (Figure 26-2E), a fraction 1 This has been shown by G. D. Guthrie and R. L. Sinsheimer. FIGURE 26-2 (opposite) . Electron micrographs of growth of T2 virus inside the E. coli host cell. A. Bacillus before infection. B. Four minutes after infection. C. Ten min- utes after infection. The thin section photographed includes the protein coat of T2 which can he seen attached to the bacterial surface. D. Twelve minutes after infection. New virus particles are starting to condense. E. Thirty minutes after infection. More than 50 T2 particles are completely formed and the host is about ready to lyse. (Cour- tesy of E. Kellenberger. Reprinted from the Scientific American, 204:100, 196E) Bacteriophage: Recombination and Genetic Maps 341 :ui CHAPTER 26 of this DNA pool is assembled into mature phage, each genome surrounded by a newly synthesized coat (head and tail) (Figure 26 2D). About 20 to 40 minutes after in- fection, the infected bacteria produce en- zymes called endolysins which rupture the bacterial cell wall and liberate infective phage into the medium. This last step completes the lytic cycle of a bacteriophage — the only one possible for intemperate or virulent phages, such as T. For temperate phages, this is one of the two possible cycles — upon entering a bacterium, the alternative is integrating with the bacterial chromosome as a prophage, thereby making the bacterium lysogenic and immune. Even in this event, remember, prophage occasionally dissociates from the chromosome and replicates to produce in- fective phage liberated by lysis of the host cell. Virulent Phages Methods for assaying the amount of phage present in a solution are based upon the virus's capacity to lyse sensitive bacteria. In one commonly-used method, the surface of an agar-containing plate is heavily seeded with sensitive bacteria which, upon incuba- tion, will grow to form a continuous and somewhat opaque lawn. When a few in- temperate phage particles are mixed with the sensitive bacteria before incubation, each particle enters a different bacterium, grows there, subsequently lyses the host, and releases up to several hundred daugh- ter phage. These particles proceed to at- tack bacteria near the original host, caus- ing them to lyse later. The repetition of this cycle produces a progressively increas- ing zone of lysis that is detected as a clear- ing or plaque in the bacterial lawn. Under these conditions each plaque is derived from one ancestral phage and a count of plaques therefore corresponds to a count of phage in the infecting sample. The detailed appearance of a plaque de- pends upon the medium, host, and phage. When other factors arc controlled, dilTcrent mutants of a phage may produce plaques with characteristically dilTcrent morphology. Plaque differences can involve size, turbid- ity, presence or absence of a halo, nature of the edges, and — when a dye is added to the agar in the plate — color. The investi- gator can. therefore, detect and maintain phage mutants affecting plaque type. Ge- figure 26-3. Plaques produced by parental and recombinant phage types. Progeny phage of a cross between hr+ and h ' r were tested on a mixture of suitable indicator bacteria. The small clear and the large turbid plaques are made by the parental types of phage prog- eny (h r f and h • r, respectively). The large clear and the small turbid plaques are produced by the recombinant types of progeny (h r and h + r+, respectively) . (Courtesy of A. D. Hershey.) Bacteriophage: Recombination and Genetic Maps 343 netically, phages can differ according to the hosts they are able to infect, and mutants occur which change the range of hosts at- tacked. Therefore, phage mutants affect- ing host range can also be detected and maintained. One can obtain a strain of intemperate T phage that is mutant both for host range, h, and plaque type, r. Sensitive bacteria are infected by a single phage particle con- taining both markers to determine the muta- tion frequencies to the wild-type alleles (h + and r+); wild-type phages are used to de- termine the mutation rates to each of the two kinds of mutant alleles. When the sen- sitive bacterial strain is exposed to a highly concentrated mixture of the double mutant (h r) and wild-type (h+r+) phages, so that some of the multiply-infected cells carry both phage types, not only do the parental types (h r and h+ r+ ) occur among the progeny, but the recombinant types {h r and h r+ ) occur in frequency too high to result from mutation (Figure 26-3). Con- figure 26-4. A recombination map of T4D. Filled-in areas represent min- imal lengths for genes. The symbols for phage components represent the typical morphological products present in lysates of mutant-infected E. coli. (Courtesy of R. S. Edgar; see F. W. Stahl, et al.. Genetics, 50:539-552, 1964.) :M4 chapter 26 sequently, such experiments prove thai ge- netic recombination occurs between phage panicles in a multiply-injected cell. From the relative frequencies with which differenl recombinants appear among phage released Erom cells multiply-infected with a series of different mutants (this procedure is known as "crossing" genetically-different phages), the genetic map of a phage can be con- structed. When this is done for T4. the mutant loci are found to be arranged in a single closed linear order, that is. a circle (Figure 26-4). Because the T4 plaque mutant r lyses rapidly, it produces a larger plaque with sharper margins than the phage with the wild-type allele, r Mixed infections with r and r+ phages usually yield progeny that produce plaques of one or the other type. Two per cent of the observed plaques, how- ever, are mottled; that is, they appear partly r and partly r+. When mottled plaques are picked and their phage content tested, they produce progeny that make either r or r" type of plaque. Since both parental types are present, the mottled plaque could not have been initiated from a single phage haploid for the r locus. Mottled plaques are not caused by infection with clumps of phage particles; moreover, these plaques are not initiated by a phage carrying an un- stable r mutant, since unstable r phages pro- duce phenotypically — and genotypically — sectored, not mottled, plaques. From these results and others, it has been proved 6 that the two per cent of T4 phage producing mottled plaques in mixed infections are het- erozygous for a short region of the phage genome that includes the r locus. A single phage particle can be hetero- zygous for several loci, provided they are GENOTYPE rl or rill Mutants rll Mutants -+ PLAQUES FORMED ON HOST STRAIN B K r r r None + + i k.i RE 26-5. Behavior of r mutants of T-even phages in the B and K strains of E. coli. located far enough apart. In fact, it is un- likely that any phage is completely haploid. Regions present in diploid condition are said to be redundant. Redundancy appears to be accomplished in two ways 7: Either both regions are part of one DNA double- helix {terminal redundancy) or the extra re- gion is present as a separate segment of double-helix DNA {internal redundancy) . Genetic Fine Structure of <£T4 B The r mutants occur in three distinct re- gions of the T4 genetic map — rl, rll, and rill. The r mutants in all three regions produce plaques when E. coli strain B is used as host. However, mutants in the rll region are unique in that they cannot form plaques when their host is strain K.12 of E. coli (which happens to be lysogenic for lambda), whereas the rl and /•/// mutants and r+ phages can (Figure 26-5). Thus, among r mutants, only those in region II have this restriction in host range. The host-range restricted rll mutant is useful since it can be employed as a selec- tive marker. The mutation frequency from After the work of M. Delbruck and W. T. Bai- ley, and of A. D. Hershey and R. Rotman. '•See A. D. Hershey and M. Chase (1951); see also A. H. Doermann and L. Boehner (1963). 7 See G. Streisinger, R. S. Edgar, and G. H. Den- hardt (1964). s The following discussion is based mainly upon the work of S. Benzer (1955, 1957). Bacteriophage: Recombination and Genetic Maps 345 /•// to r//+ can be determined readily by plating T4/7/ on strain K12, since only mu- tants to /•+ will form plaques (r+ is "se- lected" on strain K12). A large number of rll mutants that have a low mutation fre- quency (sometimes as low as one per 10s phages) can be obtained. T4/7/ mutants can be divided into two classes, A and B, on the basis of their behavior after mixed infection of strain K12. When K12 is infected with an rll phage from each class, growth of the phage and lysis of the host occurs. This behavior suggests that the rll region is com- posed of two subregions, A and B, and the products of both are required to produce the normal r+ phenotype. Mutants defective only in the A subregion presumably can still make normal B product, and vice versa. In a bacterium multiply-infected with one phage mutant in A and another in B, the B and A products produced by the mutants can co- operate— that is, show complementation — to produce the r+ phenotype (Figure 26-6). If the two different rll mutants in a mul- tiple infection of strain K12 are located in the same subregion — region A, for example — they will be unable to produce the r + phenotype by complementation since neither phage can produce normal A product. In such cases the phage cannot grow and the host will not lyse unless the infecting mu- tants have lesions far enough apart in region A to permit wild-type progeny to result from recombination between them. Two mutants, rl and r2, arising independ- ently in the same subregion. may fail to re- combine with each other; however, rl may recombine with a third independent mutant, r3, even though mutant r2 does not. These results suggest that mutant r2 has a long de- ficiency or deletion that includes all or part of the region defective in rl and r3. Such deletion mutants are never found to revert to r+. Other mutants which do revert and which give no evidence of having long de- ficiencies are considered to be point mutants. Of the more than 1500 spontaneously-occur- ring rll mutants which have been typed, FUNCTIONAL COMPLEMENTATION A B x x y r X r » NO COMPLEMENTATION figure 26—6. The occurrence or nonoccurrence of comple- mentation between different rll mutants. X z r X r 346 CHAPTER 26 i k,i ki 26—7. Genetic map <>/ the ill region o) phage T4. The breaks in the map militate segments as defined by the ends ot deletions. The order oj the segments has been determined as shown. I he order ot mutants within any one segment has not been determined, but all give recom- bination with each other. The hollow circles and other tilled-in symbols represent different types of phenotypic effects. (Courtesy of S. Benzer and S. P. Champe, Proc. Mat. Acad. Sci., U.S.. 47:1030-1031. 1961.) A CISTRON O^H^OO f AP AP 859 79W 130 129 80 vf ~o — o -o- 1008 70e G O -o -oooooo- -oooooooo P 5b9 IV 147 1034 632 33 11 H NB UV NT F En 19H A 91 67fc9 272 392 88 S 12 3099 1176 **. ^ooooo- ooojr hCKX>B- KKX) — o- 8C 380 UV NT BC 117 AP HB 577 113 UV 159 -o- E0 -o- F P 635VJ SN )S G 1310 585 (lb C 70 79 AP 181 lb 135 13b -o-o -o -o- -o- NT 470 -ooooo -ooo -o- -ooo 385 157 1181 583 P 3fa IGbO 566 EM 71 Wl C 311 279 531 13)6 1702 G JO 59613] 1314 H8 37] I55 I82, ISl 11 129 .'85 NT J 518 3JI -oooo- o- -o P NT UV 98 508 231 Bfc 65 B4b2 ^(XKKKKK)^CX>- P I960 NT 979 NB 83 281 68; EDb U/ 50 IH Bt lala Blalb Bacteriophage: Recombination and Genetic Maps 347 -¥o- -CK><>0~CK>0-0<>0<><>^K>0<>CK><><>0 NA N AP SN NT 1011 SI) 1814 H 27 7H 3\% 47 341 23 ,221 EM 100 3 24 11 249 58 55 14 Ml 58 24 -o-o IJO -O -<>0000-O0-0-OOCK> -OO-OO -( 486 228 1430 NT A N F MP & EM H NB 658 N NT NT NT 1518 NT 97 56 84 102 56 164 52 42 4777 14 71 313 108 88 <*P.\ HB309 -ooooo- -o o-ooo -oooooooA- -ooo-oaoo^,, P UV 511 J H 547 50 G 1883 %5 H 1513 425 804 603 1470 UV N F UV EM UV 1130 733 447 ^^ 14 1 255 201 18 178 14 79 II 16 214 20 94 018 lOAPbt 101 -o- -o -oooooooo -oAo -CKXT ^C/u H F 795 960 KB EH U J 1945 1345 573 P 2215 103 1322 UV 1404 2232 AP EM H 865 SD 8 235 31 1112 50 548 148 5 122 76 64 51 61 72 -o-o -oooo 50 C 5D EM N 110 145 36 131 135 21 B CISTRON 1071 qn J44 UV 3b0 010 072 -oo^oAoooooo -o-ooooooo -ooo oo -^^556 117 N 263 NT B 858 uv AP SD P J 17 150 44 63 176 6 71 110 BHbl 1870 P 8 G P 114 P 5N 261 8 188 370 441 4 55 19 87 42 103 BHo.2. BHal B3 84 8 O EM 6 -Boo -o -o OAP AP J0I6 50 EM 35 S3 13 8 7 Blal B() (HAITI K 26 E. coli have been isolated. Some of these arc ultraviolet inducible, others are not. All (seven) of the viruses whose prophages arc inducible occupj different chromosomal loci located in one region of the host chronic some (Figure 26-10). These phages also differ from each other in that a host lysogenic for one can still be infected by any o\' the others Hemophilus influenzal' is transformable and can also be infected by the UV-inducible, temperate phage HP1. The genetic material o\ this phage can be introduced into nonlyso- genic Hemophilus competent for transforma- tion in three ways: by injection from intact phage; by infection with pure phage DNA; and by infection with pure bacterial DNA "' The preceding discussion is based primarily upon work of F. Jacob and E. L. Wollman (1957). of A. D. Kaiser and F. Jacob (1957). and of F. Jacob and A. Campbell. carrying the prophage, isolated from lyso- Lvns. In each case, the host cell either be- comes lysogenic or lyses and liberates mature phage progenj . Although the immunity properties oi' pro- phage are controlled by a small portion of the phage's total DNA content — the c, re- gion— this region alone docs not include all the genetic information or specifications needed for prophage to become mature, in- fective phage. Presumably most, if not all, the other phage loci arc essential for the pro- duction of mature phage, complete with its protein parts. Using unlabeled host cells, crosses can be made between genetically-different strains of lambda, one unlabeled and the other labeled with the isotopes C,:: and N15. Following density-gradient ultracentrifugation, the dis- 11 See W. Harm and C. S. Rupert (1963). d2i do di8 di6 di4 di7 d34 d22 C CO I d32 d30/' / / im FIGURE 26-9. Diagrammatic representation of the linkage group of the tem- perate bacteriophage X. The upper diagram shows the linear arrangement of various markers for host range or plaque size or type. The d symbols refer to specific defective mutants. The e region is marked by a thicker line and is shown enlarged in the loner diagram. It is composed of three sub-regions, (;,, (,, Co. Im refers to the segment controlling immunity. (After F. Jacob and J. Monod. ) Bacteriophage: Recombination and Genetic Maps 351 tribution of labeled parental DNA is de- phages. The simplest explanation for this termined both among the parental and the result is that recombinants are formed after recombinant genotypes. In such experi- breakage and rejoining of the parental DNA ments,1- discrete amounts of the original strands. Such results do not support — but parental DNA are found in the recombinant do not exclude — phage recombinations oc- curring via a copy-choice mechanism accord- ing to which recombinant phage DNA is expected to be (1961). labeled material. 12 See M. Meselson and J. J. Weisle (1961), and G. Kellenberger, M. L. Zichichi. and J. J. Weigle expected to be made of entirely new, un 424 Lac, Galb 82 X 434 361 21 !466 ---I 1 I I I I I H- figure 26-10. Part of the E. coli linkage map showing the location of certain inducible prophages. SUMMARY AND CONCLUSIONS The morphology and lytic cycle of the virulent T-even phages of E. coli are discussed, and their genetic material identified chemically as DNA. After multiple infection with T4 phages carrying different genetic markers, genetic recombinants are found among the progeny. From the results of such phage crosses, a recombination map can be constructed for T4 in which the genes are arranged linearly in a single circle. Recombinant phages are often diploid for a short region between the recombinant markers. The genetic fine structure of the rll region of <£T4 is revealed by studies of mutation, complementation, and genetic recombination. Their data suggest the hypothesis that the smallest recombinational unit in phage is one nucleotide. Genetic recombination also occurs between mutants of a temperate phage. The single linkage group of lambda is not circular. Immunity to superinfection is deter- mined by the cx region of the phage genetic map. Sometimes, if not always, phage recombination involves parental strands which have broken and rejoined. REFERENCES Benzer, S., "Fine Structure of a Genetic Region in Bacteriophage," Proc. Nat. Acad. Sci., U.S., 41:344-354, 1955. Reprinted in Papers on Bacterial Viruses, Stent, G. S. (Ed.), Boston: Little, Brown, 1960, pp. 209-219. Benzer, S., "The Elementary Units of Heredity," pp. 70-93, in A Symposium on the Chemical Basis of Heredity, McElroy, W. D., and Glass, B. (Eds.), Baltimore: Johns Hopkins Press, 1957. Benzer, S., "On the Topography of the Genetic Fine Structure." Proc. Nat. Acad. Sci , U.S., 47:403-415, 1961. Benzer, S., "The Fine Structure of the Gene," Scient. Amer., 206 (No. l):70-84 1962. Brenner, S., Streisinger, G., Home, R. W., Champe, S. P., Barnett, L., Benzer, S., and Rees, M. W., "Structural Components of Bacteriophage," J. Mol. Biol., 1:281- 282, 1959. 352 CHAPTER 26 A. D. Hershey. about I960. Seymour Benzer, in 1961. Davidson, P. F., Freifelder. D.. Hede, R., and Levinthal, C, "The Structural Unity of the DNA of T2 Bacteriophage," Proc. Nat. Acad. Sci., U.S., 47:1123-1129, 1961. Doermann, A. H., and Boehner, L., "An Experimental Analysis of Bacteriophage T4 Heterozygotes, I.," Virology, 21:551-567, 1963. Guthrie, G. D., and Sinsheimer, R. L., "Observations on the Infection of Bacterial Protoplasts with the Deoxyribonucleic Acid of Bacteriophage ^>X174," Biochim. Biophys. Acta, 72:290-297, 1963. Hanafusa, H.. Hanafusa, T.. and Rubin, H., "The Defectiveness of Rous Sarcoma Virus, II. Specification of RSV Antigenicity by Helper Virus," Proc. Nat. Acad. Sci., U.S., 51:41-48, 1964. Harm. W., and Rupert, C. S., "Infection of Transformable Cells of Haemophilus in- fluenzae by Bacteriophage and Bacteriophage DNA," Zeit. f. Vererbungsl.. 94: 336-348, 1963. Hayes, W., The Genetics of Bacteria and their Viruses, New York: J. Wiley & Sons, 1964. Helinski. D. R., and Yanofsky, C, "Correspondence between Genetic Data and the Position of Amino Acid Alteration in a Protein," Proc. Nat. Acad. Sci., U.S., 48:173-183, 1962. Hershey, A. D.. and ( hase, M., "Genetic Recombination and Heterozygosis in Bacterio- phage." (old Spring Harb. Sympos. Quant. Biol., 16:471-479, 1951. Reprinted in Papers on Bacterial Viruses, Stent, G. S. (Ed.), Boston: Little, Brown. 1960, pp. 179-192. Bacteriophage: Recombination and Genetic Maps 353 Hershey, A. D., and Chase, M., "Independent Functions of Viral Protein and Nucleic Acid in Growth of Bacteriophage," J. Gen. Physiol., 36:39-54, 1952. Reprinted in Papers on Bacterial Viruses, Stent, G. S. (Ed.), Boston: Little, Brown. I960, pp. 87-104. Jacob, F., and Wollman, E. L., "Genetic Aspects of Lysogeny," pp. 468-500, in A Symposium on the Chemical Basis of Heredity, McElroy, W. D., and Glass, B. (Eds.), Baltimore: Johns Hopkins Press, 1957. Jacob, F., and Wollman, E. L., "Viruses and Genes," Scient. Amer., 204 (No. 6): 92-107, 1961. Kaiser, A. D., and Jacob, F., "Recombination Between Related Temperate Bacterio- phages and the Genetic Control of Immunity and Prophage Localization," Virology, 4:509-521, 1957. Reprinted in Papers on Bacterial Viruses, Stent, G. S. (Ed.), Boston: Little, Brown, 1960, pp. 353-365. Kellenberger, G., Zichichi, M. L., and Weigle, J. J., "Exchange of DNA in the Re- combination of Bacteriophage A," Proc. Nat. Acad. Sci., U.S., 47:869-878, 1961. Kilkson, R., and Maestre, M. F., "Structure of 72 Bacteriophage," Nature (Lond.), 195:494-495, 1962. Meselson, M., and Weigle, J. J., "Chromosome Breakage Accompanying Genetic Re- combination in Bacteriophage," Proc. Nat. Acad. Sci., U.S., 47:857-868, 1961. Mosig, G., "Genetic Recombination in Bacteriophage T4 during Replication of DNA Fragments." Cold Spring Harb. Sympos. Quant. Biol., 28:35-42, 1964. Rubenstein, I., Thomas, C. A., Jr., and Hershey, A. D., "The Molecular Weights of T2 Bacteriophage DNA and its First and Second Breakage Products," Proc. Nat. Acad. Sci., U.S., 47:1113-1122, 1961. Stahl, F. W., Edgar, R. S., and Steinberg, J., "The Linkage Map of Bacteriophage T4," Genetics, 50:539-552, 1964. Streisinger, G., Edgar, R. S., and Denhardt. G. H., "Chromosome Structure in Phage T4. I. Circularity of the Linkage Map," Proc. Nat. Acad. Sci., U.S.. 51:775-779, 1964. See last portion of Supplement V. QUESTIONS FOR DISCUSSION 26.1. Is the hole in the tail of T-even phages large enough for the passage of one or two double helices of DNA? Explain. In what respect does your answer concern the manner in which phage DNA enters into a bacterial host? 26.2. In studying phages, what are the advantages of using bacteria growing on a solid, rather than in a liquid, culture medium? 26.3. Do you think it would be feasible to study the genetic basis for different morphological or for different protein components of a phage? Explain. 26.4. What is meant by a phage cross? Describe how you would know that you made one. 26.5. Does the finding— that complete progeny phages are liberated after pure DNA from 4>X174 infects a protoplast — mean that all the information for making 6X174 DNA and X\14 protein is contained in the phage's DNA? Explain. 26.6. Are the hypotheses of phage recombination by breakage and by copy-choice mutually exclusive? Explain. 354 CHAPTER 26 26.7. In what respects is d>\ similar to and different from an I particle? 26.8. A temperate phage able to transduce any known chromosomal marker in E. coli is known. Would you be able to locate the chromosomal site for its prophage? I \plain. 26.9. Does the finding — that a single phage particle can transduce a bacterial fragment carrying not onlj a bacterial marker hut two linked prophages — have any bear- ing upon the essentiality of the entire phage genome being present for infection and or the production o\ phage progeny? Explain. 26.10. How can you distinguish a <£T4 mutant in the ill region from one in the rl or the rill region'.' 26.1 I. Describe how the cis-trans test is used to show functional complementation be- tween two mutants in phage. 26.12. What would you expect to be the near-maximum number of nucleotides trans- duceable by a phage still capable of phage activity? On what is your opinion based? 26.13. If the average protein-specifying gene were 2000 nucleotides long, how many different proteins could be specified by ^>T4? By <£X174? 2(^.14. What do you consider to be the most remarkable feature of X174? 26.15. Mutants which show functional complementation in the pan-2 region of Neu- rospora can be arranged in the same linear order by complementation and by genetic recombination. Is it necessarily true that both maps will also be iden- tical for other regions? Explain. 26.16. What is a functional genetic unit, or cistron? How is your answer related to its length in nucleotides? 26.17. What have you learned in this chapter regarding the chemical scope of the genetic unit of recombination? Of function? Chapter 27 BACTERIAL EPISOMES AND GENETIC RECOMBINATION T: (he location and genotype of F determines the kind of male sexuality which occurs in Esch- erichia. In this bacterium, conjugation leads to new combinations of either or both the chromosomal genes and the extrachromoso- mal episomal genes. Let us consider the sequence of events in F's genetic recombination. When an Hfr strain reverts to F + , the F particle integrated into the Hfr chromosome is somehow liber- ated from it, or deintegrated. The particle then enters the cytoplasm, replicates, and thereafter is infectious. In some subsequent generation, the F particle may reintegrate into a chromosome, making it Hfr. Before an F particle can integrate, it apparently must synapse with the chromosome — the attrac- tion between F and the chromosome prob- ably being the same as between a segment of donor chromosome and the host chromosome in transformation, conjugation, or transduc- tion. In transformation, the integrating do- nor loci must be homologous to those re- placed in the recipient cell. (Most likely, this homology is also required for the inte- gration of chromosome fragments introduced by conjugation or transduction.) Since F can integrate at a variety of loci, it appears likely that F has segments of DNA homolo- gous to a variety of chromosomal regions. The homologous segments which F presum- ably contains may have been present 'ini- 355 tially," or they may have been obtained at the time of previous deintegrations. (It is not known whether the first F originated as an offshoot of the bacterial chromosome or entered the bacterium from the outside.) If a free F particle sometimes carries an extra segment of chromosomal DNA somehow obtained at the time of deintegration, one should be able to find a type of free F parti- cle which, when introduced into an F- strain, shows a high affinity for a specific chromo- somal region. Recall that temperate phages capable only of restricted transduction show such a restriction, although the wild-type F particles do not. Consider next the following results from experiments l concerned with the expecta- tions mentioned. An F^ strain (carrying F extrachromosomally) gave rise to an Hfr strain, P4x, whose chromosomal markers were arranged in the following sequence: O (origin or lead point) -Pro-TL-Thi . . -Gal-Lac-SF (sex factor place of attach- ment). Crosses of P4x by F~ produced F~ progeny, except for Lac recombinants (which are usually Hfr males because of the close linkage of F and Lac). From P4x arose a new strain, P4x-1, hav- ing these characteristics: 1. With respect to chromosomal loci, it was identical to P4x in the order of arrange- ment and in the times of entry (determined by interrupted conjugation). For example, both transferred Pro at about six minutes, TL at about twenty minutes, and Lac last. 2. Recipients showed a frequency of re- combination for chromosomal loci lower than the frequency of recombination when P4x was the donor. For example. Pro re- combinants were 0.3 to 0.5% with P4x-1 as donor and 4.8% with P4x as donor. 3. Interrupted conjugations revealed that 1 The following discussion is based primarily upon the work of E. A. Adelberg and S. N. Burns (I960), and F. Jacob and E. A. Adelberg (1959). 356 « ii \r iir 27 mum of its recombinants for Pro 01 ll be haved us males. 4. The male factor was linked neither to the above-mentioned loci nor to an) other chromosomal marker showing recombination. .v Like tree F, the male factor entered the F cell about five minutes after con- jugation began. 6. Treatment with acridine orange elimi- nated the male sex factor, converting the cells to F-. These findings prove that the P4x-1 male sex factor — called F' — is located extrachromo- somally. Although F could attach at any one of a number of different chromosomal sites pro- ducing Hfr chromosomes differing in O point position and in direction of transfer. F' at- tached at a particular locus near Lac in such a way that chromosomal loci were always transferred in the same order and direction. Since P4x-1 transferred its chromosome more frequently than the typical F+ (F- containing) male, the chance of F' associat- ing with the chromosome near Lac seems to be greater than the total chance of F integrat- ing at any one of numerous different loci. P4x-1. on the other hand, transferred the chromosome less frequently than P4x, sug- gesting the possibility that F is not fully inte- grated in P4x-1 ordinarily, although it is in P4x. This difference would also explain why P4x-1 had free F\ whereas P4x did not. since chromosomal F prevented the estab- lishment of free F. Ordinarily, F' seems to exist as a kind of exogenote which synapses with the chromosome just before or after conjugation is initiated. The mechanism of chromosome mobilization seems to involve a recombinational event between the episomc F' and the chromosome. When F' was transferred to F~ cells as an extrachromosomal particle, the recipient cells were converted to males that, relativelv often, could transfer their chromosome in the same sequence as P4.\ and P4.\-l males, suggesting that the chromosome o[ the ordi- nary F cell has a segment of DNA near Lac homologous to a segment carried by F'; that is. F' possesses a chromosomal segment which can pair with a particular chromo- somal locus. As mentioned, treatment with an acridine dye eliminated the extrachromosomal F' par- ticles from the P4x-1 strain and converted it to F— . Such an F strain conjugates with males carrying either F or F' extrachromo- somally. In both cases the F strain was relatively often converted into a donor (male) that transferred its chromosome in the same sequence as P4x and P4x-1 males. Clearly, then, the F~ — derived from P4x-1 via acridine orange — carries a chromosome which has retained a segment of F near Lac, the portion retained held in common by F and F'. In so far as the F portion of the particle is concerned, then, F and F' were not detectably different in these experiments. Since F' is found to be approximately twice the size of F, one can think of F' as being an F particle with an extra, particular piece of chromosome attached. The preceding suggests that F' can carry chromosomal DNA apparently still capable of replication in its new location. Let us suppose that this chromosomal segment is also still able to function normally. F' may fail to show a phenotypic effect for a normal chromosomal locus because it contains one or more (as yet unidentified) chromosomal markers. (Probably less than 1% of the chromosomal loci have been identified; more- over, there may be chromosomal regions whose only function is the maintenance of and recombination with episomes.) The very existence of F', however, encourages a search for still different F particles to which a known chromosomal marker might be attached. Bacterial Episomes and Genetic Recombination 357 Using Lac F cells and a Lac+ strain of Hfr with F integrated very close to Lac rare recombinants are obtained from inter- rupted conjugations which receive Lac+ too early. Certain of these recombinants have the following properties: 1. They receive only F and Lac+. 2. They are unstable and, occasionally, give rise to Lac~F~ individuals; hence, the original recombinant must be a merozygote carrying both Lac + and Lac~ alleles. 3. When crossed to Lac~F~ cells, they simultaneously transfer both F and Lac + with 50% or higher frequency. This trans- fer starts soon after conjugation begins, just as in the case of free F (or F'), and is unlinked to other chromosomal markers. Thus, F-Lac+ behaves as a free single unit. 4. The recombinant transfers its chromo- some in the same sequence as, but with a lower frequency (%o) than, the original Hfr line. These frequencies are exactly those found in comparing P4x-1 with P4x. 5. The F-Lac+ element can be trans- mitted in a series of successive conjugations, each recipient possessing the properties of the original recombinant. All these characteristics are most simply explained by an F particle which carries a chromosomal piece bearing Lac+ deintegrat- ing in the original Hfr strain. The attached Lac+ piece is known, moreover, to contain three cistrons governing the synthesis of /?- galactosidase, /3-galactoside permease, and the repressor for this system. From subse- quent integrations and deintegrations, one can also obtain F-Lac~ particles — composed of F-Lac with a Lac~ point mutation. Finally, another Hfr, with F integrated close to Pro, is found to produce an F-Pro particle whose properties are analogous to those of the F-Lac particle. If, however, an F-Lac + particle enters a cell containing a deletion in the Lac region, the particle, besides transfer- ring Lac ' , behaves like F in that it transfers random chromosomal markers with the low frequencies characteristic of ordinary F by F~ crosses. Since F can integrate at a variety of loci, these results suggest (and it turns out to be true) that upon deintegration any one of a variety of normal chromosomal loci can be- come a part of the genotype of cytoplasmic F and can replicate and function in the extra- chromosomal state. Particles like F' and F-Lac — substituted sex factors — represent a third type of male sex factor characterized by serving as inter- mediate donors of the usual chromosomal markers. It can be hypothesized that F, F' and other substituted sex factors are nor- mally small, double-helix, DNA ring chro- mosomes. If such small ring chromosomes were integrated in the E. coli chromosome by a single crossing over, the product would be a bigger (Hfr) ring. The now-integrated F particle would cause the enlarged ring chromosome to be open, usually at one end of F, and would mobilize the chromosome during conjugation. If the chromosome were opened at some internal position in F, part of F would be at the O point and part at the opposite terminus. Some evidence suggests that openings of this type sometimes occur. F may deintegrate by an internal crossing over, producing free ring F and the ring E. coli chromosome. It should be em- phasized that the suggested circular model of F is based upon no evidence and is nothing more than speculation at present. When F is not integrated, only one or a few F particles are present per E. coli chro- mosome— at least in cells that have carried F for some time. (This case is similar to the situation in which organisms with more than one chromosome are regulated by some mechanism in the cell which permits each chromosome to replicate only once a gen- ;r>s CHAPTER 27 eration.) Populations of donors, grown to saturation density in aerated broth or cul- tured on agar overnight, can lose their donor phenotype temporarily and behave as genetic recipients. Since they retain their sex factor yet behave as F cells, they are known as /•" phenocopies ( see p. 317). If a /^/c~F~ phenocopj carrying F is mated with an F^ male carrying F-Lac + , exconjugants can be obtained that carry both types of F. Soon, however, in some experiments, one or the other F particle persists in the progeny. This adjustment shows that there is some regula- tion of the number of F partieles per nuclear body. Hfr which are F~ phenocopies, do not tolerate the presence of an introduced autonomous sex factor. F not only mobilizes the entire chromo- some in Hfr cells, but it also mobilizes merogenotes — in an F-merogenote trans- fer - — as shown in the preceding discussion. This latter process has also been called sex- duction, F-duction, or F-mediated transduc- tion. With respect to such transduction, sub- stituted sex factors resemble Adg, just as F resembles A. One can construct a haploid Hfr containing two attached F factors so integrated that, upon mating, the chromo- some is present in two pieces — 1 4 and -'{. its length — each with an F at the end, and both merogenotes capable of being transferred to the F~ cell. An F-merogenote carrying the markers Pur, V6, and Lac transfers the merogenote so that the entry order — deter- mined by studies of spontaneous and artifi- cial interruptions of mating — is 0-Pur-V6- Lac-F. In all these respects, then, whole chromosome and merogenote mobilization by F appear identical, differing only with regard to the length of the genetic segment transferred. Although the merogenote markers can -' The remainder of this section is based upon A. J. Clark and E. A. Adelherg (1962). A. M. Camp- hell (1962). and W. Hayes (1964). sometimes integrate into the chromosome, F-linked merogenotes can also persist and replicate without integration, forming clones of merozygotes or partial diploids. The longer the merogenote — when it consists of 49? or more of the genome — the more un- stable it is. Extrachromosomal F can pass from male to female during conjugation. Deintegration and integration of F can result in a two- directional flow of chromosomal genes be- tween F and the chromosome in the same cell. When F is at a chromosomal locus, the chromosomal genes are rendered mobile so that chromosomal genes can be trans- ferred to another cell. F can also enter an F~ cell by transduction, in which case it is recovered only in the free state, even if the donor was Hfr. (During phage growth, fragmentation of the chromosome takes place at or close to the ends of the integrated F element.) Consequently, the F particle is directly involved in genetic recombinations within and between bacteria which involve F itself and chromosomal genes. Promoters A promoter is a genetic element which pro- vides one or more of the special conditions needed for genetic transfer via conjugation. If a genetic element (like F) performs all the functions of a promoter, including mobiliza- tion of the entire chromosome or of a mero- genote, it is called a sex factor. Sometimes a promoter (such as F) promotes only its own transfer (into F~ cells). At other times a promoter and a whole chromosome or a merogenote are transferred in linkage. F-Lac promotes the transfer of both the merogenote and the chromosome. After ultraviolet treatment of F-Lac heterogenotes, some individuals are found no longer able to transfer Lac or chromosomal markers. Ap- parently a mutation in F resulted in a loss of one or more promoter functions. When Bacterial Episomes and Genetic Recombination 359 Salmonella or Shigella act as recipients in crosses with F+ or Hfr E. coli, F is trans- ferred but sometimes is unable to act as a sex factor until it is sent back into F~ E. coli. Such results show that promoter functions can be temporarily inhibited or unexpressed, depending upon the host genotype. * Resistance Transfer Factors A genetic agent — the resistance transfer fac- tor, RTF — has been found in Shigella.3 RTF is a promoter which causes conjugation and the transfer of a series of different, linked, drug-resistance genes; it is, therefore, a sex factor. Under special circumstances, RTF can mobilize chromosomal loci. It can be transferred from Shigella to Escherichia or Salmonella independently of the chromo- some, and such recipients can be cured of RTF by acridine dyes. RTF promotes its own transfer which starts within one minute of mixing the parents, and it replicates au- tonomously. Although RTF remains trans- ferable when introduced into an F+ or Hfr cell, F-promoted chromosome transfer in Hfr cells is reduced one hundredfold, and F- promoted merogenote transfer and the trans- fer of free F are completely inhibited by cer- tain RTFs. When RTF is spontaneously lost, these F functions are restored. Other RTF strains have no effect on F function. F produces an antigen on the cell surface which must be present for <£f2 to attack males. When RTF and F are both present, a new RTF-antigen replaces the F-antigen. When a cell is infected by RTF its chromo- somal markers are mobilized one hundred times more frequently when its chromosome carries a segment of F than when it does not. These observations suggest that at least a partial genetic homology exists between RTF and F. In view of this and other evidence, it is concluded that RTF has some relation- ship with the chromosome, although it may 3 See T. Watanabe (1963). not be able to assume a stably integrated state. RTF has been called an episome by its discoverers. * Colcinogenic Factors ' Many strains of enteric bacteria (Escher- ichia, Salmonella, Shigella, for example) pro- duce one or more highly-specific, antibiotic substances called colicins. Colicins are bac- tericidal but not bacteriolytic agents; a thou- sandth of a microgram of colicin can kill a million sensitive E. coli cells. More than a dozen groups of colicins are known; each group is designated by a different capital let- ter; each adsorbs to a different receptor site at the cell surface. Different colicins belong- ing to the same group can be distinguished by other characteristics. Colicins have a high molecular weight; two of them, colicin K and colicin V, have been purified and iden- tified as lipocarbohydrate-proteins. These purified colicins seem to be the same as the O antigen of the bacteria."1 The antigen molecule can be separated into a lipocar- bohydrate and a protein fraction, the latter having all the colicidal activity. A cell able to produce a colicin is colicino- genic; one without this ability is noncolicino- genic. Colicinogeny is stable and can be trans- mitted through thousands of cell generations. Although it can be lost spontaneously, spontaneous acquisition of colicinogeny has never been observed. Consequently, it is reasonable to suppose that colicinogenic bacteria possess genetic material — colicino- genic (col) factors — that govern the syn- thesis of different colicins. Not only are these factors transmitted to progeny via vege- tative reproduction, but new strains can be- come colicinogenic through bacterial conju- gation or phage-mediated transduction. Only a small fraction of the bacteria in a 4 This section follows the work of P. Fredericq (1963). See also W. Hayes (1964). "See W. F. Goebel (1962). CHAPTER 27 colicinogenic culture actually produce coli- cins. Colicin is lethal to the bacterium syn thesizing it. but colicinogenic cells which do DOl \ icld colicin are viable and immune when exposed to corresponding colicin. Colicin synthesis can be induced in nearly all cells of a colicinogenic culture after exposure to ultraviolet light, nitrogen mustard, or hydro- gen peroxide. Thus, in several of the prop- erties mentioned, colicinogeny and lysogeny are similar, the col factors behaving in these properties like the prophages of temperate phage. When a strain is both lysogenic and colicinogenic, induction often releases either phage or colicin but not both. Bacteria can be protected against colicins via immunity or resistance. A colicinogenic bacterium, although immune to the corre- sponding colicin, can possess receptor sites which make it susceptible to colicins of other groups. Certain other genes confer resist- ance to whole groups of colicins by causing the loss of receptors. As noted, colicin K and the O antigen are apparently identical. When the receptor sites for colicins and for virulent phages are studied, in a number of cases colicin and phage are found to share receptor sites; for example, receptor sites are shared by colicin K and #T6; colicin E and ^BF-23; colicin C and T\ or <^>T5. Since virulent phage attach to receptors by means of a protein located at the tip of their tails, colicin and tail-tip protein appear to he very similar. In serological tests, however, colicin and phage sharing a common receptor have not been found to exhibit any cross reaction, and colicinogeny does not confer immunity to infection by a site-sharing phage. When bacteria are exposed to the protein coat, or phage ghost, of 4>T2, all protein synthesis (and possibly that of RNA and DNA, too) is halted. The same effect is produced by colicin. Bacteria, however, may recover after exposure to phage ghosts or to colicin ( if the latter is removed by enzymatic diges- tion). Sometimes, virulent phage can kill its host without reproducing; in such cases. the constituent responsible lor lethality is a tail-tip protein. When T6 is the killer, the lethal protein has the same X-ray inactiva- tion curve, specificity, and receptor site as colicin K. Since the lethal protein of T6 is very simi- lar to colicin K. it seems reasonable that T6 and col K are homologous with respect to at least one gene. Col K can be thought of as a virulent phage missing that portion of the genome required to lyse the cell and to give rise to particles whose infectivity is inde- pendent of conjugation, yet with enough of the phage genome persisting to make the cell colicinogenic. Labeling experiments show that three col factors studied contained DNA in the amount of 4 to 7 times 10' nucleotide pairs — about one-tenth the amount in F or phage. During conjugation, E. coli F+ cells trans- fer col El with high frequency, so that col El can exist autonomously. Autonomous col factors can arrive in the F~ cell as early as 2U minutes after conjugation is initiated. Since Hfr E. coli carrying col V, col I, or col E2 do not transfer them in linkage, these col factors give no evidence for an attached state. In Salmonella, col I in the absence of F is transferable via conjugation. Although cells carrying only col El cannot transfer it during conjugation, col El cells infected by col I transfer both col factors. Consequently, col I promotes the transfer of col El. When Salmonella harbors only col I, transfer of the chromosome in conjugation occurs but is rare. When such cells are also infected with col El, chromosome transfer increases one hundredfold. As a result, in Salmonella col I promotes the transfer of col El, and col El promotes the transfer of the chromosome. Col I is, therefore, a sex factor. Although they do not cure colicinogeny, acridine dyes inhibit the transfer of col factors. When F col~ is crossed with F~ col I, Bacterial Episomes and Genetic Recombination 361 F is transferred to the F^ col I conjugant, phages, ,; F factors, RTF factors, and col fac- and col I is transmitted to the F+ col~ con- tors. It is still too early, however, to make jugant with high efficiency. However, when any kind of definite suggestion as to the pre- F+ col I is crossed with F~ col~ , col I is cise nature of the evolutionary interrelation- transferred at a low rate, so that F interferes ships, if any, among these various types of with transfer of col I from the same parent. episomal elements. The preceding shows a number of simi- ,. „ , „ „ „ _ , . . r . , ° See also E. Seaman, E. Tarmy and J. Marmur lanties among temperate and virulent (1964). SUMMARY AND CONCLUSIONS A two-directional flow of F and chromosomal genetic material occurs between extra- chromosomal F and the chromosome. Chromosomal markers carried by free F retain their capacity for replication and phenotypic action. F', other F-merogenotes, RTF factors, and some col factors behave like episomes. Col factors have several characteristics in common with F and the provirus stage of temperate phages. Their products, colicins, resemble the lethal, tail-tip protein of virulent phages. Accordingly, it is possible that F, temperate phage, virulent phage, and col factors are related in evolution. Certain bacterial episomes promote their own transfer and/or transfer of other genetic material via conjugation. In Salmonella, col I promotes the transfer of col El which. in turn, promotes the transfer of the chromosome. RTF promotes the transfer of linked drug resistance markers and can inhibit certain promoter functions of F. Since they can initiate conjugation, F and its derivatives, col I, and RTF are sex factors. REFERENCES Adelberg, E. A., and Burns, S. N., "Genetic Variation in the Sex Factor of Escherichia Coli," J. Bact., 79:321-330, 1960. Reprinted in Papers on Bacterial Genetics, Adelberg, E. A. (Ed.), Boston: Little, Brown, 1960, pp. 353-362. Campbell, A. M., "Episomes," Advances in Genetics, 11:101-145, 1962. Clark, A. J., and Adelberg, E. A., "Bacterial Conjugation," Ann. Rev. Microbiol., 16:289-319, 1962. Fredericq, P., "On the Nature of Colicinogenic Factors: A Review," J. Theoret. Biol., 4:159-165, 1963. Goebel, W. F., "The Chromatographic Fractionation of Colicine K," Proc. Nat. Acad. Sci., U.S., 48:214-219, 1962. Hayes, W., The Genetics of Bacteria and their Viruses, New York: J. Wiley & Sons, 1964. lacob, F., and Wollman, E. L., "Episomes, Added Genetic Elements" (in French). C. R. Acad. Sci. (Paris), 247:154-156, 1958. Translated and reprinted in Pa- pers on Bacterial Genetics, Adelberg, E. A. (Ed.), Boston: Little, Brown, 1960, pp. 398-400. Maas, R., "Exclusion of an F Lac Episome by an Hfr Gene," Proc. Nat. Acad. Sci., U.S., 50:1051-1059, 1963. Seaman, E., Tarmy, E., and Marmur, J., "Inducible Phages of Bacillus subtilis," Bio- chemistry, 3:607-613, 1964. Watanabe, T., "Infectious Heredity of Multiple Drug Resistance in Bacteria," Bact. Rev., 27:87-115, 1963. ( HAPTER 27 QUESTIONS FOR DISCUSSION 2". I. Do .ill matings transfer F particles oi one genotype or another? Explain. 27.2. Discuss the relationship between the transmission ol tree F particles and a seg- ment of the male chromosome. 27.3. Discuss the realit) of a bacterial "chromosome" and its linear arrangement. 27.4. B\ what series of events can you explain the origin of strain P4.\-l from P4x? 27.5. From which particular Htr strain of E. coli could you obtain an F-Pro (pro- line) merogenote? How'.' 27.6. How do \ou suppose episomes originate? 27.7. Are integrated episomes and episomal derivatives generally able to break chro- mosomes? Explain. 27.8. It has been found that c/>P2. a temperate phage which normally integrates at a particular chromosomal locus, position I, loses the extreme preference for position I when liberated from a strain carrying it in position II. How can you explain this finding? 27.9. Discuss the statement: "The ability to pass from the integrated to the free state or vice versa is possessed by every gene in a bacterial cell." 27. 1U. A transducing phage can carr\ two closely-linked prophages obtained from a doubly-lysogenic host. What conclusions can you draw with regard to the nu- cleotide content of a mature phage and a prophage? 27.11. The Lac gene can either be chromosomal when integrated into the chromosome, or extrachromosomal when attached to free F. Should such a gene be consid- ered an episome? Why? 27.12. How would you locate the position of the UV-inducible prophage of <£434 in the E. coli linkage map? 27.13. How would you locate the prophage site of a noninducible (by ultraviolet light or zygote formation ) phage? 27.14. Lactic dehydrogenase contains a single-strand sequence of about 33 deoxyribo- tides. Should this portion of the enzyme be considered genetic material? Ex- plain. 27.15. During conjugation. F is reported to enhance the transfer to the F~ cell such substances as lactose, ultraviolet-irradiation products that induce A prophage, and a A repressor. Is F acting as a promoter in these cases? Explain. Chapter 28 RNA AS GENETIC MATERIAL I n previous chapters, the DNA- containing phages were the only viruses discussed in detail. In this chapter, we study another group — vi- ruses that contain no DNA and are entirely, or mainly, ribonucleoprotein in content. Members of this group include many of the smaller viruses that attack animals (causing poliomyelitis, influenza, and encephalitis, for example), many viruses that attack plants (such as the tobacco mosaic and the turnip yellow mosaic viruses), and the small RNA- containing bacteriophages1 (f2, MS2, R17. and others ) . These phages are all extremely similar, but not identical. They are the same size, shape, and molecular weight; they cross react serologically, having similar coat proteins; all attack only male (Hfr or F+) E. coli. The usual host for the influenza virus is the mammalian cell. This virus consists of a helical ribonucleoprotein core surrounded by a lipoprotein membrane. It was shown that the lipids in the envelope of the influenza virus are derived mainly from pre-existing lipids of the host cell, and that the composi- tion of the lipids varies with the strain of the host cell. The outer membrane of the virus is apparently derived from the cell membrane and applied when the virus leaves the cell. After infection by the virus, normal cellular growth continues for several hours. There- 1 See T. Loeb and N. D. Zinder (1961). J. E. Davis and R. L. Sinsheimer ( 1963), and S. Mitra. M. D. Enger, and P. Kaesberg (1963). 363 fore, most of the RNA, protein, and DNA synthesized are normal cellular products and bear little relation to the growth of the virus. Using the drug, actinomycin-D, to inhibit normal cellular RNA synthesis, one can demonstrate a specific synthesis of viral RNA. Moreover, with the closely related Newcastle disease virus, which grows in the cytoplasm, one can show that the new (viral) RNA appears in the cytoplasm and not. as in normal cells, in the nucleus. It is probable, therefore, that the internal viral RNA and protein, as well as other viral antigens, such as the hemagglutinating factors, are made under the direction of the virus inside the cell. Several genetically-different, haploid strains of influenza virus have been isolated; for example. SWE (with markers, a c) and MEL (with markers. AC). When a mix- ture of the two strains is used to multiply- infect a chick's egg membranes, the mixed infections give progeny particles which, when tested, yield pure clones not only of the parental genotypes but also of stable recom- binant types {A c or a C). Since other ex- planations can be excluded, the results prove that genetic recombination occurs also be- tween RNA-containing viruses.2 Genetic re- combination has also been demonstrated for the poliomyelitis virus. Whereas the occur- rence of genetic recombination in influenza may require incorporation of two or more pieces of viral RNA into a single particle, recombinant polio-virus RNA seems to occur in one piece.3 Consequently, although the details of recombination between RNA vi- ruses are unknown, more than one mecha- nism may be involved. No evidence has been obtained for the occurrence of genetic recombination among viruses that attack plants. In the case of the tobacco mosaic virus (TMV), infection is - Based upon the work of F. M. Burnet and others. •See G. K. Hirst (1962). 364 CHAPTER 2S brought about experimental!} by rubbing a sample of virus on the leaf surface. Even when a high concentration of virus is used, onlj a small fraction of the virus particles (one in 10'') find and penetrate susceptible cells and give rise to a detectable lesion. For figure 28-1. Electron micrographs of to- bacco mosaic virus (I Ml) showing its gen- eral configuration (top) and its hollow core {middle). The bottom photo shows a particle whose protein has been partially removed by treatment with detergent, leaving a thinner strand of RNA. (Courtesy of R. G. Hart.) this reason, it is difficult to multiply-int'ect a tobacco cell, and experiments testing for genetic recombination are probably negative due to the lack of mixed infections. The TMV particle is a cylinder 3000A long and about 80A in radius (Figure 28-1, top). It has a molecular weight of about 40 times 106, of which 38 times 106 is pro- tein and 2 times 10,; is RNA. The outer dimensions of TMV are attributed to the helical aggregation of about 2200 identical protein subunits, each of which has a molec- ular weight of about 18,000 and contains 158 amino acids in a single polypeptide chain (Figure 28-2). A cross section of the TMV particle shows a hollow core about 20A in radius (Figure 28-1, middle); the protein subunit, therefore, adds about 60A to the radius. The RNA in a particle (Fig- ure 28-1, bottom) is typically a single, un- branched strand consisting of some 6400 nucleotides threaded through the protein subunits at a radius of 40A. Internally the RNA is normally covered by about 20A and externally by about 40A of protein subunit. Since the protein subunits are arranged in a gently-pitched helix (49 subunits per three turns), the RNA forms a helix of the same pitch. When TMV in water is treated with phenol, the protein of the virus is extracted into the phenol, leaving the single RNA molecule intact in the water. If the tobacco plant is exposed to RNA molecules with pro- tein thus removed, the frequency of infection is about 500 times less than the frequency obtained with an equal number of whole virus particles; typical TMV progeny (com- plete with TMV protein coats) are produced. Repeated phenol treatments do not further decrease the infectivencss of the RNA. and no amount of protein can be detected in the preparations. RNasc, on the other hand, completely destroys the infectivencss of the RNA fraction but not the infectiveness of the whole virus. It must be concluded, RNA as Genetic Material 365 1 Acetyl N-Ser+Tyr — ► Ser -► lieu— • 5 >Thr- NH2 10 1 5 ♦Thr — ►Pro— ► Ser -►Glu -►Phe -►Vol -►Phe -►Leu — ►Ser-^Ser ->^ 30 NH2 s- Ale* — Asp*— Thr •*-CySI- l-*-Asp-*- 25 Leu* NH2 20 — lieu-* — Leu-* — Glu-* — //«c* — Pro*-/isp*-a/o *-Try* -Ala *T » Leu— ►Gly — ••Asp NHZ — ►Glu- 35 -►Phe— » NH2 >Glu— NH2 NH2 40 ^_ -►Thr — ►Glu — ►Glu — ►Ala — ► Arg^-^Thr — ►Val- NH2 -►Glu- 45 ♦Vol -v 60 s- Vol-*— Thr-*— Vol" NH2 • — Glu" •— Pro-* — 55 Ser-* 50 nhz -Pro-* — Lys* — Try-*— Vol-* — Glu-*— Ser *— Phe- NH2 *— Glu*- -Arg J V Arg ^ Phe -►Pro — ►Asp 65 -►Ser— 1 •Asp- 70 ^— N ► Phe— ► Lys ! — ►Vol — ►Tyr — »— ► Ser — ► Ala- 125 -►Asp— • ■lleu- NH2 130 ♦>asD-^'ei> — Weu-»VD' — ►G/tv— ►Leu — ►lleu- — *(ArgV-»Gly -v 150 ** Leu*— Gly* — Ser' *— Ser- •-Ser*- 145 ^-~. 140 NH2 -Glu-*— Phe-* — Ser* — Ser*— (Arg)*— Asp*— Tyr •*— Ser- •—Gly* -Thr 4f V Val— ►Try— ►Thr —►Ser 155 -►Gly— ► Pro- 158 ►Alo — ►Thr figure 28-2. Amino acid sequence in the protein building block of tobacco mosaic virus (TMV). There are 158 amino acids in the sub-unit, the encircled residues indicate the points of digestion by trypsin. (Courtesy of A. Tsugita, D. T. Gish, J. Young, H. Fraenkel-Conrat, C. A. Knight, and W . M. Stanley, Proc. Nat. Acad. ScL, U.S., 46:1465, 1960.) therefore, that pure virus RNA is infective and carries all the genetic information neces- sary for its replication.* These experiments also prove that TMV protein plays no part — other than protecting the RNA and increas- ing the infectivity — in the replication either of the RNA genetic material or of itself. This conclusion is tested by what is called reconstitution experiments in which, under appropriate conditions, we can first separate the protein and RNA of TMV and then have them recombine and demonstrate the high infectiveness of the original virus. Using two genetically-different strains of this virus — the standard (TMV) and Holmes rib grass (HR) — a highly-infective virus con- 4 RNA isolated from a number of small animal viruses and from coliphages f2, MS2, etc., is also infective. (See Vol. 27, Cold Spring Harb. Sym- pos. Quant. Biol., 1962, and D. E. Engelhardt and N. D. Zinder, 1964.) taining the RNA of TMV and the protein coat of HR can be constructed. The prog- eny obtained are typically TMV with both TMV RNA and protein coat. The recipro- cal construct, a virus with HR RNA and TMV protein, produces HR progeny typical in both RNA and protein. Thus, only the RNA of a TMV particle specifies the RNA and protein of the progeny virus.5 Muta- tions can be induced in TMV by many of the agents known to be mutagenic for DNA. Such results and others prove that the bio- logical activity of the RNA depends upon its primary (nucleotide content) and not its sec- ondary (coiling pattern) structure. The complete amino acid sequence in the protein building block of tobacco mosaic 5 The genetic experiments described for TMV are based largely upon work of H. Fraenkel-Conrat and R. C. Williams (1955), A. Gierer (1960). G. Schramm, and others. 366 ( II \l- I IR 2MS2 — called the "plus" strand — as a template to synthesize the comple- mentary "minus1' RNA strand in vivo. The double-stranded product — called the replica- tive form — is used in vitro as a natural tem- plate by the same or another RNA synthetase to synthesize "plus" strands." In this con- nection it should be noted that the infective forms of a wound virus obtained from sweet clover and a reovirus associated with the respiratory and enteric tracts of animals, in- cluding man, have double-stranded RNA as their genetic material. 10 s By I. Haruna, K. Nozu, Y. Ohtaka, and S. Spie- gelman (1963), and by C. Weissmann, L. Simon, and S. Ochoa (1963). See D. Baltimore (1964). °See C. Weissmann et al. ( 1964). 10 See P. J. Gomatos and I. Tamm (1963). SUMMARY AND CONCLUSIONS RNA is the sole carrier of genetic properties in certain viruses. Some animal RNA viruses can undergo genetic recombination. Mature RNA viruses carry either single- stranded or double-stranded RNA. Viral RNA replication involves RNA synthetase and the formation of complementary RNA chains. RNA as Genetic Material 367 REFERENCES Baltimore, D., "/// vitro Synthesis of Viral RNA by the Poliovirus RNA Polymerase." Proe. Nat. Acad. Sci., U.S.. 51:450-456, 1964. Burnet. F. M.. and Stanley, W. M. (Eds.), The Viruses; Vol. 1. General Virology, 609 pp.; Vol. 2, Plant and Animal Viruses, 408 pp.; Vol. 3, Animal Viruses. 428 pp.. New York: Academic Press, 1959. Davis, J. E., and Sinsheimer, R. L., "The Replication of MS2. 1. Transfer of Pa- rental Nucleic Acid to Progeny Phage," J. Mol. Biol., 6:203-207, 1963. Engelhardt, D. E., and Zinder, N. D.. "Host-Dependent Mutants of the Bacteriophage f2. III. Infective RNA. Virology, 23:582-587, 1964. Finch, J. T., "Resolution of the Substructure of Tobacco Mosaic Virus in the Electron Microscope," J. Mol. Biol., 8:872-874, 1964. Fraenkel-Conrat. H., and Ramachandran, L. K., "Structural Aspects of Tobacco Mo- saic Virus," Advances in Protein Chemistry, 14:175-229, 1959. Fraenkel-Conrat, H., and Williams, R. C, "Reconstitution of Tobacco Mosaic Virus from Its Inactive Protein and Nucleic Acid Components," Proc. Nat. Acad. Sci., U.S., 41:690-698, 1955. Reprinted in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs. N.J.: Prentice-Hall, 1959, pp. 264-271. Gierer, A.. "Ribonucleic Acid as Genetic Material of Viruses," in Microbial Genetics, Hayes, W., and Clowes, R. C. (Eds.), Cambridge: Cambridge University Press, 1960. pp. 248-271. Gomatos, P. J., and Tamm, I., "Animal and Plant Viruses with Double-Helical RNA," Proc. Nat. Acad. Sci.. U.S., 50:878-885. 1963. Hirst, G. K., "Genetic Recombination with Newcastle Disease Virus, Polioviruses, and Influenza," Cold Spring Harb. Sympos. Quant. Biol., 27:303-309, 1962. Home, R. W., "The Structure of Viruses," Sci. Amer., 208 (No. l):48-56, 170-171, 1963. Hudson, W. R., Kim, Y. T., Smith, R. A., and Wildman, S. G., "Synthesis of Tobacco Mosaic Virus Infectivity by Cell Free Extracts," Biochim. Biophys. Acta, 76:257- 265. 1963. Loeb, T., and Zinder. N. D., "A Bacteriophage Containing RNA," Proc. Nat. Acad. Sci.,' U.S.. 47:282-289, 1961. Mitra, S., Enger, M. D., and Kaesberg, P., "Physical and Chemical Properties of an RNA from the Bacterial Virus R17." Proc. Nat. Acad. Sci.. U.S., 50:68-75, 1963. Reddi, K. K., "Studies on the Formation of Tobacco Mosaic Virus Ribonucleic Acid. V. Presence of Tobacco Mosaic Virus in the Nucleus of the Host Cell," Proc. Nat. Acad. Sci., U.S., 52:397-401, 1964. Richter, A., "Structure of Viral Nucleoproteins," Ann. Rev. Microbiol., 17:415-428, 1963. ' Shipp W and Haselkorn, R.. "Double-Stranded RNA from Tobacco Leaves Infected with TMV," Proc. Nat. Acad. Sci., U.S., 52:401-408, 1964. Singer, B., and Fraenkel-Conrat, H., "Studies of Nucleotide Sequences in TMV-RNA. I.' Stepwise Use of Phosphodiesterase," Biochim. Biophys. Acta. 72:534-543, 1963. Spirin, A. S., "Some Problems Concerning the Macromolecular Structure of Ribonucleic Acids,"' Progr. Nucleic Acid Res.. 1:301-345, 1963. 368 < mm- i ik 28 rsugita, \. Gish, IV l . Young, J., Fraenkel-Conrat, H.. Knight, C. A., and Stanley, W. M.. "The Complete Amino Acid Sequence of the Protein of Tobacco Mosaic Virus," Proc. Nat. Acad. Sci., U.S., 46:1463-1469, I960. Weissmann, c .. Borst, P., Burdon, R. H.. Billeter, \1. A., and Ochoa, S.. "Replication oi Viral RNA, IV. Properties of RNA Synthetase and Enzymatic Synthesis ol MS2 Phage RNA." Proc. Nat. Acad. Sci.. U.S.. 51:890-897, 1964. QUESTIONS FOR DISCUSSION 28.1. What conclusions can you draw from the observation that within a day or so alter infection with a single particle of TMV, the cell can produce about 50,000 viral nucleic acid molecules and about 100 > 106 protein subunits? 28.2. Using pure TMV RNA for infection, how could you test whether this RNA contains information for manufacturing TMV protein? 28.3. Compare transformation with infection by pure virus nucleic acid. 28.4. Discuss the view that cancerous growths may originate as a result of virus in- fection activating RNA replication. 28.5. What conclusions can you draw from the observation (see reference to J. E. Davis and R. L. Sinsheimer) that the RNA of the parental MS2 phage is rou- tinely excluded from the progeny phage? 28.(i. What are the similarities and differences in the behavior of DNA polymerase and RNA synthetase? 28.7. Discuss the possible advantages and disadvantages of double-stranded rather than single-stranded RNA as the genetic material of a mature virus. Chapter *29 EXTRANUCLEAR GENES T! (he DNA rabbit poxvirus and certain RNA viruses, which are restricted to the cytoplasm of cells normally possessing a definite nuclear membrane (and hence a definite nucleus), are clearly autonomous extranuclear genes. To what extent do extranuclear genes occur and what is their relationship with particular chromosomal genes? Various operational tests — chemical, re- combinational, mutational, phenotypical, and replicative — can be applied to identify an extranuclear component as genetic material. Recombination is the first operational method used in our search for extranuclear genes. To detect recombination, the extra- nuclear gene must produce a recognizable phenotypic effect. To provide the required phenotypic alternatives, changes involving either the kind or the quantity (or both) of such a gene must occur. Drosophila How do we actually proceed to look for an extranuclear gene in Drosophila? Starting with different nonoverlapping phenotypic al- ternatives which occur generation after gen- eration under the same environmental con- ditions, a series of crosses is made to test whether the occurrence of the alternatives is associated with the presence of one or more particular chromosomes (X. Y, II, III, IV). If it is, the phenotypic alternatives are prob- ably due to some genetic factor linked to, and hence located in, a chromosome. (Addi- tional appropriate crosses and cytological 369 studies will reveal the precise nature of the nuclear gene change. Since the vast ma- jority of carefully analyzed gene-based traits are located in chromosomes, the search for an extranuclear gene will usually be unsuc- cessful. ) But consider the genetic alternatives for resistance and susceptibility to C02 (gas). Although wild-type Drosophila adults can be exposed to pure COL» for as long as 1 5 min- utes and recover without apparent effect, flies of other strains almost invariably are killed by such exposure. Using marked chromo- somes, COo-sensitivity is found to be un- linked to any chromosome of the normal genome. In fact, by appropriate crosses it is possible to replace each of the chromo- somes in the sensitive strain by a correspond- ing chromosome of the resistant strain. After this is done, the flies produced are still sensitive to COo! Possibly, the sensitive strain carries an additional nonhomologous nuclear chromosome which it transmits inde- pendent of the usual ones. In the progeny of hybrids derived from sensitive and resist- ant lines the COo-sensitivity trait does not segregate, which indicates that if such a supernumerary chromosome exists, it cannot occur singly (in the individual hybrid for sensitivity) or as a pair (in flies of the pure sensitive strain). Although cytological examination reveals no additional nuclear chromosome, this finding is not a conclusive argument against a nuclear locus for COo- sensitivity, for, according to recombinational evidence, "chromosomes" so small they es- cape cytological detection are known to exist. (A phenotypic change in corn is associated with the presence of readily detectable, su- pernumerary, heterochromatic, "B" chromo- somes.) Although the sensitive female regularly transmits COL>-sensitivity to some progeny, the sensitive male does so only under special circumstances. It might be possible that a nuclear gene for sensitivity is somehow ex- 370 ( II M'TER 29 eluded from a nucleus destined to be in a sperm but not from one destined to be in an egg, although it seems much more reasonable to attribute the nontransmission of C'(). -sen- sitivity through the sperm to the rather minute amount of cytoplasm in a sperm as compared with the amount in an egg. It is therefore highly probable that CO_.-sensitivity in Drosophila is due to the presence of a particle called sigma. Other studies show that sigma contains DNA,1 is mutable, and has many of the characteristics of a virus including infectivity by experimental means. - Since sigma is not visible, its location within the cell remains somewhat of a mystery. Certain sigma and episome characteristics are similar. (Melanotic tumor incidence in Drosophila may also depend upon the pres- ence of an episome-like particle. :; ) Consider another trait of Drosophila (p. 1 10) — females mated to normal males giving rise almost entirely to females. This trait has a genetic basis; is not transmitted by males; is infective; is not linked to the usual chromosomes; and proves to be intimately associated with the presence of a spirochaete in the blood. Maize None of the examples just mentioned demon- strates conclusively the existence of both in- tracellular and extranuclear genes. They do serve to illustrate, however, possible results of a search for such genes which starts with a study of genetic recombination. The de- sirability of making a direct correlation be- tween potentially extranuclear genes and ob- jects observable in the cytoplasm is clear. Continuing the search for extranuclear genes, let us restrict our attention to cyto- plasmic components which seem to be nor- 1 See N. Plus (1963). Much work on sigma has been done by P. L'He- ritier, G. Teissier, and co-workers. 3 See C. Barigozzi (1963). mal constituents of present-day cells, disre- garding their normality when they or their precursors first arose. Many plant cells contain cytoplasmic bod- ies called plastids. Green plastids (due to chlorophyll) are called chloroplasts; white plastids are called leucoplasts. Immature plastids are small and colorless. In the ab- sence of sunlight, chloroplasts lose their pig- ment and become leucoplasts; the process is reversed when the plastids are again exposed to sunlight. In corn, mutants of chromosomal genes can interfere with the sequence of reactions leading to the manufacture of chlorophyll. One such nuclear gene prevents plastids from producing any chlorophyll at all, so that a type of leucoplast incapable of becoming green occurs. A seedling that possesses the appropriate mutant nuclear genotype will not be green; will grow only until it exhausts the food supply in the seed; will die because photosynthesis of sugar cannot occur in the absence of chlorophyll. Nuclear genes that produce albino seedlings act, therefore, as lethals. Certain corn plants have mosaic leaves, with stripes of green and white (Figure 29-1 ).4 Although the leucoplasts of the white parts are incapable of becoming green, the white parts survive by receiving nourish- ment from the green parts. Is this mosai- cism based upon nuclear genes causing differ- ent portions of the leaf to follow different paths of development? Were striping due to a nuclear gene acting upon differentiation, such a gene would have to be transmitted through the male or female gamete inde- pendent of the whiteness or greenness of the tissue giving rise to the reproductive struc- tures. Sometimes an ear of corn is derived from an ovary that is expected to be mosaic be- 4 The following account is based primarily upon work ot M. M. Rhoades. Extranuclear Genes 371 figure 29-1. Marcus M. Rhoades (in J 959) examines striped corn plants in the foreground. Unstriped corn plants are in the background. cause it originated partly from green and partly from white tissue. When the kernels in such an ear are planted in rows corre- sponding to their positions in the cob, the result is not all green, all white, all striped, nor a random mixture of these types, but groups of green and albino seedlings (Fig- ure 29-2). This outcome suggests that striping actually occurs in the ovary and per- sists in the cob. Other tests of this strain show that the greenness or whiteness of a seedling is independent of the color of the parental part forming the pollen used to produce the seed. Moreover, appropriate crosses show that none of the genes in the paternal or maternal chromosomes is in- 372 CHAPTER 29 volved. The striping effect, therefore, is not due to a nuclear gene acting differently in different tissues. Since the pollen grain is not known to carrj plastids, and since the only deciding factor proves to be the color of the tissue giving rise to the ovary, it can be concluded that only the nature of the plastids within different ova is important in determining seedling color in this case. All these facts suggest that plastids are derived only from pre-existing plastids and that daughter plastids have the same color capaci- ties as the parent plastid. This hypothesis is subjected to further test by examining the cytoplasm of cells located at the border between white and green tissue. I hese cells are found to contain mature plastids o\' both fully green and completely white types, whereas cells within a green sector contain only green mature plastids. and cells in a white sector contain only leuco- plasts. Thus, even when the two kinds of mature plastids are present in the same cell, they have no influence upon each other but develop according to their innate capacities. If a zygote (or other cell) containing both kinds of plastids produces daughter cells which happen to receive only "white" or only "green" plastids, these daughter cells will give rise to sectors of white and of green tissue. From the results presented (in addi- tion to other evidence not mentioned ) it can FIGURE 29-2. Groups of albino and non-albino seedlings from kernels plained in rows corresponding to their positions in a cob produced on a green-white striped plant. Extrcmuclear Genes 373 figure 29-3. Normal (above) and kappa-containing (right) Par- amecium. (Courtesy of T. M. Sonneborn.) be concluded that plastids do not arise except from plastids. Consequently, since they are self-replicating, mutable, and capable of rep- licating their mutant condition, plastids ap- parently contain at least one cytoplasmic gene. Although DNA is present in chloro- plasts,5 this substance has not yet been proved to be the basis for the genetic alterna- tives under discussion. As already men- tioned, the chlorophyll trait is also influenced by nuclear genes. Thus, this trait is con- trolled by both the plastid and the nuclear genotypes. In another study, a cross of two all-green corn plants gives some progeny which are green-and-white striped. The striped plants "See R. Sager and M. R. Ishida (1963). and M. Edelman, C. A. Cowan. H. T. Epstein, and J. A. Schiff (1964). prove to be homozygous for a recessive nu- clear gene, iojap (//), for which their parents were heterozygous. Since colorless plastids in ova of striped plants remain colorless in subsequent generations, even in homozygotes for the normal allele, the colorless plastid is not due to interference by if if in the biosyn- thetic pathway leading to the production of chlorophyll pigment. The only simp'e ex- planation for this effect is that, in the pres- ence of if if, an cxtranuclear gene located in the plastid and essential for chlorophyll pro- duction is somehow induced to mutate to a form no longer able to perform this function. The results convincingly demonstrate that mutation of an cxtranuclear gene ecu be in- duced by a nuclear gene. A similar case, in which a nuclear gene controls chlorophyll production by mutating plastid genes, is known in the catnip. Nepeta. 374 ( II Vl'TF.R 29 FIGURE 29-4. Simplified representation of mi- cronuclear events occurring during conjugation in Paramecium. Each conjugant has a single diploid micronucleus (A), which following meiosis produces four haploid nuclei (B). Three of these disintegrate (C). and the re- maining nucleus divides once mitotically ( D ) . The conjugants exchange one of the haploid mitotic products (E), after which fusion of haploid nuclei occurs (E) so that each of the conjugants, which later separate, contains a single diploid micronucleus. Paramecium 8 Kappa particles (and the similar lambda or mate-killer particles) are located in the cyto- plasm of certain strains of the protozoan Paramecium. Hundreds of kappa particles can easily be seen in a single cell (Figure 29-3). They contain DNA (and very prob- ably RNA) and are self-reproducing. Indi- viduals containing kappa are called killers, since animal-free fluid obtained from cultures of killer paramecia will kill sensitive (kappa- tree ) individuals. '■ The following discussion of Paramecium is based mainly upon the work of T. M. Sonneborn and co-workers. Mutant kappa particles are known to pro- duce modified poisons. Kappa is liberated into the medium once it develops a highly retractile granule which sometimes appears as a "bright spot"" under the microscope. One "bright spot" kappa particle is enough to kill a sensitive individual. Kappa has a specific relationship to its host, in that a par- ticular dominant host gene (K) must be present for kappa to maintain itself, that is. reproduce. Killer individuals homozygous for the recessive host allele (A ) cannot main- tain kappa, and after 8 to 15 divisions, kappa particles are lost and sensitive individuals result. Because it is infective and not typi- cally found in paramecia, kappa seems to be a foreign organism of some kind. Since lambda can be grown in vitro, lambda and kappa are considered to be bacterial endo- symbiotes of Paramecium.7 Although the cytoplasmic bacterial endo- symbiote kappa can be transmitted from one generation of Paramecium to the next, its distribution to the next generation depends upon the mechanism by which the new gen- eration is initiated. Two such mechanisms — asexual and sexual — are described briefly with special reference to kappa-transmission. A typical Paramecium contains a diploid micronucleus and a highly polyploid (about 1000N) macronucleus (or megamaleus) . When the parent divides asexually by fission, two daughter paramecia are produced. Both micronucleus and macronucleus replicate and separate; when fission is completed, both daughter cells are chromosomally identical to each other and to their parent cell. Although the cytoplasmic contents are not equally ap- portioned to the daughters, a killer parent will normally produce two killer daughters since each receives some of the hundreds of kappa particles present in the parental cyto- plasm. Successive fissions by the killer daughters will produce a clone of chromo- 7 See W. J. van Wagtendonk, J. O. D. Clark, and G. A. Godoy (1963). Extranuclear Genes 375 somally identical killer individuals. Simi- larly, successive fissions of a sensitive Para- mecium will produce a clone of sensitive individuals. A new generation can also be formed sex- ually. All members of a clone are of the same mating type. When different mating- type clones are mixed, a mating reaction occurs which involves individuals of different mating types sticking together to form larger and larger clumps of paramecia. After this clumping, pairs — each member a different mating type — undergo conjugation. During conjugation (Figure 29-4) the micronucleus of each mate undergoes meiosis to produce four haploid products, three of which subse- quently disintegrate. The remaining nucleus divides mitotically to produce two haploid nuclei. Next, one of the two haploid nuclei in each conjugant migrates into the other conjugant where it joins the nonmotile hap- loid nucleus to form a single diploid nucleus in each conjugant. The macronucleus dis- integrates during conjugation. After conjugation the two paramecia sepa- rate and produce the exconjugants of the next generation. Since each conjugant con- tributes an identical haploid nucleus to each fertilization micronucleus, both exconjugants are chromosomally identical — as can be proved by employing various marker genes. (When the conjugants are homozygous for different alleles, the exconjugants are identi- cal heterozygotes. ) The diploid micronucleus in each exconjugant divides once mitotically; one product forms a new macronucleus, while the other remains as the micronucleus. Since all conjugants happen to be resistant to killer action, we can study the consequence upon kappa-transmission of mating a killer with a sensitive individual. The cytoplasmic interiors of conjugants are normally kept apart by a boundary probably penetrated only by the migrant haploid nuclei so that little or no cytoplasm is exchanged. Conse- quently, the exconjugants have the same kappa-condition as the conjugants; namely, one is a killer and one is a sensitive indi- vidual. Under special conditions, however, a wide bridge forms between the conjugants allowing the cytoplasmic contents of both mates to flow and mix (Figure 29-5). The extent of the cytoplasmic mixing can be con- trolled experimentally. When cytoplasmic mixing between killer and sensitive conju- gants is sufficiently extensive, kappa particles flow into the sensitive conjugant and both exconjugants are killers. Consider how specific nuclear genes are distributed in conjugation. If each conjugant is a micronuclear heterozygote, Aa, which one of the four haploid nuclei produced by meiosis — A , A , a, or a — will survive depends on chance. Accordingly, whether the cyto- plasms of the conjugants mix or not, both exconjugants will be A A 25% of the time, Aa 50% of the time, and aa 25% of the time. Note again that both exconjugants are identical with respect to micronuclear genes, and that both will give rise to clones phenotypically identical with respect to the micronuclear gene-determined trait under consideration. When dealing with a trait determined by a cytoplasmic particle like kappa, however, the result can be different. In this particular example, the cross of a figure 29-5. Silhouettes of conjugating Para- mecium. A. Normal, no cytoplasmic mixing. B. Wide bridge, permitting cytoplasmic mix- ing. 376 CHAPTER 29 sensitive individual with a killer produces exconjugants whose type depends upon the occurrence or nonoccurrence of cytoplasmic mixing. Kappa has special significance because ii shows how a symbiotic microorganism can become so well adapted to its host, that it becomes part o\' the host's genetic system and determines some of the host's traits. 1 ike kappa, the rickettsial organism causing Rocky Mountain spotted fever is visible and transmitted through the cytoplasm of carrier cells. These rickettsiae, as well as sigma and the spirochaete already discussed, also determine certain traits of their hosts. Even though each of these organisms seems to be foreign to its host, we cannot be sure whether it was originally a parasite or symbiont. Could some of the now-foreign organisms located intracellularly have been originally part of the normal gene content of a cell? This question is particularly pertinent when viruses are considered. From what has been discussed in previous chapters, it is clear that all viruses cannot be classified as either being or arising from foreign infective agents. Present-day virulent phages seem to be acting as foreign organisms when they lyse their bacterial hosts. But, the lytic ca- pacity of a phage depends upon both its genotype and its host's, and, under some genotypic conditions, lysis is quite rare. Determining the normality or abnormality of present-day viruses is even more difficult when temperate phages are considered; not only are they less lytic (yet capable of trans- duction), but the very genes characterizing their prophages seem to be associated with part of a normal bacterial chromosome. As more is learned about viruses, and phage in particular, our understanding of what is ge- netically "normal." and what is "foreign," will undoubtedly undergo drastic revision. v As knowledge of the genetics of viruses and • Sec A. Campbell (1961). their "hosts" increases, we will be in a better position to postulate how they originated. Chlamydomonas Chlamydomonas reinhardi is a unicellular plant with two flagella and a single chloro- plast. By means of mitotic cell division, it can reproduce ascxually to produce clones. No sexual reproduction is observed between members of a clone, but if members of two different clones are mixed together, individ- uals from different clones may pair, fuse, and produce zygotes. After two divisions, the zygote produces four cells, each of which can be isolated to give rise to separate clones. When a sample from each of the four clones is mixed with different portions of a fifth clone, two of the four show mating (and are called sexual type + ) and two do not (being, therefore, of — sex). Moreover, when por- tions of the four sibling cultures under test are combined in pairs, we find that individu- als of any -f- culture can mate with individ- uals in any — culture. Combinations of two + or two — cultures, however, show no mating. No morphological difference between -f and — individuals has been detected. Among the first four cells produced from a zygote, two are + and two are — . This outcome suggests that the zygote is diploid; that it carries a pair of genes for sex (which we can call mt+ mt~); and that meiosis oc- curs in the next two divisions. As a conse- quence of this meiosis, a 1 : 1 ratio of mt+ : ml ~ is found among the haploid products. Mating type, therefore, behaves like a trait based upon two different alleles of a single nuclear gene. The wild-type Chlamydomonas is sensi- tive to streptomycin. After exposure to streptomycin, a number of preadaptive mu- tants are found which make the individual resistant to 100 p.g streptomycin per vial, al- though they are still sensitive to 300 /*g per vial. Since crosses with a streptomycin- Extranuclear Genes 377 figure 29-6. Electron micrographs of {A) mouse heart mitochondria (X60,000), (B) Neurospora mitochondria prepared to show cristae with elementary particles attached (X67,200), and (C) the outline of Neuro- spora mitochondrial DNA (X ,29,300). (Cour- tesy of Dr. Walther Stoeckenius, The Rocke- feller Institute, New York.) sensitive strain segregate two sensitive to two resistant — just as the alleles for mating type segregate — a chromosomal basis for the mutation is indicated. These chromosomal mutants are of the sr-100 class. Strepto- mycin also acts not as a selective agent but as a mutagen to produce individuals — called sr-500 mutants — resistant to 500 jxg strepto- mycin per vial. Sr-500 mutants do not show meiotic segregation in the haploid F, or the backcross progeny. Moreover, in crosses with ss (streptomycin-sensitive) individuals, the sr-500 factor is usually received by all progeny when the sr-500 parent is of mating type + (m/+), and by none of the progeny when the sr-500 parent is — (mt~). These and other characteristics demonstrate that sr-500 is caused by a "nonchromosomal" ,. • as ;.■ ■^<:k-.--..- gene; that is, a gene which does not comply with the usual transmission rules for chro- mosomal genes in sexually reproducing in- dividuals. Streptomycin also induces a large number of auxotrophic mutants whose basis also proves to be in nonchromosomal genes. Exceptional zygotes occur which contain the nonchromosomal genes from both par- ents. Using such zygotes, the results of single factor crosses and of reciprocal two- factor crosses show that the nonchromo- 378 CHAPTER 29 soma! genes recombine in postmeiotic di\ isions. These studios ' suggest the existence of an extensive extrachromosomal gene system in Chlamydomonas. Some of these genes may be located in the DNA of the chloroplast; this DNA is reported to have a base ratio dis- tinctly different from Chlamydomonas nu- clear DNA. Mitochondria Mitochondria (Fig. 29-6) are organelles consisting of a smooth continuous outer membrane and an inner membrane which folds inward to form double layers or cristae.10 The outer membrane probably controls permeability; the inner membrane, its cristae, and elementary particles contain most of the insoluble respiratory enzymes whose function provides the main source of energy for the cell. Mitochondria which ap- pear to be dividing transversely have been observed; it is very likely that most, if not all, mitochondria arise from the division of pre-existing mitochondria." DNA is a nor- mal component of mitochondria.12 This DNA (Figure 29-6) has a unique buoyant density and is probably double-stranded.' ; Certain strains of yeast produce tiny colo- nies on agar. When such organisms are crossed with normal-sized individuals, we obtain a two normal to two tiny ratio in the meiotic products after segregation. Such tiny strains due to mutant nuclear genes are called segregational petites. When normal yeast cells are treated with an acridine dye (eutlavin). numerous petite colonies arise.11 When these strains, called vegetative petites, are crossed with normal yeast, the petite ,JSee R. Sager (1965). 10 See D. F. Parsons (1963). and D. E. Green (1964). » See D. J. L. Luck (1963). J- See M. Chevremont (1963). and G. Schatz, E. Haslbrunner, and H. Tuppy (1964). ,:!See D. J. L. Luck and E. Reich (1964). 14 See B. Ephrussi (1953). phenotype does not segregate regularly. The ease with which vegetative petites are pro- duced by acridine dyes and their subsequent failure to segregate properly provide good evidence that they are caused by extrachro- mosomal mutants. The characteristic slow growth of petites is attributable to the ab- sence of respiratory enzymes known to reside in the mitochondria. Although no change in mitochondrial morphology has been detected in petites. it is clear that the presence of certain mitochondrial enzymes is controlled by chromosomal as well as by extrachromosomal genes. It has not yet been proved, however, that these extrachro- mosomal genes are located in the mitochon- drial DNA. In Neurospora a slow-growing strain, poky, when crossed with a wild strain, fails to show segregation and is unlinked to any chromosome.17. The poky phenotype is ap- parently due to a mutant of an extrachromo- somal gene. In poky individuals certain en- zymes normally present in the mitochondria are altered and so is mitochondrial morphol- ogy. Fusion of hyphae from wild type and poky strains produces a heterocytosome — a mixture of the two kinds of cytoplasm. Such fused hyphae are wild-type at first but later become poky, with the nuclear geno- type having no effect upon the outcome. Is this outcome due to selection favoring an apparently detrimental extrachromosomal gene? We do not know. As in the case of petites, the location of the extrachromosomal genes involved in poky is unknown. Centrosomes, Kinetosomes, and Kinetoplasts The centrosome is an organelle often found at each pole of a spindle, particularly in ani- mal cells. A granular structure — the cen- triole — is sometimes seen within it; similar granules can sometimes be seen within the centromere (Figure 29-7). The granules '■"•See M. B. Mitchell and H. K. Mitchell (1952). Extranuclear Genes 379 1 / A B figure 29-7. The centromere and its gran- ules in corn. (Courtesy of A. Lima de Faria, "Compound Structure of the Kinetochore in Maize," J. Hered., 49:299, 1958.) in the centromere and centrosome stain the same (both seem to contain DNA); so does the material surrounding these granules. In the living cell centromere and centrosome have a similar appearance. The granules within the centromere are apparently thick- enings of the DNA thread which passes from one chromosome arm to the other. Centromeres are sometimes attracted to each other and to the centrosomes, and at anaphase the centromeres migrate toward the centrosomes. The centrosome, too, has the capacity to move, as demonstrated by its preanaphase movement and its behavior after a sperm has penetrated an egg. Thus we see that the morphological similarity of centromere and centrosome is paralleled by their behavioral similarities. From these facts some sort of kinship is suggested 1,; between centrosome and centro- mere. This view is strikingly supported by the finding ,7 that during the meiotic divi- sions leading to the formation of certain mollusc sperms, certain chromosomes de- generate and release "naked" centromeres. These now-free centromeres group together with the centrosome and thereafter duplicate centrosomal behavior and appearance ex- actly. In effect, then, the freed centromeres become extra centrosomes. The preceding circumstances suggest that the centromere and centrosome may be two states of a presently — or previously — existing episome. The change from one episomal state to an- other is probably influenced by the presence of a nuclear membrane and by various other genetic factors present in highly organized cells, as well as by the occurrence of muta- tions in the autonomous or integrated state of the episome. At the base of each cilium or flagellum is a granular body, the kinetosome, which is re- sponsible for the motion of the organelle. Considerable evidence has been brought forth that kinetosomes contain DNA and are homologous to centrioles (or centrosomes). Perhaps kinetosomes, too, are episomes or episomal derivatives. It has been suggested that the episome F functions like a centro- mere. Could it be that the movements at- tributed to F and the centromere and centro- some have the same basis as the flagellar and ciliary movement produced by kinetosomes? In Trypanosoma, DNA and a histone-like protein are found in the kinetoplast, a large cytoplasmic organelle associated with mo- tility as well as mitochondria. DNA repli- cation occurs synchronously in nucleus and kinetoplast. Kinetoplasts can be damaged irreversibly if treated with acridine dyes. Since the kinetoplast apparently has an ap- preciable DNA content, additional molecular information about it should prove quite valu- able. 16 Originally by C. D. Darlington. F. Schrader. and others. 17 See A. W. Pollister and P. F. Pollister (1943). 380 CHAP I IK 29 Tracy M. Sonnhborn. about I960. SUMMARY AND CONCLUSIONS Nucleated cells may contain an extensive system of extranuclear genes. In some cases the extranuclear genes seem to be foreign organisms (viruses, kappa, and probably sigma ) ; in other cases they appear to be associated with normal constituents of the cell (plastids, mitochondria, centrosomes, kinetosomes, and kinetoplasts). The char- acteristics of centrosomes, kinetosomes, and centromeres suggest a past or present epi- somal interrelationship. The DNA in plastids and mitochondria has yet to be directly connected with the activity of specific genes. Nuclear and extranuclear genes are already known to be interrelated in two ways: the former can mutate the latter, and both may interact in the production of a par- ticular phenotype. REFERENCES Barigozzi, C, "Relationship Between Cytoplasm and Chromosome in the Transmission of Melanotic Tumours in Drosophila," in Biological Organization, pp. 73-89, New York: Academic Press, 1963. Beale, G. H., The Genetics of Paramecium Amelia, Cambridge: Cambridge Univer- sity Press, 1954. Campbell, A., "Conditions for the Existence of Bacteriophage," Evolution, 15:153- 165, 1961. Chevremont, M., "Cytoplasmic Deoxyribonucleic Acids: Their Mitochondrial Locali- zation and Synthesis in Somatic Cells Under Experimental Conditions and Dur- ing the Normal Cell Cycle in Relation to the Preparation for Mitosis," Sympos. Int. Soc. for Cell Biol., 2:323-331, 1963. Extranuclear Genes 381 Edelman, M., Cowan, C. A., Epstein, H. T., and Schiff, J. A., "Studies of Chloroplast Development in Euglena, VIII. Chloroplast-Associated DNA," Proc. Nat. Acad. Sci., U.S., 52:1214-1219, 1964. Ephrussi, B., Nucleo-Cytoplasmic Relations in Micro-Organisms, Oxford: Clarendon Press, 1953. Green, D. E., "The Mitochondrion," Scient. Amer., 210 (Jan.) :63-74, 152, 1964. L'Heritier, P., "The Hereditary Virus of Drosophila," Adv. Virus Res., 5:195-245, 1958. Jinks, J. L., Extrachromosomal Inheritance, Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1964. Luck, D. J. L., "Genesis of Mitochondria in Neurospora crassa," Proc. Nat. Acad. Sci., U.S., 49:233-240, 1963. Luck, D. J. L., and Reich, E., "DNA in Mitochondria of Neurospora crassa," Proc. Nat. Acad. Sci.. U.S., 52:931-938, 1964. Mitchell, M. B., and Mitchell, H. K., "A Case of 'Maternal' Inheritance in Neurospora crassa," Proc. Nat. Acad. Sci., U.S., 38:442-449, 1952. Parsons, D. F., "Mitochondrial Structure: Two Types of Subunits on Negatively Stained Mitochondrial Membranes," Science, 140:985-987, 1963. Plus, N., "Action de la 5-Fluoro-Desoxyuridine sur la Multiplication du Virus a de la Drosophile," Biochim. Biophys. Acta, 72:92-105, 1963. Pollister, A. W., and Pollister, P. F., "The Relation Between Centriole and Centro- mere in Atypical Spermatogenesis of Viviparid Snails," Ann. N.Y. Acad. Sci., 45:1-48, 1943. Rhoades, M. M., "Plastid Mutations," Cold Spring Harb. Sympos. Quant. Biol., 1 1 :202- 207, 1946. Rhoades, M. M., "Interaction of Genie and Non-Genie Hereditary Units and the Phys- iology of Non-Genie Inheritance," in Encyclopedia of Plant Physiology, Ruhland, W. (Ed.), Vol. 1, pp. 19-57, Berlin: Springer Verlag, 1955. Sager, R., "Genes Outside the Chromosome," Scient. Amer., 212 (No. l):70-79, 134, 1965. Sager, R., and Ishida, M. R., "Chloroplast DNA in Chlamydomonas," Proc. Nat. Acad. Sci., U.S., 50:725-730, 1963. Schatz, G., Haslbrunner, E., and Tuppy, H.. "Deoxyribonucleic Acid Associated with Yeast Mitochondria," Biophys. Biochem. Res. Commun.. 15:127-132, 1964. Seecof, R. L., "CO- Sensitivity in Drosophila as a Latent Virus Infection." Cold Spring Harb. Sympos? Quant. Biol., 26:501-512. 1962. Sonneborn. T. M., "The Role of the Genes in Cytoplasmic Inheritance," Chap. 14, pp. 291-314, in Genetics in the 20th Century, Dunn, L. C. (Ed.), 1951. Sonneborn, T. M., "Kappa and Related Particles in Paramecium," Adv. Virus Res., 6:229-356, 1959. Sonneborn, T. M., "The Gene and Cell Differentiation," Proc. Nat. Acad. Sci., U.S., 46:149-165, 1960. van Wagtendonk, W. J., Clark, J. A. D., and Godoy, G. A., "The Biological Status of Lambda and Related Particles in Paramecium aurelia," Proc. Nat. Acad. Sci., U.S., 50:835-838, 1963. 382 CHAPTER 29 QUESTIONS FOR DISCUSSION 29.1. What is revealed aboul nucleocytoplasmic interrelationships from the study of F? Of temperate phages? 2l>.2. What evidence can you present thai ((). -sensitivity is due to a virus rather than a normal chromosomal gene? 21>..V In proving the existence o\ extranuclear genes, which operations (recomhina- tional. mutational, functional, chemical) were utilized? Did our decision in- clude their capacity for self-replication? Why? 29.4. Discuss the genetic control of chlorophyll production in corn. 29. s. Do you think that the study of nucleocytoplasmic interrelations in Paramecium has any bearing upon differentiation processes in multicellular organisms? Explain. 29.6. As an experimental organism for genetic investigation, what are the unique ad- vantages of Paramecium? 29.7. Certain paramecia are thin because of a completely recessive nuclear gene, th. What is the phenotypic expectation for the clones derived from exconjugants of a single mating of ++ by +th? How would cytoplasmic mixing affect your expectation? Why? 29.8. According to the definition of a chromosome given on page 8, would you consider kappa to be or to contain a chromosome? Explain. 29.9. Keeping in mind the difficulties of proving the existence of extranuclear genes, which do you think represents the primary genetic material in cellular organ- isms, nuclear or extranuclear genetic material? Explain. 29.10. Discuss the statement (p. 253) that DNA "seems to be absent in the cytoplasm." 29.1 1. Do you think the evidence presented that sex in Chlamydomonas is based pri- marily upon a single pair of genes is conclusive? Justify your answer. 29.12. Report any evidence obtained since this account was written (November, 1964) that the DNA in chloroplasts or in mitochondria has a phenotypic effect. Chapter 30 THE GENETIC CONTROL OF MUTATION C hapters 11, 12, and 14 dealt with different units of muta- tion, ranging from the larg- est— genomic — changes to the smallest — gene — changes. Although various external- ly-applied environmental agents can produce mutations of all kinds (Chapters 13 and 16), we would like to know to what extent the genotype regulates its own mutability. Mitosis is so precise that ordinarily the genotype prevents the occurrence of genomic and single, whole-chromosome changes in successive generations of cells. Mitotic rate and spindle orientation are controlled by genes, just as are other aspects of mitosis. In Ascaris, the genotype seems to prevent mutation in a polycentric chromosome by suppressing the action of all but one centro- mere. In meiosis, crossing over occurs at points that correspond precisely in two nonsister strands, so that crossovers containing de- ficient or duplicated segments are avoided.1 In this way intrachromosomal euploidy is maintained even though recombination be- tween homologs is permitted. Synapsis and chiasma formation in meiosis help distribute the homologs in a way that prevents the gain or loss of whole chromosomes, that is, aneu- somy. Evidence for the genetic control of synapsis is provided by collochores in Dro- sophila (p. 185) and by the genes for asyn- apsis found in maize (p. 190), many other 1 See, however, G. E. Magni (1963). 383 plant species, and Drosophila. Genes pro- ducing spindles that diverge at the poles during meiosis are known in both Drosophila and maize. In general, the synthesis of new genes is usually regulated to prevent sub- stitution of improper genetic raw materials for the proper ones, assuming both types are present in the cell at the same time. It might be argued that the examples given demonstrate only that reduced mutability is an inevitable consequence of normal cell operation. Although present genotypes may appear to play a passive role, mitosis and meiosis are not intrinsic properties of genes or cells, and, therefore, during the course of evolution the selection of genes for carry- ing out these activities must have been an active process aimed at reducing mutability; that is, genes that could maintain genetic- stability and permit replication and genetic recombination via sexuality must have been favored. Though the genetic controls so far men- tioned lead to a reduction in mutability, it should be realized that the genotype also per- mits genetic changes to occur in the follow- ing controlled or regulated ways: 1 . The ploidy changes in a sexual cycle — from diploid to haploid and back again — are under genetic control. 2. Mutational changes increase with mito- tic activity (p. 193 ) ; since the rate of mitosis is under genetic control (many cancer cells are mutants whose mitotic rate has in- creased), the genotype controls mutability in this way. 3. We have already mentioned (Chapter 11) certain modifications of meiosis — un- doubtedly under genetic control also — lead- ing to ploidy changes in the next generation. 4. Even within the somatic tissues of a multicellular organism, controlled genetic change is permitted in cells whose chromo- somes become polyploid (as in human liver), or highly polynemic (as in the Dipteran larval salivary gland). ;st ( II U'TER 30 5. We have also noted that in Ascaris (p. 189) changes occurring in somatic tissues lead to the formation of a number of small chromosomes from a single large one. 6. The frequency of nondisjunction lead- ing to aneusomy has been shown to depend both on the amount and distribution of heterochromatin and on the types of chro- mosomal rearrangements present. There- fore, to the extent that the genotype regulates its heterochromatin and rearrangements, it is also regulating the incidence of nondis- junction. 7. Similarly, the arrangement of meiotic products in Drosophila oogensis (Chapter 1 2 ) acts to eliminate dicentrics produced by crossing over in paracentric-inversion hetero- zygotes. 8. Finally, the arrangement of the chro- mosomal material and the metabolic activity of the cell (as it influences the amount of water and oxygen present, for example) are other ways in which mutability is influenced or regulated by the genotype itself. The preceding discussion dealt mainly with the prevention or regulated occurrence of intergenic changes. Does the genotype regulate the occurrence of point mutation? Consider the spontaneous point-mutation frequencies for two lines of the same species of Drosophila — one living in a tropical and the other in a temperate climate. If the genotype were at the mercy of temperature in the wild, we would expect the tropical form to have a higher frequency of spontane- ous point mutation than the temperate form. However, when both lines are grown at the same temperature in the laboratory, the trop- ical form has a lower mutation rate than the temperate one. This result provides good evidence that the tropical form has genetically suppressed (or the temperate form has genetically enhanced) its mutational re- sponse to temperature. Consequently, in na- ture the two forms probably show less differ- ence in mutation frequency than would be expected with the differences in temperature. Other strains of Drosophila melanogaster col- lected from various regions have ditferent spontaneous point-mutation frequencies. Some o\' this may be due to differences in the mutability of their isoalleles (p. 59); part may be due also to a general control of mutability by the genotype, for some strains contain mutator genes which can increase the general point-mutation frequency as much as tenfold. Of course, other alleles of mutator genes can be considered general suppresssors of point mutability. Certain organisms (bacteria, for example) have mutants which make the individual generally less mutable to a given mutagen. Since the organisms most advanced in evolution contain more genetic material per cell than less advanced forms, the most advanced forms probably have se- lected genotypes which reduce their sponta- neous mutation rate to avoid overmutation. Activator and Dissociation in Maize 2 The triploid endosperm (p. 26) in kernels from some corn plants are white, others are colored, and still others are white with col- ored speckles. At first, it might seem as if we were dealing with a high mutation fre- quency of a gene from a "colorless" to a "colored'' allele. It is found, however, that the white phenotype results from the pres- ence of two genes adjacent to or very near each other on the same chromosome. If these two loci are separated or dissociated from each other by chromosomal breakage that removes a particular one of the two loci, the mutant cell and all its daughter cells with the remaining locus will be colored. The locus removed, called Dissociation, Ds, causes breakage in chromosome regions near it and is probably in a heterochromatic por- tion of the chromosome. If Ds is never dis- sociated from the adjacent locus, the kernel Based upon work of B. McClintock. The Genetic Control of Mutation 385 is white; if it dissociates during kernel for- mation, the kernel shows colored sectors or dots on a white background; if Ds is moved before the kernel forms, the kernel and later generations of plants arc completely colored. In some plants the colored specks are large, due to the movement of Ds early in develop- ment; in other plants they are small due to the movement of Ds later in development, when very few additional cell divisions take place. The mutation involved here is the loss or re- moval of Ds via breakage. The change in color apparently is not a mutational but a phenotypic event which enables the detection and proof of the mutational event. This phenotypic effect is dependent upon the rela- tive position of genes; that is, it is a position effect; the presence of Ds next to the gene for color suppresses color formation; its ab- sence permits the gene for color to produce color. Although the breaks which Ds causes are usually near Ds in the chromosome, they are not always at the same locus. For this rea- son and also because breaks can occur simultaneously in other chromosomes (due to spontaneous events or to the presence of other Ds genes located in them), Ds need not be lost after breakage but may transfer from one chromosomal position to another in the same or a different chromosome. As the result of the movement of Ds to new positions, the number of Ds factors present in the endosperm can increase in successive generations. When the number of Ds genes in a given region of a chromosome increases, the region breaks more and more frequently. Ds transposed to another chromosome can cause breaks near its new location. Thus, whenever Ds moves, a mutation has occurred. Such relocations of Ds often sup- press the phenotypic effect of a gene located near the new locus of Ds. As long as Ds remains in its new position, the new pheno- type is produced, thereby simulating a stable point mutation of the gene near Ds. More- over, each time Ds is lost from such a loca- tion, the new phenotype of the adjacent gene reverts to the old phenotype. If these trans- positions are frequent, they may be incor- rectly scored as point mutations of an un- stable, mutable allele of the neighboring gene. If Ds rarely moves, one might incorrectly score the new phenotype produced by its neighboring gene as a rare mutational change in the neighboring gene. Although it is still not definite how many events scored as point mutations of a given gene are position effects due to suppression or release resulting from a change in linear gene neighbors, not all point changes can be such position effects, of course, since differences among the genes involved in position effects must first arise by mutation. The ability of Ds to cause breaks in chro- mosomes is controlled by Activator, Ac, genes. Ac does not have to be on the same chromosome as Ds and usually is not; it also seems to be located in heterochromatin. By appropriate crosses kernels can be obtained whose endosperm contains none, one, two, or three Ac genes in addition to one Ds gene located near a pigment-producing gene (Fig- ure 30-1 ). In the absence of Ac, no specks are produced, and the kernel is completely white. This observation proves that Ds can- not cause chromosome breakage (and is not relocated in other ways) in the absence of Ac. Moreover, as the dosage of Ac increases from one to three, the colored spots become smaller and smaller. Thus, Ac also acts to delay the time of Ds action. Here, then, is a case in which the genotype regulates its own mutability — Ac not only determines the ability of Ds to produce breakages but regu- lates the time in development when breakage is to occur. Ac is clearly acting as a regu- lator gene. This kind of gene may be im- portant in cyclical metabolic processes as well as in cellular differentiation and em- bryonic development. Factors like Ac are fairly common in maize, and the phenotypic :fcN(i CHAPTER 30 instability of various loci in other flower- ing plants, terns, fungi, and bacteria may be due to similar factors. * Segregation Distortion in Drosophila Drosophila melanogaster homozygous for the II chromosome mutants, cinnabar (en) and brown (bw). have white eyes because en en and bw bw prevent the formation of the brown and the red pigments, respectively, which together comprise the dull-red eye color of the wild type. When the test cross 4- + en bw ' by en bw en bw 9 is made, the progeny typically occur in the approxi- mate phenotypic ratio of one white to one + . If, however, the unmarked II chromo- figure 30-1. The effect of Activator on the action of Dissociation. (A) No Ac is present. The kernel is colorless due to the continued presence of Ds, which inhibits the action of a neighboring pigment-producing gene. (B) One Ac factor is present. Breaks at Ds occur early in kernel devel- opment, leading to large colored sectors. (C) Two Ac factors arc present. Time of Ds action is delayed, producing smaller sectors which appear as specks. (D) Three Ac factors are present. Ds action is so delayed that relatively few and tiny specks are produced. (Courtesy of B. McClintock.) The Genetic Control of Mutation 387 some comes from certain natural popula- tions, this cross ■ produces 93 to 99% (in- stead of about 50% ) + progeny. More- over, this atypical ratio is not associated with any increase in egg mortality. It is con- cluded, therefore, that the two kinds of male gametes (H — h and cnbw) must be func- tionally unequal in number at the time of fertilization, suggesting that the segregation ratio 1 -\ — \- : 1 en bw is somehow distorted prior to gamete formation. Analysis of the segregation distortion phenomenon reveals a genetic factor, Segregation-Distorter , SD — present in the otherwise wild-type II chro- mosome— located in the heterochromatic re- gion of the right arm near the centromere of II. The en few-containing chromosome, therefore, carries SD+. SD causes some kind of genetic change, probably at or near SD+ in the homologous chromosome, which results either in the loss of the en few-con- taining homolog or the inability of sperm carrying that chromosome to be used in fertilization. The net result is that the SD- containing chromosome is recovered in ex- cess in the F,. When, however, the SD + -containing chro- mosome carries certain inversions involving IIR absent in the homolog, the SD/SD + male shows no segregation distortion. Con- sequently, for SD to prevent the appearance of SD+ in the progeny, SD and SD+ must "pair." SD+ alleles vary in their sensitivity to any particular SD; SD alleles differ in their ability to affect a given SD+ region. The original SD SD+ combination gives a constant amount of distortion, indicating that the SD line is stable. Every SD-bear- ing chromosome recombinant for the (prob- ably heterochromatic) tip of the right arm of the II chromosome becomes less stable. The decrease in stability is reflected by varia- tions in ability to distort. Consequently, the 3 As noted originally by Y. Hiraizumi and J. F. Crow. stable line must have a modifying gene, Stabilizer of SD, St(SD), at the tip of the right arm of II. Stabilization occurs whether St(SD) is in cis or trans position relative to SD. Since the markers purple (pr) and en closely span both the centromere and the SD locus, one can study recombinants for the regions near SD. The results show that a locus is present in the right arm of II — near SD but farther from the centromere — whose presence is essential for SD operation. This locus is Activator of SD, Ac(SD), which must be in cis position for SD to function. Since it is found that crossing over in the SD-Ac(SD) region is reduced, it is hypothe- sized that a small rearrangement exists in this region. Usually, SD-Ac(SD) causes a ge- netic change in the corresponding — presum- ably nonrearranged — SD+ segment of the homolog. An individual can be synthesized, however, with one II chromosome, contain- ing almost all the hypothesized rearrange- ment without SD, whose homolog carries SD, Ac(SD) and a segment of the nonrear- ranged region. Under these conditions, seg- regation distortion occurs against the SD- containing chromosome instead of against its homolog. Although the F, of the usual heterozygous SD male occur in a distorted ratio (the father distorts, or shows segregation distor- tion via his progeny), the Fj from hetero- zygous SD females do not. An SD SD + male can distort when outcrossed to an at- tached-X female. Surprisingly, when his 5X>-containing sons are tested (these re- ceived the father's X), they do not distort. It would appear that a distorting male condi- tions his X chromosome so that sons re- ceiving it cannot distort. When a distorting male is mated to an unrelated SD+ SD + female having separate X's, all the SD-con- taining sons can distort since each received an unchanged maternal X. (Note that the daughters carry one unchanged maternal and :;ss CHAITIR 30 one changed paternal X.) Among the SD- containing sons the daughters produce, the half receiving the unchanged maternal X can distort, whereas those receiving the changed paternal X cannot. When either of these kinds oi males are outcrossed to SP S!) females, all SD-containing sons receive an unchanged maternal X and. therefore, can distort. Females producing .VD-carrying sons of which only half distort, arc said to he conditioned and to show conditional dis- tortion: the mechanism of conditional distor- tion is unknown. The case of SD shows some similarity to the Ac-Ds example discussed.4 SD causes some kind of genetic change regulated by Ac(SD). The activity of SD is modified by St(SD) and is also conditioned by the X chromosome. SD provides an excellent example of the genetic control of muta- bility. SD was initially obtained from a natural population that showed no distortion because SD's detrimental effect on the transmission of its homolog was suppresed by a combina- tion of factors. One was X-chromosome conditioning fostered by inbreeding. In this population selection apparently favored the retention of SD+ alleles resistant to distor- tion as well as inversions or other structural changes involving IIR which, in heterozy- gous condition, prevented pairing and, hence, distortion. SD is an example of meiotic drive, a force capable of altering gene fre- quencies in natural populations by the pro- duction of functional gametes which do not carry segregants in a one-to-one ratio. Episomes and Viruses as Mutagens Suppressed or variegated phenotypic effects are known which are due to the placement of heterochromatin near euchromatin. In Drosophila. such position effects are frequent after structural changes in chromosomes. 4 See L. Sandler and Y. Hiraizumi (1961). Some cases of phenotypic suppression in- volve special genetic elements — for example, Segregation Distortion in Drosophila and Dissociation in corn — associated with hetero- chromatin. Some of these genetic factors are capable of causing breakage and of changing their location in the genome. Since such factors and episomes have certain character- istics in common, both should be studied and compared with regard to phenotypic sup- pression, organelle movement, and chromo- somal breakage. The spontaneous mutation frequency from auxotrophy to prototrophy is known for a large number of alleles for various markers in Salmonella. When auxotrophic bacteria are infected with transducing phage grown on the same genetic strain or on a bacterial strain carrying a deletion (deficiency) for the gene under test, the frequency of proto- trophs is significantly increased/' Genes in- duced to revert to prototrophy in this way are called seljers. Although the mechanism of reversion is not fully understood, the presence of a transducing fragment which synapses in a region near a selfer gene some- how stimulates the mutability of the selfer. Consequently, phage enhances the mutabil- ity of bacterial genes. We have already seen (p. 373) that the mutability of an extra- nuclear gene is under control of nuclear genes. In this connection we should also note the following results involving the higher ani- mals, including man: 1 . The addition of Rous sarcoma virus to normal rat cells in tissue culture produces an increased incidence of chromosome breakage over the control level. 2. After herpes simplex virus is innocu- lated into established human tissue-culture lines, the incidence of chromosomal break- age increases. "•See M. Demerec (1963), and A. L. Taylor (1963). The Genetic Control of Mutation 389 3. All patients 8 with clinical measles (rubeola) have a high incidence of chromo- some breakage in the white blood cells by the fifth day after onset of the rash. Chromo- some breaks occur in 33 to 72% of the cells examined, and all chromosomes are break- able at numerous positions, although the unions between ends produced by breakage resulting in structural rearrangements are of low frequency. 4. After infection of human cell lines in vitro with the simian virus SV.1(,, large num- bers of chromosomal mutants are detected ' including chromosome loss, chromosome breakage, and gross rearrangements like di- centrics, rings, and (probably) transloca- tions. The frequency with which these in- ,; Studied by W. W. Nichols, A. Levari, B. Hall. and G. Ostergren (1962). 7 See P. S. Moorhead and E. Saksella (1963). volve different chromosomes is apparently not random. At least seven other viral in- fections in man are associated with an in- creased incidence of various chromosomal rearrangements in white blood cells. We do not know whether these effects are due to a general metabolic effect of the pres- ence, functioning, or replication of viral nucleic acids; to a specific episomal-like fea- ture of these viruses; or to some other fac- tor or combination of factors. In any case, viruses can induce mutations in cells of higher organisms in vivo and in vitro, and it is pos- sible that normally-present extranuclear genes can also do so. Clearly then, the genetic control of mutability involves extra- nuclear genes, episomes, and ordinary chro- mosomal genes; each is hypothetically ca- pable of affecting its own mutability as well as each other's. SUMMARY AND CONCLUSIONS The spontaneous occurrence of genomic and of single, whole-chromosome mutations is suppressed by the genotypic control of the processes of mitosis and meiosis. Struc- tural rearrangements in chromosomes are suppressed by the precision of synapsis and crossing over. Such controls are possible because of the linear arrangement of genes in the chromosomes. In certain cases involving the production of polyploid and poly- nemic chromosomes and several monocentric chromosomes from a polycentric chro- mosome, genetic change is genotypically regulated. Point-mutation frequencies also are regulated genotypically, as shown by the general control of mutation response to temperature changes or to mutagenic agents; by the occurrence of mutator genes; and by genes which produce chromosome breakages that can lead to losses, shifts, and transpositions and, therefore, position effects. Regulator genes control the operation of genes that cause breakages or other mutations. Trans- ducing phages induce point mutations; viruses that attack higher animals can also effect chromosomal breakage. REFERENCES Demerec, M., "Selfer Mutants of Salmonella typhimurium" Genetics. 48:1519-1531, 1963. Magni, G. E.. "The Origin of Spontaneous Mutations During Meiosis." Proc. Nat. Acad. Sci., U.S., 50:975-980, 1963. McClintock, B., "The Origin and Behavior of Mutable Loci in Maize," Proc. Nat. Acad. Sci., U.S., 36:344-355, 1950. Reprinted in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs. N.J.: Prentice-Hall. 1959, pp. 199-209. 390 CHAPTER 30 McClintock, B., "Controlling Elements and the dene.'' Cold Spring Harb. Sympos Quant. Biol., 21 : 197-216, 1 956. Moorhead, P, S., and Saksela, E., "Non-Random Chromosomal Aberrations in SV,()- rransformed Human ( ells," J. ( ell < omp. Physiol., 62:57-83, 1963. Nichols, W. \V.. 1 evan, A.. Hall, B., and Ostergren, G., "Measles-Associated Chromo- some Breakage. Preliminary Communication," Hereditas, 48:367—370, 1962. Peterson, P. A.. "The Pale Green Mutable System in Maize," Genetics, 45:115-133, I960. Sandler, 1 ., and Hiraizumi, Y., "Meiotic Drive in Natural Populations of Drosophila Melanogaster. \ 111. A Heritable Aging Effect on the Phenomenon of Segrega- tion-Distortion," Canad. J. Genet. Cytol., 3:34-46, 1961. Li\lor, A. 1... "Bacteriophage-Induced Mutation in Escherichia coli," Proc. Nat. Acad. Sci., U.S., 50:1043-1051, 1963. QUESTIONS FOR DISCUSSION 30.1. How is the precision of the mitotic and meiotic processes related to the muta- bility of the genetic material? 30.2. Defend the statement that meiosis and mitosis are not intrinsic properties of genes. 30.3. If whole genome changes represent a class of mutation, should the changes in ploidy which occur in gametogenesis and fertilization be considered mutations? Why? 30.4. Does the activity of Dissociation provide evidence for the genetic control of mutability? Explain. 30.5. What evidence can you present that variegation is not always due to the effect of a single pair of genes, one with unstable alleles? 30.6. What characteristics of Dissociation resemble those of the episome F? 30.7. Discuss the mechanisms by which segregation distortion is suppressed in nat- ural populations of Drosophila. 30.8. Discuss the hypothesis that the SD phenomenon involves an episome-like agent. 30.9. In what respects is SD similar to and different from Ds? 30.10. Do you suppose that all viruses cause a significant increase in the frequency of chromosomal breakage? Explain. Chapter 31 THE MOLECULAR BASIS OF MUTATION T |he spontaneous mutation fre- quency is influenced by natu- rally-occurring physical muta- gens such as ultraviolet light and ionizing radiations (Chapter 13) and probably by nat- urally-occurring chemical mutagens (Chap- ter 14); mutability is also under consid- erable genetic control (Chapter 30). The correlation found betwen mutagenicity and the wavelength of ultraviolet light suggests that nucleic acids are involved in the muta- tion process (p. 261). What, then, is the detailed molecular basis of spontaneous and induced mutation? Mutagens and Antimutagens In E. coli auxotrophic for tryptophan, the spontaneous mutation rate from sensitivity to resistance to infection by 4>T5 or ^>T6 is found to depend upon whichever component of the nutrient medium is made the limiting factor for growth.1 The highest rate is ob- tained when the growth-controlling factor in the medium is tryptophan; the lowest rate is obtained when the growth-controlling factor is lactate. This result indicates that the spontaneous mutation rate depends upon the physiological or biochemical state of the organism — a view also supported by the effect of temperature changes upon the spontaneous mutation rate (p. 192) — and suggests that chemical sub- stances added to a culture of E. coli growing 1 Based upon work of A. Novick and L. Szilard (1951). 391 in a medium limited in one essential nutrient would have a pronounced mutagenic effect. By testing various substances in concen- trations that produce no appreciable killing of tryptophan-limited bacteria, many purines and purine derivatives are found to be muta- genic. The most mutagenic is caffeine; theophylline is nearly as effective; azaguanine is mutagenic, and — to a lesser degree — ade- nine. In contrast, no pyrimidines or their derivatives are mutagenic under the same conditions. If purine ribosides such as aden- osine or guanosine are added to the medium containing any one of several purine muta- gens, the mutagenic activity is completely suppressed.- Thus, for example, adenosine completely suppresses the mutagenicity of adenine or caffeine. Clearly the purine ribo- sides are acting as antimutagens — just as an- oxia or catalase are antimutagens so far as chromosomal breakage (p. 182) or point mutations (p. 192) produced by X rays are concerned — and are not acting as selective agents against induced mutants. On the other hand, pyrimidine ribosides, deoxyaden- osine, and deoxyguanosine either are not at all antimutagenic to purines and their derivatives, or they are much less efficient than the purine ribosides. Theophylline is mutagenic under aerobic but not anaerobic conditions. Adenosine, however, is present in significant concentra- tions in bacteria growing anaerobically but not detectable in bacteria growing aerobically. Under anaerobic conditions, adenosine is ap- parently a normally-present antimutagen that counteracts the effect of purines added in the medium. Note that anacrobiosis has no effect on the ultraviolet-induced mutation rate and makes gamma radiation less effec- tive only because of the reduction in oxygen; such a result is consistent with the finding that extra adenosine has no antimutagenic effect on either ultraviolet or gamma radia- tion mutagens. -See A. Novick (1956). 392 CHAPTER 31 The spontaneous mutation rate is reduced bj the presence of purine ribosides | adeno- sine or guanosine) but not pyrimidine ribo- sides (uridine or cytidine) in the medium. With adenosine in the medium, the sponta- neous rate is reduced to about one third its original value. Moreover, the spontaneous rate is lower under anaerobic than aerobic conditions, as would be expected from the metabolic production or lack of utilization of adenosine. Finally, it should be noted that all the purine mutagens increase the mutation rate to T5 resistance more than to T6 resistance, although the reverse is true for ultraviolet and gamma radiation. From these results it seems reasonable to distinguish two kinds of mutagens — a purine type and a radiation type — which produce two different kinds of mutations. We can postulate that under the experimental condi- tions described, about two thirds of the spon- taneous mutation rate is produced by the action of some purine type of substance pro- duced spontaneously during the normal me- tabolism of the cell. It may seem surprising that only purines, their analogs, and purine ribosides affect the mutation rate. This find- ing, however, may be related to the par- ticular mutants studied — namely, those re- sistant to phages T5 or T6. Such muta- tions may depend more upon changes in purines than pyrimidines; other mutations may prove to be relatively pyrimidine-sensi- tive and purine-insensitive. Even though it is not clear how the pu- rines and their ribosides accomplish their mutagenic and antimutagenic effects, two general conclusions are warranted: 1 . A considerable portion of the sponta- neous mutation rate is the normal con- sequence of the cell's biochemical activity in producing mutagens and antimutagens. 2. A connection exists between mutation rate and nucleic acid metabolism. I hough the mutation rate is directly con- nected only with purines and their ribosides, supported h\ the fact that thymine is muta- genic when withheld from bacteria requir- ing it. it is reasonable to suppose that an indirect connection also exists with DNA and its precursors. Mutational Spectra In discussing the genetics of the rll region of the 4>T4 genetic map, it was mentioned (p. 345) that the more than 1500 sponta- neously-occurring mutants tested involved changes in one or more of about 300 diller- ent sites in the rll region. This statement. of course, implies that some mutation sites must have been involved more than once. In fact, the number of times different sites are involved in mutation varies considerably. In terms of DNA, this variability must mean that certain nucleotides, singly or in groups, are much more likely to undergo sponta- neous mutation than others, so that muta- tional "hot spots" must occur. Since recombination studies permit the analysis of the rll region at the level of the nucleotide, the DNA of T4 can serve as ma- terial for studies leading to a clearer defini- tion of mutation on the molecular level. Note that even-number T phages (T2, T4, T6) have 5-hydroxymethyl cytosine (Figure 19-3 on p. 255) or a derivative of it, in- stead of cytosine in their DNA; in all other respects this DNA is typical. It has already been noted (p. 283) that 5-bromo uracil (Figure 21-4) can substitute for thymine — and only thymine — in the synthesis of DNA in vitro. What will be the mutational conse- quences of incorporating 5-bromo uracil into T4 DNA.' Addition of 5-bromo uracil to the normal culture medium of E. coli before infection with T4, does not necessarily result, alter 3 The discussion Following is based largely upon the work of S. Benzer and E. Freese (1958), and subsequent work by E. Freese and co-workers. The Molecular Basis of Mutation 393 H2N *N' NH, Proflavin figure 31—1. Two acridine dyes. H,C N H,C TM' Acridine Orange .CH N I CH3 infection, in the incorporation of this base analog in T4 DNA, since thymine can be synthesized by the bacterium and it — rather than the analog — may be used preferentially or exclusively in the synthesis of phage DNA. Sulfanilamide, itself not mutagenic, inhibits synthesis of folic acid, which in its reduced form (tetrahydrofolic acid) is required for enzymatic methyl transfer reactions. There- fore, sulfanilamide is added to the culture medium to assure that no thymine is synthe- sized from uracil. The medium is supple- mented with a variety of essential chemical substances already containing methyl and hydroxymethyl groups but not with the de- oxyribotides of thymine or of 5-hydroxy- methyl cytosine. (The deoxyribotide of 5- hydroxymethyl cytosine is omitted to prevent its possible conversion to an analog of thy- mine which might be incorporated in prefer- ence to the 5-bromo uracil.) In this way, the bacterium can function properly as a phage host. Under these conditions, 5-bromo uracil is highly mutagenic in the rll region. A com- parison of 5-bromo uracil-induced and spon- taneously-occurring rll mutants reveals that the induced mutants also occur in clusters on the genetic map, although the hot spots are in different positions. Moreover, con- trary to the spontaneous mutants, very few of those induced are of the gross (internu- cleotide) type, and almost all are subse- quently capable of reverse mutation to, or near the r+ phenotype. Although the mutational spectra (p. 192) for 5-bromo uracil, other chemical mutagens. and spontaneous mutants are all different at the nucleotide level, the exact chemical basis for the induced mutations cannot be specified with any certainty, because any given mutagen may be producing its effect via any of several different metabolic path- ways. Clearly, the molecular basis for muta- genic action is best studied using the shortest possible physical pathway between chemical mutagen and gene.4 Thus, it is preferable to treat sperm rather than oocytes with a chemical mutagen, and more desirable to expose phage or transforming DNA to the mutagen directly, rather than indirectly, via its host. Mutation Involving Whole Nucleotides Since the genetic material is a linear array of nucleotides, let us consider the possible changes to whole nucleotides at the basis of mutations. One or more nucleotides can be added, deleted, substituted for, inverted, or transposed to a new position with or without inversion. All these nucleotide rearrange- ments ought to be possible for single- stranded nucleic acid, except inversion, which requires double-stranded nucleic acid to maintain strand polarity. Whole-nucleo- tide changes can be produced by breaking the polynucleotide backbone at two or more places, followed by rearrangement of the fragments. Breakage of the backbone (and loss of DNA's ability to act as primer-tem- plate) occurs especially often after expo- sure to an ionizing physical mutagen, and leads to deletions and other rearrangements 4 As noted by I. H. Herskowitz (1955). 39 l CHAPTER 31 Uracil OH O^N- H Adenine i k.i kf, 31-2. Tautomers of uracil and adenine. of the already-formed, "old" gene material. Single, whole-nucleotide changes can also be produced by chemical mutagens without in- volving breakage. Molecules of chemical mutagens such as the acridities (Figure 31-1) can be inserted between successive nucleotides of a strand.1' A single, inter- calated molecule of chemical mutagen is apparently able to spread the strand length- wise 3.4A. When this chain is used as tem- plate, an entire nucleotide may be added to the complementary chain at the position oc- cupied by the molecule of mutagen. The possibility also exists that an unbound nu- cleotide or other naturally-occurring sub- stance may intercalate with similar results. This mutagenic mechanism involves changes both in the old and the new DNA strands. In the presence of Mn , the comple- mentary strand made by DNA polymerase from a DNA template in vitro is a mixture of deoxyribo- and ribotides,7 provided, of course, that the appropriate riboside 5' tri- 5 See H. Harrington (1964). and C. G. Mead (1964). ySee L. S. Lerman (1963 J. ' As shown by P. Berg and co-workers (see refer- ence on p. 290). phosphates are included as substrates. The incorporation of U and ribose into a comple- mentary strand is expected to be mutational. Since salts of manganese are highly muta- genic in bacteria,8 it appears that such an incorporation may occur in vivo. Sub-Nucleotide Mutations Mutation can involve the sugar, phosphate, or base portions of a nucleotide. Although deoxyribose is the only sugar detected in the DNA of microorganisms, one cannot ex- clude the possibility that an occasional ribose occurs in an otherwise typical DNA strand. Ribose can be part of a DNA chain synthe- sized in vitro if whole ribotides are incorpo- rated by the method described in the pre- ceding paragraph. Some evidence has been obtained for the incorporation of arabinose into DNA of mammalian cells in culture. As mentioned, ribosides give no evidence of acting as DNA mutagens; in fact, purine ribosides are sometimes antimutagenic. Nevertheless, we cannot rule out the possi- bility that some agents act as mutagens either by adding an O at the 2' position of deoxy- ribose in an already-formed DNA sequence H As shown by M. Demerec and co-workers. The Molecular Basis of Mutation 395 or by removing the 2' O of ribose in an RNA sequence. The phosphate part of a nucleotide can be changed by substituting P32 for P. When single-stranded DNA viruses such as $X174 and 4>S13 incorporate P''1', a single radioac- tive decay of P;- to S is sufficient to inacti- vate them. On the other hand, about ten decays are required to inactivate T2 and similar phages containing double-stranded DNA. One simple explanation of such suicide experiments is that each decay breaks the backbone of the polynucleotide in which it occurs, a single decay in one backbone sometimes leading to a nearby break in the backbone of a complementary strand if one is present. According to this explanation, scission of single- or double-stranded DNA (or RNA, presumably) is sufficient for in- activation. Base changes in old genes. Consider, next, changes involving the base portion of a nucleotide. We have already seen that base changes can result from the substitu- tion of one whole nucleotide for another. We are now interested in the possible ways the base portion of a nucleotide already in a genetic sequence may be changed chemi- cally. Certain atoms in each of the bases in DNA and RNA can assume several different ar- rangements; that is, each base can exist in several tautomeric forms. In previous dis- cussions and diagrams, the most likely tau- tomer of each base — that is, its (=0) keto or amino (NHo) form — was assumed to occur. In the tautomers of uracil and ade- nine shown in Figure 31-2, the alternatives differ in the positions at which a hydrogen atom is attached. The less likely tautomers exist in the ( — OH) enol or imino (NH) form. Although the usual amino tautomer of adenine pairs with thymine, one of its less common imino tautomers can pair with cytosine (Figure 31-3). Reciprocally, a rare imino tautomer of C can pair with A, forming two H bonds. A rare enol tautomer of T can pair with G, forming three H bonds, and the same T':G pair can be formed when the purine is in an uncommon tautomeric state. In each of these cases, then, a tauto- meric shift has made a new purine-pyrimi- diiie base pair possible.9 Tautomeric shifts may play an important role in spontaneous mutation. The relative frequencies of the different tautomeric alternatives depends upon several factors, including pH. The un- usual base pairs, A:C and T|G, can also occur after ionization of any one of the bases. Chemical changes in the old bases can also occur after treatment with chemical '•' See reference on p. 277 to J. D. Watson and F. H. C. Crick (1953c). Adenine ^ Oh Adenine Cytosine FIGURE 31-3. Tautomeric shift of adenine which could change its complementary base from thymine to cytosine. Upper diagram shows adenine before, and lower diagram after, undergoing a tautomeric shift of one of its hydrogen atoms. (After J. D. Watson and F. H. C. Crick.) 396 ( II M'TF.R 31 mutagens. Nitrous acid (HNO_) is muta- genic tn I \1\ . 12 and 14 phages, bacteria. yeast, and to transforming DNA. This mu- tagen removes NH_. from — that is. deami- nates — purines and pyrimidines both in DNA and RNA. Deaminatcd A becomes hypo- xanthine (Figure 21-4) (which then pairs with C); deaminatcd C becomes U (which then pairs with A): and deaminatcd G be- comes xanthine (which still pairs with C but with only two H bonds). Exposure of 4>T4 to low pH induces point mutations. /// vitro, low pH causes depuri- nation — the complete removal of all G and A — which results in apurinic acid. Al- though the backbone of apurinic acid may break after a return to higher pH. it is likely that the point mutations produced in <£T4 by low pH are caused either by an incorrect replacement of bases or by the formation of a complement with the complementary nu- cleotide deleted. The absorption of ultraviolet (UV) light by nucleic acids depends primarily upon the presence of chromatophoric groups (special groups containing double bonds). When free bases are treated with UV, pyrimidines are found to be more liable to chemical change than purines. One common change is the addition of water to the double bond between the number 4 and 5 carbon atoms of pyrimidines. The photoproduct in the case of cytosine is shown in Figure 31-4A. Although the photoproduct reverts to cyto- sine upon heating or acidification, it may persist frequently enough /'// vivo to weaken H-bonding between C and G, thereby lead- ing to localized areas of strand separation or denaturation. Supporting this view is the observation that UV disrupts H-bonding in native double-stranded DNA. UV also changes thymine at the same position as cytosine, by breakage of the 4-5 double bond; in this case two thymines unite to form a dimer (Figure 31-4B). The UV-initiated hydration of C is expected — via the weakening of H-bonding between C and G — to increase the likelihood of T dimeriza- tion. Thymine dimers form not only be- tween T's on different strands — thereby producing cross links between DNA chains — but also between adjacent T's on the same strand. ( Interchain crosslinking is also pro- duced by the mutagenic antibiotic, mito- mycin C.) Interchain dimerization prevents chain separation and also blocks replication. A. NH2 N lf"H I H Cytosine NH- Photoproduct figure 31-4. Effect of ultraviolet light upon DNA pyrimidines. ( The // atoms attached to ring C atoms are shown.) O o QANAH N^O H H Thymine Monomers J800A UV * H— N 2400A \ II cr >r ii ii isr x) H H Thymine Dimer The Molecular Basis of Mutation 397 whereas intrachain dimerization interferes with the proper base-pairing of T with A, leading eventually to the formation of incor- rect complements. Dimerization may explain how UV de- stroys the primer-template activity of single- stranded DNA, causes mutations in X174, and destroys transforming DNA. UV radia- tion has two opposite effects, however, de- pending upon the wavelength employed: At 2800A UV radiation tends to form di- mers from monomers, whereas at 2400A it tends to form monomers from dimers. In fact, DNA inactivated as primer-template by 2800A UV radiation is partially restored to activity by subsequent exposure to 2400A radiation. Similarly, the transforming ac- tivity of Hemophilus DNA inactivated by 2800A can be partially reactivated by subse- quent irradiation at 2390A. With large doses of 2800A, about 50% of the biological inactivation — as measured by transforming ability — can be attributed to T dimer for- mation, one inactivating "hit" being equiva- lent to one dimer formed for each 160 nu- cleotides.10 The replicational consequences of intra- strand dimer formation can be studied in vitro}1 After various single-stranded DNA primer-templates are exposed to 2800A ra- diation, the products of synthesis are sub- jected to nearest-neighbor analysis. As ex- pected, the frequency of the AA sequence decreases in proportion to the TT sequences dimerized; the dinucleotide sequences con- taining G, especially GG, increase in fre- quency. These results strongly indicate that T dimers in vivo decrease the chance that the complementary AA sequence will incor- porate opposite a TT sequence in the tem- plate and suggest — but do not prove — that these A's are often replaced by G's. '" As found by R. B. Setlow and J. K. Setlow (1962). 11 As shown by R. B. Setlow. W. L. Carrier, and F. J. Bollum (1963). As mentioned, photorecovery from UV- induced dimer formation occurs after expo- sure to UV radiation of shorter wavelength. In the presence of light of certain longer wavelengths — blue light — a particular en- zyme system has been found which can break T dimers- — including those which are inter- chain— into monoroers. Such a case illus- trates chemoplwtorecovery (p. 191). Since recovery from a mutagenic UV treatment is only about 50%, UV radiation probably produces mutations in other ways than dimer formation. Since large doses of UV can cause breaks in the DNA backbone in vitro, this effect is probably another mutagenic pathway in vivo. Although thymine dimers block DNA syn- thesis in vitro and in vivo, certain strains of E. coli are UV-radiation-resistant and even in the dark can recover to resume DNA syn- thesis. Such a recovery in these cells does not involve the splitting of thymine dimers; instead, the dimers are, by some mechanism, removed from the DNA (the acid-insoluble fraction) and appear in the acid-soluble fraction. In an irradiated, sensitive strain, which cannot synthesize DNA in the dark, the dimers remain in the insoluble phase and remain photorecoverable.1- Other work 13 suggests that intrastrand thymine dimers are removed from the DNA of resistant cells enzymatically, and that corrected DNA is reconstructed from information on the com- plementary strand. Such an error-correcting mechanism would be biologically important for the preservation of DNA. Finally, it should be noted that although dimerization of 5-bromo uracil is difficult, if not impossible, uracil dimers can be made by UV irradiation. Consequently. UV is expected to be mutagenic to RNA by the same mechanisms as it is to DNA. Tautomeric shifts, physical and chemical 12 See R. B. Setlow and W. L. Carrier (1964). 13 See R. P. Boyce and P. Howard-Flanders (1964). 398 < IIU'TFR 31 NORMAL T:A AT MUTATED CHAIN SEPARATION T:A / \ T A etc. at' / \ A T etc. REPLICATION 1 CHAIN SEPARATION a': C / A' etc. C T iC / \ T' C etc. REPLICATION 2 TRANSITION c : G C | TRANSVERSION (T»GI G figure 31—5. One postulated sequence of events leading to transition or transversion. mutagens, and low pH may ultimately cause base substitution. Replacement of one pu- rine by another purine (A«-»G) or one pyrimidine by another pyrimidine (T<-»C) is called transition; replacement of a purine by a pyrimidine or the reverse (for example A«-*C or TeG) is called transversion.14 Both kinds of substitution should be possible at the nucleotide and subnucleotide level. What is the sequence of events involved in a transition or transversion? A particular base pair, T:A, exposed to a mutagen may become T:A' (Figure 31—5). For exam- ple, suppose that at the time of strand sep- aration, A' specifies C (instead of T), and at the next division C acts normally to spec- ify G. The net result is that the original A strand eventually produces a second-genera- tion strand carrying G; this is a transition. (Or, given the original pair G-C, a mutagen may produce C' which specifies A instead of G which, in turn, specifics T. The net change from C to T in this example is also a transition.) If A:T becomes A:T', T' specifies C, and C specifies G, the overall result is that T is replaced by G, and a trans- version occurred. Another possible mechanism for intranu- 14 Following the terminology of E. Freese. cleotide base changes requires the members of a base pair to undergo rotational substi- tution by breaking their bonds to sugar, ro- tating 180 . and rejoining.1 Thus, after rotational substitution, which may be a fre- quent consequence of ion action, C-G be- comes GjC, the resultant double transver- sion being mutant. Note that in all the mechanisms mentioned for producing base changes, the transitions and transversions were initiated by a change in the old base. Base changes in new genes. Base analogs are incorporated into DNA in vitro when they are present as deoxyriboside 5'-triphos- phates (p. 284). For example, uracil (U), 5-bromo uracil (BU), or 5-fiuoro uracil (FU) can be substituted for T only; 5- mcthyl, 5-bromo, or 5-fluoro cytosine can be substituted for C only; hypoxanthine can be substituted for G only. BU, 5-chloro uracil, and 5-iodo uracil can replace some of the T in DNA of bacteria, phages, and human cell lines and are also highly muta- genic. 5-Bromo deoxyuridine (BUdR) is a more efficient mutagen than BU, probably because it is more readily converted to the triphosphate condition and interferes less with the formation of U or C. Consider the kinds of mistakes which may occur involving BU. Since the usual tauto- mer of BU (like T) is in the keto state, this tautomer is usually incorporated as the com- plement of A. The rare enol tautomer of BU (like T), however, can pair with G, forming BU;G. Consequently, two kinds of mistakes of incorporation of BU are pos- sible: formation of the A:BU and the GjBU pairs. Once part of a DNA strand, the BU of an A:BU pair can continue to specify A. so that no mistakes in replication occur. If, however, such a BU assumes its rare enol state and accepts G as its comple- ment, the grandparental A will be replaced '•"'See H. J. Muller, E. Carlson, and A. Schalet ( 1961 ). The Molecular Basis of Mutation 399 by G in a transition resulting from a mistake in replication. Since the BU in GjBU is usually in the keto form at the time of the next replication, it usually accepts A as its complement, resulting in the G to A transi- tion. Therefore the pyrimidine BU is ex- pected to produce purine transitions in both directions ( A <-* G ) . Errors involving BU can be studied in vitro. The copolymer of A and BU, dABU , can be made and used as primer-template in an extended 30 to 100% synthesis with TPPP, T2 gave mottled plaques whereas <£X174 gave only nonmottled plaques. What do these results suggest about DNA structure and the molecular basis of mutation? 31.6. S. Zamenhof and S. Greer found that heating E. coli to 60° C is mutagenic. What molecular explanations can you suggest for this result? 31.7. Chemical substances carrying one, two, or more reactive alkyl groups (CnH2u , , ) are called mono-, bi-, or polyfunctional alkylating agents; many of these are mutagenic. Depending upon the particular alkyl group, DNA can be altered in its phosphate or base portions. Under what circumstances would you ex- pect the use of alkylating agents as mutagens to be unsuitable for determining the molecular basis of mutation? The Molecular Basis of Mutation 403 31.8. How permanent must a change in a nucleotide be before it can be considered a mutant? 31.9. Would you consider the substitution of P1-' for P in the phosphate of a nucleo- tide a mutation? Why? 31.10. What do you think of the statement that the only way we have of detecting changes in individual genes is by the phenotypic changes they produce? 31.11. E. coli contains a locus capable of conferring resistance to c/>Tl irradiated with ultraviolet light. Substituting 5-bromodeoxyuridine for the thymidine in the phage, however, removes this protection. What do you suppose is the product and mechanism of action of the bacterial locus involved? Chapter 32 GENE ACTION AND POLYPEPTIDES D kURING interphase, the nucleus plays a very active and es- sential role in the normal metabolism of the cell. Let us make the oversimplified assumptions that chromosomal genes are the only nuclear components es- sential for normal metabolism, and that all of the features of metabolism unique to cells are the consequence of gene action. On this basis, then, all aspects of the phenotype that are genetic in origin result from biochemical effects of the genes. Because a cell contains a great variety of chemical substances, we expect one gene-initiated biochemical reac- tion to lead to others which, in turn, will lead to still others, forming a kind of tree whose branches represent successive chemi- cal reactions. Since all the branches would be affected by the initial, gene-caused bio- chemical change, we should find many differ- ent chemical, physiological, and morphologi- cal consequences of the initial change in the fully-developed cell or individual. It is not surprising, therefore, that a given genetic change usually has many different effects upon the phenotype and that most, if not all, mutants have manifold or pleiotropic ef- fects (Chapter 6). In tracing these pleio- tropic effects back toward their origin, we would expect the many different end effects to be the consequence of fewer earlier-pro- duced effects. Moreover, we would expect their more primary causes to be based up- on metabolic changes — changes sometimes identifiable with modifications of particular 404 chemical substances such as hemoglobin or pituitary hormone (Chapter 6). With this orientation in mind, we can be- gin a study of the biochemical basis of gene action — biochemical genetics. Information regarding the biochemical basis of gene ac- tion may be gained by studying a trait such as pigmentation, which, because it is describ- able in chemical terms, may require rela- tively few steps back to arrive at or near the primary, gene-caused biochemical changes. Alcaptonuria In man, a rare condition detectable at birth affects the color of urine. Though normal in color when passed, it soon darkens on contact with air and turns from light to dark brown and finally to black. This character- istic persists throughout the life of the indi- vidual. Family, pedigree, and population studies reveal that normal parents can have affected children of either sex, and that affected chil- dren appear with a much higher incidence when their parents — both normal — are re- lated. From the frequency of those affected within families, and the finding that the blackening of the urine is expressed fully or not expressed at all, we can conclude that affected individuals are homozygous for a single pair of completely-recessive, auto- somal genes. The blackening is due to the oxidation of a substance in urine called alcapton or homogentisic acid whose chemical descrip- tion is 2,5-dihydroxyphenylacetic acid (Fig urc 32-1). The disease is called alcapto- nuria ' and affected individuals, alcaptonu- rics. It should also be noted that several pedigrees have been found in which appar- ently the same phenotype is attributable to the action of a single, dominant gene. Since biochemical studies of dominant alcaptonura have not been extensive, our attention is 1 The account following is based upon the work of A. E. Garrod and subsequent investigators. Gene Action and Polypeptides 405 henceforth restricted to the recessive form of this disease. Alcaptonuria is clearly an inborn error of metabolism and results in the daily excretion of several grams of alcapton. A study of the biochemistry of alcaptonurics shows that, of numerous substances tested, only alcapton appears in abnormal quantities in the urine or blood, and that the reducing properties of the urine can be attributed entirely to the alcapton it contains. These results suggest that we have traced the pedigree of causes back to, or very close to, the primary effect of the gene. If alcapton is a substance produced by the abnormal gene, it should be absent in homo- figure 32-1. Sequence of chemical substances involved in the formation and metabolism of alcapton. 4 = homogentisic oxidase, 5 = isomerase, 6 - hydro- lase. 3 involves two reactions — first an oxidation to 2,5-dihydroxyphenylpyruvic acid, then oxidative decarboxylation. H— C H— C ■ . C— H C— H H— C H— C C H— C— H H-CNH2 COOH Phenylalanine OH C— H i C— H C H— C— H H-CNH2 COOH Tyrosine o H— C CHo H C— COOH H HOOC-CH HC • COOH Fumaric acid H— C' H— C OH I . ' H H C H— C— H c=o COOH p-OH phenylpyruvic acid I3 OH I H— C H— C ^ C— H H i OH C C— COOH H Alcapton (Homogentisic acid) ■> CO, + H20 406 ( HAPTER 32 zygotes for the normal allele. When alcap- tonurics are fed five grams of alcapton, approximately this additional amount is ex- creted in the urine. But when normal individuals are fed the same quantity of al- capton, none is found in the urine. If. how- ever, normal individuals are fed eight grams of alcapton, some is found in the urine. We can conclude from these observations that normal people have the ability to metabolize alcapton to another form which does not change color upon exposure to air and that this ability has been lost, apparently com- pletely, by alcaptonurics. The abnormal gene, therefore, does not produce its effect by forming alcapton as a unique substance. Alcapton seems to be a normal metabolic product which does not accumulate in normal individuals because it is rapidly metabolized, but which does accumulate in alcaptonurics. The blood of alcaptonurics proves to be de- ficient in a normally present enzyme which catalyzes the conversion of alcapton by oxi- dation to a noncolor-producing substance. This enzyme, homogentisic oxidase, is, in fact, missing in the liver of the alcaptonuric; thus, it must be this enzyme which is changed in alcaptonurics. It is clear, therefore, that alcapton is not produced by the gene for alcaptonuria but is a normal metabolic intermediate. Since it is not part of the normal diet, alcapton should have chemical precursors. If such a precursor of alcapton is added to the diet of alcaptonurics, it should be converted to alcapton which, in turn, would be excreted in increased amounts. When alcaptonurics are fed an excess of glucose, the amount of al- capton found in the urine is unchanged, in- dicating that glucose is not a precursor of alcapton. But, if p-OH phenylpyruvic acid or if either of the amino acids, tyrosine or phenylalanine, is increased in the diet of alcaptonurics, their excretion of alcapton is increased almost correspondingly. We can, therefore, postulate that alcapton has a series of chemical precursors (Figure 32-1 ). In the scheme illustrated, phenylalanine is con- verted to tyrosine by the addition of an oxy- gen to the top carbon; tyrosine is converted to p-OH phenylpyruvic acid by replacing the amine (NHL. ) group by an oxygen; p-OH phenylpyruvic acid is converted by still other chemical reactions to alcapton. Normally, alcapton is converted to acetoacetic acid by a process which involves the oxidation and splitting-open of the benzene ring; it is the first step in this conversion which fails in alcaptonurics. This hypothesized pathway from phenylalanine through alcapton to ace- toacetic acid has been confirmed in subse- quent work and seven enzymatically-cata- lyzed steps have been identified. It should be realized, however, that tyro- sine, an essential component of protein, can also partake in biochemical pathways other than the one leading to alcapton (Figure 32-1 ). For example, tyrosine is part of the pathway of chemical reactions leading to melanin formation; and so tyrosine, by a different chemical pathway, is also a precur- sor of melanin. Albinism (lack or absence of melanin) could be caused genetically by the defective production of an enzyme nec- essary for the conversion of tyrosine to mel- anin. In another disease due to a single, rare, recessive gene, affected individuals are feebleminded or of lower than normal mental ability and have other phenotypic changes including light pigmentation. This pleio- tropism is directly correlated with the pres- ence of phenylpyruvic acid in the urine of affected individuals. The normal conversion of phenylalanine to tyrosine fails to occur in such individuals; instead, the amine in phenylalanine is replaced by an oxygen (thus forming a keto group), so that phenylpyruvic acid is produced (Figure 32-2). Diseased persons are therefore called phenylpyruvics or phenylketonurics (Chapter 15 ). The dis- ease, phenylketonuria, can be partially allevi- Gene Action and Polypeptides 407 FIGURE 32-2. Formula for phenylpyruvic acid. H I H— C C— H H— C. JZ— H C H— C— H c=o I COOH ated or circumvented if phenylalanine — which is essential to proteins — is reduced in the diet to an amount sufficient for protein synthesis but insufficient for any appreciable quantity to be converted into phenylpyruvic acid. Since tyrosine is also needed for hu- man protein, it must be present in sufficient quantity in the diet of phenylketonurics. Finally, it should be noted that a parahy- droxylase, which converts phenylalanine to tyrosine and is normally present in the liver (where most of the phenylalanine is nor- mally metabolized and oxidized), has not been found in phenylketonurics. Inborn metabolic defects are of great help in identifying the places where genes direct metabolic processes. They also permit the determination of precursors of a genetically- defective step and aid in the study of chains of biochemical reactions and metabolic path- ways. For example, if mutant 1 cannot form substance Y but accumulates substance X, and if mutant 2 can only form Y when X is supplied, then X must be a precursor of Y (Figure 32-3). Biochemical genetics is of special inter- est in another respect. In the cases most thoroughly investigated, one can trace the pedigree of causes back to a point where only one effect of the gene is detected, for exam- ple, as in alcaptonuria. It is quite improb- able that further study of the gene for al- captonuria will reveal another phenotypic effect which, when tested adequately, will prove to be produced independent of the effect upon homogentisic oxidase. Present findings suggest, therefore, that this gene acts upon the phenotype only in one, pri- mary way. One Gene-One Primary Effect Hypothesis The study of biochemical genetics in this chapter (and also in Chapter 6) leads us to hypothesize that each gene has only a single, primary, phenotypic effect and that all the pleiotropic effects of a gene stem from this single activity. If the hypothesis of one gene-one primary phenotypic effect is sub- stantiated, we may be able to determine the size or scope of the genetic material whose action produces a single, primary effect. Such information would reveal the nature of the functional genetic unit, but it should be real- ized that this kind of information will de- figure 32-3. Determination of precursors using mutant genes. A accumulates X but makes no Y. B makes no X but will make Y if X is supplied. C is the normal pathway. mutant 1 B mutant 2 I added — Y IDS CHAPTER 32 pend upon what is considered to be primarj and what is identified as a phenotypic effect. An additional implication of the one gene- one primary effect hypothesis is that it we know the nature oi a primar\ etleet. it should always be the result of one gene. To test this prediction, it is accessary to decide which particular aspects of the phenotype arc primary effects of gene action. The cases just discussed indicate that a mutant gene has a primary effect upon the catalvtic ability of an enzyme. If we assume that the catalytic abilities of all enzymes re- sult from the primary action of genes, it should be possible to study any particular enzyme and find that its catalytic ability can be altered or abolished as the result of mu- tation. Experimental support for this one enzyme-one gene hypothesis would also provide specific — though limited — support for the more general concept of one gene-one primary phenotypic effect. One Enzyme-One Gene Hypothesis Not only is Neurospora a good organism for studying genetic recombination (p. 124), but it has certain very favorable characteris- tics for biochemical studies. Neurospora can manufacture all the components it needs to exist and to reproduce from a basic, very simple, food medium consisting of water; an array of inorganic salts; sources of nitrogen, phosphorus, sulfur; various trace elements; a carbon and energy source such as a sugar; and a single vitamin, biotin. From these raw materials it can synthesize all twenty different amino acids, all essential vitamins (except biotin), purines and pyrimidines, and every- thing else needed for its total activity. Ac- cording to the hypothesis under considera- tion, it should be possible to induce mutants that change the catalytic ability of enzymes, thereby blocking various chemical syntheses. Previous work has established that the last step in the synthesis of vitamin Bx, or thiamin, is normally accomplished by the enzymatic combination of a particular thi- azole with a particular pyrimidinc. If the catalytic action of every enzyme depends upon the primary action of genes, it should be possible to induce a mutation in the gene that normally specifies this B,-forming en- zyme. If the mutant no longer produces the active B,-forming enzyme, no B, will be made. Since B. is required for growth, the mutant mold will be auxotrophic and require B, in its diet to grow. An experiment can be performed -' in which asexually-produccd, haploid spores (p. 26) are treated with a mutagenic agent such as X rays or ultraviolet light. The treated spores are then grown on the basic medium supplemented with vitamin Bi. The spores that grow include prototrophs for B, as well as auxotrophs which obtain their B, from the culture medium. Once the spores have grown sufficiently, a portion of each of the growths is placed on a basic, minimal medium supplemented with the particular thiazole and pyrimidine that are the immedi- ate precursors of vitamin B,. (All other imaginable nutritional factors except B, it- self can also be added, but they will have no effect on the outcome. ) Cultures which fail to grow on a medium which contains the immediate precursors of Bi are clearly de- fective for the enzyme that catalyzes the last step in B, synthesis. Stocks of such cultures are made from samples growing in the pres- ence of B,. To test for and localize the genetic basis for the B, auxotrophy, each of these haploid strains is crossed to a haploid strain normal for B, synthesis. The diploid hybrid is formed and undergoes meiosis (pp. 124-126). producing a sac containing eight haploid ascospores. Each spore is removed and grown on a Bi-supplementcd minimal medium. If the haploid strain under test were Bi-deficicnt because of a mutation, transplants of each of the eight haploid cul- - Based upon work of G. W. Beadle and E. L. Tatum. Gene Action and Polypeptides 409 tures to a B,-frce minimal medium would produce exactly four that can grow and exactly four that cannot. As expected from our hypothesis, mutants are found which lack B] and do not contain this final enzy- matic activity. If for a given mutant, a number of spore sacs are tested as described, the locus of the mutant relative to the centromere of the chromosome in which it is located can be mapped (see Figure 9-10, p. 125). When no chiasma occurs between the loci of the mutant and the centromere, segregation of normal ( + ) and mutant (th) alleles occurs at the first meiotic division and — because the last two divisions in the ascus are tan- dem to the first — the eight ascospores occur in the relative order, + + + +//? th th th. However, when a single chiasma occurs be- tween the mutant and the centromere, segre- gation occurs in the second meiotic division, and the ascospores occur in the relative order, -f + th th + -f- th th. If a record is kept of the order of the spores in each ascus, the first and second division segregation arrangements can be identified after the spores are grown and their geno- types determined. It should be recalled that if 20% of all sacs show second-division segregation (two + spores alternating with two th spores), then 20% of the tetrads had a chiasma between the mutant and the cen- tromere, and the mutant is located ten map units from the centromere. When a number of separately-occurring point mutants, defective in the enzyme which catalyzes the last step in Bi synthesis, are localized this way, all are found to be on the same chromosome and approximately the same distance from the centromere. This result suggests that the catalytic ability of a particular enzyme is the result of the action of a particular gene. For the efficient detection of biochemical mutants in Neurospora, certain modifica- tions are made in the procedure already outlined. Potentially-mutant spores are grown on a medium supplemented with all substances which might conceivably be in- volved in biochemical mutation. Growing cultures are then transferred to a basic (minimal) medium containing no additions, where failure to grow indicates that the mu- tant culture has lost the ability to synthesize some component added to the basic medium. The specific ability lost is determined by testing for growth in a basic medium supple- mented, in turn, with the individual enrich- ing components of the complete medium. Techniques have been developed also to eliminate nonmutant strains selectively. Thus, spores given an opportunity to grow for a short time on a minimal medium can be subjected either to filtration, which sepa- rates the larger, (growing) nonmutant cul- tures from the smaller, (nongrowing) mu- tant ones, or to an antibiotic which kills actively-growing cultures but has less or no effect on nongrowing ones. In this way, the sample later tested for mutants can be mu- tant-enriched. It is even possible to find mutants for unknown growth factors by sup- plementing the culture medium with extracts of normal strains of Neurospora containing various substances, both known and un- known, needed by the mold. The same mu- tants requiring unknown growth factors can then be used in the specific assays needed for the isolation and identification of such substances. Such improvements in the techniques for detecting biochemical mutants in Neurospora expedite additional tests of the postulated enzyme-gene relationship. Two more tests are described briefly. The fiFst deals with the final step in the synthesis of the amino acid tryptophan and involves the catalyzed union of indole (in the substrate indole- glycerol phosphate) and the 3-carbon amino 410 (HAPTI.K 32 acid, serine. h\ the enzyme, tryptophan syn- thetase. Separately occurring tryptophan- requiring point mutants arc obtained which are blocked in the final synthetic step. All oi 25 mutants qualifying prove to be located on the same chromosome and at about the same locus. The second test involves the final step in the synthesis of adenine, cata- lyzed by the enzyme, adenyhsuccinase, which removes succinic acid from adenylo- succinic acid to leave adenine. Of 137 in- dependently occurring point mutations with little or no adenylosuccinase activity, all prove again to be on the same chromosome and at about the same locus. The genes specifying different enzymes are different, each occupying separate loci in the genome. These results and similar ones for other enzymes in Neurospora offer strong support for the hypothesis that the catalytic ability of all enzymes is under gene control. More- over, the addition of B,, tryptophan, or adenine to the diet of mutants defective in the enzymes directly responsible for their respective syntheses makes the mold com- pletely or almost completely normal, pro- viding good evidence that the genes in- volved have only one function to perform — determining the catalytic ability of one enzyme. If a gene had more than one primary effect, nutritionally overcoming one defect would not be expected to produce normality or near-normality in all cases. Because in all these cases the enzymatic de- fect is due to a defect only in one specific, localized area of the genetic map, the total catalytic ability of an enzyme seems to be the result of the primary action of a single gene. One Gene-One Polypeptide Hypothesis All enzymes are protein, at least in part, and the catalytic ability of an enzyme is known to be due to its protein content and often added co-factors. Proteins are composed of amino acids (Figure 32-4) linked to each other by peptide bonds between carboxyl and amino groups to form polypeptide chains. The catalytic ability of an enzyme depends upon the number and kinds of amino acids contained, their order in the polypeptide, the number of polypeptide chains, the way in which the parts of a poly- peptide chain are arranged relative to each other, and the way in which the different polypeptide chains in a protein are arranged relative to each other. The enzyme, tryptophan synthetase, found in Escherichia coli, can be treated in vitro so that it dissociates into two proteins, that is, two polypeptide chains. Neither single chain has the usual enzymatic activity but, when the two chains are reassociated, nor- mal enzymatic activity is restored. Clearly, to have the specific enzymatic action both chains need to be joined. Since the two chains are so easily dissociable and reas- sociable, probably no complex gene-directed physical or chemical change is needed to join them together. Therefore, the basis for the catalyzing ability of the enzyme must reside primarily in the nature of the poly- peptide chains which, when joined, make not just any enzyme, but tryptophan synthetase in particular. This reasoning leads to the suggestion that each chain might be the re- sult of the primary action of a different gene. A number of bacterial mutants lacking tryptophan synthetase activity can be ob- tained.1 Some of them are defective in one polypeptide chain, and others are defective in the second. All the mutants causing de- fects in one chain are found to be recom- binationally separable from those producing defects in the other, although adjacent areas of the genetic map are involved. In this case we have the choice of considering the two adjacent areas either as a single func- tional gene or as two separate genes. Be- cause the nature of this enzyme seems to :; Based upon the work of C. Yanofsky. Gene Action and Polypeptides 411 depend upon what each of these two genetic areas does individually, it is considered that two genes are involved, and that each gene completely specifies a polypeptide chain. The union of the two chains comprising tryptophan synthetase may somehow be re- lated to the two genes involved being adja- cent. figure 32-4. The twenty types of common amino acids. GLYCINE (GLY) ALANINE (ALA) VALINE (VAL) ISOLEUCINE (ILEU) LEUCINE (LEU) H H O H H O H H O H H O H H O 1 1 II 1 1 II 1 1 II 1 1 II 1 1 II H — N— C— C— O— H | H — N— C— C- | -O— H H— N — C — C— O — H | H— N— C — C — O— H | H — N — C— C— O— H | H H — C— H I 1 /H H — C — C— H 1 /-H H — C — C — H H — C — H 1 1 ^u 1 ■ i_i 1 ^.H H ' H 1 H H — C— C — H H — C— H 1 H — C — H 1 1 ^H H — C — H H H— C — H 1 H 1 H LYSINE (LYS) ARGININE (ARG) HISTIDINE (HIS) PROLINE (PRO) SERINE (SER) H H O 1 1 II H— N — C— C- 1 -O— H H H O 1 1 II H— N — C — C — O— H 1 H H O 1 1 II H— N— C— C— 0— H 1 H o 1 II H — N — C— C— O — H / \ H H O 1 1 II H— N— C — C— O— H 1 H — C — H 1 H — C — H 1 H — C— H 1 H — C — H H — C — H 1 H — C— H 1 H — C — H 1 N — H 1 H — C— H 1 C = C— H 1 1 H — N N \// C 1 H-C-H H-C-H \ / C / \ H H H— C— H 1 O— H H— N — H 1 C=NH 1 H THREONINE (THR) ASPARTIC ACID (ASP) ASPARAGINE (ASN) GLUTAMIC ACID (GLU) GLUTAMINE (GLN) H H O 1 1 II H— N — C — C — O- | -H H- H H O 1 1 II -N — C— C- | -O- -H H- H H O 1 1 II -N — C— C- | -O- -H H- H H O 1 1 II -N — C — C- -O- -H H- H H O 1 1 II -N — C — C— O— H H— C— O— H 1 H— C — H | H— C — H 1 1 H — C — H I l H — C — H ■ 1 H — C — H 1 1 c=o 1 c=o 1 H— C — H 1 H— C — H i i H 1 O — H 1 H — N — H 1 c = o 1 O — H 1 c = o 1 H— N— H CYSTEINE (CYS) H H O I I II H — N — C— C — O— H I H — C — H I S — H METHIONINE (MET) H H O I I II H— N— C — C— O— H I H — C— H H- H — C- I H -H TYROSINE (TYR) H H O I I II H — N — C— C— O— H TRYPTOPHAN (TRY) PHENYLALANINE (PHE) H H O H H O I I II I I II H — N — C — C— O— H H— N— C— C— O— H I I H — C — H H— C— H I XH I c=c X N — H 412 CHAPTER 32 What bearing have these results upon the genera] hypothesis o\ one gene-one primary phenotypic effect? Although the general hy- pothesis is unaffected, the specific hypothesis to test it — one enzyme-one gene — should be made more comprehensive and should be stated as one polypeptide-one gene, meaning that the eomposition of a polypeptide chain is completely determined by one gene. Ac- cording to the general hypothesis, then, the primary effect of at least some genes is to specify completely the amino acid content of a polypeptide. If the one polypeptide- one gene hypothesis is correct, we expect every polypeptide chain in every protein — including proteins that are not enzymes — to be completely specified by the primary and solitary action of a single gene. Biochemical Genetics of Hemoglobin In man, hemoglobin 4 is a protein with a molecular weight of about 66,700. In the horse (and probably in man) the shape of the molecule is spheroidal; its dimensions are 55 by 55 by 70 A; and it is composed of two dimers. Each dimer is composed of two identical polypeptide chains and the polypeptides in the two dimers are usually different. Each of the four monomeric chains contains about 140 amino acids and has a molecular weight of about 1 7,000. The chains partly coil to form what are called right-handed helices, and different chains are coiled about each other in a reg- ular way. An iron-containing heme group fits into a pocket on the outer surface of the coil of each chain. In the whole hemoglobin molecule, therefore, there are four heme groups — one for each of the chains — and a total of about 560 amino acids. Since the heme groups are not involved in the varia- tions to be considered, we shall henceforth 4 Based upon the work of V. M. Ingram, L. Paul- ing. H. A. Itano, H. Lehrmann. J. V. Neel. M. F. Perutz, and others. be concerned only with the protein, or globin, part of the molecule. Hemoglobin isolated from normal adults contains three components: A (or A,), AL., and A;!. The A component, called hemo- globin A (Hb-A) comprises about 90% of the total hemoglobin and the A2 compo- nent (Hb-Aj) about 2.5%. The remain- ing percentage of about 7.5 is due to the A.t component, probably Hb-A that has be- come chemically altered during aging of the red blood corpuscles. Hemoglobin A. In vitro, Hb-A can be dissociated into the two kinds of homo- dimer. and can be reassociated to reform the Hb-A tetramer."' Since the monomers are called «A and /i\ the reversible reaction can be written «A p£ ^± a A + /32v . The globin part of the molecule can also be par- tially digested with trypsin, which specif- ically cleaves the peptide bonds between the carboxyl group of lysine or arginine and the amino group of other amino acids. This digestion produces 28 smaller polypeptides, or peptides, in duplicate (since there are two chains of each type), plus an undigested core composed of about 25% of the orig- inal globin. The 28 peptides can be sepa- rated from each other since, on filter paper, they migrate at different rates when the digest containing them is subjected to an electrical field and various solvents. This treatment results in separate spots — "finger- prints"— on the filter paper for each of the peptides (Figure 32-5); each peptide (fingerprint) is given a different number and then analyzed for amino acid content. Peptide 4, for example, normally contains eight amino acids in the following sequence: Val-His-Leu-Thr-Pro-G/M-Glu-Lys. . . .6 The core of globin can be digested with chymotrypsin and fingerprints obtained of 5 See also G. Guidotti, W. Konigsbcrg. and L. C. Craig (1963). ,; These abbreviations are explained in Figure 32-4. Gene Action and Polypeptides 413 a • • . A i..- ,1 Hb-S .0 o 0 ?• 26 + Hb-A figure 32-5. The "fingerprints" of hemoglobin obtained after trypsin treatment. (Courtesy of V. M. Ingram, from C. Baglioni, Biochim. Bio- phys. Acta, 48:392-396, 1961.) its peptides. When these and other analyt- ical procedures are carried out, the sequence of all the amino acids in the «A and /3A chains can be determined (Figure 32-6). Note that the Val in peptide 4 is at the N- terminus of the (3A chain. Persons heterozygous for the gene for sickling have the "sickle cell trait," readily detected when their red blood corpuscles are exposed to an oxygen pressure very much lower than normal; persons homozy- gous for this mutant have "sickle cell ane- mia," and their red cells sickle even when the oxygen pressure is not so drastically re- duced. The hemoglobin of both types of persons has been fingerprinted and analyzed for amino acid content. The hemoglobin of the mutant homozygote is identical with 414 CHAPTER 32 TRY CLY LYS VAL Cl.Y ALA HIS ALA CLY CUI ALA LEU MET PHE LEU- THY CLY LYS VAL CUI ALA LEU LEU LEU VAL- VAL LYS LYS CLY HIS LYS PRO ASN CLY MET VAL ALA ASP PRO THR SER LEU ASP CLY PHE SER CLU PHE PHE ARC CLN VAL LYS LYS CLY HIS ALA SER CLY HIS SER LEU ASP 1'Ht HIS PRO PHE TYR THR LYS LEU THR ASN ALA VAL PKE SER ASP CLY LEU MET PRO ASN ALA LEU SER ALA LEU LYS CLY THR PHE ALA THR VAL ASP PRO VAL ASP PRO 130 CLY ALA VAL VAL LYS CLN TYR ALA ALA GUI VAL SB AiJk LEU PHE LYS ASF LEU SER AU BIS VAI. ■THE VAL LEU THR SER LYS TYR ARC ASP ALA LEI 140 ALA HIS LYS TYR HIS 120 1 10 PRO THR PHE CLU LYS CLY PHE HIS HIS ALA LEU VAL CYS VAL LEU VAL ASN CLY LEU LEU PRO THR PHE CLU ALA PRO LEU HIS ALA ALA LEU THR VAL LEU LEU CYS HIS SER LEU LEU figure 32-6. The amino acid sequences of the a and fj peptide chains of Hh-A. The amino acids enclosed by solid lines are identical and occupy corresponding positions along the peptide chains. The amino acids are numerated sequentially from the N -terminus. {Reproduced by permission of Dr. Vernon M. Ingram.) hemoglobin A, with the exception that the sixth amino acid in peptide 4 has valine sub- stituted for glutamic acid (the particular amino acid italicized in the sequence given earlier in the text) (see also Figure 32-5). The heterozygote produces both this type of abnormal hemoglobin, called hemoglobin S (Hb-S) and hemoglobin A. Other studies of the gene for sickling (see Chapter 6 and p. 209) show that all of its manifold pheno- typic effects are traceable through a pedi- gree of causes to this single amino acid sub- stitution in the /iA chain. Another mutant, known to be located on the same chromosome as the gene for sick- ling, produces hemoglobin C which differs from hemoglobin A by replacing the same glutamic acid in the fiA chain, this time with lysine. Still another genetic change pro- duces another hemoglobin, hemoglobin G. The amino acids in all the trypsin-produced peptides are the same as in hemoglobin A, except that the seventh one from the N-ter- minus in peptide 4 is glycine instead of glutamic acid. In this case, then, the amino acid sequence in peptide 4 is: Val-His-Leu- Thr-Pro-Glu-G/y-Lys. . . . Here, an amino acid in a different position — position 7 — in the BA chain is changed. In hemoglobin E a glutamic acid — normally found in position 26 on the fix chain (see Figure 32-6) — is replaced by lysine; this probably is the only change in the whole molecule. The preceding is evidence that different mutants cause single amino acids located in different positions in the flx chain to be re- placed by other single amino acids. Al- though the precise genetic basis for the dif- ferent mutants is unknown, the available evidence strongly indicates that all these Gene Action and Polypeptides 415 mutants are on the same chromosome. In fact, the simplest explanation is that the flx chain is specified by a single gene whose various mutants cause the different single amino-acid substitutions mentioned. Still other kinds of adult hemoglobin A are found. In some cases the amino acid sequence in the a chain is modified, as in hemoglobin I (in which a change occurs in position 16) and in hemoglobin "Hopkins- 2." Since an adult homozygous for hemo- globin A has a molecule describable as - Lys His ► Tyr His ► Arg Vol ► Glu Glu ► Lys Glu +- Glu-NHj 16 30 57 58 68 116 Lys ► Asp Glu >~ Glu-NH, Gly >- Asp His +- Tyr Asp-NHj P- Lys Glu ► Lys 416 CHAPTER 32 marj effect of gene action. We would also like to decide whether one or more genes are involved in hemoglobin A synthe- sis. Several lines of evidence point to an independent specification of a and fi chains: 1. Mutants that change the f3 chain (pro- ducing hemoglobins S, C\ E, or G ) produce no change in the a chain. 2. Mutants that change the a chain (pro- ducing hemoglobins 1 or Hopkins-2 ) produce no change in the (3 chain. Further evidence consistent with the inde- pendent specification of a and /3 chains comes from the study of individuals pos- sessing both Hopkins-2 and S hemoglobins. Such known individuals have one parent like themselves and the other of normal blood type (hemoglobin A). Also, these individuals have siblings with Hopkins-2 but not S hemoglobin, and others with the re- verse. Consequently, such Hopkins-2 + S persons cannot be monohybrid and must be dihybrid, since the abnormal hemoglobins can occur separately or together in different siblings. Because the number of siblings who must be recombinant is quite large, the two loci are either unlinked or, if linked, cannot be very close together. We can write the genotype of these dihybrids as aUo~aA /?s/iA. The two a and the two (3 chains in a given globin molecule are identical, even in heterozygotes. Since the Hopkins-2 -f S in- dividuals are dihybrid for mutants at widely separated loci, it seems reasonable that the two a chains specified by gene «""- (that is, a"""-) or by gene aA (a£) are produced independently of the two (3 chains specified by gene /is fjSf) or by gene /iA (pA). If so, either product of the two different a- specifying genes should be found joined to either of the two different products of the /^-specifying genes. Accordingly, the dihy- brid under discussion is expected to have all four of the following types of globin: an"--^. o'1"-'.^. a£j8f, «\i). This is found to be the case.7 Hemoglobin A2. The Hb-Au tetramer present in normal adults can be dissociated into two dimers and fingerprinted sepa- rately. One dimer is identical with a.}\ the other dimer is called 8;}-. The 8A- chain is very similar to the /iA chain, only 4 (or possibly 8 ) amino acid differences occur- ring among the 146 residues. Hb-AL. is rep- resented as «A8A-. Certain individuals are found s to produce only about half the nor- mal amount of Hb-AL.. In place of the missing component is an equal amount of a new hemoglobin Hb-B2. When finger- printed, the chains of Hb-B;. prove to have two a and two 8 chains in the tetramer. Further analysis shows that the a chains are a;} — that is, normal — but a single amino acid is substituted in the 8 chain which probably involves a change of Gly -> Arg at position 16. Hb-Bj can, therefore, be written «.A8^2. It should be noted that a person with Hb-B;. makes normal Hb-A, thus leaving the f3A chain unaffected. The 8 chain is presumably specified by a gene, 8, which is nonallelic to either the aA or f3A genes. Moreover, a person that makes both Hb-A;. and Hb-BL. presumably is genetically hybrid _SA2gB2. One study involved a man who made some Hb-S (o£j8f) and Hb-B-, (a£&f») and married a normal woman (with Hb-A, Hb-A;.). They had six children; those that made Hb-S did not make Hb-B;.., and vice versa. Since one of the hemoglobin defects in the father's hemoglobin was present in each of his parents, the simplest explana- tion is that the 8 and \3 genes are linked and that no crossover recombinants occurred in 7 See H. A. Itano and E. A. Robinson (1960). and ( . Baglioni ( 1963). 3 Bj K- Ceppellini. Gene Action and Polypeptides 417 his children. The father's genotype would aA #S8A- be, therefore, = As expected, he aA ftA8"°- actually made four kinds of hemoglobin — Hb-A, Hb-S, Hb-A,, Hb-B,. Finally, we find that, as expected, heterozygotes for Hb-I make not only Hb-A and Hb-1 but also make Hb-AL. and a defective Hb-A2 composed of al6A-. Dimers of the a chain apparently combine in a random way with dimers of the ft and 8 chains. Hemoglobin F. The hemoglobin of the fetus hemoglobin F has two a chains like those in adult hemoglobin A. Accordingly, persons with mutant aA genes make Hb-F whose a chains have the same abnormality as has Hb-A. The other two chains in Hb-F are different from the ft, 8, and a chains and are called y chains; thus, normal hemoglobin F is a%y\. The arrnno acid sequence in the yF chain of Hb-F is given in Figure 32-8. Homozygotes for the sick- ling gene can make hemoglobin F which is apparently normal, a^yfj so that a change in the ft chains has no effect on the y chains. Some known abnormal types of Hb-F are believed to be altered in the y chain. It is very likely, therefore, that a separate gene, yp, specifies yF chains and has allelic alter- natives. Hb-A appears in the fetus as early as the 20th week and gradually replaces Hb-F; even at parturition, however, there is still some Hb-F in the blood. The change from Hb-F to Hb-A means that during de- velopment the yF gene has its action turned off, so to speak, and gene ftA has its action turned on. The preceding evidence indicates that four genes are involved in the manufacture of fetal and adult hemoglobin, namely — a, ft, 8, and y. All results support the view that each kind of polypeptide chain in he- moglobin is completely specified by a unique gene. Because heterozygotes for mutants SER ALA VAL ALA THR ILBJ LEU TRY GLY LYS VAL ASN VAL LEU TRY CLY LYS VAL ASN VAL ASP CLU VAL GLU ASP ALA GLY CLY CLU GLY GLY GLU LEU CLY A*G LEU LEU VAL VAL TYR PRO LEU GLY ARG LEU LEU VAL VAL TYR PRO • GLY LEU SER THR • • SER PHE ALA CLY LEU VAL LYS LYS GLY HIS ALA LYS VAL LYS PRO ASN GLY MET LEU VAL LYS LYS GLY HIS ALA LYS VAL LYS PRO ASM GLY MET SER ALA SER ASP PRO THR GLY PHE SER PHE PHE ARG CLN THR TRY PHE PHE ARG GLN THR TRY GLY LEU ALA ALA HEU LYS HIS LEU ASP ASP LEU LYS GLY THR PHE ALA HIS LEU ASP ASP LEU LYS CLY THR PHE ALA LEU HIS CYS LEU HIS CYS LYS LEU HIS VAL ASP PRO LYS LEU HIS VAL ASP PRO LEU LEU GLY ASN ' LEU LEU GLY ASN ■ LYS HIS ALA LEU ALA ASP ALA VAL CLY ALA VAL VAL LYS GLN TYR ALA ALA CLN VAL PRO PRO THR PHE GLU LYS GLY PHE HIS HIS ALA LEU VAL CYS VAL LEU VAL ARG SER SER LEU ALA LYS HIS ALA LEU ALA 140 146 TYR HIS TYR HIS ALA VAL CLY ALA GLN VAL PRO THR PHE GLU LYS GLY PHE HIS ALA LEU VAL VAL LEU VAL •" figure 32-8. The amino acid sequence of the ft peptide chain of Hb-A and of the y peptide chain of Hb-F. The amino acids enclosed by solid lines are identi- cal and occupy corresponding positions along the peptide chains. The amino acids are numerated sequentially from the N-terminus. (Reproduced by permis- sion of Dr. Vernon M. Ingram.) ■4 1S CHAPTER 32 Present Genes Myoglobin Gene a yF p 8Ai Ancestral Gene FIGURE 32-9. Gene duplication and intragenic mutation hypothesis for the molecular evolu- tion of myoglobin and hemoglobin. which specify hemoglobin are rare, and be- cause the linkage maps of man are so in- complete, it is difficult to learn the precise relative positions of the various nonalleles. For the same reasons, it is difficult to study the allelism of hemoglobin mutants which affect the same chain. * Molecular Evolution of Hemoglobin !l Present-day myoglobin, a protein in muscle, is composed of a single chain of 155 amino acids which partly forms a right-handed a helix and carries a single heme group on its surface. When the amino acid sequence of myoglobin is compared with that of the a or (3 chain of hemoglobin, a large number of differences are found. After accounting for the difference in chain length, however, a number of places still remain where the same amino acid occurs on both types of chain. These similarities probably explain why both types of chain have the same three-dimensional arrangement. Though it •'See V. M. Ingram (1961). and C. B. Anfinsen (1959). is possible that some, If not all. of these similarities are due to convergent evolution by unrelated genes, we can postulate as the basis for the observed chain similarities that the genes specifying these present-day chains have a common gene ancestor (Figure 32- 9). According to this hypothesis, an ancestral gene, a, must have been duplicated in the genome by one of the mechanisms discussed in Chapter 12, since present-day species have separate loci for the specification of myoglobin and hemoglobin. Subsequent mutations of one gene could give rise to the present-day locus for myoglobin production, whereas mutations of the other gene could give rise to the ancestral locus for hemo- globin chains. Such mutations might re- sult in the addition or — more likely — the removal or substitution of amino acids sin- gly or in groups. This common-origin hy- pothesis is supported by the finding that the hemoglobin of the lamprey consists of a single polypeptide chain with a molecular weight of about 17,000 and that hagfish he- moglobin appears to be a similar monomer or possibly a dimer with a molecular weight of about 34,000. Since all known hemoglobins of verte- brates, except for the lamprey, have a he- moglobin chain that starts with a Val-Leu sequence, they may all be products of mu- tants of a. Accordingly, it is suggested that the ancestral gene for hemoglobin is a and that after a arose, it mutated to an allele whose polypeptide product could form a dimer, since dimerization enhances a's ef- ficiency as an oxygen carrier. Suppose, next, that the a locus became duplicated and that one of the resultant loci mutated to y, which produced y chains, which, in turn, formed not only dimers but also te- tramers with the a dimers. The tetramer would be a fetal-type hemoglobin a2y2. Te- trameric hemoglobin is presumably more ef- ficient than dimeric hemoglobin. Gene Action and Polypeptides 419 From which gene, a or y, did the gene for (3 chains arise? Since the (3A chain is known to differ from both the aA and yF chains by about 21 to 23 amino acids, we are told nothing about which of the last two was the ancestral type of the (3 chain. Al- though just about as many mutants involv- ing the a as the f3 chain have been discov- ered, only those affecting the (3 chain occur in the population with any appreciable fre- quency. This finding, together with the similarities between vertebrate a chains mentioned earlier, suggests that in the te- tramer changes in the « dimer produce a greater selective disadvantage than those in the (3 chain. Remember that a change in gene «A modifies both fetal and adult he- moglobin. It may also be that certain a chain changes result in loss of ability to form tetramers. The homotetramer of a, aA, may not be possible, although (3 chains can form f3A and y chains can form y|\ We, thus, conclude that the ancestral (3 gene was probably derived from one of the products of a duplication of the y gene. As mentioned previously, the (3A and 8A2 chains differ in less than ten amino acids. Presumably, the (3 gene was duplicated in the genome, and one of the two resultant genes mutated to become the 8 gene; that the duplication is recent is suggested by the small number of amino acid differences be- tween the (3A and 8A* chains; by the appar- ent persistence of linkage of the f3A and 8 A- genes; and by the restriction in the occur- rence of AL.-like hemoglobin to the primates. In summary, it is likely that by means of gene duplication and intragenic mutations, the ancestral gene, a, gave rise to the myo- globin gene on one hand and the gene se- quence, a -> y -» f3 — » 8, on the other. Since polypeptides are apparently primary products of gene action, the study of poly- peptides should considerably advance our understanding of the molecular basis of evo- lution. SUMMARY AND CONCLUSIONS The biochemical activities necessary for the existence of protoplasm are controlled by the nucleus, presumably by the genes it contains. These chemical reactions occur in sequences that form many-branched, metabolic pathways leading to the chemical, phys- ical, physiological, developmental, and morphological aspects of the phenotype. Be- cause of this branching most, if not all. genes have pleiotropic effects. The phenotypic differences produced by different alleles can be traced back toward the gene by a pedigree of causes. Such studies demonstrate that genes produce their effects at the metabolic level. The study of inborn errors of metabolism in man demonstrates that by their influence upon enzymes genes control various steps in biochemical sequences, and in these cases, the effect on enzymes appears to be the primary and the only consequence of gene action. In view of these experimental results, a one gene-one primary effect relationship is hypothesized — that a gene produces only one primary effect and that any primary effect is the result of the action of a single gene. The specific hypothesis, one enzyme- one gene, proposed as a test of the general hypothesis, is supported by biochemical and genetic studies of auxotrophy (for B^ tryptophan, adenine, and other nutrients) in Neurospora. The biochemical genetics of tryptophan synthetase in E. coli and of hemoglobin require that the hypothesis of "one enzyme-one gene" be generalized to "one polypep- 420 CHAPTER 32 tide-one gone." Thus, we consider thai the amino acid content o\ each polypeptide is specified completely In the primary action of a single gene. The one gene-one pri- mary effect hypothesis is therefore supported, and it is concluded that one way tor a gene to act in a primal \ \\a\ is to specify polypeptide amino acid content. Conse- quently, the \\a\ is opened for the study of evolution at the biochemical level. The molecular evolution of hemoglobin is discussed. REFERENCES Anfinsen, C. B., The Molecular Basis of Evolution, New York: J. Wiley & Sons, 1959. Baglioni, C, "Correlations between Genetics and Chemistry of Human Hemoglobins," Chap. 9. pp. 405-475, in Molecular Genetics, Part I, Taylor, J. H. (Ed.), New York: Academic Press, 1963. Beadle, G. W., and Tatum, E. L., "Genetic Control of Biochemical Reactions in Neuro- spora," Proc. Nat. Acad. Sci., U.S., 27:499-506, 1941. Reprinted in Classic Papers in Genetics, Peters, J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, 1959, pp. 166-173. Beam, A. G., "The Chemistry of Hereditary Disease," Scient. Amer., 195:126-136, 1956. Bonner, D. M.. and Mills, S. E., Heredity, 2nd Ed., Englewood Cliffs., N.J.: Prentice- Hall, Inc., 1964. Guidotti, G., Konigsberg, W., and Craig, L. C, "On the Dissociation of Normal Adult Hemoglobin," Proc. Nat. Acad. Sci., U.S., 50:774-782, 1963. Harris, H., Human Biochemical Genetics, Cambridge: Cambridge University Press, 1959. Hsia, D. Y.-Y., Inborn Errors of Metabolism, Chicago: Year Book Publishers, 1959. Human Genetics, Cold Spring Harb. Sympos. Quant. Biol.. 29, 1965. Ingram, V. M., "Gene Evolution and the Haemoglobins," Nature, 189:704-708, 1961. Reprinted in Papers on Human Genetics, Boyer, S. H., IV (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, 1963, pp. 164-175. Itano. H. A., and Robinson, E. A., "Genetic Control of a- and /i-Chains of Hemo- globin," Proc. Nat. Acad. Sci., U.S., 46:1492-1501, 1960. Kendrew, J. C, "Myoglobin and the Structure of Proteins," Nobel Prize Talk, Science, 193:1259-1266, 1963. Nance, W. E., "Genetic Control of Hemoglobin Syntheses," Science, 141:123-130, 1963. Wagner, R. P., and Mitchell, H. K.., Genetics and Metabolism, 2nd Ed., New York: J. Wiley & Sons, 1964. Yanofsky, C, and Crawford, L. P., "The Effects of Deletions, Point Mutations, Re- versions and Suppressor Mutations on the Two Components of the Tryptophan Synthetase of Escherichia Coli," Proc. Nat. Acad. Sci.. U.S., 45:1016-1026, 1959. Reprinted in Papers on Bacterial Genetics, Adelberg, E. A. (Ed.), Boston: Little, Brown. 1960, pp. 384-394. See Supplement VI and the first portion of Supplement VII. Gene Action and Polypeptides 421 Harriet Ephrussi-Taylor (see p. 296), Boris Ephrussi, and Leo Szilard (see p. 391 ) at Cold Spring Harbor, N.Y. in 1951. (Courtesy of the Long Island Biological Asso- ciation.) QUESTIONS FOR DISCUSSION 32.1. List five diseases in man caused by inborn errors of metabolism. 32.2. In what respect can an inborn error of metabolism be cured? 32.3. Do all mutations produce inborn errors of metabolism? Explain. 32.4. What evidence can you present that genes control different steps of a biosyn- thetic sequence of reactions? 32.5. Is a study of mutation completely or partially dependent upon the concept of a functional genetic unit? Explain. 32.6. Do you suppose that proof of the one gene-one primary effect hypothesis would reveal anything about the chemical properties of a gene? Explain. 32.7. From which of these areas of investigation would you expect to obtain the most information regarding the gene — morphology, physiology, biochemistry? Why? 32.8. Do you think that the concept of a functional genetic unit has any consequences for the practice of medicine? Explain. 32.9. In what way is the study of a functional genetic unit related to or dependent upon mutation and the genetic recombinational unit? 32.10. Is it a significant fact that a glutamic acid in hemoglobin A is replaced by an- other amino acid (valine, lysine, or glycine) in hemoglobins S. C. G. and E? Explain. 32.11. What are the disadvantages of using human beings as material for investigation of the gene? 32.12. Using Neurospora, design an experiment to detect crossing over within a gene. 32.13. Is the one gene-one primary function hypothesis equivalent to the one polypep- tide-one gene hypothesis? Why? 32.14. What evidence can you give for rejecting the hypothesis that a functional ge- netic unit is equivalent to a single genetic recombinational unit? 32.15. Would you expect a chemical substance specified in a primary way by a gene to be composed of linearly-arranged parts? Why? 32.16. Can you apply the term cistron to the one or more recombinational units that determine whether glutamic acid or lysine is located at a particular place in hemoglobin? Why? 122 CHAPTER 32 32.1/. What do genes actuall) do? Has your answer am hearing upon the concept of a gene? Explain 32. IS. Is it proper to use the term, allele, to describe functional genetic units rather than recombinational genetic units? Explain. 32.19. What conclusions can you draw from the observation that most of the amino acid substitutions involved in abnormal hemoglobins appear to lie on the sur- face rather than in the interior of the fully-folded hemoglobin molecule? 32.20. How do you interpret the finding that during hemoglobin synthesis, short-term exposures to the isotope Fe59 reveal its incorporation into Hb-A and Hb-AL. but not into Hb-A;i? 32.21. As indicated by arrows in the upper row of the accompanying diagram, trans- plantation of eye anlage (imaginal discs) between D. melanogaster larvae pure for dull-red ( + ) or bright-red — vermillion (v) or cinnabar (en) — eye color genes produces adults with the eye colors indicated in the lower row. From these results what can you conclude about the biochemical reactions leading to the production of brown eye color pigment? Chapter 33 POLYPEPTIDE SYNTHESIS AND RNA T: |he metabolism of an organ- ism is regulated primarily by proteins whose purposes are both structural (to make subcellular organ- elles), and catalytic (to make enzymes). Since the structure and function of an or- ganism is so dependent upon protein, it is not surprising that one of the primary ef- fects of genes is to specify the amino acid content of polypeptides (Chapter 32). Restricting our attention to DNA — the genetic material in most kinds of organisms — we ask, what can DNA do, or have done to it, which will result in the formation of particular polypeptide chains? Since we are dealing with conserved DNA, that is, DNA which remains part of a chromosome or other structure, whatever DNA does must happen in situ Since DNA is not protein, it cannot be an enzyme and probably does not act as a catalyst in producing its effect on polypeptide formation. However. DNA might possibly serve as a kind of template for specifying a polypeptide. Because ribose is more reactive than deoxy-D-ribose, RNA is less stable than DNA. Consequently, be- ing more inert, DNA is a more stable tem- plate than RNA. We already know that after strand separa- tion, each DNA strand serves as a template for the formation of a complementary strand. If DNA is also used as a template for gene functioning, then the four different bases — A, T, G, C — usually found in DNA must play a role in determining the nature of the 423 templates formed. In other words, the in- formation for the genetic specification of a polypeptide by a template mechanism would have to be contained in the bases of DNA. The nature of a polypeptide depends ulti- mately upon its amino acid content. Many polypeptide chains contain one or more of each of the twenty amino acids commonly found in organisms (Figure 32-4, p. 411), and the number and sequence of these build- ing blocks of polypeptides vary. Since both polypeptides and DNA are linear structures, the mechanism by which a sequence of DNA nucleotides serve as a template for specify- ing an amino acid sequence may be rela- tively simple to visualize. Ribosomes The synthesis of hemoglobin occurs in the cytoplasm of mammalian red blood corpus- cles— cells which no longer have a nucleus. Since the cytoplasm is the only site for pro- tein synthesis in this case, it is desirable to consider the structural components of the cytoplasm in some detail. For all cells which have been examined — plant, animal, and microorganismal — electron micrographs re- veal numerous bodies, called ribosomes (see Figure 1-3 and Suppl. IX Fig. 2), in their cytoplasm. These are particularly abundant in cells actively synthesizing protein and are also found in the nucleus and in chloroplasts. Ribosomes isolated from ruptured cells are characterized by their sedimentation rate — measured in the ultracentrifuge and ex- pressed in terms of sedimentation units, s. The fewer the s units, the smaller the par- ticle, although the relationship is not linear; therefore, s values reflect ribosome size. E. coli has four discrete ribosomal units: 30s, 50s, 70s, and 100s. The two basic sizes are 30s and 50s, the larger units being com- posites of the basic ones, as indicated in Figure 33-1. Both the 30s and 50s par- ticles contain about 64% RNA and 36% protein by weight. (Animal ribosomes are 424 < IIAI'TER 33 WHOLE PARTICLE 0 00 oooo Size 2(30s) 2(50s) 2170s) « KlOOs) Molecular Weight x 10* 0.85 0.15 1.80 0.15 2.8 0.2 5.9 1.0 RNA CONTENT Size Molecular Weight x 10' 16s 0.55 0.10 23s 1.15 0.20 FIGURE 33-1. Characteristics of E. coli ribosomes. 50% RNA by weight.) The smaller units aggregate to form the larger ones when Ml: or other divalent cations are added. Mammalian ribosomes behave similarly, al- though the basic particles, 40s and 60s, are somewhat larger than those in E. coli. The mammalian 80s particle (homologous to the E. coli 70s particle ) results from the com- bination of one 40s and one 60s ribosome. About 80% of the RNA in a cell is con- tained in ribosomes. (Small amounts of RNA are also reported in mitochondria.) Ribosomal RNA is single-stranded and has a relatively high molecular weight: 0.55 ± 0.10 X 106 for the 16s RNA component of the 30s particle and 1.15 ± 0.20 X 10° for the 23s RNA in the 50s particle; the number of nucleotides in 16s and 23s RNA is about 1000 and 2000. The 23s RNA, however, is not to be considered a dimer of 1 6s RNA because of evidence ' that their genetic de- rivations are different. Moreover, the RNA of ribosomes complexes best with denatured homologous DNA, suggesting that the ribo- somal RNA of different organisms differs in base sequence, if not base content. The synthesis of the RNA of certain mam- malian viruses is similar to 18s and 30s ribosomal RNA of mammalian cells in that 1 See S. A. Yankovsky and S. Spiegelman (1963), and S. Spiegelman (1964). both viral RNA replication and ribosomal RNA synthesis are inhibited by the drug puromycin. On the other hand. RNA viruses such as TMV are ribonucleoproteins whose protein portion is composed of a num- ber of identical subunits (Figure 28-2), whereas the 30s ribosome contains ten, prob- ably all different, polypeptide chains with a molecular weight of about 30,000. Clearly, the protein structure of ribosomes is more complicated than that of RNA viruses. After radioactive amino acids are injected into the body, tissues which synthesize pro- teins rapidly can be examined at intervals. - When a large dose of labeled amino acid is injected, the ribosomes are labeled almost immediately. When a minute dose of la- beled amino acid is injected it is expected to be used up rapidly in protein synthesis; the label in the ribosome increases quickly at first but then decreases. Finally, the labeled amino acid which moves out of the ribosomes is actually incorporated into pro- tein, for example, hemoglobin. These ex- periments give us clear evidence that ribo- somes are associated with protein synthesis. Since the amino acid sequence in hemoglobin is found to be a primary effect of the func- - The following is based on the work of P. C. Zamecnik and co-workers, and of M. Rabinovits and M. E. Olson. Polypeptide Synthesis and RNA 425 tioning of DNA genes (Chapter 32), it is necessary to understand how a DNA tem- plate that remains in the nucleus of a retic- ulocyte orders the amino acid sequence of hemoglobin manufactured in the cytoplasm. Clearly, if the DNA functions as a template in this respect, it must be doing so indi- rectly. DNA might be used to make another tem- plate, neither DNA nor protein, which can leave the nucleus and enter the cytoplasm where it will be used for protein synthesis. RNA fits this description, being a nucleic acid which also has a four-symbol code (A, U, C, and G) in which uracil (U) occurs in place of thymine (T) . Such a mechanism would require the four-symbol code of DNA to be transcribed directly into the four- symbol code of RNA — that is, it would in- volve a problem of transcription. It would also require RNA to carry information trans- lated into polypeptide sequences — that is, it would involve a problem of translation. Therefore, the multiple hypothesis is sug- gested that DNA nucleotide sequence is transcribed into RNA nucleotide sequence which, in turn, is translated into amino acid sequence. Messenger RNA Under normal circumstances, a considerable amount, if not all, of the RNA in higher or- ganisms is synthesized in chromosomes and then transferred to the nucleolus. Subse- quently, using radioactive tracers, RNA can be detected entering the cytoplasm. On the other hand, no evidence is found for a flow of RNA from the cytoplasm to the nucleus. These results are consistent with the hy- pothesis under consideration. The relationship between RNA synthesis and DNA can be studied in bacteria. RNA is synthesized in bacteria after infection with a DNA phage whose base ratio differs from that of the host DNA. The RNA manufac- tured after phage infection is different from the RNA manufactured prior to infection; its base ratio depends upon that of phage, since only the RNA synthesized after infec- tion can base pair in vitro with strand-sepa- rated phage DNA to form a hybrid double strand — one RNA and one DNA. (Hybrid RNA-DNA molecules have a unique specific density and, therefore, can be identified in the ultracentrifuge tube; they are also rela- tively resistant to RNase.) Also, freshly- made nuclear RNA from normal cells can form a complex with chromosomal deoxy- ribonucleoprotein.! Such results suggest the existence of a direct base-for-base de- pendence of nucleus-synthesized RNA and nuclear DNA. As mentioned, RNA complementary to phage DNA is made after a DNA phage in- fects its host. This phage-specific RNA is found to attach to a small percentage of already-formed ribosomes, suggesting that at least some ribosomes do not permanently carry a template of RNA (obtained from the DNA template) containing information for the specification of an amino acid sequence. Such ribosomes are capable of receiving seg- ments of RNA which carry the information for making phage-specific polypeptides. Thus, a type of RNA, called messenger RNA or mRNA, is synthesized. mRNA carries information for gene action from phage DNA to the ribosome. Presumably the mes- senger RNA causes the assembly of various amino acids at the ribosome where they are joined to form polypeptides. Messenger RNA is also found and functions in normal, uninfected cells.4 The RNA genetic material of MSX174 can be present in the RF and only one in the mature phage. 33.9. Compare the terminal nucleotides of TMV and sRNA. What can you infer from this comparison? 33.10. What conclusions can you draw from the observation that although the ribo- somal RNAs from Pseudomonas aeruginosa and Bacillus megaterium are indis- tinguishable, the DNA is 64% G + C in the former, and 44% in the latter? 33.11. Although the average cell of the adult rat liver probably divides less often than once a year, it synthesizes an amount of protein equivalent to its own con- tent every six or so days. In bacteria, on the other hand, the time required to double the protein content is roughly equal to the generation time. Com- pare the turnover of messenger RNA in bacteria with that expected in adult rat liver cells. 33.12. When native RNase is treated with urea and sulfhydryl reagents, its disulfide bonds are broken and the enzyme unfolds into an inactive linear form. When O.. is bubbled slowly through a solution of this denatured enzyme, the disulfide bonds reform and enzymatic activity resumes. What do these results tell you about the genetic basis for the folding of polypeptides? ( 'hapter .'54 GENETIC AMINO ACID CODING I f single ribotides in messenger RNA were translated into dif- ferent amino acids, only four amino acids would be specified or coded. Since there are twenty common amino acids, we are presented with the problem of how RNA codes for amino acids. To resolve this problem, we can assume that an amino acid is coded by a sequence of two nucleotides — a situation comparable to having an alphabet of four letters and a language of two-letter words. In this case, assuming the RNA code can be read only in one direction, we would have four times four, or sixteen, pos- sible doublets (words). (Unidirectional reading seems reasonable since a single strand of RNA is polarized just as a single strand of DNA.) However, sixteen dou- blets are still too few to specify twenty amino acids, so other assumptions must be made. We might hypothesize that a given doublet encodes more than one kind of amino acid, in which case the code would be ambiguous. Alternatively, we could as- sume an amino acid is coded by a sequence of three messenger ribotides — a triplet. Such a triplet code would give us four times four times four, or sixty-four, different, uni- directional sequences — more than enough to encode twenty amino acids. Should more than one triplet encode the same amino acid, the code would be degenerate. Thus, this introductory discussion suggests that a sequence of two or three ribotides encodes an amino acid — that is, acts as a codon. 436 Other characteristics of messenger RNA may affect amino acid coding. For exam- ple, since the number of consecutive ribo- tides can be in the hundreds or thousands, no spacing — that is. no non-nucleotidc punc- tuation— is provided to indicate where one codon stops and the next begins. Conse- quently, we are dealing with what is called a comma-free code. Suppose six ribotides are arranged linearly in positions 123456. If triplet 123 specifies amino acid A and 456 specifies amino acid B, errors are pos- sible due to overlapping triplets 234 or 345. The problem of overlapping codons can be avoided if only successive doublets or tri- plets are read starting at one distinct point on messenger RNA. In this case, the punc- tuation is provided by the mechanism for reading the code. The rll Region and the Code The genetic fine structure of the rll region of T4 has already been discussed in Chap- ter 26. We recall that the rll region is composed of two genes (or cistrons), A and B, both of which must function cor- rectly to yield the r+ phenotype. From the last chapter, it is inferred that these genes produce messenger RNA which specifies the two different polypeptide chains required for the r + phenotype. In the case of hemo- globin, the protein gene product is readily collected and analyzed, but the genetic basis for globin variants is difficult to study; the converse is true for the r+ phenotype. In other words, even though the presumed pol- ypeptide chains involved in producing r + have not been detected, the genetic basis for rll mutants can be readily determined. We would, of course, prefer to study a sys- tem whose genetic and polypeptide conse- quences both are easily investigated; never- theless, other genetic studies ' of the rll 1 The discussion follows the work of F. H. C. Crick. L. Barnett. S. Brenner, and R. J. Watts- Tobin (1961), and of others. Genetic Amino Acid Coding 437 region may reveal additional information about gene action and the RNA code. The A gene has been mapped into six major segments (A, through A6); the B gene into ten (B, through B,(l), all num- bered consecutively from left to right. Since complementation occurs, a point mutant in any one of the A segments has no effect on B function, and vice versa. A large number of point mutants can be induced in the B, and Bj regions by chemical mutagens ex- pected to cause transitional or transversional base substitutions. In some of these mu- tants all B activity is lost, and in others some product with B activity is detected. As ex- pected, those mutants assumed to involve base substitutions can be reverted to normal B activity by subsequent treatment with chemical mutagens which should produce the reverse transition or transversion. On the other hand, Bx or B2 mutants produced by acridines always completely inactivate B gene function and are not reverted by mu- tagens assumed to cause base substitutions. Such an occurrence is expected if acridines usually act as mutagens by causing the ad- dition or loss of one or more whole nucleo- tides (see p. 394). A large number of B-inactivating, acri- dine-induced mutants located in the Bt and BL. segments are obtained. After recombi- nation between such mutants, progeny phage which carry two to six different acridine- induced point mutants are recovered. Some of the doubly-mutant phages still show no B activity, but others do. If a complete series of different double-mutant combina- tions is made, a consistent pattern is ob- served. To interpret it we shall assume that a given single mutant is either + or — , that is, has either gained or lost one or more nucleotides. We shall also assume that a codon has more than two nucleotides and that the code is nonoverlapping; in other words, it is read in successive codons. Operationally, one isolates a " — " mutant as a suppressor mutant of a presumed "+" mutant, and vice versa. By isolating a se- ries of "suppressors" and a series of "sup- pressors of suppressors," one gets a series of + and of - mutants. It is not known whether "+" mutations or "— " mutations represent nucleotide additions. Accord- ingly, a double mutant of - — or + + still causes B to have no B activity, since the reading of codons starts to be out of phase with the first mutant and continues out of phase even beyond the second mu- tant. If the mutant loci are widely sepa- rated, we do not expect a double-mutant combination of + - or - ■ + to produce any B activity, since all the codons between the mutants are read incorrectly — out of phase, even though we expect those before the first and after the second mutant to be read correctly — in phase. If, however, a H or 1- mutant combination involves nearby nucleotides, it is possible that one or only a few codons — those between and including the mutants — will be read incor- rectly. The pattern reveals that any given mutant can be classified either as + or — and that only double-mutant combinations of H or (- produce some B activity — provided that the two mutants in the Bj-B;. segments are near each other. These assumptions can be tested another way. If a few incorrect codons still permit some B activity, it should be possible to increase the number of mutational errors of the same type (all - or all +) until the number of nucleotides subtracted or added equals the number in a codon. Should this point be reached, the nucleotides beyond the last mu- tated codon would be read correctly — in phase — and some B activity might be re- stored. Accordingly, phages carrying three, four, five, and even six different — (or -f) mu- tants are constructed. Some of the three or six multiple - (or + ) mutants have B activity; other combinations, like four — 438 CHAPTER 34 and one I (or four -)- and one ) mu- tants also show B activity. None is found if the mutants fail to add up to three or a multiple of three. These results demon- strate that the message from gene B is trans- lated via successive, nonoverlapping codons and that a codon is most probably three successive nucleotides. The proposed mo- lecular model of sRNA whose turn is pre- sumably made by three unpaired nucleo- tides (see Figure 33-2, p. 428) is consistent with this statement. Apparently, the triplet codon of DNA is transcribed into a com- plementary triplet codon of messenger RNA, which, in turn, is translated into an amino acid brought into position by an sRNA mol- ecule bearing a complementary triplet co- don. The triplet codon in DNA and a unique triplet in sRNA are therefore ex- pected to be identical, except that T in the former is U in the latter. In r + , genes A and B are separated by a spacer which results either in separate messenger RNAs for each gene or a non- meaningful segment between the A and B parts of a single messenger RNA. One par- ticular deletion, number 1589, removes most of region A.-, and all of regions A,;, Bu and B2. Such phage particles show no A but partial B activity. Whether or not r + makes separate messenger RNAs for the A and B genes, the spacer denoting the end of the A message and the start of the B message (or the reverse, the end of B mes- sage and the start of A message) may be absent in phages carrying deletion 1589. Consequently, this mutant may make only one continuous strip of messenger RNA containing the base complements of those parts of the A and B genes still present. This possibility can be tested as follows: single + (or — ) acridine-induced mutants in the A region are introduced by recom- bination into phages carrying deletion 1589. In each case the B gene is rendered inac- tive. In other words, B gene activity is now vulnerable to mutants in the A gene. This finding supports the view that dele- tion 1589 enables two genes to form one messenger RNA (whether or not they do so in r+ ) and that if the reading is out of phase due to a nucleotide addition or sub- traction in A. all subsequent codons — that is, those in gene B — will be misread. This result also suggests that the codons in ill messenger RNA are always read from A toward B, the order in which the genes are usually represented in genetic maps (as in Figure 26-4, on p. 343). Other evidence supports this interpretation. For example, when we make deletion 1589 phage doubly- mutant in the A region, the B gene is in- activated in some cases, but in others some B activity is detected. When + and — are assigned the mutants, we find that only -f- - double mutants in A can restore B activity; — — or -| — \- combinations cannot. More- over, as expected, the sequence -\ or - + and the distance between these two mutants make no difference. Since deletion 1589 has some B activity, it must be associated with the deletion of some multiple of three nucleotides. Al- though the codon cannot be less than three nucleotides, it can be a multiple of three if, for example, each + (or — ) mutation added (or lost) two nucleotides. In this case the codon will be six nucleotides. We can test for the size of the codon by com- bining mutant 1589 with different medium- sized deletions in the A gene. Assuming that the two breakage points involved in such deletions occur at random, then only one third of the A deletions should remove exact multiples of three nucleotides, only one sixth should remove exact multiples of six, and so on. Therefore, a test of these deletions should show one sixth which per- mit the B gene to function if the codon is six nucleotides long; one third should per- mit B to function if the codon is three nu- cleotides long. Tests show, in fact, that a Genetic Amino Acid Coding 439 little more than one third of these moderate- sized A deletions permit 1589 mutants to show B activity. Consequently, these re- sults strongly suggest that the coding unit is a triplet. What can we hypothesize about the na- ture of the spacer that normally interrupts the A and B genes in r+? If the DNA se- quence is interrupted at the ends of each gene, A and B, by short amino acid se- quences (see p. 276), transcription will be physically interrupted, thereby furnishing a starting and a stopping point for the for- mation of messenger RNA and, similarly, polypeptides. Another possibility is the oc- currence in the DNA between the A and B genes of a sequence (or a multiple) of three nucleotides whose complement in mes- senger RNA has no complement in the pre- sumed unique triplet of any sRNA. This untranslatable mRNA codon would make no amino acid sense and is, therefore, called a no sense or nonsense codon. According to this hypothesis, genes A and B normally form one continuous strip of messenger RNA, whose translation produces two sep- arate polypeptides. How many of the 64 triplet codons are nonsense? Genetic studies of the rll re- gion strongly suggest that relatively few tri- plets are nonsense. Consequently, most tri- plets probably code for amino acids, and, since only twenty kinds of amino acids com- monly occur, the same amino acid can be coded by more than one codon. Thus, we are apparently dealing in vivo with a degen- erate triplet code. If the base-pairing of sRNA with messenger RNA is strictly ac- curate— that is, exactly complementary — there will be more than twenty kinds of sRNA, several of them carrying the same amino acid. Alternately, if there are only twenty sRNA types, and the base-pairing with messenger RNA triplets is inaccurate, a given sRNA will base-pair with different (but somewhat similar) messenger RNA tri- plets. Both of these mechanisms for de- generacy may apply. In any event, most mutants involving base substitutions prob- ably produce sense — that is, code for a dif- ferent amino acid — and therefore produce missense codons. It is possible to determine the nucleotide basis for certain point mutants in the rll region.- Suppose that the DNA strand used to make messenger RNA in r+ has a G re- placed by A in a particular r point mutant. If this mutant phage does not lyse the K12 strain of E. coli because its messenger RNA, containing a U instead of a C, is abnormal, a defective r+ product results. Although 5-fluoro uracil (FU) is not mutagenic when added to the diet of K12, it can be used as a substitute for U when RNA is synthe- sized. When FU substitutes for U in mes- senger RNA, an sRNA molecule may some- times mistake it for C (see discussion of BU on p. 398). If such a mistake is made, the sRNA paired with abnormal messenger RNA will contain G and be the sRNA that transports the amino acid normally found in r+ product. Consequently, the amino acid correct for r+ will be incorporated to form some r+ product, and the host cell will lyse. Therefore, r mutants which can lyse only when FU is added most probably have G on their r+ DNA strand used to make mes- senger RNA, and C on the complementary strand. Those mutants which do not lyse in the presence of FU may have T, A, or C at this locus in the DNA strand used for transcribing messenger RNA. Using vari- ous chemical mutagens as well as FU, it is often possible to determine when T, A, or C is present in the transcribed strand. Sometimes a single bacterial mutant si- multaneously suppresses the effects of point mutants at a number of other nucleotide sites. Suppose that in some of these cases, all the suppressed point mutants have the - See S. P. Champe and S. Benzer (1962). 440 ( II \PTER 34 same triplet modified by the same base sub- stitution, resulting in the incorporation of the same incorrect amino acid into the dif- ferent polypeptide products. These effects can be suppressed by a mutant which mod- ifies the specificity of an enzyme responsi- ble for activating and attaching an amino acid to sRNA. Such a modification may sometimes cause the sRNA to transport an incorrect amino acid to the ribosome carry- ing the abnormal messenger RNA; this amino acid may be the one normally in- corporated at that position in the polypep- tide product. Consequently, mutants which make incorrect messenger RNA may still form the correct protein product, if com- pensated by the additional error of having sRNA carry a specific wrong amino acid. In a limited way, such suppressor mutants cause an alteration in the code for amino acids.3 Identification of Codons The mechanism of protein synthesis can be studied in vitro by using a suspension of ruptured cells. Such a cell-free system is prepared from E. coli plus the addition of triphosphates of the ribosides of A, G, C, and U as well as all twenty of the amino acids in their L forms. The synthesis of protein can be readily detected if one of the added amino acids is radioactive — valine, for example, which becomes incorporated into protein. This incorporation can be stopped by the addition of DNase, which halts the production of messenger RNA by destroying the DNA. In the absence of new messenger RNA, protein synthesis stops. That the DNase effect concerns the pro- duction of messenger RNA is demonstrated by the absence of valine incorporation when sRNA or ribosomal RNA is added to the Such mutants are reported by S. Benzer and S. P. Champe and by A. Garen and O. Siddiqi in Proc. Nat. Acad. Sci.. U.S.. 48:1114-1127, 1962. system and by the resumption of valine in- corporation when messenger RNA obtained from washed ribosomes is added to the sys- tem. This added messenger RNA can also come from other sources. For example, E. coli extracts can be used to synthesize hemoglobin under the direction of RNA from rabbit reticulocytes, and the RNA of coliphage f2 will stimulate amino acid in- corporation into protein, part of which at least is the coat protein of the phage.4 Using such a cell-free system derived from bacteria, we can also study whether the addition of synthetic polyribotides has any effect on protein synthesis. First, a homopolyribotide containing U is added; the polyuridylic acid causes L-phenylalanine to be incorporated into protein.5 More- over, it is found that: 1. The protein formed is poly-L-phenyl- alanine 2. No other amino acid is incorporated in substantial amounts (However, if the Mg+ + concentration is altered or if streptomycin is added, significant amounts of leucine are incorporated. The explanation for this is unknown.) 3. Phenylalanine linked to sRNA is an intermediate in this process. These results surely mean that wherever an appropriate sequence of LPs appears in nor- mal messenger RNA, the protein being syn- thesized will usually incorporate L-phenyl- alanine. This discovery is the first crack in the RNA code; in other words, this is the first determination of a sequence of mes- senger RNA nucleotides which specifies the incorporation of a particular amino acid into protein. When the synthetic polyribotide of U is mixed with the synthetic polyribotide of A in a way likely to make the strands base- 4 See D. Nathans. G. Notani, J. H. Schwartz, and N. D. Zinder (1962). rSee M. W. Nirenberg and J. H. Matthaei (1961). Genetic Amino Acid Coding 441 pair or wrap about one another, incorpora- tion of phenylalanine is partially or com- pletely reduced. Thus, the synthetic poly- mer is most effective in vitro when single- stranded/1 as is messenger RNA in vivo. How the presence of different bases in the same synthetic polyribotide affects amino acid incorporation into protein can also be studied. Using polynucleotide phosphoryl- ase, which has riboside diphosphates as sub- strate, polyribotides containing two or more different ribotides can be synthesized in vitro. Nearest-neighbor analysis of the het- eropolymer confirms that the ribotides are actually in a random linear array. The early analyses 7 were greatly expedited be- cause polyphenylalanine is insoluble in the cell-free system. In practice, then, an excess of uridylic acid was used in the synthesis of any mixed polynucleotide to obtain the later-synthesized protein as a precipitate from which the amount and kind of amino acids — in addition to phenylalanine — could be analyzed. Thus, to synthesize polyuri- dylic-adenylic acid, polyuridylic-cytidylic acid, and polyuridylic-guanylic acid, five times as much uridine diphosphate was used as the diphosphates of adenosine, cytidine, or guanosine, respectively. To make mixed polynucleotides containing UAC, UCG, or UGA, ten times as much uridine diphosphate was used as the riboside diphosphates of A, C, or G. For example, when a mixed polyribotide containing U and C is added to the cell- free system which is then tested to deter- mine whether an amino acid besides phen- ylalanine is incorporated into protein, pro- line and serine are among the amino acids incorporated. The code letters for these amino acids include, therefore, at least one ribotide of C. In the same way, we can 8 See M. F. Singer. O. W. Jones, and M. W. Nirenberg (1963). 7 By S. Ochoa and co-workers, and by M. W. Nirenberg and co-workers. also determine the effects of other mixed polyribotides on amino acid incorporation. Some amino acids such as alanine and argi- nine require the use of three different nu- cleotides for coding; thus, the coding ratio (the number of nucleotides required to code one amino acid) is at least three. No amino acid is found which requires the presence of all four types of nucleotides. From these results, it is hypothesized that triplets of nu- cleotides in synthetic messenger RNA are translated into amino acids; that is, a triplet RNA code occurs also in the in vitro stud- ies. When the proportions of uridylic acid and cytidylic acid in a mixed polynucleotide are varied, more serine than proline is incorpo- rated when there is an excess of uridylic acid. However, when the excess is cytidylic acid the reverse occurs — more proline than serine is incorporated. In terms of triplets, serine must be specified by 2U 1C and proline by 1U 2C. Note that neither the sequence of nucleotides in the triplet nor the order in which they are read is deter- mined from such results. In other words, although the messenger-RNA triplet code letters are 1U 2C for proline, we cannot say whether the sequence is UCC, CUC, or CCU. (The first and last triplets are dif- ferent, since the single-stranded messenger RNA molecule is translated in one direction only.) Starting with ribotides of U and C in the relative frequencies 5:1, the relative fre- quencies of different triplets in the synthe- sized polymer can be predicted. UUU should occur with a frequency of % times % times •;,;, or 1L' %1(J. Although three ar- rangements are possible for the code letters 2U 1C, any particular sequence should oc- cur with a frequency of % times % times ',-,, or 2%i6; any one of the three possible arrangements of 1U 2C should occur with a frequency of % times \; times %, or %16, whereas CCC should occur with a 442 ( II AIM I K 34 frequency of 54ie« These particular se- quences are. respectively, in the relative frequencies 125:25:5:1. Consequently, it a triplet code is the correct one. studies of this particular polyribotide for protein synthesis, should reveal incorporation of five times more phenylalanine than serine and twenty-five times more phenylalanine than proline. Although the results obtained using various synthetic polymers sometimes differ by a factor of two or so from those presently expected, the overall agreement is excellent and offers very strong support for a triplet RNA code. The existence of a triplet code is also supported by the finding that the messenger RNA is approximately 450 ribotides long B while the a or ft chain which it specifies in hemoglobin is 150 or so amino acids long. All the synthetic polyribotides tested thus far for messenger RNA activity in protein synthesis contain an excess of U for tech- nical reasons. As mentioned earlier, the protein product is mainly polyphenylala- nine, insoluble in the cell-free system and therefore readily collected and quantitatively analyzed for phenylalanine as well as other amino acids. Such studies reveal triplet code letters for nineteen amino acids. For exam- ple, three triplets code for leucine — 1A 2U, 1C 2U, and 1G 2U — demonstrating, as ex- pected from our previous discussion, that in vitro, at least, the code is degenerate. Degeneracy also occurs for asparagine which has 2 A 1U and 1C 1A 1U as co- dons, and isoleucine with codons 1A 2U and 2A 1U. The triplet code letters for tyrosine are 1A 2U. But is the actual sequence AUU, UAU, or UUA? Short sequences of ribo- tides (oligoribotides) can be lengthened at their nucleoside (3') ends by polynucleo- tide phosphorylase. A mixture of AUU and AAU oligoribotides (the base at the 8 As shown by T. Staehlin, F. O. Wettstein. H. Oura, and H. Noll (1964). 5' end is always written first in the se- quence) is lengthened at the 3' end with uridylic acid residues. When the length- ened, mixed polyribotide — AUUU . . . U or AAUUU . . . U — is tested for poly- peptide synthesis, it is found that phenyl- alanine and tyrosine are incorporated in significant amounts and that no significant amounts of isoleucine (which also has the code letters 1A 2U) or of asparagine and lysine (whose code letters are 2 A 1U) are incorporated. Therefore, the code sequence for tyrosine is probably AUU. Another method of attack for determining base sequence in codons makes use of the mutations causing single amino acid substi- tutions in hemoglobin (see Figure 32-7. p. 415), TMV, tryptophan synthetase, and other proteins. Those mutations occurring spontaneously or with mutagens expected to produce single base substitutions are as- sumed to involve single base changes. In TMV, a mutant causes tyrosine (AUU) to be replaced by phenylalanine (UUU); the mutant apparently causes a single base change from A to U. In tryptophan syn- thetase, a mutant replaces tyrosine (AUU) by cysteine (GUU) and presumably involves a change from A to G. In hemoglobin- Mi.,,.,,,,,, the a chain (see p. 414) has the histidine (1A 1U 1C) at position 58 changed to tyrosine (AUU). If only a single base change — from C to U — has occurred, then the codon for histidine must start with A and is either ACU or AUC. In hemoglobin Zurich, the amino acid at position 63 is changed from histidine (ACU or AUC) to arginine (1G 1C 1U). This change is prob- ably from A to G, so that the first base in the arginine codon is G, and the codon is either GCU or GUC. A continuation of this kind of analysis has made it possible '•' to assign complete U-containing nucleotide sequences to the codons for nineteen of the twenty amino acids. These sequences (listed 9 For T. H. Jukes. Genetic Amino Acid Coding 443 in the second column of Figure 34-1 ) are consistent with the triplet code letters, base sequence studies in vitro, and 87 of 93 known single amino acid substitutions as- sumed to have resulted from single base changes. The disagreements with respect to amino acid replacements are relatively few, and for the most part are probably due to an incomplete knowledge of all triplet code letters and to the inclusion of cases in which two or three base changes occurred succes- sively or simultaneously in replacing one amino acid by another. The ability of synthetic polyribotides with- out U to result in amino acid incorporation can also be studied using agents (trichlor- acetic acid, for example) that precipitate proteins otherwise soluble in the in vitro sys- tem. When homopolyribotides of A, C, or G and mixed polymers with these bases are synthesized and tested, a large number of new triplet code letters without ITs are found. For example, poly A makes poly- lysine; poly C makes polyproline. Guanine- rich polynucleotides do not work well, prob- ably because of the secondary structure due to guanine-guanine interactions. Based on the nucleotide sequences given to the U- containing codons and using sequences which will not duplicate those given to the codons of other amino acids, the base sequences in these new triplets without U's are assigned and listed in the third column of Figure 34-1. In studying the incorporation of amino acids into protein in vitro, one must use very Amino Acid U-containing Non U-containing Shared codons1 codons2 doublets Ala CUG CAG, CCG OG Arg GUC GAA, GCC G'C Asn UAA, CUA CAA •AA, C'A Asp GUA GCA G«A Cys GUU Glu AUG AAG A*G Gin AGG, AAC Gly GUG GAG, GCG G'G His AUC ACC A-C lieu UUA, AAU Leu UAU, UUC, UGU U'U Lys AUA AAA A«A Met UGA Phe UUU Pro cue CCC, CAC OC Ser CUU ACG Thr UCA ACA, CGC •CA Try UGG Tyr AUU Val UUG 'Sequence proposed by T. H. Jukes. ^Sequences given are fitted to those in footnote 1 , or ovoid duplicc tion of a sequence for another amino acid. figure 34-1. Tentative in vitro messenger RNA codons for amino acids. (After Wahba, A. J., et al., 1963; see reference at end of chapter. See also M. R. Bernfield and M. W. Nirenberg, Science, 147:479-484, 1965.) Ill ( IIM'TER 34 long oligoribotides, for example, a chain of 500-1000 uridylic acids. Although poly U greatly stimulates phenylalanine incorpora- tion into protein, the single trinucleotide UUU does not. Recall, however, that early steps in protein synthesis require the activa- tion and attachment of the amino acid to a specific sRNA molecule. This "charged" sRNA binds to the ribosome and, as directed by the messenger RNA, is incorporated at the end of the growing peptide chain. Poly U causes Phe-sRNA to be bound to ribo- somes; other polynucleotides cause other specific charged sRNAs to be bound. One can synthesize or isolate oligoribo- tides and test them for their in vitro ability to bind specifically charged sRNAs to ribo- somes.1" (According to convention, a tri- ribotide of U with a 3'-terminal phosphate is designated UpUpUp and one with a 5'- terminal phosphate, pUpUpU.) When pUpUpU, pApApA, and pCpCpC are tested, they are found to direct the binding of Phe-, Lys-. and Pro-sRNA, respectively; dinucleotides have no effect. Moreover, trinucleotides with 5'-terminal phosphate are more active than those with no terminal phosphate, and trinucleotides with 2'-(3')- terminal phosphate are inactive. From other work 2U 1G is known to be a code word for valine. The order of the bases can be investigated using poly UG, dinucleotides, the trinucleotide GpUpU, and its sequence isomers UpGpU and UpUpG. The binding of C14-Val-sRNA to ribosomes is found to be directed both by poly UG and GpUpU but not by UpGpU, UpUpG, or dinucleotides. GpUpU has no effect upon the binding of sRNAs, corresponding to 17 other amino acids, to ribosomes. There- fore, we conclude from these results that a code word for valine is GpUpU, and we predict that a GUUGUUGUU . . . GUU polymer will stimulate only valine incor- 10 See M. Nirenberg and P. Leder (1964). and P. Leder and M- W. Nirenberg (1964). poration into protein. Similar work showed that UpUpG is a code word for leucine and possibly UpGpU a code word for cysteine. Although there will undoubtedly be cor- rections and additions to the codons in Fig- ure 34-1 (some contradictory base sequence results are obtained using the different in vitro methods described ), examination of the codons listed reveals a common feature to some of the degeneracy already detected. For example, two of the codons for leucine have U at both ends. In other words, they share the same doublet, so that their codons can be written U • U, in which • can be A or G. Although the base sequences in alanine's two codons without U are postu- lated, both contain a C and a G, as does the U-containing sequence, so that one can refer to a C • G shared doublet, in which • probably can be U, A, or C. These and other doublets are listed in the figure. The meaning of such shared doublets in the de- generate in vitro RNA code is not yet clear, nor is it known to what extent triplets with- out U code in messenger RNA in vivo. That the frequency of some of the amino acids in protein remains nearly constant A 4- T when there are large shifts in the — — !— — C -f- G ratio is evidence for the existence of degen- eracy in vivo. Leu-sRNA of E. coli can be separated into three types, each with dif- ferent coding properties in vitro.11 The first type responds preferentially to poly UC, the second type responds to poly U and copoly- mers rich in U (including poly UC), and the third responds preferentially to poly UG. The discovery that leucine is carried by dif- ferent sRNAs provides an explanation for the observations in vitro that the coding unit for leucine is degenerate ( at least four dif- ferent triplets serve to encode it ) and that the UUU codon is ambiguous (since it is a codon for both leucine and phenylalanine). Assuming that there is only one DNA locus 11 See G. von Ehrenstein and D. Dais (1963). Genetic Amino Acid Coding 445 per type of sRNA molecule, the finding '- of approximately forty sites in E. coli DNA which are complementary to sRNA indicates not only the presence of degeneracy at this level but the extent to which it occurs. Al- ready twenty-nine specific sRNAs for six- teen amino acids have been detected in E. coli.li Although DNA and the polypeptides it specifies are both linear, it is important to determine whether the exact linearity of the polypeptides is dependent upon the exact linearity of the DNA; that is, whether coli near ity exists. This possibility can be tested using ten <£T4 mutants that produce incomplete head protein molecules. These mutants map in a linear sequence, as deter- mined by recombination studies. When the head protein of each mutant is analyzed, the length of the portion of the molecule made is exactly proportional to the map distance from one end of the gene. This finding 14 is proof of colinearity. !-By H. M. Goodman and A. Rich (1962), and D. Giacomoni and S. Spiegelman (1962) (see reference on p. 433). 13 See J. Goldstein, T. P. Bennett, and L. C. Craig (1964). 14 By A. S. Sarabhai, A. O. W. Stretton. and S. Brenner (1964); see C. Yanofsky, B. C. Carlton. J. R. Guest, D. R. Helinski, and U. Henning (1964), and M. E. Reichmann (1964). Despite the degeneracy and ambiguity noted, is the code basically the same for all organisms; that is, is the code essentially uni- versal? It was already mentioned that some- thing very similar to rabbit hemoglobin can be synthesized in a cell-free system derived partly from rabbit reticulocytes and partly from E. coli. As also mentioned, RNA iso- lated from phage f2 directs the synthesis of its coat protein in extracts of E. coli. This RNA also leads to the synthesis of f2 coat protein in extracts of Euglena gracilis. The DNA from the animal viruses polyoma and vaccinia is infective in competent Bacil- lus subtilis; that is, mixing the virus DNA with the bacteria produces intact virus particles which can infect the normal animal host. In certain animal cell-free systems which are stimulated by exogenous RNA messages, synthetic polynucleotides have many of the incorporation properties that they have in bacterial cell-free systems. Fi- nally, a marked correlation exists between C + G content and the percentages of cer- tain amino acids incorporated into protein in a variety of organisms. All these results support the hypothesis that even if there are mutational modifications, only one basic code for polypeptide synthesis exists in all present-day organisms. SUMMARY AND CONCLUSIONS In vivo study of the rll region of <£T4 reveals that the genetic code for amino acids is read in one direction — probably from one fixed point of messenger RNA— very likely in successive triplets. Such work suggests that the code is degenerate and almost all of the possible codons make sense. Studies of polypeptide synthesis in vitro using natural and synthetic messenger RNA, of mutants involving single amino acid substitutions, and of sRNA binding to ribo- somes in vitro support these hypotheses. Such work also permits the assignment of base sequences to the triplets which code in vitro. In vivo, DNA and the polypeptide it specifies are colinear; the RNA code is basically universal. 146 CHAPTER 34 REFERENCES Campbell, A., "line Structure Genetics and its Relation to Function," Ann. Rev. Mi- crobiol., 17:49 ''it. 1963. Champe, S. P.. and Benzer, S.. "Reversal of Mutant Phenotypes by 5-Fluorouracil : An Approach to Nucleotide Sequences in Messenger-RNA," Proc. Nat. Acad. Sci., U.S.. 48:532-546, L962. Crick. F. H. C. "The Genetic Code," Scient. Amer.. 207:66-74, 176, (Oct.) 1962. Crick, F. H. C "The Recent Excitement in the Coding Prohlem," Progr. Nucleic Acid Res.. I : 163-217. 1963. Fraenkel-Conrat, H.. "The Genetic Code of a Virus," Scient. Amer., 211 (Oct.) : 47- 54, 142. 1964. Goldstein, J., Bennett, T. P., and Craig, L. C, "Countercurrent Distribution Studies of E. coli sRNA." Proc. Nat. Acad. Sci., U.S., 51:119-125, 1964. Goodman, H. M., and Rich, A., "Formation of a DNA-Soluble RNA Hybrid and its Relation to the Origin, Evolution, and Degeneracy of Soluble RNA," Proc. Nat. Acad. Sci.. U.S., 48:2101-2109, 1962. Grunberg-Manago. M.. "Polynucleotide Phosphorylase," Progr. Nucleic Acid Res., 1: 93-133, 1963. Jukes, T. H., "Coding Units and Amino Acid Substitutions in Proteins," pp. 485-497, in Informational Macromolecules, Vogel, H. J., Bryson, V., and Lampen, J. O. (Eds.), New York: Academic Press, 1963. Leder, P., and Nirenberg, M. W., "RNA Codewords and Protein Synthesis, III. On the Nucleotide Sequence of a Cysteine and a Leucine RNA Codeword," Proc. Nat. Acad. Sci., U.S., 52:1521-1529, 1964. Nathans, D., Notani, G., Schwartz, J. H., and Zinder, N. D., "Biosynthesis of the Coat Protein of Coliphage /2 by E. coli Extracts," Proc. Nat. Acad. Sci., U.S., 48:1424- 1431, 1962. Nirenberg, M. W., "The Genetic Code: II," Scient. Amer., 208 (March) : 80-94, 190, 1963. Nirenberg, M.. and Leder, P., "RNA Codewords and Protein Synthesis," Science, 145: 1399-1407, 1964. Nirenberg, M. W., and Matthaei, J. H., "The Dependence of Cell-Free Protein Syn- thesis in E. Coli upon Naturally Occurring or Synthetic Polyribonucleotides," Proc. Nat. Acad. Sci., U.S., 47:1588-1602, 1961. Reichmann. M. E.. "The Satellite Tobacco Necrosis Virus: A Single Protein and its Genetic Code," Proc. Nat. Acad. Sci., U.S., 52:1009-1117. 1964. Singer, M. F., Jones, O. W., and Nirenberg, M. W., "The Effect of Secondary Struc- ture on the Template Activity of Polyribonucleotides," Proc. Nat. Acad. Sci., U.S., 49:392-399, 1963. Synthesis and Structure of Macromolecules, Cold Spring Harb. Sympos. Quant. Biol., 28, 1964. von Ehrenstein, G., and Dais, D., "A Leucine Acceptor sRNA with Ambiguous Coding Properties in Polynucleotide-Stimulated Polypeptide Synthesis," Proc. Nat. Acad. Sci.. U.S., 50:81-86. 1963. Wahba, A. J., Gardner, R. S., Basilio, C. Miller, R. S., Speyer, J. F., and Lengyel. P.. "Synthetic Polynucleotides and the Amino Acid Code. VIII," Proc. Nat. Acad. Sci.. U.S.. 49: i 16-122, 1963. Genetic Amino Acid Coding 447 Speakers {I. to r.) M. W. Nirenberg, F. Lipmann, and S. Ochoa at a sym- posium on the RNA code held January, 1962, at Indiana University. Weisblum, B., Gonano. F., von Ehrenstein. G.. and Benzer, S.. "A Demonstration of Coding Degeneracy in the Synthesis of Protein," Proc. Nat. Acad. Sci., U.S., 53:328-334, 1965.' Woese, C. R., Hinegardner, R. T., and Engelberg, J., "Universality in the Genetic Code," Science, 144:1030-1031, 1964. Yanofsky, C, Carlton, B. C, Guest, J. R., Helinski, D. R., and Henning, U., "On the Colinearity of Gene Structure and Protein Structure," Proc. Nat. Acad. Sci., U.S., 51:266-272, 1964. See Supplement X. Other references can be found at the end of Dr. Crick's Nobel Prize Lecture. QUESTIONS FOR DISCUSSION 34.1. Do you expect the genetic code for amino acids to be the same in all free- living organisms? Explain. 34.2. Compare the replication of an RNA virus with the replication of a polypeptide chain. 34.3. Prepare a report on advances in our understanding of the genetic code since the present account was written (November 1964). 34.4. What evidence can you present that the attachment of messenger RNA to the ribosome does not involve extensive complementary base pairing? 34.5. Give evidences that messenger RNA is single-stranded. I »S CHAPTER 34 34.6. What raw materials arc needed to make a mixed polyribonucleotide in the ab- sence ol a primer'.' In the presence of a primer? 34.7. Do you suppose the first genetic code was or was not degenerate? Explain. 34.8. Work out the relative frequencies of the triplet code letters, UUU, UUA, AAU, UAC, AAA, CCC, in the specific sequences given from a polymer synthesized from ribotides of U. A, and C in the relative amounts of 6, L, and 1, respec- tively. 34.9. Using large quantities of riboside diphosphates of A, U, G, and C in the rela- tive proportions of 4, 3, 2, 1, and polynucleotide phosphorylase to synthesize a mixed polyribotide, give the proportion of sequences in the polyribotide for the following types (all read in one direction only): (a) doublets AU; AC; CA (b) homotriplets; heterotriplets (c) quadruplet AUCG 34.10. Make a list of the minimal requirements for the functioning and reproduction of the simplest free-living organism you can imagine; estimate the minimum number of nucleotides required to perform these functions assuming the genetic material is RNA; assuming it is DNA. Compare your estimates with the num- ber of nucleotides in TMV and X174. What are your conclusions? 34.11. Devise experiments which permit the collection of essentially pure sRNA carry- ing phenylalanine; sRNA carrying lysine. 34.12. Cysteine, while still attached to its normal sRNA type, is converted to alanine by reduction with Raney Nickel. Using synthetic polyribotides, design a di- rect test of the hypothesis that sRNA functions as an adapter in specifying the fit of amino acids on a template. 34.13. How can you explain the observation by L. Grossman that ultraviolet irradia- tion of polyuridylic acid not only results in a marked depression in incorpora- tion of phenylalanine in an in vitro protein synthesis but is accompanied by an increased incorporation of serine? What relation has your explanation to the observation that polyuridylic acid can normally code not only for phenylalanine but for leucine? 34.14. How can you explain the observation by G. E. Magni (see reference on p. 389) that normad mutations (reversions) for certain UV-induced mutants in yeast occur six to twenty times more frequently during meiosis than mitosis? 34.15. A number of rll point mutants can be classified as resulting from transitions A:T-» G:C or G'-C -* A:T according to their reversibility after treatment with various chemical mutagens. Addition of FU to the nutrient medium does not produce the r+ phenotype in any of the mutants which supposedly carry the (,:(' pair at the mutant DNA site. On the other hand, some r+ phenotype is produced by FU in 17 of the 46 mutants presumed to carry A:T at the mutant site. What conclusions can be drawn? 34.16. What bearing does the observation that a hemoglobin chain is always synthe- sized beginning at the N-terminus have upon the alternatives of one-complement and two-complement transcriptions? 34.17. Of the possible 64 unidirectionally-read triplets using AUGC, how many have one or more U's? No U's? Chapter 35 REGULATION OF GENE SYNTHESIS A' s revealed by in vitro studies, DNA synthesis in biological systems requires the follow- ing: primer-template DNA; the nucleoside 5'-triphosphates of A, T, C, and G; Mg+ + ions; and DNA polymerase in an aqueous solution of proper pH and temperature. In an extended in vitro synthesis of DNA using E. coli DNA polymerase, we permit the re- action to proceed in a largely uncontrolled, unregulated manner until the supply of one of the raw materials is exhausted or until some other factor becomes limiting. The synthesis of DNA in vitro can be controlled, however, by changing one or a combination of the required factors. For example, we may choose to omit one of the triphosphates containing a base present in the primer- template. Reducing the amount of such a triphosphate or utilizing one which contains a base analog can control the rate and/or amount of DNA synthesized. The synthetic reaction can be slowed down or even partly reversed by excessive additions of pyrophos- phate. In other words, there are a large number of ways by which the biochemist can regulate the synthesis of DNA in vitro. Such knowledge should be quite valuable in helping us answer the question: In pres- ent-day organisms how is gene synthesis regulated in vivo? This question presupposes that DNA syn- thesis in vivo is regulated, and ample evi- dence, some of it already presented, sup- ports such a view. The regulation of DNA 449 synthesis at the cellular level is revealed by observations that DNA synthesis occurs dur- ing interphase and ceases during nuclear division. Evidence of regulation at the genomic level is provided by the fact that when DNA synthesis stops, the nucleus is euploid for DNA (± about 10%) — even if the nucleus fails to divide and comes to contain a multiple (polyploid or polynemic) genome content. DNA synthesis is also regulated chromosomally since largely het- erochromatic chromosomes replicate at a different time than largely euchromatic ones, and intrachromosomally since the hetero- chromatic and euchromatic regions within a chromosome are synthesized at different times during interphase. It should be noted that the occurrence and amount of "natural dAT" in different crabs is probably gene controlled.1 DNA Synthesis in Uninfected and Phage-lnfected Bacteria Let us discuss further the regulation of gene synthesis by exploring the biochemical path- ways (summarized in Figure 35-1 ) so impor- tant in the synthesis of the four usual types of deoxyriboside triphosphates in uninfected E. coli. Since the hypothesis that genes specify all protein synthesis in an organism is now generally accepted, whenever these reactions involve enzymes, the control of gene synthesis by gene action is also in- volved. In the presence of deoxyribosidase or re- ductase, the riboside diphosphates of A, G, C, and U are converted to the corresponding deoxyriboside diphosphates by removal of the O at the 2' position.2 The energy source for this reaction is ATP. The dTP is syn- thesized from d\JP by adding a methyl group at the 5 position in the presence of thymidy- late synthetase. (A thymine-requiring strain of E. coli lacking this enzyme is known. !See M. Smith (1963). -' See A. Larsson (1963). 450 ( HAM I.R 35 dHMCP i CPP _J\ dCP k dTP i upp _JTl i dUP^_ dHMCP kinase deoxyribosidase dHMCPP kinase dHMCPPP ATP dCP kinase • ATP dAP kinase ATP dGP kinase | ATP dTP kinase dCPP 3 . dAPP ATP deoxyribosidase ATP dUPP kinase _»» dUPPP ATP dUPPPase glucosyl transferases t glucosyl DNA d — deoxyribo G = guanosine U = uridine 2 = hydroxy meth ylase HMC = hydroxyn lethyl cytidine A = adenosine ATP = adenosine 5'triphosphole 3 = dCPPPdCPPase C cytidine T = thymidine 1 = thymidylate synthetase figure 35-1. Enzymatic pathways leading to DNA synthesis in E. coli. Interrupted arrows denote reactions occurring and wavy line to reaction blocked in cells infected with T-even phages. {After M. J. Bessman, 1963.) Such T-requiring E. coli can, however, syn- thesize thymidylic acid — thymidylate — when infected with c/>T2 because, as was proven, T2 carries the information to make a chro- matographically different but similar-acting thymidylate synthetase. ) The c/UP involved has at least two sources. One source, cVUPPP, loses pyrophosphate through the action of dUPPPase; another, c/UPP, loses one phosphate (orthophosphate) and be- comes i/UP. The deoxyriboside 5'-mono- phosphates of C, A, G, and T are phos- phorylated to the 5'-diphosphate condition by specific phosphorylating enzymes — nu- cleoside monophosphate kinases — in the fol- lowing reaction: i-yp . AT-p nucleoside monophosphate kinase c/XPP + ADP ("adenosine 5'-diphosphate) where X is the nucleoside of C, A, G, or T. (It should be noted that the dGP, c7TP, and <7CP kinases formed in E. coli infected with virulent phage are different from those produced in uninfected cells.) The deoxy- riboside 5'-diphosphates produced are most probably converted to the 5'-triphosphate condition by means of other specific phos- phorylating enzymes — nucleoside diphos- phate kinases. The reaction is as follows: <7XPP + ATP iVXPPP + ADP. nucleoside diphosphate kinase Regulation of Gene Synthesis 451 This formula completes a summary of the pathways involved in producing the deoxy- riboside triphosphates required for replica- tion of E. coli DNA. In the preceding discussion we noted that an infecting virulent phage carries spec- ifications for the manufacture of a specific thymidylate synthetase and probably for specific nucleoside monophosphate kinases. This conclusion suggests that virulent phages carry instructions for making a number of specific proteins. Within two minutes after injection of T phage DNA, phage specific RNA appears; within four minutes phage specific proteins appear; and within six min- utes phage DNA is synthesized — five times faster than is DNA in uninfected cells. Host DNA is destroyed soon after infection by T-even phages. Although the mechanism is not completely clear, it is thought to in- volve a new DNase that appears after phage infection. Roughly half an hour after in- fection 100 to 200 new phages are produced and liberated by lysis. These activities lead us to hypothesize that, after virulent phage infection, all DNA and messenger RNA syn- thesis in the bacterial cell is directed by the phage DNA. That E. coli DNA contains C, whereas the T-even phages contain 5- hydroxymethyl cytosine {HMC) to which glucose is attached in different ratios for dif- ferent T-even phages, suggests another test of this hypothesis. Within several minutes after infection with T-even phage, dCP is converted to rfHMCP by a hydroxymethylase. This en- zyme is newly produced, since uninfected cells or cells infected with T5 (which has no HMC in its DNA) have no hydroxy- methylase activity. Through the action of kinase — also produced only in T-even in- fected cells — dHMCP is phosphorylated to dHMCPPP. All the nucleoside monophos- phate kinase activity for G, T, and HMC in phage-infected cells may be due to a sin- gle new enzyme. In T2-infected cells an- other new enzyme appears which splits pyrophosphate away from c/CPPP and ortho- phosphate away from dCPP converting these to dCP which, as described, is the substrate for making t/HMCP. The dephosphorylat- ing activity of this enzyme, dCPPP-dCPPase, is 60 times greater than the kinase phos- phorylating activity and has no effect upon dHMCPPP. Such a mechanism seems to be adequate in excluding C from T-even phage DNA. The biochemical pathways described in uninfected and T-even infected cells result in the synthesis of dHMCPPP, dAPPP, dGPPP, and dTPPP— the raw materials re- quired for DNA polymerase action in the synthesis of T-even phage DNA. As is true for other phage-induced enzymes, DNA polymerase shows a high level of activity in phage-infected cells. Is a new DNA poly- merase developed in response to T2 phage infection, a different one from the E. coli DNA polymerase formed in uninfected cells? The DNA polymerases in samples from uninfected and T2-infected cells are found to differ in antigenic properties, mi- gration during chromatography, and inac- tivation sensitivity to specific chemicals. Moreover, E. coli DNA polymerase can use native, double-stranded DNA as primer- template, whereas the polymerase from the infected cells is virtually inert. Given single- stranded primer-template, E. coli polymer- ase can increase the amount of DNA ten- to twentyfold, whereas the polymerase from infected cells produces less than 100% in- crease. We may conclude, therefore, that a new DNA polymerase, T2 DNA polymer- ase, is formed in T2-infected E. coli. As mentioned, T2, T4, and T6 have dis- tinctly different glucose distributions in the HMC of their DNAs. The glucose residues are added to HMC in the DNA polymer through the action of enzymes called glu- 452 CHAPTER 35 cosyl transferases. These enzymes transfer glucose from uridine diphosphate glucose (UPP glucose) — not shown in Figure 35- 1 — to HMC residues in DNA. Such en- zymes are not found in uninfected or T5- infected cells and are clearly phage-induced. Note again that the glucosyl transferases act on polydeoxyribotid.es. We have seen, therefore, that after T- even phage infection, new enzymes are in- duced to carry out syntheses unique to viral DNA production and to neutralize host en- zymes which would be antagonistic to this process. New enzymes are also known to supplement the action of the host's enzymes to speed up synthesis of viral DNA. (For example, phages induce production of a dif- ferent thymidilate synthetase than their host's. ) These results not only indicate that viral DNA leads to the destruction of the host's DNA, but they provide us with some insight as to where DNA synthesis is regu- lated genetically in the T-even phage-/:, coli system. Variation in Genetic Nucleic Acid Components The base-ratio of double-stranded DNA containing A, T, C, and G can be estimated from its buoyant density in the ultracentri- fuge and from its denaturation (melting) temperature. A discrepancy in the base- ratios found by these methods :t for the DNA of phage PBS 1 (and also PBS 2) is explained by the finding that all the T in the phage DNA is replaced by U, the base composition frequencies being A = 0.359, U = 0.359, G = 0.134, and C = 0.147. The host of this phage, Bacillus subtilis, has T not U in its DNA. Appar- ently information which the phage carries in its own genome incorporates i i r 35 corporated into the progeny of R17.'" an RNA phage. Genetic DNA rran8criptioP Genetic RNA The mate-killer (ma) particle in Parame- cium, like the similar bacterial endosymbi- otes lambda and kappa (see p. 374), de- pends upon the micronuclear genes of its host for its maintenance. In this case, two unlinked dominant genes, M, and Af.. are involved, either one independently capable of supporting growth and replication of the mu particles. Sensitive, non-mate-killer in- dividuals containing either A-/,, M-, or both do not spontaneously generate mu particles, so these genes do not form mu particles di- rectly. If after conjugation of a mate-killer, its M dominant genes are replaced and the resulting exconjugant is m, m, /??•_. raL., the mu particles (which are visible and contain DNA) and the mate-killer phenotype are lost some 8 to 1 8 fissions later. This de- layed loss of mu particles is abrupt, since a cell has either a large number of particles or none. Consequently, it is suggested that the M genes control mu particle existence by their gene products, called metagons. Numerous tests of this hypothesis yield confirmative results. It has also been found in Paramecium that: 1 . A single metagon is sufficient to sup- port numerous mu particles 2. Metagons rarely, if ever, replicate 3. Metagons can be transferred via a cy- toplasmic bridge from one member of a conjugating pair to the other 4. In the absence of an M gene, the met- agons are diluted in successive fissions 5. Normally, about 1,000 metagons are present in each mate-killer individual. Moreover, since ribonuclease destroys them, RNA is an essential constituent of metagons. Mu particles are destroyed exactly one fis- 10 See D. B. Ellis and W. Paranchych (1963). sion after the metagons are eliminated from a cell by ribonuclease. Metagons can be synthesized two fissions after ribonuclease treatment, provided an M gene is present." Subsequent evidence shows that metagons are messenger RNA with a high proportion of G -f- C. Accordingly, the existence of an endosymbiont is regulated by its host's messenger RNA. Not only can other para- mecia be infected with metagons, but the very different protozoan Didinium can ac- quire both metagons and mu by eating par- amecia which contain them. Metagonic RNA recovered from Didinium or parame- cia can hybridize with DNA from M-con- taining paramccia and, to a lesser extent, m-containing paramecia but not with DNA from Didinium. Therefore, we conclude that Didinium contains no M genes. Never- theless, the metagons not only persist but multiply in Didinium.1- These results sug- gest that in Paramecium, metagonic RNA is somehow inhibited from replicating al- though it persists as a messenger for a rather long time. In this respect the RNA meta- gon resembles the DNA in an abortive transduction. In Paramecium, the RNA metagon is behaving like an RNA virus generated but incapable of replication; in Didinium, like one incapable of being gen- erated but capable of being replicated. It is important for us to learn as much as pos- sible about the nature and origin of the met- agon-replicating enzyme and the mechanism that inhibits or prevents metagon replica- tion under particular circumstances. A host's genetic control of RNA replication may sometimes involve the action of inter- ferons— proteins (presumably synthesized through the intermediary functioning of mes- senger RNA) which prevent replication of certain viruses.1'5 11 The preceding discussion is based upon work of I. Gibson and G. H. Beale ( 1963). 12 See I. Gibson and T. M. Sonneborn (1964). 1!See R. Z. Lockhart, Jr. (1964). Regulation of Gene Synthesis 455 The preceding evidence indicates that in vivo transcription of genetic DNA can pro- duce genetic RNA. We have already noted that RNA can produce DNA by transcrip- tion in vitro (p. 289). The Rous sarcoma virus (RSV) infects chick embryo cells; this is an RNA virus, and infectious RNA can be isolated from Rous sarcoma cells. When DNA synthesis is inhibited soon after ex- posure of cells to RSV, the production of progeny virus is prevented; if the inhibition occurs later, however, virus progeny are pro- duced. RNA-DNA hybridization experi- ments reveal that upon infection with RSV, the chick cell synthesizes DNA homologous to the viral RNA.14 This DNA is not pres- ent before infection and is not homologous to RNA unrelated to RSV RNA. It has been suggested that this new DNA is the provirus stage of RSV and is comparable to the prophage stage of lambda. This is ap- parently a case of in vivo transcription from genetic RNA to genetic DNA, and perhaps also of transcription in the reverse direction. 14 See H. M. Temin (1964). SUMMARY AND CONCLUSIONS Some of the biochemical pathways leading to the synthesis of bacterial and phage DNA are outlined. These pathways involve a large number of specific enzymes. Since enzymes are directly specified by gene action, we have gained some insight into the genetic control of gene synthesis. The occurrence in DNA of bases other than A, T, C, and G has given, or is expected to give, further insight into this matter. Further investigation of factors which determine polynucleotide composition, length, and single- or double-strandedness are also needed before we can fully understand how genetic nucleic acids are regulated in vivo. In vivo, genetic DNA can be transcribed to genetic RNA (metagons) and genetic RNA (Rous sarcoma virus) can be transcribed into apparently-genetic DNA. REFERENCES Bessman, M. J., "The Replication of DNA in Cell-Free Systems," Chap. I, pp. 1-64, in Molecular Genetics, Part I, Taylor, J. H. (Ed.), New York: Academic Press, 1963. Cairns, J., "The Chromosome of Escherichia coli," Cold Spring Harb. Sympos. Quant. Biol., 28:43-47, 1964. Cohen, S. S., "On Biochemical Variability and Innovation," Science, 139:1017-1026, 1963. Ellis, D. B., and Paranchych, W., "Synthesis of Ribonucleic Acid and Protein in Bac- teria Infected with an RNA Bacteriophage," J. Cell. Comp., Physiol., 62:207-213, 1963. Fong, P., "The Replication of the DNA Molecule," Proc. Nat. Acad. Sci., U.S.. 52: 641-647, 1964. Gibson, I., and Beale, G. H., "The Action of Ribonuclease and 8-Azaguanine on Mate- Killer Paramecia," Genet. Res. (Camb.), 4:42-54, 1963. Gibson, I., and Sonneborn, T. M., "Is the Metagon an m-RNA in Paramecium and a Virus in Didinium?," Proc. Nat. Acad. Sci., U.S., 52:869-876, 1964. Holland, J. J., "Depression of Host-Controlled RNA Synthesis in Human Cells During Poliovirus Infection," Proc. Nat. Acad. Sci., U.S., 49:23-28, 1963. 45(J CHAPTER 35 Kornberg, A.. Enzymatic Synthesis of DNA, New York: J. Wiley & Sons. 1962. I arsson. A., "Enzymatic Synthesis of Deoxyrihonuclcoticles. III. Reduction of Purine Ribonucleotides with an Enzyme System from Escherichia coli B," J. Biol. C'hem.. 238:3414- 3419. 1963. Lockhart, R. /... Jr.. "The Necessity for Cellular RNA and Protein Synthesis for Viral Inhibition Resulting from Interferon," Biochem. Biophys. Res. Commun., 15:513- 518. 1964. Reddi, K. K... "Studies on the Formation of Tobacco Mosaic Virus Ribonucleic Acid, III. Utilization o\ Ribonucleosides of Host Ribonucleic Acid." Proc. Nat. Acad. Sci.. U.S.. 50:419-425, 1963. Roscoe, D. H.. and Tucker, R. G., "The Biosynthesis of a Pyrimidine Replacing Thy- mine in Bacteriophage," Biochem. Biophys. Res. Commun.. 16:106-110, 1964. Smith. M.. "Deoxyribonucleic Acids in Crabs of the Genus Cancer," Biochem. Biophys. Res. Commun.. 10:67-72, 1963. Stevens, J. G., and Groman, N. B., "A Nucleic Acid Analogue Dependent Animal Virus," Biochem. Biophys. Res. Commun., 10:63-66, 1963. Takahashi. I., and Marmur, J., "Replacement of Thymidylic Acid by Deoxyuridylic Acid in the Deoxyribonucleic Acid of a Transducing Phage for Bacillus suhtilis," Nature, London, 197:794-795, 1963. Temin, H. M., "Homology between RNA from Rous Sarcoma Virus and DNA from Rous Sarcoma Virus-Infected Cells," Proc. Nat. Acad. Sci., U.S., 52:323-329, 1964. QUESTIONS FOR DISCUSSION 35.1. Why should Kornberg and his associates fail to observe extended DNA synthesis in extracts of T2-infected cells to which the deoxyriboside 5'-triphosphates of A, T, G, and C had been added? How would you proceed to obtain the desired synthesis? 35.2. What have we learned about the genetic control of nucleic acid synthesis from in vitro studies? 35.3. Do you suppose there is a genetic control over the tautomeric states a base in DNA may exhibit? Justify your opinion. 35.4. Whereas laboratory synthesis of nucleosides can produce a mixture of « and ft configurations (which differ in the way the parts are folded or pointed relative to each other), all the nucleosides in DNA have the ft configuration. How can you explain this difference? 35.5. Outline experiments designed to throw light upon the genetics of ( a ) DNA polymerase ( b ) RNA synthetase (c) RNA polymerase (d) polynucleotide phosphorylase 35.6. Which do you think came first in evolution, biochemical pathways leading to DNA or to RNA synthesis? Explain. 35.7. Many mutants induced by nitrous acid in TMV show a mutant phenotype but no change in the amino acid sequence of their protein coat. Suggest ways in which such mutants produce their phenotypic effects. 35.8. Does the work with metagons suggest an origin for viruses? Explain. Chapter 36 REGULATION OF GENE ACTION-OPERONS E xtensive study of any organ- ism reveals a large number of alternative traits with a genetic basis. Some of these alternatives result from the presence or absence of ge- netic material (for example, in Paramecium "cytoplasmic DNA" can depend upon the presence of kappa); other alternative traits involve the relocation of genetic material (for example, changes in episomal state or the inversion of a chromosomal segment). But the presence, absence, and location of genetic material do not describe the mecha- nism operating on the affected cell or organ- ism, or the ways genetic material performs a function. We are, therefore, especially interested in studying those alternative traits resulting from some action by or involving genetic material. Self-replication, one action typi- cal of what has been defined as genetic material, must have some phenotypic conse- quences due to the removal of gene pre- cursor material from the pool of metabolic substances and to the presence of new ge- netic material. (Gene control mechanisms which act via gene replication were consid- ered in the previous chapter.) Evidence for the occurrence of gene ac- tion without gene replication is provided in numerous cases, including abortive trans- duction and highly functional cells which never divide again (neurons, for example). We already know that whenever phenotypes are dependent upon protein synthesis, gene 457 action requires the formation of messenger RNA. Using DNA to make complemen- tary DNA — that is, for gene replication — may inhibit its use in making comple- mentary messenger RNA — that is, for gene functioning via polypeptide synthesis. This may be true even if DNA polymerase uses the major groove and RNA polymerase the minor groove of double-helix DNA. That gene action is sometimes controlled by ge- netic means was demonstrated by finding regulator genes (such as Activator in maize, p. 385). We should not exclude the possi- bility that DNA genes can produce pheno- typic effects using mechanisms other than DNA replication and messenger RNA for- mation. Gene action can be controlled nongenet- ically. A series of enzymatic reactions is usually required to produce a particular metabolic end product. In many cases the end product inhibits the functioning of one of the first enzymes in the pathway. Such end product inhibition of an enzyme, very widespread in bacteria, provides immediate and sensitive control of the rate of synthesis of many metabolites. Enzyme inhibition by end product is one example of controlling gene action by a feedback mechanism. The possibility also exists that gene action can be regulated more directly — at a stage prior to protein synthesis. Recall (p. 357) that the Lac segment of the E. coli genetic map contains three re- combinationally separate genes. The y + gene specifies the structure of the enzyme (3-galactoside permease; z + specifies the structure of the enzyme (H-galactosidase (certain alleles of z cause the synthesis of a modified, enzymatically inactive protein, called Cz. identified by its specific antigenic characteristics); the third gene, /' + , specifies the synthesis of a repressor substance which prevents y+ and z + from producing per- mease and galactosidase. In the presence of lactose (which supplies the substrate 158 ( IIAPTFR 36 upon which those enzymes act), however, the repressor substance made by i ' is in- activated, so that the formation of enzymes b\ v • and z ' becomes possible. Lactose functions as an inducer. Therefore. E. coli of genotype y z •' i + cannot produce per- mease or galactosidase constitutively ( in the absence of lactose) but can do so induc- tively ( in the presence of lactose ) . This example serves as a model to explain the genetic basis for many cases of induced en- zyme formation. In this instance, the feed- back system controls the production but not the activity of certain enzymes. The order of these genes in the linkage map is: TL . . . Pro . . . (Lac) yzi . . . Ad . . . Gal. Note that all three Lac genes specify unique substances. Because i+ produces a repressor substance which, in the absence of lactose, is capable of plei- otropic effects — that is, of phenotypic sup- pression of both >'+ and z+ — /+ can be called a regulator gene. Consider the consequences of certain mu- tations in the Lac region. Mutants capable of synthesizing permease and galactosidase constitutively can have the genotype y+ z+ i in which the specific repressor is not pro- duced, and their y+ and z+ genes can act under all circumstances. E. coli hybrid for the Lac region can be produced by intro- ducing into F~ cells, F merogenotes carry- ing the Lac region (p. 357). Thus, we can obtain an E. coli individual whose chromo- some is y+ z i (which by itself would make permease and Cz protein constitutively), and whose F-Lac particle is y+ z+ i+ (which by itself would make permease and galactosidase only inductively). In the hy- brid, no products are formed in noninduced bacteria (in the absence of lactose), al- though all three (permease, galactosidase, and Cz protein) are formed in induced bac- teria (exposed to lactose). We can con- clude, therefore, that a single r gene can manufacture a repressor substance which prevents the products of both normal and mutant y and z genes from being formed constitutively but not inductively, whether or not these genes are located in the same chromosome segment. In other words, the repressor substance is dillusible and can act at a distance. Various lines of evidence in- dicate that the repressor substance is pro- tein.1 An allele of /+, /* — called a super- repressor — prevents the v and z loci from functioning even in the presence of lactose. Apparently the superrepressor substance is insensitive to lactose, and the v and z loci cannot be derepressed. Another mutant has been found which permits both v f and z+ products to form constitutively, and may, therefore, be a mu- tation of /+. Let us call this mutant allele /'■''. When an F-Lac particle of the geno- type y+ z+ /' is placed in a cell with a chro- mosome y+ z i (which by itself is found to produce permease and Cz protein consti- tutively), no Cz protein is formed consti- tutively in noninduced bacteria. Contrary to the assumption that /' is a mutant of i+ , /' must actually be i+ since it produces a repressor capable of repressing Cz protein formation constitutively. In what respect, then, is the F-Lac particle mutant? Sup- pose the F-Lac particle is mutant at an ad- ditional locus, o + . The new allele, or, would permit only the y and z loci in the same chromosome (or particle) to act con- stitutively regardless of which allele of i is present in the cell. Assuming this is so and ignoring gene order for the present, then the F-Lac particle is genotypically y + z+ o'/ + , while the genotype of the chromosome can be written y+ z o+ i. According to this new hypothesis, the hybrid ought to pro- duce permease and galactosidase constitu- tively and, in addition, to produce Cz pro- tein inductivelv. This is found to be true; 1 See A. Garen and N. Otsuji ( 1964). and M. E. Balis. J. S. Salser, and A. Elder (1964). Regulation of Gene Action — Operons 459 the results obtained with this genotype are summarized in Figure 36-1 along with those of other hybrids which contain both the or and o+ alleles. For example, y + z o+ i+ /F-Lac y+ z + oc i + produces y + and z+ enzymes but no Cz substance constitutively, and it produces all three in induced bacteria. Partial pheno- typic analyses are available for two other genotypes. Thus, y z+ o+ i + /F-Lac y+ z oc i+ produces Cz protein but no galactosidase in noninduced bacteria, but it produces both of these in induced bacteria; v+ z o+ i+ /F-Lac y z+ oc i+ produces galactosidase constitutively, and galactosidase and permease inductively. We may conclude, therefore, that these results confirm the hypothesized existence of an operator gene, o + , and that this gene is the one sensitive to the repressor sub- stance produced by the regulator gene, /+. When the repressor substance is produced and not rendered inactive by lactose in- ducer, the repressor reacts with o + ; this reaction prevents both y and z alleles from operating. When the mutant allele i is present, no repressor is produced, o+ is unaffected, and v and z alleles are capable of acting constitutively. However, a mutant allele of o+, namely o'\ is insensitive to the repressor substance. Consequently, regard- less of the genotype with respect to i, the v and z alleles can act constitutively. Note that the behavior of the v and z alleles de- pends on which particular allele of o is in the same chromosome or F particle; in other words, it depends on which allele of o is linked in the cis position. Thus, the con- stitutive mutant of operator, o'\ has a plei- otropic effect only on other genes in the cis position. Recombinational studies show that the locus of or is between z and i. The operator gene does not seem to pro- duce any unique product which can be de- tected cytoplasmically. Therefore it can be considered a gene whose primary job is not GENOTYPE NON-INDUCED BACTERIA INDUCED BACTERIA Chromosome F-Lac P G Cz P + + y z o i + + C.+ y z o i 33 36 nd + + 4- y z o i + + c + y z o i 50 no nd + + .+ y z o i + c.+ y z o i — o 30 + + + y z o i + c + y z o i nd 60 _ P G Cz P 100 270 100 100 330 100 — 100 400 100 300 — P = Permease G = Galactosidase Cz P= Cz Protein figure 36-1. Crosses, and their results, involv- ing the Lac region of E. coli. nd = not detect- able, — = not tested. 4<;<> CHAPTER 36 REGULATOR GENE REPRESSOR SUBSTANCE T_n_r OPERON _V^ Operator Structural gene genes . O . A B WSAA/S/N Messenger RNA Proteins f\J\j Metabolite removes repressor figure 36-2. Relationships between a regulator gene and the operator and structural genes of an o per on. to specify a chemical product (such as an amino acid sequence in a polypeptide ) but to control the function of other genes. Consequently, operator genes can be called genes for function in contrast to those which specify chemical structures and are accord- ingly called genes for structure (Figure 36- 2). An operator gene coordinates the expres- sion of linear gene neighbors. In our model the genes controlled are related in that both structural genes affect the biochemical path- way involving lactose utilization. This sit- uation suggests that there is, at least in some cases, a unit of gene function, intermediate in size between the gene and the chromo- some, which can be called an operon. An operon is a linear group of genes whose structural activity is coordinated by a func- tional gene, or operator, located at one end? An operon probably represents the length of genetic material whose complementary RNA comprises one strip of messenger RNA; the operon, therefore, may well be a unit of transcription. When the operon - The discussion of operons and operator genes is based mainly upon work of F. Jacob, D. Perrin, C. Sanchez, and J. Monod (1960), and F. Jacob and J. Monod ( 1961 ). is functioning, a strip of messenger RNA is produced containing information for all the structural genes in the operon. The messenger RNA is then translated starting from one end. A leucine auxotroph, leu 500, is caused by an operator mutant, o' , in the leucine operon of Salmonella typhimurium.3 Using transduction to replace this mutant with the wild-type operator o+ restores leucine pro- totrophy. When leu 500 (o' ) individuals are plated on complete medium without leu- cine, some large colonies are formed as a result of reverse mutation to o+. Some small colonies, however, are also produced. When tested by transduction, the small col- onies all prove to have a common property — all are mutant at the same locus, sup- pressor of leu 500 (su leu 500), located outside the leucine operon between the tryp- tophan and cysteine operons. These mu- tants of su leu 500 are of point mutation or deletion type; they no longer completely suppress o' , and, consequently, partial leu- cine prototrophy is restored. Su leu 500 is not the normal regulator gene for the leu- cine operon; when su leu 500 is deleted, :iSee F. H. Mukai and P. Margolin (1963). Regulation of Gene Action — Operons 461 the normal regulatory factors for leucine ap- ply demonstrating that the normal leucine regulator gene is still present. It is hypoth- esized that normally su leu 500 is a regu- lator gene for one or more other operons. When the leucine operon is o + , the sup- pressor substance made by su leu 500 has no effect on the operon. The mutation to o' makes the leucine operon susceptible to the foreign repressor produced by su leu 500, the repressor acting here as a super- repressor. Mutation at the su leu 500 locus partially restores the susceptibility of o' to its own repressor and, it is inferred, also permits the operons normally regulated by su leu 500 to act constitutively. The mutant or first appeared in a culture treated with 5-bromo uracil. Its pattern of reverse mutation by 2-aminopurine strongly suggests that the production of oT involved the transition of a single base-pair. Induc- tion of mutations in su leu 500 by 2-amino- purine indicates that simple base alterations may result in a change in the nature of the su leu 500 repressor so that oT is no longer repressed by it. These conclusions lead us to believe that: 1. Regulator genes for different operons differ from one another by relatively few nucleotides in the region that spec- ifies their repressor's action upon an operator 2. Operator genes for different operons differ from one another by relatively few nucleotides. Operons (such as Lac) whose gene prod- ucts are proteins needed for special diges- tive or catabolic reactions, or for special structural or other biological purposes, are often normally repressed by repressor sub- stances produced by regulator genes. To function, such operons must be derepressed. Other operons, however — particularly those whose enzyme products are used more rou- tinely in metabolism, especially in synthetic or anabolic reactions — apparently are not ordinarily repressed but are functional. In such cases regulation of operon function is sometimes accomplished by a repressor pro- duced when the end product of operon ac- tion combines with the gene product of a regulator gene. Other mechanisms of con- trolling operon action via feedback systems have been suggested. At the level of the gene, we see that any structural gene in an operon can be perma- nently "turned off" by mutations within it. In such a gene, deletions which do not in- volve three or groups of three nucleotides are also expected to turn off other structural genes in the operon whose translation on the ribosome occurs later. We have seen that the functioning of an entire operon can be regulated or changed by the alleles present at the operator locus or at normal and for- eign regulator loci. In principle, such operon regulation could occur either at its transcription or translation stage. For ex- ample, a DNA base substitution at the end where messenger RNA synthesis starts might cause that end to be susceptible to its own or another repressor substance and prevent transcription. A deletion which does not comprise a multiple of three nucleotides might occur at the end of the operon whose messenger RNA sequence is translated first. In this instance, the messenger RNA would be produced but would make complete non- sense. After normal messenger RNA is formed the operon might also be turned off by digestion of the messenger, by a suppres- sor which blocks translation of the mes- senger,4 or by failure of the translated pro- teins to be liberated from the ribosome. Intraoperon regulation can occur at the translation stage,5 since different proteins 4 See E. Orias and T. K. Gartner (1964). ■^See Y. Ohtaka and S. Spiegelman (1963). B. N. Ames and P. E. Hart man (1964). and the refer- ences to R. Byrne, et al. ( 1964) on p. 432, and to M. Nirenberg and P. Leder (1964) on p. 446. 462 ' hap 1 1 K 36 encoded in a messenger RNA are produced of another? This spacing might be attrib- in different quantities. uted to a nonsense triplet or to an intra- The "operator*" gene might sometimes molecular base-pairing that results in seg- prove to be nothing more than the nucleotide ments of a single-stranded messenger RNA sequence which begins a scries of structural having a secondary structure (being "double- genes in an operon. However, the question stranded" as in sRNA).'; remains: What is it in messenger RNA which serves as "spacer" — that is, what marks the .. c y. c c- „ r\ xx/ ,„„oc. . „ , », ,A/ r •■ See M. r. Singer, O. W. Jones, and M. W. termination of one polypeptide and the start Nirenberg ( 1963). SUMMARY AND CONCLUSIONS DNA genes can be classified as genes for structure or genes for function. A gene for function serves as an operator controlling the expression of the structural genes which are its linear neighbors. This whole complex of genes comprises an operon which is probably a unit of transcription. Some operons are normally nonfunctional; they are controlled by a regulator gene which produces a repressor which, in turn, represses the operator and hence, the operon. Derepression of such operons can be accomplished by: 1. An inducer (often the substrate for the first enzyme) which removes the repressor 2. Mutation of the regulator gene (which changes the repressor) 3. Mutation of the operator (which renders it insensitive to the repressor). Other operons are normally functional and are rendered nonfunctional by repressors. Repression may result from mutations in regulator or operator genes which make an operator gene susceptible to the repressor of a foreign regulator gene Regulator genes, inducers, and operons interact at both the transcriptional and trans- lational levels in feedback systems of various types to regulate the production of proteins. REFERENCES Allen, J. M. (Ed.), The Molecular Control of Cellular Activity, New York: McGraw- Hill. 1961. Ames, B. N., and Hartman, P. E., "The Histidine Operon," Cold Spring Harb. Sympos. Quant. Biol., 28:349-361, 1964. Balis, M. E., Salser, J. S., and Elder, A., "A Suggested Role for Amino-acids in Deoxy- ribonucleic Acid," Nature, London, 203:1170-1171. 1964. Cellular Regulatory Mechanisms, Cold Spring Harb. Sympos. Quant. Biol., 26, 1962. Gallant, J., and Spottswood, T., "Measurement of the Stability of the Repressor of Alkaline Phosphatase Synthesis in Escherichia coli," Proc. Nat. Acad. Sci., U.S., 52:1591-1598, 1964. Garen, A., and Otsuji, N., "Isolation of a Protein Specified by a Regulator Gene," J. Mol. Biol., 8:841-852, 1964, Regulation oj Gene Action — Operons 463 x^cv Jacques Monod, in 1964. Gorini, L.. "Antagonism Between Substrate and Repressor in Controlling the Forma- tion of a Biosynthetic Enzyme," Proc. Nat. Acad. Sci., U.S., 46:682-690, 1960. Jacob, F., and Monod, J., "Genetic Regulatory Mechanisms in the Synthesis of Pro- teins," J. Mol. Biol., 3:318-356, 1961. Jacob, F., Perrin, D., Sanchez, C, and Monod, J., "The Operon: A Group of Genes Whose Expression is Coordinated by an Operator" (in French), C. R. Acad. Sci. (Paris), 250:1727-1729, 1960. Translated and reprinted in Papers on Bacterial Genetics, Adelberg, E. A. (Ed.), Boston: Little, Brown, 1960, pp. 395-397. Mahler, I., Neumann, J., and Marmur, J., "Studies of Genetic Units Controlling Ar- ginine Biosynthesis in Bacillus subtilis," Biochim. Biophys. Acta, 72:69-79, 1963. Mukai, F. H., and Margolin, P., "Analysis of Unlinked Suppressors of an o° Muta- tion in Salmonella," Proc. Nat. Acad. Sci., U.S., 50:140-148. 1963. Ohtaka, Y., and Spiegelman, S., "Translational Control of Protein Synthesis in a Cell- Free System Directed by a Polycistronic Viral RNA," Science, 142:493-497, 1963. Orias. E., and Gartner, T. K., "Suppression of a Class of rll Mutants of T4 by a Sup- pressor of a L^c-'Operator Negative' Mutation," Proc. Nat. Acad. Sci., U.S., 52: 859-864, 1964. Riley, M., and Pardee, A. B., "Gene Expression: Its Specificity and Regulation," Ann. Rev.' Microbiol., 16:1-34, 1962. Singer, M. F., Jones, O. W., and Nirenberg, M. W., "The Effect of Secondary Struc- ture on the Template Activity of Polyribonucleotides," Proc. Nat. Acad. Sci., U.S., 49:392-399, 1963. Stent, G., "The Operon: On its Third Anniversary," Science, 144:816-820, 1964. Sypherd P S., and Strauss, N., "The Role of RNA in Repression of Enzyme Synthesis," Proc. Nat. Acad. Sci., U.S., 50:1059-1066, 1963. 404 CHAPTER 36 QUESTIONS FOR DISCUSSION 36.1. What evidence can you cite thai one strip oi messenger RNA can he long enough to contain several structural genes of an operon? 36.2. In what respects are the alleles o . 0 . and oc similar? In what respects arc the) different? 36.3. How does an operator gene differ from a regulator gene? 36.4. Do you suppose that the nucleotide sequence is longer in an operator gene than in other genie members of an operon? Why? 36.5. Does prophage behave as though it contains one or more regulator genes? Does prophage behave as a repressed operon? Explain. 36.6. Some workers classify genes into three types: those for structure; those for regulation: and those for operation. Do you believe this distinction is basic? Do you believe it is useful? Explain. 36.7. What specific kinds of phenotypic effects may operons produce in man? 36.8. In what ways do genes function? 36.9. What is your present concept of the "gene"? 36.10. Discuss the view that in E. coli, rll deletion 1589 involves an operon. 36. 1 1 . What conclusions could you draw from the report that with respect to the bacterial linkage map certain operons are arranged clockwise and others counter- clockwise? 36.12. What support can you give to the hypothesis that o° mutants involve nucleotide deletions? 36.13. Apply the operon concept to hemoglobin production in man. 36.14. Are all regulator genes suppressor genes? Explain. Chapter *37 REGULATION OF GENE ACTION-GENE CONTROL SYSTEMS IN MAIZE I "n discussing the genetic control of mutation (Chapter 30), we considered the case of Activator and Dissociation in maize (pp. 384-386). The control of chromosomal breakage in the Ac-Ds system was found to be associated with the way this system controls the func- tioning of genes located in cis position near Ds — Ds being controlled by Ac, a regulator gene. Based on our knowledge of regulator gene-operator gene systems in bacteria, we can postulate that Ds functions as an opera- tor gene for the Ac regulator gene. Let us analyze another case ' of what originally appeared to be an unstable gene in maize. The pericarp of a corn kernel encloses the seed containing the embryo (Figure 2-8, p. 25). Although embryo tissue is formed by the offspring generation, the pericarp is formed by the parental gen- eration. Some plants are completely red and produce completely red pericarps; other plants are striped with red, and striping ap- pears also in the pericarp; still others are completely nonred. A plant which shows medium variegation of red (therefore called medium variegated) produces kernels of the type shown in Figure 37-1. In the random sample of kernels shown, about 6% have full red color. From this result (and others ) it appears that the parent of a medium variegated pericarp has about 6% mutant 1 Based upon work of R. A. Brink and co-workers. 465 kernels. Genetically, these results can be attributed to mutation at a locus P on chro- mosome 1, and we can expect nonred indi- viduals to be P" P", and medium variegated individuals producing some full red kernels to be heterozygous for P" . The other allele, P'\ would be unstable and in somatic cells frequently mutate to a red-producing gene. If we accept this hypothesis, large red sectors of the stem and leaves would be due to mu- tations of this unstable allele which occur early in development of the shoot; small sectors would be due to later mutations. It has been found, however, that medium variegated individuals produce not only red but also light variegated mutants. The pa- rental type and the two mutant types of ears can be seen in Figure 37-2. The light variegated kernels (lights) have about half as many sectors mutant as have the medium variegated kernels (mediums). The results of test crossing mediums (Pv P" ) are shown in Figure 37-3. As expected, half the offspring are nonred figure 37-1. A random sample of kernels from a medium variegated pericarp ear. (Courtesy of R. A. Brink; photograph by The Calvin Company reprinted by permission of McGraw-Hill Book Co.. Inc.. from Study Guide and Workbook for Genetics by 1. H. Hersko- witz. Copyright, 1960.) 'wV '■ 166 CHAPTER 37 ? A LIGHT (Mutant) MEDIUM (Parental) FULL-RED (Mutant) figurf. 37-2. Corn ears showing medium variegated pericarp (parental type) {A), and the mutant types light variegated (B) and full red (C). {Courtesy of R. A. Brink; reprinted by permission of McGraw-Hill Book Co., Inc., Horn Study Guide and Workbook for Genetics by I. H. Herskowitz. Copyright, I960.) (/' P ■ ). In the remaining half, of those with various degrees of red. 90% are me- diums {P' P"); about 6% are ///// red (reds); and 47c are lights. The similar frequency of reds and lights indicates that these two mutants may somehow be related in origin. Reds by P" P" produce offspring which are all red. if colored at all. The red allele is stable in this cross. Occasionally, medium ears show the two mutants, light and red. as twin patches of kernels (Figure 37-4). This situation sug- gests that reds and lights are not merely re- lated to one another in origin but that they are complementary. In other words, one has gained something the other has lost in the mutation process. In view of the results with Ac and Ds, a new genetic hypothesis can be presented to explain pericarp variega- tion (Figure 37-5). Note that the gene symbols are changed. The way stocks of these strains are maintained, all variegated genotypes are heterozygous for P" , the stable gene (on chromosome 1 ) for nonred peri- carp. The variegated allele is considered a dual structure, containing P' , the top domi- nant allele for red and Mp, Modulator, which suppresses red pigment production. Since a P' Mp combination suppresses red pigmentation, mediums are produced. P' alone produces stable, full red; P' Mp. plus an additional Mp somewhere else {trans- posed Modulator), produces lights. Con- sider the results obtained when certain lights {Pr Mp P" plus transposed Mp) are test crossed by P" P" . Half the offspring are nonred {P" P"); the other half are colored — about half of them lights (genetically sim- ilar to the light parent), half mediums (sim- ilar to the light parent but lacking the trans- posed Mp) with a few reds (cases where Mp is transposed from P' Mp, leaving P' alone). The mechanism for transposing Mp away from P' Mp is considered the same as INBRED MEDIUM VARIEGATED PERICARP 50% Colored Medium Variegated 90% Red 6% Mutant Light 4% Variegated INBRED COLORLESS PERICARP 50% Colorless (Discarded) figure 37-3. Results of test crossing medium variegated pericarp with colorless peri- carp. Regulation of Gene Action — Gene Control Systems in Maize 467 AJbO V figure 37-4. Twin patch of mutant kernels, full red and light variegated, in a medium variegated pericarp ear. (Courtesy of R. A. Brink; photograph LIGHT by The Calvin Company reprinted by VARIEGATED permission of McGraw-Hill Book Co., Inc., from Study Guide and Workbook tor Genetics by I. H. Herskowitz. Copy- right, I960.) that for transposing Ds. This situation is illustrated in Figure 37-6, where the medium parent cell chromosomes (P ' Mp and P" ) are shown already divided, but daughter strands are still connected at the centromere. Normal division would have produced two daughter cells — each carrying P' Mp P" , each giving rise to medium sectors. But when transposition occurs as a result of two or more breaks or some other mutational mechanism, and, consequently, the Mp in one daughter strand is transposed into a non- homologous chromosome (hollow bar), then the daughter cell which receives the trans- posed Mp may be the one carrying Pr Mp, while the other daughter cell will carry only P' . Thereafter, normal mitosis of the cell containing P' alone will produce reds; and mitosis of the sister cell will produce lights. These cells will then become adjacent mu- tant patches in a medium background (Fig- ure 37-4). Red by nonred (Pr/Pu X P"/P") should produce roughly one-half nonred and one- half red. As mentioned previously, it does. Reds do not have Mp adjacent to P'\ Light by nonred (> Mp P" plus transposed Mp x P" P") always produces half nonred. When transposed Mp is located in a non- homologous chromosome, about one quarter of Fj will be light and one quarter will be medium, as mentioned in the previous para- graph. Mp can move away from WWp yet remain in the same chromosome at a new position, so lights may still have their transposed Mp on chromosome 1. In this instance, backcrossing will produce more than one quarter lights in F, and correspond- ingly fewer mediums. Other properties of Mp have been dis- covered. Mp may become fixed at the P locus so that a medium becomes a stable nonred form. Transposed Mp may occupy a variety of sites; two of these have already been mentioned (linked and no-longer-linked to chromosome 1 ). In 57 out of 87 cases, transposed Mp was found still linked to PHENOTYPE GENOTYPE Medium Variegated P Mp / P Mutants Red Light Variegated r w P /P P Mp / P + Transposed Mp / - figure 37-5. New genetic hypothesis for pericarp variega- tion. 4(>S ( IIAPTER 37 chromosome I. the allele having moved less than 50 crossover units from /'. In the re- maining 30 eases Mp was found transposed to one of five different nonhomologous chro- mosomes. Of the 57 eases where trans- posed Mp was still linked to chromosome 1, 37 showed Mp within five crossover units of /'; 10 showed Mp within 5 to 15 units; and the remainder showed Mp to be farther away. Hence. Mp tends to move from the P locus by short rather than by long jumps. This situation suggests that contact between old and new sites may be required for shifts and transpositions of Mp. Reds sometimes revert to variegated. In such cases an Mp is found transposed near P' . The frequency of such reversions from red to variegated can be studied after intro- ducing an Mp locus various distances from P' in a P'-containing chromosome. As shown in Figure 37-7. the closer to P' the introduced Mp is, the greater is the fre- quency of reversions. In summary, medium mutates to red by loss of Mp from its posi- tion near the P' locus. In this process, com- plementary lights are produced possessing an extra Mp — a transposed Mp. The medium type is reconstituted by the return of a trans- posed Mp near the P' locus. Two additional points need to be made. Changes in phenotype involving reds, me- diums, lights, and nonreds arc not muta- tions at the P locus. These changes are the phenotypic consequences of mutations in- volving the transposition of Mp and are, in this respect, much like the changes which follow the transposition of Ds. Transposi- tion of Mp to another locus may change the phenotype the recipient locus produces. For example, a "mutation" to the waxy pheno- type was observed in a particular medium variegated individual whose chromosome 9 carried an allele for the starchy phenotype. The waxy phenotype was unstable and fre- quently "mutated" back to starchy. Tests showed that Mp had been transposed to the starchy locus, which then produced the waxy phenotype; furthermore, the reversions to starchy were the result of Mp's transposition away from this locus. All the phenotypic changes dependent upon the presence of Mp (and Ds) strongly resemble position effects. No evidence has been presented thus far that the transposition of Mp is genetically DIVIDED PARENT DAUGHTER CHROMOSOMES CELLS (Medium Variegated Phenotype) CO-TWIN PHENOTYPES P Mp P Mp > p Mp i * . P Mp | Red | Light Variegated figure 37-6. Transposition of Modulator and the origin of twin sectors. Regulation of Gene Action — Gene Control Systems in Maize 469 controlled. It has been found, however, that the relative frequency with which Mp transposes away from P' Mp is 1 00 in the absence of a transposed Mp; about 60 in the presence of one transposed Mp; and about 5 in the presence of two transposed Mp's. Thus, the transposition of Mp from P' Mp is controlled by the presence of trans- posed Mp. Topographical Relations of Controlling Elements - In the Ac-Ds system, Ac controls Ds not only by regulating Ds transposition by break- age or some other mechanism which may involve contact, but in other ways, as de- tected by the kind of phenotypic effect Ds produces on its linear gene neighbor to one side. The capacity for transposition is pos- sessed by both the regulator gene {Ac) and the operator gene (Ds), a feature unknown in the bacterial systems. In bacterial systems the regulator and operator genes may be close to each other or they may be a considerable map distance apart. In view of the transposability of both elements in maize, it can be hypothesized that both elements are at times adjacent or in close linkage. Consequently, P' Mp can be interpreted P' op}I" RMp where opiIp is the operator and RMp the regulator gene. In general, it can be hypothesized that both an operator gene and a regulator gene are lo- cated close to a structural gene when only the presence of the regulator gene is known with certainty. This general hypothesis is tested in the following way. The gene for bronze, Bz, is located in chromosome 9 and has a com- pletely recessive allele, bz, which produces no color. If a transposition of Ac occurs near Bz, resulting in a variegated bronze color, this locus is now assumed to be Bz opAc RAc. If a two-element control sys- 2 This section is based upon the work of B. Mc- Clintock (1961, 1962, 1963). PERCENT RECOMBINATION P - Mp 2.6 4.3 7.6 12.0 42.0 VARIEGATED SECTORS PER 1000 KERNELS 15 11 8 0.2 figure 37-7. Effect of distance of Mp from P upon transposition rate of Mp to P. tern actually exists at this locus, three kinds of subsequent transpositional events are possible: 1. Transposition of both opAc and Ru. (This operation should release Bz from the control system, and RAc should prove absent from the vicinity of Bz.) 2. Transposition of opAr only. (This should also release the Bz gene from control by RAr; but the latter should still be near Bz and capable of regulat- ing opAc located at other sites in the genotype.) 3. Transposition of RAc only. (This should leave the Bz opAr locus still under the control of RAc in its new location. ) Experimental results confirm these expec- tations. (Were there a one-element control system, the third alternative could not oc- cur.) Since other results demonstrate that Rac Rmp and st}n other R genes are dif_ ferent (each regulates its own type of op gene), numerous and different op-R systems occur in maize. The transposability of op explains why a given op can become the operator gene of a variety of loci previously uncontrolled by that op. Additional work with corn is needed to determine: 1. Whether transposition of op or R in- volves an addition or a replacement of an already present op or R locus 470 < HAITI R 37 2. How the two kinds of controlling loci arise, and whether their origin is inter- dependent 3. Whether the two elements of a con- trolling system usually are adjacent or separate 4. The nucleotidic basis for and the bio- chemical mode of action of, op-R gene- controlling systems. Paramutation The R locus in chromosome 10 of maize produces anthocyanin pigment in the aleu- rone and certain vegetative parts of the plant. Two alleles — stippled. R', and marbled, R'"'1 — are aleurone-spotting genes found oc- casionally in strains of corn native to Peru and neighboring countries. Heterozygotes for one of these mutants are phenotypically as expected if the other R allele is also from one of these South American countries. Mutant homozygotes are also as expected. However, when the other R allele in a hetero- zygote with these mutants is native to other countries, exceptional phenotypes result. This suggests that a control system has evolved in populations where R*' or R'"h occur which does not operate properly when these mutants are heterozygous with foreign R alleles. The exceptional result from "foreign /?" R*' hybrids is that pigmentation is reduced or suppressed in 100% of the R -containing progeny of a test cross. Moreover, all of the /?-containing descendants continue to show the same pigment suppression — even though R8t is no longer present. For this reason. Rat and /?'"'' are said to be para- mutagenic. They induce foreign R alleles to undergo paramutation to form paramutant alleles. Paramutation can occur somatically. An R factor which becomes a paramutant may itself become weakly paramutagenic to other R alleles. The pigment spotting action •See R. A. Brink (1960). of R ' and R"'1' is separable from their para- mutagenic ability. The paramutagenic ac- tion of R ' is not depleted by exercising this function repeatedly in R Rsl plants; never- theless, no evidence has been found for the occurrence of a cytoplasmic element released by a paramutagenic allele and taken up at the paramutable locus.' Paramutational events may occur in or- ganisms other than maize. The conditional segregation-distortion phenomenon observed in Drosophila (pp. 387-388) resembles paramutation. Since it is considered to in- volve a system which controls gene action, paramutation does not imply a modifica- tion of the chemical composition of genetic material. Additional studies should further elucidate the nature of paramutation. Superregulatory Mechanisms ' A single control system is capable of regu- lating the action of many genes in develop- ment. One of the systems studied — (Spm, Suppressor-mutator) — according to B. Mc- Clintock (1963) ". . . serves as a model of the mode of operation of one type of superregulatory mechanism. Such a system can activate or inactivate particular genes in some cells early in development, and activate or inactivate other genes later in development. It can turn on the action of some genes at the same time that it turns off the action of others. It can adjust the level of activity of a particular gene in dif- ferent parts of an organism. . . . The con- trolling elements of the examined systems may represent foreign, nonessential, epi- somelike components that have been inte- grated into the maize genome; or, on the other hand, they may be true chromosomal components of present-day maize, whatever their evolutionary origins and histories may have been." 4 See R. A. Brink. J. L. Kermickle, and D. F. Brown (1964). •See B. McClintock ( 1963). Regulation of Gene Action — Gene Control Systems in Maize 471 SUMMARY AND CONCLUSIONS A large number of cases involving the control of gene action in maize are known. Those analyzed prove to comprise several regulator gene-operator gene systems com- parable to those found in bacteria. In maize both genes of this two-element control system are transposable to new loci. Paramutation probably involves a gene control system. Other, more complicated, control systems (superregulatory. for example) also exist. Royal Alexander Brink, in 1961. REFERENCES Brink, R. A.. '"Very Light Variegated Pericarp in Maize," Genetics, 39:724-740, 1954. Brink, R. A., "Paramutation and Chromosome Organization," Quart. Rev. Biol., 35: 120-137, 1960. Brink, R. A.. Kermickle. J. L.. and Brown, D. F., "Tests for a Gene-Dependent Cyto- plasmic Particle Associated with R Paramutation in Maize," Proc. Nat. Acad. Sci., U.S., 51:1067-1074, 1964. McClintock, B., "Some Parallels between Gene Control Systems in Maize and Bacteria," Amer. Nat., 95:265-277, 1961. McClintock, B., "Topographical Relations between Elements of Control Systems in Maize," Carnegie Inst. Wash. Yearb., 61 ( 1961-1962) :448-461, 1962. McClintock, B., "Further Studies of Gene-Control Systems in Maize," Carnegie Inst. Wash. Yearb., 62 ( 1962-1963) : 486-493, 1963. van Schaik, N. W., and Brink. R. A., "Transpositions of Modulator, A Modifier of the Variegated Pericarp Allele in Maize," Genetics, 44:725-738, 1959. 172 CHAM i k 37 QUESTIONS FOR DISCUSSION 37.1. Could you detect the transposition oi a \fodulator to a locus near /'"'.' Explain. 37.2. Whj does a light variegated individual of maize, with a transposed Modulator on the same chromosome as /' Mp produce among the F, offspring more than one quarter lights and less than one quarter mediums when this individual is hack- crossed to nonred (P" /'" ')? 37.3. Could genes similar to Modulator he the cause of relatively rare "mutants" of amorphic, hypomorphic. and neomorphic types? Upon what do you base your opinion? 37.4. Do experimental results with corn have any hearing upon the hypothesis that the operator gene is nothing more than the initial portion of the nucleotide sequence o\' a transcriptional unit of DNA? Explain. 37.5. Is paramutation a normal mechanism for controlling gene action? Explain. 37.6. Compare the mechanisms for controlling gene action in maize and Salmonella. Chapter *38 REGULATION OF GENE ACTION-POSITION EFFECT IN DROSOPHILA S |ome of the chromosomal rear- rangements induced in Dro- sophila by ionizing radiations have the same, or nearly the same, points of breakage, and many nearly-identical rear- rangements are associated with the occur- rence of the same phenotypic change. Moreover, the new phenotype is transmitted whenever the rearrangement is, and is often similar to that produced by a known allele located at or near one of the breakage points. In such cases, the change in phenotype seems to be directly connected with the mutation of a gene known to be located at or near a point of breakage. We cannot claim that chromosome break- age in or adjacent to a gene automatically changes it to a particular allele because other breaks at this locus partake in other types of rearrangements which do not pro- duce such a phenotypic change. For the same reason, it is untenable to presume that the radiation which caused the break, simul- taneously produced a minute deficiency or duplication of the affected locus. An im- portant feature of this phenotypic change, therefore, is its disassociation with breakage. However, the change may have something to do with the broken ends that join — oc- curring only when the broken end carrying a given locus unites with the broken ends 473 from certain specific loci in the genome. If we accept this view, then we would expect the gene at the broken end to produce one phenotypic effect when united with certain broken ends and another phenotypic effect when joined to others. In other words, the phenotypic effect of a gene may be modified when it has new linear neighbors. Such a phenotypic change is called position effect (pp. 384-385). Presumably position effect changes the working of a gene without changing the gene itself. Position effect, therefore, may be one of the phenotypic consequences of mutations involving struc- tural rearrangements, even though not a mu- tation itself. A gene showing position effect is pre- sumed to be chemically and physically un- changed in any permanent way. Since genes located some distance from a point of break- age sometimes show position effects, position effect can spread somehow along the chro- mosome and affect the functioning of a gene whose immediate linear neighbors have not been switched. This spreading effect is fur- ther reason for dismissing explanations of position effect based solely upon breakage or upon other mutational changes connected with ionization. If the physico-chemical nature of a gene showing position effect is unchanged, two predictions can be made. First, the gene in a position-effect rearrangement should re- sume its original function upon being placed near its former genie neighbors in the chro- mosome. This can be tested experimentally either by irradiating individuals carrying re- arranged chromosomes and examining the progeny for structural changes that reverse this rearrangement, or by moving (by cross- ing over) the gene showing position effect to a normal chromosome. In both cases it is found that the gene returns to its orig- inal position and phenotypic effect. The second prediction is that a normal gene placed in the rearranged position by means 474 ( II M' I IK 3X of crossing over should exhibit the position effect. It docs. In Drosophila, position effects often ac- company rearrangements that bring genes in euchromatin close to those in heterochro- matin. Placing a gene normally located in a dichromatic region near or in a hetero- chromatic region often produces a special, wavering position effect which is expressed in the phenotype as a mosaic or variegated characteristic. Thus, for example, if by paracentric inversion, the gene for dull-red eye color on the X chromosome, vv+ (nor- mally located in euchromatin) is placed in the heterochromatin near the centromere, the result is mottled eye color, white and dull-red speckles. Such variegation is re- duced, however, if, by breeding, an extra Y chromosome or another heterochromatin- rich chromosome is added to the genotype. It is not altogether clear yet how this sup- pression of variegation is brought about. The only requirement for the occurrence of position effect is an appropriate change of a gene's linear neighbors. Breakage merely provides a way of obtaining such changes. Other mechanisms — such as cross- ing over — which change the relative posi- tions of genes should also reveal position effects. Let us see if a crossover system ' which will produce a position effect can be devised. An X-linked mutant in Drosophila, Bar (B), reduces the number of facets (om- matidia ) in the compound eyes, thereby nar- rowing the normally ovoid eye to a slit. When the normal and the /tar-containing chromosomes are studied in nuclei of larval salivary glands, it is found that about seven successive bands in the normal chromosome are duplicated in tandem in the Bar chro- mosome. Let us designate such a single region as abedef. Consequently, a normal 1 Based upon investigations of A. H. Sturtevant. H. J. Muller. C. B. Bridges, and others. female contains abedef abedef and a homo- zygous Bar female abedef abedef abedef abedef. In normal (+ -f ) females, homol- ogous letters (parts) of the two homologs synapse and crossing over takes place be- tween corresponding letters. In homozygous Bar (B B) females, proper synapsis and normal crossing over can also occur, but in this case a potentially different sequence of events will cause synapsis to occur incor- rectly— the left region in one chromosome will pair with the right region of the second (Figure 38-1), leaving the other two re- gions unsynapsed. If this oblique synapsis is followed by normal crossing over any- where in the paired region (as shown be- tween b and c in the figure), the crossover strands will be abedef and abedef abedef abedef. The former strand has this region only once — and will therefore be normal ( + ) — whereas the latter has this region three times. If an egg containing the one- region crossover is fertilized by an X-bear- ing sperm of a normal-eyed male, the zygote will produce a daughter having normal eye shape, thereby demonstrating that Bar has reverted to + . This result can be checked in a subsequent generation by examining the salivary gland chromosomes. If an egg containing the three-region crossover is similarly fertilized, a female will be produced having four of these regions, three in one homolog and one in the other. What will be the phenotype of such a female? Does it make any difference phenotypically whether these regions are grouped two and two (as in homozygous Bar) or three and one? Note that the genie neighbors of the four regions are different when two regions are present on each homolog than they are when one homolog has three regions and the other has one. Since position effect occurs, this gene neighbor difference may result in different phenotypes. Although we do not know what the poten- Regulation of Gene Action — Position Effect in Drosophila 475 tial position effect phenotype should be, we can look for any significant variation from the number of eye facets expected. The normal ovoid eye of females and males ( + + and + Y ) contains more than 200 facets. The homozygous Bar female (B B) and hemizygous Bar male (BY) have about 68 ommatidia per eye. The heterozygous female (+ B) has about 150; Bar on one chromosome is incompletely dominant to -f in the homolog. From the cross -f-/Y 6 by B B 9 , the typical F, females are + B with about 150 ommatidia per eye. As men- tioned earlier, reversions to the + condition by crossing over in an obliquely synapsed tetrad could be detected as a + -j- female. The complementary crossover — a triple re- gion chromosome — with a normal chromo- some would produce a female whose eye might have less than 68 facets or have more than 68 but less than 150 facets. The design of the experiment is not yet complete, however. Since we do not know how often a chromosome showing the po- tential position effect will be produced in meiosis, two other possible causes of excep- tional eye shape must be eliminated. If the cross made is -f Y by B fi, a sperm carry- ing two X's (because of nondisjunction in the father) that fertilizes an egg with no X (because of nondisjunction in the mother) will produce a +/+ daughter counted as one of the exceptional types. Although such zygotes would be extremely rare, they would bcdef a b c d e f PHENOTYPE Normal Female (+/+) abcdefa bcdef abcdefabcdef abcdefab cdef I "~v — a by\c d e f a Homozygous S Bar Female (B/B) bcdef FIGURE 38-1. Diagrammatic representation of the normal and the Bar region of the X chromosome and the consequences of crossing over after oblique synapsis ( / ) . Meiotic Products from Crossing Over Indicated abcdefabcdef abcdef abcdefabcdef a b c d e f abcdefabcdef Bar ? Normal Bar 176 CHAPTER 38 be recognized, however, if the • chromo- some carried as a marker the gene for yellow bodj color, y; such nondisjunctionall) pro- duced daughters would be yellow, not gray, in body color. (In this wa\ we would also be able to recognize any female progeny resulting from cultures contaminated by flies of the yellow stock.) Consequently, the cross that should be made is y -\-B Y by y+ B y ■ B. For clarity the gene symbol for ovoid eye is now given as +". The other event that should not be al- lowed to confuse the results is mutation at or near the B locus. The exceptional pheno- types sought (ovoid and unknown eye shapes ) will always be produced after cross- ing over in the region of Bar. The B B female can be made dihybrid for genes near and on either side of B — near enough (less than ten crossover units apart) to avoid double crossovers between them. On the X chromosome linkage map Bar is located at 57.0; forked bristles (/) at 56.7; and carna- tion eye color (car) at 62.5. Accordingly, the females constructed are: y + /+ B car v + / B car + We can now eliminate from further con- sideration any unusual eye shape that is non- crossover between the loci for / and car. All exceptional phenotypes of interest will be crossovers between / and car; normally, crossovers in this region will be present in 5.8% (62.5 minus 56.7) of F! daughters. To identify the crossover daughters (which will be either nonlorkcd noncarnation or forked carnation), the males used will have to be yf -{-"car Y. The actual cross then is: y / + " car Y 6 by y+ f+ B car y f B car+ 9 When the experiment is performed, about one daughter in two thousand is ovoid-eyed and carries a crossover between / and car; a similar percentage of crossover daughters have very narrow eyes, called Ultrabar (Fig- ure 38-2), each eye containing about 45 facets. The two types of exceptional Hies are equally frequent, as would be expected of the reciprocal products of the hypothe- sized crossing over. Moreover, Ultrabar females contain a triple region in one X and a single region in the other X, as pre- dicted and revealed by examining the sali- vary glands of their F\. Any argument that the Ultrabar phenotype results from a mutation — not a position effect — that is somehow dependent upon a simultaneously- occurring crossing over is disqualified by obtaining females which carry both excep- tional types of X and by occasionally find- ing perfectly typical Bar chromosomes in their progeny. These Bar chromosomes prove to be the product of a crossing over between the single region of one chro- mosome and the middle region of the triple-dose homolog (Figure 38-3). We may conclude, therefore, that four regions aligned in different ways — by crossing over figure 38-2. Compound eye of Drosophila. Left: Ultrabar; center: Bar; right: normal. Regulation of Gene Action — Position Effect in Drosophila All "Triple + " A /^\ y+ f i i_ i r + y f rR\ f a b c d e f a X f car ULTRABAR FEMALE FIGURE 38-3. Production of Bar chromosomes by crossing over in Ultrabar females. "Triple + " A y+ \ + abcdef abcdef abcdef car + ( \ abcdef abcdef V" f y+ f i i_ i r y+ f abcdef abcdef abcdef MEIOTIC PRODUCTS and not by mutation-producing chromo- some breaks — produce different pheno- types. From a B (double region) chromosome, it is also possible to obtain a few chromo- somes nonrecombinant for bordering mark- ers but + (single region) or Ultrabar (tri- ple region). This unusual circumstance is brought about by an intrachromosomal ex- change within the double loop that is formed when the two members of the tandemly du- plicated region synapse with each other. Similarly, intrachromosomal exchange within an Ultrabar chromosome can yield -f and B chromosomes. - -See H. M. Peterson and J. R. Laughnan (1963). Another way of detecting position ef- fects involving crossing over is possible. If the genotype of a Drosophila female is y a^b spl y + a b + spl + , both the a and b loci are heterozygous and the mutants are in different homologs or in trans position (Figure 38-4). If crossing over occurs be- tween these loci, the resulting crossover chromosomes would be y a^ b+ spl+ and y + a b spl. When both these crossover chro- mosomes are present in the same individ- ual, both mutants (a and b) are in the same homolog or in cis position. Both the trans and cis heterozygotes have the same number of loci on each chromosome. In the for- mer instance, however, a and b+ (and a + ITS CHAP MR 38 CIS + + ab TRANS + b a + FIGURE 38-4. Cis and tram positions for di- hybrid linked v<'"<,,>- and b ) would be linear neighbors; in the latter ease a and /' (and a{ and b~ ) would be linear neighbors. If the trans dihybrid had one phenotypic effect, the cis dihybrid obtained from it by crossing over had an- other, and if, by crossing over, the cis form reverted to the trans form and restored the old phenotypic effect, position effect would be considered proved. This cis-trans test for position effect should have the best chance of yielding a positive result when the two pairs of genes involved are adjacent to or very close to one another. If the genes are very close together, crossing over will rarely occur be- tween them; large numbers of progeny would have to be scored to assure at least one crossover. The members of the multiple allelic series at the white locus on the X chromo- some of Drosophila are all located at approximately 1 .5 on the crossover map. Al- though the hybrid composed of vv" (apricot) and vv (white), vv"u\ produces pale apricot eye color, this result does not prove that wa and vv are alternatives of the same gene. Suppose vv" and vv are mutants of separate but similarly-acting genes located close to- gether, one at position 1.49 (w"+) and one at 1.51 (vv" ). The crossover data, being finite and somewhat variable, could acci- dentally have placed them both at locus 1.5. If wa* and vv+ are close but separate loci, the trans dihybrid should yield the cis di- hybrid by crossing over. To test whether vv"' and vv^ are sepa- rate loci. Drosophila females which carry an attached-X chromosome with y w spl on one arm and y vv" spl on the other are bred. Recall that the use of attached-X's permits the recovery of two of the four strands involved in each crossing over (p. 122). The attached-X genetic system some- times yields both complementary crossover types in the same gamete. Figure 38-5 (left side) shows schematically a portion of this attached-X as it would appear in the tetrad stage at the time of the crossing over and indicates the standard genetic map loca- tion of the y and spl markers. When a fe- male with pale (dilute) apricot eye color and this chromosome is crossed with a Bar- containing male, the non-Bar F, daughters (who carry a paternally-derived Y) are usu- ally noncrossovers and have pale apricot- colored eyes like their mother. Crossovers between the region containing the white lo- cus and the centromere produce either white or apricot daughters. Barring mutation, these phenotypes would be the only ones expected if vv" and vv were alternatives of the same gene. But if vv"* is separate from vv + , the for- mer can be distinguished by a new symbol, apr+ . If apr+ lay to the left of vv+ (as shown in the left portion of the figure), apr and vv would each have its own -+- allele in the other arm of the parental attached-X and, consequently, the female parent would have to be a trans heterozygote with respect to these loci. A rare crossing over between these loci would produce the crossover at- tached-X shown at the right of the figure. As a result, the two mutants would be in the cis position. When large numbers of daughters from the attached-X females are examined, sev- eral are found to have dull-red eyes. It is essential to determine whether these flies are mutant or the result of a change from the '■'• Based on the work of E. B. Lewis. Regulation of Gene Action — Position Effect in Drosophila 1 1.5 3.0 479 figure 38-5. Crossing over between apricot and white /'// attached-X chromosomes. + + H h apr + apr w > spl trans to the cis form. To do this, we de- tach— that is, separate the arms of — the at- tached-X's in the dull-red-eyed exceptional flies (by collecting the products of the oc- casional crossing over that occurs between the attached-X and the Y in the hetero- chromatic regions near their centromeres) and determining the genes carried in each de- tached arm. The finding that one arm can always be represented as y apr+ w+ spl+ and the other as y+ apr w spl offers strong support for the view that the dull-red ex- ceptional females were cis heterozygotes, and that apr lies to the left of w on the X chromosome, as shown in the figure. (It is instructive to work out the arrangement of the markers after crossing over between apr and w on the assumption that apr is to the right of w.) Proof that the exceptional dull-red fe- males result from position effect rather than mutation is obtained by mating these ex- ceptional females and occasionally obtain- ing daughters with pale apricot eye color. In these new exceptional daughters the orig- inal gene arrangement is found restored by crossing over. The phenotypic difference between pale apricot and dull red is undoubtedly the re- sult of position effect, since the only differ- ence between the cis and trans conditions is in the arrangement of the genetic mate- rial. This phenomenon is therefore termed a cis-trans position effect. To detect such an effect it is necessary to separate two very closely linked genes. Prior to the experi- ment the genes used had been considered alleles because of their closeness on the ge- netic map and their similar phenotypic ef- fects, but observing their cis-trans position effect proved they were nonalleles occupy- ing different loci. When the other genes making up the "white multiple allelic se- ries" (Chapter 5) are investigated, some are found to be allelic to w and others to apr. Some, however, are allelic to neither, and appropriate crossing over studies show that the "white region" on the X is a nest of five (perhaps more) separate, linearly ar- ranged genes with similar effects. Other regions in the genome are now known where two or more genie alternatives — previously considered allelic — prove to be pseudoallele, that is, prove to be nonallelic when subjected to the cis-trans test. In addition to pseudoallelism in Aspergillus, other microorganisms, and corn, examples of pseudoallelism include cases involving color in cotton; lack of tails in mice; lozenge and vermilion eye colors in Drosophila. Another case of pseudoallelism in Dro- sophila 4 involves nonalleles whose func- tions differ somewhat more than apr and w. 4 See E. B. Lewis (1963) for complete discussion. 4S0 CHAPTER 38 figure 38-6. Drosophila melanogaster males: normal (A), bithorax (B), post- bithorax (C). and bithorax postbithorax (D). (Courtesy of E. B. Lewis; reprinted by permission of McGraw-Hill Book Co., Inc., from Study Guide and Workbook for Genetics by I. H. Herskowitz. Copyright, 1960.) The normal wild-type fly (Figure 38-6A) has small club-shaped balancers (halteres) located on the posterior part of the thorax. One of the pseudoalleles, bithorax (bx), converts the haltere into a large wing-like structure (Figure 38-6B); another, called postbithorax (pbx). appears to do much the same thing (Figure 38-6C). But close examination reveals that these two recessive pseudoalleles really have different func- tions. Bithorax converts the front portion and postbithorax the hind portion of the haltere into a wing-like structure. Flies ho- mozygous for both mutants demonstrate these changes in a fully developed second pair of wings (Figure 38-6D). What are the cis-trans effects for bx and pbx? The cis form (H — \- bx pbx) has normal balancers, whereas the trans form (bx -\-/+ pbx) shows a slight postbithorax effect, providing another example of cis- trans position effect and demonstrating the nonallelism of these genes. The map dis- tance between these loci is 0.02. These examples of pseudoallelism appar- ently involve separate but closely linked functional genes; they do not seem to in- volve intragenic recombination of the type Regulation of Gene Action — Position Effect in Drosop/iila 481 that occurs within the A or B cistron of the ill region of <£T4. The possibility exists, however, that recombination does occur within 5 as well as between functional gene units in Drosophila, although mechanisms other than crossing over may also be in- volved. Consider the morphology of Drosophila chromosome regions which contain pseudo- alleles. The white series is associated with a double band (doublet) in the salivary gland chromosome; apr may be in one band, w in the other. The vermilion series is as- sociated with another doublet in the X chro- mosome, whereas the bithorax series (com- posed of five separate pseudoallele loci) is connected with two doublets. (This last fact demonstrates what is proved by other data — that a band may contain more than a single gene.) The great number of dou- blets in salivary chromosomes suggests that genes located in these regions are pseudo- allele. The origin of adjacent loci with similar types of action can be accounted for in sev- eral ways. One explanation is that during the course of evolution, adjacent genes pro- ducing different effects mutated to alleles which performed similar, presumably advan- tageous, functions. A second explanation might be that rearrangements brought to- gether widely separated nonalleles with sim- ilar functions. Though both of these ex- planations may be sufficient for some of the cases found, it seems more likely that most adjacent and similar genes arose as duplica- tions that occurred one or more times (as in the bithorax case) in the ways described in Chapter 12 (see also pp. 418-419). After duplication linearly adjacent genes — origi- nally identical — would have become some- what different from each other functionally by mutation. 5 See W. J. Welshons and E. S. Von Halle ( 1962), and A. Chovnick, A. Schalet. R. P. Kernaghan, and M. Krauss (1964). What causes position effects? With re- spect to gene action, genes that are linear neighbors are perhaps more likely to be de- pendent upon each other than upon their alleles in a homologous chromosome (usu- ally located a considerable distance away). This kind of dependency might be subject to position effect when the relative positions of heterochromatin and euchromatin are shifted by breakage. Position effects due to structural changes might be particularly common in species whose chromosomes or chromosome parts take on special positions in the nucleus relative to each other. Two facts — that during nuclear division Dro- sophila chromosomes show somatic syn- apsis, and that somatic synapsis is found in the giant interphase nuclei of salivary gland and other cells — suggest that at the time of gene action different chromosomes and their parts are arranged so that the products of gene action may be formed or used in particular sequences. Due to the presence of heterozygous reciprocal trans- locations in Oenothera (Chapter 17), chro- mosomal parts show a very orderly arrange- ment in the circle of 14 chromosomes formed during meiosis. Here also, a new arrangement of chromosomal parts might disturb functional sequences and produce position effects. As a matter of fact, posi- tion effect is known to occur in Oenothera. The molecular basis for position effects in Drosophila is yet to be established. In studying systems that control gene action the results with bacteria and maize lead us to hypothesize that position effects in Dro- sophila can result if the production or the translation of messenger RNA for a given cistron is influenced by a change in its gene neighbors. Since protein synthesis occurs in the nucleus, the possibility also exists that position effects can result from sequential intranuclear reactions which are influenced by the diffusion and, hence, the concentra- tion of the protein products of gene action. 1S2 CHAPTFR 38 SUMMARY AND CONCLUSIONS Phenotypic changes can result when the same genetic material is arranged in different ways. 1 he shuffling ol genes which produces such position effects may he brought about b\ structural changes in chromosomes and by crossing over. linear nests of genes with similar effects have probably arisen by one or more duplications in situ o! an ancestral gene, followed by mutations that led to differentiation in their effects. Position effect is attributed to a change in one or more of the following: 1. Production of a given messenger RNA 2. Translation oi this messenger RNA 3. Interactions between proteins resulting from the intranuclear translation of mes- senger RNAs. REFERENCES Bridges. C. B.. "The Bar "Gene" a Duplication." Science. 83:210-211, 1963. Reprinted in Classic Papers in Genetics. Peters. J. A. (Ed.). Englewood Cliffs, N.J.: Prentice- Hall. 1959, pp. 163-166. Chovnick. A.. Schalet. A.. Kernaghan. R. P., and Krauss. M., "The Rosy Cistron in Drosophila melanogaster: Genetic Fine Structure Analysis." Genetics, 50:1245- 1259. 1964. Lewis. E. B.. "The Pseudoallelism of White and Apricot in Drosophila melanogaster," Proc. Nat. Acad. Sci., U.S., 38:953-961. 1952. Lewis. E. B.. "Genes and Developmental Pathways." Amer. Zool., 3:33-56, 1963. Muller. H. J., Prokofyeva-Belgovskaya. A. A., and Kossikov, K. V.. "Unequal Cross- inszover in the Bar Mutant as a Result of Duplication of a Minute Chromosome Section." C. R. (Dokl.) Acad. Sci.. U.R.S.S., N.S.. 1 ( 10) :87-88, 1936. Peterson. H. \L. and Laughnan. J. R.. "Intrachromosomal Exchange at the Bar Locus in Drosophila." Proc^Nat. Acad. Sci., U.S.. 50:126-133, 1963. Sturtevant. A. H.. "The Effects of Unequal Crossingover at the Bar Locus in Dro- sophila," Genetics, 10:117-147, 1925. Reprinted in Classic Papers in Genetics, Peters. J. A. (Ed.), Englewood Cliffs, N.J.: Prentice-Hall, 1959, pp. 124-148. Welshons. W. J., and Von Halle. E. S., "Pseudoallelism at the Notch Locus in Dro- sophila." Genetics. 47:743-759. 1962. QUESTIONS FOR DISCUSSION 38.1. If a previously unknown phenotype appears at the same time as a qualitative or quantitative change in the genetic material, can you determine whether the effect is due to mutation or to position effect? Explain. 38.2. Would you expect to find position effects in most sexually reproducing organ- isms? Why? 38.3. Is crossing over ever unequal? Explain. 38.4. Should pseudoalleles be considered subgenes (parts of one gene) rather than separate, nonallelic genes? Explain. 38.5. Does position effect require pseudoallelism for its detection? Explain. Is the reverse true? Explain. Regulation of Gene Action — Position Effect in Drosophila 483 38.6. What genotypic steps are required to prove that a region shows a cis-trans position effect? Explain. 38.7. Can one of the steps in question 38.6 be a lack of all or part of the region under investigation? Why? 38.8. What crosses would you make to test whether two recessive mutants in Dro- sophila. apparently alleles of the X-linked gene v+ (normal allele of vermilion eye color), are pseudoalleles? 38.9. Using appropriate genetic markers, draw a tetrad configuration which would permit you to identify strands which have undergone intrachromosomal ex- change in the Bar region of the X chromosome of Drosophila melanogaster. 38.10. Using the operon concept, explain how a paracentric inversion in the X chro- mosome of Drosophila might result in a mottled eye color phenotype. 38.11. Suggest a molecular explanation for the cis-trans position effect observed in the white region of the X chromosome in Drosophila. 38.12. Can position effect occur in haploids? Why? Chapter *39 REGULATION OF GENE ACTION-DOSAGE COMPENSATION Mammals Because DNA synthesis occurs almost with- out interruption in the relatively uncoiled chromosome of E. coli, we may suppose that, in general, uncoiled chromosomes can synthesize complementary DNA and RNA. On the other hand, DNA synthesis does not occur during mitosis or meiosis when the chromosomes are highly con- densed. Consequently, we may hypothesize that a chromosome cannot function in DNA or RNA synthesis while coiled. If these premises are acceptable, we may consider that during interphase the presence of heav- ily Feulgen stained, clumped, chromosomal material (chromatin knots or karyosomes) is an indication of coiled chromosomal ma- terial and. therefore, the total absence of genetic activity in such bodies. That certain chromosomes or chromo- somal regions are highly clumped or coiled while others are not is probably correlated with their different times for DNA replica- tion. Chromosomes or chromosomal re- gions which are normally coiled and stained (eupycnotic), relatively overcoiled and over- stained (hyperpycnotic), and relatively un- dercoiled and understained (hypopycnotic) differ in their time of DNA replication. Abnormal staining, heteropyknosis, is one of the characteristics of heterochromatin (p. 155). The gene action hypothesis presented 484 above can be tested by comparing the in- terphase activity of genes when a given chromosomal region is normally coiled and when it is overcoiled. Manx of the diploid interphase nuclei in human males show a small karyosome touching the nuclear mem- brane; in the human female the same cells show a similar but much larger karyosome. Because the size of this chromatin knot dif- fers in each sex. the extra karyosome ma- terial in the female is called sex chromatin, or the Barr body (after its discoverer). The presence of sex chromatin in individ- uals aneusomic with regard to sex chromo- somes has been investigated. The maxi- mum number of separate Barr bodies found are: none in XY and XO individuals; one in XX, XXY, XXYY; two in XXX, XXXY; and three in XXXX individuals. Cells with less than the maximum number have fewer and, accordingly, larger Barr bodies. We may conclude, therefore, that the maximum number of Barr bodies is one less than the number of X chromosomes in a diploid individual. Tetraploid cells of a male probably have no valid Barr body; the larger karyosome reported is attributed to the union between the small karyosomes of the two X's. That tetraploid cells of a female have two Barr bodies suggests the maximum number of Barr bodies is deter- mined by the balance between the number of X chromosomes and the number of auto- some sets. One X chromosome balanced by two sets of autosomes does not give rise to a Barr body. However, each X in ex- cess of this balance clumps and is either detected as a separate Barr body or is fused with other excess X's to form larger bodies. Hence the maximum number of Barr bodies is equal to x-(p 2). where x is the num- ber of X chromosomes, and p is the ploidy or number of sets of autosomes.' Note that the Y chromosome has no influence upon 1 According to D. G. Harnden. Regulation of Gene Action — Dosage Compensation 485 the number of Barr bodies. Consequently, the genes for sex, particularly those located in the Y, do not suppress Barr body forma- tion in normal or abnormal males. According to the hypothesis under dis- cussion, each excess X is rendered hyper- coiled and functionally inactive. It is, how- ever, still capable of being replicated dur- ing interphase, although the replication is delayed. - That no normal tissue in a fe- male ever has 100 per cent of its nuclei showing a Barr body may very well be partly due to errors in cytological observa- tion. It is also possible that some of the cells that fail to show a Barr body are rep- licating the X chromosome involved. As a consequence of Barr body formation, males and females — whether normal or abnormal — apparently have similar numbers of func- tional X chromosomal genes per diploid number of autosomes. :; In other words, basically males and females may not be very different after all, at least at the functional X chromosome level. The human X chromosome (Figure 10- 7, p. 139) contains a gene necessary for production of the enzyme glucoses-phos- phate dehydrogenase ( G-6-PD ) . One X chromosomal mutant fails to produce this enzyme. Males with the normal X there- fore can and those with the mutant X can- not make this enzyme. Since the Y chro- mosome has no effect on the production of this enzyme it carries no allele for this gene. When individually tested,4 red blood cor- puscles of genetically pure, normal males and females show the same amount of G- 6-PD activity. If both normal alleles op- erated in the female as they did in the male, we would expect the red blood corpuscles 2 See M. M. Grumbach, A. Morishima, and J. H. Taylor (1963). -See M. F. Lyon (1962). 4 See E. Beutler, M. Yeh, and V. F. Fairbanks (1962), and D. R. Davidson, H. M. Nitowsky, and B. Childs (1963). of a normal female to produce twice as much enzyme as those of a normal male. When the red blood cells of females hetero- zygous for the X-linked mutant are studied, however, some are found to be normal and others deficient with respect to G-6-PD; no corpuscles of intermediate activity are found. These results prove that such hu- man females are functional mosaics for this locus. Some of their red blood cor- puscles are derived from nucleated cells in which the normal gene is nonfunctional, the defective locus functional; others come from cells in which the mutant gene is nonfunc- tional, the normal locus functional. The al- ready-mentioned results obtained with nor- mal females support the general conclusion that euploid females can express only one allele of this locus in any cell, with some- times the maternally-derived, sometimes the paternally-derived locus being operational. In support of this conclusion is the finding that in human females not carrying the G- 6-PD mutant, otherwise diploid cells, either X0, XX, XXXY, or XXXX, all produce the same amount of G-6-PD. Since X-linked muscular dystrophy is due to a rare mutant of an X-limited gene, usu- ally only males have this disease. Muscular dystrophy is closely associated with certain enzymatic and histological abnormalities. Studies of females known to be mutant het- erozygotes and showing subclinical and clin- ical muscular dystrophy "' reveal two popu- lations of muscle fiber — one normal and the other dystrophic, a result best attributed to functional mosaicism, just as described for the G-6-PD locus. The same kind of re- sult is obtained for at least five other X- linked genes. On the basis of these genetic results and the cytological studies of Barr bodies, we can conclude that a female nor- mally has an appreciable portion of one X chromosome inactivated in many diploid so- 5 See C. M. Pearson, W. M. Fowler, and S. W. Wright (1963). ts<; CHAPTER 39 matic cells. Since the normal male has no Barr body, it is the condensed X which is inactivated. Human sex chromatin is not present at fertilization; it first appears at about the twelfth day of development. From Barr body cytology and the inac- tivation of a half dozen or so different loci, we cannot determine how much of an X chromosome is inactivated in a normal hu- man female. Still other X chromosome loci whose gene action can be studied in sepa- rate individual cells, need to be discovered. A gene whose action gives rise to a product diffusible between cells will be of little or no use in determining the length of the in- activated segment. Sex chromatin occurs in many mammals H besides human beings. In the mouse, al- though sex chromatin is absent one of a female's X's is heteropycnotic during mito- sis. One locus has been found which fails to show the inactivation expected,7 suggest- ing the presence of an X chromosome re- gion not heteropycnotic in the female and, therefore, not routinely inactivated. Since reciprocal X-autosome translocations occur in mice, we may ask whether such rearrange- ment has any effect upon the functioning of the rearranged autosomal genes. This ques- tion can be studied in females heterozygous for such a translocation when the nontrans- located autosomal homolog carries suitable recessive alleles of genes whose loci span a large portion of the linkage group. In some cases the phenotype is that of the nor- mally dominant allele present in the struc- turally rearranged autosome; in others, a mottled or variegated phenotype results. Moreover, according to studies of different rearrangements between a given autosome and the X, in the latter cases, the portion of the body showing the recessive pheno- type decreases as the distance from the au- '■See M. L. Barr (1959). •See L. B. Russell (1963). tosomal locus to the point of union with the X increases. Consequently, autosomal loci can be inactivated by translocation to the X, the greater the distance from the breaking point, the less inactivation. Since breakage in an autosome occurs at several positions, the decreasing strength of inac- tivation has been found to proceed in either direction. Thus, inactivation of autosomal loci in X-autosome rearrangements can ex- plain cases of variegated-type (V-type) po- sition effects. In two instances, the autosomal break points involved in rearrangement with the X were located in slightly different positions; both, however, were close to a given gene. In one case, the normal autosomal allele was inactivated and produced variegation; in the other no variegation occurred. Re- sults from the investigation of the latter case strongly suggest the nonsuppressed wild-type gene was in an autosomal frag- ment which had joined an X region inca- pable of causing inactivation. These studies also suggest — as does the already-mentioned finding of an apparently unsuppressible X locus — that, normally, in the segment re- placed by the autosomal fragment some X chromosome genes near the point of break- age are always functional. There are, then, two possible reasons for nonvariegation of an autosomal gene attached to the X chro- mosome: attachment to a portion of X in- capable of causing inactivation and excessive distance from a portion of X capable of causing inactivation. In light of the preceding discussion, per- haps the same genetic material can be het- erochromatic or euchromatic, the primary determinant being the degree of coiling. Chromosome hypercoiling during interphase seems to prevent the functioning of the gene contents. In mammals, this system appar- ently compensates for the difference in gene dosage existing between male and female. Regulation of Gene Action — Dosage Compensation 487 That is, the system provides dosage com- pensation with respect to some X-limited loci.8 Note that in mammals parts of certain chromosomes can become more or less per- manently fixed in their functioning.9 The mosaicism just discussed in mice results in patches of tissue of different phenotypes. When a chromosome segment is function- ally turned off (or on) in a given cell, the descendant cells are similarly turned off (or on), despite the intervening occurrence of mitosis. Consequently, the result is a patch of tissue of similar phenotype. This turn- ing off occurs more than a week after fer- tilization in mice and, as mentioned, in human beings. Moreover, it at least some- times fails to occur in the germ line. For example,1" although one X in an adult fe- male rat is hyperpycnotic in somatic tissues, the XX-bivalent in the oocyte is isopycnotic; that is, both members stain similarly. It is therefore hypothesized 1T that a gene can exist in two functional forms, active or in- active, and that one form will persist despite intervening nuclear divisions until specific conversion to the other. For some maize genes, it is found that a given state during gametogenesis is continued in the zygote. In the endosperm, for example, a maternally- derived gene has been found to continue activity whereas the paternally-derived allele is inactive. Drosophila In Drosophila, as in man, typical males have one X chromosome and typical females two. Many of the X chromosome loci have es- sentially the same phenotypic effect in males as in females; that is, many loci show dos- sSee M. F. Lyon (1962). 9 See also R. DeMars (1963). 10 See S. Ohno. W. D. Kaplan, and R. Kinosita (1960). 11 By M. Lyon and by D. Schwartz. age compensation. Thus, the eye color of the apr Y male and apr/apr female is es- sentially identical. Dosage compensation applies not only to hypomorphic mutants (p. 194) like apricot but also to their wild- type alleles. Consequently, both wild-type males and females have, for example, the same dull-red eye color. Other X chromo- some loci in D. melanogaster showing dos- age compensation are y; ac; sc; sn; g; f; and B. Only partially compensated for, how- ever, is fa, since females show somewhat more effect than males. Hw (hairy wing), we (eosin), and W (ivory) show no com- pensation, nor do any autosomal genes, nor bb which has a locus in both the X and the Y chromosomes and, therefore, is usually present in paired condition in both males and females. The cytological basis for dosage compen- sation is difficult to study in ordinary so- matic cells of Drosophila because of the small size of diploid nuclei. Since the so- matically synapsed polynemic X's in the fe- male larval salivary gland and other nuclei appear identical, dosage compensation in Drosophila does not seem to be based upon differences in chromosome coiling between homologs. Perhaps this means that the alleles on both X chromosomes of a fe- male are equally functional, and that in Drosophila dosage compensation is accom- plished by a somewhat different mechanism than it is in mammals. Whereas the single X and the paired X's in salivary cells of males and females have DNA in approxi- mately a one-to-two ratio as expected, the single X seems to contain just about as much RNA and protein as the double X. Dosage compensation therefore involves the "stepping-up" of gene action by single X's, the suppression of gene action by double X's, or both. A number of genetic studies throw some light upon dosage compensa- tion in Drosophila. and their results can be i.N.S ( IIAI'TER 39 described in terms of the apr allele. Fe- males having apr in triple dose (the extra apr [OCUS is carried in another chromosome ) have darker apricot eyes than apr apr fe- males, demonstrating not only the hypo- morphic nature of the mutant but also the direction of dosage compensation — namely, suppression of eye pigment formation in apr apr females and the eonsequent level- ing-ofi of pigment formation at the level produced by one apr locus present in the X chromosome of a male. Males carrying an extra apr locus have apricot eyes even darker than those of females with triple apr. Since XO and XY males and XX, XXY, and XXYY females — all pure for apr — have the same eye color, the Y chromo- some cannot be responsible for dosage com- pensation. In addition, males transformed from females (X"p' X""'. tratra) with or without a Y chromosome have the same eye color as an X'"" Y male. If maleness as such prevents the suppression of gene action leading to dosage compensation, the transformed-from-female male with a double dose of apr should have a darker apricot eye than a male with a single dose of apr. Thus, dosage compensation is not dependent upon male or female phenotype. What is the genotypic basis for dosage compensation in Drosophila? That the male (transformed female) with two full X chromosomes shows dosage compensation, while the male with one Xopr and another apr — either in a grossly deleted X or in- serted into an autosome — does not, suggests that the X chromosome itself contains dosage compensator genes, a single dose of such genes present in males and a double dose in females. One complete X chromosome ap- parently carries several dosage compensators to suppress one apr gene. In a regular X""' Y male, this suppression permits only apri- cot eye color. Note that females with the apr region in one X deleted, and apr present in the other have a light apricot eye color. Pure apr males or females hyperploid for different short segments of X can be ob- tained and scored for eye color and sex. Such experiments show that: 1 . Different X segments have a positive or negative dosage compensating effect on eye color with an overall effect of suppression 2. The effects of these segments on eye color are not correlated with their ef- fects on sex differentiation. Moreover, the net dosage compensation ef- fects exerted by individual segments are dif- ferent when other loci showing dosage compensation are investigated. In conclu- sion, therefore, we find: 1. No correlation between a set of com- pensator genes and their effect on sex differentiation 2. Suppression of different genes exhibit- ing dosage compensation either by dif- ferent groups of dosage compensator genes or by the same groups of com- pensator genes whose action varies with the locus to be compensated. X-linked genes or alleles without dosage compensation may be so new in terms of evolution that dosage compensator genes may not yet have had an opportunity to be- come established. Supporting this view is evidence that eosin (W ) and ivory (w1) which do not show dosage compensation are nonallelic to apr '- and, therefore, may be mutants of a more recently-evolved locus. Additional support comes from the study of mutants in the X chromosome of D. psendo- obscura which is V-shaped with one arm homologous to the X, the other to the left arm of chromosome III of D. melanogaster. If we assume that most mutants are hypo- morphs, we find more mutants in the arm homologous to the melanogaster X showing 12 From work of M. M. Green (1959). Regulation of Gene Action — Dosage Compensation 489 the same degree of phenotypic effect in both males and females (probably representing dosage compensation) than we find in the arm homologous to the melanogaster III L.13 As a working hypothesis we can suggest that the suppression involved in dosage com- pensation in Drosophila is closely associated with messenger RNA. The messenger RNA (or the protein product) of compensator genes could either interact with and inac- i:iThe preceding discussion of dosage compensa- tion in Drosophila to a great extent follows the work of H. J. Muller (1950). tivate messenger RNA of the genes they compensate, or they could act directly upon the gene to be compensated interfering with its production of messenger RNA. This suppression might be based upon the mes- senger RNAs of all dosage compensators having some common nucleotide sequence (p. 461), either the complement or the equivalent of a deoxyribotide sequence in the locus being suppressed. Perhaps the operator genes of the loci being compensated are suppressed by the regulator dosage com- pensator genes. SUMMARY AND CONCLUSIONS Dosage compensation for X-linked genes is accomplished by suppression of gene function. In human beings, in mice, and probably in all mammals having sex chromatin, this suppression is associated with heteropycnosis and, hence, chromosome coiling. Al- though some differentiation in a chromosome is temporary, dosage compensation and other nuclear phenomena demonstrate that the chromosome can become more or less permanently fixed in some way related to its function. In Drosophila, another mechanism is responsible for the regulation of gene action leading to dosage compensation. It is hypothesized that this process involves the transcription, translation, or protein products of messenger RNA; perhaps dosage com- pensator genes are regulator genes which control operator genes of the loci being compensated. REFERENCES Barr, M. L., "Sex Chromatin and Phenotype in Man," Science, 130:679-685, 1959. Beutler, E., Yeh, M., and Fairbanks, V. F., "The Normal Human Female as a Mosaic of X-Chromosome Activity; Studies Using the Gene for G-6-PD-Deficiency as a Marker," Proc. Nat. Acad. Sci., U.S., 48:9-16, 1962. Cock, A. G., "Dosage Compensation and Sex-Chromatin in Non-Mammals," Genet. Res. (Camb.), 5: 354-365, 1964. Davidson, R. G., Nitowsky, H. M., and Childs, B., "Demonstration of Two Popula- tions of Cells in the Human Female Heterozygous for Glucose-6-Phosphate De- hydrogenase Variants," Proc. Nat. Acad. Sci., U.S., 50:481-485, 1963. DeMars, R., "Sex Chromatin Patterns and the Lyon Hypothesis," Science, 141:649- 650, 1963. Grumbach, M. M., Morishima, A., and Taylor, J. H., "Human Sex Chromosome Ab- normalities in Relation to DNA Replication and Heterochromatinization," Proc. Nat. Acad. Sci., U.S., 49:581-589, 1963. Human Genetics, Cold Spring Harb. Sympos. Quant. Biol.. 29, 1965. Lyon, M. F., "Sex Chromatin and Gene Action in the Mammalian X-Chromosome," Amer. J. Hum. Genet., 14:135-148, 1962. 490 ( ii aim ir 39 McKusick, V. A.. On the X Chromosome of Man. Washington, D.C.: American In- stitute of Biological Sciences. 1964. Muller. H. J., "Evidence oi the Precision of Genetic Adaptation," I he Harvey Lectures (1947-1948), Ser. 43:165-229, Springfield. 111.: Chas. C. Thomas, 1950. Ex- cerpted in Muller. H. J., Studies in Genetics, Bloomington, Ind.: Indiana Univ. Press. 1962, pp. 152-171. Ohno. S.. Kaplan, W. D., and Kinosita. R.. "On Isopycnotic Behavior of the XX- Bivalent in Oocyte of Rattus norvegicus," Exp. Cell. Res., 19:637-639. I960. Pearson. C. M.. Fowler, W. M., and Wright, S. W., "X-Chromosome Mosaicism in Females with Muscular Dystrophy," Proc. Nat. Acad. Sci.. U.S., 50:24-31, 1963. Russell. L. B.. "Mammalian X-Chromosome Action: Inactivation Limited in Spread and in Region of Origin." Science, 140:976-978, 1963. Russell. L. B.. "Another Look at the Single-Active-X Hypothesis." Trans. N.Y. Acad. Sci.. Ser. II. 26:726-736, 1964. Russell, L. B., "Genetic and Functional Mosaicism in the Mouse," Sympos. Soc. Study Devel. and Growth, 1964. Stern, C, "Dosage Compensation — Development of a Concept and New Facts," Canad. J. Genet. Cytol., 2:105-118. 1960. Welshons. W. J., "Cytological Contributions to Mammalian Genetics," Amer. Zool., 3:15-22, 1963. QUESTIONS FOR DISCUSSION 39.1. If essentially all of one X chromosome in a human female is functionally turned off in somatic tissues, what is the disadvantage of having 9 = X 0 and 6 = X Y? 39.2. Almost without exception, piebald or tortoise-shell cats are females. Explain. 39.3. What kind of autosomal genes would be unsuitable for the experiments with mice described on p. 486? 39.4. Discuss the phenotypes expected among females heterozygous for the X-linked mutant for hemophilia. 39.5. Compare the dosage compensation mechanism in mice with that in Drosophila. 39.6. How can studies of twins test the hypothesis of dosage compensation? 39.7. What explanations can you give for the occurrence of females heterozygous for the X-linked mutant for red-green colorblindness but who are colorblind only in one eye or in portions of one eye? For identical female twins, one colorblind and one not colorblind? 39.8. Ocular albinism is an X-linked mutant that causes the retina of human males to be colorless. What phenotype would you expect in heterozygous females if the mutant is (a) Completely recessi\c' (b) Completely dominant? (c) Partially dominant? Such females actually contain patches of both albino and normally pigmented retina. What bearing does this finding have upon questions (a), (b). (c) above and to vour answers? Regulation of Gene Action — Dosage Compensation 491 39.9. An X-linked mutant in the human male prevents perspiration in any part of the body. What do you suppose would happen to a heterozygote for this mutant who dusted her body with a dry mixture of starch and iodine and then entered a hothouse? 39.10. How can you account for the fact that X0, XX, and XXX women (or XY and XXY men) are not phenotypically equivalent? Does your answer hold true for the female mouse (the X0 is usually fertile)? 39.11. What do you think of the hypothesis that dosage compensation in mammals starts because one chromosome (or part of one) is precociously used as tem- plate? 39.12. What can you conclude from the observation that in human females heterozygous for an abnormal X — the mutant X being either rod shaped but shorter or longer than normal, or ring shaped — the abnormal X appears in Barr bodies more fre- quently than the normal X homolog? 39.13. Discuss the significance of work with the toad Xenopus (discussed on p. 427) relative to dosage compensation. Chapter 40 REGULATION OF GENE ACTION-ITS MOLECULAR BASIS IN HIGHER ORGANISMS N' one of the cases discussed in Chapters 37 through 39 have yet been analyzed far enough to specify the biochemical elements of the control mechanisms involved in the regula- tion of gene activity. In this chapter, addi- tional examples of gene action regulation in higher organisms are explored in an effort to gain some insight into its molecular basis. Dipteran Polynemic Chromosomes ' Evidence obtained from radioautography strongly suggests that the ordinary chromo- some of higher organisms is polynemic or polytenic; that is, it contains more than one double helix of DNA per chromatid.-' Inter- phase chromosomes of larval Diptera which are highly polynemic may be considered merely more extreme examples of a normal tendency toward polynemy. At various times during the growth and differentiation of a dipteran cell containing highly polynemic chromosomes, different cross-bands "puff out" (Figure 40-1 ) and later "unpuff" in a regular sequence. Al- though the sequences vary, they are char- acteristic of different larval tissues. Puffing may be interpreted as a local unwinding of the chromosome and its DNA. In Drosoph- ila and the midge Chironomus, a puff region 1 See W. Beermann (1962). W. Beermann and U. Clever ( 1964), and H. J. Becker ( 1964). -See W. J. Peacock (1963). 492 synthesizes more RNA than an equivalent Qonpuff region; in the salivary gland cells of Rhyncosciara and Glyptotendipes larvae. the amount of DNA synthesized in puffed regions is greater than in equivalent non- pulTcd regions. In certain species of Chironomus, the cyto- plasm of cells in one lobe of the larval sali- vary gland contains (protein secretion?) granules due to a gene located near one end of chromosome IV. Such granules are ab- sent in other species. In cells forming gran- ules, a puff is also found near the tip of chromosome IV but none is found in non- granule containing cells not even those of the same gland. Moreover, larvae produced by an interspecific mating of granule formers and nonformers have some granules and are cytogenetically hybrid; in other words, one homolog does and the other does not have this puff. Injecting larvae with the pupation hor- mone ecdysone causes specific bands to puff and others to unpuff. It has also been shown that the RNA synthesized in a puff does not have A = U or C = G (in terms of quan- tity) and is probably messenger RNA. We may conclude, therefore, that puffing (un- winding DNA) is directly associated with gene activity. In a dipteran salivary cell, only about 20 per cent of the bands ever seem to puff, which indicates that not all genes are functional in every nucleus. Conserved vs. Nonconserved DNA What is the nature and fate of DNA syn- thesized in "excess"? The following evi- dence indicates that DNA can leave the nucleus because it is in excess or for other reasons: 1. In certain organisms (for example, the fungus gnat Sciara) some chromo- somes are regularly eliminated from the nuclei of certain cells. 2. In Drosophila, DNA is extruded from the nuclei of nurse cells in the ovary. Regulation of Gene Action — Molecular Basis in Higher Organisms SEGMENT CHROMOSOME B 493 PUFF POST- PUFF 50 figure 40-1. Puffing and unpuffing in a region of a salivary gland chromosome of Rhyncosciara. {Courtesy of G. Rudkin.) 3. All the cells in certain testicular tubes of the grasshopper Melanoplus cliffer- entialis normally disintegrate and lib- erate large quantities of DNA.:; 4. The oocyte of a newt appears to have in its nucleoplasm and nucleolus DNA — unassociated with the nucleolus or- ganizer— which presumably is not re- tained in the nucleus.4 5. A similar situation is reported "' in the oocyte of the dipteran Tipula: Within this cell's nucleus is a body which con- tains about 50 per cent of all the nu- clear DNA present. Not only is this DNA synthesized at a different time a See A. Lima tie Faria and T. Nordqvist ( 1962). 4 See M. Izawa, V. G. Allfrey. and A. E. Mirskv (1963). 5 See A. Lima de Faria (1962). from the DNA in the chromosomes, but this body and its DNA contents disappear at diplonema. 6. Germinating wheat seeds and growing roots of wheat and corn contain a double-stranded DNA of low molec- ular weight (10"') which shows turn- over; that is, it is metabolically labile. ,; Such DNA differs from stable, high molecular weight DNA by its higher G + C content. Though it may be concluded that DNA is sometimes released to the cytoplasm, no data from the studies mentioned indicate that this material has any of several known or assumed properties of genetic material 0 See M. Sampson, A. Katoh. Y. Hotta, and H. Stern (1963). (94 CHAPTER 40 (replication, mutation, recombination) once it has left the chromosome. As a result, whether this DNA is genetic material re- mains an open question. It should be realized that DNA which leaves the nucleus may serve an extranuclear function quite different from the function DNA performs within the nucleus. For example, nucleus-derived, cytoplasmically located DNA may serve as raw material for synthesis of nuclear DNA Such may be the fate of the DNA in nonfertilizing sperm which disintegrate in the cytoplasm of in- sect eggs fertilized by polyspermy (more than one sperm entering but only one ferti- lizing the egg). In Drosophila. the DNA of the nurse cells which surround the de- veloping oocyte enters the cytoplasm of the oocyte and presumably serves as raw ma- terial for future DNA synthesis. The same fate is suggested for DNA phagocytosed by fibroblasts and white blood cells in vivo, since phagocytosis of DNA by mammalian cells in tissue cultures is followed by the appearance of this DNA in the nucleus. That DNA loss may be associated with dif- ferentiation is indicated by: the loss of some chromosomal material during chromosome diminution in Ascaris; the differential elim- ination of chromosomes by the two sexes of Sciara; and the decrease in DNA in the salivary gland cells of the snail Helix as the secretion product is manufactured. It has also been suggested ' that cytoplasmic DNA may act as a messenger. It is clear from the preceding discussion that all the DNA in a nucleus may not al- ways remain there to perform the usual functions of nuclear genetic material. In this respect, then, there are two kinds of DNA: the one conserved as part of the chro- mosome (which serves as genetic material); the one not conserved (which may or may not be genetic material). In bacteria, the •See P. B. Gahan (1962). and J. J. Holland and B. J. McCarthy (1964). nonconservation of nonintegrating or deinte- grating DNA involved in transformation, conjugation, or transduction has already been mentioned. Lampbrush Chromosomes M Amphibian oocytes have giant "lampbrush" chromosomes (Figure 40-2), whose appear- ance is due to the lateral projection of numerous pairs of loops from the main chro- mosomal axis. Each loop is asymmetric — one end being thicker than the other. In addition to normal pairs of loops, some giant granular loops are found which con- tain a thin axial thread continuous with the main chromosomal axis and having a dense, contorted region at the thinner end of the loop; a coarsely granular matrix at the thicker end. When newt lampbrush chromosomes are exposed to tritium-labeled (H'A) uridine, autoradiographs reveal that incorporation into a pair of giant granular loops occurs in a definite sequence, starting at the thin end of the loop and proceeding around the loop in about 10 days. These results demon- strate the sequential synthesis of RNA by different portions of the loop. Other evi- dence proves that: 1 . The loops contain DNA 2. The RNA synthesis observed is DNA- dependent 3. Agents which inhibit nuclear RNA synthesis (such as actinomycin D) lead to disappearance of the loops (and inhibition of puffing in insect polynemic chromosomes ) . Such results lead to the hypothesis that a loop (like a puff) is a temporarily unwound portion of the chromosome thread. As the thick portion of a loop completes its syn- thetic activity, it presumably winds up and " Based upon work of W. R. Duryee, of J. G. Gall and H. G. Callan ( 1962), and of M. Izawa. V. G. Allfrey. and A. E. Mirsky (1963). Regulation of Gene Action — Molecular Basis in Higher Organisms 495 M. . ' - # ">-- i figure 40-2. Giant "lampbrush" chromo- somes of the amphibian oocyte. A. Unfixed chromosomes of Triturus viridescens in saline solution, phase contrast, 540X- B. Semidia- grammatic view of the central chromomere axis with paired lateral loops. {Courtesy of J. G. Gall.) reforms part of the main nonsynthesizing chromosome axis. At the same time more of the main axial thread unwinds to pro- duce the thin end of the loop which pro- ceeds to synthesize RNA. In any event, it is clear that the morphology of a chromo- somal site is closely related to its ability to synthesize RNA. The large number of loops present in lampbrush chromosomes indicates that a large number of chromosomal sites are syn- thesizing RNA. Although almost all the granules (chromomeres) that occur along the main axis of lampbrush chromosomes have loops, at any given time only about 2 per cent of the bands (comparable to chro- momeres) show puffs in giant polynemic chromosomes. A comparison of the pro- tein, the DNA, and the RNA content of liver, dipteran salivary, and lampbrush chro- mosomes also suggests that the oocyte chro- mosome is synthetically active at many loci. It is noteworthy that the DNA content of a lampbrush chromosome is about four times that of a regular chromosome from a diploid cell of the same newt, and that the nucleo- 496 CHAPTER 40 pkism and nucleoli of the oocyte contains an amount of aonconserved DNA equal to that present in the chromosomes. Most ol the RNA present in the mature amphibian oocyte is synthesized in the lampbrush stage. over 90 per cent being ribosomal RNA. The lampbrush type of chromosome structure has also been reported in the growing oocytes oi animals ranging from mollusks to mam- mals, in the pigeon, in the onion.' and in the Drosophila spermatocyte Y chromo- some.10 It may well be more widespread than previoush suspected. Histones The evidence presented strongly suggests that in a large variety of organisms, chro- mosome uncoiling and consequent DNA un- masking are requisites for the utilization of DNA as a template. The DNA of a mature T phage is highly coiled in the phage head. After T-phage attachment the DNA unwinds and enters the host, where it is immediately available for use as a template. Thus, phage DNA when functional is probably uncoiled and not complexed with protein. These conditions probably also hold true for func- tional chromosomal DNA in E. coli. In contrast to bacterial and phage DNA, the chromosomal DNA of most types of cells is usually joined with basic proteins (such as histone or protamine) to form deoxyribonu- cleoproteins (p. 253). According to the view already presented, perhaps the union of DNA with histone causes coiling, which, in turn, results in gene inactivation. The influence of histones on gene action will be considered further after we have explored some of their properties. When DNA has been removed from nu- cleohistone, the histone can be separated by electrophoresis, ultracentrifugation, and chromatography into numerous subfractions, ;'See B. R. Nebel and E. M. Coulon (1962). 10 See W. Beermann. O. Hess, and G. F. Meyer (1963). indicating that a given type of cell probabl) contains a heterogeneous population of hun- dreds of kinds of histone molecules. These molecules are relatively small — having a molecular weight of 3,500 to 74. 000 — and differ in amino acid composition. One class is relatively rich in lysine (and proline) and poor in arginine; another is relatively rich in arginine and poor in lysine. DNA can combine with these histone fractions to re- constitute nucleohistone whose DNA seems to be as fully complexed with histone as the DNA in native nucleohistone. The deoxyribonuclcohistone reconstituted by combining pure DNA and purified chro- mosomal histones is about 35A in diameter. The histone in such a nucleohistone seems to cover the DNA uniformly; it may be bound spirally around the DNA and occupy one if not both the grooves of the DNA double helix. Other arrangements, how- ever, have not yet been ruled out. Interphase chromatin of somatic cells seems to contain a uniform structural unit. As seen through the electron microscope, this unit is a rod approximately 160A in diameter containing two double helices of DNA (each 20A in diameter). The two double helices are coiled about each other paranemically (p. 268) and the histone seems to occupy the space between and around them.11 Histones are probably syn- thesized in the nucleolus, which is reported to contain ribosomes. When DNA is complexed with histones its melting or denaturation temperature rises. In this respect, then, histones stabilize DNA.1- Lysine-rich histones increase the temperature required to melt half the DNA of a pea from 70 C to 81 C; arginine-rich nucleohistones half-melt their DNA at 71°C. An approximately linear relation exists be- 11 See V. Luzzati, and A. Nicolai'eff (1963). and J. Bonner and P. O. P. Ts'o (1964). ]-See also S. Felsenfeld. G. Sandeen. and P. H. von Hippel (1963). Regulation of Gene Action — Molecular Basis in Higher Organisms 497 tween the lysine content of a histone and the melting temperature of the DNA in a nucleo- histone. DNA fully complexed with prota- mine (salmine or clupein) melts at the same temperature as pure native DNA. Since an appreciable portion of chromo- somal DNA must be used to carry informa- tion for the synthesis of non-histone pro- teins, not every portion of this DNA can be complexed with a specific type of histone. Consequently, a given type of histone must be able to complex with several different sequences of DNA. Although the DNA of duck erythrocytes appears to be almost completely complexed with histone, not every cell has all its DNA histone-complexed. In peas, the fraction of the total DNA not complexed with histone is 5% for the developing cotyledon; 20% for the embryo; 30% for the apices. These results suggest that actively-growing and protein-synthesizing cells have more non- complexed DNA than differentiated cells. The stabilizing effect that histones have upon DNA may involve the suppression of DNA as a template, perhaps by inhibiting DNA strand separation. Isolated nuclei can synthesize comple- mentary RNA by using the DNA as a tem- plate; much of this synthesized RNA is mes- senger RNA; some of it is transfer RNA. In isolated calf thymus nuclei, arginine-rich histone added to the incubation medium not only reduces the uptake of thymidine into DNA, but strongly inhibits the synthesis of RNA; ia lysine-rich histones are much less inhibiting. Selective removal of histones from isolated nuclei results in a two- to four- fold increase in RNA synthesis; the newly made RNA is probably messenger RNA, although its base composition is different from that of the messenger RNA normally synthesized. These results strongly indi- cate that the use of DNA as template for is See V. G. Allfrey, V. C. Littau. and A. E. Mirsky (1963). both DNA and RNA synthesis is inhibited by histones, and the release of DNA from histones can lead to the production of hitherto-repressed messenger RNA. Some evidence 1! suggests that the histone control of genetic activity in vivo is essen- tially preserved not only in isolated nuclei but in isolated chromatin as well. Chro- matin isolated from pea embryos, 20 per cent of which is not complexed with his- tone, is able to carry out DNA-dependent synthesis of RNA from the four usual ribo- side triphosphates through the action of RNA polymerase; removal of histone in- creases RNA synthesis about 500 per cent. Fully complexed DNA in a reconstituted nucleohistone either does not support DNA- dependent RNA synthesis at all or does to a considerably reduced extent. Finally, it should be noted that DNA fully complexed with arginine-rich protamine is fully active in DNA-dependent RNA synthesis. Pea cotyledons synthesize a specific pea seed reserve globulin not produced in other pea plant tissues such as buds or roots. Chromatin isolated from pea cotyledons will, in vitro, produce the messenger RNA used in a ribosomal system to manufacture this globulin. On the other hand, chromatin isolated from pea buds will not lead to synthesis of this protein; removal of histone from pea bud chromatin, however, yields DNA which supports globulin synthesis. Thus, we see that the gene for globulin syn- thesis is normally repressed in the bud by histone.15 Histones may affect gene action by an- other mechanism.1'' When already-formed histones are acetylated at their ends, they permit complementary RNA synthesis in 14 From work of J. Bonner, R. C. C. Huang. R. V. Gilden and co-workers. '•See J. Bonner, R. C. C. Huang, and R. V. Gilden (1963). "See V. G. Allfrey, R. Faulkner, and A. E. Mirsky (1964). 198 ( IIAPTF.R 40 vitro, whereas nonacetylated histones sup- matin contains a smaller percentage of press RNA synthesis. That histone acetyla- acetylated histone than dispersed (function- tion may be a mechanism for controlling ing) chromatin. cene action in vivo is supported by the find- .- D .,,-,,, v/ ,• AmT4 carries the information for the production of several proteins (noted in Chapter 35). Some mutants of protein- specifying genes can direct the synthesis of an altered protein. Some of these altered proteins function as well, or nearly as well, as the unaltered wild-type proteins in hosts grown at normal temperatures (about 25 °C) but become inactive in hosts grown at higher temperatures (about 40°C). Two types of such temperature-sensitive , ts, mutants have been found for deoxycytidylate hydroxy- methylase, dHMCase (Figure 35-1 and p. 451). Although both show reduced dHMCase activity at low temperatures as compared with the wild-type enzyme, their response to a shift (from normal to high) in the temperature at which the infected host is grown is different. In one type of mutant, the altered enzyme is inactivated by heat denaturation at any time during the eclipse 1 The following account is based largely upon the work of R. S. Edgar, M. Susman. G. H. Denhardt. L. Boice, and co-workers. See R. H. Epstein. et al. (1964). 501 period in which the temperature shift is made; the other type of mutant produces an altered dHMCase inactivated only if the temperature shift is made before the first third of the eclipse period and resistant to shifts made later. These results suggest that the former type of enzyme can be denatured by heat after it is synthesized; that the latter type is temperature-sensitive only during its synthesis. It has been found that the ts mutants — really conditional lethals — occur in roughly half of all phage genes. The loci of these genes have been mapped, and their pheno- typic effects studied at chemical, physio- logical, and morphological levels. The re- sults are summarized in Figure 26-4 (p. 343). We see that the circular phage ge- nome is organized into blocks containing genes with common functions. These blocks include the following: Mutants DO Common Characteristics DA Cannot initiate DNA synthesis (hence, normal alleles initiate DNA synthesis) Start DNA synthesis, but cease after a short time DD Delay DNA synthesis "tail fiber" Form (otherwise normal) particles lacking tail fibers "head" Form particles with heads missing (free tails are present in lysates) tail' Form particles with tails missing or incomplete "tail These results strongly suggest that the genes in any block function at the same or nearly the same time, and that the sequence of different blocks of genes may reflect the sequence of events in phage replication and maturation. The thymidylate synthetase (p. 449) locus is apparently an exception to this sequence - since it is located well within the tail fiber region rather than in the '-' According to E. H. Simon and I. Tessman (1963). 502 CHAPTER 41 earl) region as expected. Because the total Dumber of genes in phage is apparently Ear greater than the sum of structural proteins in the mature phage plus the enzymes needed to make phage DNA (Chapter 35), it has been suggested that many of the genes that do not specify the manufacture of phage DNA or protein nevertheless play roles in particle morphogenesis. Developmental Genetics of Amphibia The zygotic nucleus of a fertilized frog egg can be removed by microsurgery.'1 The enucleated cell cannot normally perform the functions of maintenance, growth, division, and differentiation; lacking the normal chem- ical reactions to carry on such functions, it eventually undergoes degeneration. That the cell's metabolic failure is attributable to the loss of the nucleus rather than to the operation itself, is proved by the normal be- havior zygotes show after undergoing sim- ilar operations without being enucleated. More important, however, is the observation that when the same (or a similar) nucleus is replaced in a second operation, normal zygotic activity resumes. Nuclei from blas- tula. gastrula, and later embryological stages can be transplanted into enucleated zygotes. Such experiments reveal that the later the stage supplying the nucleus, the more ab- normal the development, demonstrating that in the course of embryogenesis, nuclei are progressively less able to promote complete, normal development. Nuclei can be transplanted between dif- ferent species of frog. Rana pipiens nuclei which have multiplied in the cytoplasm of R. sylvatica eggs are unable to promote gas- trulation when retransferred to eggs of their own species. Since this limitation persists through repeated transfers to enucleated eggs,4 we can conclude that upon exposure • Based upon work of R. W. Briggs and J. T. King. 4 See J. A. Moore (1960). to cytoplasmic factors the chromosomes' ability to function can become permanently fixed (p. 487). Injection of small amounts of various pro- tein fractions (albumin or his tone) from adult frog liver cells into zygotes of the same species stops cell division and arrests de- velopment at the late blastula stage." At about the same time, chromosomes become essential for further development. Although new cytoplasmic ribosomes do not appear until the later tail-bud stage and new RNA is first detected at the gastrula stage, ,; protein synthesis — using messenger RNA and ribo- somes synthesized before fertilization — be- gins with fertilization. Differentiation and Transcription The similarities and differences among the populations of nucleic acids in various tis- sues of the mouse can be assessed by the formation of double-stranded structures from single-stranded DNA or RNA complexed with single-stranded DNA entrapped in agar. Competition reactions among radioactively labeled and unlabeled sets of molecules fail to show any differences in DNA polynucleo- tide sequences, providing additional evidence for the same DNA content in all somatic cells. On the other hand, large differences are found among rapidly labeled RNA mole- cules isolated from different organs, as ex- pected if differentiation is associated with the production of different populations of mRNA in different kinds of differentiated tissues.7 Other RNA-DNA hybridization experiments s show that the mRNAs from three growth phases of Bacillus subiilis are derived from distinctly different groups of loci, supporting the concept that differential transcription of the genome occurs during morphogenesis. ■See C. L. Markert and H. Ursprung (1963). ,; According to D. D. Brown and J. D. Caston. "See B. J. McCarthy and B. H. Hoyer (1964). "See R. H. Doi and R. T. Igarashi (1964). Gene Action — Growth, Differentiation, and Development 503 Genetic Regulation of Mitosis The genetic control of the structural and physiological features of nuclear and cell division is exemplified in corn and Dro- sophila by mutants which modify spindle shape during meiosis (p. 383) and in a snail by alleles which determine the orientation of the spindle during mitosis. (During snail cleavage, if the spindle becomes oriented one way, a shell with a right-handed coil results; when it becomes oriented the other way, a shell with a left-handed coil is produced.) At present, however, our interest is restricted to the biochemical control of mitosis, espe- cially its genetic basis. The microspore of the lily remains in interphase for several weeks. During this time thymidine kinase activity starts at a specific time and lasts no more than 24 hours. This observation and others indi- cate 9 that thymidine kinase (needed for DNA replication preceding mitosis) is not always present in the cell but is newly formed for this purpose, and destroyed or inactivated after it has completed its function. This cy- clical behavior system resembles induced en- zyme systems in bacteria (p. 458). Should this system prove representative, it would mean that many of the problems of inter- phase and mitosis concern cyclically-regu- lated gene action. Recall (p. 487), how- ever, that during a cell generation not all gene action is cyclical. Viral Regulation of Growth and Differentiation Phages regulate the growth and differentia- tion of their bacterial hosts, at least in part, by the messenger RNA produced using the phage genome. In cells infected by virulent T-even phages, the host materials for DNA and protein synthesis are taken over to syn- thesize viral DNA and protein. Temperate phages and various episomes also turn on 9 See Y. Hotta and H. Stern (1963), and H. Stern and Y. Hotta (1963). or off certain of the host's genes, resulting in modification of cellular growth and/or differentiation. Two cancer-inducing viruses, polyoma and SV-40, are each capable of permanently altering the properties of mouse fibroblast cells grown in tissue culture. The characteristics acquired by the virus-infected cells appear to involve latent properties of the cell. For example, infected cells can regain their ability to synthesize collagen suppressed in the uninfected state. We may therefore hypothesize that certain genes are functionally turned on in virus-infected cells, and that the cellular transformations ob- served are functional and not mutational genetic events.10 Somatic Cell Mating In tissue cultures and sometimes in vivo, successive cell fusions occur between uni- nucleated cells infected with viruses (mea- sles, varicella, herpes, and some mxyo- viruses) and noninfected cells — in a process called polykaryocytosis — to produce giant multinucleated cells which may contain thou- sands of nuclei. These fusions are postu- lated to be associated with an alteration of the cell surface by infecting virus. Poly- karyocytes are characterized by clumped nuclei.11 Several mouse tissue culture lines are unique in that each has some chromosomes with a characteristic morphology. After certain pairs of such cell lines are mixed and grown together, uninucleate hybrid cells are produced whose initial chromosome number is approximately the sum of those of the two parent lines and includes chro- mosomes morphologically characteristic of each line.1- Over the course of several months, clones of these hybrid cells show 111 See N. Sueoka and T. Kano-Sueoka (1964), and G. J. Todaro, H. Green, and B. D. Goldberg (1964). "See B. Roizman (1962). i-See B. Ephrussi and S. Sorieul (1962). 504 (HAP I IK 41 some reduction in chromosome number — probablj because o\' nondisjunction. Al- though the /'// vivo frequency o\' somatic cell mating in mammals is unknown, one possible example of somatic cell mating has been reported in cattle." This case involved a pair of twins, whose members both showed erythrocyte mosaicism due to the presence of genetically different tissues which formed antigenicallv different blood cells. At three years of age one twin had blood 10 per cent of which represented his own genotype and 90 per cent the genotype of his co-twin. At eight years of age, however, this twin had three blood types: the two "parental" types, each representing two per cent, and a "hy- brid'* type representing 96 per cent of the cell population. Somatic cell mating is known to occur in filamentous fungi such as Aspergillus and Penicillium.14 This parasexuality involves the formation of diploid nuclei by rare, prob- ably accidental, nuclear fusions in a multi- nucleate mycelium containing haploid nuclei. The diploid nuclei formed multiply side by side with haploid nuclei and undergo chro- mosome loss or "segregation" by means of mitotic crossing over and or nondisjunction. Further study of somatic cell mating and subsequent chromosome segregation may provide valuable information with regard to differentiation. RNA and Antibodies ' In the rat, the production of antibodies in- volves the following stages. Young plasma- blasts, which divide about every ten hours, have free-floating ribosomes and a poorly developed endoplasmic reticulum. After exposure to antigens these cells begin to synthesize ribosomes and mRNA at a high •'•See W. H. Stone. J. Friedman, and A. Fregin (1964). See also H. Harris and J. F. Watkins (1965i. 14 See G. Pontecorvo (1958). '-See G. J. V. Nossal (1964). rate, and the endoplasmic reticulum under- goes extensive development. Each plasma- blast undergoes about nine successive divi- sions— each successive division taking longer — to produce a clone of mature plasma cells which do not divide. Whereas the plasma- blasts produce a great deal of RNA and pro- tein— mainly structural proteins and enzymes — the mature plasma cell synthesizes mainly protein, 90 to 95 per cent of which is anti- body. The plasma cell nucleus is shrunken and dense and the nucleoli seem to disap- pear. The first antibody molecules a given cell produces have a molecular weight of about a million (19s); later ones are smaller, with a molecular weight of only 160,000 (7s). The small antibodies are 7s gamma globulins — tetramers composed of a pair of identical B chains of 20,000 molecular weight and a pair of identical A chains of 50,000 to 60,000 molecular weight. With rare ex- ceptions each cell makes one type of anti- body, even when other plasma cells in the lymph nodes are synthesizing other anti- bodies. It is not known how antibody is released from the cell to neutralize an anti- gen. Since little or no antigen enters the antibody-producing cell, it is possible that mere surface contact with the antigen is suffi- cient to start a cell into antibody synthesis. It is still too early to specify the detailed roles of the antigen and the genotype in the production of specific antibodies. RNA in Differentiation and Learning Developmental changes can be induced in a growing cell by the introduction of RNA or RNA-containing compounds. Although se- rum albumin is not produced by mouse as- cites tumor cells in vitro, such cells acquire the ability to manufacture this protein after exposure to RNA isolated from normal mouse or calf liver. Using RNA, several strains of cancer cells can be induced to synthesize such enzymes as tryptophan pyr- Gene Action — Growth, Differentiation, and Development 505 rolase and glucose-6-phosphatase."; The type of protein synthesized by the recipient cell seems to have some if not all of the specificity produced by the RNA-donor cell. Some of the introduced RNA seems to func- tion as messenger RNA for at least an hour. When the responses to stimuli do not in- volve learning, the neurons of rats show an increase in nuclear RNA but no shift in base ratios. However, when rats are placed in a learning situation (involving balance), not only does nuclear RNA increase but also the A/U ratio increases and C decreases. The RNA content and base ratios can be studied in single cortical neurons of right- handed rats forced to use the left hand to obtain food. Neurons serving the learning side not only show an increased RNA con- tent but an increased — — — ratio and a C + U G + C decreased ratio as compared with A + U that of control neurons in the contralateral part of the same cortex. In Parkinson's disease profound changes in RNA base ra- tios arise in the nervous tissue; interference with RNA synthesis in the brain sometimes impairs learning in rats. These results strongly suggest that the learning process is associated with production of messenger RNA.17 Both protein and RNA synthesis can be stimulated by estrogens (in uterine tissue); testosterone (in the prostate gland); and by a flowering hormone, presumably a sterol (in a plant bud). The flowering hormone also reduces the histone to DNA ratio, which suggests that steroids can bring about the removal of histone from chromatin.18 Low concentrations of thyroxine in a cell- free system prepared from rat liver will stimulate the incorporation of amino acids 16 See M. C. Niu, C. C. Cordova, L. C. Niu, and C. L. Radbill (1962), and A. H. Evans (1964). 17 See H. Hyden and E. Egyhazi (1963, 1964). 18 See reference to J. Bonner and P. O. P. Ts'o on p. 499. See also T. H. Hamilton (1964). into protein. This effect is dependent upon the presence of mitochondria and an oxidiz- able substrate, is independent of DNA-de- pendent RNA polymerase activity or mRNA synthesis, and seems to involve the transfer of sRNA-bound amino acid to protein syn- thesizing ribosomes.1'' Differentiation in Paramecium -° Although Paramecium is normally a single animal, or singlet, double animals, or dou- blets, occur. Singlets and doublets repro- duce true to type through numerous fissions. A doublet can also conjugate with two sin- glets and each singlet exconjugant regularly produces singlet clones and the doublet exconjugant, a doublet clone. The singlet- doublet difference cannot be due to micro- nuclear genes since exconjugants are iden- tical in this respect. This same phenotypic result is obtained even when a cytoplasmic bridge lasts long enough to permit an ex- tensive exchange of cytoplasm. Conse- quently, the difference between doublet and singlet does not have a basis in any cyto- plasmic component free to migrate. Other evidence seems to exclude the macronucleus from being involved. The only portion of the cell unaccounted for then is the immo- bile 0.5 micron-thick outer layer of ecto- plasm, the cortex. In one experiment, after cytoplasmic bridge formation between a singlet and doublet, a rare free singlet exconjugant was found bearing a conspicuous extra piece of cortex. The doublet exconjugant, on the other hand, showed a corresponding nick in its cortex. The extra piece in the sin- glet later flattened out and, after fission, one of the two daughter cells gave rise to a clone phenotypically intermediate between singlets and doublets. This natural grafting of only a small piece of a Paramecium's 19 See L. Sokoloff, C. M. Francis, and P. L. Camp- bell (1964). -°See T. M. Sonneborn (1963, 1964). 506 CHAPTER 41 oral segment gave rise to a strain having a complete extra oral segment including an extra vestibule, mouth, and gullet. Other studies reveal that various experimental modifications of visible cortical organization are perpepuated during cell reproduction, and that certain visible cortical structures initially absent, do not arise de novo. These and other results establish the im- portance of the cortex in differentiation. The cortex is not completely autonomous, however, since some nuclear genes are known to determine visible cortical struc- tures or their morphogenesis. As already mentioned, a small additional piece of cor- tex can give rise to cortical changes of greater degree. Clearly then, the nature and action of the cortex is dependent not only upon its own composition but upon nuclear genes and their products as well as metabolism in general. It should be noted that double-stranded DNA has been re- ported in human erythrocyte membranes. This DNA has a molecular weight of about I0,; and a G + C content of approximately 39-429f. Its homogeneity (and its possi- A 4- T ble higher r ratio) suggest that this C + G is not merely adsorbed DNA. At present, the mode of operation of the cortex can only be described in general, largely specu- lative, terms. "The much more difficult task for the future is to define and specify in molecular terms the decisive structures, gradients, and inductor-response systems and to reveal how specific absorption, ori- entation, and activation of migratory mole- cules leads to visible morphogenesis and ge- netic stability of cell organization." (T. M. Sonneborn, 1963). -1 See L. Philipson and O. Zetterqvist (1964). SUMMARY AND CONCLUSIONS The synthesis of mature ^>T4 progeny is regulated by blocks of parental viral genes which function at the same or nearly the same time and which are arranged in the circular linkage map in a sequence reflecting successive stages of phage morphogenesis. Nuclear transplantation, somatic cell mating, antibody-antigen, and biochemical em- bryological investigations are revealing the genetic and molecular bases of differentia- tion and development. Studies of thymidine kinase indicate that mitosis involves cyclical gene action. Base- specific RNA plays an intracellular role in learning and an intercellular role in differ- entiation. Steroids seem to be involved in RNA production and histone distribution. Although the chemical mechanism is unknown, studies of Paramecium reveal the importance of the cortex in differentiation, and emphasize that morphogenesis depends upon both the nuclear genetic material and the remainder of the protoplasmic and metabolic environment. REFERENCES Cytogenetics and Developmental Genetics, Amer. Zool., 3 (No. 1), 1963. Demerec, M., "Clustering of Functionally Related Genes in Salmonella typhimurium" Proc. Nat. Acad. Sci., U.S., 51:1057-1060, 1964. Differentiation and Development, Boston: Little, Brown, 1964; and J. Exp. Zool., 157, No. 1, 1964. Doi, R. H., and Igarashi, R. T., "Genetic Transcription during Morphogenesis," Proc. Nat. Acad. Sci., U.S., 52:755-762, 1964. Gene Action — Growth, Differentiation, and Development 507 Ephrussi, B., and Sorieul, S., "Mating of Somatic Cells In Vitro," pp. 81-97, in Ap- proaches to the Genetic Analysis of Mammalian Cells, Merchant, D. J., and Neel, J. V. (Eds.), Ann Arbor: Univ. Michigan Press, 1962. Epstein, R. H., et al., "Physiological Studies of Conditioned Lethal Mutants of Bac- teriophage T4D," Cold Spring Harb. Sympos. Quant. Biol., 28:375-394, 1964. Evans, A. H., "Introduction of Specific Drug Resistance Properties by Purified RNA- Containing Fractions from Pneumococcus," Proc. Nat. Acad. Sci., U.S., 52:1442- 1449, 1964. Hamilton, T. H., "Sequences of RNA and Protein Synthesis During Early Estrogen Action," Proc. Nat. Acad. Sci., U.S., 51:83-89, 1964. Harris, H., and Watkins, J. F., "Hybrid Cells Derived from Mouse and Man: Artifi- cial Heterocaryons of Mammalian Cells from Different Species." Nature. London. 205:640, 1965. Hotta, Y., and Stern, H., "Molecular Facets of Mitotic Regulation, II. Factors Under- lying the Removal of Thymidine Kinase," Proc. Nat. Acad. Sci., U.S., 49:861-865, 1963. Hyden, H., and Egyhazi, E., "Glial RNA Changes During a Learning Experiment in Rats," Proc. Nat. Acad. Sci.. U.S., 49:618-624. 1963. Hyden, H., and Egyhazi, E., "Changes in RNA Content and Base Composition in Cortical Neurons of Rats in a Learning Experiment Involving Transfer of Hand- edness," Proc. Nat. Acad. Sci., U.S.. 52:1030-1035, 1964. Locke, M. (Ed.), Cytodifferentiation and Macromolecular Synthesis, New York: Academic Press, 1963. Markert, C. L., and Ursprung, H., "Production of Replicable Changes in Zygote Chro- mosomes of Rana pipiens by Injected Proteins from Adult Liver Nuclei," Develop. Biol., 7:560-577, 1963. McCarthy, B. J., and Hoyer, B. H., "Identity of DNA and Diversity of Messenger RNA Molecules in Normal Mouse Tissues," Proc. Nat. Acad. Sci., U.S., 52:915— 922, 1964. McElroy, W. D., and Glass, B. (Eds.), A Symposium on the Chemical Basis of De- velopment, Baltimore: The Johns Hopkins Press, 1958. Moore, J. A., "Serial Back-Transfers of Nuclei in Experiments Involving Two Species of Frogs," Develop. Biol., 2:535-550, 1960. Niu, M. C, Cordova, C. C, Niu, L. C, and Radbill, C. L., "RNA-Induced Biosyn- thesis of Specific Enzymes," Proc. Nat. Acad. Sci., U.S., 48:1964-1969, 1962. Nossal, G. J. V., "How Cells Make Antibodies," Scient. Amer., 211 (Dec.) : 106—1 15, 154, 156, 1964. Philipson, L., and Zetterqvist, O., "The Presence of DNA in Human Erythrocyte Mem- branes," Biochim. Biophys. Acta, 91:171-173, 1964. Roizman, B., "Polykaryocytosis Induced by Viruses," Proc. Nat. Acad. Sci., U.S., 48: 228-234, 1962. Sokoloff, L., Francis, C. M., and Campbell, P. L., "Thyroxine Stimulation of Amino Acid Incorporation into Protein Independent of Any Action on Messenger RNA Synthesis," Proc. Nat. Acad. Sci., U.S., 52:728-736, 1964. Sonneborn, T. M., "Does Preformed Cell Structure Play an Essential Role in Cell Heredity?" Chapter 7, pp. 165-221, in The Nature of Biological Diversity, Allen, J. M. (Ed.), New York: McGraw-Hill. 1963. 50S | I, tPTER 41 Sonneborn. T. M.. "The Differentiation ol Cells," Proc. Nat. Acad. Sci., U.S., 51: 9 1 5-929. 1964. Stern. H.. and Hotta. Y., "Regulated Synthesis of RNA and Protein in the Control ol Cell Division," Brookhaven Sympos. Biol.. 16:59-72. 1963. Stone. W. H.. Friedman, J., and Fregin. A.. "Possible Somatic ( ell Mating in Twin Cattle with Erythrocyte Mosaicism," Proc. Nat. Acad. Sci.. U.S., 51:1036-1044. 1 964. Sueoka. N., and Kano-Sueoka, T., "A Specific Modification of Leucyl-sRNA of Escherichia coli alter T2 Infection," Proc. Nat. Acad. Sci., U.S., 52:1535-1540, 1 964. Symposium on Macromolecular Aspects of the Cell Cycle, J. Cell. Comp. Physiol.. 62 (Suppl. 1 ), 1963. Todaro. G. J.. Green. H.. and Goldberg. B. D.. "Transformation of Properties of an Established Cell Line by SV40 and Polyoma Virus," Proc. Nat. Acad. Sci., U.S., 51:66-73, 1964. Waddington, C. H., New Patterns in Genetics and Development, New York: Columbia University Press, 1962. Weiler, E., "Immunologically Determined and Competent Cells Are Affected Differently by Actinomycin D," Science, 144:846-849, 1964. QUESTIONS FOR DISCUSSION 41.1. Apply the concept of the operon to the development of <^>T4. 41.2. The work on nuclear transplantation by R. W. Briggs and T. J. King, J. A. Moore, and others has proved that development sometimes involves irreversible changes in the nucleus. Should such changes be termed mutations? Explain. 41.3. In the snail Limnea peregra, self-fertilization of pure-line individuals whose shell coils to the right, dextrally, or to the left, sinistrally, produces progeny all of which coil as their parents. A cross of dextral 9 by sinistral $ yields all dextral Fj which, when self-fertilized, yield all dextral progeny in F2- After self-fertiliza- tion, however, % of the F2 give rise to dextral F3 and Vi of the FL, to sinistral F3. The reciprocal cross, dextral 6 by sinistral 9, yields all sinistral F,. The Ft produces F;_» and F:{ phenotypically the same as the reciprocal cross. Give a genetic explanation for these results. Are cytoplasmic genes involved? Explain. 41.4. T. Yamada has found that isolated prospective ectoderm gives rise only to epi- dermal cells when cultured in vitro in standard medium, but forms mesodermal tissues if a protein fraction from bone marrow is added to the medium. To what can you attribute these results? 41.5. Suggest an explanation for the dedifferentiation which chondrocytes in vertebral cartilage undergo when grown in vitro. 41.6. Defend the following statement of J. D. Ebert (1963): "The principal theme, coursing through and underlying research in embryology today, is the impact of genetics on development. More than at any time in the past half-century, molec- ular embryology is clearly the logical extension of molecular genetics." 41.7. What conclusions can you draw from the following evidence concerning Down's syndrome as found in the Victoria region of Australia: The incidence of the syndrome varies year by year, with peaks occurring about every five years; about 40% of the cases are clustered geographically, more occurring in urban than in rural areas. Chapter 42 THE ORIGIN AND EVOLUTION OF GENETIC MATERIAL I n considering the nature and effects of genes, the origin and evolution of genetic material has been neglected. Before taking this up in detail, it would be desirable to reconsider the nature of presently-known genetic ma- terials, DNA and RNA. The replication of either type of nucleic acid involves the use of a single or double polynucleotide strand as a template for com- plementary monomers subsequently joined to form complementary polynucleotides. The nucleic acid properties responsible for its ability to serve as a template must in- clude a specific physical configuration of linearly-arranged monomers as well as a specific pattern of net electrical charges. Although the utilization of the nucleic acid template for the formation of a complement is a relatively passive process with regard to the polynucleotide strand, it is an active process if we consider the highly specific action of nucleic acid polymerase or syn- thetase. Since the nucleic acid fiber which serves as a template is mostly passive, it is not surprising that nucleic acid can be used as a template by different enzymes, pro- vided that the raw materials collected on the template have suitable physical and elec- trical properties. That nucleic acids are used as templates for the formation of poly- mers not their complements, is exemplified by DNA used as template by DNA-depend- ent RNA polymerase to make RNA, and by RNA used in vitro and in vivo as a tem- 509 plate to make DNA. Whether nucleotides other than those in DNA and RNA, or still other substances, make use of the nucleic acid template in a similar manner is yet to be determined. (Note that the basic pro- teins in nucleoproteins may be associated with nucleic acids through a template mechanism, at least in part.) The simplest and the broadest working hypothesis, there- fore, is that all the functional characteristics of genes depend upon the linear sequence of nucleotides and upon the ways this poly- mer is used as a template by various sub- stances and enzymes; in other words, the only junction of genetic material is to serve as a template. Although nucleic acid is self-replicating, the process is apparently not accomplished in one step. In fact, two replications seem to be required before a given strand can be duplicated. The first replication produces a complementary strand; the second repli- cation produces a copy of the first strand. We are probably justified in thinking of self-replication in this way for the follow- ing reasons: 1 . The self-replication of single-stranded DNA and RNA viruses must be con- sidered a two-step process. 2. Double-stranded nucleic acid may preferentially or exclusively replicate one of the complements. 3. One of the two strands in a double helix may be defective (mutant) and incapable, at least in some places, of both replication and self-replication; however, its normal, complementary chain would be capable of both. Consequently, we can now define genetic material as any template whose use eventu- ally results in its self-replication and which either retains this ability after mutation or is a mutant of a template which has this ability. We can also consider genetic ma- terial to be any substance which produces 510 CHAPTER 42 the same phenotypic effects as known ge- netic material. Should pure DNA in vitro be considered to be genetic material? Originally, the identification of genetic material depended upon its presence in organisms and its pro- duction of a phenotypic effect; we can now dispense with these requirements. Pure vi- rus DNA in a test tube should be consid- ered genetic material even though no longer within an organism, recombining, mutating, replicating, or performing any phenotypic function. This statement is valid on the basis that such DNA either is known to or expected to possess genetic properties when introduced into an organism. DNA syn- thesized in vitro is physically and chemically almost identical to chromosomal DNA. It is capable of: 1. Replicating itself and some of its vari- ants 2. Undergoing strand separation and re- combination 3. Producing a phenotypic effect by ge- netic transformation. We may conclude, therefore, that DNA synthesized in vitro also fulfills the require- ments of our definition for genetic material. The simplest biological synthesis of DNA or RNA requires the presence of nucleoside triphosphates; an enzyme (DNA polymer- ase or RNA synthetase); and water — at the correct pH — containing the ions necessary to activate the enzyme. It seems unlikely that the first gene-like material had these numerous and specific requirements for rep- lication. We may hypothesize that in the course of evolution, the first really success- ful genetic material resembled RNA rather than DNA; although DNA (by lacking an O at the 2' position) is more stable than RNA as a template, it is the RNA polyribotide which serves as a carrier (as sRNA) for amino acids. This amino-acid carrying ability may have led to the synthesis of the polymerizing enzymes essential for rapid gone synthesis, and lor the synthesis of other proteins (including enzymes) which stab- ilize and preserve the chemical integrity of the genetic material. The preceding discussion leads to the question of the origin of genetic material on earth. Were RNA (and or DNA) and proteins present during the early stages of genetic evolution? Would their presence in early evolution correlate with existing knowledge about the course of chemical evolution on earth? Eras of Chemical Evolution ' Era I. We now understand that at an early prebiotic stage in its history — some 4 bil- lion years ago — the earth had a reducing atmosphere rich in water, hydrogen, meth- ane, and ammonia, but poor in free oxygen and carbon dioxide. Using mixtures of these and similar compounds predicated to have been present in such a reducing at- mosphere plus a source of energy (such as electrical discharges, sunlight and ultravio- let light, microwaves, ultrasonic vibrations, heat, high energy electrons, X rays, and proton irradiation), it is possible to produce in the laboratory a large number of simple radicals and organic molecules. Moreover, since a projectile propelled through a gas and into a liquid can cause the formation of a large number of complex chemical com- pounds, it is very likely that in prebiotic times chemosynthesis was also induced by meteorites.- Some of the compounds syn- thesized experimentally in a "primitive" atmosphere include alanine, glycine, glutamic acid, aspartic acid, acetic acid, formic acid, proprionic acid, lactic acid, succinic acid, some fatty acids, urea, some sugars, phos- phoric acid, adenine, and uracil. Although we are not yet able to determine which of 1 See article by H. Gaffron in M. Kasha and B. Pullman (1962). -See A. R. Hochstim (1963). The Origin and Evolution of Genetic Material 511 the energy sources were primarily respon- sible, there cannot be any doubt that syn- thesis and accumulation of a great variety of organic molecules took place in the oceans, making an "organic soup.,, During this era, whose length has not been estab- lished, most of the free hydrogen escaped from the earth's atmosphere. Era II. Its atmosphere was also a re- ducing one with only traces of free oxygen. Initially, the same energy sources were available for chemosynthesis in this era as in Era I. As traces of oxygen escaped to the atmosphere, ultraviolet rays from the sun converted the oxygen into ozone. Since ozone absorbs ultraviolet light, the ozone layer in the atmosphere acted as a blanket so that the main chemosynthetic energy from the sun was visible light and heat. A study of the comparative biochemistry of present higher plants and animals, bac- teria, and many viruses, shows that all are intimately associated with the same 20 or so amino acids. Accordingly, protein and nucleic acid are, perhaps, the most durable chemical features of the earth, having ex- isted for more than a billion years. In the presence of excess aspartic acid and glu- tamic acid, temperatures of 200°C or less can be used 3 in a dry heat synthesis to pol- ymerize amino acids into proteinoids, poly- mers containing, in peptide linkage, all or most of the amino acids common to pro- teins. Proteinoids are linear polymers with a molecular weight of up to 10,000, show weak catalytic activity, and, for the most part, are indistinguishable from natural pro- teins or polypeptides of similar size. Al- though proteinoids are nonantigenic, in hot water they tend to form spheres about two microns in diameter. Since the spheres swell and shrink as the sodium chloride con- centration of the medium is changed, we are reminded of osmotic behavior. Sometimes 3 See S. W. Fox (1960, 1964). such microspheres undergo a kind of fission and show a double-layered outer membrane in electron micrographs. Homopolymers and copolymers of certain amino acids can also be produced by dry heat. On the basis of such evidence and rea- soning, it is expected that during Era II complex organic substances were synthe- sized— polypeptides, nucleotides, carotenes, polyphosphates, pigments, and porphyrins. It is also expected that adsorption and prim- itive catalysis occurred involving surfaces of clays and or polypeptides. Era III. This era is assumed to be mainly anaerobic with only trace amounts of free oxygen and some carbon dioxide. It is suggested that during this period syn- thetic cycles evolved (as did specific catal- ysis and photochemistry) on the surfaces of large organic molecules. As the last evo- lutionary step of this era, primitive enzymes and genes also arose, leading to the first or- ganism. Some recent research and speculation 4 may throw more light upon the evolution and interdependence of polynucleotides and proteins. As mentioned, it is likely that proteins and mononucleotides were already present at the start of Era III. In the pres- ence of dehydrating agents, water is re- moved and nucleotides are joined to form polynucleotides with molecular weights of up to fifty thousand. Note that such a pol- ymer is made without the use of an enzyme. The rate of such a nonenzymatic synthesis of polyuridylic acid has been found 5 to in- crease more than tenfold in the presence of polyadenylic acid, which suggests that the latter homopolymer can serve as a template during the nonenzymatic synthesis of the for- mer homopolymer. A hypothetical scheme has been proposed to form DNA by the re- 4 See article by A. Rich in M. Kasha and B. Pull- man (1962). 5 By G. Schramm. H. Grotsch. and W. Pollmann (1961). 512 ( IIAI' I I K 42 action oi glyceraldehyde, acetaldehyde. am- monia, oxaloacetic acid, glycine, and t'ormyl residues. Molecular evolution leads not only to greater complexity but to greater stability of molecules. Accordingly, separate pro- tein and polynucleotide chains might join to form a more stable complex. (We know that a DNA-histone complex stabilizes DNA; double-strandedness and polynemy could also be nucleic acid stabilizing fac- tors.) The final protein-nucleic acid com- plex, however, need not have started with a protein or a polynucleotide. Single nu- cleotide-single amino acid units might have occurred which polymerized to form a poly- peptide first and a polynucleotide later or the reverse. Regardless of the manner in which the nucleic acid-protein complex evolved, such a complex must already have entailed a primitive code by which nucleo- tides and amino acids code for each other. Since nucleic acids make better templates for replication than proteins, the nucleic acid portion of a nucleoprotein became ge- netic material. As chemical evolution pro- ceeded, the number of nucleotides specify- ing a single amino acid could have increased from one to the three — the size of our pres- ent codon. Even if the evolution of proteins and pol- ynucleotides was independent for some time, it seems clear that the two substances be- came interdependent in their later evolution. In view of the relative chemical inactivity of nucleic acids, we may hypothesize that one primary result of their evolution was the stabilization of enormous numbers of protein molecules and protein cycles which arose in Era III; a second primary re- sult of nucleic acid evolution led to pro- tein replication by ribonucleic acid. In other words, chemical evolution seems to have been largely a matter of protein evo- lution during which nucleic acids came to serve as stabilizers of and templates for pro- tein synthesis. Such an evolution would be expected it nucleic acids were first formed in an environment whose organic compo- nents were largely protein. Since nucleic acids do not include Sulfur (and many other elements) in their basic makeup, they lack the proteins' chemical drive for stability. The stabilization of proteins was probably further enhanced by retaining the infor- mation in DNA, rather than RNA whose use became more and more restricted to the translation process. As the nucleic acid transcription and translation processes evolved, it surely became advantageous to step up the rate of these reactions through the use of nucleic acid polymerizing en- zymes. It also became advantageous to protect nucleic acids from peroxides formed in the environment by radiation. Thus, it is likely that the nucleic acids which en- coded catalase protected the nucleic acid directly and protein synthesis indirectly. Since prebiotic and biotic chemo-evolution is largely describable in terms of protein structure and function, the present view — that genetic nucleic acids played a lesser role — is already generally accepted and not at all novel. The subservient role of nu- cleic acids is generally evidenced in present day organisms not only in the requirement of GPP in protein synthesis and of APP for the transport of energy, but also of UPP and CPP for the transport of monomers in the synthesis of carbohydrates and lipids.'1 Because of the intimate relationship be- tween amino acids and genetic material, we need to learn more about the evolution of all kinds of organic compounds, especially energy-rich compounds (such as ATP), cat- alysts (such as iron-containing compounds), and energy-capturing compounds (such as chlorophyll). "See R. E. Eakin (1963). The Origin and Evolution of Genetic Material 513 In considering the origin of the first gene, we should keep in mind the possibility that its nongenic predecessor might have been capable of self-replication to some degree, but might not have been able to replicate any of its mutant forms. The search for information about nongenetic systems with some but not all the properties of genetic material is clearly highly necessary and de- sirable. Subgenic chemicals may occur in present- day cells. Constituents of the cytoplasm which contain DNA and are able to self- replicate include chloroplasts, mitochondria, the centriole, and the kinetosome. If the DNA in these structures is mutable and still able to self-replicate, it can be classed as cytoplasmic or extranuclear genetic mate- rial. Experimental study of these and other organelles is expected to reveal details of their chemistry. We would also like to know a great deal more about the synthesis of RNA genes; how the metagonic RNA of Paramecium replicates in Didinium; and whether it is still self-replicating after mu- tation. Answers to such questions would be valuable in enabling us to speculate more fruitfully upon the nature of pregenetic and primitive genetic materials. Eras IV and V. The early environment in these eras was probably much the same as in Era III, but after an increase followed by a decrease in carbon dioxide, a large in- crease in the amount of free oxygen took place. Though we do not yet have suf- ficient information to decide precisely which pathways led to the chemical evolution of the first gene, we do have some evidence concerning the subsequent history of genes in organisms. The only genetic material found exclusively in free-living organisms today is DNA; this substance is found in all such organisms, be they unicellular or multicellular, plant, animal, or microorgan- ismal. Whether or not types of genes other than DNA and RNA have ever existed, DNA genes must have a definite advantage for survival — after all, they have persisted as the main genetic material for about a billion years, approximately the period the evolution of plants and animals have been separate. It is likely that the formation of chromosomes with telomeres, centromeres, and polynemy, as well as the establishment of special methods of separating daughter and homologous chromosomes (by mitosis and meiosis) and of recombining them (fertilization) were innovations involving DNA which occurred some time prior to the divergence of the plant and animal kingdoms. Evolution must have led to the transcrip- tion of only one of the two complementary polynucleotide strands in producing com- plementary RNA and viruses with single- stranded nucleic acids, and must have oc- curred in the genetic code and the appara- tus for translation. The quantity of DNA per organism and the basic proteins which regulate gene activity as well as the mate- rials which serve to regulate mutability must also have evolved. We may reasonably suppose that an evo- lution in gene activity also took place. On primitive earth it is likely that large amounts of different, more or less complex, organic materials accumulated and remained unde- graded before the advent of the first gene- containing organisms. As the organisms used up these resources in their metabo- lism, however, there would have been a se- lection in favor of those mutants capable of synthesizing such organic materials from simpler organic, or from inorganic, compo- nents.7 This hypothesis means that natural selection acted in favor of those mutant genes which specified the synthesis of a com- ponent no longer available in the environ- "See N. H. Horowitz (1945). 514 CHAPTER 42 ment. [ndependence of the environment would also have been advanced by the phys- ical association of genes involved in different portions o\' a given biochemical sequence. this independence leading eventually to the selection of mutant genes whose function, other than self-replication, was to regulate the functioning of other genes. Thus, in addition to genes for structure, evolution might well have produced genes for regula- tion— operator and regulator genes — and genes for synthesizing specific basic proteins for the regulation of chromosome coiling, replication, and functioning. Cosmic Chemo-evolution In our search for information regarding either pregenetic. preorganismal evolution or postgenetic, postorganismal evolution, we do not have to confine ourselves to this planet. The universe is about ten billion years old; the earth is roughly half this age. Because the universe contains an infinite number of stars (suns) with planets, there must be vast numbers of suns the size of our own that have planets about the same size as the earth and at comparable distances from their suns. Some of these planets are surely younger, others surely older than our own. What is the possibility that a chemical and biological evolution similar to ours took place on other planets? The answer to this question de- pends, of course, upon their chemical com- position. Most of the universe is hydrogen and helium (most of the earth's hydrogen es- caped from our atmosphere in Era I, as al- ready mentioned). Of the remaining ele- ments, the universe has abundant oxygen and nitrogen and is, in fact, richer than the earth in carbon — the element essential for organic compounds which have played such an integral role in chemical and biological evolution on earth. It is therefore likely that numerous places in the universe do exist where a chemical evolution of biological in- terest might have been successfully initiated. Since the relative scarcity of carbon makes the earth a rather poor place for such an evolution (which nevertheless occurred), most surely the universe contains numerous planets in early stages of chemical evolution, early stages of biological evolution, as well as planets older than our own. which very probably have more advanced types of or- ganisms. Evidence has been obtained for the pres- ence of organic radicals such as CH, CN, CC, and CO in comets, and for organic mole- cules of an asymmetric type on Mars. As- tronomers have also reported variations in the color and texture of Mars with changes of season, which strongly suggest that Mars, with an atmosphere thinner than the earth's, contains appreciable quantities of organic matter, although we are not yet able to de- termine whether their origin is preorganismal or organismal. Further information about the chemistry of our sun and its planets will undoubtedly be provided by telescopes orbiting far into and above our atmosphere. Plans for in- terplanetary research now underway include sending additional instruments to or near various planets in our solar system. Such missions will be designed to record the detailed chemistry of our neighboring planets and, of course, to detect the presence of organic compounds, of organisms, and of DNA and RNA. We have already sent radio signals into space in an attempt to contact other organisms capable of receiving and or replying. In any space mission it is of utmost im- portance to avoid the accidental transplanta- tion of terrestrial genotypes to other planets; if a single bacterium such as E. coli were placed on a planet containing a suitable medium, its progeny would occupy a volume the size of the earth in about 48 hours. Such The Origin and Evolution of Genetic Material 515 an unscheduled transplantation would doubt- less be disastrous to any future plans for studying either the preorganismal evolution of organic compounds or any indigenous organisms. As a safeguard against such con- tamination, objects sent beyond our atmos- phere are sterilized. Recall also that the impact of a missile on a planet may cause the production of organic chemical sub- stances. Which heavenly objects likely to be in- vestigated in the near future are interesting from the point of view of preorganismal and organismal evolution? We have already mentioned Mars. Consider Venus, whose unknown surface is hidden completely by an opaque, highly reflecting cloud layer con- taining abundant carbon dioxide and water. Although estimates of Venus' temperature vary widely (its surface is usually considered to be dry and hot), we can assume that or- ganic compounds have evolved there even if biological evolution has been impossible. After studying its chemistry in sufficient de- tail, perhaps we might wish to colonize Venus, first by placing a chlorophyll-con- taining microorganism in its outer atmos- phere. In a short time, such an organism, by utilizing great quantities of atmospheric components for growth and reproduction, might radically change the climate of Venus. Our own satellite, the moon, has no at- mosphere and probably no water. There- fore, the presence of earthlike life there today is almost out of the question. The moon, however, may be as old as the earth and may have had an organic and even a bio- logical evolution similar to our own before losing its atmosphere. So it will be interest- ing to analyze samples of the lunar surface and, particularly, its subsurface material. It has been suggested that the moon acts as a gravitational trap for fossil spores drifting between planets. Although improbable, the very possibility of an interplanetary gene flow is too important to ignore in the exploration and exploitation of space. Planetary re- search has many motivations; but the search for evidence of chemical evolution, DNA and RNA, and organisms — life of any type — would seem to be among the most signifi- cant. SUMMARY AND CONCLUSIONS The earth has undergone a chemical evolution. This process resulted in the synthesis of most types of compounds and cycles of synthesis found in present-day organisms. It also resulted in the production of proteins, polynucleotides, and nucleoproteins. These latter compounds evolved to a stage where a nucleic acid was able to replicate itself as well as some of its changed forms and so could be identified as genetic ma- terial. Chemo-evolutionary drive on earth is based largely upon protein since nucleic acids provide much less chemical diversity; therefore we hypothesize that nucleic acid evolution (which presumably went from RNA to DNA) was subservient to evolution directed primarily toward the stabilization and synthesis of protein. DNA has been the primary genetic material on earth for about a billion years. During this time DNA and its associated materials have undergone a structural evolu- tion leading to the establishment of chromosomes and mechanisms for genetic recom- bination and regulation of mutability. Also, genes have probably undergone a func- tional evolution, which proceeded from genes which serve structurally (specifying the synthesis or organization of nongenetic compounds) to those which serve functionally (regulating gene action). 516 CHAPTER 42 REFERENCES Abelson, P. H.. "Extra-Terrestrial Life," Proc. Nat. Acad. Sci., U.S., 47:575-581, 1961. Blum, H. F., "On the Origin and Evolution of Living Machines," Amer. Sci., 49:474- 501, 1961. Calvin, M.. "The Origin of Life on Earth and Elsewhere," Ann. Int. Med., 54:954-976, 1 96 1 . Clark, F., and Synge, R. I.. M. (Eds.), The Origin of Life on the Earth, New York: Pergamon Press. 1959. Eakin. R. E., "An Approach to the Evolution o\ Metaholism," Proc. Nat. Acad. Sci., U.S.. 49:360-366, 1963. Fox. S. W.. "How Did Life Begin." Science. 132:200-208, 1960. Fox, S. W., "Experiments in Molecular Evolution and Criteria of Extraterrestrial Life," Bio-Science. 14:13-21, 1964. Fox, S. W. (Ed.). The Origins of Prebiological Systems, New York: Academic Press, 1 964. Green. D. E.. and Hechter, O.. "Assemhly of Membrane Subunits," Proc. Nat. Acad. Sci.. U.S.. 53:318-325. 1965. Hochstim, A. R.. "Hypersonic Chemosynthesis and Possible Formation of Organic Compounds from Impact of Meteorites on Water," Proc. Nat. Acad. Sci., U.S.. 50:200-208, 1963. Horowitz, N. H., "On the Evolution of Biochemical Syntheses," Proc. Nat. Acad. Sci., U.S.. 31 : 153-157, 1945. Huang. S.-S., "Life Outside the Solar System." Scient. Amer., 202 (No. 4) : 55-63, 1960. Kasha, M.. and Pullman, B. (Eds.), Horizons in Biochemistry, New York: Academic Press, 1962. Keosian, J., The Origin of Life, New York: Reinhold Publ. Corp., 1964. Lederberg, J., "Exobiology: Approaches to Life Beyond the Earth," Science, 132: 393-400. 1960. Lederberg, J., and Cowie. D. B., "Moondust," Science, 127:1473-1475, 1958. Miller, S. L., and Urey, H. C. "Organic Compound Synthesis on the Primitive Earth." Science. 130:245-251, 1959. Oparin, A. I., The Origin of Life on Earth, 3rd Ed., New York: Academic Press, 1957. Oparin, A. L, The Chemical Origin of Life, Springfield, 111.: Chas. C. Thomas, 1964. Penrose, L. S., "Self-Reproducing Machines." Scient. Amer., 200 (No. 6): 105-1 14, 1959. Ponnamperuma, C, Lemmon, R. M., Mariner, R.. and Calvin, M., "Formation of Adenine by Electron Irradiation of Methane, Ammonia and Water." Proc. Nat. Acad. Sci., U.S., 49:737-740, 1963. Sagan, C, "The Planet Venus." Science, 133:849-858. 1961. Sagan, C, "On the Origin and Planetary Distribution of Life," Rad. Res.. 15:174- 192, 1961. Sinton, W. M., "Further Evidence of Vegetation on Mars," Science, 130:1234-1237. 1959. Tax, S. (Ed.), The Evolution of Life, Vol. 1 of Evolution after Darwin, Chicago: University of Chicago Press, 1960. The Origin and Evolution of Genetic Material 517 Wolstenholme, G. (Ed.), Man and His Future, Boston: Little, Brown & Co., 1963. See Supplement VII. QUESTIONS FOR DISCUSSION 42.1. Which do you think came first in evolution, the gene or what we now call the "gene product"? Explain. 42.2. Do you believe that the genetic material on earth has undergone a biochemical evolution? A structural evolution? A functional evolution? Why? 42.3. Do you believe there are "superhumans" on other planets? Why? 42.4. Do you suppose that in the future we will need to be as careful in avoiding contaminants from other planets to our own. as we now are in avoiding the reverse? Why? 42.5. In what respects would you expect the environment of Venus to be changed if a photosynthesizing microorganism were introduced into its atmosphere? 42.6. What information would you seek from landing on the moon? Mars? Venus? 42.7. What characteristics would you expect of genes from other planets? 42.8. Is protein genetic material? Justify your answer. 42.9. What is your present opinion of the assumption, stated on p. 10. that genetic material arises only by the replication of pre-existing genetic material? 42.10. What is your definition of a gene? Of genetics? Appendix ELEMENTARY BIOMETRICAL INFERENCES I. Introduction: Statistics and Parameters II. Discrete Variables A. Range of Statistics Expected from a Parameter Involving One Variable B. Range of Parameters Expected from a Statistic Involving One Variable C. Specific Probabilities Expected from Parameters Involving One Variable 1. Rules of probability a. The addition rule b. The multiplication rule 2. The binomial expression D. Comparing Observed with Expected Statistics 1. The binomial test of a parameter involving one variable 2. The confidence interval test of a parameter involving one variable 3. Chi-square test of a parameter involving one variable 4. Chi-square test of a parameter involving two or more variables E. Comparisons Between Statistics 1. Involving one variable a. Observed difference vs. expected standard deviation b. The plus-minus test c. Contingency table approach to the chi-square test 2. Involving two or more variables Contingency table approach to the chi-square test III. Indiscrete Variables A. Parameters and the Normal Curve The normal curve vs. the binomial distribution B. Statistics Expected from a Normal Curve 1. Distribution of individual statistics 2. Distribution of the means expected for groups of statistics 3. Testing hypotheses regarding y. a. The r test b. The t test c. Confidence intervals for \i d. Comparison of Xi and X2 IV. The Power of the Test 519 520 A HIM NDIX I. INTRODUCTION: STATISTICS AND PARAMETERS There are numerous occasions when one may wish i<> arrive a1 some genetic con- clusion i»n the basis of experimental data. Whenever these data arc subject to chance variation, it is necessary to make use ol biometrical ideas and techniques in order to draw the mosl precise conclusions. Let us consider, therefore, some of the basic principles and methods which are likely to be valuable in a study of genetics. (The Table of Contents at the beginning of this chapter will make it easier to find the sec- tion that describes a particular biometrical technique.) A statistic is a measurement obtained from a sample. A sample can be consid- ered as having been drawn from an ideal population composed of an infinite number of measurements. Whereas the measure- ments of a sample are statistics, the meas- urements of the ideal, infinitely large, popu- lation are expressed in terms of parameters. The difference between a statistic and a parameter can be illustrated with a penny. Let the ideal population be composed of the results of an infinite number of tosses. In this ideal population one would expect the coin to fall heads up 50% of the time, and tails up 50% of the time. The population can be characterized in terms of a para- meter, the probability of heads up, ex- pressed as p = 0.5. If one actually takes a sample of this infinite population by toss- ing a penny a finite number of times, one obtains the statistic, the frequency of heads up relative to the total number of tosses. (liven a parameter, one may want to predict the range of statistics expected to comprise a sample (Figure A-1A). Alter- natively, one might like to be able to deter- mine from a statistic the range of para- meters from which this statistic could have been obtained by sampling (Figure A-1B). PARAMETERS 12 3 4 5 6 7 9 10 11 12 12 1 2 3 4 5 6 7 STATISTICS 9 10 11 12 13 o ol o2 figure A— I. Biometrical procedures to be discussed with respect to discrete variables (see text for explanation), o = observed, e = ex- pected. Arrows show direction of prediction. One may want to determine the probabili- ties (i.e., parameters) that different alterna- tives will occur in samples drawn from an ideal population (Figure A-1C). One may wish to compare the statistics expected (e) in a sample with those actually obtained (o) (Figure A-1D). And finally, using a parameter, one may wish to compare two groups of statistics (ol and o2) (Figure A- IE). Methods for making these and other comparisons are presented here. Heads vs. tails, black vs. white, smooth vs. rough, and tall vs. short all involve discrete variables which are measured by enumeration, since the outcomes or alter- natives fall into discontinuous, easily dis- tinguished and separable, classes. On the other hand, the statistics of wreight, height, and intelligence are all quantitative, con- tinuous, or indiscrete variables. The differ- ence between the two lies in the number of alternatives possible in each case; there is an infinite variety of alternatives possible in the indiscrete case, but only a limited number of outcomes in the discrete one. This difference disappears, however, once the outcomes are tallied. For example, al- Elementary Biometrical Inferences 521 though the number of different weights pos- sible in the range of weights between fat and skinny people is infinite, weights are scored with a scale whose number of pos- sible readings is limited. In other words, an infinite variety of outcomes must always be scored or measured in a finite number of ways. So far as statistics are concerned, the only difference between indiscrete and discrete outcomes is the possible occurrence of a much larger number of scored out- comes in the former case. In either group of outcomes, scoring a statistic requires the use of some measuring device, be it the eye, ear, finger, etc., very often in combination with a ruler, photoelectric cell, and so forth. We will study first statistics and para- meters for discrete outcomes (small number of classes) and then those for indiscrete outcomes (large number of classes). It should be emphasized at this point that the accuracy of the conclusions reached from the use of biometrical procedures de- pends upon four major factors: (1) imagina- tion and flexibility, (2) proper sampling methods, (3) accurate recording of statis- tics, and (4) correct choice and use of bio- metrical procedures. It is unreasonable to expect that good biometrical technique can overcome poor data; the biometrical analy- sis becomes more efficient the closer one ad- heres to the first three factors in carrying out experiments. II. DISCRETE VARIABLES A. Range of Statistics Expected from a Parameter Involving One Variable (Figure A-1A) One often formulates a hypothesis in terms of the probability that an event will occur. It is also often desirable to know the kind of result one would obtain were this hy- pothesis tested. For example, common sense suggests that "unbiased" pennies tossed in an unbiased manner have equal likelihood of falling heads up or tails up. Let heads up be considered a success. We can state as a hypothesis (Ho) that the pa- rameter p, the probability of success, is 50r; , or 0.5, of all the times the coin falls flat. Note that there are only two alternatives involved — success and failure. Since 50% of the time we would expect failure, the probability of failure is 1 — p. One need only use a single variable, probability of success, to describe all the outcomes pos- sible. (If one were to toss an unbiased die, there would be 6 different and equally pos- sible outcomes, and 5 variables. But if one considered as a success only when the die falls "one" up, then there would be only one variable and we could state as an hypothesis that p = %.) What kind of statistics would one expect to obtain from actual tosses of an unbiased penny? Clearly the result will depend upon whether 1, 2, or many trials, i.e., tosses, are made. Expected range of f values Let us represent the number of successes by X, the total number of trials or size of sample by N, and the proportion of success by f. Therefore, ^ = f , our statistic. Suppose one collected many relatively large samples. What f values would result? It has been shown that this can be determined by using the expression P(l ~ P) which is called the standard deviation of p, or sp. If the value of N (p) (1 — p) is equal to or greater than 25, it is found that 95% of the f values obtained lie between p — 1.96 sp and p + 1.96 sp. If one stated that f can have only the values included in this 95% confidence inter- val, he would be right 95% of the time and 522 APPENDIX figure A-2. 95% confidence limits (I) for f based on a sin- gle-variable parameter, p. To determine confidence intervals, find p on the vertical scale. Move right to the intersections with the two curves indicating the sample size. Finally, read down to determine on the hori- zontal scale the confidence lim- its of f . ( 2 ) For p based on a single-variable statistic, f. To determine confidence intervals, find f on the horizontal scale. Move upward to the intersec- tions with the two curves indi- cating the sample size. Finally, read left to determine on the vertical scale the confidence lim- its of p. {Courtesy of the Bio- metrika Trustees.) wrong 5% of the time. In the penny- tossing example (p = 0.5), if N = 100, sp is approximately 0.05 and 95% of the time we would expect f to be in the interval 0.4 - 0.6. If many samples of N = 100 are drawn, one can state that 95% of all f's will lie in the interval 0.4 — 0.6. If one draws a single sample of N = 100, it can be stated that f will be in 0.4 — 0.6 and we would have a 95% chance of being right and a 5% chance of being wrong. Why should one resign himself to the handicap of being wrong 5% (or any per cent) of the time? In order to be right 100% of the time one would have to admit that, 5% of the time, f can lie outside the 95% confidence interval. In the example this would mean that 5% of the time f may lie anywhere between 0 (no successes) and 0.4 and between 0.6 and 1.0 (all successes). To be 100% correct, to have 100% con- fidence, one would have to predict f to range between 0 and 1. However, electing to be 100% right also means that all other values of p would also have an expected range of f's from 0 to 1. Accordingly, the 100% range does not provide different ex- pectations of f for different values of p; it provides no power at all to discriminate between different p values. However, by being willing to be wrong 5% of the time, the range of expected f's (when p = 0.5 and N = 100) can be reduced from (0 — 1) to (0.4 - 0.6). And were p = 0.3 and N = 100, f would be roughly between 0.2 and 0.4 95% of the time. Accordingly, accepting a 5% chance of being wrong per- mits one to have different statistical expec- tations for different p values. In genetics and biology in general, researchers usually agree to the use of the 95% confidence in- terval both for statistics and parameters. Using the expression given on page 521, one can calculate the different values of sp for numerous combinations of p and N. The 95% range for f can be determined Elementary Biometrical Inferences 523 from these calculations. For convenience, the 95% ranges for f for various values of p and N are plotted in Figure A-2. For values of N not shown, one can interpolate between curves. Note that if N were in- finitely large, f would equal p and for any given value of p, the range would become wider as N decreased. B. Range of Parameters Expected from a Statistic Involving One Variable (Figure A-1B) If one had no notion what the parameter for the chance of a successful toss of a penny should be, one could make an in- ference about the p value from the sta- tistics obtained. An estimate of the un- known parameter, p, can be obtained from the statistic f. Suppose that 100 tosses of a penny yield 30 successes. The value f = 0.30 is a single statistic. The single best estimate of p is f. From the single f value, the best estimate is p = 0.30. However, it should not be surprising if p were really 0.31, 0.29, or some other nearby value. What would also be valuable to know is the range of p values likely when f = 0.3 and N = 100. This range can be determined by calculating f(l - f) N which is the standard deviation of f, or sf. The values lying between f — 1.96 sf and f + 1.96 sf make up the 95% confidence interval of p, because 95% of the time we would expect this particular sample to have a p value in this interval. If we say that p cannot be outside this range, we will be wrong only 5% of the time. In the present case, sf is about 0.05 and the 95% con- fidence interval of p is roughly 0.20 to 0.40. If one asserts that p must lie between 0.20 and 0.40 he will be wrong only about 5% of the time. By reading upward and then to the left, one may use Figure A-2 to determine the 95% confidence intervals of p for different values of f. PROBLEMS A. 1. You suspect that the sex ratio of the fruit fly Drosophila is 0.5 d" c? and 0.5 9 9. Let success be cf • What range of successes might you expect with 95% confidence from an un- biased count of 100 flies? 250 flies? 1000 flies? What is happening to your con- fidence limits as sample size increases? What does this mean? A. 2. You expect to draw a sample in which N = 100. What is the 95% range for f when the hypothesis is p = 0.5? p = 0.3? p = 0.1? How does the range of f change according to the hypothesized p values? A. 3. You expect 8 different equally-fre- quent types of gametes to be pro- duced by a certain trihybrid. Only one of these is of interest to you. If you sample 50 gametes, what range, in numbers of these interesting gam- etes, are you likely to obtain? A. 4. Under certain conditions, white-eyed Drosophila males do not mate very readily with red -eyed females. If the chance of mating is 10%, about how many opportunities for mating should you provide to be reasonably sure that 5 matings will occur? A. 5. A student finds 25 brown-eyed flies among 100. Determine with 95% confidence the true probability of a fly's being brown-eyed. A. 6. Using Figure A-2, determine the 95% confidence limits of p when f = 0.60, and N = 100, 250, and 1000. A. 7. After meiosis of the genotype Aa Bb in Neurospora you obtain 100 asci. If you assume independent segrega- tion, how many ascospores do you 524 MM' I \DI\ expect to have the following genetic constitution :.!/->? .1/' plus - In such cases one has no difficulty in deciding upon the probability of success. At other times one does not know the probability ot success, and this parameter must then be determined. 1. Rules of Probability a. The addition rule. Sometimes a suc- cess can occur in two or more different ways, each way excluding the others. What is the total probability of success in such cases? If on a single toss of a die the probability of a "one" is J/£ and the probability of a "two" is l^, then the expectation or prob- ability of either a "one" or a "two" is Y + Y — Y- In general, the probability that one of several mutually exclusive suc- cesses will occur is the sum of their indi- vidual probabilities. If the probability that an event will succeed is p, and the probability that it will fail is q, then the probability of either success or failure is p + q. But if it is certain that the event must either succeed or fail, then p -f- q = 1, p = 1 — q, and q = 1 — p. b. The mi duplication rule. Sometimes over-all success depends upon the occur- rence simultaneously or consecutively of two or more successes, and the occurrence (or failure) of one success in no way influ- ences the occurrence (or failure) of the others. If the probability of "one" in the toss of a die is Y and if the probability of another "one" in a second toss is also Y* then the probability of "one" on the first and "one" on the second is ^ X ^ = 3^6- In general, the probability that all of several independent successes will occur is the product of their separate probabilities. 2. The Binomial Expression Given a parameter involving only one variable, one can determine the exact prob- abilities of obtaining specific combinations Elementary Biometrical Inferences 525 of successes and failures by expanding the binomial expression (q + p)N- If a "one" on a die is a success and the die is tossed 5 times, the probabilities of 0 ones, 1 one, 2 ones, 3 ones, etc., among the 5 tosses are given by successive terms of the expansion of the binomial (f+5)' In this expression % represents the proba- bility of not obtaining a one on a single trial, }/q the probability of obtaining a one, and the exponent 5 the number of trials. The expansion is shown in Table 1, below. Note that each result is possible, each hav- ing its own exact probability of occurrence. PROBLEMS A.ll. If you roll a die three times, what is the probability of obtaining (a) three "fours" in succession? (b) "One," "two," and "three" in that order? A. 12. If you roll two dice at the same time in a single trial, what is the prob- ability of obtaining a total of eleven? Two? Seven? A. 13. What is the chance that a simul- taneous toss of a penny, a nickel, a dime, a quarter, and a half-dollar will fall : (a) All heads or all tails? (b) 3 heads and 2 tails? A. 14. What is the exact probability (using an unbiased penny) of a run of tosses which: (a) Starts with 2 heads and ends with 3 tails? (b) Has 4 successive heads? (c) Has 5 successive tails? A. 15. What is the exact probability of 10 successes, if p = i^ and N = 1.5? A. 16. How often will you expect to obtain less than 3 successes if p = 34 and N = 5? A. 17. You have just etherized Drosophila which are the progeny of a cross be- tween ci+ ci and ci+ ci. What is the probability that there is only 1 ci ci fly among the first 3 flies chosen at random? Among the first 5 flies chosen at random? A. 18. Following independent segregation of Aa Bb Cc, an ascus is formed. What is the probability that if two asco- spores are chosen at random they will be A B C? a be? Either A B C or a b c? A. 19. An albino (aa) man of blood type MN marries a heterozygote for al- binism (Aa) also of MN blood type. They plan to have 4 children. If you assume independent segregation, what is the exact probability that they will have: (a) No albinos? (b) 2 nonalbino children with MN blood type? (c) 3 children with M blood type? Table 1 ®Hm^m+ [(55 - 50) - Y2Y X«ii — 50 (4^)2 , (4^)2 50 ~l~ 50 40.5 50 50 = 0.8 The probability is obtained from a chart of X2 (Figure A-3) under one degree of free- dom. (The number of degrees of freedom for such a test is one less than the number of classes; that is, it equals the number of variables). Thus, from Figure A-3 one finds that the probability lies between 0.35 and 0.40. The difference between what is observed and what is expected according to the null hypothesis is nonsignificant. Therefore, one may accept the hypothesis. The chi-square method is an approxima- tion and is valid for relatively large samples only. Its use requires that no class have an expected value of less than 2 and that most of the expected values be at least 5. 4. Chi-square Test of a Parameter Involving Two or More Variables The x2 test is applicable to parameters in- volving more than 2 alternative outcomes, hence involving two or more variables. 528 APPENDIX PROBA B I LI T Y .Ok) CD ^1 O <_n U U> O O o o b e ° o O o P o o PROBABILI TY figure A-3. 77j^ ^- aA/d t distributions. To read the chart with a ^" value of 17 based on 7 degrees of freedom, the vertical line corresponding to a v2 value of 17 is followed upward until it intersects the curve corresponding to N - 7. Directly to the left of this point the probability, .017, is read off. With the chart inverted, probabili- ties for the t distribution are read in exactly the same way. The probability given is the probability of a numerically greater deviation. {Courtesy of J. F. Crow; from Genetics Notes, Burgess Publ. Co.) Elementary Biometrical Inferences 529 The chi-square test can be used to deter- mine whether a sample is consistent with an hypothesized 9:3:3: 1 ratio, for ex- ample. If a 9:3:3:1 ratio were being tested, the ideally expected numbers in a group of 80 individuals would be 45, 15, 15, and 5, respectively. Since there are four classes, there are three variables or degrees of freedom. If we observed 40, 20, 12, and 8, respectively, we would calculate 2 (40-45)2 (20 - 15)2 X(3) — " Ar 45 15 (12 - 15)2 (8 - 5)2 + " IT - + — 5- " = 4'6 * (The term 3^» the Yates' correction, is not applicable if there is more than one degree of freedom.) Since the probability lies between 0.20 and 0.25, the difference is nonsignificant, and one accepts the null hypothesis. It is interesting to note that the prob- ability of obtaining a x value equal to or greater than 0.004 for one degree of free- dom, 0.1 for two degrees of freedom, etc. is 0.95. It follows the probability of ob- taining x2 values smaller than these must be 0.05. Such low values in an actual test indicate that the agreement between ob- servation and expectation is suggestively better than expected. The question of whether the data represent authentic ran- dom samples may be legitimately raised in such cases. PROBLEMS A. 20. A person with woolly hair marries a nonwoolly-haired individual ; they have 8 children, 7 woolly-haired and 1 nonwoolly-haired. Test the hy- pothesis that woolly hair is due to a rare, completely dominant gene. A.21. Given the data in A. 20, test the hy- pothesis that woolly hair is due to a completely recessive mutant. A. 22. A penny is tossed seven times. One time it falls on edge, five times it falls heads, and once it falls tails. Is this an "honest" coin? A. 23. A test cross produces 57 individuals of A phenotype and 43 of A' pheno- type. Is one pair of genes involved ? A. 24. Given the data in A. 23, test the hypothesis that one parent is a dihybrid and that the A phenotype is obtained only when two particular nonalleles are present. A.25. In a sample of 540, X = 90. What is the value of chi-square if you hypothesize that p = 34? Do you accept this hypothesis? A. 26. Among 60 individuals the pheno- types are 8 A, 12 B, 20 C, and 20 D. Test the hypothesis that: (a) A B C D are in the relative proportion 1:3:3:9. (b) All four phenotypes have an equal chance of occurring. (c) The ideal ratio is 1A : 3B : 5C : 7D. A. 27. A random sample from a natural population contains 65 AA, 95 Aa, and 40 aa individuals. Test the hypothesis (after consulting Chapter 15) that: (a) The frequency of a in the population gene pool is 0.5. (b) This sample is consistent with the population being in genetic equilibrium for this locus, if you assume that the observed gene frequency for a is also the population frequency. E. Comparisons Between Statistics (Figure A-1E) 1. Involving One Variable a. Observed difference vs. expected standard deviation. Suppose that a sample (A) pro- vided 20 males and 30 females, whereas a 530 AIMM.NDIX different -ample t H -a\e 30 ni. ih- .lixl 20 females. I- there a significant difference in the frequency of males in the two samples? t\ = 0.40 and fB = 0.60; XA = 50, NB = 50.) We have no ex- pectation as to what px or pB should be. According to the mill hypothesis these two samples have the same parameter, px. Our besl estimate of px is fx, obtained by pooling the results of both samples and obtaining 50 loo = 0.50. We next calcu- late li<>\\ large the difference between the observed f's is, relative to the total standard deviation that one would expect if fx were obtained in each of the two samples, NA and NB- This calculation can be made from the expression: IB - fA V fxO -fx) +fx(l -fx) X N. 0.20 \ 0.5 X 0.5 0.5 X 0.5 = 2.0 50 50 (The subtraction in the numerator should be made to give a + result, i.e., one should obtain the absolute value of the remainder.) It has been shown that if Nx is greater than 30, values of 2.0 or more will occur by chance only 5% of the time. We conclude, therefore, that the two samples under test are on the borderline of being statistically different at the 5% level of significance. 1). The plus-minus test. Suppose a par- ticular treatment is to be tested for its capacity to change a statistic. Suppose, moreover, that one does not care just how much change is being induced as compared with how much is occurring spontaneously. (The treatment might produce only a very small change; under these circumstances, two tremendously large samples, one con- trol and the other treated, would be neces- sary to obtain a statistically significant difference between their measurements.) Wli.it can be done i- to arrange a series of paired observations in which the members ot a pair are as similar as possible in order to make the measurement of difference as sensit i\ e as possible. Imagine, for example, that one wishes to determine whether feeding a salt to the de- veloping Drosophila male has any effect upon the sex ratio of his progeny. Each test consists of scoring the sex of the prog- eny of two single pair matings, in which one male has and the other has not been treated. Assume that the experiment is performed in an unbiased manner and that the results are as follows: Paired Obser- Un Sex Ratio (o*oV9 9) ± Test Un- vation treated Treated treated Treated 1 0.47 0.46 + 2 0.48 0.47 + 3 0.49 0.48 + 4 0.50 0.50 No Test 5 0.46 0.44 -f- - 6 0.51 0.50 + 7 0.48 0.47 + - One proceeds to test the null hypothesis that the treatment has no effect upon the ¥x sex ratio. In accordance with this view, there would be an equal chance for the un- treated and treated members of a pair of observations to have the higher sex ratio (that is, to be scored + ); consequently the Ho is p = ]/2- There are only 6 tests of the Ho, since one test gave the same sex ratio for both untreated and treated. The probability that the relevant 6 untreated shall be all successes or all failures is, ac- cording to the null hypothesis, 2(}/£)6, or ^2- or about 3%. (The chance that t he remain- ing 5 tests will be like the first is (%) , or also about 3%.) Accordingly, one rejects the null hypothesis at the 5% level of signi- ficance. The statistical test indicates that Elementary Biometrical Inferences 531 the untreated and treated do not have the same parameter. Upon examining the data, one will conclude that the sex ratio is lower following salt treatment than when such a treatment is omitted. (One cannot determine from these delta whether salt raises the number of females or lowers the number of males. One finds only a differ- ence in sex ratio as a function of the presence or absence of salt, the actual mech- anism of the effect remaining unknown.) c. Contingency table approach to the chi- square test. Assume that XA = 3 and NA = 6 in sample A, and XB = 5 and NB = 18 in sample B. Are these statistics different at the 5% level of significance? To determine this, one tests the null hypothesis that both samples have the same parameter (p). However, the value of p is completely unknown. If a con- tingency table is constructed, it will give the most likely values of X (and hence N — X), a common p for both samples being under- stood. Having determined these ideally- expected values, one can then proceed as before to calculate chi-square. The observed data are arranged as shown in Figure A-4A. The best estimates for the values expected according to the un- known p are shown in B. To obtain the value expected in the shaded box in A, for example, multiply together the totals at the end of its column and row and divide by the number NA + NB. This value (6 X 8/24) is 2. Since we are dealing with x2» recall that it is usually safe to require that no class have an expected frequency less than 2 and that most expected values be at least 5. Note that the other expected values in B can be obtained in a similar manner; this procedure, however, is unnecessary since all the other values are fixed by the mar- Classes Success Failure Totals Samples Totals A B 3 - 5 8 3 13 16 6 18 24 2 6 8 4 12 16 6 18 24 A. Actual Data B. Expected Data C. Difference (A-B) '/a —Vi -Vi Vi 2 6 'A 4 'A 12 Sum of Values in E =X2 D. C with Yates' Correction E. D Squared Expected figure A-4- 2X2 contingency table. rv.i-2 APPENDIX Samples R S T M N O P 5 2 4 1 1 Classes 4 3 6 13 Outcomes 3 4 4 11 8 5 2 15 20 14 16 50 4.40 3.08 3.52 1 1 5.20 3.64 4.16 13 4.40 3.08 3.52 1 1 6.00 4.20 4.80 15 20 14 16 50 A. Actual Data B. Expected Data 0.60 1.08 0.48 1.20 0.64 1.84 -1.40 0.92 0.48 2.00 0.80 2.80 C. Difference (A-BI <1 <1 <1 ^1 <1 <1 <1 <1 <1 4 6 <1 7.84 4.8 C Squared Expected figure A-5. 3X4 contingency table. ginal totals, which must be the same in B as in A. Accordingly, there is only one degree of freedom (one variable) in the 2X2 contingency table formed. Difference table C is then constructed, the values of which are identical in crisscross position and always total zero. Each of the values in C is made less extreme (closer to zero) by 3^, to comply with Yates' correction. This is shown in D. Each of the corrected differ- ences in D is squared and divided by the corresponding expected value shown in B. The sum of the four values obtained (1/2 + i/6 + |/4 + J/12) is chi-square. In the present case chi-square is less than 1 (but more than 0.004) and has a prob- ability greater than 10%. The null hy- pothesis is thus accepted, namely, that the two samples are not statistically different at the 5% level of significance. 2. Involving Two or More Variables Contingency table approach to the chi- square test. Sometimes the data in a sample fall into more than two classes or outcomes, and more than two such samples are to be compared. This involves "num- ber of classes — 1" variables as well as "number of samples — 1" variables. The total number of variables equals the prod- uct of these two sources of variability. The number of degrees of freedom is equal to the total number of variables, which is always Elementary Biometrical Inferences 533 (number of rows — 1) times (number of columns — 1) in a contingency table. Suppose three samples were scored four alternative ways to give the results shown in Figure A-5A. The procedure followed is the same as that already described for the 2 X 2 or four-fold table (note that Yates' correction is not applicable in any larger table). There are 6 degrees of freedom. If one tests at the 5% level, Figure A-3 shows that x%) has to be greater than 12.5 if one is to reject the null hypothesis, namely, that all the samples and types can be represented by the same parameters. Moreover, finding that x2 is less than 1.6 would mean that the same parameters would produce samples varying this little from the ideally expected values only 5% of the time. In that case one would reject the samples as being random, suspecting that there was some hidden bias in the collection and/or the scoring of the data. The decision that neither obtains can be seen from Figure A-5D. Consequently, one accepts the null hypothesis that these samples are not statistically different at the 5% level of significance. Assume, however, that chi-square had been 14.1 in the preceding example. One would reject the null hypothesis at the 5% level but could accept it at the 1% level of significance (meaning that these samples have more than 1%, but less than 5%, chance of having the same parameters). Assuming that such a result was obtained in an unbiased manner it might be due to the fact that (a) the null hypothesis is true but one happened to collect data (as will happen by chance one time in 20) which varied at least this much from those ex- pected, or (b) the null hypothesis is in- correct. Even if the hypothesis at the 5% level is rejected, one may wish to test the data further, using smaller contingency tables to determine which samples or out- comes are consistent or inconsistent with each other according to a null hypothesis. Note here that the observed values in a contingency table furnishing the larg- est contributions to chi-square are those most responsible for the rejection of the hypothesis. PROBLEMS A.28. A. 29. A cross yields 20 offspring of one type and 40 of another. A month later the same cross produces 15 of the first type and 15 of the second. Do these results differ significantly? Ten sets of identical twins are se- lected ; only one member (the same one) of each pair is given a particular drug daily for 10 days. All indi- viduals are weighed before and after this period. The changes to the nearest whole pound are as follows: Twin Pair Untreated Treated 1 + 1 + 1 2 + 2 4-1 3 -4 -3 4 + 1 + 2 5 -3 -2 6 + 3 + 2 7 -2 + 1 8 +4 +3 9 -1 0 10 + 5 -4 Analyze the results of this experi- ment statistically. A. 30. Among the women of population A are 10 blondes, 5 redheads, and 15 of other hair color. In population B there are 7, 7, and 6, respectively; whereas in population C the tally is 8, 4, 8, respectively. Are these populations the same with respect to the relative frequency of these hair color types? 534 APPENDIX A.31. An experimenl is performed four times. X is 5, 7. 10. and 1 1 when N is 8, 20, 20, 30, respectively. Are all tour results mutually consistent? A.32. Suppose the label on two packages of grass seed states that each package will germinate 40% grass type A, 35* J grass type B, 15% grass type C, .iiul It)' J weeds D. A sample from package 1 germinates 400A, 400B, 50C, and 150D. A sample from package 2 yields 390A, 410B, 70C, and 130D. Compare the contents of each package with the labelled contents and with each other. What do you conclude? A drug manufacturer receives results of using or not using his product. As a check on bias in testing he scores the control and experimental group for eye color and ABO blood type and finds the results tabulated below. AB A A. 33. Blue Brown Control 7 6 Experimental 4 8 B Blue B rown 12 10 13 O 8 Blue Brown Control 4 4 Experimental 5 2 Blue Brown 8 9 8 12 What should he conclude about bias? A. 34. Suppose women were classified in two ways: hair color and tempera- ment. I sing the results listed be- low, test the hypothesis that there is no relation between hair color and temperament. Blonde Red Brown Pugnacious 23 6 11 Quiet 26 3 31 Normal 41 9 30 figure A-6. The normal curve. III. INDISCRETE VARIABLES A. Parameters and the Normal Curve Suppose a particular measurement is the result of the action of a very large number of independent variables, each of which has approximately the same magnitude of effect on the measurement. If, then, an infinitely large number of such measurements are collected, they will be expected to have a range of values which are said to be nor- mally distributed. Figure A-6 shows the normal distribution or curve formed by plotting these measurements against the frequency with which they would be ex- pected to occur in this infinitely large popu- lation. The population mean, or "true" mean, is denoted by the parameter n. The population standard deviation, or "true" standard deviation, is represented by the lower case Greek letter sigma, a. It is known that about % of all the measure- ments in a normal curve lie within one a of the mean, and about 95% of all measure- ments lie within 2 a of the mean (being in the range n ± 2 a). Strictly speaking, the individual "measurements" or values which comprise the normal curve are also pa- rameters. The normal curve vs. the binomial dis- tribution. Suppose N = 20 and p = 3/6- If an infinitely large number of samples Elementary Biometrical Inferences 535 each of N = 20 were obtained, the exact probabilities of obtaining different numbers of successes would be expressed in the bi- nomial distribution plotted in histogram form in Figure A-7, where each class of success is represented by a column whose height is proportional to the frequency of the class. Note that there are only 21 ways to score the outcome of a set of 20 of these observations (from 0 to 20 successes), so we are dealing with 20 discrete variables. The smooth curve shown is the normal curve, which has the same mean and stand- ard deviation as the histogram. The larger the sample size, if p = %, the larger will be the number of outcomes possible per sample, and the closer the plot of the prob- ability of successes will approach the nor- mal curve. Therefore, as N increases without bound, the number of possible outcomes increases to provide us with an example of a continuous variable, whose values are said to be distributed normally. ax give absolute r values of 1.96 or greater we reject the null hypothesis when X gives a t value that equals or exceeds 1.96. This equation can be rearranged X = M + t 1.96. Since p <0.05 one rejects the null hypothesis and may conclude that the plant measured cannot, at the 5% level of significance, come from a theoretical population where m = 50 and jx = 4. Number of Successes When the absolute value of t is 1.96, this probability for X is 0.05. Since exactly 5% of the X values in a distribution characterized by the hypothesized ju and figure A-7. Histogram of probabilities for different numbers of successes for a binomial distribution (N = 20, p = W), and a normal curve with the same /x and a as the histogram. 536 APPENDIX 2. Distribution of the Means Expected for Groups of Statistics The arithmetic sample mean or average of a group of statistics comprising a sample is denoted by X (read "X bar") and is the average obtained by adding all the values of X and dividing by X. In more sym- bolic term-, X - N -A Given a population described by mean n and standard deviation jx, one can predict something about the range of X's to be expected from drawing a great many samples of size N from this population. If many samples are drawn, it will be found that the X values fall into a distribution which has a theoretical mean equal to m and which will be normally distributed with a standard deviation 1 .96. Consequently, one rejects the hypothesis. b. The t test. Frequently one may have to test some hypothetical m when 0.05. The hy- pothesized m is accepted. c. Confidence intervals for /z- Suppose one chooses to work at the 5% level of significance. If ax is known, the 95% con- fidence interval for /x = X ± 1.96 cx, or H = X ± 1.96 ax. If only sx is known, Elementary Biometrical Inferences 537 then the 95% confidence interval for m can be determined as follows. Given, as before, thatX = 68.03, sx = 3.24, and N = 9; first find the value of t which has p = 0.05 for N — 1 degrees of freedom. (For N — 1 = 8, this is about 2.3.) Hence, The value of t is then found from X - 4 = 2.3 N One rejects all values where X differs from n by more than 2.3 Sj, and accepts all values of X — m with 95% confidence that are less than 2.3 sx. Substituting, one finds X-, 1.08 = 2.3 or, X - n = 2.3 (1.08) = 2.48. Finally, the 95% confidence level for n, in the present case, is X ± 2.48, or 65.55 to 70.51. d. Comparison of Xx and X2. Suppose one selects two sample sets of corn plants, and then measures the height of each plant. The statistics obtained are: Sample 1: Nx = 9 X, = 72.44 2(X, - Xt)2 = 65.70 Sample 2: N2 = 10 X2 = 70.30 2(X2 - X2)2 = 69.50 To be tested is the null hypothesis that these two samples have the same n and the same ax. The best estimate of the unknown ax is sx, obtained from the two samples by the following formula: Sx . /Z(Xi - Xt)2 + S(X2 - X2): (Ni - 1) + (N2 - 1) -4 65.70 + 69.50 + 9 = 2.82 One derives a value of s- in the present case equal to 1.29, since it is known that N, ^N2 Xi — X.} to be 72.44 - 70.30 1.29 = 1.66. Since each X was obtained from a single sample, the number of degrees of freedom is (Nt - 1) + (N2 - 1), or 17. Because p > 0.1 one accepts the null hypothesis and may conclude that the two means are not statistically different at the 5% level of significance. If one obtains a value of t inconsistent with the hypothesis, the two samples differ either in their ju's, ffx's, or both. IV. THE POWER OF THE TEST There are two types of error involved in testing a parameter or statistic. One has already been discussed. This type of error is the rejection of the correct hypothesis 5% of the time (when working at the 95% confidence level, or the 5% level of sig- nificance) in order to reject incorrect hy- potheses. The other type of error is the incorrect acceptance of an hypothesis. Suppose f = 0.45 and N = 100. The hy- pothesis that p = 3^ is tested and found acceptable at the 5% level. But the real p might lie anywhere between 0.35 and 0.55 (see Figure A-2). If p is not 0.5 but some- where between 0.35 and 0.55, one may have accepted the wrong hypothesis. In the present case, the test is only power- ful enough to reject incorrect hypotheses where p < 0.35 or > 0.55. Had N been 1000 and f = 0.45, the discriminatory power of the test would be greater, at the 5% level, causing the rejection of any hypothesis where p < 0.42 or > 0.47. Before collect- ing statistics, it is necessary to determine whether there is adequate power to dis- criminate against alternative hypotheses. Suppose, for genetic reasons, one wishes to test whether some statistics obtained by experiment exhibit an expected 3 : 1 ratio. One may accept the hypothesis; but if a .-:*s APPENDIX theoretical 1:1 or 2:1 ratio is also ac- cepted, the tesl i> rendered rather weak and is not likely to be useful in describing the nature of the genetic events involved. One way to increase the meaningfulness of the tesl is to increase X. Another way i- to change the level of confidence. At the 1<»' , level of significance the "power" of the test is greater than at the 5% level; but there is a proportional increase in the chance of rejecting the correct hypothesis. Unless there is some special circumstance, geneticists usually work at the 5% level and increase the power of the test by increasing X. Recall, however, that the size of s or a decreases as the square root of N increases, so that a fourfold increase in X only reduces the standard deviation by a factor of 2. PROBLEMS A.35. Given jx = 8, X = 265, X = 12; test at the 5% level of significance the hypothesis that m = 11. A.36. Given ax2 = 412, N = 53, X = 142; test at the 5% level of significance the hypothesis that m = 135. A. 37. What are the 95% confidence limits for fi when ax = 4, X = 100, and X = 35? A. 38. Given the following statistics: 1, 3, 4, 5, 5, 5, 5, 5, 6, 8; calculate X, sx, and sx. A. 39. A new antibiotic was tested on pneu- monia patients with the following results: of those treated, 64 lived and 26 died (28.9% died); of those un- treated, 36 lived and 24 died (40% died). Test the hypothesis that the treatment is not effective. A. 40. A random sample of six observations drawn from a certain normal popu- lation is as follows: 0, 2, 6, 6, 8, 14. Test the hypothesis that m. the population mean, equals 10. Use the 5% level of significance. A. 1 1 . Normal barley seeds are treated with X-rays and planted. ( )f 400 seed- lings examined, 55 show sectors with visible mutation. Test the hypothe- sis that the true mutation frequency at this dosage is 10%. A. 42. Denote the length of an ear of corn by x inches. Explain exact ly what is meant when someone says "the probability of x being less than 7 is 0.05." A. 43. A random sample of 25 mice is taken from a certain mutant strain. It is hypothesized that the length of these mice is approximately normally dis- tributed. You find X equals 60 mm. and sx is 10mm. (a) Test the hy- pothesis that /u equals 61 mm. at the 5% level of significance, (b) Ex- plain what is meant by "5% level of significance" in this experiment. A. 44. Using the data of problem A. 43, find confidence limits for m = 61 mm. with 95% confidence. Explain the practical meaning of your result. A. 45. Given the following data: Sample 1 Sample 2 N = 10 N = 10 + 3.4 + 5.5 +0.7 + 1.9 -1.6 + 1.8 -0.2 + 1.1 -1.2 +0.1 -0.1 -0.1 +3.7 +4.4 +0.8 + 1.6 0.0 + 4.6 + 2.0 + 3.4 Determine whether these two sam- ples are statistically different. A. 46. Under what circumstances can one use the t table for values of r? Elementary Biometrical Inferences 539 REFERENCES Bailey, X. T. J., Statistical Methods in Biology, Xew York: J. Wiley & Sons, 1959. Falconer, D. S.. Introduction to Quantitative Genetics, Xew York: Ronald Press, 1961. Kempthorne. O., An Introduction to Genetic Statistics, Xew York: J. Wiley cV Sons, 1957. Levene, H., "Statistical Inferences in Genetics," in Principles of Genetics, 5th Ed., Sinnott, E. W., Dunn, L. C, and Dobzhanskv. Th.. Xew York: McGraw-Hill, \(>5H Chap' 29, pp. 388-418. Mather, W. B., Principles of Quantitative Genetics, Minneapolis: Burgess Publishing Co., 1964. SUPPLEMENTS I Part of a Letter (1867) from Gregor Mendel to C. Nageli s-9 II Nobel Prize Lecture (1934) of Thomas Hunt Morgan s-15 III Nobel Prize Lecture (1946) of Hermann Joseph Muller s-19 IV Nobel Prize Lecture (1962) of Maurice H. F. Wilkins s-31 V Nobel Prize Lecture (1959) of Arthur Kornberg s-60 VI Nobel Prize Lecture (1958) of George Wells Beadle s-75 VII Nobel Prize Lecture (1958) of Edward Lawrie Tatum s-88 VIII Nobel Prize Lecture (1958) of Joshua Lederberg s-98 IX Nobel Prize Lecture (1962) of James Dewey Watson s-111 X Nobel Prize Lecture (1962) of Francis H. C. Crick 5-735 SUPPLEMENT I Part of a Letter (1867) from Gregor Mendel to C. Nageli Translated from German by Leonie Kellen Piter- nick and George Piternick. Reprinted from The Birth of Genetics, Supplement to Genetics, Vol. 35, No. 5, Part 2 (1950), by permission of Genetics. Gregor Mendel (1822-1884) SUPPLEMENT II Nobel Prize Lecture (1934) of Thomas Hunt Morgan Reprinted by permission of The Nobel Foundation for Les Prix Nobel. The complete lecture is pub- lished in The Scientific Monthly for July 1935. Only the first portion is reprinted here. Thomas Hunt Morgan (1866-1945) By permission of The American Genetic Associa- tion, The Journal of Heredity, frontispiece, Vol. 24, No. 416, 1933. SUPPLEMENT VII Nobel Prize Lecture (1958) of Edward Lawrie Tatum Reprinted by permission of The Nobel Foundation for Les Prix Nobel. Published in Science, 129: 1711-1715, 1959. Edward Lawrie Tatum (1909- SUPPLEMENT VIM Nobel Prize Lecture (1958) of Joshua Lederberg Reprinted by permission of The Nobel Foundation for Les Prix Nobel. Published in Stanford Med. Bull., 17:120-132, 1959; and in Science, 131:269- 276, 1960. Joshua Lederberg (1925- SUPPLEMENT IX Nobel Prize Lecture (1962) of James Dewey Watson Reprinted by permission of The Nobel Foundation for Les Prix Nobel and Elsevier Publishing Com- pany. Published in Science, 140: 17-26, 1963. James Dewey Watson (1928- i SUPPLEMENT X Nobel Prize Lecture ( 1 962 ) of Francis H. C. Crick Reprinted by permission of The Nobel Foundation for Les Prix Nobel and Elsevier Publishing Com- pany. Published in Science, 139: 461-464, 1963. SUPPLEMENT I PART OF A LETTER (1867) from GREGOR MENDEL to C. NAGELI Highly esteemed Sir: My most cordial thanks for the printed matter you have so kindly sent me! The papers "die Bastardbildung im Pfianzenreiche," "iiber die abgeleiteten Pflanzenbastarde," "die Theorie der Bastardbildung," "die Zwischenformen zwischen den Pflanzenarten," "die systematische Behandlung der Hieracien rucksichtlich der Mittelformen und des Umfangs der Species," especially cap- ture my attention. This thorough revision of the theory of hybrids according to contemporary science was most welcome. Thank you again! With respect to the essay which your honor had the kindness to accept, I think I should add the following information: the experiments which are dis- cussed were conducted from 1856 to 1863. I knew that the results I obtained were not easily compatible with our contemporary scientific knowledge, and that under the circumstances publication of one such isolated experiment was doubly dangerous; dangerous for the experimenter and for the cause he repre- sented. Thus I made every effort to verify, with other plants, the results ob- tained with Pisum. A number of hybridizations undertaken in 1863 and 1864 convinced me of the difficulty of finding plants suitable for an extended series of experiments, and that under unfavorable circumstances years might elapse without my obtaining the desired information. I attempted to inspire some control experiments, and for that reason discussed the Pisum experiments at the meeting of the local society of naturalists. I encountered, as was to be expected, divided opinion; however, as far as I know, no one undertook to s-9 GREGOR MENDEL repeat the experiments. When, last year, I was asked to publish my lecture in the proceedings of the society, I agreed to do so, after having re-examined my records for the various years of experimentation, and not having been able to find a source of error. The paper which was submitted to you is the un- changed reprint of the draft of the lecture mentioned; thus the brevity of the exposition, as is essential for a public lecture. I am not surprised to hear your honor speak of my experiments with mis- trustful caution; I would not do otherwise in a similar case. Two points in your esteemed letter appear to be too important to be left unanswered. The first deals with the question whether one may conclude that constancy of type has been obtained if the hybrid Aa produces a plant A, and this plant in turn produces only A . Permit me to state that, as an empirical worker, I must define constancy of type as the retention of a character during the period of observation. My statements that some of the progeny of hybrids breed true to type thus in- cludes only those generations during which observations were made; it does not extend beyond them. For two generations all experiments were conducted with a fairly large number of plants. Starting with the third generation it became necessary to limit the numbers because of lack of space, so that, in each of the seven experiments, only a sample of those plants of the second generation (which either bred true or varied) could be observed further. The observations were extended over four to six generations (p. 13). Of the varieties which bred true (pp. 15-18) some plants were observed for four generations. I must further mention the case of a variety which bred true for six generations, although the parental types differed in four characters. In 1859 I obtained a very fertile descendent with large, tasty, seeds from a first generation hybrid. Since, in the following year, its progeny retained the desirable characteristics and were uniform, the variety was cultivated in our vegetable garden, and many plants were raised every year up to 1865. The parental plants were bcDg and BCdG: B. albumen yellow b. albumen green C. seed-coat grayish-brown c. seed-coat white D. pod inflated d. pod constricted G. axis long g. axis short The hybrid just mentioned was BcDG. The color of the albumen could be determined only in the plants saved for seed production, for the other pods were harvested in an immature condition. Never was green albumen observed in these plants, reddish-purple flower color (an indication of brown seed-coat), constriction of the pod, nor short axis. This is the extent of my experience. I cannot judge whether these findings would permit a decision as to constancy of type; however, I am inclined to regard the separation of parental characteristics in the progeny of hybrids in Pisum as complete, and thus permanent. The progeny of hybrids carries one or the other of the parental characteristics, or the hybrid form of the two; I have s-10 LETTER TO CARL NAGELI, 1866-1873 never observed gradual transitions between the parental characters or a pro- gressive approach toward one of them. The course of development consists simply in this; that in each generation the two parental characteristics ap- pear, separated and unchanged, and there is nothing to indicate that one of them has either inherited or taken over anything from the other. For an example, permit me to point to the packets, numbers 1035-1088, which I sent you All the seeds originated in the first generation of a hybrid in which brown and white seed-coats were combined. Out of the brown seed of this hybrid, some plants were obtained with seed-coats of a pure white color, with- out any admixture of brown. I expect those to retain the same constancy of character as found in the parental plant. The second point, on which I wish to elaborate briefly, contains the following statement: "You should regard the numerical expressions as being only empirical, because thev can not be proved rational." My experiments with single characters all lead to the same result: that from the seeds of the hybrids, plants are obtained half of which in turn carry the hybrid character (A a), the other half, however, receive the parental charac- ters A and a in equal amounts. Thus, on the average, among four plants two have the hybrid character A a, one the parental character A, and the other the parental character a. Therefore 2Ao+A+a or A + 2Aa+a is the empirical simple, developmental series for two differentiating characters. Likewise it was shown 'in an empirical manner that, if two or three differentiating characters are combined in the hybrid, the developmental series is a combination of two or three simple series. Up to this point I don't believe I can be accused of having left the realm of experimentation. If then I extend this combination of simple series to any number of differences between the two parental plants, I have indeed entered the rational domain. This seems permissible, however, because I have proved by previous experiments that the development of any two differentiating characteristics proceeds independently of any other differences. Finally, regarding my statements on the differences among the ovules and pollen cells of the hybrids; they also are based on experiments. These and similar experiments on the germ cells appear to be important, for I believe that their results furnish the explanation for the development of hy- brids as observed in Pisum. These experiments should be repeated and verified. . . I regret very much not being able to send your honor the varieties you desire. As I mentioned above, the experiments were conducted up* to and including 1863; at that time they were terminated in order to obtain space and time for the growing of other experimental plants. Therefore seeds from those experiments are no longer available. Only one experiment on differences in the time of flowering was continued; and seeds are available from the 1864 harvest of this experiment. These are the last I collected, since I had to aban- don the experiment in the following year because of devastation by the pea beetle, Bruchus pisi. In the early years of experimentation this insect was only rarely' found on the plants, in 1864 it caused considerable damage, and ap- peared in such numbers in the following summer that hardly a 4th or 5th s-11 GREGOR MENDKI. of the seeds was spared. In the last few years it has been necessary to dis- continue cultivation of peas in the vicinity of Briinn. The seeds remaining can still be useful, among them are some varieties which I expect to remain constant; they are derived from hybrids in which two, three, and four differ- entiating characters are combined. All the seeds were obtained from members of the first generation, i.e., of such plants as were grown directly from the seeds of the original hybrids. 1 should have scruples against complying with your honor's request to send these seeds for experimentation, were it not in such complete agreement with my own wishes. I fear that there has been partial loss of viability. Further- more the seeds were obtained at a time when Bruchus pisi was already ram- punt, and I cannot acquit this beetle of possibly transferring pollen; also, I must mention again that the plants were destined for a study of differences in flowering time. The other differences were also taken into account at the harvest, but with less care than in the major experiment. The legend which I have added to the packet numbers on a separate sheet is a copy of the notes I made for each individual plant, with pencil, on its envelope at the time of harvest. The dominant characters are designated as .4, B, C, D, E, F, G and as concerns their dual meaning please refer to p. 11. The recessive characters are designated a, b, c, d, e, /, g; these should remain constant in the next genera- tion. Therefore, from those seeds which stem from plants with recessive char- acters only, identical plants are expected (as regards the characters studied). Please compare the numbers of the seed packets with those in my record, to detect any possible error in the designations — each packet contains the seeds of a single plant only. Some of the varieties represented are suitable for experiments on the germ cells; their results can be obtained during the current summer. The round yellow seeds of packets 715, 730, 736, 741, 742, 745, 756, 757, and on the other hand, the green angular seeds of packets 712, 719, 734, 737, 749, and 750 can be recommended for this purpose. By repeated experiments it was proved that, if plants with green seeds are fertilized by those with yellow seeds, the albumen of the resulting seeds has lost the green color and has taken up the yellow color. The same is true for the shape of the seed. Plants with angular seeds, if fertilized by those with round or rounded seeds, produce round or rounded seeds. Thus, due to the changes induced in the color and shape of the seeds by fertilization with foreign pollen, it is possible to recognize the constitution of the fertilizing pollen. Let B designate yellow color; b, green color of the albumen. Let A designate round shape; a, angular shape of the seeds. If flowers of such plants as produce green and angular seeds by self-fertiliza- tion are fertilized with foreign pollen, and if the seeds remain green and angular, then the pollen of the donor plant was, as regards the two characters ab If the shape of the seeds is changed, the pollen was taken from lb If the color of the seeds is changed, the pollen was taken from aB If both shape and color is changed, the pollen was taken from AB s-12 LETTER TO CARL NAGELI, 1866-1873 The packets enumerated above contain round and yellow, round and green, angular and yellow, and angular and green seeds from the hybrids ab-\-AB. The round and yellow seeds would be best suited for the experiment. Among them (see experiment p. 15) the varieties AB, ABb, Aab, and AaBb may occur; thus four cases are possible when plants, grown from green and angular seeds, are fertilized by the pollen of those grown from the above mentioned round and yellow seeds, i.e. I. ab+AB II. ab+ABb III. ab+AaB IV. ab+AaBb If the hypothesis that hybrids form as many types of pollen cells as there are possible constant combination types is correct, plants of the makeup AB produce pollen of the type AB ABb " " " " " AB and Ab AaB " " " " " A B and aB AaBb " " " " " AB, Ab, a B, and ab Fertilization of ovules occurs: I. Ovules ab with pollen AB II. " ab " " AB and A b III. " ab " " AB and aB IV. " ab " " AB, Ab, aB, and ab The following varieties may be obtained from this fertilization : I. AaBb II. AaBb and Aab III. AaBb and aBb IV. AaBb, Aab, aBb, and ab If the different types of pollen are produced in equal numbers, there should be in I. All seeds round and yellow II. one half round and yellow one half round and green III. one half round and yellow one half angular and yellow IV. one quarter round and yellow one quarter round and green one quarter angular and yellow one quarter angular and green Furthermore, since the numerical relations between AB, ABb, AaB, AaBb are 1:2:2:4, among any nine plants grown from round yellow seed there should be found on the average AaBb four times, ABb and AaB twice each, and AB s-13 GREGOR MENDEL once; thus the IVth case should occur four times as frequently as the 1st and twice as frequently as the Ilnd or Illrd. If on the other hand, plants grown from the round yellow seeds mentioned are fertilized by pollen from green angular plants, the results should be exactly the same, provided that the ovules are of the same types, and formed in the same proportions, as was reported for the pollen. I have not performed this experiment myself, but I believe, on the basis of similar experiments, that one can depend on the result indicated. In the same fashion individual experiments may be performed for each of the two seed characters separately, all those round seeds which occurred to- gether with angular ones, and all the yellow ones which occurred with green seeds on the same plant are suitable. If, for instance, a plant with green seeds was fertilized by one with yellow seeds, the seeds obtained should be either 1) all yellow, or 2) half yellow and half green, since the plants originating from yellow seeds are of the varieties B and Bb. Since, furthermore, B and Bb occur in the ratio of 1:2, the 2nd fertilization will occur twice as frequently as the 1st. Regarding the other characters, the experiments may be conducted in the same way; results, however, will not be obtained until next year. . . . As must be expected, the experiments proceed slowly. At first beginning, some patience is required, but later, when several experiments are progressing concurrently, matters are improved. Every day, from spring to fall, one's in- terest is refreshed daily, and the care which must be given to one's wards is thus amply repaid. In addition, if I should, by my experiments, succeed in hastening the solution of these problems, I should be doubly happy. Accept, highly esteemed Sir, the expression of most sincere respect from Your devoted, G. Mendel (Altbriinn, Monastery of St. Thomas) Briinn, 18 April, 1867 s-14 SUPPLEMENT II THE RELATION OF GENETICS TO PHYSIOLOGY AND MEDICINE NOBEL LECTURE, PRESENTED IN STOCKHOLM ON JUNE 4, 1934 By Dr. THOMAS HUNT MORGAN DIRECTOR OF THE Wll. G. KERCKHOFF LABORATORIES, CALIFORNIA INSTITUTE OF TECHNOLOGY The study of heredity, now called genetics, has undergone such an ex- traordinary development in the present century, both in theory and in practice, that it is not possible in a short address to review even briefly all its outstanding achievements. At most I can do no more than take up a few topics for discussion. Since the group of men with whom I have worked for twenty years has been interested for the most part in the chro- mosome-mechanism of heredity, I shall first briefly describe the relation between the facts of heredity and the theory of the gene. Then I should like to discuss one of the physiological problems im- plied in the theory of the gene; and finally, I hope to say a few words about the applications of genetics to medicine. The modern theory of genetics dates from the opening years of the present century, with the discovery of Mendel's long-lost paper that had been overlooked for thirty-five years. The data obtained by de Vries in Holland, Correns in Ger- many and Tschermak in Austria showed that Mendel's laws are not confined to garden peas, but apply to other plants. A year or two later the work of Bateson and Punnett in England and Cuenot in France made it evident that the same laws apply to animals. In 1902 a young student, William Sutton, working in the laboratory of E. B. Wilson, pointed out clearly and completely that the known behavior of the chromosomes at the time of matura- tion of the germ-cells furnishes us with a mechanism that accounts for the kind of separation of the hereditary units postulated in Mendel's theory. The discovery of a mechanism, that suffices to explain both the first and the second law of Mendel, has had far-reach- ing consequences for genetic theory, espe- cially in relation to the discovery of addi- tional laws ; because the recognition of a mechanism that can be seen and followed demands that any extension of Mendel's theories must conform to such a recog- nized mechanism ; and also because the apparent exceptions to Mendel's laws, that came to light before long, might, in the absence of a known mechanism, have called forth purely fictitious modifica- tions of Mendel's laws or even seemed to invalidate their generality. We now know that some of these "exceptions" are due to newly discovered and demon- strable properties of the chromosome mechanism, and others to recognizable irregularities in the machine. Mendel knew of no processes taking place in the formation of pollen and egg- s-15 THE SCIENTIFIC MONTHLY s-16 GENETICS, PHYSIOLOGY AND MEDICINE cell that could furnish a basis for his primary assumption that the hereditary elements separate in the germ-cells in such a way that each ripe germ-cell comes to contain only one of each kind of ele- ment : but he justified the validity of this assumption by putting it to a crucial test. His analysis was a wonderful feat of reasoning. He verified his reasoning by the recognized experimental procedure of science. As a matter of fact it would not have been possible in Mendel's time to give an objective demonstration of the basic mechanism involved in the separation of the hereditary elements in the germ-cells. The preparation for this demonstration took all the thirty-five years between Mendel's paper in 1865 and 1900. It is here that the names of the most promi- nent European cytologists stand out as the discoverers of the role of the chromo- somes in the maturation of the germ-cells. It is largely a result of their work that it was possible in 1902 to relate the well- known cytological evidence to Mendel's laws. So much in retrospect. The most significant additions that have been made to Mendel's two laws may be called linkage and crossing over. In 1906 Bateson and Punnett reported a two-factor case in sweet peas that did not give the expected ratio for two pairs of characters entering the cross at the same time. By 1911 two genes had been found in Drosophila that gave sex-linked inheri- tance. It had earlier been shown that such genes lie in the X-chromosomes. Ratios were found in the second genera- tion that did not conform to Mendel's second law when these two pairs of char- acters are present, and the suggestion was made that the ratios in such cases could be explained on the basis of inter- change between the two X-chromosomes in the female. It was also pointed out that the further apart the genes for such characters happen to lie in the chromo- some, the greater the chance for inter- change to take place. This would give the approximate location of the genes with respect to other genes. By further extension and clarification of this idea it became possible, as more evidence ac- cumulated, to demonstrate that the genes lie in a single line in each chromosome. Two years previously (1909) a Bel- gian investigator, Janssens, had de- scribed a phenomenon in the conjugating chromosomes of a salamander, Batraco- seps, which he interpreted to mean that interchanges take place between homol- ogous chromosomes. This he called chiasmatypie — a phenomenon that has occupied the attention of cytologists down to the present day. Janssens' ob- servations were destined shortly to sup- ply an objective support to the demon- stration of genetic interchange between linked genes carried in the sex chromo- somes of the female Drosophila. To-day we arrange the genes in a chart or map, Fig. 1. The numbers attached express the distance of each gene from some arbitrary point taken as zero. These numbers make it possible to fore- tell how any newr character that may appear will be inherited with respect to all other characters, as soon as its cross- ing over value with respect to any other two characters is determined. This abil- ity to predict would in itself justify the construction of such maps, even if there were no other facts concerning the loca- tion of the genes ; but there is to-day direct evidence in support of the view that genes lie in a serial order in the chromosomes. What are the Genes? What is the nature of the elements of heredity that Mendel postulated as purely theoretical units? What are genes? Now that we locate them in the chromosomes are we justified in regard- ing them as material units ; as chemical bodies of a higher order than molecules? s-17 THE SCIENTIFIC MONTHLY Frankly, these are questions with which the working geneticist lias not much con- cern himself, except now and then to speculate as to the nature of the postu- lated elements. There is no consensus of opinion amongsl geneticists as to what the genes are — whether they are real or purely fictitious — because at the level at which the genetic experiments lie it does not make the slightest difference whether the gene is a hypothetical unit or whether the gene is a material particle. In either case the unit is associated with a specific chromosome, and can be localized there by purely genetic analysis. Hence, if tne gene is a material unit, it is a piece of a chromosome; if it is a fictitious unit, it must be referred to a definite location in a chromosome — the same place as on the other hypothesis. Therefore, it makes no difference in the actual work in genetics which point of view is taken. Between the characters that are used by the geneticist and the genes that the theory postulates lies the whole field of embryonic development, where the prop- erties implicit in the genes become ex- plicit in the protoplasm of the cells. Here we appear to approach a physio- logical problem, but one that is new and strange to the classical physiology of the schools. We ascribe certain general properties to the genes, in part from genetic evi- dence and in pari from microscopical ob- servations. These properties we may next consider. Since chromosomes divide in such a way that the line of genes is split (each daughter chromosome receiving exactly half of the original line) we can scarcely avoid the inference that the genes divide into exactly equal parts; but just how this takes place is not known. The anal- ogy of cell-division creates a presumption that the gene divides in the same way, but we should not forget that the rela- tively gross process involved in cell-divi- sion may seem quite inadequate to cover the refined separation of the gene into equal halves. As we do not know of any comparable division phenomena in or- ganic molecules, we must also be careful in ascribing a simple molecular constitu- tion to the gene. On the other hand, the elaborate chains of molecules built up in organic material may give us, some day, a better opportunity to picture the mo- lecular or aggregate structure of the gene and furnish a clue concerning its mode of division s-18 SUPPLEMENT III THE PRODUCTION OF MUTATIONS IF. as Darwin maintained, the adap- tiveness of living things results from natural selection, rather than from a teleological tendency in the process of variation itself, then heritable variations must, under most conditions, occur in numerous directions, so as to give a wide range of choice for the selective process. Such a state of affairs seems, however, in more or less contradiction to the com- monly held idea, to which Darwin also gave some credence, that heritable varia- tions of given kinds tend to be produced, in a fairly regular way. by given kinds of external conditions. For then we are again confronted with the difficulty, how is it that the "right kinds" of variations (i.e. the adaptive ones) manage to arise in response to the "right kinds" of con- ditions (i.e. those they are adapted to) ? Moreover, the de Yriesian notion of mu- tations does not help us in this connec- tion. On that view, there are sudden jumps, going all the way from one "ele- mentary species" to another, and involv- ing radical changes in numerous charac- ters at once, and there are relatively few different jumps to choose between. This obviously would fail to explain how, through such coarse steps, the body could have come to be so remarkably streamlined in its internal and external organization, or, in other words, so thoroughly adaptive. The older selectionists, thinking in terms of chemical reactions on a molar scale when they thought in terms of chemistry at all, did not realize suffi- ciently the ultramicroscopic random- ness of the processes causing inherited variations. The earliest mutationists failed, in addition, to appreciate the qualitative and quantitative multiplicity of mutations. It was not long, however, before the results of Baur on Antirrhi- num and of Morgan on Drosophila. sup- plemented by scattered observations on other forms, gave evidence of the occur- rence of numerous Mendelizing muta- tions, many of them small ones, in varied directions, and they showed no discover- erable relation between the type of mu- tation and the type of environment or condition of living under which it arose. These observations, then, came closer to the statistical requirements for a process of evolution which has its basis in acci- dents. In what sense, however, could the events be regarded as accidental ? Were they perhaps expressions of veiled forces working in a more determinate manner? It was more than ever evident that further investigation of the manner of occurrence of mutations was called for. If the mutations were really non-teleo- logical. with no relation between type of environment and type of change, and above all no adaptive relation, and if they were of as numerous types as the theory of natural selection would de- mand, then the great majority of the changes should be harmful in their ef- fects, just as any alterations made blind- ly in a complicated apparatus are usually detrimental to its proper functioning, and many of the larger changes should even be totally incompatible with the functioning of the whole, or, as we say. lethal. That is. strange as it may seem at first sight, we should expect most mu- tations to be disadvantageous if the the- ory of natural selection is correct. We should also expect these mainly disad- vantageous changes to be highly diversi- fied in their genetic basis. Frequency of Mutations To get exact evidence on these points required the elaboration of special genet- s-19 The Journal of Heredity ic methods, adapted to the recognition of mutations that ordinarily escape de- tection— (1) lethals, (2) changes with but small visible effects, and (3) changes without any externally visible effects but influencing the viability more or less unfavorably. It would take us too far afield to explain these techniques here. Suffice it to say that they made use of the principle according to which a chro- mosome is, as we say, "marked," by hav- ing had inserted into it to begin with one or more known mutant genes with conspicuous visible effects, to differenti- ate it from the homologous chromosome. An individual with two such differenti- ated chromosomes, when appropriately bred, will then be expected to give two groups of visibly different offspring, holding certain expected ratios to one another. If, however, a lethal mutation has occurred in one of the two chromo- somes, its existence will be made evident by the absence of the corresponding ex- pected group of offspring. Similarly, a mutated gene with invisible but some- what detrimental action, though not fully lethal, will be recognized by the fact that the corresponding group of offspring are found in smaller numbers than ex- pected. And a gene with a very small visible effect, that might be overlooked in a single individual, will have a great- ly increased chance of being seen because the given group of offspring as a whole will tend to be distinguished in this re- gard from the corresponding group de- rived from a non-mutant. In this way, it was possible in the first tests of this kind, which Alten- burg and the writer conducted, part- lv in collaboration, in 1918-19, to get definite evidence in Drosophila that the lethal mutations greatly outnumbered those with visible effects, and that among the latter the types having an obscure manifestation were more numerous than the definite conspicuous ones used in ordinary genetic work. Visible or not, the great majority had lowered viability. Tests of their genetic basis, using the newly found facts of linkage, showed them to be most varied in their locus in the chromosomes, and it could be calcu- lated by a simple extrapolative process that there must he at least hundreds, and probably thousands, of different kinds arising in the course of spontaneous mu- tation. In work done much later, employ- ing induced mutations, it was also shown (in independent experiments both of the present writer and Kerkis, and of Timofeeff and his co-workers, done in 1934) that "invisible" mutations, which by reason of one or another physiologi- cal change lower viability without being fully lethal, form the most abundant group of any detected thus far, being at least two to three times as numerous as the complete lethals. No doubt there are in addition very many, perhaps even more, with effects too small to have been detected at all by our rather crude meth- ods. It is among these that we should be most apt to find those rare accidents which, under given conditions or in given combinations with others, may happen to have some adaptive value. Tests of Timofeeff, however, have shown that even a few of the more conspicuous visible mutations do in certain combina- tions give an advantage in laboratory breeding. Because of the nature of the test whereby it is detected — the absence of an entire group of offspring bearing cer- tain conspicuous expected characters — a lethal is surer of being detected, and detected by any observer, than is the inconspicuous or invisible, merely detri- mental, mutation. Fortunately, there are relatively few borderline cases, of nearly but not quite completely lethal genes. It was this objectivity of recognition, com- bined with the fact that they were so much more numerous than conspicuous visible mutations, that made it feasible for lethals to be used as an index of mutation frequency, even though they suffer from the disadvantage of requir- ing the breeding of an individual, rather than its mere inspection, for the recogni- tion that it carries a lethal. In the earli- est published work, we (Altenburg and the author) attempted not only to find a quantitative value for the "normal" mu- tation frequency, but also to determine whether a certain condition, which we s-20 Muller: Production of Mutations considered of especial interest, affected the mutation frequency. The plan was ultimately to use the method as a gen- eral one for studying the effects of vari- ous conditions. The condition chosen for the first experiment was tempera- ture, and the results, verified by later work of the writer's, indicated that a rise of temperature, within limits normal to the organism, produced an increase of mutation frequency of about the amount to be expected if mutations were, in essentials, orthodox chemical reactions. Mutations as Chemical Reactions On this view, however, single muta- tions correspond with individual molecu- lar changes, and an extended series of mutations, in a great number of identi- cal genes in a population, spread out over thousands of years, is what cor- responds with the course of an ordinary chemical reaction that takes place in a whole collection of molecules in a test tube in the course of a fraction of a second or a few seconds. For the indi- vidual gene, in its biological setting, is far more stable than the ordinary chemi- cal molecule is, when the latter is ex- posed to a reagent in the laboratory. Thus, mutations, when taken collective- ly, should be subject to the statistical laws applying to mass reactions, but the individual mutation, corresponding to a change in one molecule, should be sub- ject to the vicissitudes of ultramicro- scopic or atomic events, and the appari- tion of a mutant individual represents an enormous amplification of such a phe- nomenon. This is a principle which gives the clue to the fact, which otherwise seems opposed to a rational, scientific and molarly deterministic point of view, that differences in external conditions or conditions of living do not appear to af- fect the occurrence of mutations, while on the other hand, even in a normal and sensibly constant environment, muta- tions of varied kinds do occur. It is also in harmony with our finding, of about the same time, that when a mutation takes place in a given gene, the other gene of identical type present nearby in the same cell usually remains unaffected though it must of course have been sub- jected to the same macroscopic physico- chemical conditions. On this conception, then, the mutations ordinarily result from submicroscopic accidents, that is, from caprices of thermal agitation, that occur on a molecular and submolecular scale. More recently Delbruck and Timofeeff, in more extended work on temperature, have shown that the amount of increase in mutation frequen- cy with rising temperature is not merely that of an ordinary test-tube chemical reaction, but in fact corresponds closely with that larger rise to be expected of a reaction as slow in absolute time rate (i.e. with as small a proportion of mo- lecular changes per unit of time) as the observed mutation frequency shows this reaction to be, and this quantitative cor- respondence helps to confirm the entire conception. Now this inference concerning the non-molar nature of the individual mu- tation process, which sets it in so differ- ent a class from most other grossly ob- servable chemical changes in nature, led naturally to the expectation that some of the "point effects" brought about by high-energy radiation like X-rays would also work to produce alternations in the hereditary material. For if even the rela- tively mild events of thermal agitation can, some of them, have such conse- quences, surely the energetically far more potent changes caused by power- ful radiation should succeed. And, as a matter of fact, our trials of X-rays, car- ried out with the same kind of genetic methods as previously used for tempera- ture, proved that such radiation is ex- tremely effective, and inordinately more so than a mere temperature rise, since by this method it was possible to obtain, by a half hour's treatment, over a hun- dred times as many mutations in a group of treated cells as would have occurred in them spontaneously in the course of a whole generation. These mutations, too, were found ordinarily to occur pointwise and randomly, in one gene at a time, without affecting an identical s-21 The Journal of Heredity gene that might he present nearby in a homologous chromosome. Radiation-Effects In addition to the individual gene changes, radiation also produced rear- rangements of parts of chromosomes. As our later work (including that with co-workers, especially Raychaudhuri and Pontecorvo) has shown, these latter were caused in the first place by hreak- ages of the chromosomes, followed after- wards by attachments occurring hetween the adhesive broken ends, that joined them in a different order than before. The two or more breaks involved in such a rearrangement may be far apart, caused by independent hits, and thus re- sult in what we call a gross structural change. Such changes are of various kinds, depending upon just where the breaks are and just which broken ends become attached to which. But, though the effects of the individual "hits" are rather narrowly localized, it is not un- common for two breaks to be produced at nearby points by what amounts to one local change (or at any rate one lo- calized group of changes) whose influ- ence becomes somewhat spread out. By the rejoining, ;n a new order, of broken ends resulting from two such nearby breaks, a minute change of sequence of the genes is brought about. More usual- ly, the small piece between the two breaks becomes lost (a "deficiency"), but sometimes it becomes inverted, or even becomes transferred into a totally different position, made available by a separate hit. Both earlier and later work by collabo- rators (Oliver, Hanson, etc.) showed definitely that the frequency of the gene mutations is directly and simply propor- tional to the dose of irradiation applied, and this despite the wave length used, whether X- or gamma or even beta rays, and despite the timing of the irradiation. These facts have since been established with great exactitude and detail, more especially by Timofeeff and his co- workers. In our more recent work with Raychaudhuri these principles have been extended to total doses as low as 400r, and rates as low as .01 r per min- ute, with gamma rays. They leave, we believe, no escape from the conclusion that there is no threshold dose, and that the individual mutations result from in- dividual "hits", producing genetic ef- fects in their immediate neighborhood. Whether these so-called "hits" are the individual ionizations, or may even be the activations that occur at lower ener- gy levels, or whether, at the other end of the scale, they require the clustering of ionizations that occurs at the termini of electron tracks and of their side branches (as Lea and Fano point out might be the case), is as yet undecided. But in any case they are, even when microscopically considered, what we have termed "point mutations," as they involve only disturbances on an ultra- microscopically localized scale. And whether or not they are to occur at any particular point is entirely a matter of accident, using this term in the sense in which it is employed in the mathematics of statistics. Naturally, other agents than photons which produce effects of this kind must also produce mutations, as has been shown by students and collaborators working under Altenburg in Houston for neutrons (Nagai and Locher) and for alpha rays (Ward) and confirmed by Timofeeff and his co-workers (Zim- mer, et al). Moreover, as Alten- burg showed, even the smaller quan- tum changes induced by ultraviolet exert this effect on the genes. They cause, however, only a relatively small amount of rearrangement of chromosome parts (Muller and Mackenzie), and, in fact, they also tend to inhibit such rearrange- ment, as Swanson, followed by Kauff- mann and Hollaender. has found. Since the effective ultraviolet hits are in the form of randomly scattered single-atom changes in the purines and pyrimidines of the chromosome, rather than in groups of atom changes, it seems likely that clusters of ionizations are not neces- sary for the gene mutation effects, at any rate, although we cannot be sure of this until the relation of mutation frequency to dosage is better known for this agent. s-22 Muller: Production of Mutations Induced and Natural Mutations Inasmuch as the changes brought about iu the genes by radiation must certainly be of an accidental nature, un- premeditated, ateleological, without ref- erence to the value of the end result for the organism or its descendants, it is of interest to compare them with the so- called spontaneous or natural mutations. For in the radiation mutations we have a yardstick of what really random changes should be. Now it is found in Drosophila that the radiation-induced mutations of the genes (we exclude here the demonstrable chromosome rear- rangements) are in every respect which has been investigated of the same es- sential nature as those arising naturally in the laboratory or field. They usually occur in one gene without affecting an identical one nearby. They are distrib- uted similarly in the chromosomes. The effects, similarly, may be large or small and there is a similar ratio of fully lethal to so-called visible gene mutations. That is, the radiation mutations of the genes do not give evidence of being more dele- terious. And when one concentrates at- tention upon given genes one finds that a whole series of different forms, or al- leles, may be produced, of a similar and in many cases sensibly identical nature in the two cases. In fact, every natural mutation, when searched for long enough, is found to be producible also by radiation. Moreover, under any given condition of living tried, without radia- tion, the effects appear as scattered as when radiation is applied, even though of much lower frequency. All this sure- ly means then, does it not, that the natu- ral mutations have in truth no innate tendency to be adaptive, nor even to be different, as a whole group, under some natural conditions than under others? In other words, they cannot be determi- nate in a molar sense, but must them- selves be caused by the ultramicroscopic accidents of the molecular and submolec- ular motions, i.e. by the individual quan- tum exchanges of thermal agitation, tak- ing this word in a broad sense. The only escape from this would be to suppose that they are caused by the radiation present in nature, resulting from natural radioactive substances and cosmic rays, but a little calculation (by Mott-Smith and the writer, corroborated by others) has shown that this radiation is quite inadequate in amount to account for the majority of mutations occurring in most organisms. But to say that most natural muta- tions are the results of the quantum ex- changes of thermal agitation, and. fur- ther, that a given energy level must be reached to produce them, does not, as some authors have seemed to imply, mean that the physicochemical condi- tions in and around the organism, other than temperature, have no influence upon their chance of occurrence. That such circumstances may play a decided role was early evident from the studies of spontaneous mutation frequency, when it was found (1921, reporte'd 1928) that the frequency in one ex- periment, with one genetic stock, might be ten times as high as in another, with another stock. And more recently we have found that, in different portions of the natural life cycle of the same in- dividual, the mutation frequency may be very different. Finally, in the work of Auerbach and Robson, with mustard gas and related substances, it has been proved that these chemicals may induce mutations at as high a frequency as a heavy dose of X-rays. In all these cases, however, the effects are similarly scat- tered at random, individually uncon- trolled, and similarly non-adaptive. It should also be noted in this con- nection that the genes are not under all conditions equally vulnerable to the mutating effects of X-rays themselves. Genes in the condensed chromosomes of spermatozoa, for example, appear to be changed more easily than those in the more usual "resting" stages. We have mentioned that, as Swanson has shown, ultraviolet exerts besides its own mutat- ing effect an inhibition on the process of chromosome breakage, or at any rate on that of reunion of the broken parts in a new viable order, while infrared, in Hol- laender's and Kaufmann's recent experi- s-23 The Journal of Heredity ments, has a contrary action. And Stadler, in his great work on the pro- duction of mutations in cereals, started independently of our own, has obtained evidence that in this material X-radiation in the doses used is unable to produce a sensible rise in the gene mutation fre- quency, though numerous chromosome breakages do arise, leading to both gross and minute rearrangements of chromo- some parts, bather the genes are more resistant in this material to permanent changes by X-rays as compared with their responsiveness to thermal agita- tion, or a break or loss must usually be produced by X-rays along with the gene change. The milder ultraviolet quanta, on the other hand, do produce gene mu- tations like the natural ones in these plants. Such variations in effectiveness are, I believe, to have been expected. They do not shake our conclusion as to the acci- dental, quantum character of the event which usually initiates a gene mutation. But they .^ive rise to the hope that, through further study of them, more may be learned concerning the nature of the mutation process, as well as of the genetic material that undergoes the changes. Controlled Mutation? No one can answer the question whether some special means may not be found whereby, through the application of molar influences, such as specific anti- bodies, individual genes could be changed to order. Certainly the search for such influences, and for increasing control of things on a microscopic and submicro- scopic scale as well, must be carried further. But there is as yet no good evi- dence that anything of the sort has been done artificially, or that it occurs natu- rally. Even if possible, there could be no generalized method of control of gene composition without far greater knowledge than we now have of the in- timate chemical structure and the mode of working of the most complicated and diverse substances that exist, namely. nucleoproteins, proteins in general, and enzvmes. The works of Sumner, North- rup and Stanley, together with those of other protein chemists, point the way in this direction, but everyone will agree that it is a long and devious system of roads which is beginning here. It is true that some cases are known of mutable genes which change selec- tively in response to special conditions. Such cases may be very informative in shedding light on gene structure, but we have as yet no indication that the al- terations of these genes, which in the great majority of instances known are abnormal genes, have anything in com- mon with ordinary natural mutations. It is also true that cases are known among bacteria and viruses of the induction of particular kinds of hereditary changes by application of particular substances, but here the substances applied are in each case the same as those whose pres- ence is later found to have been induced, and so there is every reason to infer that they have in fact become implanted in some way, that is, that we do not really have a specifically induced mutation. So far, then, we have no means, or prospect of means, of inducing given mutations at will in normal material ; though the production of mutations in abundance at random may be regarded as a first step along such a path, if there is to be such a path. So long as we cannot direct mutations, then, selection is indispensable, and progressive change in the hereditary constitution of a living thing can be made only with the aid of a most thoroughgoing selection of the mu- tations that occur since, being non-adap- tive except by accident, an overwhelming majority is always harmful. For a sen- sible advance, usually a considerable number of rare steps must be accu- mulated in this painful selective pro- cess. By far the most of these are indi- vidually small steps, but, as species and race crossings have shown, there may be a few large distinctive steps that have been, as Huxley terms it, "buffered", by small changes that readjust the organ- ism to them. Not only is this accumula- tion of many rare, mainly tiny changes the chief means of artificial animal and plant improvement, but it is, even more. s-24 Muller: Production of Mutations the way in which natural evolution has occurred, under the guidance of natural selection. Thus the Darwinian theory becomes implemented, and freed from the accretions of directed variation and of Lamarckism that once encumbered it. It is probable that, in a state of nature, most species have a not very much (though somewhat) lower frequency of gene mutation than would be most ad- vantageous for them, in consideration of the degree of rigor of the natural selec- tion that occurs in the given species. A much higher frequency would probably lead to faster genetic degenerative pro- cesses than the existing selection could well cope with. But, under conditions of artificial breeding, where selection can be made more effective, a higher mutation frequency can for a time at least be tolerated in some cases, and larger mutations also can be nursed through to the point where they become suitably buffered. Here it may become of practical use to apply X-rays, ultra- violet, or other means of inducing mu- tations, as Gustafsson especially has demonstrated for X-rays. This will be especially true in species which natural- ly undergo much inbreeding, or in which there is a well expressed haploid phase, or a considerable haploid portion of the genotype, for under these circumstances many of the spontaneous mutations that might otherwise have accumulated in the population and that could be brought to light by inbreeding, will have become eliminated before they could be found, and the natural mutation rate itself will be lower. We have above largely confined our- selves to considering the relation of the production of gene mutations to the problems of the general method of evo- lution, including that of the nature of hereditary variation, because this has been, historically, the main line of ap- proach to the subject of artificial muta- tions. It was from the first evident, how- ever, that the production of mutations would, as we once stated, provide us with tools of the greatest nicety, where- with to dissect piece by piece the physio- logical, embryological, and biochemical structure of the organism and to analyze its workings. Already with natural mu- tations, such works as those of Bon- nevie, Grueneberg, Scott-Moncrief, Eph- russi and Beadle, etc., have shown how the intensive tracing of the effects, and interrelations of effects, of just one or a few mutations, can lead to a deeper un- derstanding of the complex processes whereby the genes operate to produce the organism. But there are thousands of genes, and it is desirable to be able to choose them for study in an orderly fash- ion as we proceed with our dissection process. For this purpose we have thought that it would often be advan- tageous to produce mutations artificially in abundance, so as then to take our pick of those more suited for successive steps in our analysis. The work of Bea- dle and his co-workers on Ncurospora in recent years, followed by similar work of Malin and Fries and of others, has brilliantly shown the applicability of this method for studies of the paths of bio- chemical synthesis of amino-acids, vita- mins, purines and pyrimidines. And yet, in a sense, the surface of the subject as a whole has barely been scratched, and we may look forward with confidence to the combination of this technique with that of tracer substances and with all the other techniques of biochemistrv, physi- ology and experimental embryology, for the increasing unravelling of that sur- passingly intricate tangle of processes of which the living organism is constituted. There is no time, however, to go further into this subject here. Chromosome Analysis For we cannot neglect here a brief outline of another phase of the artificial mutation work, more specifically of in- terest to geneticists: that is, the further analysis of the properties of the chromo- somes and their parts, gained chiefly from studies in which parts have been removed, added, or rearranged. We have already spoken, in passing, of the studies of the mechanism of such structural change, in which a relatively simple gen- eral scheme lying at the basis of all such alterations has emerged: namely, break - s-25 The Journal of Heredity age first, followed by adhesion oi broken ends. It was early evident thai by the use of such rearranged chromosomes ad- ditional proof of the physical validity of the linkage maps could be obtained, and this was done (Muller and Painter). Furthermore, it has been possible to throw light on problems of crossing over, as in the demonstration (Muller, Stone, and ( MTermann > that to whatever posi- tion the centromere is moved, it causes a strong inhibition of crossing over, the strenth of which gradually diminishes with distance. Moreover, the same proves to be true of any point of dis- continuity in pairing, caused by hetero- zygosity in regard to a structural change. Such studies on crossing over, and on the pairing forces that affect segregation, are still capable of considerable exten- sion. We must remember, in speaking of the centromere and other apparently distinc- tive chromosome parts, that we have no right to infer that they are autonomous, locally determined structures, dependent on the genes of the regions in which thev are seen to lie, before observations have been made that show the effects of removing or displacing those regions. Therefore, it has in the main been neces- sary to wait for the study of induced inversions, deletions and translocations of chromosomes, before the inference could be secure that the centromere is, in most instances, such an autonomous organelle, dependent upon a gene or genes in the immediate neighborhood (but not in all instances in the neighbor- hood, as Rhoades has recentlv shown in a special strain of maize). Similarly, it has been possible to show (despite some contrary claims, the validity or invalidi- tv of which cannot be discussed here) that the free end of the chromosome, or telomere, constitutes in much material a locally determined distinctive structure. By a combined genetic and cvtological analysis of various cases of breakage and rearrangement of parts, it was found that there are distinctive, largely locally determined, regions of the chromosomes, usually most markedly developed near the centromeres, which we at first called "inactive" but which are now usually referred to as "heterochromatic." These were also found independently in pure- ly cytological studies by lleitz. It would be fascinating to enter here into a dis- cussion of the remarkable peculiarities which the cytogenetic studies have shown these regions to have — the evidence of repetition of more or less similar parts, of a tendency to conjugation between the differently placed parts, of distinc- tive cytological appearance correlated with whether or not such conjugation occurs, of inordinately high tendency to structural change, of strong influence of certain of their genes upon segregation, etc.. — and then to go on to discuss hy- potheses of their evolutionary origin and their functions. This would unfortunate- ly take us too far afield. We must, how- ever, insist upon one point — as it is not yet generally enough recognized. — name- ly, that the evidence is very strong that what, in the Drosophila chromosome as seen at mitosis, is called "the heterochro- matic region," is simply a large tempo- rary body of accessory, non-genic nuc- leoprotein, produced under the influence of one or two particular genes from among the dozen or more that constitute the whole heterochromatic region, as de- tected by genetic analysis and by the chromosome as seen at the resting stage (as in the salivary gland). And it is not these conspicuous non-genic blocks which are responsible for the other known peculiarities of the heterochro- matin, above mentioned — the function of the blocks is still undetermined. In other words, the so-called "heterochro- matin" with which the cytologist deals in studying mitotic chromosomes is a ouite different thing from, although in the neighborhood of. the heterochro- matin proper having the above described complex of properties. Moreover, it has been possible to show (Sutton-Gersh in collaboration with the author, unpub- lished) that the conspicuous nucleoli often associated with the heterochromat- in are produced under the influence of still other autonomous genes in it. that are senarafe from those for the mitotical- ly visible blocks. ■26 Muller: Production of Mutationi One of the most interesting findings which has come out of the study of Dro- sophila chromosomes that underwent re- arrangement of parts as a result of ir- radiation has been the generalization of the existence of the phenomenon known as "position effect. " This effect was first found by Sturtevant in the case of the spontaneous mutant known as Bar eye, but it was not known to what extent the effect might be a special one until numerous rearrangements could be studied. The term position effect im- plies that the functioning of a gene is to a certain extent at least dependent upon what other genes lie in its neighborhood. There is now adequate evidence that this is a general principle, applying to very many if not all the genes in Droso- phila, and that their functioning can be qualitatively as well as quantitatively conditioned by the character of the genes in their vicinity, some of the genes hav- ing much more effect than others and different genes working in different ways and to different extents. It is possible that, as Sturtevant sug- gested, the position effect is caused by the interaction between gene products in the vicinity of the genes producing them, assuming that such products are more concentrated there and under such cir- cumstances tend to react more with one another than when dispersed. However, the interpretation which we favor is that the functioning of the gene is affected by its shape and that this, in turn, varies with the strength and nature of synaptic forces acting on the region of the chro- mosome in which it lies. These might consist of forces directly exerted on the gene by other genes, whether allelic or not (Muller), or they might be resul- tants of the state of spiralization, etc., of the chromosome region, circumstances which in their turn are in part dependent on synaptic forces (Ephrussi and Sut- ton). This interpretation, in either of its variants, would explain why position effects are so much more general in Drosophila, an organism in which the synaptic forces are known to operate strongly even in somatic cells, than in other organisms tested, in which such forces are much weaker or absent in somatic 'cells. It would also fit in with the author's findings that the hetero- chromatic regions tend to have especial- ly strong, extensive, and distinctive kinds of position effects, effects varying in degree with the total amount of heter- ochromatin present in a cell, as well as with vacillating embryological factors. For these genetic findings are in con- formity with the cytological effects of heterochromatin, observed first by Pro- kofyeva, on the degree of extension, synaptic properties, etc., of euchromatin in its neighborhood, effects which she showed to be subject to similar vacilla- tions, that are correlated with the varia- tions in the phenotypically observed po- sition effects. Recent observations, both by Ephrussi and by Sutton, following suggestions :>f the author, and by Stern, also seem to point in this direction, for they show an influence, on the position effects exhibited by given parts, of the arrangement of homologous chromosome parts. If this interpretation based on gene shape should hold, it would open up a new angle of attack on the struc- ture and method of functioning of the gene, perhaps ultimately relating it to nucleoprotein composition and proper- ties. Another use to which the process of breakage and rearrangement of chromo- some parts by irradiation has been put is for the study of the effects of adding and of subtracting small pieces of chro- mosomes, in order to determine the re- lation of gene dosage to gene expression. In this way, it has been found out (1) that most normal genes are, even in sin- gle dose, near the upper limit of their effectiveness, and (2) that most mutant genes have a final effect qualitatively similar to but Quantitatively less than that of their allelic normal gene. The dominance of normal genes over their mutant alleles, then, turns out in most instances to be a special ca=e of the prin- ciple that one dose of a normal gene usu- ally produces nearly though not quite as much effect as two doses. This in turn is best understood as resulting from a long course of selection of the normal s-27 The Journal of Heredity gene and its modifiers for stability of ex- pression, when under the influence of environmental and genetic conditions which would affect the gene's operation quantitatively, i.e. in a manner similar to that of dosage changes. This does not mean that selection has specifically worked to produce dominance of the normal gene over its alleles, however, because ( 3 ) not all mutant genes be- have merely like weaker normal genes, and (4) those which the dosage tests show to produce qualitatively different effects from the normal genes seem oft- ener to escape from the principle of be- ing dominated over by the normals, just as would be expected on our hypothesis. Among the further results of gene dosage studies carried out by the use of chromosome fragments produced by irradiation, attention should be especial- ly called to the findings coming under the head of "dosage compensation." These have shown (1) that, when the dosage of virtually all genes in the X chromosome except a given one is held constant, the expression of that one is usually so very nearly the same when present in one dose as in two that no difference in the character can ordinarily be seen, and (2) that nevertheless this invisible difference has been so impor- tant for the organism that, in the course of the past natural selection, a system of modifying genes, called compensators, has been established, having the function of making the effects of the one and two doses nor- mally present in the two respective sexes much more nearly equal still, when these dosage differences in the given genes are present simultaneously with those in all the other X -chromosomal genes. Each gene seems to have acquired a different system of compensators, the interrela- tions of all together being extremely complicated. This then gives evidence from a new angle of the meticulousness of natural selection, of the very precise adaptiveness of the characters existing in a species, and of the final grade of a character having ordinarily become es- tablished through the accumulation of numerous small mutations having very complex functional relations with one another. It is in line with our previous thesis of evolution through the selection of multitudinous tiny accidental changes. When attention is concentrated on a given very circumscribed region of a chromosome, by a comparison of vari- ous induced rearrangements all of which have a point of breakage within that re- gion, other facts come to light, bearing on the problems of chromosome and gene divisibility. By means of special genetic methods, which cannot be de- tailed here, evidence has been obtained that the breaks in any such limited re- gion tend to occur at specific points, giving indication that discrete units or segments lie between these points, and thus arguing against the idea of the chro- mosome being a continuum and in favor of its genes corresponding to physical entities rather than merely to concepts arbitrarily set up for the convenience of geneticists. We are also enabled in this way to make estimates of the probable number of genes in the chromosome, as well as to get maximally limiting figures for their size. These estimates agree as closely as could have been expected with those based on previous genetic work, using entirely different methods, al- though net with the estimates based on the "sensitive volume" hypothesis. Duplications and Evolution Another finding made in studies of cases having a small fragment of chro- mosome moved, as a result of irradia- tion, to another position, was that indi- viduals are frequently able to survive and reproduce even when they have the given chromosome part present in its orieinal position as well as in the new position. In fact, it was in work of this kind that the effect of extra doses of genes was determined. Now, in some of these cases stocks could even be obtained which were homozygous for the dupli- cated piece as well as for the original piece. This led to the idea that duplica- tions of chromosome material might in this manner have become established in the previous course of evolution. When, in the analysis of a limited region of the 5-28 Muller: Production of Mutations X chromosome, including the locus of the so-called "scute" effect, it was found that there are in fact, within the normal X chromosome, two genes of closely re- lated effect ("achaete" and "scute") very close or adjacent to one another, it became evident that this was in all probability an example of the above postulated occurrence. This then showed the way, and apparently the main if not the only way (aside from the far rarer phenomena of polyploidy and "tetra- somy"), by which the number of genes has become increased during the course of evolution. By a curious coincidence. Bridges was at the same time making his studies of salivary chromosomes and finding direct cytological evidence for the existence of such "repeats," as he called them, in the normal chromosome, and he interpreted these in the same manner. In the twelve years since that time, various other clear cases of the same kind have been demonstrated. Thus, increase in gene number, brought about by the duplication of small parts of chromosomes, more usually in posi- tions near their original ones, must be set down as one of the major pro- cesses in evolution, in addition to the mutations in the individual genes. By itself, this process would not be of great importance, but it becomes important because, by allowing gene mutations to come afterwards that differentiate the genes in one position from the originally identical ones in the other position, the number of different kinds of genes is in- creased and so the germ plasm, and with it the processes of development and the organism as a whole, are eventually en- abled to grow more complex. Rearrangements of chromosome parts which do not lead to an increase in gene number can of course also occur in evo- lution, although it is unlikely that their role is so fundamental. By producing such changes in the laboratory it has been possible to find out a good deal more about what types can arise, and what their properties are. Various in- ferences can then be drawn concerning the viability and fertility that the differ- ent types would have, under varied ge- netic circumstances, and whether they would tend to become eliminated or to accumulate in a population of a given type. Some of them can be shown to have, under given conditions, an evolu- tionary survival value, both by aiding in the process of genetic isolation and in other ways, as by affecting heterosis. In this manner, evolutionary inferences have been drawn which have later been confirmed by comparison of the chromo- some differences actually existing be- tween related races, sub-species, and species. Probably of greater ultimate interest will be the results of studies of gene mu- tations occurring at individual loci. Radi- ation mutations are frequent enough to lend themselves to comparisons of the potentialities of different loci, although not nearly enough has yet been done along these lines. Similarly, a compari- son of the different mutations which can occur at the same locus can lead to very important results, especially since it has been shown that the different alleles may have every complex relationships to one another, so as even, in some cases, to reconstitute the normal type when they are crossed together. The way in which genes may change as a result of succes- sive mutations remains to be gone into at much greater length. So, too, does the question of changes in gene muta- bility, brought about by gene mutation itself. Somatic Radiation Effects The further the analysis of the genetic effects of irradiation has gone, particu- larly of the breakage "and rearrangement of chromosome parts, the more does our conviction grow that a large proportion if not the great majority of the somatic effects of irradiation that have been ob- served by medical men and by students of embryology, regeneration, and gen- eral biology, arise secondarily as conse- quences of genetic effects produced in the somatic cells The usefulness of this interpretation has been shown in recent studies of Roller, dealing with improved methods of irradiation of mammalian carcinoma. This is too large a subject s-29 The Journal of Heredity to digress upon here, but it is to be noted that it has been the analyses based in the first place on genetic and cytogenetic studies of the reproductive cell.--. a> shown by subsequent generations, which are tints helping to clear the way for an understanding of the mechanism by which radiation acts in inhibiting . th. in causing sterilization, in pro- ducing necrosis and burns, in causing recession of malignant tissue, and per- haps also, on occasion at least, in induc- ing the initiation of such tissue. During the war years, a curious con- firmation of the correctness of the above inference regarding the nature of the somatic effects of irradiation has come to light. While working with mustard gas in Edinburgh. J. H. Robson was struck with the remarkable similarity be- tween the somatic effects of this agent and those produced by X-ray and radi- um irradiation. This led him to wonder whether perhaps mustard gas might pro- duce genetic changes of essentially the same kind as those known to be brought about by irradiation. Comprehensive ex- periments were thereupon undertaken by C. Auerbach, working in collabora- tion with Robson, and (as mentioned on p. 263) she succeeded in showing that in fact this substance does produce mutations, both in the individual genes and by breakage and rearrangement of chromosome parts, such as X-rays and radium do, and in similar abundance. Other substances of the same general group were then found to have a similar effect. This constitutes the first decided break in the chemical attack on muta- tion. The fact that these findings were made as a direct result of the above in- ference, when so many previous at- tempts to produce mutations by chemi- cal means had failed, appears to provide strong evidence that these peculiar so- matic effects are in truth consequences of the more underlying ones which, when occurring in the germ cells, are analvzed bv the geneticist in his breeding tests. There are, however, some very interest- ing differences between the nature of the genetic effects of irradiation and of these chemicals, which we cannot go into here, but which give promise of allowing an extension of the genetic and somatic analyses. We see then that production of muta- tions by radiation is a method, capable of being turned in various directions, both for the analysis of the germ plasm itself, and of the organism which is in a sense an outgrowth of that germ plasm. It is to be hoped that it may also, in certain fields, prove of increasing prac- tical use in plant and animal improve- ment, in the service of man. So far as direct practical application in man him- self is concerned, however, we are as yet a long way from practicing any in- tentional selection over our own germ plasm, although like most species we are already encumbered by countless un- desirable mutations, from which no in- dividual is immune. In this situation we can, however, draw the practical lesson, from the fact of the great majority of mutations being undesirable, that their further random production in ourselves should so far as possible be rigorously avoided. As we can infer with certainty from experiments on lower organisms that all high-energy radiation must pro- duce such mutations in man, it becomes an obligation for radiologists — though one far too little observed as yet in most countries — to insist that the simple pre-, cautions are taken which are necessary for shielding the gonads, whenever peo- ple are exposed to such radiation, either in industry or in medical practice. And. with the coming increasing use of atomic energy, even for peacetime purposes, the problem will become very important of insuring that the human germ plasm — the all-important material of which we are the temporary custodians — is effec- tively protected from this additional and potent source of permanent contamina- tion. s-30 SUPPLEMENT IV Maurice H.F.Wilkins The molecular configuration of nucleic acids Nobel Lecture, December u, ig62 Nucleic acids arc basically simple. They are at the root of very fundamental biological processes, growth and inheritance. The simplicity of nucleic acid molecular structure and of its relation to function expresses the underlying simplicity of the biological phenomena, clarifies their nature, and has given rise to the first extensive interpretation of living processes in terms of macro- molecular structure. These matters have only become clear by an unprece- dented combination of biological, chemical and physical studies, ranging from genetics to hydrogen-bond stereochemistry. I shall not discuss all this here but concentrate on the field in which I have worked, and show how X-ray diffraction analysis has made its contribution. I shall describe some of the background of my own researches, for I suspect I am not alone in finding such accounts often more interesting than general reviews. Early Background I took a physics degree at Cambridge in 1938, with some training in X-ray crystallography. This X-ray background was influenced by J. D. Bernal, then at the Cavendish. I began research at Birmingham, under J. T. Randall, studying luminescence and how electrons move in crystals. My contem- poraries at Cambridge had mainly been interested in elementary particles, but the organization of the solid state and the special properties which depended on this organization interested me more. This may have been a forerunner of my interest in biological macromolecules and how their struc- ture related to their highly specific properties which so largely determine the processes of life. During the war I took part in making the atomic bomb. When the war was ending, I, like many others, cast around for a new field of research. Partly on account of the bomb, I had lost some interest in physics. I was therefore very interested when I read Schrodinger's book « What is Life? » and was struck by the concept of a highly complex molecular structure s-31 MOLECULAR CONFIGURATION OF NUCLEIC ACIDS which controlled living processes. Research on such matters seemed more ambitious than solid-state physics. At that time many leading physicists such as Massey, Oliphant, and Randall (and later I learned that Bohr shared their view) believed that physics would contribute significantly to biology; their advice encouraged me to move into biology. I went to work in the Physics Department at St. Andrews, Scotland, where Randall had invited me to join a biophysics project he had begun. Stimulated by Mullcr's experimental modification, by means of X-radia- tion, of genetic substance, I thought it might be interesting to investigate the effects of ultrasonics; but the results were not very encouraging. The biophysics work then moved to King's College, London, where Randall took the Wheatstone Chair of Physics and built up, with the help of the Medical Research Council, an unusual laboratory for a Physics De- partment, where biologists, biochemists and others worked with the phys- icists. He suggested I might take over some ultraviolet microscope studies of the quantities of nucleic acids in cells. This work followed that of Cas- persson, but made use of the achromatism of reflecting microscopes. By this time, the work of Caspersson1 and Brachet2 had made the scientific world generally aware that nucleic acids had important biological roles which were connected with protein synthesis. The idea that DNA might itself be the genetic substance was, however, barely hinted at. Its function in chromosomes was supposed to be associated with replication of the pro- tein chromosome thread. The work of Avery, MacLeod, and McCarty3, showing that bacteria could be genetically transformed by DNA, was pub- lished in 1944, but even in 1946 seemed almost unknown, or if known, its significance was often belittled. It was fascinating to look through microscopes at chromosomes in cells, but I began to feel that as a physicist I might contribute more to biology by studying macromolecules isolated from cells. I was encouraged in this by Gerald Oster who came from Stanley's virus laboratory and interested me in particles of tobacco mosaic virus. As Caspersson had shown, ultraviolet microscopes could be used to find the orientation of ultraviolet absorbing groups in molecules as well as to measure quantities of nucleic acids in cells. Bill Seeds and I studied DNA, proteins, tobacco mosaic virus, vitamin BI2, etc. While examining oriented films of DNA prepared for ultraviolet di- chroism studies, I saw in the polarizing microscope extremely uniform fibres giving clear extinction between crossed nicols. I found the tibres had been produced unwittingly while I was manipulating DNA gel. Each time s-32 19 6 2 M. H. F. WI LKI N S r i I • * 1 Fig. 1. One of the first X-ray diffraction photographs of DNA taken in our lab- oratory. This may be compared with the later photograph in Fig. 10. (Photograph with R. Gosling; DNA by R. Signer). that I touched the gel with a glass rod and removed the rod, a thin and al- most invisible fibre of DNA was drawn out like a filament of spider's web. The perfection and uniformity of the fibres suggested that the molecules in them were regularly arranged. I immediately thought the fibres might be excellent objects to study by X-ray diffraction analysis. I took them to Ray- s-33 M O LE C 1 LAB I ONFIGURATION OF NUCLEIC ACIDS tnond Gosling, who bad our only X-ray equipment (made from war-surplus radiography parts) and who was using it to obtain diffraction photographs from heads of ram spermatozoa. This research was directed by Randall, who had been trained under W. L. Bragg and had worked with X-ray diffraction. Almost immediately, ( rosling obtained very encouraging diffraction patterns (see fig. i ). One reason for this success was that we kept the fibres moist. We remembered that, to obtain detailed X-ray patterns from proteins, Bernal had kept protein crystals m their mother liquor. It seemed likely that the configuration of all kinds of water-soluble biological macromolecules would depend on their aqueous environment. We obtained good diffraction pat- terns with ON A made by Signer and Sch wander4 which Singer brought to London to a Faraday Society meeting on nucleic acids and which he gener- ously distributed so that all workers, using their various techniques, could studv it. Realization that the Genetic Material was a Pure Chemical Substance and Signs that its Molecular Structure was Singularly Simple Between 1946 and 1950 many lines of evidence were uncovered indicating that the genetic substance was DNA, not protein or nucleoprotcin. For instance, it was found that the DNA content of a set of chromosomes was constant, and that DNA from a given species had a constant composition although the nucleotide sequence in DNA molecules was complex. It was suggested that genetic information was carried in the polynucleotide chain in a complicated sequence of the four nucleotides. The great significance of bac- terial transformation now became generally recognized, and the demonstra- tion by Hershcy and Chase5 that bacteriophage DNA carried the viral genet- ic information from parent to progeny helped to complete what was a fairly considerable revolution in thought. The prospects of elucidating genetic function in terms of molecular struc- ture were greatly improved when it was known that the genetic substance was DNA, which had a well-defined chemical structure, rather than an ill— defined nucleoprotcin. There were many indications of simplicity and reg- ularity in DNA structure. The chemists had shown that DNA was a polymer in which the phosphate and deoxyribose parts of the molecule were regularly repeated in a polynucleotide chain with 3 '-5' linkages. Chargaff() discovered an important regularity: although the sequence of bases along the poly- s-34 I962 M. H. F. WI LKI N S nucleotide chains was complex and the base composition of different DNA's varied considerably, the numbers of adenine and thymine groups were always equal, and so were the numbers of guanine and cytosine. In the electron microscope, DNA was seen as a uniform unbranched thread of diameter about 20 A. Signer, Caspersson, and Hammarsten7 showed by flow-bire- fringence measurements that the bases in DNA lay with their planes roughly perpendicular to the length of the thread-like molecule. Their ultraviolet dichroism measurements gave the same results and showed marked par- allelism of the bases in the DNA in heads of spermatozoa. Earlier, Schmidt8 and Pattri9 had studied optically the remarkable ordering of the genetic material in sperm heads. Astbury10 made pioneer X-ray diffraction studies of DNA fibres and found evidence of considerable regularity in DNA; he correctly interpreted the strong 3.4 A reflection as being due to planar bases stacked on each other. The electro-titrometric study by Gulland and Jordan11 showed that the bases were hydrogen-bonded together, and indeed Gulland12 suggested that the polynucleotide chains might be linked by these hydrogen bonds to form multi-chain micelles. Thus the remarkable conclusion that a pure chemical substance was in- vested with a deeply significant biological activity coincided with a consid- erable growth of many-sided knowledge of the nature of the substance. Meanwhile we began to obtain detailed X-ray diffraction data from DNA. This was the only type of data that could provide an adequate description of the 3-dimensional configuration of the molecule. The Need for Combining X-ray Diffraction Studies of DNA with Molecular Model-Building As soon as good diffraction patterns were obtained from fibres of DNA, great interest was aroused. In our laboratory, Alex Stokes provided a theory of diffraction from helical DNA. Rosalind Franklin (who died some years later at the peak of her career) made very valuable contributions to the X-ray analysis. In Cambridge, at the Medical Research Council laboratory where structures of biological macromolecules were studied, my friends Francis Crick and Jim Watson were deeply interested in DNA structure. Watson was a biologist who had gone to Cambridge to study molecular structure. He had worked on bacteriophage reproduction and was keenly aware of the great possibilities that might be opened up by finding the molecular s-35 MOLECULAR CONFIGURATION OF NUCLEIC ACIDS structure of DNA. Crick was working on helical protein structure and was interested in what controlled protein synthesis. Pauling and Corey, by their discovery of the protein a-hclix, had shown that precise molecular model- building was a powerful analytical tool in its own right. The X-ray data from DNA were not so complete that a detailed picture of DNA structure could be derived without considerable aid from stereochemistry. It was clear that the X-ray studies of DNA needed to be complemented by pre- cise molecular model-building. In our laboratory we concentrated on am- plifying the X-ray data. In Cambridge, Watson and Crick built molecular models. The paradox of the regularity of the DNA molecule The sharpness of the X-ray diffraction patterns of DNA showed that DNA molecules were highly regular - so regular that DNA could crystallize. The form of the patterns gave clear indications that the molecule was helical, the polynucleotide chains in the molecular thread being regularly twisted. It was known, however, that the purines and pyrimidines of various dimensions were arranged in irregular sequence along the polynucleotide chains. How could such an irregular arrangement give a highly regular structure? This paradox pointed to the solution of the DNA structure problem and was resolved by the structural hypothesis of Watson and Crick. The Helical Structure of the DNA Molecule The key to DNA molecular structure was the discovery by Watson and Crick13 that, if the bases in DNA were joined in pairs by hydrogen-bonding, the overall dimensions of the pairs of adenine and thymine and of guanine and cytosine were identical. This meant that a DNA molecule containing these pairs could be highly regular in spite of the sequence of bases being irregular. Watson and Crick proposed that the DNA molecule consisted of two polynucleotide chains joined together by base-pairs. These pairs are shown in Fig. 2. The distance between the bonds joining the bases to the deoxyribosc groups is exactly (within the uncertainty of 0.1 A or so) the same for both base-pairs, and all those bonds make exactly (within the uncertainty of i° or so) the same angle with the line joining the Cx atoms of the deoxyribose (see Fig. 2). As a result, if two polynucleotide chains are s-36 I962 M. H. F.WILKIN S Fig. 2. Watson-Crick base-pairs (revised by S. Arnott). (Top): Guanine hydrogen- bonded to cytosine. (Bottom): Adenine hydrogen-bonded to thymine. The distances be- tween the ends of the Q-N3 and Q-N, bonds are 10.7 A in both pairs, and all these bonds make an angle of 520 with the Q-Cj line. joined by the base-pairs, the distance between the two chains is the same for both base-pairs and, because the angle between the bonds and the Q-Cj line is the same for all bases, the geometry of the deoxyribose and phosphate parts of the molecule can be exactly regular. Watson and Crick built a two-chain molecular model of this kind, the chains being helical and the main dimensions being as indicated by the X-ray data. In the model, one polynucleotide chain is twisted round the other and s-37 MOLECULAR (i>\l I (JURATION OF NUCLEIC ACIDS Fig. 3. (Left): Molecular model of the B configuration of DNA. The sizes of the atoms correspond to Van der Waals diameters. ( Right) : Diagram corresponding to the model. The two polynucleotide chains, joined by hydrogen-bonded bases, may be seen clearly. s-38 19^2 M. II. F. WILKIN S the sequence of atoms in one chain runs in opposite direction to that in the other. As a result, one chain is identical with the other if turned upside- down, and every nucleotide in the molecule has identical structure and envi- ronment. The only irregularities arc in the base sequences. The sequence along one chain can vary without restriction, but base-pairing requires that adenine in one chain be linked to thymine in the other, and similarly guanine to cytosine. The sequence in one chain is, therefore, determined by the se- quence in the other, and is said to be complementary to it. The structure of the DN A molecule in the B configuration is shown in Fig. 3. The bases are stacked on each other 3.4 A apart and their planes arc almost perpendicular to the helix axis. The flat sides of the bases cannot bind water molecules; as a result there is attraction between the bases when DNA is in an aqueous medium. This hydrophobic bonding, together with the base-pair hydrogen-bonding, stabilizes the structure. The Watson-Crick Hypothesis of DNA Replication, and Transfer of In- formation from one Polynucleotide Chain to Another It is essential for genetic material to be able to make exact copies of itself; otherwise growth would produce disorder, life could not originate, and favourable forms would not be perpetuated by natural selection. Base- pairing provides the means of self-replication (Watson and Crick14). It also appears to be the basis of information transfer during various stages in pro- tein synthesis. Genetic information is written in a four-letter code in the sequence of the four bases along a polynucleotide chain. This information may be transferred from one polynucleotide chain to another. A polynucleotide chain acts as a template on which nucleotides arc arranged to build a new chain. Provided that the two-chain molecule so formed is exactly regular, base-pairing en- sures that the sequence in the new chain is exactly complementary to that in the parent chain. If the two chains then separate, the new chain can act as a template, and a further chain is formed; this is identical with the original chain. Most DNA molecules consist of two chains; clearly the copying process can be used to replicate such a molecule. It can also be used to transfer information from a DNA chain to an RNA chain (as is believed to be the case in the formation of messenger RNA). Base-pairing also enables specific attachments to be made between part s-39 MOLECULAR CONFIGURATION OF NUCLEIC ACIDS of one polynucleotide chain and a complementary sequence in another. Such Specific interaction may be the means by which amino acids are attached to the requisite portions of a polynucleotide chain that has encoded in it the sequence of amino acids that specifies a protein. In this case the amino acid is attached to a transfer RNA molecule and part of the polynucleotide chain in this RNA pairs with the coding chain. Since the base-pairs were first described by Watson and Crick in 1953, many new data on purine and pyrimidinc dimensions and hydrogen-bond lengths have become available. The most recent refinement of the pairs (due to S. Arnott) is shown in Fig. 2. We now take the distance between Cj atoms as 10.7 A instead of the value used recently of 11.0 A, mainly because new data on N-H...N bonds show that this distance is 0.2 A shorter between ring nitrogen atoms than between atoms that are not in rings. The linearity of the hydrogen bonds in the base-pairs is excellent and the lengths of the bonds are the same as those found in crystals (these lengths vary by about 0.04 A). The remarkable precision of the base-pairs reflects the exactness of DNA replication. One wonders, however, why the precision is so great, for the energy required to distort the base-pairs so that their perfection is appre- ciably less, is probably no greater than one quantum of thermal energy. The explanation may be that replication is a co-operative phenomenon involving many base-pairs. In any case, it must be emphasized that the specificity of the base-pairing depends on the bonds joining the bases to the deoxyribose groups being correctly placed in relation to each other. This placing is prob- ably determined by the DNA polymerizing enzyme. Whatever the me- chanics of the process are, the exact equivalence of geometry and envi- ronment of every nucleotide in the double-helix should be conducive to precise replication. Mistakes in the copying process will be produced if there are tautomeric shifts of protons involved in the hydrogen-bonding or chem- ical alterations of the bases. These mistakes can correspond to mutations. The Universal Nature and Constancy of the Helical Structure of DNA After our preliminary X-ray studies had been made, my friend Leonard Hamilton sent me human DNA he and Ralph Barclay had isolated from human leucocytes of a patient with chronic myeloid leukaemia. He was studying nucleic acid metabolism in man in relation to cancer and had pre- s-40 1962 M. H. F.WILKIN S //"" x\ \ X.. v Fig. 4. X-ray diffraction pattern of cephalopod sperm. The DNA molecules in the sperm heads have their axes vertical. The 3.4 A internucleotide spacing corresponds to the strong diffraction at the top and bottom of the pattern. The sharp reflections m the central part of the pattern show that the molecules are in crystalline array. pared the DNA in order to compare the DNA of normal and leukaemic leucocytes. The DNA gave a very well-defined X-ray pattern. Thus began a collaboration that has lasted over many years and in which we have used Hamilton's DNA, in the form of many salts, to establish the correctness of the double-helix structure. Hamilton prepared DNA from a very wide range of species and diverse tissues. Thus it has been shown that the DNA s-41 MOLECULAR CONFIGURATION OF NUCLEIC ACIDS double-helix is present in inert genetic material in sperm and bacteriophage, and in cells slowly or rapidly dividing or secreting protein (Hamilton ctal.16). No difference of structure has been found between DNA from normal and ":*4tiftl 4~ ¥ Fig. 5. X-ray diffraction photograph of DNA fibres (B configuration) at high humid- ity. The fibres are vertical. The 3.4 A reflection is at the top and bottom. The angle in the pronounced X shape, made by the reflections in the central region, corresponds to the constant angle of ascent of the polynucleotide chains in the helical molecule. (Photograph with H. R. Wilson; DNA by L. D. Hamilton.) s-42 1962 M. II. F. WI LKI N S from cancerous tissues, or in calf thymus DNA separated into fractions of different base composition by my colleague Geoffrey Brown. We also made a study, in collaboration with Harriet Ephrussi-Taylor, of active transforming principle from pneumococci, and observed the same DNA structure. The only exception to double-helical DNA so far found is in some very small bacteriophages where the DNA is single-stranded. We have found, however, that DNA, with an unusually high content of adenine, or with glucose attached to hydroxymethylcytosine, crystallized differently. DNA Structure is Not an Artefact It did not seem enough to study X-ray diffraction from DNA alone. Ob- viously one should try to look at genetic material in intact cells. It was pos- sible that the structure of the isolated DNA might be different from that /'// vivo, where DNA was in most cases combined with protein. The optical studies indicated that there was marked molecular order in sperm heads and that they might therefore be good objects for X-ray study, whereas chromo- somes in most types of cells were complicated objects with little sign of ordered structure. Randall had been interested in this matter for some years and had started Gosling studying ram sperm. It seemed that the rod-shaped cephalopod sperm, found by Schmidt to be highly anisotropic optically, would be excellent for X-ray investigation. Rhine17, while making a study of liquid crystals from many branches of Nature, had already taken diffrac- tion photographs of such sperm; but presumably his technique was inad- equate, for he came to the mistaken conclusion that the nucleoprotein was liquid-crystalline. Our X-ray photographs (Wilkins and Randall18) showed clearly that the material in the sperm heads had 3 -dimensional order, i.e. it was crystalline and not liquid-crystalline. The diffraction pattern (Fig. 4) bore a close resemblance to that of DNA (Fig. 5), thus showing that the structure in fibres of purified DNA was basically not an artefact. Working at the Stazione Zoologica in Naples, I found it possible to orient the sperm heads in fibres. Intact wet spermatophore, being bundles of naturally ori- ented sperm, gave good diffraction patterns. DNA-like patterns were also obtained from T2 bacteriophage given me by Watson. s-43 MOLECULAR CONflGURA IION OF NUCLEIC ACIDS 1.7 A 3.4 A * :-> i » • vx^ >-v / Fig. 6. X-ray pattern of microcrystallinc fibres of DNA. The general intensity dis- tribution is similar to that in Fig. 4 but the diffraction is split into sharp reflections owing to the regular arrangement of the molecules in the crystals. Sharp reflections extend to spacings as small as 1.7 A. (Photograph with N. Chard; DNA by L. D. Hamilton.) s-44 19 62 M. H. F. WI LK I N S <-£ < Q O > o ►J << < rs J2 _^ u 2 ^j — ?N 00 ■ -3 ^C: to U 0 c< = ci s rt H ^ £ 2 * I 1 a u s. o .a o a -|N-|".-|". ^ b -|» -|vO , -|oo 0 g 0 0 -,i-|^ 0 0 0 -IN "I* 1 O -|N -|00 | (J O -|0 | N|"> — \m O O -1-1- -|N «|00 j O -IN 0 -|N O -IN O ~K> O -|"VN|f. O -|N O -IN O „|^N|f5 Z * * t- o O J2 h-1 pq >. nu a ■ ; O 0^ 3xO d 0 .3 VI On .3 ,g - d rt 3 rt S rt ' ) L> V5 ^ vi >- on >- rt o rt j Z w p: Z pq o m ^ o U s-45 MOLECULAR CONFIGURATION OF NUCLEIC ACIDS The X-ray Diffraction Patterns of DNA and the Various Configurations of the Molecule X-ray diffraction analysis is the only technique that can give very detailed information about the configuration of the DNA molecule. Optical tech- niques, though valuable as being complementary to X-ray analysis, provide much more limited information - mainly about orientation of bonds and groups. X-ray data contributed to the deriving of the structure of DNA at two stages. First, in providing information that helped in building the Watson-Crick model ; and second, in showing that the Watson-Crick pro- posal was correct in its essentials, which involved readjusting and refining the model. The X-ray studies (e.g. Langridge ct al.19, Wilkins20) show that DNA mol- ecules are remarkable in that they adopt a large number of different con- formations, most of which can exist in several crystal forms. The main factors determining the molecular conformation and crystal form are the water and salt contents of the fibre and the cation used to neutralize the phosphate groups (see Table i). I shall describe briefly the three main configurations of DNA. In all cases the diffraction data are satisfactorily accounted for in terms of the same basic Watson-Crick structure. This is a much more convincing demonstration of the correctness of the structure than if one configuration alone were studied. The basic procedure is to adjust the molecular model until the calculated intensities of diffraction from the model correspond to those observed (Lang- ridge ct al.19). As with most X-ray data, only the intensities, and not the phases, of the diffracted beams from DNA arc available. Therefore the structure cannot be derived directly. If the resolution of X-ray data is sufficient to separate most of the atoms in a structure, the structure may be derived with no stereochemical assumption except that the structure is assumed to consist of atoms of known average size. With DNA, however, most of the atoms cannot be separately located by the X-rays alone (see Fig. 7). Therefore, more extensive stereochemical assumptions are made: these take the form of molecular model-building. There arc no alternatives to most of these assumptions but where there might be an alternative, e.g. in the arrangement of hydrogen bonds in a base-pair, the X-ray data should be used to establish the correctness of the assumption. In other words, it is necessary to establish that the structure proposed is unique. Most of our work in recent years has s-46 1962 M. H. F.WILKIN S Contour interval 2C|A zero contour dashed Fig. 7. Fourier synthesis map (by S. Arnott) showing the distribution of electron density in the plane of a base-pair in the B configuration of DNA. The distribution corresponds to an average base-pair. The shape of the base-pair appears in the map, but individual atoms in a base-pair are not resolved. (The Fourier synthesis is being revised and the map is subject to improvement.) been of this nature. To be reasonably certain that the DNA structure was correct, X-ray data, as extensive as possible, had to be collected. The B configuration Fig. 5 shows a diffraction pattern of a fibre of DNA at high humidity when the molecules are separated by water and, to a large extent, behave independ- ently of each other. We have not made intensive study of DNA under these conditions. The patterns could be improved, but they are reasonably well- defined, and the sharpness of many of their features shows that the molecules have a regular structure. The configuration is known as B (see also Fig. 3); it is observed in vivo, and there is evidence that it exists when DNA is 111 solution in water. There are 10 nucleotide pairs per helix turn. There is no obvious structural reason why this number should be integral; if it is exactly so, the significance of this is not yet apparent. s-47 MOLECULAR CONFIGURATION OF NUCLEIC ACIDS Fig. 8. Molecular model of DNA in the A configuration. The base-pairs may be seen inclined 20° to the horizontal. When DNA crystallizes, the process of crystallization imposes restraints on the molecule and can give it extra regularity. Also, the periodic arrange- ment of the molecules in the microcrystals in the fibre causes the diffraction pattern to be split into sharp reflections corresponding to the various crystal planes (Fig. 6). Careful measurement of the positions of the reflections and deduction of the crystal lattice enables the directions of the reflections to be identified in three dimensions. Diffraction patterns from most fibrous sub- stances resemble Fig. 5 in that the diffraction data arc 2-dimcnsional. In contrast, the crystalline fibres of DNA give fairly complete 3 -dimensional data. These data give information about the appearance of the molecule when viewed from all angles, and are comparable with those from single s-48 I962 M. H. F.WILKIN S crystals. Techniques such as 3 -dimensional Fourier synthesis (see Fig. 7) can be used and the structure determination made reasonably reliable. The A configuration In this conformation, the molecule has 11 nucleotide pairs per helix turn; the helix pitch is 28 A. The relative positions and orientations of the base, and of the dcoxyribose and phosphate parts of the nucleotides differ considerably from those in the B form; in particular the base-pairs are tilted 20° from perpendicular to the helix axis (Fig. 8). The A form of DNA (Fig. 1) was the first crystalline form to be ob- served. Although it has not been observed in vivo, it is of special interest because helical RNA adopts a very similar configuration. A full account of A DNA will shortly be available. A good photograph of the A pattern is shown in Fig. 9. The C configuration This form may be regarded as an artefact formed by partial drying. The helix is non-integral, with about 9] nucleotide pairs per turn. The helices pack together to form a semi-crystalline structure; there is no special relation between the position of one nucleotide in a molecule and that in another. The conformation of an individual nucleotide is very similar to that in the B form. The differences between the B and C diffraction patterns are accounted for by the different position of the nucleotides in the helix. Comparison of the forms provides further confirmation of the correctness of the structures. In a way, the problem is like trying to deduce the structure of a folding chair by observing its shadow: if the conformation of the chair is altered slightly, its structure becomes more evident. The Helical Structure of RNA Molecules In contrast to DNA, RNA gave poor diffraction patterns, in spite of much effort by various workers including ourselves. There were many indications that RNA contained helical regions, e.g. optical properties of RNA solutions strongly suggested (e.g. Doty21) that parts of RNA molecules resembled DNA in that the bases were stacked on each other and the structure was helical; and X-ray studies of synthetic polyribonucleotides suggested that s-49 MOLECULAR C O N \ I G U H A I I O N OF NUCLEIC ACIDS / / * **\ m 1 " *r 0 " / # s^ y/ Fig. 9. X-ray diffraction pattern of microcrystalline fibres of DNA in the A con- figuration. (Photograph with H. R. Wilson; DNA by L. D. Hamilton.) RNA resembled DNA (Rich22). The diffraction patterns of RNA (Rich and Watson23) bore a general resemblance to those of DNA, but the nature of pattern could not be clearly distinguished because of disorientation and diffuseness. An important difficulty was that there appeared to be strong meridional reflections at 3.3 A and 4 A. It was not possible to interpret these in terms of one helical structure. s-50 1962 M. H. F. WILKINS -. V Fig. 10. Comparison of the X-ray diffraction patterns of fibres of DNA in the A configuration (left) and transfer RNA {right). The general distribution of intensity is very similar in both patterns, but the positions of the sharp crystalline reflections differ because the molecular packing in the crystals is different in the two cases. (Photograph with W. Fuller and M. Spencer; RNA by G. L. Brown.) In early work, many RNA preparations were very heterogeneous. We thought that the much more homogeneous plant virus RNA might give better patterns, but this was not so. However, when preparations of ribo- somal RNA and « soluble » RNA became available, we felt the prospects of structure analysis were improved. We decided to concentrate on « soluble » RNA largely because Geoffrey Brown in our laboratory was preparing large quantities of a highly purified transfer RNA component of soluble RNA for his physical and chemical studies, and because he was fractionating it into various transfer RNA's specific for incorporation of particular amino acids into proteins. This RNA was attractive for other reasons : the molecule was s-51 MO LECULAB CON FI G U RATION OF NUCLEIC ACIDS unusually small for a nucleic acid, there were indications that it might have a regular structure, its biochemical role was important, and m many ways its functioning was understood. We found it very difficult to orient transfer RNA in fibres. I Iowever, by carefully stretching RNA gels in a dry atmosphere under a dissecting mi- croscope, 1 found that fibres with birefringence as high as that of DNA could be made. But these fibres gave patterns no better than those obtained with other types of RNA, and the molecules disoriented when the water content of the fibres was raised. Watson, Fuller, Michael Spencer, and my- self worked for many months trying to make better specimens for X-ray study. We made little progress until Spencer found a specimen that gave some faint but sharp diffraction rings in addition to the usual diffuse RNA pattern. This specimen consisted of RNA gel that had been sealed for X-ray study in a small cell, and he found that it had dried slowly owing to a leak. The diffraction rings were so sharp that wc were almost certain that they were spurious diffraction due to crystalline impurity - this being common in X-rav studies of biochemical preparations. A specimen of RNA had given very similar rings due to DNA impurity. Wc were therefore not very hopeful about the rings. However, after several weeks Spencer eliminated all other possibilities: it seemed clear that the rings were due to RNA itself. By controlled slow drying, he produced stronger rings; and, with the refined devices wc had developed for stretching RNA and with gels slowly con- centrated by Brown, Fuller oriented the RNA without destroying its crys- talhnity. These fibres gave clearly defined diffraction patterns, and the ori- entation did not disappear when the fibres were hydratcd. It appeared that the methods I had been using earlier, of stretching the fibres as much as possible, destroyed the crystallinity. If instead, the material was first allowed to crystallize slowly, stretching oriented the microcrystals and the RNA molecules in them. Single molecules were too small to be oriented well unless aggregated by crystallization. It was rather unexpected that, of all the different types of RNA we had tried, transfer RNA which had the lowest molecular weight, oriented best. The diffraction patterns of transfer RNA were clearly defined and well- oriented (Spencer, Fuller, Wilkins, and Brown24). These improvements re- vealed a striking resemblance between the patterns of RNA and A DNA (Fig. 10). The difficulty of the two reflections at 3.3 A and 4 A was resolved ( Fig. 1 1 ) : in the RNA pattern the positions of reflections on three layer-lines differed from those in DNA; as a result, when the patterns were poorly s-52 I962 M. H. F. WILKIN S Fig. 11. Diffraction pattern of transfer RNA showing resolution of diffraction, in the regions of 3.3 A and 4 A, into three layer-lines indicated by the arrows and corres- ponding to the A DNA pattern. (Photograph with W. Fuller and M. Spencer; RNA by G. L. Brown.) oriented, the three reflections overlapped and gave the impression of two. There was no doubt that the RNA had a regular helical structure almost identical with that of A DNA. The differences between the RNA and DNA patterns could be accounted for in terms of small differences between the two structures. An important consequence of the close resemblance of the RNA structure to that of DNA is that the RNA must contain base sequences that are largely or entirely complementary. The number of nucleotides in the molecule is about 80. The simplest structure compatible with the X-ray results consists of a single polynucleotide chain folded back on itself, one half of the chain being joined to the other by base-pairing. This structure is shown in Fig. 12. While we are certain the helical structure is correct, it must be emphasized that we do not know whether the two ends of the chain are at the end of the molecule. The chain might be folded at both ends of the molecule with the ends of the chain somewhere along the helix. It is known that the amino acid attaches to the end of the chain terminated by the base sequence cyto- sine-cytosine-adenine. s-53 MOLECULAR CON l: I (JURATION OF NUCLEIC ACIDS Fig. 12. Molecular model and diagram of a transfer RNA molecule. s-54 1962 M.H. I'.WILKINS Relation of the Molecular Structure oj RNA to Function Molecular model-building shows that the number of nucleotides forming the fold at the end of a transfer RNA molecule must be three or more. In our model, the fold consists of three nucleotides, each with an unpaired base. It might be that this base-triplet is the part of the molecule that attaches to the requisite part of the coding RNA polynucleotide chain that determines the sequence of amino acids in the polypeptide chain of a protein. It is believed that a base-triplet in the coding RNA corresponds to each amino acid. The triplet in the transfer RNA could attach itself specifically to the coding triplet by hydrogen-bonding and formation of base-pairs. It must be emphasized, however, that these ideas are speculative. We suppose that part of the transfer RNA molecule interacts specifically with the enzyme that is involved in attaching the amino acid to the RNA; but we do not know how this takes place. Similarly, we know little of the way in which the enzyme involved in DNA replication interacts with DNA, or of other aspects of the mechanics of DNA replication. The presence of complementary base sequences in the transfer RNA molecule, suggests that it might be self-replicating like DNA; but there is at present little evidence to support this idea. The diffraction patterns of virus and ribosome RNA show that these molecules also contain helical regions; the function of these are uncertain too. In the case of DNA, the discovery of its molecular structure led imme- diately to the replication hypothesis. This was due to the simplicity of the structure of DNA. It seems that molecular structure and function are in most cases less directly related. Derivation of the helical configuration of RNA molecules is a step towards interpreting RNA function; but more complete structural information, e.g. determination of base sequences, and more knowledge about how the various kinds of RNA interact in the ribo- some, will probably be required before an adequate picture of RNA func- tion emerges. The Possibility of Determining the Base Sequence of Transfer RNA by X-ray Diffraction Analysis Since the biological specificity of nucleic acids appears to be entirely deter- mined by their base sequences in them, determination of these sequences is probably the most fundamental problem in nucleic acid research today. The s-55 MOLECULAR CONFIGURATION OF NUCLEIC ACIDS Fig. 13. Diffraction pattern of unoricntcd transfer RNA, showing diffraction rings with spots corresponding to reflections from single crystals of RNA. The arrows point to reflections from planes ~ 6 A apart. number of bases in a DNA molecule is too large for determination of base sequence by X-ray diffraction to be feasible. However, in transfer RNA the number of bases is not too large. The possibility of complete structure anal- ysis of transfer RNA by means of X-rays is indicated by two observations. First, we have observed (Fig. 13), in X-ray patterns of transfer RNA, sep- arate spots each corresponding to a single crystal of RNA. We estimated their size to be about 10^ and have confirmed this estimate by observing, in the polarizing microscope, bircfringent regions that probably arc the crys- tals. It should not be too difficult to grow crystals several times larger, which is large enough for single-crystal X-ray analysis. The second encouraging observation is that the X-ray data from DNA have restricted resolution almost entirely on account of disorientation of the microcrystals in DNA fibres. The DNA intensity data indicate that the temperature factor (B = 4 A) is the same for DNA as for simple compounds. It thus appears that DNA crystals have fairly perfect crystallinity and that, if s-56 I962 M.H. F. WILKINS single crystals of DN A could be obtained, the intensity data would be ad- equate for precise determination of all atomic positions in DNA (apart from the non-periodic base sequence). We are investigating the possibility of obtaining single crystals of DNA, but the more exciting problem is to obtain single crystals of transfer RNA with crystalline perfection equal to that of DNA, and thereby analyse base sequence. At present, the RNA crystals are much less perfect than those of DNA. However, most of our experiments have been made with RNA that is a mixture of RNA's specific for different amino acids. We have seldom used RNA that is very largely specific for one amino acid only. We hope that good preparations of such RNA may be obtained consisting of one type of molecule only. We might expect such RNA to form crystals as perfect as those of DNA. If so, there should be no obstacle to the direct analysis of the whole structure of the molecule, including the sequence of the bases and the fold at the end of the helix. We may be over-optimistic, but the recent and somewhat unexpected successes of X-ray diffraction analysis in the nu- cleic acid and protein fields, are cause for optimism. Acknowledgements During the past twelve years, while studying molecular structure of nucleic acids, I have had so much help from so many people that all could not be acknowledged properly here. I must, however, thank the following : Sir John Randall, for his long-standing help and encouragement, and for his vision and energy in creating and directing a unique laboratory; all my co-workers at various times over the past twelve years; first, Raymond Gosling, Alex Stokes, Bill Seeds, and Herbert Wilson, then Bob Langridge, Clive Hooper, Max Feughelman, Don Marvin, and Geoffrey Zubay ; and at present, Michael Spencer, Watson Fuller, and Struther Arnott, who with much ability, skill and persistence (often through the night) car- ried out the X-ray, molecular model-building, and computing studies; my late colleague Rosalind Franklin who, with great ability and expe- rience of X-ray diffraction, so much helped the initial investigations on DNA; Leonard Hamilton for his constant encouragement and friendly cooper- ation, and for supplying us with high-quality DNA isolated in many forms and from many sources ; Geoffrey Brown for giving me moral and intellec- s-57 LECU LAI CONFIGURATION OF NUCLEIC ACIDS tual support throughout : preparing UNA tor X-ray stu Harriet Ephrussi-Tavlor tor the privilege of collaborating with her in stud- stalhzarion ot transforming principle; the laboratory technicians, mechanics and photographers, including P.J. Cooper. N. Chard. J. Hayward, . F.Colh. lx>r, and R.Lerner, for having played a valuable pan rious su: I also wish to thank: the Medical Research Council for their far-sighted and consistent support or work j College for being our base: I.B.M. United Kingdom Limited and I.B. .d Trade Corporation and the London University :or help with computing: The Rockefeller Foundation and The British Empire Cancer Campaign for financial support: the Sloan- '. rk, and the Stazione Zoologica, Naples, for use of tacili: _ nerally, I thank : Francis Crick and Jim Watson for stimulating discussi nan Sim- mons for having refined techniques ot isolating DNA and thereby helping a great many workers including ourselves : many other workers for supplying us with DNA and RNA; and especially, Erwin ChargafFfor laying foun- dations tor nucleic acid structural studies by his analytical work and his dis- i the equality of base contents in DNA, and for generously helping us newcomers in the field of nucleic acids. i. T. Caspers> 94i) 33- Brachet. . Liege, 53 [942 T. Avery, C. M. MacLeod, and ML McCarry. J. Exp. Med.. 7 4. R. Signer and H Schwander, Heh _ 53. A. D. Hershey and M. Chase,/. Gen. '. 39- 6. E. ChargafE, Experiemia, t 201. ' R. Signer, T. Caspersson, and E. Hammars: Schmidt, DU Doppelbrechung von Karyoplasma, Zytoph& '. raplastna, Borntraeger, Berlin, 1937 9. H. O. E. Pactri, Z. Zellfr _ -_ 10. W. T. Astbury, Sytr.p. Soc. Exptl. Biol, I. I id, Cambridge Univ. Press, 1947, p. 66. 11. J. M. Gulland and D. O. Jordan, Symp . : Acid, Cambridge - midland, Cold. - 17)95. D. Watson and F. EL C re, 171 (1951 s-58 I9<52 M. H. F. W I L K I N S 14. J. D. '?•■ i F. H. C. Cr: 15. K. Hoogi- _ 22 16. L. D. Hamilton, R. K. 1 ins, G. L D. .- H Eg hi :- ? 1 ::r6. ■ '. H. F. V. - J. T. Ri: 19. R. Langridge, H. P - ton, J. -1 1 - 19. _ ".. H. F. \X 21. P. Dot] 21 (] :_ 22. A. Rich, in .- 17- : 3 - .'...: md J. D. "T 24 M. Spam S.L.Br: ic:_ 5-5 SUPPLEMENT V THE BIOLOGIC SYNTHESIS OF DEOXYRIBONUCLEIC ACID by Arthur Kornberg. Nobel Lecture, December n, 1959. The knowledge drawn in recent years from studies of bacterial transfor- mation (1) and viral infection of bacterial cells (2, 3) combined with other evi- dence (3), has just about convinced most of us that deoxyribonucleic acid (DNA) is the genetic substance. We shall assume then that it is DNA which not only directs the synthesis of the proteins and the development of the cell but that it must also be the substance which is copied so as to provide for a similar development of the progeny of that cell for many generations. DNA, like a tape recording, carries a message in which there are specific instructions for a job to be done. Also like a tape recording, exact copies can be made from it so that this information can be used again and elsewhere in time and space. Are these two functions, the expression of the code (protein synthesis) and the copying of the code (preservation of the race) closely integrated or are they separable? What we have learned from our studies over the past five years and what I shall present is that the replication of DNA can be examined and at least partially understood at the enzymatic level even though the secret of how DNA directs protein synthesis is still locked in the cell. DNA structure. First I should like to review very briefly some aspects of DNA structure which are essential for this discussion. Analysis of the composition of samples of DNA from a great variety of sources and by many investigators (4) revealed the remarkable fact that the purine content always equals the pyrimidine content. Among the purines, the adenine content may differ considerably from the guanine, and among the pyrimidines, the thymine from the cytosine. s-60 ADENINE .O / THYMINE GUANINE U,H " CYTOSWE Fig. i. Hydrogen Bonding of Bases. i However, there is an equivalence of the bases with an amino group in the 6-position of the ring, to the bases with a keto group in the 6-position. These facts were interpreted by Watson and Crick (5) in their masterful hypothesis on the structure of DNA. As shown in Fig. i, they proposed in connection with their double-stranded model for DNA, to be discussed presently, that the 6-amino group of adenine is linked by hydrogen bonds to the 6-keto group of thymine and in a like manner guanine is hydrogen-bonded to cytosine, thus accounting for the equivalence of the purines to the pyrimidines. On the basis of these considerations and the results of X-ray crystallographic measurements by Wilkins and associates (6), Watson and Crick proposed a structure for DNA in which two long strands are wound about each other in a helical manner. Fig. 2 is diagrammatic representation of a fragment of a DNA chain about ten nucleotide units long. According to physical measurements, DNA chains are on the average 10 000 units long. We see here the deoxypentose rings linked by phosphate residues to form the backbone of the chain; the purine and pyrimidine rings are the planar structures emerging at right angles from the main axis of the chain. Fig. 3 is a more detailed molecular model (7) and gives a better idea of the packing of the atoms in the structure. The purine and pyrimidine bases of one chain are bonded to the pyrimidine and purine bases of the complementary chain by the hydrogen bonds described in Fig. 1. The X-ray measurements have indicated that the space between the opposing chains in the model agrees with the calculated value for the hydrogen-bond linkage of a purine to a pyrimidine; it is too small for two purines and too large for two pyrimidines. Most rewarding from the biological point of view, the structure provides a useful model to explain how cellular replication of DNA may come about. For, if you imagine that these two chains separate and that a new chain is formed complementary to each of them, the result will be two pairs of strands, each pair identical to the original parent duplex and identical to each other. s-61 Fig. 2. Double Helical Structure of DXA (Watson and Crick Mo- del). Enzymatic approach to the problem of DNA replication. Although we have in the Watson and Crick proposal a mechanical model of replication, we may at this point pose the question: "What is the chemical mechanism by which this super molecule is built up in the cell?" Some sixty years ago the alcoholic fermentation of sugar by a yeast cell was a "vital" process inseparable from the living cell, but through the Buchner discovery of fermentation in extracts and the march of enzymology during the first half of this century we understand fermentation by yeast as a, now familiar, sequence of integrated chemical reactions. Five years ago the synthesis of DNA was also regarded as a "vital" process. Some people considered it useful for biochemists to examine the combustion chambers of the cell, but tampering with the very genetic apparatus itself would surely produce nothing but dis- order. These gloomy predictions were not justified then nor are similar pessi- mistic attitudes justified now with regard to the problems of cellular structure s-62 O Hyd rogen ^B Oxygen ^^ Carbon in ^B phosphate-ester ^^^ chain ( ! Guanine Cytosine ^ ) Adenine ( ) Thymine Fig. 3. Molecular Model of DNA ^^B Phosphorus (After M. Feughelman, et al. {?))■ and specialized function which face us. High adventures in enzymology lie ahead and many of the explorers will come from the training fields of carbo- hydrate, fat, amino acid and nucleic acid enzymology. I feel now, as we did then, that for an effective approach to the problem of nucleic acid biosynthesis it was essential to understand the biosynthesis of the simple nucleotides and the coenzymes and to have these concepts and methodology well in hand. It was from these studies that we developed the conviction that an activated nucleoside 5 '-phosphate is the basic biosynthetic building block of the nucleic acids (8). You will recall that the main pathways of purine and pyrimidine biosynthesis all lead to the nucleoside 5 '-phosphate (8); they do not, except as salvage mechanisms, usually include the free bases or nucleosides. While the 2' and 3' isomers of the nucleotides are known, they probably arise mainly from certain types of enzymatic degradation of the nucleic acids. You will also recall from the biosynthesis of coenzymes (9), s-63 o o o II II II Adenosine - O - P t OP - O -- P - O A o" I Nucleoside - O - P - O II O Fig. 4. Nucleophilic Attack of a Nucleoside Monophosphate on ATP. the simplest of the nucleotide condensation products, that it is ATP which condenses with nicotinamide mononucleotide to form diphosphopyridine nu- cleotide, with riboflavin phosphate to form FAD, with pantetheine phosphate to form the precursor of coenzyme A and so forth. This pattern has been amplified by the discovery of identical mechanisms for the activation of fatty acids and amino acids and it has been demonstrated further that uridine, cytidine and guanosine coenzymes are likewise formed from the respective triphosphates of these nucleosides. This mechanism (Fig. 4), in which a nucleophilic attack (10) on the pyro- phosphate-activated adenyl group by a nucleoside monophosphate leads to the formation of a coenzyme, was adopted as a working hypothesis for studying the synthesis of a DNA chain. As illustrated in Fig. 5, it was postulated that the basic building block is a deoxynucleoside 5 '-triphosphate which is attacked by the 3'-hydroxyl group at the growing end of a polydeoxynucleotide chain; inorganic pyrophosphate is eliminated and the chain is lengthened by one unit. The results of our studies on DNA synthesis, as will be mentioned presently, are in keeping with this type of reaction. Properties of the DNA-synthesizing enzyme. First let us consider the enzyme and comment on its discovery (8, n, 12). Mixing the triphosphates of the four deoxynucleosides which commonly occur in DNA with an extract of thymus or bone-marrow or of Escherichia coli s-64 0 o-P-o''' h2c or O-P-0 h2c 6- HO H i - Poly- nucleotide Vi 2 2 0-PtO-P-O-P-O' H„C<>' o- X -XTP o HO Fig.' 5. Postulated Mechanism for Extending a DNA Chain. would not be expected to lead to the net synthesis of DNA. Instead, as might be expected, the destruction of DNA by the extracts of such cells and tissues was by far the predominant process and one had to resort to the use of more subtle devices for detection of such a biosynthetic reaction. We used a C14- labeled substrate of high specific radioactivity and incubated it with ATP and extracts of Escherichia coli, an organism which reproduces itself every 20 minutes. The first positive results represented the conversion of only a very small fraction of the acid-soluble substrate into an acid-insoluble fraction (50 or so counts out of a million added). While this represented only a few /ujumoles of reaction, it was something. Through this tiny crack we tried to drive a wedge and the hammer was enzyme purification (13). This has been and still is a major preoccupation. Our best preparations are several thousand- fold enriched with respect to protein over the crude extracts, but there are still contaminating quantities of one or more of the many varieties of nuclease s-65 n TPPP TP n dGPPP dGP + DNA^=^DNA - n dAPPP dAP n dCPPP dCP + 4(n)PP Fig. 6. Equation for Enzymatic Synthesis of DNA. and diesterase present in the coli cell. The occurrence of what appears to be a similar DNA synthesizing system in animal cells as well as in other bacterial species has been observed (14). We must wait for purification of the enzymes from these sources in order to make valid comparisons with the coli system. The requirements for net synthesis of DNA with the purified coli enzyme (15) are shown in the equation in Fig. 6. All four of the deoxynucleotides which form the adenine-thymine and guanine-cytosine couples must be present. The substrates must be the tri- and not the diphosphates and only the deoxy sugar compounds are active. DNA which must be present may be obtained from animal, plant, bacterial or viral sources and the best indications are that all these DNA samples serve equally well in DNA synthesis provided their molecular weight is high. The product, which we will discuss in further detail, accumulates until one of the substrates is exhausted and may be 20 or more times greater in amount than the DNA added and thus is composed to the extent of 95 % or more of the substrates added to the reaction mixture. Inorganic pyrophosphate is released in quantities equimolar to the deoxy- nucleotides converted to DNA. Should one of these substrates be omitted, the extent of reaction is di- minished by a factor of greater than io4 and special methods are now required for its detection. It turns out that when one of the deoxynucleotide substrates is lacking, an extremely small but yet significant quantity of nucleotide is linked to the DNA primer. We have described this so-called "limited reac- tion" (16), and have shown that under these circumstances a few deoxy- nucleotides are added to the nucleoside ends of some of the DNA chains but s-66 P-P: (P v / — A —-" T — \ } . \s / I I n/ Y^ 7~~ c EEzzE g — ^ ) \i / I I I \s — x y\ I I I sT Y ) — T ZZZ" A — ( 5 X — ( ! I / 'S Y — ( ! Fig. 7. Mechanism for Enzymatic DNA Replication. that further synthesis is blocked for lack of the missing nucleotide. Current studies suggest to us that this limited reaction represents the repair of the shorter strand of a double helix in which the strands are of unequal length, and that the reaction is governed by the hydrogen bonding of adenine to thymine and of guanine to cytosine. When all four triphosphates are present, but when DNA is omitted, no reaction at all takes place. What is the basis for this requirement? Does the DNA function as a primer in the manner of glycogen or does it function as a template in directing the synthesis of exact copies of itself? We have good reason to believe that it is the latter and as the central and restricted theme of this lecture I would like to emphasize that it is the capacity for base pairing by hydrogen bonding between the preexisting DNA and the nucleotides added as substrates that accounts for the requirement for DNA. The enzyme we are studying is thus unique in present experience in taking directions from a template — it adds the particular purine or pyrimidine sub- strate which will form a hydrogen-bonded pair with a base on the template (Fig. 7). There are five major lines of evidence that I would like to present to support this thesis. s-67 Physical properties of enzymatically synthesized DNA. The first line of evidence is derived from studies of the physical nature of the DNA produced by the enzyme. It may be mentioned again that in these descriptions as in those of the chemical nature of DNA, to be discussed shortly, 90 — 95 % of the DNA sample comes from the substrates used in the reaction. From collaborative studies with Dr. Howard K. Schachman, to whom we are greatly indebted, it can be said that the enzymatic product is indistinguish- able from high-molecular weight, double-stranded DNA isolated from nature (17). It has sedimentation coefficients in the neighbourhood of 25, reduced viscosities of 40 deciliters per gram and, on the basis of these measurements, it is believed to be a long, stiff rod with a molecular weight of about 6 million. Upon heating the DNA, the rod collapses an'd the molecule becomes a compact, randomly coiled structure; it may be inferred that the hydrogen bonds holding the strands together have melted and this is borne out by characteristic changes in the viscometric and optical properties of the molecule. Similar results are found upon cleavage of the molecule by pancreatic deoxyribo- nuclease. In all these respects the enzymatically synthesized DNA is indis- tinguishable from the material isolated from nature, and may thus be presumed to have a hydrogen-bonded structure similar to that possessed by natural DNA. Would one imagine that the collapsed jumbled strands of heated DNA would serve as a primer for DNA synthesis? Very likely one would think not. Guided by intuition derived from everyday experience with a jumbled strand of twine one might regard this as a hopeless template for replication. It turns out that the collapsed DNA is an excellent primer and the nonviscous, ran- domly coiled, single-stranded DNA leads to the synthesis of highly viscous, double-stranded DNA (18). Sinsheimer has isolated from the tiny 0X 174 virus a DNA which appears to be single-stranded (19). Like heated DNA it has proved to be an excellent primer (18) and a favorable material in current studies (20) for demonstrating in density gradient sedimentations that it is progressively converted to a double-stranded condition during the course of enzymatic synthesis. While a detailed discussion of the physical aspects of replication is not feasible in this lecture, it should be mentioned that the DNA in the single- stranded condition is not only a suitable primer but is the only active form when the most purified enzyme preparations are used. With such coli prepa- rations, the native, double-stranded DNA is inert unless it is heated or pre- treated very slightly with deoxyribonuclease. Bollum has made similar ob- servations with the enzyme that he has uurified from calf thymus (21). s-68 Substitution of analogues in DNA synthesis. The second line of evidence is derived from studies of the activity of the substrates when substitutions are made in the purine and pyrimidine bases. From the many interesting reports on the incorporation of bromouracil (22), azaguanine (23) and other analogues into bacterial and viral DNA, it might be surmised that some latitude in the structure of the bases can be tolerated provided there is no interference with their hydrogen bondings. When experi- ments were carried out with deoxyuridine triphosphate or 5-bromodeoxy- uridine triphosphate, it was found that they supported DNA synthesis when used in place of thymidine triphosphate but not when substituted for the triphosphates of deoxyadenosine, deoxyguanosine or deoxycytidine. As already described (24), 5-methyl- and 5-bromocytosine specifically replaced cytosine; hypoxanthine substituted only for guanine; and, as just mentioned, uracil and 5-bromouracil specifically replaced thymine. These findings are best interpreted on the basis of hydrogen bonding of the adenine-thymine and guanine-cytosine type. Along these lines it is relevant to mention the existence of a naturally oc- curring "analogue" of cytosine, hydroxymethyl cytosine (HMC), which is found in place of cytosine in the DNA of the coli bacteriophages of the T-even series (25). In this case the DNA contains equivalent amounts of HMC and guanine and, as usual, equivalent amounts of adenine and thymine. Of addi- tional interest is the fact that the DNA's of T2, T4 and T6 contain glucose linked to the hydroxymethyl groups of the HMC in characteristic ratios (26, 27, 28) although it is clear that in T2 and T6 some of the HMC groups contain no glucose (27). These characteristics have posed two problems regarding the synthesis of these DNA's which might appear to be incompatible with the simple base-pairing hypothesis. First, what mechanism is there for preventing the inclusion of cytosine in a cell which under normal conditions has deoxy- cytidine triphosphate and incorporates it into its DNA? Secondly, how does one conceive of the origin of the constant ratios of glucose to HMC in DNA if the incorporation were to occur via glucosylated and non-glucosylated HMC nucleotides? Our recent experiments have shown that the polymerase reaction in the virus-infected cell is governed by the usual hydrogen-bonding restric- tions but with the auxiliary action of several new enzymes developed specifi- cally in response to infection with a given virus (29, 30). Among the new enzymes is one which splits deoxycytidine triphosphate and thus removes it from the sites of polymerase action (30). Another is a type of glucosylating s-69 DNA A T G C A + G T + C A + T G + C M. phle primer product 0.65 0.66 0.66 0.65 '•35 '•34 '•34 '•37 1. 01 0.99 0.49 0.48 E, coli primer product 1. 00 1.04 0.97 1. 00 0.98 0.97 1.05 0.98 0.98 1. 01 0.97 1.02 Calf thymus primer product I.14 1. 12 1.05 1.08 0.90 0.85 0.85 0.85 1.05 1.02 1.25 1.29 Bacteriophage Tz primer product '•31 i-33 1.32 1.29 0.67 0.69 0.70 0.70 0.98 1.02 1.92 1.90 A-T Copolymer 1.99 '•93 4° Fig. 8. Chemical Composition of Enzymatically Synthesized DNA with Different Primers. enzyme which transfers glucose from uridine diphosphate glucose directly and specifically to certain HMC residues in the DNA (30). Chemical composition of enzymatically synthesized DNA. The third line of evidence is supplied by an analysis of the purine and pyrimidine base composition of the enzymatically synthesized DNA. We may ask two questions. First, will the product have the equivalence of adenine to thymine and of guanine to cytosine that characterize natural DNA? Sec- ondly, will the composition of the natural DNA used as primer influence and determine the composition of the product? In Fig. 8 are the results which answer these two questions (31). The experiments are identical except that in each case a different DNA primer was used: Mycobacterium phlei, Escherichia coli, calf thymus and phage T2 DNA. In answer to the first question it is clear that in the enzymatically synthesized DNA, adenine equals thymine and guanine equals cytosine so that the purine content is in every case identical to the pyrimidine. In answer to the second question it is again apparent that the characteristic ratio of adenine-thymine pairs to guanine-cytosine pairs of a given DNA primer is imposed rather faithfully on the product that is synthe- sized. Whether these measurements are made with isotopic tracers when the net DNA increase is only 1 % or if it is 1 000 % the results are the same. It can be said further that it has not been possible to distort these base ratios by using widely differing molar concentrations of substrates or by any other means. In the last line of Fig. 8 is a rather novel "DNA" which is synthesized under conditions that I will not describe here (18, 32). Suffice it to say that after very long lag periods a copolymer of deoxyadenylate and thymidylate s-70 (A-T) develops which has the physical size and properties of natural DNA and in which the adenine and thymine are in a perfectly alternating sequence. When this rare form of DNA-like polymer is used as a primer, new A-T polymer synthesis starts immediately and even though all four triphosphates be present, no trace of guanine or cytosine can be detected in the product. The conclusion from these several experiments thus seems inescapable that the base composition is replicated in the enzymatic synthesis and that hydrogen-bonding of adenine to thymine and guanine to cytosine is the guiding mechanism. Enzymatic replication of nucleotide sequences. The fourth line of evidence which I would like to cite is drawn from current studies of base sequences in DNA and their replication. As I have suggested already, we believe that DNA is the genetic code; the four kinds of nucleotides make up a four-letter alphabet and their sequence spells out the message. At present we do not know the sequence; what Sanger has done for peptide sequence in protein remains to be done for nucleic acids. The problem is more difficult, but not insoluble. Our present attempts at determining the nucleotide sequences (33) will be described in detail elsewhere and I will only summarize them here. DNA is enzymatically synthesized using P32 as label in one of the deoxynucleoside triphosphates; the other three substrates are unlabeled. This radioactive phosphate, attached to the 5-carbon of the deoxyribose, now becomes the bridge between that substrate molecule and the nucleotide at the growing end of the chain with which it has reacted (Fig. 9) . At the end of the synthetic reaction (after some io16 diester bonds have been formed), the DNA is isolated and digested enzymatically to yield the 3' deoxynucleotides quantitatively. It is apparent (Fig. 9) that the P atom formerly attached to the 5-carbon of the deoxynucleoside triphosphate substrate is now attached to the 3-carbon of the nucleotide with which it reacted during the course of synthesis of the DNA chains. The P32 content of each of the 3' deoxynucleotides, isolated by paper electrophoresis, is a measure of the relative frequency with which a particular substrate reacted with each of the four available nucleotides in the course of synthesis of the DNA chains. This procedure carried out four times, using in turn a different labeled substrate, yields the relative frequencies of all the sixteen possible kinds of dinucleotide (nearest neighbor) sequences. Such studies have to date been carried out using DNA primer samples from six different natural sources. The conclusions are: 1) All 16 possible dinucleotide sequences are found in each case. s-71 Fig. 9. Method for Determining Sequences in DNA. \>B SYNTHESIS (by polymerase) (by micrococcal DNase and splenic diesterase) 2) The pattern of relative frequencies of the sequences is unique and re- producible in each case and is not readily predicted from the base compo- sition of the DNA. 3) Enzymatic replication involves base pairing of adenine to thymine and guanine to cytosine and, most significantly: 4) The frequencies also indicate clearly that the enzymatic replication produces two strands of opposite direction, as predicted by the Watson and Crick model. These studies and anticipated extensions of them should yield the dinucleo- tide frequencies of any DNA sample which can serve as an effective primer for enzymatic replication and thus provide some clues for deciphering the DNA code. Unfortunately this method does not provide information about trinucleotide frequencies but we are hopeful that with the improvement of enzymatic tools for analysis and chromatographic techniques for isolation some start can be made in this direction. Requirement for four triphosphates and DNA for DNA synthesis. Returning to the earlier-stated requirement for all four deoxynucleoside triphosphates and DNA in order to obtain DNA synthesis, we can now regard and understand these requirements as another and final line of evidence for hydrogen bonding. Without added DNA there is no template for hydrogen bonding and without all four triphosphates synthesis stops early and abruptly for lack of a hydrogen bonding mate for one of the bases in the template. s-72 Summary. The enzymatic approaches to the problem of DNA replication and the properties of the DNA-synthesizing enzyme purified from Escherichia coli have been sketched. The unifying and basic generalization about the action of this enzyme is that it catalyzes the synthesis of a new DNA chain in re- sponse to directions from a DNA template; these directions are dictated by the hydrogen bonding relationship of adenine to thymine and guanine to cytosine. The experimental basis for this conclusion is derived from the ob- servations of: (i) The double-stranded character of the enzymatically synthe- sized DNA and its origin from a single stranded molecule, (2) the pattern of substitution of analogues for the naturally-occurring bases, (3) the replication of the chemical composition, (4) the replication of the nucleotide (nearest neighbor) sequences and the antiparallel direction of the strands, and (5) the requirement for all four deoxynucleoside triphosphates (adenine, thymine, guanine and cytosine) and DNA for DNA synthesis. In closing may I repeat what was said at the banquet last night: Any credit for the work cited here is shared by my colleagues in New York, Bethesda, Saint Louis and Stanford, and by the whole international community of chemists, geneticists and physiologists, which is truly responsible for the progress in nucleic acid biochemistry. REFERENCES. 1. O. T. Avery, C. M. MacLeod and M. McCarty, J. Exptl. Med. yg, 137 (1944); R. D. Hotchkiss, in "The Chemical Basis of Heredity" (W. D. McElroy and B. Glass, editors), p. 321 (1957), Johns Hopkins Press, Baltimore. 2. A. D. Hershey, Cold Spring Harbor Symposia Quant. Biol. 18,. 135 (1953). 3. G. W. Beadle, in "Chemical Basis of Heredity" (W. D. McElroy and B. Glass, editors), P- 3 (I957). Johns Hopkins Press, Baltimore. 4. E. Chargaff, in "Nucleic Acids" (E. Chargaff and J. N. Davidson, editors), Vol. I, P- 3°7 — 371 (i955). Academic Press, New York. 5- J. D. Watson and F. H. C. Crick, Nature 171, 737 (1953); Cold Spring Harbor Sym- posia Quant. Biol. 18, 123 (1953). 6. M. H. F. Wilkins, Biochem. Soc. Symposia (Cambridge, England) 14, 13 (1957). 7. M. Feughelman, R. Langridge, W. E. Seeds, A. R. Stokes, H. R. Wilson, C. W. Hooper, M. H. F. Wilkins, R. K. Barclay, and L. D. Hamilton, Nature 175, 834 (i955)- 8. A. Kornberg, in "The Chemical Basis of Heredity" (W. D. McElroy and B. Glass, editors), p. 579 (1957), Johns Hopkins Press, Baltimore; Rev. Mod. Physics 31, 200 (1959). S-73 9. A. Kornberg, in "Phosphorus Metabolism" (W. D. McElroy arid B. Glass, editor)' P- 392 ('951). Johns Hopkins Press, Baltimore; Advances in Enzymol. 18, 191 (1957)- 10. D. E. Koshland, Jr., in "The Mechanism of Enzyme Action" (W. D. McElroy and B. Glass, editors), p. 608 (1954), Johns Hopkins Press, Baltimore. 11. A. Kornberg, I. R. Lehman and E. S. Simms, Federation I>roc. 15, 291 (1956). 12. A. Kornberg, Harvey Lectures jj, 83 (1957 — J958)- 13. I. R. Lehman, M. J. Bessman, E. S. Simms and A. Kornberg, J. Biol. Chem. 233, 163 (1958). 14. F. J. Bollum and V. R. Potter, J. Am. Chem. Soc. 79, 3603 (1957); C. G. Harford and A. Kornberg, Federation Proc. 17, 515 (1958); F. J. Bollum, Federation Proc. 17. *93 (!958); 18, 194 (1959)- 15. M. J. Bessman, I. R. Lehman, E. S. Simms and A. Kornberg, J. Biol. Chem. 233, 171 (1958). 16. J. Adler, I. R. Lehman, M. J. Bessman, E. S. Simms and A. Kornberg, Proc. Nat. Acad. Sci. (U.S.A.) 44, 641 (1958). 17. H. K. Schachman, I. R. Lehman, M. J. Bessman, J. Adler, E. S. Simms and A. Kornberg, Federation Proc. 17, 304 (1958). 18. I. R. Lehman, Ann. N. Y. Acad. Sci. 81, 745 (1959). 19. R. L. Sinsheimer, J. Mol. Biol. 1, 43 (1959). 20. I. R. Lehman, R. L. Sinsheimer and A. Kornberg (Unpublished observations). 21. F. J. Bollum, J. Biol. Chem. 234, 2733 (1959). 22. F. Weygand, A. Wacker and Z. Dellweg, Z. Naturforsch. 7 b, 19 (1952); D. B. Dunn and J. D. Smith, Nature 174, 305 (1954); S. Zamenhof and G. Griboff, Nature 174. 306 (1954)- 23. M. R. Heinrich, V. C. Dewey, R. E. Parks, Jr., and G. W. Kidder, J. Biol. Chem. 197. J99 (1952). 24. M. J. Bessman, I. R. Lehman, J. Adler, S. B. Zimmerman, E. S. Simms and A. Kornberg, Proc. Nat. Acad. Sci. (U. S. A.) 44, 633 (1958). 25. G. R. Wyatt and S. S. Cohen, Biochem. J. 55, 774 (1953). 26. R. L. Sinsheimer, Science 120, 551 (1954); E. Volkin, J. Am. Chem. Soc. 76, 5892 (1954)- 27. R. L. Sinsheimer, Proc. Nat. Acad. Sci. (U. S. A.) 42, 502 (1956); M. A. Jesaitis, J. Exp. Med. jo6, 233 (1957); Federation Proc. 17, 250 (1958). 28. G. Streisinger and J. Weigle, Proc. Nat. Acad. Sci. (U. S. A.) 42, 504 (1956). 29. J. G. Flaks and S. S. Cohen, J. Biol. Chem. 234, 1501 (1959); J. G. Flaks, J. Lich- tenstein and S. S. Cohen, J. Biol. Chem. 234, 1507 (1959). 30. A. Kornberg, S. B. Zimmerman, S. R. Kornberg and J. Josse, Proc. Nat. Acad. Sci. (U. S. A.) 45, 772 (1959)- 31. I. R. Lehman, S. B. Zimmerman, J. Adler, M. J. Bessman, E. S. Simms and A. Kornberg, Proc. Nat. Acad. Sci. (U.S.A.) 44, 1191 (1958). 32. C. M. Radding, J. Adler and H. K. Schachman, Federation Proc , 19,307 (i960). 33- J- Josse and A. Kornberg, Federation Proc, 79,305 (i960). s-74 SUPPLEMENT VI GENES AND CHEMICAL REACTIONS IN NEUROSPORA by George W. Beadle. Pasadena, California, California Institute of Technology. Nobel Lecture, December n, 1958. On this occasion of sharing the high honor of a Nobel Award with Edward L. Tatum for our ". . . discovery that genes act by regulating chemical events", and with Joshua Lederberg for his related ". . . discoveries concerning the organization of the genetic material of bacteria", it seems appropriate that I sketch briefly the background events that led to the work on Neurospora that" Tatum and I initiated in 1940. I shall leave to my co-recipients of the Award the task of describing in detail the developments in Neurospora that followed our first success, and the relation of this to the rise of bacterial genetics, which has depended largely on studies of genetic recombination following con- jugation and transduction. I shall make no attempt to review the entire history of biochemical genetics, for this has been done elsewhere (2, 13, 22, 23). Anthocyanins and Alcaptonuria. Soon after de Vries, Correns and Tschermak "rediscovered" Mendel's 1865 paper and appreciated its full significance, investigators in the exciting new field, which was to be called genetics, naturally speculated about the physical nature of the "elements" of Mendel and the manner of their action. Renamed genes, these units of inheritance were soon found to be carried in the chromosomes. One line of investigation that was destined to reveal much about what genes do was started by Wheldale (later Onslow) in 1903. It began with a genetic study of flower pigmentation in snapdragons. But soon the genetic observations began to be correlated with the chemistry of the anthocyanin and related pigments that were responsible. The material was favorable for s-75 both genetic and chemical studies and the work has continued to yield new information ever since and almost without interruption. Many workers and many species of plants have been involved (2, 4, 13, 22, 23). It became clear very soon that a number of genes were involved and that they acted by somehow controlling the onset of various identifiable and specific chemical reactions. Since an understanding of the genetics helped in interpreting the chemistry and vice versa, the anthrocyanin work was well known to both geneticists and biochemists. It significantly influenced the thinking in both fields and thus had great importance in further develop- ments. A second important line of investigation was begun even earlier by the Oxford phvsician-biochemist Sir Archibald E. Garrod. At the turn of the century he was interested in a group of congenital metabolic diseases in man, which he later named, "inborn errors of metabolism". There are now many diseases described as such; in fact, they have come to be recognized as a cate- gory of diseases of major medical importance. One of the first inborn errors to be studied by Garrod was alcaptonuria. Its most striking symptom is blackening of urine on exposure to air. It had been recorded medically long before Garrod became interested in it and important aspects of its biochemistry were understood. The substance respon- sible for blackening of the urine is alcapton or homogenetisie acid (2,5-dihydr- oxyphenylacetic acid). Garrod suggested early that alcaptonuria behaved in inheritance as though it were differentiated by a single recessive gene. By 1908 a considerable body of knowledge about alcaptonuria had accumu- lated. This was brought together and interpreted by Garrod in his Croonian lectures and in the two editions of his book, "Inborn Errors of Metabolism", which were based on them (11). It was his belief that alcaptonuria was the result of inability of affected individuals to cleave the ring of homogentisic acid as do normal individuals. He believed this to be due to absence or inactivity of the enzyme that normally catalyzes this reaction. This in turn was depen- dent on the absence of the normal form of a specific gene. Thus Garrod had clearly in mind the concept of a gene-enzyme-chemical- reaction system in which all three entities were interrelated in a very specific way. In the 1923 edition of "Inborn Errors" (11) he wrote: "We may further conceive that the splitting of the benzene ring of homo- gentisic acid in normal metabolism is the work of a special enzyme, that in congenital alcaptonuria this enzyme is wanting . . ." Failure to metabolize an intermediate compound when its normal pathway s-76 is thus blocked by a gene-enzyme defect was a part of the interpretation and accounted for the accumulation and excretion of homogentisic acid. Garrod recognized this as a means of identifying an intermediate compound that might otherwise not appear in sufficient amounts to be detected. He also clearly appreciated that alcaptonurics would be used experimentally to explore the metabolic pathways by which homogentisic acid was formed. He summarized a large body of evidence indicating that when normal precursors of homogentisic acid are fed to alcaptonurics there is an almost quantitative increase in homogentisic acid excretion. In this way evidence was accumulated that phenylalanine, tyrosine and the keto acid analog of the latter were almost certainly the direct precursors of homogentisic acid. Despite the simplicity and elegance of Garrod's interpretation of alcap- tonuria and other inborn errors of metabolism as gene defects which resulted in inactivity of specific enzymes and thus in blocked reactions, his work had relatively little influence on the thinking of the geneticists of his time. Bate- son's "Mendel's Principles of Heredity" and a few other books of its time discuss the concept briefly. But up to the 1940's, no widely used later text book of genetics that I have examined even so much as refers to alcaptonuria. It is true that a number of other workers had seriously considered that genes might act in regulating chemical reactions by way of enzymes (2, 13, 17, 21, 23). But there was no other known instance as simple as alcaptonuria. It is interesting — and significant, I think — that it was approximately 50 years after Garrod proposed his hypothesis before it was anything like fully verified through the resolution into six enzymatically catalyzed steps of phenyl- alanine-tyrosine metabolism via the homogentisic acid pathway, and by the clear demonstration that homogentisate oxidase is indeed lacking in the liver of an alcaptonuric (17). Perhaps it is also well to recall that it was not until 1926 that the first enzyme was isolated in crystalline form and shown in a convincing way to consist solely of protein. Eye Pigments of Drosophila. I shall now shift to a consideration of an independent line of investigation that ended up with conclusions very much like those of Garrod and which led directly to the work with Neurospora that Tatum and I subsequently began. In 1933, Boris Ephrussi came to the California Institute of Technology to work on developmental aspects of genetics. During his stay he and I had many long discussions in which we deplored the lack of information about s-77 the manna in which genes act on development. This we ascribed to the fact that the classical organisms of experimental embryology did not lend them- selves readily to genetic investigation. Contrariwise, those plants and animals about which most was known genetically had been little used in studies of development. It would be worth-while, we believed, to attempt to remedy this situation by finding new ways experimentally to study Drosophila melanogaster — which, genetically, was the best understood organism of the time. Tissue culture technics seemed to offer hope. In the spring of 1935 we joined forces in Euphrussi's section of l'lnstitut de Biologie physio-chimique in Paris, resolved to find ways of culturing tissues of the larvae of Drosophila. After some discouraging preliminary attempts, we followed Ephrussi's suggestion and shifted to a transplantation technic. It was our hope that in this way we could make use of non-autonomous genetic characters as a means 01 investigating gene action in development. Drosophila larvae are small. And we were told by a noted Sorbonne authority on the development of diptera that the prospects were not good. In fact, he said, they were terrible. But we were determined to try, so returned to the laboratory, made micro- pipettes, dissected larvae and attempted to transfer embryonic buds from one larva to the body cavity of another. The results were discouraging. But we persisted, and finally one day discovered we had produced a fly with three eyes. Although our joy was great with this small success, we immediately began to worry about three points: First, could we do it again? Second, if we could, would we be able to characterize the diffusible substances responsible for interactions between tissues of different genetic types? And, third, how many non-autonomous characters could we find? We first investigated the sex-linked eye-color mutant vermilion because of the earlier finding of Sturtevant that in gynandromorphs genetically vermilion eye tissue often fails to follow the general rule of autonomy (20). Gynandromorphs may result if in an embryo that begins development as a female from an egg with two X chromosomes, one X chromosome is lost during an early cleavage, giving rise to a sector that has one X chromosome and is male. If the original egg is heterozygous for a sex-linked gene, say ver- milion, and the lost chromosome carries the normal allele, the male sector will be genetically vermilion, whereas the female parts are normal or wild type. (Other sex-linked characters like yellow body or forked bristles can be used as markers to independently reveal genetic constitution in most parts of the body.) s-78 Yet in Sturtevant's gynandromorphs in which only a small part of the body including eye tissue was vermilion, the appearance of that tissue was usually not vermilion but wild type — as though some substance had diffused from wild-type tissue to the eye and caused it to become normally pigmented. It was on the basis of this observation that Ephrussi and I transplanted vermilion eyes into wild type larvae. The result was as expected — the trans- planted eyes were indeed wild type. At that time there were some 26 separate eye-color genes known in Droso- phila. We obtained stocks of all of them and made a series of transplants of mutant eyes into wild-type hosts. We found only one other clear-cut non- autonomous eye character. This was cinnabar, a bright red eye color, like ver- milion but differentiated by a second chromosome recessive gene. We had a third less clear case, claret, but this was never entirely satisfactory from an experimental point of view because it was difficult to distinguish claret from wild-type eyes in transplants. The vermilion and cinnabar characters are alike in appearance; both lack the brown pigment of the wild-type fly but retain the bright red component. Were the diffusible substances that caused them to develop brown pigment when grown in wild-type hosts the same or different? If the same, reciprocal transplants between the two mutants should give mutant transplanted eyes in both cases. If two separate and independent substances were involved, such reciprocal transplants should give wild-type transplanted eyes in both in- stances. We made the experiment and were much puzzled that neither of these results was obtained. A cinnabar eye in a vermilion host remained cinnabar, but a vermilion eye in a cinnabar host became wild type. To explain this result we formulated the hypothesis that there must be two diffusible substances involved, one formed from the other according to the scheme: -* Precursor -*■ v+ substance -*■ cn+ substance -* Pigment . . . where v+ substance is a diffusible material capable of making a vermilion eye become wild type and cn+ substance is capable of doing the same to a cinnabar eye (9). The vermilion (v) mutant gene blocks the first reaction and the cinnabar (en) mutant gene interrupts the second. A vermilion eye in a cinnabar host makes pigment because it can, in its own tissues, convert the v + substance into cn + substance and pigment. In it, the second reaction is not blocked. This scheme involves the following concepts: a. A sequence of two gene-regulated chemical reactions, one gene identified with each. s-79 b. The accumulation of intermediates prior to blocked reactions. c. The ability of the mutant blocked in the first reaction to make use of an intermediate accumulated as a result of a genetic interruption of the second reaction. The principle involved is the same as that employed in the cross- feeding technic later so much used in detecting biosynthetic intermediates in microorganisms. What was later called the one gene-one enzyme concept was clearly in our minds at this time although as I remember, we did not so designate it. Ours was a scheme closely similar to that proposed by Garrod for alcap- tonuria, except that he did not have genes that blocked an adjacent reaction in the sequence. But at the time we were oblivious of Garrod's work, partly because geneticists were not in the habit of referring to it, and partly through failure of ourselves to explore the literature. Garrod's book was available in many libraries. We continued the eye-color investigations at the California Institute of Technology, Ephrussi having returned there to spend part of 1936. Late in the year, Ephrussi returned to Paris and I went for a year to Harvard, both continuing to work along similar lines. We identified the source of dif- fusible substances — fat bodies and malpighian tubercules — and began to devise ways of determining their chemical nature. In this I collaborated to some extent with Professor Kenneth Thimann. In the fall of 1937 I moved to Stanford, where Tatum shortly joined me to take charge of the chemical aspects identifying the eye-color substances. Dr. Yvonne Khouvine worked in a similar rcle with Ephrussi. We made progress slowly. Ephrussi and Khouvine discovered that under certain conditions feeding tryptophane had an effect on vermilion eye color. Following this lead, Tatum found — through accidental contamination of an asceptic culture containing tryptophane and test flies — an aerobic Bacillus that converted tryptophane into a substance highly active in inducing formation of brown pigment in vermilion flies. He soon isolated and crystallized this, but its final identification was slowed down by what later proved to be a sucrose molecule esterified with the active compound. Professor Butenandt and co-workers (6) in Germany who had been col- laborating with Professor Kuhn on an analogous eye-color mutant in the meal moth Ephestia, and Amano et al. (1), working at Osaka University, showed that v+ substance was kynurinine. Later, Butenandt and Hallmank (5), and Butenandt et al. (7) showed that our original cn+ substance was 3-hydroxykynurenine. s-80 Thus was established a reaction series of the kind we had originally con- ceived. Substituting the known chemical, it is as follows: H H C— C—COOH i i H NH2 Tryptophan I ' O H H II I I -— C— C— C— COOH H NH s/ CHO NH I O H H II I I — C— C— C— COOH N-Formylkynurenine I I H NH, Kynurenine NH, en O H H II I I -C — C — C — COON I I H NH, ,, . 3-hydroxykynurenme OH Brown Pigment A New Approach. Isolating the eye-pigment precursors of Drosophila was a slow and dis- couraging job. Tatum and I realized this was likely to be so in most cases of attempting to identify the chemical disturbances underlying inherited ab- normalities; it would be no more than good fortune if any particular example chosen for investigation should prove to be simple chemically. Alcaptonuria was such a happy choice for Garrod, for the chemistry had been largely s-81 worked out and the homogentisic acid isolated and identified many years before. Our idea — to reverse the procedure and look for gene mutations that in- fluence known chemical reactions — was an obvious one. It followed logically from the concept that, in general, enzymatically catalyzed reactions are gene- dependent, presumably through genie control of enzyme specificity. Although we were without doubt influenced in arriving at this approach by the antho- cyanin investigations, by Lwoff's demonstrations that parasites tend to Income specialized nutritionally through loss of ability to synthesize sub- stances that they can obtain readily from their hosts (i.S), and by the specula- tions of others as to how genes might act, the concepts on which it was based developed in our minds fairly directly from the eye-color work Ephrussi and 1 had started five years earlier. The idea was simple: Select an organism like a fungus that has simple nutri- tional requirements. This will mean it can carry out many reactions by which amino acids and vitamins are made. Induce mutations by radiation or other mutagenic agents. Allow meiosis to take place so as to produce spores that .ire genetically homogeneous. Grow these on a medium supplemented with .m array of vitamins and amino acids. Test them by vegetative transfer to a medium with no supplement. Those that have lost the ability to grow on the minimal medium will have lost the ability to synthesize one or more of the substances present in the supplemented medium. The growth requirements of the deficient strain would then be readily ascertained by a systematic series of tests on partially supplemented media. In addition to the above specifications, we wanted an organism well suited to genetic studies, preferably one on which the basic genetic work had already been done. Neurospora. As a graduate student at Cornell, I had heard Dr. B. 0. Dodge of the New York Botanical Garden give a seminar on inheritance in the bread mold Neurospora. So-called second division segregation of mating types and of albinism were a puzzle to him. Several of us who had just been reviewing the evidence for 4-strand crossing over in Drosophila suggested that crossing over between the centromere and the segregating gene could well explain the result. Dodge was an enthusiastic supporter of Neurospora as an organism for genetic work. "It's even better than Drosophila", he insisted to Thomas Hunt MORGAN, whose laboratory he often visited. He finally persuaded Morgan s-82 to take a collection of Neurospora cultures with him from Columbia to the new Biology Division of the California Institute of Technology, which he established in 1928. Shortly thereafter when Carl C. Lindegren came to Morgan's laboratory to become a graduate student, it was suggested that he should work on the genetics of Neurospora as a basis for his thesis. This was a fortunate choice, for Linde- gren had an abundance of imagination, enthusiasm and energy and at the same time had the advice of E. G. Anderson, C. B. Bridges, S. Emerson, A. H. Sturtevant and others at the Institute who at that time were actively interested in problems of crossing over as a part of the mechanism of meiosis. In this favorable setting, Lindegren soon worked out much of the basic genetics of Neurospora. New characters were found and a good start was made toward mapping the chromosomes. Thus, Tatum and I realized that Neurospora was genetically an almost ideal organism for use in our new approach. There was one important unanswered question. We did not know the mold's nutritional requirements. But we had the monograph of Dr. Nils Fries, which told us that the nutritional requirements of a number of related filamentous fungi were simple. Thus encouraged, we obtained strains of Neurospora crassa from Lindegren and from Dodge. Tatum soon discovered that the only growth factor required, other than the usual inorganic salts and sugar, was the recently discovered vitamin, biotin. We could not have used Neurospora for our purposes as much as a year earlier, for biotin would not then have been available in the quantities we required. It remained only to irradiate asexual spores, cross them with a strain of the opposite mating type, allow sexual spores to be produced, isolate them, grow them on a suitably supplemented medium and test them on the un- supplemented medium. We believed so thoroughly that the gene-enzyme- reaction relation was a general one that there was no doubt in our minds that we would find the mutants we wanted. The only worry we had was that their frequency might be so low that we would get discouraged and give up before finding one. We were so concerned about the possible discouragement of a long series of negative results that we prepared more than thousand single spore cultures on supplemented medium before we tested them. The 299th spore isolated gave a mutant strain requiring vitamin B6 and the 1 085th one required Bi. We made a vow to keep going until we had 10 mutants. We soon had dozens. Because of the ease of recovery of all the products of a single meiotic process s-83 n Neurospora, it was a simple matter to determine whether our newly induced nutritional deficiencies were the result of mutations in single genes. If they were, crosses with the original should yield four mutant and four non-mutant spores in each spore sac. ["hey < 1 i. - lie. E- l_ \. I^ilerbcrg. cenl sciences u ill be a cogent challenge to the intellectual powers of our su< That bacteria and their genetics should now . s relevant to general h; fresh cycle in our scientific outlook. When ght oi at all, thev have often been re.. [ byway of evolution, their complexity and their homology with other or- ganisms grossl) underrated. "Since Pasteur's startling reries r the important role played bv microbes in human affairs, micro- science has always suffered from its eminent practical applications. By far the majority of the microbiological studies were undertaken to answer questions connected with the well-being of mankind" (50). The E „..-■ a lemic biology from medical education has helped sustain this dis- ci. Happily, the repatriation of bacteria and viruses is only the first measure of the re- cnt of medicine's debt to bu g mparative biochemistry has consum- mated the unification of biology revitalized by Darwin one hundred years ago. Throughout the living world u ; :nmon set of struc- tural units — amino acids, coenzymes, nucleins, carbohydrates and so forth — from which every g :sm builds itself. The same holds for the fundamental process of biosynthesis and of _\ metabolism. The exceptions to this rule thus command special interest as mean- ingful tokens of biological individuality, the replacement of cytosine by hydroxymethyl ne in the DN g .2). N ::rition has been a special triumph. Bac- teria which required no vitamins had seemed simpler than man. But deeper insights 61) interpret nutritional simplicity as a greater power of synthesis. The requirements of more g nisms comprise just those me- tabolites they cannot synthesize with their own enzymatic machinery. Species differ in their nutrition: if species are :eir gent les must con- trol the Hosynthetic steps which are re s-98 hederberg: Ger. in nutritional patterns. This syllogism, so evi- dent once told, has been amplified by Beadle and Tatum from this podium. Its implications for experimental biology and medicine are well known: among these, the methodology of bacterial genetics. Tatum has related how his early experience with bacterial nutrition reinforced the foundations of the biochemical gene- • turospora. Then, disregarding the common knowledge that bacteria were too simple to have genes, Tatum took courage to for the genes that would indeed control bacterial nutrition. This conjunction marked the start of my own happy association with him, and with the fascinating challenges of bacterial genetics. Contemporary genetic research is predicated on the role of DNA ^s the genetic material, of enzyme proteins as the cell's working tools, and of RNA as the communication channel between them Three lines of evidence substantiate the genetic function of L are related to bacterial genetics; the third and most general is the cytochemical observa- tion of DNA in the chromosomes, which are undeniably strings of gents. But chromo- somes also contain other constituents besides DNA : we want a technique to isolate a chro- mosome or a fragment of one, to analyze it, and to retransplant it to verify its functional capacity. The impressive achievements of nu- clear transplantation (29) should encourage the audacity needed to try such experiments. The constructive equivalent to chromosome transplantation was discovered by a bacteriolo- gist thirty years ago (20) . The genetic impli- cations of the "pneumococeus transformation'" in the minds of some of Griffith's successors were clouded by its involvement with the gummy outer capsule of the bacteria. How- : .43. Avery and his colleagues had shown that this inherited trait was transmitted from one pneumococcal strain to another by DNA. The general transmission of other traits by the same mechanism (25) can only mean that DNA comprises the genes ( : To reinforce this conclusion. Hershey and Chase (23) proved that the genetic element of a bacterial virus is also DNA. Infection of a host cell requires the infection of just the DNA content of the adsorbed particle. This DNA controls not only its own replication in the production of new phage but also the speci- ficity of the protein coat, which governs the serological and host range specificity of the :'.-i :- : .:. -. . •-: least in some small viruses, RNA also displays genetic functions. However, the he- reditary autonomy of gene-initiated RN the cytoplasm is now very doubtful — at least some of the plasmagenes that have been pro- posed as fulfilling this function are now better understood as feedback-regulated systems of substrate-transport (81, 65.-1 The work of the past decade thus strongly supports the simple doctrine that genetic infor- mation is nucleic, ie_ is coded in a linear se- quence of nucleotides. This simplification of life may appear too facile, and has furnished a tempting target for agnostic criticism (37 - - . 74). But, while no scientific theory would decry continual refinement and amplification, such criticism has little value if it detracts from the evident fruirfulness of the doctrine in ex- : --rr.tr.--. it.zr. The cell may, of course, carry information other than nucleic either in the cytoplasm or, accessory to the polynucleotide sequence, in :r.t :r- rr_- :■ rrt: Z; :r._:!t:: :.-.:''. rrr.i :.;.-. : .: been invoked, without being more precisely defined, in many recent speculations on cyto- diff erentiation and on such models of this as antigenic phase variation in SalmoneUs - 52, 56, 47). Alternative schemes have so much tv. :.-.: ■:.-.-. i- . - ::: : :.: . ':.::::::.:.:.: v-i.t that they are more likely to concern the r egula- . r. :: ztr.:: :_~:\ -■ zr.ir. :: ~:~. : :r.t - DXA AS A SUBSTANCE The chemistry deserves to be ex- posed by apter craftsmen (86, 31, 13) and I ir.i". rr.trt.. :t:iz.~.J.-t r-ti ::t :;:-: -.; :■ biological implications. A segment c illustrated in Fig. 1. This shows a linear poly- ~;: -■■■■:. it - n-.z-r.t : ri.r.\ r.t -..: j unit: — O — PCX— O i.t-.\t- rr.\ - _:t c c The carbon atoms are conventionally num- ztrti 1:: ri.r.z: rt.r. •.:.':.::.::.::.::.-.- i-rr Stanford Medical Bulletin ^-/ S V/H S V w c^ _ y / N C — N \ / V C / y \ s ~-N / N C // \ O H " — c JL\ v/ \ c . / / H\ / \ x / O H " — L 1\ v/ / Fig. i. — Primary structure of DNA — a segment of a polynucleotide sequence CGGT. From (13). ose ring of deoxyribose, which is coupled as aw N glvcoside to one of the nuclein bases: adenine, guanine, cytosine, or thymine, sym- bolized A. G, C, or T, the now well-known alphabet in which genetic instructions are com- posed. With a chain length of about 10,000 residues, one molecule of DNA contains 20,000 "bits of information," comparable to the text of this article, or in a page of newsprint. Pyrophosphate - activated monomer units (e.g., thymidine triphosphate) have been iden- tified as the metabolic precursors of DNA (31). For genetic replication, the monomer units must be assembled in a sequence that re- flects that of the parent molecule. A plausible mechanism has been forwarded by Watson and Crick (87) as a corollary to their struc- tural model whereby DNA occurs as a two- stranded helix, the bases being centrally ori- ented. When their relative positions are fixed by the deoxyribose-phosphate backbones, just two pairs of bases are able to form hydrogen bonds between their respective NH and CO groups; these are A : T and G : C. This pairing of bases would tie the two strands together for the length of the helix. In conformity with this model, extensive analytical evidence shows a remarkable equality of A with T and of G with C in DNA from various sources. The two strands of any DNA are then mutually complementary, the A, T, G, and C of one strand being represented by T, A, C, and G, respectively, of the other. The information of one strand is therefore equivalent to, because fully determined by, the other. The determi- nation occurs at the replication of one parent strand by the controlled stepwise accretion of monomers to form a complementary strand. At each step, only the monomer which is com- plementary to the template would fit for a chain-lengthening esterification with the adja- cent nucleotide. The model requires the un- raveling of the intertwined helices to allow each of them to serve as a template. This might, however, occur gradually, with the growth of the daughter chain — a concept em- bedded in Fig. 2 which symbolizes the new Cabala. The discovery of a single-stranded configuration of DNA (85) makes complete unraveling more tenable as an alternative model. For the vehicle of life's continuity, DNA may seem a remarkably undistinguished mole- cule. Its over-all shape is controlled by the uni- form deoxyribose-phosphate backbone whose monotony then gives X-ray diffraction pat- terns of high crystallinity. The nucleins them- selves are relatively unreactive, hardly differ- ent from one to the other, and in DNA intro- verted and mutually saturated. Nor are any of the hydroxyls of deoxyribose left unsubstituted in the polymer. The structure of DNA befits the solipsism of its function. The most plausible function of DNA is ultimately to specify the amino acid sequence in proteins. However, as there are twenty amino acids to choose among, there cannot be a one: one correspondence of nucleotide to amino acid. Taking account of the code dupli- cation in complementary structures and the need to indicate spacing of the words in the code sequence, from three to four nucleins s-100 Lederberg: Genetics may be needed to spell one amino acid (19). While a protein is also defined by the se- quence of its monomeric units, the amino acids, the protein molecule lacks the "aperiodic crystallinity" (80) of DNA. The differentiae of the amino acids vary widely in size, shape, and ionic charge (e.g., H;NCH/CH>CH,CH, •; COOHCH2CH2-; HOC8H4-CH,-; CH3-, H*) and in the case of proline, bond angles. Fig. 2. — The scheme of Watson and Crick for DNA replication. "Unwinding and replication proceed pari passu. All three arms of the Y ro- tate as indicated" ( 14). The biological action of a protein is, therefore, attributable to the shape of the critical surface into which the polypeptide chain folds (73). The one-dimensional specificity of the DNA must therefore be translated into the three-di- mensional specificity of an enzvme or anti- body surface. The simplest assumption would be that the amino acid sequence of the ex- tended polypeptide, as it is released from the protein-building template in the cytoplasm, fully determines the folding pattern of the complete protein, which may, of course, be stabilized by nonpeptide linkages. If not, we should have to interpose some accessory mech- anism to govern the folding of the protein. This issue has reached a climax in speculations about the mechanism of antibody formation. If antibody globulins have a common se- quence on which specificity is superimposed by directed folding, an antigen could directly mold the corresponding antibody. However, if sequence determines folding, it should in turn obey nucleic information. As this should be independent of antigenic instruction, we may look instead to a purely selective role of antigens to choose among nucleic alternatives which arise by spontaneous mutation (8, 50). The correspondence between amino acids and clusters of nucleotides has no evident basis in their inherent chemical make-up and it now appears more probable that this code has evolved secondarily and arbitrarily to be trans- lated by some biological intermediary. The coding relationship would then be analogous to, say, Morse-English (binary linear) to Chi- nese (pictographic). Encouragingly, several workers have reported the enzymatic reaction of amino acids with RNA fragments (22, 75). Apparently each amino acid has a different RNA receptor and an enzyme whose twofold specificity thus obviates any direct recognition of amino acid by polynucleotides. The align- ment of amino-acyl residues for protein syn- thesis could then follow controlled assembly of their nucleotidates on an RNA template, by analogy with the model for DNA replication. We then visualize the following modes of in- formation transfer: (1) DNA replication — assembly of comple- mentary deoxyribonucleotides on a DNA template. (2) Transfer to RNA by some comparable mechanism of assembling ribonucleotides. Our understanding of this is limited by uncertainties of the structure of RNA (16). (3) Protein synthesis: (a) Aminoacylation of polynucleotide fragments; (/?) Assembly of the nucleotidates on an RNA template bv analogy with step (1);. (c) Peptide condensation of the amino acid residues. Some workers have suggested that RNA is 101 Stanford Medical Bulletin replicated in step (3) concurrently with pro- tein synthesis, in addition t<> its initiation from DNA. The chief difference in primary structure be- tween 1 )NA and RNA is the hydroxylation of I .1 the ribose, so that a reactive sugar hy- droxyl is available in RNA. This may prove to be important in the less ordered secondary structure of RNA. and in its function as an intermediary to protein. It remains to be de- termined whether the aminoacyl rrucleotidates are esterified at <. V or at C ' which is also avail- able in the terminal residue. From this resume we may observe that the DNA backbone con- stitutes m\ inert but ngul framework on which the differential nucleins are strung. Their spa- tial constraint lends specificity to the pattern of hydrogen bonding exposed at each level. This extended pattern is a plausible basis for replication; it is difficult to visualize any re- agents besides other nucleotides to which this pattern would be relevant. These conditions are quite apt tor a memory device — rubber and guncotton are poor choices for a computing tape. DNA AND BACTERIAL MUTATION The ignis fatuus of genetics has been the specific mutagen, the reagent that would pene- trate to a given gene, recognize and modify it in a specific way. Directed mutation has long been discredited for higher organisms and the "molar indeterminacy" of mutation estab- lished both for its spontaneous occurrence and for its enhancement by X-rays (68). However, the development of resistance apparently in- duced by drugs revived illusions that bacterial genes might be alterable, an inference that would inevitably undermine the conception of "gene" for these organisms. No wonder that the mechanism of drug resistance has excited so much controversy (89)! What sort of molecule could function as a specific mutagen, a reagent for a particular one ot the bacterium's complement of genes, which can hardly number less than a thousand targets5 On the nucleic hypothesis, the small- est segment capable of this variety would be a /;ev different sites are more readil) discovered than those ol >■ would be. (5) Induction. Exposure ol lysogenic bacteria to small doses of ultraviolet light causes the prophage to initiate a lytic cycle with the appearance first ol vegetative, then of ma- ture phage (62). Hfr bacteria make no analogous response. However, the kinetics of the reversion, Hfr ~~ * F+, has not been carefully studied. The genetic function of bacteriophages is further exemplified by transduction whereby genes are transferred from cell to cell by the intervention of phage particles (42,91). In our first studies we concluded that the bacterial genes were adventitiously carried in normal phage particles (92, 66, 83). Further studies favor the view that the transducing particle has a normal phage coat but a defective phage nucleus. This correlation has suggested that a gene becomes transducible when a prophage segment is translocated to its vicinity (2, 9, 60). Transduction focuses special attention on the phenomenon of specific pairing of homol- ogous chromosome segments. Howsoever a transduced gene is finally integrated into the bacterial genome, at some stage it must locate the homologous gene in the recipient chromo- some. For in transduction, as in sexual recom- bination, new information is not merely added to the complement; it must also replace the old. This must involve the confrontation of the two homologues prior to the decision which one is to be retained. Synapsis is even more puzzling as between chromosomes whose DNA is in the stabilized double helix and then further contracted by supercoiling. Conceivably gene products rather than DNA are the agency of synaptic pairing. The integration of a transduced fragment raises further issues (41). The competing hy- potheses are the physical incorporation of the fragment in the recipient chromosome, or the use of its information when new DNA is repli- cated. The same issues still confound models of crossing over at meiosis in higher forms; s- once again the fundamentals of chromosome structure are needed for a resolution. VIRUS VERSUS GENE The homology of gene and virus in their fundamental aspects makes their overt differ- ences even more puzzling. According to the simplest nucleic doctrine, DNA plays no active role in its own replication other than furnish- ing a useful pattern. Various nucleotide se- quences should then be equally replicable. What then distinguishes virus DNA, which replicates itself at the expense of the other path- ways of cellular anabolism? For the T-even phages, the presence of the unique glucos- ylated hydroxymethylcytosine furnishes a partial answer (12). However, other viruses such as a display no unique constituents; fur- thermore, as prophage they replicate coordi- nate^ with bacterial DNA. Does the virus have a unique element of structure, either chemical or physical, so far undetected? Or does it instruct its own preferential synthesis by a code for supporting enzymes? THE CREATION OF LIFE The mutualism of DNA, RNA, and pro- teins as just reviewed is fundamental to all contemporary life. Viruses are simpler as in- fective particles but must, of course, parasitize the metabolic machinery of the host cell. What would be the least requirements of a primeval organism, the simplest starting point for pro- gressive replication of DNA in terms of pres- ently known or conjectured mechanisms? They include at least: (1) DNA. (2) The four deoxyribotide pyrophosphates in abundance. (3) One molecule of the protein, DNA po- lymerase. (4) Ribotide phosphates as precursors for RNA. (5) One molecule of the protein RNA po- lymerase. (6) A supply of the twenty amino acyl nucle- otidates. (a) Failing these, each of the twenty en- zymes which catalyze the condensa- tion of an amino acid and correspond- 106 Lederberg: Genetics ing RNA fragments together with sources of these components. (7) One molecule of the protein aminoacyl- RNA polymerase. In principle, this formidable list might be reduced to a single polynucleotide polymer- ized by a single enzyme. However, any scheme for the enzymatic synthesis of nucleic acid calls for the coincidence of a particular nucleic acid and of a particular protein. This is a far more stringent improbability than the sudden emergence of an isolated DNA such as many authors have suggested, so much more so that we must look for alternative solutions to the problem of the origin of life. These are of two kinds. The primeval organism could still be a nucleic cycle if nucleic replication oc- curs, however imperfectly, without the inter- vention of protein. The polymerase enzyme, and the transfer of information from nucleic acid to protein, would then be evolved refine- ments. Alternatively, DNA has evolved from a simpler, spontaneously condensing polymer. The exquisite perfection of DNA makes the second suggestion all the more plausible. The nucleoprotein cycle is the climax of bio- chemical evolution. Its antiquity is shown by its adoption by all phyla. Having persisted for — io9 years, nucleoprotein may be the most durable feature of the geochemistry of this planet. At the present time, no other self-replicat- ing polymers are known or understood. Never- theless, the nucleic system illustrates the basic requirements for such a polymer. It must have a rigid periodic structure in which two or more alternative units can be readily substituted. It must allow for the reversible sorption of spe- cific monomers to the units in its own se- quence. Adjacent, sorbed monomers must then condense to form the replica polymer, which must be able to desorb from the tem- plate. Primitively, the condensation must be spontaneous but reliable. In DNA, the sorp- tion depends on the hydrogen bonding of nu- clein molecules constrained on a rigid helical backbone. This highly specific but subtle de- sign would be difficult to imitate. For the more primitive stages, both of biological evolution and of our own experimental insight, we may- prefer to invoke somewhat cruder techniques of complementary attachment. The simplest of these is perhaps the attraction between ionic groups of opposite charge, for example, NH: and COO which are so prevalent in simple organic compounds. If the ingenuity and craftsmanship so successfully directed at the fabrication of organic polymers for the practi- cal needs of mankind were to be concentrated on the problem of constructing a self-replicat- ing assembly along these lines I predict that the construction of an artificial molecule hav- ing the essential function of primitive life would fall within the grasp of our current knowledge of organic chemistry. CONCLUSIONS The experimental control of cellular geno- type is one of the measures of the scope of genetic science. However, nucleic genes will not be readily approached for experimental manipulation except by reagents that mimic them in periodic structure. Specifically in- duced mutation, if ever accomplished, will then consist of an act of genetic recombination between the target DNA and the controlled information specified by the reagent. Methods for the step-wise analysis and reassembly of nucleic acids are likely to be perfected in the near future in pace with the accessibility of nucleic acid preparations which are homo- geneous enough to make their use worth while. For the immediate future, it is likely that the greatest success will attend the use of biological reagents to furnish the selectivity needed to discriminate one among innumerable classes of polynucleotides. Synthetic chemistry is, however, challenged to produce model poly- mers that can emulate the essential features of genetic systems. REFERENCES 1. Anderson, T. F., "Recombination and Segre- gation in Escherichia coli" Cold Spring Harbor Symposia Quant. Biol., 23:47, 1958. 2. Arber, W., "Transduction dcs caracteres Gal par le bacteriophage Lambda," Archives des Sciences, Soc. Phys. Hist. Nat. Geneve, 1 1 : 259, 1958. 3. Baron, L. S., Carey, W. F., and Spilman, W. M., "Hybridization of Salmonella Species by Mating with Escherichia coli" Abst. yth Int. 107 Stanford Medical Bulletin Microbiol. (Stockholm), pp. 50 51, 1 i^sH. 4. Benzer, S., "The Elementary Units of He- redity," in The Chemical Basis of Heredity, W. I). McElroj and B. Glass (eds.), Baltimore: The Johns llopkmv Press, k/^-. pp. 7°-93- s. Bryson, \'., and Szybalski, W., "Microbial Drug Resistance," Adv. Genetics, 7:1, t955« 6. Burnet, F. M., Biological Aspects of Infec- tious Disease, Cambridge: Cambridge Univer- sity Press, [940. -. Burnet, F. M., Virus as Organism, Cam- bridge, Mass.: Harvard University Press, 1945. 8. Burnet, sir MacFarlane, The Clonal Selec- tion Theory of Immunity ( Abraham Flexner Lec- tures, 1958), Nashville: Vanderbilt University Press, 1959. 9. Campbell, A., "Transduction and Segrega- tion in Escherichia coii k-12," Virology, 4: 566, [957. 10. Cavalli-Sforza, L. L., and Lederberg, J., "Genetics of Resistance to Bacterial Inhibitors," in Symposium on Growth Inhibition and Chemo- therapy, Int. Cong. Microbiol. (Rome), 1953, pp. [08-42. 11. , "Isolation of Preadaptive Mutants in Bacteria by Sib Selection," Genetics, 41 : 367, 12. Cohen, S. S., "Molecular Bases of Para- sitism of Some Bacterial Viruses," Science, 123: 653, 1956. 13. Crick, F. H. C, "The Structure of the Hereditary Material," Scient. Am., 151:54, 1954. 14. Delbrikk, M., and Stent, G. S., "On the Mechanism of DNA Replication," in The Chemi- cal Basis of Heredity, W. D. McElroy and B. Glass (eds.), Baltimore: The Johns Hopkins Press, 1957, pp. 699-736. 15. Demerec, M., Hartman, Z., Hartman, P. E., Yura, T., Gots, J. S., Ozeki, H., and Glover, S. W., Genetic Studies with Bacteria, Washing- ton, D.C.: Carnegie Inst. Publ. 612, 1956. 16. Doty, P., Boedtker, H., Fresco, J. R., Hasel- korn, R., and Litt, M., "Secondary Structure in Ribonucleic Acids," Proc. Nat. Acad. Sc, 45 : 4s-- '959- 17. von Euler, H., "Einfluss des Streptomycins auf die Chlorophyllbildung," Kem. Art., 9: 1, '447- 18. Freese, E., "The Difference Between Spon- taneous and Base-Analogue Induced Mutations of Phage T4," Pros. Nat. Acad. Sc, 45 : 622, 1959. 19. Golomb, S. W., Welch, L. R., and Del- brikk, M., "Construction and Properties of 8-1 Comma-Free Codes," Biol. Medd. Can. Vid. Sels{., 23 : 1, 1958. 20. Griffith, F., "The Significance of Pneumo- coccal Types," /. Hyg., 27 : 113, 1928. 21. Haas, F. L., and Doudney, C. O., "A Rela- tion of Nucleic Acid Synthesis to Radiation-In- duced Mutation Frequency in Bacteria," Proc. Nat. Acad. Sc, 43 : 871, 1957. 22. Hecht, L. I., Stephenson, M. L., and Zamec- nik, P. C, "Binding of Amino Acids to the End Group of a Soluble Ribonucleic Acid," Proc. Nat. Acad. Sc, 45: 505, 1959. 23. Hershey, A. D., and Chase, M., "Independ- ent Function of Viral Protein and Nucleic Acid in Growth of Bacteriophage," /. Gen. Physiol.. 36:39- 195'- 24. Hiiota, Y., and Iijima, T., "Acriflavine as an Effective Agent for Eliminating F-Factor in Escherichia coli K-12," Nature, 180: 655, 1957. 25. Hotchkiss, R. D., "The Genetic Chemistry of the Pneumococcal Transformations," Harvey Lect., 49 : 124, 1955. 26. Jacob, F., and Wollman, E. L., "Sur les pro- cessus de conjugaison et de recombinaison chez Escherichia coir. I. L'induction par conjugaison ou induction zygotique," Ann Inst. Pasteur, 91 : 486, 1956. 27. , "Genetic and Physical Determina- tion of Chromosomal Segments in Escherichia coli," Symp. Soc Ex per. Biol., 7: 75, 1958. 28. , "Les episomes, elements geneti- ques ajoutes," C. R. Acad. Sc, Paris, 247: 154, 1958. 29. King, T. J., and Briggs, R., "Serial Trans- plantation of Embryonic Nuclei," Cold Spring Harbor Symposia Quant. Biol., 21 : 271, 1956. 30. Kluyver, A. J., and van Niel, C. B., The Microbe's Contribution to Biology, Cambridge, Mass.: Harvard University Press, 1956. 31. Romberg, A., "Enzymatic Synthesis of De- oxyribonucleic Acid," Harvey Lect., 53 : 83, 1959. 32. Knight, B. C. J. G., "Bacterial Nutrition," Med. Res. Council (Brit.), Spec. Rep. Ser. No. 210, 1936. 33. Lederberg, E. M., "Allelic Relationships and Reverse Mutation in Escherichia coli," Ge- netics, 37 : 469, 1952. 34. Lederberg, E. M., and Lederberg, J., "Ge- netic Studies of Lysogenicity in Escherichia coli," Genetics, 38: 51, 1953. 35. Lederberg, J., "Gene Recombination and Linked Segregations in Escherichia coli," Ge- netics, 32 : 505, 1947. 36. , "Problems in Microbial Genetics," Heredity, 2 : 145, 1948. 08 Lederberg: Genetics 37. , "Bacterial Variation," Ann. Rev. Microbiol., 3:1, 1949. 38. , "Genetic Studies with Bacteria," in Genetics in the 20th Century, L. C. Dunn (ed.), New York: The Macmillan Company, 1951, pp. 263-89. 39. , "Prevalence of Escherichia coli Strains Exhibiting Genetic Recombination," Science, 114: 68, 1951. 40. , "Cell Genetics and Hereditary Symbiosis," Physiol. Rev., 32: 403, 1952. 41. , "Recombination Mechanisms in Bacteria," /. Cell & Comp. Physiol., 45 (Suppl. 2): 75, 1955- 42. , "Genetic Transduction," Am. Scientist, 44 : 264, 1956. 43. , "Conjugal Pairing in Escherichia coli," /. Bad., 71 : 497, 1956. 44. , "Comments on Gene-Enzyme Re- lationship," in Enzymes: Units of Biological Structure and Function, O. H. Gaebler (ed.), New York: Academic Press Inc., 1956, pp. 161- 69- 45. , "Viruses, Genes and Cells, Bad. Rev., 21 : 133, 1957. 46. , "Sibling Recombinants in Zygote Pedigrees of Escherichia coli," Proc. Nat. Acad. Sc, 43 : 1060, 1957. 47. , "Genetic Approaches to Somatic Cell Variation: Summary Comment," /. Cell. & Comp. Physiol., 52 (Suppl. 1): 383, 1958. 48. , "Extranuclear Transmission of the F Compatibility Factor in Escherichia coli," Abstr. yth Int. Cong. Microbiol. (Stockholm), pp. 58-60, 1958. 49. , "Bacterial Reproduction," Harvey Led., 53: 69, 1959. 50. , "Genes and Antibodies," Science, 129 : 1649, 1959. 51. Lederberg, J., Cavalli, L. L., and Leder- berg, E. M., "Sex Compatibility in Escherichia coli," Genetics, 37: 720, 1952. 52. Lederberg, }., and Edwards, P. R., "Sero- typic Recombination in Salmonella," /. Im- munol., 71 : 232, 1953. 53. Lederberg, }., and Lederberg, E. M., "Replica Plating and Indirect Selection of Bac- terial Mutants," /. Bad., 63 : 399, 1952. 54. , "Infection and Heredity," Symp. Soc. Growth and Develop., 14: 101, 1956. 55. Lederberg, J., Lederberg, E. M., Zinder, N.D., and Lively, E. R., "Recombination Analysis of Bacterial Heredity," Cold Spring Harbor Symposia Quant. Biol., 16: 413, 1951. 56. Lederberg, J., and lino, T., "Phase Varia- tion in Salmonella," Genetics, 41 : 743, 1956. 57. Lederberg, J., and Tatum, E. L., "Gene Recombination in Escherichia coli," Nature, 158 : 558, 1946. 58. , "Sex in Bacteria: Genetic Studies, 1 945-1 952," Science, 118: 169, 1954. 59. Luria, S. E., and Burrous, J. W., "Hybrid- ization Between Escherichia coli and Shigella," /. Bad., 74:461, 1957. 60. Luria, S. E., Fraser, D. K., Adams, J. N., and Burrous, J. W., "Lysogenization, Transduc- tion, and Genetic Recombination in Bacteria," Cold Spring Harbor Symposia Quant. Biol., 23 : 71, 1958. 61. Lwoff, A., "Les facteurs de croissance pour les microorganismes," Ann. Inst. Pasteur, 61: 580, 1938. 62. Lwoff, A., Siminovitch, L., and Kjeld- gaard, N., "Induction de la production de bac- teriophages chez une bacterie lysogene," Ann. Inst. Pasteur, 79 : 815, 1950. 63. McElroy, W. D., and Glass, B. (eds.), The Chemical Basis of Heredity, Baltimore: The Johns Hopkins Press, 1957. 64. Meselson, M., and Stahl, F. W., "The Repli- cation of DNA in Escherichia coli," Proc. Nat. Acad. Sc, 44: 671, 1958. 65. Monod, J., "Remarks on the Mechanism of Enzyme Induction," in Enzymes: Units of Biological Structure and Function, O. H. Gaebler (ed.), New York: Academic Press Inc., 1956, pp. 7-28. 66. Morse, M. L., Lederberg, E. M., and Leder- berg, J., "Transduction in Escherichia coli K-12," Genetics, 41 : 142, 1956. 67. , "Transductional Heterogenotes in Escherichia coli," Genetics, 41 : 758, 1956. 68. Muller, H. J., "The Production of Muta- tions," Les Prix Nobel en 1946, Stockholm, 1948, pp. 257-74. 69. Nagel, E., "The Meaning of Reduction in the Natural Sciences," in Science and Civilization, R. C. Stauffer (ed.), Madison: University of Wis- consin Press, 1949, pp. 99-138. 70. Nanney, D. L., "Epigenetic Control Sys- tems," Proc. Nat. Acad. Sc, 44: 712, 1958. 71. Novick, A., "Mutagens and Antimuta- gens," Broo\haven Symposia in Biology, 8 (Mutation): 201-15, Washington, D.C.: Office of Tech. Serv., U.S. Dept. Commerce, 1956. 72. Novick, A., and McCoy, A., "Quasi-Ge- netic Regulation of Enzyme Level," in Physio- logical Adaptation, Washington, D.C.: Amer. Physiol. Soc, 1958, pp. 140-50. 73. Pauling, L., "Molecular Structure and In- termolecular Forces," in The Specificity of Sero- logical Reactions, K. Landsteiner, Cambridge, 109 Stanford Medical Bulletin M.tsv: Harvard University Press, 19451 pp- 93- 74. Pirie, N. W., "Sour- Aspects oi the < )ngins oi Life Considered in the Light oi the Moscow Intcrn.ition.il Symposium," ICSU Rer., 1:40, 1959. 75. Thins. J., Berg, P., Ofengand, E. J., Berg- ni.inn. F. 11.. and Dieckmann, M-, "The Chemi- cal Nature oi the RNA Amino Acid Compound Formed by Amino Acid-Activating Enzymes," Wat. Acad. Sc, 45 : $19, [959. 7<>. Provasoli, 1... Hutner, S. H.. and Pintner, I. J., 'Destruction of ChloroplastS by Streptomy- cin." Cold Spring } I arbor Symposia Quant. Biol., 16 : 1 1 }, 1 95 1 . 77. Richter, A. A.. "Determinants of Mating Type in Escherichia coli," Ph.D. Dissertation, University ol Wisconsin (University Microfilm, Ann Arbor, Mich. ), 1959. 78. Ryan, F. J., and Lederberg, }., "Reverse- Mutation in Leucineless Neurospora," Proc. Nat. . lead. Sc, 32 : 163, 1946. - j. Sanger, F., Les Prix Nobel en 1958, Stock- holm, K)SO. 8< ■. Schrodinger, E., What Is Life? Cambridge: Cambridge University Press, 1944. Si. Spiegelman, S., Lindegren, C. C, and Lindegren, G., "Maintenance and Increase of a Genetic Character by a Substrate-Cytoplasmic In- teraction in the Absence of the Specific Gene," Proc. Nat. Acad. Sc, 31 : 95, 1945. 82. Stent, G. S., "Mating in the Reproduction oi Bacterial Viruses," Adv. Virus Research, 5: 95, 1958. 83. Stocker, B. A. D. S., Zinder, N. D, and Lederberg, }., "Transduction of Flagellar Charac- ters in Salmonella," f. Gen. Microbiol., 9: 410, 1593. 84. Tatum, E. L., and Lederberg, J., "Gene Recombination in the Bacterium Escherichia colt," j. Bad., 53: 673, 1947. 85. Tessman, I., "Some Unusual Properties of the Nucleic Acid in Bacteriophages S13 and *I'X 1 74," Virology, 7: 263, 1959. Todd, Sir Alexander, "Synthesis in the Study of Nucleotides," Les Prix Nobel en 1957, Stockholm, pp. 119-33, 1 95^- 87. Watson, J. D., and Crick, F. H. C, "The Structure of DNA," Cold Spring Harbor Sym- posia Ouanr. Biol., 23: 123, 1953. 88. Wollman, E. L., Jacob, F., and Hayes, W., "Conjugation and Genetic Recombination in Escherichia coli K.-12," Cold Spring Harbor Sym- posia Quant. Biol., 21 : 141, 1956. 89. Wolstenholme, G. E. W., and O'Connor, C. M. (eds.), Ciba Foundation Symposium on Drug Resistance in Microorganisms, London: J. and A. Churchill, Ltd., 1957. 90. Yanofsky, C, and Crawford, I. P., "Effects of Deletions, Point Mutations, Suppressor Muta- tions and Reversions on the Two Components of Tryptophane Synthetase of Escherichia coli," Proc. Nat. Acad. Sc, 45 ( in press), 1959. 91. Zinder, N. D., "Bacterial Transduction," /. Cell. & Comp. Physiol., 45 (Suppl. 2): 23, 1955- 92. Zinder, N. D., and Lederberg, }., "Genetic Exchange in Salmonella," J. Bad., 64 : 679, 1952. 93. 0rskov, F., and 0rskov, I., unpublished observations. a. No reader who recognizes deoxyribonucleic acid will need to be reminded what DNA stands for. b. One might be tempted to write: "One DNA molecule = one gene." However, the quanta of factorial genetics, based on mutation, recombina- tion, and enzymatic function are all smaller than the DNA unit of molecular weight — 6 X i°6 (4). There is increasing evidence that such a molecule is a natural unit rather than an artefact of fragmentation. c The experimental work from my laboratory summarized in this paper has been generously supported by research grants from the National Institutes of Health, U.S. Public Health Service, the National Science Foundation, the Rockefeller Foundation, the Wisconsin Alumni Research Foundation, the University of Wisconsin, and, most recently, Stanford University. It is also a pleasure to record my thanks to the Jane Coffin Childs Fund for Medical Research for a research fellowship which supported my first association with Professor E. L. Tatum. s-110 SUPPLEMENT IX James D.Watson The involvement of RNA in the synthesis of proteins Nobel Lecture, December 11, ig62 Prologue I arrived in Cambridge in the fall of 195 1. Though my previous interests were largely genetic, Luria had arranged for me to work with John Kendrcw. I was becoming frustrated with phage experiments and wanted to learn more about the actual structures of the molecules which the geneticists talked about so passionately. At the same time John needed a student and hoped that I should help him with his X-ray studies on myoglobin. I thus became a research student of Clare College with John as my supervisor. But almost as soon as I set foot in the Cavendish, I inwardly knew I would never be of much help to John. For I had already started talking with Francis. Perhaps even without Francis, I would have quickly bored of myoglobin. But with Francis to talk to, my fate was sealed. For we quickly discovered that we thought the same way about biology. The center of biology was the gene and its control of cellular metabolism. The main challenge in biology was to understand gene replication and the way in which genes control protein synthesis. It was obvious that these problems could be logically attacked only when the structure of the gene became known. This meant solving the structure of DNA. Then this objective seemed out o£ reach to the interested geneticists. But in our cold, dark Cavendish lab, we thought the job could be done, quite possibly within a few months. Our optimism was partly based on Linus Pauling's feat1 in deducing the a-helix, largely by following the rules of theoretical chemistry so persuasively explained in his classical The Nature of the Chemical Bond. We also knew that Maurice Wilkins had crystalline X-ray diffraction photo- graphs from DNA and so it must have a well-defined structure. There was thus an answer for somebody to get. During the next eighteen months, until the double-helical structure be- came elucidated, we frequently discussed the necessity that the correct struc- ture have the capacity for self-replication. And in pessimistic moods, we of- ten worried that the correct structure might be dull. That is, it would s-111 [Q62 |. 1). WATSON suggest absolutely nothing and excite us no more than something inert like collagen. The finding of the double helix2 thus brought us not only joy but great relief. It was unbelievably interesting and immediately allowed us to make a serious proposal3 for the mechanism of gene duplication. Furthermore, this replication scheme involved thoroughly understood conventional chemical forces. Previously, some theoretical physicists, among them Pascual Jordan4, had proposed that many biological phenomena, particularly gene replica- tion, might be based on still undiscovered long-range forces arising from quantum mechanical resonance interactions. Pauling5 thoroughly disliked this conjecture and firmly insisted that known short-range forces between complementary surfaces would be the basis of biological replication. The establishment of the DNA structure reinforced our belief that Pau- ling's arguments were sound and that long-range forces, or for that matter any form of mysticism, would not be involved in protein synthesis. But for the protein replication problem mere inspection of the DNA structure then gave no immediate bonus. This, however, did not worry us since there was much speculation that RNA, not DNA, was involved in protein synthesis. Introduction The notion that RNA is involved in protein synthesis goes back over twenty years to the pioneering experiments of Brachet and Caspersson6 who showed that cells actively synthesizing protein are rich in RNA. Later when radio- active amino acids became available, this conjecture was strengthened by the observation7 that the cellular site of protein synthesis is the microsomal component, composed in large part of spherical particles rich in RNA. Still later experiments8 revealed that these ribonuclcoprotcin particles (now con- veniently called ribosomes), not the lipoprotein membranes to which they arc often attached, are the sites where polypeptide bonds are made. Most ribosomes arc found in the cytoplasm and correspondingly most cellular protein synthesis occurs without the direct intervention of the nuclear-lo- cated DNA. The possibility was thus raised that the genetic specificity pres- ent in DNA is first transferred to RNA intermediates which then function as templates controlling assembly of specific amino acids into proteins. We became able to state this hypothesis in more precise form when the structure of DNA became known in 1953. We then realized that DNA's s-112 RNA IN THE SYNTHESIS OF PROTEINS genetic specificity resides in the complementary base sequences along its two intertwined chains. One or both of these complementary chains must serve as templates for specific RNA molecules whose genetic information again must reside in specific base sequences. These RNA molecules would then assume 3 -dimensional configurations containing surfaces complementary to the side groups of the 20 specific amino acids. X-ray Studies on RNA and RNA-contaiuiug Viruses The direct way to test this hypothesis was to solve the RNA structure. Al- ready in 1952, I had taken some preliminary X-ray diffraction pictures of RNA. These, however, were very diffuse, and it was not until I returned to the United States in the fall of 1953 that serious X-ray studies on RNA began. Alexander Rich and I, then both at the California Institute of Technology, obtained RNA samples from various cellular sources. We9 were first very encouraged that all the RNA samples, no matter their cellular origin, give similar X-ray diffraction pattern. A general RNA structure thus existed. This gave us hope that the structure, when solved, would be interesting. Our first pictures already showed large systematic absence of reflections on the merid- ian, suggesting a helical structure. But despite much effort to obtain native undegraded high molecular weight samples, no satisfactory X-ray diffraction pattern was obtained. The reflections were always diffuse, no evidence of crystallinity was seen. Though there were marked similarities to the DNA pattern, we had no solid grounds for believing that these arose from a similar helical molecule. The problem whether RNA was a one- or several-chained structure remained unanswered. We then considered the possibility that RNA might have a regular struc- ture only when combined with protein. At that time (1955) there was no good evidence for RNA existing free from protein. All RNA was thought to exist either as a viral component or to be combined with protein in ribo- nucleoprotein particles. It thus seemed logical to turn attention to a study of ribonucleoprotein particles (ribosomes) since upon their surfaces protein was synthesized. Our hope again was that the establishment of their structure would reveal the long-sought-after cavities specific for the amino acids. Then we were struck by the morphological similarity between ribosomes and small RNA-containing viruses like Turnip Yellow Mosaic Virus or Poliomyelitis Virus. By then (1955-1956) I was back in Cambridge with s-113 [962 J. D.WATSON Crick to finish formulating some general principles on viral structure10. Our main idea was that the finite nucleic acid content of viruses severely restrict- ed the number of amino acids they could code for. As a consequence, the protein coat could not be constructed from a very large number of different protein molecules. Instead it must be constructed from a number of identical small sub-units arranged in a regular manner. These ideas already held for Tobaa 0 Mosaic Virus, a rod-shaped virus, and we were very pleased when I ). I . 1 ). C Jaspar1 ', then working with us at the Cavendish, took some elegant diffraction pictures o\' Bushy Stunt Virus crystals and extended experimental support to the spherical viruses. Structural Studies on Ribosomes At that time almost no structural studies had been done with ribosomes. They were chiefly characterized by their sedimentation constants; those from higher organisms12 in the 70S-80S range, while those from bacteria13 appeared smaller and to be of two sizes (30s and 50s). Because the bacterial particles seemed smaller, they seemed preferable for structural studies. Thus when Alfred Tissieres and I came to Harvard's Biological Laboratories in 1956, we initiated research on the ribosomes of the commonly studied bacteria Esche- richia coli. We hoped that their structure would show similarities with the small spherical RNA viruses. Then we might have a good chance to crys- tallize them and to eventually use X-ray diffraction techniques to establish their 3 -dimensional structure. Ribosome sub-units But from the beginning of our Harvard experiments, it was obvious that ribosome structure would be more complicated than RNA virus structure. 1 Vpending upon the concentration of divalent cations (in all our experi- ments Mg •"), 4 classes of E. coli ribosomes were found, characterized by sedimentation constants of 30s, 50s, 70s, and 100s. Our first experiments in io~4 M Mg revealed 30s and 50s ribosomes. At the same time Bolton14, at the Carnegie Institute of Washington employing higher Mg1-4" levels, saw faster sedimenting ribosomes and suggested that they were observing ag- gregates of the smaller particles. Soon after, our experiments15 revealed that, as the Mg++ concentration is raised, one 30s particle and one 50s particle s-114 RNA IN THE SYNTHESIS OF PROTEINS 0 0 CD 2(30S) 2(50S) 2(70S) K100S) M.W. x 10 0.85±0.15 1.80 ±0.15 2.8 ± 02 5.9 ± 1.0 All particles are composed of 64 °/o RNA and 36°/o protein Fig. I. Diagrammatic representation of E. coli ribosomc sub-units and their aggrega- tion products. (The molecular weight data are from Tissieres et al.ls) Fig. 2. Electron micrograph of negatively stained E. coli ribosomes (Huxley and Zu- bay15). Two particle types are predominant: (i) 70s containing two sub-units of unequal size, and (2) 100s consisting of two 70s ribosomes joined together at their smaller (30s) sub-units. combine to form a 70s ribosomc. At still higher Mg++ concentrations, two 70s ribosomes dimerize to form a 100s ribosomc. (Figs. 1 and 2). Ribosomes from every cellular source have a similar sub-unit construction. As with E. coli ribosomes, the level of divalent cations determines which ribosomes exist. Bacterial ribosomes seem to require higher Mg++ levels in order to aggregate into the larger sizes. Conversely they break down much s-115 196- J- D. WATSON faster to the 30s and 50s forms when the Mg' level is lowered. It is often convenient16 when using mammalian ribosomes to add a chelating agent to rapidly break down the Sos ribosomes (homologous to the 70s ribosomes of bacteria) to their 40s and 60s sub-units. Bacterial ribosomes are thus not significantly smaller than mammalian ribosomes. It is merely easier to ob- serve the smaller sub-units in bacterial systems. Ribosomal RNA Already in 1958 there were several reports17 that ribosomal RNA from higher organisms sedimented as two distinct components (18s and 28s). We thought that the smaller molecules most likely arose from the smaller sub- unit while the faster sedimenting RNA came from the larger of the ribo- somal sub-units. Experiments of Mr. Kurland18 quickly confirmed this hunch. The E. cell 30s ribosome was found to contain one RNA chain (16s) with a molecular weight of 5.5 X io5. Correspondingly a larger RNA mol- ecule (23s) of mol. wt. 1.1 X io6 was found in most 50s ribosomes (Fig. 3). Ribosome proteins Analysis of the protein component revealed a much more complicated pic- ture. In contrast to the small RNA viruses, where the protein coat is con- structed from the regular arrangement of a large number of identical protein molecules, each ribosome most likely contains a large number of different polypeptide chains. At first, our results suggested a simple answer when Drs. Waller and J. I. Harris analysed E. coli ribosomes for their amino terminal groups. Only alanine, methionine, with smaller amounts of serine, were present in significant amounts. This hinted that only several classes of protein molecules were used for ribosomal construction. Further experiments of Dr. Waller19, however, suggested the contrary. When ribosomal protein frac- tions were analysed in starch-gel electrophoresis, more than 20 distinct bands were seen. Almost all these proteins migrated towards the anode at pH 7 confirming the net basic charge of ribosomal protein20. A variety of control experiments suggested that these bands represent distinct polypeptide chains, not merely aggregated states of several fundamental sub-units. Moreover, the band pattern from 30s ribosomes was radically different from that of 50s proteins. As yet we have no solid proof that each 70s ribosome contains all the s-116 RNA IN THE SYNTHESIS OF PROTEINS various protein components found in the total population. But so far, all attempts by Dr. Waller to separate chromatographically intact ribosomes into fractions with different starch-gel patterns have failed. The total protein component of a 70s ribosome amounts to about 9 X 105 daltons. Since the end group analysis suggests an average mol. wt. of about 30,000, approx- imately 20 polypeptide chains are used in 50s construction and 10 for the 3 os ribosome. It is possible that all the polypeptide chains in a 30s particle are different. Waller already has evidence for 10 distinct components in 30s ribosomes and the present failure to observe more in the 50s protein fraction may merely mean that the same electrophoretic mobility is shared by several polypeptide chains. We believe that all these proteins have primarily a structural role. That is, they are not enzymes but largely function to hold the ribosomal RNA and necessary intermediates in the correct position for peptide bond formation. In addition a number of enzymes are bound tightly to ribosomes. As yet their function is unclear. One such is a bacterial ribonuclease, found by Elson21 to be specifically attached to 30s ribosomes in a latent form. No ribonuclease activity is present until ribosome breakdown. Dr. Spahr22 in our laboratories has purified this enzyme, shown its specificity and from specific activity measurements, concludes that it is present on less than one in twenty Ribosomes 30 S 50 S M.W. x 10"6 085 i 0.15 1.80 ± 0.15 RNA MW x 10 16 S 0.55 ± 0.10 2(16S) Fig. 3. Molecular weights of RNA isolated from E. coli ribosomes. (This picture is diagrammatic and does not represent the true conformation of ribosomal RNA.) s-117 [962 J.D.WATSON 30s particles. It is clear that this enzyme it present in .1 free active form, would be rapidly lcth.il to its host cell. Thus its presence in latent form is expected. Hut why it is stuck to ribosomes is still a complete mystery. Chemical Intermediates in Protein Synthesis Our early experiments with ribosomes were almost unrelated to the efforts of biochemists. At that time our research objects seemed very different. The enzymologically oriented biochemists hoped to find the intermediates and enzymes necessary for peptide bond formation. On the contrary, those of us with a genetic orientation wanted to see the template and discover how it picked out the correct amino acid. Very soon, however, these separate paths came together, partly because of a breakthrough in the nature of the ammo acid intermediates, and partly from an incisive thought by Crick. The biochemical advances arose from work in Paul Zamccnik's laboratory at the Massachusetts General Hospital. There was developed a reproducible in vitro system23 containing ribosomes, supernatant factors, and ATP which incorporated amino acids into protein. Using these systems Hoagland made two important discoveries. Firstly, he24 showed that amino acids are ini- tially activated by ATP to form high-energy AA-AMP complexes. Sec- ondly, he demonstrated25 that the activated amino acids arc then transferred to low molecular weight RNA molecules (now known as soluble or transfer RNA), again in an activated form. These amino-acyl-sRNA compounds then function as the direct intermediate for peptide bond formation (Fig. 4). It had previously been obvious that amino acid activation would have to occur. However, Hoagland's second discovery (in 1956) of the involvement of a hitherto undiscovered RNA form (sRNA) was unanticipated by almost everybody. Several years previously (in 1954), Leslie Orgel and I spent a quite frustrating fall attempting to construct hypothetical RNA structures which contained cavities complementary in shape to the amino acid side groups. Not only did plausible configurations for the RNA backbone fail to result in good cavities, but even when we disregarded the backbone, we also failed to find convincing holes which might effectively distinguish between such amino acids as valine and isolcucinc. Crick, at the same time (early 1955) sensed the same dilemma, and suggested a radical solution to the paradox. He proposed26 that the amino acids do not combine with the tem- plate. Instead each should first combine with a specific adaptor molecule, s-118 R N A IN THE SYNTHESIS OF PROTEINS (a) AA +■ ATP — AMP ~ AA + PP (b) AMP~ AA + SRNA — AA~ SRNA 4- AMP (c) (AA ~ SRNA)n ■+- GTP Ri bo somes AA,-AA2 AAn + GDP (GMP') Fig. 4. Enzymatic steps in protein peptide bond formation. Steps (a) and (b) are catalyzed by single enzymes. The number of enzymes required in (c) is unknown. capable of selectively interacting with the hydrogen bonding surfaces pro- vided by RNA's purine and pyrimidine bases. This scheme requires at least twenty different adaptors, each specific for a given amino acid. These are very neatly provided by the specific sRNA molecules. Soon after Hoag- land's discovery of sRNA, many experiments, particularly by Hoagland and Paul Berg27, established that the sRNA molecules are in fact specific for a given amino acid. It thus became possible to imagine, following Crick's reasoning, that the ribosomal template for protein synthesis combined not with the amino acid side groups, but instead with a specific group of bases on the soluble RNA portion of the amino-acyl-sRN A precursors. Participation of Active Ribosomcs in Protein Synthesis Very little protein synthesis occurred in the cell-free system developed by the Massachusetts General Hospital Group. Only by using radioactive amino acids could they convincingly demonstrate amino acid incorporation into proteins. This fact, initially seemed trivial and there was much hope that when better experimental conditions were found, significant net synthesis would occur. But despite optimistic claims from several laboratories, no real improvement in the efficiency of cell-free synthesis resulted. Some exper- iments (1959) of Dr. Tissieres and Mr. Schlessinger28 with E. coli extracts illustrate well this point. At 30°C, cell-free synthesis occurs linearly for 5-10 minutes and then gradually stops. During this interval the newly synthesized protein amounts to 1-3 y of protein per mg of ribosomcs. Of this about one third was released from the ribosomes, the remainder being ribosomal bound. Cell-free synthesis in E. coli extracts requires the high (~ io~2 M) Mg++ s-119 [962 J.D. W A 1 SON 1 v( Is which favor the formation of 70s ribosomes from their 30s and 50s sub-units. I ollowing incorporation, those ribosomes possessing nascent poly- peptide chains become less susceptible to breakdown to 30s and 50s ribo- somes. When cell-free extracts (following synthesis) are briefly dialyzed against 10 4 M Mg , about No-90% of the 30s and 50s ribosomes become tree. There remain, however, 10-20% of the original 70s ribosomes and it is upon these i stuck • ribosomes that most ribosomal bound nascent protein is located. This firstly suggests that protein synthesis occurs on 70s ribosomes, not upon free 30s or 50s ribosomes. Secondly, in the commonly studied E. coli extract, only a small ribosomc fraction is functional, Tissieres and Schles- singer named these particles « active ribosomes » and suggested, they con- tained a functional component lacking in other ribosomes. Each active nbosome synthesizes on the average between 15,000 and 50,000 daltons of protein. This is in the size range of naturally occurring polypeptide chains. Thus while we remained unsatisfied by the small net synthesis, sufficient synthesis occurs to open the possibility that some com- plete protein molecules are made. This encouraged us to look for synthesis of /?-galactosidase. None, however, was then found29 despite much effort. Another important point emerged from these early (1959) incorporation studies with E. coli extracts. Addition of small amounts of purified dcoxy- ribonuclcase decreased protein synthesis to values 20-40% that found in untreated extracts28. This was completely unanticipated, for it suggested that high molecular weight DNA functions in the commonly studied bacte- rial extracts. But since a basal synthetic level occurs after DNA is destroyed by deoxyribonuclcase, the DNA itself must not be directly involved in peptide bond formation. Instead, this suggested synthesis of new template RNA upon DNA in untreated extracts. If true, this would raise the possibil- ity, previously not seriously considered by biochemists that the RNA tem- plates themselves might be unstable, and hence a limiting factor in cell-free protein synthesis. Metabolic Stability of Ribosomal RNA All our early ribosomc experiments had assumed that the ribosomal RNA was the template. Abundant evidence existed that proteins were synthesized on ribosomes and since the template must be RNA, it was natural to assume that it was ribosomal RNA. Under this hypothesis ribosomal RNA was a s-120 RNA IN THE SYNTHESIS OF PROTEINS collection of molecules of different base sequences, synthesized on the func- tioning regions of chromosomal DNA. Following their synthesis, they combined with the basic ribosomal proteins to form ribosomes. We thus visualized that the seemingly morphological identical ribosomes were, in fact, a collection of a very large number of genetically distinct particles masked by the similarity of their protein component. Then there existed much suggestive evidence that ribosomal RNA mol- ecules were stable in growing bacteria. As early as 1949, experiments showed that RNA precursors, once incorporated into RNA, remained in RNA. Then the distinction between ribosomal and soluble RNA was not known, but later experiments by the ribosomc group of the Carnegie Institute of Washington and at Harvard indicated similar stabilities of both fractions. These experiments, however, did not follow the fate of single molecules, and the possibility remained that a special trick allowed ribosomal RNA chains to be broken down to fragments that were preferentially re-used to make new ribosomal RNA molecules. Davern and Mcselson30, however, ruled out this possibility by growing ribosomal RNA in heavy (I3C,I5N) medium, followed by several generations of growth in light (I2C, I4N) medium. They then separated light from heavy ribosomal RNA in cesium formate density gradients and showed that the heavy molecules remained completely intact for at least two generations. This result predicts, assuming ribosomes to be genetically specific, that the protein templates should persist indefinitely in growing bacteria. Experiments Suggesting Unstable Protein Templates But already by the time of the Davern 6V Meselson experiment (i959)> evidence began to accumulate, chiefly at the Institut Pasteur, that some, if not all, bacterial templates were unstable with lives only several per cent of a generation time. None of these experiments, by themselves, were con- vincing. Each could be interpreted in other ways which retained the concept of stable templates. But taken together, they argued a strong case. These experiments were of several types. One studied the effect of sud- denly adding or destroying specific DNA molecules. Sudden introduction was achieved by having a male donor introduce a specific chromosomal region absent in the recipient female. Simultaneously the ability of the male gene to function (produce an cnzymatically active protein) in the female cell s-121 I 9 f. 2 |. D. W ATS ON was measured. Riley, Pardee, Jacob, and Monod31 obtained the striking finding that 0-galaCtOsidase, genetically determined by a specific male gene, began to be synthesized at its maximum rate within several minutes after gene transfer. Thus the steady state number of /^-galactosidase templates was achieved almost immediately. Conversely when the /:. colt chromosome was inactivated bv decay of32P atoms incorporated into DNA, they observed that active enzyme formation stops within several minutes. It thus appeared that the ribosomal templates could not function without concomitant DNA function. At the same time, Francois Gros discovered32 that bacteria grown in 5- fluorouracil produced abnormal proteins, most likely altered in amino acid sequences. s-Fluorouracil is readily incorporated into bacterial RNA and its presence m RNA templates may drastically raise the mistake level. More unexpected was the observation that following 5-fluorouracil addition the production of all normal proteins ceases within several minutes. Again this argues against the persistance of any stable templates. Unstable RNA Molecules in Phage Infected Cells At first it was thought that no RNA synthesis occurred in T2 infected cells. But in 1952 Hershcy33 observed that new RNA molecules are synthesized at a rapid rate. But no net accumulation occurs since there is a correspondingly last breakdown. Surprisingly almost everybody ignored this discovery. This oversight was partly due to the tendency, still then prevalent, to suspect that the metabolism of virus infected cells might be qualitatively different from that of uninfected cells. Volkm and Astrachan34 were the first (1956) to treat Hershcy's unstable fraction seriously. They measured its base composition and found it different from that of uninfected E. coli cells. It bore a great resemblance to the infect- ing viral DNA which suggested that it was synthesized on T2 DNA tem- plates. Moreover, and most importantly, this RNA fraction must be the tem- plate for phage specific proteins. Unless we assume that RNA is not involved in phage protein synthesis, it necessarily follows that the Volkin-Astrachan DNA-like RNA provides the information for determining amino acid se- quences in phage specific proteins. Not till the late summer of 1959 was its physical form investigated. Then Nomura, Hall, and Spicgelman35 examined its relationship to the already s-122 RNA IN THE SYNTHESIS OF PROTEINS characterized soluble and ribosomal RNA's. Immediately they observed that none of the T2 RNA was incorporated into stable ribosomes. Instead, in low Mg++ ( io-4 M) it existed free while in io~2 MMg++ they thought it became part of 30s ribosomal like particles. At the same time, Mr. Risebrough in our laboratories began studying T2 RNA, also using sucrose gradient centrifuga- tion. He also found that T2 RNA was not typical ribosomal RNA. In addi- tion, he was the first to notice (in early spring i960) that in io~2 M Mg++, most T2 RNA scdimented not with 30s particles but with the larger 70s and ioos ribosomes. His result leads naturally to the hypothesis that phage protein synthesis takes place on genetically non-specific ribosomes to which are attached metabolically unstable template RNA molecules. Independently of our work, Brenner and Jacob motivated by the above-mentioned metabolic and genetic experiments from the Institut Pasteur, were equally convinced that condi- tions were ripe for the direct demonstration of metabolically unstable RNA templates to which Jacob and Monod36 gave the name messenger RNA. In June of i960, they travelled to Pasadena for a crucial experiment in Mesel- son's laboratory. They argued that all the T2 messenger RNA should be attached to old ribosomes synthesized before infection. This they elegantly demonstrated37 by T2 infecting heavy (I3C and I5N) labeled bacteria in light (I2C and I4N) medium. Subsequent CsCl equilibrium centrifugation revealed that most of the T2 messenger RNA was indeed attached to « old » ribosomes, as was all the ribosomal bound nascent protein, labeled by pulse exposure to radioactive amino acids. Demonstration of Messenger RNA Molecules in Uninfected Bacteria We were equally convinced that similar messenger RNA would be found in uninfected bacteria. Its demonstration then presented greater problems, be- cause of the simultaneous synthesis of ribosomal and soluble RNA. Francois Gros had then (May i960) just arrived for a visit to our laboratory. Together with Mr. Kurland and Dr. Gilbert, we decided to look for labeled messenger molecules in cells briefly exposed to a radioactive RNA precursor. Exper- iments with T2 infected cells suggested that the T2 messenger comprised about 2-4% of the total RNA and that most of its molecules had lives less than several minutes. If a similar situation, held for uninfected cells, then during any short interval, most RNA synthesis would be messenger. There s-123 [962 |. I). WATSON would be no significant accumulation since it would be broken down almost .is tast .is it was made. Again the messenger hypothesis was confirmed38. The RNA labeled dur- ing pulse exposures was largely attached to 70s and 100s ribosomes in 10 2 M Mg . In low Mg (10 * M), it came offthe ribosomes and sedimented free with an average sedimentation constant of 14s. Base ratio analysis revealed |)NA like RNA molecules in agreement with the expectation that it was produced on very many DNA templates along the bacterial chromosome. Soon afterwards, Hall and Spiegelman39 formed artificial T2 DNA; T2 messenger RNA hybrid molecules and in several laboratories40, hybrid mol- ecules were subsequently formed between E. coli DNA and E. coli pulse RNA. The DNA template origin for messenger RNA was thus established beyond doubt. The Role of Messenger RNA in Cell-Free Protein Synthesis It was then possible to suggest why deoxyribonuclcasc partially inhibits amino acid incorporation in E. coli extracts. The messenger hypothesis prompts the idea that DNA in the extract is a template for messenger RNA. This newly made messenger then attaches to ribosomes where it serves as additional protein templates. Since deoxyribonuclcasc only destroys the ca- pacity to make messenger, it has no effect upon the messenger present at the time of extract formation. Hence, no matter how high the deoxyribonu- clcasc concentration employed, a residual fraction of synthesis will always occur. Experiments by Tissieres and Hopkins41 in our laboratories and by Berg, Chamberlain, and Wood42 at Stanford confirmed these ideas. First it was shown that addition of DNA to extracts previously denuded of DNA significantly increased amino acid incorporation. Secondly, RNA synthesis occurs simultaneously with in vitro protein synthesis. This RNA has a DNA like composition, attaches to ribosomes in io~2 M Mg++, and physically resembles in vivo synthesized messenger RNA. Furthermore, Tissieres showed that addition of fractions rich in messenger RNA stimulated in vitro protein synthesis 2-5 fold. More striking results came from Nircnbcrg and Matthaei43. They reasoned that in vitro messenger destruction might be the principal cause why cell-free systems stopped syn- thesizing protein. If so, preincubated extracts deficient in natural messenger should respond more to new messenger addition. This way they became s-124 RNA IN THE SYNTHESIS OF PROTEINS able to demonstrate a 20-fold increase in protein synthesis following ad- dition of phenol-purified E. coli RNA. Like Tissieres' active fraction, their stimulating fraction scdimented heterogencously arguing against an effect due to cither ribosomal or soluble RNA. More convincing support came when they next added TMV RNA to prcincubated E. coli extracts. Again a 10-20 fold stimulation occurred. Here there could be no confusion with possible ribosomal RNA templates. Even more dramatic44 was the effect of polyuridylic acid (like TMV RNA single stranded) addition. This specifi- cally directed the incorporation of phenylalanine into polyphenylalanine. With this experiment (June 1961) the messenger concept became a fact. Direct proof then existed that single stranded messenger was the protein template. Presence of Messenger RNA in Active Ribosonies In in vitro systems ordinarily only 10-20% of E. coli ribosomes contain attached messenger RNA. This first was shown in experiments of Rise- brough45 who centrifuged extracts of T2 infected cells through a sucrose gradient. Ribosomes containing labeled messenger were found to centrifuge faster than ordinary ribosomes. Similarly, Gilbert46 showed that these faster scdimenting ribosomes are « active », that is, able to incorporate amino acids into proteins. A fresh cell-free extract was centrifuged through a sucrose gradient. Samples along the gradient were collected and then tested for their ability to make protein. A complete parallel was found between « activity » and the presence of messenger. Furthermore, if an extract is centrifuged after it has incorporated amino acids, the nascent protein chains also sediment attached to a small fraction of fast sedimenting ribosomes45. These ribosomes still contain messenger RNA. For when the messenger molecules are destroyed by ribonucleasc (ribosomes remain intact in the presence of y amounts of ribonuclease), the ribosomal bound nascent protein sediments as 70s ribosomes. The nascent protein is thus not attached to messenger RNA but must be directly bound to ribosomes. Binding ofsRNA to Ribosomes Experiments by Schweet47 and Dintzes48 show that proteins grow by step- wise addition of individual amino acids beginning at the amino terminal end. s-125 I i)() 2 |. 1). WAT SO N \ / \ / (0) N— AA, AA2 AA3 cf — "N AA4 C i-r Oh o SRNAa SRNAb H ,° \ / (b) \ AA, AA2 AA3 AA4 C5 + SRNAa SRNAb \ ig. s. Stepwise growth of a polypeptide chain. Initiation begins at the free NH2 end with the growing point terminated by a sRNA molecule. Since the immediate precursors arc amino-acyl-sRNA molecules, their re- sult predicts that the polypeptide chain is terminated at its carboxyl growing end by an sRNA molecule (Fig. 5). To test this scheme, we began some studies to see whether sRNA bound specifically to ribosomes. Cannon and Rjrug49 tirst examined binding in the absence of protein synthesis. They showed that in 10 2 MMg+ f each 50s sub-unit of the 70s ribosome revcrsibly bound one sRNA molecule. The same amount of reversible binding occurs with amino-acyl-sRNA or with free sRNA and in the presence or absence of protein synthesis. Protein synthesis, however, effects the binding observed in io-4 M Mg4_+. In the absence of protein synthesis no sRNA remains ribosomal bound when the Mgf+ level is lowered from 10 2 M to io~4 M. On the contrary, fol- lowing amino acid incorporation, sRNA molecules become tightly fixed to the « stuck » 70s ribosomes, whose nascent polypeptide chains prevent easy dissociation to 30s and 50s ribosomes. One sRNA molecule appears to be attached to each stuck ribosome. Prolonged dialysis against io~4 M Mg++ eventually breaks apart the stuck ribosomes. Then all the bound sRNA as well as almost all the nascent protein is seen attached to the 50s component supporting the hypothesis that these bound sRNA molecules arc directly attached to nascent chains (Fig. 6). Direct proof comes from recent exper- iments in which Gilbert50 used the detergent duponol to further dissociate the 50s ribosomes to their protein and RNA components. Then the nascent protein and bound sRNA remained together during both sucrose gradient ccntrifugation and separation on G200 Sephadcx columns. Following ex- s-126 RNA IN THE SYNTHESIS OF PROTEINS posurc, however, to either weak alkali or to hydroxylamine, treatments known to break amino-acyl-bonds, the sRNA and nascent proteins move separately. The significance of the reversible binding by non-active (no messenger) ribosomcs is not known. Conceivably inside growing cells, all ribosomes have attached messenger and synthesize protein. Under these conditions, only those sRNA molecules corresponding to the specific messenger se- quence can slip into the ribosomal cavities. But when most ribosomes lack messenger templates, as in our in vitro extracts, then any sRNA molecule, charged or uncharged, may fill the empty site. All evidence suggests that covalent bonds are not involved in holding nascent chains to ribosome. Instead it seems probable that the point of firm attachment involves the terminal sRNA residue, bound by Mg++ dependent 70S \ Messenger RNA (poly U) Active complex n x 70 S + M -RNA + S- RNA-aa. 100 S — 200 S S-RNA Polypeptide chain + + \ S- RNA-aa. cCQXQ- SDS S-RNA poly <{>-ala Fig. 6. Diagrammatic summary of ribosome participation in protein synthesis. (The active complex is pictured in Fig. 7.) s-127 [962 J. D. W A I SON mdan forces to a cavit) in the 50s ribosome. Extensive dialysis against 5 to-5 M Mg (which leaves intact 30s and 50s ribosomes) strips the n.is,\nt chains off the 50s ribosomes50*51. The released polypeptides sediment about 4s and it" the latent rihonuclcasc is not activated, most likely still have terminally bound sKNA. When the Mg1 ! level is again brought to io~2 M main released chains again stick to ribosomes. Movement of the Messenger Template over the Ribosomal Surface At any given time, each functioning ribosome thus contains only one nas- cent chain. As elongation proceeds, the NH3-tcrminal end moves away from the point of peptide bond formation and conceivably may assume much of its final three-dimensional configuration before the terminal amino acids are added to the carboxyl end. The messenger RNA must be so attached that only the correct amino-acyl-sRNA molecules are inserted into position for possible peptide bond formation. This demands formation of specific hy- drogen bonds (base-pairs?) between the messenger template and several (most likelv three) nucleotides along the sRNA molecule. Then, in the pres- ence o( the necessary enzymes, the amino-acyl linkage to the then terminal sRNA breaks and a peptide bond forms with the correctly placed incoming amino-acyl-RNA (Fig. 5). This must create an energetically unfavorable environment for the now free sRNA molecule, causing it to be ejected from the sRNA binding site. The new terminal sRNA then moves into this site completing a cycle of synthesis. It is not known whether the messenger tem- plate remains attached to the newly inserted amino-acyl-sRNA. But if so, the messenger necessarily moves the correct distance over the ribosomal sur- face to place its next group of specific nucleotides in position to correctly select the next amino acid. No matter, however, what the mechanism is, the messenger tape necessarily moves over the ribosome. They cannot remain in static orientation if there is only one specific ribosomal site for peptide bond formation. Attachment of Single Messenger RNA Molecules to Several Ribosomes Addition of the synthetic messenger poly U to extracts containing pre- dominantly 70s ribosomes creates new active ribosomes which sediment in s-128 RNA IN THE SYNTHESIS OF PROTEINS S-RNA Growing polypeptide chain Messenger RNA Fig. 7. Messenger RNA attachment to several ribosomes. (This illustration is schemat- ic since the site of messenger attachment to ribosomes is not known.) the 150-200S region52. Fixation of a single poly U molecule (mol. wt. = 100,000) to a 70s ribosome (mol. wt. = 3 X io6) should not significantly increase ribosomal sedimentation. Nor is it likely that a very large number of poly U molecules have combined with individual ribosomes. In these experiments, the molar ratio of fixed poly U to 70s ribosomes was less than }. Instead, the only plausible explanation involves formation of ribosomal aggregates attached to single poly U molecules. The 300 nucleotides in a poly U molecule of mol. wt. ~ io5 will have a contour length of about 1000 A if the average internucleotide distance is 3.4 A. Simultaneous attachment is thus possible to groups of 4-8 ribosomes (diameter ~ 200 A) depending upon the way the messenger passes over (through) the ribosomal surface. This estimate agrees well with the average aggregate size suggested by the sedimentation rate of the « active » complexes. Sedimentation of extracts after incorporation reveals most polyphenylalanine attached to the rapidly sed- imenting « active » ribosomes. Single messenger molecules thus most likely move simultaneously over the surfaces of several ribosomes, functioning on each as protein templates (Fig. 7). A progression of increasingly long polypeptide chains should be attached to successive ribosomes depending upon the fraction of the mes- senger tape to which they were exposed. When all the messenger has moved across the site of synthesis, some mechanism, perhaps itself triggered by a s-129 [962 |. 1). W A I SON specific template nucleotide sequence must release the finished protein. The now vacant ribosome then becomes competent to receive the free end of another (or perhaps even the same) messenger molecule and start a new cycle ot protein synthesis. The realization that a single messenger molecule attaches to many ribo- SOmes resolves a bothersome paradox which accompanied the messenger hypothesis. About 2-4% of E. coli RNA is messenger40-53. Its average sed- imentation constant ot 14s54 suggests an average molecular weight about 500,000. This value may be too low since it is very difficult to completely prevent all enzymatic degradation. There thus must be at least 6-8 70s ribo- somes for every messenger molecule. It was very difficult to believe that only 10-20% of the ribosomes functioned at a given moment. For, under a variety of conditions, the rate of protein synthesis is proportional to ribo- some concentration55. Instead, it seems much more likely that, in vivo, almost all ribosomes are active. During the preparation of cell extracts, however, many ribosomes may lose their messenger and become inactive. If true, we may expect that use of more gentle techniques to break open E. coli cells will reveal larger fractions of fast-scdimenting active material. Already there are reports56 that over 50% of mammalian reticulocyte ribosomes exist as ag- gregates of 5-6 80s particles. Furthermore, it is these aggregated ribosomes which make protein, both /'// vivo and in vitro. Template Lifetime Under the above scheme a messenger molecule might function indefinitely. On the contrary, however, the unstable bacterial templates function on the average only 10-20 times. This fact comes from experiments done in Levin- thal's laboratory57 where new messenger synthesis was blocked by addition of the antibiotic antinomycin D. Preexisting messenger (Bacillus subtilus growing with a 60 minute generation time) then broke down with a half- life of 2 minutes. Correspondingly, protein synthesis ceased at the expected rate. A mcchanism(s) must thus exist to specifically degrade messenger mol- ecules. Several enzymes (polynucleotide phosphorylasc and a K+ dependent diesterase) which rapidly degrade free messenger are active in bacterial cell extracts58. They function, however, much less efficiently when the messenger is attached to ribosomes59. Conceivably, a random choice exists whether the free forward-moving end of a messenger tape attaches to a vacant ribosome, s-I30 RNA IN THE SYNTHESIS OF PROTEINS or is cnzymatically degraded. If so, this important decision is settled by a chance event unrelated to the biological need for specific messengers. Conclusion We can now have considerable confidence that the broad features of protein synthesis are understood. RN A's involvement is very much more complicated than imagined in 1953. There is not one functional RNA. Instead, protein synthesis demands the ordered interaction of three classes of RNA - ribo- somal, soluble, and messenger. Many important aspects, however, remain unanswered. For instance, there is no theoretical framework for the riboso- mal sub-units nor, for that matter, do we understand the functional signif- icance of ribosomal RNA. Most satisfying is the realization that all the steps in protein replication will be shown to involve well-understood chemical forces. As yet we do not know all the details. For example, are the DNA base-pairs involved in messenger RNA selection of the corresponding amino- acyl-sRNA? With luck, this will soon be known. We should thus have every expectation that future progress in understanding selective protein synthesis (and its consequences for embryology) will have a similar well- defined and, when understood, easy-to-comprchend chemical basis. Acknowledgment I have been very fortunate in having the collaboration of many able students and colleagues. The Ph.D. thesis work of Dr. C. G. Kurland, Dr. David Schlessingcr and Dr. Robert Riscbrough established many ideas reported here. Equally significant have been experiments by Drs. Kimiko Asano, Michael Cannon, Walter Gilbert, Francois Gros, Francoisc Gros, Johns Hopkins, Masayasu Nomura, Pierre Francois Spahr, Alfred Tissiercs, and Jean-Pierre Waller. The visit of Francois Gros in the spring of i960 was crucial in focusing attention on messenger RNA. Most importantly, I wish to mention my lengthy and still continuing successful collaboration with Alfred Tissieres. Since i960, I have the good fortune to also work closely with Walter Gilbert. s-131 [962 I. D.VL \ 1 s ' I N i.l. Pauling and R. B. Corey, Proc. Natl. Acad. Sci. U.S., 37 ( ">>> 1 ) 235. I D Watson and F. H. C. Crick, Nature, 171 (1953) 737- 3. J. D. Watson and 1 . 11. C. Crick, Nature, [7] I [953) 9^4- , P. [ordan, Physik. '/... 39 (1938 711. 1 he reader is also referred to the disc ussion of possible implications ol long-range forces in biology, by II. J. Muller in Ins [946 Pilgrim Trust Lecture, Proc. Roy. London, B ( 1947). 5. A sample of Pauling's views is found in Ins note with M. Delbriick, Science, 92 •40) 77- 6. I. Brachet, Arch. Biol. Liege, 53 (1942)207; T. Caspersson, Naturwiss., 29(1941) 7. 11. Borsook, C. 1 . Deasy, A. J. Haagen-Smit, G. Keighley, and P. H. Lowy, /. Biol. Chan., 187 ( [950) 839; and T. Hultin, Exp. Cell Res., r (1950) 37r>- t. |. \\ . 1 ittlefield, I . B. Keller, J. Gross, and P. C. Zamecnik,/. Biol. Chan., 217 (1955 ) hi; V. G. Alltrey, M. M. Daly, and A. E. Mirsky, J. Gen. Physiol, 37 153 [57. 9. A. Rich and J. D. Watson, Nature, 173 ( 19.S4) 995; A. Rich and J. D. Watson, Proc. Natl. . lead. Sci. I IS., 4° ( *954) 759- 10. F. 1 1. C. < 'nek and J. D. Watson, Nature, 177 ( 1956) 473 ; F. H. C. Crick and J. D. Watson, Ciba Foundation Symposium, « The Nature oj Viruses*, 1957. 11. D. L. D. Caspar, Nature, 177 (1956) 475. 12. M. 1 . Peterman and M. G. Hamilton, J. Biol. Chem., 224 (1957) 7-5' P. O. Tso, J. Bonner, and J. Vinograd, J. Biophys. Biochem. Cytol., 2 (1956) 725. 13. 11. K. Schachman, A. B. Pardee, and R. Y. Starrier, Arch. Biochem. Biophys., 38 (1952) 245. 14. E. T. Bolton, B. 1 1. 1 [oyer, and D. B. Ritter, Microsomal Particles and Protein Syn- thesis, Pergamon Press, New York, 1958, p. 18. is. A. Tissieres and J. 1). Watson, Nature, 182 (1958) 778; A. Tissieres, J. I). Watson, 1). Schlessinger, and B. R. Hollingworth,/. Mol. Biol, 1 (1959) 221; C. E. Hall and 1 1. S. Sl.iyeter, /. Mol. Biol., 1 ( 1959) 329; H. E. Huxley and G. Zubay,/. Mol. Biol., 2 ( i960) 10. t6. 1 1. I amfrom and E. R. Glowacki, /. Mol. Biol., 5 (1962) 97; P. O. Tso and J. Vinograd, Biochim. Biophys. Acta, 49 (1961) 113. [7. B, 1 1. ill and I*. Doty,/. Mol. Biol., 1 (1959) in, U.Z. Littauer and H. Eiscnbcrgcr, Biochim. Biophys. Acta, 32 (1959) 320; S. M. Timasheff, A. Brown, J. S. Colter, and M. Davies, Biochim. Biophys. Acta, 27 (1958) 662. is. C. G. Kurland, J. Mol. Biol., 2 (i960) 83. 19. J. P. Waller and J. I. Harris, Proc. Natl. Acad. Sci. U.S., 47(1961) 18. 20. P. F. Sp.dir, /. Mol. Biol., 4 (1962) 395. 21. D. Elson, Biochim. Biophys. Acta, 27 (1958) 216 and 36 ( 19S9) 372. 22. P. 1;. Spahr and B. R. Hollingworth,/. Biol. Chetn., 236 (1961) 823. 23. J. W. Littlefield, E. B. Keller, J. Cross, I'. Q Zamecnik,./. Biol. Chew., 217 (1955) 1 1 1 ; J. W. Littlefield and E. B. Keller, / Biol. Chem., 224 (19S7) 13; P. C. Za- mecnik and E. B. Keller, /. Biol. Chem., 209 (1954) 337; E. B. Keller and P. C. Zamecnik,/. Biol. Chem., 221 ( 1956) 4s. s-132 RNA IN THE SYNTHESIS OF PROTEINS 24. M. B. Hoagland, P. C. Zamecnik, and M. L. Stephenson, Biochim. Biophys. Acta, 24 (i957) 215. 25. M. B. Hoagland, M. L. Stephenson, J. F. Scott, L. I. Hecht, and P. C. Zamecnik, J. Biol. Chan., 231 (1958) 241. 26. F. H. C. Crick, Symp. Soc. Exptl. Biol, 12 (1958) 138. 27. P. Berg and E. J. Ofengand, Proc. Natl. Acad. Sci. U.S., 44 (1958) 78. 28. A. Tissicres, D. Schlessinger, and F. Gros, Proc. Natl. Acad. Sci. U.S., 46 (i960) 1450. 29. F. Gros and D. Schlessinger, unpublished experiments. 30. C. I. Davern and M. Mesclson, J. Mol. Biol., 2 (i960) 153. 31. M. Riley, A. Pardee, F. Jacob, and J. Monod, /. Mol. Biol., 2 (i960) 216. 32. S. Naono and F. Gros, Compt. Raid., 250 (i960) 3889. 33. A. D. Hershey, J. Dixon, and M. Chase, /. Gen. Physiol, 36 (1953) 777. 34. E. Volkin and L. Astrachan, Virology, 2 (1956) 149. 35. M. Nomura, B. D. Hall, and S. Spiegclman, /. Mol. Biol, 2 (i960) 306. 36. F. Jacob and J. Monod, /. Mol. Biol., 3 (1961) 318. 37. S. Brenner, F. Jacob, and M. Meselson, Nature, 190 (1961) 576. 38. F. Gros, H. Hiatt, W. Gilbert, C. G. Kurland, R. W. Risebrough, and J. D. Wat- son, Nature, 190 (196 1) 581. 39. B. D. Hall and S. Spiegelman, Proc. Natl. Acad. Sci. U.S., 47 (1961) 137. 40. M. Hayashi and S. Spiegelman, Proc. Natl. Acad. Sci. U.S., 47 (1961) 1564; F- Gros, W. Gilbert, H. Hiatt, G. Attardi, P. F. Spahr, and J. D. Watson, Cold Spring Harbor Symp. Quant. Biol., 26 (1961). 41. A. Tissicres and J. W. Hopkins, Proc. Natl. Acad. Sci. U.S., 47 (1961) 2015. 42. M. Chamberhn and P. Berg, Proc. Natl. Acad. Sci. U.S., 48 (1962) 81 and W. B. Wood and P. Berg, Proc. Natl. Acad. Sci. U.S., 48 (1962) 94. 43. M. W. Nircnbcrg and J. H. Matthaci, Biochem. Biophys. Res. Commun., 4 (1961) 404. 44. M. W. Nircnbcrg and J. H. Matthaei, Proc. Natl. Acad. Sci. U.S., 47 (1961) 1588. 45. R. W. Risebrough, A. Tissicres, and J. D. Watson, Proc. Natl. Acad. Sci. U.S., 48 (1962) 430. 46. W. Gilbert, /. Mol. Biol, 6 (1963) 374. 47. J. Bishop, J. Leahy, and R. Schweet, Proc. Natl. Acad. Sci. U.S., 4.6 (i960) 1030. 48. H. Dintzes, Proc. Natl. Acad. Sci. U.S., 47 (1961) 247. 49. M. Cannon, R. Krug, and W. Gilbert, /. Mol. Biol., 7 (1963) 360. 50. W. Gilbert, /. Mol. Biol, 6 (1963) 389. 51. D. Schlessinger and Francoise Gros, /. Mol. Biol., 7 (1963) 350. 52. S. H. Barondes, M. W. Nirenberg, Science, 138 (1962) 813; G. J. Spyridcs and F. Lipmann, Proc. Natl. Acad. Sci. U.S., 48 (1962) 1977; W. Gilbert, /. Mol. Biol., 6 (1963) 374- 53. S. S. Cohen, H. D. Barner, and J. Lichtenstein, /. Biol. Chan., 236 (1961) 1448. 54. R. Monier, S. Naono, D. Hayes, F. Hayes, and F. Gros,/. Mol. Biol, 5 (1962) 311; K. Asano, unpublished experiments (1962). 55. O. Maaloe, Cold Spring Harbor Symp. Quant. Biol, 26 ( 1961 ) 45 ; F. C. Neihardt and D. Fraenkel, Cold Spring Harbor Symp. Quant. Biol, 26 (1961) 63. s-133 [962 J. D. WATS O N ,<, K.Gierer, J. Mol.BioL, 6 (196} [48; J R. Warner, P. M. Knopf, and A. Rich.Frw. Natl. Acad. Set. U.S., 49(1963) >--■ I c-vinth.il, A. Keynan, and A. 1 liga, Proc. Natl. Acad. Set. I '.S., 48 ( 1962) 163 1. 58. II. Sekiguchi and S. S. Cohen,/. Biol.Chem., 238 (1963) 349; D- Schlessinger and P. l . Spahr, /. Biol. Chm.t 238 (1963) 6. $o. R. Gesteland and J. D. Watson, (will be published, 1963). s-134 SUPPLEMENT X Francis H. C. Crick On the genetic code Nobel Lecture, December 11, 1962 Part of the work covered by the Nobel citation, that on the structure and replication of DNA, has been described by Wilkins in his Nobel Lecture this year. The ideas put forward by Watson and myself on the replication of DNA have also been mentioned by Kornberg in his Nobel Lecture in 1959, covering his brilliant researches on the enzymatic synthesis of DNA in the test tube. I shall discuss here the present state of a related problem in information transfer in living material - that of the genetic code - which has long interested me, and on which my colleagues and I, among many others, have recently been doing some experimental work. It now seems certain that the amino acid sequence of any protein is deter- mined by the sequence of bases in some region of a particular nucleic acid molecule. Twenty different kinds of amino acid are commonly found in protein, and four main kinds of base occur in nucleic acid. The genetic code describes the way in which a sequence of twenty or more things is determined by a sequence of four things of a different type. It is hardly necessary to stress the biological importance of the problem. It seems likely that most if not all the genetic information in any organism is carried by nucleic acid - usually by DNA, although certain small viruses use RNA as their genetic material. It is probable that much of this information is used to determine the amino acid sequence of the proteins of that organism. (Whether the genetic information has any other major function we do not yet know.) This idea is expressed by the classic slogan of Beadle: « one gene -one enzyme », or in the more sophisticated but cumbersome terminology of today: « one cistron-one polypeptide chain ». It is one of the more striking generalizations of biochemistry - which surprisingly is hardly ever mentioned in the biochemical text-books - that the twenty amino acids and the four bases, arc, with minor reservations, the same throughout Nature. As fir as I am aware the presently accepted set of twenty amino acids was first drawn up by Watson and myself in the summer of 1953 in response to a letter of Gamow's. In this lecture I shall not deal with the intimate technical details of the s-135 [962 P.H. C. CRICK problem, if only for the reason that 1 have recently written such a review1 which will appear shortly. Nor shall I deal with the biochemical details of messenger KNA and protein synthesis, as Watson has already spoken about these. Rather I shall ask certain general questions about the genetic code and ask how tar we can now answer them. 1 et us assume that the genetic code is a simple one and ask how many- bases code tor one amino acid? This can hardly be done by a pair of bases, .is from tour different things we can only form 4x4=16 different pairs, whereas we need at least twenty and probably one or two more to act as spaces or tor other purposes. However, triplets of bases would give us 64 possibilities. It is convenient to have a word for a set of bases which codes one ammo acid and I shall use the word «codon» for this. This brings us to our first question. Do codons overlap? In other words, as we read along the genetic message do we find a base which is a member of two or more codons? It now seems fairly certain that codons do not overlap. It they did, the change of a single base, due to mutation, should alter two or more (adjacent) amino acids, whereas the typical change is to a single amino acid, both in the case of the « spontaneous » mutations, such as occur in the abnormal human haemoglobin or in chemically induced mutations, such as those produced by the action of nitrous acid and other chemicals on tobacco mosaic virus2. In all probability, therefore, codons do not overlap. This leads us to the next problem. How is the base sequence, divided into codons? There is nothing in the backbone of the nucleic acid, which is perfectly regular, to show us how to group the bases into codons. If, for example, all the codons are triplets, then in addition to the correct reading of the message, there are two ///correct readings which we shall obtain if we do not start the grouping into sets of three at the right place. My col- leagues and [3 have recently obtained experimental evidence that each section of the genetic message is indeed read from a fixed point, probably from one end. This fits in very well with the experimental evidence, most clearly shown 111 the work of Dintzis4 that the amino acids arc assembled into the polypeptide chain in a linear order, starting at the amino end of the chain. This leads us to the next general question: the size of the codon. How many bases are there in any one codon? The same experiments to which I have just referred3 strongly suggest that all (or almost all) codons consist of a triplet of bases, though a small multiple of three, such as six or nine, is not completely ruled out by our data. We were led to this conclusion by the study of mutations in the A and B cistrons of the r„ locus of bacteriophage s-136 ON THE GENETIC CODE T4. These mutations are believed to be due to the addition or subtraction of one or more bases from the genetic message. They are typically produced by acridincs, and cannot be reversed by mutagens which merely change one base into another. Moreover these mutations almost always render the gene completely inactive, rather than partly so. By testing such mutants in pairs we can assign them all without exception to one of two classes which we call -j- and — . For simplicity one can think of the + class as having one extra base at some point or other in the genetic message and the — class as having one too few. The crucial experiment is to put together, by genetic recombination, three mutants of the same type into one gene. That is, either ( -\- with -f- with -+- ) or ( — with — with — ) . Where- as a single -f or a pair of them ( + with -(- ) makes the gene completely inactive, a set of three, suitably chosen, has some activity. Detailed examina- tion of these results show that they are exactly what we should expect if the message were read in triplets starting from one end. We are sometimes asked what the result would be if we put four -|-'s in one gene. To answer this my colleagues have recently put together not merely four but six -f-'s. Such a combination is active as expected on our theory, although sets of four or five of them are not. We have also gone a long way to explaining the production of « minutes » as they are called. That is, combinations in which the gene is working at very low efficiency. Our detailed results fit the hypothesis that in some cases when the mechanism comes to a triplet which does not stand for an amino acid (called a « non- sense » triplet) it very occasionally makes a slip and reads, say, only two bases instead of the usual three. These results also enable us to tie down the direc- tion of reading of the genetic message, which in this case is from left to right, as the Tn region is conventionally drawn. We plan to write up a detailed technical account of all this work shortly. A final proof of our ideas can only be obtained by detailed studies on the alterations produced in the amino acid sequence of a protein by mutations of the type discussed here. One further conclusion of a general nature is suggested by our results. It would appear that the number of nonsense triplets is rather low, since we only occasionally come across them. However this conclusion is less secure than our other deductions about the general nature of the genetic code. It has not yet been shown directly that the genetic message is co-linear with its product. That is, that one end of the gene codes for the amino end of the polypeptide chain and the other for the carboxyl end, and that as one proceeds along the gene one comes in turn to the codons in between in the s-137 [962 F.H.C. CRICK Linear order in which the amino acids are found in the polypeptide chain. Tins seems highly likely, especially .is it has been shown that in several sys- tems mutations affecting the same amino acid are extremely near together on the genetic map. The experimental proof of the co-linearity of a gene and the polypeptide chain it produces may be confidently expected within the next year or so. There is one further general question about the genetic code which we can ask at this point. Is the <:odc universal, that is, the same in all organisms? Preliminary evidence suggests that it may well be. For example something very like rabbit haemoglobin can be synthesized using a cell-free system, part of which comes from rabbit reticulocytes and part from Escherichia C('// \ This would not be very probable if the code were very different in these two organisms. 1 low ever as we shall see it is now possible to test the univer- sality of the code by more direct experiments. In a cell in which DNA is the genetic material it is not believed that DNA itself controls protein synthesis directly. As Watson has described, it is be- lieved that the base sequence of the DNA - probably of only one of its chains - is eopud onto RNA, and that this special RNA then acts as the genetic messenger and directs the actual process of joining up the amino acids into polypeptide chains. The breakthrough in the coding problem has come from the discovery, made by Nirenberg and Matthaei6, that one can use synthetic RNA for this purpose. In particular they found that polyuridylic acid - an RNA in which every base is uracil - will promote the synthesis of polyphenylalanine when added to a cell-free system which was already known to synthesize polypeptide chains. Thus one codon for phenylalanine appears to be the sequence UUU (where U stands for uracil: in the same way we shall use A, G, and C for adenine, guanine, and cytosinc respec- tively). This discovery has opened the way to a rapid although somewhat contused attack on the genetic code. It would not be appropriate to review this work in detail here. I have discussed critically the earlier work in the review mentioned previously1 but such is the pace o{ work 111 this field that more recent experiments have already made it out ot date to some extent. However, some general con- clusions can safely be drawn. The technique mainly used so far, both by Nirenberg and his colleagues'' and by Ochoa and his group7, has been to synthesize enzymatically « ran- dom 1 polymers of two or three of the four bases. For example, a polynu- cleotide, which I shall call poly (U,C), having about equal amounts of s-138 ON THE GENETIC CODE uracil and cytosinc in (presumably) random order will increase the incor- poration of the amino acids phenylalanine, serine, leucine, and proline, and possibly threonine. By using polymers of different composition and assum- ing a triplet code one can deduce limited information about the composition of certain triplets. From such work it appears that, with minor reservations, each polynu- cleotide incorporates a characteristic set of amino acids. Moreover the four bases appear quite distinct in their effects. A comparison between the triplets tentatively deduced by these methods with the changes in amino acid se- quence produced by mutation shows a fair measure of agreement. Moreover the incorporation requires the same components needed for protein syn- thesis, and is inhibited by the same inhibitors. Thus the system is most unlikely to be a complete artefact and is very probably closely related to genuine protein synthesis. As to the actual triplets so far proposed it was first thought that possibly every triplet had to include uracil, but this was neither plausible on theoret- ical grounds nor supported by the actual experimental evidence. The first direct evidence that this was not so was obtained by my colleagues Bretscher and Grunberg-Manago8, who showed that a poly (C,A) would stimulate the incorporation of several amino acids. Recently other workers9-10 have reported further evidence of this sort for other polynucleotides not con- taining uracil. It now seems very likely that many of the 64 triplets, possibly most of them, may code one amino acid or another, and that in general sev- eral distinct triplets may code one amino acid. In particular a very elegant experiment11 suggests that both (UUC) and (UUG) code leucine (the brackets imply that the order within the triplets is not yet known). This general idea is supported by several indirect lines of evidence which cannot be detailed here. Unfortunately it makes the unambiguous determination of triplets by these methods much more difficult than would be the case if there were only one triplet for each amino acid. Moreover, it is not possible by using polynucleotides of « random » sequence to determine the order of bases in a triplet. A start has been made to construct polynucleotides whose exact sequence is known at one end, but the results obtained so far are suggestive rather than conclusive12. It seems likely however from this and other unpub- lished evidence that the amino end of the polypeptide chain corresponds to the « right-hand » end of the polynucleotide chain - that is, the one with the 2', 3' hydroxyls on the sugar. It seems virtually certain that a single chain of RN A can act as messenger s-139 igCM F.H.C.CRICK RNA, since polv U H a angle chain without secondary structure. If poly A is added to polv U. to form a double or triple helix, the combination is inactive. Moreover there is preliminary evidence9 which suggests that sec- ondarv structure within a polynucleotide inhibits the power to stimulate protein synth It has vet to be shown bv direct biochemical methods, as opposed to the indirect genetic evidence mentioned earlier, that the code is indeed a triplet :npts have been made from a study of the changes produced by muta- tion to obtain the relative order of the bases within various triplets, but my own view is that these are premature until there is more extensive and more reliable data on the composition of the triplets. Evidence presented bv several groups8-911 suggest that poly U stimulates both the incorporation ot phenvlalanine and also a lesser amount of leucine. The meaning ot this observation is unclear, but it raises the unfortunate possibilitv ot ambiguous triplets: that is, triplets which may code more than one amino acid. However one would certainly expect such triplets to be in a minoritv. It would seem likely, then, that most of the sixtv-four possible triplets will be grouped into twenty groups. The balance of evidence both from the cell- tree system and from the study of mutation, suggests that this does not occur at random, and that triplets coding the same amino acid may well be rather similar. This raises the main theoretical problem now outstanding. Can this grouping be deduced from theoretical postulates? Unfortunately, it is not difficult to see how it might have arisen at an extremelv early stage in evolu- tion by random mutations, so that the particular code we have may perhaps be the result of a series of historical accidents. This point is of more than abstract interest. It the code does indeed have some logical foundation then it is legitimate to consider all the evidence, both good and bad, in any attempt to deduce it. The same is not true if the codons have no simple logical connec- tion. In that case, it makes little sense to guess a codon. The important thing is to provide enough evidence to prove each codon independently. It is not yet clear what evidence can safely be accepted as establishing a codon. What is clear is that most of the experimental evidence so far presented falls short of proof in almost all cases. In spite of the uncertainty of much of the experimental data there are certain codes which have been suggested in the past which we can now reject with some degree of confidence. b-140 ON THE GENETIC CODE Comma-less triplet codes All such codes arc unlikely, not only because of the genetic evidence but also because of the detailed results from the cell-free system. Two-letter or three-letter codes For example a code in which A is equivalent to O, and G to U. As al- ready stated, the results from the cell-free system rule out all such codes. The combination triplet code In this code all permutations of a given combination code the same amino acid. The experimental results can only be made to fit such a code by very special pleading. Complementary codes There are several classes of these. Consider a certain triplet in relation to the triplet which is complementary to it on the other chain of the double helix. The second triplet may be considered either as being read in the same direc- tion as the first, or in the opposite direction. Thus if the first triplet is UCC, we consider it in relation to either AGG or (reading in the opposite direction) GGA. It has been suggested that if a triplet stands for an amino acid its comple- ment necessarily stands for the same amino acids, or, alternatively in another class of codes, that its complement will stand for no amino acid, i.e. be nonsense. It has recently been shown by Ochoa's group that poly A stimulates the incorporation of lysine10. Thus presumably AAA codes lysine. However since UUU codes phenylalanine these facts rule out all the above codes. It is also found that poly (U,G) incorporates quite different amino acids from poly (A,C). Similarly poly (U,C) differs from poly (A,G)9-10. Thus there is little chance that any of this class of theories will prove correct. Moreover they are all, in my opinion, unlikely for general theoretical reasons. A start has already been made, using the same polynucleotides in cell-free systems from different species, to see if the code is the same in all organisms. Eventually it should be relatively easy to discover in this way if the code is universal, and, if not, how it differs from organism to organism. The prelim- inary results presented so far disclose no clear difference between E. coli and mammals, which is encouraging10'13. s-141 [962 I. II. C. C RICK At the present tunc therefore, the genetic code appears to have the fol- lowing general properties: (1) Most if not all codons consist of three (adjacent) bases. (2) Adjacent codons do not overlap. (3) The message is read m the correct groups of three by starting at some fixed point. (4) The code sequence 111 the y^cne is co-linear with the amino acid se- quence, the polypeptide chain being synthesized sequentially from the ammo end. (5) In general more than one triplet codes each amino acid. (6) It is not certain that some triplets may not code more than one amino acid, i.e. they may be ambiguous. (7) Triplets which code for the same amino acid are probably rather sim- ilar. (S) It is not known whether there is any general rule which groups such codons together, or whether the grouping is mainly the result of histor- ical accident. (9) The number of triplets which do not code an amino acid is probably small. (10) Certain codes proposed earlier, such as comma-less codes, two- or three-letter codes, the combination code, and various transposible codes arc all unlikely to be correct. (11) The code in different organisms is probably similar. It may be the same in all organisms but this is not yet known. Finally one should add that in spite of the great complexity of protein syn- thesis and in spite of the considerable technical difficulties in synthesizing polynucleotides with defined sequences it is not unreasonable to hope that all these points will be clarified in the near future, and that the genetic code will be completely established on a sound experimental basis within a few years. s-142 ON THE GENETIC CODE The references have been kept to a minimum. A more complete set will be found in the first reference. 1. F. H. C. Crick in Progress in Nucleic Acid Research, J. N. Davidson and Waldo E. Colin (Eds.), Academic Press Inc., New York (in the press). 2. H. G. Wittmann, Z. Vererbungslehre, 93 (1962) 491. A. Tsugita, J. Mol. Biol, 5 (1962) 284, 293. 3. F. H. C. Crick, L. Barnett, S. Brenner, and R. J. Watts-Tobin, Nature, 192 ( 1961 ) 1227. 4. M. A. Naughton and Howard M. Dintzis, Proc. Natl. Acad. Sci. U.S., 48 (1962) 1822. 5. G. von Ehrenstein and F. Lipmann, Proc. Natl. Acad. Sci. U.S., 47 (1961) 941. 6. J. H. Matthaei and M. W. Nirenberg, Proc. Natl. Acad. Sci. U.S., 47 (1961) 1580. M. W. Nirenberg and J. H. Matthaei, Proc. Natl. Acad. Sci. U.S., 47 (1961 ) 1588. M. W. Nirenberg, J. H. Matthaei, and O. W. Jones, Proc. Natl. Acad. Sci. U.S., 48 (1962) 104. J. H. Matthaei, O. W. Jones, R. G. Martin, and M. W. Nirenberg, Proc. Natl. Acad. Sci. U.S., 48 (1962) 666. 7. P. Lengyel, J. F. Speyer, and S. Ochoa, Proc. Natl. Acad. Sci. U.S., 47 ( 1961 ) 1936. J. F. Speyer, P. Lengyel, C. Basilio, and S. Ochoa, Proc. Natl. Acad. Sci. U.S., 48 (1962)63. P. Lengyel, J. F. Speyer, C. Basilio, and S. Ochoa, Proc. Natl. Acad. Sci. U.S., 48 (1962) 282. J. F. Speyer, P. Lengyel, C. Basilio, and S. Ochoa, Proc. Natl. Acad. Sci. U.S., 48 (1962) 441. C. Basilio, A. J. Wahba, P. Lengyel, J. F. Speyer, and S. Ochoa, Proc. Nail. Acad. Sci. U.S., 48 (1962) 613. 8. M. S. Bretscher and M. Grunberg-Manago, Nature, 195 (1962) 283. 9. O. W. Jones and M. W. Nirenberg, Proc. Natl. Acad. Sci. U.S., 48 (1962) 2115. 10. R. S. Gardner, A. J. Wahba, C. Basilio, R. S. Miller, P. Lengyel, and J. F. Speyer, Proc. Natl. Acad. Sci. U.S., 48 (1962) 2087. 11. B. Weisblum, S. Benzer, and R. W. Holley, Proc. Natl. Acad. Sci. U.S., 48 (1962) 1449. 12. A. J. Wahba, C. Basilio, J. F. Speyer, P. Lengyel, R. S. Miller, and S. Ochoa, Proc. Natl. Acad. Sci. U.S., 48 (1962) 1683. 13. H. R. V. Arnstein, R. A. Cox, and J. A. Hunt, Nature, 194 (1962) 1042. E. S. Maxwell, Proc. Natl. Acad. Sci. U.S., 48 (1962) 1639. I. B. Weinstein and A. N. Schechter, Proc. Natl. Acad. Sci. U.S., 48 (1962) 1686. s-143 AUTHOR INDEX Supplements not indexed; page numbers in bold face refer to photographs. Abelson, P. H., 516 Adams, J. N., 378 Adelberg, E. A., 277, 303, 304, 315, 320, 321, 327, 328, 329, 337, 338, 355, 358, 361, 402, 420, 463 Akinrimisi, E. O., 303 Alexander, P., 186, 199 Allen, J. M., 462 Allen, M. K., 300, 304 Allfrey, V. G.. 431, 432, 493, 494, 497, 498, 499 Allison, A. C, 212 Ames, B. N., 461, 462 Anderson, T. F., 318 Anfinsen, C. B., 418, 420 Arber, W., 336, 337 Auerbach, C, 162, 226 August, J. T., 433 Avery, O. T., 295, 303 Bachmann, B., 29 Bacq, Z. M., 186 Baglioni, C, 413, 416, 420, 435 Bailey, W. T., 344, 539 Balboni, E. R., 24 Balis, M. E., 458, 462 Baltimore, D., 366, 367 Bangham, A. D., 113 Barigozzi. C, 370, 380 Barnett, L., 351, 436 Barr, M. L., 486, 489 Barratt, R. W., 145 Basilio, C, 446 Bateson, W., 49, 53 Bauer, H., 162 Baumiller, R. C, 219 Bautz, E. K. F., 432 Beadle, G. W., 127, 145, 408, 420 Beale. G. H., 380, 454, 455 Beam, A. G., 176, 420 Becker, H. J., 492, 498 Becker, Y., 426, 432 Beermann, W., 492, 496, 498 Beers, R. F., Jr., 264 Belling, J., 151, 157, 162, 228 Bender, M. A., 186 541 Bendich. A.. 276. 277 Bennett, D.. 83 Bennett. T. P.. 445, 446 Benzer, S., 344. 346. 351, 352, 392, 401, 439, 440, 446, 447 Berg, P., 289. 290, 394, 428, 429, 434 Bernfield. M. R.. 443 Bertsch, L. L., 290 Bessman, M. J., 290, 450, 455 Beutler, E., 485, 489 Billeter, M. A., 368 Binnington. J. P., 186 Birnstiel, M. L., 427, 432 Bladen, H. A., 431, 432 Blakeslee. A. F., 151. 157, 162, 228 Bloch, D. P., 499 Blum, H. F., 516 Boehner, L., 344, 352 Boice, L., 501 Bollum, F. J., 288, 290, 397 Bolton, E. T., 304 Bonner, D. M., 420, 434 Bonner, J.. 425, 496, 497, 499, 505 Borek, E., 429 Borst, P., 368 Boveri, T., 14 Boyce, R. P., 397, 401 Boyer, S. H., IV, 420 Brachet, J., 12, 28 Bradshaw, A. D.. 28 Brehme, K. S., 29, 162 Brenner, S., 339, 351, 432, 436, 445 Brewen, J. G.. 182, 186 Bridges, C. B., 29. 95, 98, 99, 105, 113, 154, 162, 474, 482 Briggs, R., 502, 508 Brink, R. A., 465, 466, 467, 470, 471 Brown, D. D., 427, 432, 502 Brown, D. F., 470, 471 Bryson, V., 290, 446 Bunker, M. C. 162 Burdette, W. J., 162 Burdon, R. H., 368 Burnet, F. M., 363, 367 Burnham, C. R., 127 Burns, S. N., 327. 328, 355, 361 Burton, K... 290 Busch, H., 500 Byrne, R., 431, 432, 461 Cairns, J., 327, 328. 453, 455 Calvin, M., 516 Callan, H. G., 494, 499 Campbell, A., 336, 337, 350, 358, 361, 376, 380, 446 Campbell, P. L., 505, 507 Canellakis, E. S., 289, 290 Carlson, E., 398, 402 Carlson, J. G.. 181 Carlton. B. C. 445, 447 Carothers, E. E.. 19, 28 Carrier, W. L.. 397, 402 Caston, J. D.. 502 Cavalieri, L. F.. 289. 290 Ceppellini. R., 416 Cerhova. M., 290 Chamberlin, M., 289, 290 Champe, S. P., 346, 351, 439, 440, 446 Chandler, B., 426, 432 Chandler, B. L.. 275, 277 Chang, R. S., 500 Chargaff. E., 264, 286, 291 Chase, M., 340, 344, 352, 353 Chesley, P., 83 Chevremont. M., 378, 380 Childs, B., 485, 489 Chipchase, M. I. H., 427, 432 Chovnick, A., 481, 482 Chu, E. H. Y., 186, 226 Clark, A. J.. 328, 358, 361 Clark, F., 516 Clark. J. O. D., 374, 381 Claus, W. D., 186 Cleland, R. E., 9, 16, 228, 231, 232, 239 Clever, U., 492, 498 Clowes, R. C, 367 Cobb, V., 4 Cock, A. G., 489 Coe, E. H., Jr., 142 Cohen, S. S., 431, 453, 455 Cohn, W. E., 264, 432 Cooper, K. W.. 185, 186 Cordova, C. C, 505, 507 Coulon, E. M., 496, 499 Cowan, C. A., 373, 381 Cowie. D. B., 336, 516 Cox, E. C, 434 Craig, L. C, 412, 420, 445, 446 Crawford, L. P., 420 Creighton. H. S.. 126, 128 Crick. F. H. C, 267, 268. 271. 277, 395, 432, 436, 446, 447 Crow, J. F., 66. 199. 210, 212, 217, 218, 219, 226. 387, 528 Dais. D., 444, 446 Darlington, C. D., 12. 28, 379 Davidson, D. R., 485. 489 Davidson, E. H.. 499 Davidson, J. N., 264. 432 Davidson, P. F.. 352 Davies, D. R.. 304 Davis, J. E., 363, 367. 368 Dawson, M. H., 295 Day, P. R., 29, 144 Deering. R. A., 401 DeGeorge, F. V., 86 >42 INDI \ Delbrikk, M . J44 iv Mars. K , Demerei \! !9, 173, 526, )32, 194 506 Denhardt, c. IF. 344, 353, 501 DeRobertis, I D P DaVries, M . 228, 239 Dintzes, H., 429 Dobzhansky, I li .. 39, 70, 86, 162. 212, 213, 216.217, 218, 221, 246, 250, ^;" Dodson, I t) . 59 Docrmann. \ H , 344, 352 Doi, V M . 42^ Doi, R. M . 502, 506 Dins. P., 250. 301. 304 Dunn. li. ;". 79, 80, 81, 83, 84, 250, 539 Duryee, W. R . 494 Dunnebacher, T. H.. 500 I akin. R I .. 512, 516 Ebert, J. D., 508 Edelman, M.. 373. 38 1 Edgar, R S., 343, 344. 353. 501 I dwards, J. H.. 64. 66 1 gyhazi, I ., 505, 507 Ehrlich, P. R.. 250 Ehrman, 1 ... 246. 250 I igner, J . 301, 304 1 Ider, \.. 458, 462 Ellis, D, B., 454. 455 I Won. IF. 431. 434 I merson, R. A.. 127. 145 l merson, S., 22S I ngelberg, J., 447 I ngelhardt, D. E., 365. 367 M. I). 363. 367 I phrussi, B., 421. 503. 507 Ephrussi-Taylor, H„ 296. 304, 381, 121 Epling, C . 243 I pstein, H. T.. 373. 381 Epstein, R. H.. 501. 507 I tans, A. IF. 505. 507 Fairbanks. V. F . 485. 489 aei I) s . 66, 539 Falkow, s. 304 I ancher, H . 289, 290 Faulkner, R.. 497, 498 uson-Smith, M. A.. 160 I elsenfeld, S., 496. 499 l iala, J.. 431. 434 I inch, I. F. 367 Fincham, J. R. s . 29, 144 Flax, F (... 434 I lemming. VV.. I 2 s 12. 28. 127. 199. 212. 264 P 453, 455 I owler, w. \F. 485, 1 ox. \1 s 299, 300, 304 I ox, S. W., 511, 516 1 raenkeK onrat, IF. 365, 366. J67, 368, 446 Francis, ( \F. 505, 507 l rancis, I ., 83 l rasei \ I 145 I redericq, P., 359, 361 I reese, I . >1'2. 398, 399. 401 I reese, I B . J99, 401 I regin, A., 504, 508 I reifelder, I).. 352 Freire-Maia, N.. 226 I renster, J. IF. 498, 499 I riedman, F. 504. 508 Furth, J. J.. 432. 433 Gabriel, M. F.. 12. 28. 127. 199. 212. 264 Gaffron, H.. 510 Gahan. P. B.. 494, 499 Gall, J. G.. 494. 495. 499 Gallant. J.. 462 Gardner. R. S., 446 Garen, A.. 440. 458. 462 Garnjobst. F.. 145 Garrod. A. E.. 404 Gartner. T. K... 461. 463 Gates, R. R.. 39 Gay. H.. 432 Geiduschek. E. P.. 432, 433 Gellert, M.. 304 German. J. F.. III. 176 Giacomoni, D., 428. 433, 445 Gibson, F, 454, 455 Gierer. A., 365. 367 Gilbert, W., 429, 433 Gilden. R. V.. 497, 499 Giles, N. FF. 186, 226 Gish. D. T., 365, 368 Glass, B., 337, 351, 353, 507 Gluecksohn-Waelsch. S., 86 Godoy, G. A.. 374, 381 Goebel. W. F., 359. 361 Gold. M.. 429 Goldberg. B. D., 503, 508 Goldhaber. P.. 500 Goldman, M., 432 Goldschmidt, R. B.. 86. 87, 113 Goldstein, J., 445, 446 Gomatos. P. J., 366, 367 Gonano, F., 447 Gooch. P. C. 186 Goodman. H. M., 445, 446 Gordon. J. A.. 304 Gorini, F., 463 Gowen. J. W.. 212 Green. D. F.. 378. 381. 516 Green, FF. 503. 508 Green, \F. 278 Green, M. ( .. 140 Green, M. FF. 336. 426, 433 Green, M. M.. 488 Greer. S.. 402 Grell, I IF. 29 Griffen, A. B., 162 Griffith, 1 .. 295 (, Ionian. N. IF. 452. 456 ( rlliss. I .. 433 Gross. P. R.. 500 Grossman, F.. 448 Grotsch. H.. 51 I Grumbach, M. M.. 485. 489 Grunberg-Manago, M.. 446 Griineberg, H., 86 Guest, J. R., 445. 447 Guidotti, G., 412. 420 Gunsalus. I. C. 337 Gurdon. J. B.. 427, 432 Guthrie. G. D.. 340, 352 Habermann, U.. 290 Habermannova. S.. 290 Hadorn. E., 70, 71. 86 Haldane. J. B. S.. 101 Hall. B.. 389. 390 Hall. B. D.. 432. 433 Hamburger. V.. 79 Hamilton. T. H.. 505, 507 Hanafusa, H., 352 Hanafusa. T., 352 Hannah-Alava, A., 113 Hardy, G. H.. 212 Harm, W.. 350, 352 Harnden, D. G., 484 Harrington. H.. 394. 401 Harris. H.. 420, 504, 507 Harris, W., 212, 213 Hart, R. G.. 364 Hartman, P. E., 332. 333, 334. 461. 462 Haruna, F. 366 Harvald. B.. 89 Haselkorn. R.. 367 Haskell. G.. 29 Haslbrunner. E.. 378. 381 Hauge. M., 89 Hayashi. M.. 425. 426. 432, 433 Hayashi. M. N.. 426, 432, 433 Hayes, W.. 317. 319. 328. 329, 352. 358. 359. 361. 367 Hechter. O.. 516 Hede. R.. 352 Heitz, E.. 162 Helinski, D. R.. 349. 352. 445, 447 Henning. U.. 445. 447 Herriot. R. M., 304 Hershey, A. D.. 340. 342. 344, 352, 353 Herskowitz, F, 76 Herskowitz. I. H., 9, 16. 29. 80. 81. 84. 183. 219. 226, 393, 401. 465, 466, 467, 480 Authors 543 Herskowitz, J.. 76 Hess, O., 496. 498 Hiatt, H., 433 Hinegardner, R. T.. 447 Hiraizumi, Y.. 387, 388, 390 Hirst, G. K., 363, 367 Hoagland. M. B.. 433 Hochstim. A. R., 510, 516 Hogness. D. S.. 336, 338 Holland, J. J.. 453, 455, 494, 499 Hollaender. A.. 176 Holm. R. W.. 250 Holtz. A. M.. 70, 86 Hopkins. J. W.. 427. 434 Horn, E. C 499 Home, R. W.. 351, 367 Horowitz. N. H., 513. 516 Hotchkiss. R. D., 304 Hotta, Y.. 493, 499. 503, 507, 508 Howard-Flanders. P., 397, 401 Hoyer. B. H, 304. 502. 507 Hsia, D. Y.-Y.. 420 Hsu, T. C. 159 Huang. R. C. C. 425. 497, 499 Huang. S.-S., 516 Hurwitz, J.. 426, 429, 432. 433 Hyden, H., 505, 507 Hymer, W. C, 431, 434 Hudson, W. R., 367 Igarashi, R. T., 502, 506 Ingerman, M. L., 277 Ingram, V. M, 412. 413. 414, 417. 418, 420 Inhorn. S. L.. 162 Isaacson, R. G., 148 Ishida. M. R., 373, 381 Itano, H. A., 412, 416. 420 Izawa, M., 493. 494. 499 Jackson. J. F.. 290 Jacob, F.. 319, 321, 323, 327, 328, 329, 335, 337, 338. 350, 353, 355, 361, 432, 460, 463 Janaki-Ammal, E. K., 12 Jehle. H. 16. 28. 277 Jenks. W. P.. 304 Jinks, J. L.. 381 Johannsen. W. L.. 1. 12. 13 John. B.. 28. 146 Johnston, A. W.. 160 Joklik. W. K., 426. 432 Jones, O. W.. 441, 446, 462, 463 Jukes. T. H, 442. 446 Kaesberg, P., 363, 367 Kaiser, A. D., 336, 338, 350, 353 Kallman, F. J., 86 Kammen, H. O., 289, 290 Kaplan, W. D., 487, 490 Karpechenko, G. D.. 248 Kasha, M., 510. 511. 516 Katoh. A., 493, 499 Kano-Sueoka, T., 503, 508 Kaufmann, B. P.. 29. 154, 174. 175 Keeler, C. E., 4 Kellenberger, E., 293, 336, 337. 340. 351. 353 Kempthorne, O.. 539 Kendrew. J. C. 420 Keosian, J., 516 Kermickle. J. L., 470, 471 Kernaghan, R. P.. 481, 482 Kessinger. M. A., 212. 213 Khorana. H. G.. 290 Kiesselbach, T. A., 29 Kihara, H.. 251 Kilkson. R.. 339, 353 Kim. Y. T.. 367 King. T. J., 502, 508 Kinosita, R.. 487, 490 Knight. C. A.. 365. 368 Konetzka. W. A.. 327, 329 Konigsberg. W.. 412. 420 Kornberg. A.. 280, 290, 291, 402. 456 Kossikov, K. V.. 482 Kostyanovskii, R. G., 276, 277 Krakow. J. S., 289, 290, 426, 434 Krauss, M., 481, 482 Krieg. D. R.. 401 Krieger. H., 226 Kubitschek, H. E.. 401 Kuff. E. L., 431, 434 Kurland. C. G.. 433, 434 Lampen, J. O.. 290. 446 Landauer. W.. 79. 86 Landsteiner, K., 57. 58 Lane, D.. 301. 304 Larsson, A.. 449, 456 Laughlin, J. S.. 183 Laughnan. J. R.. 477, 482 Leboy, P. S., 434 Leder, P., 444, 446, 461 Lederberg, E. M.. 307. 309, 315, 335, 338 Lederberg, J.. 307, 309, 310. 313, 314. 315, 330, 334, 335, 338, 516 Lee-Huang. S.. 289. 290 Lehrmann. H.. 412 l.emmon. R. M.. 516 Lengyel. P., 446 Lennox, E., 332 Lerman, L. S.. 296. 302. 304. 394, 402 Levan, A., 389, 390 Levene, H., 539 Levin, J. G., 431. 432 Levine, L.. 304 Levine. P.. 57 Levinthal. C., 352 Lewis, E. B.. 478. 479. 480. 482 Lewis. K. R., 28. 146 L'Heritier. P.. 370, 381 Li, C. C., 212 Lima de Faria. A.. 379, 493, 499 Lindsley. D. L.. 29 Lipmann, F.. 434, 447 Lipsett. M. N.. 304 Litman. R. M.. 303. 304 Littau. V. C. 497, 498 Littauer, U. Z., 429 Locke. M.. 507 Lockhart, R. Z.. Jr.. 454, 456 Loeb, T.. 318. 363. 367 Luck, D. J. L.. 378. 381 Lunt, M. R.. 290 Luria. S. E.. 334 Luzzati. V.. 277. 496. 499 Lyon. M. F.. 485. 487. 489 Maas. R.. 361 MacDowell. E. C, 83 MacLeod. C. M.. 295. 297, 303 Maestre. M. F., 339, 353 Magni. G. E.. 383. 389, 448 Maheshwari. N.. 425 Mahler, I., 463 Makino, S„ 12 Mandel. M.. 304 Margolin. P., 460, 463 Mariner. R.. 516 Markert. C. L.. 502, 507 Marmur. L. 301. 304, 361, 452, 456, 463 Mather. W. B.. 539 Matthaei. J. H.. 440. 446 Mayer, E., 250 Mazia, D.. 12 McCarthy. B. L. 304. 336, 494, 499, 502, 507 McCarty. M., 295, 303 McClintock, B., 126, 127, 128, 384, 386, 389, 390, 469, 470. 471 McConkey. E. H., 427. 434 McElroy. W. D.. 337. 351. 353. 507 McKusick. V. A.. 113. 139. 490 McLeisch. J.. 28 Mead. C. G.. 394. 402 Melechen. N. E.. 294. 308 Mendel. G., 31. 39. 42. 53. 138 Merchant. D. J.. 507 Merrill. D. L, 250 Meselson. M.. 271, 274. 277, 351. 353. 432 Metz. C. W., 236 Meyer. G. F.. 496. 498 Miescher, F.. 253. 264 Miller. R. S.. 446 Miller, S. L.. 516 Mills, S. E., 420 r\Dl \ \lnskv. \ E.. 12, 28, 431, 4'>2. ... . ■ ;»v 199 Mitchell, H K 29 J78, 381, 420 Mitchell, M B . 378, 381 Mitra, S U '■ ;"~ Mohr, i) l .. 40 Monod, J.. 350, 460, 163 Montagu, A B6 Moohr, J W., 432 Moore, J. A.. 503, 507. 508 Moorhead, P S 389, 390 Morgan, D. I.. Jr.. 173 Morgan, 1. H., 91, 99, 128 Morishima, \.. 485, 4Nl> Morse, M. I .. 335, 338 Morton, N I . 217. 2ll>. 226 Mosig, Ci . 353 Muench, K.. 42^ Mukai, F. H.. 460, 463 Muller, H. J., /.v. 29. 176, 183, 186, 194. 195. 198. 199, 200. 217. 21K. 219. 226. 398, 402. 474. 482. 489. 490 Miintzing, A.. 226 Nagata, T.. 328 Nakamoto. T.. 433 Nance. W. E.. 420 Nathans. D., 440. 446 Nebel. B. R.. 496. 499 Neel, J. \.. 40. 412. 507 Neumann. J.. 463 Nov. man. H. H.. 86 Newmever. D.. 145 Nichols! W. W.. 389. 390 Nicolaieff. A.. 496. 499 Nirenberg. M. W.. 431. 432, 440, 441. 443. 444. 446. 447, 461, 462. 463 Niu. L. C, 505. 507 Niu. M. C, 505, 507 Nitowskv. H. M., 485, 489 Noll. H.. 442 Nordqvist. T.. 493. 499 Norris. A. T.. 428. 434 Nossal. G. J. V.. 504. 507 Notani. G., 440. 446 Novick. A.. 391. 402 Novitski. E.. 29 Nowinski. W. W., 28 Nozu. K.. 366 Ochoa. S.. 366. 368. 426, 434. 441. 447 Oehlkers. F., 228 Ohno. S.. 4K7. 490 Ohlaka. > .. 366, 461. 463 Olson. M. 1 .. 424 Oparin. A. !.. 516 Onas. J . 461. 463 Osborn. F.. X6 Osborn. R. H., 86 Osier. I. I . 200 Ostergren, G., 389, 390 Otsuji, N., 458. 462 Ottolenghi, I . 247 Oura, H . 442 Ozeki, H., 334. 338 Painter. T. S.. 162 Paranchych, W., 454. 455 Pardee. A. B., 463 Parke. W. ( .. 277 Parsons. D. F.. 378. 381 I'.tssano. k.. 1S6. 226 Patau. K.. 162. 176 Patterson. J. T.. 239 Pauling. I... 412 Pavlovsky, O.. 246, 250 Peacock. W. J.. 275. 277. 492, 499 Pearson, C. M., 485. 490 Penrose, L. S.. 516 Perkins. D. D.. 145 Perrin. D.. 460. 463 Perutz. M. F.. 412 Peters, J. A.. 12. 28. 39, 99. 113, 128. 146. 162. 199, 212, 277, 303. 315, 367, 389. 420. 482 Petersen. G. B.. 290 Peterson, H. M.. 477, 482 Peterson, P. A.. 390 Philipson. L.. 506, 507 Pina. M.. 278 Plus, N.. 370. 381 Pollister. A. W.. 379, 381 Pollister. P. F.. 379, 381 Pollmann. W.. 511 Pond. V.. 186 Ponnamperuma. C 516 Pontecorvo, G., 504 Potter. V. R.. 264 Prokofyeva-Belgovskaya. A. A., 482 Puck. T. T., 180. 186 Pullman. B.. 510. 511. 516 Raacke. I. D.. 431. .434 Rabinovits, M., 424 Race. R. R.. 66 Radbill. C. I... 505, 507 Ragni. G., 297 Ramachandran. L. K... 367 Rapoport. I. A.. 276. 277 Rasmuson, M.. 213 Ravin. A. W.. 304 Reddi. K. K.. 367. 453. 456 Rees. M. W.. 351 Reich. E., 378, 381. 426, 434 Reichmann. M. E.. 445, 446 Renner. ().. 228 Rhoades. \1 M., 21, 28. 29. 127, 142. 175. 370. 371, 381 Rich. A., 429. 434. 445. 446. 511 Richardson. C. C, 291 Richter. A.. 367 Riley. M.. 463 Ris. H.. 275. 277 Risebrough, R. W.. 433 Robinson. E. A.. 416. 420 Roeser. J.. 327. 329 Roger. M.. 304 Roizman. B.. 503. 507 Roscoe, D. H.. 452, 456 Rosencranz. H. S.. 276. 277 Rotman. R.. 344 Rownd. R.. 304 Rubenstein. I.. 353 Rubin. H., 352 Ruddle. F. H.. 12 Rudkin, G.. 493 Rudnick. D.. 79 Ruhland. W.. 381 Rupert. C. S., 350. 352 Russell, L. B.. 162, 176, 486, 490 Russell. W. L., 88 Ryan. F. J., 29 Saez. F. A.. 28 Sagan, C, 516 Sager. R., 373. 378, 381 Saksella, E.. 389, 390 Salser, J. S.. 458. 462 Salyers. A. A., 277 Sampson. M.. 493, 499 Sanchez, C 460. 463 Sandeen. G.. 496, 499 Sander. C. 303 Sandler, L.. 388. 390 Sang. J. H.. 86. 88 Sanger. R.. 66 Sarabhai. A. S., 445 Sarkissian. I. V.. 212. 213 Schalet. A.. 200, 398, 402, 481, 482 Schatz. G.. 378, 381 Schiff, J. A.. 373, 381 Schildkraut. C. L.. 291. 301, 304 Schrader, F.. 12. 379 Schramm, G., 365. 51 I Schreil, W. H. G., 293 Schull. W. J.. 40 Schulman. H. M.. 434 Schultz. J., 253 Schwartz. D.. 487 Schwartz. J. H.. 440. 446 Schweet. R.. 429 Seaman, E., 361 Seecof, R. L.. 381 Sekiguchi. M., 431 Setlow. J. K... 397, 402 Setlow. R. B.. 397 Shettles, L. B.. 109. 113 Shipp. W.. 367 Shirven. R. M.. 277 Sia, R. H. P.. 295 A uthors 545 Siddiqi. O. H., 300, 304. 440 Siebke, J. C, 290 Simon, E. H., 501 Simon, L., 366 Singer, B., 366, 367 Singer, M. F.. 441, 446, 462, 463 Sinnott, E. W.. 39. 539 Sinsheimer, R. L., 291, 340, 352, 363, 367, 368 Sinton, W. M.. 516 Slatis. H., 218 Smellie, R. M. S., 434 Smith, D. W., 162 Smith, M.. 449, 456 Smith, P. E., 83 Smith, R. A., 367 Snell, G. D., 83 Snoad, B., 28 Sonneborn, T. M., 373, 374, 380, 381, 454, 455, 505, 507. 508 Spahr. P. F., 429 Sparrow, A. H.. 186 Spencer, J. H., 286, 291 Speyer, J. F., 446 Spiegelman, S., 366, 424, 425, 426, 427, 428, 432, 433 Spirin. A. S., 367, 434 Spottswood, T., 462 Sobels, F. H., 186 Sokoloff, L., 505, 507 Sorieul, S., 503, 507 Spassky, B., 246, 250 Spector, W. S., 13 Spiess, E. B., 213 Sprague. G. F.. 29, 213 Stadler, L. J., ix, 187 Staehlin, T., 442 Stahl. F. W., 271, 274, 277, 343, 353, 402 Stanier, R. Y., 337 Stanley, W. M., 365, 367, 368 Stebbins, G. L., 250 Steffenson, D. M., 155, 162, 276 Steinberg, J., 353 Steiner, R. F., 264 Stent, G. W.. 277, 337, 338, 351, 352, 353, 463 Stern, C, 40, 60, 79, 128, 194, 213, 219, 490 Stern, H., 493, 499, 503, 507, 508 Stevens, A., 426, 434 Stevens, J. G., 452, 456 Stocker, B. A. D., 334 Stoeckenius, W., 377 Stone, W. H., 504, 508 Stone, W. S., 239 Strauss, N., 463 Streisinger, G., 344, 351. 353, 402 Stretton, A. O. W., 445 Strickberger. M. W.. 29 Strickland, W. N.. 24 Sturtevant, A. H., 103. 113, 145, 146. 228. 474. 482 Sueoka, N.. 277. 288. 327. 329, 503. 508 Sugiyama, T., 366 Suomalainen, E., 162 Susman, M., 501 Sutton, W. S., 28 Swanson, C. P.. 13, 28 Swartz, M. N., 402 Synge, R. L. M., 516 Sypherd, P. S.. 463 Szilard. L.. 391, 402, 421 Szybalska. E. H., 297, 304 Szybalski, W.. 297. 302. 304 Takahashi. I.. 452. 456 Tal. M.. 431. 434 Tamm. I.. 366. 367 Tarmy. E., 361 Tatum, E. L., 310, 315. 408. 420 Tax, S.. 516 Taylor. A. L., 320, 321, 324, 329, 388, 390 Taylor, 1. H., 290. 401, 420, 455, 485, 489 Teissier, G., 370 Temin, H. M., 455. 456 Temin. R. G.. 199, 218, 226 Tergazhi. B. E., 402 Tessman. I., 402. 501 Therman. E., 162 Thoman, M. S., 324, 329 Thomas. C. A., lr.. 353 Thurline, H. C. 115 Todaro. G. J.. 503, 508 Tolmach, L. J., 296, 302, 304 Trautner, T. A., 402 Ts'o, P. O. P., 303, 496, 499, 505 Tsugita. A., 365, 368 Tucker, R. G., 452, 456 Tuppy, H., 378, 381 Urey, H. C. 516 Ursprung. H.. 502, 507 Van Beneden, E., 28 van Schaik, N. W., 471 van Wagtendonk, W. J.. 374. 381 Vogel. H. 1.. 290. 446 von Ehrenstein. G.. 444, 446, 447 Von Halle, E. S., 481, 482 von Hippel, P. H., 496, 499 Wacker, A., 402 Waddington. C. H., 86, 508 Wagner, H. P.. 162 Wagner. R. P., 29. 420 Wahba. A. 1.. 443, 446 Wallace. B., 218, 227 Wallace. E. M., 23, 50. 103. 106. 151, 156 Warner. J. R.. 429, 434, 435 Watanabe, T.. 359, 361 Watkins, J. F., 504, 507 Watson, J. D.. 267, 268, 271. 277, 395, 433, 434 Watts-Tobin, R. J., 436 Weigle, J. J.. 336, 337, 351, 353 Weijer, J., 29 Weiler, E., 508 Weinberg, W.. 213 Weisblum. B., 447 Weiss. S. B.. 426. 432. 433 Weissmann. A., 14 Weissmann. C. 366. 368 Welshons. W. 1.. 481. 482, 490 Wettstein. F. O., 442 Wexler. I. B., 66 White. M. J. D., 162, 239 Whiting. P. W.. Ill, 113 Wiener, A. S., 66 Wildman. S. G., 367 Wilkins, M. H. F., 268, 434 Williams, R. C., 365, 367 Wilson. E. B., 28 Woese, C. R.. 447 Wollman. E. L., 319. 321. 323. 327, 328, 329, 335, 337, 338, 350, 353. 361 Wolstenholme. G. E. W., 516 Wright. S.. 86. 213 Wright. S. W., 485, 490 Yamada. T., 508 Yankovsky, S. A.. 424, 427, 434 Yanofsky, C, 349. 352. 410. 420. 445, 447 Yeh, M., 485, 489 Yoshikawa, H.. 327, 329 Young, 1., 365, 368 Zamecnik, P. C., 424, 434 Zamenhof, S.. 402 Zelle, M., 313 Zetterqvist. O., 506, 507 Zichichi, M. L.. 351, 353 Zinder, N. D.. 318. 330, 334, 337, 338, 363, 365, 367, 440, 446 Zubay, G., 427, 434 SUBJECT INDEX Stir pic niv Ms not indexed; />4 hemizy gosis, 93 hemoglobin, 424. 431, 435. 440, 442, 445 \ 412 4I(> \ 416 417 huvlicmic.il genetics of, 412- 41S chains. 111. 117 I . 417 molecular evolution of. 418- 419 hemophilia. 95, 1 19, L39 Hemophilus, 2l>7. 350. 397 hermaphroditism. 1 I 1 heterocaryon, 27 heterochromatin, 156, 384. 388, 4~4. 481. 484. 486 heterocytosome, 378 heterogamety, 95 heterogenote, 332 heteroploidy, 151-155 heterosis (hybrid vigor). 208- 212. 236 heterozygosity, enforced. 229 and inbreeding, 206-208 heterozygote (hybrid), 35 hexagonal plate. 339 hexaploid, 151 Hexaptera, 149, 150 hexasomic, 158 Hfr, 318, 319, 320. 321, 322, 323 histidine, genetics of, 333 histogram, 535 histories, 252. 379. 496-498 histochemistry, 260 Holmes rib grass. 365 homogenote. 338 homogentisic acid. 404. 405 homopolymer, 288. 426 homozygote, 34 hormones, 83, 107. 492, 505 horse. 41 2 host range of phage. 343 "hot spots" of mutation. 392 human, cell line transformed. 297 chromosomes. 10. 159, 160 DNA base content, 266 genetics. J7 J9, 107-110. 217- 225 aneusomy, 158. 159, 160 map ot X. 130 and radiation. 180, 222-226. 222 twins, 36, 38. 74-79. 76 Inbrid. interspecific. 246 249. 251 inviabilit) and sterility. 245 vigor (heterosis). 208 212. 236 hydrogen bond. 209, 284 bydroxymethylase, 451, 501 11\ menoptera, 110 111 hyperploidy, 168 hyphae, 26, 27 hypomorph, 194. 487. 488 hypoploid, 167 hypostasis, 52. 74. 103 hypoxanthine, 2S4. 396 imaginal disc (anlage), 82 imago. 24 imidazole ring. 253 imino form. 395 immunity, bacterial. 331. 350, 360 inborn errors of metabolism. 405 inbreeding, 206-208 incipient species, 244 indiscrete variables. 534-537 inducer, lactose as. 458 induction, embryonic, 84 infection and sex, 3 17 inhibition, end product, 457 inosinic acid pyrophosphorylase. 297 insulin. 4. 220 integration, 298-300. 300, 322, 355 intelligence. 79 interference, 136 interferon, 454 interphase. 8. 9, 17, 18 intersex. 103-105, 103 introgression, 249 inversion. 168, 169, 170, 173, 237 in natural populations. 242- 243, 243 X. 196 iodine-131, 227 iojap, 373 ionization, 179 isoalleles, 59-60, 384 isogenic strains, 70 isolation, reproductive. 243-249 limson weed (see Datura) juvenile amaurotic idiocy, 204- 205. 205, 219 kappa. 373, 374-376 karyosomes (chromatin knots), 484 karyotype. Drosophila, 95, 216, 236 human. 160, 161 kernel. 25, 26, 386 keto form. 254 killers. 374-375 kinetoplast and kinetosome, 379. 513 Klinefelter*s syndrome, 108 l.iu segment. 457-459 lactic dehydrogenase. 362 lambda particle in Paramecium. 374 larkspur (Delphinium). 248, 249 larva. 24 law of parsimony (Occam's rule). 33 lawn, bacterial. 308 lysozyme, 339 learning and RNA, 504 leprosy. 227 lethal equivalents, 218 lethals, 69, 70, 80, 183, 195-197, 216. 217 balanced. 229, 230 conditional. 501-502 leptonema, 16, 20 leucoplasts, 370 life cycles. 23-27 life on other planets, 514 lily, 16, 17, 19, 503 Limnea (snail), 503, 508 linkage, and crossing over. 116- 126 groups, limitation in. 136 map ( see maps ) . 131 liver cells in tissue culture, 500 loads, balanced vs. mutational, 218-219 locus, 122 locust DNA base ratio. 266 lysis and lysogen, 331 lysosome. 7 lytic cycle. 341 macronucleus (meganucleus), 374 maize (Zea; see corn) malaria. 209 male sterile, 216, 217 manganese. 394 manifold effects (see pleiotro- pism ) maps, chromosome, 185 crossover (or linkage). 132, 139, 140, 141, 142, 143, 144 genetic, of E. coli, 314, 324- 326, 351 of lambda. 350 of T4. 343, 346-347 Mars. 514. 515 master plate. 307. 308, 309 mate-killer. 374. 454 mating, assortive. 206 nonrandom, 205-208 reaction. 375 reciprocal. 32 sib-, 207 Maxy technique, 198 mean, arithmetic, 63 Subjects 551 measles, 78, 389 medulla. 107 megasporocvte and megaspores, 25, 26' meiosis, 15-23. 156. 157 and gene pairs, 47, 48 in Neurospora, 124, 125, 126 meiotic drive, 388 Melandrium, 115 Melanoplus. 493 melanotic tumors and episomes, 370 merogenote, 320 meromixis and merozygote, 320 messenger RNA (mRNA), 425- 431, 430 (see also RNA) metagon, 454. 5 1 3 metamorphosis. 24 metaphase, 8, 9, 17, 18, 20, 21 methyl green stain, 261 methylation, DNA and RNA, 429 micronucleus, 374 microspectrophotometry, 261 microspores and microsporocytes, 25 migration, 203, 242 mistakes, incorporation and rep- lication, 398 mitochondria, 7, 377, 378, 505, 513 mitomycin C, 396 mitosis, 5, 6, 8, 9, 10 radiation effect on, 184-185 regulation of. 503 mitotic rate and mutations, 222, 383 Modulator, 466-469 molecular evolution of hemo- globin, 418-419 mollusc chromosome, 496 monad, 16 monoecious, 25 monohybrid, 45 monomers, 280 monosomic. 107-108, 156 moon, 515 morphological barriers, 245 mosaicism, 106-108, 106, 161, 370-373, 371, 474 erythrocyte, 504 functional, 485 mosquito, 246 mosses, 1 1 1 moths, 95, 107 mottled vs. sectored plaques, 344 mouse, 69, 83, 84, 88, 107-110, 1 14, 479, 502 aneusomy, 160 crossover map, 140-141 heart mitochondria, 377 position effect in. 486-487 tissue culture. 503 multigenic (polygenic) traits, 62-65 muscular dystrophy. 485 mutagens and antimutagens, 192, 225. 391-392 mutant, 5 (see also gene; muta- tion) bacterial, isolation of, 294 constitutive, 459 detection of. 149-151, 195-198 detrimental nature. 193-194 lethal (see lethals) in natural populations, 216, 217 pre- or postadaptive, 306-309 sub- and supervital, 216, 217 suppressor, 437-440 temperature sensitive, 501 visible, 198 mutation (see also mutant), 5 in bacteria, 306-309 defined operationally, 399-400 and evolution, 220-221 frequency, 192, 204 gene, 190 genetic control of, 383-389 germinal, 222-225 intergenic, 191 molecular basis of, 391-400 point. 189-198, 345, 392-399 and populations. 202 repair from, 397 rate, 192, 193, 224-225, 309, 392 reverse, 393, 468 somatic, 221-222 at specific loci, 192 spontaneous, 179, 192, 224 sub-nucleotide, 394-399 time of. 193 and ultraviolet, 396-397 whole nucleotide, 393-394 mutational damage, 221-225 mutational "hot spots." 392 mutational spectra, 192, 392-393 mycelium, 26 Mycobacterium, 452 myoglobin, 418-419 natural selection, 202-205, 245, 513 nearest-neighbor analysis, 286- 288, 397 Neisseria, 297 neomorph, 194 neurons and RNA, 505 Neurospora (bread mold), 26. 27, 124, 125, 126. 408-410 bibliography. 29 crossover map, 144 mitochondria. 377, 378 newt, 493, 495 Nicotiana (tobacco), 60-62 nitrous acid as mutagen. 396 nondisjunction, 95-98, 159, 384, 475-476 nonlysogenic (sensitive) strain. 331 normal curve. 534 nuclear body in E. coli, 293 nuclear membrane, 6, 7 nuclear transplantation. 502 nucleic acid, 253 as genetic material (see genetic material; DNA; RNA) hybrid, 425 methylated, 429 subservient role of, 512 nucleolus. 6, 7, 427, 493, 495, 496 organizer. 11, 427 nucleoplasm, 6 nucleoprotein. 263, 496-498 nucleoside, 257, 258, 263 phosphate kinases, 450 nucleotide, 257, 263 sequence in TMV, 366 nucleus, 6, 7, 9, 25, 26 null hypothesis, 526 nullosomic, 159 nurse cells. 24, 492, 494 octoploid, 151 ocular albinism, 490 Oenothera (evening primrose), 151, 228-236. 228, 481 oligonucleotide, 287 ommatidia, 474, 476 one gene one enzyme, 408-410 one gene one polypeptide, 410- 418 one gene one primary effect, 407- 408 onion, 9, 496 oocyte, 24, 160, 223, 493-496 oogenesis, 24 oogonia, 24 operons, 457-462, 469-470, 489 514 Ophryotrocha, 1 1 1 organic bases, 253-256 origin (O). point of. 320 ovaries and ovarioles, 24 oxidase, homogentisic, 406 oxygen and mutation. 182 ozone. 5 1 1 P,. PL„ 32 pachynema. 16, 17, 19, 173 panmixis, 205 paracentric. 166, 167 parahydroxylase, 407 INDEX Paramecium, 373, 374, '575, 376. 454. 505 506 parameter, 520 paramutation, 470 paranemic and plectonemic, 268. 49h pansexuality . 504 Parkinson's disease, 505 parthenogenesis, 152 pea, garden, 10, 31, 42. 116 118 gene action in. 497 sued. 1 19 pedigree, 73, 95 ot causes. 72 method. 37 symbols, 36 penetrance and expressivity. 72- 74 Penicillium, 504 pentasomic, 158 pentose. 257 pericarp. 25, 26 variegation. 465-469 pericentric. 167, 237 perifertilization stages, 193 peroxide. 192. 360, 512 persistence. 219 perspiration. 49 1 petites. segregational or vegeta- tive. 378 phage {see bacteriophage) phagocytosis, 298 phenocopy. 4. 317. 358. 439-440 phenogenetics. 79 phenol. 364 phenotype. 3. 52 phenotypic ratios. 49-52. 49 phenylketonuria. 205, 206. 406- 407 phenylpyruvic acid, 407 phen\l thiocarbamide (PTC), 214 phocomelia. 4, 80 phosphate, pyro- and ortho-, 279, 450. 451 phosphodiesterase, splenic, 281- 282. 285 phosphorous-32 as mutagen, 395 photoproduct. ultraviolet, 396 photorecovery, 191, 397 pigeon chromosome. 496 pine. 251 pinocytosis, 7. 298 pistil. 26. 61 planetary research. 515 plaque (clearing). 342-344. 342 plasm. iblast, 504 plastids, 370 373 pleiotropism. 70 72. 70. 404, 406 ploidy. 151 plus-minus test. 5 30 pok\ in Neurospora. 378 Pneumococcus I Diplococcus), 295. 452 point mutants and phenotype, [93 195 polar nuclei. 24. 25, 26 polariu. strand. 2(>(). 270, 271 poliomvelitis. 77, 78. 363. 453 pollen. 25. 26. 60-62 poly (A + U). 289 polj UiA + T). 289 polyamines. 339 Polydactyly. 72. 73 polydeoxyrihotide, 259. 260 polykaryocytosis, 503 polymer. 260 polymorphism. 218. 241 polvnemv (polyteny). 153, 154, 155. 383. 481 polynucleotide, 263 phosphorylase, 431. 441. 442 polypeptide, amino (N-) end, 429 carboxyl (C-) end. 429 and gene action. 404-419 synthesis and RNA. 423-431 polyploidy. 151. 383 polyribotide. 263 homo-. 443 mixed. 441 polysaccharide, 295. 296 polysomes (polyribosomes, ergo- somes). 430, 431. 435 polysomic. 157, 158 population genetics. 201-212, 216-225 population mean, 534 population method. 37 position effect, 385, 468 in Drosophila. 473-481 potato, 246-247 Potentilla (cinquefoil ). 243 poultry. 55, 56, 68, 79-81, 80, 81, 93. 94, 266 power of the test, 537-538 precursors determined. 405, 406. 407 presence-absence view, 57 probability. 134. 136, 524 proboscis. 111, 112 proflavin, 393 promotor, 358 prophage. 331 prophase. 6. 9, 16. 17, 18. 19, 20, 21 protamine, 252, 497 protein, 410, 496-498 protein evolution. 512 protein of TMV, 364, 365 proteinoids. 5 I 1 Proteus, 336 protoplast. 340 prototroph, 295 provirus. 338 Pseudomonas, 336, 435 pupa. 24 pure line. 1. 150 purine. 253, 254, 256 2-amino-, 399 dimethylamino-. 256 6-methylamino-, 255. 256 ribosides as antimutagens, 391- 392 puromycin, 424 putrescine, 339 pyrimidine. 253. 254, 255 quadrivalent, 248 quail. Japanese. 88 quantum, 192 Queen Victoria, 101 rll region. 344-349. 344, 392- 393, 436-440, 481 rabbit, 4, 57. 67. 71 races, 241-249 rad, 181 radiation, and genetic load. 223- 225 and mutation. 179-185 radish, 248 Rana (frog), 502 Raney Nickel. 448 rats. 487. 504, 505 ravelase (unravelase). 302 recessive, 35 reciprocal translocation. 170. 183, 234, 237-238 recombination, unit of. 302 reconstitution experiments. 365 red blood corpuscles, 57, 71, 413. 485 redundancy, 344 regression, 64. 65 regulatory, super-, mechanisms. 470 replica plating, 307-311. 308, 309, 310, 311 replication, 8. 10. 449 (see also DNA; RNA) replicative form (RF), 366, 426 repressor substance, 457 reproduction. 5 vegetative. 292, 293 reproductive barriers. 244-245 reproductive isolate. 206 reproductive potential (see fit- ness, biological ) repulsion and coupling. 118 resistance to colicins. 360 resistance transfer factors (RTF). 359 restitution, 164-166 reticulocyte. 425, 435. 440, 445 retinoblastoma. 204. 219 Rhesus (Rh) factor. 58 Subjects 553 Rhyncosciara, 492. 493 RNase (ribonuclease), 263, 364, 425, 431, 435 denatured and renatured, 435 RNA (ribonucleic acid), 262-263 and antibodies, 504 coding for amino acids, 436- 445 and differentiation, 504 double-stranded, 366 as genetic material, 263, 363- 366 and genetic recombination, 363 and learning, 504 messenger (mRNA), 438, 481, 489, 502, 503, 504, 505 methylation of. 429 polymerase, DNA-dependent, 426 RNA dependent, 366, 426 and polypeptide synthesis, 423- 431 in puffs, 492 replicase, 366 ribosomal, 426-427, 496 secondary structure, 428, 462 soluble (sRNA) (transfer or adapter), 427-428, 428, 429, 430, 444 synthesis regulation, 453-454 synthetase, 366 as template for DNA, 289 terminology, 262 ribonucleoprotein, 262 ribonucleotides in DNA, 289 ribose. D-, 257, 262 2'-deoxy-D-, 257 riboside, 263 ribosomes, 423-425, 424, 429- 431, 496 poly-, 430, 431 ribotide, 262 oligo-, 442-444 rickettsia, 376 roentgen (r) unit, 180 rotational substitution, 398 roundworm (see Ascaris) rye, 246 salmon DNA bases, 266 Salmonella, 297, 330-334, 333, 334, 358, 360, 388, 460-461 schizophrenia, 79 Sciara (fungus gnat), 492, 494 scission of DNA, 302 sea urchin, 152, 266, 500 seasonal barriers, 245 sedimentation unit (s), 423 seedlings, corn, 371-372, 372 segregants, bacterial, 313 segregation, 156 of alleles, 31-39 alternate, 231-233, 232 independent. 42-49 distortion, 386-388, 470 selection, 65, 202-205, 513 coefficient, 204 selective markers, 313-314 selfers, 388 self-sterility, 60-62, 61 semilethals, 216, 217 Serratia, 323 sex, 5 in bacteria, 312. 317-327 in Chlamydomonas, 376 chromatin, 484-486 chromosomes, 90-98 determination, 102-112 factor, 357-358 index, 105 -linked genes, 90-98 mosaics, 106, 107 ratio, human, 109-110 types, abnormal, 103, 104, 105 sexduction, 358 sexual barriers, 245 sexuality, importance of, 112 para-, 504 sheath, 339 Shigella, 358 sib matings, 207 sibling, 75 species, 245 sickle-cell anemia, 71, 72, 209, 413-416, 413 sigma in Drosophila, 370, 376 sigma (population standard devi- ation), 534 silks, 26 silkworm, 10 singlet or doublet Paramecium, 505 sister strands, 122 skin color, 51, 56 snail, 111, 494, 503, 508 snake venom diesterase, 282, 366 snapdragon (Antirrhinum), 54, 69 Solanum. 246-247 Solenobia. 152 soluble RNA (sRNA) (see RNA) somatic cell mating, 503-504 somatic line, 15, 221-222 somites, 83 sonication, 273, 280. 299 Sordaria ftmicola, 148 Spartina, 247-248 speciation, 203, 241-249 sperm, 5, 24, 180, 184 human, 109 nuclei, 26 spermadine, 339 spermatheca, 24 spermatid and spermatocyte. 24 spermatogenesis, 24 spermatogonia, 24 sperm iogenesis, 109 spindle, 6. 9 achromatic, 155 divergent, 383 orientation, 503 spirochaete, 1 10, 370, 376 sporophyte, 25 spreading effect, 473 standard deviation, 521-523 standard error, 536 Staphylococcus, 336 statistic, 520 sterility, 245 mutants, 216, 217 stillbirths, 208 strand recombination in vitro, 301-303 strand synthesis, direction of, 286 streaking method, 294 streptomycin. 299. 306. 307, 309, 376, 440 strontium-90, 224-225 style, 26 subvitals, 216, 217 suicide experiment, 327, 395 sulfanilamide, 393 superfemale and -male, 103-105, 103 superposition, 5 1 superrepressor, 458 surface-volume relations, 153 symbols, 116 sympatric, 243 synapsis, 16, 153, 298, 481 oblique, 474, 475 synergid nuclei, 25, 26 t test, 528, 536 tasters and nontasters, 214 tarweeds, 246 T-(tau) test, 536 tautomerism, 394, 395 taxonomy, and DNA hybrids. 302, 305 telomere. 164 telophase. 8. 9, 17, 18. 21 temperature and mutation, 192, 402 tempo preference, 78, 79 teosinte, 251 test cross, 46 testosterone, 505 tetrad, 16 tetrad analysis, 147-148 tetramers, 412, 504 tetraploid. 105. 151. 158, 484 tetrasomic. 157 thalassemia, 39 theophylline as mutagen, 391 55 i INDEX thiamin < vitamin B,), 108 109 thymidine, 2^~. 2.">s kinase. 503 thymidylate synthetase, 44i>. 501 thymidylic acid, thymine, 254, 2"> dimers, 396, J97 mutagen, 392 thyroxine, 505 tinv yeast colonies, 378 Tipula, 4l>3 tissue culture. 84, 180, 427 toad (Xenopus), 427, 4lM tobacco (Nicotiana), 60-62 I \l V i tobacco mosaic \ irus ). $40, 363 365, 304, 365, 442. 453 toes. 72, 7:5. "4 tomatoes. I 29 traits, qualitative \s. quantitative, 62-65 trans position. 349, 477. 478 transcription. 425. 454-455. 461. 512 differential, 502 and differentiation, 502-503 one-complcmcnt. 426 unit. 460 transduction, genetic. 330-336, 333, 334, 388 transfer RNA {see RNA) transformation, genetic, 295-300, 303, 336 transformer, 102, 103, 488 transition. 398 translation. 425. 461. 481. 512 translocation. 237-238 half-. 172. 237 reciprocal. 170, 171, 175, 183, 234 236, 234, 486 transplantation of eye anlage, 422 transposition. 174, 466, 469 trans\ersion. 398 triplet. 462 triploid. 104. 151 trisomic. 156 Triturus. 495 trivalent, 156. 248 tropical VS. temperate. 384 1 1 \ panosoma, 379 trypsin, 289, 412 413 triptophan, swithctase. 410 411. 442. 504 pyrrolase. 504 tuberculosis. 77, 78, 266 tulips. 247 Turner's syndrome. 107 twin method. 37 twin patches of kernels, 467 twins. 36, 38. 74-79. 76 in cattle. 504 typhoid, mouse, 330 ultracentrifugation, 273, 274 ultraviolet light (UV). 179. 360 -induction. 335, 350 and mutation. 192-193, 396- 397 on polyuridylic acid. 448 univalent. 16 unselected markers. 313, 314 uracil. 255, 262, 284 5-bromo, 284, 392-393, 453 in DNA, 452, 453 dimers, 397 5-fluoro, 431 tautomers. 394 variables, discrete, 520, 521-534 indiscrete. 534-537 variance, 63, 535 variegated (V-type) position ef- fect. 474, 486 vegetative nucleus. 26 ventral receptacle. 24 Venus. 514. 515 viability, 69, 70 {see also mutant) Vibrio, 336 Vicia jaba (broad bean). 453 virus {see also bacteriophage) herpes. 388. 503 infectious bovine rhinotrache- itus, 452 influenza, 363 lipo-. 500 measles. 389, 503 morphology. 339, 340 as mutagens. 388-389 polio-. 363. 453 polyoma. 445. 503 pox-, rabbit. 369 pro-. 455 regulatory effects. 503 reo-. 366 Rous sarcoma. 388. 455 simian. SV4„. 389. 503 tobacco mosaic (sec TMV) turnip yellow mosaic, 340, 363 vaccinia, 266. 426. 445 wound. 366 visible light. 179 visibles. 198 vitamin B, (thiamin). 408-409 war, nuclear, 225 water shrimp (Artemia). 152 watermelon. 25 1 Watson-Crick model of DNA, 267 {see also DNA) wheat. 155. 251, 493 white region, 478, 479 wild type. 23 woolly hair. 38 X-linked muscular dystrophy, 485 xanthine, 396 Xanthomonas. 297 Xenopus (toad), 427, 491 Xg locus, 139 X ray diffraction pattern, 267 X rays and mutation. 180-185, 180, 181 yak, 245 yeast, 266, 378, 396, 452 zygonema. 16 zygote, 5 zygotic induction, 335