An Inf&mQthnat S&rm of

K

X

Molecular Genetics Part I

Molecular Biology

tn Inlrrnalional Scries of Monographs and Textbooks

Edited by

NATHAN O. KAPLAN

Graduate Department of Biochemistry

Brandeis University

Waltham, Massachusetts

HAROLD A. SCHERAGA

Department of Chemistry Cornell I nil ersity Ithaca, New York

Volume 1. Harold A. Scheraga. Protein Structure. 1961

loliinir 2. Stuart A. Rice and Mitsuru Nagasawa. Polyelectrolyte Solutions: A Theoretical Introduction. With a contribution by Herbert Morawetz. 1961

Volume 3. Sidney Udenfriend. Fluorescence Assay in Biology and Medicine. 1962

Volume 4. J. Herbert Taylor (Editor). Molecular Genetics, i In two parts. I Part I — 1963. Part II in preparation.

Other Volumes in Preparation

Molecular Genetics

Edited hij J. HERBERT TAYLOR

Departments of Botany and Zoology

Columbia University

New York, New York

Part I

1963

ACADEMIC PRESS • NEW YORK AND LONDON

Copyright© 1963, by Academic Press Inc.

all rights reserved

no part of this book may be reproduced in any form

by photostat, microfilm, or any other means,

without written permission from the publishers.

ACADEMIC PRESS IXC. Ill Fifth Avenue New York 3. X. Y.

United Kingdom Edition

Published by

ACADEMIC PRESS IXC. (London) Ltd.

Berkeley Square House

Berkeley Square, London, W.l.

Library of Congress Catalog Card Number 62-22U

PRINTED IN THE UNITED STATES OF AMERICA

Contributors to Part I

CoRR.\DO Baglioni,* Department of Biology, Masm{:hmetts Institute of Technology, Cambridge, Massachusetts

Maurice J. Bessmax, McCollum-Pratt Institute, The Johns Hopkins University, Baltimore, Maryland

John Bishop. Department of Biochemistry, University of Kentucky College of Medicine, Lexington, Kentucky

R. J. Britten, Carnegie Institution of Washington, Department of Ter- restrial Magnetism, Washington, D.C.

H. Fr.aenkel-Conrat, Viru^ Laboratory, University of Califcyrnia, Berkeley, California

Ernst Freese,! Department of Genetics, University of Wisconsin, Madi- son, Wisconsin

Karl G. Lark, Department of Microbiology, St. Louis University School of Medicine, St. Lauis, Missauri

B. J. ^McCarthy, Carnegie Institution of Washington, Department of Terrestrial Magnetism, Washington, D.C.

R. B. Roberts, Carnegie Institution of Washington, Department of Ter- restrial Magnetism, Washington, D.C.

Richard Sch"«t:et, Department of Biochemistry, University of Kentucky College of Medicine, Lexington, Kentucky

J. Herbert Taylor, Department of Botany and Department of Zoology, Columbia University, New York, New York

Charles A. Thomas, Jr.. The Johns Hopkins University, Baltimore, Maryland

A. TsuGiTAji Virus Laboratory, University of California, Berkeley. California

Elliot Volkin. Biology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee

* Present address: International Laboratorv for Genetics and Biophysics. Xaples, Italy.

t Present address: Laboratory of Molecular BiologA-, National Institutes of Health— XIXDB, Bethesda. Man>-land.

t Present address : Research Institute for Protein, Osaka University, Osaka, Japan.

Preface

Althuugli gc'iK'tics is one of the youngest subdivisions of l^iology, it has come to occupy a central jiosition and to serve as a focal point not only for the interests of biologists, but also for diverse groui)s who use desig- nations such as biochemists, physical chemists, biophysicists, virologists, geneticists, and cytologists or cell biologists. A consideration of the groups involved indicates that the emphasis in this work has shifted from the classical approaches of genetics to a molecular view of the mechanism of heredity; therefore "Molecular Genetics" is an appropriate title. Many of the problems are being considered from the molecular point of view even though the structures and mechanisms involved cannot yet be precisely defined in molecular terms.

In a rapidly developing field of science, the ability to keej) informed of progress becomes a major problem. When the groups involved have such diverse backgrounds, interests, and locations among the various established disciplines, the task becomes even more difficult. Fortunately there is no dearth of scientific meetings and special symposia which consider various aspects of the ill-defined area to be included under molecular genetics. Many of these meetings and discussions result in publications. However, a need was felt for a more consistent and thor- ough coverage, which would highlight some of the interesting and, hope- fully, significant advances, provide a useful reference source, and allow our interested colleagues, who work in these or related areas, to bring themselves up to date. The material should also be useful for teaching and training at the graduate and upper undergraduate levels. Only time can render a verdict on the success of these aims.

As indicated by the designation of this as Part I, several imi)or- tant topics in molecular genetics have been missed or touched lightly. Although plans are still indefinite we hope to fill the gaps in a second and possibly third volume. Some of the most exciting concepts are still not sufficiently ciystallized to warrant recording even in an advanced work such as this. As some degree of stabilization occurs, and willing and able writers are found, we trust the gaps will be closed. Nevertheless, a unity and narrative form were kept in mind in planning the volume. The first five chapters are concerned with DNA — its replication and organization into chromosomes, control mechanisms in its replication, and its mutability. The next two chapters deal with the biosynthesis of RNA, its role in translating and transferring the genetic information.

viii prp:facf,

and the sites of its synthesis, organization, and function. The last three chapters deal with the genetic aspects of protein synthesis and the tran- scription of the genetic code into functional entities. A few of the areas not covered will be obvious; for example, exciting conceptual develop- ments on control mechanism in the transcription of the genetic code arc touched lightly. The cell-free synthesis of UNA, which was advancing so rapidly as this volume was written, is yet to be covered in detail. Another topic, genetic recombination, which is of deep concern and great significance to geneticists, has been largely ignored, jicrhaps with good reason. A major conceptual revolution, with some indication that the topic can be treated in molecular terms, is now in the making.

The chapters were not intended as comprehensive reviews, but repre- sent personal accounts of developments best known to each writer, who in all cases was involved in making significant advances in the respective areas. However, there was an attempt to present a balanced view, to cite related findings and most of all to write a readable and up-to-date account. In a field which moves as fast as molecular genetics, publica- tions tend to be outdated before they are off the press. To circumvent this, in jiart at least, we hope we have covered some material of lasting importance, presented the newest developments known to us while writ- ing, and in some cases looked into the future with a bit of speculation that will hopefully prove stimulating and i^roductive.

The Editor is indebted to the contributors who paused in their research long enough to make their contribution and who responded to the stated objective of finishing tlie manuscripts together so that all of the chapters would have the same degree of timeliness, except for the limitations imposed by the nature of the material covered. For expediting the work and for skillfully managing many of the technical details, the publishers are to be given much of the credit. To all of those who contributed by supplying unpublished figures and data or made material in press available to the various authors, special credit is due. ]\Iany of them are acknowledged in the various chapters, but for contributions of unpublished material to Chapter II, I wish to thank Joseph G. Gall of the University of Minnesota, George D. Pappas and Philip W. Brandt of Columbia University, and Oscar Tj. Miller, Biology Division, Oak Ridge National Laboratory.

J. Herbert T.wlor

Contents of Part I

CONTRIBUTORS TO PART I PREFACE

Chapter I The Replication of DNA in Cell-Free Systems

Maurice J. Bessmax

I. Introduction 1

II. Synthesis of Deoxyribonucleotides 3

III. Phosphorylation of Deoxyribonucleotides 12

IV. Synthesis of DNA by Escherichia coli Polymerase 17

V. Synthesis of DNA by Calf Thymus Po^-merase 50

VI. Synthesis of DNA in Bacteriophage-Infected Escherichia coli ... 54

VII. Conclusions 60

Acknowledgments 60

References 60

Chapter II The Replication and Organization of DNA in Chromosomes

J. Herbert Taylor

I. Introduction 65

II. A Morphological Picture of tlie Chromosome 67

III. Replication of DNA 72

IV. Exchanges between DNA Subunits and Genetic Recombination . . 82 V. Sequences of Replication of DNA in Chromosomes 88

VI. Chromosome Reproduction in Relation to DNA Replication and

Protein Synthesis 89

VII. Molecular Models for Organization of DNA into Chromosomes . . 96

References 108

ix

coxtp:xts

Chaplcr III

The ()r<i;aiiizati()ii of DNA in Bactcriojjliage and Bacteria Charles A. Thomas, Jr.

I. Bacteriophage DNA 113

II. Bacterial DNA 141

Acknowledgments 148

References 148

Chapler IV

Cellular Control of DNA Biosvnthesis

Karl G. Lark

I. Introduction 153

II. The Role of DNA Precursons in Controlling DNA Synthesis ... 162

III. Role of Protein or RNA Synthesis in Controlling DNA Synthe.sis . . 173

IV. The Physical State of DNA as a Factor Controlling its Replication . 180 V. Other Factors Affecting DNA Biosynthesis and the DNA Cycle . . 190

VI. Summary, Conclusion, an<l Opinions 197

Acknowledgments 200

References 200

Chapter V

Molecular Mechanism of Mutations

Ernst Freese

I. Introduction 207

II. Structure, Dui)licati<)n, and Alteration of tlie Hereditary Material . . 215

III. Biochemistry of Mutagens and Theory of Base Alterations . . . 217

IV. The Extent of Point Mutations 248

V. Mutagenic Specificity 252

References 264

CONTENTS XI

Chapter VI

Biosynthesis of RNA in Relation to Genetic Coding

Problems

Elliot Volkin

I. Introduction 271

II. Evidence Implicating an RNA Species as a Carrier of Genetic

Information 273

III. Mechanism of Messenger RNA Function Based on in Vivo

Experiments 277

IV. Attempts at Isolation of a Messenger RNA 279

V. In Vitro Synthesis of RNA 280

VI. RNA Involvement in in Vitro Protein Sj-nthesis 281

VII. The Triplet Code 284

VIII. Summary 286

References 286

C/iapfer VII

Kinetic Studies of the Synthesis of RNA and Ribosomes

R. B. Roberts, R. J. Britten, and B. J. McCarthy

I. Introduction 292

II. Properties of Ribosomes in Vitro 299

III. Fractionation of Ribosomes 308

IV. Metabolic Pools 311

V. Kinetics of Incorporation into RNA of Ribosomes 315

VI. Kinetic Studies of Phenol-Extracted RNA 329

VII. Kinetics of Incorporation into Protein 331

VIII. Composition of Newly Formed RNA 335

IX. Discussion 342

X. Epilogue 350

References 350

Chapter VIII Protein Synthesis in Relation to Gene Action

Richard Schweet and John Bishop

I. Introduction 353

II. Formation of Amino Acryl-RNA Compounds 355

Xll CONTENTS

III. Steps Rpqiiirinp Participation of the Ribosome 362

IV. Information Transfer in Protein Synthesis 380

References 400

Chapter IX

Correlations between Genetics and Chemistry of Human Hemoglobins

CORRADO BaGLIONI

I. Introduction 405

II. The Separation of Human Hemoglobins by Chromatography

and Electroi)horesis 406

III. Normal Hemoglobins 409

IV. Abnormal Hemoglobins 421

V. The Control of the Synthesis of Hemoglobin 456

Acknowledgments 469

References 469

Chapter X

Contributions from TMV Studies to the Problem of Genetic Information Transfer and Coding

A. Tsu(;iTA AND H. Fraenkel-Conrat

I. Introduction 477

II. TMV-Protcin 479

III. TMV-RNA 487

IV. Viral Infectivity 489

V. Natural Strains of TMV 491

VI. Chemically Produced Strains 497

VII. The Coding Problem 509

References 517

AUTHOR INDEX 521

SUBJECT INDEX 540

Chapter I

The Replication of DNA in Cell-Free Systems^

MAURICE J. BESSMAN

McCollum-Pratt Institute,

The Johns Hopkins University,

Baltimore, Maryland

I. Introduction 1

II. Synthesis of Deoxyribonucleotides 3

A. Origin of the Deoxj'ribosides 3

B. Synthesis of Thymidylic Acid 6

C. Synthesis of 5-Hydroxymethyldeoxycytidyhc Acid (dHMP) . . 10

III. Phosphoiylation of Deoxyribonucleotides 12

A. Deoxynucleoside Monophosphate Kinases 12

B. Deoxynucleoside Diphosphate Kinases 16

IV. Synthesis of DNA by Escherichia coli Polymerase 17

A. Measurement of DNA Synthesis 17

B. Requirements of the Reaction 20

C. The Limited Reaction 23

D. Nature of the Primer 25

E. Nature of the Product 27

F. Primer-Product Relationships 29

G. Synthesis of Deoxyadenylate-Deox>'thymidylate and Deoxyguanyl-

ate-Deoxycytidylate Polymers 40

H. Substitution of Analogs 46

V. Synthesis of DNA by Calf Thymus Polymerase 50

VI. Synthesis of DNA in Bacteriophage-Infected Escherichia coli ... 54

A. Bacteriophage-Specific Polymerase 54

B. Glucosylation of 5-Hydroxymethylcytosine (HMC) 56

VII. Conclusions 60

Acknowledgments 60

References 60

I. Introduction-

Most organisms can synthesize all the constituents of the DNA macro- molecule from small molecular weight precursors which are also the raw

' Contribution No. 371 of the McCollum-Pratt Institute, The Johns Hopkins University, Baltimore 18, Maryland.

^ The following abbreviations have been used in this chapter : RNA, DNA : ribonucleic acid, deoxjTibonucleic acid; RNase, DNase: ribonuclease, deoxjTibo-

1

I MATRICF-: J. niv'^SMAX

materials for the synthesis of many other celhilar constituents. It is probably the synthesis of deoxyribose or a derivative of it which is the point of diversion from this non-sjiccific pool of precursors to the specific sequence of reactions culniiiiatinfj; in the jjolynierization of the deoxy- ribonucleoside triphosphates to form the DNA molecule. Since several of the later chapters refer to specific reactions in this metabolic chain,

CDP

UDP-

dHMP -^dHDP— -^ dHTP-

dCMP

dAMP

dGMP

-f

dUMP

dCDP dCTPmN

dADP

dGDP

dTMP -dTDP

dUDP

dATP

dGTP

dTTP

dUTP

DNA ^ glucosyl DNA

Fig. 1. Enzymatic reactions in E. coli leading to the sj'nthesis of DNA. ,

specific to cells infected with T-even phages. /WW, blocked in cells infected with T-even phages.

tlie current concepts of the pathways of enzymatic synthesis of deoxy- ribonucleotides and their phosphorylated derivatives will be discussed briefiy with the intention of providing an immediate reference for the ensuing chapters. Figure 1 shows several of the reactions suggested as intermediate steps leading to the synthesis of DNA.

In the cell-free system, the replication of DNA imi)lics the net synthesis of exact replicas of the "primer" molecules. For this reason, only those systems sufficiently iiurificd to dcmon.-^tiatc both net .synthesis

nuclease; AMP, ADP, ATP: adenosine monophosphate, adenosine diphosphate, adenosine triphosphate; CMP, CDP, CTP: cytidine monophosphate, etc.; GMP, GDP, GTP: guanosine monophosphate, etc.; HMP, HDP, HTP: hydroxj-methyl- cytidine monophosi)hate, etc.; IMP, IDP, ITP: inosine monophosphate, etc.; TMP, TDP, TTP: thymidine monophosphate, etc.; UMP, UDP, UTP: uridine monophos- phate, etc.; dAMP, dADP, d.\TP: deoxyadcnosinc monophosphate, etc.; DPN, DPXH: diphosphoi)yridino and its reduced form; TPN, TPNH: triphosphopyridine and its reduced form; F.\D, FADH;: flavin adenine dinucleotide and its reduced form; THFA: tetrahydrofolic acid; Pi, PPi : inorganic ortlu)i)hos])hate and pyro- phosphate ; tris: tris(hj'dro.\ymethyl)aminomethane.

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS d

and a requirement for primer DNA will l)e considered in this chapter. The influence of primer on the synthetic reaction is a necessary but certainly not a sufficient condition for replication since, as will be discussed later, there are some situations in which the synthetic DNA reflects the composition and structure of the added primer, and other primer-dependent reactions in which the product is clearly different from the primer.

II. Synthesis of Deoxyribonucleotides

A. ORIGIN OF THE DEOXYRIBOSIDES

Recent experiments with cell-free systems support the earlier indica- tions from tracer experiments in the whole animal that deoxyribosyl compounds are synthesized by a direct reduction of some ribosyl deriva- tives. Hammarsten et al. (1950) suggested that the N^^-labeled deoxy- cytidine they isolated from rat liver DNA arose by a direct conversion of the injected N^^-cytidine without cleavage of the ribosidic linkage, since it was known from the results of several laboratories that free cytosine is not a good precursor of DNA cytosine. Experimental evidence in support of this contention was furnished by the work of Rose and Schweigert (1953), who injected C"-cytidine, labeled in both the base and the ribose moieties, into rats and found that the ratio of counts in the base and deoxyribose in the DNA deoxycytidine was the same as the ratio of counts in the base and ribose of the injected nucleoside. Either cytidine was converted to deoxycj^tidine without cleavage of the ribosidic linkage, or cleavage did occur and the pool sizes of the free cytosine and ribose in the rat were exactly the same. The latter possi- bility seems too remote to have been seriously entertained. These results were confirmed by Roll et al. (1956) and Reichard (1957) and extended to include chick embiyo (Reichard, 1958, 1959), Ehrlich ascites cells (Edmunds, 1958) and Neurospo7'a (McNutt, 1958a). Cell-free extracts of chick embryo (Reichard, 1959, 1960), Escherichia coli (Reichard and Rutberg, 1960; Reichard et al, 1961), and Novikoff hepatoma (Moore and Hurlbert, 1960) were shown to carry out the reduction of ribosyl compounds to deoxyribosyl derivatives and, as is usually the case, it has been through the use of purified, cell-free systems that insight into the mechanism of the interconversion has become possible. Reichard et al. (1961) had suggested from isotope dilution experiments and kinetic data that the conversion of cytidine to deoxycytidine nucleotides occurred at the diphosi:)hate level and required ATP, ^Vlg'^^, and TPN. Bertani et al. (1961), using a fraction purified 100-200- fold from crude extracts of Escherichia coli. have presented clear evidence that the reduction occurs

4 MAIRICK .1. HKSSMAN

at the diphosphate level and show that optimal synthesis requires the participation of ATP, IMg**, and reduced lipoic acid. These results are presented in Table I. It is interesting to note that reduced lipoic acid is an essential component of the system and cannot be replaced by TPNH, DPNH, glutathione, ascorbic acid, 2,3-mercaptopropanol, or oxidized lipoic acid. The authors suggest that the cfTect of lipoic acid may be

TABLE I

KEyl IKEMENTS FOR dCDP FORMATION WITH A

PiRiFiED Fraction from Escherichia coli "•''

		Deoxycytidinc
		phosphates
		formed
Omission	Addition	(m^moles)
None	None	2.75
Tritium-labeled CDP	Tritium-labeled CMP (0.04 Aimole)	0.25
Tritium-l.-iholcd CDP	Tritium-labeled CTP (0.04 ^mnle)	0.7:5
ATP	None	O.OG
MgCl=	None	0.03
Lipoic acid	None	0.05
Lipoic acid	TPNH (O.H/xmole)	0.04

" From Bertani el al. (1961).

'' The complete reaction mixture contained 0.04 /imole of CDPH^ (3 X 10" cpm per Mmole), 0.5 ^mole of ATP, 1.5 /umoles of MgClo, 0.2 yumoles of reduced DL-lipoic acid, 2.5 Aimoles of tris-HCl buffer, pH 7.85, 0.1 /umole of mercaptoethanol, and 0.11 mg of frac- tion B in a final volume of 0.1 ml.

indirect, mediated through a flavin coenzyme, but they could not rei)lace lipoic acid by FAD or FADHo. The requirement for ATP suggests the participation of an "activated" intermediate during the reductive process, possibly a phosphorylated, pyrophosphoiylated, or adenylated derivative at the 2'-position (Bertani et al., 1961). This same enzyme fraction catalyzes the reduction of UDP to deoxy-UDP, again at the diphosphate level (Table II).

The specificity of this enzyme in respect to nuclco.^ide diphosphates has not been described and the question may be asked whether the same enzyme or similar enzymes account for the net formation of purine deoxyribonucleotides. Indications that the purine deoxyribosyl com- pounds are fonued directly from ribosyl derivatives have been reported from experiments in whole cells (Roll cf a/.. 1956; McNutt, 1958b) and in cell extracts (Reichard, 1960). This probably represents the major pathway for the fonnation of the purine deoxyribonucleotides, but the possibility of a contribution by the transdeoxyribosidase of McNutt

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

TABLE II

Formation of Deoxyuridine Phosphates from Different Uridine Phosphates"''

Deoxyuridine phosphates formed Substrate (niMmoIes)

UMP (28 mMmoles) 0 . 003

UDP (36 mMmoles) 0.340

UTP (38 niMmoles) 0.024

" From Bertani et al. (1961).

'' The reaction mixture contained the different C'^-uridine nucleotides (11 X 10" cpm per Mniole) together with 0.5 ^niole of ATP, 1.2 ^umoles of MgClo, 0.34 /xmole of reduced lipoic acid, 2.3 Aimole of tris-HCl buffer, i)H 7.85, 0.1 yumole of mercaptoethanol, and 0.16 mg of fraction B, in a final volume of 0.1 ml.

(1952) should not be overlooked. This latter enzyme catalyzes the following reaction:

Pyrimidine deoxyriboside + purine ^ purinedeox.\riboside + pyrimidine

The importance of this reaction in the de novo synthesis of purine deoxy- ribonucleotides remains to be evaluated.

A pathway for the formation of deoxyribose, not involving nucleoside derivatives, was described by Racker (1952), who demonstrated the condensation of acetaldehyde and glyceraldehyde-3-phosphate to form deoxyribose-5-iihosphate. This reaction, catalyzed by enzymes from E. coli or animal tissues (deoxyriboaldolase) could be coupled to the phosphodeoxyribomutase of Manson and Lampen (1951) and the nucleo- side phosphorylase of Friedkin and Kalckar (1950) and Friedkin and Roberts (1954) to form deoxyribonucleosides according to the following scheme:

aldolase

Acetaldehj'de + glyceraldeh3de-3-phosphate ;:^ deoxyribose-5-phosphate

mutase

Deoxyribose-5-phosphate ;=^ deoxyribose-1-phosphate

pho.iphorylase Deoxyribose- 1 -phosphate -|- base ^ deox^Tibomicleoside -|- phosphate

The equilibrium of the aldolase reaction is very unfavorable toward synthesis, suggesting that the significance of this reaction may lie in the degradation of deoxyribose-5-phosphate.

In analogy to ribonucleotide synthesis in which 5-phosphoribosyl-l- pyrophosphate is the active pentose intermediate (Kornberg et al., 1955; Remy et al., 1955), one could reason that a deoxyribose analog of this compound might participate in deoxy ribonucleotide synthesis:

O MAURICE J. HESSMAX

kinase Dt^oxyril)08e-5-pho8phato + ATP — >

5-phosphodeoxyribosyl-l-p>Topho8phat€ + AMP

5-Phospho(leoxyribosyl-l-pyrophosphate + base — » deoxjTiucleotide + p^Tophosphate

Although these reactions were looked for (Kornbcrg, 1957) no evidence for such a pathway was found.

B. SYNTHESIS OF THYMIUYLIC ACID

Tracer experiments in whole-cell systems indicated tiiat tlie general pathway for the synthesis of thymidylic acid involved the condensation of an "active" 1 -carbon unit, probably a derivative of folic acid, with a derivative of undine or cytidine (for review see Crosbie, 1960). Friedkin and Koiiiherg (1957) described a cell-free preparation from E. coli which converted dcoxyuridylate to thymidine triphospiiate in the presence of tetrahydrofolic acid (THFA), serine, and ATP. The properties of the system are described in Table III. The role of

TABLE III

CONVERSIO.N OF DeoXYURIDYLIC AcID TO THYMIDINE TrIPHOSPH.^TE"

Cpm

on Xorit

Experiment 1

Complete system* 1()50

zero time 24

minus THFA 47

- serine 480

- ATP 51

- extract 212

- Mg 30

- pyridoxal phosphate 1500

Experiment 2 Complete system

+ serine + THFA 187U

- serine + THFA 470 Complete system

+ serine + hydroxymothyl THFA lfi5()

- serine + hydroxymothyl THF'A 1210

° From Friedkin and Kornberg (1957).

'' The romplete system (^.325 /^liters; pH 7.4) contained 0.011 /xmoles of deoxyuridine- 5'-phosphate labeled with P^^ (2500 cpm), 2.3 ^nioles of ATP, 5.7 Mmoles of MgCh, 11.4 /iimoles of L-serine, 0.1 /xniole of THFA or A'^^-hydrox^-methyl THFA, 0.5 ^mole of cysteine, 2.5 ^moles of inorganic orthophosphate, 0.1 /tmole of pyridoxal phosphate, 21 /iimoles of tris, thymidylate kinase, and an extract of E. coli treated with Dowex-1- formate.

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 7

serine can be explained by the observation that the active 1 -carbon unit wliich ultiniately becomes the methyl group of thymine can be derived from the /3-carbon of serine (Ehvyn and Sprinson, 1954). The requirement for serine could be replaced by synthetic hydroxy- methyltetrahydrofolic acid which suggested that this was the active 1 -carbon intermediate generated in this system. These observations were extended to animal tissues by Phear and Greenbcrg (1957) and Humphreys and Greenberg (1958) and to a serineless mutant of E. coll by Birnie and Crosbie (1958). The bulk of the evidence impli- cates deoxyuridine monophosphate as the immediate precursor of thy- midylate in these systems. Blakely (1957), however, has suggested from his experiments that deoxyuridine rather than the corresponding nucleo- tide is the acceptor of the methyl group.

TABLE IV

Stoichiometry of Thymidylate Formation from Deoxyuridylate"'^

	Increase of	Dihvdrofolate	Thymidj'late	Dihydrofolate
Incubation	absorbancv	formed	formed
period	at 340 niM	(m/imole)	(ni/umole)	Thymidylate
10	0.100	17.2	19.2	0.89
20	0.180	28.2	28.2	1.00
30	0.211	33.0	33.4	0.99
40	0.225	35.3	36.0	0.98
50	0.232	36.3	37.5	0.97
60	0.235	36.7	40.0	0 . 92

" From Wahba and Friedkin (1961).

'' The incubation mi.xture consisted of 0.040 ^mole of dUMP (labeled with P'^, 80,000 cpm); 0.15 /xmole of dl, L-tetrahydrofolate; and enzj^me (0.2 mg of protein) in 1 ml of a mi.xture containing tris-HCl buffer, pH 7.4, 40 pinoles; 2-mercaptoethanol, 100 )umoles; ethylenediaminetetraacetate, 1.0 pinole; MgCU, 25 pmoles; and formaldehyde, 15 ^moles.

In addition to its role as carrier of the 1-carbon unit, tetrahydrofolic acid has been implicated as the reducing agent in the methylation reac- tion (Friedkin, 1959a; Humphreys and Greenberg, 1958). Direct evidence for this has been presented by Friedkin (1959a), who has reported the transfer of tritium from tritiated tetrahydrofolic to thymidylic acid. In addition, Wahba and Friedkin (1961) have recently described the stoichiometric oxidation of tetrahydrofolic to dihydrofolic acid during the synthesis of thymidylate from dcoxyuridylate. For each mole of thymidylate formed, one mole of dihydrofolate appeared. These results are reported in Table IV. Friedkin (1959b) has proposed a scheme for the methylation of deoxvuridvlate bv N^'-N'" methvlenetetrahvdrofolic

8

MAUKIOE .1. BESSMAN

acid. In Fig. 2, (looxyuridyiate is seen to cleave the methylene group at N= with the formation of a labile intermediate between this methylene group and the 5-position of the uracil ring. An intramolecular oxidation- reduction would yield thymidylate and dihydrofolate in equal amounts.

I

deoxyribose phosphote

H

I

C-glut

II

0

J^ CH.

I

deoxyribose phosphate

i

H

C-glut

OH

0

0' ^N*

I

deoxyribose phosphate

Fig. 2. Postulated mechanism for the synthesis of thymidylate. (From Fried- kin, 1959b.)

Humphreys and Greenberg (1958) independently proposed a similar mechanism for this reaction.

An interesting aspect of the synthesis of thymidylic acid is the increase in the capacity of E. coli infected with certain bacteriophages to convert deoxyuridylate to thymidylate. Flaks and Cohen (1957)

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 9

originally reported an increase in this activity in extracts of cells infected with bacteriophage T6 and suggested the name "thymidylate synthetase" since it was not known whether one or more enzymes was required to transfer the 1-carbon unit and cany out the reductive step (Flaks and Cohen, 1959b). These same investigators (1959b) demonstrated a 7-fold increase in thymidylate synthetase after infection with bacteriophage T2 and showed that the conversion of deoxyuridylate to thymidylate required formaldehyde and tetrahydrofolic acid; it did not proceed through 5-hydroxymethyldeoxyuridylate as an intermediate. Earlier ex- periments of Earner and Cohen (1954), in which they showed that a thymine-requiring strain of E. coli, IST", could synthesize thymine after infection with bacteriophage, can be explained by later experiments of these same authors (1959) in which they demonstrated that thymidylate synthetase (missing in 15T-) is induced by bacteriophage infection. This would support the idea that the infecting phage particle carries in the information for the induction of thymidylate synthetase, and Cohen (1961) investigated whether the enzyme foraied after infection is the same as the enzyme in the normal cell. A comparison of thymidylate synthetase in extracts of E. coli B, Bt" infected with T2 and Bt- infected with T5 showed that the enzymes from all three sources were inhibited to the same extent by varying concentrations of 5-flu6rodeoxyuridylate or D-arabinofuranosyl-5-fluorouracil-5'-phosphate, two powerful inhibi- tors of thymidylate synthetase. The active sites of the enzymes from these three sources are thus indistinguishable by this test. Recently, however, Greenberg et al. (1962b) have demonstrated that thymidylate synthetase from T2-infected E. coli has different chromatographic properties from the thymidylate synthetase of uninfected cells.

Cohen et al. (1958) have shown that 5-fluorodeoxyuridylate is a powerful non-competitive inhibitor of thymidylate synthetase, and at a concentration of 2 X 10"*^ M (a ratio of substrate to inhibitor of 1000) from 35-55% inhibition was observed. Heidelberger et al. (1957) have reported the anti-tumor activity of 5-fluorouracil and presented evidence for the inhibition of thymine synthesis in whole cells by this analog. The inhibition of thymidylate synthetase by 5-fluorodeoxyuridylate could explain this observation.

As mentioned previously (Section I,A), Bertani et al. (1961) have indicated that the conversion of uridylate to deoxyuridylate occurs at the diphosphate level. The bulk of the evidence indicates that the methylation of deoxyuridylate to thymidylate occurs at the monophos- phate level. This would mean that deoxyuridine diphosphate would have to be converted to deoxyuridine monophosphate before methylation. Bertani et al. (1961) have found that deoxy-UDP is phosphoiylated to

10 MAliRICE J. liESSMAN

(leoxy-l^TP in their system and that a specific UTPase is present which converts deoxy-UTP to deoxy-UMP and pyrophosi)hate. They suggest the following scheme for synthesis of deoxythymidylate in E. coli:

TiMP

T

UMP deoxyUMP

11 T

UDP -> deoxy UDP

11 11 -PP

UTP deoxy UTP

Another reaction which can jirovide deoxy-UMP rlirectly for deoxy- TMP synthesis is the deamination of deoxycytidylic acid. Scarano

(1958) has described a specific deoxycytidylate deaminase in sea urchin eggs and rabbit (Scarano and Tolarico, 1959) and Maley and Malcy

(1959) have extended this work to rat tissues and ciiick embryo. Keck et al. (1960) and Flaks and Cohen (1959b) reported the induction of a specific deoxycytidylate deaminase in T2-infected E. coli. Thus, in any of these systems the following pathway is possible:

deaminase

CDP -> dCDP ^ dCMP -^ dUMP-^TMP

Friedkin (1959b) has published an interesting account of the unique position of thymidylic acid in nucleotide metabolism.

Recently, Mandel and Borek (1961) reported that the methyl group of the thymine isolated from the ribonucleic acid of E. coli arises from the methyl group of methionine. This interesting observation would suggest that the small amount of thymine in RNA and the thymine in DNA are synthesized by two discrete pathways.

C. SYNTHESIS OF 5-HYDROXY.METHYLDEOXYCVTmYLIC ACm (dHMP)

The discoveiy that cytosine is completely replaced by tlio unique pyrimidine, 5-hydroxymethylcytosine (HMC), in the DNA of the T-even bacteriophages (Wyatt and Cohen. 1953) prompted an investigation of the enzymatic steps leading to the synthesis of this comi)ound. Tracer experiments with whole cells had implicated cytosine as a precursor of the pyrimidine ring and the /?-carbon of serine as a precursor of the hydroxymethyl group (Cohen and Weed, 1954) and eliminated methio- nine as the one-carbon donor (Green and Cohen. 1957). Flaks and Cohen (1957, 1959) first reported the synthesis of hydroxymethyldeoxycytidy- late in cell-free extracts of T6r-'-infected E. coli and showed that analogous to thymidylate synthetase the system required tetrahydrofolic acid and fomialdehyde. Tetrahydrofolic acid appears to act catalytically

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

11

since the amount of dHMP formed is greater than the amount of tetrahydrofohite j)resent. Dihydrofolic acid is inactive. The folic acid derivative presumably acts only as a one-carbon carrier in this system and has no function as a reducing agent as it does in thymidylate syn-

H

N. Ji HC-CH,-N^[ ^C-glut

^X^^N^ ^ II

OH N5 ^ n

NHg/

deoxyribose phosphate

H

OH

-N^r^C-glut

- I ^ ^ II

0

NH, ^^i

'HOH

N

I

deoxyribose phosphate

\

H HgNi^ Y CHg

^^ N^ H I'

NHj N^^^I-CHgOH

OH

O^N I deoxyribose phosphate

Fig. 3. Postulated mechanism for the synthesis of o-hydroxymethyldeoxycytidy- late. (From Friedkin, 1959b.)

thesis. Friedkin (1959b) has presented a hypothetical scheme for the reaction between deoxycytidylate and formyl tetrahydrofolic acid (Fig. 3). This scheme differs from the mechanism proposed for thymidylate (Fig. 2) in that no intramolecular shift of hydrogens occurs and hy- drolysis of the intermediate leads directly to the hydroxymethyl deriva- tive and free tetrahydrofolate. Deoxycytidylate hydroxymethylase, as

12 MAURICE J. bp:.ssman

the enzyme is called, is specific for dCMP and will not react with CMP, deoxycytidine, cytidine, or cytosine. It is distinct from thymidylate synthetase since it is much more stable in crude extracts of infected cells than the latter enzyme and it is not inhibited by 5-fluorode- oxyuridylate.

In respect to its distribution, deoxycytidylate hydroxymethylase has only been found in significant levels in extracts of E. coli infected with T2, T4, or T6 bacteriophages. It is not detectable in uninfected cells, in the viruses, or in cells infected with a virus which does not contain 5-hydroxymethylcytosine. Recently, Pizer and Cohen (quoted in Cohen, 1961), using very highly labeled formaldehyde in the enzyme assay, have shown that less than 1 molecule of active deoxycytidylate hy- droxymethylase is present per uninfected bacterial cell. These calcula- tions are based on a molecular weight of 64,000 ± 9000, and a turnover number of 1200 for the enzyme. It would appear then, that deoxycy- tidylate hydroxymethylase is either present in uninfected bacterial cells in an inactive form and is activated after infection or its formation is induced by injection of tiie viral DNA. Experiments designed to test these two alternatives tend to favor the latter, although no definitive statement can be made.

III. Phosphorylation of DEOXYRiBONUCLEOxroES

A. DEOXYRIBONUCLEOSIDE MONOPHOSPHATE KINASES

The name deoxyribonucleotide kinase has been given to a group of enzymes which catalyze the following reaction:

Deox>Tibonucleoside monophosi)hate + ATP ^

deoxyrihonucleoside diphosphate + A DP

This classification is misleading because it ascribes a degree of specificity to this group of enzymes which is not warranted. Both of the enzymes in this category, which have been purified and studied in detail (deoxy- cytidylate and deoxyguanylate kinases), catalyze the phosphorylation of ribo- as well as deoxyribonucleotides.

The occurrence of phosphorylated deoxyribonucleotides in animal tissues was reported by Potter et al. (1957), who identified the di- and triphosphates of deoxythymidine and deoxycytidine in the acid-soluble fraction of thymus glands. Interestingly enough, these authors found no evidence for the polyphosphates of the corresponding purine deoxyribo- nucleosides. Sable et al. (1954) observed that rat liver mitochondria actively oxidizing glutamate could fonn dADP and dATP from dA^NIP. The demonstration of adenylate kinase in liver mitochondria (Siekevitz

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 13

and Potter, 1953) suggests that Sable's observation is probably related to the reports of Klenow and Anderson (1957) and Klenow and Lichtler (1957) that adenylate kinase (myokinase) can use adenine deoxyribo- nuclcotides in place of ribonucleotides. This possibility is supported by the observation of Lehman et al. (1958) that heated extracts of E. coli |iartially retain their ability to phosphoiylate deoxyadenylate while losing their capacity to phosphorylate other deoxyribonucleotides. Heat stability is a characteristic property of muscle adenylate kinase (Colo- wick and Kalckar, 1943). Hecht et al. (1954) reported the phosphoryla- tion of dCMP to dCDP and dCTP in a cytoplasmic fraction of rat liver and found that dTTP was formed from dTMP in extracts of regenerating liver. Ochoa and Heppel (1957) have observed the phosphoiylation of (lAAIP and dCMP in extracts of Azotobacter. Bollum (1958) and Canel- lakis and Mantsavinos (1958) reported the enzymatic synthesis of all four deoxyribonucleoside triphosphates in extracts of regenerating rat liver, and Kielley (1961) has studied these reactions in mouse liver.

It is interesting that normal rat liver, which can catalyze the phos- l)horylation of dAIVIP, dCMP, and dGMP by ATP, has at most a very low capacity to phosphoiylate dTMP. The rapid increase in thymidylate kinase activity in regenerating rat liver has been demonstrated in several laboratories (Bollum and Potter, 1959; Canellakis, 1959; Canellakis et al., 1959; Smellie, 1960). This increase is inhibited by post-operative X-irradiation of the whole animal (Bollum et al, 1960). Weissman et al. (1960) and Smellie (1960) have studied the sequence of reactions:

Deoxythymidine ^ dTMP ^ dTDP -^ dTTP

They claim that there is a sequential induction or appearance of thymi- dine kinase, thymidylate kinase, and thymidine diphosphate kinase in regenerating rat liver and in an L-strain of fibroblasts grown in tissue culture.

Gray et al. (1960) and Smellie (1960) claim there is a factor in normal rat liver which inhibits thymidylate kinase when added to extracts of regenerating liver. It is heat-labile, nondialyzable, and has no effect on dAMP, dOAIP, and dCMP kinases. It will be interesting to learn more of this inhibitor which could play a role in regulating dTTP formation.

Bianchi et al. (1961) have reported that dTDP is not an intennediate in dTTP synthesis in mouse lymph nodes and leukemic spleen. They suggest that a pyrophosphate group is transferred from ATP according to the following reaction :

dTMP + ATP -> dTTP + AMP

14 M.MKirK J. nKSSMAN

This would 1)(' (he only instance in the synthesis of deoxyi ihoiiuflco-idi ti'iphospiiafcs in which a transfer of the intact pyropiiosphate supposedly occurs, and the I'caction should \)r studied with the i)urifie(l enzyme. The small amount of (ri'I)!' found in the cliromatogranis (Bianchi ct al., 19()1 I was attrihutcd to the hydrolysis of dTTP. The other possibility (which 1 fa\()n is that dTDP is an intermediate hut does not accumulate in the reaction mixture because another enzj'me, nucleoside diphosphate kinase, is present in the crude extract which converts dTDP to the triphosphate. This would be in keeping with their conditions of incuba- tion which include an ATP-generating system. The continued conversion of ADP to ATP by the generating system would tend to drive the reac- tion to the right, lience favoring the foi-mation of dTTP and decreasing the accumulation of dTDP:

ATP generating system

i dTMF + ATP ^ dTDP + ADP — + ATP

w

(ITTP + ADP —

We have evidence (^1. Hienhoif and AI. Bessman, in prejjaration ) that partially purified extracts of regenerating rat liver catalyze the i:)lios- |)liorylation of dTMP to the corresponding dii)hosphate.

The interesting observation (Hiatt and Bojarski, 1960) that intra- peritoneal injections of thymidine stimulate the formation of thy- midylate kinase in rat liver was explained by these same authors (Bojarski and Hiatt, 1960) as a stabilization of thymidylate kinase during preparation of the extract. They showed that preincubation of a liver extract resulted in a rapid decrease of thymidylate kinase activity wiiich could be prevented by including thymidylate (or thymidine at 10 times the concentration) in the preincubation medium. This information has been useful in preparing extracts for further fractionation.

Romberg (1957) and Lehman et al. (1958a) described the phos- phoiylation of dAMP, dCMP, dGMP, and dTMP by a partially purified extract from E. coli. It became apparent that more than one enzyme was responsible for the iihosphorylation of these four different nucleotides when Lehman et nl. (1958a) observed that heated extracts lost their ability to phosphorylate dCMP, dOMP, and dTMP while partially retaining their ability to phosphorylate dAMP. An indication that the other three activities w^re due to separate enzymes was the observation of Hurwitz (1959) that acetone fractionation resulted in relative eni'ich- ments of dCMP, dGMP, and dTMP kinases in different fractions.

The only report of the study of a ])urified kinase to date is that of

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 15

Maley and Ochoa (1958), who described the properties of the deoxy- cytidylate kinase in Azotobacter vinelandii. This enzyme was purified 800-fold over the crude extract and was shown to catalyze the following reaction:

dCMP + ATP -' (K^DP + A DP

The other three nucleotides (dGMP, dTMP, or dAMP) are not sub- strates for this enzyme.

The stoichiometry of the reaction is shown in Table \. In Expt. 1, a less purified enzyme was used and it can be seen that botli dCDP and dCTP were formed probably due to the presence of a nucleoside diphos- phate kinase which is known to be present in Azotobacter. In the more purified fractions (Expts. 2 and 3) only the diphosphate was formed. The same enzyme appears to catalyze the phosphorylation of the ribo compound, cytidylic acid, since the ratio of the activities toward the two substrates remains constant during the purification. Cytidylate is actually phosphoiylated at about 1.5 times the rate of deoxycj'tidylate by this enzyme at equal concentrations of the two compounds.

Stoichiometry	OF	TABLE V Reaction: ATP + dCMP ^	ADP + dCDP-	.h
Expt. No.	ATP		dCMP		ADP	dCDP	dCTP
1 2 3	-4.72 -4.54 +0.71		-4.35 -3.79 +0.70		+4.28 +4.24 -0.70	+3.05 +3.08 -0.73	+ 1.15 0 0

" From Maley and Ochoa (1958).

*" All samples contained 100 Mmoles of tris buffer, pH 7.5, and 20 yumoles of MgCh in a final volume of 1 .0 ml and were incubated at 30°C. The remaining components and conditions were as follows: Expt. 1: 9 Mmoles of ATP, 9 /^moles of dCMP, and 10 ng of enzyme of specific activity 60 with incubation for 15 hours; Expt. 2: 10 /xmoles of ATP, 5.8 /xmoles of dCMP, and 2.3 Mg of enzj'me of specific activity 175 with incubation for 3 hours; Expt. 3: 4.5 ^moles of ADP, 1.1 /imoles of dCDP, and 2.3 ^g of enzyme of specific activity 175 with incubation for 18 hours. Results are expressed as /imoles decrease or increase in reactants and products.

We have purified a deoxyguanylate kinase (M. Bessman, unpub- lished) from extracts of E. coli approximately 400-fold which appears to catalyze the following reaction.

dGMP + ATP ^ dGDP + ADP

Ion exchange chromatography reveals that more than 95% of the phosphoiylated product is the diphosphate. Again, ribo-GlVIP is also phosphoiylated but dCMP and dT]\IP are not.

Both these purified enzymes (deoxycytidylate and deoxyguanylate kinases) phosphoiylate the ribo- as well as deoxyribonucleotide. The

16 MAURICE J. BESSMAN

specificity seems to reside in the structure of the base and does not involve the sugar moiety. The name deoxynucleotide kinase is thus not appropriate in its strictest sense.

An interesting feature of the deoxynucleotide kinases of E. coli is that they are markedly influenced by bacteriophage infection. Activities of dTMP and dGAlP kinases rapidly increase after T2, T4, or T6 infec- tion, and dTMP, dGAIP, and dCMP kinases increase after T5 infection (Kornberg et al., 1959; Bessman, 1959). In addition, a new kinase, un- detectable in the uninfected cell, which catalyzes the phosphoiylation of hydroxymethyldeoxycytidylate is formed after infection with T2, T4, and T6 bacteriophages (Kornberg et al., 1959; Somerville et al., 1959). Recently it has been shown that the dGMP and dTMP kinases formed after bacteriophage infection are different from their counterparts in the normal cell (Bessman and Van Bibber, 1959; Bello et al., 1961a, b). They can be separated from the kinases of the uninfected cell by chroma- tography on DEAE cellulose columns and have different properties. For example, dGMP kinase from T2-infected E. coli is absolutely specific for the deoxyribonucleotide. It will not phosphoiylate ribo-GMP, whereas the enzyme from the uninfected cell will.

Purification of dGMP kinase from T2-infected cells has led to a concomitant purification of dTMP and 5-hydroxymethyldeoxycytidylate kinases (Bessman and Bello, 1961). Since the ratios of the three activities do not change over a 200-fold range in purification these authors investi- gated the possibility that all three activities resided in the same protein. Results of heat-inactivation studies and competitive inhibition measure- ments have led to the tentative conclusion that the three kinases are associated with one protein, although the data can be interpreted equally well by assuming the activities are associated with three veiy closely related proteins. Preliminaiy studies with the T5-induced kinases show that the dCMP, dGMP, and dTMP kinases, which increase after infec- tion, chromatograph in the same place on columns of DEAE cellulose, suggesting a close relationship if not identity of these three activities. Again, the phage-induced dCMP kinase differs from the dCMP kinase of normal cells by requiring deoxyribonucleotides as substrates. Ribo-GMP or CMP are not phosphoiylated. Thus, in this case and in the T2- infected cell, the name "deoxyribonucleotide" kinase is truly applicable. These are the only two members of the so-called deoxyribonucleotide kinases which are specific for the deoxy compounds.

B. DEOXYNUCLEOSIDE DIPHOSPHATE KINASES

The phosphorylation of deoxyribonucleotides most likely proceeds through the corresponding diphosphates to the triphosphate level. In the

I. REPLICATION OF DNA IN CELL-P^REE SYSTEMS 17

two cases cited above which were studied with purified enzymes (dCMP and dGMP kinases from E. coli) the diphosphate accumulates in the reaction mixture and no triphosphate is fomied. However, as discussed above, crude extracts of E. coli, regenerating rat liver, leukemic cells, Ehrlich ascites cells, thymus tissue, etc., form predominantly triphos- phates in the presence of ATP and a deoxyribonucleoside monophosphate. These two observations taken together mean there must be an enzyme (s) which catalyzes the phosphorylation of the deoxyribonucleoside diphos- phates formed as intermediates during the incubation. However, no specific deoxyribonucleoside diphosphate kinases have been reported, with the possible exception of the dTDP kinase alluded to in the report of Weissman et al. (1960). It is not unlikely that the nucleoside diphos- phate kinases described by Berg and Joklik (1953) and Krebs and Hems (1953), which catalyze the phosphorylation of the ribonucleoside diphos- phates, also react wdth the corresponding deoxy compounds:

(deoxy)micleoside diphosphate + ATP ;=i (deoxy) nucleoside triphosphate + ADP

Kirkland and Turner (1959) have shown that, in addition to the phos- phorylation of IDP and UDP originally demonstrated by the above investigators, purified extracts of peas also phosphoiylate CDP and GDP.

Davidson (1959), in a reinvestigation of the specificity of pyruvate kinase, has demonstrated that GDP can replace ADP in the reaction. It is possible that dGDP also could be phosphorylated in this system since Klenow and Anderson (1957) presented evidence for the replace- ment of ADP by dADP using crystalline pyruvate kinase from rabbit muscle. It is clear that the specificity of these enzymes with respect to the deoxy compounds should be investigated.

IV. Synthesis of DNA by Escherichia coli Polymerase

A. MEASUREMENT OF DNA SYNTHESIS

1. Acid Insolubility

The insolubility of DNA in dilute acids affords a convenient means of separating the macromolecule from low molecular weight deoxy- nucleotide derivatives which are soluble in acid solution. For the ensuing discussion, we shall define "DNA synthesis" to mean the incorporation of acid-soluble deoxyribonucleotide derivatives into covalent linkage in an acid-insoluble product. This is purposely a broad definition which includes "exchange reactions" if they occur as well as reactions in which

18

MAURICE J. BESSMAN

there is a net synthesis of polymer. Arguments for a narrov^ definition have an emotional rather than practical basis.

The most sensitive measure of DNA synthesis involves ooxyribonu- cleotides labeled with radioactive isotopes of carbon or ph«nhorus.

Base-deoxyribose-P'--P'--P'-

Any combinations of these labels may be present in the su-trate, but it should be borne in mind that the two terminal phosphoruiitoms will not become part of the DNA molecule. This technique i- extremely sensitive and has been used to measure veiy minute renrions. For example, the incorporation of 1 ^/j.mo]e of dcoxyribonucleode with a specific activity of 1 X 10^^ cpm per jumole would give 100 pm above background. By suitable wash procedures, the blank values en be kept to a relatively low level. For routine assays, specific act'ities two orders of magnitude lower (1 X 10''' cpm per /imole) may b(»mployed. For situations in which substantial amounts of DNA nve been synthesized, the acid-insoluble precipitate may be used or direct measurements of deoxyribose, phosphorus, or ultraviolet extirtion.

2. Viscosity Measurements

"The viscosity of dilute DNA sohitions cxci-cmIs that ol iiy other widely studied natural or synthetic substance. . . ." (Doty, 195V Pouyet

Itp

50

450

150 250 350

Time in minutes Fig. 4. Measurement of DNA synthesis by viscosity. (From Lehman. 1>!).)

.,//•

I. REPLICATION OF DNA IX CELL-FREE SYSTEMS

19

(1952) otained a value of 50 deciliters (dl(/gm for the intrinsic viscos- ity usin a Couette viscometer, and several other measurements of various ^NA samples put the range between 40 to 60. This proi)crty is very u>iul for kinetic measurements of DNA synthesis since the whole reactioiiiuay be carried out in a viscometer. Figure 4 demonstrates the increas( in specific viscosity of a reaction mixture j)rimed with calf thymus' )NA. This technique is applicable to systems in which the newly synthesjcd DNA represents a substantial increase over the initial DNA content..! the reaction mixture.

3. Meaurements of Hypochromwity

Kunz (1950) observed an increase in the ultiaviolet extinction of solutioF of DNA during enzymatic digestion with dcoxyribonuclease and utized this "hyi)erchromic eft'ect"' as a means to assay this enzyme. This hverchromic effect is due to the fact that the ultraviolet extinction of a nuleic acid preparation is le.ss than the sum of the extinctions of its indivifud nucleotides. The theoretical basis for this effect is under in-

220

240

mp

260

280

Fi 5. Absorption -sitectra of an equiniolar mixture (phosijliorus basis) of poly- adeny: acid and polyuridj-lic acid. The upper cur\'e refers to the mononucleotides obtami by alkaline hydrolysis of the mixture. The middle curve is that for the sepanuly measured polymer spectra. The lower curve is the measured curve for the nKture of polymers. (From Warner, 1957.)

18

M.\rRUE J. nESSMAN

there is a net synthesis of i)oIynier. Arguments for a nanowcr definition ha^■e an emotional rather than practical basis.

The most sensitive measure of DNA synthesis imolvcs deoxyrihoiiu- cleotitles hihelcd with radioactive isotopes of cai-hon or i)hos])liorus.

Base-deoxyribose-P'--P3--P'2

Any combinations of these labels may be present in the substrate, but it should be borne in mind that the two terminal phosphorus atoms will not become part of the DNA molecule. This technique is extremely sensitive and has been used to measure veiy minute reactions. For example, the incorporation of 1 /x^mole of deoxyribonucleotide with a specific activity of 1 X 10"^ cpm j^er /^mole would give 100 ci)m above background. By suitable wash procedures, the blank values can be kept to a relatively low^ level. For routine assays, specific activities two orders of magnitude lower (1 X 10'' cpm per /xmole) may be employed. For situations in w^hich substantial amounts of DNA have been synthesized; the acid-insoluble precipitate may be used for direct measurements of deoxyribose, phosphorus, or ultraviolet extinction.

2. Viscosity Measurements

"The viscosity of dilute DNA solutions exceeds that of any other widely studied natural or synthetic substance. . . ." (Doty, 1957). Pouyet

Itp

50

450

150 250 350

Time in minutes

Fig. 4. Measurcniont of DNA syntliesis by visco.sity. (From Lehman, 1959.)

I. REPLICATIOX OF DNA IN CELL-FREE SYSTEMS

19

(1952) obtained a valiR- of 50 deciliters (dl)/gin for the inti-in.sic viscos- ity using a Couettc viscometer, and several other measurements of various DNA samples put the range between 40 to 60. This property is veiy useful for kinetic measurements of DNA synthesis since the whole reaction may be carried out in a viscometer. Figure 4 demonstrates the increase in specific viscosity of a reaction mixture primed with calf thymus DNA. This technique is applicable to systems in which the newly synthesized DNA represents a substantial increase over the initial DNA content of the reaction mixture.

3. Measurements of Hypochromicity

Kunitz (1950) observed an increase in the ultraviolet extinction of solutions of DNA during enzymatic digestion with deoxyribonuclease and utilized this ''hyperchromic effect" as a means to assay this enzyme. This hyperchromic effect is due to the fact that the ultraviolet extinction of a nucleic acid preparation is less than the sum of the extinctions of its individual nucleotides. The theoretical basis for this effect is under in-

£(P)

	1 1 1 1 1 1 1 —	I	1 ■
12	\ Nucleotides ^..^^ /""^		-
10	^ r\ \ / \		-
8	\ \ Polymers -,. / / \ \ " /^~\\ "Y separately^/ \\		-
6	\ \ // /^ ^\	\	-
4	v^ // / Mixture of ^s^.^// / polymers	^	-
2	1 1 1 1 r 1 1	1	N.

220

240

m^

260

280

Fig. 5. Absorption spectra of an equimolar mixture (phosijhorus basis) of poly- adenylic acid and polyuridylic acid. The upper curve refers to the mononucleotides obtained by alkaline hydrolysis of the mixture. The middle curve is that for the separately measured polymer spectra. The lower cur\'e is the measured cur\"e for the mixture of polymers. (From Warner, 1957.)

20 MAUUICK J. UESSMAX

vestigation (cf. Shii^ar, I9()()i hut it is clear tliat there are two distinct contributions to the over-all hyjujchroniicity of DNA solutions. The first is maintained through the hydrogen bonding between the base pairs in the twin-stranded Watson-Crick model and accounts for approxi- mately 50% of the total hypochromicity. The remainder is maintained through the jihosphodiester bridges between the individual nucleotides (Sinsheimer, 1954; de Garilhe and Laskowski, 1956; Michelson, 1958). An illustration of this is shown in Fig. 5.

The decrease in extinction during polymerization of mononucleotides can be made the basis of an assay for DNA synthesis. The method is not suitable for measurements of small increments of DNA since the effect is relatively small and in addition to the absorption of the nucleotides themselves there is the contribution of the primer DNA. The method has been used successfully in measuring unprimed reactions and reactions in which extensive synthesis of polymer occurs (see Section IV,G).

B. REQUIREMENTS OF THE REACTION

1. Forward Reaction

The synthesis of DNA catalyzed by purified fractions of E. coli polymerase requires the presence of all four of the deoxyribonucleoside triphosphates oonimonly found in DNA, Mg ions, and a highly polymer- ized preparation of DNA (Bessman et at., 1958a). In Table VI, the

TABLE VI Requirements for Deoxynucleotide Incorporation' into DNA"*

System mMinoles

Complete system 0 . 50

Omit dCTP, dGTP, dATP <0.01

OmitdCTP <0.0I

Omit dGTP <0.01

Omit dATP <0.01

Omit Mg++ <0.01

Omit DNA <0.01

DNA pretrcated with DNase <U . 0 1

" P>om Bessman et al. (1958a).

*" The complete system contained 5 m^moles of dTTP (dTP'^PP, 1.5 X 10* cpm per A/mole), dATP, dCTP, and dGTP, 1 Mmole of MgCU, 20 fimoles of glycine buffer, pH 9.2, 10 fig of DNA, and 8 /ig of "polymerase" fraction V in a final volume of 0.30 ml. The incul)ation was carried out at 37°C for 30 minutes.

effect of omitting any one or combinations of these constituents is reported. Within the limits of this particular assay, omission of any one

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 21

of the deoxyribonucleoside triphosphates of adenine, cytosine, or gua- nine prevents DNA synthesis. Also, there is an absolute requirement for Mg ions and DNA. Preincubation of the DNA with DNase (5 /j,g, 37°C, 30 minutes) completely destroyed its priming ability. In a similar experiment, using a different P''--labeled nucleotide as an index of DNA synthesis, it could be shown that dTTP is also required.

2. Reverse Reaction

Concentrations of inorganic pyrophosphate equal to that of the individual triphosphates (1.6 X 10"^ M) have no effect on the synthetic reaction rate. However, at pyrophosphate concentrations 200-fold higher (3 X 10"'' M) the reaction rate is inhibited 50% and pyrophosphate is incorporated into deoxyribonucleoside triphosphates, as shown in Table VII. Unlike the synthetic reaction the reversal proceeds well in the absence of three of the four triphosphates. Omission of all four, however, reduces the rate of pyrophosphorolysis considerably.

TABLE VII Evidence for Reversal of Reaction"

P*2_(Jeoxynucleoside triphosphate'' Sj'stem (m/xmoles)

Complete system 3 . 47

Omit DNA <0.02

Omit enzyme <0.02

OmitdATP 2.34

Omit d ATP, dTTP 2.14

Omit dATP, dTTP, dGTP 1 . 34

Omit dATP, dTTP, dGTP, dCTP 0 . 20

Replace P^^P^z with Pi32 <0 . 02

Replace DNA with DNase-treated DNA <0.02

" From Bessman et al. (1958a).

* These values are corrected for a zero-time value of 0.05 myumole. The complete sys- tem contained 5 mMmoles each of dATP, dTTP, dGTP, and dCTP, 1 jumole of P'^pa (5 X 10^ cpm /miole), 10 ng of DNA, 2 Mmole of MgCl2, 20 /imoles of tris buffer, pH 7.5, and 0.8 fig of "polymerase" fraction VII in a final volume of 0.3 ml. After incubation at 37°C for 30 minutes the mixture was chilled and cold solutions of albumin (0.05 ml of 10 mg per ml), perchloric acid (0.25 ml of 1 A^), and "carrier" PP (0.5 ml of 0.1 M, pH 7) were added. The nucleotides were adsorbed on charcoal which was then washed, plated, and counted.

That pyrophosphate is actually a product of the reaction is shown in Table VIII. For each mole of dTMP incorporated into DNA, one mole of pyrophosphate is released. The stoichiometry of this reaction and the

22

MATHICK ,T. nKSSiMAN

H,C O

Poly- nucleotide

H,C

O— P— O— P— O— P-O

I /-

o o

-XTP

Yic. 6. Postulated mechanism for extending a DNA chain. (From Kornberg, I960.)

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 23

TABLE VIII Liberation of Inorganic Pyrophosphate"'*
Estimation	Control' (m^moles)		Experimental (ni/imoles)	A (m/xmoles)
DNA-C" incorporation P32P32 released Pi'2 released	1 2 7		25 22 11	24 20 4

nTPPP	r dTP ]
:dAPPP+™^-™^-	dGP dAP
n dCPPP	dCP

" From Bessman et al. (1958a).

* The labeled substrate was dTP32p32P-2-C".

' dATP was omitted.

requirements for the system have led to the following equation for the synthesis of DNA (Romberg, 1960) :

+ 4 (n) PP

The mechanism of the reaction is viewed by Koraberg (1960) as the nucleophilic attack of the terminal 3'-hydroxyl group of the DNA chain on the nucleotidyl phosphorus atom of the deoxyribonucleoside triphos- phate with the elimination of pyrophosphate. By this mechanism, the DNA chain is lengthened from the nucleoside and as depicted in Fig. 6. This mechanism is consistent with all the present data.

C. THE LIMITED REACTION (Adler et al., 1958)

In the previous section (IV,B,1), a requirement for all four deoxy- ribonucleoside triphosphates was shown to be necessaiy for DNA syn- thesis. However, using highly labeled substrates, a small but significant incorporation of single deoxyribonucleotides can be demonstrated (Table IX). This incorporation amounts to approximately 0.1% of the rate of the complete system and occurs with any of the four deoxyribonucleoside triphosphates. The reaction proceeds rapidly at first and then falls off, suggesting that there are preferred sites of attachment which become saturated. Several lines of evidence suggest that this reaction is not due to the presence of small amounts of the other deoxyribonucleotides in the enzyme preparation, in the added DNA, or in the labeled triphos- phate; and the possibility that the other triphosphates are formed by pyrophosphorolysis of the DNA during the incubation has been ruled out.

When a product of this limited reaction (containing a labeled deoxy- ribonucleotide) is isolated, washed free of acid-soluble deoxyribonucleo- tides and hydrolyzed with snake venom phosphodiesterase, almost all

^^^;!iss;,s^^--js6Hi^^

24

MAURICE J. BESSMAN

TABLE IX Incorporation ok Singi>k Deoxynucleotides into DNA°

Additions''

DNA-P" (m/xmoles)

dCTP + dGTP + TTP + dATP

dCTP + dGTP + TTP

dCTP + dGTP

dCTP

dCTP (DNA omitted)

3300

15.7

5.1

2.5

0.0

» From Kornberg (1959).

* The incubation mi.xture (0.3 ml) contained 5 niMinoles of dCP^'PP (7.1 X 10^ cpm per Ajmole), and of the other deoxynuclcoside triphosphates where indicated, 1 ^umole of MgCU, 20 iumoles of glycine buffer, pH 9.2, 10 ng of DNA, and 3 ng of enzj^me frac- tion VI. The incubation was carried out at 37°C for 30 minutes.

of the radioactivity is released before appreciable amounts of the DNA has been rendered acid-soluble (Fig. 7). Since this enzyme has been shown to hydrolyze polynucleotides (ribo- and deoxyribo-) in a stepwise fashion starting from the nucleoside end (Laskowski et al., 1957; Bow- man, 1957; Khorana et al., 1958; Singer et al., 1958), it would appear

lOO-i

Optical Density

60 Time in Minutes

Fig. 7. Action of venom phosphodiesterase on limit product. (From Adler et al., 1958.)

that the added deoxyribonucleotides are localized at the ends of the DNA chain. On the average, one deoxyribonucleotide is added per DNA chain. It may be asked whether each of the deoxyribonucleotides is added to specific DNA chains terminating with a specific base or whether dCTP for example will add to DNA chains tenninating in any of the deoxyribonucleotides. Results of nearest neighbor analyses

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

25

(see Section IV,F) favor the latter interpretation since it can be shown that each of the four deoxyribonucleotides attaches to chains ending in each of the bases. This analysis also rules out the possibility that the incorporation of single deoxyribonucleotides is due to the fomiation of homopolymers of the individual deoxyribonucleotides (poly-dCMP, for example), and demonstrates that the deoxyribonucleotide is attached to the DNA in covalent linkage through its phosphorus atom and not by some nonspecific adsorption or inclusion, etc.

What is the significance of the limited reaction? One possibility is that this is the prototype of the reaction that occurs when all four deoxyribonucleotides are present. If DNA replicates by adding deoxy- ribonucleotides to its end, these new units must loop back in some manner to allow the directing influence of the DNA chain to be exei-ted. But this imposes an additional requirement, that is some means of break- ing this loop when the DNA has been duplicated. We know that in reac- tions in which there has been a 20-fold synthesis over the DNA originally present as primer, the molecular weight of the product is not increased 20-fold.

Another possibility is that the chains of the double helix of primer DNA as isolated from natural sources are of unequal length (possibly through damage during preparation) and the limited reaction is the repair of these broken ends. This could mean that the sequence of deoxy- ribonucleotides added to the shorter ends would be dictated by the longer end. In keeping with this notion is the recent observation (A. Kornberg, personal communication) that in the limited reaction using the d-AT polymer as primer (see Section IV,G) dATP or dTTP is incorporated, whereas dGTP and dCTP are not. This would indicate that there is some specificity imposed by the primer even for the limited additions to the ends of the chains. This "repair" mechanism would not be pertinent to a scheme for the replication of DNA which does not involve a covalent linkage at any stage between primer and product.

A third possibility is that the enzyme which carries out the limited reaction is different from E. coli polymerase. This possibility seems remote.

D. NATURE OF THE PRIMER

So far, using highly purified fractions of E. coli polymerase, it has not been possible to synthesize a product containing all four deoxyribo- nucleotides in the absence of a primer. The priming DNA can be ob- tained from many sources, e.g., E. coli, calf thymus, salmon spenii, bacteriophage, crab testes, wheat germ, etc., but at present no chemically synthesized product has been able to substitute for the natural polymer.

n

24

MAURICE J. BESSMAN

TABLE IX Incorporation of Single Deoxynucleotides into DNA"

Additions'"

DNA-P32

(m/xinoles)

dCTP + dGTP + TTP + dATP

dCTP + dGTP 4- TTP

dCTP + dGTP

dCTP

dCTP (DNA omitted)

3300

15.7

5.1

2.5

0.0

" From Kornberg (1959).

* The incubation mixture (0.3 ml) contained 5 mAimoles of dCP^'PP (7.1 X 10' cpm per ^raole), and of the other deoxynucleoside triphosphates where indicated, 1 yumole of MgClo, 20 /xmoles of glycine buffer, pll 9.2, 10 fxg of DNA, and 3 ng of enz\'me frac- tion \T. The incubation was carried out at 37 °C for 30 minutes.

of the radioactivity is released before appreciable amounts of the DNA has been rendered acid-soluble (Fig. 7). Since this enzyme has been shown to hydrolyze polynucleotides (ribo- and deoxyribo-) in a stepwise fashion starting from the nucleoside end (Laskowski et al., 1957; Bow- man, 1957; Khorana et al., 1958; Singer et al., 1958), it would appear

lOOn

60 Time in Minutes

Fig. 7. Action of venom i)hosijhodiesterase on limit product. (From Adler et al., 1958.)

that the added deoxyribonucleotides are localized at the ends of the DNA chain. On the average, one deoxyribonucleotide is added per DNA chain. It may be asked whether each of the deoxyribonucleotides is added to specific DNA chains terminating with a specific base or whether dCTP for example will add to DNA chains terminating in any of the deoxyribonucleotides. Results of nearest neighbor analyses

I. EEPLICATION OF DNA IN CELL-FREE SYSTEMS 25

(see Section IV,r) favor the latter interpretation since it can be shown that each of the four deoxyribonucleotides attaches to chains ending in each of the bases. This analysis also rules out the possibility that the incorporation of single deoxyribonucleotides is due to the foraiation of homopolymers of the individual deoxyribonucleotides (poly-dCMP, for example), and demonstrates that the deoxyribonucleotide is attached to the DNA in covalent linkage through its i)hosphorus atom and not by some nonspecific adsorption or inclusion, etc.

What is the significance of the limited reaction? One possibility is that this is the prototype of the reaction that occurs when all four deoxyribonucleotides are present. If DNA replicates by adding deoxy- ribonucleotides to its end, these new units must loop back in some manner to allow the directing influence of the DNA chain to be exerted. But this imposes an additional requirement, that is some means of break- ing this loop when the DNA has been duplicated. We know that in reac- tions in which there has been a 20-fold synthesis over the DNA originally present as primer, the molecular weight of the product is not increased 20-fold.

Another possibility is that the chains of the double helix of primer DNA as isolated from natural sources are of unequal length (possibly through damage during preparation) and the limited reaction is the repair of these broken ends. This could mean that the sequence of deoxy- ribonucleotides added to the shorter ends w^ould be dictated by the longer end. In keeping with this notion is the recent observation (A. Kornberg, personal communication) that in the limited reaction using the d-AT polymer as primer (see Section IV,G) dATP or dTTP is incorporated, whereas dGTP and dCTP are not. This would indicate that there is some specificity imposed by the primer even for the limited additions to the ends of the chains. This "repair" mechanism would not be pertinent to a scheme for the replication of DNA which does not involve a covalent linkage at any stage between primer and product.

A third possibility is that the enzyme which carries out the limited reaction is different from E. coli polymerase. This possibility seems remote.

D. NATURE OF THE PRIMER

So far, using highly purified fractions of E. coli polymerase, it has not been possible to synthesize a product containing all four deoxyribo- nucleotides in the absence of a primer. The priming DNA can be ob- tained from many sources, e.g., E. coli, calf thymus, salmon spenii, bacteriophage, crab testes, wheat germ, etc., but at present no chemically synthesized product has been able to substitute for the natural polymer.

26

MArRICE J. BESSMAN

As shown in T:il)le VI, DNA treated with l)Nas;e loses its capacity to support synthesis. Other polymers which have been tested (Bessnian et al., 1957) and found ineffective as primers arc apurinic acid, turnii) yellow mosaic virus RNA, and a polythymidylic acid polymer (kindly provided by H. Gobind Khorana).

The important feature of the priming DNA seems to lie in its primary structure since many treatments designed to destroy the hydrogen bond- ing between the specific base pairs of double helices (Watson and Crick, 1953a) not only do not destroy priming ability but in some instances actually enhance it. In Table X, for example, native DNA is compared

TABLE X Enzymatic Synthesis of Viscous DNA from Non-Viscous Primers"

			Heated thymus	<AX174
		Thymus DNA	DNA	DNA
Reaction
Rate		1 X	2 X	2 X
Extent		Tenfold	Tenfold	Tenfold
Viscosity (dl/gm)
Primer		45	—	1
Product		41	20	22
Heated product		<1	—	<1
Hj'perchromic effect (	%)
Primer		30	—	10
Product		30	—	30

" From Lehman (1959).

to heat-denatured DNA and "natural" single-stranded DNA (Sins- heimer, 1959). It can be seen that the "denatured" or single-stranded DNA primes synthesis at twice the rate and to the same extent as the natural DNA. Furthermore, the products of the reaction are in all cases viscous even though the heated, and single-stranded DNA are non- viscous preparations.

An interesting observation (Lehman, 1959) is the improvement of calf thymus DNA as a primer by treating it with minute amounts of pancreatic DNase:

MgDNA Mg DNase

= 2 X lO'^

DNA treated in this manner primes syntheses at 2 to 3 times the rate of untreated primer. It is possible that this treatment provides more free 3'-hydroxyl groups for condensation with the deoxynucleoside triphos- phates. Alternatively, the pretreatment could provide weak points

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 27

throughout the molecule by hydrolysis of phosphodiester linkages, thus facilitating strand separation and enhancing priming efficiency.

With some highly purified fractions of E. coli polymerase, native DNA acts veiy poorly as a primer compared to heated DNA or DNA slightly hydrolyzed with DNase. This raised the question (Kornberg, 1960) of whether native DNA (double-stranded) can function at all as primer. To test this possibility rigorously, E. coli polymerase prepara- tions free of nuclease activity must be used; unfortunately, at present, none are available.

E. NATURE OF THE PRODUCT

1. Chemical Structure

The deoxynucleotides incorporated into DNA by E. coli polymerase are present in 3'-5' phosphodiester linkage (Bessman et al., 1958a). A product labeled with C"-thymidylate was digested with deoxyribonucle- ase and chromatographed according to Sinsheimer (1954). Two dinucleo- tides, cytidylate-thymidylate and thymidyl-thymidylate, were isolated and characterized as dinucleotides with typical 3'-5'-phosphodiester bridges. Thus it is clear that covalent linkages are fomied during the incorporation and the four deoxyribonucleotides add to each other. This rules out the possibility, for example, that homopolymers of the indi- vidual deoxyribonucleotides are the sole products. Besides having the nonnal internucleotide linkages, the product reacts with diphenylamine to give the characteristic color for deoxyribose (Dische, 1955). This is shown in Table XI, where the values for deoxy ribonucleotide incorpora-

	TABLE	XI
	Net Synthesis	OF DNA"
		Control		A
Expt. No.	Estimation (no enzyme)	Complete	(Mmoles)
1	P'^ incorporation	0.000	0.280	0.280
	Optical density	0.193	0.458	0.265
	Deoxypentose	0.187	0.395	0.208
2	Optical density	0.060	0.628	0.568
3	Optical density	0.052	0.583	0.531
4	Optical density	0.051	0.642	0.591
5	Optical density	0.041	0.889	0.848

" From Bessman et al. (1958a).

tion (as measured by P^-), ultraviolet extinction, and deoxyribose con- tent of a synthetic DNA are compared.

28

MAURICE J. BESSMAN

By these criteria, natural DNA and the j^roduct synthesized with E. coli polymerase are identical,

2. Physical Structure

Pliysical nioasiirenients of the DNA synthesized by E. coli polymerase became feasible when net synthesis of product was accomplished. In Table XI, several preparations of DNA are shown in which increases

TABLE XII Physical Properties of Enzymatically Synthesized DNA"

	Primer	Product	Heated at 100°C for 15 minutes
Property	Primer	Product
Sedimentation coefficient Intrinsic viscosity (dl/gm) Molecular weight	25 40-50 8 X 10«	20-25 15-45 4-6 X 10«	20 <1	14 <1

° From Lehman (1959).

up to 20 times that of the added primer indicate that at least 95% of the total DNA represents newly synthesized material. These products were

30	r		X	X X
25	-	x / •	"^ •
20	X / X /
15	/• XL X /			X Thymus DNA
10	X / x/« x/ x/« /•			• Enzymatic Product
5	/• /•	1		1 .J

Time in Hours

Fig. 8. Increase in ultraviolet absoriition of DNA upon digestion with pancreatic deoxyribonuclease. (From Lehman, 1959.)

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 29

analyzed by Schachman et al. (1958) and in eveiy parameter tested they resembled native calf thymus DNA. In Table XII the viscosities, sedimentation constants, and molecular weights for synthesized DNA and calf thymus DNA (used as primer) are compared, and it can be seen that the two materials are closely related. jMolecular weights of several preparations range from 4 to 6 million and the viscosity and sedimentation data indicate that the deoxynucleotide units are organized in the macromolecule to form relatively stiff rods with effective volmnes greater than would be expected from single polynucleotide chains with freedom of rotation at each internucleotide link. If the product is heated at 100°C for 15 minutes (Table XII) there is a sharp drop in viscosity with a relatively small decrease in sedimentation. When the product is digested with DNase there is a 30% increase in extinction, or hyper- chromic effect (Fig. 8). These latter two observations support the concept of an organized rigid arrangement of deoxyribonucleotides in the super- structure of the native molecule.

F. PRIMER-PRODUCT RELATIONSHIPS

As mentioned in the introduction, the tenn "replication" is taken to mean the synthesis of a DNA which is in every way indistinguishable from the primer. In this sense it would be better to use the word "tem- plate" rather than primer to describe the small amount of DNA required for the synthesis of more DNA since this implies a copying mechanism rather than initiating mechanism as the function of the added DNA although, the two functions go hand in hand, except in some instances mentioned later. "Primer" will be used throughout this discussion for convenience, but it should be borne in mind that the primer used in DNA synthesis is unique in biochemistiy in that it serves both as initia- tor and template.

So far we have seen that in respect to gross chemical structure and physical profile the synthetic DNA closely resembles the primer. We would like to be able to determine the exact sequence of deoxyribo- nucleotides in the product and compare this to the primer molecule; but, at present, techniques are not available for doing this even for relatively small polynucleotides. Valuable information has been obtained toward elucidating the fine structure of the synthetic DNA by determining its base composition and dinucleotide frequencies.

1. Base Composition

Two general methods have been used to determine the base composi- tion of the newly formed DNA. In the first procedure four reaction mixtures identical except for the radioactive deoxyribonucleoside tri-

30 MMRICE .T. BKSSMAX

phosphate are incubated separately, and the incorporation of label into the acid-insoliiblo product is measured in each. This method has the advantage of beino; applicable to very small syntheses and has been used to measure base ratios of products foi'med vciy early in the reac- tion. In the second method, i)roducts obtained from large-scale syntheses are hydrolyzed and chromatographed to determine base content directly. It is important that the primer represent only a small percentage of the total DNA analyzed so that routinely 10- to 20- fold net syntheses are employed in this procedure.

In Table XIII, the base compositions of several different DNA preparations are shown. First, in all cases, the percentage compositions of the bases in the products are very similar to these values for the respective primers. Second, the content of adenine approximates the con- tent of tliymine and the guanine and cytosine contents are about equal. Thus, purines ecjual pyrimidincs. These relationships are in keeping with the proposed structure for DNA.

With the use of the radioactive assay it was possible to show that the comi)osition of the product is established early in the reaction. In Table XIV it can be seen that when only 2% increase in DNA synthesis had occurred in the reaction primed with Mycohacterinm phlei DNA, the ratio of thymine to cytosine incorporation was equal to that of the primer. This ratio remained constant with u]) to 1637^ increase in DNA. The same is true for the reactions primed with calf thymus and Aerobacter aerogenes DNA. The increase in ratios during the latter part of the reaction can be explained by the formation of a polymer of deoxyadeny- late and deoxythymidylate fomied late in the incubation. This polymer will be discussed in Section IV,G.

An important consideration in establishing what factor (s) is respon- sible in determining the ultimate composition of the product is whether the relative concentrations of the individual deoxj^-ibonucleoside triphos- phates in the incubation mixture influence their relative concentrations in the product. That this is not the case is indicated in Table XV, where it can be seen that although the relative concentration of dTTP and dATP were varied singly and together over a 5-fold range, the composi- tion of the product remained constant. Thus, the determinant of product composition seems to lie exclusively in the structure of the primer DNA and the concentrations of the individual deoxyribonucleoside triphos- phates do not play an important role here. It is clear from experiments cited earlier that elimination of one of the triphosphates entirely from the incubation mixture effectively stops DNA synthesis except under spe- cial circumstances. Therefore, although the sfnictvre of the product is not influenced by the concentration of the substrates, the rate of syn-

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

31

o	o
+	4-
<	H

H	o
+	+
<	o

o

o

^ ~ -'si

-t< 00 O 00		o	CD ?D 05 O	00 '^ o o	o o
00 00		1	ss	T 7 O 00
o o O 00	o	o	o o, GO -; C5 o	^ o	00 -M ^^ o	o
^ o	_	o	o ^	-H O	■O -H	~*

00 t^

o o

ro CO —1 3

T I

o t^ 50 t:

c; GO

o o

V

V

bO

o3

o

h fc. O !»

s ^ 'E s ^ ^ 2 ^ § III a

tq O E-i cz}

°c	jS	Oi	a;
	0.
8) 3	a I a.	o3	a 41 13
0	_fl		T3 C
0		o3 <	03
03	bl)	y,
(M	y3	P	1
M	r;	T3	■1^
"0			>^.
	-a	j2
OJ	-c	c3	^
0	c3	'o	-kJ
	0	c	0	e3
	C!	00		^
0	00 CO	c3	"3
-^J	•^^		S	>
jj	^	-^-	,-*-.
			'■J
0		00		CD
><	0;
«		m	IV-	^

a bC

3 >

3 =3

r- GO

C M

CO CO c

r»i >-■ _,

.S'S

3^ .2

>. -<^ 3

_, -3 J2

^^ '^ -^

s c 0

c ' c

C t- .3

e

X.'

-C o

O j3

>^ «

• '-3 00 c. ■- o c ^

-2 >-. "^

03 03

03 gj

.22 "S

si I

>

Sex

■JO	>v
in-	0
u
t^ <	to

CO -^

32

MAURICE J. BESSMAN

TAHLE XIV

Effect of Extent of Net Synthesis on Composition OF Enzymatically Synthesized DNA"

		Increase	T	C
Expt.	Time	in DNA	incorporated	incorporated
No.	(hours)	(%)	(m/imoles)	(m/imoles)	T/C
Mycobacterium phlei DNA
Primer					0.49
Product 1		2	0.047	0.11	0.42
2	1.0	35	0.97	2.05	0.47
	2.0	86	2.13	5.03	0.43
	4.0	163	4.20	9.54	0.44
	6.0	—	10.25	9.88	1.04
	7.0	—	24.20	12.10	2.00
Calf thymus DNA
Primer					1.24
Product 3		8	0.40	0.28	1.43
4	0.5	63	3.08	2.43	1.27
	1.0	197	9.10	7.72	1.20
	3.0	534	31.60	20.60	1.53
Aerobacter aerogenes DNA
Primer					0.82
Product 5	0.5	18	0.74	0.93	0.79
	1.0	56	2.59	2.89	0.91
	3.0	344	15.00	17.75	0.84
	4.0	406	16.62	21.00	0.80

" From Lehman et al. (1958a).

thesis of the DNA could be influenced by lowering the concentration of any one of the four deoxyribonucleoside triphosphates.

2. Nearest Neighbor Anahjsis

Although it is not yet possible to determine the sequence of nucleo- tides in DNA, it has been possible to examine the frequency of nucleo- tide pairs in products prepared from several different primers. The tech- nique as developed by Josse et al. (1961) involves the following operations :

a. Synthesis of four identical products each one containing a different P^--labeled dcoxynucleotide. This is accomplished by setting up incuba- tion mixtures containing the following deoxyribonucleoside triphosphates:

1. dATP32 + dTTP + dGTP + dCTP

2. dATP -1- dTTP32 4- dGTP -|- dCTP

3. dATP + dTTP + dGTP32 + dCTP

4. dATP + dTTP + dGTP + dCTP^^

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 33

TABLE XV

Effect of Relative Substrate Concentrations ON Composition of Synthetic DNA"

Products

Substrates (relative molar concentration)

Net A + T A + G

dCTP dGTP dTTP dATP synthesis G + C C + T

1.0 1.0 1.0 1.0 IIX 1.98 1.00

1.0 1.0 0.2 1.0 6X 1.82 1.04

1.0 1.0 0.2 0.2 6X 1.82 0.97

T2 bacteriophage DNA as primer 1 . 92 0 . 98

" From Lehman et al. (1958a).

b. Degradation of each of the products in such a way that the P^^ introduced into the DNA in the form of the specific deoxyribonucleotide is now part of the particular deoxyribonucleotide to which the labeled triphosphate was attached. This is accomplished by incubating the indi- vidual products with micrococcal DNase (Cunningham et at., 1956), and splenic phosphodiesterase (Heppel et al., 1953; Hilmoe, 1960).

c. Separation of the deoxyribonucleotides and estimation of their P^- content. This is accomplished by paper electrophoresis (]\larkham and Smith, 1953).

For a detailed description of this technique, the reader is referred to the original paper of Josse et al. (1961). A brief description of a typical determination follows: Four different reaction mixtures con- taining one labeled and three unlabeled triphosphates are incubated and the reaction is terminated by the addition of carrier DNA and perchloric acid. Let us consider the product containing dAMP^-. If we examine Fig. 9 we see that dATP^- (designated as Y-nucleotide) adds to the growing deoxyribonucleotide chain by forming a 3'-5'-phosphodiester bridge with the terminal deoxynucleotide (designated X-nucleotide). This will occur at specific points in the newly synthesized DNA and these positions are presumably dictated by base-pairing with the template DNA. The product is washed free of acid-soluble unreacted deoxy- ribonucleoside triphosphates and degraded enzymatically as shown in Fig. 9. Using the same designation as for synthesis, it can be seen that since the enzymes specifically hydrolyze the phosphodiester bridges between the 5'-carbon and the phosphorus, the P^^ from the incorporated dAMP^- (Y) now ends up on the 3'-position of the neighboring deoxy- ribonucleotide (X). These 3'-deoxyribonucleotides are separated and analyzed for P^^ content and the fraction of each of the four possible dinucleotide pairs containing dAMP as the 5'-deoxyribonucleotide can

34

MAURICE J. BESSMAN

be calculated. This value, multiplied by the molar proportion of dAMP in the particular nucleic acid studied (this may be determined independ- ently or it can be derived from the nearest neighbor data) gives the fraction of the particular dinucleotide in the DNA. The jirocedure is repeated for the other three reaction mixtures and the frequencies of all possible dinucleotides may be determined.

P--P--P

SYNTHESIS

(by polymerase)

DEGRADATION

(by micrococcal DNase and splenic diestcrasci

Fig. 9. Synthesis of a P^"-Iabeled DNA chain and its subsequent enzymatic degradation to 3'-deoxyribonucleotides. The arrows indicate the linkages cleaved by micrococcal DNase and calf spleen phosphodiesterase, yielding a digest composed exclusively of 3'-deoxvribonucleotides. (From Josse et al., 1961.)

Table XVI illustrates the data obtained from a typical nearest neighbor analysis. In this experiment, Mycobacterium phlei DNA was used as primer and the actual radioactivity j)resent in each of the frac- tions is recorded. In the columns labeled "sequence" are reported the dinucleotides represented by the incorporated label. In the column.'^ labeled "fraction" are the values calculated as the fraction of the total counts in a particular reaction digest found in the particular 3'-deoxy- ribonucleotides. When these fractions are multiplied by the base-incorpo- ration factors (mole fraction of the particular base in the given DNA) the nearest neighbor frequencies recorded in Table XVH result. Earlier work showing that the base composition of the newly synthesized DNA reflected that of the primer was consistent with the scheme in Fig. 10 in which each strand of the proposed Watson-Crick double helix acts as

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

35

p::

rt

«

fa

p^

-3 « ;i .^

«2 ^' X

C^l 3: 00 -H o 00 00 :o o o " — ri r: o

o o o o o

CO o t^ o o

rt CO O 'M t^

•* -* O 00 M

o o o o

H < U O

t^ -f •* lO o

00 CO ^ --C o

^ ^ 't Ol o

o o o o —

O C 3 O O

CI o CO --C 00

'* lO 1- C'. --C

CO c^j i^ 't oc

O O O isj a a a a H < O O

t^ rt< c; o o

no 31 t^ 1- O

--^ — ' -M re o

o o o o —

^ !M M CO C

b^ b^ b^ 'c;:^

lO O 00 — o

1^ -r t^ o o

o — 1 CO 't o

O O O O -H

< < < <

Pi ^ "^ Pi

H <0 O

H < O O :^

36

MAUKICE J. BESSMAN

TABLE XVII Nearest Neighbor Frequencies ok Mycobacterium phlei DNA"'*"

Reaction No.	LalH'k'd triphosphate	Isolated 3'-deoxyriboniicleotide'"
Tp	Ap	Cp	Gp
1	dATP32	a TpA 0.012	b ApA 0.024 I	c CpA 0.063 II	d Gp.\ 0.056 III
2	dTTP"	b TpT 0.026 I	a ApT 0.031	d CpT 0.045 IV	c GpT 0.060 V
3	dGTP32	e TpG 0.063 II	f ApG 0.045 IV	g CpG 0.139	h GpG 0.090 VI
4	dCTP'2	f 0.061 III	e 0.064 V	h 0.090 VI	g 0.122
	Sums	0.162	0.164	0.337	0.337

« From Josse et al. (1961).

* Chemical analysis of the base composition of the primer DNA indicated molar proportions of thymine, adenine, cytosine, and guanine of 0.165, 0.162, 0.335, and 0.338, respectively.

'^ Identical Roman numerals designate those secjuence frequencies which should be equivalent in a Watson and Crick DN.\ model with strands of opposite polarity; identical lower case letters designate sequence frequencies which should be equivalent in a model with strands of similar polarity.

a template for the synthesis of a daughter strand with complementaiy bases. The nearest neighbor frequencies of Table XVII provide inde- pendent evidence for this scheme. ApA and TpT sequences are equal as are GpG and CpC. In examining the other sequences, it is important to realize that the predicted pairings of dinucleotides will depend on whether the two chains of the double helix run in the same or opiiosite directions, that is, have similar or opposite polarity. This difference is made clear in Fig. 11 where a short segment of the replicating chain is depicted. The expected pairing of dinucleotide-scquenccs as a conse- quence of similar and opposite polarity in the nucleic acid chains are shown, and it is obvious that decidedly different pairings would be expected depending on which of the two possibilities obtained. In Table

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

37

T A G A

::::::vQ ::::::Q

T A G A

A T C T

Q::::.]	A
S::::::	T
{B--^	C
{IJ::::::j	T

Fig. 10. Proposed scheme of replication of a Watson and Crick DNA model. Bold-lined polynucleotide chains of the two daughter molecules represent newly synthesized strands. (From Josse et al., 1961.)

T A G		A T C
A		T

OPPOSITE POUM^ITY

TpA (.012) = TpA (.012) ApG (.045) = CpT (.045) GpA (.0(55) = TpC (.061)

T		A
A		T
G		C
A		T

SIMILAR POLARITY

TpA (.012) - ApT 1.030 ApG (.0,5) - TpC 1.061) GpAi,.0o5) = CpT (.045)

Fig. 11. Contrast of a Watson and Crick DNA model with strands of opposite polarity with a model with strands of similar polarity. The predicted matching nearest neighbor sequence frequencies are different. Values in parentheses are sequence frequencies from the experiments with M. phlei DNA. The strands repre- sented here are the newly synthesized strands of Fig. 10; for ease of comparison, they are aligned as if they were complementary strands of the same double helix. (From Josse et al., 1961.)

38 MALKK'E J. LiESSMAN

XVII, the expocted pairings based on strands of opposite polarity are designated by Roman numerals whereas the i)airings based on similar polarity are designated i)y lower case letters. It is clear that the fit is in excellent agreement with the predictions based on oj)posite polarity and not with the other possibility. This represents an elegant proof of the original proposal that the DNA molecule is composed of strands of opposite polarity (Watson and Crick, 1953a).

This technique was extended to other DXA types with the following observations :

1. Nearest neighbor sequence frequencies are the same in pi'oducts primed with heated or unheated DNA.

2. The nearest neighbor sequence freciuencies of DNA from A ami A-dg (temperate bacteriophages) were very similar to those of the host, whereas the frequencies in T2, T4, and T6 (virulent bacteriophages) were markedly different from those of the host.

3. The dinucleotide frequencies of DNA from most sources were unique to the species and did not confonn to a calculated random distri- bution of the bases.

4. The dinucleotide frequencies of the DNA preparations from sev- eral tissues of the same organism showed no significant variation, nor did the DNA preparations from normal and neoplastic tissues of the same organism (Swartz et al., 1962).

5. The distribution of sequence frequencies in 16 DNA samples analyzed conform to a mechanism for DNA replication involving pairing of adenine to thymine and guanine to cytosine and are in keeping with the opposite polarity of the DNA strands.

6. The nearest neighbor sequence frequencies of a jM'oduct primed with native calf thymus DNA or primed with newly synthesized "calf thymus" DNA (containing less than 5% of the original native DNA) were the same. From this it may be inferred (though not proved) that the sequence frequencies in the product are indicative of the sequence frequencies in the native DNA.

An interesting application of the nearest neighbor analysis is the recent investigation (Swartz et nl., 1962) of whether both strands of the double-strand fonn of DNA from the small bacteriophage 0X174 (Sinsheimer, 1959) can serve as templates in DNA rejilication. When the single-stranded form of 0X174 is used to prime a reaction in wliicli (^acli strand of the original DNA replicates once, the composition of the newly formed strand should be complementaiy to the original. An ex|)erinient was devised in which DN'A synthesis was allowed to proceed until 20% of the original DN'A was replicated, thus minimizing the chance of multiple rounds of replication. The product of this reaction was analyzed

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

39

TABLE XVIII

Composition of Products after Limited and

Extensive Replication of (j>X DNA"

Composition determined by nearest neighbor analysis

20% synthesis

600% synthesis

Base

Composition Predicted determined from

by chemical chemical

analysis analysis

Observed

Predicted'' Predicted*

from from

chemical 20%

analysis synthesis Observed

A	0.246	0.328	0.310	0.287	0.276	0.271
T	0.328	0.246	0.242	0.287	0.276	0.293
G	0.242	0.185	0.202	0.214	0.224	0.213
C	0.185	0.242	0.246	0.214	0.224	0.224

" From Swartz et al. (1962).

'' Based on unlimited replication.

for base composition by the nearest neighbor technique (for details of the use of nearest neighbor analysis for determining base composition see Josse et al., 1961) and the results reported in Table XVIII show that the observed values agree very well with the predicted values. Based on this "limited" synthesis, the base composition of product formed after

TABLE XIX

Nearest Neighbor Frequencies" of </)X DNA in

Limited and in Extensive Replication*

	Limited replication	Extensive replication (600%)
	(20%)
		Pred icted from
Nearest neighbor		limited
sequence	Observed	replication"^	Observed
ApA, TpT	0.101,0.069	0.085,0.085	0.085,0.099
CpA, TpG	0.096, 0.048	0.072,0.072	0.070, 0.070
GpA, TpC	0.054,0.064	0.059,0.059	0.058, 0.065
CpT, ApC	0.052,0.069	0.061, 0.061	0.064, 0.058
GpT, ApC	0.047, 0.068	0.057, 0.057	0.053, 0.053
GpG, CpC	0.040,0.053	0.046, 0.046	0.041, 0.045
TpA	0.061	0.061	0.059
ApT	0.072	0.072	0.075
CpG	0.045	0.045	0.045
GpC	0.061	0.061	0.061

" E.xpressed as decimal proportions of 1.000.

* From Swartz et al. (1962).

" Based on unlimited replication.

40 MAURICE J. BESSMAN

extensive synthesis (6-fokl) was calculated and c()m])arc(l to the experi- mental findings. Again there is good agreement between the observed and predicted values. The calculation was based on the assumption that both strands replicate equally and this appears to be borne out by the data. The results were extended by subjecting the products of "limited" and "extensive" synthesis to nearest neighbor analysis. Based on the assumption that both strands replicate equally well, values were calcu- lated for nearest neighbor frequencies in the DNA from "extensive" synthesis and the data in Table XIX support this contention very well. These elegant experiments further support the proposed scheme for replication (Fig. 10) in which both strands of the primer DNA act as templates through which specific base-pairing is accomplished.

A few of the applications of nearest neighbor analysis have been mentioned here. It is clear that until a new technique is devised for determining longer sequences in the DNA molecule, the nearest neighbor approach will provide the most revealing information about the fine structure of DNA.

G. SYNTHESIS OF DEOXYADENYLATE-DEOXYTHYMIDYLATE AND DEOXYGUANYLATE-DEOXYCYTIDYLATE POLYMERS

(Schachman et al., 1960)

We have already seen an exception to our generalization that all four deoxyribonucleoside triphosphates are required for DNA synthesis in that single deoxyribonucleotides can add to the ends of DNA chains. In this section, two reactions catalyzed by E. coli polymerase, w^hich not only do not require all four triphosphates, but in addition may proceed in the absence of primer, will be discussed.

1. The d-AT Polymer

When E. coli polymerase is incubated with all four triphosphates in the absence of a primer there is an extensive lag period, 3 to 4 hours, in which no detectable reaction occurs. This period of inactivity is followed by a rapid synthesis of a high molecular weight polymer of deoxyadenylate and deoxythymidylate. The kinetics of the reaction are shown in Fig. 12, where three independent techniques were used to follow the reaction. In Expt. 1, plots of the rate of incorporation of P^- from dATP''- and the rate of increase in viscosity are superimposable. In Expt. 2, the hypochromicity curve is the mirror image of the P^- in- corporation. In Fig. 13, the kinetics of DNA synthesis in the complete system is compared to the synthesis of the d-AT polymer. The complete system contained, beside polymerase and calf thymus DNA, all 4 tri- phosj)hates. Polymer formation as measured bv viscositv increase starts

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

41

•»-• tAu o<?3 ;d X+isuap iD-Dudo

			y	^
crimen		>^	/	N	\.
a.	/	/^				'N	>-
Jt	p-*-*	...•——	•*-—••		••••-••'	^•— ••
{\							N
						— — <

- s_ o

n4- c-i o BO o ^

o— o X4IS03SIA oiipods

So 3 e

8 a

22 «

-a a X

3 O kO

t- a)

rt q; o3

-i a

O -rr tc

.2 -s •- ^

73 03 CO t^

a ^

2 o

O Ml

2 & a ^

fe

0.6-

Isp

No triphosphates

dATP

TTP

dCTP+dGTP

150 250 350

Time in minutes

Fig. 13. Synthesis of DNA and deox^-adenylate-thymidylate copolymer, with thymus DNA as primer. (From Lehman, 1959.)

+ 0.1 0.0

-0.1 0.2

^0.3

o

CM 0.4

d

<J0.5

0.6 0.7 0.8

20 ^0 60 60 100 120 140 160 Time in minutes

Fig. 14. Comparison of primed versus unprimed synthesis of d-AT copolymer. (From Schachman et al., 1960.)

42

r		S>€k&A-
		'^^	\
\				>
4	°nmec
i				L/npn'j	-ned
I
1	J	(
\	J					\	V
					\.

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 43

with no appreciable lag and continues linearly. When dGTP or both dGTP and dCTP are omitted from the incubation, there is a 4-hour lag followed by a period of rapid synthesis. If no triphosphates are added to the incubation mixture, or just those indicated in the last curve in the figure, no synthesis occurs. The lag period can be reduced somewhat by raising the enzyme concentration; but it can be eliminated completely by priming the reaction with synthetic d-AT polymer (Fig. 14). Only d-AT polymer itself is an effective primer in the reaction, and other DNA preparations have no effect in reducing the lag period.

When the "limited" reaction (Section IV,C) is investigated, using d-AT as the primer, small but significant quantities of deoxyadenylate or deoxythymidylate are added to the ends of the polymer. It is of interest that neither dCIVIF nor dGMP can add to the ends of the d-AT polymer, suggesting that even in the limited reaction specific base-pairing is required. This requirement for specific base-pairing favors the "repair" hypothesis suggested as one of the possible explanations of the "limited" reaction (Section IV,C).

Analysis of the d-AT polymer shows it to be a rodlike organized structure with a molecular weight ranging between 10'^' and 10". It is composed exclusively of equal proportions of deoxyadenylate and deoxy- thymidylate residues, and an interesting feature is that these residues

TABLE XX Hydrolysis of d-AT Copolymer to S'-Deoxyribonucleotides"'^

	Substrates for preparation of d-AT copolymer		Products of hydrolysis
	3 '-Deoxyadenylate 3 '-Deoxythymidylate
Expt. No.	cpm	cpm in polymers (%) cpm	cpm in pohmers (%)
1 2	dAPPP and dTP^^pp dAP'^PP and dTPPP	•»670 46	100 23 <0.5 8510	<0.5 100

° From Schachman et al. (1960).

''d-AT copolymer was hydrolyzed to 3'-deoxyribonucleotides by micrococcal DNase and splenic phosphodiesterase until more than 96% of the P'- was released in the form of deoxj'ribonucleoside monophosphates.

are arranged in a strictly alternating sequence. This is demonstrated in Table XX where the results of a nearest neighbor analysis show that each deoxythymidylate residue is linked by a phosphodiester bridge to a deoxyadenylate residue and vice versa. This regularly alternating sequence has been used to explain the obser^'ation that the d-AT co-

44

MAURICE J. BESSMAN

polymer does not show the hysteresis typical of DNA preparations after heating and cooling (Fig. 15). It has been suggested that the regularly alternating sequence of deoxyadenylate and deoxythymidylate in each strand increases the opportunity for correct alignment of the bases in opposite strands during the cooling period (Marmur and Hoty, 1959; Schachman et al., 1960).

1.4

o

>

1.2

O

c oi.l

'■^

a o

1.0

o Orig ® Secc	inol heotin	?	®i) f
)nd heating		«.
			j
OmmJdm^^^P^	1 ®]

40

50 60 70

Temperature in ^C.

80

Fig. 15. Melting of d-AT copolj-mer. Plotted on the ordinate is the relative absorbancy at 260 m/i at t° compared to the value at 25°C. The abscissa gives the temperature. The first heating cycle is indicated by open circles, and the solution after coohng in the stoppered Beckman cell was heated a second time, giving the data indicated. (From Schachman et al., 1960.)

Attempts to elucidate the mechanism of synthesis have so far met with little success. For example, in the unprimed synthesis of d-AT polymer, at any stage during the synthesis only two types of molecules are observed. One type has a molecular weight in the order of 10' and is obviously made up of the original triphosphate molecules, and the other has a molecular weight in the order of 10'^ and must represent completed molecules. No intermediates have been found between these two classes of molecules, as indicated in Table XXI where it can be seen that the sedimentation coefficients and reduced viscosities are constant throughout the early and later phases of the synthesis.

Recently, Sueoka (1961) reported the presence of a minor com- ponent in the DNA of crab testes {Cancer borealis) which had an

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 45

unusually low density suggesting a very low content of cytosine and guanine. A sample of this material was used to prime a reaction using E. coli polymerase and the product was analyzed for nearest neighbor sequence frequencies (Swartz et al., 1962). Over 95% of the product had a sequence made up of alternating deoxyadenylate and deoxythymidylatc

TABLE XXI Relationship between Extent of Reaction and Size of Product"''

				Properties of product
	Extent of reacti	ion
				Sedimentation	Reduced
Expt.	Viscosity	Hypochromicity	coefficient	viscosity
No.	(%)		(%)	(S)	(gm/100 ml)-i
1	8		5	29	26
	37		46	28	26
	100		100	27	24
2	36		28	19	23
	80		91	19	25
	100		100	20	26

" From Schachman el al. (1960).

^ Reaction mixtures, 6 ml, were incubated directh' in viscometers; at selected times after the viscosity of the solution had been measured, aliquots were removed, and the enzymatic action terminated by the addition of NaCl-citrate solution to give a concen- tration of 0.2 M NaCl-0.1 M sodium citrate. Absorbancy measurements of these samples as compared to the original gave a measure of the extent of the reaction (hypochromic- ity). The samples were then dialyzed, and the reduced viscosity, Vsp/C, in (gm/100 ml)~' and sedimentation coefficient were measured in 0.02 Af sodium citrate.

residues. This naturally occurring d-AT polymer differs from the unprimed d-AT polymer produced by E. coli polymerase in that it does have a small but significant content of deoxyguanylate and deoxycy tidy late residues accounting for less than 5% of the total deoxyribonucleotides.

2. The d-GC Polymer

Much less infonnation is available about an interesting polymer composed exclusively of deoxycytidylate and deoxyguanylate residues. This material is similar to the d-AT copolymer in that it is composed of only two deoxyribonucleotides and is formed after a considerable lag period in an unprimed reaction catalyzed by purified E. coli polymerase. However, this product is distinctly different in one important respect. Whereas d-AT is a copolymer of alternating residues of deoxyadenylate and deoxythymidylatc, the d-GC polymer is made up of separate strands containing either deoxyguanylate or deoxycytidylate. This was shown

46 MAURICE J. BESSMAN

clearly by .Io.xsjc et al. (1961) by iicaix'st iici<i;lib()r analysis. Within the liinits of the niothods, (k'uxycytitlylatc was only found next to deoxy- oytidylatc. No jiliosphodicster linkages were evident between deoxy- ^nanylate and deoxycytidylate. Besides this dif'fei'ence, d-dC j)olyniei' is distinct fi'oni d-AT in having uneciuai i)roj)ortions of deoxycytidylate and deoxyjiuanylate. Usually, the molar ratio of cKlMP to dCMP is greatei' than I.

H. SUBSTITUTION OF ANALOGS

An analog is any purine oi" pyriniidine base not normally found in the DNA of a particulai- oi-ganism. P"or example, although 5-inethyl- cytosine is a normal constituents of wheat germ DNA, it would be con- sidered an analog if incorjjorated into E. coli DNA, because it is not normally detected there. Several analogs have been synthesized as the deoxyribonucleoside triphosphates and tested for incorporation into DNA (Bessman et al, 1958b). When primed with E. coli DNA, the following analogs support DNA synthesis: uracil, hyjioxanthine, 5- bromouracil, 5-bromocytosine, and 5-methylcytosine. Xanthine was not incorporated. Recent experiments from Romberg's laboratory (personal comnnmication) have shown that 5-fluorocytosine and 5-fluorouracil are incorporated into DNA whereas A^-methyl 5-fluorocytosine is not. 5- Hydroxymethylcytosine, the pyrimidine base found exclusively in the T-even bacteriophages (Wyatt and Cohen, 1953), is also incorporated into DNA by E. coli polymerase primed with DNA prepared from several different sources including calf thymus, E. coli and phage ^X174 (Korn- berg et al, 1959).

According to the restrictions imposed by the base-pairing rules of the Watson-Crick model (1953b), these base analogs should be incorpo- rated into specific positions in the DNA molecule. A diagram of these relationships is shown in Fig. 16. The specificity in replacing a given base by an analog was tested by determining which deoxyribonucleoside triphosphate could be omitted from a reaction mixture in the presence of a given analog and still support DNA synthesis. For example, in order to find out which substrate could be replaced by deoxyuridine triphos- phate, four reactions were set up, each lacking one of the normal triphosphates. DNA synthesis proceeded only in the reaction mixture lacking dTTP, suggesting that dUTP was replacing the dTTP require- ment. Table XXII contains the results of several such experiments. It can be seen that in each case the given analog substitutes for the purine or pyrimidine base predicted by the Watson-Crick model. Table XXIII shows direct evidence for the incorporation of deoxyuridylate into DNA. The results are consistent with a specific substitution of deoxyuridylate

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

47

for deoxythymidylate since the data indicate that these two nucleotides compete with each other during tlie synthesis. Results of a nearest neighbor analysis on a product synthesized with dUTP*- showed that deoxyuridylate was linked to each of the other deoxyribonucleotides.

Hydrogen Bonding of Adenine to Thymine

Vi-H--

._..0

---H

N = C,

,N

/

W

OSUGAR

R
-CH3	Thymine
- H	Urocii
-Br	Bromuraci!

Hydrogen Bonding of Guanine to Cytosine

H

N

,H--

.-N

\ C — N

C-H /

^N^C.

R'
-H	Cytosine
-CH3	Methylcytosine
-Br	Bromcytosine

oSUGAR

r"
- NHg	Guanine
-H	Hypoxanthine

Fig. 16. Base-pairing relationships of normal constituents and analogs as pre- dicted from the Watson and Crick model for DNA. (From Bessman el al, 1958b.)

An interesting study of the effect of 5-bromouracil substitution on replication of DNA has recently been reported (Trautner et al., 1962). The d-AT polymer was used to test the frequency with which mistakes in base-pairing occur during DNA synthesis by noting the frequency with which dGMP was incorporated into the d-AT product. Highly radioactive dGTP^- (specific activity =1 X 10^ cpm/^mole) provided the basis for an assay that was capable of detecting the incorporation of

48

MAURICE J. BESSMAN

TABLE XXII Replacement of Natural Bases by Analogs in Enzymatic Synthesis of DNA"

				Natural base omitted
	Control value''
Expt.		Thjinine	Adenine	Guanine	Cytosine
No.	(m/iiraoles)	Base analog used		(% of control) <^
1	0.50	Uracil	H	4	6
la	0.88	Uracil				3
2	0.43	5-Bromouracil	97	2	4
2a	0.42	5-Bromouracil				4
3	0.51	5-Bromocytosine		4	4	118
3a	0.40	5-Bromocvtosine	4
4	0.58	5-IMethylcytosine		2	3	185
4a	0.52	5-Met.hylc3tosine	2
o	0.37	Hypoxanthine		3	25	5
oa	0.27	Hypo xanthine	4

" From Bessman et al. (1958b).

'' Control values are m/xmoles of radioactive deoxynucleotide incorporated into DNA in the absence of analog. Incubation mixtures contained in 0.3 ml, 5 m^imoles each of TTP, dATP, dCTP, and dGTP; 2 Mmoles of MgClj; 20 Mmoles of potassium phosphate (pH 7.4); 10 Mg of calf-thymus DNA; and 1 /ug of enzyme fraction VII-R. Experiments were performed at 37°C for 30 minutes. Labeled substrates were: dCP'^'PP in E.xpts. 1, 2, and 5a; TP^^PP in Expts. la, 3, 4, and 5; and dOP^^PP in E.xpts. 2a, 3a, and 4a.

" The percentage value represents the fraction of the labeled substrate incorporated when the analog (5 mMmoles) was used instead of a natural base. All bases, natural or analog, were supplied as the deoxjnucleoside triphosphates. Values of 5% or below are near the limit of detect ability and are of questionable significance.

TABLE XXIII

I.\CURPUR.\TION OF P'^-DeOXYURIDYLATE INTO DNA"

Substrates'*

P^^-deo.xynucleotide

incorporated

(m^moles)

U + T + C + A+Ci Omit T Omit C Omit A OmitG

U+T+C+A+G Omit U

0 14 0 . 38 0.01 0.01 0.01

0.37 0.77

« From Bessman et al. (1958b).

* U = dUTP; U = dUP^^pp. '1^ = '["pi>. ^l = TP32PP; C = dCTP; A = dATP; G = dGTP.

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 49

1 dGMP residue per 10^ nucleotide units. Within the limits of the assay, no significant incorporation of dGMP occurred in a reaction mixture made up of d-AT primer, dOTP^^-, dATP, dTTP, Mg^ and E. coli polymerase. However, when a polymer composed of dAMP and 5- bromodeoxyuridylate (d-ABU) was used in place of d-AT as the primer, there was a significant incorporation of dGMP. For example, in one experiment less than 0.10 /xfimoles of dGMP^- was incorporated using d-AT as a primer, whereas 2.82 ^^moles of dGMP''- was incorporated using d-ABU as primer.

Mistakes in base-pairing could result from tautomeric shifts in the bases during replication (Watson and Crick, 1953b). The specific muta- genic effect of 5-bromouracil which is incorporated into DNx\ in place of thymine (Dunn and Smith, 1954) has been attributed to its facilitated tautomerization to a fonii w^hich could pair with guanine instead of adenine (Watson and Crick 1953b; Freese, 1959). If guanine only substi- tuted for adenine in the d-ABU polymer described above, a nearest neighbor analysis of a product prepared dGMP^- as the labeled deoxy- ribonucleotide should reveal only one labeled 3'-deoxyribonucleotide, that is, 5-bromodeoxyuridylate. When Trautner et al. (1962) analyzed the product they found that all three possible 3'-deoxyribonucleotides were labeled (Table XXIV). These data show there are just as many

TABLE XXIV

Nearest Neighbors of the Guanixe Residues Incorporated

INTO THE d-ABU Polymer"

		Radioactivity	in the isolated 3'-deoxynucleotides
Expt. - No.		BU		A		G
cpm	% of total	cpm	% of total	cpm	% of total
1 2	122 227	36.3 45.3	56 83	16.7 16.7	158 190	47.0 38.0

" From Trautner et al. (1962).

GG pairs as GBU pairs in the product and about one-half as many GA pairs. This would suggest that the mutagenic effect of 5-bromouracil is not entirely explained on the basis of a mistake in pairing with adenylate instead of guanylate, but that the analog exerts a more complicated effect. The authors are careful in pointing out that these results in vitro with the d-ABU polymer do not necessarily represent the in vivo effect of 5-bromouracil on whole DNA. Also, the argument outlined above is predicted on the basis that the d-AT (and hence d-ABU) polymer

50 MAURICE J. BESSMAN

is composctl of 100% alternating deoxyadenylate-deoxythymidylate residues.

Another interesting experiment of Trautner et al. (19()2) was the preparation of a product from a reaction mixture primed with K. coll DNA in which 100% of the dTMP residues were rcphiced by 5-bronio- deoxyuridyhite (Baldwin et al., 1961). A nearest neighbor analysis of this DNA revealed that its dinucleotide sequences were indistinguishable (±2%) from those in normal E. coli DNA. These results are not in keeping with a report of Shapiro and Chargaff (1960), based on the analysis of an acid-hydrolyzed product that marked changes in sequences in E. coli DNA result from substitution of 5-bromouracil.

The general conclusion that may be drawn from the studies on the incorporation of base analogs into DNA is that, in vitro, purine and pyrimidine bases supplied as deoxyribonuclcoside triphosphates usually are incorporated into DNA at positions predicted by the base-pairing relationships of the Watson-Crick model. This does not mean that these same analogs supplied as free bases, deoxyribonucleosides, or deoxyribo- nucleotides will be incorporated into DNA in vivo since these compounds may be excluded by being blocked at any one of the several steps leading to the deoxyribonuclcoside triphosphates (Kalle and Gots, 1961 ; Fried- kin and Kornl)crg, 1957; Bessman et ol., 1958b). Another interesting method of exclusion is based on the specific hydrolysis of the deoxynu- cleoside triphosphates by enzymes such as dCTPase (Zimmerman and Kornbcrg, 1961; Koerner et al., 1960) and dUTPase (Bertani et al., 1961; Greenberg and Somerville, 1962a). These two enzymes are pre- sumably involved in keeping cytosine out of the DNA of T-evcn bacteriophages and uracil out of E. coli DNA, respectively.

V. Synthesis of DNA by Calf Thymus Polymerase

There have been numerous reports of the synthesis of DNA in vitro catalyzed by extracts prepared from higher organisms. With the excep- tion of veiy few, these have been done with imfractionated preparations which contain several activities related and unrelated to the direct reaction. These investigations, although important in establishing an over-all picture of the fate of labeled nucleic acid precursors in cell homogenates cannot provide the detailed information necessary to define the pathways and mechanisms of DNA synthesis. Bollum (1959a,b, 1960a,b) has investigated the synthesis of DNA using a partially puri- fied enzyme from calf thymus glands. His experiments will be discussed as representative of the properties of DNA synthesis in animals although the generalization may not be applicable to all systems. No obvious

I. REPLICATION OF DXA IN CELL- FREE SYSTEMS 51

differences between the synthesis of DNA in the calf thymus system and the less purified regenerating rat liver system described by Bollum and Potter (1958) and Bollum (1958) are evident and therefore remarks will be confined to the former.

At the outset it may be stated that the reaction catalyzed by calf thymus polymerase (purified about 50-fold) is very similar to the one from E. coll. The requirement for all four deoxyribonucleoside triphos- phates, jNIg ions, and DNA is evident in Table XXV. Omission of any

TABLE XXV

Requirements for Thymidine Triphosphate Incorporation

WITH Calf Thymus Polymerase"

HMTMP incorporated into DNA Reaction mi.xture (MMmoles)

Complete* 100

Minus dATP 9

Minus dGTP 13

Minus dCTP 28

Minus dATP, dGTP, dCTP 2

Minus Mg++ 0

Minus DNA 0

" From Bollum (1960a).

''The complete reaction mixture contained the following concentrations: deox>Tibo- nucleoside triphosphates, 25 MHioles in each (H^-TTP containing 6700 cpm per m/xmole) ; MgCl2, 10 mmoles; glycine buffer, pH 9, 25 mmoles; 25 jug DNA (salmon sperm); and 195 Mg of fraction C protein. Final volume, 0.20 ml. Incubated 20 minutes at 37°C.

of these constituents reduces the rate of incorporation. The reaction is inhibited by high concentrations of pyrophosphate and it can be shown that DNA synthesis is reversible in this system by measuring the forma- tion of deoxyribonucleoside triphosphates in the presence of polymerase, DNA, Mg ions, and pyrophosphate. With the partially purified enzyme, net synthesis of DNA can be demonstrated if heated DNA is used as the primer. The products formed resemble the primer with respect to adenine-cytosine or thymine-cytosine content as demonstrated in Table XXVI. Again, the failure of native DNA to prime the reaction is demon- strated in this table.

This reduced priming capacity of native DNA was investigated by Bollum (1959a,b) who showed that even at early stages in the purifica- tion the calf thymus polymerase uses denatured DNA much more efficiently than native DNA as primer. This is seen clearly in Fig. 17 where several heated and unheated DNA preparations are compared. Except in the case of salmon sperm DNA, the unheated or native DNA preparations are strikingly deficient as primers as compared to the heated

52

MAURICE J. BESSMAN

preparations. As mentioned previously, Kornberg (1960) has pointed out that some preparations of E. coli polymerase prefer denatured DNA preparations as primers. Bollum has suggested that the reason a dena- tured ])rinu'r re(iuirement is more easily demonstrated with calf thymus polymerase than witii the corresponding enzyme from E. coli is that thymus glands contain much less nuclease than E. coli and thus poly- merase preparations prepared from this tissue have relatively less nuclease activities earlier in the purification procedure. As mentioned

TABLE XXVI Comparison of Puimer-Product Base Ratios with Calf Thymus Polymerase"

		Treatment
						Primer	Product
	Primer DNA	Time (miii)	(T	emp	°C)	ratio	ratio
		Experiment	1
A/C	Tl	2		99		—	1.3
	T2	2		99		1.9	1.8
	T4	2		99		1.9	1.9
	T6	2		99		1.9	1.8
	Calf thymus	2		99		1.3	1.8
	Diplococcus pneumoniae	10		99		1.6	1.6
		Experiment	2
T/C	Calf thymus	2 15 13 15		99 68 78 98		1.37	1.38 Non-priming 1.30 1.37
		30		0,	OH-		1.39
		30		0,	H+		1.36

" From Bollum (1959b).

before (Section IV,E) the presence of nuclease in the enzyme prepara- tions could convert non-priming (native) DNA to priming (denatured) DNA thus obscuring a requirement for denatured DNA preparations. The implication in all these studies is that the natural primer for DNA synthesis is single-stranded DNA. As in the E. coli system, the single- stranded DNA from bacteriophage ^X174 is an excellent primer for calf thymus polymerase and requires no prior heating to manifest its activity. Other means of denaturing DNA such as acid or alkali treat- ment are also effective in converting non-priming to priming DNA (Table XXVI).

Calf thymus and E. coli polymerase differ in that no "unprimed" reactions have yet been demonstrated with the mammalian enzyme. Calf thymus polymerase does not catalyze the unprimed synthesis of the

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS

53

d-AT or d-GC polymers. These reactions were difficult to discern at first, even with the highly purified E. coli enzyme, until the optimal condi- tions were worked out. It is possible that further attempts to synthesize these products with more purified enzyme preparations from calf thymus under suitable conditions will resolve this diff"erence.

in

^ 12 5 o

E

3.

E

Salmon sperm DNA (CC/BR)

Calf thymus DNA . (Worthington)

/

/

Heated

75

50

25

Calf thymus DNA (Volkin via KSD)

T2 DNA (Volkin via KSD)

180 2^0 0 Time at 35°C

Fig. 17. Comparison of heated and unheated DNA preparations with respect to priming capacity. The question mark indicates the uncertainty concerning the manufacturer's statement that "certain less gentle modifications" are used in the preparative procedure. (From BoUum, 1959a.)

A second difference between the two enzymes, more difficult to resolve, also involves primer. So far, no chemically synthesized polymer has served as primer for E. coli polymerase, and DNase digests of primer DNA are totally inactive in this capacity. Calf thymus polymerase, on the other hand, can use DNase digests as well as synthetic polynucleo-

54 MAURICE J. BESSMAN

tides as primers for (looxyiihoimclcotidc incorporation (Bolluin, 1960b, c). PolyriboiUK'k'otidcs cannot serve as i)riniers, nor ean sonicated DNA unless it is heated after sonication to melt out interchain hydrogen bonds.

Preliminaiy studies with a series of deoxythymidylate homopolymers of general structure (pT)„ wliere n^S showed that single deoxyribo- nucleotides (incubated as triphosphates) added to the ends of these synthetic oligonucleotides. All of llic four triphosphates were effective but dGMP and dTAIP tended to form multiple products. Since the incubation mixtures contained relatively large excesses of primer to deoxyribonucleoside triphosphate, the most likely event was the addition of single deoxynucleotides to the ends of the chains. The added deoxy- ribonucleotides were removed from the polymer by snake venom phosphodiesterase suggesting that authentic phosphodiester linkages were established between the oligonucleotide and the incorporated deoxyribo- nucleotide. Further investigations using chemically synthesized poly- thymidylates as primer (Bollum, 1962) have extended this earlier work and have shown that oligonucleotides with greater than two nucleotide residues are required. The 3'-hydroxyl group of the terminal dcoxynu- cleoside residue must be free for the condensation since polymers con- taining an acetylated 3'-hydroxyl group are inactive as primers. Since single deoxyribonucleotides add to the ends of chains in this reaction, the mechanism seems similar to the "limited" reaction catalyzed by E. coli polymerases (Section IV,C). The difference is that the latter enzyme requires polymerized DNA and the deoxynucleotides added seem to conform to the base-pairing relationships of the primer.

These observations are part of the general investigation undertaken by Bollum into what is the minimum length of jjrimer required for replicative or hydrogen bond directed synthesis where base-pairing relationships are obligatory. Preliminary observations by Furlong and Bollum (cited in Bollum, 1962) suggest that replicative synthesis begins when chain length exceeds 20 monomer units.

Bollum has suggested (personal communication) that the addition of specific deoxyribonucleotides to the ends of chemically prepared oligo- nucleotides might provifle a useful means of synthesizing specific sequences for testing in other biological systems.

VI. Synthesis of DNA in Bacteriophage-Infected Escherichia coli

A. BACTERIOPH.\r,E-SPECIFTC POLYMERASE

The dramatic appearance of several new enzymatic activities soon after infection of E. coli with the T-even bacteriophages has recently

I. KEPLICATION OF DNA IN CELL-FREE SYSTEMS 55

been reviewed by Cohen (1961). Many of the:<e new activitie.s exj)lain several observations concerning the DNA metabolism of the infected cell. For example, the induction of a specific enzyme which catalyzes the hydroxymethylation of deoxycytidylate (Flaks and Cohen, 1957, 1959a) accounts for the presence of this unique pyrimidine base in bacteriophage DNA. Also, the marked increase in activity of a dCTPase (Zinnnerman and Romberg, 1961; Koerner et al., 1960) may account for the apparent decrease in dCMP kinase activity in extracts of bacteriophage-infected cells. Romberg et al. (1959), in a study of the effect of bacteriophage infection on DNA synthesis, observed a 10-fold stimulation of poly- merase activity in extracts of T2-infected cells but noticed that the observed increase was dependent on the type of primer used for the assay. That is, the increased activity after infection could only be dem- onstrated if heated DNA was used as primer. Recently, Aposhian and Romberg (1961, 1962) have conducted a careful study of the polymerase in extracts of infected cells and provide conclusive evidence that the increased activity after infection is due to the formation of a new polymerase.

Polymerase from T2-infected cells was purified 600-fold and was distinguishable from E. coli polymerase in the following respects:

1. Levels of antiserum against E. coli polymerase causing 100% inhibition have no effect on T2 polymerase and vice versa.

2. 7;-Chloromercuribenzoate (1.7X10"^^^^) inhibits T2 polymerase completely whereas E. coli polymerase is inhibited 27%.

3. T2 polymerase and E. coli polymerase are easily separable by chromatography on a column of phosphocellulose.

4. The two polymerases have distinctly different primer requirements. Except for this last property the differences between the two enzymes

should have no consequences for the in vitro synthesis of DNA. Thus, the T2 enzyme is similar to the E. coli enzyme in that it requires all four deoxyribonucleoside triphosphates, Mg ions, and DNA. It can also incorporate base analogs into DNA in the same manner as the E. coli enzyme provided the analogs are supplied as the deoxyribonucleoside triphosphates.

The most striking property of the T2 enzyme is its characteristic dependence on heated DNA first observed in crude extracts of infected cells. The E. coli enzyme and T2 polymerase are compared in Table XXVII; it may be seen that heated DNA is a better primer for the T2 polymerase-catalyzed reaction but is actually somewhat worse for the E. coli polymerase reaction. That this difference in behavior is not due to some unknown factor in either the purified T2 or E. coli polymer- ases is shown by the mixture of the two preparations. In this case the values equal the sums of the individual incubations.

56 MAURICE J. be:ssman

TABLE XXVII Primer Differences of T2 and Escherichia coli Polymerases"

		Po	Ivmerase
	(C'H.	deo.xj'niK'leotide incorporated)
	T2		E. coli T2 + E. coli
Primer'"	(m/imoles)		(m/xmoles) (m/imoles)
Expt. 1
Unheated salmon sperm DNA	0.175		0.995 1.30
Heated salmon sperm DNA	0.635		0.492 1.04
Expt. 2
Unheated salmon sperm DNA	0.249		2.00 2.11
Heated salmon sperm DNA	1.30		1.20 2.84

" From Aposhian and Kornberg (1962).

* In Experiment 1 , 0.09 ng of T2 jmI ymerase or 0. 10 mS of E. coli polymerase was used. In Experiment 2, twice the amount of enzj'me was used. The labeled substrate was C'-dGTP.

How these properties of the induced polymerase are significant in the infectious process is not known, but it would not be surprising to learn that these differences are essential in allowing the phage to synthesize DNA in an alien environment.

B. GLUCOSYLATION OF^ 5-HVDROXVMETHYLCYTOSINE (HMC)

The composition of the DNA in T-even bacteriophages i.>^ unicjue in that it contains 5-hydroxymethylcytosine instead of cytosine (Wyatt and Cohen, 1953), and it also contains glucose (Jcsaitis, 1954; Volkin, 1954; Sinsheimer, 1954). Flaks and Cohen demonstrated that a new enzyme, deoxycytidylate hydroxymethylase (which accounted^ for the synthesis of this viral pyrimidine; see Section II,C) was induced in the bacteriophage-infccted system. Likewise, Kornberg et al. (1959, 1961; Zimmerman ef a/.. 1962) have indicated that new enzymes capable of glucosylating HMC-containing DNA were formed soon after infectio^i The elegant experiments of Lehman and Pratt (1960) using a highly purified phosphodiesterase wliicli could degrade the normally DNase- resistant bacteriophage DNA to mononucleotides (Lehman, 1960) de- lineated the .'Structures and distribution of the glucosylated HMC- nucleotides isolated from T2, T4, and T6 DNA preparations. A compila- tion of their data is shown in Table XXVIIL If we assume that a specific enzyme is required for each type of glucosylated nucleotide, six enzymes (two for each phage) should be demonstrable. So far, five different enzymes have been purified and characterized (Zimmemian ef al., 1962;

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 57

Josse and Kornberg, 1962). The only activity not yet found is the enzyme which forms diglycosyl-HAIC units in T2 DNA.

UDP-glucose (Caputto et al., 1950) has been shown to be the coenzyme reciuii'ed for tlie transfer of glucose to the HAIC residues in each of the reactions tested (Kornberg et al., 1959). The enzymes will not glucosylate the free deoxyribonucleotide or deoxyribonucleositle

TABLE XXVIII Analysis of HMC"-Residues in Bacteriophage DNA*

	T2	T4	T6
Glucose linkages	(^f)	(%)	(%)
Non-glucosylated	25	0	25
a-Glucosvl	70	70	3
/3-Glucosyl	0	30	0
Diglucosyl	5	0	72

" HMC = h3'drox3'meth3'lc3'tosine. * Data from Lehman and Pratt (1960).

triphosphate, but require HMC residues in intact DNA. Conversely, synthetic glucosyl hydroxymethyldeoxycytidine triphosphate in the pres- ence of the other three triphosphates is much less effective than the non-glucosylated triphosphate in supporting DNA synthesis in extracts of infected cells (Koerner et al., 1960). The generalized reaction for the a-glucosyl and /?-monoglycosyl transferase may be formulated:

UDP-glucose + HMC-DNA -^ UDP + glucosyl-HMC-DNA

The HMC-DNA used as acceptor for the glucosyl units is prepared by incubating E. coli polymerase with 5-hydroxymethyldeoxycytidine tri- phosphate, the other three deoxyribonucleoside triphosphates, Mg ions, and primer DNA. Several different types of primer, including DNA from E. coli, bacteriophage, and calf thymus have been used with equal success. Alternatively, naturally occurring HMC DNA may be used as acceptor provided the heterologous transferase is being studied. For ex- ample, T2 DNA (containing 25% unglucosylated HMC residues) is a good substrate for T4 or T6 a-transferase but will not accept glucose using the purified T2 transferase. As a general rule, no transferase will catalyze the further glucosylation of its homologous DNA.

The T2, T4, and T6 a-glucosyl transferases are three very closely related proteins which have many properties in common. They are differ- ent in certain respects as is evident from an inspection of Table XXIX. The T4 /^-transferase is readily distinguishable from any of the a-transferases.

58

MAURICE J. BESSMAN

TABLE XXIX Comparison of the Mo\o(;hcosyl Traxsferase"

					Transferase
Propert\'		T2a			TAa			TGa			T4/3
Retiuiremciit for
sulfhydrvl		+			+			+			0
Iiihil)it.i()ii l)y Mg"*""*"		+			+			+			0
Inhibition 1)\- tris		0			0			0			+
Km-. DNA	3	X 10-5	M	2	5 X 10-5	M	3	X 10-5	M	3	3 X 10-5 M
K„: UDP-ghicose	3.0	X 10-5	M	0	7 X 10-5	M	1.5	X 10-5	M	2 0	1 X 10-5 J/ 5 () X 10-5 M'
Behavior in pliosjjhate
buffer:
Activity at pH (5 . 5 Activity at pH 7.0		4.0			4.0			1.0			0.8
Ghicosyhition of T2
DNA		0			+			+			+
Glucosylalion of 'r4
DNA		0			0			0			0
Glucosylation of T()
DNA		0			0			0			0

" From Josse and Kornberg (1962).

^ In phosphate buffer.

■^ In Ammediol l)uffer (2-amino, 2-methyl, 1,3-propanedioI).

A study of the reversibility of the reaction with the T2 a-transferase (Zimmerman et al., 1962) indicates that glucosyl residue.^ may be removed from glucosylated Hj\IC DNA in the presence of UDP but the equilibrium strongly favors the forward reaction:

A' =

(glucosyl-HMC)(UDP) (HMC)(rDP-glucose)

200 2000

Thus, the extent of glucosylation is probably not limited by equilib- rium considerations since 25% of the available sites in T2 DNA are unglucosylated, and incubation in the presence of T2 a-transferase and UDP-glucose does not further glucosylate it.

What then does control the number and positioning of the glucosyl residues in the different DNA types? The simplest hypothesis is that specific base sequences determine the addition of glucosyl units. Recent experiments by Josse and Kornberg (1962) do not tend to support this notion although they do not exclude it. First of all, as had already been observed by Kornberg et al. (1961), T4 ^-glucosyl transferase can glucosylate 100% of the available HMC residues in DNA. We know

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 59

that T4 DNA contains 707o o- and 307^ /?-gliico.syl linkages. Second, we would expect that natural T4 DNA which has had some a-glucosyl residues removed with the a-transferase and UDP (running the reaction in reverse) should accept only a-glucosyl units, and similarly /3-degluco- sylated DNA should accept only |8-glucosyl units. The data in Table XXX, however, show that no matter which of the enzymes is used to

TABLE XXX

Addition of Glucose to "De-Glucosylated" T4 DNA"

Transferase used to	Transferase used to	Extent of
remove glucose	restore glucose	addition
None	T4a	<0.001
	T4^	<0.001
T4a	T4a	0.024
	T4/3	0.024
T4^	T4a	0.021
	T4^	0 . 040

" From Josse and Kornberg (1962).

remove the glucose units, both enzymes can put the glucose back in. This would indicate that the sequence around the glucosyl units does not direct the synthesis. Third, a polymer containing only dGMP and H]\IC nucleotides was synthesized using d-GC polymer as primer. This new polymer presumably had the structure d-GH]\lC with all the guanine residues in one strand and all the H]\IC residues in the other. When this polymer was used as an acceptor for glucosylation, all four of the specific monoglucosyl transferases were active. This would mean that all the enzymes were capable of glucosylating an HMC-residue bounded on both sides by another HMC residue.

If base sequence is not the determinant for insertion of glucose residues, where else may we look? Josse and Kornberg (1962) point out that the secondaiy or tertiary structure of the macromolecule may be pertinent here. They indicate the heat-denatured DNA is markedly altered in respect to rate of glucosylation and in some instances extent of glucosylation. Indeed, the decrease in rate of glucosylation is so well correlated with the "melting" of DNA that Zimmerman et al. (1962) suggest the possibility of using the rate of glucosylation as a rapid, sensitive index for measuring heat denaturation of DNA.

At present, the mechanism which directs the positioning of glucosyl units throughout the DNA molecule is unknown. Equally mysterious is the function of glucose in bacteriophage DNA.

60 mavricp: .i. hkssman

VII. Conclusions

Tlio pathway of DNA synthesit; in E. coli and calf tliyinus studied in vitro arc similar in several respects. We may considci- the formation of deoxyribonucleotides as the initial reaction in the metabolic sequence leadino; specifically to DNA. The reduction of lihonucleotides to the corresponding deoxyrihosyl dcrixatives, the phosphorylation of these derivatives and their sui)se(iuent polymerization into the macionioh'culc are being studied as individual reactions in purified systems.

The biological aspects of DNA synthesis, such as cellular growth and pi'oliferation, genetic continuity, and metabolic control, are depend- ent on the individual enzymatic steps in this pathway and are therefore influenced by the physical i)roperties of the enzymes and tlie chemical properties of the reactions. Lord Kelvin said, "When you can measure that of which you speak, and express it in numbers, you know something about it." The detailed biochemical investigations of the individual steps in this sequence are providing the "numbers'" indispensable in arriving at a satisfactory descrijition of the biological pi'ocess and are establish- ing a foundation for the formulation of hyj^otheses necessary to this area of research.

ACKNOW^LEDG M EN TS

I should like to thank Dr. Arthur Kornberg, Dr. John Josse, and Dr. Fnnl J. Bollum for letting me use material which had not yet been published.

Thanks are also due to Mrs. Mary E. Backer for expert assistance in tlu- preparation of the manuscript, and to the people in my laboratory for li\ing with mo (luring this period.

This manuscript was written during the tenure of a grant (C-4088) from the U. S. Pubhc Health Service.

References

.^dler, J., Lehman, I. R., Bessman, M. J., Simms, E. S., and Kornberg, A. (1958).

Proc. Nail. Acad. Sci. U.S. 44, 641. Aposhian, H. V.. and Kornberg, A. (1961). Federation Proc. 20, 361. Aposhian, H. V., and Kornberg, A. (1962). ./. Bml. Chern. 237, 519. Baldwin, R. L.. Inman, R. B., and Wake, R. G. (1961). Federation Proc. 20, 354. Barner, H. D., and Cohen, S. S. (1954). J. Baclcriol. 68, 80. Barner, H. D., and Cohen, S. S. (1959). J. Biol. Chem. 234, 2987. Bello, L. J., Van Bibber, M. J., and Bessman, M. J. (1961a). J. Biol. Cfum. 2:36,

1467. Bello, L. J, Van Bibl)pr, M. J., and Bessman, M. J. (1961b). Biochim. el Biophy><.

Acta 53, 194. Berg, P., and Joklik, W. K. (1953). Nature 172, 1008.

Bertani, E.. Haggmark, A., and Reichard, P. (1961). ./. Biol. Chem. 236, PC 67. Be.s.sman, M. J. (19.59). ./. Biol. Chem. 234, 2735. Bessman, M. J., and Hello, I.. .1. (1961). ./. Bi>,l. Chem. 236, P(^ 72.

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 61

Bessman, M. J., and Van Bibber, M. J. (1959). Biochem. Biophys. Research Com-

muns. 1, 101. Bessman, M. J., Lehman, I. R., Simms, E. S., and Kornberg, A. (1957). Federation

Proc. 16, 153. Bessman, M. J., Lehman, I. R., Simms, E. S., and Kornberg, A. (1958a). J. Biol.

Chem. 233, 171. Bessman, M. J., Lehman, L R., Adler, J., Zimmerman, S. B., Simms, E. S., and

Kornberg, A. (1958b). Proc. Natl. Acad. Sci. UH. 44, 633. Bianchi, P. A., Butler, J. A. V., Crathorn, A. R., and Shooter, K. V. (1961). Bio-

chim. et Biophys. Acta 53, 123. Birnie, G. D., and Crosbie, G. W. (1958). Biochem. J. 69, IP. Blakely, R. L. (1957). Biochim. et Biophys. Acta 24, 224. Bojarski, R. B., and Hiatt, H. H. (1960). Nature 188, 1112. Bollum, F. J. (1958). J. Am. Chem. Soc. 80, 1766. Bollum, F. J. (1959a). J. Biol. Chem. 234, 2733. Bollum, F. J. (1959b). In "The Cell Nucleus" (J. S. Mitchell, ed.), p. 60. Academic

Press, New York. Bollum, F. J. (1960a). J. Biol. Chem. 235, 2399. Bollum, F. J. (1960b). J. Biol. Chem. 235, PC 19. Bollum, F. J. (1960c). Federation Proc. 19, 305. Bollum, F. J. (1962). J. Biol. Chem.. in press. Bollum, F. J., and Potter, V. R. (1958). J. Biol. Chem. 233, 478. Bollum, F. J., and Potter, V. R. (1959). Cancer Research 19, 561. Bollum, F. J., Anderegg, J. W., McElya, A. B., and Potter, V. R. (1960). Cancer

Research 20, 138. Bowman, H. G. (1957). Nature 180, 1182.

Canellakis, E. S. (1959). Proc. 4th Intern. Congr. Biochem., Vienna, 1958, p. 75. Canellakis, E. S., and Mantsavinos, R. (1958). Biochim. et Biophys. Acta 27, 643. Canellakis, E. S., Jaffe, J. J., Mantsavinos, R., and Krakow, J. S. (1959). J. Biol.

Chem. 234, 2096. Caputto, R., Leloir, L. F., Cardini, C. E., and Paladini, A. C. (1950). J. Biol. Chem.

184, 333. Cohen, S. S. (1961). Federation Proc. 20, 641. Cohen, S. S., and Weed, L. L. (1954). J. Biol. Chem. 209, 789. Cohen, S. S., Flaks, J. G., Earner, H. D., Loeb, M. R.. and Lichtenstein, J. (1958).

Proc. Natl. Acad. Sci. U.S. 44, 1004. Colowick, S. P., and Kalckar, H. M. (1943). J. Biol. Chem. 148, 117. Crosbie, G. (1960). In "The Nucleic Acids" (E. Chargaff and J. N. Davidson, eds.).

Vol. Ill, pp. 337-342. Academic Press, New York. Cunningham, L., Catlin, B. W., and de Garilhe, M. P. (1956). /. Am. Chem. Soc.

78, 4642. Davidson, E. A. (1959). Biochim. et Biophys. Acta 33, 238. de Garilhe, M. P., and Laskowski, M. (1956). J. Biol. Chem. 223, 661. Dische, Z. (1955). In "The Nucleic Acids" (E. Chargaff and J. N. Davidson, eds.),

Vol. I, p. 285. Academic Press, New York. Doty, P. (1957). /. Cellular Comp. Physiol. 49, 27. Dunn, D. B., and Smith, J. D. (1954). Nature 174, 305. Edmunds, M. (1958). Federation Proc. 17, 215. Elwyn, D., and Sprinson, D. B. (1954). J. Biol. Chem. 207, 467. Flaks, J. G., and Cohen, S. S. (1957). Biochim. et Biophys. Acta 25, 667.

62 MAURICE J. BESSMAN

Flaks. J. G., and Cohen, S. S. (1959a). J. Biol. Chrm. 234, loOl.

Flaks, J. C, and Cohon, S. S. (1959b). J. Biol. Chcm. 234, 2891.

Flaks, J. G., Lii-htonstein, J., and Cohen, S. S. (1959). J. Biol. Chcm. 234, 1507.

Freese. E. (1959). J. Mol. Biol. 1, 87.

Fricdkin, M. (1959a). Federation Proc. 18, 230.

Friedkin, M. (1959b). In "The Kinetics of Cellular Proliferation" (F. Stohlman, Jr.,

cd.), p. 97. Gruno & Stratton, New York. Friedkin, M., and Kalckar, H. M. (1950). J. Biol. Chem. 184, 437. Friedkin. M., and Kornbcrg, A. (1957). In "The Chemical Basis of Heredity"

(W. D. Mcp]lroy and B. Glass, eds.), j)].. 609-613. John.s Hopkins Press, Balti- more, Maryland. Friedkin. M.. and Roberts, D. (1954). J. Biol. Chem. 207, 245. 257. Gray, E. D., Weissman, S. M.. Richards, J., Bell, D., Keir, H. M., Smellie, R. M. S.,

and Davidson, J. N. (1960). Biorhim. el Biophys. Acta 45, 111. Green, M., and Cohen, S. S. (1957). J. Biol. Chem. 225, 387.

Greenberg, G. R., and Somerville, R. L. (1962a). Proc. Natl. Acad. Sci. U.S. 48, 247. Greenberg, G. R., Somerville, R. L., and De Wolf, S. (1962b). Proc. Natl. Acad. Sci.

U.S. 48, 242. Hammarsten, E., Reichard, P., and Saluste. E. (1950). J. Biol. Chem. 183, 105. Hecht. L. I., Potter, V. R., and Herbert, E. (1954). Biochim. et Biophys. Acta 15,

134. Heidelberger, C, Chandhuri, N. K., Danneberg, P., Mooren, D., Gricsbach, L.,

DiKschinsky, R., Schuitzer, R. J., Plever, E., and Scheimer, J. (1957). Nature

179, 663. Heppel, L. A., Markam, R., and Hilmoe, R. (1953), Nature 171, 1151. Hiatt, H. H., and Bojarski, T. B. (1960). Biochem. Biophys. Research Communs. 2,

35. Hilmoe, R. J. (1960). J. Biol. Chem. 235, 2117.

Humi>hreys. G. K., and Greenberg. D. M. (1958). Arch. Biochem. Biophys. 78, 275. Hurwitz, J. (1959). J. Biol. Chem. 234, 2351. Jesaitis, M. A. (1954). Microbial Genetics Bull. 10, 16. Josse, J., and Romberg, A. (1962). J. Biol. Chem. 237, 1968. Josse, J., Kaiser, A. D., and Romberg, A. (1961). J. Biol. Chem. 236, 864. Kalle, G. P., and Gots, J. S. (1961). Biochim. et Biophys. Acta 51, 130. Reck, R., Mahler, H. R., and Fraser, D. (1960). .4/t/i. Biochem. Biophys. 86, 85. Rhorana, H. G.. Tener, G. M., Razell, W. E., and Markham, R. (1958). Federation

Proc. 17, 312. Rielley, R. R. (1961). Biochim. et Biophys. Acta 53, 150. Rirkland, R. J., and Turner, J. F. (1959). Biochem. J. 72, 716. Rlenow, H., and Anderson, B. (1957). Biochim. et Biophys. Acta 23, 94. Rlenow, H., and Lichtler, E. (1957). Biochim. et Biophys. Acta 23, 6. Koerner, J. F., Smith, M. S., and Buchanan, J. M. (1960). J. Biol. Chem. 235, 2691. Romberg, A. (1957). In "The Chemical Basis of Heredity" (W. D. McElroy and

B. Glass, eds.), p. 589. Johns Hopkins Press, Baltimore. Romberg, A. (1959). Harvey Lectures Ser. 50, 83. Romberg, A. (1960). Science 131, 1503.

Romberg, A., Lieberman, I., and Simms, E. S. (1955). J. Biol. Chem. 215, 389. Romberg, A., Zimmerman, S. B., Romberg, S. R., and Josse, J. (1959). Proc. Natl.

Acad. Sci. U.S. 45, 772. Romberg, S. R., Zimmerman, S. B., and Romberg, A. (1961). J. Biol. Chcm. 236,

1487.

I. REPLICATION OF DNA IN CELL-FREE SYSTEMS 63

Krebs, H. A., and Hems, R. (1953). Biochim. et Biophys. Acta 12, 172.

Kunitz, M. (1950). J. Gen. Physiol. 33, 349.

Laskowski, M., Hagerty, G., Laurila, U. (1957). Nalure 180, 1182.

Lehman, I. R. (1959). Aim. N.Y. Acad. Sri 81, 745.

Lehman, I. R. (1960). J. Biol. Chem. 235, 1479.

Lehman, I. R., and Pratt, E. A. (1960). J. Biol. Chem. 235, 3254.

Lehman, I. R., Bessman, M. J., Simms. E. S., and Komhorg. A. (1958a). ./. Biol.

Chem. 233, 163. Lehman, I. R., Zimmerman, S. B., Adler, J., Bessman, M. J., Simms, E. S., and

Kornberg, A. (1958b). Proc. Natl. Acad. Sci. U.S. 44, 1191. McNutt, W. S. (1958a). J. Biol. Chem. 233, 189. McNutt, W. S. (1958b). J. Biol. Chem. 233, 193. MacNutt, W. S. (1952). Biochem. J. 50, 384. Maley, G. F., and Maley, F. (1959). J. Biol. Chem. 234, 2975. Maley, F., and Ochoa, S. (1958). J. Biol. Chem. 233, 1538.

Mandel, L. R., and Borek, E. (1961). Biochem. Biophys<. Research Communs. 6, 138. Manson, L. A., and Lampen, J. 0. (1951). ./. Biol. Chem. 191, 95. Markham, R., and Smith, J. D. (1953). Biochem. J. 52, 552. Marmur, J., and Doty, P. (1959). Nature 183, 1427. Michelson, A. M. (1958). Nature 182, 1502.

Moore, E. C., and Hurlbert, R. B. (1960). Biochim. et Biophys. Acta 40, 371. Ochoa, S., and Heppel, L. A. (1957). In "The Chemical Basis of Heredity"

(W. D. McElroy and B. Glass, eds.), pp. 633-637. Johns Hopkins Press, Balti- more. Phear, E. A., and Greenberg, D. M. (1957). /. Am. Chem. Soc. 79, 3737. Potter, R. L., Schlesinger, S., Buettner-Janusch, V., and Thompson, L. (1957). J. Biol.

Chem. 226, 381. Pouyet, J. (1952). Compt. rend. acad. sci. 234, 152. Racker, E. (1952). /. Biol. Chem. 196, 347. Reichard, P. (1957). Acta. Chem. Scand. 11, 11. Reichard. P. (1958). Biochim. et Biophys. Acta 27, 434. Reichard. P. (1959). ./. Biol. Chem. 234, 1244. Reichard, P. (1960). Biochim. et Biophys. Acta 41, 368. Reichard, P., and Rutberg, L. (1960). Biochim. et Biophys. Acta 37, 554. Reichard, P.. Baldesten, A., and Rutberg, L. (1961). J. Biol. Chem. 236, 1150. Remy, C. N., Remy, W. T., and Buchanan, J. M. (1955). ./. Biol. Chem. 217, 885. Roll, P. M., Weinfeld, H., and Carrol, E. (1956). J. Biol. Chem. 220, 455. Rose, I. A., and Schweigert, B. S. (1953). J. Biol. Chem. 202, 635. Sable, H. Z., Wilber, P. B., Cohen, A., and Kane. M. R. (1954). Biochim. et Biophys.

Acta 13, 156. Scarano, E. (1958). Biochim. et Biophys. Acta 29, 459. Scarano, E., and Tolarico, M. (1959). Boll. soc. ital. hiol. sper. 35, 97. Cited in Maley

and Maley (1959). Schachman, H. K., Lehman, L R., Bessman, M. J., Adler, J.. Simms, E. S., and

Kornberg, A. (1958). Federation Proc. 17, 304. Schachman, H. K., Adler, J., Radding, C. M., Lehman, I. R., and Kornberg, .\.

(1960). J. Biol. Chem. 235, 3242. Shapiro, S., and Chargaff, E. (1960). Nature 188, 62. Shugar, D. (1960). In "The Nucleic Acids" (E. Chargaff and J. N. Davidson, eds.).

Vol. Ill, p. 39. Academic Press, New York. Siekevitz, P., and Potter, V. R. (1953). J. Biol. Chem. 200, 187. Singer, M. F., Hilmoe, R. J., and Heppel, L. A. (1958). Federation Proc. 17, 312.

64 MAURICE J. BESSMAN

Sinshcinu'r. R. L. (1954). J. Biol. Chcm. 208, 445.

Sinshoimor, R. L. (1954). Science 120, 551.

Sin.-^hoimor, R. L. (1959). J. Mol. Biol 1, 43.

Smellio. R. M. S. (1960). Biochcm. J. 77, 168.

Somorvillo, R., Ebisuzaki, K., and Grccnberg, G. R. (1959). Proc. Null. Acad. Sci.

VS. 45, 1240. Sueoka, X. (1961). J. Mol. Biol. 3, 31.

Swartz, M. N., Trautner, T. A., and Romberg, A. (1962). /. Biol. Chem. 237, 1961. Trautner. T. A., Swartz, M. N., and Kornbcrg, A. (1962). Proc. Natl. Acad. Sci. U.S.

48, 449. Volkin, E. (1954). J. Am. Chem. Soc. 76, 5892.

Wahba, A. J., and Friedkin, M. (1961). J. Biol. Chem. 236, PC 11. Warner, R. C. (1957). J. Biol. Chem. 229, 711. Watson, J. D., and Crick, F. H. C. (1953a). Nature 171, 737. Watson, J. D., and Crick, F. H. C, (1953b). Cold Spring Harbor Symposia Quant.

Biol. 18, 23. Weissman, J. M., Smellie, R. M. S., and Paul, J. (1960). Biochim. et Biophys. Acta

45, 101. Wyatt, G. R., and Cohen, S. S. (1953). Biochem. J. 55, 774. Zimmerman, S. B., and Romberg, A. (1961). /. Biol. Chem. 236, 1480. Zimmerman, S. B., Romberg, S. R., and Romberg, A. (1962). J. Biol. Chem. 237,

512.

i

Chapter II

The Replication and Organization of DNA in Chromosomes

J. HERBERT TAYLOR

Department oj Botany and Department of Zoology, Columbia University, New York, New York

I. Introduction 65

II. A Morphological Picture of the Chromosome 67

A. Mitotic Chromosomes 67

B. Giant Chromosomes 68

III. Replication of DNA 72

A. Chromosomal Level 72

B. Molecular Level 77

IV. Exchanges between DNA Subunits and Genetic Recombination . . 82

A. Chromosomal Level 82

B. Molecular Level 86

V. Secjuences of Replication of DNA in Chromosomes 88

VI. Chromosome Reproduction in Relation to DNA Replication

and Protein Synthesis 89

A. General Comments 89

B. Types of Chromosomal Aberrations Produced at Various

Stages of the Cell Cycle 90

C. Healing and Reunion of Broken Chromosomes 92

VII. Molecular Models for Organization of DNA into Chromosomes ... 96

References 108

I. Introduction

Reproduction is a unique characteristic of living systems. Obviously it can occur at many levels of organization. The largest hiatus in our understanding of reproduction involves problems of the establishment and maintenance of organization above the ordinary molecular level. Most of the enzymatic reactions of the cell can be demonstrated in cell-free systems and at least one virus (tobacco mosaic) can be dis- sociated and reassembled. However, we are far from being able to reconstruct the simplest cell, much less a multicellular organism. In principle the problem is not beyond the realm of possibility for we have the concept that all of the complexity is traceable to properties of the

65

66 J. HERBERT TAYLOR

constituent molecules and their disposition in space and time. Therefore, we attack the problem by tr>-ing to understand the organization of the cell's constituent molecular systems.

In molecular genetics we might begin by trying to understand the reproduction and organization of the genetic stmctures themselves — the chromosomes. The tenii is used here in its broadest sense to designate the entities in which is recorded the major part of the genetic infomiation of cells and the more complex viruses. The smallest of these appear to consist of a single molecule (Chapter IIIi. Whether we refer to a large chromosome as a molecule or not may be a matter of semantics, but evidence indicates that it has some of the properties usually associated uith molecules — a definite size, atomic integrity over long periods of time, and a regular sequence of its constituent parts which are bonded together by rather stable chemical bonds.

A basic hypothesis of molecular genetics is that the synthesis of DXA is a process of replication, i.e.. that new molecules are made by copying a complementary' sequence of the various nucleotides from an original template which is not destroyed in the process. Much of the e\'idence supporting this concept has been presented in Chapter I. The demon- stration of the atomic integrity of two DXA subunits in a chromosome (Taylor et al., 1957) and later in the various particles of DXA isolated from chromosomes (Meselson and Stahl. 1958) was a significant con- tribution to this concept, but these discoveries posed a larger question — a problem of organization, packaging, and sorting that may not be solved soon. The conserved subunits in the DNA particles are ver\- likely the individual polynucleotide chains, although this interpretation has been challenged (Cavalieri and Rosenberg. 1961a). Likewise the two con- serv'ed subunits of a chromatid are almost certainly the same as those in its fragments (the isolated DNA molecules or particles). Some of the questions that are unresolved concern the continuity of the DXA along the chromosome, the nature of linkers if these exist, the singleness or degree of multiplicity of the chromonema. and the pattern of packaging, folding, or coiling in the various structures that a chromosome assumes during the phases of the cell cycle. In addition, the mechanisms which regulate function and replication almost certainly have a basis in the structure and organization of a chromosome, but the details arc largely unknown.

Three possible arrangements include most of the conceptual pictures which have been presented: (/) a multistranded complex of DX'A. pro- tein, and perhaps other constituents with many DX'A double helices forming the axis of the chromonema: i2) some regular arrangement of molecules of DX''.\ or of nucleoprotein particles linked together in tan-

II. KKPLICATIOX OF DNA IN CHROMOSOMES 67

dem with non-DXA linkers, or alternately attached by their ends to a shorter central axis; and (3) one long continuous DNA double helix with its associated protein folded and coiled to form the visible chro- inonema of a typical chromosome.

First a morphological picture of a chromosome will be presented briefly. Tiien the evidence on replication at the chromosomal and molec- ular levels will be summarized. After examining the information avail- able on the jiroperties of the subunit*:, their exchanges and reunion dur- ing genetic recombination, and their repair following radiation-induced breakage, the results will l)e interpreted in terms of a molecular model of a chromosome.

II. A NIoKl HDl.OGICAL PlCTlKE OF THE ChKO.MOSOME

A. MITOTIC CHROMO.^^OMES

Chromosomes present a changing pattern in relation to the cell divi- sion cycle, and during interphase when some of the most important and significant changes are occurring they present little structure which has yielded to analysis with either the light or the electron microscope. Perhaps it will be appropriate to begin a consideration of the chromosome at anaphase when it is composed of a single chromatid. In some favor- able preparations these rather flexible rods can be seen to consist of helically coiled chromoneraata. Usually they appear to have a single chromonema, but in some fixed preparations, probably due to treatment with acid fixatives, the chromatids appear double. In these there may be two helically coiled chromonemata (half-chromatids) usually ver\' poorly separated as if the coils were interlocked and perhaps the axes otherwise bonded together. Fixation in hot water may also reveal the doubleness, but in vivo a doubleness has never been demonstrated. In the living state anaphase chromatids appear to be solid cylinders. An exception may occur in certain coccids which have chromosomes with a multiple or diffuse spindle attachment (Hughes-Schrader. 19401 and in certain symbiotic protozoa of termites (Cleveland, 1949 •. In the sec- ondary' spermatocytes of the coccids the half-chromatids may be com- pletely separated. In the symbiotic protozoa mentioned above half- chromatids are reported to be separated in some cells and not in others of the same species. Whether these move apart during fixation or actually separate in the living cell is not clear. In addition, it is not clear that the half-chromatids in these cases are comparable to the half-chromatids of other organisms. However, it may be noted that the agents (fixatives) which supposedly produce this partial separation would break hydrogen

68 J. HERBERT TAYLOR

1)1)11(18 as well us (k'nalui'c pi'otcins, hut would not he expected to break ('o\'alent bonds.

At the end of anajjhase theic may be a further eondensation of the chromosomes but foUowinji; this phase they expand somewhat in telophase and the coiled chromonemata become less tightly coiled. The chromo- somes then ai)pear to fuse and nearly fill the refoi'iued inteiphase nucleus. As tiie nuclei grow, and particularly in cells that are not to divide again, the chromosomes may not continue to fill the nucleus. Their disposition in most interphase nuclei is difficult to follow either by the use of the light microscope or the electron microscope.

In dividing cells there is typically a gap (G'l) of a few hours before DNA replication begins (see Chapter IV for further details). Synthesis (*S period) then occurs over a period of several hours followed by another gap (G2) before the next division.

By the earliest pro])hase stages at which chromosomes can be dis- tinguished, they can be seen to consist of two chromatids which continue to shorten during prophase. The chromonemata can be seen to be helically coiled in some types of cells. The gyres increase in diameter and their number apparently decreases during prophase as the chromatids shorten.

The coils usually show best in meiotic chromosomes. In addition, the chromomeres of the classical cytologists, which have sometimes been equated with poorly resolved coils, are most striking at the prophase of the first meiotic division. Although chromomere patterns along the chromonemata vary in cells at different stages of development or in different cells in a species (Lima-de-Faria et al., 1959), the mapping of pachytene and dijilotene chromosomes indicates that chromomeres are significant structural features comparable to the bands along the giant chromosomes of Diptera. The ultimate shape and dimensions of chro- monemata, in other words, the packing pattern of the nucleoprotein in the chromosomes, is still largely unknown. Investigation is difficult because the structural details are generally below the level of resolution of the light microscope and often above the size level at which the thin sections required for resolution in electron microscopy can give a com- plete picture. An additional difficulty appears to be an instability of any tertiary structure in the nucleoprotein com))onents of chi'omosomes.

B. GIANT CHROMOSOMES

From a study of structures such as prophase and anaphase chro- mosomes one would easily reach the conclusion that there are many strands along their axes at the molecular level. A chromonema with a diameter of 0.1-0.3 /j, is several orders of magnitude larger than the extended nucleohistone fiber of 30 A. However, in a few special cases

II. REPLICATION OF DNA IN CHROMOSOMES 69

chromosoiiies elongate far beyond the length ob.servcd at early meiotic prophase (leptotene). The giant chromosomes of the Diptera provide one example, but their structure is very complex and they contain many chromonemata. In addition, since these giant chromosomes are end products of differentiation, there may be some reason to question their exact homology with mitotic chromosomes. However, in the oocyte of amphibians a very much elongated chromosome exists which does revert to a regular mitotic type at the following division (Fig. lA). These chromosomes are 500-800 /,<. long in diplotene (Gall, 1956), but change into chromosomes only 15-20 p. in length at later stages. In addition to this shortening by a factor of 40-50 along the axis there are numerous loops which are retracted into the main body of the chromosome. When extended these would give the chromosome a contour length of several centimeters. A few years ago Callan and McGregor (1958) showed that DNase (deoxyribonuclease) would rapidly sever the loops of these isolated lampbrush chromosomes. Although some cytologists had long maintained that the loops were part of the chromatid axis, this was the first clear demonstration that the genetic material, DNA, passed through and maintained the continuity of the loops. In addition to the DNA of the loops, the chromomeres along the axis, which consist of two closely associated chromatids, were known to be Feulgen positive and therefore to contain much of the DNA. This elongation to several hundred times the length of a metaphase chromosome in itself leads one to suspect that much of the bulk of even the early meiotic prophase chromosomes is due to some type of folding or coiling of the much longer axis of one or a few DNA strands. The difficulty of determining the structure of the loops lies in the fact that a great amount of RNA and protein is attached to them. In spite of this neither RNase (ribonuclease) nor proteases will sever the loops. Recently Miller (1962) has been able to dissolve off this protein and RNA coating with concentrated KCl solution and to reveal the fine fibril, presumably DNA, which remains. Electron micrographs reveal its diameter to be sometimes as small as 40-50 A (Fig. IB). Gall (1962) has been able to obtain evidence, by the rate at which the loops are broken by DNase, that the structural axis is com- posed of a single DNA double helix. Since the loops are part of the axis of single chromatids, these elongated structures which are among the largest chromosomes in the animal kingdom are shown to have a single DNA double helix as their linear component. DNase also produces breaks between the chromomeres along the axis of the paired chromatids with the kinetics predicted for a structure held together by a pair of DNA double helices. If all of these results are correctly interpreted we finally have the demonstration that the polymer maintaining the linear

70

J. HERBEin lAVI.OK

Fig. 1A. Ph.ajsi.' pholognipli of ualixed laiupbrush bivalent chromosome from an oocyte of the newt, l^riturus viriilcscois (x650). Present evidence indicates that DNA is the axial element which maintains the continuitj' of the material in each of the loops extending from the main axis. (Original photograph contributed by J. G. Gall.)

II. REPLICATION OF DNA IN CHROMOSOMES

71

Fig. IB. Electron micrograph of a .small portion of a loop of a lampbrush chromosome. Most of the protein and RNA has been dissolved away with a satu- rated solution of KCl after placing an isolated chromosome on Formvar film. The dissolution was interrupted by adding saturated (NH4):S04 at pH 5.1 followed by staining with uranyl acetate (x 75,000). Thinner portion of axial fiber is 40-50 A in diameter. (Original photograph contributed by O. L. Miller.)

integrity of a chromatid is, at some stage anyway, DNA, and tliat a single pair of polynucleotide chains from this axis. Linkers along the axis are not ruled out but neither are any revealed by these experiments. However, long stretches of peptides or RNA which might serve as link- ing material appear unlikely from the fact that proteases and RNase do not break the loops.

In addition to the much extended chromosomes of the amphibian oocytes, other oocytes and spermatocytes of many species have chro- mosomes of a less extreme lampbrush type. These chromosomes as well

72 J. llKHUKRr TAYLOR

as those of the amj^hibian oocyte were called lainphiush ciiromosomes because some early cytologists who studiccl them thought the extensions from the axis were not loops hut bristles like llio.se of a brush. Actually they may be bristles instead of loops at the earlier stages, during pachytene, for example (Tayloi-, 1962). This point has not been estab- lished. However, Moses (1956) demonstrated that an electron-dense axial element exists in the meiotic chromosomes and that when fixed in neutral osmium tetroxide the lateral extensions were fine filaments 50- 100 A in diameter which are often paired where they radiate from the axis. That they contain DNA could be demonstrated by staining adjacent thick sections of the same chromosome by thj Feulgen reaction. Moses (1960) proposetl a model which is consistent with this structure as well as that observed in the nuclei of amphibian oocytes described above.

The axial element cannot be demonstrated after diplotcne by which time the bristles, if that is what they are rather than loops, are being folded or otherwise packaged in the condensing chromosome. Further details of the chromosome's organization is not discernible with the elec- tron microscope as it approaches the metaphase state.

III. Replication of DNA

A. CHROMOSOMAL LEVEL

Until AVatson and Crick (1953a) pi-oposed a specific model for DNA, ideas concerning its mechanism for transmission of genetic information were rather vague. However, following the demonstration by Avery et al. (1944) that the transforming principle of Pneumococcus was DNA, the idea that DNA was the genetic material began to influence the design of experiments. The finding of a constancy in amount of DNA per genome (Boivin et al, 1948; ]\Iirsky and Ris, 1949) and finally the experiments by Hershey and Chase (1952) which indicated that only the DNA of phage entered the bacterial cell made the case more con- vincing. Watson and Crick (1953a) took the information available on bond distances and bond angles, on the most likely tautomeric fonns for the bases, on the X-ray diffraction pictures which suggested a helix, and the chemical analyses, which indicated that DNA contained a 1:1 ratio of adenine to thymine and of guanine to cytosine, and constructed a model in which the purine and pyrimidine bases were directed toward the axis of the helix. By allowing two hydrogen bonds between each of these base pairs in opposing antiparallcl chains which rotated to form a double helix with one gyre per 10 nucleotide pairs, a stable rather stiff molecular model was obtained with the phosphoric acid residues at the periphery of the structure.

II. REPLICATION OF DNA IN CHROMOSOMES 73

The complementarity of the two chains and the specificity of the hydrogen bonds suggested a mechanism for precise replication (Watson and Crick, 1953b). If the chains came apart, each contained the hydrogen bonding sites to line up the nucleotides of a complementary chain. This type of replication would yield two identical molecules (Chapter I). With this proposal of a precise model, the foundations were laid for experimental tests. Techniques and suitable biological materials had to be found. By taking the most optimistic view of all the findings, namely, (1) that DNA was the genetic material of the cell, (2) that all of the DNA of a chromosome was exactly reproduced once each division cycle,

(3) that the proposed structure was correct for all of the DNA, and

(4) that the Watson-Crick scheme for its replication was correct, one could reach the conclusion that one chain of each original DNA molecule in a chromosome or in a whole cell would be passed to each daughter nucleus at mitosis. A new complementary chain would likewise be built along each and passed on to the descendants. If a label (radioactive isotope) were present only in the original DNA or available only for the synthesis of the new chains, the distribution could be predicted. The possibilities were formally stated by Delbriick and Stent (1957): (1) if the original double helix remained intact and somehow produced a completely new two-stranded molecule, the scheme would be conserva- tive; {£) if the chains separated without breakage and each served as a template for the formation of a complementary chain the scheme would be semi-conservative; and (3) if the chains were broken or destroyed in the process the scheme would be dispersive.

Attempts to follow the distribution of carbon- 14-labeled DNA in whole cells at anaphase by autoradiography might have provided a par- tial answer. At least a clear result in a cell with a few chromosomes could have distinguished between a conservative scheme versus a semi-con- servative or dispersive replication. However, the published results (Plaut and IMazia, 1956) did not provide a clear answer although they were interpreted by the authors to favor a conservative scheme.

To get a definitive answer a techniciue that could follow the distribu- tion of label at the chromosomal level was required. This was provided for large chromosomes by autoradiography with tritium-labeled thy- midine, first prepared and used by Taylor et al. (1957). Because of its low energy beta emission tritium gives better resolution for autoradi- ography than any other isotope. In addition, compared to carbon-14 its shorter half-life (12.3 years) allows one to prepare compounds with a much higher specific activity. Since thymidine, which was labeled by an exchange reaction (Taylor, 1960a), is also a selective label for DNA, other components of the cells would not be labeled. Because the intra-

/4 J. HKHMKKT lAVLoH

crllular i)i)()l of thyiiiidinc and it.- (lcii\-ati\-('s i> >iiiall in nio.-t cells, (ho rcniDval of the lal)el after one eycle of re))li('ation would allow one to follow the distribution of labeled DNA through several division cycles and tliei"el)y distinguish a >enii-('onsei\'ati\-e replication from the disper- sive one.

The results for root cells of \'ici(i weic clear (Fig. 2A). After one or a part of a cycle of duplication with thyniidine-H'', the two chromosomes descended from an original unlabeled one were both labeled (Taylor et al., 1957). After one more duplication in the absence of the labeled

Fiu. 2A. Aulonuliognuu of Vicia cliroino.sunie.s at division after one replication in which thymidine-H' was incorporated into the DNA. Fig. 2B (see facing page). .\utoradiogram of a .<5imilar sroup of chromosomes after one replication with thymi- dine-H^ and one additional iiiiluatioii iii an environment free of labeled precursors of DNA (X 3200).

•
4 It m 1	•
* *	«
B

Duplication with labeled thymidine

1st c-metaphase after labeling

Duplication without labeled thymidine

2nd c-metaphase after labeling

Fig. 2B. For legend see page 74.

Fig. 2C. Diagrammatic representation of the distribution of subunits of DNA during chromosome duphcation. Although the subunits are not resolved by micro- scopic examination, they are represented and distributed as indicated by the observed pattern of labeling in Figs. 2A and B. Broken lines represent labeled subunits and unbroken lines represent unlabeled subunits. The dots represent grains in the autoradiograms. (From Taylor et al., 1957.)

75

76 .1. IIIlKHl.Hr TAM.UU

thymidine, the labeled ciiroiiiosoniet> pi'oduced one lahdeil and one un- labeled descendant- (Fig. 2B). This is the result predielfd for senii- conservative icplieation of DNA (Fig. 2C). The examination of other species of plants, Hcllcvcdia (Taylor, 1958a), Crepi-s (Taylor, 1958b), and Allium (Gall, 1958a) confirmed the regular semi-conservative be- havior at the chromosomal level. The fii'st exi)eriments were monitored by the use of colchicine. It was used to accumulate cells at division and to identify the cells which had gone through two cycles. By placing roots in colchicine after labeling and leaving them in this solution until labeled tetraploid cells appeared, we could identify chromosomes which had gone through one cycle of duplication after the one in which they were labeled. These cells regularly showed one labeled chromatid ]:)aire(l with one unlabeled chromatid (Fig. 2B). Chromatid exchanges between the labeled and the unlabeled chromatids tended to obscure the picture, but these later proved of value in analyzing the organization of chromatids (see below).

A report by LaCour and Pelc (1958) criticized the use of colchicine and reported that it had an effect on the segregation of the labeled DNA in Vicia chromosomes. However, a rc-cxamination of the problem by Woods and Schaircr (1959) failed to show any effect of colchicine. In addition the studies on Bellevalia and Crepis (Taylor, 1958a,b) wei'e carried out with and without colchicine. The results with respect to segregation of labeled DNA were the same in both instances. Rare instances of apparent equal segregation of labeled DNA after the second replication in colchicine-treated cells were reported by Taylor (1958a). These could result from exchanges among the four chromatids which were the descendants of a pair of labeled chromatids. Colchicine, which probably keeps such pairs together, would be expected to increase the frequency of this event. In addition an unequal labeling of sister chro- matids may appear at the first division in the first cells to reach division. Usually these have relatively less radioactivity than the majority of labeled division figures. With a low number of grains the significance of any apparent asymmetiy of labeling in sister chromatids is difficult to assess. Recent studies (Taylor, unpublished) with thymidine-H'' of higher specific activity than was available in early experiments reveal what appears to be a significant asymmetry in the labeling of sister chromatids. This is interpreted as evidence for the out-of-phase replica- tion of the two subunits of a chromatid.

Examination of two manmialian species (human and Chinese ham- ster) and an insect (grasshopper) has revealed the semi-conservative distribution of tntium-labeled DNA (Taylor, 1960b and unpublished).

Forro and A\'ertheimer (1960) have also carried out extensive studies

II. REPLICATION OF DNA IN CHROMOSOMES 77

on the segregation of tritium-labeled DNA in whole cells of Eschenchia coli. Cells of E. coll 15T" were exposed to thymidine-H-^ of high specific activity for less than one division time and then removed from the label and allowed to divide a number of times. The distribution of labeled DNA was very heterogeneous in clones derived from single young cells. The pattern of segregation indicated that two large DNA-containing units regularly separate at cell division. These units may then be per- petuated intact for many successive cell divisions, but are subject to fragmentation at a finite probability. These observations are consistent with the view that the two subunits represent the single chromosome, which replicates in a semi-conservative fashion, but is subject to an occasional sister chromatid exchange as observed in higher forms.

These results certainly rule out any dispersive mechanism of replica- tion, but they cannot be taken as definitive proof of a semi-conservative scheme at the molecular level. Because of the uncertainties concerning the structure and organization of chromosomes it was conceivable, although highly unlikely, that the two subunits of DNA in a chromo- some represented a higher level of organization than that of the molecule of DNA.

B. MOLECULAR LEVEL

The first evidence that replication might be semi-conservative at the molecular level was provided by Levinthal (1956). In an autoradio- graphic study of the distribution of phosphorus-32-labeled DNA from phage T2, he obtained evidence that about 40% of the DNA, which remained in one piece when the phage was broken by osmotic shock, was replicated semi-conservatively. The particles obtained from cells infected with P^--labeled phage yielded DNA fragments after similar treatment with one-half as much P^- as the original phage. Doubts concerning the validity of the results were raised when other methods of analysis failed to reveal these large pieces (see Chapter III ; also Davison and Levinthal, 1961). It now appears that the large pieces were probably some sort of artifact and therefore the significance of the original experi- ments remains unclear. However, more convincing results at the molec- ular level were soon presented.

Meselson et al. (1957) published a technique for separating DNA molecules of different densities in a cesium chloride gradient at high centrifugal force. They soon applied the technique to a study of the distribution of labeled DNA during replication in Escherichia coli (Meselson and Stahl, 1958). The cells were grown for a number of generations in N^^-labeled medium. The heavy DNA of these cells could be banded separately from the regular N'^-containing DNA (Fig. 3).

a

b

EXP. NO.

GENERATIONS

0.3

0.7

1.0

I.I

1.5

1.9

2.5

3.0

4.1

Oandl.9 mixed

Oond 4. mixed

Fir;. 3. ((/) I'ltraviolet absorption photograpli.s .showing DXA hand.s rrsulting from density-gradient centrifugation of ly.sates of E. cuh .sampled at various times after the addition of an e.xcess of N^ substrate to a culture of cells labeled with

78

II. REPLICATION OF DNA IN CHROMOSOMES 79

They clianged a rapidly growing culture of E. cull labeled with N'"' to a medium in which the nitrogen source contained the lighter isotope, N^"*. They took samples at intervals as replication continued. When lysates of these samples of cells were centrifuged in the Spinco Model E and photographed at equilibrium, as shown in Fig. 3, the behavior of the labeled DNA undergoing replication was revealed. A new band of DNA with a density intermediate between the N^^- and N^''-labeled material appeared first. By the end of one cycle of replication all of the DNA was in the band with this intermediate density, while the original mate- I'ial with a high density had disappeared. As growth continued a band corresponding in density to N^'^-DNA began to appear. This N^^-DNA continued to increase in amount, but the band of intermediate density did not disappear. These results indicated very clearly that replication is semi-conservative at the molecular level (Fig. 4). ]\Ieselson and Stahl (1958) made another observation of significance. When the DNA of intermediate density was heated for 30 minutes at 100°C and centrifuged again, it formed two separate bands. Therefore, the DNA of intermediate density was composed of hybrid particles in which the separable strands had different densities. The question raised was difficult to answer. Did the replication involve the unwinding of the two polynucleotide chains of a Watson-Crick double helix? Originally this problem of how two chains with one twist every 34 A could separate was one of major con- cern. Further consideration and some calculations of the energy require- ments and the time that might be required for unwinding have indicated that the problem was more imaginary than real. Levinthal and Crane (1956) showed that the energy requirement and the speed to be expected, if the double helix rotated like a cable, were well within the limits available for cellular processes. More recent experiments and calculations have indicated tiiat no problem exists if the replicating units are similar in size to the isolated particles of DNA (Miller. 1961 ; Geiduschek, 1961).

Similar results for DNA of mammalian cells were soon provided. Djordjenic and Szybalski (1960) and Simon (1961) used bromouracil as a density label for DNA and obtained evidence for a semi-conservative replication in human cells (HeLa strain) in culture. Evidence was also

X'^. Each photograph was taken after 20 hours of centrifugation at 44,770 rpm in a .sohition of CsCI; the density of the solution increases to the right. The time of sainphng was measured in units of generation time as indicated on the right, (b) Microdensitometer tracings of the DNA bands adjacent to each photograph. The density of each band can be compared by reference to the two lowest frames which show the .separation of the mixtures indicated on the right. (From Meselson and Stahl, 1958.)

do

J. HKKHKHT TAYLOR

ORIGINAL

PARENT

MOLECULE

FIRST

GENERATION

DAUGHTER

MOLECULES

SECOND

GENERATION

DAUGHTER

MOLECULES

Fig. 4. Illustration of the segrogation of DXA chains according to the scheme proposed by Watson and Crick. If the original labeled DNA contains two chains of high density (X"), the daughter molecules after one replication in a medium containing X'^ would have one chain with X" and one with X". After a second rephcation in the X" medium one-half of the molecules would still have one chain with X'^ and one with X", but the other half would have two chains with only X". All molecules after one generation in the X" medium should have a density inter- mediate between the original X" DX"A and the X" DX^A as shown in the fourth frame of Fig. 3. (From Meselson and Stahl, 1958.)

available that distribution in this same cell line was semi-conservative at the chromosomal level (Taylor, 1960b). More recent studies have shown that the particles of intermediate density from human cells can also be separated into a hea\y and light subunit (Chun and Littlefield, 1961).

A semi-conservative replication has also been demonstrated for phage A (Meselson and Weigle, 1961), phage T7 (Meselson, 1960), and for the alga Chlymadomonas (Sueoka, 1960).

The most serious challenge to the simple interpretation of expcri-

11. REPLICATION OF DNA IN CHROMOSOMES SI

merits on semi-conservative replication which indicate separation of the two complementary chains of a double helix comes from a series of papers by Cavalieri and Rosenberg (1961a,b,c). These authors main- tain, primarily on the basis of the kinetics of molecular weight reduction by enzymes, that DNA extracted from rapidly growing cells is four- stranded. In particular their results indicate that the particles of DNA isolated from E. coli such as those used by Meselson and Stahl (1958) are composed of paired double helices held together by hyj^thetical biunial bonds. According to their interpretation these are the bonds broken by heating and the separation of hybrid DNA observed by Meselson and Stahl involved not the unwinding of the polynucleotide chains, but the separation of a heavy Watson-Crick double helix from a double helix of lower density. Later Hall and Cavalieri (1961) pub- lished electron micrographs of E. coli DNA which indicated that at least some particles of DNA consisted of paired double helices. In addition, the mass per unit length of the particles, calculated from measurements in the electron micrograjihs and molecular weights determined by the light-scattering technique, indicated that the particles contained between 3 and 4 polynucleotide chains. On the other hand, there are a number of experiments which clearly indicate that agents known to break hydrogen bonds cause the separation of the strands of DNA particles. Among these are heat, low or high hydrogen ion concentration, and agents that compete for hydrogen-bonding sites or destroy the water layer of hydration, such as urea and formamide (Doty, 1960; Marmur and Doty, 1961; and Schildkraut et al., 1961). The separated strands will recom- bine if the population of molecules is not too heterogeneous. In addition, hybrid molecules can be produced by heating and slowly cooling DNA's with different isotopic densities.

Whatever may prove to be the significance of these conflicting inter- pretations of experiments on strandedness of different DNA samples, the Watson-Crick scheme for replication, in which specific pairing of com- plementary bases is the principal role of the template, is likely to dominate our thinking and the design of experiments in the foreseeable future (see Chapters III, IV, and V). The semi-conservative mode of replication is likewise not a trivial quirk of nature. A phenomenon so universal would not be preserved in evolution so consistently unless it played a fundamental role in reproduction. The only basis on which it makes any sense is the separation of complementary units as a part of the process of replication. One of the most puzzling jiroblems of chro- mosome reproduction as stated earlier is the packaging and sorting of the tremendously long chromosomal strands if the axial element is a single DNA helix. A model that folds and forms temporary H linkers

82 J. HKIIHKHT TAYLOR

durinp; replication has been propo.scd (Taylor, 1962) to provide a sorting and control mechanism. Fragmentation of such a structure could give the nmltistranded DNA rejwrtcd for rapidly proliferating cells (Cavalieri

and Rosenbei'i;, 19()l;i.l)). Fui-thcr details of the model will he presented in Section VII.

IV. Exchanges between DNA Subunits and Genetic Recombination

A. CHRO.MOSOMAL LEVEL

Sister chi-omatid exchanges had been indicated by the behavior of ring chromosomes in maize (McClintock, 1938; Schwartz, 1955) but the autoradiograms of tritium-labeled chromosomes (Fig. 2) provided unmis- takable visual evidence for their occurrence in other species (Taylor, 1958a, 1959a). Their frequency and pattern could be determined in well- spread preparations of chromosomes at the second division after labeling Exchanges were not revealed in the first division because all chromatids were labeled. When the exchanges became visible they had to be interpreted in terms of events during two duplication cycles — the one in which labeling occurred and the succeeding one.

An analysis was made from root cells of Bellevalia, which has four pairs of large chromosomes (Taylor, 1958a). Attention will be directed to the exchanges in the largest chromosome of the complement. Since we are dealing with diploid cells, there are two of the large homologous chromosomes, each composed of a single chromatid before duplication. Each chromatid contains two subunits of DNA that extend throughout the length— these are the units conserved during replication. Aftei' duplication the chromosomes consist of two chromatids, each with two subunits, one new labeled subunit and one unlabeled original. If the subunits are analogous to the polynucleotide chains of the Watson- Crick double helix, they will be unlike. Reciprocal exchanges between the two chromatids will require the union of labeled subunits with unlabeled subunits in each chromatid because only like chains can join (Fig. 5). Since each chromatid will have a labeled subunit along the entire length at the following metaphase, there will be no visible evidence of the exchange. How(>ver, when the two subunits separate and replicate at the next duplication (referred to below as the second interphase), one chromatid will have the labeled subunit up to an exchange point and beyond that locus its sister chromatid will contain the labeled subunit. The exchanges will now be revealed by autoradiography at the succeed- ing metaphase, because silver grains will be over one chromatid from the end to an exchange and over its sister chromatid from that point to

11. REPLICATION OF DNA IN CHROMOSOMES 83

another exchange. The nietaj)hase chroniosoine derived from tlie other sister chromatid of this originally labek'd j)air will show an exchange at the same locus (the twin exchanges referred to hereafter). The two metaphase chromosomes descended from the other liomolog in these diploid cells will have twin exchanges at different loci if an exchange occurred during its labeling, i.e., the exchanges will be at different loci because they are independent events.

Unpaired (single) exchanges will appear as a result of exchanges during or following the duplication at the second interphase. These will be independent events in all four chromosomes and will be revealed im- mediately when the chromosomes reach metaphase.

The tendency of exchanges to occur as twins was evident in the first sample of cell examined (Table I). In fact these were more numerous than expected compared to the single, unpaired exchanges. If exchanges at the second interphase occurred at the same frequency as in the first interphase, i.e., the one in which the label was incorporated, the ratio would be one twin pair (product of a single reciprocal exchange) to each two single exchanges produced at the second interphase. Since there are four reproducing chromosomes at second interphase if all descendants of the original two homologs are considered, the chance for single exchanges is increased by a factor of two compared to first interphase.

When the two subunits of a pair of chromatids are unlike as assumed in the prediction of the 1 : 2 ratio of twin and single exchanges, all of the exchanges in the first interphase will produce twins and all of those in the second interphase will produce singles. On the other hand, if the two subunits were alike as would be true if each were represented by a Watson-Crick double helix, only one-fourth of the exchanges of the first interphase would yield twins and, of course, no twins would be produced from exchanges at the second interphase. Therefore, the pi'edicted ratio would be one twin pair to each 10 singles (Taylor, 1958a).

The results of the first experiment revealed a 2:1 ratio, i.e., more twins than predicted but a ratio not at all compatible with results pre- dicted in a chromosome consisting of identical subunits. The deviation appeared to be the result of an increased exchange frequency at the interphase when labeling occurred or a decreased one at the second interphase. The latter proved to be correct. Colchicine was present at the second interphase, but not duiing most of the first. It was applied just before the labeled cells, which were being analyzed, reached division. By placing the roots in colchicine at the same time labeling began, i.e., along with the thymidine-H'', the ratio was shifted in the predicted direction, i.e., 1:1 (Table I). In a third experiment the colchicine was added 2 hours before labeling began. In this instance the ratio was

84

J. IIKRHKRT TAYLOR

DUPLICATION with Ihymidine-H^

1st DIVISION

DUPLICATION wi Ihout t hymidine-H ■

2nd METAPHASE

Vrt

1/2 breoks, reunion ond seporotion

single

L i

K t ■i i "■ ir 1'

1 A

J

complete breoks, exchange and restricted reunion

singles at 2nd metaphase

connplete breaks,

exchange and random reunion

Fig. 5. DitiKnims sliowing the predicted results of chromosome reproduction and sister chromatid exchange in which DNA was labeled during one replication with thymidine-H^; the dashed lines represent labeled subunits and the chromosomes shown in solid black represent regions showing grains in autoradiograms. The unbroken lines represent unlabeled subunits and the chromosome regions shown in outline represent regions without grains in the autoradiograms as shown in Fig. 2. (a) The predicted results of exchanges involving only one subunit in each chromatid (these were undetected in the experiments and perhaps do not occur). (6) The predicted results of exchanges when the two subunits are restricted in their reunion by a difference in sense analogous to the two chains of a DX.\ double helix.

II. REPLICATION OF DNA IN CHROMOSOMES 85

very nearly the predicted one, and furthermore, the absolute frequency of exchanges decreased in the first interj:)hase as colchicine was added at earlier times. The frequency of exchanges at the second interphase remained nearly constant in all tliree experiments.

TABLE I Frequency of Single and Twin Exchanges in the Tetraploid Chromosome Sets of Bellevalia romana'^

	Number of chromosomes	Twin exchanges	Frequency per	chromosome
Experiment''	Single First exchanges interphase	Second interphase
1 2 3	72 (18 cells) 112 (28 cells) 96 (24 cells)	36 39 14	15 1.00 36 0 . 70 26 0 . 29	0.21 0.28 0.27

" Data for only the largest chromosome are given.

'' Data for the first two experiments are from Taylor (1958a, b).

The conclusion reached ^vas that the t\vo subunits of a chromatid are unlike. In addition, the exchanges appear to occur during DNA replication, for colchicine must be applied at that time in order to get the maximum effect on their frequency. While these results strongly suggest a chromosome model in which the subunit is a single polynucleo- tide chain of the DNA double helix, they do not exclude models in which the axis of the chromosome is a tandemly linked sequence of DNA helices (Taylor, 1959a,b).

Sister chromatid exchanges have been observed in all species analyzed to date. These include several genera of plants, Vicia, Bellevalia, Crepis, Allium, two mammalian species, Chinese hamster, and human cells in culture, and an insect, the grasshopper (Taylor, 1958a,b; 1959a; 1960b; and unpublished). None of these except Bellevalia have been analyzed for the occurrence of twin exchanges. From what is known of the survival of ring chromosomes in the Diptera, it appears that sister chromatid exchanges may be absent or occur at a very low frequency in this group.

(Exchanges are assumed to occur with an efjual frequency during the first duplica- tion with thymidine-H'' and during the second duplication without thymidine-H^. All exchanges in the first duplication produce twins and the frequency of single exchanges to twin pairs is 2:1; see text for explanation.) (c) The predicted frequency of single and twin exchanges if the two subunits reunite at random. (Since single exchanges could result from events at both the first and second duphcation, the frequency of single exchanges would be higher, at least 10:1.) (From Taylor, 1958a.)

86 J. IIKRHKHT TA-^'LOR

Ilowfwr, to date it has not been pos.-ihic to test this hypothesis directly hy means of hihelin'j:; these I'athci' >iii;ill chroiiiosoines.

li. .\n)i.i:(rLAK lkvel

The exchanges seen at the h'xcl (if the chromosome would not he detectable at the molecular le^-el 1).\" any means now known. For example if the chromosoinc wei'c a tandcmly liiikc(| sci'jcs of DXA particles with molecular weii^hts of 10-20 X 10'', a single large cliromosome would con- tain more than 100,000 of these particles. Therefore, one or two exchanges per chi'omosome per i'ej)lication cycle even if all exchanges involved breakage of the DXA would not be detectable at the molecular level. One must conclude that the chromosom(> is comi)Osed of two remai'kably stable structural units. This stability may l)e more ai)parent than real because the chromosonu' may have higher orders of structural organiza- tion and repair mechanisms to mend breaks (see Section VII). In any case, the DXA of bacteriophages appears to be considerably less stable in linear organization during transmission. Kozinski (1961) has reported a dispersion of labeled phage T4 DXA that would indicate much more frequent breakage than appears to occur in large chromosomes (see Chapter III for further details). The two subunits of phage T7 appear to be much more stable in their linear organization. Most of the subunits remain intact through several replications and Meselson (1960) could demonstrate that the distribution of DXA was semi-conservative. Phage A also undergoes exchange rather infrequently compared to phage T4. Replication has been shown to be semi-conservative, but exchanges occur often enough to be detectable by density-labeling experiments (Meselson and Weiglc, 1961; Kellenberger et nl.. 1961).

Meselson and Weigle (1961) used the hea\y isotopes of nitrogen and carbon to label one parent. In crosses of these with phages containing the light isotopes and appropriate genetic markers, exchange could be shown even between phage chromosomes which did not undergo replica- tion. At low multiplicity of infection (0.1 phage/cell) the original DX^A was distributed to progeny phage pai'ticles as predicted if the original "phage chromosome" was composed of two subunits of DX^A and at the first replication one heaAy subunit was contributed to each new chromosome. These phage chromosomes of hybrid density were not dispersed by exchange during the subsequent replications enough to obscure the essential semi-conser\-ativ(> distribution. However, an actual breakage and exchange even of non-replicating "phage chromosomes" could be demonstrated at a high multiplicity of infection (5-7 ])hage/ cell) with appropi-iate genetic markers. Phage were also recovered which

ir. REPLICATION OF DNA IN CHROMOSOMES 87

contained chromosomes with only one of the original labeled subunits. By the use of the genetic markers phage could also be recovered with the densities expected if they contained a single hea\y subunit that had undergone an exchange between the two markers. The markers used were mi (at one end of the genetic map) and c which is 5 units away. The remainder of the map is 12.5 units. The density of both the con- served particles (containing unreplicated chromosomes) and the semi- conserved particles with a crossover between the markers was that predicted if the chromosome lost an average of about 14% of its heavy kibel, i.e., breaks occurred at random between the loci c and mr.

Kellenberger et al. (1961) found mutants of phage A having different densities. The density differences were apparently due to insertion of extra DNA at the b2 locus in one mutant and at the b5c locus in the other. The markers are 7.5 units apart, i.e., about 43% of the map. The two mutant phages could be separated by centrifugation in a cesium cliloride gradient. Recombination between the two markers yielded par- ticles denser than either parent and also lighter than either parent. When they produced the light phage, labeled its DNA with P^-, and crossed it with the heavy phage, most of the P'^- transmitted remained with the light parental particles. However, the particles recombinant for the density markers contained considerably more P'^- than non-recom- binant hea\y particles. They interpreted the results to show clearly a breakage and exchange of DNA associated with recombination. They also interpreted the evidence to indicate that most of the P'^- transferred to the particles non-recombinant for the density markers was received by breakage and exchange distal to the two markers.

These experiments make a significant contribution to our understand- ing of the mechanism of recombination in phage chromosomes and perhaps in larger chromosomes as well. The experiments are comparable to the classic experiments performed by Stern (1931) and Creighton and McClintock (1931) which demonstrated that genetic recombination in higher organisms involved an exchange of chromosome segments. They go further than the chromosome experiments in that they demonstrate that copy choice in replication is not the only mechanism of recom- bination. In fact there is still no direct evidence that copy choice occurs. However, it is not ruled out by the experiments and is still a convenient mechanism to explain aberrant segregation, perhaps as a part of break- age and exchange events. By putting together the evidence available on physical exchange between chromatids (sister chromatid exchange made visible in tritium-labeled chromosomes) and these recently re- ported data on phage chromosomes, we can formulate tlie hypothesis that

88 J. HERBERT TAYLOR

a variable amount of exchange characteristic for each type of genetic organization occurs in those forms with the smalh-st chi-oniosomos as well as those with the largest ones.

V. Sequences of Replication of DNA in Chromosomes

As already noted, DNA replication is a discontinuous process during the cell cycle. In cells of higher forms the period of synthesis may be 6 to 8 hours or perhaps longer in some cells. Even in cells where replica- tion is completed in much less time, the period is long compared to the time required to synthesize most large molecules. Therefore, the ques- tion of the sequence of replication and its control has often been raised. Evidence at the molecular level is very scarce. The mechanism in which nucleotides are added at the 3' OH ends by DNA polymerase (Chapter I) would appear to require a sequential addition. In addition, the failure to find particles with the whole range of densities when a density label is introduced into a replicating system has led to the sug- gestion that there are few growing points at any moment during the S period. However, the only direct evidence relating to the sequence of replication has been obtained at the level of the chromosome by use of tritium-labeled DNA. These studies now clearly indicate that there are control mechanisms that operate to control the replication among chromosomes within a single nucleus as well as the sequence or order of replication within a single chromosome. This type of control probably has a structural basis (see Section VII on chromosome models).

The first indication of asynchronous replication within chromosomes was the observation that the regions around the centromeres were repli- cated late in root cells of Crepis (Taylor, 1958b). Later Lima-de-Faria (1959) found that the X-chromosome of a grasshopper continued its replication longer than the autosomes. Taylor (1960c) reported a rather complex control in cells of Chinese hamster involving both sex chromo- somes and autosomes. Lima-de-Faria (1959) had suggested that the late replicating DNA was located in heterochromatin, but the two X-chro- mosomes of the Chinese hamster female were different in their replica- tion. These observations suggested that the pattern of replication was a genetic trait, and like other genetic factors could mutate. If late replica- tion was a characteristic of heterochromatin, this state was also mutable or subject to change at certain stages of differentiation (Lyons, 1961 ; Morishima et al., 1962). Recent studies (Morishima et al., 1962) have shown that the heterochromatic X-chromosome of the female human complement is late in completing its replication compared to its homol- ogous mate and to most of the autosomes.

II. REPLICATION OF DNA IN CHROMOSOMES 89

By the iis^c of i)ulse-labeling techniques with thymidine-H^ asyn- chronous hibeling of sectors of chromosomes can be demonstrated in a variety of cells with relatively large chromosomes. However, of perhaps more significance is the observation that a large chi'omosome can be labeled at many sites simultaneously. For example, the Y-chromosome and the late replicating arm of an X-chromosome in the cells of Chinese hamster can be labeled at points along the whole length in 10 minutes of a total replication period of SYz hours for these particular chromosomes (Taylor, 1960c). The late replicating X-chromosome of the human com- plement can likewise be labeled along the whole length in a 10-minute period although it, like the sex chromosomes of the hamster, can be labeled in isolated sectors when supplied with the isotope very near the end of the S period (Morishima et al., 1962). This rapid labeling of DNA along the whole length of chromosomes was also noted in root cells of Bellevalia (Taylor, 1959a). These observations must mean that a large chromosome has several and probably a very large number of sites where DNA replication may be occurring simultaneously.

VI. Chromosome Reproduction in Relation to DNA Replication and Protein Synthesis

A. general comments

Chromosome reproduction involves replication of DNA and synthesis of proteins at least in higher organisms. Synthesis of RNA may be involved directly in the process but this has not been demonstrated so clearly as in the case of the other two polymers. DNA replication can be dissociated from protein and RNA synthesis in bacteria. By using a mutant of Escherichia coli that requires arginine, uracil, and thymine for growth, Maal0e and Hanawalt (1961) showed that DNA would increase by 40% or more without additional protein or RNA synthesis. This would be the amount formed if all nuclei in synthesis should finish but not begin another round of replication. By restoring arginine and uracil to allow a short period of protein and RN'A synthesis, another replication could then be initiated. Cohen (1962) has found a change in state of the DNA after this transition period. If the DNA is isolated be- fore protein synthesis is allowed to occur, in cells starved for arginine and uracil, the isolated DNA can be precipitated with streptomycin. After the period of protein synthesis and the initiation of a new cycle of replication, the DNA isolated from these cells is no longer precipitated by streptomycin. It is perhaps significant that the single-stranded DNA of ^X174 is not precipitated by streptomycin. Cavalieri and Rosenberg (1961a) also reported a change in state of E. coli DNA when cells were

90 J. IIKRBERT TAYLOR

starved for thymine for 45 minutes and then returned to a complete medium. DNA sanijilos folliM-ted before and for a short period after restoring thymine had a nioh'cuhir \vci<;ht twice that of DNA isohited from cells ahout one-half houi' after restoring thymine. They also found an associated change in the kinetics of degradation by DNase which they interpreted as a change from four-stranded DNA to two-stranded DNA. Further information on changes in state of DNA is given in Chapters III and IV.

B. TYPES OF CHROMOSOMAL ABERRATIONS PRODUCED AT VARIOUS STAGES OF THE CELL CYCLE

Methods for detecting changes of chromosomes in relation to the duplication cycle are quite limited. One of the most revealing techniques is the study of changes produced by ionizing radiations. Soon after the mutagenic effects of X-rays had been demonstrated, there were attempts to determine the time of chromosome duplication by the type of aberra- tions observed after irradiation at different stages of the nuclear cycle in root cells (]\Iather and Stone, 1933). In early experiments most of the aberrations observed were the type expected if the chromosomes were broken before reproduction. Their conclusion was that chromosomes w^re single during most or all of interphase and reproduction occurred late in the division cycle. Their work was criticized by Huskins and Hunter (1935) on the basis that the stage irradiated could not be determined accurately, and that cells observed might not have been at the first division after treatment. Huskins and Hunter and also Mather (1937) improved on the design of the early experiments by selecting microspores of higher plants as experimental material. After meiosis the haploid microspores formed pass through a long interphase of several days to 2 or 3 weeks depending on the species and temperature of development. By irradiating in this interphase and analyzing cells at the first microspore division, tliey could be certain that the chromosomes were observed at the first division after treatment. However, Huskins and Hunter (1935) reported that the chromosomes of Trillium were already double when irradiated. Matliei' (1937) fovuid tliat treatment at eaily intcM'phase resulted in only chromosomal type aberrations, i.e., those produced before production of the two chi'omatids. At later stages of interphase single chromatids of a pair could be broken. Huskins and Hunter (1935) contended, along with others, that the degree of subdivision of a chromosome could not be revealed by irradiation because the passage of an ionizing paiticle could sever several suhunits at once. Another possibility is tiiat a chromosome broken befoi'c repi'oduction may i-emain open and react with other broken ends after reproduction.

II. REPLICATION OF DNA IN CHROxMOSOMES 91

AVhilc both of these possibilities may be realized in some instances, the reproduction of chromosomes determined by the types of aberrations produced shows a good correlation with other events that would be expected to occur during reproduction. Although both chromatids of a chromosome may be broken simultaneously (isolocus breaks) the break- age of one without the other being affected reveals a kind of reproduction or change of state. Although broken ends may not undergo reunion immediately, experimental evidence indicates that most breaks rejoin or the ends exchange with other ends produced by breaks close by within an hour or less in metabolically active cells (Wolff and Luippold, 1956a).

Sax (1941) studied the change in state of chromosomes of Trades- cantia microspores by analyzing the types of aberrations in cells fixed at hourly intervals after X-radiation. The change from single chromo- somes to double chromosomes occurred between 32 and 26 hours before metaphase. When the time of DNA synthesis was determined in these same cells by autoradiographic studies (Taylor, 1953) the two events were shown to coincide. Later a study of DNA synthesis by autoradi- ography and cytophotometric techniques at this stage and in the subse- quent replication in the generative nucleus again showed a close correlation of the change to doubleness to radiation and the increase of DNA (Moses and Taylor, 1955). Thoday (1954) found a similar correlation in root cells of Vicia. The change in state did not appear to be abrupt but a gradual one spread over several hours. The experi- ments were not designed in such a way that they could reveal whether this difference was due to an asynchrony in the cell poinilation or a gradual change of the chromosome complement of a single cell as its DNA was replicated. Recent experiments indicate that the change may be rather abrupt and may occur before DNA replication (Wolff, 1961 ; J. H. Taylor, unpublished). However, the abruptness of the change may vary with the cell type. In the cells studied DNA replication extends over a period of several hours and is sometimes asynchronous with respect to whole chromosomes or sectors of these. In Vicia and BeUevalia root cells, all chromosomes are labeled along most of their length nearly simultaneously, but in cells from other sources (Crepis roots, Chinese hamster fibroblasts, human leucocytes, and grasshopper gonial cells) individual chromosomes or sectors are asynchronous in their replication as indicated by labeling with thymidine-H^ (Taylor, 1960c; Morishima et (iL, 1962). A recent report by Chu and Giles (1961) shows that a chromatid exchange and a chromosome exchange can occur in a single cell of tlie human where the labeling mentioned above indicates asynchrony.

All of tlic information available indicates that the changed state,

92 J. nKHMKin' tavi.ok

from that in wliich chroinosonici^ act as single units in the formation of aberrations to the state in which they react as two units, is correhited in time with DNA replication. The change either occurs progressively during synthesis or rathci- ahiuptly preceding syntlicsis. The earlier discrepancies reported by lluskins and Hunter (1935) and by Mather (1937) are very likely the result of variations in the biological material. The time of synthesis of DNA has not been reported foi- tlic two genera used, Trillixim and Allium. However, in Tradescantid thi' change and synthesis both occur in late interphase, while in Tiilbaghin both DNA synthesis and the change in state occur in very caiiy interphase (Taylor, 1958c).

In meiosis where chroniosonie ivpi-oduction has often been supposed to occur in the long prophase of the first meiotic division, the change in state from single to double occurs in interphase (Mitra, 1958; Sauerland, 1956). Likewise DNA rei)lication and the synthesis of basic proteins of the nucleus are correlated with the change (Taylor, 1959c; De, 1961). The synthesis of basic proteins (histones) of chromosomes is apparently correlated with DNA replication in other cell types (Alfert, 1955; Bloch and Godraan, 1955; Gall, 1959). In all of these studies the correlation of changes in response to ii'radiation and the synthesis of DNA and proteins lack the precision in timing for specific chromosomes or regions of chromosomes that would be desirable. Further discussion of a change in state of DNA in relation to its replication cycle will l)e found in Chapter IV and the last section of this chapter.

Another type of aberration induced })y radiation at limited stages of the cell cycle are half-chromatid exchanges. They can be produced only in mitotic prophase (Sax, 1957) and in the stages after pachytene in meiosis (Mitra, 1958; Grouse, 1954). Although these connections are strong enough to lead to breakage of chromatids as they stretch in anaphase, the bridges do not persist through the subsequent interphase and usually do not become chromatid aberrations at the next division (Ostergren and Wakonig, 1954). After anaphase the chromatids revert to a state in which they act as if composed of single axial elements in breakage and reunion rather than doul)le as in the pi'evious jirophase. However, there is some evidence that they can become double to radia- tion breakage before DNA synthesis begins. See the last section of this chajjter for a model that could b('lia\'e in this fashion.

C. HEALING AND REUNION OF BROKEN rHROMO.'^OMES

The synthesis of the major components of chromosomes is restricted to one-half or less of the cell cycle in many cells of higher organisms. Yet, the repair or reunion of broken chi'omosomes can occur at other

II. REPLICATION OF DNA IN CHROMOSOMES 93

stages of the cell cycle. Pei'haps the clearest case is the breakage of chromosomes in the prophase stages in Lilmm after a demonstration of DNA replication is no longer possible. These breaks produce chroma- tid bridges in abundance (Mitra, 1958). Other examples are the produc- tion of chromatid exchanges in pollen tubes and in the late interphase in roots. Breaks that are produced before DNA replication in the G-^ stage also undergo reunion or exchange befoi'e i-eplication. This is indicated by the yield of chromosome type aberrations in some cells and perhaps more convincingly by the results of variations of intensity and fractionation of dose in radiation studies (Sax, 1939; Wolff and Luippold, 1956a). The yield of one-hit aberrations, i.e., single breaks, is linear with dose, but exchanges that require interaction between ends produced from separate breaks increase almost as the square of the dose. If the dose is given over a longer period so that most breaks are repaired before another is produced close enough for interaction, the curve of yield to dose approaches linearity. On the other hand, one-hit aberrations show no change with variations in rate, but are proportional to total dose. There is also a decrease of the two-hit type when the dose is given in two fractions separated by an interval long enough for nearly complete healing of lesions. Wolff and Luippold (1956b) obtained evidence for two kinds of aberrations in Vicia roots, a group that undergoes reunion in about 1 minute and a second group that rejoins within 1 to 2 hours. Since these experiments were carried out with Vicia seeds soaked for 18 to 24 hours in water, many of the cells are presumably in G^. Exposing the cells to tritiated thymidine revealed no DNA synthesis (Wolff, 1960). These results indicate that broken chromosomes can rejoin before DNA i-eplication. By exposing the roots to chloramphenicol (300/xg/ml) the incorporation of glycine-C^* was reduced by 13-33%. When the chloramphenicol was given before or after a single dose of radiation it had no effect on aberrations, but when given for the period between two doses it prevented the usual split dose effect, i.e., the breaks from the first dose remained open and capable of exchanging with those produced by the second. Aureomycin had a similar effect, but penicillin did not prevent rejoining. On the basis of these experiments Wolff (1960) maintains that protein synthesis is necessary for chromosome reunion and supposes that the linear structure is maintained by protein. Since DNA replication could not be demonstrated during the period of rejoining, he assumes DNA is not involved.

However, recent experiments (Taylor et al., 1962) indicate that DNA replication may be involved in chromosome reunion. Fluorodeoxyuridy- late is a potent inhibitor of thymidylate synthetase. When fluorodeoxy- uridine (FUDR) is presented to cells, many of tlu>m ha\-e enzymes for

94 J. JIEHHEHT TAYLOR

its phosphorylation. The mu-lt'osido inonoiiliosphatc hinds with the enzyme that normally convei'ts deoxyui'idylate to thyniidylate and bloeks synthesis. In the ahsnicc of thymidine tripliosijhate, DNA repli- cation ceases (sec Chapter II. The root cells of Vicia faba are quite sensitive to the inhibitor. At a coiicciiti-ation of 10 "il/ the rate of cell division drops to a low Ic\t1 within (5 hours, l)ut tiie cells which arrive at division have intact chromosomes. At higher concentrations {10''' M and higher) the chromosomes show extensive damage. The damage appears about 3 hours after immersion of roots in the inhibitor. Lesions, which appear as non-staining gaps in the late prophase, metaphase, and anaphase chromosomes, are the first visual signs of damage. Some of the gaps are complete discontinuities of the chromatids and free fragments are the results. Frequently, but not regularly, both chromatids are affected at the same locus. Many of the last cells to arrive at division show an extensive shattering of all or most of the chromosomes. The lesions apparently occur in cells near the end of their DNA replication cycle when the enzyme, thymidylate synthetase, is quickly blocked. Such cells have no mechanism to prevent their progress through division even though their DNA replication is not complete. The lesions can be pre- vented by supplying exogenous thymidine along with the inhibitor. In addition, the lesions can be cured by giving the cells thymidine at least 1 hour before anaphase, when they are presumably in late interphase or early prophase. Bromodeoxyuridine can be substituted for thymidine in the healing of these lesions. Evidently, DNA replication, which requires all four nucleotides including thymidylate, is necessaiy for healing the lesions.

An interesting feature of the fragmentation is tlie absence of chroma- tid exchanges or any aberrations that will produce bridges at anaphase. This observation led to the idea that DNA replication might be required for reunion of chromosomes broken by radiation. A test of the hypothesis appeared possible because a concentration of 10" M stopped cell division and presumably DNA replication without yielding visible chromosome damage. By giving cells a dose of 2o r of X-irradiation 4 hours before fixing, anaphase bridges were produced in 6.4% of the cells. If the cells had been placed in 10"'^ ilf FUDR for 30 minutes before irradiation and were grown in the same solution until fixed 4 hours later, only 1.1% of the anaphases had bridges. By supplying tliyniidiiie or bromodeoxy- uridine no-' M) to the irradiated cells treated with FUDR, the per- centage of bridges could be restored to nearly the control level, 5.4 and 5.5%, respectively. From tliese experiments the hypothesis was projiosed that reunion of broken chromosomes requires DNA replication. The process was visualized as involving two steps. The first would be a very

II. REPLICATION OF DNA IN CHROMOSOMES

95

rapid one that does not require high energy bonds, i.e., base pairing of two single polynucleotide chains (Fig. 6). The two chains of the DNA double helix were assumed to break at different points in many cases. The union by base pairing could be a very rapid process. The second step would presumably be the healing of the lesion by the growth of two antiparallel DNA chains, each copying the other. If the growth occurs

Fig. 6. Diagram of the steps involved in the reunion of broken chromosomes according to the hypothesis described in the text. Fragments of two replication units of different broken chromosomes are shown. The lines represent the subunits of chromatids, presumably single nucleotide chains. (From Taylor et al.. 1962.)

by the action of the enzyme DNA polymerase (Chapter I) only chains with 3' OH groups at the end can grow. If the breaks leave a 3' phos- phate group at the end, an enzyme might be required to chop the chain and prepare it for growth. In addition, about one half of the unions by base pairing would be expected to occur between non-growing ends, i.e., those ending in a 5' phosphate or 5' OH group. In these the shorter complementary chains ending in a 3' OH might be expected to grow and bypass the original point of union of nongrowing ends. This could result in a reopening of the break and result in new exchanges or to a union

96 J. HERBERT TAYLOR

throu<ili till' I'ontiiiiK'd ^i-owtli of the 3' ends. One of the most interesting consequences of such a mechanism of reunion of broken ends would be the elimination of single strands of DNA over a part of the replicating unit. Perhaps more interesting is the idea that it would j)n)vide a mecha- nism for copy choice ovvv short segments in combination with breakage and exchange between homologous chromatids. Such a mechanism of reunion could exjilain all of the types of aberrant recombination reported in ascomycetes (Olive, 1959; Kitani et a/., 1961; Case and Giles, 1958; Rizet et ai, 1961) and perha|)s also in phages (P>igar, 1961 ».

The basis for the above hypothesis is somewhat weakened by subse- quent studies of reunion of radiation-induced breakage. Chromatid ex- changes apjiarently can occur in the presence of the block by FUDR. This would suggest that many of the breaks do not involve DNA or that another component, probably protein, is able to bring about I'eimion. Such a mechanism is suggested by the model of chromosome structiu'e presented in the following section.

VII. Molecular Models for Organization of DNA INTO Chromosomes

Models which allow us to explain data, to express more concretely certain concepts, and to plan and design future experiments are useful. However, models have the disadvantage that they tend to freeze our thinking in such a way that significant data may be neglected. When we are fully aware of such disadvantages models can be useful, even when the information for making them is limited. Several recent attempts have been made to construct models of a chromosome (Schwartz, 1955; Bloch, 1955; Taylor, 1957; Ris, 1957, 1961; Gall, 1958b; Freese, 1958; Taylor, 1958d; Steffensen, 1959; Stahl, 1961). All of these are useful in explaining some data and in ciystallizing concepts, but all are deficient in other respects.

Any model of a chromosome based on the present information should be composed of one, or at most a few, DNA double helices that extend the length of a chromatid. These must function as two subunits during replication at all levels of organization. There is also considerable evidence that the chromatid functions as two subunits in non-reciprocal recombination (Olive, 1959) and in the induction and segregation of mutants induced by base analogs (Strelzoff, 1962). However, in chemi- cally and radiation-induced breakage and exchange as well as I'eciprocal recombination and sister chromatid exchanges it acts as a single unit except during prophase. On the other hand, the pattern and frequency of sister chi'oinatid exchanges indicate that, even though the exchanges

II. REPLICATION OF DNA IN CHROMOSOMES 97

are only between whole chromatids, each chromatid is composed of two unlike subunits (Taylor, 1958a). When all of the evidence is considered, the basis for this "twoness" would appear to be the complementary chains of the DNA double helix with some tertiary structure added (see below).

However, evidence is accumulating whicli indicates that chromosomal fibrils are double at many stages in the division cycle. For a review of the evidence from electron microscopy see papers by Ris (1957, 1961), Kaufmann et al. (1960), and Pappas and Brandt (19611. In interphase nuclei fibrils about 100 A in diameter are seen in cells of many species. Often these are paired. The pairs of 100 A fibrils could very well repre- sent the two chromatids. However, the 100 A fibrils which appear in spermatids before the chromosomes have been reproduced may still contain two DNA double helices, for according to Ris (1961) these change to 40 A fibrils in most spemi nuclei. Although the decrease in diameter could result from a change in the protein with consequent variations in the pitch of a superhelix of the DNA double helix, it is possible that a significant characteristic of genetic material is being overlooked in a preoccupation with models based on single DNA double helices. The studies of Cavalieri and Rosenberg (1961a, b) on the physical properties of isolated DNA and Hall and Cavalieri's (1961) studies on the mass per unit length of DNA fibrils isolated from rapidly dividing E. coll also indicate that DNA from dividing cells is four-stranded while that from nondividing cells is two-stranded, i.e., is composed of single double helices. Is it possible to reconcile these observations with those which indicate that the two complementary chains of a single DNA double helix are the subunits for replication, mutation, and per- haps recombination?

A model based on a single DNA double helix which extends through the length of a large chromosome without interruption has some of the features required by our present data, but other necessary features would appear to be missing. One of these is the packing or folding of several centimeters of DNA for manipulation within the dimensions of a cell. The DNA or "chromosome" of phage T4 is more than 50 /x long when fully extended (Chapter III). The chromosome of E. coli would be perhaps 50 times this length (more than 2.5 mm). Some of the largest chromosomes might be more than a meter in length (Taylor, 1957). Even if these prove to be composed of several strands of DNA, they would still be several centimeters in length. The most useful models to visualize the folding and unwinding of such a long piece of DNA are those based on a suggestion by Freese (1958). The essential feature of the model is a DNA double helix with a regular sequence of linkers alternating in the

98 J. in:RRERT taylor

two chains and located opposite a gap in the complementaiy chain. For modifications and other features of the model see papers by Taylor 0958(1; 1959a. h; 1962) and Kcllenbcrgcr (1960). Linkers have not yet been identified in analyses of DNA and indeed this can hardly be expected if they separate segments with molecular weights of several millions. Since such linkers can only be found by a search directed specifically to the problem and then probal)ly only after improvement in present te('hni(iues, a little sj^eculation on the nature of the linkers may be useful in the interim.

In addition to a capacity for folding and efficient packaging which might be provided for by irregularities such as odd bases, for example, methylaminopurine (Dunn and Smith, 1955 and 1958), a chromatid must have many growing points for replication and yet be closed to replication for long periods of function. The initiation and control of replication would appear to require operator sites to explain the observed control in order or sequence and to prevent more than one replication at each locus. To provide growing points breaks might be made by an enzyme, but to explain how the two subunits of a chromatid can remain intact or undergo only a few exchanges during replication the breaks must be produced in a very precise manner which would require regular structural sites for initiation of replication. To provide the necessary controls one may have to imagine several specific inducers for replication, the speci- ficity of which need not reside in the type of linkers but in the neigh- boring nucleotides. In addition, a minimum of four types of linkers appears to be required. The first type of linker needed is one to join 5' OH or 5' phosphate groups at the ends of two polynucleotide chains with a reversal in polarity (Fig. 7). This will be referred to as a 5' linker. Their complementary chains will be assumed to be joined by a 3' linker. The chromatid is assumed to consist of a series of tandemly linked segments (replication units) of DNA with a 3' OH group linked to a 3' OH group and a 5' end linked to another 5' end at each operator site. Any irregularity in single chains of the DNA double helix wouM appear to be inadequate, for these would tend to be redistributed by inversions or other chromosomal aberrations in such a way as to prevent the subunits from maintaining their alternating sequence of linkers. Without such a sequence there would be numerous mistakes in sorting during replication and the subunits would not maintain their integrity. With linkers of the type proposed above, chromosomal aberrations might change the length of segments but would not destroy the alternating sequence of 3' and 5' linkers.

The 5' linkers would always be closed and in non-replicating DNA the 3' linkers would also be closed. One of the first steps preparatoiy to

II. REPLICATION OF DNA IN CHROMOSOMES

99

LINKER

LINKER

r^

^V/

^X/

r%j

/VJ

Fig. 7. A portion of a chromosome preceding replication. The plectonemically coiled strands represent molecules of DNA joined in tandem array by means of special sites or linkers in each polynucleotide chain. These linkers are assumed to occur in pairs at sites in which there is a reversal in polarity so that two 3' and two 5' ends of chains are linked by 3' linkers and 5' linkers, respectively.

replication would be the enzymatic opening of 3' linkers. However, a necessary' preliminary step would be the establishment of H linkers (Fig. 8) which are assumed to stabilize the chromatid by holding it in a folded ladder-like arrangement (Taylor, 1962). In the model the H linkers are shown as some type of polymer joining alternate 5' linkers along two axes. These would represent the half-chromatids, but during

100

J. HERBERT TAYLOR

5' LINKER

Fig. 8. Before or during replication the fhromosome is assumed to be stabilized by the addition of H linkers to form a double axis, each of which with its attached DXA chains will become a chromatid by the time rejilication is complete. The stabilizing material is r(>presented as a polymer attacluni to alternate 5' linkers. Although shown as a continuous chain for simplicity in illustration, it probably would consist of many small subunits attached in some way as yet unknown.

II. REPLICATION OF DNA IN CHROMOSOMES 101

and following replication would form part of the linear axi.s of the two sister chromatids. A new set would be formed on each chromatid at prophase when it became double, i.e., when half-chromatid bridges can be induced. In rapidly dividing cells these linkers might be closed most of the time, but in cells where the chromosome acts as a single unit in breakage and reunion these linkers would have to be assumed to be open or rather labile. Since protein synthesis appears to play a role in the reunion of chromosomes, the H linkers for the present will be assumed to consist of polypeptides.

As soon as the H linkers are stabilized the chromosome should be capable of acting as a double structure in breakage and exchange. In Vicia. this appears to occur for all chromosomes before DNA replication begins or at the beginning of the S period (see Section VI). In other species where there is a marked asynchrony of replication, the stabiliza- tion would be expected to occur in some chromosomes or sectors of chromosomes before its completion in others.

Replication is assumed to begin at a pair of the 3' ends, perhaps by the insertion of a 5' linker and the closure of the 3' ends by a new and different 3' linker not susceptible to the enzyme which is operating to open the regular 3' linkers. Since this 3' linker is necessary only during the replication cycle, we will refer to it as the 3'R linker. It must be changed to a regular 3' linker before another replication can occur. Some evidence suggests that the change may occur during anaphase or telophase (see below). To allow rotation and untwisting of the DNA helices the 3' linkers should be open on the side opposite the insertion of the new 5' linkers. If there w^ere an out-of-phase replication of the two subunits of a chromatid, unwinding would appear to be facilitated. However, in an organism where the rate of DNA replication may be a factor in limiting growth, E. coli for example, the cell must have evolved some control so that both subunits can be replicating simultaneously.

When two adjacent chains finish replication we will assume that the 3' ends of the two new chains are now closed with a 3'R linker (Fig. 9). The result would be a replicated package consisting of a pair of double helices attached to one axis of the original chromatid. The comple- mentary chains attached to the other axis would now replicate and form other pairs of helices attached to a single axis. When both chroma- tids w'ere complete, each would consist of a series of the four-stranded packets of DNA attached to a linear axis with a length that might very well be manipulated in the dimensions available in a cell (Fig. 10). For example, in E. coli if the units of replication had a molecular weight of 2.4X10*"' (Cavalieri and Rosenberg, 1961a), and were attached 40 A apart along the axis, the over-all length of the chromosome would be only

102

J. HERBKH'r TAVI.OH

5' LINKER

3" LINKER

Fig. 9. During replication the 3' linkers are assumed to be oi)ened by appropriate enzymes. The two subunits of the chromosome consisting of tlie alternate 5' linkers and the H linkers with the attached DNA chains are then held together only liy the hydrogen bonds in Uic DXA. Two clKiins of adjacenl iikiIccuIcs are assumed to l)egin replication i»y the insertion of a new r>' linker.

II. REPLICATION OF DNA IN CHROMOSOMES

103

3' LINKER

Fig. 10. As each pair of chains begins repHcation, a new 3'R hnker is assumed to be inserted which is resistant to the attack of the enzyme that opens the regular 3' Hnkers. When the pairs of chains finish rephcation their growing 3' ends are assumed to be closed by the addition of another 3'R linker. The two new chromatids would consist of these pairs of DNA molecules attached to a single axial element. Another axial element (half-chromatid) would be added to each during the succeed- ing late interphase or prophase.

im J. HHRUHRT TAYLUK

3 or 4 .u. A larjio ohromosomo would likewijio bo of a size so tliat one or two setc? of coils in the ribbon formed by the ladder-like arrangement of DXA helices would produce the structure seen with the light micro- scope (Taylor. 1959d: Taylor ct ai. 1957 1.

In G: each chromatid should consist of a single axial element made o\ alternating H linkei-s and .V linkei's. DXA extracted at this stage should be a very poor primer for DXA polymerase unless the chains in replicated units were broken to provide growing points. By some stage of prophase when half-chromatid exchanges can be induced the new set of H linkers may be assumed to have formed and become st^ibilized so that each chro- matid has two axial elements. By late anaphase or early telophase some chromatids become separable into two subunits (in most chromosomes which are twisted the plectonemic coiling of the two axial elements, half-chromatids, prevents complete separation^ The separation can be brought about by agents which would be expected to break hydrogen bonds and perhaps denature proteins, but would not be expected to break covalent bonds. This is the basis for the idea that the 3'R linkei-s may be opened at this time and presumably replaced by regular 3' linkers at a later st^*ige. If that is the correct interpretiition. the DX^A might now be converted to a potentially active primer if the enzyme for opening the regular 3' linkers were available. However, there is no assurance that this is the time of the change even if the model is correct. We know that diromosomes may undergo duplication without passing through anaphase. However, in the cases studied there is an endomitotic reorgani- zation of some sort before a second duplication occui"s.

The model can be used to account for most or all of the presently knouii properties of chromosomes. It is based on morphological and biochemical evidence drawn from a variety of sources where information in all cases is at best fragmentary. The morphological counterparts are the lampbrush chromosomes and the structures found by Pappas and Brandt (1960: and personal communication^ in the interphase nucleus of amoeba (Fig. W. The structures consist of remarkably well- preserved helical filaments clustered as if attached by one end to an axial element. I interpret these to be G._. chromosomes of the type shown in the model (Fig. lOV The helical filaments have a g\'re diameter of about 3(XV3oO A and an over-all length of 3000-4000 A. The fiber wound into this helix is about 120 A in diameter and has a periodicity indicating that a smaller fiber of about 40 A is wound into a helix to form this r20-A fiber. At certain loci the coiled filaments are separated enough to show that some at least are double, i.e.. consists of two strands of lo- 20 A (Fig. 12V Measurements indicate that the ultimate helical filaments would have a contour length of 3-4 «. Each of these two associated

II. REPLICATION OF DNA IN CHROMOSOMES

105

Fig. 11. Helical structures revealed by the electron microscope in interphase nuclei of Amoeba proteus. These helices appear to be attached along an axis which is assumed to represent a portion of a chromosome after replication according to the hypothesis in the text. (Photograph from Pappas and Brandt, 1960.)

!()()

J. HKHUKHT 'I\\'\I.()H

Fig. 12. Higher magnification of the hohcal structin-es .shown in Fig. 11. At jjlaces the helices (aryuLcs) can be seen to be double and the fibrils wound into the large helix have a periodicity of density indicating that this fibril is formed from a helically coiled ultimate fibril of about 30 A. (Original photograph contributed by George D. Pappas and PhiUip W. Brandt.)

11. REPLICATION OF DNA IN CHROMOSOMES 107

strands (DNA double helices perhaps with some associated histone) would have a molecular weight of 6-8X10" (Taylor, 1960b) and presumably would represent the replicating unit in amoeba.

The lampbrush chromosomes seen in meiosis may be a similar type of structure except that the axial element has a much greater density to the electron beam CMoses, 1956, 1960). Recently Bullivant (1962) has demonstrated helically coiled filaments in metaphase chromosomes of mammalian cells. The gyre diameter is about 75 A and these helices occasionally appear to be coiled into a super helix comp.irable to those found in the amoeba. The chromosomes in amphibian oocytes have loops, the axis of which appear to be single DNA double helices (Miller, 1962). These could be the pairs of replication units completely uncoiled which have opened up into rings. When stretched along the axis they break (Callan, 1956) at what would be the 3'R linkers in the model.

The pairs of replication units in the model could also correspond to the four-stranded DNA units reported by Cavalieri and Rosenberg (1961a,b). If a chronuUid built according to the model is stripped of its protein (including the H linkers) during Go, its weak points will be the 5' linkers. The separated units will consist of rings in which one polynucleotide chain is interrui)ted with two free 5' ends. If one assumes some type of biunial bonds the two double helices may remain paired. In rapidly dividing cells with a veiy short G^ stage most of the DNA might exist in this state. Upon heating the rings could open and allow unwinding of the polynucleotide chains, with a decrease of the molecular weight by a factor of two and the separation of chains of different densities. Thus the observations of Meselson and Stahl (1958) on the separation of hybrid molecules into single chains could apply to four- stranded DNA as well as two-stranded DNA.

The model has at certain stages some characteristics of a model composed of tandemly linked rings, a model which has been suggested l)y Stahl (1961) to explain recombination events in higlier organisms. However, at other stages the model acts like a side-chain model for purposes of genetic recombination (Taylor, 1962).

Very little can be said at this time about the nature of the hypo- thetical 3', 3'R, and 5' linkers. The H linkers have been assumed tenta- tively to be peptides, but it now appears that the other linkers cannot be peptides. At the chromatid level the best evidence comes from the report (Callan and McGregor, 1958; (lall, 1958b) that proteases and RNase do not sever the loops of the lampbrush chromosomes while DNase quickly disrupts them. In addition, isolated deproteinized DNA is not significantly degraded by proteases. A recent report by Bendich and Roscnkranz (1962) indicates that sperm DNA contains about one

108 .1. HKHUKRT TAYLOR

phosphoseriiK' reisidiu's pvv thousand nucleotides. This suggests that the reguhir 3' linkers or perhaj)s the 3'R linkers might consist of two l')hosphoserine residues coupled to the terminal nucleoside of DNA chains by an ester linkage in the same way that amino acids are coupled to transfer RXA. A diphosphate bridge could then couple two chains with a reversal in i)olarity. The linkage might rei)resent an activated state ready for the initiation of replication or for a change of the linker if it should i)rove to be the 3'R linker. The 5' linker should be formed by covalent bonds which might be as stable as the i)hosphodiester linkages along the iiolynucleotide chains. Although there are several possible types of molecules that could serve as linkers, it would be premature to make further guesses at this time.

Although the model presented is designed primaiily to explain the replication and sorting of subunits in chromosomes, the traffic control problem in template function might be mentioned. How is a chromosome able to open for replication without functioning as a template for RNA synthesis as ap{)ears to be true for the heteropycnotic X-chromosomc and jierhaps other such genetically inactivated regions on autosomes? Or what is more common, how does it oj:)en for RNA replication without serving as a template for DNA synthesis when all the enzymes and precursors for replication are operating at other sites in the nucleus? The cell could solve this problem by having special ojierator sites which are opened for RNA replication, l)ut wiiich could not serve as starting ends for DNA polymerase. Let us supjjose that a DNA polymerase is available to repair random breaks in DNA by adding to free 3' OH ends and that the same or a similar polymerase can begin to copy at the proposed operator sites for DNA replication. However, the special sites opened for RNA synthesis would not be available as starting loci for these DNA polymerases because of the occurrence of some odd base, or other end group, for example. Such a mechanism would seem to be necessary to simplify the problem of traffic control in the operation of a template which can serve either for RNA synthesis or DNA synthesis. The principal difference would be that DNA replication requires opening and unwinding of the two chains of the DNA. RNA replication either can occur without strand separation or the chains can open at special sites and then close again without destroying the integrity of the DNA.

References

.Vll'crt, M. (1955). St/mposium on Fine Structure of Cells. Leklcn, 195.'t. Intern. I'nion Biol. Set. Publ. Ser. B, 21, 157. P. Noordhoff, Groningen, Netherlands.

Avery, O. T., MacLeod, C. M., and MrCarty, M. (1944). J. Exptl. Med. 79, 137.

Bendich, A., and Rosenkranz, H. S. (1962). In "Progress in Nucleic Acid Research" (W. E. Cohn and J. N. Davidson, eds.), Vol. I. Acaileniic Press, New York. In press.

II. REPLICATION OF DNA IN CHROMOSOMES 109

Bloch, D. (1955). Proc. Natl. Acad. Sci. Wash. 41, 1058.

Bloch, D., and Godman, C. G. (1955). J. Biophys. Biorhcm. Cytol. 1, 17.

Boivin, A., Vendrely, R., and Vendrely, C. (1948). Compt. rend. acad. set. 226, 1061.

Bullivant, S. (1962). Personal communication; State of New York, Dept. of Health,

Albany, New York. Callan, H. G. (1955). Symposium on Fine Structure of Cells, Leiden, 1954. Intern.

Union Biol. Sci. Publ, Ser. B, 21, 89. P. Noordhoff, Groningen, Netherlands. Callan, H. G., and McGregor, H. C. (1958). Nature 181, 1479. Case, M. E., and Giles, N. H. (1958). Cold Spring Harbor Symposia Quant. Biol.

23, 119. Cavalieri, L. F., and Rosenberg, B. H. (1961a). Biophysics J. 1, 337. Cavalieri, L. F., and Ro.senberg, B. H. (1961b). Biophys. J. 1, 317. Cavalieri, L. F., and Rosenberg, B. H. (1961c). Biophys. J. 1, 323. Chu, E., and Giles, N. H. (1961). Proc. Natl. Acad. Sci. U£. 47, 830. Chun, E. H. L., and Littlefield, J. W. (1961). J. Mol. Biol. 3, 668. Cleveland, L. R. (1949). Trans. Am. Phil. Soc. 39, 1. Cohen, S. S. (1962). Personal communication; Dept. of Bioch. Univ. Penn. School of

Med., Philadelphia. Creighton, H., and McChntock, B. (1931). Proc. Natl. Acad. Sci. U.S. 17, 492. Crouse, H. V. (1954). Science 119, 485.

Davison, P. F., and Levinthal, C. (1961). Atin. Rev. Biochem. 30, 641. De, D. N. (1961). The Nucleus 4, 1. Delbruck, M., and Stent, G. S. (1957). In "The Chemical Basis of Heredity"

(W. D. McElroy and B. Glass, eds.), pp. 699-736. Johns Hopkins Press,

Baltimore. Djordjenic, B., and Szybalski, W. (1960). J. Exptl. Med. 112, 509. Doty, P. (1960). Harvey Lecture 55, 103-139.

Dunn, D. B., and Smith, J. D. (1955). Biochem. J. 60, xvii (abstract). Dunn, D. B., and Smith, J. D. (1958). Biochem. J. 68, 627. Edgar, R. S. (1961). Virology 13, 1.

Forro, F., and Wertheimer, S. A. (1960). Biochem. et Biophys. Acta 40, 9. Freese, E. (1958). Cold Spring Harbor Symposia Quant. Biol. 23, 13. Gall, J. G. (1956). In "Mutation," Brookhaven Symposia in Biology No. 8, pp. 17-

32. Brookhaven National Laboratory', Upton, New York. Gall, J. G. (1958a). Personal communication; Dept. of Zool., Univ. Minn., Minne- apolis. Gall, J. G. (1958b). In "The Chemical Basis of Development" (W. D. McElroy and

B. Glass, eds.), pp. 103-135. Johns Hopkins Press, Baltimore. Gall, J. G. (1959). /. Biophys. Biochem. Cytol. 5, 295.

Gall, J. G. (1962). Personal communication; Dept. Zool., Univ. Minn., Minneapolis. Geiduschek, E. P. (1961). Proc. Natl. Acad. Sci. U.S. 47, 950. Hall, C. E., and Cavalieri, L. F. (1961). /. Biophys. Biochem. Cytol. 10, 347. Hershey, A. D., and Chase, M. C. (1952). J. Gen. Physiol. 36, 39. Hughes-Schrader, S. (1940). Biol. Bull. 78, 312.

Huskins, C. L., and Hunter, A. W. S. (1935). Proc. Roy. Soc. B117, 22. Kaufmann, B. P., Gay, H., and McDonald, M. R. (1960). Intern. Rev. Cytol. 9, 77. Kellenberger, E. (1960). 10th Symposium Soc. Gen. Microbiol., pp. 39-66. Kellenberger, G., Zichichi, M. L., and Weigle, J. (1961). Proc. Nail. Acad. Sci. U.S.

47, 869. Kitani, Y., OHve, L. S., and El-Ani, A. S. (1961). Science 134, 668.

iiiiiiii

Mil

ift**i>imit*K<v:< «•-•.

110

.1. IIKHUKirr TAYLOR

Kozinski, A. W. (liKH). Vndlagy 13, 121.

LaCour, L. V.. an.i Pole, S. K. (1958). Nature 182, 506.

Levinthal. C. (1956). Proc. Xnll. Arml. Sci. U.S. 42, .3!)4

Lovinthal, C. and Crano, H. R. (1956). Proc. Natl. Acad. Sci. U.S. 42, 436.

Linia-dc-Faria, .\. (1959). J. BiopJiys. Biorheni. Cytol. 6, 457.

Lima-do-Faria, A., Sarvella, P., and Moni.-^. K. (1959). Hcnilila.^ 45, 467.

Lyons, M. E. (1961). Nature 190, 372.

Maal0e, O., and Hanawalt, P. C. (1961). ./. Mol. liiol. 3, 144.

McClintock, B. (1938). Genetics 23, 315.

Mather, K. (1937). Proc. Roy. Soc. B124, 97.

Mather, K., and Stone, L. H. A. (1933). J. Civuvt. 28, 1.

Marmur, J., and Doty, P. (1961). J. Mol Biol. 3, 585.

Meselson, M. (1960). In "The Cell Nucleus" (J. 8. Mitcholl. cd.), i-p. 240-245.

Academic Press, New York. Meselson, M., and Stahl, F. W. (1958). Proc. Natl. Sci. 1.8. 44, 671. Meselson, M., and Weigle, J. (1961). Proc. Natl. Acail. Sci. i'Ji. 47, 857. Meselson, M., Stahl, F. W., and Vinograd, J. (1957). Proc. Natl. Acad. Sci. U.S.

43, 581. Miller, I. R. (1961). J. Mol. Biol. 3, 229. Miller, O. L. (1962). Personal connnunication : Biol. Div., Oak Ridge Natl. Lab.,

Oak Ridge, Tenn. Mirsky, A. E., and Ris, H. (1949). Nature 163, 666. Mitra, S. (1958). Genetics 43, 771. Morishima, A., Grumbach, M. M., and Taylor, J. H. (1962). Proc. Natl. Acad. Sci.

U.S. 48, 756. Moses, M. J. (1956). J. Biophys. Biochem. Cytol. 2, 215. Moses, M. J. (I960). 4th Intern. Kongr. Elektro)ienmikroskopie. Berlin, 1958, pp.

199-211. Moses, M. J., and Taylor, J. H. (1955). Exptl. Cell Res. 9, 474. Ostergren, G., and Wakonig, T. (1954). Botan. Notiser 4, 357. Olive, L. S. (1959). Proc. Natl. Acad. Sci. U.S. 45, 727. Pappas, G. D., and Brandt, P. W. (1960). /fth Intern. Kongr. Elek tronenmikroshopie ,

Berlin, 1958, pp. 244-246. Plaut, W. S., and Mazia, D. (1956). J. Biophys. Biochem. Cytol. 2, 573. Ris, H. (1957). In "The Chemical Basis of Heredity" (W. D. McElroy and B.

Glass, eds.), pp. 23-62. .Johns Hopkins Press, Baltimore. Ris, H. (1961). Can. J. Genet. Cytol. 3, 95. Rizet, G., Lissouba. P., and Moussoar, J. (1961). Bull. soc. fran{-. physiol. vegetalc

6, 175. Sauerland, H. (1956). Chromosoma 7, 627.

Sax, K. (1941). Cold Spring Harbor Symposia Quant. Biol. 9, 93. Sax, K. (1939). Proc. Natl. Acad. Sci. U.S. 25, 225. Sax, K. (1957). Quart. Rer. Biol 32, 15.

Schildkraut. C. L.. Marmur, J., and Doty, P. (1961). ./. Mol Biol 3, 595. Schwartz, D. (1955). ./. Cellular Comp. Physi(d. 45, Suju'l. 2. 171. Simon, E. H. (1961). ./. Mol Biol 3, 101. Stahl, F. W. (1961). J. chim. phys. 58, 1072-1077. Steffensen, D. (1959). "Structure and P'unction of Genetic Elements," Brookhaven

Symposia in Biology No. 12, pp. 103-124. Brookhaven National Laboratory,

Upton, New York.

I

Taylor, Taylor, Taylor. Taylor, Taylor, Taylor, Taylor, Ta.vlor,

U Taylor, Taylor,

pp. Tavlor,,

Taylor, .

Tavlor, ,

II. REPLICATION OF DNA IN CHROMOSOMES

111

li.

Stern, C. (1931). Biol. Zentr. 51, 547.

Strelzoff, E. (1962). Zeitsch. Vererb. 93, 301.

Siieoka, X. (1960). Proc. Natl. Acad. Sci. U.S. 46, 83.

Taylor, J. H. (1953). Exptl. Cell Research 4, 169.

Taylor, J. H. (1957). Arn. Naturalist 91, 209.

Taylor, J. H. ('l958a). Genetics 43, 515.

Taylor, J. H. (1958b). Exptl. Cell Research 15, 350.

Taylor, J. H. (1958c). .4m. J. Botany 45, 123.

Taylor, J. H. (1958d). Sci. American 198, 36.

Taylor, J. H. (1959a). Proc. 10th Intern. Congr. Genet.. Montreal, 1958, pp. 63-78.

Taylor, J. H. (1959b). Oklahoma Conj. Radioisotopes in Agr., 1959, U.S. Atomic

Energy Comm. TID-757S. pp. 123-129. Taylor, J. H. (1959c). Am. J. Botany 46, 477. Taylor, J. H. (1959d). In "Symposium on Molecular Biology" (R. E. Zirkle, ed.),

pp. 304-320. Univ. Chicago Press, Chicago. Taylor, J. H. (1960a). Advances in Biol, and Med. Phys. 7, 107. Taylor, J. H. (1960b). In "Cell Physiology of Neoplasia" (M. D. Anderson Hosp.

and Tumor Inst.), jip. 547-575. Univ. Texas Press, Austin. Taylor, J. H. (1960c). J. Biophys. Biochem. Cytol. 7, 455. Taylor, J. H. (1962). Intern. Rev. Cytol. 13, 39. Taylor, J. H.. Woods, P. S., and Hughes, W. T. (1957). Proc. Natl. Acad. Sri. U.S.

43, 122. Taylor, J. H., Haut, W. F., and Tung, J. (1962). Proc. Natl. Acad. Sci. V.S. 48, 190. Thoday, J. M. (1954). Neiv Phytologist 53, 511. Watson, J. D., and Crick, F. H. C. (1953a). Nature 171, 737. Watson, J. D., and Crick, F. H. C. (1953b). Nature 171, 964. Wolff, S. (1960). Am. Naturalist 44, 85. Wolff. S. (1961). Radiation Research 14, 517 (abstract).

Wolff, S., and Luippold, H. E. (1956a). In "Progress in Radiobiology" (J. S. Mitchell, B. E. Holmes, and C. L. Smith, eds.), pp. 217-221. Oliver and Boyd, Edinburgh Wolff. 8., and Luippold, H. E. (1956b). Proc. Natl. Acad. Sci. U.S. 42, 510. Woods, P. S., and Schairer, M. V. (1959). Nature 183, 303.

110 .1. IIKUKKKl lAVLoK

Kozinski, A, W. (I'tOl). r/co/oy;/ 13, 121.

LivCour. L. v.. :.iul IMc S. H. (1958). Xalurc 182, 506.

Levinthal, C. (1956). I'mc. Xall. Acad. Sci. V.S. 42, 394

Lovinthal, C, and Crane, H. R. (1956). Pror. Xntl. Arml. Sci. UJ>. 42, 436.

Lima-dc-Faria. \. (1959). J. Biophys. liiochcm. Ct/lol. 6, 457.

Lima-de-Faria, A., Sarvdla. P., and Morri.><. H. (1959). Hcifiltltis 45, 467.

Lyons, M. E. (1961). Xahirc 190, 372.

Maaldo, O., and Hanawalt, P. C. (1961). ./. Mol. Bt„l .3, 144.

McClintock, B. (1938). Genetics 23, 315.

Mather, K. (1937). Proc. Roy. Sor. B124, 97.

Mather, K., and Stone, L. H. A. (1933). J. (hurt. 28, 1.

Marmur. J., and Doty, P. (1961). J. Mol. Biol. 3, 585.

MeseLson, M. (1960). In "The Cell Nucleus" (J. S. Mitrhdl. ci.). iip. 240-245.

Afademic Press, New York. Meselson, M., and Stahl, F. W. (1958). Proc. Xall. Scl. i'Ji. 44, 671. Meselson, M., and Weigle, J. (1961). Proc. Xatl. Acad. Sci. U.S. 47, 857. Mesel.-^on. M.. Stahl. F. W.. and Vinograd, J. (1957). Proc. Xatl. Acad. Sci. L\S.

43, 581. Miller, I. R. (1961). ./. Mol. Biol. 3, 229. Miller, 0. L. (1962). Personal communication; Biol. Div., Oak Ridge Natl. Lab.,

Oak Ridge, Tenn. Mirsky, A. E., and Ris, H. (1949). Xalmr 163, 666. Mitra, S. (1958). Genetics 43, 771. Morishima, A., Grumbach, M. M.. and Taylor, J. H. (1962). Proc. Xatl. Acad. Sci.

U£. 48, 756. Moses, M. J. (1956). J. Biophys. Biochcm. Cytol. 2, 215. Moses, M. J. (1960). .'fth Intern. Kongr. Elcktroncnmikroskopie. Berlin, 195S, pp.

199-211. Moses, M. J., and Taylor. J. H. (1955). Exptl. Cell Res. 9, 474. Ostergren, G., and Wakonig, T. (1954). Botan. Xotiser 4, 357. Ohve, L. S. (1959). Proc. Xall. Acad. Sci. l'.S. 45, 727. Pappas, G. D., and Brandt, P. W. (1960). -y//) Intern. Kongr. El( kironrnmikroskopie,

Berlin, 1958, pp. 244-246. Plaut, W. S., and Mazia, D. (1956). ./. Biophys. Biochcm. Cytol. 2, 573. Ris, H. (1957). In "The Chemical Basis of Henniity" (W. D. IVIcElroy and B.

Glass, eds.). pp. 23-62. Johns Hopkins Press, Baltiin(jr(>. Ris, H. (1961). Can. J. Genet. Cytol. 3, 95. Rizet, G., Lis.souba, P., and Mou.s.sear, J. (1961). Bull. .soc. frang. physiol. vegelale

6, 175. Sauerland, H. (1956). Chromosonta 7, 627.

Sax, K. (1941). Cold Spring Harbor Symposia Quant. Biid. 9, 93. Sax, K. (1939). Proc. Natl. Acad. Sci. U.S. 25, 225. Sax, K. (1957). Quart. Rev. Biol. 32, 15.

Schildkraut. C. L., Marmur, J., and Doty, P. (1961). ./. Mol. Biol. 3, 595. Schwartz, D. (1955). J. Cellular Comp. Physiol. 45, Supjil. 2, 171. Simon, E. H. (1961). J. Mol. Biol. 3, 101. Stahl, F. W. (1961). J. chim. phy.^. 58, 1072-1077. Steffensen, D. (1959). "Structure and Function of Genetic Elements," Brookhaven

Symposia in Biology No. 12, i)p. 103-124. Brookhaven National LaboratorA-,

Upton, New York.

II. REPLICATION OF DNA IN CHROMOSOMES 111

Stern, C. (1931). Biol. Zentr. 51, 547.

Strelzoff, E. (1962). Zeitsch. Vererb. 93, 301.

Sueoka, N. (1960). Proc. Natl. Acad. Sci. U.S. 46, 83.

Taylor, J. H. (1953). Exptl. Cell Research 4, 169.

Taylor, J. H. (1957). Am. Naturalist 91, 209.

Taylor, J. H. (1958a). Genetics 43, 515.

Taylor, J. H. (1958b). Exptl. Cell Research 15, 350.

Taylor, J. H. (1958c). Am. J. Botany 45, 123.

Taylor, J. H. (1958d). Sci. Amencan 198, 36.

Taylor, J. H. (1959a). Proc. 10th Intern. Congr. Genet.. Montreal. 1IJ5S. pp. 63-78.

Taylor, J. H. (1959b). Oklahoma Con}. Radioisotopes in Agr., 1059, U.S. Atomic

Energy Comm. TID-7578, pp. 123-129. Taylor, J. H. (1959c). Am. J. Botany 46, 477. Taylor, J. H. (1959d). In "Symposium on Molecular Biology" (R. E. Zirkle. ed.),

pp. 304-320. Univ. Chicago Press, Chicago. Taylor, J. H. (1960a). Advances in Biol, and Med. Phys. 7, 107. Taylor, J. H. (1960b). In "Cell Physiology of Neoplasia" (M. D. Anderson Hosp.

and Tumor Inst.), pp. 547-575. Univ. Texas Press, Austin. Taylor, J. H. (1960c). J. Biophys. Biochem. Cytol. 7, 455. Taylor, J. H. (1962). Intern. Rev. Cytol. 13, 39. Taylor, J. H., Woods, P. S., and Hughes, W. T. (1957). Proc. Natl. Acad. Sci. U.S.

43, 122. Taylor, J. H., Haut, W. F., and Tung, J. (1962). Proc. Natl. Acad. Sci. U.S. 48, 190. Thoday, J. M. (1954). New Phytologist 53, 511. Watson, J. D., and Crick, F. H. C. (1953a). Nature 171, 737. Watson, J. D., and Crick, F. H. C. (1953b). Nature 171, 964. Wolff, S. (1960). Am. Naturalist 44, 85. Wolff, S. (1961). Radiation Research 14, 517 (abstract).

Wolff, S., and Luippold, H. E. (1956a). In "Progress in Radiobiology" (J. S. Mitchell, B. E. Holmes, and C. L. Smith, eds.), pp. 217-221. Oliver and Boyd, Edinburgh Wolff, S., and Luippold, H. E. (1956b). Proc. Natl. Acad. Sci. U.S. 42, 510. Woods, P. S., and Schairer, M. V. (1959). Nature 183, 303.

Chapter III

The Organization of DNA in Bacteriophage and Bacteria

CHARLES A. THOMAS, JR.

The Johns Hopkms University, Baltimore, Maryland

I. Bacteriophage DNA 113

A. The Organization of DNA in the Intact Phage Particles . . . . 114

B. The Structure of the Liberated Phage DNA Molecule .... 121

C. The Genetic Structure of Phage 131

D. The Distribution of Parental Phage DNA to Progeny Phage ... 138 II. Bacterial DNA 141

A. Cytology of Bacteria 142

B. Growth and Transfer Experiments 142

C. The Structure of DNA Liberated from Bacteria 145

D. The Genetic Structure of Bacteria 145

E. Conclusions 148

Acknowledgments 148

References 148

I. Bacteriophage DNA

The extensive development of the field of bacteriophage genetics has led to an increasing interest in the structure and organization of bacterio- phage DNA. Since bacteriophage are the smallest biological units which are susceptible to thorough genetic analysis, and since their DNA is best described in molecular or chemical terms, a better understanding of the structure of phage DNA promises to lead to a chemical interpreta- tion of genetic recombination, replication, and transcription. In addition to the genetic functions of DNA, this molecule has the important ability to assume a folded, compact configuration such as found in chromosomes, spores, and sperm heads. A form of this tertiary structure of DNA is found in the compact arrangement of the DNA inside bacteriophage. One may hope that the study of the organization of DNA in the phage head will lead to an understanding of the principles and requirements of this folding process.

It is convenient to divide these studies into those performcMl on the intact virus particles, and those on the liberated DNA molecules.

113

114 CHARLES A. III(iMA>. .IK.

A. THK OUCANIZATION OK DNA l.\ lllK INTACT 1'HA(;K I'AKTICLES

1. Stmlics on Purtldlh/ Oriented Pre p(iratie>n:^

If the DNA in the phage hi'ail were organized in a i^jjecial manner, it would be likely that this would be revealed by a preferential align- ment of regions of the molecule with respect to the long axis of the ])hage.

When a concentrated suspension of anisometric particles is subjected to a stress, the long axes of the particles are preferentially aligned in the direction of the stress. It was l)y this technique that tobacco mosaic virus (TMW) was first aligned (Bernal and Fankuchen. 1941). By allowing a drop of concentrated T2 phage suspension to dry under a cover slip on a microscope slide, Bendet, Goldstein, and Lauffer (1960) fountl that there was a negative birefringence with respect to the direc- tion of flow, while TM\' displayed a j:)Ositive birefringence in agreement with the studies of others. Since the birefringence of TMV is known to be positive (Franklin, 1955), while the birefringence of DNA is nega- tive (Wilkins, Gosling, and Seeds, 1951), it was concluded that there is a preferential alignment of DNA molecules in the direction of the long axis of the phage particle. As the authors point out, there is a great danger that DNA is ejected as a result of shearing the concentrated suspension and certain precautions were taken to make this less likely.

A more quantitative study of this same phenomenon was done by Gellert (1961, 1962) employing flow-birefringence techniques. These ex- periments revealed that the rotary diffusion constant, and hence the fraction of particles that could be considered perfectly oriented at any rate of shear, was essentially the same for intact phage particles and for purified ghosts. The sign of the birefringence w^as exactly opposite: positive for the ghosts, and negative for the intact phage (Fig. 1). By subtracting these values, and knowing the fraction of particles that could be considered perfectly aligned with the flow, Gellert concludes that the phage DNA is so organized in the head that 10 to 15*^ of it appears to be perfectly aligned with the long axis of the phage, if the remainder is considered to be arranged at random. The form birefrin- gence associated with the slightly anisometric cavity in which the DNA resides would account for only a small fraction of this. Thus it appears that there is an internal organization of the phage DNA. and this organ- ization is reflected in the birefringence.

In principle more detailed information can be had by X-ray diffrac- tion studies on oriented phage preparations. Such a study was made by North and Rich (1961). Their X-ray diffraction patterns show a high degree of orientation, and a veiy strong reflection at 3.4 A which is

III. BACTKRIOFHAliE UNA AND BACT?:K1AL UNA

115

characteristic of DNA. In an effort to rule out the l)os;^ibility that these reflections came from extra-particular DNA, they studied the 24 A equatorial spacing as a function of relative humidity. Their results show that the fibers containing phage do not show the continuous in-

60	■		T4B GHOSTS	0
40	■	X *
20	. /
BIREFRINGENCE o	x\
-20	■ \
-40		^	T4B

1000 3000 5000

SHEAR (sec-)

7000

Fig. 1. The birefringence of phage particles and ghosts. This figure shows the way in which the birefringence (mole"^ Uter"') changes as a function of shear gradient for intact T4 phage (lower curve) and phage ghosts (upper curve). If the DXA were completely randomized within the phage head, the birefringence of intact phage would be expected to superimpose approximately with that of the ghosts. (From Gellert, 1962.)

crease in this side-to-side packing distance as the relative humidity approached 100%. With fibers of purified DNA this spacing increases indefinitely as the fiber dissolves. In this regard, the fibers containing phage are similar to DNA-protamine complexes. At present it is not

Fig 2 \ T2 pliage i-articli' Rcntly disniptrd by shear. Many equivalent i.ietures show the presence of a bulge, which is variable in size and generally about 5000 A from the broken phage coat. The marker indicates a length of 1 /t. (From Ruben- stein, 1960.)

116

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 117

possible to conclude from these studies what proportion of the DNA is oriented parallel to the phage tail.

Similar studies have been made by Maestre and Kilksom (1962), who took great care to avoid and to evaluate DNA ejection. Their results for T2 and M-5 (a long-tailed phage) showed clear 3.36 A reflections. The M-5 pictures also showed a strong 34 A meridional spac- ing and a 52 A equatorial spacing. This confirms the notion that a significant fraction of the DNA is oriented parallel to the phage tail.

2. Partially Disrupted Phage Particles

During the course of making electron micrographs of phage particles, occasionally a disrupted particle can be seen with filamentous DNA emerging from it. The first studies of this kind were done by Fraser and Williams (1953). A tangled coil of filaments can be seen very close to the protein ghost from which it presumably emerged. Rubenstein (1960) noticed that phage particles could be broken by hydrodynamic shear on the electron microscope grid in such a way as to draw out the DNA from a tear in the head membrane. This treatment reproducibly resulted in a long DNA fiber, presumably consisting of many DNA molecules, with a characteristic blob or knot of material near the middle of the strand. It is probably safe to conclude that this characteristic knot is a result of some residual tertiary structure of the DNA. But so far no direct information can be drawn about the tertiary structure itself. It is possible that many types of internal organization could result in such fibers with bulges. On the basis of careful measurements of these strands and a consideration of packing requirements, Ruben- stein made the suggestion that there exists a tertiary structure based on the formation of cables from the single Watson-Crick duplex strands (Fig. 2).

3. The Electron Microscopy of Phage Maturation

An important contribution to this problem was made by Kellenberger et al., 1959; Kellenberger, 1962). By sectioning phage-infected bacteria at intervals during intracellular growth, they find that by the 10th minute each cell contains about 5 to 10 dense particles having the characteristic shape of phage particles but somewhat smaller in size. While these dense particles do not form in the presence of chloramphenicol, they are found in abundance shortly after chloramphenicol is removed. On lysis of the cell, these condensed particles are not found as intact phage, nor are there enough "empty phage heads" to account for their number. One cannot be completely sure that these condensed units are truly "mem- braneless" for there is some evidence that the membranes may be broken

118

CHARLES A. THOMAS, JR.

•,j£ ..

./

Fig. 3. Organized or condensed phage DXA. Tlie den.se area.s within the cell are thought to be phage DNA which has been folded and organized prior to the assem- bly of the head membrane around it. (From Kellenl)erger et al. 1959.)

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 119

down during lysis (Koch and Hershey, 1959j. Kellenberger concludes that these "membraneless dense particles" represent DNA which has been organized into a compact form as the result of the action of a still unknown "organization principle" (Fig. 3). To what extent this ability to form a folded configuration is a property of the DXA mole- cule, and to what extent it relies on the presence of other molecules such as the polyamines (Ames et al., 1958) or the internal proteins (Levine et al., 1958) is a matter of conjecture, as is the role of the head mem- brane, if present. It seems certain that the internal protein does not bind to free DNA in the presence of 0.13/ NaCl, nor is it transferred to progeny phage (^linagawa. 1961 1. It is of some importance to note that proflavine, an acridine dye which is known to alter the structure of free DNA (Lerman, 1961). also interferes with the intracellular organ- ization of the compact form (Kellenberger et al., 1959).

4. The Question of Single-Chain Regions

The thorough studies of Sinsheimer (1959a,b) have led to the con- clusion that the small bacteriophage. 0X174, contains a single molecule of DNA which is not in the form of the Watson-Crick duplex. This phage is about 10 times more sensitive to the decay of incorporated P^- atoms, 10 times more sensitive to X-rays, and 25 times more sensitive to UV than T2 phage (after correcting for the 77-fold difference in DNA content) (Tessman et al., 1957; Tessman, 1959). This increased radio- sensitivity is in accord with the hypothesis (Stent and Fuerst. 19601 that the lethal event in phage is the complete scission of the DNA molecule; <^X, having only a single chain, would be expected to be broken by every P^- disintegration. If increased radiosensitivity. and in particular sensitivity to P^- decay, reflects single-chain construction of the viral DNA, then these kinds of experiments could yield useful information on the DNA in various phages. The only other phage known to have a high sensitivity to incorporated P^- decay (also X-rays and UV) is phage a studied by Graziosi and his collaborators (Celano et al., 1960). However, the DNA excised from this phage has all of the char- acteristic properties of a duplex molecule such as that taken from T2 or T4 (Cordes et al, 1961; Aurisicchio et al., 1960). In this connection it is important to note that the simple interpretation of P^- damage as a "direct" effect needs revision in view of the fact that the efficiency of killing can be decreased substantially by the addition of a free radical trapping agent (Matheson and Thomas. 1960).

A promising approach to the problem of detecting single-chained DNA emerges from the studies of Mayor and Hill (1961). These authors

120 CHARLES A. THOMAS, JR.

(leinonstrate that ^X piiage particles <;tain bright red with acriflinc orange, wliile T2 phage stain yellow-green (Anderson et al., 1959).

A vciy sensitive method for the detection of single-ehain regions has been perfected l)y Levine, Murakami, Van Vunakis, and Grossman (1960), based on the complement fixation that attends the reaction of an antibody to DNA which is specific to denatured DNA.

In conclusion, then, one can say that the molecules liberaterl from T2 and T4 have very little or no single-chain regions while the majority of ^X DNA is single-chained. Molecules of intermediate type have not yet been identified. At jiresent it is not known whether some hydrogen bonds are broken during the folding of the DNA into its compact form.

5. Speculations on DNA Organization

If the DN^A is uniquely folded into a compact structure prior to the assembly of the head membrane about it, as is suggested by the work of Kellenberger, it is reasonable to suspect that at least pavi of the ability to form specific folds is due to some structural features along the DNA molecule. The presence of such regions along the DNA mole- cule which are concerned only with the determination of the tertiary folding have been suggested by ]\Iahler and Fraser (1961). From the chemical point of view, there is the unexplained observations of Dunn and Smith (1959) who find about 1 mole of 6-methylaminopurine for eveiy 200 moles of adenine in T2 phage. This corresponds to about 500-600 of such strange bases per phage particle. Since such residues could not form the normal hydrogen bonds with thymine, these bases, if present, might cause an alteration in the nonnal duplex structure of the molecule. It is interesting to note that the average number of internal protein molecules per phage is about 300 calculated on the basis of the amount of this protein, about 5% (Levine et al., 1958; Minagawa, 1961), and a molecular weight of 15,000 which is based on methionine content (Rubenstcin, 1960). In order to compact the DNA into the small volume of the phage head about 300 to 600 turns or folds must be made. The coincidence in the magnitude of these numbers leads one to suspect that they result from some special requirements of the organization of the DNA in the phage particle.

In experiments on the reassembly of protein subunits around TMV- RNA, it seems clear that the RNA plays an important role in the reassembly process. First, because the reconstitution process can occur at higher pH values than wnll permit the reassembly of the protein units alone, and second, because the length of the TMV rod assumes a unique value — presumably corresponding to the length of the coiled RNA molecule (Fraenkcl-Conrat, 1962; Fraenkel-Conrat and Singer, 1959;

in. BACTERIOPHAGE DNA AND BACTERIAL DNA 121

Fraenkel-Conrat and Williams, 1955). This is as might be expected in view of the fact that the polyribonucleotide chain is completely sur- rounded by the protein subunits in the completed TMV particle (Klug and Caspar, 1960).

This experience with TMV leads one to inquire whether the organ- ization of the phage head is strongly dependent on the enclosed nucleic acid. The globular form and the large amount of DNA to be enclosed suggests that not all of the DNA can be in contact with the protein sub- units of the head membrane. The existence of "density mutants" in phage lambda indicate that this phage can exist in a number of different foiTns differing in DNA content (Kellenberger et al., 1961a,b). A vari- able DNA content seems to be displayed by the different strains of A-dg transducing phage (Weigle, Meselson, and Paigen, 1959). Likewise the existence of deletion mutations in T4, which seem to be missing a segment of the linkage map, suggests a variability in DNA content (Nomura and Benzer, 1961). One would expect that this difference in DNA content would result in a significant perturbation in the packing arrangement of the DNA. The fact that viable phage particles can be constructed in spite of this alteration in DNA structure suggests that there is some degree of structural independence between the compact nucleic acid and the protein membrane.

In conclusion, it seems clear that the DNA inside the head of bac- teriophage has a special organization, but the requirements for this organization are not overly strict because phage particles can be con- structed with slightly variable amounts of DNA. The studies of Kellen- berger would indicate that the folding and "condensing" of the DNA precedes the assembly of the head membrane around the packet of nucleic acid. As to the details of the folding and how folding allows for rapid injection, we remain ignorant.

B. THE STRUCTURE OF THE LIBERATED PHAGE DNA MOLECULE

1. The DNA jrom <f>Xl74

The first successful isolation and characterization of the DNA mole- cule liberated from any bacteriophage was accomplished by Sinsheimer (T959a,b). Measurements of the DNx\ content by colorometric tests indicated the virus contained 25.5% DNA. Light-scattering molecular weight determinations yielded a value of 6.2 X lO*' daltons for the intact virus. This requires the DNA content to be 1.6 X lO'' daltons. The DNA was extracted and purified by phenol extraction and the molecular weight repeatedly measured by light scattering under various conditions, giving values between 1.6 and 1.8 X 10'\ By assuming the absence of

122 CHARLE.s A. THOMAS, JR.

any significant systematic error in the light-scattering molecuhir weight detennination. Sinsheinier concluded that there was a single molecule of DNA in each \-irus jjarticlc. A vai'iety of chemical and phy.-ical experiments indicates that this DNA molecule does not have the chai- actoristic secondary structure of DNA's from other phages. A large fraction of the amino groups is free for reaction with foinialdehydc, and the extinction coellicient does not show a characteristic sharp in- crease with temperature that is characteristic of DNA's having the duplex structure. It would appear that the molecule consists of a single polynucleotide chain. This molecule is resistant to the single-chain specific exonuclease from Escherichia coll (Lehman, 1960) anrl there- fore has no free 3' end — a finding which suggests a circular form (Fiers and Sinsheimer. 1962).

2. The Antonulioijraphii of T2 and T4 DNA

Turning now to the T phage, the first information that one would like is the number and size of the DNA molecules making up the phage genome. The molecular weight of the DNA in the even phages is so large that the standard methods of molecular weight determination cannot be trusted; however, autoradiography has proved useful. When P^--labeled DNA molecules are mixed with a sensitive photographic emulsion, it is possible to detect single molecules of DNA because they give rise to a cluster of tracks in the develoi)cd emulsion. Each track corresponds to the path of a single /?-ray emitted by a decaying P"*- atom. The first experiments of this kind on T2 DNA resulted in two populations of stars, the first arising from DNA molecules comprising 36-40% of the phosphorus of the intact phage particles, and the re- mainder in molecules one-fifth this size or smaller (Levinthal, 1956; Levinthal and Thomas, 1957; Thomas, 1959). The validity of this distribution became questionable as a result of the work of Davison, who demonstrated the extreme sensitivity of phage DNA molecules to fragmentation by hydrodynamic shear (Davison, 1959). This work sug- gested the possibility that larger molecules might have existed, and that they were brok(>n during the dilution and mixing with the melted emulsion.

Further doubt was cast on the autoradiographic results by the work of Hershey and Burgi (1960). By employing a chromatographic colunm which was able to fractionate large DNA molecules primarily with respect to molecular weight (Mandell and Hershey, I960), these authors were able to show that unbroken molecules rechromatographed uniformly overlapping the original material, even though the sample selected for rechromatography was taken from the edge of the elution profile. In

III. IU('TKKIOPHA(iE DNA AND BACTKRIAL UNA 123

short, they were unable to demonstrate any iieterogeneity in tlie original unbroken material. These results were in contrast to those obtained with shear-broken molecules, which could be demonstrated to show a certain variability in size. Thus, it seemed likely that whatever the molecular weight of the unbroken material was, it was uniform and not as indi- cated by the autoradiography up to that time.

This situation was finally resolved by the use of autoradiography and column fractionation by Rubenstein, Thomas, and Hershey (1961). The general scheme of these experiments is shown in Fig. 4.

Phage were labeled with P^-, then mixed w^ith a large excess of cold carrier phage. The mixture was extracted with phenol following a pro- cedure known to liberate unbroken molecules. The resulting DNA preparation was loaded onto the chromatographic column and the excess P^- from the growth medium was washed out with the loading solution. This avoids pelleting the labeled phage which probably results in some radiation damage to the DNA. The DNA is then eluted and, as seen in Fig. 4, the radioactive DNA and the carrier DNA elute in an over- lapping peak, which means that as far as chromatographic properties are concerned, the labeled and unlabeled molecules are the same. A single fraction is then selected for gentle mixing with the nuclear emulsion. The resulting population of star sizes shown in Fig. 4E is the same as the population of star sizes that arise from intact phage particles shown in Fig. 4B. Thus, there is a single molecule of DNA which comprises all of the phosphorus of the phage.

By knowing the specific activity of the P^- in the growth medium, it is possil)le to calculate that the minimum molecular weight of this mole- cule is 130 X 10'\ while the phosphorus content per plaque-forming unit sets an upper limit of 160 X lO*'. This conclusion is supported by the results of Davison et al. (1961), who showed that drojjlets containing phage particles or DNA molecules containing equal amounts of P''- had an equal number of star-forming units. This molecule is not broken down by heating to 70°C, proteases, or column treatment. The kinetics of breakage (Burgi and Hershey, 1961 ) indicate that there are not a few preferential weak points along the molecule which are more sus- ceptible to shear.

If the mixed DNA preparation is subjected to stirring before loading onto the column, both the lal)eled and unlabeled molecules are eluted at lower salt concentrations (Fig. 4D). When a sample of this material is examined in the nuclear emulsion, the star population is seen to have a lower mean value (Fig. 4F). By an examination of a number of dif- ferent fractions of shear-broken preparations, it was concluded that the molecule breaks successively into approximately halves and then into

124

CHARLES A. THOMAS, JR.

"• '"»«	I—,..,,
A ORKICK Pf	*Gt
""♦''
ON*
X ~	vx
0	1
...„..„..„.	""T

1^	B	J"	PHAGE 2nir\- 90
10 n(r)	f	^^	n	r • II 1 f =0 178 q =0955 ^ N"=65 3
5	tr		to
o	/			Ln-i

30 35

TUBE NUMBER

	15	E			AMGlE MOLtCOLtS
					s	Xrnn=90
					r = iC 4
			,,		f =0 170
n(	10 r)				^^	9=0 946 N'=64 7
	5	-	/		\
	n		vT ^^	\ ■

Fig. 4. The results of the autoradiographic experiment. A. The experimental plan. B, E, and F. The frequency distribution of star sizes; the number of stars having r rays plotted against the star size r. C and D. DXA chromatograms of intact molecules and broken molecules, respectively. The sohd line is optical den- sity; the broken hne, P\ The shaded fractions were .^elected for autoradiography. The star population arising from these samples is shown in E and F. (From Ruben- stein et al, 1961.)

approximately quarter-molecules when stirred at the critical rates of shear. This is just what one would expect for a linear molecule, while a circular structure would be expected to break repeatedly into fragments which would be substantially smaller than one-half. When a hoop-

III. BACTERIOPHAGE DNA AND BACTERIAL DXA 125

shaped molecule is broken once by shear, its maximum extension would now be twofold greater than in the closed form. This would mean that it would be susceptible to further breakage. The hydrodynamics of this breakage process has been examined by Levinthal and Davison flQei).

The structure of To DXA is probably equivalent to that of T2 since the critical stirring speeds that are required to break the T5 molecule are the same as those required to break a T2 DNA fragment of 84 X 10^ daltons (Hershey et al, 1962). In the case of T5, P^-, and T7 DXA, the sedimentation coefficients are the same as that found for T2 fragments of the same DXA content, if we assume a single molecule per phage particle. However, the sedimentation behavior of these large molecules is not understood, and it would be risky to infer molecular structure from sedimentation coefficients at this time.

The fact that the smallest and largest known bacteriophages have been shown to contain a single molecule of DX'A leads one to the obvious hypothesis that all bacteriophage contain a single molecule of DXW. and that this unitary structure is a fundamental feature of the function and construction of virus particles.

3. The Mass per Unit Length — the Xumber of Polynucleotide Chains in the Molecule

While the autoradiographic results indicate that there is a single linear molecule of DXA liberated from the phage particle, no conclu- sions can be drawn about the number of polynucleotide chains making up the molecule. Therefore, it would be appropriate to review the evi- dence that the T2 DX"A molecule is a linear duplex of the Watson- Crick type. The X-ray fiber photographs of DX^A extracted from T2 and T7 phage show the same characteristic pattern as do DXA's from a wide variety of other sources, except that T2 DX^\ is unable to form the A form as the relative humidity is reduced; a fact which may be related to the presence of glucose on the hydroxymethylcytosine residues (Ham- ilton et al., 1959). Thus it would seem safe to conclude that at least a certain fraction of these phage DX^A's have the structure derived for the B form of DXA (Langridge et al, 1957, 1960a,b). The equatorial spacing corresponding to the side-to-side packing of the molecules in the fiber increases in a smooth fashion from a minimum value of 20 A at 50% relative humidity to more than 30 A as the relative humidity approaches 100% (Xorth and Rich, 1961). This is in accord with expectation for the packing of helical duplex molecules of diameter 20 A or less and makes unlikely the suggestion that there is a stable associa- tion between two duplex molecules, unless one made veiy special requirements on the packing and on the manner in which hypothetical

12(1 CHAKLKS A. IHOMAS, .IK.

linkages hctwoon molecules changed with changing humidity. The jjack- ing re(iuii-ements of helical duplexes have been examined carefully hy Mai-vin ct <;/.. (19()ll. .\ii analy.><is of the l;itticc spaciiigs and (he iiia.-^.^^ per unit length of the rodlike i)article.s ha,s been made by Luzzati (t (il. (19G1). Unfortunately, the preparations were iii.ammalian and chicken DNA, not b.acteriophage DNA; ne\-cilh('lcss. their icsult> arc proliably jiertinent. 'i'hey conclude that the particles in solution haxc the mass per unit length that one would expect for a Watson-C "lick duplex, and that this is so under a \"aiiety of conditions of conceiitialion and ionic strength. Taking these results together implies that the mas-s per unit length for T2 DNA would be approximately 210 d dtons/A. For the T2 molecule this would correspond to a sti'ucture (V2 n. in contour length.

This conclusion gains strong sui)poi't fi-om the direct electron micro- scopic detei-mination of the thickness of the phage DNA molecule by Beer and Zobel (1961). These authors stained the DNA with uranyl acetate, deposited the molecules on a cai'bon grid witli polystyrene latex spheres, and then shadowed with platinum. The molecuU^s were seen to run through the shadow of the s])heres. In this region the uranyl-stained molecule can be seen and its thickness measured. Outsirle the shadow of the sphere the height of the molecule can be measured from the shadow length and known shadow angle. This gives two inde- pendent measurements on the same molecule. Both give about 20 ± 5 A for the diameter of the molecule. Since the entire molecule can be seen in overlapping pictures (Beer, 1961), one would like to know whether all of the molecule is in the form of a 20 A fiber. The vicissitudes of the substrate prevent a confident answ(M' to this question; the best one can say is that there are no obiuous si)ecial features along the limited number of molecules so far seen.

Another way to settle the (juestion of the mass per unit length of the DNA molecule is to measure its total length, and to comb'ne this number with the molecular weight determination.

A direct observation on the total length of T2 DNA molecules has been made by Beer (1961, 1962), who has perfected techniques which allow the molecule to be picked up and stretched out on the electron microscope grid. Using these procedures he has been able to measure 30 different molecules from samples of varying fi-actional length. When the observed length is divided by the fractional length of the chromato- graphic fraction, the calculated total lengths range from 30 to 90// with an average of 49//,. Similar studies on T3 (Bendet et al., 1962) and T7 (Davi.son and Freifelder, 1961; Beer, 1962) have i)een made.

Another approach to this problem was nuide by Cairns (19621, who

1 t I' ' ■ • 4	-<«
* «
% ' '
> : [
• • # •	• •
*

Fig. 5. Tritium autoradiograph of a T2 DNA moleculp. Thp markpr imlioates a length of 10 m. (See Cairns, 1962.)

127

128 CHARLES A. TIKiMAS, JR.

laiiolod T2 phajz;e witli tritiiiin and then mixed with a hirge excess of unhibeled carrier phage to prevent breakage l)y shear (hiring extraction. When this 1)X.\ is streaketl on a ghiss slide and then covered witii stripping fihn. the localized grains produced by the weak tritium decays are seen to lie in more or less straigiit lines, the lengths of which are approximately ")4 /i (Fig. 5).

Taken togi'ther. these studies indicate tliat the DXA molecule liberated from the jihages so far studied haw the length and mass that woulil be expected for a single helical duplex.

4. The Frohlcni of Wccik Points or Polynucleotide Chdin Interruptions

Ila\-ing considered the evidence that T2 and T4 ])hage DXA is a single linear duplex, one would now like to know if there are any special non-nucleotide residues which unite distinct polynucleotide chains. As mentioned above, stability tests and the kinetics of shear breakage seem to indicate no i)referential weak points along the molecule. Therefore, if non-nucleotide linkages exist, they must be at least as strong as the inter-nucleotide linkages themselves. But what about interruptions in the polynucleotide chain; could there be a few breaks located at special or random points in the duplex? Some information on this point has been obtained by sedimentation equilibrium studies done on the unbroken molecules, and the polynucleotide chains produced by the denaturation of the unbroken molecule (Thomas and Berns, 1961; Berns and Thomas. 1961). In these experiments the unbroken molecules were denatured by any of three different methods, in the i)resence of 1% CHoO. The formaldehyde reacts rapidly with the amino groups exposed by the denaturation, and effectively prevents the reassembly of the poly- nticleotide chains. This stabilization by reaction with CHoO is important because the separated polynucleotide chains have a strong tendency to reassociate in solution unless the amino groups are blocked. "When a mixture of undenatured whole molecules and denatured whole molecules is banded in CsCl density gradient (in the presence of 1% CHoO), one gets two well-separated, sharp bands. From the variance of these distri- butions one may calculate a molecular weight value. It turns out that the calculated molecular weight of the polynucleotide chains is just one-half the molecular weight of the undenatured material. This is true irrespec- tive of the mode of denaturation. These results are summarized in Fig. 6. This could only be the case if there were no interruptions in the poly- nucleotide chains. The method should be able to detect a single interrup- tion if it occurred in the middle third of the molecule. Recent findings indicate that the polynucleotide chains from either end of the molecule have slightly different densities — a fact which would increase the

111. BACTERIOPHAGE DNA AXD BACTERIAL DXA

129

8^

(x I0~^)

1.0

10

<

.05

02

.01

Fig. 6. A comparison of the band forms of T2 and T4 whole molecules with that of formaldehyde-reacted polynucleotide chains derived from them. The Gaus- sian bands produced by native and denatured whole DNA molecules are shown on the same graph by plotting the log of the band height vs the band width. The whole duplex molecules fall on the steepest curve, while the polynucleotide chains fall on a curve which has a twofold lower slope; a fact which indicates that their relative molecular weight is lower by a factor of 2. (From Berns and Thomas, 1961.)

130 CHARLKS A. THOMAS, JR.

ability to drtect nicks l)y |)r()(luciii<r a still liroaWcf l)aii(l (Thomas and Pinkcrton, 1902 1.

Two control cxpci'inicnts should l)i' niciitionciL If a tew interruptions in the polynucleotide chains are jjroduced l)y the limited reaction with DNase, then the shorter chains are easily observed by the disproi)or- tionately broad band produced by the single polynucleotide chains, whereas the undenatured duplex has about the same band wirlth as before. In order to ride out the i)ossibiIity that formaldehyde is forming cross-links between polynucleotide chains, N"-N^^ hybrid molecules were produced by reannealing. On subse({uent dcnaturation in the presence of CH^-O, the chains of different density were cleanly separable, which meant that the formaldehyde was not causing them to be linked together (Thomas and Berns, 19621.

When the continuity of the molecule is maintained by a single chain, one would expect that its tensile strength would be one-half that of the duplex. Thus, interrupted molecules might lie nioi-c fragile to shear than uniform dujilexes. The major conse(iuence of the decay of an incorporated P'*- is the destruction of the phosphodiester linkage and thus the interruption of the polynucleotide chain. Therefore, it is possible to produce artificially interrupted molecules in this way. This expecta- tion was confirmed l)y Levinthal and Davison (1961), who found that T2 molecules which had suffered a substantial decay of P^- were broken at lower rates of shear than was required to break undamaged molecules. Since interruptions apparently do increase the shear fragility of the molecule, this indicates that there were no interruptions in the orig- inal molecule in those regions which can be broken by shear. Since the maximum tension is developed in the middle, these experiments suggest that ther(> ai'e no natui'al interi'uptions near the middle ])art of the molecule.

The sedimentation \-elocity analysis of the chromatographic fi-actions of the first breakage product of whole molecules (Burgi and Hershey, 1961 1 shows that there is a single, ajiproximately Gaussian distribution of molecular sizes produced by stirring. This distribution is not altered by prolonged stirring at the critical s])eed, and would only be produced if the most probable point of breakage were in the center of the molecule, and if the probability of fracture decreased in an approximately normal manner on each side of the middle. This would indicate that the proba- bility of the breakage is governed by the hydrodynamic factors in the region of the stirring blade, rather than weak points located at a special location. The same kind of analysis was applied to the breakage of half-molecules by stirring at a higher (critical I s])eed. This produced a normal distribution of molecular lengths about otie-quarter the original

III. BACTKKI()l'HA(iE DNA AND BACTERIAL UNA 131

size. On the basis of these arguments, and others based on the critical rates of shear required to break molecules of different sizes, these authors come to the conclusion that there are no weak points located at either special or random locations along the molecule.

5. Conclusions

The studies on the DNA from T2 and T4 show that there is a single, stable, linear DNA molecule which comprises all of the nucleic acid in the phage particle. Insofar as is known, all bacteriophage may contain a single molecule of nucleic acid. The length of the molecule (in T2) is approximately what one would expect for a duplex molecule. When the polynucleotide chains are separated by denaturation in the presence of fomialdehyde, the high molecular weight of the polynucleotide chains indicates the absence of interruptions in the continuity of the chains. Thus, the bacteriophage DNA molecule could be a simple duplex over its entire length.

C. THE GENETIC STRUCTURE OF PHAGE

A central concept that has developed in microbial genetics is that the genetic structure, as revealed by the compiled data from recombina- tion experiments, is related in an understandable way to the physical structure of the DNA molecule, and that the jihysical structure is reflected by the processes of synthesis and recombination in the form of a genetic linkage map. The processes of "synthesis" and ''recombina- tion" are themselves not understood, and the clarification of these events are among the major objectives of molecular genetics. In this section some features of the genetic structure of phage will be considered in their relation to the physical structure of the phage DNA molecule.

1. The Circular Genetic Map in T2 and T4

On the basis of a sensitive linkage test, Streisinger anfl Bruce (1960) have come to the conclusion that all of the known genetic markers in T2 and T4 are linked to form a continuous genetic map. In these experiments three markers were employed, two of which were known (or could be demonstrated) to be linked, and a third the linkage of which was to l)e tested. By selecting for observation a c'ass of progeny phage which showed recombination between the two markers that were known to be linked, they could then examine these phage for the third character. If this third character were totally unlinked, one would expect to find it associated with the most proximal of the linked pair with a frequency depending solely on the relative multiplicities. The experiment can be done with greater sensitivity by comparing the results

132

CHARLES A. THOMAS, JR.

of two crosses, the first (Cross I) with the multiplicity of the parent bearing the markers a, b, and c about 10 times that of the minority parent a'b'^r*. A second cross (Cross II) was performed with exactly the ftainc, but opposite multiplicities. In Fig. 7 the relationships are depicted.

		CROSS I				CROSS H
	obc > + + +			abc < +++
	+	+ +	Majority - Parent Minority ' Porent	+	+ +
Parent
Majority Parent	a	b c	a		b	c

Fig. 7. The types of crosses which give a sensitive test for hnkage in phage. The recombinant type indicated should be equal in the two crosses if the c marker were unlinked from the linked a and b markers. The recombinant type shown in Cross I was actually substantially higher than that shown in Cross II. (From Strei- singer and Bruce, 1960.)

In this figure the recombinant type is indicated (for convenience only) by a copy-choice scheme. In these two crosses the progeny phage will contain a certain number of phage of the recombinant type indicated. If this number is divided by the total number of a 6^ recombinant types, irrespective of whether they contain c or c^, the resulting ratio should be identical in both crosses — if the c marker is truly unlinked. In all experiments that were perfoniied in both T2 and T4, the ratio obtained in Cross I was substantially higher than that obtained from Cross II — a fact which indicated that the c marker is linked to the other two. These conclusions are completely free of assumptions concerning either the kinetics of replication or of recombination. Using this technique with various markers thought to be unlinked, it was demonstrated that they were all linked to each other. This unitaiy genetic map has been con- finned using the many temperature-sensitive mutants which can be found distributed over the entire map. The closer spacing of these markers makes the sensitive test described above unnecessary (Edgar, 1961).

By a continuation of the type of crosses done by Streisingcr and Bruce, it was found by Streisinger, Edgar and Harrar (1961) that markers previously thought to be at the extremes of the genetic map were themselves linked to each other. Thus the genetic map in T2 and T4 must be considered to be formally circular. This is in contrast to the linear structure of the excised T2 or T4 DNA molecule. At present it is not yet known whether any other smaller virulent phages have a

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 133

circular map. In the exploration of the genetic structure of any organism one is restricted to that region of the map which contains recognizable markers. The available markers in lambda are in a region which is only about 20 units long, and there is no evidence of circularity.

The discovery of the circular map in T2 and T4 makes it quite certain that additional map segments will not be added as additional markers are discovered. Thus one can state with confidence that the map in T2 is known, unitary, and circular. The DNA molecule liberated from the phage, containing essentially all of the nucleotides in the virus, is unitaiy, but almost surely linear. Thus, at present, one would like to know the relationship between the circular map and the linear molecule. Perhaps each DNA molecule has genes in the same order but the beginning and end points are different. Such a circularly permuted molecule would make the genetic results easily understood. Alterna- tively, the molecule and map could be unique but two (or an even number of) cross-overs always occur during a recombinational event (Stahl, 1962). Some of these possibilities can be tested. At the present time, one is limited by the fact that no sensitive transformation assay for T2 segments has been developed.

2. Transformation by Lamhda-dg DNA

In the case of phage lambda the situation is more favorable as a result of the work of Kaiser and Hogness (1960) and Kaiser (1962) who have demonstrated a transformation system with lambda-dg, and lambda DNA. When a bacterial strain which is lysogenic for lambda is induced by ultraviolet light, the vegetative phase of growth is initiated and new phage particles are synthesized and subsequently released. Approximately one in 10" of these new particles is lambda-dg. These phage are different from ordinary lambda in a number of ways. This phage particle can transduce the galactose genes from the bacteria in which it was grown to the bacteria it infects. It is defective in the sense that it is missing a sizable fraction of its genetic map (Arber, 1958) and it has a veiy small chance of establishing the lysogenic condition compared wth normal lambda. Independently arising lambda-dg phage can be shown to differ in their density when banded in CsCl, indicating that they have a variable DNA content (Weigle, Meselson, and Paigen, 1959). It is currently thought that these lambda-dg phage are the result of a substitution of bacterial genes for some of the phage genes. This exchange appears to be confined to a special region of the lambda map, and appears to be unequal in the sense that the resulting lambda-dg may have somewhat more (or somewhat less) total DNA than ordinary lambda.

134 rHARLKS A. THOMAS, ,IH.

Kaisrr and llogncs.s purilicd the transducing pha^c hy making use of a property of hunbda-dg, namely, that it has a (HfTcrcnt density than ordinary hinil)da. The DNA was then extracted fioni these phage by the phenol technique, and the resulting DNA could be shown to produce Gal"" colonies when added to appropriately lysogenized Gal" cells in the presence of "helper" phage. While the details of this cxpci-i- ment are somewhat technical, the net result is a transfontiation of the recipient cell to Gal^ which is scored on EMB-galactose plates. A further examination of these cells has revealed that they are Gal"" because they have become lysogenic for lambda-dg. If this is true, these transformants should also contain other genes (outside the dg region) as well. This was tested by looking for the cotransfer of r and mi genes which are located at some distance from the end of the dg region. It was found that bacteria which had received the Gal"^ also had a high proba- bility of receiving both the c and mi genes as well. This suggested strongly that the molecule which was responsible for the transformation contained the entire lambda-dg chromosome.

3. Transformation with DNA frorn Ordinary Lambda

These studies on lambda DNA were continued by Kaiser (1962), who developed an assay for ordinaiy lambda DNA; in this case the assay was based on the production of infectious phage particles rather than the transformation of a recipient cell. Again the role of "helpei'" phage is important. A non-lysogenic strain of E. coli K12 was infected with helper lambda (bearing the i^-'* marker) at a multiplicity of 10. After a suitable period for adsorption, the phage-infected bacteria were washed and then mixed with the diluted DNA solution. This DNA was prepared by phenol extraction of a purified preparation of lambda phage bearing the appropriate genetic markers. After a period of ap])i()xiniately 40 minutes DNase was added to digest any extracelhilai- DXA and appropriate dilutions of this mixture were plated on indicator bacteria to characterize the progeny phage genetically. These indicator strains will not record jihage which are produced by the "helper" jihage, but they will record phage bearing one or more genetic markers furnished by the DNA. The map of the lambda phage from which the DN.\ was prepared is given in Fig. 8. It was found that when the DNA contrii)uted genes, it contributed all of the markers on the above map. There was the apparent simultaneous transformation of all of the markers, and in this regard it was similar to phage infection. Thus the infecting DNA molecule comprises the entire chromosome of the phage. When this DNA preparation is stirred, the molecules are broken near the middle, causing them to sediment more slowly in a sucrose density gradient. The ratio

III. BACTERIOPHAGE DNA AND BACTERIAL DNA

135

of the sedimentation constant estimated before and after stirring indicates that the molecular weight is lower by a factor of 2. The bio- logical assay of the stirred DNA now shows that most of the phage progeny contain the i^ marker, but very few now contain the h marker; there is no longer the simultaneous transfer of all markers. By replating the resulting phage, it is possible to examine all the markers shown in the map in Fig. 8. It turns out that both the s and mi markers are

"1 I I i I — I — I — I — TTi — I — I — r 0 /lo

SHEAR

20

Fig. 8. The map of lambda. Tlie re.sult of hydrodynamic shear is to ehminate the simultaneous transformation of both h and i genes. The simplest interpretation of these experiments is depicted by the arrow labeled "shear."

transferred with essentially the same frequency as is the i gene. Thus it appears that stirring, a treatment which breaks the molecule near its mid-point, actually causes a separation of the map between the h and s markers. This would be expected if the map sequence and the nucleotide sequence were equivalent. The fact that both the map and the molecule are broken near their midpoints suggests that the available markers are distributed over most of the length of the molecule.

8.0 9.0 10.0 1 1.0 12.0 13.0

DISTANCE FROM AXIS OF ROTATION (cm)

Fig. 9. The separation of DNA fragments bearing the i^ gene (O) from unbroken molecules bearing the pair hi^ (%) by sucrose gradient centrifugation. (From Kaiser, 1962.)

136 CHARLES A. THOMAS, JR.

By employing tlie s^ucrosc gradient sedimentation technique, it was actually possible to physically separate the unbroken molecules bearing the h and i^ genes from the broken inolccules whicii can transfoi-in the i^ gene only. This is shown in Fig. 9.

If extensions of this kind of exiu'rinient can be made to the otiier genes in lambda, and to other genetic systems, the results may prove what everyone has hoped: that the map sequence and the nucleotide sequence is the same, and that genetic structure is a simple manifesta- tion of physical structure.

4. The Structure of Heterozygous Phage Particles

When a bacterial cell is infected with two bacteriophage T2, one of which is a mutant and the other wild type, approximately 2% of the resulting progeny will contain both alleles. These heterozygote particles were discovered and analyzed by Hershey and Chase (1951). In general these heterozygotes (HETS) will give rise to both types of phage on further growth, and this is the way in which they are detected. So far as is known it is possible to find particles which are heterozygous for any marker tested, although the r (rapid lysing) marker is very con- venient because HETS will produce plaques containing both r and r^ phage. These plaques have a characteristic mottled appearance. Pairs of markers which are closely linked are often heterozygous for both markers simultaneously, while an unlinked marker shows only a slightly higher than usual (2%) chance of also being heterozygous. Further work showed that phage which were heterozygous for a central marker were in general recombinant for markers on either side of the heterozygous region (Levinthal, 1954; Trautner, 1958), a fact which indicated that these structures could play an important part in the formation of recombinants. These findings were extended by Edgar (1957), who was able to demonstrate that there was a higher probability of recombination between closely linked markers when cells were infected with heterozygotes (for these markers) than was the case for inixedly infected bacteria. Perhaps the most important of Edgar's finding was that the heterozygote particle gave rise to a distribution of recom- binants which suggested that the heterozygote persisted or multipled within the infected bacteria in such a way as to produce recombinant particles at any time during the latent period. Thus heterozygotes were pictured as a more permanent feature of the phage DNA, and not as structures which were obliged to segregate on replication. These findings led to the investigation of the fine structure of the heterozygous region by Doermann and Boehner (1962), who developed techniques to test for six different closely linked r (rll) nuitations and a distant tu marker.

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 137

Phage bearing 6 markers were crossed with wild type, and the mottled plaques were resuspended and replated. The r-type plaques were then tested to learn exactly which of the 6 r's are represented on the plate, for these r's must have been originally found in the heterozygous phage. They found that these heterozygous particles must have contained all 6 r+ markers (otherwise they would not have been detected as mottled plaques), and in general 1, 2, 3, 4, 5, or 6 r's usually in contiguous groups. For example, if a, b, and c were present and d, e, and /, were absent it was concluded that the heterozygous segment must have terminated in the region between c and d. By further examination of these heterozygotes which had a single termination in the region covered by the 6 r markers, Doermann and Boehner demonstrated that r markers near the termination of a string of r's were not present in the mottled plaques as frequently as markers near the extremes of the r region. This would only be expected if the existence of a termination somehow caused the markers near the termination to be copied less frequently (in preference to the wild type alleles). From these experiments Doermann and Boehner conclude that the physical structure of these heterozygotes is probably not a simple duplex DNA molecule, but rather a more complicated structure which actually has more DNA than a non- heterozygous particle. Their proposal is shown in Fig. 10.

iiiiiiiii I iii^iiiiiiiiiiii \

t - )iiiiiiiiiiiiiiiiiiiiiiiiiiim

'"" ' ^< i ,

< iiiiiiMMiiqin imi ' ^^-^^^

Fig. 10. Proposed structures for heterozygous DNA molecules. (From Doermann and Boehner, 1962.)

This model receives some support from the findings of Doermann (1961). When progeny phage containing recognizable heterozygotes are banded in CsCl and fractions taken, the heterozygous particles are 15 times as numerous in the denser edge of the band, indicating that they could indeed contain more DNA than the average phage particle, or that there is some other special reason for these particles to be denser.

The termination of the genetic strand in the heterozygous region and the finding that HETS have an increased density in CsCl suggests that this discontinuity should be recognizable by physicochemical ex-

Li2JS-

jyUM: .SL

nasi

TOR m

Bs:

.1^

t

--2r

jmir. ^T5E2tgS- fl

138 CHARLES A. THOMAS, JR.

periment!;!. So fai-. tlic handing studies of Bcrns and Thomas (1901) indicate that tlie polynucleotide chains are continuous and nonintei'- ruptt'd. Howevei'. tlieii' conclusions refer to the nidjority of the molecules. Moreover, their techni(|Ue would not detect intei-rujjtions near the ends of the molecule. Finally, an o\'erlap structure of the kind shown in Fig. 10 could result from the closure of a linear molecule on itself to form a circular structure, a hypothesis which would allow only a single heterozygous region of this fonn per molecule. Therefore, at this point it is not clear whether there is a conflict between the genetic and physical experiments or not.

5. Conclusions

At the present time one may picture the phage DNA molecule as comprising all of the DNA in the phage particle. After release the T2 molecule is found to be linear and over 50 /x in length. The genes in lambda appear to be arranged on a linear molecule and can be separated by breaking the molecule by shear; one expects (but as yet no evidence) that the genes in T2 are in map sequence along the T2 molecule. There are two aspects of the genetic structure of T2 which as yet have no physical interpretation: (1) the formal circularity of the T2 linkage map, and {^) the existence of a partly stable heterozygous region which displays genetic discontinuities. It is important to note that the com- bined length and mass determinations of the T2 DNA molecule only allow one to conclude that the majority of the molecule is duplex. Direct visualization in the electron microscope supports this conclusion. It is still possible that as much as 10% of the molecule may have a more special or elaborate structure which may perhaps correspond to the physical heterozygous region, although it must be said with equal weight that there is at the moment no physical evidence in favor of these regions.

I). THK DISTRIBUTION OF PARF:NTAL PHACE DNA TO PROGENY PHAGE

With the finding that the DNA in mitotic chromosomes is in the form of two subunits which, upon replication, separate and become associated with newly synthesized subunits (Taylor et al., 1957; Mesel- son and Stahl. 19o8a,b), attention turned quite naturally to the question of the physical and chemical structure of these subunits. In view of the small amount of DNA in bacteriophage, it seemed likely that phage chromosomes were of a relatively simple structure. This has to a large extent proved so, and has been the subject of previous sections of this chapter. In this section the manner in which the DNA from parental phage segregates and is distributed into i)rogeny i)hage molecules will be

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 139

considered. This distribution of parental DNA should be related to the structure of the phage molecule and its mode of replication.

1. Transfer of </>.r DNA

This virus shows no or very little transfer of parental DXA to progeny phage (Sinsheimer, 1961).

2. Transfer of T7 DNA

When eoli are infected with T7 phage labeled with N^'^ and C", the parental DNA that is labeled with these heavy isotopes is transferred to progeny phage particles. When the progeny phage are examined by density gradient centrifugation, there appears a band corresponding to phage with 24% of their atoms substituted by heavy isotopes (Mesel- son, 1960). Since T7 phage are about 50% DNA (Putnam, 1953; Davison, 1962) this means that some phage particles are formed which have half newly synthesized DNA and half parental DNA. The total DNA content of this phage is approximately 20 X 10" daltons (Davison and Freifelder, 1962) and there is evidence to support the notion that it is in the form of a single molecule (Thomas and Pinkerton, 1962). // the biological formation of hybrid T7 phage could be accomplished it could now be established that these phage DNA molecules which have hybrid density are simple duplexes of polynucleotide chains, and if it could be shown that the heavy and light subunits cannot be separated by breaking the molecule by shear, but that separation of the density label does occur by denaturation and separation of the polynucleotide chains, then it would be permissible to conclude that the transferred subunits are single polynucleotide chains, and that they are associated with a newly synthesized polynucleotide chain to form a hybrid duplex.

3. Transfer of T2 or T4 DNA

In contrast to the results of T7 the parental DNA of a T2 or a T4 phage is broken into smaller pieces and incorporated into the DNA molecules of many progeny phage. The transfer is dispersive. The early work on this subject has been reviewed by Sinsheimer (1960). These experiments by Levinthal (1956), Hershey and Burgi (1956), and Stent, Sato, and Jerne (1959) jiresented evidence that a sizable fraction (40 to 60%) of the parental DNA was transferred to a relatively few progeny phage particles in units which were as large as 15 to 20% of the parental phage. The remaining parental DNA appeared to be even more highly dispersed to a large number of phages in pieces no larger than 0.3%. This transfer of apparently very large pieces of DNA suggested that at least part of the phage chromosome was replicating semi-conservatively.

140 ( HAHI.KS A. THOMAS, JR.

This view was supported by the findings of both Stent et al., and I.evintJKd that these rehitively hirp;e pieces were not further dispersed on transfer to a second <j;('n('ratioii of jji^o^cny phage.

With the a(lv(>nt of density hd)eHng techni(iues, two groups })egan to reinvestigate these pi-oblenis: Kozinski, using o-bromodeoxyuridine (BU), and Roller and Meselson employing N^^ and C^-^ (Kozinski. 1961 ; Roller, 1961). In the case of Kozinski's ex])eriinents, normal P'*--labeled phage were allowed to singly infect "heavy" bacteria growing in a medium containing 5-bromodeoxyuridine. Progeny phage were collected and DNA extracted by the phenol procc^lure. This DNA was banded in CsCl and the P'*- (which came from the parental DNA) was now found to be in a band whicii overlapped the band for totally new or "heavy" DNA. This experiment can be done in reverse as a result of perfecting methods to grow highly infective, uniformly BU-labeled phage (Kozin- ski, 1962). In Roller's experiment "heavy" phage are allowed to infect "light" bacteria and the progeny DNA molecules are examined by centrifugation in CsCl. Again the conclusion is the same, namely, that all of the parental label is now found in molecules having a density which is essentially the same as that of the newly synthesized "light" DNA. Although it is difficult to estimate the upper limit to the size of the transferred pieces, it seems clear that only a small fraction of the parental DNA now resides in progeny molecules where it makes up as much as 10% of the total molecule. Following breakage by shear (Roller, 1961) or sonication (Kozinski, 1961) the parental label now assumes a position in the band corresponding to molecules made up of half-hea\y and half-light DNA. When the majority of the DNA has been sheared to a molecular weight of about 10 X 10® (as estimated by sedimentation coefficient), the majority of the parental DNA is found at the half-heavy position. One would like to conclude directly that these fragments are hybrid duplexes, in view of the fact that the majority of the phage molecule is duplex. However, the possibility still exists that the parental subunits now preferentially reside in some special regions of the molecule, the structure of which is not known, and the existence of which cannot completely be ruled out at this time.

Unfortunately, there are no published experiments in which the density labeling techniques have been employed to examine a second generation of phage. There is at present very little information available from these experiments on the distribution of the sizes of the transferred pieces, and even less on the perpetuation of this distribution during another cycle of infection.

There is a conflict between the earlier results and those derived from the density labeling experiments just mentioned. It could be one of

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 141

degree, however, for all experiments indicate that the transfer is dis- persive, and that the difference is in the degree of dispersion. It is not unlikely that the degree of dispersion depends on physiology of the infected bacteria. At this point there has been only limited effort to reconcile these apparent differences. The transfer studies using P''- autoradiographs have recently been reinvestigated by Kahn (1961), who finds in general smaller and more heterogeneous parental pieces in the transferred phage than was found by Levinthal.

4. Transjer of Lambda DNA

At the present time there have been no published experiments which have attempted to trace the distribution of parental DNA into progeny lambda DNA molecules, although density gradient studies have been done on whole phage. Meselson and Weigle (1961) have shown that the progeny lambda phage collected from a single cycle of infection by an N^^-C^^-labeled parent produce some infectious phage which have a density corresponding to phage particles made of totally new protein and totally heavy, parental, DNA ("conserved" phages), and another peak having the density expected for phage having half parental and half newly made DNA enclosed in newly made protein ("semi- conserved" phages). Phages with recombinant genotype were shown to contain parental DNA because they were found at certain positions between those expected for fully conserved and totally light or newly synthesized phage. Experiments of a similar nature have been performed by Kellenberger, Zichichi, and Weigle (1961a) using phage which have natural density difference as a result of a difference in DNA content (Kellenberger, Zichichi, and Weigle, 1961b). Again phages of recom- binant type w^ould be shown to contain preferred quantities of parental DNA. These experiments have been discussed more fully in Chapter II. It is important to note here, however, that the transfer in lambda is quantitatively different from that found in T2 or T4 phage. The parental DNA is not dispersed into pieces which are small (compared to the total DNA content of the lambda phage) and dispersed into many progeny phage.

II. Bacterial DNA

At the present time we have much less precise knowledge of the organization of bacterial DNA's than we do of bacteriophage DNA's. Again what information is available has come from the interpretation of cytological studies, growth and transfer experiments, the study of purified DNA, and genetic experiments.

142 CHAHLKS A. THOMAS, .IK.

A. rvToi.ocv or nACTERiA

The cxccllcut i-c\-ir\v l)y KcllciilKiLici' ( 19()()l has coNcicd this subject in (li'lail, and only sonic niajoi- coiichisions will he considcicd here. He concludes that in bacteria the I)X.\ is ()r<4anize(| into well-defined areas calieil iHicleaf x-acuoles which are x'aiiahlc in shape ami do not appear to have a nicinhi'anc. These vacuoles aic lilled with a fibrillar material about 20-60 A in di.ainetci-; the dense chromosomes which are sometimes seen are thought to be a coagulation artifact. As yet, electron micro- graphs are not able to reveal anything about the number of DNA mole- cules in the fibiils, or anything about the linkages between DNA molecules, although some recent work in this area appears promising (Kleinschmidt, Oehatia, and Zahn, 1960).

1^. OROWTH AND TRAN.SFER EXPERIMENTS

Following the experiments of Meselson and Stahl (1958a,b) which demonstrated that the rei)lication of E. roli DNA took place by the formation of a structural unit which was one-half old and one-half newly synthesized material, subsequent work has been aimed at (/) dis- cerning the largest structure which will replicate by this semi-conserva- tive fashion, and (3) identifying precisely the structure of the DNA molecule which contains equal amounts of old and new material. The first of these topics has been investigated by Forro and Wertheimer (1960) by tritium autoradiography. His experiments have been consid- ered in Chapter II. For the purposes of this discussion the salient point is that by using young, small cells of E. coli 15T", the DNA seems to be synthesized as though the bacteria contained only tii'o DNA-containing structures which can survive many rounds of re|)lication. These findings have been verified by Van Tubergen and Setlow (19611 using tritium autoradiography. Exponinitially growing cultures of coli which probably have two nuclear vacuoles were found to have foi<r structures which were conserved during further replication. This would be just as expected if each nuclear vacuole contained a single long DNA molecule, or a molecule punctuated by a special non-DNA unit ("linkers") at various points (Kellcnl)erger, 1960; Frecse, 1958). Thus, while the details of the structure of such a unit are still unknown, one may consider the "nuclear vacuole" as seen in electron micrograi)hs to consist of two DNA-con- taining units which may (but rigorous proof is still lacking) be the two polynucleotide chains of a very long DNA molecule.

Further evidence on this point has come from the study of an E. coli which simultaneously recpiires thymine, arginine, and uracil (Maal0e and Hanawalt, 1961 ; llanawalt ct a/., 1961). The .starting point for tlie.se

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 143

investigations is the thyniinelcss death phenomena discovered by Cohen and Earner (1954). When thymine is abruptly removed from a mutant strain of coli (15T ) wliich has a requirement for thymine, the cells gradually lose their viability in an exponential fashion for over four decades. In the case of the triple mutant T-A-U' thymineless death pro- ceeds essentially as before when the cells are incubated in a medium lacking T, A, and U except that there is a residual fraction (3% to 10%) of the cells which is immune to thymineless death. This fraction cen be increased to 100% of the cells by a 90-minute preincubation in a medium containing T, but lacking A and U. These authors make the plausible interpretation that this preincubation period allows the cells which have begun DNA synthesis to complete the templating of a genome. Before another round of templating can begin, there are some essential steps which require arginine and uracil, which were missing from this medium. This suggests that thymineless death is the result of attempted DNA synthesis which can take place in the absence of thymine only at a very slow rate and with a high incidence of errors, some of which are lethal.

If it were true that only a completion of a round of replication were permissible in the presence of thymine, and in the absence of arginine and uracil, then the resulting molecules after being broken during ex- traction should have hybrid density if 5-bromouracil were added in place of thymine. This hypothesis would predict that no fully heavy molecules would be found even on prolonged incubation with 5-bromo- uracil. On banding the extracted DNA, two distinct, symmetrical bands were found which corresponded to the density of normal DNA and to DNA of hybrid density. Thus, segments of the expected form were found. This conclusion is weakened somewhat by the control experi- ments which indicated that it was difficult to form uniformly BU-Iabeled DNA; a fact which is probably related to an inhibitoiy effect of bromouracil (Maal0e, 1962).

Taking these results, together with those of Forro, strongly suggests that the bacterial chromosome is a single structure of fixed DNA con- tent, and that the two independently labeled units that emerge from the studies of Forro are united in equal proportions in such a way as to produce hybrid material in the density gradient. Since the hybrid DNA can be broken extensively by sonication and yet retain its hybrid density, one is forced to admit that traverse breaks do not separate the subunits (Rolfe, 1962). This is exactly what one would expect if the two inde- pendent labeling units were the complementary polynucleotide chains making up the duplex DNA molecule. On the other hand, one would expect the same thing from a pair of duplex molecules united at frequent

144 CHARLES A. THOMAS, JR.

intervals, one of which was a pair of old polynucleotide chains, and the other made of two new chains. Such a structure has been proposed (Cavalicri and Rosenberg, 1961a,b). This structure, consisting of a pair of duplexes, would be thought to arise from anotlicr unsi)ecified kind of replication scheme. The essential feature of this proposal is that it requires the biologically produced density hybrid to be of four i)oly- nucleotide chains rather than two. These four chains are pictured as being arranged in two conventional duplexes which are united l)y fre- quent inter-duplex links. After one round of replication this structure is pictured as one old and one new duplex, whereas the conventional interpretation argues that a single duj)lex now consists of one old and one new polynucleotide chain. To accept this hypothesis would demand many adjustments in the interi)retation of a wide range of experiments, and a full discussion would be ([uite lengthy, and largely hypothetical.

If the biologically produced hybrid molecules are of this doublc- du]ilex structure, then it seems clear that the bonds which unite the dui)lex strands must have almost the same thermal stability as the hydrogen bonds that unite the polynucleotide chains, for the experiments of Schildkraut, Marmur, and Doty (1961) show that the biologically produced hybrid is converted uniformly with time to heavy and light subunits as the native structure of the molecule (as evidenced by density) is destroyed. Since a 20-minute heating at 90.2°C fails to result in the separation of subunits, while an equivalent heating at 93.8°C does, we must conclude that the hypothetical inter-duplex linkages must be as resistant to heat treatment as the hydrogen bonds, but at the same time no more resistant.

Another special condition must be placed upon these hypothetical linkages between duplex molecules, namely, that of species specificity. DNA samples from different species cannot be renatured together to form molecules of hybrid density, whereas those from the same or a closely related species can (Schildkraut et al., 1961). If the double duplex structure did exist before denaturation, it must have recovered during the renaturation process because transforming DNA has a similar absorbance-temperature profile, and the same rate of thermal inactiva- tion of biological markers as the original DNA, facts which strongly indicate that the basic structure of the molecules is the same before heat denaturation and after renaturation (Marmur and Doty, 1961).

Thus, it is still possible to think of the biologically produced hybrid in terms of a doubled-duplex molecule, provided that one is willing to allow the inter-duplex linkages to have veiy special properties which are almost the same as the hydrogen bonds which unite the polynucleotide chains. On the other hand, all of the above experiments are easily under-

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 145

stood in tenns of a biologically produced hybrid molecule made of a single duplex consisting of one new and one old polynucleotide chain.

C. THE STRUCTURE OF DNA LIBERATED FROM BACTERIA

The great sensitivity of bacteriophage DNA to mechanical breakage by shear means that one should proceed with caution in the interpreta- tion of any molecular weight determined by whatever method as repre- senting a structural or functional subunit of the bacterial genome. The inadvertent fragmentation by stirring or shaking already has proved to be a stumbling block in the extraction of phage DNA ; it will be a more acute problem with the extraction of bacterial DNA's which in prin- ciple at least could be 50 to 100 times longer than T2 phage DNA molecule. Among the recent investigators who have been studying the extraction of high molecular weight DNA from bacteria one has obtained preparations with sedimentation coefficients of 22 to 29 S corresponding to molecular weights greater than 8 million (Marmur, 1961), while others find 1.2 to 2.4 million, which is thought to be the molecular weight of DNA subunits in vivo (Cavalieri et al., 1961).

An interesting kind of DNase has been reported by Bernardi et al. (1960, 1961), Bernardi and Sadron, 1961) which reduces the apparent light-scattering molecular weight to about 500,000 and no further. If this proves to be the enzymatic cleavage of unique regions of the DNA molecule, it will prove a useful tool for the examination of DNA organization.

D. THE GENETIC STRUCTURE OF BACTERIA

The study of bacterial conjugation has led to two major ideas: (1) that there does exist a physical structure called the bacterial chromosome which can be transferred from a donor to a recipient cell and which contains all of the genes that have so far been recognized, an {£) that there can exist certain strains that contain genetic elements (which contain DNA), which are not physically united with the chromosome; these independent elements, which can exist in a number of different forms, are called episomes. Jacob and Wollman (1957, 1958a, b, 1961) have suggested that the bacterial chromosome is originally a circular structure which is broken by the introduction of the fertility factor (F^^) to produce the various Hfr strains which are capable of trans- ferring genes (by conjugation) with high frequency to the recipient cell. By examining the order and time sequence of the entiy of a number of genes, it was found that each Hfr strain had the same gene order, but a diiferent starting point. Such a situation might arise if these Hfr strains were formed by the fracture of the circular chromosome at

146

CHARLKS A. THOMAS, JR.

various points. This picture of the circuhir ehroinosonic was greatly stroiij^theiu'd hy the (hscovery of certain Hfr strains which had a ?'rr?/ high frequency of transfer (Vhf) of the donor chromosome (Tayloi' ami Ad(>lherg. 1900). These strains are able to transfer tiie entire male genome to the female cells in approximately 111 minutes. Using three different \'hf strains, these authors were able to demonstrate more conclusively the contimious linkage of all known chromosomal markers and identify more precisely the separation (in minutes) between markers distributed aix)un(l the map. In Fig. II their maj) determined l)y kinetic experiments is shown.

ser/gly

ade

his

Fig. 11. Tho kinetic linkage map of E. coli. The map represents the circular linkage group of the F"^ ancestor of three Vhf strains. Total injection requires about 111 minutes; the various markers enter sequentially and at intervals shown (in minutes). (From Taylor and Adelbcrg, 1960.)

The genetics of episomes is a subject of some beauty and detail. The reader is referred to the excellent review by Campbell n962). An im- portant kind of episome is the F+ unit which may be united with any of a number of different segments. This episome may be transferred from a male to a female bacteria by a process called sexduction which is analogous to the transfer of the entire chromosome just mentioned (Jacob et al., 1960). This transfer can be accomplished to cells other than coli and in this wav Scrratia cells can be infected with a coli

III. BACTERIOPHAGE DNA AND BACTERIAL DNA

147

episome (Carey et al., 1960). Wlien Serratia cells hearing the E. coli F-lac episome are cultured and the DNA is examined in the CsCl density gradient, a small amount of DNA which has a density corresponding to E. coli DNA now appears as a small peak close to the peak which is the majority of the Serratia DNA. This small amount of DNA whicli has the density and presumably the composition of E. coli DNA is absent in uninfected Serratia DNA (Fig. 12A) (Marmur et a/., 1961 I.

A

S marcescens

EPISOME INFECTED S. marcescens

E coh F

■* DENSITY

Fig. 12. Baad profiles of unfractionated DNA from parental and recombinant strains. Microdensitometer tracings of ultraviolet absorption photographs of samples at equilibrium in a CsCl density gradient at 44,770 rpm. The ordinate represents DNA concentration as a function of the distance from the axis of rotation. The band of buoyant density 1.742 gm/cm^ (N'Mabeled Pseudomonas aeruginosa DNA) is used as a reference standard. (From Marmur et al., 1961.)

148 CllAKLKS A. THOMAS, .IR.

E. CONCLUSIONS

At the pro!>ent time we can say very little al)Out the j)hysieal struc- ture of the DNA in bacteria. Genetically the j^enonie ajipear.s circular, hut this inference is based on the finding of many different Hfr strains which themselves luive a linear genetic structure. The genetic evidence, which has led to the idea of indei)endent genetic elements called epi- soines, seems to be confirmed by tiie first density gradient equilibrium ex]K'riments.

Acknowledgments

The author has j^rofited by many holijful conversations with members of the Biophysics Department, Johns Hopkin.s University. He is particularly appreciative of the help and criticism of Drs. Irwin Rubenstein and Michael Beer in the prepara- tion of this manuscript.

References

Ames, B. X., Dubin, D. T., and Rosenthal, S. M. (1958). Science 127, 814. Anderson, E. S., Armstrong, J. A., and Niven, J. S. F. (1959). 9lh Symposium Soc.

Gen. Microbiol, p. 224. Arber, W. (1958). Arch. sci. (Geneva) 11, 259. .\urisicchio, S., Chiozzotto, A., Coppo, A., Donini, P., Frontali, C, and Graziosi, F.

(1960). Nuovo cimento 18, Suppl. 2, 197. Beer, M. (1961). /. Mol. Biol. 3, 263. Beer, M. (1962). Personal communication. Beer, M., and Zobel, R. (1961). J. Mol. Biol. 3, 717-726. Bendet, I., Goldstein, D. A., and Lauffer, M. A. (1960). Nature 187, 781. Bendet, I., Schachter, E., and Lauffer, M. A. (1962). J. Mol. Biol. 5, 76. Bernal, J. D.. and Fankuchen, I. J. (1941). J. Gen. Phyxiol. 25, 111. Bernardi, G., Champagne, M., and Sadron, C. (1960). Nature 188, 228. Bernardi, G., Champagne, M., and Sadron, C. (1961). Biochim. et Biophys. Acta

49, 1. Bernardi, G., and Sadron, C. (1961). Nature 191, 809. Berns, K. I., and Thomas, C. A., Jr. (1961). J. Mol. Biol. 3, 289. Burgi, E., and Hershey, A. D. (1961). J. Mol. Biol. 3, 458. Cairns, J. (1962). J. Mol. Biol. 3, 756. Campbell, A. (1962). Adranccs in Genet. 11, 101.

Carey, W., Spilman, W., and Baron, L. S. (1960). Bacteriol. Proc. p. 73. Cavaheri, L. F., and Ro.senberg, B. H. (1961a). Biophys. J. 1, 317. Cavalieri, L. F., and Rosenberg, B. H. (1961b). Biophy.-^. J. 1, 323. Cavalieri, L. F., and Rosenberg, B. H. (1961c). Biophys. J. 1, 337. Cavalieri, L. F., Deutsch, F., and Rosenberg, B. H. (1961). Biophys. J. 1, 301. Celano, A., Aurisicchio, S., Coppo, A., Donini, P., and Graziosi, F. (1960). Nuovo

cimento 18, Suppl. 2, 190. Cohen, S., and Barner, H. D. (1954). Proc. Natl. Acad. Sci. U.S. 40, 885. Cordes, S., Epstein, H. T., and Marmur, J. (1961). Nature 191, 1097. Davi.son, P. F. (1959). Proc. Natl. Acad. Sci U.S. 45, 1560. Davison, P. F. (1962). Personal communication.

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 149

Davison, P. F., and Freifoldrr, D. (1961). Personal communication.

Davison, P. F., and Freifelder, D. (1962). Unpublished experiments.

Davison, P. F., Freifelder, D.. Hede, R., and Levinthal, C. (1961). Proc. Natl. Acad.

Sci. U£. 47, 1123. Doermann, A. H. (1961). Personal connuunication. Doermann, A. H., and Boehner, L. (1962). In press. Dunn, D. B., and Smith, J. D. (1959). 4th Intern. Congr. Biochem., Vienna, 1958,

p. 72. Edgar, R. S. (1957). Ph.D. Thesis, University of Rochester, Rochester, New York. Edgar, R. S. (1961). Personal communication. Falkow, S., Marmur, J., Carey, W., Spilman, W., and Baron, L. S. (1961). Genetics

46, 703. Fiers, W., and Sinsheimer, R. L. (1962). 6th Arm. Meeting Biophys. Soc, St. Louis.

Abstr. Forro, F., Jr., and Wertheimer, S. A. (1960). Biochim. et Biophys. Acta 40, 9. Fraenkel-Conrat, H. (1962). Personal communication.

Fraenkel-Conrat, H., and Singer, B. (1959). Biochim. et Biophys. Acta 33, 359. Fraenkel-Conrat. H.. and Williams, R. C. (1955). Proc. Natl. Acad. Sci. US. 41,

690. Frankhn, R. E. (1955). Biochim. et Biophys. Acta 18, 313. Eraser, D., and Williams, R. C. (1953). Proc. Natl. Acad. Sci. U.S. 39, 750. Freese, E. (1958). Cold Spring Harbor Symposia Quant. Biol. 23, 13. Gellert, M. (1961). Intern. Biophys. Congr., Stockholm, p. 288. Abstr. Gellert, M. (1962). Personal communication.

Graziosi, F. (1961). Intern. Biophys. Congr., Stockholm, p. 292. Abstr. Hamilton, L. D., Barclay, R. K., Wilkins, M. H. F., Brown, G. L., Wilson, H. R.,

Marvin, D. A., Ephrussi-Taylor, H., and Simmons, N. S. (1959). J. Biophys.

Biochem. Cytol. 5, 397. Hanawalt, P. C, Maal0e, O., Cummings, D. J., and Schaechter, M. (1961). J. Mol.

Biol. 3, 156. Hershey, A. D., and Burgi, E. (1956). Cold Spring Harbor Symposia Quant. Biol.

21, 91. Hershey, A. D., and Burgi, E. (1960). J. Mol. Biol. 2, 143. Hershey, A. D., and Chase, M. (1951). Cold Spring Harbor Symposia Quant. Biol.

16, 471. Hershey, A. D., Burgi, E., and Ingraham. L. (1962). In press. Jacob, F., Schaeffer, P., and Wollman, E. L. (1960). 10th Symposium Soc. Gen.

Microbiol, p. 67. Jacob, F., and Wollman, E. (1957). Compt. rend. acad. sci. 245, 1840. Jacob, F., and Wollman, E. (1958a). Symposium Soc. Exptl. Biol. 12, 75. Jacob, F., and Wollman, E. (1958b). Compt. rend. acad. sci. 247, 154. Jacob, F., and Wollman. E. (1961). "Sexuality in Bacteria." Academic Press, New

York . Kahn, P. (1961). Ph.D. Thesis, Yale University, New Haven, Connecticut. Kaiser, A. D. (1962). J. Mol. Biol. 4, 275.

Kaiser, A. D., and Hogne.ss, D. S. (1960). /. Mol. Biol. 2, 392. Kellenberger, E. 1960. lOlh Symposium Soc. Gen. Microbiol, pp. 39-66. Kellcnbcrger, E. (1962). Advances in Virus Research 8, 2. Kellenberger, E., Sechaud, J., and Ryter, A. (1959). Virology 8, 478. Kellenberger, G., Zichichi, M. L., and Weigle, J. (1961a). Proc. Natl. Acad. Sci.

U.S. 47, 869.

150 CHARLES A. THOMAS, .IR.

KellculxTKcr, CI., Zicliichi. M. L.. aiul W.ikIc. J. (I'JGlb). J. Mai. Btul. 3, 3<.«).

Kleinschmidt. A., Gchatia. M.. and Zalin, R. K. (1960). Kolloid-Z. 169, 156.

Klug, A., and Casper, D. I.. I). (11)60). Ailnnices in Virus Research 7, 225.

Koch. G., and Hor.shoy, A. D. (IDSO). ./. Mol. Biol. 1, 260.

Kozin.ski, A. W. (lOGl). ViroUHju 13, 124.

Kozinski, .\. \V. (1!K)2). I'cisuiial t-onununicalion.

I.an^M-idiie, K., Seeds, W. E., Wilson, H. R., Hooper. C. W., Wilkins. M. H. F., an<l

Hamilton, L. D. (1957). J. Biochcm. Biopliys. Ci/h,!. 3, 767. Lanjiridfie. R., Seeds, W. E., Wilson, H. R., Hooix-r, C. W., \\'ilkins, M. H. V., and

Hamilton, L. D. (1960a). J. Mol. Biol. 2, 19. Langridge, R., Seeds, W. E., Wilson, H. R., Hooper, C. W., Wilkins, M. H. 1-"., and

Hamilton, L. D. (1960b). J. Mol. Biol. 2, 38. Lehman, I. R. (1960). J. Biol. Chem. 235, 1479. Lerman, L. J. (1961). J. Mol. Biol. 3, 18.

Levine, L., Barlow, J. L., and Van Vunakis, H. (1958). Viroloi/i/ 6, 702. Levine, L., Murakami, W. T., Van Vunakis, H., and (irossmaii. L. (1960). Proc.

Natl. Acad. Sci. U.S. 46, 1038. Levinthal, C. (1954). Genetics 39, 169. Levinthal, C. (1956). Proc. Natl. Acad. Sci. U.S. 42, 394. Levinthal, C, and Davison, P. (1961). J. Mol. Biol. 3, 674.

Levinthal, C, and Thomas, C. A., Jr. (1957). Biochim. et Biophys. Acta 23, 453. Luzzati, v., Nicolaieff, A., and Mas.son, F. (1961), J. Mol. Biol. 3, 185. Maal0e, O. (1962). Personal communication. Maal0e, O., and Hanawalt, P. C. (1961). J. Mol. Biol. 3, 144. Maestre, M. F., and Kilksom, R. (1962). Nature, in press. Mahler, H. R., and Fraser, D. (1961). Nature 189, 948. Mandell, J. D., and Hershey, A. D. (1960). Anal. Biochetn. 1, 66. Marmur, J. (1961). J. Mol. Biol. 3, 208. Marmur, J., and Doty, P. (1961). J. Mol. Biol. 3, 585. Marmur, J., Rownd, R., Falkow, S., Baron. L. S., Schildkraut, C, and Doty, P.

(1961). Proc. Natl. Acad. Sci. U.S. 47, 972. Matheson, A. T., and Thomas, C. A., Jr. (1960). Virology 11, 289. Marvin, D. A., Spencer, M., Wilkins, M. H. F., and Hamilton, L. D. (1961). J. Mol.

Biol. 3, 547. Mayor, H. D., and Hill, N. O. (1961). Virology 14, 264. Meselson, M. (1960). In "The Cell Nucleus" (J. S. Mitchell, ed.), j). 240. Academic

Press, New York. Meselson, M., and Stahl, F. W. (1958a). Proc. Natl. Acad. Sci. U.S. 44, 671. Meselson, M., and Stahl, F. W. (1958b). Cold S/iriug Harbor Syjnpo.'^ia Quant. Biol.

23, 1. Me.selson, M., and Weigle, J. (1961). Proc. Natl. Acad. Sci. l'.S. 47, 857. Minagawa, T. (1961). Virology 13, 515. Nomura, M., and Benzer, S. (1961). J. Mol. Biol. 3, 684. North, A. C. T., and Rich, A. (1961). Nature 191, 1242. Putnam, F. W. (1953). Advances in Protein Chem. 8, 175. Rolfe, R. (1962). J. Mol. Biol. 4, 22.

Roller, A. (1961). Ph.D. Thesis, California Institute of Technology, Pa.sadena. Ruben.stein, I. (1960). Ph.D. Thesis, University of California, Los Angeles. Rubenstein, L, Thomas, C. A., Jr., and Hershey, A. D. (1961). Proc. Natl. Acad.

Sci. U.S. 8, 1113.

III. BACTERIOPHAGE DNA AND BACTERIAL DNA 151

Schildkraut, C, Marmur, J., and Doty, P. (1961). ./. Mul. Biol. 3, 595.

Sinsheimer, R. L. (1959a). J. Mul. Biol. 1, 37.

Sinsheimer, R. L. (1959b). J. Mol. Biol. 1, 43.

Sinsheimer, R. L. (1960). In "The Nucleic Acids" (E. Chargaff and J. N. Davidson,

eds.), Vol. Ill, p. 187. Academic Press, New York. Sinsheimer, R. L. (1961). dth Ann. Meeting Biophys. Sac, St. Louis. Abstr. TA 9. Stahl, F. W. (1962). Personal communication.

Stent, G. S., and Fuerst, C. J. (1960). Advances in Biol, and Med. Phys. 7, 2. Stent, G. S., Sato, G. H., and Jerne, N. K. (1959). J. Mol. Biol. 1, 134. Streisinger, G., and Bruce, V. (1960). Genetics 45, 1289.

Streisinger, G., Edgar, R. S., and Harrar, G. (1961). Personal communication. Studier, F. W., and T'so, P. (1961). Personal communication. Taylor, A. L., and Adelberg, E. A. (1960). Genetics 45, 1233. Taylor, J. H., Woods, P. S., and Hughes, W. L. (1957). Pmc. Natl. Acad. Sci. U.S.

43, 122. Tessman, I. (1959). Virology 7, 263.

Tessman, I., Tessman, E. S., and Stent, G. S. (1957). Virology 3, 209. Thomas, C. A., Jr. (1959). J. Gen. Physiol. 42, 502. Thomas, C. A., Jr., and Berns, K. I. (1961). J. Mol. Biol. 3, 277. Thomas, C. A., Jr., and Berns, K. I. (1962). J. Mol. Biol., in press. Thomas, C. A., Jr., and Pinkerton, T. C. (1962). Gen. Mol. Biol. 5, 356. Trautner, T. A. (1958). Z. Verbungslehre 89, 264. Van Tubergen, R. P., and Setlow, R. B. (1961). Biophys. J. 1, 589. Weigle, J., Meselson, M., and Paigen, K. (1959). J. Mol. Biol. 1, 379. VVilkins, M. H. F., Gosling, R. G., and Seeds, W. E. (1951). Nature 167, 759.

Chapter IV

Cellular Control of DNA Biosynthesis

KARL G. LARK

Department of Microbiology,

St. Louis University School of Medicine,

St. Louis, Missouri

I. Introduction 153

A. The DNA Cycle 155

B. The Level of Biochemical Control 160

II. Role of DNA Precursors in Controlling DNA Synthesis .... 162

A. The Direct Measurement of Deoxynucleotides or Precursors

Thereof and Their Relation to the DNA Cycle 163

B. The Measurement of Enzj^me Activity Associated with the

Production of DNA Precursor 170

III. Role of Protein or RNA Synthesis in Controlling DNA Synthesis . . 173

IV. The Physical State of DNA as a Factor ControlHng Its Replication . . 180

A. Is It Necessary for DNA to Exist in a Special Form in Order for It to

Be Duphcated? 183

B. Is DNA Synthesis Compatible with Simultaneous RNA S\Tithesis? . 186

C. Is DNA Structurally Stabihzed within the Cell and Is This a

Necessary Part of DNA Replication? 188

D. Is the Replication of DNA Physically Ordered, with Respect to the

Macromolecule Itself, with a Beginning, a Middle, and an End? . 158

V. Other Factors Affecting DNA Biosynthesis and the DNA Cycle . . 190

A. The Localization of DNA Synthesis 190

B. Variation in the DNA Cycle 191

C. Effects of Temperature and Radiation on the DNA Cycle . . . 192

D. The Effect of the Nutrient Environment 194

E. The Dissociation of the Cell Division Cycle from DNA Synthesis . 197 VI. Summary, Conclusion, and Opinions 197

Acknowledgments 200

References 200

I. Introduction

The biosynthesis of DNA is a distinctive characteristic of cell pro- liferation rather than of the broad span of phenomena commonly- associated with actively functioning cells. Cells which are rapidly metabolizing may often exhibit little or no DNA synthesis, whereas division requires a previous replication of genetic material if competent daughter cells are to be produced. The mechanisms controlling the

153

154 KARL G. LARK

biosynthesis of DNA are of interest in the study of pi'ohh'nis of (Uffer- entiation as well as those studies concerned witli the maintenance of genetic continuity, since absence of DNA synthesis may prevent, and thus control, further cell division without impairing cell function.

Individual cells possess the ability to control DNA synthesis. During a normal cycle of cellular division the content of DNA, as measured chi'mically and genetically, is exactly doubled. Tlie existence of this situation has been inferred for some time from the knowledge that during multiplication the DNA content of individual cells of a given species or clone remains remarkably constant and that the chromosome number of the vast majority of individuals arising in a clone also does not in- crease or decrease. Instances in which a dividing cell divests itself of genetic material as a result of asynchrony between cell division and duplication of DNA are the exception rather than the rule; cells in which DNA synthesis and cell division have ceased usually resume DNA synthesis before division recommences.

The extent to which intracellular DNA synthesis is controlled has become apparent in recent years, tlirough use of cellular autoradiogra- phy, quantitative microspectropliotometry, and synchronous culture techniques. As a result of these studies it has become evident that most animal, plant, or protozoan cells, as well as some bacteria, have a defined period during their life cycle during which DNA is synthesized.

The living cell therefore exhibits two types of control over the bio- synthesis of DNA: one, quantitative; the other, temporal. The amount synthesized is always equal to that initially present and the time of synthesis is restricted to a defined interval during the division cycle.

Although the evidence for this control of DNA synthesis has been available for some time, few systematic attempts have been made to understand the mechanism of control which governs the "DNA cycle." This has been due, in part, to a natural desire for further evidence to substantiate the generality of such a cycle and, in part, to a lack of knowledge of the biochemical mechanisms underlying the synthesis of DNA. Until recently, microscopic observation was found to be of little use in studying the interphase nucleus, in which DNA is synthesized.

In recent years the enzymatic basis for DNA synthesis has become known, and work has concentrated upon the study of the in vivo replication mechanism of this genetic material. In addition, physico- chemical information on the structure of the molecule, spurred on by the development of new physical techniques and genetic concepts, has led to a detailed knowledge of its structure.

Finally, a more detailed knowledge of the mechanisms whereby DNA controls the svnthesis of cellular material has led to an ever-increasing

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 155

interest in the role of DNA in regulating the cell's physiological behavior.

A large amount of information is now available relating to the control of DNA biosynthesis, but little of it has been obtained with this end in view. In what follows, material from different sources, investi- gated originally in a variety of contexts, is discussed. This will lead to the inevitable criticism that results are not strictly comparable and that, therefore, general conclusions may not be drawn. However, what follows presents a definition of the problem, the current status of its investigation, and the direction in which it appears that further investi- gation can profitably be directed.

It is intended to represent one individual's analysis of the situation rather than a comprehensive review of the literature.

A. THE DNA CYCLE

Recent studies have led to the general conclusion that during the division cycle of higher cells there exists a period following mitosis during which DNA synthesis does not take place (Gi), a subsequent

DNA Synthesis

Division

Fig. 1. A representation of tlie DNA synthetic cycle in relation to the cell division cycle. Division, or the end of mitosis, marks the beginning of the cycle. A gap, Gi, occurs before the onset of DNA synthesis, S. DNA synthesis is followed by another period, Gi, prior to the onset of mitosis. (Adapted from Stanners and Till, 1960.)

period of DNA synthesis during which the DNA content of the inter- phase nucleus is doubled (S), a second period during which DNA is not synthesized iG-z), and, finally, a period of mitosis {M) (Howard and Pelc, 1953) (Fig. 1). Although the duration of these periods may vary from system to system, they are often remarkably constant from cell to cell within the same system. [In bacteria, it has been possible to demon- strate that DNA is synthesized throughout most of the cell cycle (Schaechter et al., 1959; Young and Fitz-James, 1959; jVIcFall and Stent, 1959; Abbo and Pardee, 1960). However, recent evidence has indicated that even in these systems a DNA duplication cycle may exist

156

KARL G. LARK

(Maalpc and Haiiawalt, 1961; Hanawalt et al., 1961; Goldstein et al., 1959). It seonis possible, therefore, that such systems may represent situations in which Gi and Go may be of extremely short duration relative to >S. This is made more probable by the finding that G'l or Go may be extended in such systems by simple chemical or physical treat- ments which also synchronize the cell population (Lark and ]\laal0e, 1956; Earner and Cohen, 1956; Maruyama, 1956; Scott and Chu, 1958; Maruyama and Lark, 1959).]

The existence of a cycle such as that in Fig. 1 has been demonstrated with autoradiography (Howard and Pelc, 1953; Lajtha et al., 1954; Hornscy and Howard, 1956; Taylor et al., 1957; Firket and Verly, 1958; Mendelsohn et al., 1960; Painter and Drew, 1959; Stanners and Till, 1960; Prescott, 1960; Edwards et al, 1960; Sisken and Kinosita, 1961) by labeling cells at a given time with a radioactive precursor of DNA (usually tritiated thymidine) and then measuring the period of time required for the first labeled mitotic figures to appear subsequently as well as the proportion of cells so labeled or the rate at which radioac- tivity, incorporated during interphase, appears in metaphase mitotic figures. These measurements yield the time elapsing between the end of

^250-

0 2 4 6 8 10 12 14 16 18 20 Hours after addition of tritiated thymidine

Fig. 2. The peak metajihase grain count values from autonuiiograplis of cells sampled at various times after addition of thymidine-H^ to L-strain mouse cells growing in tissue culture. The upper curve represents a film exposure of 15 days; the lower, 8. Peak counts for each sample were obtained using the peak grain count values from a frequency distribution of the grain count for each sample. (Number of metaphases as a function of their inilividual grain count.) G2 is the period before an increase in grain count is observed. Tiie grain count rises throughout S, after which it is maintained at a constant value. (From Stanners and Till, 1960.)

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 157

DNA synthesis and the beginning of mitosis as well as the duration of the period of DNA synthesis. (For a detailed review of the design and interpretation of such experiments see Taylor, 1960a.)

Figure 2 is an example of such an experiment in which exponential multiplying L cells (tissue culture strain originally derived from the mouse) were labeled with thymidine-H^ and the peak grain count at metaphase measured at intervals thereafter. As may be seen, no counts were recorded for 3^ hours after the onset of labeling ; this is the period (Go) between the cessation of DNA synthesis and metaphase. The num- ber of grains/metaphase nucleus then rises for 6-7 hours after which no increase in labeling is observed; this, then, is the period of DNA synthesis, or S, during which incorporation of thymidine can occur. The number of grains will be low for those nuclei which were leaving S when thymidine-H'^ was introduced. The grain number per nucleus will rise until a maximum value is reached which represents nuclei which have passed through their entire S period in the presence of the radio- active isotope. The duration of S is thus given by the length of time during which the grain count/nucleus rises. vSince the generation time of these cells was 20 hours and mitosis occupies a period of 1/3 hour, Gi may be estimated as between 9 and 11 hours. Note the sharply defined appearance of labeled mitoses and of the saturation of the grain count, indicating the similar behavior of different individuals of the random population.

The cycle of DNA synthesis may also be reconstructed from micro- spectrophotometric measurements of the DNA content of individual cells sampled at random (Swift, 1950; Walker and Yates, 1952; Alfert, 1955; Bloch and Godman, 1955; Vendrely and Vendrely, 1956; Richards et al, 1956; McDonald, 1958; Walker and Mitchison, 1958; Kimball and Borka, 1959; Smith et al, 1959; Painter et al, 1960; McLeish and Sunderland, 1961; Utsume, 1961). Such reconstructions have been sub- stantiated by determination of the DNA content of individual cells at known intervals following mitosis and by measurements of synchronous cultures.

Figure 3 represents an example of such an experiment using embry- onic mouse heart cells grown in tissue culture. Two observations are noteworthy: (a) the similarity between the DNA cycle obtained by this method and that illustrated in Fig. 2, and (6) the precise control of the quantity of DNA within the cells, the number of cells containing amounts of DNA less than 22 or greater than 44 units being negligible.

Finally, experiments on synchronous cell populations have measured the total amount of DNA in the population and assumed that this would reflect the course of synthesis by a single cell (Ogur et al, 1953; Hotch-

158

Fraction of interphase time

Fig. 3. The DNA content of cells from mouse heart, growing in tissue culture. DNA content was measured microspectrophotometrically on Feulgen-stained cells. The number of cells on the ordinate contain the amount of DNA shown on the abscissa. The lower curve was constructed from this distribution a.ssuming an exi)onential rate of cell growth and that a constant amount (previously determined) of the division cj^cle was devoted to interphase. (From Richards ct al., 1956.)

ki.ss, 1954; Young and Fitz-James, 1959; Iwanmra, 1955; Earner and Cohen, 1956; Lark and Maal0c, 1956; ]\Iaruyania, 1956; Scott and Chii, 1958; Maruyania and Lark, 1959; Foster and Stern, 1959; Newton and

IV. CELLULAR CONTROL OF DNA BIOSYNTHESLS

159

Wildy, 1959; Nygaard et at., 1960; Williamson and Scopes, 1960; Abbo and Pardee, 1960; Burns, 1961; Terasima and Tolmach, 1962). Figure 4 (open circles) shows the course of DNA synthesis as well as of mitosis in dividing microspores of the anther of Trilliiun (Erikson, 1948; Stern, 1960b; Hotta and Stern, 1961b). Chemical analyses were performed on samples of isolated microsj^ores. It should be noted that the amount of DNA synthesized is exactly equal to that initially present.

.E 0.5 -

U)	4) m	«	n	1
a	<n O			«
JZ	o -c	u	«	(/>
a	SL Cl	.c		n
o	a. o	u.	c
	O t".	o
F	o E	c o	>>	«
"to		T3 C	o LU	c

0) w

*" o

Fig. 4. DX.\ syntliesis in tlie microspores and deoxyrilaoside production in the antlier of Tnllium. Synchronous de\eloiiment of tlie microspores occurs and as a result DNA doul^les during a defined period of interphase. A double peak of deoxy- riboside production hy the anther is observed at this time. The possibility exists that the second peak may represent an overshoot in microspore DNA precursor produc- tion (see text, p. 163). (From Hotta and Stern, 1961b.)

All of these experiments serve to demonstrate the temporal control outlined in Fig. 1. A similar situation is sometimes seen within a single cell containing more than one type of nucleus or in a single, very large nucleus. Thus, in the ciliate Euplotes (Gall, 1959) synthesis of DNA in the large U-shaped macronucleus (which is a functional nucleus in

^y

Ay

'^'

>^/r

r'

t/A

./ /j/. /./ ',.

/ ,* ' , / I .t. I ...

//'

■'/■• ." , 7 ,'r • '' r ,'• W\i'Y

■-> v<vtn > 4 \5iK >.\A '.'vvv V'^'^ -^^\ ^s^^. ^».«^'=^>^^

,SK*

W

l^\

I

•t^l

t s\>

^

A

\

I

X

V*v ^ ^>\ V

jv a

V

iiVlVSlW . C vVi -J -\\

160 KARL G. LARK

contrast to the micronuclcus which maintains tiie genetic continuity of the cell) starts simultaneously at the ends of the two lobes and proceeds toward the center. Similarly, in paramecia, synthesis of DNA is initiated simultaneously in the macro- and micronuclei, although synthesis is completed in the latter within a much shorter time than the former (Woodard et al., 1961). This suggests that the initiation and rate of DNA synthesis in both the macro- and micronuclcus may be controlled by a similar mechanism. (However, see Section V,A below.)

The above examples, though only a few of the large number to be found in the literature, illustrate the type of control in which we are interested.

B. THE LEVEL OF BIOCHEMICAL CONTROL

From our present biochemical knowledge, what possibilities exist which could serve as a basis for a control mechanism?

Figure 5 is a diagrammatic representation of the sequence of bio- chemical steps in the synthesis of DNA. The experimental basis for this sequence has been discussed extensively in the first chapter of this book. Mechanisms for controlling the biosynthesis of DNA may, in theory, be encountered at several levels.

The requirement for deoxynucleotide triphosphates as the substrate for polymerization makes the production of these precursors an im- portant point of control in the synthesis of DNA (Bessman et al., 1958). (Control, at a level less specific than this, would simultaneously affect other synthetic processes such as RNA and protein synthesis. For this reason only more specific control mechanisms will be considered.) The synthesis of these molecules, in turn, may be controlled at:

(a) the conversion of ribonucleotides to dcoxyribonucleotides (Rei- chard et al., 1961) and the formation of thymidylic acid (Fried- kin, 1959), and/or

(6) the phosphorylation of deoxy ribonucleotides to the trij)hosphate form (Lehman et al., 1958).

Thus, the rates at which tlie triphosphates are formed are governed by several enzyme reactions subject to control through feedback inhibition or repression (Pardee, 1961). Alternatively, competitive factors may be operative; for example, a precursor substrate such as a ribonucleotide may have been depleted by the synthesis of compounds other than DNA.

The actual polymerization of DNA presents a variety of opportuni- ties for the operation of a control mechanism. This reaction requires the participation of the "old" or pre-existent DNA which acts as both a

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS

161

AMP-	dAMP dADP
GMP	' -»- dGMP -*- dGDP dATP'
	\
CMP.	dCMP dCDP dGTP

UMP

TMP

TDP

\

dCTP

TTP J

+

"PRIMER" DNA

conserved units?

DNA

semi-

conserved

un i t

DNA

semi -conserved

Fig. 5. Schematic representation of the biochemical sequence of events involved in the biosynthesis of DNA. The ribonucleoside monophosphates — adenylic acid (AMP), cyt3^dilic acid (CMP), guanyhc acid (GMP), or uridylic acid (UMP)— are converted to deoxyribotides dAMP, dGMP, dCMP, or TMP (thymidylic acid) which are, in turn, converted to the di- and triphosphate derivatives (dADP, dGDP, dCDP, TDP, and dATP, dGTP, dCTP, and dTTP). These are polymerized under the direction of "primer" DNA to form new DNA. The primer DNA is derived from pre-existing DNA and may represent the conserved unit of this semi- conserved molecule. Thus, if the pre-existent DNA carries a semi-conserved density label such as N", the product DNA will carry a density label distribution (as shown), demonstrating the semi-conservative nature of DNA replication.

primer and a template in the formation of new material (Romberg, 1959). In doing this, the old and new polynucleotide strands are not confused; there is one for one pairing of strands and these remain as separate entities which may be distinguished by a variety of physical techniques (Levinthal, 1956; Taylor et al., 1957; Meselson and Stahl, 1958; Mendelsohn, 1960; Sueoka, 1960; Forro, 1960; Simon, 1961). This "semi-conservative" method of replication appears to extend to aggre- gates of DNA molecules which may be as large as chromosomes (the evidence for which has been discussed in Chapter II).

162 KARL C. LARK

\iy ciitcriiiu; a |)riiiu'r slate DXA may initiate its own i-eplicatioii which may again be terminated when primer is no longer avaihible. During DNA replication the conversion to primer may l)e a complicated process which may involve structural changes in the molecule itself and in its association with other molecules such as protein and KNA.

At this level, then, control of DNA biosynthesis may involve the synthesis of RNA or protein or both.

Finally, it should l)e realized that restrictions on the proximity of the various components of the polymerizing reaction may play an important role. Thus, if the various enzymes involved in DNA synthesis (kinase and polymerase) are formed in the cytoi)lasm, they must enter the nucleus in order to utilize the DNA primer.

II. Role of DNA Precursors in Controlling DNA Synthesis

In order to synthesize DNA the cell nuist produce the necessary pre- cursor deoxyribotides. In vitro enzymatic studies have shown that these must be present in the triphosphate form and that all of the nucleotides found in the DNA which is being synthesized must be present simul- taneously. There is eveiy reason to believe that the results of these studies are valid in vivo as well as in vitro.

For example, bacteria, upon infection with bacteriophage, commence the rapid production of large quantities of viral DNA which continues until the cell lyses (Cohen, 1947; Hershey et al, 1953). It has been shown that a primaiy step in the conversion of the cell into such a "DNA factory" is the production of many new enzymes needed for the synthesis of new types of phage-specific deoxynucleotides (Flaks and Cohen, 1957; Kornberg et al., 1959), as well as the production of increased quantities of pre-existing enzymes also associated with DNA synthesis. All of these are necessary for producing the deoxynucleotide triphosphates required for DNA synthesis. As would be expected, such phage-infected cells contain large quantities of acid-soluble purine and pyrimidine deoxynucleotides, whereas uninfected cells contain small amounts of only the pyrimidine deoxynucleotides (O'Donnell et al., 1958). Moreover, to further enhance the efficiency of the system the enzymatic balance of the cell is shifted to emphasize the synthesis of DNA rather than RNA (Cohen, 1947). Although this system may be regarded as one in which DNA synthesis continues "out of control" it exemplifies a biological system in which the cell follows the predictions of the enzyme chemists, with respect to precursor production and DNA biosynthesis.

Granted that the production of DNA pi-ecursors is a necessary step

IV. CELLULAR CONTROL OF DNA BIORYNTHESLS 163

in DNA synthesis, to what extent does such precursor production control DNA biosynthesis during the division cycle of the normal cell?

To date, experiments to answer this question have utilized synchro- nized cell systems of various types. Two types of experimentation may 1)0 considered:

A. THE DIRECT MEASUREMENT OF DEOXYNUCLEOTIDES OR PRECURSORS THEREOF AND THEIR RELATION TO THE DNA CYCLE

The most striking example of a control mechanism which may rest on precursor production is that of DNA synthesis in the microspores of the anthers of Lilium longijorum. and Trillium erectum. The lily system has proved useful because of the fact that the length of the flower bud enclosing the anthers is correlated with the synchronized activities of the microspore (Erikson, 1948). In Trillium, microspore development is also synchronized with respect to time. By sampling anthers from a given bud at different times, or by selecting buds of the proper length, it is possible to examine large samples of material at known stages of the division cycle and to study the synthesis of DNA by the microspores (Taylor and McMaster, 1954; Foster and Stem, 1958). When this is done, it can be shown that the deoxyribosides of guanine, adenine, cytosine, and thymine appear in the fluid surrounding the microspores shortly before DNA synthesis and mitosis (Stem, 1960a). These disappear abruptly as DNA is made (Fig. 4). It has been shown that such material originates from the destruction of a polydeoxy- ribotide in extra sporageneous tissue (Stem, 1961) and that the deoxy- nucleosides thus formed are rapidly converted in the microspores or the microspore walls into triphosphates, the "half-life" of which is very short (Stern, 1960b; Hotta and Stern, 1961a). Recent evidence has demonstrated that the ability of microspores to phosphorylate thymidine increases immediately following the appearance of deoxyribosides in the extracellular fluid. It therefore ajipears that these phosphoiylating enzymes (necessary to produce deoxynucleotide triphosphates) may be induced by deoxyribosides (Hotta and Stern, 1961b). Although it has not been possible to demonstrate directly the conversion of the extra- cellular deoxyriboside material into DNA, the data at present are consistent with a mechanism by which the presence of deoxyribosides initiate the synthesis of DNA, presumably by making a pool of pre- cursor material available.

One difficulty, however, is that two cycles of deoxyriboside production and disappearance occur for each step of DNA synthesis, a pattern which is most striking in experiments with Lilium. This may arise from a slight shift in the phase of synchrony among anthers from a single

1()4 KAUI, (;. LARK

1)11(1 iSU'iii, 19(12 1 or may inilicatc a I'olc of dcoxyi'ibosidi's otlitT than as a pn'cui'sor I'oi' DXA (SttTii, 19(iOa). If the former is true, a similar difforciu'e shoiikl ai)pear in the syntliesis of DNA which should then occur as a two-step process. Tliis has not been observed. Moreover, the amount of DNA synthesized is less than the deoxyriboside pool or the amount of DNA broken down to create the i)Ool (tiiis loss of DNA luay indicate the total size of the deoxyriboside pool formed) (Stern, 1960a).

These difficulties appear to be explained by the recent work of Takats (1962) who has demonstrated that the deoxyriboside material which is derived from the breakdown of tapetal DNA is not used in the synthesis of microspore DNA. He labeled the tapetum with thymidine- H-' at a time when no DNA was being synthesized in the microspores and subsequently traced the breakdown and reutilization of this radioactive material by means of autoradiography. It was found that the degraded material was taken up in the microspore walls but not into microspore DNA. He suggests that this material could sen'e to induce the formation of a second deoxyriboside or -tide pool (this might explain the double peak of acid-soluble material prior to DNA synthesis) which is then used to supply the necessary precursors for microspore DNA synthesis. This idea is similar to that of Hotta and Stern (1961), who have sug- gested that this pool of material is responsible for the induction of nucleotide phosphoiylating enzymes in the anther.

It would appear, then, that control of DNA synthesis in the lily anther may be the result of a precursor pool induced by the products from the degradation of tapetal DNA. The induction of a second pool of deoxyriboside material is further suggested by the finding, by Stem and Hotta (1962), of two types of DNA in cells of the anther wall, one type having the base ratios and properties of the microspore DNA. Breakdown of this material could easily provide a pool of precursors for the synthesis of microspore DNA.

A situation similar to that found in the division of lily microspores may exist during the development of fertilized eggs and embryos of amphibia and Echinodermata. Previous evidence (Hof!"-J0rgensen and Zeuthen, 1952; Gregg and L0trup, 1955; Finamore and Volkin, 1958) had indicated a high extra nuclear (cytoplasmic) content of DNA-like substances in developing amphibian embiyos. It was presumed that this material ser\'es as a source of nuclear DNA. It has been shown (Grant, 1958; Kuriki and Okazaki, 1959) that a large portion of this "excess DNA" is composed of deoxynucleotides which are present in amounts far in excess of the nuclear DNA. The presence of all of the pyrimidine and purine deoxynucleotide triphosphates was noted. Thus this material could serve as either a reservoir of pi-ccursors for DNA synthesis or as an

IV. CELLULAR CONTROL OP DNA BIOSYNTHESIS 165

inducer for the synthesis of nuclear deoxyribotides. No correlative studies were made between the utilization of these compounds and nuclear DNA synthesis.

It has also been shown that sea urchin embryos contain large amounts of acid-soluble deoxyribosidic compounds (Sugino et al., 1960; Sugino, 1960) amounting to more than 5 or 6 times the intracellular DNA content. In these studies it was shown that the DNA content of the cell doubled during the cell division cycle. However, no studies of the time- course of DNA synthesis as related to the disappearance of acid-soluble deoxyribosidic compounds were made. The acid-soluble material appears to be derived from cytidine, which would indicate that it cannot alone serve as an adequate reservoir of DNA precursor material. Nevertheless, it is possible that further, more detailed studies on this system could reveal a relationship similar to that found in amphibian embryos. That this may be the case is indicated by the immunity of echinoderm embryo development to agents inhibiting purine or pyrimidine deoxynucleoside biosynthesis (Karnofsky and Bevelander, 1958; Karnofsky and Basch, 1960). This immunity persists until the blastula stage and indicates that until this time the developing embryo is not dependent on de novo purine or pyrimidine biosynthesis. It is possible that further studies in these systems of the time relationship between DNA synthesis and deoxyribo- tide utilization may substantiate a product-precursor relationship be- tween the two.

In these systems, then, the possibility exists that changes in the quantity of precursor material may control the synthesis of DNA.

Other situations in which deoxyribosides, or deoxyribotides, accumu- late quantitatively prior to DNA synthesis are not known. IMaterial from a number of different sources have been analyzed for their content of deoxyribosides or deoxyriboside derivatives (Schneider, 1955, 1957; Schneider and Brownell, 1957; Rotherham and Schneider, 1958, 1960; Potter and Schlesinger, 1955; Potter et al., 1957; Potter and Buettner- Janusch, 1958; LePage, 1957; Ord and Stocken, 1958; Parizek et al, 1958; Davidson and Bishop, 1957; Okazaki and Okazaki, 1958; Okazaki et al, 1959; Okazaki, 1959; O'Donnell et al, 1958; Okazaki et al, 1960; Lark, 1961). With one exception (Kanazir, 1954), the amount of material was only a small fraction (from 0.2-5%, depending upon the system) of the DNA content of the cell. Moreover, with the exception of two observations only pyrimidine deoxynucleosides or deoxynucleotides have been reported. In one of these exceptions where deoxyadenosine triphos- phate was observed in a rat carcinoma (LePage, 1957), this amounted to less than 10% of the quantity of pyrimidine deoxynucleosides or deoxy- nucleotides (Schneider, 1957).

ItiO KARL C. I.ARK

Although a hugi' (|uaiitity of precursors might not he fouutl in a popuh\tion of randomly Uividing cells, cells undergoing synchronous division and DXA synthesis might well be expected to show such an accumulation. This is apparently not so. In regenerating liver' an increase in deoxynucleoside and deoxynucleotide content is ohseived but the amounts of these are still extremely small and the quantity present at 24 hours (presumably the onset of DXA synthesis) are not strikingly greater (at most o09r) than the amounts found in hepatomas (Schneider and Brownell. 1957; Potter et al., 1957; Rotherham and Schneider, 1958 1 or regenerating liver after 48 hours (following 1)N.\ synthesis). Thus the increase observed appears consistent with the over- all increase in DXA synthetic activity and cell division in this tissue.

Finally, it has not been demonstrated that all of such material serves as precursoi-s for DXA synthesis (Hecht and Potter. 1956; Potter and Buettner-Janusch, 1958). Especially significant is the fact that no purine deox\'riboside material is observed.

On the other hand, in regenerating liver and in hepatomas the pro- portion of deoxynucleotides is increased relative to nucleosides and blockage of DXA synthesis following irradiation will result in accumula- tion of deoxj'pyrimidines in the tissue and excretion of such material in the urine (Ord and Stocken, 1958; Parizek et al., 1958; Davidson and Bishop, 1957; ^Main et al, 1957). It seems probable, therefore, that these substances are acting as precursors (see discussion of Fig. 7 below).

Bacteria have also been found to contain a small pool of dcoxy- pyrimidine compounds (Okazaki and Okazaki, 1958; Okazaki et a/.. 1959; Okazaki, 1959; O'Donnell et o/., 1958; Okazaki ct ai. 1960; Lark. 1961). To date, almost all of these have been shown to be thymidine derivatives and evidence is accumulating which indicates that they may act as intermediates in the transport or activation of certain sugars (Strominger and Scott. 1959; Okazaki. 1959; Okazaki, 1960; Baddiley and Blumson. 1960). With one exception, no deoxypurine or derivatives thereof have been observed. |In the case of E. coli B. Kanazir. (19541 has reported the existence of large amounts of deoxyguanosine within the acid-soluble pool. Other workers, using this and other bacterial strains, have not found any deoxypurines. The quantity of material reported was sufficiently great (0.2 ^.V mg dry weight of cells) that it would seem probable tliat this material would have been observed by others.]

'In liver, following removal of a portion of the organ, tissue regeneration occurs. When DXA synthesis is followed during this process no SA'nthesis is observed until 24-36 hours, at which time a striking synchronous increase in DXA content is observed. For this reason, regenerating liver i^resents an ideal biochemical system for studying certain jirohlems in tlie control of DX.\ l>iosynthesis (Barnum <7 nl.. 1057).

IV. CELLULAR CONTROL OF DNA BIOSYNTHESLS

167

As with regenerating liver, l)acteria do not exhibit a gross build up of deoxynuc'lcotides i)rior to synchi-onous DXA synthesis nor are purine deoxyriboside derivatives accumulated in even measurable quantities during any portion of the cell cycle (Lark, 1961). ^Moreover, those slight changes in the dcoxynucleotide pool which are observed have been shown to be unrelated to DNA synthesis since they may be retained as discontinuous changes under conditions leading to continuous DXA synthesis (see Fig. 6). In addition, if a step of DNA synthesis is blocked, no measurable excess of deoxynucleotides is accumulated.

Again, the sudden onset of DNA synthesis in Salmonella typhi- murium, which results from raising the temperature from 25° to 37°C,

Minutes

Fig. 6. Acid-soluble deoxyribotide (DOR) s^iithesis in relation to discontinuous DNA synthesis in sj^nchronized bacteria. Deoxj-ribotides are given as milligram equivalents. Deox5Tibotides are sj^nthesized discontinuously about (5 minutes) before the onset of DNA sj^nthesis (A). The quantitj^ of material synthesized is less than 3% of the intracellular content of DNA. If the discontinuity of DNA sj^nthesis is removed (B) by the addition of deo.xyribosides to the medium during a previous cj'cle of synthesis (see text, p. 196), the pattern of subsequent deo.xyribotide synthesis is unaltered. (From Lark, 1961.)

DNA (xlO"5)

12 3 4 5

2 3 4 5

Time in minutes

Fig. 7. The pattern of acid-soluble thymidine utilization by E. coli B. A trace of tritiated thymidine was added to the growth medium of the bacteria, and was rapidly utilized. Cells taken at the intervals indicated were analyzed for their intracellular acid-soluble thymidine. The distribution in each of several components is given. As may be seen, those components which have been shown to represent the major steady state pool constituents, TDP rhamnose, and TDP-X- (both thymidine (iiphospho sugars) do not appear to serve as immediate DNA precursors. The minor pool constituents, TMP, TDP, or TTP, appear to turn over (incorjiorate and lose radioactivity) rapidly as would be expected. (Unpublished data of Dr. R. Okazaki presented here by his kind permission.)

168

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 169

is not preceded by a rapid increase in intracellular deoxyribosidic compounds (Lark, 1961).

Okazaki has shown that the major components of the deoxynucleo- tide pool in E. coll do not normally serve as an immediate precursor for the bulk of the intracellular DNA. Thus, labeled thymidine incorporated into thymidine diphospho sugar compounds does not turn over as rapidly as thymidine mono-, di-, or triphosphates (Fig. 7). Studies on Lacto- bacillus acidophilus have shown that such thymidine compounds can eventually serve as a source of material for DNA synthesis; however, the rate of transfer of label is quite slow (Okazaki et al., 1959). Experi- ments of this type are extremely valuable in distinguishing whether a compound found in a cell is, in fact, serving as an immediate precursor for DNA synthesis.

Feinendigen et al. (1961) have obtained a similar result to that of Okazaki by studying the rate of incorporation of cytidine and of thymidine into DNA of animal cells. They used autoradiographic techniques and demonstrated that cytidine entered a pool from which it could be withdrawn over a long period for use in DNA synthesis. This pool could not be affected by exogenously supplied material. In contrast, thymidine entered DNA via a much smaller pool, susceptible to changes in the outside medium.

With the exception of the lily anther system and developing am- phibian or echinoderm embryos, the majority of data strongly support a picture in which deoxynucleoside or deoxynucleotide precursors are supplied for DNA synthesis upon demand, neither accumulating to any great extent before synthesis nor diminishing greatly during synthesis. This conclusion is supported by autoradiographic experiments which have demonstrated that the pool which can dilute exogenous thymidine or deoxycytidine during incorporation into DNA is extremely small (Painter et al, 1960; Taylor, 1960b).

Little is known about the control of DNA synthesis resulting from the conversion or lack of conversion of ribotides to deoxyribotides. However, in vitro studies have indicated that a singular type of feed- back inhibition mechanism may exist (Reichard, 1960; Reichard et al., 1961). Purine or pyrimidine deoxyriboside triphosphates have been found to inhibit the in vitro enzymatic conversion of nucleotide monophos- phates to deoxynucleotide phosphates. IMoreover, this inhibition is not specific for the particular type of purine involved, i.e., deoxyadenosine triphosphate will inhibit the conversion of cytidylic acid to deoxycytidine monophosphate. Recent studies have shown that this system may be operative in vivo as well as in vitro. Several workers have found that high concentrations of deoxynucleosides in the medium will inliibit the

170 KARL C. LARK

synthesis of DNA in vivo in l)oth l)act('iia and animal cells (Piusoff, 1959; Morris and Fischer. 19G0; Lark. 19b0; Longer and Klenow, 1960). Longer and Klenow (1960) have shown that whereas deoxyadenosine will inhibit the synthesis of DNA by ascites tumor cells, this inhibition can be overcome by the addition of deoxygiianosine. The inhibition is accompanied by the accumulation, intracellularly, of deoxyadenosine triphosphate (Munsch-Peterson, 1960). Other studies have indicated that this is responsible for the block in DNA synthesis (Overgaard- Hansen and Klenow, 1961; Klenow, 1961). On the other hand, these effects have been demonstrated to date only in the presence of extremely high concentrations of deoxy nucleosides in the medium, whereas purine deoxy riboside triphosphates have not been observed to accumulate in normally growing cells. For these reasons it is not clear whether this method of regulation is used to control the supply of precursors needed for DXA biosynthesis.

The possibility also exists that the rate-limiting ste]) in the synthesis of DNA precursors may be the availability of ribonucleotides from which deoxyribonucleotidcs are formed. A system of this type has been obtained from phage-infected cells by Cohen et al. (1961), who have studied the in vitro conversion of RNA to deoxyribonucleotides. They have noted the fact that RNA may play a key role in the control of DNA and protein biosynthesis (see Section III below) in that an RNA molecule may serve as substrate in the synthesis of the former or tem- plate in the synthesis of the latter but cannot be both template and substrate at the same time. They point out, however, that such a control mechanism can only operate in systems where ribonucleotides are limited in availability. This does not appear to be the case in bacteria. In synchronized bacteria, in which DNA is being synthesized in steps, large amounts of ribonucleotides are present at all times during the cell cycle, (Maruyama and Lark, 1962) and bacteria, which synthesize DNA and RNA at different rates, adjust tlic intracellular concentration of ribonucleotides to corresiiond to these rates (Franzen and BinkU^y. 1961 ; Smith and ]Maal0e, 1962 1. In all cases, the intracellular ribonucleotide content is ample to provide material for DNA synthesis, and DNA and protein synthesis i)roceed simultaneously.

B. THE MEASIREMENT OF ENZYME ACTIVITY ASSOCIATED M'lTH THE PRODUCTION OF DNA PRECURSOR

This indirect approach has been used by several workers interested in the enzymatic basis of the control of DNA synthesis.

In interpreting data of this typo, it must be remembered that the

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 171

nhility to rleinonstrate active enzyme in extracts of a tissue does not mean that the same enzyme is necessarily active in vivo. On the other hand, the absence of a necessaiy enzyme may indicate a negative type of control over DNA synthesis (pi'ovided the extraction procedure would preserve the enzyme activity).

The isolation and identification of the enzymes invoh'cd in the production of DNA precursors have been carried out using a wide variety of biological sources (see Chapter I). As was pointed out for the case of bacteriophage, it has been possible to demonstrate the de novo synthesis of various enzymes necessary for the production of specific types of DNA.

Few attempts have been made, however, to correlate pertinent altera- tions in intracellular enzyme patterns with the DNA cycle. We have already noted the induction of phosphorylating enzymes in the micro- spores of the lily anther (Hotta and Stern, 1961b). The time sequence was such that it was suggested that activity had been induced by the deoxyriboside substrates. A more detailed study has been made of a similar situation in regenerating liver.

The liver of adult rats contains all of the enzymes necessary to convert deoxy adenylic acid, deoxycytidylic acid or deoxyguanylic acid to the corresponding triphosphates (Hecht et ciL, 1954; Hecht and Potter, 1956; Canellakis et al, 1959; Gray et al, 1960). However, enzymes necessary to accomplish the phosphorylation of thymidine are lacking (Bollum and Potter, 1959; Canellakis and JNlontsavinos, 1958; Canellakis et a?., 1959; Gray et al, 1960; Hiatt and Bojarski. 1960). The latter appear to be present in newborn animals and are lost as the animals grow older (Hiatt and Bojarski. 1960). As a result, liver slices, as well as extracts of liver from normal adult animals, cannot synthesize DNA in vitro.

Following partial hepatectomy the necessary thymidylic kinases can be isolated from the regenerating liver (Bollum and Potter, 1959; Canel- lakis et al, 1959; Gray et a/., 1960; Hiatt and Bojorski, 1960; Bianchi et al. 1961) at the time at which DNA synthesis commences.

It is natural to assume that it is the production of these enzj^mes and the consequent production of the required thymidine triphosphates that leads to resumption of DNA synthesis. In examining the question of the control of DNA biosynthesis, however, several questions arise:

a. Does control lie in the production of a single enzyme or of a sequence of enzymes? Recent evidence indicates that the enzymes for thymidine production may not be produced until regeneration com- mences (Hiatt and Bojarski, 1960). Moreover, a careful examination of the time of appearance of thymidine, thymidine monophosphate, and

7'>

KARL G. LARK

lliyiuidiiic (lipliospliatr kiiiasi's lia.s slunvii that these enzynies a])i)eai- sequentially in that order during regeneration (Weissman et al., 1960). Similar results were obtained by the same workers studying enzyme production by L cells of mouse fibroblasts grown in tissue culture. In this case, the sequential appcai-ance of enzymes was observed upon transfer of cells to fresh medium. Finally, it has been shown that production of these enzymes by normal liver can be induced by injecting animals with thymidine (Hiatt and Bojarski, 19()0). Tims, it would appear that a sequence of enzymatic steps, or rather a metabolic pathway, is con- cerned and that this pathway may be normally repressed or inhibited. In this connection, it is interesting to note that nomial liver contains an extractable substance which will specifically inhilnt the tliyniidylic kinase system (Gray et al., 1960).

h. Is the establishment of this pathway the sole controllinsj; factor in DNA biosynthesis? If so, one would expect that the maximum rate of DNA synthesis would not be achieved until the entire enzymatic mechanism for the synthesis of thymidine triphosphate was optimally established. Instead, evidence exists which indicates that the maxinnmi enzyme activity (both thymidylic kinase and polymerase) is reached after the rate of DNA synthesis has already begun to decline (Bollum and Potter, 1959; Bianchi et nl, 1961). Similarly, when DNA synthesis ceases in regenerating tissue, precursors continue to accumulate (Wil- liamson and Gaschlbauer, 1961). Thus, if synthesis of thymidine tri- phosphate is responsible for the onset of DNA synthesis, the latter does not cease as a result of an absence of the former.

Canellakis et al. (1959) observed the presence of enzymes in normal liver, which degrade uridine or thymidine compounds. A reduction in their activity in regenerating liver was thought to contribute to the conti'ol of DNA synthesis. In the same system an increase in thymidylate kinase activity was observed. However, the observed changes in activi- ties of all of these enzymes were not cyclic, whereas DNA synthesis took place in three cycles occurring at 24, 48, and 72 hours. Clearly, if only the production of these enzymes controlled DNA biosynthesis tlicir activity should also have varied in a cyclic manner.

From these considerations it would appear that although the jiroduc- tion of enzymes needed for the synthesis of DNA precursors may be necessary, it is not a sufficient condition to explain the degree of control observed.

c. Are other control mechanisms also operating? Evidence from X-radiation studies has indicated that it is possible to block thymidylate kinase synthesis by irradiating regenerating liver at 6 hours whereas this is impossible at 16 hours. In the latter case, however, DNA synthesis

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 173

is blocked (Bolluni ct ciL, 1960). Thus, it appears that with respect to radiosensitivity, at least two possible control steps may be differentiated, one of which does not involve the enzymatic production of precursors. It is not known whether both are operational in non-irradiated cells.

III. Role of Protein or RNA Synthesis in Controlling DNA Synthesis

The relative synthesis of RNA, DNA, and protein within growing cells has been studied with increasing intensity in recent years. Such studies have stressed the importance of the interrelationship between the synthesis of different types of macromolecules. Several studies have been carried out attempting to elucidate the role that protein or RNA synthesis may play in the synthesis of DNA.

It is obvious that protein synthesis must be maintained in dividing cells if each cellular unit is to receive a sufficient complement of enzymes to function properly. Absence of certain enzymes, as we have seen in the preceding section, may help to account for the lack of DNA synthe- sis in certain situations. This may be the case in systems such as regenerating liver and the microspores of the lily anther (discussed above), which are unable to synthesize DNA and may be induced to regain this capacity. Although experiments with such systems are illus- trative of a possible control mechanism, it is questionable whether cells which are rapidly dividing and synthesizing DNA lose the enzymes controlling DNA synthesis during the non-synthetic portion of the DNA cycle.

Measurements of the persistence of such enzyme systems in bacteria have shown their presence, in adequate amounts, in cells in which protein or DNA synthesis, as well as cell division, have been blocked for periods greater than a division cycle (Billen, 1960a; Doudney and Billen, 1961; Billen, 1962a). In a synchronized bacterial system, exhibit- ing periodic DNA synthesis, enzymes which convert thymidine mono- phosphate into thymidine triphosphate double in activity during a defined, short period of the division cycle. Although this increase in activity does not occur in the presence of chloramphenicol, cyclic DNA synthesis persists and enzyme activity is maintained (Maruyama and Lark, unpublished data).

In contrast to these findings, work with plants and animal cells has shown that enzymatic activity may disappear during a period of several days (or a few generations) and that some enzymes may be lost more easily than others. Thus, upon transfer of HeLa or L cells to fresh medium, the intracellular activities of thymidine kinase, thymidylic kinase, and thymidine diphosphate kinase increase sequentially (Weiss-

174 KARL C. LARK

man <■/ <//.. H)()()i. 1 lowcvci', as cell .^i-owili pio^ix'sso, the enzyme ac- tivities arc ay;aiii ledured, thymidine diphospliokinase ])v'w^ lust more rapidly (80''/ of il> actixity disappearm.^ in 24 hours). Addition of thy- midine to the medium appears to stabilize the enzymes ( W'eissman ct al., 1960; Bojarski and lliatt, 19()()l. A similar seciuence of event> has been observed in regenerating liver. Such loss of enzymatic activity does not appear to be uni(iue to the nucleotide kinases since similar findings have been observed with glutamyl transferase (Fottrell and i*aul, 19()1). Similarly, phosphoiylating activity which has appeared in the micro- spores of Tnllium shortly before DNA synthesis also disappears again within a short time (Hotta and Stern, 1961b). Thus, loss of enzymatic activity during the course of the DNA cycle could occur.- However, in no case has the loss of acti\ity been sufficiently rapid to account for cessation of DNA synthesis within 4-10 hours (by the end of the syn- thetic period) nor can the observed dependence of enzyme stability on the medium composition and/or the conditions of growth be reconciled with a precise control mechanism which halts DNA synthesis after an exact duplication of the intracellular DNA content. Thus, although enzj'me biosynthesis ajipears to exei'cise a negative tyjic of conti'ol over the initiation of DNA biosynthesis, as in the case of regenerating liver, such a mechanism cannot account for the degree of control observed.

A more important influence in the regularly dividing cell may be the synthesis of proteins or histones necessary to stabilize the structure^ of the new DNA. It is quite possible that DNA synthesis is linked to the formation of such proteins by a mechanism whereby synthesis of the former cannot proceed without simultaneous or piior syntliesis of the latter (see Section TV).

To answer these (luestions, a more general approach to the role of protein synthesis in regulating DNA formation has been devised by several workers who have studied the effects of inhibiting protein synthe- sis on subse(iuent DNA synthesis. Two tyi)es of inhibition have l)een cmjiloyecl: {(i) starvation for an essential amino acid, and ih) addition of an agent which inhibits protein synthesis.

It has been generally observed that bacteria starved for amino acids required for growth synthesize a small amount of DNA, 50-150% of the pre-existing amount, after which all further synthesis ceases (Pardee and Prestidge, 1956; Earner and Cohen, 1957; Goldstein et al, 1959; Okazaki and Okazaki, 1959; Gros and Oros, 1958; Billen, 1961; Maal0e and Hanawalt, 19611. In such starved cells, all demonstrable protein

"Recently the i)os.sil)ility ha.s been raised lliat tlie observed lark of enzyme activity in system.s .sueh as regenerating liver may he due to the conversion of intracellular enzyme from a soluble into a particulate form (Bianclii d dJ., 1061).

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 175

and RNA synthesis ceases. From these observations it was concluded that synthesis of protein and/or RNA was required for sustained DNA synthesis (Goldstein et al., 1959; Maal0e and Hanawalt, 1961). More careful investigation of the synthesis of DNA in cells starved for an essential amino acid revealed that it could proceed to a limited extent in individual cells before ceasing. (In these studies the cells were also starved of uracil; however, the similarity of the behavior of the culture as a whole to experiments of others in which only amino acids were limited makes it most probable that the effects observed were the direct result of amino acid starvation.) The amounts of DNA made (ca. 50% of the initial cell content), as well as the number of cells which at any time during starvation could be shown (by autoradiography) to retain the capacity for DNA synthesis, were consistent with the hypothesis that each cell continued to finish the DNA replication cycle upon which it had embarked (Hanawalt et al.. 1961). Experiments in which DNA formed under these conditions is labeled by means of a density marker are also consistent with this hypothesis (Hanawalt et al., 1961 ; Nakada and Ryan, 1961). It was suggested that such cells could not commence a second cycle of DNA replication without a short respite from the starvation to which they were subject. It was concluded that protein and/or RNA synthesis were required to initiate a new cycle of DNA replication. The evidence presented appears conclusive in showing that DNA synthesis can proceed at least to a limited extent in the absence of RNA and protein synthesis. Moreover, the hypothesis that a cell engaged in DNA synthesis will finish a cycle of replication despite the absence of RNA or protein synthesis is extremely attractive.

Subsequent experiments have shown, however, that the amount of DNA made will depend upon the growth conditions prior to starvation and that, under certain conditions, cells can more than double their content of DNA or carry on synthesis beyond the end of one cycle. Moreover, quantitative autoradiography indicates that those bacteria which are able to retain their ability to synthesize DNA for a long time may have a relatively slow rate of DNA synthesis (Schaechter, 19611. Thus, an alternative possibility is that all of the cells retain the capacity to synthesize a certain amount of DNA upon deprivation of an essential amino acid but some cells exhaust this capacity quickly, whereas others take a longer time.

Another possibility, which has not been properly resolved, is that under certain starvation conditions bacteria may complete one replica- tion cycle in addition to that in which they were engaged when starva- tion began (possibly by using an endogenous reservoir of the essential amino acid).

I7() KAHL (i. LARK

The bacteria investigated have been shown to synthesize DNA throughout the majority of their (li^•ision cycle. When a siniihir experi- ment was cari'ied out on animal cells grown in tissue culture, which have a defined period, N, of DNA synthesis, they were found, after 48 hours of amino acid deprivation, to be arrested in a state in which the DNA content of the nucleus was that characteristic of cells which have just divided (Paul, 1962) (see Fig. 8).

Similarly, when an amino acid analog, ethionine, was fed to rats, it was found that the DNA content of individual liver cells was one-half that (4C -^ 2C) found in control animals or in animals in which the inhibition was ovei'come by feeding methionine (Holzner et al., 1959). The results of this latter expei-iment are similar to the finding of Schneider et al. (1960) that injections of certain amino acid analogs will result in a loss of ability to incorporate thymidine-H'' into regen- erating liver. This may be due to inability to form active DNA- synthesizing enzymes, or possibly to an effect similar to that observed in bacteria starved for amino acids. Unfortunately, no enzyme assays were carried out and thus it is impossible to distinguish beween the two hypotheses (Schneider et al., 1960). Similar observations to these were made by Lieberman and Ove (1962), who studied DNA synthesis in cells from rabbit kidney cortex.

Further experiments w^ould seem to be required in the animal cell systems in order to determine at what stage in d such cells are halted; whether or not individual cells are all halted in the same stage; and the number of DNA cycles which can take place following the onset of starvation. However, these findings would tend to support the hypothesis that amino acid deprivation allow'ed cells to finish DNA duplication and division but prevented initiation of a new cycle of DNA biosynthesis.

The assumption that protein synthesis is required for the initiation of a new DNA synthetic cycle needs to be examined in greater detail. Although the effects of chloramphenicol on E. coli (Billen, 1959b; Neid- hardt and Gros, 1957; Doudney, 1961a) tend to support the observations on cells starved for amino acids, frequently the magnitude of the chlor- amphenicol effect does not parallel that of amino acid starvation. This is exemplified by experiments with Lactobacillus acidophilus (Okazaki and Okazaki, 1959) for which it has been shown that deletion of essential amino acids from the medium results in suppression of DNA synthesis following a pattern observed in the starved auxotrophs of E. coli described above (residual synthesis of 80-100% of existing intracellular content). However, if RNA and protein synthesis were inhibited by a combination of uracil starvation and the addition of chloramphenicol, no inhibition of DNA synthesis was observed. A similar inhibition of RNA

IV. CELLULAB CONTROL OF DNA BIOSYNTHESIS

177

o o

0 10 20 30 40 50

UNITS DNA

Fig. 8. Microspectrophotometric distribution of the DXA content of individual I. cells (clone P2) grown for 48 hours in (A) complete medium, (B) medium from wliich leucine was omitted, and (C) medium from which glutamine was omitted (10 units of DNA equal to approximatelj^ 4 fifig). In the absence of either leucine or glutamine nuclei accumulate which cannot double their DNA complement. (Unpub- lished data of Dr. J. Paul printed here by his kind permission.)

and protein synthesis could be achieved by simply prolonging the period of uracil starvation, and again no diminution of DNA synthesis was observed (in both cases synthesis of DNA was greater than 600% of the initial intracellular content). As an example of a possible type of ex- planation which could be advanced, the authors pointed out that amino acid-conjugated deoxynucleotides might be required for DNA synthesis.

178 KARL C. I.AHK

It- has also Ui'cii louml that in syiichroiious ciiHuics of hactciia in which the DNA synthetic cych' consists of a step of >ynth('>is foHowcd by a ])('riotl of no synthesis, chloi'aniphcnicol blocks synthesis of cyto- plasmic protein but will allow DNA synthesis to continue until the intfact'llular content of the latter is more than iloubleil. Moreover, DNA continues to l)e synthesized (to a reduced extent) in steps separated by periods ecjual to those found in non-treated cultures (Mai'tiyama anrl Lark. 1961). Periodic DNA synthesis was followed undei- these condi- tions for three cycles (diu-ing which the cellular DNA content doubled. This would argue that protein synthesis did not control the initiation of (or the timing of) a new step of DNA synthesis.

In these experiments, the rate of DNA .synthesis (that is, the amount made in a definite time interval) was found to dejiend upon the time at whii'h chloramphenicol was added to the culture. Thus, it appeare(| that the rate of DNA synthesis was increased if protein synthesis was allowed to proceed during the initial stages of the cell cycle.

A similar study was nuide on bacteria with synchi'onous division cycles in which no cyclic synthesis of DNA occurred (Doudney, 1960). In this system DNA synthesis was inhibited only if chloramphenicol was added at a certain stage of the cell cycle. The author concluded that DNA synthesis depended upon protein synthesis occurring during that particular phase of the cell cycle. This conclusion, however, appears imwarranted since cells subjected to chloram])henicol 10 minutes earlier were perfectly capable of synthesizing DNA for the next 40 mimites, that is, during and after the period in which it was postulated that protein synthesis was required. This occurred despite the fact that l)rotein synthesis was blocked.

An extension of experiments on the starvation of amino acirl auxo- trophs has been carried out by Billen (1960b), who has shown that the recovery of the ability to synthesize DNA upon subsecjuent refeeding can be inhibiterl by irradiation or by chloi'amphenicol. As observed by others, chloramphenicol was not found to be as effective as amino acid starvation in blocking DNA synthesis. However, it blocked resumption of DNA synthesis previously inhil)ited by amino acid starvation. These experiments suggest that during the process of amino acid starvation proteins necessary for DNA synthesis are lost oi' destroyed and that chloramphenicol or irradiation may destioy or inhibit the mechanism for replacing these.'* Experiments have shown that cells thus treated retain all of the active enzymes necessary for DNA .synthesis, i.e., all of the kinases and the polymera.se (Billen, 1960b, 1962a). Therefore, if a

^ Thi.s might occur in a inaiinor similar to tliat in which sulfur-starved cells lose alcofiol-sohiblo ]irotcins (Kolx-rts cl (iL. 19.5.5).

IV. CELLrLAR CONTROL OF DNA BIOSYNTHESIS 179

critical i)rotcin is lost or destroyed, it would have to be of a type other than those involved in the enzymatic synthesis of DNA. Cells in which DNA synthesis has been delayed as a result of irradiation or nitrogen mustard treatment also do not resume this activity if protein synthesis is inhibited (Harold and Ziporin, 1958; Billen, 1959a,b; Harold and Ziporin, 1959; Draculic and Errera, 1959) and, a^ain, either chlor- aniplienicol or starvation of an essential amino acid will pioduce similar effects (Billen, 1961). Extracts of such irradiated cells also do not appear to have lost their enzymatic capacity for in vitro DNA synthesis (Doudney and Billen, 1961).

Thus, although it appears probable that protein synthesis is normally not necessary to allow a cycle of DNA synthesis once initiated to run to completion, further experimentation is required before it is possible to conclude that protein synthesis must occur in order to initiate a new cycle of synthesis. There can be no doubt that deprivation of amino acids from a cell which requires them will inhibit DNA synthesis. How- ever, it has recently been shown (Gallant and Suskind, 1962; Kellen- berger et ciL, 1962) that this inhibition can be reversed by the addition of chloramphenicol alone. As the period of starvation is increased, how- ever, it becomes more difficult to reinstate DNA synthesis either by addition of chloramphenicol or by readdition of amino acids. This second, irreversible effect may represent a true requirement for protein synthesis.

Amino acid starvation has also been shown to block RNA synthesis. ]\Ioreover, evidence has accumulated which indicates that only certain types of RNA may accumulate in the presence of low concentrations of chloramphenicol and that such RNA may be unstable (Neidhardt and Gros, 1957; Astrachan and Volkin, 1958; Britten et al, 1962; Nomura and Watson, 1959). Further investigation (Aronson and Spiegelman, 1961 a,b) has shown that the amount of RNA synthesized, its stability, and perhaps the type of RNA will depend on the concentration of chloramphenicol used and may depend on the amino acid composition of the medium. Thus, under certain conditions, RNA will accumulate in chloramphenicol-treated cells, whereas in others it may be broken down. Radiation will also inhibit the synthesis of RNA as well as of protein. Thus, in several of the experiments cited above, the controlling factor may be the synthesis of a particular type of RNA rather than protein.

This is illustrated by several experiments.

It has been shown that the prolonged synthesis of protein and RNA, under conditions in which DNxA. synthesis is prohibited, results in an increase in the amount of DNA which can be subsequently made in the absence of protein and RNA synthesis. This has been done with a

180 KAKl, (i. LAHK

mutant of E. coll B, T" liist^, rc(iuiriiig thyniiiK' ami histidine, starved of thymine in the presence of histidine, and subsequently starved of histidine in the presence of thymine. From the data it appears that the amount of new DNA synthesized appears to be proportional to the net amount of RNA which can be synthesized. Chloram])hcnicol could be shown to block this increased capacity for DNA .synthesis (Nakada, 1960).^

Bacteria starved of an amino acid (tryi>toj)lian I arc unable to synthe- size DNA in the presence of chloramphenicol and tryptophan. Upon preincubation with amino acid, starved cells recover the ability to syntlicsizc DNA in the presence of the drug. The rate at which this ability is recovered has been correlated with the synthesis of RNA fDoudney, 1961b). A similar result was obtained if DNA synthesis was blocked as a result of UV irradiation rather than amino acid starvation (Doudney, 1961a). In both of these experiments some correlation could, of course, be made with the recovery of protein synthesis which is known to depend on RNA synthesis. The frequency of ultraviolet induced mutations (sliown to be maximal if DNA synthesis is allowed to proceed as rapidly as possible) also appears to be greatest if the synthesis of RNA is not inhibited (Doudney and Haas, 1961).

Finally, it has been demonstrated that the genetic inability to syn- thesize RNA in the absence of amino acids (Stent and Brenner, 1961) also extends to the synthesis of DNA (Kellcnberger et al., 1962).

Although these findings imply that RNA rather than protein synthe- sis may be necessary for the initiation of DNA synthesis, the results of Okazaki and Okazaki (1959) argue that this is not the case, since they obtained DNA synthesis without RNA synthesis when lactobacilli were grown in the absence of uracil. It has been suggested (Kellenbergcr et al., 1962) that both RNA and DNA synthesis are subject to the control mechanism involving amino acid participation.

IV. The Physical State of DNA as a Factor Controlling Its Replication

The requirement for pre-existent DNA to act both as a template and a primer in the synthesis of DNA has been well established by m ?'?Y/-o studies (see Chapter I). The availability of intracellular DNA to act in this capacity may play an important role in controlling its own biosyn- thesis. This has been suggested by the following experiments:

^ Allhough the investigator noted a correlation between the amount of protein synthesized and the amount of DNA which can be made, this appears to be errone- ous, since the comparison is made between net contents of DNA rather than newly synthesized DNA. From the data, the correlation with newly synthesized RN.A appears to be more significant.

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 181

Cultures of synchronized bacteria {Alcaligenes faecalis), in which DNA synthesis occurs in steps, will cease to synthesize DNA if large amounts of deoxyadenosine are added to the medium (presumably as a result of precursor starvation; see Section II.A). If this is done immedi- ately before a burst of DNA synthesis (to cells which presumably have completed all the necessary preparations for DNA synthesis) DNA synthesis is blocked. Reversal of inhibition at different periods there- after leads to an eventual resumption of DNA synthesis which occurs during a period in which a control (non-inhibited) culture undergoes a second cycle of DNA synthesis. The timing at which this eventual DNA synthesis takes place is independent of when inhibition is initiated or reversed (Lark, 1960), and is illustrated in Fig. 9. Non-synchronized exponential cultures of this organism treated in this way initiate DNA synthesis immediately upon reversal of inhibition and proceed to synthe- size DNA exponentially at the normal rate. These results have been interpreted as evidence for the existence of certain allowed periods of DNA synthesis. The lack of correspondence between these periods and the periods for which DNA precursor production is inhibited led to the suggestion that control of these periods depended upon the conversion of DNA to a primer state rather than upon the production of precursor material. This was further emphasized by the finding that DNA pre- cursors, when added to the growth medium, could only influence the DNA cycle if added during an allowed period of synthesis.

Cultures of E. coll starved of an essential amino acid cease produc- tion of DNA. Ability to synthesize this macromolecule is regained upon readdition of the essential amino acid. Reacquisition of this synthetic ability is inhibited by both chloramphenicol and irradiation. Extracts of cells so inhibited exhibit adequate activity of all of the enzymes neces- sary to synthesize DNA (Billen, 1962b) . The lack of in vivo synthesis has been interpreted as an inability of intracellular DNA to enter into the primer state.

The slime mold Physarum polycephalum undergoes synchronized mitosis. Guttes and Guttes (1961) studied the ability of a "well-fed" Plasmodium to support DNA synthesis in an implanted plasmodium which had been previously starved. DNA synthesis in the former took place 3-4 hours after mitosis, whereas in starved plasmodia it occurs after 6-9 hours. If a starving plasmodium was implanted into one which was well fed at a period in which neither was synthesizing DNA, the well-fed Plasmodium could initiate and support DNA synthesis of the implant despite the fact that it was not itself synthesizing DNA. They interpreted their results to indicate that DNA synthesis in the well-fed plant was controlled by some factor other than the substrate and enzyme conditions required for DNA synthesis and suggested that the ability of

182

KARL G. LARK

■8
■ 4 ^ a> E 3 C	;	-X	0 A
• •	1	t t
•1	A	0 A
	20	40 60	80
8			■^
■4 < Q ■2	^	y 0' , .-.X-,---»-A-^	0-° a'
	t	t t
■1	X	0 ^

20 40 60 80 Minutes after elution Fig. 9. Pattern of DNA synthesis and cell division following removal of the

inhibitor deoxyadenosine (DOA). % #, control; X X, DOA; O O, no

DOA after 35 minutes; A A, no DOA after 50 minutes. DOA was added at a high

concentration (10"" il/) immediately before the onset of DNA synthesis. Inhibition was reversed at different times (as indicated) by the addition of adenosine. Resumii- tion of DNA synthesis occurred at the same time as the second step of DNA synthesis in the control culture despite reversal of inhibition at different times. In contrast, inhibition of cell division continued in proportion to the length of time DOA was present. The timing of DNA synthesis appears to be independent of the DNA content of the cell, or of whether or not a previous cycle of DNA sj^nthesis was completed. (From Lark, 1960.)

DNA to act as primer was the controlling factor. (Recijirocal implant experiments demonstrated that starved plasmodia did not contain an inhibitor of DNA synthesis.)

In all of these studies the hypothesis of jirimer control was suggested because of the inadequacy of alternative hypotheses. In no case was a lirimcr-c'ontrollcd mechanism directly demonstrated.

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 183

The dependence of i)rinier activity on the physical state of DNA may be inferred from in vitro studies of the primer requirements of cell-free polymerizing systems. The availability of primer for DNA duplication may involve several factors: (a) Is it necessary for DNA to exist in a special form in order for it to be duplicated? (5) Is DNA synthesis compatible with simultaneous RNA synthesis, and if not, to what extent are' these two processes incompatible? (c) Is DNA structurally stabilized within the cell and is this a necessary part of DNA replication? If so, factors needed for this stabilization may control its replication, (d) Is the replication of DNA physically ordered, with respect to the macro- molecule itself, with a beginning, a middle, and an end? If so, a control mechanism may involve the appropriate conditions for initiation of synthesis at one end of the molecule, maintenance of synthesis through- out the molecule, and termination of synthesis once the end of the mole- cule is reached. The onset and cessation of synthesis may also involve the activation of information contained within the molecule itself.

Little experimental evidence is available to answer these questions. Thus, the following discussion will be largely speculative.

A. IS IT NECESSARY FOR DNA TO EXIST IN A SPECIAL FORM IN ORDER FOR IT TO BE DUPLICATED?

In vitro studies of the deoxynucleotide triphosphate polymerase have shown that enzymes isolated from a variety of sources require primer in order to act (see Chapter I). The ability of DNA to function as a primer in these reactions varies. In some systems native DNA appears to function efficiently as a primer, whereas in others heated or denatured DNA is required.

It appears difficult, therefore, to reach a general decision from such experiments as to whether the jiolymerizing reaction utilizes a special form of DNA as primer. Certainly, in some systems, denatured DNA ai:)pcars to be required.

It has been shown that denaturation of DNA by heat involves strand separation (Meselson and Stahl, 1958). Similar separation occurs as a result of DNA duplication in that one-half of the parent DNA molecule is found together with an equal amount of new material in a "hybrid," semi-conserved, daughter molecule (Meselson and Stahl, 1958; Sueoka, 1960; Taylor et al., 1957; Forro and Wertheimer, 1960). Despite these findings, the presence during the DNA cycle of a distinct molecule or species, which could serve as primer DNA, has seldom been observed. In one experiment (Cavalieri and Rosenberg, 1961c), it was possible to isolate a special species of DNA shortly before DNA replication in a culture of E. coli T in which discontinuous DNA synthesis had been

184 KARL G. LARK

initiated by means of previous thymine starvation. Tliis material was characterized by a greater susceptibility to DNase than normal DNA. The kinetics of DNase action were interpreted as evidence that the molecule contained half of the number of component strands usually present (Cavalieri and Rosenberg, 1961a). In all of its characteristics it resembled DNA obtained at other times from the same organism but heated in cesium chloride. On the basis of their findings these workers suggested that there exist two types of DNA molecules during the rejilication cycle: before DNA synthesis the molecule is cleaved and subsecjuently rebuilt during the replication process. Further studies of this type would be desirable, since no evidence was obtained to indicate whether or not the observed change in state of the DNA was reversed following DNA synthesis. Evidence was presented to show that DNA existed in a different state in non-proliferating as compared to dividing cells (Cavalieri and Rosenberg, 1961b).

Indirect evidence for a change in the state of DNA during its repli- cation has been obtained by a study of the action spectrum for the inactivation of bacteriophages by ultraviolet light. Studies on the inactivation of viruses believed to contain a single- as opposed to double-stranded DNA molecules have demonstrated that those with single-stranded DNA have a different action spectrum from those with two-stranded DNA. A shift in the action spectrum of two-stranded DNA virus to that characteristic of a single-stranded DNA virus occurs following host infection and initiation of virus replication. As virus multiplication proceeds this spectrum shift is reversed (Setlow, 1960; Setlow and Boyce, 1960; Setlow and Setlow, 1960). These results have been interpreted as evidence that during replication viral DNA passes through a single-stranded or "denatured" stage.

A change in the physical state of intracellular bacterial DNA im- mediately following replication has been observed by Goldstein and Brown (1962). They found that newly synthesized DNA is more resistant to sonic disintegration than the bulk of cellular DNA. The nature of their experiment excluded a determination of whether this state was achieved prior to replication.

Recent analysis of DNA isolated at intervals during the division cycle of synchronized bacteria (.4. faecalis) has revealed the presence of small amounts of material whose density corresponds to that of heated or denatured DNA. The amount of material observed (which is sensitive to DNase and can be heat renatured) is usually less than 10% of the total DNA content of the culture (Lark, Cavalieri, and Rosenberg, unpublished results) .

Much less, or no such material, is found in cultures of randomly dividing cells. The "denatured DNA" arises when the cells are lysed with

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 185

Diiponol to obtain their DNA. The temperature at which lysis is carried out and the m^racellular environment of the DNA at the time of lysis, appear to be important in determining the quantity of "denatured DNA" obtained. The evidence indicated that the intracellular environment is detemiining the temperature at which denaturation of the cellular DNA takes place. During the DNA cycle the amount of denatured material isolated varies. Further studies in this and other systems should facili- tate the investigation of control mechanisms in which primer require- ments are involved.

All of the evidence available indicates that in the process of replica- tion the DNA molecule must change its state (splitting the molecule in half). However, it is not certain that this does not occur as an automatic result of the replication process instead of as a first step. From the experiments cited, it would appear that the latter may be true. Such a step may then represent an important controlling factor in the biosyn- thesis of DNA.

There is no evidence for the existence of a biological system capable of converting native into primer DNA. It is tempting to assign such a role to DNase and an increase in activity of this enzyme has been shown shortly before DNA synthesis occurs in regenerating liver (Brody and Balis, 1959). In vitro studies have shown that a brief treatment with this enzyme results in strand separation measured as a large increase in the reactivity of the nucleoside bases with formalin fSarkar and Dounce, 1961; Sarkar, 1961). Moreover, DNA treated in this way becomes active when tested in a biological system requiring denatured DNA as primer (Sarkar and Dounce, 1961; Dounce et al., 1961; Mont- savinos and Canellakis, 1961). Unfortunately, it appears that during such treatment the molecule suffers several random scissions, a result which appears to be incompatible with the lack of observed breakage in the replication of the chromosome.^ It is, of course, possible that specific DNase action (Laskowski, 1961) might cause non-random scissions com- patible with the structural integrity of DNA.

A more attractive possibility is suggested by the observations on the conditions under which the "denatured DNA" is obtained from cells. Changes in the intracellular environment of DNA may lower the temperature of denaturation into the physiological range. When this occurs, a portion of the DNA molecule would suddenly denature and synthesis would be initiated. Such a hypothesis could easily explain the timing of DNA synthesis and would account for the obsen^ed effects of temperature upon the DNA cycle (see Section V.C below).

In vitro studies would indicate that such an idea is feasible. Hama-

^ DNA, in which scissions have occurred as a result of X-radiation in vivo, undergoes degradation upon replication (Kas and DracuHc, 1961).

186 KARL O. LARK

ifiu'hi and ( Ji'icluschck (19(32) have i-('iwrtcd tlic dcnaturation of DNA at teniperatures as low as 25° in the presence of certain anions and Keir et al. (1962) have noted that DNA can serve as primer at slightly alkaline pH's (8-9) without having been heated.

B. IS DNA SYNTHESIS COMPATIHLE WIVU SI.MI LTANKOUS RNA SVNTHKSIS?

Is DNA, which is s('r\-in<i as a functional tciiiplate for the formation of cellular proteins, simultaneously able to act as a template for its own synthesis? Much recent evidence has accumulated to show that in the synthesis of proteins, RNA serves as a messenger, to carry the necessary information between the DNA template and the ribosome which serves as the site of protein synthesis (Brenner et al., 1961 ; Gros et al, 1961).

Moreover, it has been shown both in vivo and in vitro that this RNA is formed in combination with DNA and will recomplex with it (Spiegel- man et al., 1961; Geiduschek et al., 1961; Hall and Spiegelman, 1961). In one instance an RNA-DNA-protcin complex thus formed was quite stable (Bonner et al., 1961).

Under some conditions the ability to replicate DNA may be related to the ability to synthesize messenger RNA. Bacteria which require thymine, arginine, and uracil for growth cease to synthesize DNA when starved of arginine and uracil in the presence of excess thymine. When arginine and uracil are returned to the medium, DNA synthesis resumes gradually after a characteristic lag. This lag in ability to synthesize DNA is independent of the actual synthesis of DNA since it occurs irrespective of whether or not thymine has been omitted from the medium. An identical lag in the ability to induce ^-galactosidase forma- tion is also observed (Roberts, 1960), which is also independent of the actual occurrence of DNA synthesis. (On the other hand, gross protein and RNA synthesis are reinitiated instantly with no apparent lag.) It is possible, therefore, that DNA in such cells is in a physical state which renders a portion of it unabh^ to serve as a template in either RNA or DNA synthesis, oi" alternatively, that its activity is non-s])ecifically re- pressed for a short period of time.

Experiments on DNA and RNA synthesis in amino acid-starved auxotrophs of E. coli have indicated that synthesis of both can be regu- lated by a common, amino acid-dependent, control mechanism (Kellcn- berger et al., 1962). This mechanism would appc^ar to operate on the ■primer rather than the precursor level.

No in vitro experiments have been reported, however, in which it was sought to determine whether the synthesis of "messenger" RNA is inhibited by the simultaneous synthesis of DNA or vice versa. Some

IV. CELLILAR CONTROL OF DNA BIOSYNTHESIS 187

evidence has been obtained which indicates that heat-denatured DNA, which will act as primer in self-replication, is unable to function in the in vitro synthesis of messenger RNA (Burma et al., 1961 ; Ochoa et al., 1961; Eisenstadt et al, 1962).

That DNA and RNA synthesis are mutually exclusive in vivo is indicated by the experiments of Sisken (1959) and of Prescott and Kimball (1961). The latter workers studied the synthesis of RNA, DNA, and protein in the macronucleus of the ciliate Euplotes. As mentioned previously, DNA synthesis in this case proceeds via a DNA-synthesizing band, commencing at the edge of each lobe of the U-shaped macronucleus and continuing until the two bands meet (Gall, 1959). The nucleus then divides. They demonstrated that this DNA-synthesizing band is devoid of RNA and incapable of synthesizing RNA. It is only this part of the nucleus which was shown to have these two characteristics. Whether the progression of the synthesizing band is responsible for the cessation of RNA synthesis and dissociation of RNA-DNA complexes, or the result of it, is not known. It is obvious that if the latter were true, dis- sociation of the RNA-DNA complex and cessation of RNA synthesis could be a controlling factor in DNA synthesis.

Such a mechanism could explain the observed dependence of DNA synthesis on the ability of cells to carry on RNA synthesis (see Section III) since RNA synthesis could be blocked in such a way as to prevent dissociation of a DNA-RNA complex, or to favor reassociation. The recent findings of Aronson and Spicgelman ( 1961a, b) would indicate that chloramphenicol might, under certain conditions, create a situation in which this could occur. The formation under these conditions of RNx\- DNA complexes could also explain the reduction of ultraviolet-induced mutation.-^ which occurs in the absence of RNA synthesis (Witkin, 1959; Doudney and Haas, 1961), since complex formation with RNA and subsequent lack of DNA replication might help to stabilize the structure of DNA until ultraviolet-induced lesions could be repaired. On the other hand, mutations induced by a change in base structure (resulting from alkylating agents) should not be affected, as is indeed the case (Strauss and Okuba, 1960). Finally, the observed dependence of "thymineless death"« (Gallant and Suskind, 1961, 1962) and DNA synthesis (Kellen- berger et al, 1962) on conditions leading to RNA synthesis could also be explained as a stabilization of DNA (which had undergone lesions as the result of attempted synthesis in the absence of thymine) by complex- ing with RNA. In all of these situations, RNA synthesis may be visualized as resulting in a turnover, and therefore dissociation, of RNA-

* This occurs when bacteria which require thymine for growth are incubated in media lacking thymine.

188 KARL G. LARK

DNA conipU'xcs witli the concurrent release of the DNA molecule, which if unstable may then untlei-<z;o change.

C. IS DNA STRUCTURALLY STABILIZED WITHIN THE CELL AND IS THIS A NECESSARY PART OF DNA REPLICATION?

^luch evidence has accumulated to demonstrate that DNA exists within the cell in the form of nucleoprotein (Bloch, 1958). We have already discussed the evidence which exists for the role of protein in controlling DNA synthesis. De (1961) has shown that histone is synthe- sized sinmltaneously with DNA in the chromatin of interphase plant cells. Additional protein and histone synthesis was demonstrated in the nucleolus during Go, following DNA rei)lication (see also Woodard et al., 1961). In the studies cited above with Euplotes, it was noted that the synthesis of DNA and of protein or histone (incorporation of thymidine and of histidine) were extremely active in the same region of the synthe- sizing band. This region of active synthesis, however, was extremely low in its content of both DNA and protein. The region of the band immedi- ately adjacent contained large amounts of both. It thus would appear that the synthesis of both DNA and protein proceed simultaneously. An interesting point in this connection is that this protein can apparently be synthesized in the absence of RNA, since this region of the macro- nucleus is devoid of this nucleic acid. Thus, this nuclear protein may be synthesized by a special mechanism which may, as pointed out earlier, be chloramphenicol-insensitivc (Nathans and Lipmann, 1961). The role of this protein is unknown. However, it has been suggested that such protein may serve to hold several DNA molecules together, linking them into a chromatid (Dounce, 1959; Kellenberger, 1960). This might account for the marked instability of such large chromatids which, al- though they can be shown to exist within the cell, break down, upon extraction, into smaller pieces (Burgi and Hershey, 1961). The molecular weight of the latter suggests a frequent and uniform occurrence within the chromatid of weak links. It has been suggested that the synthesis of DNA may be controlled by the rate of synthesis of these "linkers" (Kellenberger, 1960).

D. IS THE REPLICATION OF DNA PHYSICALLY ORDERED, WITH RESPECT TO THE MACROMOLECULE ITSELF, WITH A BEGINNING, A MIDDLE, AND AN END?

No evidence exists to indicate whether the DNA molecule, the chromatid, or portions of the latter are synthesized in a linear progres- sion or not. The evidence for a preferentially timed synthesis of individ- ual chromosomes has been discussed in a previous chapter (Chapter II). Exjieriments on density labeled chromatin indicate that the proportion

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 189

of the chromatin in the process of being synthesized is extremely small compared to that which is either already synthesized completely or that which is not yet synthesized. Thus, only a small portion of the material is being replicated at any time (Sueoka, 1960; Meselson and Stahl, 1958). Whether this portion moves from one end of the chromatid to the other is not known. That this may be the case is indicated indirectly by the experiments of Rudner (1960) on the induction of mutations by 2-aminopurine incorporated into a synchronized culture of bacteria. In this case, appearance of mutants occurred in steps, whereas DNA was synthesized continuously. This would indicate that synthesis of particu- lar loci in the DNA molecule occurred at precisely timed intervals.

It is difficult to see how sequential synthesis of DNA can be avoided if strand separation is to occur without frequent breakage or crossing over which would destroy the semi-conservative pattern of chromosome duplication which has often been observed. If DNA synthesis occurs in small bursts covering the rapid synthesis of one segment (molecular weight ca. 10") after another of the chromatid and if these are to be stabilized bj^ linkers of some type, it also would appear reasonable to suppose that a stabilized type of bonding (perhaps covalent) must occur at each end of the segment, to prevent subsequent rupture of terminal hydrogen bonds and possible loss of the semi-conservative pattern. Such bonding has been shown to result from ultraviolet irradiation and it is possible that it could occur in the absence of irradiation. If this were the case, initiation of synthesis could be an all-or-none event dependent upon breakage of such critical bonds (Marmur and Grossman, 1961).

A key question concerning control via a point of initiation may be: What happens to cells in wdiich DNA synthesis is interrupted before completion? When synthesis is resumed will it resume at the point of the chromatid where it was interrupted, or at the original initiation point? We do not know whether, once initiated, the conditions necessary for DNA synthesis remain imposed upon the cell until such synthesis can be completed. To study this, it is necessary to have a system in which damage to DNA or other cell constituents does not occur during the blockage of synthesis. The use of deoxyadenosine inhibition for autoradiographic studies of chromosome synthesis in higher cells would appear extremely fruitful from this point of view.

In general, it is obvious that if DNA synthesis is initiated at one end of the molecule, special conditions will be needed to distinguish the beginning and end of the molecule and to jirepare it for initiation of synthesis.

Such a statement implies that the molecule itself carries specific information which differentiates one end from the other. If this dif-

190 KARL (!. LARK

fert'utiation is to ])lay a coiilrolliii'i; role in DNA hiosyntlicsis then the availability of thi.s information must be under some on-off type of control, similar to the oonti-ol of operator <j;enes by oytoplasmie repressors (see Jacob and Monod, 19611. However, a siniilai' t>|)c of cytoplasmic repressor would not appear to be involved, since conditions leading to the accumulation of RNA, DNA, or i)roteiii within tlic cell do not affect the timing of DNA synthesis (Maruyania and Lark. I9()li. whicli is also independent of tlu^ synthesis or lack of synthesis of DNA itself (Lark, 1960).

Guttes and Guttes (1961) have suggested that the nucleolus may play the role of an "oiierator" since its synthesis is initiated simul- taneously with that of DNA.

V. Other Factors Affecting DNA Biosynthesis and the DNA Cycle

A. THE LOCALIZATION OF DNA SYNTHESI.S

The in vivo synthesis of DNA depends upon its environment within the cell. Thus, fragments of DNA (resulting from radiation damage) which are left in the cytoplasm as micronuclei can no longer serve as foci for DNA synthesis although the parent nucleus still does (Das and Alfert, 1961b). In ciliates different nuclei such as the macronucleus and the micromicleus have been found to follow different patterns of DNA synthesis, the micronucleus entering its period of DNA synthesis when the macronucleus has finished and when the cell is beginning to divide (McDonald, 1962; Prescott ef al, 1962).

These data would indicate that the nucleus represents a unique environment for DNA .synthesis; and it has been shown that specific DNA precursors such as thymidine, when taken into the cell from the environment, will rapidly apjiear in their free form, or as nucleotides (thymidine mono-, di-, or triphosiihate), in the nucleus, but are not found in the cytoplasm (Crathorn and Shooter, 1960). Howe\-ci'. good evidence has been accumulated demonstrating that the deoxyribonucleo- tide polymerase is found in the cytoplasm, but 7wt in the nucleus of mammalian cells (Smellie of a/.. 1959; Smellie and Eason, 1961; Prescott et al., 1962a). In l)ringing together all of these components refjuired for the synthesis of DNA, a controlling factor may be the permeability of the nuclear membrane.

A further factor may be the association within th(> nucleus of enzyme, substrate, and primer. Billen 1 1962b) has recently demonstrated that in E. coli the polymerase is complexed witli the DNA. It is not known whether the formation of this complex is a controlling factor in DNA

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 191

synthesis. E. coli is known to synthesize DNA throughout most of its cell cycle (see Introductionj. However, the same type of complex is found at all times in synchronous cultures of A. jaecalh, which synthe- size DNA discontinuously (Billcn and Lark, unpublished data).

B. VARIATION IN THE DNA CYCLE

Whereas rapidly dividing cells in tissue culture often exhibit a pre- cisely defined cell cycle, the same is not necessarily true for cells growing in the animal. Mendelsohn et al. (1960) have shown that transplanted tumor tissue shows a large variation in the length of the period preceding DNA synthesis (GJ, whereas both .S and Go remain rather constant over the cell population. Studies on the intestinal epithelium and the ear epidermis of the mouse (Quastler and Sherman, 1959; Quastler, 1960; Sherman et al., 1961) have shown a similar variation in G'l, as opposed to S or Go. Moreover, different cell types from the same animal showed differences in the average length of their Gi periods. Differences were also observed in S and Go, but these were much less striking. In addition, it was possible to show that the decision as to whether an intestinal epithelial cell should cease dividing and differentiate (move into the villi) or to continue proliferation, was taken during the early portion of the cell cycle before DNA was synthesized. On the other hand, cells of the epidermal layer of the ear often ceased division in the Ga (Gelfant, 1960) as well as the Gi (Gelfant, 1961) stages. However, the low tem- perature at which these cells are growing may account for a mitotic block arresting cells in G2.

Upon injury, cells may move out of both Gi and G2 to commence DNA synthesis and/or division. This has been demonstrated with cells of the epidermal layer of the mouse ear and a similar situation may exist in tissue of the eye (Harding and Srinivason, 1961). It is, of course, known to exist in liver as evidenced by the events involved in regenera- tion (Barnum et al., 1957). Injury would appear in these cases to pro- voke a humoral response capable of initiating DNA synthesis and hence cell proliferation. Thus, partial hepatectomy of one animal will result in liver proliferation in another animal which is placed in parabiosis with it (Bucher et al, 1951).

It is interesting to note that during liver regeneration a good deal of the variability between the DNA cycles of individual animals (as evi- denced by increasing asynchrony of DNA synthesis) may be reduced by the use of inbred strains of the same age which have been partially hepatectomizcd at the same time. This would indicate that the inherent rate of progression through Gi may be related, in some way, to a diurnal

192 KARL O. LARK

cycle of the animal as a whole and thus be under sonic humoral control mechanism which does not involve the actual synthesis of DNA itself.

Althoupih variation among individual cells is much less for animal cells growing in culture, Tcrasima and Tolmach (1962) have noted that most of the randomization observed at M in a synchronized population of HeLa cells occurs during d. Moreover, animal cells growing in culture are halted in Gi when the medium in which they are growing is exhausted (Whitfield and Rixon, 1959). The length of the DNA cycle may also change for an entire cell population when the growth conditions in culture are altered. When this occurs, the change is expressed as a lengthening or shortening of Gi rather than S or Gn (Richards et al., 1956). A similar situation may occur in bacteria where it has been shown that major increases in growth rate brought about by enriching the growth medium are accompanied by an increase in the average DNA content per nucleus (Schaechter et al, 1958). This would indicate that more DNA is now being synthesized in the early stages of the cell cycle thus raising the average DNA content per nucleus.

The data available would indicate that under most conditions, Gi, the portion of the DNA cycle prior to synthesis, is the most susceptible to change.

C. EFFECTS OF TEMPERATURE AND RADIATION ON THE DNA CYCLE

Although several workers have noted the effects of temperature on the division cycle, few studies have been made of the effect of heat or cold on DNA synthesis and these have yielded little information. Zeuthen and Scherbaum (1954) synchronized tetrahymena by repeatedly shifting the temperature from 29° to 34°C. At the latter temperature cell division appears to be blocked, whereas DNA synthesis continued. Subsequent experiments demonstrated that during such repeated cycles, the cells doubled their content of DNA without undergoing division (Scherbaum, 1957; Scherbaum et al., 1959) but that upon being returned to their optimal growth temperature, division took place before further DNA synthesis was resumed (thus reducing the DNA content). These experiments indicate that the temperature block (in this case a heat block) was operative during the pre-raitotic, or Ga period.

Randomly dividing bacteria (S. typhimurium) grown at 25°C have the same content of DNA as when grown at 37°, although the rate of synthesis at the former temperature is about half of that at the latter (Lark and Maal0e, 1956). However, when cells grown at 25°C are shifted to 37°, a sudden burst of DNA synthesis occurs equal to approxi- mately 30% of the cellular content of DNA. Synthesis is then arrested for a short period. Repeated application of such temperature shifts at

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 193

properly spaced intervals results in phasing of DNA synthesis and in synchronization of division. These results indicate that the population of cells growing at 25° can be divided into two classes on the basis of their ability to synthesize DNA when placed at 37° and that their DNA cycles may be divided into portions of different temperature dependence. A similar result was found with another bacterial system {B. mega- therium) which continued to synthesize DNA at a rapid rate for a short period of time when shifted from 34° to 15° (Falcone and Szybalski, 1956). In this case, all other cell growth parameters were immediately affected by the lower temperature. Returning the temperature to 34° resulted in synchrony of cell division and phasing of DNA synthesis (Delamater, 1956). This again would indicate the existence of at least two portions of the DNA cycle with different temperature dependencies.

Hotchkiss (1954) found that the competence of pneumococci to be transformed by DNA was increased by a short period of incubation at 25°. This resulted in a synchronization of cell division and a phasing of competence. The latter, however, occurred in cycles whose frequency was somewhat different from those of cell division. Whether the cycles of competence are related to DNA synthesis, to adsorption of transforming DNA (Lerman and Tolmach, 1957), or to the ability of the transforming DNA to successfully survive within the cell until it is integrated (Fox and Hotchkiss, 1960) is not known. Further investigation of this system to ascertain the fate of the "infecting" DNA as a function of the host cell DNA cycle should be extremely profitable.

The DNA cycle of HeLa cells synchronized by cold shock differs from that of randomly dividing cells not subjected to cold (Smith et al., 1959). These results indicate that tw^o populations of cells may be created by their treatment — one whose DNA is synthesized as in cells not exposed to cold (toward the end of the cycle), and the other com- posed of cells in which Gi has been almost eliminated and DNA is synthesized immediately after division.

The meager information on the effects of temperature on the DNA cycle indicates that different portions of the cycle may have different temperature coefficients and that the DNA cycle may be dissociated to a certain extent from the cell division cycle.

As mentioned previously, similar suggestions have resulted from studies with X-radiation. In general, exposure to X-radiation appears to block cells in G2 and prevent mitosis. Cells which have not, at the time of irradiation, synthesized DNA proceed to do so before the cycle is halted although, in some cases, the onset of DNA synthesis is delayed (for reference see Seed, 1961; Whitmore et al.. 1961; and Quastler, 1962).

194 KARL n. LARK

lIowi'NiT, ii'crnt t'\'ifloii('e has (k'lnun.-^tratiMl that ladiatioii can, uiidcr certain circunii-tanccs, initiate DNA synthesis in onion root ti])S before the normal onset of S (Das and Alfert, 1961a). Moreover, the rate of DNA synthesis early in 5 can be stimulated in both plant and animal cells by X-radiation (Alfert and Das, 1961). A similar ob.servation was made on irradiated intestinal epithelium (Sherman and Quastler, 1960) who observed that cells in N matuicd early following irradiation. These results are consistent with tlie hypothesis that some chromosomes are synthesized earlier in S than otliers and that replication of umhiplicated chromosomes may he initiated by irradiation. Presumably tliis could occur by conversion of their DNA to a primer state.

More infomiation should become available with the study of the effects of physical changes in the environment of cells that have been synchronized by selective methods. Such studies have been initiated by Terasima and Tolmach (1962), who have studied the effect of tempera- ture and X-radiation on cultures of HeLa cells synchronized by a selective method. Preliminary results have indicated that a period of reduced temperature (29°C) during Gi, or S will delay mitosis, whereas Go does not appear to be as temperature dependent. X-radiation during S w'ill prolong this period of DNA synthesis, whereas radiation during Gi is without effect. In general, radiation was found to delay or block mitosis. Most striking, perhaps is the finding that cells about to enter S are most sensitive to the lethal effect of X-radiation and tliat tliis sensitivity can be greatly increased by the addition of dcoxyadenosine late in Gi to block DNA synthesis and entry into S (Terasima and Tolmach, 1962).

Measurement of the rate of DNA synthesis during S in synchronized HeLa cells has shown that it is not constant (Terasima and Tolmach, 1962). This result con-fii'ms the indirect observations of Howai'd and Dewey (1961) and Painter et al. (1960), who came to a similar conclu- sion on the basis of variation in the grain count of cell labeled during random growth.

D. THE EFFECT OF THE NUTRIENT ENVIRONMENT

Temporaiy deprivation of thymidine will result in synchronization of cell division and DNA synthesis in both bacteria (Earner and Cohen. 1956) and in animal cells, i.e., HeLa (Rueckert and Mueller, 1960). In the latter case, cell division ceases upon treatment with the thymidine analog fluorodeoxyuridine (FI'^DR), whereas attempted DNA synthesis continues for a short period resulting in the formation of a "DNA-like material." Readdition of thymidine results in resumption of DNA synthesis in two steps, just before and shortly after cell division. This

IV. CELLULAR CONTROL OF DXA BIOSYNTHESLS 195

result is similar to the effect of cold shock on such cells (Smith et al., 1959).

These results serve to (listinf2;uish two classes of cells in the randomly dividing population. Presumably these differ with respect to when in their DNA synthetic cycle they have been subjected to the inhil)itoiy agent.

This has been further clarified by Taylor et al. (1962), who have demonstrated that thymidine deprivation will only block mitosis if it occurs early in *S. On the other hand, cells subjected to FUDR late in S proceed into mitosis with a concomitant fragmentation of the chromo- some. Removal of the thymidine block and addition of thymidine as late as 1 hour before anaphase will prevent fragmentation. This may indicate that incorporation of thymidine into specific sites in the nuclear material may be necessary even during Go if the integrity of the chromosome is to be maintained. This has been substantiated by Hsu and Somers (1962). They demonstrated that the number of chromosome breaks increased in proportion to the length of time prior to mitosis (that is, in Go or 5) that cells were subjected to low concentrations of FUDR. The number of breaks increased continuously showing no sudden rise in number which would reciprocally represent a transition from »S to G-..

Under certain conditions alteration of the nutritional environment may alter the DXA cycle without l)locking cell growth or division. Only a few such experiments will be discussed.

Bacteria {E. coli) when grown exponentially under conditions of balanced growth appear to synthesize DNA throughout the majority of their division cycle (Schaechter et al, 1959). However, when ribosides are added to the medium DXA synthesis becomes cyclic, occurring in steps separated by periods in which almost no synthesis takes place (Lark and Lark, 1962). No phasing was observed in cell division, although the over-all growth rate was increased. This result is related to the findings of Kjeldgaard et al., (1958», who observed a delay before the rate of DNA synthesis was increased as a result of shifting bacteria to a medium in which growth occurred at a faster rate. No phasing of DNA was reported at this time, although the data indicate that some may in fact have occurred. In this case, phasing of nuclear division was reported. Subsequent studies (Maal0e, 1961) have shown that during the transition period in which the synthetic rate is eventually increased, the cell popu- lation becomes heterogeneous, some cells commencing to synthesize DNA at a more rajiid rate than others. These results indicate that fol- lowing a "shift up" into a richer medium two classes of cells may be eventually distinguished. Maal0e (1961) has interpreted this to indicate that cells must complete a given pliysiological pattern of events (either

196 KARL C. LARK

the DNA rt'plication cycle it'^elf or sonic other ]);ir;iinctci-c()iit rolling replication rate) before increasing the DNA synthetic rate. However, a phasing of synthesis as observed with libonucleotides would intlicate that the DNA cycle of individual cells may be composed of at least two components, which differ markedly in their response to nutritional changes in the medium. Such an explanation would be consistent with the observed effect of temperature on such cells.

The stimulation of DNA synthesis as a result of wounding (see Sec- tion V,B) is characterized by a stimulation of cells often far removed from the area of trauma. This suggests the ingestion by these cells of some compound or comi)ounds capable of stimulating DNA synthesis. These may be cellular breakdown products. On the other hand, ingestion of foreign substances may give rise to a similar stimulation. Dutton et al. (1962) have studied the stimulation of DNA synthesis which occurs upon addition of a specific antigen to spleen cells taken from an innnunizcd animal. It would appear probable that the results of further detailed study of this system could be applied with profit to the study of regenerating systems, especially with reference to the ability of humoral factors to stimulate specific tissues or organs.

Deoxynucleosides and tides have been observ^ed to increase the rate of division and mitosis of cells (Karpfel et al., 1959; Firshein and Braun, 1960; Butros, 1959; Gruelich et al, 1961) as well as of DNA synthesis (Firshein, 1961). A study of the effect of deoxynucleosides on the timing of DNA synthesis in synchronized bacteria (Lark, 1960) demonstrated that addition of low concentrations of the four deoxyribosides (deoxy- cytidine, thymidine, deoxyguanosine, and deoxyadenosine) would destroy periodic DNA synthesis. This effect only occurred if these compounds were present intracellularly during a period of DNA synthesis (ap- proximately 30% of the division cycle). Deoxyribosides are normally not found within such cells (see Section II), which have been shown to contain deoxynucloetides only. The observed effects were interpreted to indicate: (a) that during the period of DNA synthesis the control of the synthetic process could be affected by these substances, presumably as a result of their competition with nucleotides engaged in base-pairing with primer (this might extend the period of DNA synthesis) ; and (b) that the time of initiation of a period of DNA synthesis was related, in some manner, to the time at which a previous cycle had been completed. It was assumed that a lack of absolute synchrony would lead to a dif- ferent response by individual cells to the competitive effects of deoxy- ribosides tending to further randomize the population and destroy synchrony.

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 197

E. THE DISSOCIATION OF THE CELL DIVISION CYCLE FROM DNA SYNTHESIS

We have already noted the use of inhibitors to study the dependence of DNA synthesis on precursor production (Section II) or on synthesis of protein and RNA (Sections II and IV). Such experiments have also proved extremely useful in studying the dissociation of DNA synthesis from mitosis, or cell division.

Mercaj^toethanol has been used to block the development of centrioles in fertilized sea urchin eggs, an event known to precede DNA synthesis. Nevertheless, DNA synthesis (if the agent is added in interphase) will proceed (Bucher and Mazia, 1960). On the other hand, if cells in meta- phase are blocked by this agent, DNA synthesis is also blocked, indi- cating that DNA in the condensed form of the metaphase chromosome cannot duplicate. In this case, an interesting phenomenon may occur. The centers divide and form a quadripolar figure and upon the release of inhibition the egg divides into four, each of the nuclei receiving only half of its normal complement of DNA. Nevertheless, this appears to be adequate since such cells go on to further division and development. The DNA deficiency is made up before the next division since the cells go through a cycle in which the required number of centriolar units are restored (Mazia, 1960). It would be of interest to know whether the DNA so made, is made in two cycles of normal duration, including Gi, S, and G\.. or whether the cell immediately replenishes the necessary material in one big step.

As pointed out in this section (see Section V,B and C), under certain conditions, both temperature and nutritional changes can be used to dissociate the cell division cycle from DNA synthesis. Similarly, inhibi- tors of protein synthesis in bacteria may block cell division while allow- ing a DNA cycle to proceed (Maruyama and Lark, 1961), whereas an inhibitor of DNA synthesis (deoxyadenosine) will block DNA synthesis while permitting a further cycle of cell division (Lark, 1960).

Finally, repeated observations of synchronous cultures have shown that cycles of cell division and DNA synthesis may often be dissociated under conditions of continued cell growth and proliferation (Lark, 1962).

VI. Summary, Conclusion, and Opinions

To date, the most complete body of experimental results concerning the regulation of DNA biosynthesis has been concerned with the role of precursors. From the data described above, there seems to be little doubt that production of DNA precursors may act as part of a mechanism for

198 KARL G. LARK

initiatin^ij; DXA biosynthesis, and that rcjjicssion of ccrlain cnzynics may lead to a lack of such synthesis. In correhiting the production of DNA precursors with DNA biosyntiiesis two rehitionships may be encountered: The first, which may exist in germinal and embryonic tissue, could involve the induction of a metabolic pathway as a result of accumulation of large quantities of DNA precursors. Tn this situation, the sudden IM-oduction of precursor may set in motion a train of events leading to the initiation of DNA biosynthesis. The second, which may exist in cells from adult organisms, in vegetative cells, and in bacteria, is one in which prochiction of precursors occurs as a result of derepression and the machinery for producing DNA precursors is eitlier synthesized or acti- vated upon demand. The two systems differ only in that in the first, a stimulus applied at one end of the reaction chain forces biochemical events toward initiation of biosynthesis, whereas in the second, the initiation of biosynthesis forces a need for mechanisms which supply the necessary precursor materials. In both cases the mechanism for supply- ing the necessary precursors may be the same; however, in the first this mechanism may act as stimulus to initiate biosynthesis.

In no case, however, can such mechanisms exercise more than a crude control over the synthesis of DNA. It is impossible to explain the termination of synthesis and its reinitiation with no api)arent disap- pearance of the enzymatic mechanisms evoked. Nor can the stoichi- ometry of DNA biosynthesis, resulting in the exact duplication of all of the material initially present, be accounted for. Finally, the rather specific pattern of time at which individual chromosomes are duplicated (discussed in Chapter II) cannot be controlled by such a mechanism.

An additional, more intimate and precise control appears to be operative which is capable of providing the stimulus for activating the machinery for precursor production.

The role of the primer in maintaining such a control is very attrac- tive, but has yet to be investigated in vivo. Nevertheless, it would appear that it could play an all-important role in the control of biosynthesis; a compelling reason for assuming this, is that it is only at this level that the stoichiometry of DNA synthesis could be controlled.

Any controlling role of the DNA molecule in its own biosynthesis would appear to be complex indeed. The experiments discussed in Sec- tions III and IV above would indicate that, among other things, the interaction of this macromoleculc with others such as RNA and protein will play an im]:)ortant role in determining its availability to act in its own biosynthesis. Before such a mechanism can be established, experi- ments must be carried out on in vivo systems to establish the existence of a primer state and a means of quantitating its jiarticipation in bio-

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 199

synthesis. At present such experiments involve measurements carried out on material which has been extracted from cells in a given state of growth or division. However, the present state of microspectrophoto- metric techniques suggests that it may be possible to carry out physico- chemical investigations of macromolecules in situ. If a role of primer in controlling biosynthesis can be established, the problems of the control of DNA biosynthesis will have only begun. It will then be necessary to establish the mechanisms whereby non-replicating material is converted to the primer state. The experiments discussed indicate that both protein and RNA synthesis may play an important role in this respect. The state of the template which directs RNA and protein synthesis may depend upon its interaction with these macromolecules. Moreover, it ap- pears that DNA biosynthesis requires concomitant histone biosynthesis. The timing of DNA biosynthesis appears to depend upon the termina- tion of the previous period during which synthesis could take place. This suggests that some "timing reaction" is initiated upon cessation of DNA synthesis or upon cessation of a particular state ("primer") associated with DNA biosynthesis. The existence of such a reaction remains to be either proved or disproved, but the ability of chemical and physical changes in the environment to affect the DNA cycle of different cells to different extents suggests that such a mechanism exists and may be operative during G^ but not during S, the period of DNA synthesis. A possible mechanism has been suggested in Section IV,A above.

The above conclusions concern intracellular control mechanisms operating at the molecular level. However, it is obvious that these in turn must be affected by the external environment of the cell and by its internal structure. The experiments discussed in Section V are only a small sample of the data available. However, they serve to illustrate the complexity of this problem. Within the cell, molecular components of the control mechanism must be brought together. Changes in the environ- ment, often produced by neighboring cells, may initiate or halt the synthetic cycle. The period elapsing between mitosis and the onset of DNA synthesis appears to be extremely important in deciding the destiny of a cell. The most immediate effect of this decision may be the ability to synthesize DNA.

The significance of this aspect of the control of DNA biosynthesis may, perhaps, be viewed as follows: A series of molecular events within the cell control the functioning and duplication of DNA in a precise manner. The result of this control will determine whether the capacities of the cell are directed toward some particular purpose, other than duplication, or to replication of the cell itself. Changes produced by the interaction of the cell and its environment may be influential in

200 KARL G. LARK

tihifting the balance of control to emphasize either i)roliferation or func- tion. The control of DNA biosynthesis may thus present the key to the problem of cellular (lifferentiation.

Acknowledgments

The autlior wishes to thank Drs. R. Okazaki and J. Paul for making available their results prior to publication. He is indebted to Drs. V. Bonifas, H. Stern, L. J. Tolmach, and P. M. B. Walker for their comments on the manuscript. Unpublislied work of the author, presented above, was sujjported by grant E-1391 of the National Institutes of Health, Bethcsda, Maryland.

References

Abbo, F. E., and Pardee, A. B. (1960). Biochim. el Biophys. Acta 39, 478.

Alfert, M. (1955). Stjmpofsiuni oti Fine Struclure of Cells, Leiden, 1054, p. 157.

Alfert, M., and Das, N. K. (1961). Proc. Am. Soc. Cell. Biol. 1, 6.

Aronson, A. I., and Spiegelman, S. (1961a). Biochim. et Biophys. Acta 70, 70.

Aronson, A. I., and Spiegelman, S. (1961b). Biochim. et Biophys. Acta 70, 85.

Astrachan, L., and Volkin, E. (1958). Biochim. et Biophys. Acta 32, 449.

Baddiley, J., and Blumson, N. L. (1960). Biochim. et Biophys. Acta 39, 376.

Barner, H. D., and Cohen, S. S. (1956). J. Bacteriol. 72, 115.

Barner, H. D., and Cohen, S. S. (1957). J. Bacteriol. 74, 350.

Barnum, C. P., Jardetzky, C. D., and Halberg, F. (1957). Texas Repts. Biol. Med.

15, 135. Bessman, M. J., Lehman, I. R., Simms, E. S., and Romberg, A. (1958). J. Biol.

Chem. 233, 171. Bianchi, P. A., Crathorn, A. R., and Shooter, K. V. (1961). Pruc. Intern. Congr.

Biophys., Stockholm, 1061, p. 230. Billen, D. (1959a). Nature 184, 174. Billen, D. (1959b). Biochim. et Biophys. Acta 34, 110. Billen, D. (1960a). Nature 187, 1044. Billen, D. (1960b). J. Bacteriol 80, 86. Billen, D. (1961). Pathol. Biol. 9, 758. Billen, D. (1962a). Biochim. et Biophys. Acta, in i)ress. Billen, D. (1962b). Biochem. Biophys. Research Communs. 7, 173. Bloch, D. P. (1958). In "Frontiers in Cytology" (S. L. Palay, ed.), p. 113. Yale

Univ. Press, New Haven, Connecticut. Bloch, D. P., and Godman, G. C. (1954). J. Histochem. and Cytochem. 2, 455. Bloch, D. P., and Godman, G. C. (1955). J. Biophys. Biochem. Cytol. 1, 17. Bojarski, B., and Hiatt, H. H. (1960). Nature 188, 1112. BoUum, F. J., and Potter, V. R. (1959). Cancer Research 19, 561. Bollum, F. J., Andregg, J. W., McElya. A. B., and Potter, V. R. (1960). Cancer

Research 30, 138. Bonner, J., Huang, R. C, and Maheshwari, M. (1961). Proc. Natl. Acad. Sci. U.S.

47, 1548. Brenner, S., Jacob, F., and Meselson, M. (1961). Nature 190, 576. Britten, R. J., McCarthy, B. J., and Roberts, R. B. (1962). Biophys. J. 2, 83. Brody, S., and Balis, M. E. (1959). Cancer Research 19, 538. Bucher, N. R. L., and Mazia. D. (1960). J. Biophys. Biochem. Cytol. 7, 651. Bucher, N. R. L., Scott, J. E., and Aub, J. C. (1951). Cancer Research 11, 457.

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 201

Burgi, E., and Hershey, A. D. (1961). J. Mol. Biol. 3, 458.

Burma, D., Kroger, H., Ochoa, S., Warner, R., and Weill, J. (1961). Proc. Natl.

Acad. Sci. US. 47, 749. Burns, V. W. (1961). Exptl. Cell Research 23, 582. Butros, J. M. (1959). Exptl. Cell Research 18, 318.

Canellakis, E. S., and Montsavinos, R. (1958). Biochivi. et Biophys. Acta 27, 643. Canellakis, E. S., Jaffe, J. J., Montsavino.s, R., and Krakow, J. S. (1959). J. Biol.

Chem. 234, 2096. Cavalieri, L. F., and Rosenberg, B. H. (1961a). Biophys. J. 1, 317. Cavalieri, L. F., and Rosenberg, B. H. (1961b). Biophys. J. 1, 323. Cavalieri, L. F., and Rosenberg, B. H. (1961c). Biophys. J. 1, 337. Cohen, S. S. (1947). Cold Spring Harbor Symposium Quant. Biol. 12, 35. Cohen, S. S., Barner, H. D., and Lichtenstein, J. (1961). J. Biol. Chem. 236, 1448. Crathorn, A. R., and Shooter, K. V. (1960). Nature 187, 614. Das, N. K., and Alfert, M. (1961a). Proc. Natl. Acad. Sci. UJi. 47, 1. Das, N. K., and Alfert, M. (1961b). Proc. Am. Soc. Cell Biol. 1, 42. Davidson, J. N., and Bishop, C. N. (1957). Brit. J. Radiol. 30, 367. De, D. N. (1961). The Nucleus 4, 1.

Delamater, E. D. (1956). Symposium Soc. Gen. Microbiol. 6, 215. Doudney, C. O. (1960). /. Bacterial. 79, 122.

Doudney, C. O. (1961a). Biochem. Biophys. Research Communs. 4, 218. Doudney, C. O. (1961b). Biochein. Biophys. Research Communs. 5, 405. Doudney, C. O., and Billen, D. (1961). Nature 190, 545.

Doudney, C. O., and Haas, F. L. (1961). Genetics 45, 1481.

Dounce, A. L. (1959). Ann. N .Y . Acad. Sci. 81, 794.

Dounce, A. L., Sarkar, N. K., and Kay, E. R. (1961). J. Cellular Comp. Physiol. 57, 47.

Draculic, M., and Errera, M. (1959). Biochim. et Biophys. Acta 31, 459.

Dutton, R. W., and Pearce, J. D. (1962). Nature 194, 93.

Dutton, R. H., Eady, J. D., and Bulman, H. X. (1962). 16th Ann. Stjmposium on Fundamental Cancer Research, Houston, Texas, 1962, in press.

Edwards, J. L., Koch, A. L., Youcis, P., Freese, H. L., Laite, M. B., and Donalson, J. T. (1960). J. Biophys. Biochem. Cytol. 7, 273.

Eisenstadt, J. M.. Kayeyama, T.. and Novelli. G. D. (1962). Proc. Natl. Acad. Sci. U.S. 48, 652.

Erikson, R. 0. (1948). Am. J. Botany 35, 729.

P\alcone, G., and Szybalski, W. (1956). Exptl. Cell Re.'iearch 11, 486.

Feinendigen, L. E., Bond, V. P., and Painter, R. B. (1961). Exptl. Cell Research 22, 381.

Finamore, F. J., and Volkin, E. (1958). Exptl. Cell Research 15, 405.

Firket, H., and Verly, W. G. (1958). Nature 181, 274.

Firshein, W. (1961). /. Bactenol. 82, 169.

Firshein, W., and Braun, W. (1960). J. Bacteriol. 79, 246.

Flaks, J. G., and Cohen, S. S. (1957). Biochim. et Biophys. Acta 25, 667.

Forro, F., Jr., and Wertheimer, S. A. (1960). Biochim. et Biophys. Acta 40, 9.

Foster, T., and Stern, H. (1958). Science 128, 653.

Foster, T., and Stern, H. (1959). J. Biophys. Biochem. Cytol. 5, 187.

Fottrell, P., and Paul, J. (1961). Biochem. J. 80, 17.

Fox, M. S., and Hotchkiss, R. D. (1960). Nature 187, 1002.

Franzen, J. S., and Binkley, S. B. (1961). J. Biol. Chem. 236, 515.

202 KARL G. LARK

FriiHlkin, M. (1959). /// "The Kinetics of Cellular riolifcnitioa" (F. Stolilman, cd.),

p. 97. Gninc & Slratton, New York. Call. J. G. (1959). J. Biophys. Biochem. Cylol. 5, 295. Callant, J., and 8u.skind, S. A. (1962). Bactetiol. Proc. 61, 181. Gallant, J., and Suskind, S. R. (1962). Biuchim. et Biophys. Acta 55, 627. Gciduschck. E. P., Nakanioto, T., and Weiss, S. B. (1961). Pn>r. yall. Arad. Sci. U.S.

47, 1405. GeKant, S. (1960). Ann. N.Y. Acad. Sci. 90, 536. Gelfant, S. (1961). Proc. Am. Soc. Cell Biol. 1, 67.

Goldstein, A., and Brown, B. J. (1962). Biochim. el Biophys. Ada 53, 19. Goldstein, A., Goldstein, D. B., Brown, B. J., and Clion. S. (1959). Biuchim. et

Biophys. Acta 36, 163. Grant, P. (1958). J. Cellular Cump. Physiol. 52, 227. Gray, E. D., Weissman, S. M., Richards, J., Bell, D., Koir, H. M., Smellie, R. M. S.,

and Davidson, J. N. (1960). Biochim. et Biophys. Acta 45, 111. Gregg, J. R., and L0trup, S. (1955). Biol. Bull. 108, 29. Gros, F., and Gros, F. (1958). Exptl. Cell Research 14, 104. Gros, F., Hiatt, H., Gilbert, W., Kulad, C. G., Rissebrough, R. W., and Watson, J. D.

(1961). Nature 190, 581. Gruelich, R. C., Cameron, I. L., and Thrasher, J. D. (1961). Proc. Natl. Acad. Sci.

U.S. 47, 743. Guttes, E., and Guttes, S. (1961). Proc. Am. Soc. Cell Biol. 1, 79. Hall, B. D., and Spiegelman, S. (1961). Proc. Natl. Acad. Sci. U.S. 47, 137. Hamaguchi, K., and Geiduschek, E. P. (1962). J. A771. Chem. Soc. 84, 1329. Hanawalt, P. C, Maal0e, O., Cummings, D. I., and Schaechter, M. (1961). /. Mol.

Biol. 3, 156. Harding, C. V., and Srinivason, B. D. (1961). Exptl. Cell. Research 25, 326. Harold, F. M., and Ziporin, Z. Z. (1958). Biochim. et Biophys. Ada 28, 492. Harold, F. M., and Ziporin, Z. Z. (1959). Biochim. et Biophys. Acta 29, 439. Hecht, L. I., and Potter, V. R. (1956). Cancer Research 16, 999. Hecht, L. I., Potter, V. R., and Herbert, E. (1954). Biochim. et Biophys. Ada 15,

134. Hershey, A. D., Dixon, J., and Chase, M. (1953). J. Gen. Physiol. 36, 777. Hiatt, H. H., and Bojanski, T. B. (1960). Biochem. Biophys. Research Communs.

2, 35. HofT-j0rgensen, E., and Zeuthen, E. (1952). Nature 169, 245.

Holzner, J. H., Barker, T., and Popper, H. (1959). J. Natl. Cancer Inst. 23, 1215. Hornsey, S., and Howard, A. (1956). Ann. N.Y. Acad. Sci. 63, 915. Hotchkiss, R. D. (1954). Proc. Natl. Acad. Sci. U.S. 40, 49. Hotta, Y., and Stern, H. (1961a). J. Biophys. Biochem. Cytol. 9, 279. Hotta, Y., and Stern, H. (1961h). ./. Biophys. Biochem. Cytol. 11, 311. Howard, A., and Dewey, D. L. (1961). Exptl. Cell Research 24, 623. Howard, A., and Pelc, S. R. (1953). Heredity 6 (Suppl.), 261. Hsu, and Somers, (1962). Abstr. 2nd Meeting Am. Soc. Cell Biol. Iwamura, T. (1955). J. Biochem. (Tokyo) 42, 575. Jacob, F., and Monod, J. (1961). J. Mol. Biol. 3, 318. Josse, J. (1962). oth Intern. Congr. Biochem., Moscoio, 1961, in press. Kanazir, D. (1954). Biochhn. et Biophys. Acta 15, 592.

Karnofsky, D. A., and Basch, R. S. (1960). ./. Biophys. Biochem. Cytol. 7, 61. Karnofsky, D. A., and Bevolander, G. (1958). Proc. Soc. Exptl. Biol. Med. 97, 32. Karpfel, Z., Saska, J., and Drasil, V. (1959). Nature 183, 1600.

IV. CELLULAR CONTEOL OF DNA BIOSYNTHESIS 203

Kas, E., and Draculic, M. (1961). Biochem. Biophys. Research Commnns. 5, 152.

Keir, H. M., Binnie, B., and Smellie, R. M. S. (1962). Biochem. J. 82, 492.

Kellenberger, E. (1960). Symposium Gen. Microbiol. 10, 39.

Kellenberger, E., Lark, K. G., and Bolle, A. (1962). Proc. Natl. Acad. U.S., in press.

Kimball, R. F., and Borka, T. (1959). Exptl. Cell Research 17, 173.

Kjeldgaard, N. O., Maal0e, 0., and Schaechter, M. (1958). J. Gen. Microbiol. 19, 607.

Klenow, H. (1961). Biochem. Biophys. Research Commnns. 5, 156.

Romberg, A. (1959). Harvey Lectures 53, 83.

Kornberg, A., Zimmerman, S. B., Romberg, S. R., and Josse, J. (1959). Proc. Natl.

Acad. Sci. U.S. 45, 772. Ruriki, Y., and Okazaki, R. (1959). Embryologia 4, 337. Lajtha, L. G., Oliver, R., and Ellis, F. (1954). Brit. J. Cancer 8, 367. Lark, R. G. (1960). Biochim. et Biophys. Acta 45, 121. Lark, R. G. (1961). Biochim. et Biophys. Acta 51, 107. Lark, R. G. (1962). 22nd Intern. Congr. Physiol. Sci.. Leiden. 1962. p. 766. Lark, K. G. LTnpublished data.

Lark, C, and Lark, R. G. (1962). Biochim. et Biophys. Acta 55, 401. Lark, R. G., and Maal0e, O. (1956). Biochim. et Biophys. Acta 21, 448. Lark, R. G., Cavalieri, L., and Rosenberg, B. H. (1962). Unpublished data. Laskowski, M., Sr. (1961). In "The Enzymes" (P. D. Boyer, H. Lardy, and R. Myer-

back, eds.), Vol. V, p. 123. Academic Press, New York. Lehman, I. R., Bessman, M. J., Simms, E. S., and Romberg, A. (1958). /. Biol.

Chem. 233, 163. LePage, G. A. (1957). J. Biol. Chem. 226, 135.

Lerman, L. S., and Tolmach, L. J. (1957). Biochim. et Biophys. Acta 26, 68. Levinthal, C. (1956). Proc. Natl. Acad. Sci. U.S. 42, 394. Lieberman, I., and Ove, P. (1962). /. Biol. Chem. 237, 1634. Longer, L., and Rlenow, H. (1960). Biochim. et Biophys. Acta 37, 33. Maal0e, 0. (1961). Cold Spring Harbor Sympo.sia Quant. Biol. 26, 45. Maa!0e, O., and Hanawalt, P. C. (1961). J. Mol. Biol. 3, 144. McDonald, B. B. (1958). Biol. Bull. 118, 71. McDonald, B. B. (1962). /. Cell Biol. 13, 193.

McFall, E., and Stent, G. (1959). Biochim. et Biophys. Acta 34, 580. McLeish, J., and Sunderland, N. (1961). Exptl. Cell Research 24, 527. Main, R. R., Cole, L. J., and EHis, M. E. (1957). Nature 180, 1287. Marmur, J., and Grossman, L. (1961). Proc. Nat. Acad. Sci. U.S. 47, 778. Maruyama, Y. (1956). J. Bacteriol. 72, 821.

Maruyama, Y., and Lark, R. G. (1959). Exptl. Cell Research 18, 389. Maruyama, Y., and Lark, R. G. (1961). Exptl. Cell Research 25, 161. Maruyama, Y., and Lark, R. G. (1962). Exptl. Cell Research 26, 382. Mazia, D. (1960). Ann. N.Y. Acad. Sci. 90, 455. Mendelsohn, M. L. (1960). J. Natl. Cancer Inst. 25, 485. Mendelsohn, M. L., Dohan, F. C, Jr., and Moore, H. A., Jr. (1960). J. Natl. Cancer

Inst. 25, 477. Meselson, M., and Stahl, F. W. (1958). Proc. Nat. Acad. Sci. U.S. 44, 672. Montsavinos, R., and Canellakis, E. S. (1959). Cancer Research 19, 1239. Morris, N. R., and Fischer, G. A. (1960). Biochim. et Biophys. Acta 42, 183. Munsch-Peterson, A. (1960). Biochem. Biophys. Research Commims. 3, 392. Nakada, D. (1960). Biochim. et Biophys. Acta 44, 243. Nakada, D., and Ryan, F. J. (1961). Nature 189, 398. Nathans, D., and Lipmann, F. (1961). Proc. Natl. Acad. Sci. U.S. 47, 497.

204 KARI. <;. I-AHK

Npi(lli;inlt. F. CJ., and (Iros, F. (1957). liiorhrni. ct Hu>i>hy><. Acta 25, 513.

Newton, A. A., and Wildy, P. (1959). Exptl. Cell Research 16, 624.

Nomura, M., and Watson. J. D. (1959). ./. Mol. Biol. 1, 204.

Xygaard. 0. F., Gultos, S., and Hu.scli. H. P. (1960). BiocJiim. el Binphys. Aria 38,

298. Orhoa, S.. Hunna. I)., Krofirr, H., and Woill, .J. (1961). Proc. Natl. Acad. Sci. U.S.

47, 749. O'Donnrll, J. F.. Mackal, R. P.. and F:vans, E. A., Jr. (1958). J. Biol. Chem. 233,

1523. Ofiur, M., Min("kl(>r, S., and McCIary, D. O. (1953). J. Bacteriol. 66, 642. Okazaki, R. (1959). Biochem. Biophys. Research Commons. 1, 34. Okazaki, R. (1960). Biochim. et Biophys. Acta 44, 478. Okazaki, R., and Okazaki, T. (1958). Biochim. et Biophys. Acta 28, 470. Okazaki, R., Okazaki, T., and Kiiriki, Y. (1959). Biochim. et Biophys. Acta 33, 289. Okazaki, R., Okazaki, T., and Kuriki. T. (1960). Biochim. el Biophys. Acta 38, 384. Okazaki, T., and Okazaki, R. (1959). Biochim. et Biophys. Acta 35, 434. Ord, M. G., and Stockon, L. A. (1958). Biochim. el Biophys. Acta 29, 201. Overgaard-Hansen. K., and Klenow. H. (1961). Proc. Natl. Acad. Sci. U.S. 47, 680. Painter, R. B., and Drew, R. M. (1959). Lab. Invest. 8, 278.

Painter, R. B., Drew, R. M., and Giauque, B. G. (1960). Exi)tl. Cell Rescarcli 21, 98. Pardee, A. B. (1961). Symposium Soc. Gen. Microbiol. 11, 19. Pardee, A. B., and Prestidge, L. S. (1956). J. Bacteriol. 71, 677. Parizek, J., Arient, M., Dienstbier, Z., and Skoda, J. (1958). Nature 182, 721. Paul, J. (1962). Personal Communication. Pogo, A. O., Cordero Funes, J. R., and Mordah, J. (1960). Exptl. Cell Research 21,

482. Poschki.s, K. E., Bartuska, D., Zogerman, J., Goddard, J. W., and Cantarow, A.

(1959). Cancer Research 19, 561. Potter, R. L., and Buettner-Janusch, T. (1958). J. Biol. Chem. 233, 462. Potter, R. L., and Schlesinger, S. (1955). J. Am. Chem. Soc. 77, 6714. Potter, R. L., Schlesinger, S., Buettner-Janusch, V., and Thomjison, L. (1957). J.

Biol. Chem. 226, 381. Prescott, D. M. (1960). Exptl. Cell Research 19, 228.

Prescott, D. M., and Kimball, R. F. (1961). Proc. Natl. Acad. Sci. U.S. 47, 686. Prescott, D. M., BoUum, F. J., and Kluss, B. C. (1962). J. Cell Biol. 13, 172. Prescott, D. M., Kimball, R. F., and Carrier, R. F. (1962). /. Cell Biol. 13, 175. Prusoff, W. H. (1959). Biochem. Pharmacol. 2, 221. Quastler, H. (1960). Ann. N.Y. Acad. Sci. 90, 580. Quastler, H. (1962). In "Chemical and Biological Effects of Radiation" (Hais-

sinsky, ed.), in press. Quastler, H., and Sherman, F. G. (1959). Exptl. Cell Research 17, 420. Reichard, P. (1960). Biochim. el Biophys. Acta 41, 368. Reichard, P., Canellakis, Z. N., and Canellakis, E. S. (1961). J. Biol. Chew. 2,36,

2514. Richards, B. M., Walker, P. M. B., and Deeley, E. M. (1956). Ann. AM'. Acad.

Sci. 63, 831. Roberts, R. B. (1960). Ann. N.Y. Acad. Sci. 88, 752. Roberts, R. B., Abelson, P. H., Cowie, D. B., Bolton, E. T., and Britten. R. J.

(1955). Studies of biosynthesis in E. coli. Cainegie Inst, of Wash. Publ. 607. Rotherham, J., and Schneider, W. C. (1958). ./. Biol. Chew. 232, 853.

IV. CELLULAR CONTROL OF DNA BIOSYNTHESIS 205

Rotherham, J., and Schneidor, W. C. (1960). Biochim. et Biophys. Acln 41, 348.

Rudner, R. (1960). Biochcm. Biophys. Research Comrnuns. 3, 275.

Rueckert, R. R., and Mueller, G. C. (1960). Cancer Research 20, 1584.

Sarkar, N. K. (1961). Arch. Biochem. Biophys. 93, 328.

Sarkar, N. K., and Bounce, A. L. (1961). Biochim. et Biophys. Acta 49, 160.

Schaechter, M. (1961). Cold Spring Harbor Symposia Quant. Biol. 26, 53.

Schaechter, M., Maal0e, O., and Kjeldgaard, N. 0. (1958). J. Gen. Microbiol. 19, 592.

Schaechter, M., Bentzon. M. W., and Maal0e, O. (1959). Nature 183, 1207.

Scherbaum, O. (1957). Exptl. Cell Research 13, 24.

Scherbaum, O. H., Lauderbach, A. L., and Jalin, T. L. (1959). Exptl. Cell Research

18, 150. Schneider, J. H., Cassir, R., and Chordikion, F. (1960). J. Biol. Chem. 235, 1437. Schneider, W. C. (1955). J. Biol. Chem. 216, 287. Schneider, W. C. (1957). J. Natl. Cancer Inst. 18, 569.

Schneider, W. C, and Brownell, L. W. (1957). J. Natl. Cayicer hist. 18, 579. Scott, D. B. M., and Chu, E. (1958). Exptl. Cell Research 14, 166. Seed, J. (1961). Nature 192, 944.

Setlow, R. B. (1960). Biochim. et Biophys. Acta 39, 180. Setlow, R. B., and Boyce, R. (1960). Biophys. J. 1, 29. Setlow, R. B., and Setlow, S. K. (1960). Proc. Natl. Acad. Sci. U.S. 46, 791. Sherman, F. G., and Quastler, H. (1960). Exptl. Cell Research 19, 343. Sherman, F. G., Quastler, H., and Wimber, D. R. (1961). Exptl. Cell Research 25,

114. Simon, E. H. (1961). /. Mol. Biol. 3, 101. Sisken, J. E. (1959). Exptl. Cell Research 16, 602.

Sisken, J. E., and Kinosita, R. (1961). J. Biophys. Biochcm. Cytol. 9, 509. Smelhe, R. M. S., and Eason, R. (1961). Biochcm. J. 80, 39P. Smelhe, R. M. S., Keir, H. M., and Davidson, J. N. (1959). Biochim.. el Biophys.

Acta 35, 389. Smellie, R. M. S., Gray, E. D., Keir, H. M., Richards, J., Bell, D., and Davidson,

J. N. (1960). Biochim. et Biophys. Acta 37, 243. Smith, C. L., Newton, A. A., and Wildy, P. (1959). Nature 184, 107. Smith, R., and Maal0e, O. (1962). Biochim. et Biophys. Acta, in press. Spiegelman, S. B., Hall, D., and Storck, R. (1961). Proc. Natl. Acad. Sci. U.S. 47,

1135. Stanners, C. P., and Till, J. E. (1960). Biochim. et Biophys. Acta 37, 406. Stent, G. S., and Brenner, S. (1961). Proc. Natl. Acad. Sci. U.S. 47, 2005. Stern, H. (1960a). 18th Symposium on Growth, New York. 1959. p. 135. Stern, H. (1960b). Ann. N.Y. Acad. Sci. 90, 440. Stern, H. (1961). /. Biophys. Biochem. Cytol. 9, 271. Stern, H. (1962). Personal communication. Stern, H., and Hotta, Y. (1962). Personal communication. Strauss, B., and Okubo, S. (1960). J. Bacterial. 79, 464.

Strominger, J. L., and Scott, S. S. (1959). Biochim. et Biophys. Acta 35, 376. Sueoka, N. (1960). Proc. Natl. Acad. Sci. US. 46, 83. Sugino, Y. (1960). Biochim. et Biophys. Acta 40, 425. Sugino, Y., Sugino, X., Okazaki, R., and Okazaki, T. (1960). Biochim. et Biophys.

Acta 40, 417. Swift, H. H. (1950). Physiol. Zool. 23, 169. Takats, S. (1962). Am. J. Botany 49, 748.

206 KARL C. LARK

Taylor, J. II. (1960a). Advances in Biol, and Mai. PIn/s. 7, 107.

Taylor, J. II. (1960b). Ann. N.Y. Acad. Sci. 90, 409.

Taylor, J. H. (1960c). J. Biophya. Biochem. Cylol. 7, 455.

Taylor, J. H., and MoMa.stor, R. D. (1954). Chromosoma 6, 489.

Taylor. J. H., Wood.-:, P. S., and Hughes, W. L. (1957). Proc. Natl. Acad. Sci. U£.

43, 122. Taylor, J. H., Haul, \V. F., and Tung, J. (1962). Proc. Xatl. Acad. Sci. UJi. 48, 190. Terasima. T.. anil Tolmach, L. J. (1961). Nature 190, 1210. Terasinia, T., ami Tolmach, L. J. (1962). Unpublislied results. Utsumi, K. R. (1961). Gann. 52, 203.

Vendrely, R., and Vendrely, C. (1956). Intern. Rev. Cytol. 5, 171. Walker, P. M. B., and Yates, H. B. (1952). Proc. Roy. Soc. B140, 274. Walker, P. M. B., and Mitchison, J. M. (1957). Exptl. Cell Research 13, 167. Wei.s.sman, S. M., Smellie, H. M. >S., and Paul, J. (1960). Biochim. et Biophys. Acta

45, 101. Whittield, J. F., and Rixon, R. H. (1959). Exptl. Cell Research 18, 126. Whitmore, G. F., Stanners, C. P., Till, J. E., and Gulyas, S. (1960). Biochim. et

Biophys. Acta 45, 101. Whitmore, G. F., Stanners, C. P., Tdl, J. E., aii.i Gulya.s, S. (1961). Biochim. et

Biophys. Acta 47, 66. Williamson, D. H., and Scopes, A. W. (1960). Exptl. Cell Research 20, 338. Williamson, M. B., and Gaschlbauer, W. (1961). J. Biol. Chem. 236, 1463. Witkin, E. M. (1956). Cold Spring Harbor Symputiia Quant. Biol. 21, 123. Witkin, E. M. (1959). Proc. 10th Intern. Congr. Genet. Montreal, 1058, p. 280. Woodard, J., Gelber, B., and Swift, J. (1961). Exptl. Cell Research 23, 258. Young, I. E., and Fitz-James, P. C. (1959). Nature 183, 372. Zeuthen, E., and Scherbaum, O. (1954). Proc. Symposium Colston Research Soc.

7,41.

Chapter V

Molecular Mechanism of Mutations

ERNST FREESE^

Department of Genetics, University of Wisconsin, Madison, Wisconsin

I. Introduction 207

A. Classification of Heritable Variations 207

B. Development of Genetic Systems and Discovery of the First Mutagens 209

C. The Genetic System of rlj Mutants of T2 or T4 Phages . . . .211

D. Further Subdivision of Mutations 212

II. Structure, Duphcation, and Alteration of the Hereditary Material . . 215

A. Structure and Replication of DNA and RNA 215

B. Structure and Duplication of Chromosomes 216

C. Classification of Base Alterations in Nucleic Acids 216

III. Biochemistry of Mutagens and Theory of Base Alterations .... 217

A. Inhibition of Nucleic Acid Precursors 218

B. Incorporation of Base Analogs 219

C. Dyes 224

D. Chemical Alteration of Resting Nucleic Acid 225

E. Radiations 240

F. Unclassified Mutagens 245

G. Spontaneous Mutations 245

IV. The Extent of Point Mutations 248

A. Genetic Mapping of Mutations 248

B. Spontaneous and Induced Reverse Mutations 249

C. The Kinetics of Mutation Induction 250

V. Mutagenic Specificity 252

A. The Specificity of Forward Mutations 252

B. The Specificity of Reverse Mutations 253

References 264

I. Introduction^

A. CLASSIFICATION OF HERIT.\BLE VARIATIONS

The concept of mutation has played an important role ever since the large variety of biological species had been attributed to evolution from

^Present address: Laboratory of Molecular Biology, National Institutes of Health— NINDB, Bethesda, Maryland.

'Abbreviations used throughout this paper: A, adenine; C, cytosine; HMC, 5-hydrox\'methylcytosine ; MeC, 5-methylcytosine ; G, guanine; T, thymine; H, hypoxanthine ; U, uracil; X, xanthine; G-C, guanine-cytosine pair with anj^ of the

207

208 ERNST FREESE

a coiiuiion ()i'i<!;in. A new species was thought to conic about by the gradual change of an old species. Laniaick (1809) atti'ibutcd this change to both the crossing of existing fofnis and the inheritance of acciuircd ciiaracters, to which tiie organism had achipted during its Ufetinie. Dar- win (1859), however, explained evolution by natural selection and reasoned that new forms arise by chance irresjiective of whether they are useful, useless, or deleterious to the organism. But convincing evi- dence, showing that ac(iuired (adapted) characters are not indefinitely inherited, accumulated only slowly; for example, the spontaneous nature of mutations in bacteria has been proved only recently (Luria and Dcl- briick, 1943). Phenotypic variations of similar organisms could then be divided into those which are heritable and those which are non-heritable and for some time this proved the only experimentally verifiable distinc- tion. During this period, the w^ord "mutation" was synonymous with all "heritable variations."

A further classification of heritable variations became feasible only when Mendel's laws were rediscovered at the beginning of the twentieth centuiy and when many crosses w^ere examined genetically and cyto- logically (see scheme in Fig. 1). One learned to distinguish nuclear from extranuclear (cytoplasmic) inheritance and their variations. The herit- able nuclear variations were found to involve the alteration of the chromosomes, either their number in a nucleus (polyploids, aneuploids, etc.) or their information content. Alterations of the information content could be further divided into those that occur through a recombination of information already present in two pieces of genetic material and those that occur by a mutation of the information within one chromo- some and do not require the interaction of genetic material. This present meaning of the word "mutation" is much more restricted than the original one.

The bipartition into nuclear and extranuclear variations, shown in Fig. 1, is reasonable for organisms having a defined nucleus, and it is useful, because the nucleus and the cytoplasm can often be separated by natural or artificial means. But as is common to all classifications one encounters difficulties. For certain microorganisms, e.g., bacteria, a nuclear membrane may not exist and thus a division into nuclear and extranuclear variations may be without meaning. Nevertheless, both bacteria and viruses contain most of their hereditaiy information in one- dimensional genetic structures (genomes) which must correspond to physical entities. The existence of such a physical entity containing

cytcsines; dCMP, dcoxycytidinc monophosphate; AP, 2-aniinopiirinp; BU, 5- bromouracil; BUdR, 5-bromodcoxyuridine; EES, ethyl ethanesulfonatp; HA, hy- droxylamine; NA, nitrous acid.

V. MOLECULAR MECHANISM OF MUTATIONS

209

nucleic acids has been proved for several microorganisms. We shall name this structure a chromosome in all those cases in which it contains DNA, even though the DNA may not be visible by Feulgen staining. One can

Extranuclear (or extrachromosomal) = cytoplasmic

Chromosome

number

( polyploids aneuploids )

Information content

Recombination

( includes transformation, transduction, etc.)

Mutation

Larger alteration

Fig. 1. Classification of heritable variations.

then distinguish chromosomal and extrachromosomal variations; for bacteria the latter are often called cytoplasmic variations. Since bipar- tition of chromosomal variations into mutation and recombination re- mains the same as above, the subsequent discussion about mutations refers to both macro- and microorganisms.

B. DEVELOPMENT OF GENETIC SYSTEMS AND DISCOVERY OF THE FIRST MUT.AGENS

The early genetic experiments were performed on higher organisms such as garden peas (Mendel, 1865) and Oenothera (De Vries, 1909).

210 ERNST FREESE

De Vries discovered tlir (ir.<t spontaneous "mutation," wliirli \va!< later shown to be a rcciproeai translocation. These higher organisms were useful for discovering new genetic phenomena and for investigating, both eytologieally and genetically, gross changes of the genetic material. With the desire to examine in detail the smaller chromosomal variations it became necessary to employ simpler genetic systems of organisms possessing a shorter generation time, thus enabling the experimenter to grow large populations and to statistically analyze genetic traits over many generations. The utilization of these systems acc{)mj)anied the advancement of specific knowledge about recombination and mutation.

An important step in this direction was made by developing the genetics of the fruit fly, Drosophila. This genetic test system has been so refined that the frecpiency of x-ray-induced lethal mutations could be measured (]\Iuller, 1927), i.e., changes in the X-chromosomc that are lethal to the progeny receiving this chromosome. Radiation-induced nuitations have also been observed in barley by Stadler (1928). Since then all kinds of radiations have been examined.

The first chemical mutagens were found only much later; for exam- ple, urethan + KCl which induces chromosomal breaks in Oenothera (Oehlkers, 1943), nitrogen and sulfur mustards CAuerbach and Robson, 1946), formaldehyde (Rapoport, 1946), and diethylsulfate (Rapoport, 1947a,b) which induce mutations in Drosophila. Since then many weak and a few strong chemical mutagens have been found.

The next step toward simpler genetic systems was the use of micro- organisms, e.g., Neurospora, Aspergillus, yeast, and bacteria, which could be grown on a defined medium and for which biochemical mutants could be isolated, i.e., organisms that need the addition of a certain nutrient to the minimal medium. For each of these systems the induction of mutations both by radiations and chemicals has been examined. If the functional difference between the mutated and the unmutated organ- ism is known one can distinguish forward mutations, which result in the loss of a biochemical function, and reverse mutations, which restore this function. Reverse mutations usually occur much less frequently than forward mutations but their frequency can be measured more easily, because omission from the medium of the extra nutrient, necessary for the growth of the mutant, selects for the revcrtant type. This selective procedure enables one to determine the frequency of revertants even in diploid cells for which the measurement of forward mutations is difficult. [For a review of recent methods and results for bacteria see Ryan (1961).]

A most important step toward an understanding of mutagenic mech- anisms was the development of bacteriophage genetics. Since Hershey

V. MOLECULAR MECHANISM OF MUTATIONS 211

and Chase (1952) had shown that phage T2 injects essentially only its DNA into the bactci-iiim it was possible to correlate genetic observations directly with alterations of the bacteriophage DNA. Other viruses have also played a significant role in this development but the most carefully analyzed tobacco mosaic virus has so far resisted all attempts to effect a genetic analysis, i.e., to obtain recombination between two mutants. Another excellent .system for mutagenic studies should be that of transforming principle (i.e., DNA). DNA could be treated by powerful reagents which cannot be used for phages or cells since they also denature protein ; some of these are so specific that only one of the four bases is attacked. So far no genetic region has been examined in sufficient detail to permit a simple analysis of induced mutations both in the for- ward and the reverse direction.

C. THE GENETIC SYSTEM OF /'// MUTANTS OF T2 OR T4 PHAGES

One particular phage system, the rll mutants of phage T4, plays an important role in the following discussion and will therefore be described in more detail. This system has been developed by Benzer (1955) and by Chase and Doermann (1958). The phages are assayed by plating them together with about 2 X 10^ sensitive bacteria, Escherichia coli B, (abbreviated B), on a nutrient agar plate. Each phage that infects and develops in a bacterium produces about 200 new phages which are re- leased by lysis of the bacterium. These phages in turn infect more bacteria and thus a chain reaction produces a halo or "plaque" in the turbid bacterial meadow that is formed after several hours of incubation. Phages of "standard type" produce small plaques with fuzzy edges while rll mutant phages form large plaques with sharp edges. If one phage gives rise to both standard and mutant progeny a "mottled plaque" is formed and one thus has a method to detect forward mutations. Reverse mutations can be selectively measured in a different bacterium, E. coli K12 (A) (abbreviated K), on which rll mutants cannot grow while standard type mutants or revertants can. The plaque types of the phages on the two bacterial strains are summarized in the diagram.

Bacteria

K

p, \ r-^ = Standard type Standard Standard

rJI type

Chase and Doermann (1958) have used this selective technique to measure the recombination frequencies between several rll mutants and

212 ERNST FREESE

thu^ to dctciininc in two- and tlircc-factor crosses the relative distance and order of the mutations in tlic /•// i'c<j;i()n of the phage genome (cliromosome). Benzer (1961) lias been able to niaj) a much larger num- ber of mutants by using a set of rll mutants which contain a large alteration within the rll region. Such a set is shown in Fig. 2. Each rll

ht cM hr in rll Iv rl h m M M la In M

" Ml ■! I — I ■ —Mnr-i I I I I

■ Slondord

I^ZZZ r 1272 BZZZH r 1241 r 4 3 r PTI r PB242 r A 105 r 638

S««in«n( Al A2 A3 A4 A3 A6 B

Fig. 2. At the top, the rll region is shown compared with the entire genetic map of the phage. This map is a composite of markers mapped in T4 and the related i:)hage T2. Seven segments of the rll region are defined by a set of "dele- tions" beginning at different points and extending to the right band (and possiblj' beyond, as indicated by shading). (From Benzer, 1961, p. 404.)

mutation can thus be rapidly located in one or more of the segments demarcated by the alteration, if the mutants are crossed, by a simple spot test, against each mutant in the set. This ingenious trick reduces the amount of labor tremendously. A genetic map of spontaneous mutants is shown in Fig. 3. The same method has been used to determine rapidly the relative location of a large number of mutants induced by different means and to decide which of them could recombine with each other and which could not. Results of such examinations will be described in this paper.

D. FURTHER SUBDIVISION OF MUTATIONS

In the attempt to subdivide mutations further, "chromosomal muta- tions" and "gene mutations" have been defined. But the word "gene" referred here to the older concept of a gene being at the same time the functional, recombinational, and mutational unit. Recent examinations of the genetic fine structure have proven that a functional unit contains many mutational units and can l^o finely subdivided by recombination;

V. MOLECULAR MECHANISM OF MUTATIONS

213

214 ERNST FREESE

tlic best illu>tr;iti()n of this fart is Fi<i;. 3. One may thus idciitil'v the word "jione" with only one concept, and best with a functional unit (even that requires certain qualifications). "Gene mutations," then, ai'e muta- tions that alter only one functional unit while "chromosomal mutations" affect several functional units at the same time. In order to define a functional unit ojierationally one can use the "cis-tmns" test by which mutations that cannot cross-feed are attributed to one "cistron" (Benzer, 19571. For this test two homologous chromosomes, each cairying one mutation (/ra/?s-position), are i)laced in a conniion cytoplasm; when this combination does not show the nonnal functional activity the two mutations are said to belong to the same cistron. The corresponding cis test, in which both mutations are on one chromosome while the homol- ogous chromosome is non-mutant, serves to establish that normal func- tional activity is obtained when one chromosome is not mutated. Sometimes the cis-tmns test is not sufficient to demarcate the boundaries of a functional unit, for example, when cross-feeding is not possible because the two homologous chromosomes are too distant. Another complication is illustrated by the case in which it is not possible to decide by qualitative growth tests whether the functional activity of the hetero- caryon is normal or subnormal and caused by another repair mechanism, i.e., "complementation" (see Demerec and Hartmann, 1959). In these cases it is necessaiy to measure quantitatively the functional (usually enzymatic) activity. Nevertheless, a suitably refined cis-trans test pro- vides, in most cases, a practical definition of a functional unit (e.g., the genetic material which determines the structure of a peptide).

As another subdivision of mutations which is useful for detailed genetic experiments, one can distinguish "large alterations" and "point nmtations." A mutation is a large alteration if the corresponding nuitant cannot recombine, and produce the standard type, with at least two other mutants which in turn are able to recombine with each other. A chromosomal mutation is a large alteration; but also a gene mutation can be a large alteration, for examjilc, if a larger piece of a gene (DNA) is missing. In contrast, one would like to define a point mutation as one in which only one nucleotide pair has been altered (including deletion or addition). Since it is impossible to prove at the presc^nt time that a given mutant has undergone such a small change the following oi^cra- tional definition will be used: A 'point mutation is one which does not give evidence of a larger alteration in the above-mentioned crosses with as many independently isolated mutants of the same phenotype as are available, and usually it does revert spontaneously or by induction. Obviously the precise definition of the "point" depends both on the number atul kinrl of mutants ci-ossed and on the smallest distance on

V. MOLECULAR MECHANISM OF MUTATIONS 215

the genetic map for which recombinants are still detectable experi- mentally or possible molecularly. The unit of recombination may be as small as two adjacent nucleotides, since Helinski and Yanofsky (1961) have found two very closely linked but recombining mutations of the tryptophan synthetase region for which the same amino acid is changed in the corresponding protein. The condition of reversibility is introduced in the above definition since practically all mutants of small genetic extent do revert and since one imagines that the change of a single nucleotide pair can revert occasionally. Yet, the reversion frequency of some mutations (e.g., deletions, insertions) may be unobservably small in the genetic system used. The lack of reversion does not prove that more than one nucleotide pair has been altered. In fact, some rll mutants were thought to be irreversible until their reversion was observed by induction. Conversely, a mutation should not be called a point mutation merely because it reverts. Unless one knows that it has been induced by a chemical which induces nothing but point mutations in other genetically well-studied systems a strong argument cannot be made, for in some cases the revertants may be the result of suppressor mutations (see Section V,B).

II. Structure, Duplication, and Alteration of THE Hereditary Material

We wish to determine which molecular groups of the hereditary material are changed in a mutation and how these changes come about. But before we can discuss molecular changes we need to know the molecular structure of the hereditary material itself, i.e., the molecular structure of the chromosome and the mode of its replication.

A. structure and replication of DNA AND RNA

The molecular structure of DNA, as proposed by Watson and Crick (1953), is now well established, thanks to the careful studies by Wilkins (1956), Spencer (1959), and Miles (1961). Watson and Crick (1953) also proposed that DNA duplicates by the separation of the two com- plementary strands and the formation of an exact new complement along each parental strand. This hypothesis is supported by studies with density labeling (Meselson and Stahl, 1958), with autoradiography (see Taylor, Chapter II), and with DNA synthesis in cell-free systems (see Bessman, Chapter I). I shall regard this mode of DNA duplication as an established theory.

Single-stranded, information-carrying DNA and RNA probably repli- cate also by such a complementary base-pairing mechanism. Hence most

216

ERNST FREESE

of the following considerations should also apply to mutations that occur in such systems.

B. STRUCTURE AND DUPLICATION OF CHROMOSOMES

Since tlie hereditaiy information of chromosomes is presumably car- rii'd by DNA it would be desirable to know how the DNA is arranged in a chromosome. One possibility would be to assume that a chromosome contains just one long DNA molecule. But as the DNA apparently has to unwind during its duplication it seems more likely that occasionally one of the two DNA strands is (or becomes) interrupted in order to provide a swivel. The other DNA strand at the swivel site could be either continuous or interrupted by some non-DNA material. Details about the structure and duplication of chromosomes can be found in the article by Taylor (Chapter II). In view of the present lack of knowledge about the precise chromosomal structure it seems wiser to refrain from trying to explain large chromosomal alterations in molec- ular terms and to limit ourselves to the small changes that are limited to a small region of DNA.

C. CLASSIFICATION OF BASE ALTERATIONS IN NUCLEIC ACIDS

All possible kinds of base alterations in DNA are outlined in Fig. 4. The corresponding changes of single-stranded nucleic acid can be

	• •	• •	original DNA
12 3^5678	9	10
• ••••O**	• f	• •	replacement
4 5 6 7 I
	• •		deletion
4 5 7
• ••••0««	• f	• i • «	insertion >
4 5 6 7
1 1 inversion

12 3

7 6 5 4 ! 8 0 10

Fig. 4. Classification of tlie base pair changes in DNA. Each circle (#) repre- sents a base. A hollow circle (O) is a base dilTerent from that in the original standard type DNA.

V. MOLECULAR MECHANISM OF MUTATIONS

217

obtained by covering one of the two strands. Each of the classes of base alteration could be further subdivided, because mutagens may prefer- entially attack one or the other of the bases or base pairs. So far only the specificity of agents inducing replacements is sufficiently understood to warrant such a subdivision; for reasons that will become clear later it is useful to distinguish "transitions" and "transversions" (Freese, 1959b) as shown in the tabulation.

Original base

Ultimate base

Name of base change

Purine Pyrimidine Purine Pyrimidine

Other purine Other pyrimidine Any pjTimidine Any purine

Transition Transition Transversion Transversion

For double-stranded DNA the possible base ])air ciianges are the following:

A^ T^	"^ C	= transitions
A	C
T	G
:	I	= transversions
T^ A ^	^G

In order to find out how many and which nucleotide pairs of DNA have been altered in a mutation one would like ideally to make a sequence analysis of both normal and mutant DNA or at least to determine chemically that portion of DNA which has been changed. At present this is not possible; instead we have to proceed indirectly by comparing genetic and chemical observations on the effects of different mutagens on DNA.

III. Biochemistry of Mutagens and Theory of Base Alterations

Apart from radiations there are many chemicals that show some mutagenic effect. It would be impossible to mention all these agents. Rather, I shall point out various categories and elaborate in more detail the mutagenic action of those agents whose chemical effects are better understood.

In principle one can distinguish agents that affect only replicating nucleic acid and others that exert their primaiy action on resting nucleic acid, although the ultimate fixation of the mutations may also require

' See list of abbreviations in footnote 2.

218

ERNST FREESE

leplication. Most agents have been tet^ted only on cells or organisms and it is not clear whether they affect DNA directly or only indirectly, e.g., by alteration of a nucleic acid jirecursor which thus becomes mutagenic.

A. IMIIBITION OF NUCLEIC ACID PRFX'URSORS

There are many agents that interfere with the normal synthesis of nucleic acid precursors (for review see Handschumacher and Welch, 1960) ; in addition, a number of mutants have been isolated that are blocked in one of the pathways toward nucleic acids. Some of the agents or mutants interfere with the synthesis of purines, others of pyrimidines, and again others are more specific and affect only the synthesis of one base, e.g., thymine, adenine, or guanine, sometimes only at the nucleo- tide level. Some of the inhibitors of the nucleic acid pathways that have shown a significant mutagenic effect are smnmarized in Table I. The

Mutagenic Agents and

TABLE! Mutations Which Interfere with the Normal Formation of Nucleic Acid Precursors

Agent or mutant	Inhibits		Mutagenic effect observed in	Reference
5-Aminouracil	Thymine	Escherichia coli E. coli, biochemistry	Greer, 1958 Handschumacher and Welch, 1960
Azaserine	Purines	E. E.	coli coli, biochemistry	Iyer and Szybalski, 1958, 1959 Handschumacher and Welch, 1960
Benzimidazole	Purines	E.	coli	Novick, 1956 Greer, 1958
Caffeine	Purines	E. coli E. coli, biochemistry E. coli Drosophila	Novick, 1956 Koch and Lamond, 1956 Glass and Novick, 1959 Andrew, 1959
8-Ethoxycaffeine	Thymine	Allium Vicia faba	Kihiman, 1955 Levan, I95I
Ethyl urethan	Purines	Oenothera Drosophila	Oehlkers, 1949, 1953 Vogt, 1950
6-Mercaptopurine	Purines	E.	coli	Greer, 1958
5-Nitroquinoxaline	Purines	E.	coli	Greer, 1958
Paraxanthine	Purines	E.	coli	Novick, 1956
Tetramethyl uric acid	I>urines	E.	coli	Novick, 1956
Theobromine	Purines	E.	coli	Novick, 1956
Theophylline	Purines	E.	coli	Novick, 1956
Thymine deficiency	-	E.	coli	Coughlln and Adelberg, 1956 Kanazir, 1958
Uracil in excess	-	E. coli (chromosome breaks)	Deysson, 1932

V. molp:cular mechanism of mutations 219

strongest mutagen in this table is azaserine wliich inhibits purine syn- thesis (see Handschumacher and Welch, 1960). But this chemical is also a very labile alkylating agent; it may therefore owe its mutagenicity to its alkylating ability rather than to its inhibitory action. Similarly urethan is both an inhibitor of pyrimidine synthesis and a mild alkylat- ing agent. But in this case Rogers (1957) has shown that in mice thymine inhibits the production of chromosome breaks by urethan. All other growth inhibitors are rather weak mutagens in bacteria and induce chromosome breaks in higher organism. The most careful quantitative experiments about the mutation induction have been done by Novick and Szillard (see Novick, 1956). These authors have also shown that even adenine increases the frequency of mutations while the normal purine nucleosides have an anti-mutagenic effect. These observations suggest that the frequency of mutations increases because the lack of one base either causes a chromosome break, and thus a large alteration, or it increases the frequency of pairing mistakes, through which one normal nucleic acid base is incorporated in the place of another one. However, it is just as likely that the inhibitor causes the increased formation of another base analog which is the actual mutagen, perhaps by being incorporated into the nucleic acids in place of the normal bases. ]Many such natural base analogs have been observed in RNA [although mostly for soluble RNA, e.g., Littlefield and Dunn (1958)] and con- ditions of thymine starvation cause the accumulation of A"-methyl- adenine in DNA (Dunn and Smith, 1955). In addition, a naturally produced mutator factor has been genetically proved to exist in a certain strain of Salmonella typhi murium (Miyake, 1959).

B. IXCORPORATION OF BASE ANALOGS

Apart from naturally occurring base analogs some artificial analogs are incorporated into RNA and into DNA. Here we are only concerned with those analogs that have shown a strong mutagenic effect. 5-Bromo- uracil (BU), 5-chlorouracil, and 5-iodouracil can replace thymine in DNA (bacteria: Weygand, Wacker, and Dellweg, 1952; Zamenhof and Griboff, 1954; phages: Dunn and Smith, 1954; Litman and Pardee, 1960; human cell lines: Szybalski and Djordjevic, 1959; Hakala, 1959; Eidin- off et al, 1959), while 2-aminopurine (AP) (Wacker et al, 1960a, Gott- schling and Freese, 1961; Rudner, 1961) is incorporated to such a small extent that it has not been possible to determine chemically which base it replaces. 2,6-Diaminopurine is also highly mutagenic (Freese, 1959a; Benzer, 1961) but its effect has not been studied in great detail. These base analogs are much less inhibitory for bacterial and phage growth than other purine or pyrimidine analogs ; yet they are much more muta-

220 ERNST FREESE

genie. For thir> reason tlicii- iiiuta<icnic cfft'el lias been alti'ihutcd to base pairing mistakes that occur as a cons('(|uence of tlicii- incorpoi'ation into DXA (Freese, 1959a).

1. 5-Bnymourncil (/^f'l or o-Hroinodcoxj/uriflinc (Bl'dR)

Figure 5A illustrates that BU can pair with adenine (A) as well as thymine (T) can. When T is lacking, addition of BU (or better, BUdR) to the growth medium leads to the (luantitative replacement of T by BU in DNA. In most cases extensive death has been reported under such

H ^ N >-H O Br

\9/56\ /45\

N— C4 iN H— N3 6C — H

/ \3 2/ \2 1/

— C N-C C— N

\ \ <i )c-

Adenine 5-Bromouracil

(normal amino state) (normal keto state)

(A)

\ p-^ p-\

N-C N— H N C— H

/ \ / \ /

— C N— C C— N

\ \ / \

N-H O C-

i '

Guanine 5-Bromouracil

(normal amino state) (rare enol state)

(B)

Fig. 5. Base pairing of 5-bromoiiiacil (BU). A. Normal base pair of BU with A. B. Rare ba.se pair of BU with G.

growth conditions, but when phages (Stahl, 1961) or transforming prin- ciple (Szybalski et al, 1960) are kept in the absence of light veiy little death occurs. In other organisms BU may have side effects on vital cell constituents different from DNA. Occasionally a base pair change in DNA is induced when BU loses its hydrogen atom in the 1 -position and pairs with guanine (G) instead of A (Fig. 5B). This ])airing "mistake"

V. MOLECULAR MECHANISM OF MUTATIONS

221

may occur either when BU enters DNA, as a "mistake in incorporation" (Fig. 6A), or after it has been incorporated into it, as a "mistake in replication" (Fig. 6B). Hence, both kinds of base pair transitions, G-C into A-T and vice versa, should be inducible with this agent. Although BU-induced nuitations apparently are due to the Bl' incoi'poration, the

B

Fig. 6. The induction of base pair transitions by o-bromouracil. A. Mistake in incorporation. During DNA duplication dBUTP rarely may pair "by mistake" with a G present in DXA and the dBUMP is incorporated into the new DNA strand. This initiates a base pair change since in the next DNA duplication dATP pairs with this BU, A being the "normal" complementary base of BU, and dAMP is incorpo- rated into the new strand. Thus eventually an A-T pair appears in some of the progeny DNA in place of the original G-C pair. B. Mistake in replication. BU first is incorporated into DNA by pairing with its normal complementar.y base A. This increases the propensity of this base pair to mutate since BU undergoes pairing mistakes more frequently than T. In one of the subsequent DNA duplications dGTP may pair with BU and the incorporation of dGMP into the new DNA strand leads to the eventual appearance of a G-C pair in place of the original A-T pair.

above scheme shows that the frequency of forward mutations is not necessarily proportional to the amount of BU incorporated (Litman and Pardee, 1956) since the ratio of mistakes in incorporation to mistakes in replication may vaiy. For the induction of reverse nuitations in bacteria such a proportionality has been shown (Strelzoff, 1962i.

The higher frecjuency of pairing mistakes which BU apparently undergoes as compared to T is probably due to the higher electronega- tivity of Br as compared to that of the methyl group in T. The pyrim- idine ring gets poorer in electrons and apparently can lose the hydrogen in the 3-position more easily than T, thus giving rise to the ionized or tautomeric form of BU. This suggests the possibility that BU induces

222 ERNST FREESE

mistakes in incoipoi'atioii (( l-C — * A-Ti inoic ficcniciitly than mistakes in replication since in the lattci' ease BU is already a inemhei' of a DNA strand and the electi'onetiativity of its bromine may become |)aitially reduced by electron-donatino; nei^lil)or <2;i-oups.

BUdR is a more etiicicnt mutaiicn than HV (Freese, 1959a; I.itmaii and Pank^e, 1960) probably because it is more readily converted into the deoxynucleotide triphosphate and interferc^s less with the formation of ni'acii and cytosine. Both BU and Bl'dlv induce point nuitations (Benzer and Freese, 1958; Frccse, 1959a).

The jiroposed mechanism of nuitagenesis is supported by several other obser\-ations. "NMien ])liap:e infected bacteria are exposed to BUdR and diluted into a medium containint^ excess thymidine, befoi'e any intact phages are formed, indueed mutations are still observed in the lysate (Litman and I'aidee, 1960). The same is observed wlien jirotein synthesis is stopped by chloramphenicol and BUdR is present only while the phage DNA multiplies (Litman and Pardee, 1959; Brenner and Smith, 1959). This shows that the induction of mutations by BUdR does not reciuire some fixation process and does not depend on protein syn- thesis. The observation of mottled plaques after BUdR mutation induc- tion shows that only one base of a DNx\ pair is replaced at a time and that DNA apparently duplicates by the separation of the complementary strands (Pratt and Stent, 1959), as has been assumed.

When DNA contains BU, its biological activity is much more sensi- tive to chemical attack. This has been observed by its sensitivity to UV (Litman and Pardee, 1960; Kozinski and Szybalski, 1959; Greer, 1960; Lorkiewicz and Szybalski, 1960), to UV and visible light (Stahl, 1961), and to X-rays (Djordjevic and Szybalski, 1960; Kaplan and Tomlin, 1960). This additional effect of BU incorporation is apparently only lethal and not mutagenic. The same applies to the increased lethality of BU-containing jihages wlien treated by hydroxylamine (Freese et nl., 1961b).

2. 2-Aminopnrine (AP)

AP is mutagenic, apparently by a quite different mechanism. It can pair in its normal tautomeric form with two bases, with T by two hydro- gen bonds and with C by one hydrogen bond (Freese, 1959a) (see Fig. 7). The pairing involving two hydrogen bonds should occur much more frequently tlian that with one liydrogen bond since in the latter case the two protonless nitrogens repel each other by their electron clouds and their relative steric ])osition is not fixed, making it more difficult for the polymerizing enzyme to link the nucleotide to the growing DNA chain before it has s('parate(j again. Thus AP should l)e incorporated

V. MOLECULAR MECHANISM OF MUTATIONS 223

most frequently in the place of A which in itself would not yet be mutagenic. But whenever AP pairs with C, either in its normal tauto- meric state or by two hydrogen bonds after it has undergone a tautomeric shift, a base pair cliange results. The frecjuency of the pairing "mis- takes" between AP and C, per base incorporated, should be much larger

N-C N H-N C-H

/ \ / \ /

/ N— C C— N

N-H O ^

/ H

2-Aminopurine Thymine

(A)

H

H\ N H H— N H

C \ / \ /

N— C N N C— H

/ \ / \ /

N— C C— N

N— H O ^

/ H

2-Aminopurine Cytosine

(B) Fig. 7. Normal (A) and rare (B) base pairing of 2-aminopurine.

than between BU and G ; this may explain why, in spite of its very small incorporation, AP is highly mutagenic. Analogous to the mechanism explained for BU in Fig. 6, AP is expected to induce base pair transitions in both directions. It does induce point mutations (Freese, 1959b; Demerec, 1960).

In bacteria, kinetic studies on the induction of reverse mutations by BD and AP, have been used in the attempt to distinguish the direction of the base pair transitions, by the time and pattern of mutant increase after a pulse of the mutagen (Rudner, 1960, 1961; Strelzoff, 1962). The authors observed that for some biochemical mutants the frequency of revertants increases at some time after the pulse, within one generation, and then remains constant in later generations: this indicates that the reverse mutation was induced by a mistake in incorporation. In contrast,

224

ERNST FREESE

other iiuilaiits willi tlii' .^aiuc hiocliciiiical rccniiivnu'iit sliowed a con- tinued increatsc of revertants throughout several generations, indicating that the reverse mutation was induced by a mistake in replication. P'or a given mutant the time at which the first induced mutants showed up was difterent for AP tiian for BI), in the case of Sabnonella, while it was the same in the case of A', coli; this may be caused by a difference in the methods used. But the o\er-all pattern of either cessation or continued increase of mutation induction was, for a given mutant, the same with both mutagens, in agreement with the expectations. The precise timing, however, of the appearance of revertants could also depend on the func- tional activity of the DNA strand in wliicli the first pairing mistake occurs.

C. DYES

Man}' different basic dyes have been used for the cytological staining of nucleic acids. Some of them, whose structures are shown in Fig. 8, have also been tested for their effect on living oi'ganisnis; proflavin and

H,N

Proflavin

NH,

HoCv

N I H,C

Acridine orange

N

I

CH

CH,

HoCv

I

N' I CH,

^CH.j

HoCn

N I H,C

CH,

NH,

Methylene blue Toluidine blue

Fig. 8. Structure of dyes often rmi)loyed in biological experiments.

acridine orange are two examples of "acridine dyes." Undoubtedly many other compounds with structures similar to those in Fig. 1 have biological effects. For exam])le, the similarity to some aromatic carcinogens is noteworthy.

A direct attachment of these dyes to nucleic acids is the probable cause of mutations. However, there are some other biological effects wliich will be mentioned briefly. The i)resence of acriflavin in the growth mediinn of various organisms eliminates ncjn-chromosomal elements, e.g.,

V. MOLECULAR MECHANISM OF MUTATIONS 225

the pi factor in E. coli (Hirota and Tijima, 1957; Hirota, 1960). Pro- flavin prevents the assembly of phage T4 (De Mars et al., 1953) at concentrations at which the uninfected bacterium can still duplicate and even the phage DNA can still multiply. Certain dyes (methylene blue, toluidene blue, and proflavin) when exposed to light have a strong lethal effect on phages, provided they are inside the phage (Clifton, 1931; Yamamoto, 1956, 1958; Kaufman and Hiatt, 1959; Hiatt, 1960). This "photodynamic effect" presumably is due to the oxidation of some DNA bases, mediated by the dye and stimulated by light; for the presence of oxygen increases the effect and the presence of reducing agents (sulf- hydryl groups) reduces it.

A mutagenic effect so far has been reported only for acridine dyes. Acrifiavin (a mixture of proflavin and 2,8-diamino-lO-methylacridinium chloride) is mutagenic for bacteria (Witkin, 1947), and this as well as other dyes are mutagenic for Allium (Bauch, 1948; D'Amato, 1950). Proflavin is mutagenic for phages; when it is removed at some time after phage infection normal phages are produced but an abnormally high frequency of mutants is found (De Mars, 1953; Brenner et al., 1958). Many other acridine dyes similarly show a mutagenic effect on phages (Orgel and Brenner, 1961).

The molecular mechanism by which acridine dyes induce mutations is still unknown. Acridine orange and probably other acridines can stack along nucleic acids. They apparently attach to the phosphate backbone (Beers et al, 1958; Bradley and Wolf, 1959) except for polyadenylic acid for which some attachment may involve the adenine bases (Steiner and Beers, 1958). Lerman (1961) has recently found that proflavine greatly increases the viscosity of DNA; from this and X-ray data he concludes that acridines can be sandwiched between two purine bases and thus cause some binding of the dye at concentrations at which the stacking is still unimportant. With these data at hand and a particular coding model in mind, Brenner et al. (1961) have proposed that acridines induce only deletions or insertions of one base pair in DNA.

D. CHEMICAL ALTERATION OF RESTING NUCLEIC ACID

1. Nitrous Acid

According to Schuster (1960) nitrous acid deaminates, with decreas- ing frequency, the bases G, C, and A in DNA and according to Schuster and Schramm (1958) with about equal frequencies in RNA. Its muta- genic effect was first analyzed for tobacco mosaic virus and its RNA by Mundry and Gierer (1958), for bacteria (Kaudewitz, 1959), phage T2 (Vielmetter and Wieder, 1959), and many other DNA- and RNA-

226

ERNST FREESE

containing systems, including transfomiing principle (Litman and Ephi-ussi-Taylor, 1959).

Nitrous acid induces point mutations in T4 phages (Freese, 1959c; Benzer, 1961) and in yeast, i.e., Schizosaccharomyces pombe (Gutz, 1961).

The effect of the deaminations on the pairing properties of the differ- ent bases is depicted in Fig. 9. One can conclude (Freese, 1959c) that

^\

\

/

.c-c

\ / N— C

/ \

— C N- C

\ ^

N— H-

/

H

Hypoxanthine

H

H. ^x N— H-

C ^ / \ C-C \ / \

N-C x\

/ \ / — C N— C

H

/ -N H

\ /

-N C— H

/

Cytosine

(a)

— O H

■H— N C— H

\ / C — N

/ \

o c—

/

Adenine

Uracil

H

II H ^ C

C ^^ G

III A C

T i

	0 1 III '
G G / III <r c u ^	^ C A II U ^ A
	II U

(b)

«x.

N—

■ N O

\ /

c— c

/ \

— c

C N— H

Vc'

o

Xanthine

H

/ H-N H

\ /

N C— H

\ / /C-N

o c

/

Cytosine

(c)

Fig. 9. Base pairing and base pair changes of deaminated bases. A is deaminated to hypoxanthine (H) which now pairs with C in place of T. C is deaminated to uracil (U) which now pairs with A in place of G. G is deaminated to xanthine (X) which still pairs with C, but by one less hydrogen bond.

V. MOLECULAR MECHANISM OF MUTATIONS 227

nitrous acid should induce only, or predominantly, base "transitions" and that the deamination of G should not be mutagenic since the resulting base xanthine could still pair with cytosine. To prove this Vielmetter and Schuster (19601 compared at different pH's the rates of the deamination of the three DXA bases with the rate, per time unit, of mutation induc- tion. They found that the deamination rates of A and C decrease with increasing pH as rapidly as the mutagenic rate while that of G decreases about one-third as fast. This shows that the deamination of G cannot be mutagenic. Since the lethal rate of nitrous acid treatment also de- creases slowly with increasing pH these authors propose that the deam- ination of G is lethal. This conclusion is not compelling since nitrous acid exerts side reactions in addition to deamination; one can see this when dC]MP is treated with nitrous acid in vitro and then chromato- graphed. Furthermore, Geiduschek (19611 has shown that nitrous acid causes cross-linking of DXA which would be almost certainly lethal for phages. The pH dependence of all these side reactions has not yet been measured. Another side reaction is the lethal effect of nitrous acid on protein; for T2 or T4 phages this is a relatively small effect, since only about 1 out of 10 lethal hits (for definition see Section IV,C) prevents the injection of phage DXA (Harm, 1960; E. B. Freese and E. Freese, 1961).

The induction of mottled plaques after nitrous acid treatment of standard type phage T2 and of non-mottled mutant plaques after treat- ment of phage </)X174 has been used as a further proof that T2 has double-stranded and <^X174 has single-stranded DX'A (Tessman. 1959).

2. Hydroxylamine (HA) and Hydrazine

Both hydroxylamine (HA) and hydrazine react with certain pyrim- idine bases but in spite of their chemical similarity show very different chemical specificitj'.

The effect of hydrazine on uracil has been long known (Fosse et al., 1924) ; it has been examined again for UMP and also for C]\IP by Baron and Brown (1955). Hydrazine breaks the ring of uracil and cytosine, giving rise to pyrazolone and 3-aminopyrazole respectively, while the remaining urea remains attached to the sugar and in the presence of water is subsequently hydrolyzed off. Treatment of RX"A by anhydrous hydrazine produces "ribo-apyrimidinic acid" free of pyrimidines (Take- mura, 1957) and the treatment of DX'A produces the corresponding "apyrimidinic acid" (Takemura. 1959). The author finds a small amount of thymine remaining on this DX'A. In contrast, treatment of the nucleo- tides in a 1.0 M aqueous solution of pH 8.5 gives a reaction (decrease of UV absorption at 276 m^n) with dT:\IP. U:\IP. and BUdR while

228 ERNST FREESE

dCMP or (leoxyliydioxynu'tliylcvtidylatc iilII.MPi sliuw no effect (Freese et ai. 19611)1. It should be noted tliat the ])K value of hydrazine is about 8.4; the j)!! apparently influences its chemical specificity api)recial)ly.

The chenucal effect of hi/dro.vylamine (HA) on the i)yriniidine bases has been observed only recently. The strongest effect (measured by the decrease of UV absorption at 260 or 276 mp.) of a 1 M solution at pH 7.5 is on C. (MPv. dCMP, and U, UR, UMP; HMC or dHMP .^liows less effect, while methyl-C, T, and dTAlP react too little to give any signifi- cant change of UV absorption (Freese et al., 1961a,b). The diffeience between the methylated and unniethylated uracil and cytosinc may be due to the sliglit electron-donating ability of the methyl grou]) or to steric effects. 5-Bromouracil or its deoxyriboside reacts very rapidly owing to the electronegative bromine (Freese et al., 1961b). Again the |)H controls the chemical specificity (the pK of HA is 6.2) such that at 1)H 6.15 the reactivity of CMP/U.MP in RNA is 30 and at pH 9.15 the ratio is 1/8 (Schuster, 1961).

The reaction of HA with U or UMP causes the irreversible loss of UV absorption in the 260-280 m/.i. range, breaks and removes the pyi-im- idine ring, producing pliosphoribosyl urea and 5-isoxasolone (Schuster, 1961). The reaction with the cytosine com|X)unds causes the loss of the 260-280 m/x absori)tion; but exposure to low pH return the UV absorj)- tion partially or completely. The compound obtained after exposure to low pH is not C but the hydroxy lamino compound IV in Fig. 10 (Freese et al., 1961b). The exact structure of the intermediate non-UV-absorbing

H H \/ N 1		H OH H OH \/ \/ N N 1 1
1	r n	1 H ' N C—^ N C
1 II +2NH,OH-»> C C^ ' ^N-^ ^H /	-	+ NH,OH-».| 1 NHOH-*- | || +NH,OH / /

I n ni n?

Fic. 10. Reactions of liydroxylamine \vit)i cytosine. Reaction II -^ III in\()lves the li))oration of an amino group. Reaction III ^ IV is slow at neutral l)ut rai)i(l at low pH.

compound obtained after reaction with DNA is not known but the analysis of 1-methylcytosine by Brown and Schell (1961) indicates that it is the dihydroxylamine compound given as formula III in Fig. 10.

V. MOLECULAR MECHANISM OF MUTATIONS 229

Since for the DNA bases the strongest reacting base is cytosine and since the mutagenicity of this reaction can be explained, it seems probable that the major mutagenic effect of hydroxy lamine in DNA is due to the alteration of C. Both the unstable compounds (II or III) or the final compound IV miglit be responsible for base pair changes. In order to be sure of the details the frequencies of the tautomeric states of the molecules in Fig. 10 should be known. In any event, the hydroxyl- amino group should be more electronegative than the amino group (be- cause of the electronegative oxygen) and hence at least the molecule III or IV should be more frequently than C in the tautomeric state in which a hydrogen atom is on the 3-position nitrogen. This tautomeric form cannot pair with G but it can make at least one (and probably two) hydrogen bonds with A. Thus one should expect that the effect of hydroxylamine on DNA induces predominantly the base pair transitions (Freese et al., 1961b).

The specific effect of HA also explains why the reaction of hydroxylamine with DNA has such a small lethal effect compared to the strong mutagenicity.

The reaction of hydroxylamine with phage T4 is somewhat more complex since at low salt concentration HA mainly kills, apparently by some reaction with the phage coat, and the mutagenic effect becomes predominant only at high salt concentrations (>lilf NaCl) (Freese et al., 1961a). These authors have also shown that for short reaction times most ?'-type mutants appear in forai of mottled plaques while for longer treatment the fraction of pure r plaques increases; one would expect this since for longer treatment the likelihood increases that a second reaction occurs in the non-mutated chromosomal strand which causes it to produce non-viable or also r-mutant phages. HA induces point mutations in phage T4 (Freese et al., 1961a).

5-Bromouracil and its deoxyriboside react very rapidly with hy- droxylamine, lose their bromine and their UV absorption irreversibly. It is probable that this involves a succession of reactions but this has not yet been examined thoroughly. T4 phages containing 5-bromouracil in place of thymine are rapidly inactivated by hydroxylamine without any concomitant increase in the frequency of mutations. Similarly, hydrazine shows a much weaker mutagenic effect than hydroxylamine. It seems that the breakage of the pyrimidine ring is much more fre- quently lethal than mutagenic. The mutation frequencies of treatment

230 ERNST FREESE

1)V liydiM/.iiic air coiiipai-ahlc to oi' lowci' tliaii those of tl-catiiicnt hy low |)11 oi- hy cthyhitin^ agents (Fivese ct (il., HMilh). This ohsei'vation a-j;ain iiKhcatcs that the luutagenic effect of HA is predominantly caused hy its I'caction with C.

3. Exposure to Low pll

When one exposes DNA to solutions l)eIow pH 4 some bases (whose pK value for the amino groups is in this pH range) become positively- charged and strand separation occurs (Marmur et al, 1961) which may be lethal. At or above pH 4 this effect is negligible, but low pH can affect DNA by two additional reactions wliich sliould be active even above pH 4.

a. Dcpurhiation. The complete removal of A and G by low ])H has been reported by Tamm et al. (1952). Although these measurements have only been made in the pH ranges 1.6 to 4 it is clear that some depuri- nating effect must remain at higher pH, especially at elevated tempera- tures. At pH 1.6 the authors observed that G is liberated initially more readily than A. This should be measured again for pH ranges near 4.2, at which the mutagenic experiments are usually done.

No other effect of low pH on the nucleic acid bases has been reported, and extensive hydrolysis by low pH has often been used to isolate and quantitate the DNA bases.

b. Backbone Breakage. The sugar-phosphate backbone of DNA could be labile to treatment at low pH. This must be a minor effect since extensive treatment at pH 1.6 by the above authors gave "apurinic acid" without too much backbone breakage. However, when this apurinic acid is exposed to solutions at high pH all sugar-phosphate bonds which have no base attached to the sugar, apparently break, for the phosphate appears in the dialyzate (Tamm et al, 1953). Even at neutral pH this breakage occurs, although very slowly. It is apparently caused by "/3-elimination" (Bayley et al, 1961). Since the biological material after exposure to low pH is returned to neutral pH some of these breaks may occur.

Exposure of T4 phages to pH 4.2 induces mutations most of which arc point mutations (Fi'ecsc, 1959('l. It seems unlikely that such a small change can be induced by a break in the DNA backbone; a break should rather induce larger alterations or, more frequently, be lethal to the phage. Hence it is probable that the observed point mutations have been induced by depurination without any subsequent backbone breakage.

The gap left by the purine removal might be filled again before the DNA replicates. No such mechanism is known so far. Alternatively, the gap could remain empty until the DNA replicates. DNA replication

V. MOLECULAR MECHANISM OF MUTATIONS

231

could then stop at this site, which would be merely lethal, or any one of the four DNA bases could get incorporated, into the new comple- mentary strand, or finally the gap could be left out and thus cause a

transversions

transition

no base pair change

Fig. 11. Base pair replacements which could be induced by the removal of a guanine base from DNA.

small deletion (Freese, 1959c). The consequences of the incorporation of one of the four bases are illustrated in Fig. 11 for the case of a G removal ; in this case one would expect both transitions and transversions.

4. Exposure to High Temperatures

Similar to the effect of DNA exposure to too low pH a temperature increase above the "melting temperature" causes strand separation (Doty et ah, 1959). Again, in addition to this effect heat apparently causes some depurination of DNA (Greer and Zamenhof, 1962) and hence must also induce chain breakage. Zamenhof and Greer (1958) showed that heating to 60°C (below the melting temperature) is muta- genic for the E. coli strain W6. Since the extent of these mutations has not been genetically determined it is not known whether they are "point mutations" or larger alterations. Temperature effects on the pro- duction of mutations have also been found for Tradescantia (Sax, 1937) and Drosophila (Plough, 1941).

5. Alkylating Agents

There are many agents that carry one, two, or more alkyl groups in a reactive form; these are called mono-, bi-, or polyfunctional alkylating agents. The various groups on a polyfunctional agent may either act separately or they may cause the cross-linking of molecules. Figure 12 shows the chemical structure of the most commonly used classes of alkylating agents.

232 ERNST FREESE

All alkylating agents that prevent cell duplication have in common their reactivity with nucloophilic, i.e., slightly negatively charged, groups (Stacey et ai, 1958). The various chemical reactions by which alkyla- tion can occur have been described, e.g., by Price (1958), and the rela- tive reactivities, with respect to alkylation, of various chemical groups

Al-Cl

1. s Sulfur mustard:

\ both alkyl groups halogenated, bifunctional;

"^ ^^ only one alkyl group halogenated, monofunctional

Al— CI

2. CI — Al— N Nitrogen mustard:

\ all tnree alKyl groups halogenated, trifunctional;

two alkyl groups halogenated, bifunctional; only one alkyl group halogenated, monofunctional

O

II

3. Al— O— S— O— Al Dialkyl sulfates

II O

o

II 4 Al— O— S— Al Alkylalkane sulfonates

II O

H H

5. H— C-C — R Epoxides

O

H H

6. HC — CH Ethyleneimines

N I R

H H 7 HC — CH /3 -Propiolactone

I I

o— c=o

8. '*"N=N— R Diazo compounds

9. R— O' Other reactive oxygens

Fig. 12. Different t.vpes of most coinnionly used alkylating agents.

have been compared by Ross (1958); most reactive are sulfhydryl and thioester groups. When isolated DNA is treated in solution the reac- tivity of alkylating agents depends on their charge, negatively charged molecules, like chlorambucil, being least reactive. But when nucleopro- teins are used (sperm heads) all alkylating agents can readily react with phosphate groups and this reactivity compares well with their lethal

V. MOLECULAR MECHANISM OF MUTATIONS 233

effect on whole organisms (Alexander and Stacey, 1958). These authors propose that polyfunctional mustards exert their lethal effect by cross- linking of DNA molecules rather than merely by breaking the sugar- phosphate backbone of DNA. But the latter mechanism does not seem negligible since monofunctional alkylating agents also have a lethal effect.

Most alkylating agents have some mutagenic effect, whether they are mono- or polyfunctional. Their mutagenic strength and specificity, as compared to the lethal effect, depend on the structure of the alkyl group. Table II gives a list of mutagenic alkylating agents. Since they have mostly been tested on whole cells or organisms it is not known whether they act directly on DNA or the chromosomes, or whether their effect is indirect. In contrast to the necessary fixation of UV-induced mutations that requires protein synthesis no such fixation seems neces- sary for the mutagenic effect of at least certain alkylating agents; with diethyl sulfate and epichlorohydrin, inhibition of protein synthesis ac- tually increases the frequency of mutations in bacteria (Strauss and Okubo, 1960).

Only few attempts have so far been made to correlate the chemical specificity of alkylating agents with their mutagenicity. An example is the trifunctional agent, triethylenemelamine, which reacts with free pyrimidine bases; the reaction product is still mutagenic with thymine but not with cytosine (Lorkiewicz and Szybalski, 1961). The authors propose that triethylenemelamine is mutagenic by altering a DNA pre- cursor into a mutagenic base analog. This is apparently not the only effect since Fahmy and Fahmy (1958) reported that triethylenemelamine induces dominant lethals in Drosophila by causing large chromosomal alterations.

Most extensively analyzed with respect to both chemical effects on DNA and mutagenic effects are the methylating and ethylating agents. Such agents are dimethyl sulfate (DMS), diethyl sulfate (DES), methyl methanesulfonate (MMS), ethyl ethanesulfonate (EES), ethyl methane- sulfonate (EMS). They all act as monofunctional agents even when they cany two functional groups like DMS and DES since each group alkylates separately.

DNA can be altered in five different ways by these alkylating agents (see Fig. 13).

1. The strongest reaction is the alkylation of the phosphate groups of nucleic acids, which has been measured for many alkylating agents (Alexander, 1952; Reiner and Zamenhof, 1957; Stacey et al., 1958). The phosphate triester thus formed is unstable and for the most part hydro- lyzes to return the free alkyl group (Alexander and Stacey, 1958). If

234

ERNST FREESE

2 tn

-2 MfA
a, -H 05
>^ .	t- 05 T}< in	•<J> ir> fv,
T3 "O "O	05 05	05 05 ;q
	r^ *-H	rt rt (J)
-. ^	-. ^"-H
ark ark erga	*J to	port less ISS,
aa^	as	ail

T3 -a" '^

^s

O (U

5^

i2 ^

o ^ a*

M 2 JS

(0 3

Q 2

Q > U

Q>

X

o

X o

2 U

\ /

a:

X

o

o o

X

u

■z

			X
	u		y^
d^			g-
u	u*		X
x^	«c	x"	<-^\
H	x_	u	X/
w	o'	z	u

O -5,

U

x" u

X

o

X/^

o

x/ u

2X

V. MOLECULAR MECHANISM OF MUTATIONS

235

J3	CT>
a	ITS
tn	O
	»^
c	.
a	(0
>,	<u
B	"Z

OT o> ^ -a

§§

q, u

2

6

q, x* u

Qi -H 1=0 Z^ '^ S

•0 05 2

y "^ w

<U t» >, 3 n S"

U

u

x; 05

rt ITS

fa o»

S "a!

§ 3 rt

0)	0) u	;h	oi o	<u
s	s s	U	2 ^	s
rt	eri oj	rt	rt ^2	rt
			^^ d JZ
J2 H	22	a	S-i-H
a	a a	tn	a^ M	a
O 0)	o o	o	o ^ 'I'	O I)
O d	03 CO	u	to 2 ■<=	m bp
o o	s	o G "	o n
iH x:	u u	0)	!-r -ii Cfl	t, J=
Q ft	QQ	Z	Q > W	Q Oi

U w

q, af u

2;j

>o

y

o

o

u

I

z

/ \

I =!

0)

a o o ei

o

a o u

ft

I

2 1"

N >>

^ "O J=

2 "1 fa

i2 I >,

:>. ^ -^

N O rt

CO >J fa

236 ERNST FREESE

enough alkyl groups icniain attached uj) to the time at which tlie DNA attempts to duplicate, tlie duplication might be inhibited. The attached alkyl group might conceivably (but not likely) intei'fere with the DNA (hiplication in such a way that some non-com])lementary base would be

iiicoi-porated into the new strand.

Al Al

I I

1. S— Ph-S-Ph-S-Ph— S-Ph-S— Ph

I I I I I

B B B B B

Al Al

' II >

S — Ph-S-Ph— S— Ph-S Ph-S— Ph

I I I I " I

B B B B B

Al Al

I I

S— Ph— S— Ph— S— Ph— S-Ph— S— Ph

I I I I I

B B B B B

I I

Al Al

Al Al

I I

S— Ph— S — Ph— S— Ph— S— Ph-S— Ph I III

B B B B B

B 1	B	B 1
1 Al		Al

5. S-Ph-S Ph— S— Ph— S-Ph— S— Ph

I "ill

B B B B B

I I

Al Al

Fig. 13. Various reactions of alkylating agonts with DXA (see text for ex- planation).

2. The phosphate triester can occasionally also hydrolyze between the sugar and the phosphate and thus the DNA backbone gets broken. The relative frequencies with which the alkyl group becomes removed or the chain is broken are not known. Chain breakage might induce larger alterations or be lethal, but probably does not induce point mutations.

3. Some of the bases are alkylated (see Fig. 13). Treatment of DNA

V. MOLECULAR MECHANISM OF MUTATIONS

237

with DMS mainly produces 7-methyIguaninc (Reiner and Zamenhof, 1957). In addition 1-niethyladenine, some 3-mcthyIadenine and 1,3- dimethyladenine, and some cytosine derivative are formed (Brookes and Lawley, 1960a). Ethylation of the bases is less effective but some 7-ethylguanine has been found after treatment of DNA with DES (Bautz and Freese, 1960). The preferential formation of 7-alkylguanine has also been found for nitrogen mustard (Brookes and Lawley, 1960b). Lett et al. (1962) argue that the alkylation of the base does not occur directly but by transalkylation, the alkyl group changing from the phosphate to the base; for our purposes it is unimportant by which mechanisms the bases are alkylated. The alkylated base might in- hibit DNA duplication or cause base pairing mistakes during DNA duplication.

4. Depurination. The alkylation of purines in the 7-position (see Fig. 14) gives rise to quaternary nitrogens which are unstable. Either

CH,

I

CH,

I

O

N N NH,

I '

R

m

IV

NH

NH,

NH,

+ROH +H+

Fig. 14. Guanine removal by ethylation. I and II: Mesomeric states of 7-ethyl- deoxyguanosiue. Ill: Liberation of the ethyl-group. IV. Decay into 7-ethylguanine and deoxyribose.

the alkyl group itself hydrolyzes away from the purine or else the alkylated purine separates from the deoxyribose leaving it "depuri- nated." The liberation of ethylated and methylated purines from DNA

238 ERXST FREESE

has Ihtii ()I)srrvc,l ( I^autz aii.l Fivcsc, 1960; Lett ct ni. 1962). The g'dp might intc'i'tVic wilh 1)XA (luplication oi' cause tlic incorporation of a wrong base.

5. The depiii-inalcd DNA is labile at higli i)ll and as sliown in Section II,C,3 for the (le|)ui'inati()n l)y low pli, may occasionally break even at neutral i)II. Again this might induce larger alterations or be lethal, but ]irobably does not induce jjoint mutations.

Any one of these five different effects could be lethal. Treatment of DNA with MM8 or DMS rapidly methylates the phosjihates and the purine bases (Reiner and Zamenhof, 1957) and leads to the depoly- merization of DNA (Lett et (iL, 19621. These agents are, however, very slightly mutagenic (per lethal hit) for bacteriophages, in contrast to the ethylating agents (Loveless, 19.59; Bautz and Freese, 1960). This seems to exclude points 4 and 5 as the major cause of lethality, since depurination by alkylation should be as mutagenic as by low pH treatment. Of the remaining three mechanisms — 1, 2, and 3 — the hy- drolysis of the DNA backbone (2) seems the most likely mechanism for the major cause of lethality by methylation especially since Loveless (1959) has shown that the phage titer still decreases after the treatment with MMS has been stopped by sodium thiosulfate or by dilution.

The inactivation by ethylating agents proceeds much more slowly than for mcthylating agents, but for phages the frequency of induced mutants per lethal hit is much larger for ethylating than for methylating agents. Since the amount of esterification of the phosphates in DNA is about the same as for methylation the high mutagenicity indicates that DNA backbone breakage by mechanism 2 is much less for ethylated than for methylated DNA. In agreement with this expectation Lett et al. (1962) find that EMS treatment of DNA does not influence its viscosity while MMS treatment leads to a rapid degradation. The authors interpret these results differently; they believe that mechanism 5 is mainly responsible for breakage. EMS or EES (Benzer, 1961 ; E. B. Freese, 1961) induce point mutations in phage T4. As for the nnitations induced by low pH this finding makes it unlikely that they are caused by breaks of the DNA backbone. Even in maize ethylating agents mainly induce mutations of small genetic extent and not large altera- tions (Heiner et al., 1960).

For the induction of mutations mechanisms 1, 3, and 4 are then to be considered. For mechanism 1 the alkyl group on the phosphate would have to interfere in some way with the pairing properties of the bases and cause the incorporation of a w^rong base during DNA duplication. But the alkylated phosphate triester hydrolyzes and each removed ethyl group should decrease the frequency of mutations in treated phages. Since actually the contrary is observed (Bautz and Freese, 1960) mecha-

V. MOLECULAR MECHANISM OF MUTATIONS

239

nism 1 can be excluded as the major cause of mutations. Green and Krieg (1961) have furthermore shown that the hibile ethyl groups can- not indirectly induce mutations inside the bacterium since a coinfected but nonalkylated phage does not reveal any mutation induction. These authors have also observed, by single burst experiments, that EMS- treated phages seem to continue throwing out normal and mutant phages even after they have replicated; this is in agreement with both mecha- nisms 3 and 4.

In case 3 the frequency of mutants per viable phage would be inde- pendent of subsequent incubation without DNA duplication while it might increase in case 4. Such an increase has been observed after incu- bation of free phages (Bautz and Freese, 1960) and for bacteria (Strauss, 1962). Besides that, mechanism 3 is expected to induce base pairing mistakes only during DNA duplication and thereby only transitions, while mechanism 4 should give essentially the same results as treatment by low pH : the removal of guanine could result in transitions as well as non-transitions. Non-transitions have actually been found (E. B. Freese, 1961) (see Table III). Hence mutations by ethylation of DNA seem to be mainly caused by the ethylation of G, which often is removed from

TABLE m Reversion Induction of Spontaneously Reverting Mutants

Rn Mutants induced by	No. of mutants tested	Mutants found (%) Inducible Non-inducible to revert to revert by AP by neither and/or BD AP nor BD	Approximate spontaneous background of non-reversion inducible mutants
2-Aminopurine	98	98		2	2
5-Bromouracil	64	95		5	2
Hydroxylamine	36	94		6	4
Nitrous acid	47	87		13	15
pH 5, 45°C	115	77		23	15
Ethyl ethanesulfonate	47	70		30	10
Proflavin	55	2		98	-
Spontaneous	110	14		86	-

From E. Freese (1961)

240 ERNST FREESE

DXA, either (luring growth or before, when the treated DNA is incubated at elevated temperatures. Nevertheless, the alkylation of A seems to have some although a smaller mutagenic effect as will be seen in Section V,B,4.

It is interesting to note that the methylating agents also have some mutagenic effect: DMS is able to induce the reversion of certain Xeuro- spora mutants (Westergaard, 1957) and MMS induces the reversion of certain bacterial mutants (Strauss, 1962). The frequency of these muta- tions could not be increased by incubation subsequent to treatment (Strauss, 1962) ; they may have been induced by another one of the five mechanisms mentioned or by an indirect effect.

The emphasis on the explanation of point mutations should not distract from the fact that alkylating agents can also induce large alterations. For example, Xasrat et al. (1954) have shown in Drosophila that mustard gas induces trans- locations whose frequency increases with the square of the frequency of lethals: these alterations presumably are caused by a break in each of the two cliromosomes involved.

Formaldehyde may also be classified as an alkylating agent. It can react with proteins (Putnam, 1953), with RXA (Fraenkel-Conrat. 1954), and with single- but not with double-stranded DNA (Staehlin. 1958). It forms monomethylol groups ( — NH — CH...OH) with the amino groups of the non-hydrogen-bonded nucleic acid bases; these bonds can either slowly break again or the hydroxymethyl group can react with some other group to form methylene bridges (Staehlin, 1958). Even native DNA can apparently react to some extent with fomialdehyde, at least in its native state inside a phage, since Sauerbier (1960) has shown, by experiments with multiplicity reactivation, that the inactivation of phage Tl by formaldehyde is not caused by an efTect on the injection mechanism. The mutagenic mechanism of fomialdehyde (Rapoport, 1946; Auerbach and Moser, 1953) is not yet understood. Formaldehyde, if UV-irradiated shortly before use for mutagenic treatment, is more nmtagenic (Jensen et al., 1951) ; peroxides also increase its mutagenicity (Dickey et al, 1949; Sobels, 1956). This suggests that either a radical or a peroxide is the active mutagenic agent of formaldehyde, unless the combined action of the two agents produces an additional mutagenic effect.

E. RADIATIONS

1. Ionizing Radiations

The biological effects of X-rays and y-rays, a particles, protons, high energy neutrons, and electrons have been studied so extensively that only

V. MOLECULAR MECHANISM OF MUTATIONS 241

a few review articles can here be mentioned, especially the volumes of Radiation Biology (1954, 1955) and Supplement 1 of Radiation Research (1959). The radiation can have a direct effect on the chromosome, e.g., directly break it or alter one of the DNA bases, and an indirect effect, since along the track of each particle a large number of ions or radicals remain which can initiate a chain of chemical reactions. The number of ion pairs produced in a given thickness of material, i.e., the ionization density of the radiation, depends on the nature and the kinetic energy of the particles, those of low energy giving the largest ionization density (except for neutrons which at thermal energies do not excite electrons but specifically interact with certain nuclei). The biological effect also depends on the kind of cell and on the stage of its nuclear cycle at which the irradiation is given (Marquardt, 1938) . For example, chromo- somes are extremely sensitive to breakage in meioitic prophase (Sax, 1941; Glass, 1955).

The inactivation of organisms often follows a first-order reaction equation

^ = -/3iV or N = Noe-0' = N.er-^'' at

where p is proportional to the intensity of the radiation. D is the "dose," that is, the energy absorbed per gram of material ; it is usually measured in roentgens (see Marinelli and Taylor, 1954) ; K often is a dose inde- pendent constant describing the sensitivity of the organism. The fre- quency of mutants per viable organisms often increases linearly with the dose (Demerec and Sams, 1960) :

M = aD

Timofeeff-Ressovsky et al. (1935) interpreted these equations in temas of a "target theory" which states that a single "hit" of the particle on the "target," i.e., the genetic material, inactivates or mutates it. If the inactivation follows a first-order equation one calls (3t = KD = n the number' of lethal hits because than the survival is given by the zero- order Poisson term iV/jYo = e~". Formally one obtains the same first- order reaction equations if the radiation product produces some chemical which can migrate from its origin to some other site in the organism or even in the reaction vessel. But the meaning of the word "target" is different and depends on the chemical or metabolic lifetime of the radia- tion produced chemical, i.e., whether it can get to any place or is limited to the closest surroundings of the track of the ionizing particle. If the lifetime is long the radiation produces mutagenic chemicals whose con- centration dejiends on the number of ionizations; in that case the amount

242 ERNST FREESE

of absorlx^l I'licr^y, in roentgens, is the piopci- lucasui'c for the "dose." Ho\ve\-(M', if the I'ffeetive lifetime of tiie pi'dduccil I'licinical (radic-ili is short a hettei' nieasui'e foi' tlu^ dose would he tiie nuinher of ionizing parti- cles whit'h penetrate the material, multiplied by some factor that measures the ra(hation efficiency of these particles. These complications arise only when one comi)ares radiations with different ionization densi- ties. FarlieU's with high ionization density produce, per roentgen, more chromosomal aberrations than particles with h)w ionization density, probably because the chromosome has to be cut at more than one place; in contrast the induction of mutations is usually more efficient for radia- tions with low ionization density (Lea, 1955).

This shows that the word "single hit" can mean different reactions. It can mean a single chemical event or a single ionization, but it can also mean a single cut by a densely ionizing particle. A single-hit curve, without the precise knowledge of the chemical or physical reactions going on, does not permit one to decide whether the reaction is direct or indirect and how many chemical reactions occur until the actual inactivating or mutating event takes place.

The frequency of simple chromosome aberrations, e.g., deletions, is proportional to the dose (measured in roentgens). Hence they arise by "single hits." In contrast the more complicated ring dicentrics and ex- change aberrations increase approximately with the square of the dose; they are caused by two hits (Sax, 1957). The frequency of two-hit aber- rations is larger when a certain dose is given continuously than when it is fractionated, with larger rest periods. This indicates that chromo- somal breaks have the chance to heal again and fewer aberrations are produced when the healing takes place before the second hit occurs (e.g., Sax, 1957). It is particularly interesting that anoxia (Wolff and Luip- pold, 1955) as well as chloramphenicol (Wolff, 1959), an inhibitor of protein .synthesis, keep the induced breaks open longer. These observa- tions may have a significance for the understanding of the structure of the chromosome as well as of mutation and recombination.

The importance of the indirect radiation effect was recognized even before the first chemical mutagens were known. Thus Fricke and Demerec (1937) and Friedewald and Anderson (1940) showed that rab- bit papilloma virus is about 20 times more resistant in crude preparations than in a purified state. More light was shed on the indirect effect when Thoday and Read (1947) found that low oxygen concentrations reduced the frequency of chromosome breaks induced by X-radiation. This oxy- gen effect has been extensively studied and it is jirobable that radiation in the presence of oxygen causes tlie foi'mation of some peroxide radical

V. MOLECULAR MECHANISM OF MUTATIONS 243

which in turn affects the frcciucncy of chromosome breaks and maybe other mutations (see Stone, 1955). The formed peroxides may cause an increase in the frequency of chromosome breaks either because they induce breaks themselves or because they prevent the rejoining of chromosomes cut by the radiation (Baker, 1955). Reducing agents lessen the effect if added before the irradiation (Hollaender and Kimball, 1956). The major effect of the peroxides is not known; one effect on DNA could be the oxidation of the bases. In particular adenine readily reacts with peroxides giving adenine-1-A^-oxide (Stevens et al., 1958; Frederik- sen and Klenow, 1960), whose nucleotides are growth inhibitors. It is not known whether the oxidation of bases in DNA is mutagenic.

In microorganisms both forward and reverse mutations have been induced by X-rays, e.g., in Neurospora (Giles et al., 1955), in bacteria (Demerec and Sams, 1960), and in yeast (Gutz, 1961). The observations prove that X-rays also often induce point mutations.

Several phage strains are more readily inactivated in the absence than in the presence of oxygen. Alper and Ebert (1954) conclude from this and other irradiation experiments that phage is inactivated by re- ducing rather than by oxidizing radicals. This may be due to an effect on the phage protein rather than the DNA. (For a detailed review of the radiation effect on vii-uses see Gard and Maal0e, 1959; Stahl, 1959). To my knowledge there are no reports about the mutagenic effect of ionizing radiations on free viruses.

2. Ultraviolet Light

The irradiation with UV has a more specific chemical effect than ionizing radiation since the UV absorption is limited to molecules cariy- ing conjugated double bonds and each of these molecules has a special absorption spectrum with maxima at certain wavelengths. The absorp- tion maxima of the nucleic acid bases are in the range of 260 to 280 m/x. UV of this wavelength range exerts a strong direct effect on nucleic acids. This can be seen especially well for viruses; they show no lethal effect when they are introduced into preirradiated medium but are killed by a single-hit curve when directly irradiated (McKinley et al., 1926; Fisher and McKinley, 1927). The UV action spectrum of the rate of inactivation has a maximum at 260 m/x indicating that the inactivating "hits" occur directly in DNA. This is further supported by the observa- tion that two or more UV-damaged phages can still infect the same bac- terium and produce viable progeny (multiplicity reactivation, Luria, 1947). The ability of the phage to inject its DNA is inactivated much more slowly than its viability. For one needs about 10 times as many

244 ERNST FREESE

UV hit^ to inactivate the rll function (Kric^, lOoOa) or an rll marker (Docrniann ct (iL, 19')o) than to inactixatc the phage. For a review see Stalil (1959).

Attempts to induce mutations l)y the UV-irradiation of free phages have often been unsuccessful (Weigle, 1953; Latarjet, 1954) while irradiation of the infected bacteria (Latarjet, 1949, 1954; Weigle, 1953; Weigle and Dulbccco, 1953; Tessman, 1956) or even of the bacteria alone before infection (Jacob, 1954; Fraser, 1957 1 leadily produccHl UV-induced phage mutations. In some cases these ''mutations" may have been caused by a radiation-induced recombination process between the phage and the bacterial genome (Stent, 1958) excejit for bacteriophages, like T2 or T4, which do not seem to have any similarity to the host genome (very different base ratio and Hj\IC in their DNA). A very small direct effect has been reported by Tessman (1956) for phage Tl. A significant induction of reverse mutations of a T4 rll mutant by UV-irradiation of free phage has been shown by Krieg (1959b). To prove the effect it was necessaiy to inactivate the phages quite strongly and then to infect the bacteria in high multiplicity; it is not clear whether revertants can be induced by some mistakes in multiplicity reactivation. Another direct effect is reported by Ellmauer and Kaplan (1959) for the phage K of Seiratia marcescens.

In bacteria the possibility of an indirect induction of mutations by UV has been clearly established. Stone et al. (1947) showed that the UV-irradiation of bacterial media produced mutagens which induced nmtations in unirradiated organisms placed in these media within less than 4 hours. The radiation product is apparently unstable and organic peroxides or radicals have been suggested as the cause of mutations. Wyss et al. (1948) have shown that organic peroxides are mutagenic while HoOo in saline is not. But since UV also acts on the pyrimidine bases and the corresponding nucleotides within the bacterium, it is by no means clear how the major mutagenic effect of UV on bacteria comes about. That most UV-induced mutations somehow involve extrachromo- somal material is shown by the observations (Witkin, 1959) that the frequency of UV-induced mutations is reduced when in the innncdiate post-irradiation period protein synthesis is blocked (by lack of amino acids or by chloramphenicol) or altered (by the incorporation of amino acid analogs) or when the formation of purines is inhibited (by non- mutagenic concentrations of caffein). In contrast the frequency of mutations is increased by the addition of nucleic acid bases. The estab- lishment of UV-induced mutations apparently needs the presence of some unstable DNA precursors and DNA synthesis; the majority of potential UV-induced reverse nmtations are established within one bac-

V. MOLECULAR MECHANISM OF MUTATIONS 245

tcrial division following UV treatment (Witkin, 1956; Haas and Doud- ney, 1957).

The effect of UV has also been studied in other organisms; both mutations and chromosomal aberrations have been observed (e.g., Swan- son and Stadler, 1955; Kirby-Smith and Craig, 1955; Pomper and Atwood, 1955). In yeast a mutagenic spectrum of UV-induced point mutations has been detennined (Leuppold, 1962).

UV-irradiation of the pyrimidines produces unstable base analogs. The major effect is apparently the saturation of the double bond between the 5- and 6-position by water. (For a review see Shugar, 1960.) The reaction product of cytosine readily reverts at neutral pH when the irradiation is shut off. Nevertheless, this product or some derivative may be the cause of mutations. The reaction product with thymine is more stable and reverts quickly only on heating or acidification.

Recently Beukers and Berends (1960) have shown that UV-irradia- tion of concentrated thymine solutions as well as of DNA jiroduces a thymine dimer. The reaction apparently goes via a higher energetic triplet state. This dimer has also been found in the DNA of UV- irradiated bacteria (Wackcr et al., 1960b). One can readily understand the lethal effect of the dimer formed in DNA since DNA replication is probably stopped at this place; but it is an open question whether the dimer has anything to do with UV-induced mutations.

F. UNCLASSIFIED MUTAGENS

Certain mutagens are hard to classify since so little is known about their chemical site of mutagenic attack. Such agents are, for example, manganese chloride (Demerec and Hanson, 1951), whose effect depends on the pretreatment of the bacteria, and several heavy metals (Glaess, 1956) and alkaloids (Ochlkcrs, 1949) in higher organisms. Another agent is 4-nitroquinoline-l-A^-oxide, which is mutagenic for tobacco mosaic virus RNA (Endo et al., 1961) and for bacteria (Mashima and Jkeda, 1958) but it is not clear whether its mutagenicity is due to the liberation of nitrous oxide or to some other effect. Also DoO was found to induce mutations slightly in bacteria (De Giovanni, 1960, 1961) and phages (Konrad, 1960) grown in its presence.

G. SPONTANEOUS MUTATIONS

The only reasonable operational definition of "spontaneous" muta- tions is that these are mutations produced under "normal" growth conditions. Both the frequency and specificity of these mutations may therefore depend on the growth medium, temperature, pH, etc., which

246

ERNST FREESE

are noiuially cniploycd. Experience has shown that the spontaneous frequency of mutations is approximately the same for (Ufferent growth con(htions usually used in a laboratory. Nevertheless, small but signifi- cant differences may be found and sometimes have been recorded as "mutagenic" or "anti-mutagenic" effect.

Some spontaneous nnitations nuist come about l)y mistakes in the normal duplication of DNA or RNA. Actually it is surj)rising how few mistakes are made during this process. Transitions might come about

H "^ N-C^ N- -C N-C	-H-		H /
— H-	-N H c— C / \ -N ^-H C-N / \ o c— /
Adenine			Cytosine
		(rare	imino form)

H\„/N^ P H-Q^ C-H

I /C-c /:-{

N— C N— H N C— H

/ \ / \ /

— C N— C C-N

\ \ / \

^ N— H O C—

Guanine

Thymine (rare end form)

Fig. 15. Rare base pairs which could occur spontaneously and would give rise to transitions. Only the cases are shown in which the pyrimidine has undergone the tautomeric shift. The same base pairs could occur when a purine has undergone a tautomeric shift or even when anv one of the bases has been ionized.

by the tautomeric shift (or ionization) of one of the bases (Watson and Crick, 1953) which lead to the mistaken base pairing between A-C and, probably more frequently, G-T, as illustrated in Fig. 15. Transversions might come about by the mistaken pairing between two purines or two pyrimidines (Freese, 1959b). There are several such pairs possible that

V. MOLECULAR MECHANISM OF MUTATIONS

247

use one or two hydrogen bonds; in the latter case they have slightly wrong distances or angles for the two sugar (C) — base (N) bonds. Since the DNA backbone is not completely stiff it can accommodate such distortions, at least occasionally, as a mistake. The principle by which in this way transversions may come about is illustrated in Fig. 16.

transversion

t ransversion

e T

by

pairing- mistake

A- e

by pairing- mistake

T- C

	6l
^^ H
••,	i_.
""■•<	T C
	L J

Fig. 16. Spontaneous production of transversions by rare base pairing between two purines or two pyrimidines.

Initiated by the first pairing mistake, larger alterations might be induced occasionally (Freese, 1959b). Deletions of one or more nucleotide pairs might also come about when one or more nucleotides in an existing DNA strand form a loop and thus do not get copied during DNA duplication, and inversely insertions might come about when by the same mecha- nism one or more nucleotides get copied twice (Fresco and Alberts, 1960). Apart from these mistakes inherent in the structure and mode of replication of nucleic acids, spontaneous mutations may be caused by a variety of agents made by the cell or present in the medium. Agents of this kind are base analogs, peroxides, nitrous acid, etc., all of which are produced under some growth conditions and often are inactivated by special enzymes like catalase, peroxidase, etc. Some organisms have an especially high mutations rate (Drosophila: Demerec, 1937; bacteria: Treffers et al., 1954) ; jMiyake (1959) has shown that in .S. typhimurium

248 ERNST FREESE

this property is genetically controlled and can itself nnitate. Thus, "spontaneous" mutations can Ix^ conti'olled genetically and by the environment.

Some of these agents inside the cell act on duplicating DNA, others on resting DNA. The freciuency of nmtations induced by the attack of resting DNA increases with the time of exposure and may become large if resting nucleic acid is stored for a long time, as in seeds, j)ollen (Cartledge and Blakeslee, 1934), viruses, etc. Also in resting bacteria an increase in the frequency of mutations has been observed (Ryan, 1959). Such bacteria do not show any major turnover of DNA, but it is not clear whether or not a minoi- DNA turnover occurs (Ryan et al., 1961).

In general it is difficult to decide which mutagenic mechanism is mainly responsible for the spontaneous mutations of a given organism. This decision is relatively easy only in special cases in which specific environmental conditions or a genetic factor greatly influence the rate. I'sually one can only exclude certain mechanisms as unlikely in one of two ways. On the one hand, one can compare the genetic extent and the frequency distribution of spontaneous forward mutations with the extent and frequency distribution of mutations induced by agents of known chemical properties. This has been done for the rll mutants of phage T4 and very different mutability spectra have been found for spontane- ous mutations and those induced by 5-bromouracil (Benzer and Freese, 1958), i)roflavin (Brenner et al., 1958), 2-amino|mrine (Freese, 1959a), and other mutagens (Benzer, 1961). One can conclude that the major spontaneous mutagenic mechanism in phage must be different from that caused by any one of the agents used for mutation induction. On the other hand, one can induce the reverse mutation of spontaneous and induced mutations by various mutagens and thus determine the proper- ties of the base pair changes involved (see Section V).

IV. The Extent of Point Mutations

As shown in Section III, most mutagens induce in nucleic acids the change of a single base or base pair. This does not necessarily imply that such a simple change can show up as a phenotypically observable muta- tion. Conceivably, several base pairs may have to be altered before such a mutation would l)e produced. I shall now cite several experimental results which indicate that actually a single base pair change is responsi- ble for most mutations observed in phage T4.

A. GENETIC MAPPING OF MUTATIONS

The classical approach to an analysis of the extent of mutations has been to cross as many independent mutants as possible with each other

V. MOLECULAR MECHANISM OF MUTATIONS 249

and to decide whether or not recombinants can arise between them. The frequency of single and double recombinants could be used to determine the sequence and relative distance of the various mutations.

As already outlined in Section I,C, the most refined genetic map is that of the rll region of phage T4 which Benzer (1961) has evolved using a new technique of deletion mapping. This experimental system is sensitive enough to detect any recombinant that occurs. Within the genetic region of the rll functional property at least 304 different mutational sites have been found and according to the Poisson distribu- tion at least 120 more sites have not yet been detected (Benzer, 1961). Actually even more sites must exist, because so far certain mutagenic agents of a restricted specificity have been used mainly. Since the length of the rll region is at most 10* and probably of the order of 10^ nucleo- tide pairs, each mutational site can correspond at most to a few nucleo- tide pairs and recombination can occur between any two such sites. Most spontaneous and practically all chemically induced mutations so far analyzed involve only one of these sites. The induced mutations include those induced by 5-bromouracil (Benzer and Freese, 1958), proflavin (Brenner et al., 1958), 2-aminopurine (Freese, 1959a), nitrous acid (Freese, 1959c; Benzer, 1961), ethyl methanesulfonate (Benzer, 1961), low pH (Freese, 1959c), and hydroxylamine (Freese et al., 1961a). Only about 5% of the spontaneous mutations extend over several of the presently known mutagenic sites.

B. SPONTANEOUS AND INDUCED REVERSE MUTATIONS

If most mutations really correspond to the change of a single nucleo- tide pair in DNA they should occasionally mutate back to the original genotype, i.e., the original sequence of nucleotide pairs. Furthermore, in some cases this reversion should be inducible by some of the mutagenic agents. In contrast, a mutation which extends over several nucleo- tide pairs might find it difficult to revert all these nucleotide pairs simultaneously.

Actually, practically all T4-r7/ mutants induced by the aforemen- tioned chemicals can revert to the original phenotype spontaneously and most of them can be induced to revert by some of the agents. Also all but 10% of the spontaneous mutants can revert spontaneously. The question is only whether these reverse mutations to the original pheno- type arose by a genuine back mutation at the mutant site or rather by some other ("suppressor") mutation at a different (or even the same) genetic site.

Suppressor mutations will be treated in more detail later and it may suffice here to say that, for most spontaneous and induced mutants which

250 ERNST FREESE

have l)cen oxainined in sufficient detail, rcvcrtants have been found wliich l)y functional and ivcoinhinational tests are indistinguishable from the standard type phages (Freese, 19')9b; P'reese ct al., 1961b). When a new sjiectruni of rll forward nuitations was deterniined a few revertants behaved as standard tyi)e phages (Benzer, 1961). It is veiy likely, therefore, although inipossil)le to prove unequivocally by jiurely genetic means, that genuine back mutations to the standard genotyi)e exist for most /•// mutations and hence that the original cliangc in DNA was rather small.

Exceptions to this rule may he the nuitants induced by jiroflavin (Crick et nl., 1961) which spontaneously seem to revert only, or mostly, by suppressor mutations. Otherwise the existence of suppressor mutations is unimportant for the abo^'e argument.

C. THE KINETICS OF MUTATION INDUCTION

The ease with which most chemical reactions can be both initiated and stopped enables one to study the increase of the mutations frequency as a function of the number of nucleic bases attacked. Phage T4 is usually inactivated by a first-order reaction kinetics:

= g-^' = e-»

Here B is the titer of the T4 phages on bacteria B, B,, is the titer at zero time of treatment, /? is the inactivation constant and n is the number of "lethal hits," n is proportional to the number of deaminated bases (as long as that number is small) ; it is not identical to the number of deaminated bases since there should be many bases which do not have any important informational task and thus can be altered without lethal elTect.

(3 depends on the concentration of the mutagen. If that is constant in time, as for the treatment by nitrous acid, the time of treatment itself is proportional to the number of bases attacked. When, however, the mutagen decays rather rapidly,'* e.g., by hydrolysis as for many alkylating agents, the more adequate measure for the number of bases attacked is:

1 Bo n = In -5-

D

* Once the single-hit character of mutation.s of a certain kind (e.g.. forward muta- tions of a certain phenot.ype) is established one can use their frequency as another measure for the number of bases attacked. One then plots against this frequency the frequency of another kind of mutation. Such a plot has the advantage of being independent of the concentration of the mutagen or of an assumed constant corre- lation between lethal hits and bases attacked, which sometimes does not hold true (Freese et al, 1961a).

V. MOLECULAR MECHANISM OF MUTATIONS 251

If mutation induction is a one-hit phenomenon, i.e., a single chemical reaction can induce a mutation, the frequency of mutants per viable l)hage must increase linearly with the number of treated bases, that means with the number of lethal hits. If we call M the frequency of mutants per viable phage we thus have:

M = ilf 0 + ixn

where jx is a constant which depends only on the kind of mutation which one examines. It measures the probability with which the mutational site or sites are attacked by the chemical, per lethal reaction of the phage. If, in contrast, more than one chemical reaction (hit) were neces- sary to induce a mutation, M would increase with some power of ixn.

Kinetic measurements of his kind have been done with nitrous acid as mutagen for TMV (Mundry and Gierer, 1958) and phages T2 (Vielmetter and Wieder, 1959) and T4 (E. B. Freese and E. Freese, 1960), with ethyl ethanesulfonate (Bautz and Freese, 1960; E. B. Freese, 1961), with hydroxylamine (Freese et al., 1961a,b), and with low pH treatment (E. B. Freese, 1961). It was always found that the frequency of forward mutations per visible phage (e.g., from standard type to r type) increases linearly with the number of lethal hits. This indicates that each of these mutations gets induced by the chemical reaction with a single base of DNA. Not only the initial but also the ultimate change in this mutated DNA apparently involves only one nucleotide pair, since for the induction of reverse mutations again the frequency of revertants per viable phage increases linearly, as can be seen in Fig. 17 (p. 259).

The various mutagens induce about 10 *-10~'' revertants per lethal hit for highly reversion inducible mutants. This frequency is about lO'-lO"* times smaller than that of the induction of rll forward muta- tions. Since a T2 or T4 phage has about 2 X 10^ nucleotide pairs, these figures are of the right order of magnitude if one assumes that the hit of only one or any one of a few bases can induce a reverse mutation while an rll forward mutation can be induced by the alteration of any one of many bases within the rll region, and a lethal effect can be produced by the alteration of any one of even more bases. In contrast, if the alteration of more than one base were necessaiy to induce a reverse mutation, a two or more hit curve should be observed, with a gradually increasing slope, and the frequencies of reverse mutations per lethal hit should be much less (e.g., of the order 10~^").

It should be noted that quite a number of mutants do not respond to any in vitro mutagen known so far. Obviously, the above kinetic argu- ments do not hold for them.

252 ERNST FREESE

\ . Ml TAGEMC Specificity

The previous chapters have shown that I'ach mutagen has eertain chemical properties and therefore is expected to iiuhK-e only certain base pair changes to the exclusion of others. This chemical sjiecificity should correspond to a mutagenic specificity such that each mutagen can induce only certain of the possible mutations to the exclusion of others. Some selection rules should be strict ami exceptions impossible, others should merely indicate a more or less pronounced mutagenic preference of a given mutagen.

A mutagenic specificity has been actually observed for the induction of reverse mutations by various mutagens. For example, Demerec (1953) reported two mutants in E. coli of which one (leucine-dependent) responds little to UV and much to MnClj while the other (phenylala- nine-dependent) responds in the opposite way. In Xeurospora crassa several adenine-dependent mutants, controlling the same functional property, have been found to differ in their spontaneous and X-ray- induced revertibility (Giles, 1955; De Serres. 1958) and can be dis- tinguished by their pattern of response to different chemical mutagens (Westergaard. 1957). A ver>' detailed analysis of mutagenic specificity has been undertaken for the rll region of bacteriophage T4 which will be reported in more detail below. Several results with bacterial systems (Demerec. 1960; Rudner. 1960. 1961; Kirchner. 1960; Margolin and Mukai, 1961; StrelzofT, 1962) and with yeast (Leuppold. 1962; Gutz. 1961) corroborate and extend these findings.

A. THE SPECIFICITY OF FORW.\RD MUTATIONS

Each functional property is determined by a genetic region, contain- ing many sites whose change results in the loss of the function. It is not surprising, therefore, that forward mutations for different functional properties can be induced with similar frequency (same order of magni- tude), because individual site specificities average out. But if one examines the frequency of forward mutations at one particular site quite a different picture emerges. For such an analysis a large number of independently arising mutant* of the same functional property must be isolated and mapped. The results for the ;•// mutants of phage T4 show that mutations frequently recur at certain sites (hot spots). The loca- tion of these sites is specific and different for different mutagens. The highest specificity has been observed with spontaneous mutations (Ben- zer, 1957, 19611, next with the base analogs 5-bromouracil (Benzer and Freese, 1958) and 2-aminopurine (Freese, 1959a), and witli some in vitro chemicals (Freese, 1959c; Benzer, 1961). Proflavin-induced mutational

V. MOLECULAE MECHANISM OF MUTATIONS 253

sites (Brenner et al., 1958) are more scattered and are nearly all different from the spontaneous ones.

These observations show that the mutagenic specificity is more complicated than one might have expected from the chemical observa- tions. Clearly the mutability of a given nucleotide pair is influenced by other structural factors. Part of this effect may be due to the different kinds of cytosine bases that are observed in DNA. 5-Methylcytosine is found in many organisms in addition to cytosine and preferentially next to G (Sinsheimer, 1955) ; in T2, T4, T6 phages differently glucosylated 5-hydromethylcytosine bases are observed and the extent of glucosyla- tion depends on the neighbors (Lehman and Pratt, 1960; Kornberg et al., 1961). Hydroxylamine distinguishes chemically between some of these C bases (Freese et al., 1961b), hence other mutagenic mechanisms, including the spontaneous one, might do the same.

In addition, the nearest neighbors of a given base undoubtedly can have a direct influence upon its mutability, as is indicated by the sites of \evy high mutability mentioned above. Such effects may be produced for example by the difference in the hydrogen bond strength of the next neighbors, that of A-T pairs being less than that of G-C pairs, as it determines the ease with which the base pairs can open up, or by sterical effects of the next neighbor, purines being bulkier than pyrimi- dines, or by an electronic influence of the Tr-electron clouds of one base upon the next one.

It is conceivable that certain DNA sites have evolved by such means a special protection against spontaneous mutations and against certain mutagens.

B. THE SPECIFICITY OF REVERSE MUTATIONS

The aim of studying the induction of reverse mutations is to divide mutations into classes which do or do not respond to a given mutagen or which respond much and little. When forward mutations, induced by a chemical of known specificity, are used the consistency of the theory of mutations can thus be checked. In addition one can hope to determine for each or at least many mutations by which base pair change they came about; this would also enable one to infer the mutagenic mecha- nism for agents whose chemical specificity is not sufficiently understood, as well as for the spontaneous mutations.

1. Complications of the Analysis

The interpretation of experimental results concerned with the induc- tion of reverse mutations is complicated by two factors: the higher mutagenic specificity and suppressor mutations.

254 ERNST FREESE

a. The Huilwr M iiitunnir Spvnjiritii. 'V\w I'cactix'ity of a ^^ivcii nucleotide pair toward a mutagen dcjx'iids on its location in the J)NA, as we have already seen for the induction of forward mutations (Section V,A). Mutants containing the same type of base pair change at different sites may thus give ajipreciably dilTeicnt rates of reversion. However, this variable mutagenic si)ecificity can only influence the height of the mutagenic response and cannot make an A-T base pair rcs])ond to a mutagen that attacks only G or C, etc. If one finds two distinct classes of mutants, those that are much and others that are little (or not) inducible by a given mutagen, the corresponding mutations must be fundamentally different. This bipartition is most simply explained by assuming that the mutations of the two classes differ by the type of base pair cliange by which they arose; the higher mutagenic specificity would then only control the relative strength of the response within each class.

h. Suppressor Mutations. Since reverse mutations occur so rarely they can be observed only on selective media on which the original mutant cannot multiply. All strains which can grow under these selective conditions are called revertants of the original mutant. They may arise by "back mutations" to the standard genotype, i.e., the original sequence of nucleotide pairs in DNA, or by "suppressor mutations." The latter do not return the genome to the standard state but nevertheless restore the growth of the organism partially and sometimes even completely. Such suppressor mutations may occur in the same functional region as the original mutation (intragenic suppressor mutations) or in a tliffcrent functional region (extragenic suppressor inutations).

The extragenic suppressor mutations usually restore the growth of the organism only ])artially and then can be distinguished from back mutations on growth plates. Their presence is furthermore revealed readily by crosses of a revertant to the standard ty])e organism. These extragenic suppressor mutations may be non-si:)ecific or specific for the original mutational site; they have interesting i)i'operties which go beyond the scope of this paper. For the rll region with which we are here mainly concerned extragenic suppressor nuitations do not seem to exist. This follows from the fact that 10% of the spontaneous rll mutants do not revert at all, which excludes non-specific supi^ressors, and that specific extragenic suppressor nuitations have so far not been found.

I shall therefore limit the further discussion to back mutations and intragenic suppressor mutations. Let us consider the case in which the functional region contains tlu^ infoi'ination for a ])olypeptide that is the whole or a part of an enzyme. Some mutations may alter the enzyme such that the change of no amino acids in the polypeptide can repair the damage, except the restoration of the original amino acid sequence.

V. MOLECULAR MECHANISM OF MUTATIONS 255

Such a situation may hold, e.g., for a functionally important amino acid in the "active center" of the enzyme or the presence or absence of a proline which produces a kink in the protein. In these cases only a back mutation would return full functional activity. On the other hand, many mutations may cause a change of the enzyme which can be counteracted also by a second mutation at some other genetic site. Such a suppressor mutation may return the functional property of the enzyme partially or even completely. If a given mutagen should induce this second base pair change more frequently than the back mutation, most induced revertants would contain both the original mutation and an intragenic suppressor mutation.

The question now is, how can one distinguish intragenic suppressor mutations from back mutations? If the functional property is restored only to a limited extent this distinction is easy because one observes a different phenotype. For example, many rll mutants plated on bacteria K produce both standard type and tiny plaques ; at least the latter must be due to suppressor mutations. Profiavin mutants represent an extreme case since most revertants show plaques different from the standard phenotype. Practically all of these revertants are apparently caused by suppressor mutations and back mutations are very rare. Brenner et al. (1961) have proposed that profiavin induces only deletions or insertions of one base pair which rarely back mutate. Crick et al. (1961) have employed these mutations in order to provide some evidence for their coding theory.

The revertants can be examined further for their functional proper- ties by plating them on different bacterial hosts of type K and B; thus one can distinguish partial revertants from revertants with standard phenotype. Occasionally one observes revertant plaques so similar to the standard type that a clear distinction is difficult; in most cases such phages have not been used for further analysis.

The next step is to determine whether any of the remaining revertants of standard phenotype still contain both the original rll mutation and a suppressor mutation. To this end one backcrosses revertant to standard type phages and measures the frequency of rll mutants before and after the cross. Many such crosses have been made, leaving out all unclear cases of intermediate plaque types, and no evidence for further sup- pressor mutations was found in these cases (Freese, 1959b; Freese et al., 1961b; E. B. Freese, 1962). Thus it would seem that most revertants of genuine standard phenotype arose by back mutations. However, the significance of the backcrosses is limited by the fact that each standard type and revertant stock already contains spontaneous r mutations ; the crosses thus permit one to state only that the sujijiressor mutation cannot

256 ERNST FREESE

bo more distant from the original mutation than about 0.2 recombina- tion units. This is only a small portion of the rll region, which extends over about 8 recombination units. One could reduce this possible distance even further, by examining the frecjuency of the original rll mutant, among all r mutants, before and after the cross (by spot tests against the ileletion set) ; this has not been reported so far.

In another important analysis Benzer (1961) has isolated from a hot spot mutant a few revertants with standard phenotypc and determined a new forward mutational spectrum for each of them. If a revertant arose by a suppressor mutation it should give a different mutational spectrum, either leaving out the original hot spot or else, if the sup- pressor mutation itself is highly revcrtible, giving only the same hot spot and no other ones. In contrast about the same mutational spectrum was found as originally present.

Even this elaborate analysis does not uneciuivocally prove the existence of back mutations since it is possible, though unlikely, that hot spots correspond to a group of nucleotide pairs. As a further proof it seems desirable that a correlation be made between revertants and the amino acid sequence in the corresponding protein. Interestingly enough even that would not enable a complete proof of genuine back mutation since it is possible that the same amino acid can be coded by two differ- ent nucleotide combinations. Hence the ultimate proof can only come from a partial sequence analysis of the DNA around the mutational site.

In spite of the above complications several observations on the induction of reverse mutations are clear cut and demand an explanation. When a mutagen efficiently increases the frequency of revertants for certain mutants while it leaves other mutants untouched, it seems probable that the non-inducible mutants contain a base pair change which cannot be reversed by the particular mutagen. The inducible mutants can be further examined in order to decide which of the induced revertants do or do not seem to be caused by a suppressor mutation. If there are induced revertants, which by functional and recombination tests give no indication of a suppressor mutation, it seems reasonable to assume they arose by back mutations. Hence one can conclude that the particular base pair change of these mutants can be induced to back mutate by the mutagen. It does not matter, for this argument, whether none, few, or even many suppressor mutations are induced at the same time.

2. Transitions versus Other Changef;

The specificity of the induction of reverse mutations shows up most effectively for the difference between transition and non-transition mu-

V. MOLECULAR MECHANISM OF MUTATIONS 257

tants. For this investigation one can use mutagens, which are expected to induce only base pair transitions, but those in both directions (see Section III) ; such agents are AP, BUdR, and NA. AH transition mutants should be inducible to revert while non-transition mutants should not respond. The two base analogs are especially useful for these studies because simple spot tests can be employed to screen a large number of mutants; such spot tests have been described for bacteria (Szybalski, 1958) and phages (Freese, 1959b). The reversion inducibility by the base analogs has been determined for many rll mutants and the results are given in Table III (p. 239). As expected, most of the mutants induced by the transition-inducing agents AP, BUdR, HA, and NA can be induced to revert by the base analogs. Several of them have also been tested for reversion inducibility by nitrous acid and the result was positive (E. B, Freese and E. Freese, 1961). The non-inducible mutants correspond to the background of spontaneous revertants in the phage stocks.

In contrast most of the spontaneous and proflavin-induced mutants do not respond in this way. It seems clear that most of these non- base analog inducible mutants must have been caused by base pair changes different from transitions. Also about one-third of the mutants induced by low pH or EES are of this type. It is significant that for these mutants neither back mutations nor suppressor mutations are base analog inducible (except for a few mutants whose spontaneous reversion rates are so high that induction cannot be observed).

Two-thirds of the mutants, induced by low pH and EES, can be induced to revert by base analogs; they may contain a transition, or possibly some other base pair change if the base analogs should only induce suppressor mutations for some of these mutants. But EES and low pH certainl}^ can induce transitions; this is also shown by their ability to induce the reversion of transition mutants (see Section V,4).

3. Further Subdivision of Non-Transition Mutants

The mutants of Table III that cannot be induced to revert by the two base analogs nevertheless revert spontaneously and genetic mapping shows that they are point mutations. The question is which of these mutants contain a transversion, a deletion, or an insertion?

To answer this question, at least in part, one can measure the rever- sion induction by the two mutagens, low pH and EES. About 70% of the rll mutants, induced by these agents, can be induced to revert by the base analogs; therefore many of them seem to contain a transition. We have proposed before (Section III,D) that the gap produced by the depurination of DNA could result in both transitions and transversions, or in deletions. Since low pH and EES can induce transitions they

258 ERNST FREESE

TABLE IV Rates a/;3 of Reversion Induction for Non-Transition Mutants

Spontaneous backgound of _^

Mutant With pH 4. 2 With EES With nitrous acid revertants (10 )

r HI 66 70 S3 160

r 207 57 55 :S2. 5 69

r 132 55 50 S2 197

r 157 40 50 S5 175

r 315 5.0 4. 5 50.2 5.0

y 102 2.2 3.8 ;:;0. 1 15

r 105 2.4 3.5 50.1 6. 1

f 179 2.4 3.4 SO. 05 6.6

EES 66 4.0 4.5 50.1 5.3

EES 116 2.9 3.3 50.1 7.8

P 11 ~1.3 50.8 50.2 ~4.0

r 114 50.2 50.1 50.1 0.7

r 131 50.06 <0. 1 <0. 1 0,4

r 117 <0. 05 <0. 1 <0. 1 0.6

r 194 50.04 50.1 — 0.5

r 178 50.01 50.01 50.01 0.02

r 123 50.01 50.01 — 0.02

r 289 50.01 50.01 _ 0.1

EES 122 ;50. 3 <0. 2 _ 1.1

EES 64 50.1 ;c;0. 2 _ 1.3

EES 32 50.02 50.2 _ 0.3

P 12 50.1 _ _ 0.5

P 5 50.05 50.05 _ 0.9

P 14 50.04 _ _ 0.1

P 8 50.03 50.01 _ 0.02

P 13 50.03 — — 0.3

P 33 50.03 — — 0.2

P 3 50.02 50.02 — 0.02

P 4 50.01 50.02 - 0.08

P 36 50.01 — — 0.2

'From E.B. Freese (1!)61) a/P, rate of increase of the frequency of revertants per viable phage and per lethal hit, in units lO"^; r , spontaneous mutants; EES, ethyl ethane sulfonate- induced mutants; P, proflavin-induced mutants;— , not measured; 5 , smaller thEin or about equal to.

		revertants per
	■ viable phage
	in units
	• of 10'* / /
2.0	-	/ N 29 / H 24
	.	/
	/
	' /
	/ ./
	; /•
	■ ./ /
1.0	- / /• ■ / i	1
0	1 10'' phage survival	10-
	0 1 2^3 igihQi hits ^

4,:

time of treatment in min.

Fig. 17A. The induction of reverse mutations. Reversion induction of BU induced mutants by nitrous acid (see E. B. Freese and E. Freese, 1961).

Fig. 17B. The induction of reverse mutations. Reversion induction of an EES- iuduced mutant (66) by EES and low pH. This mutant does not respond to nitrous acid (NA) or to hydroxylamine (HA) (see E. B. Freese, 1961).

259

200 ERNST FREESE

should also be able to induce transvcrsioiis. W'iiether they induce dele- tions is unknown.

The non-transition nuitants were therefore tested for their reversion inducibility by EES or low pH. The results of such tests are summarized in Table IV (E. B. Freese, 1961). The values a/p in this table are the slopes of the straight line shown in Fig. 17. Each slope has been deter- mined by several points. Several spontaneous and EES induced mutants show a small increase in the frequency of revertants per viable phage, with both low pH and EES, while no such increase was observed with nitrous acid. Several induced revertants of the two EES-induced nm- tants, EES 66 and EES 116, were examined further and behaved in functional tests as well as in backcrosses as if they arose by back muta- tions (E. B. Freese, 1962). It thus seems likely that the non-transition mutants which are inducible to revert by low pH and EES have a trans- version at their mutant site.

The strongly responding nuitants of Table IV probably carry a G-C pair at their mutant site, because EES and low pH seem to remove G much more frequently than A under the conditions used for the induc- tion of mutations. The G-C pair could stem either from a T-A or a C-G pair in the standard DNA. But whether the less strongly respond- ing mutants also have a G-C pair and just react less because of some higher mutagenic specificity or whether they have an A-T pair is not clear.

Some spontaneous and most of the proflavin-induced mutants cannot be induced to revert by EES or low pH. This is complemented by the observation of Orgel and Brenner (1961) that most proflavin- or 5- aminoacridine-induced and some spontaneous mutants can be induced to revert by proflavin or 5-aminoacridine, while BU induced mutants cannot be induced by these agents to revert. If the hypothesis of Brenner et al. (1961) is correct, mutants inducible by acridine dyes should con- tain a deletion or insertion of one nucleotide pair. So far the experi- mental results do not permit a decision between this possibility and the alternative that these mutants contain a transversion with an A-T pair.

4. The Preferred Midogenie Direction of Transition Inducers

Some transition-inducing mutagens react preferentially or predom- inantly with one of the DNA bases. It should be possible to utilize these agents in order to subdivide transition mutants further according to A-T and G-C pairs at the mutant site. Thus nearly all forward mutants induced by HA and most of the transition mutants induced by

G A EES or low i)H should contain a base pair change -^ .

C T

V. MOLECULAR MECHANISM OF MUTATIONS 261

The directional effect of these mutagens can be observed when the rates of induced r(>version are deterniined for many transition nuitants. If one obtains a bii^artition into highly and weakly inducible mutants and if the results with different mutagens agree with the expectations, it seems likely that the hypothesis about the direction of base pair changes is correct.

It should be remembered, however, that the frequency of revertants is deter- mined by a selective technique. The mutagenic chemical alters only one base in a pair, and the resulting hybrid may be able to act functionally hke a rever- tant either immediately or only after one or two DNA duplications. Hence the result of measurements on induced reversions depends on the number of DNA duplications which the selective method allow^s after the first base has been altered. This functional expression may depend, furthermore, on the particular mutant (which of the two DNA strands is attacked, the functional or the other one) and on the mutagen (how much the altered base differs functionally from a normal one). The rate of reversion induction has, therefore, been measured in two ways, at least for nitrous acid (E. B. Freese and E. Freese, 1961) and hydroxylamine (Freese et al., 1961b). On the one hand, the chemically treated T4-r// phages were plated directly on the selective bacteria K (rate a//3). On the other hand, they were preadsorbed onto (UV-killed) bacteria B in which all live phages can replicate about five times; the infected bacteria were then plated on the selective bacteria K (rate a'/fi). The results in Table V show that for HA plating with preadsorption merely enhances the values found for direct plating, while for nitrous acid some rates, which were smaller than others for direct plating, were much larger after preadsorption. The cause of these differ- ences has yet to be resolved.

The rates a/fS are quite different when the same mutant is induced to revert by different mutagens. This reflects the varying degree of the mutagenic versus the lethal effect. The lethal effect is especially small for HA. If the lethal effect is large, as for EES, the difference between highly and weakly inducible mutants becomes much less pronounced.

Different mutants show greatly different rates of reversion induci- bility by a given mutagen. With HA as mutagen a clear bipartition has been observed into highly and weakly (or not at all) inducible mutants; the a/(i values of these two classes differ by more than a factor 100 (see Table V). It should be mentioned that one more AP mutant (AP 12) has been found highly inducible but has not been included in Table V for reasons given later. The same AP mutants that are most inducible by HA are also most inducible by all other agents used (though for nitrous acid only after preadsorption). But with these other agents the bipartition is not as clear-cut. It is improbable that the higher nmta- genic specificity can be responsible for the observed bipartition since HA (and nitrous acid) attack C while low pH and EES attack G and yet the results agree. Thus, most of the rarely or non-inducible mutants have an A-T pair at their mutant site since they have been induced by

202

ERNST FREESE

TABLE V Rates of Reversion Induction for Base Analog Inducible Mutants

			With		With	With	Background of
	With HA	nitrous	acid	pH4. 2	EES	spontaneous
Mutant No.	a/^	a'/ (3'	a/,i	a!/^'	a/p	a/ ,3	revertants (ICT* )
AP 114	450	~1040	0.03	160	8.0	1.4	5
AP 275	280	240	0.24	40	7.6	1.5	2
AP 72	134	158	0.03	80	5.2	1.0	1
AP 41	~ 62	~ 265	-0.03	- 28	-1.2	-2.3	-0.3
N 24	0.2	;S10	0.64	€4	;30. 1	;S0. 01	0.48
N 29	0.1	S 1	1.2	S5	0.01	0.03	0.013
HA 54	0.1	5 3	—	-	-	-	0.08
AP 61	0.08	Sl8	0.33	S4	SO. 03	SO. 01	0.13
N 19	0.06	S 6	0.16	S6	so. 01	sO. 02	0.009
HA 45	0.06	:£ 4	—	—	-	-	0.01
N 7	0.04	S 2	0. 15	S8	so. 01	so. 01	0.02
HA 11	0.04	S 8	-	-	-	—	0.02
N 12	0.03	S 2	0.21	S7	—	SO. 01	0.03
N 17	0.03	—	0. 15	—	so. 01	so. 01	0.07
N 31	0.03	£ 6	0.24	—		so. 01	0.028
HA 25	so.ooi	< 0.1	—	—	—	—	0.003
AP 156	SO. 002	S 5	0.03	S4	0.01	0.08	0.007
AP 70	SO. 01	S 6	0.01	Si	0.01	0.07	0.02
AP 83	SO. 005	—	0.07	S2	0.01	0.09	0.01

From Freese (1961)

q/^, rate of increase of the frequency of revertants per viable phage and per lethal hit, in units 10"*, measured by directly plating phages on bacteria, E. coli B (all viable phages give plaques) and on K (selective, rll mutants do not, but revertants do, grow). a/^ is the same rate but the treated phages were preadsorbed to UV killed bacteria B (survival 10"', and the infectd centers plated on K; .'\P, HA, and N mutants induced by AP, HA, and Bl', respectively;— , not measured;^ , smaller than or about equal to.

agents that produce transitions. For this conclusion it is unessential that some of the little in(iucible mutants show induced suppressor muta- tions, as for example, AP 70, which gives after EES induction only revertants that form big standard type plaques on bacteria B.

One would like to argue that the highly inducible mutants have a G-C pair at their mutant site and that the above agents can induce their back mutations. Here the problem of suppressor mutations does play an important role in the interpretation. Functional tests (on K and B) of induced revertants have shown that for mutants AP 72 and AP 275 (see Table V) most revertants have standard phenotype; for AP 41 both standard type revertants (on K and B) and others with tiny plaques on K (and plaques on B) have been observed; for AP 114 less

V. MOLECULAR MECHANISM OF MUTATIONS 263

than 50% of the revertants gave standard phenotype while the rest gave large fuzzy plaques on bacteria B. The important point is that each of these four mutants does yield at least some induced revertants of standard phenotype. When these revertants were backcrossed to standard type phages the frequency of r mutants observed in the progeny was not significantly higher than in the stocks used (usually smaller than 0.1%, Freese et al., 1961b, E. B. Freese, 1962). As outlined before these results make it likely that the revertants of standard phenotype arose by genuine back mutations. The four AP mutants in Table V must then have a G-C pair at their mutant site.

The mutant AP 12, although highly inducible, has been omitted from Table V because a functional test revealed that all induced revertants produce r plaques on bacteria B and the same was observed with another mutant of this genetic site. Revertants of AP 12, induced by the above agents, may arise only or mainly by suppressor mutations.

The data in Table V indicate that BU induces transitions from GC into AT more frequently then in the opposite direction ; this has already been explained earlier (Section III,B,1). In contrast, AP seems to induce transitions in both directions with comparable frequency. Actually all

G A other transition inducing agents predominantlv induce changes -^ .

C T The observed directional effect of the base analogs complements the earlier observation that AP is more efficient than BUdR in inducing the reversion of BU-induced mutants (Freese, 1959b) ; the opposite state- ment concerning AP mutants seems no longer justified. Similar direc- tional effects for induced reversions have been observed in Salmonella typhimurium (Margolin and Mukai, 1961) when the frequency of AP- induced revertants was measured on plates with minimal and with suboptimally enriched medium. Again, kinetic studies, on the time and mode of appearance of bacterial revertants induced by BUdR or by AP, have been used to identify the mutant base pair in a transition mutant fRudner, 1961; Strelzoff, 1962).

5. Specificity of Spontaneous Mutations

The preceding analysis on the induction of mutations also provides infoniiation about the specificity of spontaneous mutations. The result will be summarized for the mutations of phage T4:

1. About 80% of the spontaneous mutations are of the non-transition type; the rest apparently arose by transitions (see Table III).

2. Some of the spontaneous non-transition mutants, including some hot spots, seem to cany a transversion (see Table IV). The other non-

264 ERNST FREESE

transition luutaiits that are not rcN'cisihIc hy I'll^S oi- low |)I1 may have been caused by transversions, (lelclions, or inscitioiis.

3. Both transition and transversion mutants which aic \ciy suscepti- ble to reversion by agents preferentially attacking G or C also fic(|ueiitly revert spontaneously. It seems likely, therefore, that (!-(' pairs mutate spontaneously more freciuently than A-T pairs.

It must be emphasized that these tentative conclusions arc only valid for the particular phage system examined. There is no justification for a general statement about the frequency or the specificity of spontaneous mutations in all organisms. One could argue that G and C are chemically more vulnerable than A or T; but even that is true only for most of the directly acting chemical agents and wrong for certain ones (e.g., perox- ides, hydrazine). Besides, we do not know whether in a particular organism most mutations arc caused by mistakes during DNA duplica- tion or by chemical changes of DNA. A typical example for the differ- ences of similar organisms has been reported by Kirchner (1960) for Salmonella; he found that for one strain (LT2) 26% of the spontaneous mutants could be induced to revert by AP or BUdR while for another strain (LT7) 95% of the spontaneous mutants could be induced to revert by these base analogs. The second strain differs from the first one by a gene which increases the frequency of spontaneous mutations. The result indicates that the mutator gene produces a substance that induces transitions.

References

Alexander, P. (1952). Nature 169, 226.

Alexander, P., and Stacey, K. A. (1958). Ann. NY. Acad. Set. 68, 1225.

Alper, H., and Ebert, M. (1954). Science 120, 608.

Andrew, L. E. (1959). Am. Naturalisl 93, 135.

Auerbach, C. (1951). Cold Spring Harbor Sympo.'<ia Quatit. Biol. 16, 199.

Auerbach, C, and Moser, H. (1950). Nature 166, 1019.

Auerbach, C, and Moser, H. (1953). Z. Induktirr Ahslannnung.s-u. Vererbungs-

lehre 85, 479, 547. Auerbach, C, and Robson, J. M. (1946). Nature 157, 302. Baker, W. K. (1955). Brookhaven Synipo.^ia in Biol. 8, 191. Baron, F., and Brown, D. M. (1955). J. Chem. Soc. p. 2855. Bauch, R. (1948). Planta 35, 536.

Bautz, E., and Freese, E. (1960). Proc. Natl. Acad. Sci. U.S. 46, 1585. Bayley, C. R., Brammrr, K. W., and Jones, A. S. (1961). J. Chcm. Soc. p. 1903. Beers, R. F., Hendley, D. D., and Steiner, R. F. (1958). Nature 182, 242. Benzer, S. (1955). Proc. Natl. Acad. Sci. V.S. 41, 344. Benzer, S. (1957). In "The Chemical BasLs of Heredity" (W. D. McElroy and B.

Glass, eds.), p. 70. Johns Hopkins Press, Baltimore. Benzer, S. (1961). Proc. Natl. Acad. Sci. U.S. 47, 403. Benzer, S., and Freese, E. (1958). Proc. Natl. Acad. Sci. U.S. 44, 112.

V. MOLECULAR MECHANISM OF MUTATIONS 265

Beukers, R., and Beronds, W. (1960). Biochem. el Biophys. Acta 49, 181.

Bradley, D. F., and Wolf, M. K. (1959). Proc. Natl. Acad. Sci. U.S. 45, 944.

Brenner, S., and Smith, J. D. (1959). Virology 8, 124.

Brenner, S., Benzer, S., and Barnott. L. (1958). Nature 182, 983.

Brenner, S., Barnett, L., Crick, F. H. C, and Orgel, A. (1961). J. Mol. Biol. 3, 121.

Brookes, P., and Lawley, P. D. (1960a). J. Chem. Soc. p. 539.

Brookes, P., and Lawley, P. D. (1960b). Biochem. J. 77, 478.

Brown, D. M., and Schell, P. (1961). J. Mol. Biol. 3, 709.

Cartledge, J. L., and Blakeslee, A. F. 1934. Proc. Natl. Acad. Sci. U.S. 20, 103.

Chase, M., and Doermann, A. H. (1958). Genetics 43, 332.

Clifton, C. E. (1931). Proc. Soc. Exptl. Biol. Med. 28, 745.

CoughHn, C. A., and Adelberg. E. A. (1956). Nature 178, 531.

Crick, F. H. C, Barnett. L., Brenner, S., and Watts-Tobin, R. J. (1961). Nature

192, 1227. D'Amato. F. (1950). Caryologia 2, 229.

Darwin, C. (1859). "The Origin of Species by Means of Natural Selection." De Giovanni, R. (1960). Ann. N .Y . Acad. Sci. 84, Art. 16. De Giovanni, R. (1961). Z. Vererbungslehre 92, 389. De Mars, R. I. (1953). Nature 172, 964. De Mars, R. I., Luria, S. E., Fisher, H., and Levinthal, C. (1953). Arm. inst. Pasteur

84, 113. Demerec, M. (1937). Genetics 22, 469. Demerec, M. (1953). Symposia Soc. Exptl. Biol. 7, 44. Demerec, M. (1960). Proc. Natl. Acad. Sci. U£. 46, 1075. Demerec, M., and Hanson, J. (1951). Cold Spring Harbor Symposia Qua)it. Biol.

16, 215. Demerec, M., and Hartmann, P. E. (1959). Ajiu. Rev. Microbiol. 13, 377. Demerec, M., and Sams, J. (1960). Proc. Symp. Venice, 1959, p. 283. Intern. J.

Radiation Biol. Suppl. p. 283. De Serres, F. J. (1958). Genetics 43, 187.

De Vries, H. (1909). "The Mutation Theory." Open Court, LaSalle, Illinois. Deysson, M. (1952). Compt. rend. acad. sci. 234, 650. Dickey, F. H., Cleland, G. H., and Lotz, C. (1949). Proc. Natl. Acad. Sci. U^. 35,

581. Djordjevic, B., and Szybalski, W. (1960). J. Exptl. Med. 112, 509. Doermann, A. H., Chase, M., and Stahl, F. W. (1955). J. Cellular Comp. Physiol.

45 (Suppl. 2), 51. Doty, P., Marmur, J., and Sneoke, N. (1959). Brookhaven Symposia in Biol. 12, 1. Dunn, D. B., and Smith, J. D. (1954). Nature 174, 304. Dunn, D. B., and Smith, J. D. (1955). Nature 175, 336. Eidinoff, M. L., Cheong, L.. and Rich, M. A. (1959). Science 129, 1550. Ellmauer, H., and Kaplan, R. W., 1959. Natunnss. 46, 150. Endo, H. W., Miura, K., Hidaka, F., and Hiruk, C. (1961). Nature 190, 833. Fahmy, O. G., and Fahmy, M. J. (1955). J. Genet. 53, 563. Fahmy, O. G., and Fahmy, M. J. (1957). Nature 180, 31. Fahmy, O. G., and Fahmy, M. J. (1958). Discussion to the paper on the "Mutagenic

Effects of Alkylating Agents," presented by C. Auerbach in Ann. N.Y. Acad.

Sci. 68, 736. Fischer, R., and McKinley, E. B. (1927). /. Infectious Dwco.se.s 40, 399. Fosse, R., HieuUe, A., and Bass, L. W. (1924). Compt. rend. acad. sci. 178, 811.

266 ERNST FREESE

Fraenkel-Coniat , H. (1954). Biochim. el Biophys. Ada 15, 307.

Frascr, D. K. (1957). Virology 3, 527.

Frcdoriksen, S., and Klonow, H. (1960). Biochem. Biophys. Research Communs. 2, 329.

Freese, E. (1959a). J. Mol. Biol. 1, 87.

Freese, E. (1959b). Proc. Natl. Acad. Sci. UJS. 45, 622.

Freese, E. (1959c). Brookhaven Symposia in Biol. 12, 63.

Freese, E. (1961). Proc. 5th Intern. Congr. Biochem., Moscow, 1961.

Freese, E., Freese. E. B., and Bautz, E. (1961a). J. Mol. Biol. 3, 133.

Freese, E., Freese, E. B., and Bautz, E. (1961b). Proc. Natl. Acad. Sci. U.S. 47, 845.

Freese, E. B. (1961). Proc. Natl. Acad. Sci. U.S. 47, 540.

Freese, E. B. (1962). Biochem. Biophys. Research Communs. 7, 18.

Freese, E. B., and Freese, E. (1961). Virology 13, 19.

Fresco, J. R., and Alberts, B. M. (1960). Proc. Natl. Acad. Sci. U.S. 46, 311.

Fricke, H., and Demerec, M. (1937). Proc. Natl. Acad. Sci. U£. 23, 320.

Friedewald, W. F., and Andenson. R. S. (1940). Proc Soc. Exptl. Biol. Med. 45, 713.

Gard, S., and Maal0e, O. (1959). In "The viruses" (F. M. Burnet and W. M. Stan- ley, eds.). Vol. 1, Chapter V, pp. 359-418. Academic Press, New York.

Geiduschek, E. P. (1961). Proc. Natl. Acad. Sci. U.S. 47, 950.

Giles, N. H. (1955). Brookhaven Symposia in Biol. 8, 103.

Giles, N. H., De Serres, J., and Partridge, C. W. H. (1955). Ann. N.V. Acad. Sci. 59, 536.

Glaess, E. (1956). Chromo.'<oma 8, 260.

Glass, B. (1955). Brookhaven Syynposia in Biol. 8, 148.

Glass, E. A., and Novick, A. (1959). ./. Bacteriol. 77, 10.

Gottschling, H.. and Freese, E. (1961). Z. Naturforsch. 16b, 515.

Green, D. MacD., and Krieg, D. (1961). Proc. Natl. Acad. Sci. U.S. 47, 64.

Greer, S. B. (1958). J. Gen. Microbiol. 18, 544.

Greer, S. B. (1960). J. Gen. Microbiol. 22, 618.

Greer, S. B., and Zamenliof, S. (1962). J. Mol. Biol. 4, 123.

Gutz, H. (1961). Nature 191, 1124.

Haas, F. L., and Doudney, C. 0. (1957). Proc. Natl. Acad. Sci. U.S. 43, 871.

Hadorn, E., and Niggli, H. (1946). Nature 157, 162.

Hakala, M. T. (1959). Biol. Chcm. 234, 3072.

Handschumacher, R. E., and Welch, A. D. (1960). /// "Tlic Xu<-I<mc Acids" (E. Chargaff and J. N. Davidson, eds.), Vol. 3, pp. 453-526. Academic Press, New York.

Harm, W. (1960). Z. Vererbungslehrc 91, 52.

Heiner, R. E., Konzak, C. F., Nilan, R. A., and Logault, R. R. (1960). Proc. Natl. Acad. Sci. U.S. 46, 1215.

Helinski, D. R., and Yanofsky, C. (1961). Proc. Natl. Acad. Sci. U.S. 48, 173.

Hershey, A. D., and Chase, M. (1952). ./. Gen. Physiol. 36, 39.

Hiatt, C. W. (1960). Trans. N.Y. Acad. Sci. 23, 66.

Hirota, Y. (1960). Proc. Natl. Acad. Sci. U.S. 46, 57.

Hirota, Y., and Tijima, T. (1957). Nature 180, 655.

Hollaender, A., and Kimball, R. F. (1956). Nature 111, 726.

Iyer, V. N., and Szybalski, W. (1958). Proc. Natl. Acad. Sci. US. 44, 446.

Iyer, V. N., and Szybalski, W. (1959). Science 129, 830.

Jacob, F. (1954). Compt. rend. acad. sci. 238, 732.

Jensen, K. A., Kolmark, G., and Westergaard, M. (1949). Ilereditas 35, 521.

V. MOLECULAR MECHANISM OF MUTATIONS 267

Jensen, K. A., Kirk, J., Kolmark, G., and Westergaanl, M. (1951). Cold Spring

Harbor Sy7nposia Quant. Biol. 16, 245. Kanazir, D. (1958). Biochcm. et Biophys. Acta 30, 20. Kaplan, M. S., and Tomlin, P. A. (1960). Radiation Research 12, 447. Kaudewitz, F. (1959). Nal^ire 83, 1829; Z. Naturfor.sch. 146, 528. Kaufman, E., and Hiatt, C. W. (1959). Virology 9, 478. Kihlman, B. (1955). Exptl. Cell Research 8, 345. Kihlman, B. (1960). Exptl. Cell Research 20, 657.

Kirby-Smith, J. S., and Craig, D. L. (1955). Records Gen. Soc. Am. 24, 579. Kirchner, C. W. J. (1960). J. Mol. Biol. 2, 331. Koch, A. L., and Lamond, W. A. (1956). J. Biol. Chem. 219, 189. Kolmark, G., and Giles, N. H. (1955). Genetics 40, 890. Kolmark, G., and Westergaard, M. (1953). Hereditas 39, 209. Konrad, M. (1960). Ann. N.Y. Acad. Sci. 84, 678. Kornberg, S. R., Zimmerman, S. B., and Kornberg, A. (1961). ./. Biol. Chem. 236,

1487. Kozinski, A. W., and Szybalski, W. (1959). Virology 9, 260. Krieg, D. R. (1959a). Virology 8, 80. Krieg, D. R. (1959b). Virology 9, 215. Lamarck, J. B. (1809). "Philosophie zoologique." Latarjet, R. (1949). Compt. rend. acad. sci. 228, 1354. Latarjet, R. (1954). Acta Unio Intern, contra Cancrum 10, 136. Lea, D. E. (1955). "Action of Radiations on Living Cells," 2nd ed. Cambridge Univ.

Press, London and New York. Lehman, J. R., and Pratt, D. (1960). /. Biol. Chem. 235, 3254. Lerman, L. S. (1961). /. Mol. Biol. 3, 18.

Lett, J. T., Alexander, P., and Parkins, G. M. (1962). Arch. Biochem. Biophys. 97, 80. Leuppold, V. (1962). Arch. Juliiis-Klaus-Stift. Vererbungsforsch. In preparation. Levan, A. (1951). Cold Spring Harbor Symposia Quant. Biol. 16, 233. Litman, R. M., and Ephrussi-Taylor, H. (1959). Compt. rend. acad. sci. 249, 838. Litman, R. M., and Pardee, A. B. (1956). Nature 178, 529. Litman, R. M., and Pardee, A. B. (1959). Virology 8, 125.

Litman, R. M., and Pardee, A. B. (1960). Biochem. et Biophys. Acta 42, 117, 131. Littlefield, J. W.. and Dunn, D. B. (1958). Biochem. J. 70, 642. Lorkiewicz, Z., and Szybalski, W. (1960). Biochem. Biophys. Research Communs.

2, 413. Lorkiewicz, Z., and Szybalski, W. (1961). /. Bacteriol. 82, 195. Loveless, A. (1951). Nature 167, 338. Loveless, A. (1959). Proc. Roy. Soc. B150, 486 and 497. Luria, S. E. (1947). Proc. Natl. Acad. Sci. U.S. 33, 253. Luria, S. E., and Delbriick, M. (1943). Genetics 28, 491. McKinley, E. B., Fischer, R., and Holden, M. (1926). Proc. Soc. Exptl. Biol. Med.

23, 408. Margohn, P., and Mukai, F. H. (1961). Z. Vererbnngslchre 92, 330. Marinelli, L. D., and Taylor, L. S. (1954). Radiation Biol. 1, 145. Marmur, J., Schildkraut, C. L., and Doty, P. (1961). J. dc Chimie Physique p. 945. Marquardt, H. (1938). Z. Botan. 32, 401. Mashima, S., and Jkeda, Y. (1958). /. Appl. Microbiol. 6, 45. Mendel, G. (1865). Verhandl. naturforsch. Ver. Briinn, Abhandl. VI. Meselson, M., and Stahl, F. W. (1958). Proc. Natl. Acad. Sci. UjS. 44, 671.

268 ERNST FREESE

Milrs, H. T. (1%1). Proc. Xntl. Acad. Sci. l\S. 47, 791.

Miyako. T. (11)50). Genelics 45, 1, 11.

Mullor, H. J. (1927). Science 66, 382.

Mundry, K. W., and Giorer, A. (1958). Z. Vcrerbungslehre 89, 614.

Nasrat, G. E., Kaplan, W. D.. and .Vuorhach, C. (1954). Z. hiduklivc Ahslntn-

mungs-u. Vercrhitngshhre 86, 249. Novick. A. (1956). Brnnkhnven Symposia in Biol. 8, 201. Oplilk(M-.-<, F. (1943). Z. Indiiktirr Ahstammiings-it. Vcrerbungslehre 81, 313. Oehlkers, F. (1949). Ileidelherger Akad. Wiss. 9, 373. Oehlkers, F. (1953). Heredity 6 (Suppl.), 95. Orgel, A., and Brenner, S. (1961). /. Mol. Biol. 3, 762. Plough, H. H. (1941). Cold Spring Harbor Symposia Quant. Bud. 9, 127. Pomper. S., and Atwood, K. C. (1955). Radiation Biol. 2, 431. Pratt, D., and Stent, G. S. (1959). Proc. Natl. Acad. Sci. l'.S. 45, 1507. Price, C. C. (1958). .1////. AM'. Acad. Sci. 68, 663. Putnam. F. W. (1953). /// "Tlie Proteins" (H. Xenrath and K. Bailey, eds.), Vol. I,

Part B. 11. 893. Academic Press, New York. Radiation Biol. (1954. 1955). Radiation Research (1959). Supi)l. 1.

Rapoport, J. A. (1946). Doklady Akad. Xauk. S.S£.R. 54, 65. Rapoport, J. A. (1947a). Doklady Akad. Nauk. S.S.S.R. 56, 12. Rapoport, J. A. (1947). Doklady Vsesoyuz. Akad. Sel'sko-khoz Naukijn. V. I.

Lenin a 12, 12. Rapoport, J. A. (1948a). Doklady Akad. Xauk. S.S.S.R. 59, 1183. Rapoport, J. A. (1948b). Doklady Akad. Nauk. SJiJ^.R. 60, 469. Reiner, B., and Zamenhof, S. (1957). /. Biol. Chem. 228, 475. Rogers, S. (1957). J. Exptl. Med. 105, 279. Ross, W. C. J. (1958). Ann. N .Y . Acad. Sci. 68, 669. Rudner, R. (1960). Biochem. Biophys. Research Communs. 3, 275. Rudner, R. (1961). Z. Vererbung.dehre 92, 336 and 361. Ryan, F. J. (19.59). J. Gen. Microbiol. 21, (No. 3), 530. Ryan, F. J. (1962). In "Methodology in Basic Genetics" (W. J. Burdette, ed.),

Hoklen-Day, Inc., San Francisco. Ryan. F. J., Nakada, D., and Sclmeid.T, M. .T. (1961). Z. Vcrerbungslehre 92, 38. Sauerbier, W. (1960). Nature 188, 327. Sax, K. (1937). Genetics 22, 523.

Sax, K. (1941). Cold Spnng Harbor Syrnposia Quant. Biol. 9, 93. Sax, K. (1957). Quart. Rev. Biol. 32, 15.

Schuster, H. (1960). Biochem. Biophys. Research Communs. 2, 320. Schuster, H. (1961). J. Mol. Biol. 3, 447.

Schuster, H., and Schramm, G. (1958). Z. Naturfor.^ch. 136, 697. Shugar, D. (1960). In "The Nucleic Acids" (E. Chargaff and J. N. Davidson, eds.),

Vol. 3, Chapter 30, p. 39. Academic Press, New York. Sinsheimer, R. L. (1955). J. Biol. Chem. 215, 579. Sobels, F. H. (1956). Nature 177, 979. Spencer, M. (1959). Acta Cryst. 12, 59, 66. Stacey, K. A., Cobb, M., Cousens, S. F., and Alexander, P. (1958). Ann. N .Y . Acad.

Sci. 68, 682. Stadler, L. J. (1928). Science 68, 186. Staehlin, M. (1958). Biochem. et Biophys. Acta 29, 410. Stahl, F. W. (1959). /// "The Viruses" (F. M. Burnett and \V. M. Stanley, ed.s.),

Vol. 2, Chapter X, pp. 353-384. Academic Press, New York.

V. MOLECULAR MECHANISM OF MUTATIONS 269

Stahl, F. W., Crascmann, J. M,, Okun, L., Fox, E., and Laird, C. (1961). Virology

13, 98. Steiner, R. F., and Boers, R. F. (1958). Science 127, 335. Stent, G. S. (1958). Advances in Virus Research 5, 95. Stevens, M. A., Magrath, D. J., Smith, H. W.. and Brown, G. B. (1958). J. Biol.

Chem. 233, 1513. Stone, W. S. (1955). Brookhavcn Symposia in Biol. 8, 171.

Stone, W. S., Wyss, O., and Haas, F. (1947). Proc. Natl. Aca<l. Sci. U.S. 33, 59. Strauss, B. S. (1962). J. Bacterial. 83, 241. Strauss, B. S., and Okubo, S. (1960). J. Bacteriol. 79, 464. Strelzoff, E. (1962). Ph. D. Thesis, Columbia University, New York. Swanson, C. P., and Stadler, L. J. (1955). Radiation Biol. 2, 249. Szybalski, W. (1958). Ann. N.Y. Acad. Sci. 76, 475. Szybalski, W., and Djordjevic, B. (1959). Genetics 44, 540. Szybalski, W., Opara-Kubinska, Z., Lorkiewicz, Z., Ephrati-Elizur, E., and Zamen-

hof, S. (1960). Nature 188, 743. Takemura, S. (1957). J. Biochem. 44, 321. Takemura, S. (1959). Bull. Chem. Soc. Japan 32, 920. Tamm, C, Hodes, M. E., and ChargafT, E. (1952). J. Biol. Chem. 195, 49. Tamm, C, Shapiro, H. S., Lipschitz, R., and Chargaff, E. (1953). /. Biol. Chem.

203, 673. Tessman, E. S. (1956). Virology 2, 679. Tessman, I. (1959). Virology 9, 375. Thoday, J. M., and Read, J. (1947). Nature 160, 608. Timofeeff-Ressovsky, M. W., Zimmer, K. G., and Delbriick, M. (1935). Nachr. Ges.

Wiss. Gottingen Math. Physik. Kl. Fachgruppen [VI] 1, 189. Treffers, H. P., Spinelli, V., and Belser, X. D. (1954). Proc. Natl. Acad. Sci. U.S.

40, 1064. Vielmetter, W., and Schuster, H. (1960). Biochem. Biophys. Research Communs. 2,

324. Vielmetter, W., and Wieder, C. M. (1959). Z. Naturforsch. 14b, 312. Vogt, M. (1950). Z. Induktive Absta7n7fiungs-u. Vererbungslehre 83, 324. Wacker, A., Kirschfeld, S., and Trager, L. (1960a). J. Mol. Biol. 2, 241. Wacker, A., Dellweg, H.. and Weinblum, D. (1960b). Naturwiss. 47, 477. Watson, J. D., and Crick, F. H. C. (1953). Cold Spring Harbor Symposia Quant.

Biol. 18, 123. Weigle, J. J. (1953). Proc. Natl. Acad. Sci. L.S. 39, 628. Weigle, J. J., and Dulbecco, R. (1953). Experientia 9, 372. Westergaard, M. (1957). Experientia 13, 224.

Weygand, F., Wacker, A., and Dellweg, H. (1952). Z. Naturforsch. 7b, 19. Wilkins, M. H. F. (1956). Cold Spnng Harbor Symposia Quant. Biol. 21, 75. Witkin, E. M. (1947). Cold Spring Harbor Symposia Quant. Biol. 12, 256. Witkin, E. M. (1956). Cold Spring Harbor Symposia Quant. Biol. 21, 123. Witkin, E. M. (1959). Proc. 10th Intern. Congr. Genet., Montreal, 1958. p. 280. Wolff, S. (1959). Radiation Research Suppl. 1, 453. Wolff, S., and Luippold, H. E. (1955). Science 122, 231.

Wyss, O., Clark, J. B., Haas, F., and Stone, W. S. (1948). .1. Bacteriol. 56, 51. Yamamoto, N. (1956). Virus (Osaka) 6, 510. Yamamoto, N. (1958). J. Bacteriol. 75, 443. Zamenhof, S., and Greer, S. (1958). Nature 182, 611. Zamenhof, S., and Griboff, G. (1954). Nature 174, 306. Zetterberg, G. (I960). Exptl. Cell Research 20, 659.

Chapter VI

Biosynthesis of RNA in Relation to Genetic Coding Proljlenis

ELLIOT VOLKIN

Biology Division,

Oak Ridge National Laboratory,'^

Oak Ridge, Tennessee

I. Introduction 271

II. Evidence Implicating an RNA Species as a Carrier of

Genetic Information 273

A. The DNA-like RNA in Bacteriophage Infection 273

B. Gradient Centrifugation of Cell Components 273

C. Messenger RNA-Homologous DNA Hybrids 275

D. RNA Turnover 275

E. Messenger RNA in Other Biologic Systems 276

III. Mechanism of Messenger RNA Function Based on

in Vivo Experiments 277

A. Enzyme Induction 277

B. Bacteriophage Infection 278

IV. Attempts at Isolation of a Messenger RNA 279

V. In Vitro Synthesis of RNA 280

A. Polj'nucleotide Phosphorylase 280

B. RNA Polymerase 280

VI. RNA Involvement in in Vitro Protein Synthesis 281

A. Requirement for DNA-Directed RNA Synthesis 281

B. The Use of Polynucleotides in Establishing the Code .... 282

C. Correlations with Plant Virus Mutants 284

VII. The Triplet Code 284

VIII. Summary 286

References 286

I. Introduction

A tremendous concentration of research effort on the nucleic acids

from the late 1940's through the early 1960's has culminated in what many respected investigators believe to be the essentially complete

* Operated by Union Carbide Corporation for the United States Atomic Energy Commission.

271

272 ELLIOT VOLKIN

answer to iiiolccular prohlcins of iiilicritaiicc ami >p('cilic <:i()\vtli. Since their discovoiy by Miescher in 1871, the nueh'ic acids have been vari- ously inipHcated as occupyin<f a central I'ole in the processes of genetic continuity, though historically such conjectuix' has l)een based on evi- dence of circumstance in time and place (iirachet, 1933, 1937, 1940a; (^aspersson, 1941). As with so many scientific endeavois, research in the nucleic acids acquired a renewed impetus with breakthroughs in meth- odology, specifically the development of chromatographic i)iocedures (Vischer and Chargaf!\ 1947, 1948a,b; C'ohn, 1949, 19o0a,b, 1951) and the availability of radioactive elements. The establislmient of deoxyribo- nucleic acid (DNA) as the entity i-esponsible for cellular heritability has been thoroughly documented (see Chapters I, II, and IV). This chapter will be largely concerned with one aspect of ribonucleic acid (RNA) metabolism, i.e., the role of RNA in the translation of genetic information from DNA to the sites of protein synthesis.

On the basis of rigorous and detailed in vivo and in vitro experi- mentation, originally done by Zamecnik and colleagues (see review by Zamecnik, 1960), such sites for i)rotcin synthesis have been demon- strated to be the RNA-rich subcellular components of cells — the micro- somes of highly developed tissues, the ribosomes of microbial systems (see Chapter VHI). Altliough the function of the bulk of RNA in these particles is still open to question, the role of another species of RNA has been disclosed by the use of cell-free systems that synthesize protein. This is the "transfer RNA," which is involved in cariying amino acids in their aminoacyl form to the particles where they are utilized for protein synthesis. Purification and separation of transfer RNA's have revealed that a high degree of specificity exists between the transfer RNA species and particular amino acids (see Chapter VIII). This phase of specific interaction, however, is not considered to occujw the key role in detennining the ultimate explicit structure of the protein product. For this purpose, another species of RNA has been invoked. The generally accepted term for it is "messenger RNA" (Jacob and Monod, 1961), so named because of its putative intermediary role in relaying the genetic code of DNA to functional sites. Within the span of just a few years, research in this area has quickly passed from that which purported to demonstrate the biologic existence of such an RNA to that which now utilizes RNA in vitro systems to establish the exact coding relations between specific nucleotide triplets and specific amino acids incorporated into protein. At this point, the speed with which such coding information becomes available in the literature may make this writing archaic by the time it reaches final printing.

VI. RNA AND CODING PROBLEMS 273

II. Evidence Implicating an RNA Species as a Carrier of Genetic Information

a. the dna-like rna in bacteriophage infection

In 1956 Volkin and Astrachan observed that in bacteriophage- infected Escherichia coli B a minor species of RNA was synthesized whose composition appeared to be similar to the DNA of the phage but not to the host. This biologic system offers a unique advantage for such a study, since viral infection immediately eliminates the bulk of RNA synthesis, and the synthesis of a minor class of RNA having a different over-all composition can be magnified. Thus, it was possible to demon- strate the existence of phage-specific RNA even during T7 infection (Volkin et al., 1958), as well as that produced during T2 infection. The DNA composition of T7 deviates only slightly from that of the host, whereas the DNA composition of T2 is widely different. The demonstra- tion of phage-specific RNA originally took the form of a comparison between relative specific activities of the RNA mononucleotides liberated by alkaline hydrolysis, after an assimilation of P'^- orthophosphate (Volkin and Astrachan, 1956a,b, 1957; Volkin et al., 1958). The conclu- sions from these experiments, that a phage DNA-like RNA was being synthesized, has since been amply confirmed by employing a C^*-labeled precursor (Volkin, 1959), and by a comparative study of the rates of labeling of the free nucleotide pool with the rates of labeling of the RNA mononucleotides (Astrachan, 1960; Volkin, 1962).

By employing only brief isotope assimilation periods (pulse experi- ments), it was revealed that the newly synthesized RNA does not merely accumulate but undergoes constant breakdown and renewal, or turnover (Volkin and Astrachan, 1957; Astrachan and Volkin, 1958). It was demonstrated (Astrachan and Volkin, 1958) that such turnover, though always present to some extent during the intracellular production of phage, is most pronounced during the first 10-15 minutes after the initial infection. It is noteworthy that virtually all of the extensive biochemical transformations of the host take place during this initial time period (Romberg et al, 1959; Flaks et al., 1959; Bessman, 1959; Sommerville and Greenberg, 1959; Keck et al., 1960).

B. GRADIENT CENTRIFUGATION OF CELL COMPONENTS

The question of a specific RNA newly induced by bacteriophage infection has been reinvestigated in a number of other laboratories by different experimental approaches. Their investigations not only rein-

274 KLLIOT VOLKIN

force the conclusions of XOlkiii and Astracliaii hut additionally describe the subcellular localization of the new RNA, and, in one instance, dem- onstrate a bona fide coniplcxiiiij; bet\v(>en iiha<j;e-s])e('iric RNA and tlie homologous phage DNA.

By employing a sucrose gradient technifjue for the fractionation of bacterial intracellular particles, Nomura et al. (1960) showed that the SOS particles contained most of the radioactive RNA after a short-term lalnding period of T2-iiif('ctcd E. coli with P'^-. A large pai't of this radioactive RNA could be released, depending on the Mg'" concentration, yielding a nucleic acid essentially free of pi-otein with a sedimentation constant of 8S. Some of the latter exists in the "soluble" foi-m in the T2-infected cell. The T2-s]iecific RNA has a distinguishably higher elec- trophoretic mobility than normal ribosomal RNA.

Brenner et al. (1961) employed the elegant procedure of separating components of differing densities by equilil)rium centrifugation in a cesium chloride gradient. They, too, concluded that RNA with a base composition similar to that of phage DNA is associated with ribosomes and can be detached by lowering the magnesium concentration. How- ever, in this case, the new RNA is associated with 70S and lOOS ribo- somes, and when disassociated has a sedimentation constant of 12S. In accord with the findings of Volkin and Astrachan, their data clearly demonstrate a turnover of the newly synthesized RNA, whether labeling was carried out with C"-uracil or P^-04. They further demonstrated that no new ribosomes are formed after T2 infection, but that these particles become enriched for T2-specific or T2 messenger RNA. Similar experi- ments employing S^^ for protein labeling revealed that the ribosomes are also enriched for nascent protein that undergoes rapid turnover with respect to its association to the ribosomes. These data led them to the conclusion that pre-existing ribosomes synthesize most of the protein in the infected cells but under the complete direction of the messenger RNA they contain, uninfluenced by the bulk of stable ribosomal RNA. Gros et al. (1961), using sucrose gradient separation techniques, con- cluded that normal, actively growing E. coli contains a messenger RNA component, physically distinct from the bulk ribosomal RNA and trans- fer RNA. They find the active RNA associated with 70S ribosomes, which Tissieres et al. (1960) have described as the principal site of in vitro protein synthesis. Turnover of the messenger RNA in uninfected E. coli was clearly demonstrated ; in this case, however, turnover consists of a transfer of newly labeled RNA to metabolically stable ribosomal RNA and soluble RNA, rather than a breakdown of RNA to acid-soluble products.

VT. RXA AND CODING PROBLEMS 275

C. MESSENGER RNA-HOMOLOGGUS DNA HYBRIDS

Perhaps the most convincing evidence for the existence of a phagc- specific messenger RNA has been accomphshed by Hall and Spiegelman (1961). These workers demonstrated that when single-stranded T2 DNA was slowly cooled in the presence of partially purified T2-specific RNA, an RNA-DNA hybrid was formed that could be separated by equilibrium density centrifugation. Since such hybrid formation was found to be unique to the homologous pair, the conclusion seems valid that a high degree of homology exists between T2 DNA and T2-specific RNA in terms of their complementaiy nucleotide sequential patterns. A later study by Spiegelman et al. (1961) demonstrated that such DNA-RNA complexes could be detected as naturally occurring hybrids in T2- infected E. coll. Hayashi and Spiegelman (1961) have extended these observations to bacterial systems that were subjected to a shift from an enriched medium to a basal medium. These systems simulate conditions of phage infection whereby ribosomal RNA synthesis is depressed while protein synthesis and messenger RNA synthesis proceed preferentially. The RNA synthesized and labeled under these "step-down" cultural conditions possessed the ability to form RNA-DNA hybrids when slow-cooled with only their homologous DNxA-'s.

D. RNA TURNOVER

The question of turnover or metabolic instability of messenger RNA, although referred to later with respect to its functional role, warrants further comment here. The findings of Yolkin and Astrachan (Volkin and Astrachan, 1957; Astrachan and Volkin. 1959; Volkin. 1962) with T2-infected E. coli demonstrate clearly that the T2-specific RNA is labile and converted to acid-soluble products. Furthermore, these break- down products do not. to any significant extent, reach the level of nucleosides and inorganic phosphate (Volkin. 1962). The phosphorus originating in messenger RNA can eventually terminate in phage DNA but not by either an obligatory pathway or a pathway utilizing large (acid-insoluble) polynucleotide fragments (Volkin. 1962). The in vitro data of Cohen et al. (1961) would indicate that the phage-specific RNA is preferentially enzymatically hydrolyzed to ribonucleoside-5'-P04's, which are readily reutilized for phage DNA synthesis.

In growing cultures of bacteria, a turnover of messenger RNA can be demonstrated with respect to (1) its transient existence evidenced by isotopic labeling, and (3) its geographical translocation from one type of ribosomal particle to another. It should be noted, however, that there

276 KLLIOT VOLKIX

is uniform and ox-ci-wlicliniiiti; evidence that a significant measurable loss of acid-insoliihle lor liitili molecular weight) RXA ne\-er occuis in normally growing bacteria. Thus, in pulse-type experiments, after re- moval of tlic radioactive substrate, instead of a loss of label, there is a continued increase in radioactivity into total RNA for some time. Concurrently, there is a shift in labeling pattern of the RNA nucleotides in a manner tliat would indicate an over-all transition of DNA-like RNA to ribosomal, or total, RNA (Volkin and Astrachan, 1957; Hayashi and Spiegelman, 1961). These observations pose the question of the fate of messenger RNA and its jirobable transformation to stable ribosomal RNA. This aspect of the i)roblem is covered in more detail in Chapter VII. Suffice it to note here that since most organisms and tissues contain ribosomal RNA of high cytidylic and particularly high guanylic acid contents, efficient conversion of messenger RNA to ribosomal RNA could indicate that large amounts of guanylic and cytidylic acids have been added. Kitazume et al. (1962) studied the kinetics of synthesis of the dRNA [sic] in yeast subjected to differing metabolic conditions. They conclude that dRNA is an obligatoiy precursor to i-ibosomal RNA and that this transition occurs by a subtraction of nucleotides from dRNA. They further propose that instability of dRNA is not necessarily a result of its biologic action (see below) but merely results from con- version to ribosomal RNA.

E. MESSENGER RXA IN OTHER BIOLOGIC SYSTEMS

Yeas and Vincent (1960), comparing the labeling of RNA mono- nucleotides liberated as the 2'-, 3'-isomers (alkaline hydrolysis) and as the 5'-nucleotides (phosphodiesterase hydrolysis), concluded that in exponentially growing yeast an RNA with a composition like yeast DNA exists. Astrachan and Fi.-^hcr (1961) and Volkin (1962) used 30-second pulses of P'--labeling to demonstrate the existence of a DNA-like RNA in Proteus vulgaris, E. coli, anfl Psendomonas aerugi- nosa— AT/GC ratios of these organisms vary from 1.47 to 0.92 to 0.50, respectively. In such growing bacteria, if the isotope incorporation period is extended to a few minutes, the labeling pattern appears to be identical with total (or ribosomal) RNA, in accord with the observa- tions that messenger RNA is present as a minor fraction of total RNA. The occurrence of a messenger RNA in E. coli and P. aeruginosa has been independently noted by Hayashi and Spiegelman (1961). who likewise determined its presence in Bacillus megaterium, an organism with DNA of high AT/GC ratio. The observations of messenger RNA in growing E. coli by Gros et al. (1961) have been mentioned.

Some evidence exists for the presence of an RNA with a composition

VI. RNA AND CODING PROBLEMS 277

like the homologous DNA in iiianiinalian systems (Boulanger and Montreuil, 1952; Hokin and Hokin, 1954; Logan and Davidson, 1957). Here again, such observations were attainable only where relatively short periods of isotope feeding w'cre employed and the composition of newly formed RNA was an apparent one, evaluated from the specific radioactivity of its mononucleotides. With the advent of autoradiogra- phy, particularly where tritium-labeled precursors are involved, experi- mentation with a large variety of organisms makes it increasingly clear that most, if not all, RNA synthesis of the cell starts within the nucleus — the initial product may be exclusively a DNA-like RNA (see Chapter II).

III. Mechanism of Messenger RNA Function Based on in Vivo Experiments

A. ENZYME INDUCTION

Perhaps the most striking characteristic of messenger RNA function in microbial systems is its short lifetime with respect to activity. That is, if it behaves either as a template or a catalyst, it must be an extremely unstable one. Evidence by way of independent techniques suggests that inactivation of messenger RNA is a direct result of its action, and, as a matter of fact, a reasonable degree of stoichiometry between messenger RNA and biologic product has been demonstrated.

The synthesis of the enzyme ^S-galactosidase is initiated almost im- mediately at a maximum rate by addition of the inducer, while synthesis is stopped in an equally short time interval by removal of the inducer (Monod, 1958; Boezi and Cowie, 1961; Roberts et al, 1961). Essentially similar kinetics of ;8-galactosidase synthesis are obtained by adding and removing the responsible gene (Pardee ef al., 1959; Riley et al., 1960). Insertion of the ^g-galactosidase gene by bacterial conjugation results in enzyme synthesis within 2 minutes; moreover, the subsequent rate of enzyme synthesis per zygote is constant. Experiments with heavily labeled (P^-) male parental bacteria, involving ample time for gene expression after insertion, followed by various periods of P''- decay, reveal that the rate of enzyme synthesis decreases as a function of P-'- atoms decayed. Thus the integrity of the z'^ gene is constantly required for continued enzyme synthesis. These data on /?-galactosidase induction are interpreted by Jacob and Monod (1961) to mean that the structural messenger of the gene is rapidly formed and exists as a short-lived intermediate. The group at the Carnegie Institution of Washington (Roberts et al., 1961; Boezi et al.. 1961) has described similar kinetics for the induction of /3-galactosidase and has suggested a model for the

278 ELLIOT VOLKIN

iiiiliu'tion iiu'cliaiiisiii llial siiuilaily pioposcs an unstable ciizynic- forming unit that must be synthesized continuously.

Further experiments with inducible enzyme systems utilizing purine and jiyrimidine analogs not oidy confirm the functional instability of the gene messengei- but suggest that tliis intermediate is indeed a polyri- bonucleotide. The presence of o-fluorouracil results in an alteration of the ^-galactosidase protein whereby its enzymatic activity is drastically reduced, but its antigenic capacity is retained (Bussard et al., 1960). The same analog, 5-fluorouracil, causes the formation of induced alkaline phosphatase jM'otein whose rate of thermal inactivation is greatly in- creased (Naono and Gros, 1960). The important features of the analog effect is the fact that it occurs veiy quickly after administration and thereafter is reflected by a constant proi)ortion (rather than increasing) of abnormal to normal enzyme molecules. The responsible agent, as- sumed to be RNA altered by incorporation of o-fluorouracil into its structure, must be short-lived and not accumulative to account for the observed kinetics.

B. B.\CTERIOPHAGE INFECTION

The original experiments (Volkin and Astrachan, 1957) with virulent phage-infected bacteria directly determined that the phage DNA-like RXA undergoes continual turnover. However, they could not distinguish whether such turnover is an obligatory part of its biologic mechanism or whether the turnover is simply a consequence of its catabolism, insofar as no ribosomal or stable RNA accumulates in this system. Further insight was gained by the discoveiy that some DNA-like RNA was synthesized but did not undergo turnover when protein synthesis was inhibited [chloramphenicol inhibition (Astrachan and Volkin, 1959) or deprivation of a required amino acid (Volkin, unpublished observa- tions)]. This technique with T2 infection was more fully exploited (Volkin, 1960) by using as host the mutant B94 (Gollub and Gots, 1959) which requires both adenine and arginine. A prior incubation of T2- infectcd B94 cells in basal medium containing only adenine permits a definite yield of phage upon removal of the adenine and replacement with arginine alone or arginine plus deoxyadenosine (neither condition by itself will allow production of phage). During the preincubation in adenine alone, though essentially no DNA or protein is synthesized, RNA of T2 DNA-like composition steadily accumulates for about 60 to 90 minutes. Furthermore, the amount of newly formed RNA (and not any other adenine-containing cell component) can be quantitatively normalized with the subsequent phage yield. The production of T2 during the secondaiy incubation is limited, stopping after about 30

VI. RNA AND CODING PROBLEMS 279

minutes, though the cells are capable of producing a full complement of phage if again given adenine. It seems feasible, then, to interpret these results to mean that a certain degree of stoichiometry exists between the messenger RNA and its specific product and that, as a result of its action in relaying the message, the RNA is concomitantly inactivated. Since T2 production ceases so abruptly, it must be assumed that the protein product (s) involved are not limited to the T2-induced enzymes but include phage protein itself.

IV. Attempts at Isolation of a Messenger RNA

Some purification of highly radioactive RNA (presumably enriched in messenger RNA) from pulse-type experiments has been attained from both T2-infected and nomially growing E. coli by classical centrifuga- tion techniques (Volkin and Astrachan, 1956b, 1957) and by centrifuga- tion through sucrose gradients (Nomura et al., 1960). Since the over-all nucleotide compositions of these partially purified fractions were not unlike total RNA, it was evident that only a small degree of purification was made. In view of the observed heterogeneity of messenger RNA as it exists in the cell (Hayashi and Spiegelman, 1961), the inadequate results of such fractionations by centrifugation are not surprising. Spe- cific hybrid formation of homologous DNA with messenger RNA as effected by Hall and Spiegelman (1961) obviously should represent a purified RNA product. On one such occasion this was proved for the T2-specific RNA banded with T2 DNA in the cesium chloride gradient (B. D. Hall and L. Astrachan, unpublished experiments). This method, by limitations in its scale, is not readily applicable to the mass isolation of messenger RNA. An ingenious technique has now been described which has been used to isolate highly purified messenger RNA and which lends itself to use on a preparative level. Bautz and Hall (1962), employing a column of denatured T4 DNA coupled to acetylated, phosphorylated cellulose, obtained specific interaction of it with only messenger RNA, the bulk of the RNA from the T4-infected cell passing through the column. The specific RNA, adsorbed to the DNA at a temperature favorable for annealing with DNA, was recovered by heating at elevated temperatures and elution. That the method is in- herently specific is evident from the remarkable observation that mes- senger RNA specific for DNA of a T4r7/ mutant has thus been purified.

Isolation by ECTEOLA column chromatography of DNA-RNA com- plexes has been accomplished from N'eurospora extracts by Schulman and Bonner (1962) and from ovarian frog egg preparations by Finamore and Volkin (1961). This technique may serve as a basis for the mass isolation of such complexes or of the specific RNA itself.

280 ELLIOT VOLKIN

V. In Vitro Synthesis of RNA

A. POLYNUCLEOTIDE PHOSPHORYLASE

In 1955 (ii'uiilx'i'^-Mana^o and Ochoa (lisc()\H'i(Ml an enzyme, jjolynu- cleotidc iihosphoiylase. that fur the fii\><t time allowed the eell-free, enzymatic synthesis of polyriboniu'leotide. Ribonucleoside diphosphates serve as substrates in the reaction and the composition of the product formed can be manipulated according to the relative concentrations of the individual diphosphates. Thus, polynucleotides can be synthesized ranging from the homopolymers to those containing all four bases in a variety of ratios (Ochoa and Heppel, 1957). It would be difficult to imagine that this mode of enzymatic synthesis is responsible for the cellular production of messenger RNA. Thus, if rigorous biologic speci- ficity is associated with the RNA, and this specificity is a function of the nucleotide sequence, it seems unlikely that such sequences should depend on the relative pool sizes of the nucleoside diphosphate precursors. In view of the truly outstanding nature of the discovery of this enzyme, however, it is pleasing to note the eventual great usefulness of this system in the elucidation of the coding problem (see below).

B. RNA POLYMERASE

The independent discovery from the laboratories of Weiss and Gladstone (1959), Hurwitz et nl. (1960), and Stevens (1960) of the enzyme RNA polymerase has opened new avenues of approach not only to the question of the biologic synthesis of RNA but also to the entire question of specificity relations between RNA and protein. The enzyme has been detected in a variety of mammalian (Weiss and Gladstone, 1959; AVciss, 1962) and plant (Huang et al, 1960) tissues, as well as in bacteria (Hurwitz et a/., 1960; Stevens, 1960; Ochoa ef ai, 1961). RNA polymerase, like DNA polymerase (Romberg et al., 1956; Bessman et al, 1958), requires DNA as a primer, and utilizes the ribonucleoside triphosphates as substrates. A net synthesis of the product can be attained. Most importantly, the composition of the DNA primer deter- mines the composition of the RNA product. All four nucleoside triphos- phates must be present for RNA synthesis, although ATP can be polymerized to polyadenylic acid (see also Spiegelman, 1959). The fact that heat-denatured DNA (Geiduschck et al., 1961 ; Chamberlain and Berg, 1962) and single-stranded DNA of (/)X174 phage (Chamberlain and Berg, 1962) can serve as i)rimers, i)lus the decisions based on nearest- neighbor frequency analysis (Weiss and Nakamoto, 1961), point to the

VI. RNA AND CODING PROBLEMS 281

conclusion that the RNA synthesized is a true complement of the primer. This enzyme, tlicn, provides a means for the synthesis of DNA- like or messenger RNA in biologic systems.

VI. RNA Involvement in in Vitro Protein Synthesis

A. REQUIREMENT FOR DNA-DIRECTED RNA SYNTHESIS

Probably the first evidence for a synthetic pathway for RNA via nucleoside triphosphates, and, at the same time, the first indication that this RNA may be required for protein synthesis, was reported by Spiegelman (1959) with cell-free fractions from microbial systems. Spiegelman found the protoplast membrane fraction highly potent in leucine-C^'*-incorporating ability into acid-insoluble protein, exceeding by a factor of 100 that found in the high-speed pellet or supernatant fraction. There was a definite requirement for all four nucleoside tri- phosphates, the presence of these compounds eliminating a lag in incor- poration, and they could not be substituted for by the corresponding diphosphates. Nisman and Fukuhara (1959) and Nisman et al. (1961), likewise employing the E. coli membrane fraction, were able to assign a definite importance to the presence of intact DNA in the cell-free synthesis of /3-galactosidase and the incorporation of amino acids into protein. A clear-cut approach to the question of the necessity for DNA was carried out by Novelli and his co-workers (1961). By the device of either UV-irradiation of the supernatant fraction or X-radiation of whole cells, they demonstrated an absolute requirement for the addition of DNA to their cell-free system (particles + supernatant) for the syn- thesis of /3-galactosidase. Moreover, only DNA prepared from cells containing the gene for j8-galactosidase was effective. The system re- quired all for ribonucleoside triphosphates for enzyme synthesis and the authors invoked a mechanism for the synthesis of RNA as an inter- mediate, or messenger, in the transmission of information from the specific DNA to the enzyme-fomiing system.

Tissieres et al. (1960) claim that the major particle site for protein synthesis in bacterial systems is the 70S ribosomes. Tissieres has shown (Tissieres and Hopkins, 1961a,b) that the RNA synthesized in vitro from the triphosphates has similar sedimentation properties in a sucrose gradient as that observed by Gros et al. (1961) following pulse labeling in vivo. In addition, this RNA seems to be required for the amino acid incorporation observed in the 70S ribosomes (Tissieres and Hopkins, 1961b).

282 ELLIOT VOLKIN

B. THE USE OF POLYNUCLEOTIDES IN ESTABLISHING THE CODE

From the laboratories of Nirenberg (Matthaei and Nirenberg, 1961 ; Nirenberg and Matthaei, 1961; Martin et al., 1962; Nirenberg et al., 1962) and Ochoa (Lengyel et al, 1961, 1962; Speyer et al., 1962) have come the most startling applications of the microbial ribosome-siiper- natant amino acid incoi-porating systems. Nirenberg and ]\Iattliaei (1961) discovered that the addition of heat-stable RNA from different sources increased the incorporation of C^^-labeled amino acids by the ribosome-supernatant system. It is of interest that these authors note that "the correlation between the amount of incorporation and the amount of added RNA suggested stoichiometric rather than catalytic activity of the template RNA" (Nirenberg and Matthaei, 1961). In the process of testing a variety of RNA's, synthetic polynucleotides synthe- sized by polynucleotide phosphorylase were tested for stimulatoiy ac- tivity. It was found that the addition of polyuridylic acid resulted in a rather massive synthesis of poly-L-phenylalanine, and the polymeriza- tion of only this amino acid was specifically directed (Nirenberg and Matthaei, 1961). In a subsequent thorough investigation of this reaction, Nirenberg et al. (1962) determined that the phenylalanine was linked to soluble RNA as a true intermediate in the process. Polyuridylic acid associated in a double- or triple-stranded configuration is ineffective in amino acid stimulation, suggesting that active RNA is a single-stranded template (Nirenberg et al., 1962). In any case, the observations with polyuridylic acid have been fully corroborated by Lengyel et al. (1961), indicating that uridylic acid residues serve as the information centers for phenylalanine insertion into protein.

The laboratories of Nirenberg and Ochoa have fully exploited the advantages of such synthetic polynucleotides as synthetic "messengers," whereby copolymers enriched in uridylic acid and containing known proportions of one or more of the other nucleotides have been tested for stimulatory activity with virtually the entire spectrum of natural amino acids. In this fashion, relatively long sequences of uridylic acid in the chain serve as templates for the production of polyphenylalanine chains, which in turn serve as "handles" (Lengyel et al., 1961) for the attach- ment of other amino acids whose linking is directed by the specific neigh- boring base sequences. Thus, one may correlate the observed incorpora- tion of a particular amino acid with the calculated frequency ratio of the uridylic sequences to the other nucleotide sequences in the polymer. Assuming a minimum coding ratio of three nucleotide sequences for the direction of a specific amino acid insertion into protein, it has been possible thus far to establish nucleotide coding letters for a larger num-

VI. RNA AND CODING PROBLEMS 283

ber of amino acids. The ratio of three has a rather solid basis in terms of both the theoretical and experimental approach (see below). At the risk of publishing the score in the middle of the ball game, I have compiled the presently available "genetic code" for 17 amino acids in Table I. It should be noted that the sequential order of the nucleotides

TABLE I Genetic Code for Polymerization of 17 Amino Acids"

Amino acid Triplet code letter''

Ahinine lU IC IG

Arginine lUlClG

Cysteine 2U IG

Glutamic acid lU IC IG

Glycine lU 2G

Histidine lU lA IC

Isoleucine 2U lA

Leucine 2U IC, 2U IG

Lysine lU 2A

Methionine UG'

Phenylalanine UUU

ProUne lU 2C, [CCC]

Serine 2U IC, UGC, 2U IG

Threonine lU 2C

Tryptophan lU 2G

Tjrosine 2U lA

Valine 2U IG

° Compiled from the data in Lengyel et al. (1961, 1962), Nirenberg and Matthaei (1961), Martin et al. (1962), Speyer et al. (1962). * Sequential order unspecified. <= Not further specified (Martin el al, 1962).

in any triplet (with the obvious exception of UUU = phenylalanine) cannot as yet be specified. In addition, the fact that more than one triplet can specify a single amino acid shows that the code is degen- erate; that is, more than one coding unit can specify the insertion of a single amino acid. Thus, it should be clear that the use of enriched polyuridylic acid sequences as a "handle" probably weights the present code toward this nucleotide. A number of other triplets, absent in uridylic acid, could be equally specific for some of the amino acids listed.

Wood and Berg (1962) have succeeded in demonstrating a significant amino acid activating effect by enzymatically synthesized RNA w'ith a system consisting in ribosomes plus a soluble fraction that had been virtually freed of DNA. In particular, the addition of T2 phage DNA, the four ribonucleoside triphosphates and RNA polymerase resulted in a 4-fold increase in rate of reaction and a 20-foId increase in extent of

284 ELLIOT VOI.KIN

ri'action. Likewise, the syiitlu'sizcd l^NA, freed of the j)()lyinerase and DNA, eould a('ti\-ate the iiicoiporatioii of amino acids sonic 3- to 5-fold. Althouiih it could he siiown (Chanil)eiiain and Bcrfj;, 1962) that sinf^le- stranded DNA from hcat-denatiired T2 DNA, or ^X174, could prime the foi-niatit)!! of KNA by tlu' polynieiase with about v(\un\ facility as the native, or double-stranded DNA, these KNA's are completely inactive in the amino acid-incorporating system (Wood and Berg, 1962). In view of the ability of the biologically unspecific |)olynucleotides synthe- sizetl l)y polynucleotide phosphosphylase to activate amino acid incorj)o- ration in the systems of Ochoa and Nirenberg, and especially the fact that double-stranded synthetic polymers are ineffective (Nirenberg et al., 1962), an enigma seems to be created on the mechanism of activation by RNA in these systems.

C. CORREL.\TIONS WITH PLANT VIRI^S MUTANTS

If there are suspicions about the validity of establishing a genetic code on the basis of in vitro incorporation of very limited amounts of radioactive amino acids into "acid-insoluble protein," it should be noted that the code letters often are the same as those surmised from in vivo experiments involving some plant virus mutants (see Chapter X). One nitrous acid mutant contains leucine in place of a proline residue (Tsu- gita and Fraenkel-Conrat, 1960). Nitrous acid acts by deaminating C to U. Inspection of Table I reveals that such a conversion of C of the proline code to U is in line with the specification of a leucine in place of proline. In the same mutant, replacement of threonine by serine can be explained in a similar way (Tsugita, 1961). In another nitrous acid mutant (No. 167) serine is replaced by phenylalanine and glutamine by valine (Tsugita, 1961). The former conversion is readily predictable from the data in Table I (2U IC^UUU). Although no code letters have been established as yet for glutamine, it is of interest that glutamic acid (lU IC IG) fulfills the requirements for such a conversion to valine (2U IG) by deamination. Also in line with the code as determined by Nirenberg and Ochoa (Table I) is the observation that the proline- phenylalanine ratio in the proteins of wild cucumber and tobacco mosaic virus can be correlated with their respective RNA C:U ratios (Yama- zaki and Kaesberg, 1961).

VII. The Triplet Code

The question of the transfer of information from nucleic acid to the synthesis of protein involves the reading of a polymer with sequences of only four different units (nucleotides) for the .sequential ordering of some 20 different units (amino acids). The minimum coding unit then

VI. RNA AND CODING PROBLEMS 285

would be a three-letter one where 64 different sets are possible. This scheme was first proposed by Gamow (1954), who, on the basis of DNA structure, considered the code to be overlapping; that is, the letters of one triplet can be used sequentially in adjacent triplets. This coding mechanism would predict that a change in one base would result in changes in three neighboring amino acids in a protein. However, such observations that nitrous acid mutants generally are reflected in the change of only one amino acid at a time (Tsugita, 1961 ; Tsugita and Fraenkel-Conrat, 1960), and that the variety of abnormal human hemo- globins differ by only single amino acid changes, (Watson and Kendrew, 1961) seem to rule out the overlapping code. Helinski and Yanofsky (1962) and Henning and Yanofsky (1962) recently found that an ultraviolet-produced mutant of E. coli K12 contains the A protein of tryptophan synthetase differing from normal A protein by only a single amino acid substitution, glutamic acid for glycine. This same glycine is replaced by arginine in the A protein of a closely linked mutant.

Since 64 triplets can determine only 20 amino acids, it was felt that, with the exclusion of an overlapping mechanism, there must be a scheme to select the correct triplets along a continuous sequence. Crick et al. (1957) suggested that, although all possible sequences of amino acids can be coded, only those nucleotide triplets that sit side by side make sense, and the overlapping triplets so formed must be nonsense. In this way, they could demonstrate that the maximum number that can be coded is 20. Since then. Crick et al. ( 1961), on the basis of ingeniously conceived genetic experimentation with the B cistron of the rll region of bacteriophage T4, have proposed a non-overlapping triplet code where selection of the correct sequence of bases depends on reading at a fixed starting point. In this way, by starting at such a point and reading toward one direction, the insertion or deletion of a base would result in incorrect further reading of the code. However, an additional insertion or deletion (double mutant) will result in a correction of the code- reading in all sequences following the second mistake. Thus, recovery of most of the original information (in their experiments recoveiy to pseudo-wild type) can be obtained. It should be noted that by use of the proper triple mutants they obtained firm evidence for a probable coding ratio of three. Insofar as the region over which the suppressors they investigated covered about a quarter of the B cistron (the protein produced by this cistron may contain about 200 amino acids), the authors feel that the code is probably degenerate; otherwise, nonsense reading, which should have been encountered at a much closer distance, has been compensated for by the flexibility afforded by a degenerate code. It will be of interest to see whether as predicted (Crick et al.,

286 ELLIOT VOLKIN

19G1), the protein iJiodiiccd hy some of tlic ildiihic iiiiit;iiit> lias l)C'cn altered in amino acids tlirouii;li the Icii^tli of the polypcplidc chain corresponding; to the region between the two alterations in the DNA chain.

VIII. Summary

An RNA of composition siniihir to DNA has been detected in a number of different l)iologic systems. Its presence in pluige-infected bacteria is particularly ajjparent, since ribosomal RNA synthesis is preventeci in that systeuL This RNA jirobably has a sedimentation constant of 12-14S but aj^pears to be largely reversibly associated with ribosomal pai'ticles. The RNA undergoes rapid turnover, which can be measured by its molecular degradation in the ])hage-infected cell, and by its translocation to other subcellular components in actively growing bacteria. The turnover phenomenon seems to be part and parcel of its function as deduced fi"om kinetic studies. Moreover, by cori'elating the amount of such RNA synthesis witli (/) phage ])i-oduction, (2) enzyme synthesis, and (5) i?} intro amino acid incorporation, a definite stoichi- ometry has been found to exist between the amount synthesized and the protein product. Specific RNA-DNA hybrids (where the RNA has a composition similar to the DNA) can be isolated from extracts, as well as produced by "annealing" of RNA with only the homologous DNA. RNA of DNA-like composition likewise can be synthesized iti vitro with the enzyme, RNA polymerase. RNA synthesized in this manner markedly activates the incorporation of amino acids into protein in cell-free systems. In all probability such in m'^ro-synthesized, complete, RNA molecules will be shown eventually to be capable of directing the synthesis of complete, biologically active proteins. Presently, however, it is possible to use synthetic polynucleotides to demonstrate the existence of a definite specificity between the nucleotide composition and the particular amino acid incorporated into protein. In this way, a partial nucleotide triple code for most of the naturally occurring amino acids has been resolved.

References

A.straclian, L. (1960). Fcdcmiion Proc. 19, 198.

A.strachan, L., and Fisher, T. N. (1961). Federation Proc. 20, 359.

A.strachan, L., and Volkin, E. (1958). Biochim. el Biophys. Ada 29, 536.

A.strachan, L., and Volkin, E. (1959). Biochim. et Biophy.f. Ada 32, 449.

Bautz, E. K. F., and Hall, R. D. (1962). illh Annual Mcding Bi,>phi/.'<. Sor.. Wash-

inglon, D.C. Abstr. TA8. Bessnian, M. J. (1959). ./. Biol. Clwm. 234, 2735.

VI. RNA AND CODING PROBLEMS 287

Bessman, M. J., Lehman, I. R., Simnis, E. S., and Kornbrrg, A. (1958). J. Biol.

Chem. 233, 171. Boezi, J. A. Bolton, E. T., Britten, R. J., Cowie, D. B., McCarthy, B. J., Midgley,

J. E., and Roberts, R. B. (1961). Carnegie Inst. Wash. Yearbook 60, 263. Boezi, J. A., and Cowie, D. B. (1961). Biophys. J. 1, 639. Boulanger, P., and Montreuil, J. (1952). Biochim. et Biophys. Ada 9, 619. Brachet, J. (1933). Arch, biol (Liege) 44, 519. Brachet, J. (1937). Arch. biol. (Liege) 48, 529. Brachet, J. (1940a). Arch. biol. (Liege) 51, 151, 167. Brachet, J. (1940b). Compt. rend. soc. biol. 133, 88, 90. Brenner, S., Jacob, F., and Meselson, M. (1961). Nature 190, 576. Biissard, A., Naono, S., Gros, F., and Monod, J. (1960). Compt. rend. acad. sci. 250,

4049. Caspersson, T. (1941). Natunciss. 29, 33.

Chamberlain, M., and Berg, P. (1962). Proc. Natl. Acad. Sci. U.S. 48, 81. Cohen, S. S., Barne.s, H. B., and Lichtenstein, J. (1961). J. Biol. Chem. 236, 1448. Cohn, W. E. (1949). Science 109, 377. Cohn, W. E. (1950a). /. Arn. Chem. Soc. 72, 1471. Cohn, W. E. (1950b). J. Am. Chem. Soc. 72, 2811. Cohn, W. E. (1951). J. Am. Chem. Soc. 73, 1539. Crick, F. H. C, Griffith, J. S., and Orgel, L. E. (1957). Proc. Natl. Acad. Sci. U£.

43, 416. Crick, F. H. C, Barnett, L., Brenner, S., and Watts-Robin, R. J. (1961). Nature 192,

1227. Finamore, F. J., and Volkin, E. (1961). Am. Chem. Soc. Div. of Biol. Chem. Abstr.

53C. Flaks, J. G., Lichtenstein, J., and Cohen, S. S. (1959). J. Biol. Chem. 234, 1507. Gamow, G. (1954). Nature 173, 318. Geiduscheck, E. P., Nakamoto, T., and Weiss, S. B. (1961). Proc. Natl. Acad. Sci.

U.S. 47, 1405. Gollub, E. G., and Gots, J. S. (1959). J. Bacteriol. 78, 320. Gros, F., Hiatt, H., Gilbert, W., Kurland, C. G., Riseborough, R. W., and Watson,

J. D. (1961). Nature 190, 581. Grunberg-Manago, M., and Ochoa, S. (1955). J. Am. Chem. Soc. 77, 3165. Hall, B. D., and Spiegelman, S. (1961). Proc. Natl. Acad. Sci. U.S. 47, 137. Hayashi, M., and Spiegelman, S. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1564. Heiinski, D. R., and Yanofsky, C. (1962). Proc. Natl. Acad. Sci. UB. 48, 173. Henning, U., and Yanofsky, C. (1962). Proc. Natl. Acad. Sci. U.S. 48, 183. Hokin, L. E., and Hokin, M. R. (1954). Biochim. et Biophys. Acta 13, 401. Huang, R. C, Maheswari, N., and Bonner. J. (1960). Biochem. Biophys. Research

Commune. 3, 689. Hurwitz, J., Bresler, A., and Biringor. R. (1960). Biochem. Biophys. Research

Communs. 3, 15. Jacob, F., and Monod, J. (1961). J. Mol. Biol. 3, 318.

Keck, K., Mahler, H. R., and Fraser, D. (1960). Arch. Biochem. Biophys. 86, 85. Kitazume, Y., Yeas, M., and Vincent, W. S. (1962). Proc. Natl. Acad. Sci. U.S. 48,

265. Romberg, A., Lehman, I. R., and Simms, E. S. (1956). Federation Proc. 15, 291. Kornberg, A., Zimmerman, S. B., Kornberg, S. R., and Josse, J. (1959). Proc. Natl.

Acad. Sci. U.S. 45, 772.

288 ELLIOT VOLKIN

Lcngyi'l. P., Sl)(\Vfr. .1. [■'.. .iiid Orhoa. S. (liKU). l'n>r. Xiill. Ara,l. Sri. (jS. 47, 1936. Longyel, P., Speypr, J. F., Hasilio, C. .iii.l Odioa, S. (U)62). Pror. Null. Acad. Sci.

US. 48, 282. Logan, R., and Davidson, J. N. (1!)57). liidchim. et Biophys. Ada 24, 196. Martin, R. G., Matthaei, J. H., Jones, (). \V., and Nironhoifi, M. W. (1962).

Biochcm. Biophys. Research Commun.s. (>, UO. Matthaei. J. H., and Nirenbcrg, M. W. (1961). I'ror. .\(ttl. Amd. Sri. l.S. 47, 15S0. Mio.sfhor. F. (1871). Ilnppc-Scylvr's Med. Chem. I'nlers. \k 441. Monod, J. (1958). Rec. Irar. rhim. 77, 569.

Naono, S., and Gros, F. (1960). Compl. rend. ncad. sri. 250, 3527, 3889. Nircnbcrg, M. W., and Matthaei, J. H. (1961). Proc. Natl. Acad. Sci. US. 47, 1588. Nirenberg, M. W.. Matthaei, J. H., and Jone.s, O. W. (1962). Proc. Natl. Acad. Sci.

US. 48, 104. Nisman. B., and Fukuhara, H. (1959). Compt. rend. acad. sci. 250, 410. Nisman, B., Kayser, A., Demailly, J., Jr., and Genin, C. (1961). Compl. rend. acad.

sci. 252, 2619. Nomura, M., Hall, B. D., and Spiegchnan, S. (1960). J. Mol. Biol. 2, 306. Novelii, G. D., Kanieyama. T., and Ei.sen.stadt, J. M. (1961). ./. Cellular Contp.

Physiol. Suppl. 1, 58, 225. Ochoa, S., and Heppel, L. A. (1957). In "The Chemical Basis Heredity" (W. D. Mc-

Elroy and B. Gh^ss, eds.), p. 615. Johns Hopkins Press, Baltimore. Ochoa, S., Burma, D. P., Kroger, H., and Weill, J. D. (1961). Federation Proc. 20,

362. Pardee, A. B., Jacob, F., and Monod, J. (1959). ./. Mol. Biol. 1, 165. Riley, M., Pardee, A. B., and Monod, J. (1960). J. Mol. Biol. 2, 216. Roberts, R. B., Britten, R. J., and McClure, F. T. (1961). Biophys. J. 1, 649. Schulman, H. M., and Bonner, D. M. (1962). Proc. Natl. Acad. Sci. US. 48, 53. Sommerville, R., and Greenberg, G. R. (1959). Federation Proc. 18, 327. Spey.er, J. F., Lengj^el, P., Basilio, C., and Ochoa, S. (1962). Proc. Natl. Acad. Sci.

US. 48, 63. Spiegelman, S. (1959). In "Recent Progress in Microbiology," p. 81. 7th Intern.

Congr. for Microbiol., Stockholm, 1Q5S. Almquist and Wiksell, Uppsala. Spiegelman, S., Hall, B. D., and Storck, R. (1961). Proc. Natl. Acad. Sci. US. 47,

1135. Stevens, A. (1960). Biochem. Bio pit ys. ResearcJi Communs. 3, 92. Tissieres, A., and Hopkins, J. W. (1961a). Cold Spring Harbor Symposia Quant.

Biol. 26, 126. Tissieres, A., and Hopkins, J. W. (1961b). Proc. Natl. Acad. Sci. US. 47, 2015. Tissieres, A., Schlessinger, D., and Gros, F. (1960). Proc. Natl. Acad. Sci. US. 46,

1450. Tsugita, A. (1961). Protein. Nucleic Acid Enzyme (Tokyo) 6, 385. Tsugita, A., and Fraenkel-Conrat, H. (1960). Proc. Natl. Acad. Sci. US. 46, 636. Vi.scher, E., and Chargaff, E. (1947). J. Biol. Chem. 168, 781. Vischer, E., and Chargaff, E. (1948a). J. Biol. Chem. 176, 703, 715. Vischer, E., and Chargaff, E. (1948b). Federation Proc. 7, 197. Volkin, E. (1959). In "Biochemistry of Viruses" (E. Broda and W. Frisch-

Niggemeyer, eds.), pp. 7, 212. Proc. J,th Intern. Congr. Biochein., Vienna, 195S.

Pergamon Press, New York. Volkin, E. (1960). Proc. Natl. Arad. Sci. US. 46, 1336. Volkin, E. (1962). Federation Pror. 21, 112.

VI. RNA AND CODING PROBLEMS 289

Volkin, E., and Astrachan, L. (1956a). Virology 2, 149. Volkin, E., and Astrachan, L. (1956b). Virology 2, 433.

Volkin, E., and Astrachan, L. (1957). In "The Chemical Basi.s of Heredity," (W. D. McElroy and B. Glass, eds.), p. 686. Johns Hopkins Pre.ss, Baltimore. Volkin, E., Astrachan, L., and Countryman, J. L. (1958). Virology 6, 545. Watson, H. C, and Kendrew, J. C. (1961). Nature 190, 670. Weiss, S. B. (1962). Federation Proc. 21, 120.

Weiss, S. B., and Gladstone, L. (1959). J. Am. Chem. Soc. 81, 4118. Weiss, S. B., and Nakamoto, T. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1400. Wood, W. B., and Berg, P. (1962). Proc. Natl. Acad. Sci. U.S. 48, 94. Yamazaki, H., and Kaesberg, P. (1961). Nature 191, 97. Yeas, M., and Vincent, W. S. (1960). Proc. Natl. Acad. Sci. V.S. 46, 804. Zamecnik, P. C. (1960). Harvey Lectures 54, 256.

Chapter VII

Kinetic Studies of the Synthesis of RNA and Ril^osonies

R. B. ROBERTS, R. J. BRITTEN, and

B. J. McCarthy

Carnegie Institution of Washington, Department of Terrestrial Magnetism, Washington, D. C.

I. Introduction 292

A. Microsomes and Ribosomes 292

B. Effects of Magnesium 293

C. Other Influences on Ribosome Patterns 296

D. Generality of Ribosome Patterns 298

E. Ribosomes as Site of Protein Synthesis 299

II. Properties of Ribosomes in Vitro 299

A. Preparation and Purification of Ribosomes 299

B. Protein/Nucleic Acid Ratio 300

C. Protein Components 302

D. Nucleic Acid Components 305

E. Structure of Ribosomes 306

III. Fractionation of Ribosomes 308

A. Chromatography 308

B. Sedimentation Analysis 309

C. Density Gradient Banding 310

D. Electrophoresis 310

E. Two-Phase Systems 310

F. Conclusion 310

IV. Metabolic Pools 311

V. Kinetics of Incorporation into RNA of Ribosomes ..... 315

A. DEAE Chromatography 315

B. Sedimentation Analysis 316

C. Quantitative Kinetic Analy.sis 321

D. Stages of Synthesis of the Ribosomal RNA ...... 326

VI. Kinetic Studies of Phenol-Extracted RNA ....... 329

VII. Kinetics of Incorporation into Protein ....•••■ 331

A. Incorporation of C"-Leucine into the Ribosomes ..... 331

B. Effect of Chloramphenicol on Ribosome Synthesis .... 333

C. Assembly of the Ribosomes 333

VIII. Composition of Newly Formed RNA 335

A. Ambiguities 335

B. Time Course of Composition 337

C. Fractionation 338

291

292

R. H. HOHKHTS. H. .1. HKnTHN, AND H. .1. MCCAKIIIY

1). KNA Classes

E. Mechanisms for the Change in Apiiaicnt ( "oiniiosition

IX. Discussion

A. Rolo of Ribosonios in Information Traiisl'i r H. Correlation with Enzyme Induction

C. Messenger Theory

D. Homogeneity of Ribosomes

E. Site of Ribosome Synthesis

F. DXA-RXA Hybrids

(!. SeU-Replication of Ribosomes .... H. Cell-Free Synthesis of Ribo.somes

I. State of Ril)osomes in Living Cells

X. Epilogue

References

340 341 342 342 343 344 345 346 347 347 348 348 350 350

I. Introduction

A. MICROSOMES AND RIBOSOMES

The term "microsome" was originated by Claude in 1943 to describe a fraction obtained by differential centrifugation of disrupted vertebrate cells (Claude, 1943). After breakage, the cells were centrifuged at low g to sediment nuclei, cell walls, and mitochondria. The supernatant fluid, devoid of structures visible in the microscope, was then centrifuged at 100,000 g for 2 hours and the pellet so obtained was designated the microsomal fraction. In it appeared much of tlie cytoplasmic ribonucleic acid (RNA) and lipid.

Electron microscopy of this pellet showed that it consisted mainly of membranes to which were attached dense particles of about 200 A diameter. Electron micrographs of the same cells showed an extensive network of membranes designated the endoplasmic reticulum. In addi- tion, dense particles could be seen sometimes free and sometimes at- tached to the membranes. The RNA content of the pellet was attributed to its content of particles (see review, Palade, 1958). Particles from bacteria had been observed earlier (Luria et al., 1943.)

Analysis of the microsome fraction in the analytical centrifuge, especially after treatment with deoxycholate to dissolve the mem- branous material, showed the jiresence of objects having sedimentation coefficients in the range 20-lOOS (Petcrman and Hamilton, 19o7).

Similar fractions of high RNA content were obtained from bacteria, yeast, and plant cells. All showed the presence of a number of discrete peaks in the analytical centrifuge having sedimentation coefficients of 20-1008 (Schachman et al, 1952; Chao and Schachman, 1956; Ts'O et nl, 1956).

Since the microsomal fraction included l)()(li the membrane matei'ial

VII. SYNTHESIS OF RNA AND RIBOSOMES

293

and the RNA-rich j)articles as well as any protein which was sedimented by the prolonged centrifiigation, the rejjorted composition of the micro- somes was highly variable. In 1957 the term "ribosome" (originally suggested by H. M. Dintzis) was introduced to distinguish the particu- late material from the remainder of the microsome fraction (Roberts, 1958).

B. EFFECTS OF MAGNESIUM

In the same year the effect of magnesium concentration upon the sedimentation coefficients obtained for the ribosomes was recognized. Chao showed that the larger ribosomes (about SOS) obtained from yeast

Exponential (broth) E. CoU

Fig. 1. Centrifugation pattern of juice from E. coll growing in broth medium. These rapidly growing cells have a high content of ribosomes. Sedimentation from left to right observed in Spinco Model E ultracentrifuge.

2!>4

H. li. KOBERTS, R. .1. HHITTKN. AND H. J. MCCARTHY

dissociated into two coiupoiK'Hts of about 60S and 408 when the mag- nesium concentration of the sut^pending fluid was reduced from 10'^ M (Chao, 1957). Bolton et at. independently found a similar effect in the ril)08omes of Escherichia coli (Bolton et al, 1958). Ts'O et al. reported

the dissoeiatit)n of the SOS coinponciit of iiI)Osomcs of pea seedHngs into

L AJ

Ribosome pellet in I0~ M Mg.

i

Fig. 2. Centrifugation pattern of ribosome pellet in 10 ' M Mg. Only 50S and 30S ribosomes appear.

components of 60S, 40S, and 26S (Ts'O et al., 1958). A Mg/P ratio of 0.3 was required to preserve the integrity of SOS particles from pea seedlings (Edelman et al., 1960).

The sedimentation coefficient.-^ reported in early jiapers vary widely; most of the values were not corrected for viscosity or concentration.

VII. SYNTHESIS OF RNA AND RIBOSOMES

295

Fig. 3. Ribosoraes of E. coli (x 100,000). (Hall and Slayter, 1959.) (a) 308 preparation, (6) 508 preimration, (c) 708 preparation, (d) 1008 preparation.

290 H. H. ROBERTS. K. J. URITTEN, AM) U. J. MCCARTHY

\\'li('n tlicsr conH'ctions arc made the rihosonics of A', coli show three prominent groiii);> at 30.6 di IS. oO.O d= IS, and 69.1 ± IS. These are usually roinided off to 30, 50, and 7()S for identification (Tissiei-cs rt (iL, 1 909 1 .

Hihosonios extracted from E. coli at 10- A/ Mg show predominantly the 70S (or larger) form (Fig. 11, but when the extraction is carried out in 10 ' Mg or when the Mg concentration is reduced to 10 ' M after extraction th(> large particles give rise to equal numbers of 30S and 50S (Fig. 2). This fact, together with electron microscope pictures which show shapes consistent with a 3O-50S complex (Fig. 3), (Hall and Slayter, 1959; Huxley and Zubay, 1960) indicates that the 30S and 50S ribosomes associate reversibly to form 70S particles.

In addition, ribosomes of sedimentation coefficients greater than 70S are observed. One of these (al)out lOOS) appears to be a dimer of the 70S particle (Fig. 3). The reactions are not simi)le, how(>ver. An inter- mediate group (called 85S for identification) also appears and its sedimentation coefficient as well as its proportions depends on the mag- nesium concentration (Britten and McCarthy, 1959). As this 85S grouji yields 30S and 50S ])articles in equal numbers when the magnesium concentration is reduced, it appears to be due either to an altered shape or hydration of the 70S or lOOS particles or to a rapid equilibration between the 70S and lOOS forms which yields an interme<;liate sedimenta- tion coefficient.

C. OTHER INFLUENCE.S ON RIBOSOME P.\TTERNS

The sedimentation pattern also depends upon the metabolic state of the cell at the time of breaking and upon the method of breaking the cells (Fig. 4). McCarthy has shown that cells rapidly accumulate the lOOS particles when their energy supply (glucose) is exhausted and rev(>rt with equal rapidity to the normal pattern when glucose is restored (McCarthy, 1960). Similar changes also appear when mutants are deprived of a refjuired metabolite. There is little difference, however, in ribosomes taken from different phases of the growtli cycle in synchronized cultures (Britten et nl, 1960).

Prolonged incubation of growing cells in ]\Ig-free media reduces the ribosome content to less than 5%. The ribosome content is restored (fol- lowing an exponential growth curve) when the magnesium of the medium is replenished (McCarthy, 1962).

The ribosome pattern does not result simply from an equilibrium among two basic units and combination thereof. Some 30S and 50S particles are present in the cell juice even when the magnesium concen- tration is high. These are sometimes designated "native" 308 and 50S

VII. SYNTHESIS OF RNA AND RIBOSOMES

297

ribosomes in distinction to the "derived" SOS and 50S particles which are obtained from the dissociation of 70S particles. Native particles have higher specific radioactivities (after a short period of exposure to the tracer) than do the derived particles (Britten and McCarthy, 1959). Isolated native particles do not combine (Green and Hall, 1961).

uj

50 85

C D

Fig. 4. Changes in ribosome pattern caused by different metabolic conditions: A. The lOOS component is prominent when growth is stopped by lack of energy source. B. Magnesium starvation causes disappearance of ribosomes. C. Incubation with chloramphenicol results in accumulation of small particles. Ribosomes already present appear as 70S. D. Incubation with 5-fluorouracil results in accumulation of 308 and 50S ribosomes.

Furthermore, it has been shown m vivo (McQuillen et al., 1959) and in vitro (Tissieres et al., 1960) that the 70S particles carrying newly made protein do not dissociate completely when the magnesium concen- tration is reduced to 10 ■* M.

In addition to the main peaks which are readily observed smaller groups of particles sometimes appear. Many preparations show a peak at about 20S. The 20vS component is complex, containing one component

298 H. H. ROBERTS, R. .1. HRITTKX. AM) H. .1. MCfAKTHV

rich in protein and aiioliu'r rich in KNA which may he a ribosome jirccursior. The 20S particle also appeared after prolonged storage of ribosonies at 10^ M magnesium. In this circumstance a sufficient quan- tity was obtained to show that the protein/nucleic acid ratio was about 40/60 but it is not known whether this degradation product is the same as the particle observed in cell juices (Roberts anrl Duerksen, 1960; Aronson and McCarthy, 1961).

Another particle a])out 43S plays a !)rominent role as a protein deficient precursor to the 50S ribosome (see Section V). It has also been observed in the juice of cells in which DNA synthesis was prevented by lack of thymine (Rol)erts, 1960) and as a degradation jjiodiict of 50S ribosomes (Elson, 1961).

Abnormal patterns develop when cells are exposed to antibiotics or analogs. Chloramphenicol causes the accumulation of small particles deficient in protein while preventing the entry of P'*- into the normal l)articles (Pardee ef a/., 1957; Nomura and Watson, 1959; Bolton, 1959). 5-Fluorouracil causes the accumulation of SOS and 50S ribosomes which are unable to form 70S particles (Aronson, 1961).

A part of the complexity of the ribosome pattern may be due to the association of newly synthesized RNA or protein with the particles (see Section V,C and VIII). Such material might facilitate or inhibit associa- tion of the basic units. In most experiments its presence would be undetected and uncontrolled, thus giving rise to unpredictable variations.

D. GENERALITY OF RIBOSOME PATTERNS

Because of the wide range of ribosome patterns which can l)e observed in one organism it is difficult to compare the ribosomes from different sources. Those which have been most extensively studied show in common particle groups of 70-lOOS which dissociate in low concentra- tions of magnesium into two groups of 2(>-40S and 40-60S. ISIcCarthy has surveyed the ribosomes derived from a number of microorganisms and found a general similarity of patterns (McCarthy, 1959). Some minor diffcrenc(>s appeared but they cannot lie considered significant without extensive studies of the effects of growth conditions. Also, the magnesium concentration required for stability vai'ies from one organism to another.

In view of these difficulties of intcrcomparing ribosomes and because most of the studies of ribosome synthesis have been carried out with E. coli, it seems best to describe the process of ribosome synthesis as it has been observed in this organism. We hope that this will prove to be a specific example of a general process. Furthermore, we have omitted any reference to the synthesis of RNA in virus-infected cells as the relation of this process to ribosome synthesis is still obscure.

VII. SYNTHESIS OF RNA AND RIBOSOMES 299

E. RIBOSOMES AS SITE OF PROTEIN SYNTHESIS

Interest in the biosynthesis of ribosomes stems primarily from the belief that the ribosomes serve as the principal machinery of the cell for protein synthesis. The evidence favoring this view has accumulated over a period of years and is now quite convincing. Cassperson and Brachet observed that RNA was invariably present in high quantity wherever or whenever there was a high rate of synthesis (Cassperson, 1950; Brachet, 1950). Caldwell and Hinshelwood showed a quantitative correlation between the quantity of RNA and the growth rate of a wide range of different cells and different conditions (Caldwell and Hinshelwood, 1950). Subsequently, when it was found that most of the RNA was located in ribosomes, the same evidence indicated that a high concen- tration of ribosomes was correlated with a high rate of protein synthesis.

Further evidence appeared in the studies of incorporation of labeled amino acids into protein. In vivo experiments showed that the radioac- tivity of the microsome fraction has a rapid initial rise, whereas that of the soluble proteins has an initial delay suggesting a precursor product relationship (Borsook et al., 1950; Littlefield et al., 1955). In mam- malian tissues the times involved were about 15 minutes and in bacteria the times were about 15 seconds. Furthermore, with the bacteria it was possible to follow the "pulse" of radioactivity with a "chase" of non- radioactive material and to show a transfer of radioactivity from the ribosomes to the soluble protein (McQuillen et al., 1959).

In vitro experiments showed that ribosomes were an essential com- ponent of any cell-free system capable of incorporating amino acids into protein (Littlefield and Keller, 1957; Tissieres et al., 1960). In these cell-free preparations a large part of the radioactivity incorporated into peptide linkage remained associated with the ribosomes.

It is now widely accepted that ribosomes furnish the sites of protein synthesis and that the information for the assembly of the amino acids in their proper order is transferred either to the RNA of the ribosomes or to an RNA which is associated with the ribosomes while it acts as template for protein synthesis.

II. Properties of Ribosomes in Vitro

A. PREPARATION AND PURIFICATION OF RIBOSOMES

The preparation of ribosomes begins with the harvesting of the cells. A culture in the desired metabolic condition is rapidly chilled to main- tain that state and centrifuged. The resulting pellet is washed three times to replace the growth medium by a buffer suitable for subsequent

300 R. B. ROBERTS, R. J. BRITTEN, AND B. J. MCCARTHY

processing of ribosomes. The buffer most frequently used is tris(hy- droxymethyl)aminomethane 0.01 il/ pH 7.4 with magnesium added as the chloride at the desired concentration (usually 10- i\/).

The cells may then be broken by grinding the pellet with twice its weight of alumina, by repeated freezing and thawing with lysozyme, or by forcing them through a small orifice in the French pressure cell. The pressure cell is preferable for the preparation of ribosomes as it is highly efficient in breaking most cells and the DNA is degraded so that the cell juice is not viscous.

The cell juice is then diluted to less than 100 mg wet w-eight/ml and the cell walls, membranes, and any unbroken cells are removed by centrifugation for 1 minute at 40,000 rpm in the angle head rotor of the Spinco Model L ultracentrifuge. The supernatant fluid (40K I'S) is decanted and the pellet is either discarded or resuspended to recover its content of ribosomes (about 10-20%).

The crude ribosome pellet is then obtained by centrifugation of the 40K I'S. If complete recoveiy of ribosomes as small as 308 is required, 2-3 hours of centrifugation is needed. Alternatively, if only a 90% recovery of 70-lOOS ribosomes is desired, 30 minutes' centrifugation will suffice.

The first ribosome pellet is highly contaminated, partly by small bits of cell wall and membrane and partly by soluble protein and RNA. jMucli of the contaminating material will not resuspend readily and can be removed by a brief centrifugation (40K 1'). Alternate short and long cycles of centrifugation helps to reduce the contamination and the ribosome pellet becomes more colorless and transparent. By appropriate choice of the times of centrifugation, pellets rich in one or another class of ribosome can be obtained. However, this technique is not adequate for measurements of the protein or enzyme content of ribosomes. Further purification of the pellet by sedimentation through sucrose gradients or by column chromatography shows that after three cycles of sedimenta- tion in the angle head rotor as much as half the protein content of the pellet may be due to contamination. The best use of differential centrifu- gation is to prepare partially ])urified ribosomes as the starting material for other techniques of separation.

B. protein/nucleic acid RATIO

Various values reported for the protein and nucleic acid content of ribosomes are listed in Table I. As many of these experiments were carried out before the difficulties of removing protein contaminants were fully appreciated, the protein content is likely to be overestimated. There is no certainty that the protein/nucleic acid ratio varies in

VII. SYNTHESIS OF RNA AND RIBOSOMES

301

TABLE I RNA Content of Ribosomes

Source of ribosomes	RNA (%)	Reference
Calf liver	40	Peterman and Hamilton (1957)
Yeast	42	Chao (1957)
Pea seedlings	40	Ts'O et al. (1956)
Rabbit reticulocytes	50	Dintzis et al. (1958)
Escherichia coli	60	Roberts el al. (1958)
Escherichia coli	63	Tissieres et al. (1959)

different species; the reported variations may be due to variable contamination,

Ribosomes of different sedimentation constants appear to have the same protein/nucleic acid ratio. Figure 5 shows the correspondence be- tween the protein (measured by S^'') and the nucleic acid (measured by

.300

o

CD

d

200

.100

LLr-i II

80

50 1=

OC

8 12 16 20

Fraction number

Fig. 5. Sedimentation analj-sis of a ribosome pellet shows correspondence be- tween nucleic acid content indicated by optical density (O.D.), protein indicated by S", and peptidase activity.

302

U. B. HOHKHTS, H. .1. BRITTKN, AND B. J. MCCARTHY

optical density) in ribosomes separated by setlinientation. Similar results are shown in Fig. 6 using P*- and C'*-leucine. Thus the ribosomes of E. coli, whether 308, 508, 708. or 1008. all have the same nucleic acid

4000 -

3000 -

in 2000

c:

o

CO

ro CL

1000

Fraction number

Fig. 6. Sedimentation analysis of a total cell extract from E. coli cells in tris- HCl 0.01 M, pH 7.4, MgCl; 0.01 M, randomly labeled with C'-leucine and P^^Or"' . Centrifugation 90 minutes at 37,000 rpm, 4°C. Note same ratio of protein/nucleic acid in ribosomes of different sedimentation coefficients. Protein indicated by radio- activity of C"-leucine, nucleic acid indicated by P^l

content of 63% or possibly higher. Notable exceptions to this statement are the particles which accumulate during growth in chloramphenicol and the precursors of the ribosomes (see Sections V-VII).

C. PROTEIN COMPONENTS

1. Structural Elements

The protein and nucleic acid components of ribosomes are readily dissociated either by treatment with 4M urea or by removal of mag- nesium with chelating agents such as disodium cthylenediaminetetra- acetate. Bolton showed the presence of several components separable by chromatography (Bolton, 1958). More extensive studies by Waller and Harris show at least 20 components which are resolved by electrophoresis.

VII. SYNTHESIS OF RNA AND RIBOSOMES

303

Methionine (about 47%), alanine (about 377^ ), and serine (about 11%) account for most of the NHo terminal amino acids and the average molecular weight of the individual protein components estimated from the proportion of NHo terminal amino acids is 25,000 (Waller and Harris, 1961). Since the molecular weight of the total protein of the 308 particle is only 260,000, any one 30S particle could include only about 10 of the observed components.

2. Ribosomal Enzymes

Latent ribonuclease (RNase) was found in chemically isolated ribo- nucleoprotein of E. coli. The enzyme was not active until the structure of the nucleoprotein was disrupted (Elson, 1958, 1959).

.300 -

E O CO 00

Q O

.200 -

.100 -

1 1 1 1 \ n
n 1 1 ~
1 ~ 1 1 I
	-■J^	/SGaloctosidase ] i -
i—	r-l
_i
r		1 •"! r	■
"Uti J
1 1 1 4-^— lj-"

- 80

- 60

40 £

UJ

- 20

Bottom

10

20 25 Top

Fraction number

Fig. 7. Sedimentation analysis shows a correspondence between ribosomes indi- cated by optical density (O.D.) and /?-galactosidase activity. (Cowie et al., 1961.)

Latent RNase also appeared in ribosomes isolated by centrifugation or chromatography. Furthermore, the total cellular content of RNase could be attributed to the RNase of the ribosomes (Bolton, 1958). The enzyme is either localized in 30S particles exclusively (Elson and Tal,

304

R. U. ROUKKTS, H. .1. HKITTEN, AND U. J. MCCARTHY

( )

spuooas/siunoD ui AnAjjoDOipcy

_ -S ««-i

o	o o	o CD	o	o 'a-	o CJ
	^:	1		?		■i
i	L	J
2	1 1	L 1		"^^ 1	\ 1	-| -j

"1 ri^l — ID 1 r

J I I I I L

1	1	;i 1 I--!	1 1	-
3 1	1	1 1	1 1	^

O	O	O
o	O	o
rO	OJ	—

( — ) Huu 092 *D 0 0

		-ii	"oo >>
Top		"^	s:>
•^		o	" "^
CM		2	O 12^
			a
C\J			>. ^ ^ .«
U>		n	T3 °
		2
OJ		03 a S	1 *"
00		-a	P'l
*		^
Bottom		is o	o "
^Top		2 a
(M
		o	o
			fl 72
o		o	■"" -U
CM	F	'ii!	Is O X
	3 c	St	^§
CM			kC
	c	Oj	^ >.
	o	-d	^ ^
00	o	<U
	o	e3
		t-.
	U.	O	1 o
Bottom		O CI	Pa en ,
		>. "i		c' O
Top		a	-a -73
		t,^	fl fl
		o	o o
CM		§	tc CO	,fl
			»0 lO	ri
		Cl
				s
CM		03	-a 0) OJ a
00		a o	i|	o fl
			" o	o
			to «	-<-s
^		l>	1; OQ	fl
		H	o^	o
Bottom		00	9 ■?
			. fl	C/}
		fe	(U o	1 o

VII. SYNTHESIS OF RNA AND RIBOSOMES 305

1959; Spahr and Hollingworth, 1961) or equally distributed among the 30S and oOS particles (Bolton, 1959).

Deoxyribonuclease (DNase) is also present in ribosomes in a latent form (Elson, 1959).

The significance of other enzymes found in ribosome pellets is uncer- tain. The pellet is apt to be contaminated by soluble protein and adsorp- tion of enzyme by ribosomes is difficult to rule out.

A major proportion of leucine amino peptidase was found associated with ribosomes of both SOS and 50S. The profile of enzyme activity corresponds to that of ribosomes in sedimentation analysis, ruling out simple contamination of pellets (Fig. 5). This enzyme shows no latency (Bolton and :\IcCarthy, 1959).

Small traces of )8-galactosidase continue to sediment with ribosomes after prolonged washing procedures (Fig. 7). Here again the activity of the ribosome associated enzyme is partially latent. Acid phosphatase and alkaline phosphatase have also been found in E. coli ribosomes (Cowie et al, 1961).

3. Nascent Protein

Kinetic studies show the rapid appearance of newly incorporated S^^- and C^*-amino acids in ribosomes (Fig. 8). The quantity corresponds to about 0.1% of the total soluble protein, i.e., to the protein synthesized in a period of about 5 seconds (McQuillen et al, 1959). In studies of the synthesis of the structural protein of ribosomes care must be exercised to distinguish the nascent protein, which is only transiently associated with ribosomes and not destined to become a part of their structure.

D. NUCLEIC ACID COMPONENTS

Degradation of the 50S ribosomes by sodium dodecyl sulfate or by phenol releases RNA of 16 and 23S corresponding to molecular weights of 0.55 X 10" and 1.1 X 10^ The 30S ribosomes release a single com- ponent of 16S (Kurland, 1960). Aronson and McCarthy report similar results and describe further a progressive degradation to small units of 13. IS, 8.8S, and 4.4S which was brought about by heating or removal of magnesium by dialysis. The 4.4S component was also found in the RNA extracted from magnesium starved cells (Aronson and McCarthy, 1961; McCarthy and Aronson, 1961). Further evidence for small sub- units of ribosomal RNA shows in X-ray analysis (Timasheff et al., 1961).

The nucleotide composition of ribosomal RNA has been reported by several authors and is given in Table II. Bacteria of different DNA composition show a total ribosomal RNA composition which is quite

306

R. H. Ror?KRTS, R. J. HRITTKX. AND I!. .1. MCrAUrin-

'IWIiLi: II Nucleotide Composition of Various RNA's

Organism	DNA	S-RNA	70S	SOS	30S	Reference
Escherichia coli	A 24	A	—	25.6	24 . 6	Si)alir ;uk1
	T 21	U	—	22.1	21.0	TissitTcs (1959)
	G26	G	—	31.4	31.6
	C 2()	C	—	20 . 9	22.8
Escherichia coli	A 21	A li).7	—	26.4	24.3	Holt on (l'.)5i)j
	T24	U17.2	—	18.3	20 . 5
	G26	G34.2	—	34.8	31.6
	C26	0 29.1	—	20.5	23.6
Escherichia coli	A 24	A 20.5	25.1	25.4	24.8	Midgley (1962)
	T24	U16.5	20.4	19.6	21.5
	G26	G33.2	32.6	33.5	31.0
	C26	0 29.8	21.9	21.5	22.7
Bacillus subtilis	A 29	A 20.2	25.9	26.2	26.5	Midgloy (1962)
	T29	U17.6	20.8	19.3	21.6
	G21	G32.9	31.0	32.0	29 . 6
	C21	0 28.3	22.3	22.5	22.3
Proteus vulgaris	A 32	A 19.1	26.2	26.5	24.7	Midgley (1962)
	T32	U18.3	20.7	20.8	20.4
	G18	G33.3	31.4	31.4	31.9
	C18	0 29.3	21.7	21.3	23.0
Aerobacter aerogenes	A 22	A 19 . 7	25.5	25.6	25.3	Midgley (1962)
	T22	U18.8	21.1	21.2	21.5
	G28	G32.3	31.5	31.2	30.8
	0 28	029.2	21.9	22.0	22.4
Pseudomonas aeruginosa	A 18	A20.8	25.7	26.3	25.1	Midgley (1962)
	T18	U17.1	21.0	21.3	20.5
	G32	G33.8	31.6	31.2	32.8
	0 32	028.3	21.7	21.2	21.6

constant. The increased accuracy obtained by use of the isoto])e dilution technique shows a slight difference in composition between the RNA's derived from 308 and oOS ribosomes.

E. STRUCTURE OF RIBOSOMES

Table III lists the values of the physical constant of purified E. coli ribosomes as measured by Tissieres et al. Molecular weights of 0.9 X 10*^ and 1.8 X lO'' for the 30S and 50S particles are consistent with the molecular weights of 0.55 X lO'' and 1.1 X 10'^ found for the RNA, assuming an RNA content of 63%.

VII. SYNTHESIS OF RNA AND RIBOSOMES 307

TABLE III Properties of E. coli Ribosomes"

Particle	s	D	V	■n	MW*	MW^
30	30.6	2.95	0.64	0.080	0.7	1.0
50	50.0	1.91	0.64	0.054	1.8	1.8
70	69.1	1.83	0.64	0.061	2.6	3.1
100	100.0	—	0.64	0.071	— •	5.9

Key: S, sedimentation coefficient S^.w X 10^^ cm/sec; D, diffusion coefficient D20.W X 10^ cmVsec; V, partial specific volume; 77, viscosity cm/dl; MW, molecular weight X 10-8.

" From Tissieres et al. (1959).

* Calculated from S and D.

" Calculated from S and rj.

Electron microscopy of purified ribosome preparations provides the dimensions shown in Table IV (Hall and Slayter, 1959; Huxley and Zubay, 1960).

Electron microscopy of sectioned bacteria is more ambiguous. Some sections show dense regions which may be due to ribosomes, others which might be expected to show ribosomes do not. It is not at all clear whether these differences are due to differences in staining techniques or whether the ribosomes do not exist as compact spheres in the living cell (Hanzon ef a/., 1959).

	DiMENSIO:	TABLE IV vs OF E. coli Ribosomes
Particle	Shape		Dimensions (A)	Mol. wt. X 10-«
30S SOS 70S	Prolate Oblate Oblate		95 X 170 170 X 140 200 X 170	0.76 2.0 3.4

The protein/nucleic acid ratio of 37/63 corresponds closely to a ratio of 2 amino acids/nuclcotide. Furthermore, a portion of the protein can be removed (see Section III,A) leaving another portion still in the ribosome structure. Electron microscopy of ribosomes stained with uranyl acetate shows no evidence of a protein shell around a nucleic acid core (Huxley and Zubay, 1960). These findings suggest that the 2/1 ratio may not be fortuitous but a consequence of saturating the entire RNA strand w4th protein along its entire length.

The hyperchromicity of ribosomes is 40%, the same as that of the RNA after isolation from the ribosomes. Thus, some aspects of the RNA structure are not appreciably altered by the addition of protein (Schles-

308

R. B. ROBERTS, R. J. BRITTEN, AND B. J. MCCARTHY

singer, 19601. A similar conclusion is drawn from X-ray diffraction studies (Zubay and Wilkins, 1960).

III. Fractionation of Ribosomes

A. CHROMATOGRAPHY

Chromatograi)hy on DEAE cellulose columns provides a means for separating protein, S-RNA, ribosomal RNA, and DNA from ribosomes. It is particularly useful in distinguishing the precursors in ribosome synthesis.

The colunm (1 cm- X 10 cm) is prepared by packing a slurry of DEAE, thoroughly washed with tris buffer (0.01 M, pH 7.6. 0.01 M Mg), at 3-5 psi pressure. The preparation to be analyzed (whole cell juice or ribosome pellet) is adsorbed on the column from this buffer and eluted by 200 ml of the buffer in which there is a linearly increasing concentra- tion of NaCl (0-1 M).

Under these conditions ribosomes elute at 0.4 M NaCl, S-RNA at 0.5 M, and DNA at 0.6 M. Ribosomal RNA (prepared by phenol) remains adsorbed. The ribosomes emerge as 30 and 50S particles with their full complement of protein. If the magnesium concentration is

E O

(a)	10-2 M 0.40M 0.50M 1 1	-
«*««»*4»»	-«««;*/ r^5»»«.i««	>»«=•

Ci in _

o

(c)

2.5xlO-3M

0.40Mn Q50M r 060M

J i [

10 20

Fraction number

Fig. 9. Chromatography on DEAE of 70S ribosomes using different concentra- tions of magnesium. Protein content is indicated by S" radioactivity.

VII. SYNTHESIS OF RNA AND RIBOSOMES

309

decreased the ribosomes lose protein and elute at higher salt concentra- tions, as shown in Fig. 9.

The use of the DEAE column to separate ribosomes and their pre- cursors for kinetic analysis was first reported in 1957 (Roberts et al., 1958). Recent examples of this technique are given in Section V,A. Separation is achieved because of the different protein content of the ribosome and precursors.

B. SEDIMENTATION ANALYSIS

The separation of ribosomes of different sedimentation coefficients can be carried out by sedimenting a layer of ribosome suspension through a solution stabilized by a density piadient (Brakke, 1953; Britten and

Fig. 10. Analytical centrifuge pattern of SOS ribosomes purified by centrifugation through a sucrose density gradient.

310 B. B. BOBEBTS. R. J. BRITTEN. AND B. J. MCCARTHY

Robert*, 1960 •. Figure 10 shows the purity of SOS ribosomes prepared in this way. When small quantities of ribosomes are used in ver>' thin layers excellent resolution can be achieved, as shown in Section V,B below.

C. DENSITY GR.\DIENT B.\NDINC.

Since DXA. RXA. and ribosomes have different densities they can be separated by prolonged centrifugation in appropriate solutions of high density. The materials forai bands at the levels where their density equals the density of the solution. Materials differing by only a few per cent in density can be well separated since the density of the solution varies only slightly from top to bottom of the centrifuge tube. Density banding was used to detect the DXA-RXA hybrid complex (Hall and Spiegelman, 1961; Spiegelman et al.. 1961) and to show an association between DXA and ribosomes (Xisman, 1961).

Density banding also provides a unique method of separating newly formed ribosomes from old ones. The cells are first grown in media containing hea\y isotopes (C", X^^. and H-» and then transferred to media containing light isotopes. The ribosomes formed before and after the transfer can then be separated by density' gradient banding. The value of this technique is illustrated in the work of Brenner et al. in their studies of phage-infected cells (Brenner et al.. 1961 ».

D. ELECTROPHORESIS

Electrophoresis has been used occasionally in the study of ribosomes. Pardee et al. showed a difference in the electrophoretic pattern caused by growth of the cells in chloramphenicol (Pardee et al., 1957). Electro- phoresis was used to isolate the RX'A formed after phage infection (X'omura et al, 1960 ».

E. r«'0-PHASE SYSTEMS

Successful separation of ribosomes from other cellular components by partition in two-phase systems is described by Albertsson (1960i. This method has not been exploited extensively in the study of ribosomes but it gives promise of being extremely useful in the search for complexes of ribosomes with DX'A or for other studies of cellular organization.

F. CONCLL'SION

There is still no general technique for distinguishing ribosomes of slightly different composition. Hence there is no method for determining whether there are two or a thousand different types of ribosomes. The only technique applied to date is the specific precipitation of the ribo-

VII. SYNTHESIS OF RXA AXD RIBOSOMES 311

somes which earn' a particular protein by the antibody to that protein. Ribosomes have been precipitated by specific antibodies (Cowie et al., 1961; Warren and Goldthwait, 1961) but analysis shows only a marginal difference in base composition from the average (Midgley and Gold- thwait, 19611.

J\. Met.4BOLic Pools

For kinetic studies of ribosome sjTithesis to be interpretable in a quantitative way a number of conditions must be met. Among these, two are of primary' importance. First, the cells must be growing in steady-state conditions so that cellular components are not changing in their proportion. Second, the time course of the specific radioactivity of the immediate precursors for macromolecular synthesis must be known.

The first condition can be satisfied by allowing a prolonged period of exponential growth (in an adequate and unchanging medium) before the introduction of the tracer. The addition of a tracer such as S^^O^" or P^-Oi — to a medium which already contains S'-Oi" and P'^Oi intro- duces no disturbance. In principle, labeled amino acids or bases should be added to a medium which already contains the unlabeled amino acid or base without changing the concentration. In practice, however, this procedure is wasteful of the tracer, and no transient in the rate of macromolecular synthesis occurs as a result of the change from an endogenous to an exogenous supply of amino acids or bases. The cus- tomary- procedure is therefore to add a labeled amino acid or base to a culture already growing exponentially on its endogenous supply. De- ficient mutants should not be used in this way as the addition of the required (labeled) nutrient to a starved cell would initiate growth, steady-state conditions would not apply, and the results would not be meaningful.

The second condition requires a technique for the evaluation of the specific radioacti^-ity of the precursors to the macromolecules. In E. coli there exist pools of amino acids in the cell so that exogenous tracer amino acids may be diluted by endogenous amino acids upon entn,-. Also there is exchange between exogenous and endogenous amino acids so that even the external material can be diluted. Following the addition of an amino acid to the medium, it is concentrated (as much as 5000 times) in the cells and depleted from the medium. Furthermore, an amino acid may be converted to another before incoi-]X)ration into protein. These reactions would be difficult to follow in detail by any direct method.

Fortunately, when steady-state conditions prevail, the required

312

H. H. HOHKKIS, H. .1. HHII'IKN . AND H. J. MCCARTHY

in formation can be derived indirectly from simple measurements. The rate of protein synthesis per unit mass of cells is constant; thus, the total rate of entry of radioactivity into proteins of all types must measure the specific radioactivity of the precursor. Direct measurements of the pool and product jirotein have shown that the rate of entry into the protein was proportional to the specific radioactivity of the pool. This observation indicated that the amino acid pool was in fact a precursor to protein (Britten et al., 1955).

The rapid establishment of the full rate of entry of S^^04"" into protein indicated that there was no appreciable pool of sulfur compounds (McQuillen et a/., 1959). Leucine also shows prompt entry into protein (Britten et al., 1962) but a mixture of amino acids from a protein hydrolyzate may show a lag before the full rate is established. This is presumably due to the existence of quite large pools of glutamic acid, alanine, and aspartic acid.

P^-04 is far from an ideal tracer because of the large pool of phos- phate and phosphorylatcd compounds. There is a further delay intro- duced by the time required for the P''" to reach the a position of tiic nucleotides (Bolton and Roberts, 1956). In E. roll the specific radioac-

3CXX)

Uptake of lO'^M C'"- uracil at 37°C f. Co// ML- 30, 1 mg (wet) /ml

2000

%

.£ 1000

I

Time in minutes

Fig. 11. Incorporation of C"-uraciI into the metabolic pool and the HXA of E. coli. The difTerence between the two curves measures the radioactivity of the pool. The initial concentration was low; 10"' Af.

tivity of the nucleic acid precursors rises roughly as (1 — e~*/^") where Tq is 600 seconds. This slowly rising temi makes it very difficult to observe the kinetic delays introduced by subsequent smaller pools of ribosome precursors.

The nucleic acid bases provide much more suitable tracers for kinetic

VII. SYNTHESIS OF RNA AND RIBOSOMES

313

studies of RNA synthesis. In sinte of the presence of i)ools of nucleoside di- and triphosphates, exogenous purines and pyrimidines are rapidly incorporated into nucleic acid (Bolton and Roberts, 1956). Recent studies of incorporation of C'^-uracil (McCarthy and Britten, 1962) and other bases (Buclnvald and Britten, 1962) have given a much more detailed understanding of the process.

Figures 11 and 12 show two characteristic features of the incorpora- tion process. At low concentration two phases of the process are evident, one a rapid rate of incorporation into nucleic acid which persists as long as external uracil is present, followed by a much reduced rate which occurs when the uracil is utilized solely from the pool. At higher concen-

		Incorporat	on of	y
		C"*-	uracil	into RNA	/
<v					/
o	3	—			/
o to				10'' M	/ y
>->				V	/
D					/
'.O				/	/
^	2
o				/	/
_>>				/ /	s ^
;>				/ /	5x10-^
o				f /
			/	f
Q			/	/
■■5	1	~	/,	/
o			/ /
cr		/	1	1	1 1

20

30

40

Time in minutes

Fig. 12. Incorporation of C'Miracil into the RNA of E. coli from high concentra- tions of external uracil. The ordinate scales of the two curves are different.

trations the two phases are also observable. However, in this case the rate of incorporation during the second phase is more rapid, as the direct incorporation continues throughout and is supplemented during the second phase by C"-uracil drawn from the pool.

To account for the direct entry of C"-uracil together with utilization of pool material a mechanism such as that shown in the diagram below must be envisaged.

Exogenous C"-uracil

Endogenous C'--uracil

/•

NA

b\\c

314 H. B. ROBERTS. R. J. BRITTEN. AND B. J. MCCARTHY

The maxinium entry flow (a) is equal to the flow (d) which is required for the synthesis of nucleic acid (NA). Furthermore, the flow due to endogenous synthesis (e) is regulated by feedback control so that a -\- e = d. Thus the pool (S) remains constant and cannot be expanded or depleted. The pool (Pi cannot contain more than 5 seconds' supply of uracil compounds or an observable kinetic delay would appear in the entr>' of C'^-uracil into NA. Forty per cent of the exogenous uracil proceeds directly to NA and 60% enters the pool S, so that the flow (b) is roughly 607c of (a).

Labeled bases, once incorporated into the pool, are not rapidly removed by exchange with external material. Hence the incorporation of tracers into nucleic acid cannot be abruptly terminated by dilution of the exogenous tracer material. The best condition which can be achieved is the abrupt transition from phase 1 to phase 2. By allowing only a very short time before the external material is exhausted, a ratio of 20: 1 in the rates of phase 1 and phase 2 can be obtained.

The conversion of uracil to cytosine causes a further complication in the kinetics of incorporation because the proportion of the labeled material which enters nucleic acid as cytosine increases with time.

Guanine, adenine, and cytosine show a similar behavior but the pool sizes and the proportions entering directly differ, as shown in Table V.

TABLE V Incorporation Ch.\r.\cteristics of Four RNA Bases

Direct entry Pool (S) time constant

Base C^)" (minutes)'"

Uracil 40 + 5 10+2

Cj'tosine 45+5 28 + 5

Guanine 78 ± 5 3.3 + 0.5

Adenine >50 3-9<^

" Shows the initial rate of incorporation of the tracer into RNA as per cent of the ultimate rate, per unit cell mass.

* Shows the time constant (1/e) of the pool (S) derived from the exponential decay of its radioacti\ity during the second phase of an experiment at low concentration such as Fig. 11. These values are in agreement within the errors indicated with the values derived from the time const-ant with which the incorporation into RXA (at high con- centrations such as Fig. 12) approaches its ultimate rate.

« The pool of adenine compounds does not show a single time constant and a detailed analysis to resolve the individual time constants has not been carried out.

Once the complications introduced by the pool are recognized and measured, appropriate corrections to the kinetics obser^-ed in the macro- molecular components can be made. Without this understanding, the two

VII. SYNTHESIS OF RNA AND RIBOSOMES 315

phases of incorporation into nucleic acids might be misinterpreted as indicating an exchanging component of the nucleic acid.

V. Kinetics of Incorporation into RXA of Ribosomes

Kinetic studies of ribosome synthesis depend upon methods of sepa- rating various classes of macromolecules. The techniques which have proved to be most useful are chromatography and sedimentation anal- ysis. The latter technique can be used both for the analysis of the ribonucleoprotein particles and of the RXA extracted from them.

In the following paragraphs details have been drawn exclusively from the work carried out in our laboratoiy (Roberts et al., 1958; McCarthy and Aronson, 1961; McCarthy and Britten. 1962: Britten et a/.. 1962 1. The reasons for this are twofold. On the one hand, we are more familiar with our own work, and on the other, different experiments can be related more surely because they were carried out under similar conditions. Experiments in other laboratories also contribute to our understanding.

An RXA fraction of high specific radioactivity has been separated by electrophoresis (X'omura et al., 1960). Otaka et al. have observed precursor product relationships among fractions separated on DEAE (Otaka et al.. 1962 1. Gros et al. have studied an RXA component of high specific radioactivity which is dissociated from ribosomes when the magnesium concentration is reduced and which has an average sedi- mentation coeflBcient of about 12S (Gros et al.. 19611.

A. DEAE chrgmatogr-aphy

Figure 13 shows the elution patterns obtained from cells given progressively longer exposures to C^^-uracil. After breaking the cells, the walls and membranes were removed by a brief period of centrifugation. Ribosomes and their precursors were sedimented by prolonged centrifu- gation i40K 240' I leaving the S-RX'A in the supernatant fluid. This procedure is desirable because S-RXA elutes at 0.5 M X'aCl contami- nating ribosome precursors. DXA. which also elutes in this region, was removed by DX'ase.

The patterns show clearly that the C^^-uracil first appears in ma- terial eluting at 0.6 -V XaCl. After the first few minutes this component reaches a saturation level and C^*-uracil builds up in material eluting at 0.5 -1/ XaCl. As time goes on the second region approaches saturation and C^*-uracil appears in the bulk of the ribosomes which elute at 0.4 M X'aCl. The column seems to have separated two sequential precursors from the final product. This qualitative impression is confirmed by the

310

R. U. ROBERTS, R. J. BRITTEN, AND B. J. MCCARTHY

	(a)	25"	_
1.00	-	\
0.50	~^ j	\ A \ / \
			«>~U..M~..

(b)		r\	50'
-		A
	1 ^ ^	1 '"--T..	V	"XWNll

- 600

- 400

- 200

E O CVi

	(c) 100 ■•
1.00	1 'i
050

050 -

I

- 2400

- 1600

-800 £

- 10.000

- 5000

40 50

Fraction number Fig. 13. Analysis of total ribosomal pellets (40K 240 minute pellets) on a DEAE-cellulose column. Linear sodium chloride gradient from 0.2 3/ to 1.2 M in tris-HCl buffer 0.01 M containing MgCL 0.01 M, pH 7.4. Salt gradient 0.004 M/ml Volumes collected 3.6-3.8 ml. (a) 25 seconds exposure to C"-uracil, (6) 50 seconds, (c) 100 seconds, (d) 3 minutes, (e) 6 minutes, (/) 12 minutes. The three sequentially labeled peaks appear at 0.6 M, 0.5 M, and 0.4 M NaCl.

quantitative analysis given below (Section V,D). As these two pre- cursors have quite distinct elution properties, as well as other charac- teristics to be described later, we have used the names "eosome" and "neosome" to distinguish them from ribosomcs and to indicate their order in the sequence of synthesis.

B. SEDIMENTATION ANALYSIS

Most of the earlier studies of the kinetics of ribosome synthesis were carried out by means of sedimentation analysis in a magnesium concen-

VII. SYNTHESIS OF RNA AND RIBOSOMES 317

tration of 10- M, which is necessary for the preservation of large ribo- somes. At early times the specific radioactivities of the 308 and 50S ribosomes present in the extract were considerably higher than that of the 70S. These small particles, described as native 30S and 508 to dif- ferentiate them from those derived by breakdown of the 70S at lower magnesium concentrations, represent together some 10 to 20% of the ribosomal material. Their higher specific activity after a pulse was interpreted as showing a precursor relationship to the 70S ribosomes.

Figure 14 shows the analysis of extracts of four samples of cells taken after 10, 20, 50, and 55 minutes' exposure to C^^-uracil. 508 and 308 ribosomes show as shoulders in the ultraviolet absorption profile. At 10 minutes the specific radioactivities of the 508 and 30S ribosomes are at least three times that of the 70S. This difference in specific radio- activity decreases with time and has practically disappeared by 55 minutes.

When a series of cell extracts was analyzed after much shorter exposures to C^^-uracil the results were not interpretable in terms of a simple hypothesis. At early times an appreciable fraction of the radio- activity appeared in the 708 particles. The quantity was greater than would be expected in an end product. This suggested direct entry of RNA into 708 without the delay brought about by passing through the large pools of 508 and 308 precursors. Another possibility was that the labeled RNA was loosely associated with the 708 peak rather than being incorporated into the large ribosomes.

Because of the complexity encountered in the analysis of cell extracts made in 10"- M Mg*"", a search was made for conditions more suitable for the resolution of precursors from products in ribosome synthesis. Comparison was made of a pulse labeled cell extract prepared and analyzed in a range of magnesium concentrations (Fig. 15).

At 10'- M Mg*"^ (Fig. 15a) most of the ribosomes are present as 70S but most of the radioactivity runs as 508 and 308. The apparent specific radioactivity of the small ribosomes is more than three times that of the 70S. Cells broken in the presence of 3 X 10"' M Mg++ (Fig. 15b) have most of their ribosomes as 508 and 30S. At the same time about half of the radioactivity associated with ribosomes in Fig. 15a now appears in a peak of about 148. There is now little difference between the specific radioactivities of the three groups of ribosomes although there is a peak of radioactivity running between the main 508 and 30S peaks. In 10'^ M magnesium (Fig. 15c) an even higher proportion of the radioactivity appears in a 148 peak, and the faster of the other two peaks of radioac- tivity appears to lag behind the 508 ribosome peak. The distribution of radioactivity between the three objects is very similar at 10^^ M Mg^*

318

R. B. ROBERTS, R. J. BRITTEN, AM) B. J. MCCARTHY

:± E O

CO CM

d

	(a)	10'
	-	TQS SOS	30S
	1 1	i *	1
0.60	1 1 f	1 \
0.40	/ / 1 1	1	f*
	1		/ /
0.20	'k	1	/ 1 1 1

		(c) 40'
0.80	h-	70S SOS	30S
	_	f 1	1
0.60	-	^	-
0.40	-		A
	-	i K	Jr
020	\	1 1 1	1 1

- 600

- 400

200

Bottom

15 20

Top

Bottom

(d)	n ^^'
70S	\ SOS 30S
\	f" -
- 1	-
i
/ /	'\ \ j'\
1	1 1 1 1

- 800

- 600

- 400

200

3 O

o

20

Top

Fraction number

Fig. 14. Sedimentation analysis of four total cell extracts taken from cells given (a) 10-minute, (6) 20-minute, (c) 40-minute, and (d) 55-minute exposures to C"-uracil. Centrifugation 90 minutes at 37,000 rpm. Cell washing and breakage and centrifugation carried out in tris-HCl 0.01 M, pU 7.4, MgCL 0.01 M.

but the increased centrifugation time in Fig. 15d produces higher resolution. In this instance the front-running peak of radioactivity is clearly resolved from 50S ribosomes and appears to have a sedimentation coefficient of about 43S.

The resolution of early labeled RNA from the bulk ribosomes sug-

VII. SYNTHESIS OF RNA AND RIBOSOMES

319

70S 50S 30 S

^ i I

,,s 70S SOS

(b) ^ I

3xlO"'M

30S

CL

30 40

Fraction number

Fig. 15. Sedimentation analysis of a total extract from cells labeled with P'' for four generations and with C"-uracil for 2 minutes. Cells washed and extracts prepared and run in tris-HCl 0.01 M, pH 7.4 (a) containing MgCL at 10"= M, (b) 3 X 10'' M, (c) 10"' Af, (d) WM. Centrifugation : (a), (b), and (c) 90 minutes at 37,000 rpm; id) 150 minutes at 37,000 rpm, 4°C.

820

R. B. ROBKRTS, R. J. HRl'ITKN, AND B. J. MCCARTHY

gCf^ti'd that tlicsf cDiiditions of cell breakage and analysis of extracts would be suitable for the determination of the details of the kinetics of ribosome synthesis.

Therefore, analyses were made in 10 ' M magnesium of six samples of extracts prepared from P'*- steady-state labeled cells given from 30

	(a)
	30"
1200	1 - II II M II 1 1 II 11	~
800	II
400	i . T 1 /^	4 n , 1'
	' \ I .^\	A,i
		N-

	(d) 4'
1200	-
800	_ 1
400	1 \

10 20 30 40

(b) r
50S 30S
-♦ i	-
II
i	A
w . 1 i A ' \ 'J'

10 20 30 40

Fraction number

- 600

400

200

3000

- 2000

1000

10 20 30 40

Fig. 16. Sedimontation analy.sis uf six DNawe treated total extracts made and centrifuged in tris-HCl 0.01 M, pH 7.4, MgClj 10 ' M from cells steady state labeled with P" given (a) 30 seconds' exposure to C"-uracil, (b) 1 minute, (c) 2 minutes, (d) 4 minutes, (e) 7 minutes, (/) 10 minutes. Centrifugation : 175 minutes at 37,000 rpm, 4°C.

seconds to 12 minutes' exposure to C'^-uracil (Fig. 16). The extracts were treated with DNase prior to centrifugation. The P^" profile shows three main peaks of 50S and 308 ribosomes and soluble RNA.

The patterns indicate a sequence of precursors and products. At the earliest time the C^^-uracil is carried by material sedimenting in a broad band centered at about 14S. Later, C'^-uracil shows at 30S and at 438.

VII. SYNTHESIS OF RNA AND RIBOSOMES 321

Finally, the C'^-uracil accumulates in the end products, the 308 and 50S ribosomes. In addition, radioactivity builds up in S-RNA which sediments at 4S. A qualitative comparison with the sequence of events revealed by DEAE chromatography suggests that the first precursor, eosome, sediments at 14S and that neosomes are present in both the 43S and the 308 regions. The 308 and 508 ribosomes correspond to the main ribonucleoprotein peak of the column analysis.

C. QUANTITATIVE KINETIC ANALYSIS

To compare the results of experiments with theoretical models the experimental data are expressed in terms of functions ^.v. These functions and the other needed symbols are defined below.

r: Time after addition of tracer; units such that t = 1 when the cells have grown by a factor e; Q = Q^e'' gives the growth of the cells or any components. m: Effective specific radioactivity of tracer; units such that the specific radioactivity of the RNA will approach m after a long period of growth at a constant tracer concentration. X: Quantity of a component, in general. X* : Its radioactivit3^ fix'. Its specific radioactivity. E, N, R: Quantities of three specific components. T: The total of all components.

«.v=^-| = ^ (1)

This choice of units is designed to allow different experiments carried out with different growth rates and different effective tracer radioactivi- ties to be plotted on the same graph.

While ^A is adequately defined by Eq. (1), the following general definition may perhaps be useful, ^.v is the ratio of the number of labeled atoms at a given time in a given class of molecules to the number of atoms in the whole cell which would be labeled after a long period of growth during which the labeling conditions were precisely constant, ^.v is thus a measure of the number of newly synthesized molecules which occur in a given class, and has been named the "newly" synthesized fraction."

When there are two sequential precursors (E and N) to a product R

E

-

N

->

R

322

H. H. ROBERTS, R. .1. HHll IKN. AND H. J. MCCAKTIIY

<t>T = ^ — C

(2)

(3)

0£ = ^!i -oxpi-tr/A-H}

NL E{R + iV) exp [- (r/A>] - TAT exp [-iR/N)r]\ ,,, •^-^ = r t^ E(R + AO - TA^ / ^'^^

and

0/e = </>7' — <I>E — 4>.v

(5)

The dei'ivation of these equations and the restrictive conditions governing their use are given in (Britten and iVIcCarthy, 19621.

It can be seen from Eq. (1) that the ultimate value of (/> for any

0.0001

TIME-r (Defined by growth Q = OQe' )

P'la. 17. I>oK-log plot of tlip time counso of the radioactivity (expressed as <p) of the ribonuclooprotein fractions resolved by DEAE chromato^;raphy. Data from Fig. 13 and Fig. 1 of McCarthy el. al. (1962). The four curves have been calculated from Eqs. (2), (3), (4), and (5) for the ca.se E = 0.027, N = 0.068, R = 0.905.

VII. SYNTHESIS OF RNA AND RIBOSOMES

323

component is the proportion of that component. At early times both ^e and cjiT increase as t, </)y increases at t- and <j)ii increases as t\

In Fig. 17, <^.v is i)lotted for the two precursor fractions and the product resolved by chromatography, together with theoretical curves calculated from Eqs. (2), (3), and (4) for E = 2.7%, N = 6.8%. (The details of the procedure for deriving values of cj> from the experimental data are described in (McCarthy et ai, 1962). The quantitative analysis thus verifies the assumption that the three components show the kinetic features of two sequential precursors and a stable product.

The corresponding analysis of the peak resolved by sedimentation is more complicated since there are more components to consider. As a first step the 14S region is compared with other regions of the sedimenta- tion pattern by plotting its radioactivity {<j>e) together with ^r and (}>r' {<{>R' = <l>5o + (}>43 + <j>3o) (Fig. 18). Initially, all of the radioactivity

SEDIMENTATION ANALYSIS 10"^ Mg"*

C -Uraci incorporated info RNA

Fig. 18. The C"-uraoil label incorporated into total RNA (#), the eosome region 8-20S (D), and the ribosome region (O), i-e., material between SOS and 50S including neosome and ribosome. Data from Fig. 16. Tlie label in each component is plotted as <p functions <Pt, <Pe, ^n + r:

324

R. B. ROBERTS, R. J. BRITTEN, AND B. J. MCCARTHY

appears in the 14S region and the shape of the curve indicates that the total flow of material to ribosomal RNA passes through this stage. The curve for ^k-, has zero slope at zero time showing that the entry of radioactivity is delayed by a precursor and that little, if any, C^^-uracil by-passes the precursor. The values of <j)ir at early times indicate a

0.1

0.001

: SEDIMENTATION ANALYSIS
- Newly synthesized fraction of RNA		/
-	Totah«/	A-
: ^	osome / /
43 S Neos	ome 7 / d/ /
:	/ J SOS Ribosome
- /	/		Generation tinne
1 1 / 1 1 1 1	/<? 1 1 1 1 1 1 1 1		1 J 1 J 1 J 1

0.0001

0.001 0.01 1 0.1 1.0

TIME -T ( Defined by growth Q = QqS ^ )

Fig. 19. Log-log plot of the newly synthesized fraction of RNA present in the total cell e.\tract (#), eosome (A), 43S neosome (D), and 50S ribosome (O) as a function of time. Data from Fig. 16. The curve drawn for the SOS ribosomes is the calculated curve shown on Fig. 17 multiplied by 2/3 since the SOS ribo.somes account for 2/3 of the ribo.«omal nucleoprutcin. The agreement is striking.

precursor quantity of 3.5%. The quantity of precursor estimated from its final level is 2%. Although this discrepancy may indicate complexity of the first precursor, the material which sediments in the 14S region can be identified with the eosome fraction resolved by chromatography.

Figure 19 shows ^^ plotted together with 04,j, ^-,„ and 4>r- At early times 0,.i i.s proportional to r- and <^,-,o is proportional to t'\ At late times

VII. SYNTHESIS OF RNA AND RIBOSOMES

325

^43 accounts for 2/3 of the neosome fraction resolved by chromatography and ^50 accounts for 2/3 of the ribosome fraction.

The remaining 1/3 of the ribosome fraction is clearly accounted for by the 30S peak. However, this component must also include as an unresolved peak the remaining 1/3 of the neosome fraction. In Fig. 20

0.01

0.00

0.0001

0.001

TIME -T (Defined by growth Q = QQe*^)

Fig. 20. Log-log plot of the newly synthesized fraction of RNA present in total cell e.xtract (#), 308 region (D), and 43S + 50S region (Q) as a function of time. Data from Fig. 16.

<^3o is compared to ^43 + 50- If the 30S and SOS ribosomes were synthesized independently from a common pool of eosomes, and if the 30S peak included both the neosome and the ribosome stage of synthesis, then ^43+50 should be twice </)3o at all times. At early times, however, the 30S component has more than its share of radioactivity. Furthermore, at

326 R. B. ROBERTS, R. J. BRITTEN, AND B. J. MCCARTHY

early times the 43S coinponent aceounted only fur 1/3 of the neosome fraction. One interi)retation of these results is that the BOS peak inchules not only the SOS neosome static of the SOS ribosome but also a SOS neosome precursor to the 4SS.

D. STAGES OF SYNTHESIS OF THE RIBOSOMAL RXA

From the analysis of uracil-labeled cell extracts, it is clear that at early times all the radioactivity is present in a fraction, named eosome, peaking at 14S. The kinetic behavior of the com])onent has been studied and its size estimated at 2-37^ of the total ribosomal RNA. It can be shown from the same experiments that label entering nucleojirotein of higher molecular weight is delayed by a component of about the same size. Together these observations suggest that the eosome is predomi- nantly precursor to ribosomes.

In another recent study (Gros et al., 1961) the rapidity of labeling of the eosome has been invoked as a criterion of turnover. These authors have suggested that the rapid rate with which tracer enters and leaves the eosome fraction proves instability of these RNA molecules. In gen- eral, however, it is clear that the rapid labeling of a small component indicates merely that the flow into it is much larger than that required to maintain its quantity in the growing cell. The loss of label when an excess of unlabeled isotope is added shows only the already obvious fact that there is also a transfer of material out of this component. These measurements cannot be expected to distinguish between a precursor through which label flows to a product and a compound which is synthe- sized and then is broken down to its constituent parts. What we have demonstrated is that there exists a precursor-product relationship be- tween the eosome and the product ribosomes for the transfer of C"- uracil radioactivity.

This does not prove that labeled molecules initially observed as eosomes are incorporated as such into completed ribosomes. An alterna- tive involves breakdown of the eosome molecules and a synthesis of ribosomes by means of a quantitative reutilization of the labeled degra- dation products. It should be pointed out that, in common with many similar tracer studies, the latter alternative cannot be distinguished. For quantitative reutilization the breakdown has to occur in such a way that there is no mixing with other RNA precursors. The kinetics of tracer flow are most simply interpreted on the basis of precursor product relationships, although other evidence discussed in Section VIII points to greater complexity. Here we follow the simple interpretation as a first approximation.

As far as the details of ribosome synthesis are concerned, it is clear

VII. SYNTHESIS OF RNA AND RIBOSOMES

327

that three main stages may be described. The clearest demonstration derives from the rate of labeling of the main ribonucleoprotein peak in the column analysis. Its radioactivity rises initially as the cube of time. The total label present at early times is so low as to require two sequen- tial precursors. The total precursor material can be only 10% of the total RNA and a single precursor of this magnitude could not introduce such a delay into the product. The diagram shown in Fig. 21 summarizes

DEAE COLUMN ANALYSIS

Molarity of NoCI at peak

0.6 Q5 0.4

Fig. 21. Schematic diagram of the flow of RNA in ribosomal synthesis. The diagram indicates the way in which the information derived from DEAE-cellulose column analysis and sedimentation analysis has been combined to give a more com- plete scheme. The figures in the boxes represent the percentages of the total ribo- somal RNA existing in each of the states in a steadily growing cell. Movement downward is a re.sult, principally, of the completion of RNA subunits. Movement to the right is principally a result of the addition of protein.

the results obtained from chromatography and sedimentation and indi- cates the correlation between the two analyses. In the following para- graphs the evidence which supports each of the features of this diagram will be considered.

1. Eosome

The interpretation of the role of the eosome as precursor to all of the ribosomal RNA has been discussed above. Both the peak at 0.6 M in the column analysis and the broad region peaking at 14S in the sedimenta- tion analysis carry all of the radioactivity at early times and later saturate at about the same quantity of radioactivity.

328 K. H. RdBERTS, R. ,1. HRITTKN, AM) B. J. MCCARTHY

2. Flows a (ind h diid (Ik .iOS Xcnsomc Precursor to oOS RiUnsdiiics

It i:? clear from the coluiun analyrsis that there exists a neusuiue pre- cursor to 308 ribosomes since the peak eluted at 0.4 M contains both the 308 and 508 ribosomes, the radioactivity entering it is very strongly delayed and rises as r- at early times. The only location in the sedi- mentation analysis (Fig. 16) where such an object can occur is in the 308 region. Because of the lack of resolution it has been impossible independently to assess its kinetics of labeling or measure its quantity. After correction for the trails of 438 and eosome peaks in Fig. 16c and d, it appears that the C^* radioactivity in the 308 region reaches its maximum somewhat behind the P-'- peak corresponding to the 308 ribosomes. This result is quite uncertain but leaves an impression of heterogeneity in the 308 region.

3. Flows c and d and the SOS Precursor to 508 Ribosomes

Figure 20 shows that there is a greater flow into the 308 region than is required for the synthesis of the 308 ribosomes; in fact, more than half the flow from the eosome passes into the 308 region at early times. This shows that a part of the flow that ultimately reaches the 508 ribosome passes through a 308 neosome. We have chosen to indicate on the diagram that V2 of the flow to 508 passes this way for the following reasons. In the first place, the 438 radioactivity (Fig. 19) is 1/3 of the neosome radioactivity (Fig. 17) at early times. In the second place, a sedimentation coefficient of 308 suggests that I/2 of the 508 RNA is al- ready present. It is not clear whether the two 308 neosomes indicated are identical to each other. The nucleotide compositions of the 508 and 308 ribosomes differ but this difference could be made up through flow e. The total quantity of the 308 neosomes can be crudely estimated by the difference between the amount of 438 neosome (4-5%) and the total amount of neosome (7%) indicated on Fig. 17. No estimate of the relative quantities of the two objects indicated can be made if in fact they differ.

4. The 43s Neosome and Flows e and f

The 438 neosome stands out clearly in sedimentation analysis (Fig. 16) at the appropriate time. The time course of labeling shown on Fig. 19 is that of the neosome or second stage in the sequence. The radioac- tivity rises at t- initially and levels off later. At early times the curve has the shape of ^^ (Fig. 17) but only Vr^ of the magnitude of ^.v. If the total flow to the 508 ribosomes passed directly to the 438 neosome from the eosome, ^^.-s would be expected to be just % of </)\. Therefore,

VII. SYNTHESIS OF RNA AND RIBOSOMES 329

it is clear that about half the flow to 50S ribosonies passes to the 43S neosome from the eosome.

It appears certain that no eosomal RNA goes directly to the SOS ribosome because <^5o is proportional to t^ at early times. The 43S neosome is shown on the diagram (Fig. 21) as containing the full complement of RNA of the oOS ribosome. This seems likely from its sedimentation constant.

The quantity of 43S neosomal RNA can be estimated to be about 5% from the steady-state P^- radioactivity that remains in this region after the 308 and oOS contributions have been subtracted, assuming reasonable and symmetrical peak shapes. The specific radioactivity esti- mated on this basis at early times is just V2 what would be expected if the total flow to 50S ribosomes passed directly from the neosome to the 43S. This is, of course, consistent with the diagram since % of the flow should be delayed by the 308 neosome precursor to the 438.

The diagram (Fig. 21) shows a set of sequential relationships which are more complex than the two sequential precursors shown on Fig. 17. The diagram suggests that half of the radioactivity of the 508 ribosomes should rise in proportion to t* for a time while the other half should rise as t^. All of the 308 ribosome radioactivity should rise as t^.

While deviations from the three-stage model showed clearly in the time course of labeling precursors such as the 308 and 438 neosomes, the accuracy of the data is not adequate to resolve the predicted fourth power component in the final product.

These studies of the synthesis of the RNA portion of ribosomes provide only a few hints to the concurrent process of the addition of protein. The step from 438 neosome to 508 ribosome involves both a change in sedimentation coefficient and a change in the elution from DEAE, but no additional RNA is added. It seems quite obvious that this change is due to a change in the protein content. Studies using C^*-leucine to follow the synthesis of the protein are described in Section VII.

VI. Kinetic Studies of Phenol-Extracted RNA

The kinetics of incorporation of tracers can also be observed in the purified RNA extracted by phenol. After short exposures to C"-uracil some radioactivity is associated with the main peaks of 168 and 238 RNA and the rest appears as a component of average sedimentation coefficient of about 88 (McCarthy and Aronson, 1961; Gros ef al., 1961). At the earliest times, however, there is always a considerable fraction of the radioactivity of sedimentation coefficient of 168 or greater even though the specific radioactivity of the 88 region is high (Fig. 22).

330

H. H. RllRKRTS. H. J. IJRITTKN. AND H. J. MCCARlin

I']\'i(l('iitly such results ai'c not. couipatihlc with any >iin|)lc ])i'ccur.sor product relationship.

Phenol-extracted UNA may also be fractionated by chromatography on columns of methylated scrum albumin adsorbed on kieselguhr (Man- dell and Ilei-sliey, 19()()). A clear sejjaration is obtained Ix'tween soluble KNA. DNA, and hiyh molecular wei(2;ht RNA. Analyses of pulse-labeled

Bottom

20 Top Bottom 5 Fraction number

cr

20 Top

Fig. 22. Sedimentation analysis of RXA extracted from ribosomi>s («) from cells given a 30-second exposure to C'^-uracil, (b) from cells given a 30-second exposure to C"-uracil followed by 20 minutes growth in C'"-uracil at a 200-fold excess con- centration. Cent rifugat ion 230 minutes at 37,000 rpm, 4°C.

RNA slio\y two main features. At early times almost all of the radioac- tivity appears in the I'cgion of high moUn-ular weight RNA in three peaks, none of which corres])onds exactly to the two ])eaks of ribosomal RNA (Fig. 23). JVIoreover, the over-all specific radioactivity of the high molecular weight region is greater than that of soluble RNA (e.g., a ratio of 3-4 at 1 minutes) (Fig. 23).

Analysis of a number of time points in this way allows deterniinalion of the delay of entry of label into S-RNA and DNA relative to the high molecular weight RNA. Parallel experiments carried out with C^'- uracil and P'- as the label lead to a delay time of 1-2 minutes. The interpretation has been made (Midgley, 1962) that the entry of label into both S-RNA and DNA results from nucleotide matei'ial produced

VII. SYNTHESIS OF RNA AND RIBOSOMES

331

by the degradation of a part of the high molecular weight RNA fraction. Thus if the D-RNA portion of the eosome fraction were degraded after an average lifetime of some 2 minutes, the delay in labeling could reflect the utilization of these secondhand nucleotides for S-RNA and DNA

synthesis.

3000 -

a;

"^ 2000

c

=5 O

o

CVI

ro CL

1000

	S-RNA	DNA	Ribosomal	RNA
	1	i
-			-M	» r^^	-
	p32	.r-v	z/^	y V V	V
	^^"^ ^sw^^	^o)"*"'**	C^ 1	V	h^

- 600

400

- 200

Z5

o o

100

200

300

Mi. Eluote

Fig. 23. Chromatography on a column of protein-coated kieselguhr of phenol- prepared nucleic acids from E. coli labeled randomly with P^^Or'" and for 1 minute with C"-uracil. Elution with a linear gradient of NaCl in 0.04 M phosphate buffer, pH 6.7, from 0.4 3/ to 1.1 M. Flow rate 35 ml/hour.

In support of this view is the fact that the delay in S-RNA and DNA labeling by C^*-uracil is abolished by the presence of chloram- phenicol. Under these conditions, where C"-uracil accumulates in the eosome fraction (see Section VII), nucleotide analyses indicate that D-RNA is synthesized but degradation is markedly reduced (Midgley, 1962). There is then little transfer of label from D-RNA to S-RNA and DNA.

VII. Kinetics of Incorporation into Protein

A. INCORPORATION OF C"-LEUCINE INTO THE RIBOSOMES

The separation of ribosomes from their precursors by chromatogra- phy depends, presumal)ly, on differences in their proportions of protein. Thus it appears that the product ribosomes are not formed by accretion of material having a constant nucleic acid/protein ratio but that the addition of RNA and protein occurs in time separated stages. J\Iore

332

H. H. ROBERTS, R. J. BRITTEN, AND B. J. MCCARTHY

rated
	'
O	f
Q.	J
o	r
O	1
c	1
— 4000	- L' ^
<u	r
3	j
C	JAM
e	j lO-^'M
^ 2000	7 C'2 leucine
	/
c	J
J	1	1 1 1 1

Time in minutes

Fig. 24. The incorporation of C'Meucine (10" 3/) into protein in a culture of E. coli ML30. C'-leucine was added to a final concentration of 10 ' M at 1 minute.

direct evidence for these processes can be obtained by observation of the kinetics of incorporation of amino acids into ribosomes.

These experiments suffer, however, from two technical difficulties. A part of the newly formed protein found associated with ribosomes is not destined to remain in the ribosomes but is nascent protein which is sub-

•a.

(b) ,
II
1 1	1 1 1

5 O

20 30 40

Fraction number

Fig. 25. Sedimentation analysis of three total cell extracts prepared from cells labeled with P" for four generations and C"-leucine for 1 minute followed by (a) 10 seconds, (b) 2 minutes, (c) 7 minutes, 45 seconds in C'Meucine (see Fig. 24). Cells washed and extracts prepared in tris-HCl 0.01 M, pH 7.4, containing MgCL lO"' M. Centrifugation 165 minutes at 37,000 rpm, 4°C. P'= shows the location of the SOS and SOS ribosomes.

VII. SYNTHESIS OF RNA AND RIBOSOMES 333

sequently released as soluble protein (see Section I,E). In addition, that part of the newly formed protein which is destined to be a part of the ribosomes does not appear to be firmly bonded to the RNA because a considerable proportion is not eluted from DEAE. Thus chromatography cannot be used to complement the results of sedimentation analysis and the sedimentation analysis is obscured by the presence of nascent protein.

The time course of incorporation of C^^-leucine is shown in Fig. 24, taken from Britten et al. (1962). Figure 25 shows the sedimentation profiles obtained at various times after a 1 -minute period of incorpora- tion was terminated by dilution of the tracer.

The results differ markedly from those obtained with uracil as a tracer. There is no sign of the delay which would be apparent if a con- siderable pool of precursors to ribosome protein existed. Most of the C"-leucine radioactivity appears to enter the SOS and 50S ribosomes directly. In addition, a part of the tracer enters the 43S neosome and is subsequently transferred to the SOS ribosome.

B. EFFECT OF CHLORAMPHENICOL ON RIBOSOME SYNTHESIS

Chloramphenicol, which inhibits protein synthesis, provides a sub- sidiary technique for showing which stages of ribosome synthesis depend on the addition of protein. Figure 26 shows that in the presence of chloramphenicol most of the C"-uracil tracer is restricted to particles having sedimentation coefficients less than 20S. This RNA has been shown by Bolton (1959) to have ribosomal base composition.

C. ASSEMBLY OF THE RIBOSOMES

The experiments reported relating to the synthesis of ribosomal protein do not add any fundamentally new features to the sequence of ribosome synthesis. They do, however, confirm other indications that some of the steps in the sequence represent the addition of protein. The fact that a high proportion of the leucine label enters 30S and 50S ribosomes directly with delays of less than 1 minute shows that the last stages of synthesis (30S neosome^ 30S ribosome and 43 neosome-^ 508 ribosome) involve the addition of protein. This is in agreement with the fact that no new RNA is incorporated directly into ribosomes and with the low protein RNA ratio of neosomes as indicated by their column behavior.

One other feature is clear from the leucine pulse and chase experi- ments. Since the 43S receives some label at early times it is evident that not all the protein of the 50S is added in one step. In fact, the chase experiment shows that the 50S increases some 30% in specific radioac- tivity during an 8-minute chase period at the expense of the 43S. Rough

334

R. B. KOHEHT;

HUITTEN. AND B. J. MCCARTHY

800 -

—^ 600 -

c

o o

(M

ro

400 -

200

(b)	Chlora	mphenicol
■~	200	ng/ml
M 11 1 1 1 i 1 1		A"
- 1 1 !		ft / ') 1 1/
1 _ J / V V	/ i ^ 1 / y 1	1 1

- 4000

- 3000

- 2000 ^

- 1000

c

o o

o

20 30

40

Fraction number

Fig. 26. The effect of chloramphenicol on the incorporation of C"-uracil into ribosomes. Cells were grown for three generations in P^" and given a 10-minute exposure to C"-uracil. (a) Control, (h) in the presence of chloramphenicol (200 fig/ml added 2 minutes before the uracil). Cells washed and extracts prejiared in tris-HCl 0.01 M. pH 7.4, MgCL 10' i1/ in the iircsence of DNase. Ccntrifugation 160 minutes at 37,000 rpm, 4°C.

estimates made from the two leucine exi)eriinents suggest that the 43S has a protein to RNA ratio 14 to Ys t^^'^t of the 50S ribosome.

The final flow diagram, including both RNA and protein moieties of ribosomes, is shown in Fig. 27. The open and shaded areas are propor- tional to RNA and protein contents. The eosome is shown as pure RNA since there are no measurements of its protein content. The 30S neosome is shown with less than half of the ])rotein of the 30S ribosome by analogy to the 43S. This point is not really established because of the lack of resolution between objects in the 30S size range. Likewise, the 30S neosome which is precursor to the 43S is shown as a sej)arate object and this may well be an unnecessaiy complication. The 43S neosome is shown with only one-quarter of the protein of the 50S ribosome. This quantity is uncertain and it is not known what fraction of its protein enters directly in the formation of the 438 or by way of the 30S precursor to it. While these estimates of relative protein contents in neosome and ribosome are crude and preliminary, it is clear that con- version of neosome to ribosome involves only the addition of protein antl that the greater part of the ribosome protein is added in this step.

VII. SYNTHESIS OF RNA AND RIBOSOMES

335

The effect of chloramphenicol is very much what one would expect from this model. The predominant effect is the accumulation of what appears to be eosome material by both sedimentation and column analysis. Moreover, the base composition of this material is like that of ribosomes. Apparently chloramphenicol inhibits the addition of pro- tein to eosomcs. The minor peak appearing at about 24S may represent

Biosynthesis of E. Coli ribosomes

EOSOME

43S

"^

(Peak-l4S) ( Range =8S-20S)

RIBOSOME

i

|50S

70S

Isosi

l/

Areas approximate molecular weights I ~| = .55x 10^ M.W. units of RNA i = . 17x10^ M.W.unitsof protein

Fig. 27. The biosynthesis of ribosomes in E. coli. The open and shaded areas are proportional to weights of RNA and protein, respectively.

small quantities of neosome produced in the presence of chloramphenicol since even at 200 /xg/ml the inhibition of protein synthesis is not abso- lutely complete.

One feature of the general sequence of ribosome synthesis deserves further mention. The addition of RNA and protein occur in time sepa- rated stages. This is clear from the very observation of RNA rich intermediates.

VIII. Composition of Newly Formed RNA

A. AMBIGUITIES

If an end product is stable, then newly formed material should show the same com]iosition as does the total. If, however, two components of

336

K. U. ROBERTS, R. J. BRITTEN, AiNU B. J. MCCARTHY

different composition cxij^t and only one of tliem is stable the conipoisi- tion of newly foi-nicd inatei-ial will differ from the total. The newly formed material must be distinu;uislied by its content of a tracer. Thus the base ratios of newly formed KNA are usually estimated from the P-^- content of the different mich'otides obtained by hydrolysis of the RNA. This jirocedure, however, yields an "ap|)arcnt" compositit^n which may differ from the true composition of newly formed RXA.

If 5'-nucleotidcs are isolated and measured the apparent composition will be determined partly by the specific radioactivities of the different nucleotide precursors. Several authors have shown that considerable differences in the specific radioactivities of the precursors do exist for prolonged periods of time (Volkin and Astrachan, 1956; Yeas and

T 3

C\]

X

Q. O

E

=3

E

"x

o

E Ci O

Cytidylic Adenylic

Guonylic

3000

2000 ^

- 1000

in

"c

O U

rO Q.

Fraction number

Fig. 28. Separation of nucleotides released by alkaline hydrolysis of RNA by chromatography on Dowex-1. A small quantity of P"-labeled unknown material is mixed with a much larger quantity of RNA of known composition before hydrolysis. The composition of the unknown material can be calculated from the measured specific radioactivities of tho nucleotide peaks.

Vincent, 1960). When 2',3'-nucleotides are isolated, differences in the specific radioactivities of the precursors will be averaged out provided that the product nucleic acid has a random distribution of nearest neighbors. Accordingly, alkaline hydrolysis to yield 2',3'-nucleotides is customarily used.

The neighbor distribution in ribosomes is random (Bolton, 1960) but

VII. SYNTHESIS OF RNA AND RIBOSOMES 337

appreciable deviations from randomness occur in DNA (Josse et al., 1961). Alkaline hydrolysis should then release nucleotides which show an apparent composition equal to the composition of newly formed RNA if the RNA is similar in its distribution of neighbors to ribosomal RNA. If, however, the RNA is similar to DNA and contains a non-random distribution, the apparent composition will not be an accurate measure of the newly formed material.

At early times after the addition of P^- the radioactivity of the free nucleotide pool is much greater than that of the nucleic acid. Some of this radioactivity is likely to contaminate the nucleic acid. It is therefore necessary to use separation techniques capable of separating the 2',3'-nucleotides released by alkaline hydrolysis from contaminating 5' nucleotides or serious errors may result. Also, to minimize errors from losses or conversions (by deamination) during hydrolysis it is desirable to measure the radioactivity by isotope dilution (Fig. 28).

B. TIME COURSE OF COMPOSITION

Differences in the compositions of newly formed material have been observed in several situations. Volkin and Astrachan found the RNA formed after infection by T2 was intermediate in composition between the RNA of the bacteria and the DNA of the phage (Volkin and Astrachan, 1956). The formation of RNA approaching the composition of the DNA was observed in several bacteria (of quite different DNA composition) following a transition from broth to sjmthetic medium (Hayashi and Spiegelman, 1961).

The same effect also appears in the absence of intentional transients in the growth conditions. Yeas and Vincent observed the apparent com- position of the RNA formed by yeast during 5-20 minute periods of exposure to P'*'- (Yeas and Vincent, 1960). Although the data scatter widely, the apparent composition seems to be intermediate between the RNA of the cell and its DNA. Similar results were obtained in a variety of bacteria (Astrachan and Fisher, 1961).

An extensive series of measurements has been carried out by ]\Iidgley using several different organisms (jMidgley, 1962) (Table VI). The accuracy of these measurements, carried out by isotope dilution, is sufficient to allow an estimate of the time constant of the transition. Correcting for the delays introduced by the nucleotide pool, the time constant is less than 3 minutes, corresponding roughly to that of the eosome stage of ribosome synthesis. Also the composition of the newly formed material (extrapolated to zero time) is not like that of the DNA but is midway between the DNA and the RNA of the cell.

338 R. B. ROBERTS, R. J. BRITTEN, AND B. J. MCCARTHY

TABLE VI

CllANCiK IN AlTAUKNT CoMTOSITIUN OF UNA DePE.NOING

UN Pkkiou of KxrosiRK to V^'K)^

Time (min)	A	C		G	U
		Escherich	ia coli
0.5	25.3	23.2		29.8	21,7
1	24.8	23.4		30.1	21.7
2	25.0	22.9		29.5	22.6
5	2G.0	22.4		29.9	21,7
10	25.1	22.0		30.9	22,0
20	25.6	20.9		32.5	21.0
30	25.8	21.4		33.3	19.5
00	25.2	22.1		33.5	19.2
		Bacillus subtilis
0.5	25.6	23.5		27.9	23 0
1	25.5	22.8		27.2	24.5
2	25.6	23.3		27.7	23,4
4	26.2	23.7		28.0	22.1
8	26.2	22.5		29.3	22.0
14	25.5	22.1		31.5	20.9
28	25.6	22,0		31.7	20.7
00	25.5	22.1		31.4	21.0

C. FRACTIONATION

At early times after the addition of the tracer the bulk of the P^- is still in the eosome fraction. Thus the composition observed at early times should be a measure of the composition of eosomes. Ribosomes free of S-RNA can be obtained by several cycles of centrifugation in tris buffer containing 10"- M Mg. The entire eosome fraction can be released from the ribosomes by reducing the magnesium concentration to 10 ■* M.

TABLE VII Fractionation of Newly Formed RNA"

Released by Released by Not released

10-^ Mg H2O by H2O

C23,2 25.0 215

A 25.0 24.4 25.3

G29.8 27.0 32.8

U21.8 23.6 20.4

» A pellet of 70S ribosomes was prepared from cells given 2 minutes' exposure to P32O4 . The pellet was then suspended in buffer with a magnesium concentration of lO"** M or in water with magnesium concentration of 3 X 10^' -U and centrifugod 40K 240 minutes to remove ribosomes and adhering RNA.

VII. SYNTHESIS OF RNA AND RIBOSOMES

339

	r	(o) 70S Preporotion
20	-	i
	.	i 1 r"^ ^
10	_£j-l	n	i i 1 1 1 1 1	L V 1	I.

20 -

10 -

(C)

IO"'*M Mg SN

i>W-^^^

20

Froction number

30

Fig. 29. Purification of the eosome fraction. 70S ribosomes were prepared liy centrifugation from E. coli given 2 minutes' exposure to P^'Oi . A portion was analyzed on DEAE (a). Another portion was suspended in buffer with a magnesium concentration of 10"' M and centrifuged 40K 240 minutes. The pellet (6) and supernatant fluid (c) were then analyzed on DEAE. Note that the bulk of the material is sedimented but the bulk of the ?'" remains in the supernatant fluid.

840

R. B. ROBERTS, R. J. BRITTEN, AND B. J. MCCARTHY

Further centrifugation sediments tlie ribosomes leaving the eosomes in the supernatant fluid. Chromatography on DEAE provides a final puri- fication of the eosome fraction (Fig. 29). The composition of the material obtained by this procedure is intermediate between DNA and RNA as was expected.

If, however, the ribosome pellet is suspended in water containing a slightly higher concentration of Mg only a portion of the eosome fraction is released. Centrifugation then separates the free eosomes from those still adhering to ribosomes. The compositions of these two fractions differ; the free eosomes are more like DNA and those still adhering to the ribosomes are more like the final RNA product (Table VII) (Midgley, 1961).

D. RNA CLASSES

The compositions of the 308 and 50S ribosomes and the apparent composition of the newly formed RNA show some systematic relation- ships. These can be interpreted as indicating that the different types of RNA are the result of different proportions of two hypothetical RNA's

Fig. 30. Relationship between composition of DNA and various RNA fractions. The composition of DNA is plotted at the loft (D). The compositions of early labeled RNA (E), 308 ribosomal RNA (30), and 50S ribosomal RNA (50) are plotted at the positions expected from the hypothesis described in the text. At the right the lines extrapolate to the composition of a hypothetical purinc-rich RNA (R).

VII. SYNTHESIS OF RNA AND RIBOSOMES 341

which we designate D-RNA and R-RNA to avoid any implication as to their functions. D-RNA indicates an RNA having the same base com- position as the DNA and R-RNA indicates a purine rich RNA. For example, the newly formed RNA might be a mixture of equal propor- tions of D-RNA and R-RNA. The SOS ribosomes having a slightly greater purine excess can then be assumed to contain y^ D-RNA and % R-RNA. The purine excess of the 50S particle is roughly twice that of the SOS ribosome but the kinetics of synthesis indicate that the SOS neosome may be a precursor to the oOS ribosomes. Thus the oOS ribosome may be assumed to contain % D-RNA and % R-RNA.

Figure SO shows that (with the exception of Proteus vulgaris) the data are consistent with this hypothesis. The accuracy is not sufficient to rule out other possibilities (e.g., that the 50S contains no D-RNA), but this particular one is the most attractive. The reality of the two types is supported by the experiments showing the fractionation of eosomes. There is, however, no direct evidence that the SOS ribosomes contain a D-RNA component. If this hypothesis happens to be correct, then D-RNA and R-RNA would be formed initially in equal quantities but the requirement for ribosome synthesis would be 2 parts of D-RNA for 7 parts R-RNA.

E. MECHANISMS FOR THE CHANGE IN APPARENT COMPOSITION

Apparently, more D-RNA is synthesized that could possibly be needed for the production of ribosomes. This interpretation is commonly used as evidence that the D-RNA is subsequently degraded. While this hypothesis may be correct, it poses several problems and it should not be accepted as final until several other alteiTiative interpretations have been excluded.

The degradation products must be such that they are rapidly re- utilized without mixing with the large pool of mononucleotides of the cell. If this were not the case, TCA precipitable material would be con- verted to TCA soluble material when a short period of uracil incorpora- tion is terminated. The expected degradation products (from ribonuclease action) would be 2',S'-nucleotides which would require phosphorylation to 5'-nucleoside triphosphates before reincorporation. Such a process would be likely to cause mixing with the TCA soluble pool. Perhaps the degradation products are larger and are reutilized without breakdown to mononucleotides by a presently unknown process. Such products could serve as precursors for soluble RNA if the delay of entiy into S-RNA is to be attributed to the pool of D-RNA.

As these difficulties are not trivial, it is worthwhile examining some of the alternative interpretations. The concept of degradation and

ivfi H. H. ItoHKHTS, H. .1. HHITTEN, AND D. J. MCCAHTHV

rrutilization ivciuiri's tlu' inti'oductioii of the cuiicci)t of a "i)ri\'atc' pool," i.e.. a pool of nucleic acid prt'cuisoi> t^cparate from the usually observed pool of TCA !>olul)]c material. Such a "private pool" could in itself be the cause of the appaicnl change in comi)osition. If exogenous P'-O^ had a pathway giving direct entiy into D-RNA while entry into S-RNA and R-RNA was delayed by a precursor pool, the newly formed RNA would show the composition of D-RNA.

As a third alternative D-RNA could be converted in part to R-RNA by the addition of purines without extensive degradation. Such a process would also result in a change with time of the apparent composition.

At pi'esent, no certain choice can be made among these and other possible alternatives. None are completely satisfying and none are com- pletely excluded by the experimental data. The only certainty is that the reactions are complex and elusive.

IX. Discussion

A. ROLE OF RIBOSOMES IN INFORMATION TRANSFER

The preceding sections have presented facts concerning the ril)Osoraes, and tlie flow of material through three stages of ribosome synthesis. These facts provide no direct indications of the part ribosomes may play in directing the formation of i)rotein in accord with the information carried by DNA. However, they provide necessary conditions which must be met by any theory of intermediate processes in information transfer.

At the present time it is easier to provide a reasonable interpretation of a complicated set of biological facts than to demonstrate that one particular interpretation does in fact correspond to nature. For example, it is currently accepted that ribosomes provide the active sites and tem- plates for protein synthesis. The basis for this belief is (1) that the rate of protein synthesis is proportional to the ril)Osome content, (3) that ribosomes are essential for protein synthesis in cell free systems, and (3) that newly formed protein is associated with ribosomes.

However, one who had reason to believe that ribosomes were not the protein synthesizing sites might interpret the same facts as follows. (1) That a high rate of protein synthesis is required for the formation and accumulation of ribosomes, (S) that an enzyme required for protein syn- thesis is firmly bound to ribosomes, and (5) that newly formed protein is readily adsorbed by ribosomes.

The correlation of facts relating to the flow of material with another set of facts relating to the flow of information is extremely difficult. In most experiments only the average flow of material into all proteins

VII. SYNTHESIS OF RNA AND RIBOSOMES 343

or all rihosoines can be measured. Furthermore, there may be no im- mediate alteration in the flow if some of the products are faulty and nonfunctional. In contrast, the experiments dealing with information usually measure one particular function such as an enzymic activity and the rate of synthesis of the particular functional enzyme may differ markedly from the average rate of protein synthesis.

B. CORREL.\TION WITH ENZYME INDUCTION

Upon addition of the inducer the rate of synthesis of the enzyme /8-galactosidase increases 200-fold. The time constant of the induction process is as low as 2Y2 minutes when the inducer level is high. As the same time constant appears in the kinetics of "deinduction" when the inducer is diluted out it seems reasonable to attribute the kinetic char- acteristics to the formation and subsequent decay of unstable enzyme forming units (Pardee and Prestidge, 1961; Boezi and Cowie, 1961).

Attributing all of the increase in synthetic rate to an increase in the number of enzyme-forming units (EFU), and assuming that there is 1 EFU per cell before induction, the number of EFU's is about 200/cell in fully induced cells. This number is not unreasonable since ^-galactosi- dase constitutes 2% of the protein and its synthesis might be expected to involve 2% of the cells' 15,000 70S ribosomes. As the time constant for induction is 2Y2 minutes, the initial rate of synthesis must be 1.3/ sec/cell in order to build up a complement of 200/cell in this time. Since the rate of 70S ribosome synthesis is 3/sec/cell (15,000/cell, 0.02%/ sec) the cell could supply completely new ribosomes at the required rate but only by devoting nearly half its capacity for ribosome synthe- sis to that particular product. Furthermore, if the de novo synthesis of complete 70S particles were required, the kinetics of induction might be expected to show the somewhat longer delay characteristic of the appearance of new material in finished ribosomes.

On the other hand, one RNA unit of the eosome fraction might be sufficient to act as template. The minimum number of eosomes synthe- sized is 36/sec/cell (MW = about 0.15 X 10", degradation ignored) and the number needed for /?-galactosidase EFU's would be 3.6% of the total. Furthermore, the kinetics of formation of ^-galactosidase EFU's correspond to the kinetic of eosome formation.

The average molecular weight considered is of course too small to specify a protein molecule of 100,000 MW. Perhaps the RNA used as template for yS-galactosidase is considerably larger than the average, or perhaps only a part of the enzyme molecule is dependent upon induc- tion for its synthesis.

In any event the part of the DNA associated with /3-galactosidase

:M } U. H. ROHKRTS. R. J. KHITTKN, AND U. J. MCCARTHY

imist !)(,' liijj;lily active coinpaivd to the average DNA. The entire lac region only comprises a tiny part of the chromosome, yet its product must account for at least 3.6% of the gene-directed material.

This example shows the present difficulties in correlating the facts of ribosome synthesis with their probable role as enzyme-forming units. Only relative enzyme-forming capacities are measured and several wild assumptions (which can only be justified as the best guess at present) are needed to estimate the absolute number of EFIJ per cell.

C. MESSENGER THEORY

At present the most precisely foi'inulated theory of the role of ribo- somes in information transfer is the messenger hypothesis (Jacob and Monod, 1961). The kinetics of enzyme induction and deinduction in- dicate that enzyme-forming units are unstal)l(> and that template mate- rial must be continuously supplied by the gene. However, the ribo- somes which appear to be the sites of protein synthesis are stable. Ac- cordingly, Jacob and Monod postulate that an unstable messenger RNA is produced by the gene and associates temporarily with ribosomes to act as template for protein synthesis.

They predict that cells should therefore contain an RNA fraction having the following properties: (a) average molecular weight 5 X lO"' or more, (b) base composition reflecting the base composition of DNA, (c) capable of associating with ribosomes, and (d) having a high rate of turnover.

As described in Sections V and VH, material having these character- istics has been found both in experiments designed to identify "mes- senger" RNA (Brenner et oL, 1961 ; Gros et al, 1961) and in experiment^ designed to observe ribosome synthesis (McCarthy et ai, 1962). De- pending upon the objective of the experiment the same material is desig- nated messenger RNA or the eosome fraction. The quantity of this material is between 1 and 3% of the ribosomal RNA. If its molecular weight is 5 X lO-' (V, the molecular weight of the RNA of the 70S!) it would be sufficient to provide 3-9% of the 70S ribosomes with tem- plates; only a small fraction of the ribosomes could be active in protein synthesis at any one time. There remains some question whether the "turnover" is due to degradation to low molecular weight material or to incorporation into ribosomes.

The experimental evidence considered in formulating the messenger theory was heavily weighted in favor of the rapid transients observed during enzyme induction or phage infection. Quite a different predic- tion of the properties of temjilate material could be drawn. Since the amino acid composition of a series of organisms varies much less than

VII. SYNTHESIS OF RNA AND RIBOSOMES 345

does the composition of their DNA's, the composition of the template RNA might be expected to be relatively constant and, therefore, in some cells at least, unlike the DNA. Furthermore, different parts of the DNA may transmit information (i.e., direct the synthesis of RNA) at different rates. The activity of the gene determining the structure of an inducible enzyme appears to be markedly increased upon induction. In this event the composition of DNA-directed RNA should reflect the composition of the DNA weighted in proportion to the activity of the different parts.

Alternatively the whole DNA complement might be replicated as RNA, but subsequently those regions not suitable for protein forming templates would be degraded. In some instances, such as hemoglobin synthesis, the enzyme-forming units are quite stable. Templates might then be expected to show a range of stabilities. Finally, for economical operation of the cell, there should be a sufficient quantity of template material to allow all ribosomes to be active. Thus, as an alternative to the messenger theoiy, template material might be predicted to have the following properties: (a) a constant composition from cell to cell re- flecting the average amino acid composition of the proteins, (b) having a range of lifetimes depending on the stability of the particular enzyme forming unit, and (c) a transient or permanent association with the ribosomes. Material of these characteristics is well known. Ribosomal RNA from bacteria has a relatively constant composition. A wide range of stabilities of template materials can be envisaged if some templates can function only during the eosome stage while others can operate after incorporation into ribosomes.

The value of this model is that it demonstrates clearly that the template material has not yet been identified. The question is still open and cannot be resolved by kinetic studies. At best, the kinetic and com- positional studies identify material having the properties postulated for "the messenger." Even hybrid formation is not diagnostic of template material as it merely indicates a specific relationship to the DNA. Quite a different approach is needed, such as an assay system to detect tem- plate RNA by its effectiveness in synthesizing a definite protein. Niren- berg and Matthaei have made a spectacular beginning in this direction. In their cell-free system polyuridylic acid senses as template for poly- phenylalanine and polycytidylic acid for polyproline (Nirenberg and Matthaei, 1961).

D. HOMOGENEITY OF RIBOSOMES

According to the messenger theory ribosomes could be completely homogeneous except for the attached template component. On the other

346 H. H. HOHKRTS, H. .1. HHirilN, AM) H. ,1. M (■CAHIII Y

hand, tlu'if iiiiiiht he a specific type of rihusunic iov each pi'otciii ol llic cell. There are a few indications of sonic heterogeneity in the ribosomes. The nunil)cr of diffcMcnt protein coniponents .suggests that lihosonies of different prott'in coinpleinent exist (Section II,C). If the rihosonie.s contain a component approximating DNA in conii)osition, many (Hfferent types of ribosonie aiv needed to contain e\-en one coniph'te copy of the cell's DNA.

More direct evidence of heterogeneity comes from chiomatography of ribosomes. Samples taken from different parts of the elution pattern have different chromatographic jiroperties. These differences maj'' be due to the protein i)ortion of the particles as thei'e is little spread in the elution pattern of the RNA obtained from ribosomes. Slight differences between the RNA's from SOS and 50S ribosomes have been observed both by chromatography and density separations. These differences indicate that both techni(iues are sensitive to the composition of the RNA and that RNA's of widely different compositions would be detected easily if they had been ])resent.

The most promising techniciue of isolating a si^ecific class of ribo- somes is precipitation by an antibody. At present the only analysis of such RNA shows a marginal difference from the average ribosomal RNA.

Templates for different proteins would not be expected to vary appreciably in their base compositions as the amino acid distribution is relatively constant. An RNA forming the structure of the ribosomes might also be expected to be constant. Thus any separation of ribosomes based on differences in their RNA composition may be difficult to achieve and the heterogeneity of the RNA of ribosomes separated by other techniques (such as antibody precipitation) may be difficult to demonstrate.

E. SITE OF RIBOSOME SYNTHESIS

Mammalian cells are large enough for radioautographs to show that newly .synthesized RNA is localized in the nucleus. Bacteria are too small to allow such direct methods, but Caro and Forro have analyzed the distribution of radioactivity in slices of bactei'ia. These studies have shown that newly synthesized RNA, unlike the bulk of the RNA, is localized in a central region of the cell (Caro and Forro, 1961). At cor- responding times the radioactivity is found in the eosome fraction (see Section I V,B) .

The composition of the newly formed RNA suggests that a consider- able fraction has a composition like that of the cell's DNA. An enzyme has been isolated which in the i)resence of a DNA primer catalyzes the

VII. SYNTHESIS OF RNA AND RIBOSOMES 347

polymerization of ribonucleotides into a replica of the DNA (Hurwitz et al., 1960; Weiss and Nakamoto, 1961). Finally, the specific radioac- tivity of the RNA found in the first jx^llct (which contains DNA as well as membranes, walls, and unbroken cells) is often unusually high. Thus there are numerous indications that some of the newly formed RNA is synthesized in association with and under the direction of DNA.

F. DNA-RNA HYBRIDS

Another indication of the direct participation of DNA in the syn- thesis of RNA is provided by the DNA-RNA hybrids (Hall and Spiegel- man; Spiegelman et al., 1961). These complexes form when RNA and the corresponding DNA are heated and slowly cooled. No hybrids are formed unless homologous DNA and RNA are used, demonstrating that complementarity in sequence and not just similarity in composition is required. Such a condition does not seem possible unless the RNA is formed as a direct copy of the DNA.

G. SELF-REPLIC.\TION OF RIBOSOMES

By far the greater part of the ribosomal RNA has no obvious relation- ship in its composition to that of the DNA. The composition of the ribosomal RNA is relatively constant in strains showing a wide variation in the GC/AT ratio of DNA (see Section II, C). It is difficult to visualize how RNA originally DNA-like in composition could be so altered to produce a constant purine-rich composition. It is possible, however, that a limited region of the chromosome, which is similar in many bacteria is particularly active in directing the synthesis of RNA. Such RNA might also be stabilized by the prompt addition of basic proteins. If this were the case a single process could be responsible for the synthesis of both D-RNA and R-RNA. Both types would be copies of DNA and the differences would be attributed to rates of synthesis and stability.

The alternative view, that the bulk of the ribosomal RNA is formed by a different process, seems less attractive at present. An enzyme has been found which carries out the appropriate reaction, the polymeriza- tion of ribonucleotides under the direction of an RNA primer (Reddi, 1961).

When cells recover from a period of magnesium depletion, during which they lose their complement of ribosomes, the ribosomes increase exponentially. This behavior might be taken to mean that, in the absence of other limiting factors, ribosome synthesis is autocatalytic. There are, however, alternative interpretations.

At present, the self-replication of ribosomes remains a possibility. If, however, it can be established that all ribosomes contain a component

348 R. B. ROBERTS, H. .1. HHITTP:N. and B. J. MCCARTHY

whose synthesis is directed by DNA, then self-replication would be ruled out.

H. CELL-FREE SYNTHESIS OF RIBOSOMES

Several of the cell-free systems for protein synthesis have been reported to show a sensitivity to DNase and to require ribonucleoside triphosi)hates for optimal synthesis (Kameyama and Novelli, 1960; Matthaei and Nirenberg, 1961; Tissieres et al., 1960). Presumably in all these systems there is a DNA-dei)endent synthesis of RNA which allows a more prolonged period of protein synthesis.

The incorporation of labeled nucleotides into RNA has been observed both in purified enzyme preparations and in some of the cell-free systems capable of protein synthesis. The purified enzyme preparations yield RNA of random sequence (Grunberg-Manago et al., 1955), copies of a DNA primer (Hurwitz et al, 1960; Weiss and Nakamoto, 1961), or copies of an RNA primer (Reddi, 1961), depending on which enzyme is present. As these preparations are not designed to allow protein synthesis, no ribosomes are formed.

In most of the experiments using cell-free systems capable of protein synthesis the RNA concurrently synthesized has not been characterized. In one experiment P^--labeled nucleotides were added to the cell free system of Nirenberg and Matthaei. The products (analyzed by chroma- tography on DEAE) showed RNA eluting at 0.5 M NaCl and at 0.6 M NaCl. These concentrations elute S-RNA and eosomes, respectively (see Section III,A). The peak at 0.6 M NaCl, but not the peak at 0.5 M NaCl, was absent when the system had been treated with DNase. As free RNA elutes at a much higher NaCl concentration, this system seems to be capable of the synthesis of the first ribosome precursor. The composition of the eosome-like RNA showed the typical 50-50 proportions of D-RNA and R-RNA. Thus both D-RNA and R-RNA, but not S-RNA, seem to be synthesized only when DNA is present.

Since these systems have been optimized on the basis of their capacity to synthesize protein it seems likely that ways to improve their capacity for ribosome synthesis could be found. Synthesis of complete ribosomes in cell free systems is a likely development in the future.

I. STATE OF RIBOSOMES IN LIVING CELLS

It is by no means established that the ribosomes observed in the analytical centrifuge or the electron microscope exist as definite entities of the same form in the living cell. For several years indications have accumulated that ribosomes undergo certain irreversible changes at the instant of breaking the cells.

VII. SYNTHESIS OF RNA AND RIBOSOMES 349

Whether or not ribosomes appear in electron micrographs of cellular sections depend strongly on the methods of preparation. The pattern observed in the analytical centrifuge is quite different when the cells were broken by lysozyme-freeze-thaw and then put through the pressure cell from the pattern obtained when the processes were interchanged. The quantity of /?-galactosidase associated with the ribosomes can vary 10- fold, depending on the method of breaking. The ribosomes obtained by breaking the cells in the presence of 10'- M Mg and dialyzing to 10"* M are not identical to ribosomes obtained by breaking in lO'Mf Mg. The protein content of 308 ribosomes from cells broken in low magnesium concentrations is unexpectedly high. Some find ribonuclease equally distributed between 30S and 50S particles, but others find it only in the 30S particles.

Perhaps all of these anomalies indicate that the ribosome as it exists in the cell has quite a different form and the ribosome as observed in the broken cell juice depends both on the metabolic state of the living cell and on the particular conditions at the time of breakage.

Inforaiation can be derived from studies of the particles even though they may be formed at the instant of breakage as they cany reproducible samples of cellular components. In other respects there may be important differences. The low rates of protein synthesis and the difficulties of removing nascent protein from the ribosomes in vitro may be due to a less favorable configuration of the particles in vitro than in vivo. Models of cellular function and structure should not be limited to those con- ceived in terms of the little round balls observed in the electron micro- scope. Extended intertwining strands might equally well, or better, represent the organization of ribosomes in the living cell. If this were the case, then at the moment of breakage the strands might gather into their coils quite different proportions of cellular material depending on the conditions at that particular instant.

The freedom given by this point of view is particularly necessaiy in trying to picture how template RNA might be associated with a pre- existing ribosome. The template must contain 600 nucleotides at least if it is to carry information for a polypeptide strand of 200 amino acids. Thus, it is long enough to wrap three times around the 70S particle seen in the electron microscope. It is, however, quite compatible with an extended ribosome.

The extended picture is also more helpful in visualizing ribosome synthesis. The experiments of Caro indicate that newly formed RNA shows the same localization in the cell as does DNA. Thus, eosomes may initially be extended along DNA strands. The hybrid RNA-DNA complexes found by Hall and Spiegelman indicate a considerable binding

350 H. H. R()Bp:rts, r. .t. hhittkx, and b. j. mccarthv

force may exi.<t. Soon thereafter the cosomes are (lish)(l<2;e(l. j)erhaj)s l)y accumulation of piotein, and can he t'ound associaterl with ribosomes. Such a transfer would seem moi'e phiusihle if tlic recijiient I'ibosome were equally extended.

At present there are no certain facts relating to the state of organiza- tion of nucleic acids in the living bacterial cell. A model showing the DNA, the eosomes, neosomes, and I'ibosomes extended as roughly parallel strands seems equally valid or e\-en jireferablc to a model in which the DNA is packed in a nucleus surrounded by a multitude of small round ribosomes.

X. Epilogue

In this review of ribosome synthesis we have attempted to fractionate the material and to present the facts in Sections I-VIII partially sepa- rated from the interpretations of Section IX. In so doing we have tried to show that many of the currently popular hypotheses are quite ac- ceptable and not in conflict Avith the facts. They should not, however, be considered as proved until other alternative interpretations have been ruled out. In addition, we have attempted to point out the areas where additional experiments are most needed. We hope that we may have encouraged some of our readers to measure the number of enzyme- fonning units, to devise an assay for template RNA, or to carry out a cell-free synthesis of ribosomes.

References

Albertsson, P.-A. (1960). "Partition of Coll Particle.s and Macromolecules." Wiley,

New York. Aronson, A. I. (1961). Biochim. el Biophys. Acta 49, 98. Aronson, A. I., and McCarthy, B. J. (1961). Biophys. J. 1, 215. Astrachan, L., and Fisher, T. N. (1961). Federation Proc. 20, 359. Boczi, J., and Cowie, D. B. (1961). Biophys. J. 1, 639. Bolton, E. T. (1958). Carnegie Inst, of Wash. Yearbook 57, 135. Bolton, E. T. (1959). Carnegie Inst, of Wash. Yearbook 58, 276. Bolton, E. T. (1960). Carnegie Inst, of Wash. Yearbook 59, 260.

Bolton, E. T.. and McCarthy, B. J. (1959). Carnegie lust. Wash. Yearbook 58, 278. Bolton, E. T., and Robert.s. R. B. (1956). Carnegie Inst. TTV/.s/j. Yearbook 55, 122. Bolton, E. T., Hoycr, B. H., and Hitter, D. B. (19.58). In "Micro.somal Particle.s and

Protein Synthe.sis" (R. B. Roberts, ed.), p. IS. Pergamon Press, New York. Borsook, H., Deasy, C. L., Haagen-Smit, A. J., Keighly, G., and Lowy, P. H. (1950).

J. Biol. Chem. 187, 839. Brachet, J. (1950). "Chemical Embryology." Interscience, New York. Brakke, M. K. (1953). Arch. Biochem. Biophys. 45, 275. Brenner, S.. Jacob, F., and Me.selson, M. (1961). Nature 190, 576. Britten, R. J., and McCarthy, B. J. (1959). Carnegie Inst. Wash. Yearbook 58, 267. Britten, R. J., and McCarthy, B. J. (1962). Biophys. J. 2, 49.

VII. SYNTHESIS OF RNA AND RIBOSOMES 351

Britten, R. J., and Roberts, R. B. (1960). Scieyice 131, 32.

Britten, R. J., Robert.s, R. B., ami French, E. F. (1955). Pruc. Xnll. Acad. Sci. US.

41, 863. Britten, R. J.. Larlv, K. CI., and Nort'ros.s, F. (1960). Biocltint. el Biophy.s. Acta

39, 150. Britten, R. J., McCartliy, B. J., and Roberts, R. B. (1962). Biaphys. J. 2, 83. Bucliwald, M., and Britten, R. J. (1962). Biuplnjs. ./.. in press. Caldwell, P. C, and Hin.shehvood, C. N. (1950). J. Chem. Soc. p. 1415. Care, L. G., and Forro, F., Jr. (1961). Biuphyx. Biochem. Cytol. 9, 555. Cassperson, T. (1950). "Cell Growth and Function." Norton, Xew York. Chao, F. C. (1957). Arch. Biochem. Biophys. 70, 426.

Chao, F. C, and Schachman, H. K. (1956). Arch. Biochem. Biophys. 61, 220. Claude, A. (1943). Science 97, 451. Cowie, D. B., Spiegelman, S., Roberts, R. B.. and Duerksen, J. D. (1961). Proc. Natl.

Acad. Sci. U.S. 47, 114. Dintzis, H. M., Borsook, H., and Vinograd, J. (1958). //( "Microsomal Particles and

Protein Synthesis" (R. B. Roberts, ed.), p. 95. Pergamon Press, New York. Edelman, I. S., Ts'O, P. 0. P., and Vinograd. J. (1960). Biochim. et Biophys. Acta

43, 393. Elson, D. (1958). Biochim. et Biophya. Acta 27, 216. Elson, D. (1959). Biochim. et Biophys. Acta 36, 372. Elson, D. (1961). Biochim. et Biophys. Acta 53, 232. Elson, D., and Tal, M. (1959). Biochim.. et Biophys. Acta 36, 281. Green, M. H., and Hall, B. D. (1961). Biophys. J. 1, 517. Gros, F., Hiatt, H., Gilbert, W., Kurland, C. A., Risebrougli, R. W.. and Watson,

J. D. (1961). Nature 190, 581. Grunberg-Manago, M., Ortiz, P. J., and Ochoa, S. (1955). Science 122, 907. Hall, B. D., and Spiegelman, S. (1961). Proc. Natl. Acad. Sci. U.S. 47, 137. Hall. C. E., and Slayter, H. S. (1959). J. Mol. Biol. 1, 329. Hanzon, V., Hermod.s.son, L. M., and Toschi, G. (1959). J. Ultrastructiire Research

3, 216. Hayashi, M., and Spiegelman. S. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1564. Hurwitz, J., Bresler, A., and Dinger, R. (1960). Biochem. Biophys. Research Com-

muns. 3, 15. Huxley, H. E., and Zubay, G. (1960). /. Mol. Biol. 2, 10. Jacob, F., and Monod, J. (1961). J. Mol. Biol. 3, 318. Josse, J., Kai.ser. A. D.. and Kornberg, A. (1961). J. Biol. Chem. 236, 864. Kameyama, T., and Novelli. G. D. (1960). Biochem. Biophys. Research Comyyiuns.

2, 393. Kurland, C. G. (1960). J. Mol. Biol. 2, 83. Littlefield, J. W., Keller, E. B., Gross, J., and Zemecnik. P. C. (19^5). ./. Biol. Chem.

217, HI. Littlefield, J. W., and Keller, E. B. (1957). /. Biol. Chem. 224, 13. Luria, S. E., Delbriick, M., and Anderson, T. F. (1943). J. Bacterial. 46, 57. McCarthy, B. J. (1959). Carnegie Inst. Wash. Yearbook 58, 281. McCarthy, B. J. (1960). Biochim. et Biophys. Acta 39, 563. McCarthy, B. J. (1962). Biochim. et Biophys. Acta 55, 880. McCarthy, B. J., and Aronson, A. I. (1961). Biophys. J. 1, 227. McCarthy, B. J., and Britten, R. J. (1962). Biophys. J. 2, 35. McCarthy, B. J., Britten, R. J., and Roberts, R. B. (1962). Biophys. J., in press.

352 H. H. houerts, r. j. hkittkn, and h. .i. mccarthy

McQuillen. K.. Rol.crts. K. H, and Ikitt.ii. H. J. (1950). Proc. Natl. Acad. Sci. UJi.

45, 1437. Maiuh'U, J. D.. aiul Ilprshey, A. D. (1960). Anal. Biochcm. 1, 66. iMatthaci, J. H., and Xirenbcrg, M. W. (1961). Proc. Natl. Acad. Sci. UJi. 47, 1580. Midgley, J. E. (1961). Carnegie Inst. Wash. Yearbook 60, 305. Midgley, J. E. (1962). Biochim. et Biophys. Acta., in press.

Midgley, J. E., and Goldthwait, D. (1961). Carnegie Inst. Wash. Yearbook 60, 309. Nirenberg, M. W., and Matthaei, J. H. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1588. Nisman, B. (1961). Cold Spring Harbor Sytnposia Quant. Biol., in press. Nomura, M., and Watson, J. D. (1959). J. Mol. Biol. 1, 204. Xomura, M., Hall, B. D., and Spiegclman, S. (1960). J. Mol. Biol. 2, 306. Otaka, E., Osawa, S., Oota, Y., Ishihama, A., and Mit.sui, N. (1962). Biochim. el

Biophy.'i. Acta, in press. Palade, G. E. (1958). In "Microsomal Particles and Protein Synthesis" (R. B. Rob- erts, ed.). ]). 36. Pergamon Press, New York. Pardee, A. B., and Prestidge, L. S. (1961). Biochim. el Biophys. Ada 49, 77. Pardee, A. B., Paigen, K., and Prestidge, L. S. (1957). Biochim. et Biophys. Acta

23, 162. Peterman, M. L., and Hamilton, M. G. (1957). ./. Biol. Chem. 224, 725. Reddi, K. K. (1961). Science 133, 1367. Roberts, R. B. (1958). In "Microsomal Particles and Protein Synthesis" (R. B.

Roberts, ed.), p. viii. Pergamon Press, New York. Roberts, R. B. (1959). Carnegie Inst. Wash. Yearbook 58, 290. Roberts, R. B. (1960). Ann. N.Y. Acad. Sci. 88, 752.

Roberta, R. B., and Duerksen, J. D. (1960). Carnegie Inst. Wash. Yearbook 59, 231. Roberts, R. B., Britten, R. J., and Bolton, E. T. (1958). In "Microsomal Particles

and Protein Synthesis," (R. B. Roberts, ed.), p. 84. Pergamon Press, New York. Schachman, H. K., Pardee, A. B., and Stanier, R. Y. (1952). Arch. Biochem. Biophys.

38, 245. Schlessinger, D. (1960). /. Mol. Biol. 2, 92.

Spahr, P. F., and HoUingworth, B. R. (1961). J. Biol. Chem. 236, 823. Spahr, P. F., and Tissieres, A. (1959). /. Mol. Biol. 1, 237.

Spiegelman, S., Hall, B. D., and Storck, R. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1135. Timasheff, S. N., Witz, J., and Luzzati, V. (1961). Biophys. J. 1, 525. Tis.sieres, A., Watson, J. D., Schlessinger, D., and HoUingsworth, B. R. (1959). J.

Mol. Biol. 1, 221. Ti.ssieres, A., Schlessinger, D., and Gros, F. (1960). Proc. Natl. Acad. Sci. U.S. 46,

1450. Ts'O, P. 0. P., Bonner, J., and Vinograd, J. (1956). J. Biophys. Biochem. Cytol.

2, 451. Ts'O, P. O. P., Bonner, J., and Vinograd, J. (1958). Biochim. et Biophys. Acta 30,

570. Volkin, E., and Astrachan, L. (1956). Virology 2, 146. Waller, J.-P., and Harris, J. I. (1961). Proc. Natl. Acad. Sci. U.S. 47, 18. Warren, W., and Goldthwait, D. (1961). Federation Proc. 20, 144. Weiss, S. B., and Nakamoto, T. (1960). Proc. Natl. Acad. Sci. US. 46, 1020; (1961).

47, 694. Yeas, M., and Vincent, W. S. (1960). Proc. Natl. Acad. Sci. US. 46, 804. Zubay, G., and Wilkins. M. H. F. (1960). J. Mol. Biol. 2, 105.

Chapter VIII

Protein Synthesis in Relation to Gene Action^

RICHARD SCHWEET and JOHN BISHOP

Department of Biochemistry ,

University of Kentucky College oj Medicine,

Lexington, Kentucky

I. Introduction 353

II. Formation of Amino Acyl-RNA Compounds 355

A. Amino Acid-Activating Enzymes 355

B. Transfer RNA 356

C. Specificity of Amino Acyl-RNA Formation 358

III. Steps Requiring Participation of the Ribosome 362

A. Transfer of Amino Acids from Amino Acyl-RNA to Ribosomes . . 363

B. Formation of the Peptide Chain 373

IV. Information Transfer in Protein Synthesis 380

A. Information Transfer in Bacterial Protein Synthesis 381

B. Information Transfer in Nucleated Organisms 395

References 400

I. Introduction

The past ten years have seen remarkable advances in understanding the mechanism of synthesis of the major macromolecules of the cell. It appears likely that the major pathways for the synthesis of DNA, RNA, and protein are under study. Some biochemists may consider this view- point to be unduly optimistic. Certainly, a vast amount of uncertainty remains. Many new reactions and mechanisms will be discovered. These latter statements are perhaps more true for the field of protein synthesis than DNA and RNA synthesis, despite the great amount of research effort expended in this field. Even for proteins, the main outlines of a major synthetic pathway are apparent. It seems appropriate, therefore, to attempt to summarize the important steps which have laid the founda- tions for the present state of knowledge.

The comprehensive review by Borsook in 1952 summarized the avail- able information of that date. Among the important earlier advances recognized are the dynamic state of proteins and the energetic require-

^The studies reported here were supported bj^ Grant H-5293 from the United States Public Health Service and Grant G-11400 from the National Science Foundation.

353

354 HICIIAHI) SCIIWKKT AM) JOHN HISIIOP

iiit'iils for peptide IjoikI synthesis. 'I'lie use of ('"-laheled amino acids and in vitro (l)Ut not cell-free) systems is ivviewed and e\-idence that amino acid iiicorpoi'ation did repivsent pi'oteiii synthesis is gi\'en in detail. This prt)l)lem appeared ai^ain when cell-fi-ee systems were de\-eloped. A lai'^e part of the review is devoted to analogies between protein synthesis and the biosynthesis of glutaniine and glutathione or transpcptidation reac- tions, since these systems were well studied at the time. Tiic problem of exchange versus dc novo synthesis also was unresolved at that time as was the status of free jK'ptide intermediates. Both of these problems were raised in part by the "une(iual labeling" results of Anfinsen and Steinberg (19olK Some of these jioints have been clai'ifi('(l; others seem to have lost their significance for the time being.

Contributions which turned out to be significant for future de^"elo|)- ments in jirotein synthesis were rei^orted just about this time. In Borsook's laboratoiy, the importance of RNA in protein synthesis was re-emphasized by Holloway and Ripley (1952), and microsomes were sliown to be the most rapidly labeled cell fraction (Borsook cf al., 1950). Similar results were reported l)y Hultin (1950). Although a number of experiments on the incorporation of C^'*-amino acids into protein using houiogenates were reported earlier, the first experiment which clearly implicated microsomes was reported at this time. Siekevitz (1952), in Zam(>cnik's laboratory, reported that a mixture of mitochondria and microsomes from I'at li\-er incorporated C"-alanine at a rate ec^ual to that of the whole homogcnate. The role of mitochondria was recognized as providing energy, and the highest amount of incorporation was into the microsomal protein. These findings set the stage for the development of cell-free systems which provided most of the advances in the under- standing of protein synthesis in the next ten years.

An assessment of the most significant advances from 1952 through 1961 is more difficult, and reflects undoubtedly some bias on the part of the writers. However, since the details of tlicse studies and particularly the more recent ones will form the subject matter of this review, and some choices must be made to keej) the length within reasonable bounds, a sunniiaiy of these advances will be attempted. The work of Siekevitz (1952) led directly to the system of Zamecnik and Keller (1954), where microsomes plus supernatant fraction plus an ATP-generating system gave good amino acid incorporation. Studies of the function of the super- natant fraction began with the discoveiy of amino acid-activating enzymes by Hoagland (1955) and by Berg (1956). These enzymes could i)e assayed in a numb(>r of ways, but their physiological role appeared to involve the formation of amino acyl-RNA compounds (Hoagland ct (iL. 1957; Ogata and Nohara, 1957). The existence of amino acyl-RNA compounds was indicated by studies in other laboratories at this same

VIII. PROTEIN SYNTHESIS AND GENE ACTION 355

time (see review by Hoagland, 1960). However, the main line of investi- gation which hiter led to the conclusion that the amino acyl-RNA compounds were the important soluble intermediate in protein synthesis was initiated by these studies in Zamecnik's laboratory (Hoagland et al., 1957). From these results and those of others came the recognition of a particular type of RNA with a specific function, that is, transfer RNA. Studies of the transfer of amino acids from amino acyl-RNA to ribosomal protein have been reported from a number of laboratories in the past few years and are reviewed in detail below.

The development of two new systems for studying amino acid incor- poration led to some of the more recent advances. A cell-free system from rabbit reticulocytes (Schweet et al., 19o8b) provided a system where incorporation into a known, soluble protein (hemoglobin) was readily studied. Finally, the bacterial ribosomal preparation (Lamborg and Zamecnik, 1960) made available a system which permitted studies on information transfer in protein synthesis. These studies involving "messenger" RNA are reviewed in detail below.

A number of other pathways for protein synthesis have been reported and some of these are discussed below. However, in this review the trans- fer RNA to ribosome pathway is emphasized, largely because this is the only system where the intermediate stages have been defined and w^here the role of nucleic acids has been studied. Thus at least a framework, if not the details, of information transfer from gene to protein can be discussed.

II. Formation of Amino Acyl-RNA Compounds

This subject has been reviewed recently, either as part of general reviews on protein synthesis (Raacke, 1961; Berg, 1961; Campbell, 1961), or in reviews devoted to this topic specifically (Stulberg and Novelli, 1962; Hoagland, 1960). For this reason, only a brief summary of the earlier findings are presented here plus more detailed discussion of recent papers, particularly those relevant to specificity aspects of this reaction. -

A. AMINO ACID-ACTIVATING ENZYMES

These enzymes catalyze the following reactions:

1. Amino acid + ATP + enzyme ^ (amino acyl-Amp-E) + PP

2. (Amino acyl-AMP-E) + T-RNA ^ amino acyl-RNA + enzyme + AMP

3. SUM - amino acid + ATP + T-RNA ^ amino acyl-RNA + AMP + PP

" The abbreviations are : PP, inorganic pyropliospliate ; T-RNA, transfer RNA ; A, adenylic acid; C, cytidylic acid; U, uridylic acid; G, guanylic acid; DOC, deoxycholate; C-terminus, free carboxyl terminus; N-terminus, free amino terminus; TMV, tobacco mosaic virus.

356 RICHARD SCHWEET AND JOHN BISHOP

Reaction 1 can be measured by the amino aci(i-(le])en(lent incoij)ora- tiun of P^--pyrophosphate into ATP and by the formation of amino acyl liych'oxamates wlien hirge amounts of hydroxyhunine are present. The evidence for the formation of the enzyme-bound AMP-amino acid, in which tiie phosphate is linked in anhydride bond with the amino acid carboxyl, is quite conchisive. This intermediate has been isohited using substrate amounts of enzyme and its synthesis shown by 0^^ transfer experiments.

Reaction 2 has been demonstrated directly starting with tlie enzyme- AMP-amino acid complex, but is usually measured using the over-all reaction 3. The incorporation of a C^^-amino acid into T-RNA (transfer RNA, see below) is measured. This reaction is the physiological one catalyzed by these enzymes and they have recently properly been named amino acyl-RNA synthetases (Berg, 1961). The evidence that reactions 1 and 2 are catalyzed by the same enzyme is based largely on the failure to separate these activities on j^urification of the enzyme (Berg and Ofengand, 1958). Later studies are in agreement with these results, in- cluding those with an alanine enzyme where the activity for PP :?± ATP and AMP ^ ATP exchange were purified together (Holley and Gold- stein, 1959). The reversibility of reaction 3 has been demonstrated and the equilibrium constant determined from both directions. The values obtained for various amino acids indicate the high-energy nature of the amino acyl-RNA bond (Glassman et al., 1958; Lipmann et al., 1959; Preiss et al., 1959; Leahy et al., 1960). The specificity of amino acyl- RNA synthetases is discussed below. The review by Stulberg and Novelli (1962) contains further details of these reactions.

B. TRANSFER RNA

The RNA which is active for formation of amino acyl-RNA com- pounds has been called transfer RNA (T-RNA) based on its function in protein synthesis (Allen et al, 1960) or soluble RNA (S-RNA) based on its location in the cell (Hoagland, 1960). The term T-RNA has been used recently to distinguish this RNA from another type of RNA, also found in the soluble cytoplasm under some conditions, that is, messenger RNA. T-RNA is a specific type of RNA with a molecular weight of about 25,000. However, T-RNA does not appear to be either a precursor or degradation product of larger molecular weight RNA. This conclusion is based on the unique structure of T-RNA. In particular, T-RNA has known end groups as shown:

pGp RNA pCpCpA

VIII. PROTEIN SYNTHESIS AND GENE ACTION 357

The discovery of the pCpCpA end group was made by a number

of laboratories. Enzymes which remove and replace these three nucleo- tides in a very specific way have been studied (see Hoagland, 1960; Raacke, 1961). There is a rapid turnover of the end gioup of T-RNA, but the significance of this is not known. The finding of guanosine di- phosphate in alkaline hydrolyzates seems to have established this as the major terminus at the other end of the molecule (Singer and Cantoni, 1960; Zillig et al, 1960; Herbert and Canellakis, 1960). In addition, the base composition of T-RNA is unique in its content of "odd" bases. These odd bases, particularly pseudouridine, have been studied by several groups (see Hoagland, 1960 for earlier references). A very complete study of the composition of the soluble RNA of Escherichia coll has been reported by Dunn et al. (1960). They report the base composition to be (moles per 100 moles): A = 20.3, G = 32.1, C = 28.9, U = 15.0, pseudo-U = 2.1, thymine = 1.1, and 2-methyladen- ine = 0.3. Smaller amounts of other odd bases were also present. Ribo- somal RNA contains very little pseudouridine and also differs in having a lower C and higher U and A content. The soluble RNA from other species also contains high proportions of odd bases, but some differ- ences exist. More recently, Cantoni (1962) has reported similar results with rabbit liver soluble RNA, except that thymine nucleotides were not found, but methylamino C was present. Thymine may be an artifact derived from methyl C during hydrolysis (Dunn et al., 1960). The ability of RNA fractions to form amino acyl-RNA compounds was directly proportional to their pseudouridine content (Osawa, 1960). Further discussion of the relation of base composition to function will be given below in connection with the specificity of T-RNA.

The nature of the amino acid to RNA linkage has been studied. This linkage is an ester between the carboxyl group of the amino acid and the ribose hydroxyl of the terminal adenosine. It is not known whether the 2'- or 3'-hydroxyl of the ribose is the esterified group. This finding was first reported by Zachau et al. (1958) and was based on the isolation of C^'^-leucyl-adenosine after RNase digestion of C"-leucyl-RNA. This result was confinned and extended by Preiss et al. (1959) and Hecht et al. (1959). The essentiality of the ACC tenninus for amino acid attachment was shown by enzymatic studies and by periodate oxidation of the free cis-hydroxyls of the ribose of the terminal adenosine. If an amino acid was attached to the RNA, the ribose was protected from periodate action. This findings led to methods of separating amino acid- specific RNA fractions (see later).

358 RICHAHI) SCIIWKKT AM) JOIIX BISHOP

C SPI'^riFICITV OF AMINO ACVL-RNA FORMATION

Thr sp(M'ifi('ity of aiuino acid-activating enzyincs for individual aniiiio acids was demonstrated soon after tiie discovery of these enzymes. A number of enzymes which activate only a single amino acid have been isolated (reviewed by Raacke, 1961; Berg, 1961). Although analogs of the amino acid may be activated by a i)articular enzyme, in most cases only a single naturally occurring amino acid is activated by a particular enzyme. An excejition is the isoleucine enzyme studied by Bergmann ft (d. (19G1) which activated valine also when measured by PP-ATP exchange. However, even this enzyme formed only isoleucyl-RNA, a more significant assay for the physiological functioning of these enzymes (Berg et al., 1961). These results thus indicate a two-stage specificity in the action of activating enzymes. First, a specificity for the formation of the enzyme-amino acyl-adenylate complex; and second, a specificity of the complex in its reaction with T-RNA. This second specificity extends to the recognition of T-RNA molecules specific for individual amino acids.

Despite the uniformities in structure at the ends of T-RNA, evidence has accumulated that fractions specific for the attachment of individual amino acids are present. Evidence for T-RNA fractions specific for individual amino acids based on fractionation was first noted by Schweet et al. (1958a). Fractionation procedures using untreated T-RNA have resulted in partial separations of fractions specific for various amino acids. Chromatography on an anionic starch resin (Smith et al., 1959) , countercurrent distribution (Holley and Merrill, 1959), clectroi:)horesis (Lipmann et al., 1959), and chromatography on calcium phosphate columns of various types (Hartmann and Coy, 1961 ; Cantoni, 1962) have been used. None of these methods has so far resulted in isolation of a fraction specific for a single amino acid.'' Promising results have also been obtained by utilizing the change in jn'operties of a particular RNA by virtue of having its amino acid attached. Tyrosyl- and histidyl- RNA chains have been ]:)artially purified by complexing to i)olydiazo- styrene columns (Brown et al., 1959). What appears to be a nearly pure RNA fraction specific for valine has been reported by Stephenson and Zamecnik (1961). Yeast soluble RNA containing only attached valine was prepared. Aftci- oxidation by pcriodate of the RNA (the valyl-RNA was protected), and attachment of a bulky dye molecule, fractionation on DEAE-Sephadex yielded a fraction which in one experiment was

'Holley (private eommunicalion) has rri)orte(l tlie isolation of amino arid- specific T-RNA fractions using countercurrent distribution.

VIII. PROTEIN SYNTHESIS AND GENE ACTION 359

90% pure. It seems likely that individual fractions of T-RNA specific for single amino acids will eventually be isolated.

A different api)roach which has provided evidence for amino acid- specific RNA chains was first reported by Hecht et al. (1959) and Preiss et al. (1959). These authors showed that all T-RNA chains had the same terminal nucleotides (RNA-pCpCpA) and that amino acids were attached to the terminal adenosine. Since individual amino acids were bound to RNA only to a specific extent and there was no competi- tion for sites among various amino acids, amino acid-specific RNA molecules were assumed to be present (based on an unbranched RNA molecule). This approach was amplified by forming a single amino acyl-RNA (valyl-RNA) and then oxidizing with periodate. The amino acid should have protected only the specific RNA. This was tested (Preiss et al., 19591 by removing the bound amino acid with dilute alkali and showing that incorporation of only a single amino acid (C^*- valine) into the treated RNA was possible. A number of amino acids protected only their own sites against periodate oxidation.

The interest in separation of amino acid-specific T-RNA molecules stems from the hypothesis (Crick, 1958) that the nucleic acid of the ribosome contains a "code" which specifies the amino acid sequence of the protein being made. This code resides in the sequence of the RNA nucleotides. An amino acid, however, according to this idea, cannot be coded for directly by RNA. An "adaptor" molecule which can hydrogen bond with the ribosomal template coding unit is required to transport the amino acid to its proper position in the peptide chain. This adaptor is T-RNA. Each individual T-RNA must therefore carry the code or more likely the complement of the code for the single amino acid it binds. Further discussion of this concept is given below, but its conse- quences for T-RNA structure and attempts to study these experimentally are appropriate here.

One consequence of the adaptor hypothesis is that differences in amino acid-specific RNA molecules reside in their base sequence and not in secondary structure or molecular weight. This ajijiears to be correct, since T-RNA is fairly homogeneous in size, and while it has a consid- erable amount of secondary structure, evidence that secondary structure is not required for activity has been reported (see Berg, 1961, for further discussion). Several studies have now appeared which demonstrate considerable variation in the base composition of different T-RNA fractions. An ingenious combination of labeling techniques, enzymatic digestion, and isolation of labeled fragments has been used to study the nucleotide sequences near the adenosine end of T-RNA without prior

8(iO HICHARI) SCinVEKT AND JOHN BISIIOI'

separation (Berg ami Lajjjerkvist, 1962). The results shown in Table I indicate a remarkable degree of heterogeneity in T-RNA, considering that only the first few nucleotides have been analyzed. These results indicate nuuli gi eater differences in various T-RNA types than would be expected if the only differences were those in the code, which is expected to be a trinucleotide or in that size range. It is of interest to

emphasize that even the molecules which have pCpCpA as their

sixth, fifth, and fourth residues (sequences 6 and 10, Table I) are inactive

TABLE I

Terminal Nucleotide Sequences of Amino Acid-Acceptor RNA"

Per cent of RNA chains

Seciuence^ having .sequence shown

1. RNA pGpApCpCpA 6-8

2. pGpCpA 25-30

3. pGpCpU 6

4. pGpUpA 3

5. pGpCpCpU 2

6. — pGpCpCpA 4

7. pGpApG 4

8, pGpCp(UpC)pA 7-9

9. pGpCp/UpCpOp 7-12

10. pGpApCpCpCpA 4-9

11. pGpUp(CpCpU)pG 5

" Taken from Berg and Lagerkvist (1962).

* All the sequences terminate in pCpCpA as shown in the first Hne. These have

been omitted for simj)licity in the lower lines. The exact sequences shown in parentheses are not known.

in forming amino acyl-RNA compounds, when the terminal three

nucleotides are removed. The complete sequence, in this case

pCpCpApCpCpA, is required for activity. Using similar techniques combined with periodate oxidation of chains not protected by attached amino acid, it was also shown by these authors that the chains for isoleucyl-RNA and leucyl-RNA were different. However, the leucyl- RNA contained sequence 2 (Table I) which is present in 25-30% of chains and undoubtedly is not specific for leucyl-RNA. Differences in T-RNA fractions eluted from Brushite columns were found by Cantoni (1962). There was a progressive change in C plus A content and in the "odd" base content of various fractions. Although the purine to pyrimi- dine ration was 1 in all fractions, the A to U plus pseudo-U ratio departed significantly from unity in some fractions. Thus the resem- blance to T^NA noted by some authors does not appear in all fi-actions.

VIII. PROTEIN SYNTHESIS AND GENE ACTION 361

However, the possibility of a DNA-like structure in localized regions of all T-RNA molecules still remains. From the studies cited above it appears that attempts to study the code by isolating specific T-RNA fractions will be difficult.

Several other complications reinforce this viewpoint. While the evidence points to a particular T-RNA chain being able to accept only a single type of natural amino acid, it appears that more than one kind of chain can code for the same amino acid. This was demonstrated most clearly by the finding that at least 20% of E. coli leucyl-RNA had a different base sequence than the rest (Berg and Lagerkvist, 1962). Separation of two fractions from rat liver soluble RNA with leucine- acceptor activity has also been reported (Doctor et al., 1961). Similar results have been reported when T-RNA's from different species have been compared. In general, these experiments have involved studies of the formation of various amino acyl-RNA compounds using T-RNA from one species and crude activating enzyme preparations from another. Thus, different amounts of methionine-RNA were formed using E. coli T-RNA depending on whether E. coli or yeast enzymes were used (Berg et al., 1961), and guinea pig liver enzymes were quite poor for forming leucyl-RNA with E. coli T-RNA, although RNA preparations from a number of mammalian sources were approximately equivalent (Allen et al., 1960). The most detailed study of the species specificity of T-RNA has shown that the amount of a particular amino acid attached to a particular RNA varies greatly, depending on the sj^ecics from which the activating enzymes are derived and which amino acid is studied (Benzer and Weisblum, 1961). These results with crude enzyme systems are less certain than those involving separation of RNA types. However, it is clear that differences do exist among RNA chains specific for a single amino acid.

These differences may result from a non-universal code. That is, the "coding" base sequence for leucine in T-RNA from rabbits and E. coli may be different. A more likely possibility is that the specificity meas- ured in the ''species" experiments is for the reaction of T-RNA with the specific amino acyl-AMP-enzyme complex. Species specificity in this case might be expected. The discussion above indicates that a large part, if not the complete T-RNA molecule, is required for the formation of amino acyl-RNA. A smaller part of the T-RNA molecule is expected to form the coding unit (see below), and this may not be the only factor in the reaction with activating enzymes. This explanation for hetero- geneity in T-RNA chains specific for a single amino acid does not apply in the case where two leucyl-RNA chains are found in the T-RNA of the same cell and both of these appear to react equally well with a single

302 RICHARD .SCinVKKT AM) .lOllX BISHOP

activating enzyme. It is likely that a (l('<!;('nci-atc code linoi-c Ihaii one code word for a single ainiiio acidj is the exi)lanatiun here. It should Ix; possible to determine whether these RNA chains contain different codes for leucine by study of the transfer system (see below).*

The synthesis of T-RNA is a problem which is so far unresolved. If a specific sequence of bases is involved in the amino acid transfer function of T-RNA, this information might be expected to come from DNA. However, in contrast to results with "messenger" RNA, no system for synthesizing T-RNA is known and even the site of synthesis in the cell is uncertain. Herbert and Canellakis (1961) have reported experi- ments with soluble RNA which has lost ability to form amino acyl-RNA compounds after treatment with snake venom diesterase. The capacity to accept amino acids was restored by incubation with solul)le enzymes and all four (in some cases only three) ribonucleoside triphosphates.

Resynthesis of only the CCA end did not explain the restoration

of activity. However, whether a specific base sequence was synthesized, or whether large oligonucleotides were joined, is uncertain. Further studies of this system may clarify the problem of T-RNA synthesis.

III. Steps Requiring Participation of the Ribosome

A great deal of evidence has accumulated which points to the ribo- some as the site of protein synthesis. This evidence consists of kinetic studies where intact cells were incubated with a C^^-amino acid and at various time periods the specific activity of various cell fractions was examined. The ribosomes were the first to be labeled and then reached a steady state of radioactivity consistent with the hypothesis that most soluble proteins were being .synthesized in these particles (Raacke, 1961; Hoagland, 1960). When specific proteins were examined, again ribosomes were found to contain protein of highest specific activity (Rabinowitz and Olson, 1959; Peters, 1957; Siekevitz and Palade, 1960; Morris and Dickman, 1960). Studies of "nascent" enzymes on bacterial and yeast ribosomes indicate a similar situation in microorganisms (Kihara et al., 1961; Cowie et al., 1961). The labeling of specific proteins in the cell-free ribosoraal system under discussion has also been reported and constitutes the strongest evidence for protein synthesis in these systems (Schweet et al, 1958b; (;ami)bell et al, 1960; Ogata et al, 1960; Suzuka and Shimura, 1960; Kameyama and Novelli, 1960).

In this section we shall attempt to review the evidence concerning various steps whirli ](\id to the synthesis of the peptide chain in the

*Wei.sblum. Bcnzer. and Hollcv Imxc now dfrnonstratod this rosiilt (jirivate communiration).

VIII. PROTEIN SYNTHESIS AND GENE ACTION 363

ribosonie followed by the appearance of the completccl protein in the nonsedimentable fraction.

A. TRANSFER OF AMINO ACIDS FROM AMINO ACYL-RNA TO RIBOSOMES

1. Amino Acyl-RN A as an Intermediate in Protein Synthesis

The transfer of amino acids from amino acyl-RNA compounds to ribosomal protein indicated that these compounds were intermediates in protein synthesis (Hoagland et al., 1958). Evidence for a soluble inter- mediate in protein synthesis had been reported earlier by Hultin (1956). The isolation of labeled hemoglobin (Bishop et al., 1961 ; von Ehrenstein and Lipmann, 1961) and albumin (Hirokawa et al., 1961; Campbell, 1961, 1962), starting with C^^-amino acyl-RNA, provided strong evidence that these compounds could serve as intermediates for the synthesis of defined proteins.

Kinetic studies w'ith intact cells supports the role of amino acyl-RNA compounds as major intermediates in protein synthesis. The work of Hoagland et al. (1958) showed that, in ascites cells at depressed tem- peratures, labeled amino acids were initially linked to soluble RNA more rapidly than they were found in peptide linkage either in the microsome fraction or in soluble protein. Lacks and Gros (1959), using intact E. coli cells, also showed that the soluble RNA fraction was most rapidly labeled. In addition, the experiments were compatible with the idea that the label was passed on into peptide linkage when the labeled amino acid (S'^°-methionine) was diluted with a fivefold excess of non- radioactive methionine. When chloramphenicol was added to the incuba- tion, greatly reducing the rate of protein synthesis, the rate of equilibra- tion of the RNA-bound amino acid with the free amino acid pool was also reduced, although the same saturation point was ultimately reached, suggesting that transfer of the RNA-bound amino acid to protein was the normal pathway which released free RNA for further formation of RNA-bound amino acid. Lacks and Gros (1959) feel that amino acyl- RNA is an intermediate in protein synthesis, but that amino acids may also extend protein by a second pathw^ay. The hypothesis is based on the observation that the initial rate of labeling of the RNA-bound amino acid was one-third as fast as the initial rate of entiy of the label into ''protein." Roberts (1959) has indicated that the turnover rate of amino acyl-RNA was too small for these compounds to act as the only inter- mediate in protein synthesis. These results are consistent with amino acyl-RNA compounds acting as intermediates in protein synthesis in the intact cell, but suggest the existence of other pathways which could be of considerable importance. On the other hand, no other pathway for

364 KICHARD SCHWEKT AND JOHN BLSHOP

protein synthesis has been demonstrated (see discussion of membrane protein synthesis below). Amino acyl-RNA compounds were shown to be obligatoiy intermediates for hemoglobin synthesis in a cell-free system (Allen and Sclnveet, 1960).

2. The Transfer Reaction

The transfer of amino acids from amino acyl-RNA to ribosomes has now been observed in cell-free systems derived from a number of differ- ent sources. J\Iany workers have employed the livers of noniial (Hoag- land et al., 1958; Nathans and Hiilsman, 1960; jMoldave and Grossi, 1960; Takanami, 1961) or hepatectomized (von der Decken and Hultin, 1960) rats. Systems from rabbit reticulocytes (Bishop et al., 1960a; von Ehrenstein and Lipmann, 1961), from E. coli (Nathans and Lip- mann. 1961), and from pea seedlings (Webster, 1960) have also been described. The i)henomcnon, then, appears to be quite general.

a. Cofactorial Requirements for Transfer. A requirement for GTP in the transfer process was first pointed out by Hoagland et al. (1958) and has since been confirmed in a number of laboratories (Moldave and Grossi, 1960; Nathans and Hiilsman, 1960; Bishop et al., 1960a; Nathans and Lipmann, 1960; von Ehrenstein and Lipmann, 1961). In the reticulo- cyte system, a maximum rate of transfer is observed in the presence of GTP as the sole phosphorylated nucleoside (Schweet et al., 1961). Taka- nami (1961) has succeeded in purifying a liver system to the extent that ATP is inactive when it is the sole phosphorylated nucleoside present, while GTP alone is active. Some workers (Hoagland et al., 1958; Grossi and Moldave. 1960) have observed a requirement for ATP in addition to GTP. Magnesium ion is reported to be required in the pea seedling system (Webster, 1960), in rat liver systems (von der Decken and Hultin, 1960; Takanami, 1961), and at a high level in the E. coli system (Nathans and Lipmann, 1961).

A requirement for glutathione, or other sulfhydryl compounds, has been demonstrated (Nathans and Hiilsman, 1960; Hiilsman and Lip- mann, 1960) in a rat liver system. In pea seedlings, the glutathione requirement was absolute, and irreplaceable by other sulfhydiyl com- pounds (Webster, 1960) ; but in E. coli the glutathione requirement was less marked than in rat liver (Nathans and Lipmann, 1961). Again in a rat liver system, however, Hirokawa et al. (1961) found little requirement for glutathione, presumably because they used an undialyzed pH 5 supernatant (see Hiilsman and Lipmann, 1960). More diflficult to understand is the inhibitoiy effect of glutathione observed by Taka- nami (1961) in a purified rat liver system. However, the purity of commercial glutathione preparations is not always good, and Hiilsman

VIII. PROTEIN SYNTHESIS AND GENE ACTION 365

and Lipmann (1960) have demonstrated the inhibitory effect of added oxidized glutathione. A strong glutathione requirement was observed in the rabbit reticulocyte system (Bishop and Schweet, 1961a) and shown to be due to the reversible inactivation of an enzyme fraction in the absence of sulfhydiyl compounds (Bishop and Schweet, 1961b). A fully activated enzyme, freed of glutathione, could be maintained in the activated state by the addition of Versene. Under these conditions, transfer occurred in the absence of glutathione.

In summaiy, there is general agreement on a GTP requirement in the transfer reaction, although in two isolated cases ATP was required in addition to GTP for maximum transfer. Whenever it has been studied, a magnesium ion requirement has been found, although the optimum varied from system to system. It is not known whether magnesium ion is required in the transfer reaction itself, or to maintain integrity of the ribosome. Glutathione was required in most cases; in one instance, the requirement was absolute, and in one case glutathione was inhibitory.

b. Enzymatic Requirements for Transfer. An enzymatic requirement for the transfer process in the liver system was reported by Hoagland et al. (1958) to be located in the pH 5 supernatant fraction. Hiilsman and Lipmann (1960), however, have shown this to be almost completely replaceable by glutathione and other sulfhydryl compounds. Subse- quently, it appeared that a soluble enzyme fraction was, in fact, required for transfer in the liver system, but that it was present in the microsomes as they were then prepared (Nathans and Lipmann, 1960). By treating their microsome preparation with 0.5% deoxycholate (and thus preparing rihosomes) these authors were able to free their preparations of con- taminating soluble enzyme and show an almost absolute requirement for added enzyme, even in the presence of glutathione. Rabbit reticulocyte ribosomes, when prepared according to conventional methods, without resort to detergent treatment, are not associated with a lipoprotein membrane. These ribosomes did show a stimulation of amino acid trans- fer by added enzyme even in the presence of glutathione (Bishop et al., 1960a). However, rabbit reticulocyte ribosomes show an almost absolute enzyme requirement following treatment with deoxycholate (Bishop and Schweet, 1961b; von Ehrenstein and Lipmann, 1961). In the pea seedling system (Webster, 1961) and in the E. coli system (Nathans and Lipmann, 1961), an enzyme requirement has been observed with washed ribosomes, in the absence of detergent treatment. Thus, it would appear that the enzyme requirement is not an artifact of detergent treatment. Other treatments employed to show an enzyme requirement in rat liver microsomes have been a combination of two detergents (deoxycholate and Lubrol W) and a high KCl level (von der Decken and Hultin.

366 RICHARO S( IIWEKT AND JOHN BISHOP

1960). Tlie .same trctitnicnt followed hy ])i-ccipitatioii of the lihosomes with magnesium and dialysis has been used by Takanami (19()ll.

Using such purified particulate preparations, several assay systems have been developed, and some progress has been made in the purifica- tion of the nondialyzable factors required for the transfer reaction. Solubility properties, column chromatography, and heat lability indicate that the transfer factor or factors are protein and presumably an enzymatic reaction is involved (Nathans and Lipmann, 1960; Bishop and Schweet, 1961b). A heat-stable, nondialyzable transfer factor has also been reported (Grossi and Moldave, 1960). Nathans and Lipmann (1960) reported a 30- fold purification of an enzyme fraction from the rat liver 105.000 g supernatant by ammonium sulfate and acetone frac- tionation. Takanami (1961) has purified 8-fold an enzyme fraction from the same source by DEAE-cellulose chromatography. Transfer enzymes have also been purified from pea seedlings (Webster, 1961), from E. coli (Nathans and Lipmann, 1961), and from rabbit reticulo- cytes (Bishop and Schweet, 1961c).

It may be asked whether the transfer function is served by the same group of enzymes which catalyze the activation of the amino acids. This is an extremely difficult question to answer rigorously. It may be that quite different concentrations of the same group of enzymes are required for the different types of catalytic activity. Thus, the demon- stration that a certain enzyme fraction will catalyze the transfer reac- tion, but not incorporation from free amino acids (Bishop et nl., 1960a; Takanami, 1961), may simply mean that more enzyme is required to activate than to transfer the amino acids. However, relative purification data (von der Decken and Hultin, 1960) and the failure to form leucyl-RNA by a fraction which transferred from C**-leucyl-RNA to protein (Nathans et al., 1962) is highly suggestive that transfer is not a function of the activating enzymes. The final answer, however, may be found in working with the artificial system of Nirenberg et al. (1962) in which a mixture of polyuridylic acid and DNasc-trcated E. coli ribo- somes will accept only the amino acid phenylalanine from amino acyl-RNA.

The first suggestion that more than one enzyme was involved in the transfer reaction came from von der Decken and Hultin (1960), who fractionated the pH 5 supernatant of rat liver by batch treatment with hydroxylapatite, and followed the activity of various eluted fractions in the transfer of valine and tyrosine. In one particular instance a relative separation of valine and tyrosine transfer activities was observed. However, there was a strong correlation between low tyrosine transfer and hiqh amount of tyrosine activation, suggesting that Ww low transfer

VIII. PROTEIN SYNTHESIS AND GENE ACTION 367

of C^'^-tyrosine might be due to exchange of C^ --tyrosine with C^"*- tyrosyl-RNA yielding dilution of the specific activity. Such an exchange seems likely because of the long incubation (in the presence of ATP and a generating system) which was used. On the other hand, the condi- tions of the incubation, with a veiy high microsome: amino acyl-RNA ratio, were such that transfer of a single amino acid might well have been observed, in the relative absence of transfer of another. Bishop and Schwect (1961c) fractionated the 105,000 (/ supernatant of the rabbit reticulocyte lysate using salting out with ammonium sulfate, adsorption to calcium phosphate gel, and chromatography on DEAE-cellulose; they obtained two enzyme fractions which showed a 6- to 10-fold enhancement of activity when assayed together. These fractions showed small differences in relative activities in the transfer of four amino acids under conditions (3-minute assay in the absence of ATP or an ATP- generating system) which precluded the occurrence of a significant exchange reaction between RNA-bound and free amino acids. Fessenden and Moldave (1961), using rat liver ribosomes, have reported that the enzymes removed from microsomes with deoxycholate did not catalyze transfer unless a purified, heat-stable, but nondialyzable fraction of the soluble cytoplasm was added. The significance of these fractions is not apparent, since microsomes have been reported to contain transfer en- zymes and to show full activity if glutathione alone is added (Hiilsman and Lipmann, 1960). Thus, the fraction solubilized by DOC should contain all the transfer enzymes.

On the other hand, Takanami (1961), using a rat liver system, and Nathans and Lipmann (1961) w^ith E. coli, have fractionated the super- natant factor on DEAE-cellulose without observing any separation of transfer activities. It appears likely that more than one enzyme is involved in the transfer reaction. However, whether separate stages in the reaction are catalyzed by different enzymes, or whether amino acid-specific fractions have been separated is still not clear.

3. The Mechanism of the Tran.sjer Reaction

Evidence that transfer RNA is repeatedly recharged with amino acids, after discharging bound amino acid on the template, was presented by Allen and Schweet (1960). The complete reticulocyte cell-free system (starting with free amino acid) showed a considerable requirement for added transfer RNA which was fulfilled by 45-50 fig of RNA. The stimu- lation of C^*-leucine incorjioration into protein, however, was 15-20 times as much as the amount of leucine required to charge completely 50 /xg of transfer RNA. Thus, the added transfer RNA must have re- cycled many times in the course of the incu!)ation. Independently,

368 RICHARD SCHWEET AND JOHN BISHOP

Nathan.s and IjpiiKinii (1961) have reported that after discharging its amino acid, E. coli RNA coidd be reisohited (by i)henol extraction) and completely recharged.

The original "adaptor hypothesis" (Crick, 1958) called for a rela- tively small polynucleotide as the adaptor. It was suggested (Hoagland, 1960) that the relatively large size of the transfer RNA molecule was a requirement for specific interaction with the activating enzyme and that, after accepting the amino acid, the transfer RNA molecule might release a small polynucleotide fragment with the amino acid attached to it. The results cited above suggest that the transfer RNA does not break down. Hoagland and Comly (1960) have provided careful studies which point to the same conclusion. C'*-leucyl-RNA was incubated with various enzyme fractions, or microsome extracts, in the presence and absence of GTP, but in the absence of microsomes. At various times parallel assays were made of the remaining acid-precipitable amino acid, and of the competence of the complete incubation mixture in the transfer process. It was found in every case that the amount of C"-leucine transferred to peptide linkage was proportional to the amount of acid-jirecipitable C"-leucine remaining. Thus, free, acid-soluble, amino acid polynucleo- tide compounds do not appear to be significant intermediates in protein synthesis. It remains to be seen whether such intermediates are foimed in association with microsomes.

a. The Transfer of Labeled, Soluble RNA to the Ribornicleoprotem Particles. If, as discussed above, amino acyl-RNA fails to form a smaller amino acyl polynucleotide, a direct physical association of the transfer RNA and the microsomes, if only a veiy transient one, must occur during transfer. A temperature- and energy-dependent transfer of labeled, soluble RNA to ribosomes has been reported by von der Decken and Hultin (1958), by Bosch et al. (1959), and by Hoagland and Comly (1960). In all cases, a considerable non-temperaturc-dependent, non- energy-dependent transfer of labeled RNA to the 105,(XX) g pellet was also observed. The system studied in greatest detail is the rat liver system of Bosch et al. (1959, 1960; Bloemendal et al, 1960, 1961). These authors found that a P'^-'-S-RNA fi-action isolated on an Ecteola column was readily incorporated into ribosonial RNA. The fraction was able to accept C"-leucine in the presence of an activating system. After incuba- tion with P'--S-RNA, the microsomes were sedimented and the micro- somal RNA extracted with phenol and chromatographed on an Ecteola column. Three fractions were recovered, only one of which was labeled. Significantly, the labeled fraction had different chromatographic prop- erties from those of the labeled soluble RNA. Furthermore, when micro- somal RN.\ was isolated after P^--labeling in viiw, radioactivitv was

VIII. PROTEIN SYNTHESIS AND GENE ACTION 369

found in all three chromatographic fractions. Bloemendal et al. (1961) observed a similar transfer of labeled RNA to detergent-treated liver ribosomes. When the soluble RNA was prelabeled in the terminal posi- tion with C"-adenine and incubated with ribosomes, the C^*-adenine was found in internal positions in the ribosomal RNA (63% of the radio- activity was isolated as adenosine monophosphate after alkaline hydrol- ysis). It appears, then, that the transfer which occurred in this system resulted in the covalent linkage of solul)le RNA with microsomal RNA. It is not clear, however, that this reaction is related in any way to the transfer of amino acids from amino acyl-RNA. In fact, the addition of soluble enzymes and an amino acid mixture failed to stimulate the RNA transfer (Bloemendal et al., 1961).

A transfer of labeled S-RNA to ribosomes has been reported by Hoagland and Comly (1960). In this case, P''=-labeled, soluble RNA was coupled with C'^-leucinc so that the transfer of both the RNA and the amino acid could be studied. It was shown not only that the P^--labeled RNA would enter the sedimentable fraction, but also that it could be displaced by unlabeled soluble RNA. A measure of specificity was indicated by the inability of unlabeled microsomal RNA to displace the labeled soluble RNA from the microsomes. In agreement with the experi- ments of Bloemendal et al. (1961), when the C^'^-leucine was removed from the P^--RNA, no inhibition of RNA transfer was observed.

These various experiments do demonstrate that an association between soluble RNA and the ribonucleoprotein particles may be very readily achieved. The slight stimulations observed upon the addition of ATP or GTP may well indicate simply that this is not an energy-dependent step.

Since it is altogether likely that a physical association can occur between soluble RNA and the microsomes, it might be expected that when C"-amino acid-labeled amino acyl-RNA is incubated briefly with microsomes, some portion of the amino acyl-RNA will sediment with the microsomal fraction. Zamecnik (1960) has in fact observed such a transfer to the RNA of the sedimentable fraction in a rat liver system. The problem of the first intermediate in the ribosome has been studied in the reticulocyte system starting with C^^-leucyl-RNA (G. Favelukes and R. Schweet, unpublished data) to avoid complications of contam- ination with free C'^-amino acid. Incorporation into ribosomes in alkali- labile linkage was used as a measure of amino acyl-RNA binding. As shown in Table II, a GTP- and enzyme-dependent linkage was found. However, the significance of these alkali-labile counts as an inter- mediate in jirotein synthesis is dubious, since these counts could not be "chased" into alkali-stable (protein) linkage. It seems likely, there-

370 HICflAUl) SCinVKKT AM) .TOIIX HISIIOP

fore, that wliik' tlici'c must he some initial association of amino ac_\'l- HNA and ribosouR's, this is very transitory and the annuo acid is transferred from the amino acyl-RNA to an alkali-stahh' foim in the ribosonie veiy rapidly. The RNA is then fice to foi'in amino acyl-RNA once again.

TABLE II RiBosoMAL C"-Leuci.\e IN Alkali-Labu^e Li.\ka(;e"

	C"-Leucine	incorp	orated
Assay conditions'*	Labile infjimolcs)		Stable (At/intioles)
Complete Minus enzyme Minus GTP Complete plus 2 X 10^^ M puromycin	93 4 20 74		154 6 9 27

" G. Favclukes and R. Schweet (unpublished data).

'' DOC-ribosomes (2.8 mg), C'Meuoyl-RNA (0.8 mg), enzyme (3 mg) and GTV (1 /miole) were incubated at 37°C for 3 minutes.

b. The Role of GTP in the Transfer Reaction. An understanding of the role of GTP is essential to the elucidation of the mechanism of the transfer reaction; but to date little information is available. Nathans et al. (1962) have shown, in the reticulocyte transfer system, that the breakdown of GTP to GMP and pyrophosphate is several orders of magnitude lower in extent than amino acid transfer. This would appear to eliminate the breakdown of GTP to GMP as a significant factor in the transfer process. Webster (1961), using a pea seedling system, has reported a reversible breakdown of GTP accompanying amino acid transfer. Incubation of ribosomes charged with amino acids and soluble RNA resulted in the formation of amino acyl-RNA and GTP. On the other hand, Nathans et al. (1962) have observed a rapid breakdown of GTP to GDP upon mixing E. coli ribosomes and soluble enzyni(>s in the absence of amino acyl-RNA.

Several pieces of evidence which suggest a role for GTP have come from studies of "release" in the reticulocyte ribosome system (Morris and Schweet, 1961; Morris et al, 1962). In these studies it was shown that ribosomes prelabeled with C^*-leucine could be incubated in a com- plete system with C^ --amino acids and the radioactivity "chased" into soluble hemoglobin. In this system, incorporation of C^--amino acids finishes partially completed chains in the ribosome and the formation of labeled, soluble protein is a measure of these reactions. It has now been found that GTP will also i)roniote release of soluble C^''-protein from

VIII. PROTEIN SYNTHESIS AND GENE ACTION 371

prelabeled ribosomes (Table III). This formation of soluble protein occurs in the absence of incorporation of free C^'*-amino acids into ribo- somes. This result dcmon.><trates a role for GTP at a later stage than the transfer of amino acid from amino acyl-RNA. One interpretation of this finding is that GTP acts to convert intraribosomal intermediates into peptide linkage, thus producing completed or almost completed chains which are then released. It is postulated that a GTP is split in

TABLE III Requirements for Incorporation and Release"

Incorporation Release

Assay components (%) (%)

Complete system 100 100 No energy 1.5 4.1

Plus 1 Mmole ATP only 51.1 50 . 2

Plus 1 nmole GTP only 4.9 35.2

" A. Morris and R. Schweet (unpublished data).

the process of forming each peptide bond and that GTP does not act in the initial transfer step from amino acyl-RNA. However, since ribo- somes as isolated are filled with intermediates, no transfer from amino acyl-RNA is observed in the absence of GTP, since no ribosomal sites are available. Thus, an intraribosomal reaction requiring GTP is needed in order for an earlier step, the transfer from amino acyl-RNA, to occur. This mechanism is advanced as a working hypothesis to guide future studies of the role of GTP.

4. Specificity in the Transfer Reaction

A lack of specificity for the transfer RNA species is the most strik- ing aspect of this reaction. Von Ehrenstein and Lipmann (1961) demon- strated that E. coli C"-amino acyl-RNA could form hemoglobin when incubated with rabbit reticulocyte ribosomes. Bishop et al. (1961) showed that amino acyl-RNA from guinea pig liver was competent in hemo- globin synthesis with reticulocyte ribosomes. In these experiments, the labeled hemoglobin was digested with trypsin and fingerprinted. The correspondence of leucine-labeling in these peptides with whole-cell labeled hemoglobin (Table IV) provides good evidence for the synthesis of hemoglobin from this heterologous amino acyl-RNA.

Despite the ability of guinea pig liver amino acyl-RNA to act as a precursor of hemoglobin synthesis, it was observed (Bishop et al, 1961) that amino acids were transferred more readily from amino acyl-RNA to homologous ribosomes. Guinea pig liver transfer RNA was labeled

372

RICHARD SCHWEET AND JOHN BISHOP

with H'-lt'uciiu'. and rabbit roticiilocyto RNA with C"-lcucine. To eliminate posi-ible effects of contaminants of the RNA preparations, the two hibeled RNA's were mixed and incubated in j)arallel experiments with either rabbit reticulocyte or guinea pig liver ribosomes. More H^-leucine than C"-leucine was transferred to guinea pig liver ribo- somes, and more C^^-leucine that H''-leucine to reticulocyte ribosomes. Two explanations of these results are suggested. (/) There exist subtle differences between rabbit reticulocyte and guinea pig liver leucine-

TABLE IV

RAniOACTIVITY OK PEPTIDES OF AlPHA ChAIX OK RaBKIT llE.\I()(il-OBIN" ''

		Fingerpri	nt 1	Fingerprint	.2
Peptide	Intact cells	Cell-free	Intact cells		Cell-free
No.	(C	Meucine)	(HMeucine)	(CiMeucine)	(H'-	-leucyl-RNA)
1		2.5	4.6	9.2		8.7
2		8.2	5.1	3.8		1.9
3		31.1	12.5	32.2		36.9
4		7.2	6.6	6.1		6.8
5		18.2	20.4	20.3		20.4
6 + 7		1.3	2.0	1.8		1.1
8		3.2	2.3	1.4		1.6
9		0	0	0		0
10		16.4	19.8	13.5		13.0
12		12.2	26.5	11.6		9.7
13		0	0	0		0
14		0	0	0		0
Blank 1		0	0	0		0
Blank 2		0	0	0		0
Blank 3		0	0	0		0

- From Bishop et al (1961).

* Radioactivity is expressed as the percentage of the total radioactivity in all peptides. 0 denotes less than 0.1%.

transfer RNA which affect the rate of the transfer reaction without influencing the site to which the leucine is transferred. (2) Different types of transfer RNA, specific for the same amino acid, are present and have different transfer specificities, perhaps transfer to different types of sites. If, for example, two kinds of leucyl-RNA's exist, but in different relative proportions in guinea pig liver and rabbit reticulocytes, the results may be explained if, in addition, different relative frequencies of two corresponding types of template sites are found in the ribosomes of these ti.^sues.

Specificity in transfer enzymes has been reported. Transfer enzyme fractions of rabbit reticulocytes and E. coll were not interchangeable

VIII. PROTEIN SYNTHESIS AND GENE ACTION 373

(Nathans and Lipmann, 1961). The partially purified E. coli enzyme fraction already referred to was completely inactive in the transfer of amino acids from amino acyl-RNA to reticulocyte ribosomes. Sim- ilarly, rabbit reticulocyte pH 5 supernatant was inactive in the E. coli system. The significance of this type of specificity is not understood.

B. FORMATION OF THE PEPTIDE CHAIN

Understanding of the mechanism of assembly of the amino acids may be expected to yield information on the central problem of how amino acid sequence is specified. One approach has been to study the distribu- tion of a C^^-amino acid in various portions of a specific protein synthe- sized in the presence of this labeled amino acid. If the specific activity of the C"-amino acid is different in various parts of the peptide chain (unequal labeling) , information about intermediate pools or sequence of peptide bond formation can be deduced (Loftfield, 1960).

1. Labeling Patterns in Intact Cells

Unequal labeling of specific proteins has been observed in a number of systems. The clearest example, perhaps, is the unequal labeling of the ovalbumin synthesized in chicken oviduct (reviewed by Steinberg et al., 1956). Unequal labeling was clearly demonstrated by analysis of the products of (a) partial acid hydrolysis, (6) peptic digestion, and (c) sequence analysis of the hexapeptide liberated by subtilisin. The data were interpreted to indicate the existence of intermediate pools of some kind, or an ordered synthesis of the chain. It is not possible to distinguish between these possibilities as the author noted. Equal labeling of goat lactalbumin was observed by Askonas et al. (1955). This is consistent with any of the hypotheses discussed above, since the time to synthesize a lactalbumin molecule was not determined.

The a-amylase secreted by late log phase B. siibtilis is the subject of a careful study by Yoshida and Tobita (1960). C^*-leucine was added to the culture during the later stages of growth, a-amylase was isolated, and leucine was isolated from positions close to the C-teiTninus, the N-terminus, and from eight tryptic peptides. The highest specific activity leucine was found close to the C-terminus, and the lowest close to the N-terminus. Furthermore, the specific activities of the leucine in the eight tryptic peptides fell between these limits. This experiment, then, is suggestive of an ordered synthesis of the polypeptide chain of ^-amylase, starting at the N- terminal end, and finishing at the C-terminal end. It should be noted, however, that a high molecular weight precursor of amylase was present before the isotope experiment began and the results may mean that this precursor was completed.

374 KICllAHl) SCIIWKKT AND JOHN HISHOP

Tilt' lahi'linu; of liciuo^lohin in rcticulocytt's has been studied by a miinbt'i- of workers, with lathci- coullictin^ results. Muir et al. (1952) injected rats with C"-valine and deinoiistrated that the specific activity of the N-terniinal valine of isolated hemoglobin was the same as the specific activity of the valine of the remainder of the molecule. On the other hand, Ki'uh ct a/. (1907, 19()()) usino; radioactive glycine and phenylalanine have obserN-ed unef|ual labelin*^ of hemoj.!;lobin in isolated rabbit I'eticulocytes and in the livini;; rabbit. At times after the addition or injection of the labeled amino acid ranging from 1 to 4 hours in ex- periments with isolated cells, and from 5 to 58 days in the living animal, hemoglobin was isolated. The specific activity of the labeled amino acid released after a brief hydrolysis of each sam])lc with HCl was compared with the specific activity after complete hydrolysis. In the experiments with isolated cells, unequal labeling could still be detected after 4 hours, while in the living rabbit uniform labeling was not attained until between 30 and 50 days following injection. These experi- ments are difficult to understand in terms of any of the hypotheses already put forward. Since the time for synthesis of the hemoglobin molecule is estimated to be approximately 1 minute (Dintzis et al., 1958; Schweet et al., 1961), ordered synthesis could not result in signifi- cantly unequal labeling at times exceeding 20 minutes. Schapira et nl. (1960) have examined the question in another way. After incubation of reticulocytes with radioactive arginine, hemoglobin was isolated and digested with tiypsin. The digest was "fingerprinted," the arginine pep- tides identified, eluted, and hydrolyzed. The specific activity of the arginine in each of nine peptides was found to be different. It should be noted that Jonxis (1958) has reported four arginine residues per half- molecule of hemoglobin. It appears unlikely that the long-term unequal labeling was due to an analytical artifact since unequal labeling was detected in two different ways. It is possible that slow, long-term exchange reactions, or the presence of several types of hemoglobin with different turnover rates, could explain the data. Further work is needed to clarify these results.

2. Evidence for Ordered Peptide Bond Formation

As mentioned earlier, labeled amino acids rapidly appear in ribosomes in alkali-stable linkage, presumably protein (Hoagland, 1960). In intact reticulocytes, "pulse-labeling" experiments demonstrated the transitory nature of this material and indicated that hemoglobin precursors, rather than the structural protein of ribosomes, were formed (Dintzis et al., 1958; Rabinowitz and Olson, 1959). These results suggested the presence of peptides in the ribosome which were hemoglobin precursors. Similar

VIII. PROTEIN SYNTHESIS AND GENE ACTION 375

results were obtained in the cell-free system from reticulocytes (Schweet et al, 1958b).

Since earlier studies had siiown that the labeled ribosomal material could be "chased" directly into soluble hemoglobin when labeled ribo- somes were incubated in the complete cell-free system (Schweet et al., 1961; ]\Iorris and Schweet, 1961), this system was used to study the nature of the ribosomal intermediates. Ribosomes were labeled in tlie intact reticulocyte with C^'* -valine, the N-terminal amino acid of hemoglobin. These ribosomes were washed and then incubated with C^--valine in the cell-free system. Comparison of the amount of C^'*- valine in N-terminal positions and in internal positions in the soluble hemoglobin formed, indicated that the ribosomes contained incomplete chains synthesized sequentially from the N-terminal end (Bishop et al., 1960b). This was also demonstrated in intact reticulocytes by similar methods (Schweet et al., 1961).

Dintzis (1961) studied the labeling in peptides formed by the tryptic digestion of the separated chains of hemoglobin labeled in the intact cell. He showed that tryptic peptides isolated from hemoglobin after short periods of incubation with labeled leucine could be arranged in a sequence of increasing specific activity relative to the specific activities of the same peptides from hemoglobin isolated after 6 hours of incuba- tion. The peptide with the highest specific activity after short periods of labeling was near the C-terminus. These results are also suggestive of synthesis from the N-terminus is sequential fashion, although other interpretations of labeling data using intact cells are possible (see Steinberg et al., 1956). The results of Goldstein and Brown (1961) indi- cate that this mechanism may be general. A leucine-requiring strain of E. coli was starved of leucine to reduce the possibility of interference by unlabeled pools and then labeled for very short times with C"-leucine. The shorter the labeling period, the more C^*-leucine was found near the C-terminus. On the other hand, Shimura et al. (1956) have shown that shortly after injection of C'^-glycine into silkworm larvae, the N-terminal glycine of silk fibroin had the highest specific activity. A sequential synthesis of many proteins from the N-terminal end appears likely.

The significance of polypeptide chain synthesis from the N-terminal end, if any, is not known ; however, if one makes several reasonable assumptions some interesting consequences follow. The two main assump- tions are:

1. In addition to functioning to position amino acids in the proper sequence for the protein being made, the ribosome functions to position amino acids in the proper steric relation so that the peptide bond can be

376

HICIIAKI) SCIIWKKT AND JOHN UISHOI'

foriiK'd Ix'twrcn the aiuiiio ^louj) of one amino acid and the carboxyl group of the other. The projjer condition for this is shown in Fig. lA. Two carboxyl-activated amino acids arc attached to the ribosomc ready for peptide bond formation. These may be attached to transfer RNA, or in some other linkage.

NH-CH NH-CH

c=o c=o

RIBOSOME

I NH,CH

I

c=o

NH— CH

I

c = o

R/BOSOME

RIBOSOME

INTERRUPTED GROWTH

RIBOSOME

GROWTH FROM C-TERMtNUS

NKaa,

NHaa.

aa^ aa, 1 1			1 1
RIBOSOME			RIBOSOME
	GRC	)WTH FROM	N-TERMINUS

Fig. 1. Proposed mechanisin.s of pei)tide chain growth in protein synthcsi.s.

2. When the peptide bond is formed, one of the amino acids is then no longer attached to the ribosome (Fig. lA, on the right). From the previous assumption, if a third amino acid is to be added to tlie chain, it will attach to the ribosome by its carboxyl group, but then will not be able to link if it attaches to the left of amino acid R', since R' is not attached to the ribosome. It can link only by attaching to the right of amino acid R- and attaching to this amino acid.

Thus, if peptides would start to form at various points, at some stage (Fig. IB), aag (attached to the ribosome) would need to form a peptide bond with aa^ of another peptide; according to our assumption this is

Vlll. PROTEIN SYNTHESIS AND GENE ACTION 377

not possible for steric reasons. As shown in Fig. IC, if growth of the chain started at the C-temiinus, a similar situation would arise. How- ever, if the N-teniiinal amino acid is attached first by its carboxyl group and initiates chain growth (Fig. ID), the growing tip of the chain (aag) is attached and can form a peptide bond with aa4, since both are attached to the ribosome. The new growing tip at aa4 is still attached and when aag attaches, further chain growth can continue.

How is an ordered synthesis from the N-terminus arranged for? Presumably, this is a function of the "code" and is accomplished by the mechanism of interaction between messenger RNA (see Section IV) and amino acid-specific transfer RNA. The first requirement for the code in this viewpoint is that it can be transcribed only from a single starting point, corresponding to the N-temiinal amino acid. Genetic evidence for such a code has been reported by Crick et al. (1961). The mechanism by which the N-terminal amino acid is positioned on the messenger RNA- ribosome template first is not known, but several possibilities have been discussed by Goldstein (1962). What is the mechanism by wiiich the code ensures that peptides are not formed anywhere along the chain, but only sequentially from the N-teiTiiinus? A simple mechanism for providing this arises from a consideration of the mechanism of interaction between the messenger RNA in the ribosome and the incoming amino acyl-RNA. To find its proper place in the amino acid sequence, a triplet of bases in the T-RNA is supposed to hydrogen bond to a triplet of bases in the messenger RNA (see preceding discussion). This alone, however, is postulated to be too weak an interaction to permit the amino acid to remain on the template long enough to participate in further reactions. If it is assumed that in addition to bonding to the code sequence of the messenger RNA, the T-RNA of the incoming amino acid must bond to the preceding molecule of T-RNA, then sequential growth is provided for. According to this idea each T-RNA molecule, in addition to its unique coding triplet of bases, must contain two common sets of sequences which are complementary. For example, each T-RNA would have possibly an AG-rich region and a UC-rich region at different parts of the chain. AC-rich and GU-rich would work also. The AG-rich and UC-rich regions might be 10 or more nucleotides in length and might or might not be exactly complementary in sequence for the required interactions to take place. According to this hypothesis, the incoming amino acyl-RNA would have to bond both by its triplet code to the messenger RNA and by its UC-rich region to the AG-rich region of the T-RNA on the preceding amino acid for effective amino acid transfer. Tliis would be the one at the growing tip of the chain according to the previous discussion.

378 HiciiAiu) sciiwKKr and .lonx hisikh'

(^tlior inocli;inisins can he pi'opo.-cd hy wliidi the iiicoiiiiiio; amino acid could recognize that the |)rece(nn^ amino acid is j)resent. Whatever mechanism is iiostulated, however, it appeal's likely that the important characteristic of chain growth fi'om the N-terminal end is that the incoming amino acid and the i)i'('ccding one are both on the template. This provides steric conditions suitable for jieptide bond formation, and could also ensure uninterrupted chain growth, as discussed above.

3. Final Stages in Protein Sj/nthesis

The pre^•ious discussion has dealt with indirect e\-idence for a particu- lar sequential type of peptide chain growth. It should be noted that details of the mechanism of peptide bond formation and the inter- mediates directly involved in this process are as yet unknown. The role for GTP suggested eai'liei-, which is hypothetical as yet, is the first approach to this problem. Skipping over these basic questions, therefore, the next problem concerns the mechanism by which finished proteins are released from the ribosome. The best evidence that some mechanism is needed to release finished proteins from their site of synthesis comes from experiments where "nascent," enzymatically active proteins on ribosomes cannot be removed by simple washing techniques (Kihara et a!.. 1961; Cowie et al, 1961; Siekevitz and Palade, 1960).

In cell-free systems, Simkin (1958) showed that labeled liver microsomes formed soluble protein when incubated with cell sap and an energy-generating system. In the reticulocyte ribosome system, similar studies demonstrated the formation of labeled, soluble hemoglobin when labeled ribosomes were incubated in the complete system (C^--amino acids, energy, and enzymes). It was also shown that the ribosomes were not degraded under these conditions (]\Iorris and Schweet, 1961). How- ever, these experiments are essentially "chase" experiments and do not distinguish between release of completed proteins and peptide bond synthesis which results in synthesis of completed chains. Soluble enzyme fractions which increase the formation of soluble protein without increas- ing total incorporation greatly have been termed "releasing" enzymes (Lamfrom, 1961), and it has been suggested that these were species specific. Howc\-(M-, if i-ibosomal-bound proteins were nearly completed, the addition of only a few amino acids might result in tlie fonnation of soluble protein. Further discussion and evidence for this viewpoint is given in the next section. Until systems are developed where incorpora- tion into ribosomes is 7wt occurring while "release" is being measured, the problem of whether or not specific "releasing" enzymes exist remains open.

When the transfer system (from amino acyl-RNA to i)rotein) is

VIII. PROTEIN SYNTHESIS AND GENE ACTION 379

used, since ATP is not required, a better assessment of the energy re- quirements for the formation of soluble protein may be made. Bishop et al. (1961) found that little soluble protein was formed in the transfer system, with GTP as the only nucleoside trijihosphate. Howevei", this was probably due to the small amount of C^^-amino acyl-RNA used in these experiments. Formation of soluble protein using a transfer system was shown by von Ehrenstein and Lipmann (1961). However, since GTP plus a generating system was used, it is likely that ATP was formed. In any event, since chains were being formed, the "release" mechanism per se was not studied in either of these cases. Even the formation of soluble protein from labeled ribosomes with GTP and enzyme in the absence of amino acyl-RNA (Table III) does not repre- sent release alone, since it is possible that chain growth from intraribo- somal intermediates occurs here. Thus, other than the evidence which suggests that some mechanism is involved in liberating completed pro- teins from the ribosome, practically nothing is known.

Studies of the mechanism of inhibition of protein synthesis by the antibiotic puromycin have been of interest in this connection. Inhibition of protein synthesis puromycin was first reported by Yarmolinsky and de la Haba (1959). Inhibition of transfer from amino acyl-RNA to ribosomes was considered to be the site of action of puromycin. This seemed likely since puromycin has a structure analogous to the amino acyl-adenosine end of transfer RNA. A direct effect of puromycin on the ribosome resulting in the release of soluble protein has been reported for reticulocyte ribosomes (]Morris and Schweet, 1961) and liver ribo- somes (Hultin, 1961). It was later shown that release of soluble protein by puromycin was maximal in the absence of added soluble enzyme and energy (jVIorris et al., 1962). Release of soluble protein occurred at 4°C, although more slowly than at 37° C. Ribosome breakdown could not be detected under these conditions. The protein material released by puro- mycin in the absence of added enzjane was shown to be incomplete globin chains synthesized from the N-terminal end, confirming earlier conclu- sions (Section III,B,2).

In the presence of soluble enzymes and puromycin, in addition, free amino acids and small peptides are found in the TCA-soluble fraction. These have not been well characterized, but it has been suggested that they may represent earlier, intraribosomal intermediates in globin syn- thesis (Morris et al., 1962). The inhibition of protein synthesis by puromycin is considered to result from the displacement of incomplete globin chains and the postulated earlier intermediates from the ribosome. In the ]:)resence of puromycin, further formation of TCA-precipitable peptides would be blocked by a continuation of the same process. This

380 RICHARD SCHWEET AND JOHN BISHOP

mechani:>in would ix'i'iuit a I'citain aiiiouiil of |)e|)ti<lc Ixjiid I'oiniatioii in the presence of puroniycin, which might result in the formation of small peptides or earlier intermediates; Init once the larger pei)tid(' chains in the ribosome were released, no incorporation into TCA- precipitable material would occur. The locus of puromycin action is considered to be at the growing point of the chain, which is interrupted at a specific site, possibly when an aromatic amino acid appears in the sequence. This last result is suggested from the fact that neither the leucine analog of puromycin (Rabinowitz and Fisher, 1962), or the free aminonucleoside (A. Morris and R. Schweet, unpublished data) was active. The former authors have studied the effects of puromycin on intact ascites cells. Incorporation into ribosomes was inhibited, but the formation of labeled soluble protein continued. These results were attributed to release of incomplete proteins. However, no release of protein from the ribosomes was obser\-ed witli intact reticulocytes in this case (Rabinowitz and Fisher, 1962).

Inhibition of incorporation into i)rotein })y puromycin in the transfer system from E. coli has been studied (Nathans and Lipmann, 1961; Nathans et al., 1962). These authors observed degradation of amino acyl-RNA in the presence of puromycin. This de-acylation was de- pendent on ribosomes, GTP, and soluble enzymes. It is possible that the de-acylation may be a secondary phenomenon resulting from the incor- poration of the C^*-amino acid of the amino acyl-RNA into small peptides and other intermediates in the ribosome which are then released in the presence of puromycin. Material which was not free amino acid was indeed noted by Nathans et al. (1962). Thus, an effect at the first stages of transfer may reflect the primary action of puromycin on an intraribosomal event, as noted earlier in connection W'ith GTP action, where a cofactor required for transjer is believed to act on an intraribo- somal stage.

To conclude this section on "release," it should be noted that the discussion has been concerned only with the fomiation of the peptide chain sequence. Specific disulfide bond and tertiary structure formation is currently considered to follow spontaneously (Crick, 1958). The evidence for this viewpoint has been discussed by Berg (1961).

IV. Information Transfer in Protein Synthesis

In this section, an attempt will be made to review in some detail the question of how^ the information in DNA specifies a particular amino acid sequence. That DNA is the controlling factor, with some possible exceptions to be discussed, has become clear (see Berg, 1961, for a recent discussion). The mechanism bv which DNA transfers this information is

VIII. PROTEIN SYNTHESIS AND GENE ACTION 381

now considered to be via formation of a "messenger" RNA. Since the problem of DNA and messenger RNA synthesis is reviewed elsewhere (see Chapters VI and VII), the emphasis here will be on the function of DNA and RNA in i:)rotein synthesis, particularly in cell-free systems. The question of whether a linear sequence of the bases in DNA cor- responds to a linear sequence of amino acids is omitted also. The role of RNA in protein synthesis in cell-free systems will be emphasized, since studies in intact bacteria are reviewed by Volkin (Chapter VI). Finally, results with bacteria are treated separately from those using higher organisms because of the very different picture of DNA function in these two cases.

A. INFORMATION TRANSFER IN BACTERIAL PROTEIN SYNTHESIS

The evidence that DNA carries the information for protein synthe- sis in most systems (RNA viruses may be an exception) is clear. The question to be considered here is the mechanism involved. One possibility might be that DNA is itself the template for protein synthesis. This does not appear to be the case, although in intact bacteria the evidence is not particularly convincing. Rather, the ability of enucleated algae and amebae to continue protein synthesis has provided the major evi- dence for this view (see next section). The best evidence that DNA is not the template for protein synthesis in bacteria is the finding that DNA acts to make a particular type of RNA, which does appear to participate in the formation of the template for protein synthesis, as discussed below. These studies involve cell-free systems and it must be noted that the evidence using intact bacteria does not distinguish clearly between the direct action of DNA as a template, or the accepted hypothesis that DNA produces an unstable intermediate, the messenger RNA.

The evidence which has led to the present concept of an unstable "messenger" RNA as the intermediate between DNA and protein comes largely from studies of induced enzyme formation (Jacob and Monod, 1961) and phage infection (see Chapter VI). The existence of RNA which is rapidly labeled in intact cells, has the composition of DNA and can form what seem to be specific hybrids with DNA has been demonstrated. These studies have relied largely on labeling experiments and the isolation and purification of messenger RNA has not been achieved, although the report of Bautz and Hall (1962) suggests that this may be accomplished in the near future. However, crude messenger RNA isolated from the intact cell is heterogeneous wdth regard to molecular weight (see Chapter VI for detailed discussion). This hetero- geneity may reflect a structural heterogeneity related to the particular

382 RICHARD SCHWEET AND JOHN BISHOP

protein for wliich the RXA contains the sequence information. The separation of the postulated single protein-specific messenger RNA has not been achieved. Further, RNA with the i)roperties described above may be heterogeneous with regard to function and need not consist solely of RNA carrying sequence information for protein synthesis (see Hayashi and Spicgelman. 1961, for a discussion of tiiis point). The synthesis of RNA with the base composition of a primer DNA has been demonstrated in cell-free systems (Weiss, 1960; Weiss and Nakamoto, 1961; Hurwitz et ai. 1960; Stevens, 1960). The use of these "RNA polymerase" enzymes in cell-free systems has provided the most powerful evidence for the messenger RNA hypothesis.

1. Functioning of Messenger RXA in Intact Cells

Among the first clues which suggested that RNA could carry informa- tion for protein synthesis was the finding of infective TMV-RNA (Gierer and Schramm, 1956; Fraenkel-Conrat, 1956). At the time, the idea that this infective RNA played a direct role in the synthesis of TMV coat protein and other proteins was not seriously considered. Recent results in cell-free systems (see below) and the effect of mutagenic agents on TMV-RNA (see Chapter X) suggest that TMV-RNA is or contains messenger RNA. However, the molecular weight of infective T]\1V is higher than the reported values for the sedimentation constants of messenger RNA. which appear to range from 8 to 16S.^ The size of the minimum unit of T]MV which will act as messenger in cell-free systems is not known.

The other systems which have led to the messenger RNA concept also involve "new" protein i^ynthesis. The very careful studies of Volkin and Astrachan (see Chapter VI) showed that following phage infection of E. coli a small fraction of the RNA was labeled when P^- was added. This RNA had the properties of messenger RNA, particularly with respect to the base composition which resembled the DNA of the infecting phage. Recently, Volkin has shown that the synthesis of this RNA is essential for phage protein synthesis and, in fact, that some stoichiometry exists between the amount of RNA and jirotein synthesis. This kind of evidence and similar studies which, for example, show a requirement for uracil for phage-induced enzyme synthesis (Pizer and Cohen, 1961) provide strong evidence that a particular type of RNA must be synthesized prior to phage protein synthesis. However, such studies do not define the exact function of this RNA.

The function of this type of RNA, as a component of the template

'More recent results indicate the sedimentation constants of' messenger RXA to range from 8 to 308 (F. Gros, private communication).

VIII. PROTEIX SYNTHESIS AND GENE ACTION 383

actually making protein, was indicated by finding such RXA bound to ribosomes. RXA having the characteristics of messenger RXA has been reported in ribosomes of phage infected cells by Nomura et al. (1960), Gros et al. (1961a); and Brenner et al. (1961). A similar ribosome- associated RNA was demonstrated in normal (not phage-infectedj E. coli by pulse-labeling (Gros et al., 1961a). Evidence for messenger RNA in Pseudomonas aeruginosa, E. coli. and Bacillus megaterium was reported by Hayashi and Spiegelman (1961). There authors used "step- down" cultures, e.g., transferred from a rich to a limited medium, to inhibit ribosomal RNA synthesis and thus permit a more clear-cut labeling of the rapidly synthesized RNA fraction. This fraction was characterized by hybridization to specific DNA (Hall and Spiegelman, 1961). ^inall3^ the elegant experiments of Brenner et al. (1961) have shown that no new ribosomes were synthesized after phage infection, and that the newly synthesized messenger RNA, as well as newly synthesized protein, were associated with old ribosomes (present before phage infection).

The distribution of messenger RNA. in either normal or infected cells, depends on the magnesium ion concentration. Binding to ribosomes occurs in 0.01 M magnesium ion and most of the highly labeled RNA can be dissociated in 10^ M magnesium ion. In addition to this type of binding noted by several authors, Gros et al. (1961a) found a part of the messenger RNA bound to 70S ribosomes in a more stable linkage. Thus, two stages in messenger RNA fvmction may be, first, a dissociable binding to ribosomes, followed by a more stable linkage. The results of Tissieres et al. (1960) suggest that ribosomes containing the messenger bound in stable linkage are active in protein synthesis. The results of Brenner et al. (19611 are in agreement with this conclusion.

Perhaps the earliest complete "expression" of a messenger RNA concept was that of Jacob and Alonod (1961). The concept was based partly on studies mentioned above, but also on studies of induced enzjaue formation. While this latter subject is outside the scope of this review, it is not possible to completely avoid this topic in any discussion of messenger RNA. The experiments of Riley et al. (1960) illustrate the type of evidence which led to the messenger RNA concept. The structural gene (z-") for ^S-galactosidase was introduced into a recipient cell by mating, and within a few minutes the maximal rate of enzyme synthesis of the zygote was attained and continued at a uniform rate. If DNA itself is not the template, then a rapid transfer of information to the enzyme-forming site has taken place. This is similar to the entry of phage DNA into the bacterial cell noted above. However, in addition, the kinetics indicated that a constant number of active templates were pres-

384 RICHARD SCHWKKT AND .lOlIX lUSIIOl'

ent, otherwise the rate of enzyme synthesis would inci'ease with time. Two exphinations arc considei'cd: (/I stahh' teniphites are iorme(l, hut the gene heeouies inoperative after a few minutes of temphite formation; (2) unstable temphites are formed whieh break down and are reformed at some constant rate. Using P'--hibele(l donor bacteria carrying the z* gene, it was found that breakdown of the genome due to P'^" decay, as originally noted by McFall et al. (19581, led to decreases in ^-galacto- sidase synthesis. These experiments appear to support the second possi- bility. The concept of a rapidly formed intei-mediate which is the carrier of information from DNA in the case of induced enzyme formation is supported by many types of evidence. However, that this information is messenger RNA, or RNA at all, has not been shown, in contrast to the case of phage infection. Gros et nl. (1961b) have reported small increases in RNA labeling on induction, which may represent such synthesis. The evidence that an intact gene is required for /?-galactosidase synthesis based on P^- decay cannot be interpreted unequivocally, particularly since McFall (1961) has suggested that this effect results from repression of enzyme synthesis by catabolites which accumulate in slowly growing non-viable cells. Thus, the argument for an unstable template which needs to be re-formed by continuous gene action, via messenger RNA synthesis, rests largely on genetic studies and the kinetics of enzyme formation after introduction of DNA into the cell. Another powerful argument for the messenger RNA hypothesis in induced enzyme syn- thesis is that the mechanism of induction is explained on the same basis, e.g., the template is unstable and needs RNA synthesis continually; in the absence of inducer, this synthesis is repressed. Thus, the inducer acts at the genetic level to derepress messenger RNA synthesis (see Jacob and Monod, 1961, for details of this hypothesis).

In addition to the evidence given above, it has long been known that RNA synthesis is needed for protein synthesis in bacteria. In part, these studies have involved measurement of protein synthesis using purine- and pyrimidine-requiring mutants. If RNA synthesis was limited by the absence of the required base, protein synthesis was depressed (see Berg, 1961, for discussion). The interpretation of this RNA-protein relation- ship is not clear. Suggestions have been made that common intermediates for RNA and protein synthesis (amino acid-nucleotides) are present. However, more indirect relationships due to jioor cell growth (McFall, 1961) may explain this apjiarent coupling of RNA and protein syn- thesis (see Spiegelman et al., 1955, for an interesting summary of this problem). More recently, the use of base analogs has permitted addi- tional insight into the RNA-protein relationshij). In particular, 5- fluorouracil (see below), azaguanine frhantr(nme. 1959). and thiouracil

VIII. PROTEIN SYNTHESIS AND GENE ACTION 385

have been used. Naono and Gros (1960a) and Biissard et al. (1960) have studied the synthesis of ;8-galactosidase and alkaline phospha- tase in E. coli grown in the presence of the analog, 5-fluorouracil. Very soon after the introduction of the analog, the cells began to syn- thesize altered enzymes which could be characterized by their immuno- logical and enzymatic properties (yS-galactosidase) or by thermolability (alkaline phosphatase). The change from the synthesis of normal to the synthesis of altered enzyme was abrupt and apparently complete. These results were explained by the hypothesis of Chantrenne (1959) that certain nucleic acid analogs may be incorporated into the RNA of the growing cell and exert their influence on protein synthesis by altering the specificity of the RNA. Naono and Gros (1960a; see also Gros, 1960) have observed that, within 1 minute of the addition of the analog, the rates of incorporation of tyrosine and proline into protein dropped while the rate of incorporation of arginine increased. An explanation of this was found in changes in the saturation levels of these amino acids in the amino acyl-RNA pool. Gros and Naono (1961), on the other hand, have found additional changes in another RNA fraction. After incubation of bacteria in the presence of C^^-5-fluorouracil, the radioactivity was found exclusively in a fraction with the sedimentation coefficient 12-14S. After phenol extraction of the RNA, the radioactivity sedimented at 8-lOS. The suddenness and completeness of the effect provide good evidence for a rapid synthesis of the RNA fraction involved, and the sedimentation data suggest identity with messenger RNA. Thus, the changes in enzyme synthesis are explained by the synthesis of an altered messenger RNA containing the analog. The problem is somewhat com- plicated by the finding of incorporation into other types of RNA at later time periods (see Chantrenne and Devreux, 1960, for example) with azaguanine, and the variation in reversal of inhibition when guanine is added after azaguanine incubation for various times.

The final topic for discussion in this section is the "instability" of the messenger RNA. Two types of instability, which may be related, can be distinguished. The first concerns the question of whether mes- senger RNA breaks down to the nucleotide level before assimilation into ribosomal RNA, or is used at the polynucleotide level. This question is discussed in detail by Roberts (Chapter VII) and by Volkin (Chapter VI). A few observations will be made here. The RNA fraction under consideration in a particular experiment may vaiy. When, for example, the criterion for messenger RNA, is that of a rapidly labeled RNA fraction, this may contain various types of RNA. Part of this may be RNA destined for ribosome synthesis; another part may be messenger (defined as carrying infomiation for protein sequence determination).

386 UICII.MU) SCinVKKT AND JOHN HlSIIOl'

Secontl, tlie type of exporiiiient may. in fact, (Ictci'iniiic the fate of the RNA. For example, in pliage-infected cells, wlicic miclccjtides are used for synthesis of deoxynucleotides, and ribosomc synthesis is inhibited, the breakdown of messenger RNA may be emphasized over other path- ways. These points are mentioned because of the a])i:)arent disagreement in the literature on this question. In any event, it seems that under some conditions there is extensive breakdown of the labeled RNA fraction. In some cases, the specific RNA fraction is loosely bound to ribosomes, but another part is more firmly incorporated in ribosomes which have been described as "active" 70S (and also probably lOOS) ribosomes (see below). This RNA is not removed by washing in 10 ' .1/ magnesium chloride (Gros et ciL, 19611)) and may be the functionally important form of messenger RNA. Finally, a third form may involve synthetic processes which result in the labeled RNA being changed to ribosomal RNA as regards base composition. In this situation it may be func- tionally inactive.

The important ciuestion with regard to functional stability of mes- senger RNA is related to the mechanism by which it controls protein synthesis. It has been suggested (Gros et al., 1961b) that there is a stoichiometric breakdown of messenger RNA during protein synthesis. This would result if RNA breakdown were directly coupled to jieptide bond synthesis. There is little evidence to sui)port such a mechanism. Particularly, in reticulocytes (Bishop et nl., 1961, and see below) the stability of messenger RNA rules out such a mechanism. Even in bac- teria, there is evidence that a considerable variation in the functional stability of messenger RNA exists. The extreme instability in the case of phage infection and induced enzyme synthesis has been noted. How- ever, even in phage infection, at a later stage the synthesis of messenger RNA is low, while phage synthesis continues (Astrachan and Volkin, 1958).

2. Function of DNA and RNA in Cell-Free Systems

The cell-free system for incorporation of amino acids developed by Lamborg and Zamecnik (1960) was the first ribosomal system from bacteria. This cell-free system was similar in many respects to cell-free systems from animal tissues. A remarkable property of this system, not found in animal systems (see below), was inhibition by DNase. Tissieres et al. (1960) obtained uji to 70% inhibition of amino acid incorporation with DNase. RNase contamination was excluded and polyelectrolytes did not reverse the inhibition, suggesting that removal of DNA was the cause of inhibition. One of the synthetic compounds used in an attempt to restore activity was poly-A, which was ineffective. Inhibition

VIII. PROTEIN SYNTHESIS AND GENE ACTION 387

of cell-free synthesis of /^-galacto.'^idase by DNase was reported by Kameyama and Novelli (1960). Complete inhibition was obtained and, in fact, the activity of the initial jS-galactosidase present was decreased. A thorough investigation of the DNase inhibition (Matthaei and Nirenberg, 1961) showed that decreases in protein synthesis was noted only after 10-15 minutes of incorporation. The amounts of RNase and trypsin contamination of the DNase preparation were estimated and these amounts shown not to inhibit protein synthesis. The products of DNase digestion did not account for the inhibition. Thus, a require- ment for DNA for protein synthesis was demonstrated. The effect of DNA was probably not direct, since complete inhibition (particularly in the initial phases) was not attained. This was attributed to the presence of endogenous messenger RNA which was utilized in the early phases of amino acid incorporation and then destroyed, leading to a complete dependence on DNA at later stages.

Restoration of activity for protein synthesis by RNA (Nirenberg and Matthaei, 1961) as expected was the next step forward. For these studies, the authors used the crude E. coli supernatant containing ribosomes, which had been preincubated with DNase and energy. The use of the energy-generating system resulted in incorporation of C^--amino acids which presumably led to loss of endogenous messenger RNA. This latter point has not been demonstrated, but, in any event, the treatment pro- vided a system which showed almost no amino acid incorporation in the absence of added messenger RNA. Increases in amino acid incorpo- ration by addition of E. coli and yeast ribosomal RNA and by TMV- RNA were observed. The most striking increase was that given by TMV-RNA, which increased incorporation from 42 to 872 counts per minute/mg of ribosomal protein. The RNA-stimulated incorporation was inhibited by puromycin, required energy, and in general showed the properties of the usual amino acid-incorporating systems. Excess soluble RNA was present and did not account for the observed stimulations. Finally, in this paper, stimulation of phenylalanine incorporation by polyuridylic acid was observed (Nirenberg and Matthaei, 1961). These results represent the most significant advance in the understanding of information transfer in protein synthesis in recent years and many new fields are open to investigation as a result. However, the RNA types which gave increased incorporation did not resemble messenger RNA, as defined earlier, and in fact these results did not prove that DNA destruction was involved in the loss of activity.

Tissieres and Hopkins (1961) were able to remove DNA from an E. coli system by precipitation with magnesium chloride. By adding back E. coli DNA ])lus spermidine, some stimulation of amino acid incorpo-

388 RICHARD SCHWEET AND JOHN BISHOP

ration was observed. This stimulation required addition of all four nucleoside triphosphates. RNA with the sedimentation properties of messenger RNA was shown to be synthesized. In addition, these authors isolated an RXA fraction, by washing ribosomes in dilute magnesium chloride, which also stimulated amino acid incorporation. This fraction also had the sedimentation characteristics of messenger RNA. These results then suggest that amino acid incorjioration depends on the DNA- directed synthesis of messenger RNA and also that a reserve of the latter is present in the cell-free system.

Wood and Berg (1962) removed nucleic acids from an E. coli extract by ammonium sulfate and protamine fractionation. Partially purified soluble enzymes and ribosomes, when incubated with the usual com- ponents, including added soluble RNA, showed low amino acid incorpo- ration. DNase inhibited only slightly, and the incorporation appeared similar to the residual incorporation found by others to occur in the presence of DNase, e.g., due to endogenous messenger RNA. T2 phage DNA and purified RNA polymerase added to the system gave a 20-fold increase in incorporation. That the DNA effect was due to synthesis of RNA was shown by a number of experiments. These involved two- stage experiments. In the first stage, RNA synthesis was allowed to proceed, or prevented, for example, by omission of UTP. This reaction was stopped and then protein synthesis measured using the previous reaction mixture as the stimulating component. The results showed that RNA synthesized under the influence of DNA was responsible for in- creased amino acifl incorporation, although more striking increases were found when RNA was continuously generated in the amino acid- incorporating system. Since present evidence indicates that RNA poly- merase synthesized messenger RNA, it is likely that it is this type of RNA which is effective. However, as in other studies in this field, final proof of this hypothesis will require the demonstration that known, specific proteins are synthesized under the influence of a specific DNA or RNA. Such experiments appear feasible and are in progress in several laboratories. An interesting aspect of these studies was the finding that phage DNA was 5 times more active in stimulating incorporation than E. coli DNA, and that heated phage DNA and single-stranded (^X174 phage DNA were inactive. These latter DNA's are efficient primers for the RNA polymerase, and unless the RNA remained bound and unavail- able to the protein-synthesizing system, it is difficult to understand the lack of stimulation. The general concept of the role of DNA as primer in the RNA polymerase reaction is that each strand of DNA produces a complementary strand of RNA, resulting in production of RNA of the same over-all composition as the DNA primer. With single-strand

VIII. PROTEIN SYNTHESIS AND GENE ACTION 389

DNA or synthetic deoxypolymers, the complementary RNA is produced (see Chamberlin and Berg, 1962, for references and recent results). Bautz and Hall (1962) have suggested that in the intact cell only one of the two strands of DNA is active, since the T4-specific RNA they isolated did not have the composition of the DNA. The solution to this question is critical to the understanding of the mechanism of infomia- tion transfer and it is likely that these answers will come from further studies of DNA-directed protein synthesis in cell-free systems.

Studies of the cell-free synthesis of a specific protein, ;8-galactosidase, were reported by Kameyama and Novelli (1960). This cell-free system from E. coli produced a net increase in ^-galactosidase activity. Addi- tion of DNase resulted in complete inhibition of enzyme synthesis, accompanied by an 85% inhibition of amino acid incorporation into total protein (Novelli et al., 1962). After X-irradiation of E. coli syn- thesizing ^-galactosidase constitutively, the combined ribosomal and supernatant fractions were unable to synthesize ^-galactosidase. Addi- tion of DNA isolated from cells constitutive for ^-galactosidase syn- thesis, or from induced cells, repaired the lesion introduced by X-irradia- tion. DNA isolated from non-inducible cells was ineffective, providing a control for non-specific effects of added DNA. In addition, in order for DNA from inducible cells to stimulate ;8-galactosidase synthesis, the DNA must be prepared from preinduced cells. This finding would be consistent with the concept of repressors acting at the gene level (Jacob and ]\Ionod. 1961). However, the ribosomes used must also come from preinduced cells, which suggests perhaps that specific ribosomes are synthesized during preinduction. The results discussed previously have indicated a lack of specificity of the ribosome and, in fact, suggest that all types of proteins can be made in any population of ribosomes de- pending on the messenger RNA provided. Even when all the components of the system for /3-galactosidase synthesis came from preinduced cells, inducer was required for cell-free synthesis of ^-galactosidase (Novelli et al., 1962). This suggested production of repressor in the system under the influence of the i"^ gene (Jacob and ]Monod. 1961). Indeed, when all the constituents came from r cells, (constitutive mutant) inducer was not required and DNA from r cells, not induced, was inhibitoiy. This result was interpreted as indicating a binding of repressor to DNA. These results represent a remarkable advance in the understanding of induced enzyme synthesis. They differ enough from other results that confirmation in other laboratories would be desirable. Also, since there is always a considerable amount of preformed enzyme present at the start of the experiment, the usual enzyme synthesis measured is only a doubling of that already present. However, careful immunological and

390 RICHARD SCHWEET AND JOHN BISHOP

electrophorotic studios by the authors prox'ide good evidence that new enzyme synthesis has occurred in this system.

3. Function of Synthetic Polynucleotides in Cell-Free Systems

The DNase-treatcd, iirciucuhatcd cell-free system from E. coli, described earlier, was stimulated by various types of RNA (Nircnberg and Matthaei, 1961). However, it was not clear that this RNA served an infonnational function, e.g., directed the synthesis of specific proteins by specifying the sequence of amino acids. Using poly-U as the stimu- lating polynucleotide, however, only phenylalanine was incorporated. Evidence was provided that the product was polyphenylalanine. In this artificial system, therefore, the polynucleotide was sj)ecifying which amino acid was incorporated. In a subsequent publication (Nirenberg et al., 1962), it was shown that this system had many of the properties of usual incorporating systems. For example, phenylalanyl-RNA was an intermediate. Other synthetic polynucleotides were then tested in both Nirenberg's and Ochoa's laboratories. The results of these studies have been extremely fruitful, in that only certain polynucleotides stimulated the incorporation of particular amino acids. Using certain assumptions noted below, it was possible to assign particular nucleotide triplets to particular amino acids and thus provide the solution to the code. The findings have been published in some detail; references to these are found in the most recent publications (Basilio et al., 1962; Matthaei et al, 1962).

The primary basis for assignment of code letters rests on the finding that poly-U stimulates only phenylalanine incorporation. In fact, in- corporation of leucine and isoleucine were also stimulated, but to a much smaller extent. When polymers containing two or more nucleo- tides were used, they were assumed to contain random sequences. For example, the U:C polymer containing 5 times as much U as C was considered to contain 5 times as many UUU triplets as UUC, UCU, or CUU triplets. The use of a triplet is indicated as a minimum number for each coding unit by the finding that incorporation of some amino acids was stimulated only by polymers containing three different nucleo- tides. The results of Crick et al. (1961) (see Chapter VI) also indicate a triplet code. The term U2C includes all the possible arrangements of these three letters; thus the sequence is not known for any code letter. The letters in the code word are then decided by comparing the ratio of phenylalanine stimulation to the amino acid in question. For example, using the case given above, the only amino acids (besides phenylalanine) whose incorporation was stimulated by poly-UC were leucine, proline, and serine (Speyer et al., 1962). The ratios were 5, 13 and 4, respec-

VIII. PROTEIN SYNTHESIS AND GENE ACTION 391

tively. The code words assigned then were U2C for leucine and serine, and UCo for proline (calculated ratio is 25). However, proline is also stimulated by a UAC polymer. The phenylalanine/proline ratio here was 29, expected UUU/UC2 in the polymer was 36. Thus, agreement was satisfactory. However, a UCG polymer with the same UUU/UC2 ratio did not stimulate proline incorporation. The results for most of the amino acids do not contain discrepancies of this type. But not many polymers of different types with similar calculated ratios of two trip- lets have been tested. Another example of this type is tryptophan. The code word assigned by Speyer et al. (1962) and Matthaei et al. (1962) is UG2. The ratios of stimulated incorporation by UG and UCG poly- mers agree, but stimulation by UAG was not observed and should have been the same as UCG (Speyer et al, 1962). On the other hand, Matthaei et al., for the same amino acid, but using polymers of different composi- tion, confinii the assignment with UG and UGA, but not with UGC. In any event, a number of amino acids are given the same code letters by both laboratories, although some differences are noted. Certain amino acids, such as glutamic acid, were stimulated only by UAG so this word was assigned although the UUU/UAG ratio did not agree with in- corporation ratios.

Despite these problems, which may be explained in the future, some remarkable conclusions can be made from the results, and also new problems are opened for investigation. For example, some of the earlier codes are eliminated. All of the code words contain uridylic acid (see Chapter VI for code words). The question arises whether this is a real phenomenon or results from the technique used. Since it is likely that UUU is the code for phenylalanine, and poly-A and poly-C do not code for any amino acid, it is clear that polymers containing large amounts of U will give runs of phenylalanine, thus providing a matrix for other amino acids to attach to and yield an insoluble peptide. Both labora- tories have only looked for insoluble peptides. Therefore, code words containing no U may exist, but would yield stimulation into small pep- tides, soluble in trichloroacetic acid. Matthaei et al. (1962) have noted that poly-AG did not stimulate incorporation, but small peptides were not looked for. If this argument is true then even the failure of poly-A or poly-C to stimulate incorporation may not be correct. There are 37 triplets containing U and all the amino acids are coded for by 24 trip- lets. There are two words for asparagine, and threonine, and three words for leucine, according to Speyer et al. (1962). Matthaei et al. (1962) list serine as having two code words. Thus, 12 triplets with U are not accounted for. Either these are "nonsense," or further degen- eracies (more than one code word per amino acid) exist. If triplets

392 KK'II AHI) SCIIWKKT AND .loHN UISHol'

witliout U could also code, tlit- degeneracy must he even greater. On the other hand, if all codes contain U, the composition of DNA and mes- senger KNA does not corres])ond (does not contain sufficient U) and it must he supposed that i)art of these iiolymicleotides does not code, e.g., is nonsense or is used for some other function.

An important impetus in suppoi't of the validity of the code letters has come from studies of amino acid replacements in mutant jiroteins. In particular, the nitrous acid-induced mutants are of interest. They are jiroduccd hy nitrous acid treatment of infective TMV-RNA. After infection and rei)lication, mutants are isolated and the protein structure determined (see Chapter X). It is considered that most if not all of the changes are due to the action of nitrous acid on adenine and cytosine, which yield guanine and uracil, respectively, in the new RNA. The amino acid changes being known, it is possible to compare these with the change in the code words derived from cell-free systems. Tiie data have been compiled from the litei-ature in both laboratories and the agreement is good. Speyer et ol. (1962) record that 11 of 16 replace- ments agree. For example, a common mutation results in threonine

> isoleucine. The codes are UAC > UAU, which is a C to U

change and agrees with the action known for nitrous acid. Important

exceptions are the aspartic acid > serine, and serine > leucine

replacements, which involve either change of two bases or an A > U

change based on UoC for serine. However, if a second code word for serine were UCG, some of these discrepancies would disappear. Matthaei et al. (1962) include UCG as a second code word for serine, but their published data do not lend strong support to this assignment. In general, then, these results support the proposed code words, since the combined effect of the known action of nitrous acid and the proposed code words are so restrictive to the possible amino acid changes that the agreement found must have significance.

This is reinforced by the conclusions reached by Smith (1962) in a study of amino acid replacements in human hemoglobins. The question studied is whether known amino acid changes in the abnormal hemo- globins are compatible with the change of a single base in the synthetic code words. In this case there are no restrictions to the kind of base change involved and in fact almost every type must occur to account for the amino acid changes. Smith predicted the code words for four amino acids based on known ones. For example, glutamic acid had not yet been coded; in fact, the data with synthetic polynucleotides is still weak for this amino acid (Speyer et al., 1962). But since glutamic acid could be replaced !)y valine (UoCi), lysine (UA_,), and glycine (UGo) in various hemoglobins, the code word for glutamic acid must contain

VIII. PROTEIN SYNTHESIS AND GENE ACTION 393

U, A, and G. Tliis is consistent with another licmoglohin where <flutaniic acid is replaced by glutaininc, which is UGC. The code word for gluta- mine, UGC, was predicted from TMV replacement data, but so far the synthetic polynucleotide, UGC, has not stimulated glutainine incorpora- tion. Smith also considered the question of sequences of the bases in any code word, and it seems likely that U may only occupy two of the three possible positions. Further, if the sequence of one code word con- taining three bases is established, most of the others can be deduced. Amino acid replacements in proteins of various species were also com- pared, and here the situation was not as clear. Some of the changes must involve two mutations if the code letters are correct. However, this may not be too unusual. The implications of these comparisons in the evolu- tion of protein structure and function are of great interest.

One of the important conclusions from the agreement in code words derived from amino acid replacements and those derived from stimula- tion of amino acid incorporation is that the code is "universal." That is, in general the code word for an amino acid is the same in mammals, E. coli, and TMV. A similar conclusion was reached from the finding that transfer RXA from various species could be used to synthesize hemoglobin using reticulocyte ribosomes. This conclusion does not exclude differences in composition or specificity in messenger RNA from different species, but would ascribe these to non-coding portions of the molecule. Nonetheless, it is still important to show that poly-U will code only for phenylalanine using mammalian ribosomes," for example, or that a particular messenger RNA directs the synthesis of the same protein in ribosomes of two different species.

This question is also related to the question of the role of the ribo- some in protein synthesis. While ribosomes of E. coli can accept informa- tion from different types of natural and synthetic messengers, only 10% or less of the ribosomes contain labeled poly-U when this is used as messenger. The rest of the labeled poly-U is rapidly degraded, but phenylalanine incorporation continues, implicating these special ribo- somes as the "active" sites of protein synthesis (Matthaei et al., 1962; M. Nirenberg, private communication). Do these "active" ribosomes represent a special class, or are they representative of all the ribosomes of the cell? Before discussing this question, it is of interest to note that these data provide evidence for a catalytic role of the messenger RNA in the active ribosomes. Since one phenylalanine residue is incorporated per uridylic acid residue under favorable conditions, and much of the poly-U is probably never integrated into functional ribosomes, plus the

"This result has now been reported l\v H. Arnstein {Nature, in press) and confirmed in this laboratory.

394

RICHARD SCHWEET AND JOHN BISHOP

fact that tliree U's code I'ur a single amino acid, it is likely that the poly-U messenger acts catalytically, even in the cell-free system.

The data shown in Table V indicate that poly-U or TMV-RNA stim- ulated ribosomes synthesize protein in similar amounts to the original crude extracts. The somewhat lower values are undoubtedly due in part to the manipulations involved in removing DNA and endogenous mes- senger RNA from the artificial systems before assay. If we assume that ribosomes stimulated by poly-U or other added messenger do not incorpo- rate more amino acids i)cr ribosome than in the crude extract, then these

TABLE V

Incorpohatiox of Amino Acids in Cell-P^ree Systems

		Incorpo-	Total"
System	Amino acid	ration"	protein	Reference
TMV-RNA	Valine	0.4	4.5	Nirenherg and Matthaei (1961)
Poly-U	Phenylalanine	6	6	Nirenberg and Matthaei (1961)
T2 phage DNA and	Leucine plus valine	0.7	5	Wood and Berg
RNA polymerase				(1962)
Poly-U	Phenylalanine	24.6	24.6	Lengyel et al. (1961).
Original crude	Individual amino acids	—	50	Tissieres et al.
extract				(1960)
Original crude	Valine	1.0	18	Matthaei and
extract				Nirenberg (1961)
Reticulocj'te	Leucine	1.5	11	Schweet et al.
ribosomes				(1961)

" Values are given as m/umoles of amino acid per mg of ribosomal protein.

data indicate that most of the ribosomes which were making many proteins in the crude extract are now making only polyphenylalanine or a few proteins involved in TMV synthesis. Thus, from the infonnation transfer standpoint there seems to be little ribosome specificity. A point which will be returned to is the fact that reticulocyte ribosomes, which contain a stable messenger RNA, incorporate amounts of amino acids that are similar to the stimulated bacterial cell-free system (Table V). Ribosome specificity has, however, been noted by Novelli et al. (1962). Also, as noted earlier, there is an enzymatic specificity, since transfer enzymes from rabbit reticulocytes and E. coli were not inter- changeable (Nathans et al., 1962). Since it ajipears that transfer and messenger RNA are not species specific, it follows that the enzymatic specificity serves another function, not involving the template (which

VIII. PROTEIN SYNTHESIS AND GENE ACTION 395

is considered to be the ribosome-messenger RNA region with which the transfer RNA interacts). This function could involve intermediates in the ribosome beyond the amino acyl-RNA stage as noted earlier.

B. INFORMATION TRANSFER IN NUCLEATED ORGANISMS

In this section again, the main emphasis will be on studies with cell- free systems. Since the nucleus is the site of the hereditary control of protein specificity in nucleated organisms, we shall concern ourselves here with evidence and speculation regarding the method by which the nucleus exerts its control. The controlling role of the nucleus, long a subject of speculation, was first clearly indicated by the work of Hammerling (1953) with the alga Acetabularia. These experiments, ex- periments with enucleated Amoeba (reviewed by Brachet and Chan- trenne, 1956) and human amnion cells (Goldstein et al., 1960), demon- strate (1) that protein synthesis can continue for extended, but not indefinite periods of time in the absence of the nucleus, (2) that some long-term control, particularly of differentiation (as in the maturation of Acetabularia), is under nuclear control. It is impossible, from such experiments, to decide whether any fraction of protein synthesis is not under long-temi nuclear control. Furthermore, it is not clear at the moment whether protein synthesis in enucleated organisms might not be under the control of cytoplasmic DNA.

1. Synthesis of Ribosomes

Evidence concerning the site of synthesis of the ribosomes is indeed scanty. Much of the RNA of the nucleus appears to be located in ribonucleoprotein particles. In pea seedling nuclei, for example, 20% of the RNA is found in ribosomes (Rho et al., 1961). Furthermore, the ribo- somes of thymus nuclei are active in protein synthesis (Frenster et al., 1960) by pathways which are essentially the same as those followed in cytoplasmic synthesis (reviewed by Allfrey and Mirsky, 1961). It is not clear, however, whether the nuclear ribosomes are in the process of synthesis and are destined to be passed to the cytoplasm, or are engaged in the synthesis of the enzymes and other proteins found within the nucleus.

Using pea seedlings, Ts'o and Sato (1959) have observed that, after short periods of labeling, the RNA of the nuclear ribosomes was 50-100 times more radioactive than the RNA of the cytoplasmic ribosomes. If this is not due to a phenomenally rapid synthesis of ribosomes of nuclear function, it indicates that the ribosomes are synthesized in the nucleus. Schneider (1961) has observed a transfer of radioactivity from the RNA

396 RICHARD SCHWEET AND JOHN BISHOP

of prelabeled (C^^-orotic acidi thymus nuclei to unlabeled cytoplasm upon incubation of the two fractions together. The labeled RXA in the cytoplasm was found associated with microsomal and ribosomal frac- tions. It is clear that intact RXA molecules were transferred from nucleus to cytoplasm, but whether they were incorporated into the ribosomes before or after transfer is not clear. These two shreds of evidence might s^uggest a nuclear origin of cytoplasmic ribosomes.

2. Synthesis of EXA

Early enucleation studies with Acetabularia and Amoeba (Brachet, 1955) indicated that RXA synthesis was impaired in the absence of a nucleus. The more recent work of Prescott (.1960) and Goldstein et al. (1960) indicates that the inhibition of RXA synthesis may be complete in enucleated Amoeba, Acanthamoeba, and human amnion cells. Recent autoradiographic experiments (Perrj' et al., 1961a) are indicative of more specific effects. After inactivation of the nucleolus with an ultraviolet microbeam. incorporation of cytidine into cytoplasmic RXA was severely inhibited, while incorporation into the nucleus (apart from the nucleolus) was inhibited to a lesser extent. In this case. too. however (Periy et al.. 1961b) protein synthesis (as measured by H"-lysine incorporation) was inhibited, although to a lesser extent than RX'A synthesis.

X'uclear transplantation studies in Amoeba (Goldstein and Plant, 1955) showed that a transfer of RX'A from nucleus to cytoplasm oc- curred. The work of Schneider ( 1961) with isolated nuclei is in agreement with this. Kinetic studies of RXA synthesis in HeLa cells ( Feinendegen et al., 1961) and hyphae of Xeuros^p&ra crassa (Zalokar. 1960) indicate that by far the greater part of RXA synthesis in intact cells occurs in the nuclei. The elegant studies of Zalokar (1960) using Xeiiros^pora hyphae centrifuged to separate the different elements of cellular organ- ization, irithaut cell rupture, indicate that the RXA transferred from the nucleus to the cytoplasm becomes associated with the reticulum, pre- smnably with the ribosomes. Thus a nuclear origin of ribosomal RX'A is suggested, but by no means proven. The cytoplasmic synthesis of a metabolically stable ribosomal RX'A would presumably pass undetected in these studies.

Rho and Bonner (1961) have fractionated pea seedling nuclei pre- labeled with RXA precursors, and have concluded that the primary' site of RX'A synthesis is the chromatin. The newly synthesized RX'A was transferred to the nucleolus. Some autoradiographic studies are in agreement, while others are not. Reference to this work will be found in the paper of Rho and Bonner (1961).

The type of RXA synthesized in these studies is not known. In many

VIII. PROTEIN SYNTHESIS AND GENE ACTION 397

of the autoradiographic studies, it is likely that the synthesis involved ribosomal RNA, based on the large amount of synthesis. The finding of a rapidly labeled RNA with the base composition of DNA (Sibatani et al., 1962) in thymus nuclei is evidence that "informational" RNA can be synthesized in the nucleus. The demonstration of an RNA poly- merase enzyme in thymus nuclei (Weiss, 1960) is consistent with this viewpoint. The association of newly synthesized RNA with chromatin DNA (Bonner et al., 1961) using pea embryo preparations is also sug- gestive of this type of RNA, which may also be a precursor of ribosomal RNA. However, it is not yet known whether messenger RNA synthesis occurs exclusively in the nucleus, or also to a small extent in the cyto- plasm. In addition, is there destruction of messenger RNA and replace- ment into old ribosomes from nuclear messenger RNA as in bacteria? Or is the complete package made in the nucleus? The continued synthesis of protein in the absence of a nucleus and of net RNA synthesis (see Goldstein et al., 1960, for references!, does not exclude the presence of a small fraction of RNA with a rapid turnover, presumably synthesized under the direction of cytoplasmic DNA. The alternative is a stable messenger RNA integrated into the ribosome. During the lifetime of a nucleated cell, depending on the stage of differentiation, it is possible that both mechanisms play a role.

The messenger RNA hypothesis (Jacob and ]Monod, 1961) is based on phenomena observed in bacteria, and we have seen that, though no rigorous proof is available, the circumstantial evidence is largely con- sistent with it. Rapid adaptation is a characteristic, and indeed neces- sary, property of bacteria. On the other hand, although adaptation is not unknown in cells of higher organisms, the changes which normally occur are slow and often irreversible. Thus, information turnover might differ in rate and even in basic mechanism from that in bacteria.

3. Studies with Cell-Free Systeins

Studies of information transfer with cell-free systems of higher organisms have been done mainly with the reticulocyte sj^stem. This type of cell represents an extreme degree of differentiation. Measure- ments of P^- and guanosine incorporation into RNA of intact reticulo- cytes showed a turnover of RNA too low to be compatible with messenger RNA fonnation during hemoglobin synthesis (Nathans et al., 1962). Similar results were obtained in this laboratorJ^ For example, after 2 hours of incubation of intact reticulocytes, the incorporation of either P^-- or H^-cytidine into total RNA was between 0.1 to 0.5 m/xmoles per ml of packed cells, while amino acid incorporation was 6.000 to 10.000 m/^moles. (O. Favelukes and R. Schweet, unpublished

398 RICHARD SCinVKKT AND JOHN lUSHOP

data). Tims, in this extivmo case, assuming the ex])eriments ai'e testing RNA turnover adequately, tliere is no evidence for any kind of infoima- tion turnover.

The studies with cell-free systems have attempted to study this question further, and also to change the "information" in reticulocyte rihosomes, even though this might not occur in the intact cell. Lamfrom (1961) incubated the rihosomes of rabbit or sheep reticulocytes with enzyme fractions from rabbit or sheej) supernatants, and concluded that the hemoglobin tj^pe synthesized was partly a function of the species from which the ribosomes were derived. However, some hemoglobin corresponding to the species from which the soluble fraction was derived was also found. It was concluded that the pH 5 precipitate was inactive in this respect, and that the ability to change the specificity of the ribosomes was found in tlie pH 5 supernatant. Kruh et al. (1961) described a rather different situation. Using ribosomes and soluble frac- tions derived from the reticulocytes of the rabbit and the guinea pig, considerable synthesis of the heterologous type of hemoglobin was observed when the ribosomes of one species were incubated with the pH 5 precipitate of the other. Thus, Kruh et al. (1961) concluded that soluble fractions could change the type of hemoglobin synthesized, but the opposite enzyme fraction was effective. Bishop et al. (1961), using reticulocytes of the rabbit and the mouse, incubated the ribosomes of one species with a mixture of the ribosomal supernatants of both species, and found that the hemoglobin synthesized corresponded to the species which provided the ribosomes. It is possible that the use of mixed supernatants resulted in the failure of a potential specificity change to be expressed if greater affinity existed between homologous than be- tween heterologous factors. Recently, Miller and Lamfrom (1962) ex- tended these studies to duck-rabbit mixtures. In this case, hemoglobin corresponding to the ribosome species was synthesized, and in addition radioactivity was found in a protein peak which appeared to be a hyl)rid hemoglobin (composed of chains from both species). Thus, the evidence for the synthesis of hemoglobin different from the ribosome species used is conflicting, and further studies are required to decide whether the addition of supernatant fractions from one species can result in synthesis of the hemoglobin of that species when ribosomes of another species are used. In addition, no evidence that RNA is involved has been presented as yet.

Another approach to the problem of infomiation transfer has been to study whether it is necessary to provide sequence information to the ribosome in order to synthesize hemoglobin in the cell-free system. Both Ki'uh ct (tl. (1961) and Lamfrom (1961) reported that no soluble Ikmho-

VIII. PROTEIN SYNTHESIS AND CiENE ACTION 399

globin was formed when reticulocyte ribosomes were incubated with pH 5 enzymes from livers of various animals. Kruh et al. (1961) suggested that this is a failure to form the peptide chain, while Lamfrom (1961) postulated that species-specific "releasing" enzymes might be involved. Bishop and Schweet (1962) studied this question using washed reticulo- cyte ribosomes and guinea pig liver enzymes. The results with pH 5 enzyme fractions were similar to those noted above; little labeled, soluble hemoglobin was formed. However, this failure did not involve species specificity, but was related to the low incorporation observed with this enzyme fraction. When a soluble fraction from liver purified by am- monium sulfate fractionation was used, soluble, labeled hemoglobin was formed. The soluble fraction used was prepared from a liver supernatant as previously described for reticulocyte enzymes (Allen and Schweet, 19621. The amount of labeled hemoglobin formed with liver enzymes plus reticulocyte ribosomes was similar to the amount formed with reticulocyte enzymes added at a level to give the same total incorporation.

TABLE VI Pr(jperties of Escherichia coli and Reticulocyte Ribosomes"

Reticulocyte Treatment E. coli system system

Ribonuclease (1 Mg/ml) Complete inhibition Complete inhiljition

DNase (5-50 Mg/ml) 55% inhibition No effect

Chloramphenicol (10-3 il/) 91% inhibition 3% inhibition

Puromycin (10"^ M) 95% inhibition 95% inhibition

Add TMV-RXA after DNase 80-fold stimulation No effect

Internal ribosomal ribonuclease Present Absent

" Data summarized from Tissieres et al. (1960); Nathans et al. (1962); Schweet et al. (1961); and unpublished data.

Lack of inhibition by DNase (Table VT) is consistent with the concept that information transfer is not necessary for hemoglobin syn- thesis in the cell-free system as suggested by the lack of RNA synthesis in the intact cell. However, even in the absence of addition of messenger RNA, the activity of the reticulocyte ribosomes is comparable to ''stimu- lated" E. coli ribosomes (Table V). Further efforts along these lines in several laboratories involve attempts to add various types of synthetic polynucleotides or messenger RNA, such as TMV (Table VD. So far our own results with TMV-RNA have been negative. However, a striking stimulation of phenylalanine incorporation has been obtained with poly-U, confirming the report by Arnstein." Further, hemoglobin syn- thesis can be almost completely abolished by pro-incubation in the

400 KICHAHl) >('in\i:KT and ,I()11.\ UlSIlol*

fomploto tjy^Loin or by inciil)ali()n w ilh (•liclalin^ a^fiits, witlumt aricctin^ the poly-U stimulated incorporation. These results (R. Arlinghaus and R. Sclnvcct. unpublished data) indicate that honiOG;lobin and poly])hcnyl- ahmine are .synthesized on separate sites.

It seems likely that a fraction similar in function to messenger HNA is present in reticulocyte ribosomcs (G. Favclukes, li. Hardesty, and R. Schweet. unpul)iished data). Ribosomal RNA from i-cticidocytcs stimulated amino acid incorporation when added to the usual pi-ciiim- bated E. coli .system. Whether this stimulation represents even a small amount of globin synthesis or not, is not known. Our present conce])t, tliercfore, is that the liasic mechanism of information ti'ansfci- in higher organisms is similar to bacteria. However, at least in tlie reticulocyte, the messenger RXA is firmly bound in the ribosome and if it is destroyed and replaced at all, this is a slow process. The synthesis of messenger RNA probably occurs only in the nucleus.

References

Allen, E. H., and Schweot, R. S. (1960). Biochim. el Binphijs. Acta 39, 185.

Allen, E., and Schweet, R. (1962). J. Biol. Chem. 237, 760.

Allen, E., Glas,sman, E., Cordes, E., and Scliweet, R. (1960). ./. Biol. Chrm. 2,35,

1068. Allfrey, V. C, and Minsky, A. E. (1961). In "Protein Bio.^ynthrsi.s" (R. J. C. Harris,

ed.), pp. 67-81. Academic Press, New York. Anfinsen, C. B., and Steinberg, D. (1951). J. Biol. Chcm. 189, 739. Askonas, B., Campbell, P., Godin, C, and Work, T. (1955). Biochem. J. 61, 105. Astrachan, L., and Volkin, E. (1958). Biochim. et Biophys. Acta 29, 536. Basilic, C, Wahba, A., Lengyel, P., Speyer, J., and Oclioa, S. (1962). Proc. Natl.

Acad. Sci. U.S. 48, 613. Bautz, E., and Hall, B. (1962). Proc. Natl. Acad. Sci. U.S. 48, 400. Benzer, S., and Wei.'^blum, B. (1961). Proc. Natl. Acad. Sci. l'.S. 47, 1149. Berg, P. (1956). /. Biol. Chem. 222, 1025. Berg, P. (1961). Ayin. Rev. Biochem. 30, 293. Berg, P., and Lagerkvist, V. (1962). J. Biol. Chem., in press. Berg, P., and Ofengand, E. (1958). Proc. Natl. Acad. Sci. VS. 44, 78. Berg, P., Bergmann, F., Ofengand, E., imd Dieckmann, M. (1961). J. Biol. Chem.

236, 1726. Bergmann, F., Berg, P., and Dieckmann, M. (1961). J. Biol. Chem. 2.36, 1735. Bishop, J., and Schweet, R. (1961a). Federation Proc. 20, 389. Bishop, J., and Schweet, R. (1961b). Biochim. et Biophys. Acta 49, 235. Bishop, J., and Schweet, R. (1961c). Biochim. el Biophys. Acta 54, 617. Bishop, J., and Schweet, R. (1962). Biochim. el Biophys. Ada, in press. Bishop, J., Allen, E., Leahy, J., Morris, A., and Schweet, R. (1960a). Federation

Proc. 19, 347. Bishop, J., Leahy, J., and Schweet, R. (1960b). Proc. Natl. Acad. Sci. U.S. 46, 1030. Bi.shop, J., Favelukes, G., Schweot, R., and Russell, E. (1961). Nature 191, 1365. Bloemendal, H., Bo.sch, L., and Sluysor, M. (1960). Biochim. et Biophys. Acta 41,

454.

VIII. PROTEIN SYNTHESIS AND GENE ACTION 401

Bloemondal, H., Littaucr, I'. Z., and Daniel, V. (1961). Biochim. et Biophys. Acta

51, 66. Bonner, J., Huans, R., and Maheshwari, N. (1961). Proc. Natl. Acad. Sci. U.S. 47,

1548. Borsook, H. (1952). Forf.schr. Chem. org. Naturstofje 9, 202. Borsook, H., Dea^y, C, Haagcn-Smit, A., Keighley, G., and Lowy, P. (1950). J. Biol.

Chem. 187, 839. Bosch, L., Bloemendal, H., and Sluyser, M. (1959). Biochim. et BiopJtys. Acta 34,

272. Bosch, L., Bloemendal, H., and Sluy.ser, M. (1960). Biochim. et Biophys. Acta 41,

444. Brachet, J. (1955). Biochim. et Biophys. Acta 18, 247. Brachet, J., and Chantrenne, H. (1956). Cold Spring Harbor Symposia Quant. Biol.

21, 329. Brenner, S., Jacob, F., and Meselson, M. (1961). Nature 190, 576. Brown, G., Brown, A., and Gordon, J. (1959). Brookhaven Symposia in Biol. 12, 47. Bussard, A., Naono, S., Gros, F., and Monod, J. (1960). Compt. rend. acad. sci. 250,

4049. Campbell, P. N. (1962). Fifth Intern. Congr. Biochem., Moscow, 1962, Vol. 2. In

press. Campbell, P. N. (1961). Biol. Revs. Cambridge Phil. Soc. 35, 413. Campbell, P. N., Greengard, 0., and Kernot, B. A. (1960). Biochem. J. 74, 107. Cantoni, G. L. (1962). Strasbourg Conf. Private communication. Chamberlin, M., and Berg, P. (1962). Proc. Natl. Acad. Sci. U.S. 48, 81. Chantrenne, H. (1959). Biochem. Pharmacol. 1, 233.

Chantrenne, H., and Devreux, S. (1960). Biochim. et Biophys. Acta 41, 239. Cowie, D., Spiegelman, S., Roberts, R., and Duerk.sen, J. (1961). Proc. Natl. Acad.

Sci. U.S. 47, 114. Crick, F. H. C. (1958). Symposia Soc. Exptl. Biol. 12, 138. Crick, F. H. C, Barnett, L., Brenner, S., and Watts-Tobin, R. (1961). Nature 192,

1227. Dalglie-sh, C. E. (1958). Rec. trav. chim. 77, 634. Dintzis, H. M. (1961). Proc. Natl. Acad. Sci. U.S. 47, 247.

Dintzis, H. M., Borsook, H., and Vinograd, J. (1958). In "Microsomal Particles and Protein Synthe.sis" (R. B. Roberts, ed.), pp. 95-99. Pergamon Press, New York. Doctor, B., Apgar, J., and Holley, R. (1961). /. Biol. Chem. 236, 1117. Dunn, D. B., Smith, J. D., and Spahr, P. F. (1960). J. Mol. Biol. 2, 113. Feinendegen, L. E., Bond, V. P., and Painter, R. B. (1961). Exptl. Cell Research

22, 381. Fe.s.senden, J., and Moldave, K. (1961). Biochem. Biophys. Research Communs. 6,

232. Fraenkel-Conrat, H. (1956). J. Am. Chem. Soc. 78, 882. Frenster, A., Allfrey, V. C, and Mirsky, A. E. (1960). Proc. Natl. Acad. Sci. U.S.

46, 432. Gierer, A., and Schramm, G. (1956). Nature 177, 702.

Glassman, E., Allen, E., and Schweet, R. (1958). J. Am. Chem. Soc. 80, 4427. Goldstein, A. (1962). J. Mol. Biol. 4, 121.

Goldstein, A., and Brown, B. J. (1961). Biochim. et Biophys. Acta 53, 438. Goldstein, L., and Plaut, W. (1955). Proc. Natl. Acad. Sci. U.S. 41, 874. Goldstein, L., Micou, J., and Crocker, T. T. (1960). Biochim. et Biophys. Acta 45, 82.

402 RICHARD SCHWEET AND JOHN BISHOP

Gros, F. (1960). In "Tho Xucloic Ac-ids" (E. Cliargaff and J. X. Davidson, cds.),

Vol. Ill, 1)]). 409-452. Atadoniic Press, New York. Gros, F., and Xaono, S. (1961). In "Protoin Bio.syntlie.sis" (R. J. V. Harris, ed.),

ji]'. 195-206. Academic Press, Xcw York. Gros, l-\, Iliatt, H., Gilbert, W., Kurland, C. G., RisebrouKli, R. \\ ■, and Watson,

J. D. (1961a). yaturc 190, 581. Gros, F., Gilbert, W., Hiatt, H., Attardi, G., Spahr, F., and Wat.son, J. (1961b).

Cold Spring Harbor Syinposia Quart. Biol. 26, 111. Grossi. L. G.. and Moldave, K. (1960). J. Biol. Chem. 235, 2370. Hall, H. D.. and Spiofielman, S. (1961). I'roc. Natl. Acad. Sci. U.S. 47, 137. Hiinnncrling. J. (1953). Intern. Rev. Cytol. 2, 475.

Hartmann, G., and Coy, U. (1961). Biochim. el Biophy.^. Acta 47, 612. Hayashi, M., and Spiegelman, S. (1961). Proc. Natl. Acad. Sci. US. 47, 1564. Hecht, L., Stephenson, M., and Zamecnik, P. (1959). Proc. Natl. Acad. Sci. UJi. 45,

505. Herbert, E., and CancUakis, E. (1960). Biochim. et Biophys. Ada 42, 363. Herbert, E., and Canellakis, E. (1961). Biochim. et Biophys. Acta 47, 85. Hirokawa, R., Omori, S., Takahashi, T., and Ogata. K. (1961). Biochim. el Biophys.

Acta 49, 612. Hoagland, M. B. (1955). Biochim. et Biophys. Acta 16, 288. Hoagland, M. B. (1960). In "The Nucleic Acids" (E. Chargaff and J. N. Davidson,

eds.), Vol. Ill, pp. 349-408. Academic Press, XTew York. Hoagland, M. B., and Comly, L. T. (1960). Proc. Natl. Acad. Sci. U.S. 46, 1554. Hoagland, M. B., Zamecnik, P. C., and Stephenson, M. L. (1957). Biochim. et

Biophys. Acta 24, 215. Hoagland, M. B., Stephen.son, M. L., Scott, J. F., Hecht, L. I., and Zamecnik, P. C.

(1958). J. Biol. Chem. 231, 241. HoUey, R. W., and Goldstein, J. (1959). J. Biol. Chem. 234, 1765. Holley, R., and Merrill, S. (1959). J. Am. Chem. Soc. 81, 753. Holloway, B., and Ripley, S. (1952). J. Biol. Chem. 196, 695. Hiilsmann, W. C., and Lipmann, F. (1960). Biochim. el Biophys. Aria 43, 123. Hultin, T. (1950). Exptl. Cell Research 1, 376. Hultin, T. (1956). Expll. Cell Research 11, 222. Hultin, T. (1961). Biochim. et Biophys. Ada 51, 219. Hurwitz, J., Bresler, A., and Diringer, R. (1960). Biochem. Biophys. Research

Communs. 3, 15. Jacob, F., and Monod, J. (1961). ./. Mol. Biol. 3, 318. Jonxis, J. H. P. (1958). Exposes arm. biochim. med. 20, 89. Kameyama, T., and X'ovelli, G. D. (1960). Biochem. Biophys. Research Communs.

2, 393. Kihara, H., Hu, A., and Halvorson, H. (1961). Proc. Natl. Acad. Sci. U.S. 47, 489. Kruh, J., Dreyfus, J., Schapira, G., and Padieu, P. (1957). J. Biol. Chem. 228, 113. Kruh, J., Dreyfus, J., Schapira, G. (1960). J. Biol. Chem. 235, 1075. Kruh, J., Ro.sa, J., Dreyfus, J., and Schapira, G. (1961). Biochim. el Biophys. Ada

49, 509. Lacks, S., and Gros, F. (1959). ./. Mol. Biol. 1, 301.

Lamborg, M., and Zamecnik, P. (1960). Biochim. et Biophys. Ada 42, 206. Lamfrom, H. (1961). J. Mol. Biol. 3, 241.

Leahy, J., Glassman, E., and Schweet, R. (1960). J. Biol. Chem. 235, 3209. Lengyel P., Spcyer. J., and Ochoa, S. (1961). Pruc. Nail. Acad. Sd. U.S. 47, 1936.

VIII. PROTEIN SYNTHESIS AND GENE ACTION 403

Lipmann, F., Hiilsiimnn, W. C, Hartniann, G., Bonian, H., and Acs, (!. (1959). J.

Cellular Comp. Physiol. 54 (Suppl. 1), 75. Loftfield, R. (1960). Fourth Intern. Congr. Biochevi., Vienna 8, 222. McFall, E. (1961). J. Mol. Biol. 3, 219.

McFall, E., Pardee, A. B., and Stent, C. S. (1958). Biochim. et Biophys. Acta 27, 282. Matthaei, J. H., and Nirenberg, M. W. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1580. Matthaei, J. H., Jones, O., Martin, R., and Nirenberg, M. (1962). Proc. Natl. Acad.

Sci. U£. 48, 666. Miller, A., and Lamfrom, H. (1962). Federation Proc. 21, 413. Moldave, K., and Grossi, L. G. (1960). Federation Proc. 19, 351. Morris, A. J., and Dickman, S. (1960). J. Biol. Chem. 235, 1404. Morris, A. J., and Schweet, R. S. (1961). Biochim. et Biophys. Acta 47, 415. Morris, A. J., Favelukes, S., Arlinghaus, R., and Schweet, R. (1962). Biochem.

Biophys. Research Communs. 7, 326. Muir, H. M., Neuberger, A., and Perrone, J. C. (1952). Biochem. J. 52, 87. Naono, S., and Gros, F. (1960a). Compt. rend. acad. sci. 250, 8527. Naono, S., and Gros, F. (1960b). Compt. rend. acad. sci. 250, 3889. Nathans, D., and Hiilsman, W. C. (1960). Federation Proc. 19, 347. Nathans, D., and Lipmann, F. (1960). Biochim. et Biophys. Acta 43, 126. Nathans, D., and Lipmann, F. (1961). Proc. Natl. Acad. Sci. U.S. 47, 497. Nathans, D., von Ehrenstein, G., Monroe, R., and Lipmann, F. (1962). Federation

Proc. 21, 127. Nirenberg, M. W., and Matthaei, J. H. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1588. Nirenberg, M. W., Matthaei, H., and Jones, O. (1962). Proc. Natl. Acad. Sci. U.S.

48, 104. Nomura, M., Hall, B. D., and Spiegelman, S. (1960). J. Mol. Biol. 2, 306. NovelH, G. D., Kameyama, T., and Eisenstadt, J. M. (1962). Cold Spring Harbor

Symposia Quant. Biol. 26, 133. Ogata, K., and Nohara, H. (1957). Biochim. et Biophys. Acta 25, 660. Ogata, K., Hirokawa, R., and Omori, S. (1960). Biochim. et Biophys. Acta 40, 178. Osawa, S. (1960). Biochim. et Biophys. Acta 43, 110.

Perry, R. P., Hell, A., and Errera, M. (1961a). Biochim. et Biophys. Acta 49, 47. Perry, R. P., Hell, A., and Errera, M. (1961b). Biochim. et Biophys. Acta 49, 58. Peters, T. (1957). J. Biol. Chem. 229, 659.

Pizer, L., and Cohen, S. (1961). Biochim. et Biophys. Acta 53, 409. Preiss, J., Berg, P., Ofengand, E., Bergmann, F., and Dieckmann, M. (1959). Proc.

Natl. Acad. Sci. U.S. 45, 319. Prescott, D. (1960). Exptl. Cell Research 19, 29. Raacke, I. D. (1961). In "MetaboHc Pathways" (D. E. Greenberg, ed.). Vol. II, pp.

263-388. Academic Press, New York. Rabinowitz, M., and Fisher, J. (1962). ./. Biol. Chem. 237, 477. Rabinowitz, M., and Olson, M. (1959). J. Biol. Chem. 234, 2085. Rho, J. H., and Bonner, J. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1611. Riley, M., Pardee, A. B., Jacob, F., and Monod, J. (1960). J. Mol. Biol. 2, 216. Roberts, R. (1959). Carnegie Inst. Wash. Yearbook 59, 268. Schapira, G., Dreyfus, J. C, Kruh, J., and Labie, D. (1960). Biochim. et Biophys.

Acta 45, 409. Schneider, J. H. (1961). Biochim. et Biophys. Acta 47, 107. Schweet, R., Bovard, F., Allen, E., and Glassman, E. (1958a). Proc. Natl. Acad. Sci.

U.S. 44, 173.

404 RICHARD SCHWEET AND JOHN BISHOP

Schweet, K., Lamfrom, H.. and Allen, K. (1958b). Proc. Natl Acad. Sci. UJi. 44,

1029. Sohwoot, R. S., Bishop, J., and Moni.s A. (1961). Lab. Jiivc'^l. 10, 992. Shimura. K., Fukai. H., Sato, J., and Saoki, R. (1956). J. Biochcm. (Tokyo) 43, 101. Sibatani, A., do Kloet, S., AUfrcy, V., and Mirsky, A. (1962). Pruc. Natl. Acad. Sci.

U£. 48, 471. Siekevitz, P. (1952). J. Biol. Chem. 195, 549.

Siekevitz, P., and Palade, G. (1960). J. Biophys. Biochem. Cylol. 7, 619. Simkin, J. (1958). Biochcm. J. 70, 305.

Singer, M., and Cantoni, G. (1960). Biochim. et Biophys. Acta 39, 182. Smith, E. (1962). Proc. Natl. Acad. Sci. UJi. 48, 677.

Smith, K., Cordes, E., and Schweet, R. (1959). Biochim. el Biophys. Acta 33, 286. Speyer, J., Lengyel, P., Basilio, C., and Ochoa, S. (1962). Proc. Natl. Acad. Sci. U.S.

48, 441. Spiegelman, S., Halvorson, H., and Ben-Ishai, R. (1955). hi "Amino Acid Metabo- lism" (W. McEIroy and B. Glass, eds.), p. 124. Johns Hopkins Press, Baltimore. Steinberg, D., Vaughan, M., and Anfinsen, C. B. (1956). Scierwe 124, 389. Stephenson, M., and Zamecnik, P. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1627. Stevens, A. (1960). Biochcm. BiopJtys. Research Commnns. 3, 92. Stulbcrg, M. P., and Novelli, G. D. (1962). In "The Enzymes" (P. Boyer and

H. Lardy, eds.), Vol. VI. Academic Press, New York. In press. Suzuka, I., and Shimura, K. (1960). J. Biochem,. (Tokyo) 47, 555. Takanami, M. (1961). Biochim. et Biophys. Acta 51, 85. Tissieres, A., and Hopkins, J. (1961). Proc. Natl. Acad. Sci. V.S. 47, 2015. Tissieres, A., Schlessinger, D., and Gros, F. (1960). Proc. Natl. Acad. Sci. U£. 46,

1450. T'so, P., and Sato, C. (1959). Exptl. Cell Research 17, 227. von der Decken, A., and Hultin, T. (1958). Exptl. Cell Research 15, 254. von der Decken, A., and Hultin, T. (1960). Biochim. et Biophys. Acta 45, 139. von Ehrenstein, G., and Lipmann, F. (1961). Proc. Natl. Acad. Sci. U.S. 47, 941. Webster, G. C. (1960). Federation Proc. 19, 345. Webster, G. C. (1961). Federation Proc. 20, 387. Weiss, S. B. (1960). Proc. Natl. Acad. Sci. U.S. 46, 1020.

Weiss, S. B., and Nakamoto, T. (1961). Proc. Natl. Acad. Sci. U.S. 47, 694. Wood, W., and Berg, P. (1962). Proc. Natl. Acad. Sci. U.S. 48, 94. Yarmolinsky, M. B.. and de la Haba, G. L. (1959). Proc. Natl. Acad. Sci. U.S. 45,

1721. Yoshida, A., and Tobita, T. (1960). Biochim. et Biophys. Acta 37, 513. Zachau, H., Acs, G., and Lipmann, F. (1958). Proc. Natl. Acad. Sci. U£. 44, 885. Zalokar, M. (1960). Exptl. Cell Research 19, 114. Zamecnik, P. C. (1960). Harvey Lectures 54, 256. Zamecnik, P., and Keller, E. (1954). J. Biol. Chem. 209, 337. Zillig, W., Schacht.schabol, D., and Krone. W. (1960). Z. phy.'^iol. Chem. Hoppe-

Seyler's 318, 100.

Chapter IX

Correlations between Genetics and Chemistry of Human Hemoglobins

CORRADO BAGLIONI^

Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts

I. Introduction 405

II. Tlie Separation of Human Hemoglobins by Chromatography and

Electrophoresis 406

III. Normal Hemoglobins 409

A. Hb-A 409

B. Hb-F 415

C. Hb-A, 418

D. Considerations on the Amino Acid Sequence of the Human

Hemoglobin Chains 419

E. Hb-Aa 420

F. Embryonic Hemoglobin 421

IV. Abnormal Hemoglobins 421

A. Hb-A Variants— Chemical Studies 423

B. Formal Genetics of the Hb-A Variants 436

C. Genetic and Biochemical Relationships 438

D. Dissociation and Recombination of Hemoglobin 440

E. Hb-F Variants 442

F. Hb-A, Variants 444

G. The Problem of Thalassemia 449

H. Abnormal Hemoglobins from Altered Aggregation of

Peptide Chains 453

Y. The Control of the Synthesis of Hemoglobin 456

A. The Control of the Rate of Synthesis of the Peptide Chains . 456

B. The Control of the Synthesis of Fetal and Adult Hemoglobin . . 459

C. Hereditary Persistence of Hb-F 462

D. Models of Red Cell Differentiation and Synthesis of Hb-F ... 466

Acknowledgments 469

References 469

I. Introduction

Substantial advances in our knowledge of the genetics and chemistiy of the human hemoglobins have been made recently and the attention

^Present address: International Laboratory for Genetics and Biophysics, Naples, Italy.

405

406 CORRADO BAGLIONI

of several inve;<ti<2;ators has hccii (li'ixcii to this field of rcscaich. A louj^h estimate of the iiuinhcr of papcis puhlishcd within the last ten years on this subject shows this number to be close to a thousand. It is impossible to review and give reference to all these ])apers for reasons of conciseness and limitation of space. The necessaiy omissions sliould not imply lack of appreciation of the jiapers not cited; most of these references have not been included because they are reported in previously i)ublished reviews.

The rapid progress in the past few years has stimulated many authors to cover this field of research with excellent and specialized reviews. Among the many good reviews may be mentioned Itano's (1957) review- on human hemoglobins, Lehmann's (I960) review on abnormal hemo- globins, Ingram's (1961b) book ''Hemoglobin and Its Abnormalities," and Rueknagel and Neel's (1961) review on hemoglobinopathies. Gratzer and Allison (1960) have recently reviewed the related field of the animal hemoglobins. ]\Iost of these reviews report the historical developments of the research in the particular field covered, so that no attempt will be made to present a historical background in this review.

The purpose of this review is to give the reader up-to-date informa- tion on the latest developments of the genetic and chemical studies on the human hemoglobins and to correlate the work of people active in widely different disciplines: human genetics, biochemistry, hematology, and biophysics.

II. The Separation of Human Hemoglobins by Chromatography and Electrophoresis

Hemoglobin is a conjugated protein, consisting of a protein part (globin) and of a prosthetic group (heme). Mammalian hemoglobins are rather large molecules, made up of approximately 600 amino acids and with a molecular weight of 66,000 dz 2,000. Hemoglobin is contained in specialized cells, the red cells, which are nucleated in lower vertebrates and anucleated in mammals. Red cells are easily obtained by bleeding and hemoglobin solutions are usually prepared by lysis of the red cells with hypotonic solutions, after washing away the serum proteins with isotonic solutions. Over 95% of the soluble protein obtained in this way from mammalian red cells is hemoglobin. This remarkable circumstance and the ready availability of large quantities of red cells have made hemoglobin one of the preferred objects of study for the protein chemist for almost a century.

Hemoglobins obtained from different animal species have been analyzed by several investigators; the heme has been found to be

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 407

identical in all hemoglobins (protoporphyrin IX + Fe), while tlie globin varies considerably in physicochemical properties and in amino acid composition in different species. In many cases more than one hemo- globin has been demonstrated in individuals of the same species by the electrophoretic and chromatographic methods devised for protein separa- tion. That heterogeneity of hemoglobin arises as a consequence of the oxidation or reduction of the heme, can usually be excluded by converting the heme to the carbomonoxy or cyanmet derivative, both of which have the same electrophoretic charge as oxyhemoglobin. Consequently, the hemoglobin found in individuals and in species with multiple hemoglobins differ somewhat in the protein part of the molecule.

Pauling et al. discovered in 1949 the difference in electrophoretic mobility between normal human hemoglobin and hemoglobin obtained from sickle-cell anemic patients. This observation has been of funda- mental importance in the development of the concept of gene-protein relationship and has stimulated a number of workers to investigate human hemoglobins. Several abnormal hemoglobins have been discovered by the same method of electrophoretic analysis. The original Tiselius moving boundary electrophoresis technique was found, however, to be an expensive and relatively impractical method, because of the amount of work involved in a single determination and because of the expensive equipment. Simpler methods of analysis (such as paper or starch-gel electrophoresis) have been developed for the electrophoretic analysis of the human hemoglobins. These methods have been applied to the study of ''normal" human hemoglobin and of the human hemoglobin variants.

"Normal" human hemoglobin is not homogeneous in free boundary electrophoresis (Hoch, 1950). Kunkel and Wallenius (1955) isolated a major and two minor components by using starch block electrophoresis. This method of electrophoretic analysis allows the elution from the starch slab of the separated hemoglobins and the quantitative evalua- tion of each component.

The more sensitive starch-gel electrophoresis (Smithies, 1959) has now become more widely used in order to reveal minor components; the buffer system described by Poulik (1957) exhibits considerable sharpen- ing of the bands and is greatly advantageous. The sliced starch-gel is stained with the benzidine or with the o-dianisidine reagent (Smithies. 1959), which are catalytically oxidized by hemoproteins in the presence of hydrogen peroxide. The sensitivity of the method is considerably enhanced and components present in quantities of less than 1% of the total hemoglobin can be detected. The use of filter paper electrophoresis is confined to the screening of major abnormal components. Adsorption

408 CORRADO BAGLIONI

on tlie liltcr paper and tailing of the hemoglobins may j^reohide the observation of minor components ha\ing an clectrophoretic mobility lower than the major components. Filt(M- i)aper electrojihoresis is still the simplest and fastest method of analysis for the detection of major components if a lai'ge number of samples have to be analyzed. The clectrophoretic analysis on agar gels has also been ap])lied to the separa- tion of hemoglobins (see review by Gratzer and Beaven, 1961).

Chromatographic methods of analysis of human hemoglobin on ion exchangers have also been developed; these methods offer better resolu- tion and quantitative isolation of the separated components. The pres- ence of minor components in samples of human hemoglobin has been confirmed by several investigators, using different analytical techniques for protein separation. Huisman and JMeyering (1960) have standardized the chromatographic separation on caiboxymethylcellulose, described by Peterson and Sober (1956), for the qualitative analysis of human hemo- globin and for the preparative isolation of the components.

Morrison and Cook (1955) demonstrated by chromatography on the resin IRC-50 that normal adult hemoglobin is heterogeneous; Prins and Huisman (1956) also detected heterogeneity of human hemoglobin in the same way. The chromatographic method of analysis on IRC-50 has been greatly developed and improved by Allen et al. (1958) and subse- quently by Clegg and Schroeder (1959) and by Schnek and Schroeder (1961); several minor components are separated by this method of analysis.

The separation of human hemoglobins by starch block electrophoresis and by column chromatography has considerably complicated the prob- lem of the identification of components isolated by different techniques and the problem of nomenclature.

The original system of nomenclature proposed by Kunkel and Wal- lenius (1955) has been the most widely accepted. The major component isolated by starch block electrophoresis has been designated Ai by these authors; it represents approximately 90% of the total hemoglobin. The most slowly moving hemoglobin component in electrophoresis has been designated A2 and the most rapidly moving A.j; they represent 2-3% and 4-12% of the total hemoglobin, respectively. It has become cus- tomary to indicate the major component as hemoglobin A or in shortened form as Hb-A. However, the two minor components are still indicated as Hb-Az and Hb-A^. Keys to the identification of the minor components isolated by starch block electrophoresis with the components isolated by column chromatography are found in the articles of Schnek and Schroeder (1961) and of Huisman and Meyering (19601.

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 409

III. Normal Hemoglobins

A. Hb-A

Hb-A is the normal inajoi- c'oin]ioiU'nt of henioglohiii prepared from the red cells of adult individuals. Hb-A is usually defined by its electro- phoretic mobility, the isoelecti'ic i)oint being at pH 6.87, and by its rate of alkali and acid denaturation. Most of the i)hysical and chemical properties established by the protein chemists in the past for human hemoglobin can be referred to Hb-A. Rhinesmith et al. (1957a) deter- mined the N-terminal residues of Hb-A; four valine residues were ob- tained indicating the presence of four jieptide chains. Rhinesmith et al. (1957b) next found that two peptide chains on acid hydrolysis release quickly the N-tcrminal dipeptide valijl-leucine and two chains release valine only more slowly; these peptide chains have been designated a chains and p chains, respectively. The N-terminal sequence of the latter has been found to be Val.His.Leu. (Rhinesmith et al., 1958).-

The presence of two types of peptide chains in Hb-A, demonstrated by Rhinesmith et al. (1958), has been of particular importance in solving some of the genetic and chemical problems presented by the abnormal hemoglobins. The idea of a molecule composed of two pairs of identical chains agreed particularly well with the X-ray crystallographic picture of the horse hemoglobin molecule, obtained by Perutz and his collab- orators (see Perutz et al., 1960). The horse hemoglobin molecule has the shape of an ellipsoid composed of two identical parts. In each half- molecule there are two different peptide chains. This type of structure is probably common to all mammalian hemoglobins. A rather schematic representation of the hemoglobin molecule is shown in Fig. 1.

Wilson and Smith (1959) obtained pure preparations of the two peptide chains of horse hemoglobin by ion exchange chromatography in a urea gradient and by preparative boundary electrophoresis. Ingram (1959a) applied the same methods to the preparation of the peptide chains of human adult hemoglobin; two fractions were separated by free boundary electrophoresis and l)y column chromatography.

The fractions isolated by column chromatography were examined by "fingerprinting." In the fingerprinting analysis a mixture of peptides, obtained by tiyptic hydrolysis of denatured hemoglobin, is separated by paper ionophoresis, followed by paper chromatography in a perpen- dicular direction. The peptides are spread on paper sheets in character-

■ Amino acids are indicated by the first tliree letters of their names. Asparagine and gkitamine are indicated as Asp-NH2 and Gki-NH-;, respectively, in the text, but as Asn and Gin in the illustrations. Isoleucine is indicated as Ilou.

410

CORR.\DO BAGLIONI

Fig. 1. Schematic representation of the arrangement of the peptide chains in the hemoglobin molecule. (From Ingram and Stretton. 1959b; reproduced by permission of the editor of Xature.)

istic patterns and the comparison of patterns of peptides — fingerprints — allows the detection of minor differences in the composition of the proteins under analysis (see Fig. 2, A and B).

The fingerprinting method has been modified by several authors, who have adapted the same principle, i.e.. the two-dimensional separation by ionophoresis and chromatography, to different tj'pes of ionophoresis apparatuses or to different chromatographic systems. These methods of analysis yield comparable results; in the author's experience the original (Ingram's! method (1958) is to be recommended for its simplicity and practicality. The use of a solvent system different from the one orig- inally employed by Ingram ('19581 has been found advantageous (Bag- lioni. 19611.

The fingerprints of the isolated chains of Hb-A showed different patterns of peptides, which taken together made up the pattern given by whole hemoglobin (Ingram, 1958, 1959a ». The fractions were then identified with the two t>'pes of peptide chains described by Rhinesmith et al. ("1958). by determining their X-terminal sequence (Ingram. 1959a I. A great deal about the primary" sequence of Hb-A has been learned recently and the amino acid sequence of the two chains of Hb-A is now completely elucidated.

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 41 1

^ CO

I —

U4

* # n

V X

it, .

X

•- •

412

CORHADO IlAtll.loNI

I

®

€

0

t ^ W

CO

K .3^ -r

^ fo —

S t ~

2 c w

"^ - .2

5 ■- c

o 5 'C

i « o - --2 O

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY

413

It is beyond the purpose of this review to give a full account of the veiy specialized chemical methods that have been used to study the amino acid sequence of the hemoglobin chains. The peptide chains have been separated by column chromatography, according to Wilson and Smith (1959), or by countercurrent distribution (Hill and Craig, 1959). Hydrolysis by means of proteolytic enzymes has been used to degrade the peptide chains to small peptides. Trypsin was the first choice because tiypsin specifically cleaves the peptide bonds between the carboxyl group of lysine or arginine and the amino group of other amino acids. The separation and purification of the tiyptic peptides (peptides obtained

TABLE I

Tryi'tic Peptides of the a and /3 Chains of Hb-A

	a chain peptides		/3 chain peptides
Peptide	Peptide	Peptide	Peptide
number"	position''	number"	position''
3	aTp IX	4	^Tpl
9	aTp VIII, IX	5	/3Tp XIII
10	aTp VI	6	^Tp IX
11	aTp I, II	12	/3Tp II
13	aTp V	14	/3TpIV
15a	aTp III	15^	/3TpXV
16a	aTp XIV	16/3	/3Tp XIV
17a	aTp II	17/3	^Tp XIV, XV
18	aTpX	19	/3Tp VI
20a	aTp VII	20^	^Tp VII
21a	aTp VII, VIII	21^	/3Tp VII, VIII
22a	aTp VIII	22^	^Tp VIII
23	aTp IV	24	/3Tp V oxidized''
c	aTp I	25	/3TpV
		26	iSTp III
		a	^Tp II, III
		b	/3Tp VIII, IX

" Peptide numbers refer to Fig. 2B, according to Ingram (1958). Two peptides are present in each of the spots 16, 20, and 21, one from the a chain and one from the /3 chain. Peptide 17/3 overlaps with peptide 15a in Fig. 2.

* The tryptic peptides are numl)ered consecutively from the N-terminus in each chain (Gerald and Ingram, 1961). Peptide 22 happens to be free Ij^sine. The identification of the tr.yptic peptides and the correspondence with the amino acid sequence of the peptide chains (Braunitzer et al., 1961b; Konigsberg et al., 1961) are based on the amino acid composition and/or sequence of the peptides (quoted by Baglioni, 1961). Preliminary evidence exists for the identification of peptides a, b, and c; analysis of peptide e (Fig. 2) gives only b'sine (C. Baglioni, unpublished). Peptide e may possibly be lysyl-lysine.

" Peptide 24 is identical with peptide 25, except that the methionine residue in this peptide has become oxidized.

414

CORRADO BAGLIONI

J3

[f hi

hi J.

<x	o
qI	qI
CO	c f—
•	• -J

> »—

-J _j

^ ^

_J	<I
<3	c3
«£	CL
hi	«
q1	qI

>	_j
•«i	<i
c5	CO
>	>

-r	o	bX)	, '	<D
	c	a	y,	_C
w	3	"rt	(U	"o
v«	c	Vj	t3
o		n		o
73	H		a	r1
-	6	o
^	to	"~		C)
^		c^
~	^	x		c S3
<1>	w	M4	3
c	a	o	c	a
			to
«a.	o	f=^
_	■/.		T3 CJ	S
c	o	'^.	"tw	^
	O	ZT	a	c
<3				r;
	»-H	n	^	o
a;			a	K*
-tj	^	§	Q)	n

	M	0;	'o	o
(U	OJ	^	-.J
^	n	• »-«		r,
	■jj bC	<U	O a	a
rt	O	a to g	S 03 O
o	,—1	T!	H	-n
c	o	•^		o
•^	C5	o		o
^^	*^	m	03	^
	^		fl	T-l
	»^	>.	c3 J5	O	'■'''**■
<u	c	-O		to a
H
CO	tsi	CJ		CO 3	■a
u	fl	<D
		^		s	'5

C s a i: o

QO-

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 415

by enzymatic digestion with tiypsin) is a troublesome problem and present several difficulties. Liebold and Braunitzer (1959) have used column chromatography, followed by paper ionopiioresis and/or paper chromatography. Hill and Konigsberg (1960) have used countercurrent distribution, followed by column chromatography and paper ionoj)horesis.

The composition of the tiyptic peptides has been determined by Braunitzer et al. (1960) and by Hill and Konigsberg (1960). The hemo- globin chains were digested by means of proteolytic enzymes character- ized with different specificities, like chymotrypsin or pepsin, so that overlapping peptides were obtained; i.e., peptides containing either two or more tryptic peptides or parts of two tryptic peptides and bridging them together, so that their sequence could be determined (Braunitzer et al., 1961; Hill and Konigsberg, 1961). The position along the peptide chains of Hb-A of the peptides separated by fingerprinting is indicated in Table I.

By the conventional methods of the protein chemists, the partial sequences obtained have been filled with all the amino acids in definite order. Conclusive results have been published for the a chain of Hb-A by Konigsberg et al. (1961), for the a and /? chain by Braunitzer et al. (1961b), and for part of the /? chain by Goldstein et al. (1961). Many years of work of several investigators are summarized in these few lines. This work should be fully appreciated; it has required the constant application of the ingenuity and skill of several experienced investigators and has been concluded in a surprisingly short time. The two sequences of the a chain, which have been independently determined, show complete agreement. The amino acid sequences of the Hb-A peptide chains are shown in Fig. 3.

B. Hb-F

The major component of hemoglobin prepared from fetal blood is fetal hemoglobin, indicated as Hb-F. Hb-F usually accounts for 70-80% of the total hemoglobin of a full-term infant (Jonxis, 1959) ; adult hemoglobin appears in the red cells of fetuses at about the thirteenth week of fetal life (Walker and Turnbull, 1955) and increases in relative proportion until it completely replaces Hb-F by the end of the first year of postnatal life. Hb-F is, however, observed in adults affected by acquired or hereditary anemias, like thalassemia major or sickle-cell anemia. High levels of Hb-F have also been found in some healthy Negro adults (Edington and Lehmann, 1955) ; this condition has been designated "hereditary persistence of Hb-F" (Jacob and Raper, 1958).

Hb-F is characterized by an electrophoretic mobility only slightly different from that of Hb-A. Hb-F is usually characterized and meas-

416 CORRADO BAGLIONI

urcd l)v alkali (K'liaturatioii. Tlir hiji;li(M- resistance to alkali is in fact the outstaniling characteristic of Hb-F. The reaction which henioglol)in undergoes in alkali is essentially the change of hemoglobin to henio- chromogen (Drabkin and Austin, 1935) ; the rate of this change can be followed quantitatively by spectrophotometry. The methods for the identification and evaluation of Hb-F have been reviewed by White and Beaven (1959).

The end group analysis of the Hb-F peptide chains has shown the jiresence of two pairs of peptide chains with the N-terminal sequence Val. Leu. and Gly.His.Phe., respectively (Schroeder and Matsuda, 1958). The peptide chains Gly.His.Phe., which were found to be different from the p chains of Hb-A on the basis of their N-terminal sequence, were designated y chains.

Hunt (1959) has isolated l)y column chromatograi)liy the peptide chains of Hb-F and has examined the pattern of pc])tides after tryptic hydrolysis. The a chains isolated from Hb-A or Hb-F gave identical fingerprints; the fingerprints of the /? and y chains were found to be different. The chemical analysis of the N-terminal sequences of Hb-A and Hb-F and the fingerprinting analysis support the claim that these hemoglobins have a pair of peptide chains in common, the a chains, which are associated with pairs of different chains, /? chains in Hb-A and y chains in Hb-F. A sequence of the a chain of Hb-F has been reported by Schroeder et al. (1961) ; this sequence is in agreement with the sequence of the a chain of Hb-A reported by Konigsberg et al. (1961) and by Braunitzer et nl. (1961b). Schroeder et al. (1961, 1962) have determined the sequence of the y chain of Hb-F. This sequence is reported in Fig. 4 together with the sequence of the /? chain of Hb-A. Genetic evidence showing that the a chains of Hb-F and the a chains of Hb-A are under the control of the same gene has been accumulating in the past few years (Baglioni et al, 1961; IVIinnich et al., 1962), so that there is little doubt today that these peptide chains are chemically identical (see Section IV,E).

Since Hb-A and Hb-F have in common the same a chains, the higher alkali resistance of Hb-F has been attributed to a property of the y chains. The denaturation of Hb-F at alkaline pH's follows, however, a first-order kinetics, with no break in the linear plot corresponding to the presence of a resistant portion of the molecule (Charlwood et al., 1960). Hb-A is known to dissociate at pH 11.0 into symmetric subunits (see Section IV,D) of molecular weight approximately half that of hemoglobin ('Hasserodt and Vinograd, 1959). Al)ove pH 11.0. Hb-A dis- sociates into subunits of smaller molecular weight, pi'esuniably into

IX. GENETICS AND HUMAN HEMO(iLOBlN CHEMISTRY

417

■5 y ^ S^

r*i ?^_ V r! S

■<	1—
to	o
CD	c3
O	t5

<i	o
=s	=5
<i	<I
>	>
_j	_l
o	c3
h^	H^
_j	_J

_J	_j
c3	9
c5	to
qI	c3
(-^	1—
_j	CL.
3:	±

CO	CO
cL	dl
to	to

^	S
t5	to
<c	<I
t)	o
Q-	Q.
_J	_J
>	:>
-J	_i
<i	«i
..	„
:r	x:
to	to
_j	1
_j	_j
>	>
_j	_j

■>	;
X	^
CL.	of
to	to
J	^
to	to
q1	CL.
H^	h^
o	2
Q.	Q.

s	x	s c3	a
<^	3	Is	S
n	H^	«	p
w		fl		<*-!
		<1)		n
o	_^	•■=<	cS	n
d	;:;	p	O	o
	-	■r.	g

2 S.

^ H

Qi S ^- '3 I

="	~~	O ■r.	o	6
n
	^	c	u,	tf
0)		O
d	N
a>	."t;	.■:i	OJ
C	—	■r:	—	.:::

to ," O M c - - G C -- O

3 3 •= _o ^

^ ^

c3 i^H

o J: X G

C CO — O <3-

fa 5--J £ o =

*— I C ^ f;; - u

^ --^ o fc c

r^ «J w 73 I— (

CI	K
G
cS	J
O	c
H	-f
"<i<

01 J^

418 CORRADO UA(;M()N1

single chains; the changes in the 1I1)-A niolccult's exposed to ])I1 11.6 are not reversible and extensive denaturation occurs (Hasserodt and Vinograd. 1959). The dissociation of Hb-F into half-molecules occurs at higher pH's compared to Hb-A; the dissociation is completed at pH 11.6 and the Hb-F retains at this pH the spectroscopic characteristics of undcnaturcd hemoglobin for some days (Charhvood et al., 1960). Since denaturation of Hb-A occurs at pH's where the molecule appears to be dissociated into single peptide chains, it may be supposed that Hb-F is more resistant to alkali than Hb-A, because of the di.'ssociation equi- librium of Hl)-F into single chains being shifted toward higher pH's compared to Hb-A.

The nature of the chemical bonds involved in stabilizing the hemo- globin molecules is not known; it has been suggested that histidines, lysines, and tyrosines provide the active groups involved (Hasserodt and Vinograd, 1959). The different alkali resistance of Hb-A and Hb-F may be due to different chemical bonds established between the peptide chains of these hemoglobins. The higher alkali resistance of Hb-F may not be peculiar of the y chains, but may rather result from the chemical bonds established between the a and y cliains in Hb-F being different from those established between the a and [3 chains in Hb-A. A number of factors are involved in determining the higher alkali resistance of Hb-F; aging of the Hb-F, oxidation of the heme group (Matsuda et al., 1960), and exposure to pH 5.0 (Huisman, 1961) are effective in lowering the alkali resistance of Hb-F. Changes of the Hb-F molecule presumably occur under these conditions, which alter the bonds involved in stabi- lizing the quaternary structure of the molecule.

C. Hb-Ao

This minor component was first described and isolated by Kunkcl and Wallenius (1955) ; it represents approximately 2.5% of the hemo- globin of normal adults. Hb-Ao is easily separated from Hb-A since it is characterized by an electrophoretic mobility at pH 8.6 considerably lower than that of Hb-A. Hb-A and Hb-Ao have identical molecular weights and UV spectra (Kunkel and Beam, 1957) and very similar amino acid composition (Ingram and Stretton, 1961). Hb-Ao is not a chemical derivative of Hb-A (Kunkcl and Beam, 1957).

Kunkel et al. (1957) reported that the Hb-A^ level is approximately doubled in thalassemia minor, the heterozygous, nonsymptomatic fonn of the hereditary' anemia known as thalassemia major or Cooley's ane- mia. This observation has been confirmed in a larger number of individuals by other investigators (Ceppellini, 1959a; Silverstroni et al., 1957).

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 419

Ingram and Stretton (1959a) investigated the chemical nature of Hb-Ao by fingerprinting and determined the sequence of a few peptides of this hemoglobin. Few differences were observed in the fingerprint of Hb-A;., when comjiared to the fingerprint of Hb-A. The differences were localized to /? chain peptides. Muller and Jonxis (1960) and Ingram and Stretton (1961) separated the peptide chains of Hb-A^ and fingerprinted the isolated chains. Hb-A:.. appears to have a pair of chains identical to tiie a chains of Hb-A; peptide chains different from the /? chains of Hb-A are associated with the a chains in Hb-Ao ; these peptide chains have been designated 8 chains. Ingram and Stretton (1961, 1962) isolated several of the peptides of the S chain of Hb-Ao and determined their amino acid composition and some of their amino acid sequences. The S chains of Hb-Ao have been found to be surprisingly similar to the (3 chains of Hb-A. Only 4 residues out of 146 have been shown to be definitely different between these two chains (Ingram and Stretton, 1961) and there is preliminary evidence for at least 4 further changes (A. 0. W. Stretton, personal communication).

D. CONSIDER.\TIONS ON THE AMINO ACID SEQUENCE OF THE HUMAN HEMOGLOBIN CHAINS

In Figs. 3 and 4 the amino acid sequences of the a, (3, and y chains are shown together in order to point out the similarities existing between these peptide chains. In the case of the /? and y chains the similarities are striking. If one calculates the relative number of amino acids in corresponding positions along these peptide chains which are found to be identical, one can observe 42% correspondence between the a and the /? chains, 39% correspondence between the a and the y chains, and 71% correspondence between the (3 and the y chains. The y chain is certainly more similar to the /? chain than to the a chain; most of the correspondences between a and y chains are present also in the ft chain. The analysis of the correspondence can be extended to the S chain, for which no ordering of the peptides has been provided, if one assumes that peptides with identical or very similar sequence or com- position occupy a similar position along the f3 and 8 peptide chains. An upper limit for the extent of correspondence between the ^ and the S chains can be established in this way at 94%.

In 1957, when the stiTicture of the human hemoglobins and the sequence of the hemoglobin chains were not yet elucidated, Itano con- sidered the possibility that the genes for myoglobin — Hb-A, Hb-F, and Hb-Aj — were derived from a primitive gene through a process of gene duplication and independent evolution of the dujilicated genes. This hypothesis has been examined in greater detail by Ingram (1961a) on

420 CORRADO BAGLIONI

the ba^i^ of the afcuinulatiiig e\idence of siinilarities in .structure ami amino acid sequence among the human hemoglobin chains.

Ingram (1961 a) (lc\-eloi)cd an evolutionary sclieme which accounts for the similarities between the hemoglobin chains. In the course of evolution a gene or a chromosomal fragment, carrying a hemoglobin gene, has undergone duplication. Mutations of the two initially identical genes result in single or multiple amino acid substitutions, or possibly in in^•ersions of the amino acid sequence of part of the peptide chains. These events are favored or discarded in the course of natural selection. The duplicated genes then evolve independently, under the selective pressure of the environment, and may be separated by chromosomal inversions or translocations. Several duplications followed by inde- pendent evolution would account for the presence of the four hemoglobin chains, for which an independent genetic control has been shown. The primaiy sequences of the a, (3, y, and S peptide chains have been shown to be determined by different stmctural genes; these genes undergo mutational events which affect only the corresponding peptide chains (see Section IV, A, E, and F). On the contrary, the peptide chains of the minor component Hb-As have been shown to be not under independ- ent genetic control. The chemical studies on Hb-A-, (Muller, 1961) provide evidence that the peptide chains of Hb-A and of Hb-A:j are under the control of the same structural genes and have an identical primary sequence.

E. Hb-Aa

This minor component migrates in electrophoresis at pH 8.6 faster than Hb-A, but is not completely resolved from Hb-A in starch block electrophoresis (Kunkel and Wallenius, 1955). Considerable evidence was obtained by these authors that Hb-A.-, is a heterogeneous component, which may be further resolved into at least two fractions.

Hb-As is most likely a derivative of Hb-A (Kunkel and Wallenius, 1955) ; Hb-As increases in amount in old samples, particularly if Hb-A is present as methemoglobin. The rate of labeling of Hb-Ag after Fe"^^ administration was studied by Kunkel and Beam (1957) ; Hb-Aj showed a lower specific activity than Hb-A or Hb-Ao. This and several other observations have shown that the Hb-A,-, fraction, which can be further resolved by column chromatography (Schnek and Schroeder, 1961). contains a mixture of by-products of Hb-A, formed by alterations of the hemoglobin molecule during the aging of the red cells.

The fingerprint of Hb-A.-, isolated by column chromatography has shown a single difference from the finger])rint of Hb-A (Muller, 1961). A peptide reacting to the ninhydrin and sulfur test was present only in the fingerprints of Hli-A.-j and not in the fingerprints of pui'ified Hb-.\;

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 421

this peptide was shown to be glutathione on the basis of the amino acid composition and of chromatographic and electrophoretic mobihty (]\Iul- ler, 1961). Glutathione is not an integral part of the Hb-Ag molecule, although it is rather stably attached; it is indeed released upon acid or heat denaturation but it is not removed by extensive dialysis (Muller, 1961). The type of chemical bond established between glutathione and hemoglobin is not known ; aging of hemoglobin is required to bind gluta- thione, present in high concentration in red cells, to the hemoglobin molecule. The heterogeneity of the Hb-Aj component can possibly be explained by the binding of a different number of glutathione molecules to the hemoglobin molecule. Glutathione has also been found in one of the minor components isolated from cord blood hemogloinn (Muller, 1961). Undoubtedly, many instances of heterogeneity of hemoglobins are likely to be explained by similar secondary modifications of the hemoglobin molecule. Environmental factors probably play a decisive role in modifying the hemoglobin molecule.

F. EMBRYONIC HEMOGLOBIN

A hemoglobin different from Hb-F has been reported in very early fetuses. This hemoglobin has been designated primitive or embryonic type hemoglobin (Allison, 1955). Drescher and Kiinzer (1954) have reported a hemoglobin with intermediate alkali resistance in human fetuses 7-12 weeks of age. Halbrecht and Klibansky (1956) have re- ported a hemoglobin component with lower electrophoretic mobility than Hb-F, but with an Hb-F type UV spectrum in very early fetuses. Huehns et al. (1961a) have reported two hemoglobins with different electrophoretic mobilities in a survey of human fetus hemoglobins; these components were observed in the smallest fetuses examined, which had a crown-rump measurement of only 3.5 cm. No chemical characterization of these hemoglobins has, however, been reported.

The evidence for the existence of an embryonic hemoglobin is still unsatisfactory, since it has not been supported by careful chemical investigations. These investigations may rule out the possibility that the hemoglobins observed in very early fetuses correspond to artifacts or that they may in some instances represent genetic variants of Hb-F. It has been suggested by Matsuda et al. (1960), who failed to observe any hemoglobin different from Hb-F in human fetuses, that the "em- biyonic" hemoglobin is an artifact of the deterioration of Hb-F.

IV. Abnormal Hemoglobins

Abnormal hemoglobins are defined as the hemoglobins which differ in their physical and chemical properties from the normal components.

422 CORRADO BAGLIONI

They arc, with ivw fxt'cptioii;^. which will he cDiisidcrcd separately, the direct expression of mutations which aic inherited as autosomal genes.

The first hemoi^lohin ahnormality re])orted in the literature was asso- ciated with congenital methemoglobinemia (Horlein and A\'eber, 1948). These authors found that the hemoglobin from patients affected by this disease showed an abnormal absor])tion spectrum when in the form of acid methemoglobin; the disease was transmitted as a dominant factor through four generations. The first abnormal hemoglobin electropho- retically different from Hb-A was found by Pauling et al. (1949) in sickle-cell anemic patients. Hereditary transmission of this disease has been recognized for a long time, but until Neel's investigation in 1949 it was not clearly understood that the sickling gene is responsible for the mild abnormality known as sickle-cell trait in heterozygotes and for the severe sickle-cell anemia in homozygotes. Pauling et al. (1949) reached similar conclusions independently and were able to show that no Hb-A is present in homozygotes and that the heterozygotes have both Hb-A and the abnormal hemoglobin (Hb-S) in the approximate ratio 60:40.

These fundamental observations stimulated several investigators to examine patients affected by other forms of hereditaiy anemia and normal individuals, as well, for the presence of abnormal hemoglobins. Since 1950 many other abnormal hemoglobins have been discovered: Hb-C by Itano and Neel (1950), Hb-D by Itano (1951), Hb-E by Itano et al. (1954), and several others more recently (see Lehmann, 1960; Rucknagel and Neel, 1961). The enumeration of all the abnormal hemoglobins reported in recent years is outside the purpose of this review. Only short descrii)tions of few abnormal hemoglobins, which are of particular genetic or chemical interest, will be given in this review. Techniques involved in the detection of the abnormal hemoglobins have been reviewed by Lehmann (1960) and by Rucknagel and Neel (1961). The i^roblem of the nomenclature of abnormal hemoglobins has become increasingly difficult. The letters of the al])hai)et were assigned in the order of tlieir discovery to the various types of abnormal hemoglobins, with the exceptions of the letters F and S, reserved to fetal and sickle- cell hemoglobin. Since the abnormal hemoglobins discovered have out- numbered the letters of the alpliabet, it has become necessary to name the newly discovered abnormal hemoglol)ins after the place where they have been found, following the example of Robinson et al. (1956). The first abnormal hemoglobins discovered were variants of the normal adult hemoglobin, Hb-A. Subsefjuently abnormal hemoglobins have been ob- served that have been shown to be variants of Hb-F and of Hb-A:-. These abnormal hemoglobins will be considered separately fi'om the abnormal forms of Hb-A.

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTKV

423

A. Hb-A VARIANTS — CHEMICAL STUDIES

The abnormal forms of Hb-A represent the largest and best known group of abnormal hemoglobins. Most of the chemical studies have been carried out on variants of Hb-A and our knowledge of the inheritance of the hemoglobin genes and of the population distribution of the

TABLE II

Chemical Studies of the Hb-A Variants

		Abnormal
Charge		chain and
differ-		trjTJtic	Amino acid	Posi-
ence"	Hemoglobin	peptides*	substitution	tion	Reference
-4	I	aTp III	Lys -^ Asp	16	Murayama and Ingram (1959)
	K	a			Gammack et ah (1961)
	J	iS			Itano and Robin- son (1959)
	Jxrindad
	Jlreland	|8			Gammack et al.
	"Jamaica				(1961)
_2	JBaltimore	^Tp II			C. Baglioni, unpublished
	Norfolk	aTp VII	Gly -^ Asp	57	Baglioni (1962a)
	N	^			Gammack et al, (1961)
	R ( = Durham-] )	/3Tp I			Chernoff and Liu (1961)
	MBoston	aTp VII	His -> Tyr	58	Gerald and Efron (1961)
	JVl Saskatoon	i3Tp VII	His -> TjT	63	Gerald and Efron (1961)
0^	MMiIwaukee-1	^Tp IX	Val -^ Glu	67	Gerald and Efron (1961)
	MMiIwaukee-2	0			Gerald and Efron (1961)
	Mlwatc	aTp IX			Gerald and Efron (1961)

" The charge difference with Hb-A at pH 8.6 is the theoretical charge difference based on the amino acid substitution and expressed in arbitrary units.

* For the peptide nomenclature see Table I and Gerald and Ingram (1961).

'The Hb-M's are observed as methemoglobins (Fe++'*'); the formation of internal complexes between the iron and the substituting amino acids neutralizes the charge difference with oxy-Hb-A (Fe+"'"). The Hb-M's show a slightly different electrophoretic mobility compared to Hb-A (see Gerald and Efron, 1961). Table continued

424

CORRADO BAGLIONI

TABLE II* (Conlinurd)

		Abnormal
Charge		chain and
differ-		tryplic	Amino acid	Posi-
ence"	Hemoglobin	peptides'*	substituti	on	tion	Reference
	Dl'unjab	liTp XIIl	Gill	- .Glu-	MI,	121	Bafi;lioni (l!»()2l))
	( = Dcypriix)
	D„	aTp IV					Beiizcr el al. (1958)
	Drf	^Tp III					Benzer eifl/. (1958)
	Drrankfurt	li					Gammack et al. (1961)
	D	P					Chernoff and Liu (1961)
	D	P					Itano and Robin- son (1959)
	vJSan Jose	^Tpl	Glu	--Gly		7	Hill and Schwartz (1959)
	Gpniladelphia	aTp IX	Asp-	NHa-^	Lys	68	Baglioni and
	(= GAzakuoli)						Ingram (1961) and C. Baglioni, unpublished
	GhodoIuIu	aTp IV	Glu	-^ GIU-NH2	30	Swenson et al.
+2	Glbadan ( GBristoI J Gchinese L P Q	a a					(1962) Gammack et al. (1961) Gammack et al. (1961) Gammack et al. (1961) Gammack et al. (1961) Gammack et al. (1961)
	s	/3Tp I	Glu	-^Val		6	Himt and Ingram (1959)
	Stanleyville I	aTp IV					Chernoff and Liu (1961)
	Zurich	/31 p VII	His	-^ Arg		63	Muller and Kingma (1961)

* For footnote to table see p. 423.

abnormal hemoglobin genes is based on the study of the Hb-A variants. The Hb-A variants can be roughly subdivided in four groups according to their electrophoretic mobility; this classification has been found useful in classifying new types of abnormal hemoglobins. In Table II are indicated the abnonnal hemoglobin types discovered to date that have been chemically studied and shown to be variants of Hb-A.

IX. GENETICS .\ND HUMAN HEMOGLOBIN CHEMISTRY

425

TABLE II (Continued)

		Abnormal
Charge		chain and
differ-		tryptic	Amines acid	Posi-
ence"	Hemoglobin	peptides'"	substitution	tion	Reference
	C	/3TpI	Glu — * L\^s	6	Hunt and Ingram (1959)
	E	(3Tp III	Glu ^ Lys	26	Hunt and Ingram (1959)
+ 4	0 Arabia	iSTp XIII	Glu -^ Lys	121	Baglioni and Lehmann (1962)
	^Indonesia	aTp XII	Glu -^ Lys	116	Baglioni and Lehmann (1962) and C. Baglioni, unpublished

The chemical study of the abnormal hemoglobins by means of the fingerprinting method of analysis was initiated by Ingram (1956). The fingerprints of Hb-A and Hb-S showed that only one peptide in the Hb-S fingerprint occupies a position diff'erent from that in Hb-A finger- prints (see Fig. 2). After tryptic hydrolysis of the human hemoglobins a few large tryptic peptides, accounting for approximately ^ to % of the molecule, are insoluble at neutral pH. These peptides are removed by centrifugation from the tiyptic digest of the human hemoglobins prior to fingerprinting (Ingram, 1958). That portion of the hemoglobins which contains the peptides insoluble at neutral pH has been designated the ''core" of the molecule (Ingram, 1958). The "core" can be examined by fingerprinting after chymotryptic digestion (Hunt and Ingram, 1958a). Hunt and Ingram (1958a) have shown that the fingerprints of the chymotiyptic digest of the Hb-A and Hb-S "cores" are identical.

The isolation and the analysis of the peptide altered in Hb-S have allowed Ingram (1958) to establish that Hb-S differs from Hb-A by only one amino acid: a valine in Hb-S is substituted for a glutamic acid residue in Hb-A. This fundamental observation which revealed the alteration caused by a mutant gene at the molecular level undoubtedly was a landmark in the history of biology.

The next abnormal hemoglobin analyzed was Hb-C (Hunt and In- gram, 1958b) ; the same peptide displaced in Hb-S fingerprint is missing in Hb-C fingerprint, while two new peptides appear. The chemical analysis of these peptides (Hunt and Ingram, 1960) showed that the same residue substituted by a valine in Hb-S, namely, a glutamic acid, is replaced in Hb-C by a lysine residue. This new lysine offers a point of attack to trypsin and two peptides appear instead of the altered one in Hb-C fingerprints.

426 CORRADO BACLIONI

TIk' known allflisiu of the Ilh-S ;iii(l lll)-(" «;ciic> I l{aniicy, 1954) was thus coiiliiiiicd at a strictly molecular lc\-('l. 'riicsc results estab- lished a correlation between the abnoi-nial hemoglobins genes and the linear structure of the peptide chains, the synthesis of which is con- trolled by these genes.

Several other abnormal liemoglobins were analyzed in a similar way by Ingram and by others (see Table II). From the chemical studies of the peptide chains of Hb-A (Ingram, 19r)9a), it is known which pej)- tides in the fingerjirinting of this hemoglol)in belong to the « chain and which belong to the (i chain (see Table 1 1. By fingerprinting the abnor- mal hemoglobins it was possible to deteiinine the i)ej)ti(le chain altered in these hemoglobins; i.e., peptides altered in fingerpiints of Hb-S, Hb-E, and Hb-Upunjab were found to belong to the /? chain, pe|)tides altered in Hb-I, Hb-Norfolk, and Hb-Gi-Miiad^ipina were found to belong to the a chain. The alterations of the Hb-A variants have been assigned to one or the other of the Hb-A peptide chains by this method of analysis oi' by recombination experiments (Itano and Robinson, 1959; see Section IV,D). The chemical formula of an abnormal hemoglobin can be written by indicating with the hemoglobin designation the peptide cliain that is altered, i.e., Hb-S = a-^ p.^, Hb-C = a-^ po"", Hb-I = as'iSa^.

Ingram ( 1959a) suggested a one gene-one peptide chain relationship for the Hb-A peptide chains. In this view each gene controls the syn- thesis of one type of peptide chain. The structural genes determining the amino acid sequence of the peptide chains arc designated with the same symbols used to indicate the corresponding ])eptide chains; namely the gene a^ carries the information for the synthesis of a^ chains and the gene /3-^ for the sjmthesis of /?-^ chains. We can subdivide the Hb-A variants into two categories: the a chain and the /? chain variants. The Hb-S gene is defined as p^, the Hb-I gene as a', and so on.

One of the most significant results of the chemical analysis of abnormal hemoglobins is the discovery by Benzer et al. (1958) of the heterogeneity of hemoglobins, which are indistinguishable by means of physical methods of analysis. Three samples of Hb-D obtained from individuals living in different countries were analyzed and found to cany different peptide changes. These three Hb-D's are thus distinct proteins and, more important, they are certainly produced by three difTerent mutant genes. The importance of such observation should be fully evaluated: the validity of genetic, ethnological, or anthropological correlations established in the absence of a biochemical analysis at the ultimate molecular level is in fact shown to be questionable. The heterogeneity of licmoglobins having identical electi-ojihoretic and chromatographic characteristics has also been demonsti'ated for Hb-0

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 427

(Hill et al.y 1960; Baglioni and Ingram, 1961; Swenson et al., 1962) and for Hb-0 (Baglioni and Lehmann, 1962). Three types of Hb-G cariying different peptide alterations have been described and possibly several other types may exist among the several Hb-G's reported in the litera- ture. Two types of Hb-0 characterized by different alterations have been described (Baglioni and Lehmann, 1962). The heterogeneity of hemoglobin variants, which are indistinguishable on the basis of their physicochemical behavior, seems to be a rather general phenomenon. It is quite likely that several other examples of heterogeneity W\\\ be found in the near future.

All the abnormal hemoglobins examined have shown a single amino acid difference with Hb-A. Any single amino acid substitution can only produce one of four different types of electrophoretically altered pro- teins. If we arbitrarily assign the charge 0 to the normal protein, the substitution of a neutral amino acid by an acidic or a basic amino acid, as well as the substitution of a basic or acidic amino acid by a neutral amino acid, will result in a charge alteration of — 1 or -\-\, respectively. The substitution of a basic amino acid by an acidic or of an acidic by a basic will result in charge alteration of —2 or -|-2, respectively. The charge difference is expressed in arbitrary units and should be multiplied by 2 when referred to hemoglobin, since two altered peptide chains are present in the molecule.

We know many more than four electrophoretically different abnormal hemoglobins, but the abnormal hemoglobins can be roughly grouped in five classes on the basis of their electrophoretic mobility (see Table II). A small electrophoretic difference exists at pH 8.6 between Hb-OArabia, Hb-C, and Hb-E, in which a lysine substitutes for a glutamic acid. These hemoglobins should have the same electrophoretic charge if the charge difference with Hb-A results only from the type of amino acid substitu- tion. The small charge differences between these hemoglobins are prob- ably due to different contributions of the residues to the ionic charge of the protein. The charge contribution is presumably dependent on the place occupied by a residue in the spatial configuration of the protein. The charge of ionizable groups may be modified by interaction with neighboring residues.

There seems to be no correlation between the type of amino acid involved in substitutions, apart from its charge, and the change in electrophoretic mobility. The Hb-G's, for instance, present different substitutions: lysine for glutamic in Hb-Gsan .lose (Hill et al., 1960), lysine for asparagine in Hb-Gphnadeiphia (Baglioni and Ingram, 1961), and glutamine for glutamic in Hb-Gnonouiiu (Swenson et al., 1962).

The electrophoretic behavior of the abnormal hemoglobins is dis-

428 CORRADO BAGLIONI

i'Us:>ih1 luMi' only with reference to i)apcr or starch-gel electrophoresis at pH 8.6. which are most commonly used. It should be pointed out that the abnormal hemogloi)ins may show a charge difference with Hb-A variable with the pH and the nature of the electroi)horetic medium. The behavior of some abnormal hemoglobins in agar-gel electrophoresis is completely different from their behavior in starch-gel electrophoresis at similar pH's I see review by dratzer and Bcaven, 1961) ; i.e., Hb-Dp„njah and II1)-E are not resolved from llh-A by agar-gel electrophoresis at pH 6.2. wiiile they are resolved by other types of electrophoresis and by chromatography on ion exchangers. It seems rather difficult to explain the electrophoretic and chromatographic behavior of the abnormal hemoglobins in terms of charge difTerence with Hb-A only; interactions at the level of the tertiary structure of the hemoglobin molecule may possibly account for some of the discrepancies observed. The charge of the amino acids involved in substitutions and their position in the tridimensional configuration of the ])rotein determine the electrophoretic and chromatographic behavior of the abnormal hemoglobins.

1. The Distributio?} of Amino Acid Substitutions in Hb-A Variants

By determining the composition and the sequence of the peptides altered in alinormal hemoglobins, it has been j^ossible to locate the residues substituted in the Hb-A variants along the sequences of the Hb-A peptide chains. In this way one can draw a map of the chemical changes caused by known mutations of the hemoglobin genes (Fig. 5). This is not a genetic map, but since the amino acid sequence of the peptide chains is probably parallel or colinear with the nucleotide sequence of the corresponding genes, it may correspond to a genetic map of the mutations in the hemoglobin gene. Intragenic recombination is a rare event and there is no doubt that the chances of observing such an event in humans are ridiculously small; thus, there is no hope to prove the colinearity between the amino acid sequence of a hemoglobin chain and the genetic fine structure of the corresponding gene by means of recombination of alleles. Demonstration of colinearity between genetic map and amino acid sequence is now being sought in more favorable organisms, as in Escherichia coli, for the gene controlling the enzyme trytophan synthetase (Helinski and Yanofsky, 1962) and phosphatase (Garen et al., 1961) or in bacteriophage for the gene controlling lysozyme (Dreyer. 1960).

In Fig. 5 is shown schematically the position along the peptide chains of Hb-A of the amino acids substituted in the abnormal hemo- globins. There is no particular clustering of substitutions in any region of the peptide chains. It is remarkable, however, that some residues

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY

429

^(^ 0

4)-

X

	1 -X	S(|)
.2C "o
o
X	"5 o c o I CD
	1 X

'f^' X

J3

I

X

<D * O

■<* J^ '^

W 1.1

^

O 2 5 c 13 "

XI 5 5 -= ^ ;i:

(I'

	1	ti.	'fi	7>
	-2	5	-5	7
	%	■^	T(,	—
«	CO	x	X
<a	o		r	^
^	o		"*	^
H	<a	x	_-

jO	^		-	^
X	fl	■^	"-■	■z.
	"3	.s	P	a
	^	^	o
	•"	G	^	'C
	(B	.+^		^
	o	O	X	^
				X
	'*^		~	.i
	M
	^	_g	^	^
		^	c	."ti

-c "S H 4 i

— ? X ■"

a S* G " ^ Si:--? = -

lO -

c: S .S £ ., ^

o

o cs

b-

430 CORRADO BAGLIONI

appear to \)v iinoh-i'd iiu)iv than oner by an amino acid rei^lacemcnt: this is observed for residues 6, 63, and 121 of the /? chain (see Table II for references). This does not necessarily imply that mutations occur more frequently in given regions of the hemoglobin genes {hot spots). ^lutations inohahly occur in any position along the genes, with the consequent amino acid substitutions at various positions. However, abnormal hemoglobins cariying amino acid substitutions are probably synthesized at a normal rate only if the altered molecules are able to fold in the proper way. Therefore, we would not expect to observe substi- tutions involving certain amino acids because the corresponding muta- tions are particularly rare, but more likely because these mutations result in no production of hemoglobin. Higher levels of organization, the secondary and tcrtiaiy structures, which are dependent upon the primaiy structure, are in fact necessary for a protein to be stabilized.

2. The Consequences of Amino Acid Substitutions on the Secondary and Tertiary Structures of Hemoglobin

The secondary structure refers to the geometrical arrangement of the amino acids of a pejitide chain in helical and non-helical regions; the tertiary structure refers to the final folded configuration of a peptide chain. We know very little about the alterations of the secondary and tertiary structures in abnormal hemoglobins. Watson and Kendrew (1961) have recently suggested that most of the residues involved in the amino acid substitutions of the abnormal hemoglobins stick outward from the three-dimensional configuration of the hemoglobin chains, so that these substitutions would not be expected to affect the tertiar}' structure considerably.

AVatson and Kendrew (1961) have carefully compared the amino acid sequence of the sperm whale myoglobin with that of the human hemo- globin chains. The tertiary structure of the sperm whale myoglol)in has been resolved by X-ray ciystallography to a very high degree of pre- cision by Kendrew and his collaborators (Kendrew et al., 1961); of the 1200 atoms contained in the myoglobin molecule, excluding the hydro- gens, some 75% have been localized without ambiguity in a three- dimensional model of this protein. Perutz et al. (1960) have shown by X-ray crystallograjihy that the tertiary structure of the a and p chains of horse hemoglobin closely resembles that of the spenn whale myoglobin. Watson and Kendrew (1961) also have assumed that the secondary and tertiary structures of the human hemoglobin chains closely resemble that of the spenn whale myoglobin and have ideally aligned the hemo- globin peptide chains along the model of the sperm whale myoglobin.

By this comparison most of the residues involved in substitutions in

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 431

tlie abnormal hemoglobins have been found to lie on the surface of the tridimensional structure of the hemoglobin peptide chains. Few notable exceptions to this rule are represented by the amino acid residues involved in substitutions in the Hb-M's (Gerald and Efron, 1961).

The characteristic abnormality of this group of Hb-A variants is an unusual difficulty in reducing the ferric form of the iron atom in the heme group to the ferrous form. The Hb-M genes are responsible for the dominant hereditary form of methemoglobinemia; the recessive heredi- tary form of this disease is due to the absence or alteration of the enzyme diphosphopyridine nucleotide diaphorase (Scott and Griffith, 1959). The presence of Hb-M is associated with chronic cyanosis, since this hemoglobin is normally in the oxidized form as methemoglobin. Five samples of Hb-M from unrelated individuals have been examined by Gerald and Efron (1961). In three of these hemoglobins the amino acid substitution has been detected and localized along the hemoglobin chains.

Tyrosine has been found to substitute for histidine in homologous regions of the a chain in Hb-Muoston (Fig. 6), and of the /3 chain in Hb-Msaskatoon (Fig. 7). Ou tho basis of the similarities with the tertiaiy structure of the sperm whale myoglobin, Watson and Kendrew (1961) have suggested that these regions of the human hemoglobin chains are situated immediately opposite the sixth coordination position of the iron atom in the heme. The heme is known to be located in the inner part of each folded chain, between the folds of the chain; it is most likely attached to a portion of the peptide chains opposite the one altered in Hb-MBoston and Hb-M saskatoon- If, because of a gene mutation, an amino acid with a reactive side chain is substituted for an amino acid with no such side chain in the vicinity of the heme group, then the reactive side chain may complex with the oxidized heme group to give a stable complex. This is an internal complex in which the reactive group of the amino acid acts as a ligand. The complex is quite stable and for this reason the Hb-M remains in the ferric form as methemoglobin (Gerald and Efron, 1961).

That part of the hemoglobin chains where the amino acid substitu- tions leading to Hb-M have been found is presumably in helical con- figuration (Watson and Kendrew, 1961). The pitch of the helix is about 100° per residue, so that the side chains of two adjacent amino acids point toward different directions at a 100° angle. Only residues four positions apart have their side chains pointing toward the same side at a 40° angle.

Other amino acid substitutions have been observed in the same region of the peptide chains where the Hb-Mnoston and Hb-Msaskatoon substitu-

432

CORRADO BAGLIONI

c	n	Ci	01	o	^-v	^^	-r)
2		o	^	a	73	o	a
n	a;	u	■^		^-'	'^^
^	_o	O	o	n		a	—
<*,	n
1		-^	fl	<*«		*	•J

"in _^ o o

H 2

^O

.22 § C-

~ £ ^ ■^ '5 S: ^

a	o.
d	3
	O
u
>^.	bl)
- '	OJ
a	fi
(jj	1)
^	A

^ o -B c

o

C " -O

-^ o O -

§.-

-T" _C

-r C (M ^ ~ -

"2 ^

■ S - ^

2^

c .2 '^ •§ S

^ CO ;S ^' tt

^ o C

J 3 o - _w

• i: 5 Si o ~

ii7':;S^S;S2S

-c fe ^ -C

*^ o

t: o o oj

- a

O) !»

S OJ

Cj ^

hC +^ O

o

a _

bC 0)

S3 ^

rt

O o;

c3

o

d ._ i^i -S -3 ■=! o —

rr T^ -^ -^ X Cl r^

i ^ '^ ^ ^ -^

- S3

01	t>	1	3 o	s	b. rt	'S	o
n		;_	CJ	c3	03	o
c	:!	5	0) s OJ J2	.22	CO C	O
w		hr		n
	^	y^	0)	•^		o;
■o	-3	o			h	■^	m v
6			o	0) T3	^	0) r1
!-=-(		_^	S o	2.			X
	fl	a	r<			C	z^

S = K

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY

433

o	o	Ki	a	o	1 r1	o
i	o	-r!		o		0.
t;		OJ	C	^	-n	^	ci
iA	^	■*-,		-^	a	a>	«J
S W	J2	.2		-a cS is o	o & O	CO	O o
	o	c	i'	>.		+	bC
<		o 5t a;	-3	o a		+	s 0)
n	C	f=	r/I		tc	C
W		Cj	_3	^ ^	o	"

b£^ ^ — 1 ^ • r^

O ^

— ~" -^ •- _C C hH ■5 -I fe ri ^

"^ - = rt

Si :=i o

CO o3

CO 03 O

c -c

.^ O)

J2

■ S y: — ;;r

-S .2 a .s ^ ■§

^ ."ti c3 2 ^ O

"3 d (I) 7^ '5 ^

^ a^ O = o

■13 S o C r c.

^ ^ ^ O ^

CO .b c

CD CO ^

-rt a-

>■ C

0) • 73 O q; H>

T. 2: o _- o

to '.^ ~ C CO

-=! f= ~

CU 0) ^ -O

fl 15 ii -^ ■ S H ^ -5

o ^ C C ci

fl .=; i i a

CL - — - — li

57 — _ t-^ ^ tn —

n	-c	0 ^
^	ro	a; 0	0
	03
w	"^	7J 0)	G-
>	s		c

c c

H ^ ^

3 t^ 0)

s «2 ^

.5 "^

.- ^ 2 J^ c

^ ^H "^ ■— • ^

M •■= '5 ^

i-l-i X c o

a •■^- -

a-5 s

^ 03 .±

434 CORRAIH) UA(.I.1()M

tions have Ix'cii located, hut none of these substitutions causes the formation of a stable inetheinoji;lobin l)y forming; an internal complex as the Ilb-M's do. This is the situation also when the substitution involves an amino acid with a reactive side chain, like Hb-Xoi'folk. where an aspartic acid substitutes for the glycine next to the histidine in IIb-]\I„„s,„„ (Baglioni, 19()2a). The reactive side chain of the aspartic acid veiy likely does not point toward the heme group and camiot complex with the iron atom (see Fig. 6).

In the (3 chain a different Hb-M (MMiiwaukee-i) has been found to have a glutamic acid substituting for a valine in position 67, four resi- dues apart from the histidine in position 63, substituted in Hb-lNIsaskatoon- The side chain of the glutamic acid probably points toward the heme and complexes with the iron atom.

The abnormal hemoglobin H])-Zui-ich has I'ccently been investigated by ]\Iuller and Kingma (1961); in Hb-Zurich the same histidine substi- tuted by tyrosine in Hb-Msaskatoon is substituted by arginine. In Hb- Zurich apparently there is no interaction between the arginine residue in position 63 and the heme group (see Fig. 7). In order to have an interac- tion the requirement is not only the appropriate geometrical position of a residue in the secondaiy and tertiaiy configurations of the hemoglol)in chain, but also the presence of a reactive side chain. The Hb-Zurich carriers show a puzzling phenomenon: severe hemolytic crisis and inclu- sion bodies in the red cells after sulfanilamide administration (Hitzig et ciL, 1960). The presence of Hb-Zurich appears thus to be associated with a drug-induced inclusion body anemia; this abnormal hemoglobin is presumably altered following sulfanilamide administration, and pre- cipitates, forming inclusion bodies.

3. The Nucleotide Base Code and the Amr'jio Acid Su})stihitioiis in Abnormal Ilemoqlohins.

Nirenberg and Matthaei (1962) have recently discovered that poly- uridylic acid directs the synthesis of polyphenylalanine in a cell-free protein-synthesizing system, dependent upon the addition of messenger RNA. This indicated that the nucleotide sequence coding for phenyl- alanine in messenger RNA is made up of uridylic acid only. The effect of randomly ordered polyribonucleotides upon the incorporation of other amino acids has been investigated by Lengyel et al. (1961), by Matthaei et al. (1962), and by Speyer ef al. (1962), in order to determine the composition of RNA coding units. The work of the above-mentioned authors has elucidated the four letter nucleotide code, which specifies the primary sequence of amino acids in proteins. This code is pre- sumably a comma-less, non-overlapping code, as suggested by the

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 435

genetic experiments of Crick et al. (1961). These autliors have obtained evidence that the code is a three-letter, or a multiple of three letters, code. In agreement, the minimum number of nucleotides in coding units of copolymers of three different ribonucleotides, which stimulate the incorporation of some amino acids, is obviously three.

Triplet code letters have been experimentally obtained for 19 amino acids (Matthaei et al., 1962; Speyer et al., 1962). An experimental value for the glutamine coding unit is still unavailable, but the composition of this coding unit has been predicted from amino acid replacement data in nitrous acid mutants of tobacco mosaic virus (Speyer et al., 1962). Smith (1962a) has examined the relationships between the proposed triplet code and the amino acid substitutions in polypeptides or proteins of known structure, pointing out that all the known amino acid replace- ments in the abnormal human hemoglobins are consistent with a single

TABLE III The Amino Acid Code"

Amino acid	Code
Ala	CUG''
Arg	UGC
Asp-NHj	UCA, (AUA)
Asp	UGA
Cys	(GUU), (GUG)
Glu	UAG
Glu-NHo	UCG
Gly	UGG
His	UAC
lieu	UUA
Leu	UCU, AUU, (UGU)
Lys	UAA
Met	GUA
Phe	UUU
Pro	ecu
Ser	CUU, (UCG)
Thr	CUA, UUC
Try	(GGU)
T>T	UAU
Val	UUG

" The code is constructed according to Smith (1962b), assigning to glutamic acid the base sequence UAG. The triplets composition is from Matthaei e( al. (1962) and from Speyer et al. (1962). The sequence of the triplets in parenthesis cannot presently be established. This is one of the six possible codes, based on the permutations of the sequence for glutamic acid arbitrarily assigned.

* Abbreviations: A, C, G, and U are used for adenylic, cytidylic, guanylic, and uridylic nucleotide residues in RNA, respectively.

436

CORRADO BACILIONl

base chaiijic in the pi-oposcd code. Smith (]9()21)) lias shown that if an arbitrary so([iK'ni'o is assif^nt'd to the triplet cotliny for a given amino acid (i.e., the ba.sc sequcnee for glutamic acid is a.ssumed to be IWCJ). the se(|Uence of the trijilets coding for 18 amino acids can be established utilizing the known amino acid sul)stitutions in abnormal hemoglobins and in nitrous acid mutants of tobacco mosaic virus (see Table III). It is possible to construct in this way six sets of codes, based on the six permutations of the seciuencc arbitrarily assigned; one of these is likely to represent the correct code in messenger RNA. However, for some amino acids two or three alternative coding units have been rejjorted (Matthaei et al., 1962; Speyer et al., 1962). The code appears thus to be degenerate, as suggested also by the elegant genetic exjieriments of Crick et al. (1961).

In view of tlie degeneracy of the code it is of some interest to examine in detail instances of multiple amino acid replacements of the same resi- due. These replacements are evidently produced by different changes of the same coding units. Three double replacements have been found in the ^-peptide chain of Hb-A (for references and explanations see Tables II and nil:

Glu

Val

Lvs

UAG

UUG

UAA

^ Zurich g3

His

Tyr

Arg

UAC

UAU

UGC

D Punjab /3 121

Arabia

121

Glu

Glu-NH.,

UAG

-UCG

UAA

These rejilacements are in complete agreement with the jiroposed genetic code (Smith, 1962b).

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY

437

B. FORMAL GENETICS OF THE Hb-A VARIANTS

jMcndelian inheritance of the abnormal hemoglobin genes has been shown for most of the Hb-A variants (Neel, 1956). The most pertinent information on the inheritance of the hemoglobin genes is obtained from marriages where one partner possesses two abnormal hemoglobins and the other none (Rucknagel and Neel, 1961). Ranney (1954) described marriages in which both Hb-C and Hb-S were observed in one partner; only children with either Hb-S or Hb-C were found among the offspring of these marriages. The allelism of the Hb-C and Hb-S genes was con- firmed a few years later by the chemical investigation of Hunt and Ingram (1958b); the same amino acid residue was found to be substi- tuted in these two abnormal hemoglobins.

One marriage supporting allelism of the Hb-E and Hb-S genes has been described by Aksoy and Lehmann (1957). Veiy few marriages involving a partner with two chemically characterized abnormal hemo- globins have been reported. No individuals having two hemoglobins abnormal in the a chain have thus far been reported.

The low frequency with which the hemoglobins abnormal in the a chain are observed in populations suggests very small probability that

/ \

.y

^^p"

IL.

\. y

0-5

Not examined

Fig. 8. Family tree of the propositus (indicated by the arrow) with four hemo- globins. (From Atwater et al, 1960a). The genotype of the individuals examined is illustrated. The capital letters designate the hemoglobin types observed in the corresponding individuals. The projiositus and one sister are doubly heterozygous.

438 CORRADO BAC.LIONI

critical marriages involving individuals with two of these hemoglobins can be encountered.

Allelism for most of the abnormal hemojilobin genes has been postulated because of the absence of Hb-A in individuals having two abnormal hemoglobins. Individuals ha\ing Hb-S and Hb-I),.unjab or having Hb-8 and Hb-N do not show any Hb-A, even in trace amounts. The first exception to the observation that persons with two abnormal hemoglobins do not have any Hb-A was rej)orted by Smith and Torbert (19581. These authors described the occurrence in a family of Hb-!S and of a second abnoniial hemoglobin (Hb-Hopkins-2) ; in all the members of this family showing both abnormal hemoglobins, Hb-A also was found to be present. Smith and Torbert (1958) reported a three- generation pedigree showing evidence for independent segregation of the two abnormal hemoglobin genes.

Atwater et al. (1960a) have reported the presence in two sisters of four hemoglobins; in the three children of the two sisters, presumably married to normal individuals, either no abnormal hemoglobins or only one abnormal hemoglobin w-as observed (Fig. 8). Other individuals with three (Cabannes and Portier, 1959) and four hemoglobins (Raper et a/., 1960) have been reported. The existence of two hemoglobin genes is shown by the segregation of the abnormal hemoglobins in the families reported by Smith and Torbert (1958) and Atwater et al. (1960a). Atwater et al. (1960a) suggested possible schemes of inheritance of these genes; they considered it most likely that these genes are located on two different chromosomes. The possibility that the two genes are located far apart on the same chromosome should also be considered ; the small num- ber of individuals observed does not provide definite information on the position of the tw^o genes.

C. GENETIC AND BIOCHEMICAL RELATIONSHIPS

Of particular interest is the presence in two members of the family reported by Atwater et al. (1960a) of four electrophoretically distinct hemoglobins. Baglioni and Ingram (1961) have investigated the chemi- cal alterations of these four hemoglobins, which have been called Hb-A, Hb-G, Hb-C, and Hb-X, according to the order of migration in electro- phoresis. The Hb-C from this family was found to be identical to the Hb-C described by Hunt and Ingram (1960) and consequently altered in the /? chain. The Hb-G (designated Hb-Gphuadeiphia) was found to be al- tered in the a chain, where a lysine substitutes for an asparagine residue. Hb-X was found to contain the abnormal peptides characteristic of both Hb-C and Hb-Gpi.jiadeiphia; Hb-X is thus abnormal in both its a and ft peptide chains. A similar situation has been described by Itano and Robinson (1960) for the individuals of the family reported by Smith

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 439

and Torbcrt (1958) with three hemoglobins. Hemoglobin Hopkins-2 has been found to be altered in the a chain by dissociation and reassociation experiments (Singer and Itano, 1959; see Section IV); the alteration of the second abnormal hemoglobin, Hb-S, observed in the family is known to reside in the /3 chain. Hb-S and Hb-Hopkins-2 have altered electro- phoretic mobilities; Hb-S has a +2 charge and Hb-Hopkins-2 a —2 charge with respect to Hb-A. The doubly abnormal hemoglobin Q,^^Hopkins-2^^s happens to have the same charge as Hb-A. Itano and Rob- inson (1960) have shown that this hemoglobin is present in the indi- viduals with three hemoglobins of the family reported by Smith and Torbert (1958).

The occurrence of hemoglobin molecules abnormal in both peptide chains allows one to make certain deductions about the mode of assem- bly of the hemoglobin molecule. The persons with four hemoglobins are doubly heterozygous, having the genotype a'^'/a^ /^Vj^"^; as suggested by Atwater et al. (1960a) or aV«"°'"''"'"' )SVj8^ (Itano and Robinson, 1960). Apparently each gene causes the manufacture of a peptide chain charac- teristic of it, so that four types of peptide chains are produced in these persons. It has to be assumed that each gene controls the synthesis of a dimer, rather than that of single chains. Hemoglobin molecules having two different a or [3 chains, such as a^ a^ fSn^ or a.'^ (^^ (3^, have never been observed (Itano, 1956). The fact that each gene controls the synthesis of a dimer may be a consequence of the way in which the hemoglobin molecules are synthesized. A possible explanation is that the chains have to dimerize in order to be released in solution. In this view each microsomal particle has the information for only one type of peptide chain or possibly several copies of the same peptide chain; the newly synthesized chains are scarcely soluble and are released in solution only upon dimerization. The different types of dimers present assemble in a random fashion to complete hemoglobin molecules, leading to all the possible combinations (see Fig. 9).

An alternative explanation is that each chain associates with an identical partner because of steric hindrances, which prevent association of non-identical partners. It is rather difficult to hold such an explana- tion in view of the rather limited alterations in three-dimensional struc- ture that certain Hb-A variants seem to have. Hb-A and Hb-Dpunjab, for instance, differ by only one amide group; a glutamic acid in Hb-A is changed to glutamine in Hb-Dpimjab in position 121 of the ^ chain (Baglioni, 1962b). It seems unlikely that this limited alteration could prevent association of the /3'^ chains with the /J'^punjai. chains.

According to the former explanation for the specificity of the hemo- globin chain dimerization, the existence of a hemoglobin molecule having two different types of « or ^ cliains should be possible. No such molecule

440

CORRADO BAG LION I

Fic. <>. ^e'liciuatic oxplanatiou of the presence of four hemoglobins in doubly heteroz3'gous individuals (Atwater et al., 1960a; Baglioni and Ingram, 1961). The loci controlling the synthesis of the a and /? peptide chains are represented on the left. The arrows show the different dimers synthesized under the control of the a and p alleles and the random combination of these dimers in the last stage of the assemljly of the hemoglobin molecule. The letters A, G, C, and X indicate the four hemoglobin types synthesized. Hb-A results from the combination of a-/ dimers with ^jA dimers, Hb-G from the combination of a.p dimers with P-,^ dimers, Hb-C from the combination of a.,^ with p.,^ dimers, and Hb-X from the combination of a2^' with jSj^ dimers.

has yet been reported to be produced by experiments of dissociation and recombination; however, the hemoglobin molecule is not dissociated to single chains in these experiments. (Hasserodt and Vinograd, 1959).

D. DISSOCIATION AND RECOMBINATION OF HEMOGLOBIN

Field and O'Brien (1955) reported that hemoglobin dissociates reversibly at acid and alkaline pH's into subunits of molecular weight approximately half that of undissociated hemoglobin. Itano and Singer (1958) applied the acid dissociation method to the study of the dissocia- tion and recombination of mixtures of hemoglobins A, S, and C. Singer and Itano (1959) and Vinograd et al. (1959) observed transfer of C^''- labeled subunits from Hb-A to Hb-C or Hb-S upon dissociation and recombination. When equivalent amounts of Hb-A^'^ and Hb-C were recombined, one quarter of the radioactivity was found to be transferred to Hb-C. By these experiments Itano and his collaborators and Vinograd et al. (1959) have been able to show that different hemoglobins exchange subunits by a specific mechanism of dissociation and recombination, which results in the formation of hyl)rid molecules. However, formation of hybrid molecules containing two different a or ^ chains does not occur. Recombination of hemoglobins altered in different chains provides a

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 441

more direct and striking demonstration of the exchange of subunits upon recombination. Itano and Robinson (1959) recombined a mixture of Hb-I <«,'i8o^) and of Hb-S (a-^ ^J") and observed the formation of two new hemoglobins, one with the electrophoretic iriobility of Hb-A iay'' fS-y'') and the other, electroi^horetically faster than Hb-A, resulting from the combination of a/ dimers with f^-/ dimers. Observing the presence of four hemoglobins after recombination of two hemoglobins having differ- ent jieptide chains, Itano and Robinson (1959) predicted that in indi- viduals doubly heterozygous for two abnormal hemoglobins, a similar situation could be found. Their prediction was scon to be validated by the discovery of individuals with four hemoglobins (Atwater et al., 1960a; Raper et a/., 1960).

Jones et al. (1959a) have used dissociation at alkaline pH and recombination, to determine the gross chemical structure of Hb-H (see Section IV,H,1). Recombination of Hb-H with Hb-S^" (a-^* fS-^*) re- sulted in the formation of four components, which were identified as fi^^, I3i^*, ao"* ^2-'^, and ay"* /So^*, thus showing that Hb-H is an abnormal hemoglobin made up of fS'^ chains only.

Jones et al. (1959b) have shown that «2'''* subunits can be transferred from Hb-S«" (ao^* ^2^"") to Hb-F and that ao subunits from Hb-F could associate with ySo^ dimers produced by dissociation of Hb-H to give Hb-A (a2^ ^o^) ; this indicates that the a-2 dimers of Hb-A and Hb-F can be exchanged and are probably identical. Weatherall and Boyer (1962) have reported the formation of Hb-A and hemoglobin F/G (a-,^ y/) upon acid recombination of Hb-F (a^'" yo^) and of Hb-G (ao^/S.-M.

The results of the recombination experiments have suggested (Itano and Robinson, 1959) a method for identifying the altered peptide chain of any electrophoretically abnormal hemoglobin. A hemoglolnn in which the abnormal chain has been identified is recombined with an abnormal hemoglobin in which the abnormal chain is unknown. If the two hemo- globins are abnormal in different peptide chains, namely, if one has abnormal a chains and the other abnormal /? chains, new hemoglobin species, which are electrophoretically different will be produced upon recombination. If the two hemoglobins are abnormal in the same chain, no new hemoglobin species will be produced upon dissociation and recom- bination. A microscale method of analysis has been devised by Gammack et al. (1960) to determine by dissociation and recombination the altered peptide chain of an abnormal hemoglobin. By this method of analysis Gammack et al. (1961) were able to individuate the abnormal peptide chain of several abnormal hemoglobins. The dissociation and recom- bination analysis has also been applied to the study of other human hemoglobins and of animal hemoglobins.

442 CORH.MX) HACLIOXI

Huelins and Shooter (1961) have shown by recombination that Hb-Ao dissociated into «..'^ and 8-/'^= subunits and that the ay'' dimors can combine with ^o-^ diniers to give a h('nu)gU)bni ch'cti'ophoi'otically identical to II1)-A. In Di'der to obtain recombination of Ilb-A.., piolongt^l exposure to acid pi I is necessaiy and equilibrium in the exchange of sul)units is not reached after 72 hours. The rate of exchange of Hb-Aj sul)units is con- sidei'ably inci'cased by dissociation in acid buffers of higher ionic strength (Weatherall and Boyer, 19()1). These findings suggest that the chemical bonds between the subunits of Ilb-Ao are proljably different from the ones iin'oh-ed in stabilizing the (luaternary sti'ucture of Hb-A.

The ways in which the exchange of subunits occurs upon dissociation and recombination merits further experimentation. A conti'adiction exists between the fairly rapid change in molecular weight of the hemoglobin molecule at acid pH and the relatively long time required to reach equilibrium in the exchange of subunits (J. R. Vinograd, personal com- munication). Itano et (d. (1959) considered two possible mechanisms to explain the results of dissociation and reassociation experiments: (1) Hemoglol)in dissociates symmetrically into half-molecules, but only identical half-molecules recombine to re-form the original components; (3) hemoglobin dissociates asymmetrically into pairs of identical chains. Mechanism (2) is in agreement with the recombination experiments, while mechanism (/) would explain the rapid reduction in molecular weight of hemoglobin in acid solutions, which is not followed by recom- bination of subunits. The factors which may prevent aggregation of non-identical half-molecules obtained by symmetric dissociation are not understood at all. We are still far from having a clear-cut explanation of the physicochemical phenomena involved in dissociation and reas- sociation of hemoglobin molecules.

E. Hb-F VARL\NTS;

Electrophoretically abnormal fonns of Hb-F luu'e been reported by Fessas and Papaspyrou (1957), by Vella (1959), and by Fessas et al. (1961). These variant forms of Hb-F have not been chemically investi- gated, so that they are still rather undefined. They are thought to have altered y peptide chains; this would prove that the y peptide chain is under independent genetic control. The occurrence of mutant forms of a y chain gene has not been shown experimentally; the existence of this gene is predicted l)y the one gene-one peptide chain hypothesis (Ingram, 1959a).

1. The a Chains of Hb-F

Minnich et al. (1962) have reported the occui'rence in the cord blood of two Negro babies of a hemoglobin with an electrophoretic mobility

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY

443

somewhat slower than that of Hb-D; this hemoglobin was found to be alkali-resistant and to have the same N-terminal amino acids as Hb-F. One of the parents of each baby was heterozygous for Hb-Dst. Louis. Min- nich et al. (1962) postulated that the abnormal fetal hemoglobin resulted from the combination of a^st.Louis chains with y'" chains. Similar cases have been observed by Weatherall and Boyer (1962). The cord blood hemoglobins of a Negro baby have been studied by Weatherall and Baglioni (1962); column chromatography on IRC-50, according to Allen et al. (1958), showed the presence of four hemoglobins, identified by fingerprinting as Hb-F, Hb-F/Gphnadeiphia Hb-A, and Hb-Gpunadeiphia in the order of elution from the column. The alteration of Hb-Gphiiadeiphia is known to reside in the a peptide chain (Baglioni and Ingram, 1961). These results show that the a chain of Hb-A and Hb-F are under the control of the same a gene. The a peptide chains of Hb-A and Hb-F have been shown to be chemically indistinguishable (Hunt, 1959; Schroeder et al., 1961). Nomial ao^ dimers and abnormal a^^ dimers are produced in

Fig. 10. Schematic explanation of the presence of four hemoglobins in fetuses or newborns heterozygous for an a chain mutation (Minnich et al., 1962; Weatherall and Boyer, 1962). The loci controlling the synthesis of a, /3, and y peptide chains are represented on the left. The arrows show the different dimers synthesized under the control of the two a alleles and of the /3 and y genes. These dimers combine randomly in the last stage of assembly of the hemoglobin molecule to form the four hemoglobins indicated by capital letters.

444 CORRADO HACLIOXI

a'^/«^ lu'tt'i'Dzyiioiis l);il)i('s and the (rniicc> coiiiMiic rainlonilN' with fj/'' oi' y/ (liinci's in a way apparently siniilai- to that sn^^t'.stt'd to cxphain the pivscncc of I'oui- h('iiioii;lul)int5 in individuals doubly ht'tcrozygous fof an a' and a /3 chain ahnonnality (Fig. 10).

F. lib- A:; VARIANTS

1. 8 (linin ]'ariants

Ccppellini (1959a) discovered an abnoinial minor component, called Hb-Bj, which migrated in electrophoresis more slowly than Hb-Ao. Ccppellini < 1959b) suggested that Hb-Ba is related to Hb-Aa since the amount of Hb-Ao in the Hb-Ba carriers is decreased to half its normal N'alue and the sum of these two fractions equals the normal value of Hb-Aa. ^Moreover, in thalassemia heterozygotcs the amotmt of both Hb-A, and Hb-B, is doubled (Ccppellini, 1959b).

Huisman and Meyering (1960) have also describt'd a minor com- ponent with an electrophoretic mobility similar to that of Hb-Bj. Horton et al. (1961) have reported an individual apparently homozygous for the Hb-Bo gene; the chemical study of Hl)-B:. has shown that only one pep- tide is different in the fingerprint of this hemoglobin, when conijiared to the fingerprint of Hb-Ao.

The peptide chains of Hb-An and of Hb-B.. have been separat(Yl and fingerprinted by Horton et al. (1961), and the altered peptide was shown to belong to the 8 cliain. The amino acid substitution may presumably involve a glycine in position 16 of the 8^- chain, which may be substi- tuted by an arginine residue in the S"- peptide chain (A. 0. W. Stretton, personal communication) ; this suggestion is in accord with the chemical findings of Horton et al. (1961). These facts prove the existence of the 8 gene as an independent genetic entity.

2. Formal Genetics of the 8 Chain Variants

Ccppellini (1959b) reported a double heterozygous ))ropositus /3-^ 8''-/ ^s gAo i-n-ii-i-iefj ^Q .J noniial woman. Their six children were found to be either Hb-S or Hl)-B2 carriers; the alternative segregation of the j3^ and 8"- genes suggests linkage of the ^ and 8 loci, the 13^ and 8"- genes i)eing in repulsion in the propositus (see Fig. 11).

Huisman et al. (1961) have reported a family in which scn-eral mem- bers exhibit Hb-Bo plus thalassemia minor of the "high-A." type (see Section IV,G). This type of thahissemia behaves as if allelic to the (3 locus (see Ingram and Stretton, 1959b) and the abnormal gene may be indicated as [3"^^ gene. The /3'^^' and 8^- genes do not segregate in the family report(>d by Huisman et al. (1961), suggesting linkage of these

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY

445

two genes in coui)ling. It must be pointed out that in double lu'terozygotes for thalassemia and the 8^- gene, the increase of Hb-A^ level is identical, whether the two mutant genes are in coupling /3'^^^ 8^-/y8'^ S'^- (Huisman et al, 1961) or in repulsion /J^h 8VyS' S''^ (Ceppellini, 1959a).

Not enough genetic data have been accumulated to determine the extent of the linkage between the (i and the 8 loci. It seems unlikely, moreover, that enough data can be found to determine the fine genetic structure of the ft and 8 loci. The /3 and 8 genes seem, however, to be

Fig. 11. Schematic explanation of the presence of four hemoglobins in the propositus of the family reported by Ceppellini (1959b). The genotype is illustrated on the left; the (3 and 5 loci are shown in linkage as suggested by the alternative segregation of the /js and 5^^ genes. The arrows show the different dimers syn- thesized and their random combination to form the four hemoglobins indicated by capital letters.

under a common genetic control: (i) Hb-A and Hb-Ao appear in fetuses at approximately the same age and increase in concentration in a parallel way (Kunkel et al., 1957) ; {2) both Hb-A and Hb-Ao are absent from individuals homozygous for the "hereditary persistence of Hb-F" (Wheeler and Krevans, 1961) or for the Hb-Pylos gene (Fessas et al., 1962). These relationships in the behavior of the ft and 8 genes and the family studies suggest a functional and topographical linkage of these genes.

3. The a Chains of Hb-A.

Shooter et al. (1960) have shown in a person apparently homozygous for the abnormal hemoglobin Gibadan the presence of a minor component with an altered electrophoretic mobility, called Hb-Gi... Hb-Gn.adan bas

446

CORRADO BAGLIONI

\)V('\\ sliuwii to have ahiionual a cliains by dissociation and recombina- tion rxperinients (Ganiniack et a/.. 1960) and it has been suggested that Hb-G-.. is a combination of atiiha.ian chains with 8^- chains (a2Gibadan Sv^-). Keconibination of Hb-Gn.a.ian with Ilb-A,. produced some Hb-Go antl Hb-A (Huehns and Shooter, 1961).

Several reports have appeared recently in the literature of a "split" Hb-Ao in individuals who possess an abnormal hemoglobin. One of the minor components has the electrophoretic mobility characteristic of Hb-A.., while the abnormal minor component shows the same difference in electrophoretic mobility relative to Hb-A, that the abnormal major component shows relative to Hb-A (Weatherall and Boyer, 1961). Bou- lard et al. (1961) have fingerprinted the abnormal minor component present in a Hb-I carrier and Baglioni (1962a) has fingerprinted the

Fig. 12. Schematic e.xplanation of the presence of four hemoglobins in individuals heterozygous for an a chain mutation (Baglioni, 1962a). The genotype is illustrated on the left (No =: Norfolk). The arrows show the different dimers synthesized under the control of the two a alleles and of the fJ and 8 genes. The dimers combine randomly to form the four hemoglobins indicated by capital letters. The /? and 5 loci arc linked (Ceppellini, 1959h).

abnormal minor component present in a Hb-Norfolk carrier. Both Hb-I and Hb-Norfolk arc known to have abnormal a chains (Murayama and Ingram, 1959; Baglioni, 1962a). The fingerprints of the abnoraial minor components have shown the presence of « chains with alterations similar to the one reported for Hb-I and Hb-Xorfolk, respectively, and of normal S-^= chains.

The a chains of Hb-A and of Hb-A- have been shown to be chem-

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 447

ically identical (Ingram and Stretton, 1959; Muller and Jonxis, 1960); they are under the control of the same a gene. The So dimers, like the y., dimers, combine in a random way with all the types of ao dimers present in the last stage of Hb-Aa synthesis (Fig. 12).

4. The Problem of Hb-Lepore

Gerald and Diamond (1958) described a patient of Italian extrac- tion affected by a severe anemia, clinically classified as thalassemia major. The propositus, his mother, and four maternal relatives showed an electrophoretically slow moving abnormal hemoglobin called Hb- Lepore, from the name of the family, in concentration of approximately 10 to 15% ; they were found to be affected by microcytemic anemia. The father of the propositus was found to be a ft'^^/p^ heterozygote and the severe anemia in the propositus was thought to result from the interaction of the jS"^*^ gene and of the Hb-Lepore gene. Similar cases of an electrophoretically slow moving abnormal hemoglobin, associated with microcytemic anemia have been reported among Italians (Silvestroni and Bianco, 1958), Papuans (Neeb et al., 1961), and Greeks (Fessas et al., 1962). This condition seems to be relatively frequent among patients with microcytemic anemia in Greece (Fessas et al., 1962).

The abnormal hemoglobin, always present in concentration of about 10-15%, has been alternatively classified as Hb-G (Silvestroni and Bianco, 1958), Hb-Pylos (Fessas et al, 1962), or Hb-Lepore (Neeb et al., 1961). Individuals apparently homozygous for the Hb-Lepore gene or for the Hb-Pylos gene have been reported ; they are characterized by a complete absence of Hb-A and of Hb-Ao in the case described by Fessas et al. (1962) and presumably in the two cases of Neeb et al. (1961). In the heterozygotes the amount of Hb-As is reduced to about half the normal value (Fessas et al.. 1962).

Gerald et al. (1961) have briefly reported the results of their chemical studies on Hb-Lepore and Hb-Pylos; these hemoglobins give fingerprints indistinguishable from those of Hb-Ao. Moreover, some of the peptides of Hb-Lepore were analyzed and found to have an amino acid composi- tion identical to homologous peptides of Hb-Ao. Neeb et al. (1961) have also fingerprinted Hb-Lepore from Papuans and have obtained a peptide pattern indistinguishable from that of Hb-Ao. Hb-Lepore has presum- ably normal a peptide chains, since no alteration of the fingerprint pat- tern of the a chain has been observed, but mainly because normal Hb-F is present in Hb-Lepore carriers; one would expect that if any abnormal ao dimers were present, they would associate randomly with yz as well as So dimers. The abnormality in the "S" chain of Hb-Lepore has not been discovered, but it may possibly be hidden in the "core" of this

448 CORRADO BAGLIONI

hcmoglohin. whit'h is excliulcd from the fiiiG;c'r|)i'intinfz; analysis. The Hb-Lcporo aljuonnality thus seems to involve a mutation of the /? locus, resulting in comj)lete suppression of /3 chain synthesis and a simul- taneous mutation of the 8 locus, resulting in the synthesis of an electro- ]>horetically abnormal heiuoglobin. Ilb-Lepore. It is rather hard to understand why these independent mutations in the two linked genes fS and S should always be observed in c()ui)lii^g. The S"- mutation has, for instance, been observed either in couj^ling oi' in repulsion with a Z?"^'' gene (Ceppellini, 1959a; Horton et al., 1961).

It seems likely that a common genetic event is responsible foi- both the absence of fS chains and the alteration of the "8" chains; the only genetic event involving two neighboring genes at the same time is an overlapping deletion of genetic material, resulting either from mutation or from unequal crossing-over.

It has recently been suggested by Smithies et al. (1961) that a mechanism of unequal crossing-over between two haptoglobin genes could be responsible for the duplication in length of one peptide chain of the human haptoglobins. A similar mechanism can be advocated to explain the puzzling genetic and chemical findings on Hb-Lepore: in this case loss of genetic material by unequal crossing-over, may result in the formation of a hybrid ;S-8 gene, having part of the genetic material of both the j3 and the 8 genes.

The chemical findings on Hb-Lepore so far do not contradict this theory; only that part of the "8" chain of Hb-Lepore, going from the N-terminal end of the peptide chain to approximately halfway along the chain, has been shown to give a fingerprint similar to the 8 chain of Hb-Ao, while no information exists about the C-terminal portion of the "8" chain of Hb-Lepore. The possibility exists that the C-terminal part of the "8" chain of Hb-Lepore is identical to the p chain of Hb-A.^

It is known that unequal crossing-over frequently involves chromo- somal duplications like the bar duplication of Drosophila inelonngaster (Bridges, 1936) ; double-bar and normal flies appear with relatively high frequency as a result of unequal crossing-over in liomozygous bar females.

It has been suggested that the 8 gene results from a duplication of the (3 gene, followed by independent evolution (Ingram, 1961a). Profound homologies exist between the (3 gene and the 8 gene, since the primary

'Recent exi)eriment.s (Baglioni, unpubli.shed) have shown that the C-terminal portion of the 5 chain of Hb-Lepore is identical to the C-terminal portion of the P chain of Hb-A. Two amino acid replacements have been located by A. O. W. Stretton (unpubHshed) in the C-terminal portion of the S chain of Hb-.\,. These replacements are not pre.sent in the 0-8 chain of the Hb-Lepore.

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 449

sequence of these peptide chains is extremely similar (Ingram and Stret- ton, 1961). These circumstances may increase the probability of unecjual crossing-over involving the (3 and S loci.

G. THE PROBLEM OF THALASSEMIA

Thalassemia is a hereditary anemia which occurs connnonly in Medi- terranean countries. Two main forms of the disease are recognized: thalassemia minor, which occurs in individuals heterozygous for a thal- assemic gene, and thalassemia major or Cooley's anemia, in homozygotes (Valentine and Neel, 1944). Thalassemia minor is a mild hypochromic microcytic anemia which may be asymptomatic, while thalassemia major is a marked hypochromic anemia, characterized by anisocytosis and accompanied by profound alterations of skull and bones, due to bone marrow hyperplasia. Children affected by thalassemia major usually do not reach adulthood and early death in the absence of constant medical care and repeated blood transfusions is the rule. Several aspects of the thalassemic diseases have recently l)een reviewed by Lehmann (1960) and by Rucknagel and Neel (1961).

1. Hemoglobins in Thalassemia

In thalassemia major Hb-F is constantly present in amounts which, together with the hematological picture, are characteristic of this dis- ease. As much as 97% of the total hemoglobin has been reported to be Hb-F in a thalassemic patient (Rich, 1952). This Hb-F appears to be identical to Hb-F isolated from cord blood when analyzed by finger- printing (Baglioni, unpublished). In thalassemia minor the ]iroportion of Hb-F is very close to normal values.

Kunkel and Wallenius (1955), and subsequently Ceppellini (1956) and several others, have reported an increase of Hb-A. to values approxi- mately double the normal values in individuals affected by thalassemia minor. The Hb-Ao obtained from thalassemic patients has been found to be indistinguishable from Hb-Ao from normal subjects by finger- printing (Ingram and Stretton, 1961). Kunkel et al. (1957) and Carcassi et al. (1957) found several exceptions to the rule of increased Hb-Aj in thalassemia minor. This and other genetic observations have proved the heterogeneity at the molecular level of thalassemia.

Several combinations of a thalassemic gene with abnormal hemoglobin genes have been reported in the literature (see Lehmann, 1960). Affected individuals in general show a moderately severe microcytic anemia; the range of variability is, however, quite wide. Sickle-cell/thalassemia is the combination which is more frequently observed (Silvestroni and Bianco, 1952). Hb-S, Hb-F, and veiy small amounts of Hb-A may be present in

450 CORRADO BAGLIONI

sickle-cell, tlialasscmia. Since the amount of any ahnoiinal licniof^lobin is found to be lower than that of llh-A in heterozyjrotes for an abnormal hemoglobin gene, the inversion of the Hb-A abnormal iiemoglobin ratio has been thought to be characteristic of thalassemia in combination with an abnormal hemoglobin gene. This type of thalassemia has been defined "interacting" thalasst'mia and combinations of this tyjie of thalassemia with Hb-C, Hb-D, and 111)-!^ have also been described (({uoted in Leh- mann, 1960). These lu'moglol)ins have been shown to have altered {3 chains (Hb-D is presumably of the Hb-Dp„„jaij type).

Thalassemic individuals in whom the ratio Hb-A/abnormal hemo- globin is not invei-ted have been reported for the Hb-S/thalassemia combination and for the Hb-C/thalassemia combination (Zuelzer et al., 1956; Zuelzer and Kaplan, 1954). The thalassemia in these individuals is thus of a "non-interacting" tyjie. Examination of some of these indi- viduals has shown that the non-interacting thalassemia is of the non- increased Hb-As type. Interaction of the non-increased Hb-Ao type thalassemia with Hb-I, resulting in an inversion of the Hb-A/Hb-I ratio has been reported by Atwatcr et al. (1960b). Hb-I is known to be altered in the a chain (Murayama and Ingram, 1959). Thalassemic genes may thus interact with a or p mutant genes and be of the non-increased or increased Hb-Ao type, respectively.

2. Formal Genetics of Thalassemia

Several marriages of persons affected by Hb-S/thalassemia or Hb-C/ thalassemia combinations have been reported in the literature. (For a review, see Rucknagel and Neel, 1961.) In most marriages only children either heterozygous for thalassemia or heterozygous for the abnormal S or C hemoglobins are born. The thalassemia in these families is of the increased Hb-Aa interacting type. The thalassemic gene in these families behaves as if allelic to the /3^ or y8*^ genes. However, in some similar marriages, normal children, or children with thalassemia or sickle-cell trait or with sickle-cell/thalassemia, are born (quoted in Ingram and Stretton, 1959b). These findings indicate non-allelism of the thalassemic gene concerned with the fi^ or /3^ genes. The thalassemia in these families is usually of the non-increased Hb-Ao, non-interacting type.

In the marriage of the Hb-I thalassemic individual reported by Atwater et al. (1960b), only children with the thalassemic trait were i)orn. It may be pointed out that this thalassemic gene is of the non- increased Hb-Aa type and was shown to interact with the «' gene; it is not proved, because of the very small number of individuals in this family, whether this thalassemic gene behaves as if allelic to the «' gene, as the family data suggest.

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 451

3. The Molecular Basis of Thalassemia

The main pathological manifestation of thalassemia is the failure to produce red cells with a normal hemoglobin content. The total red cells production is increased (Bailey and Prankerd, 1958 1, hut in a way insufficient to compensate for the decreased hemoglol)in content of the red cells.

The decrease in hemoglobin synthesis is at the expense of Hb-A only: abnormal hemoglobins, found in heterozygous combination with thalassemia, are produced at a normal rate, so that defects of the heme synthesis or of other enzymatic mechanisms seem unlikely in these cases.

Abnormal hemoglobins are usually found in lower amounts than Hb-A in heterozygotes for an abnormal hemoglobin gene. It is assumed that the lower rate of synthesis of abnormal hemoglobins relative to Hb-A is the consequence of the gene mutation. The ratio Hb-A/abnormal hemoglobin is, however, quite variable, from close to 1 to 4 or more (see Section V,A,2). One can thus imagine instances where a hemoglobin gene mutation produces such drastic alterations that the corresponding abnormal hemoglobin cannot be made at all or is made at a very low rate.

Itano (1956), and in more detail Ingram and Stretton (1959b), have suggested that the drastic reduction in Hb-A synthesis in thalassemia may result from an alteration or an amino acid substitution (Ingram and Stretton, 1959b), not involving a change of the electrophoretic charge of the hemoglobin molecule.

Either peptide chain of Hb-A could be involved by such alteration and two types of thalassemia can be defined (Ingram and Stretton, 1959b) : (a) a chain thalassemia, allelic to the a locus, which impairs the synthesis of a chains; (b) p chain thalassemia, allelic to the f^ locus, which impairs the synthesis of (3 chains. This hypothesis provides an explanation for most of the alterations of the hemoglobin pattern in thalassemia. In the a chain thalassemia a reduced amount of ao dimers is available to combine with ^^^ and 82'^- dimers or with any abnormal type of /32 dimers, like jSo^ or j32^. Consequently, the rate of synthesis of Hb-A, Hb-Ao, and of any other hemoglobin containing a^ dimers will be equally lowered and the relative proportion of these hemoglobins will be unchanged from the normal; the a chain thalassemia is in fact of a non-increased Hb-Ao and non-interacting with /? chain abnormality type. On the contrary, the a chain thalassemia will interact with « chain abnormal genes, like a^ (Atwater et al., 1960b). The balance in the rate of synthesis of a chain versus ^, y, and 8 chains is profoundly altered in a thalassemia; /?•., yo, and 80 dimers are produced in large excess over

4')2 COKUADO HA(.I.I()X1

tlu' ;n-ailal)l(' a-., diinci".^. Aggregation of identical dinici's may in this case result in the formation of abnormal hemoglobin molecules, like ilh-H (^/) and Ill)-Bart's (y,^') (soc Section TV. 111. in /? chain thalassemia there is a deficit in the /? chain production; in the heterozygote f3-^/ft'^^\ for instanci', oidy the /f^ gene is ()|)erating and the i-ate of synthesis of f3 chains is lowered to approximately lialf the normal \'alue. Since the rate of synthesis of 8 chains is not affected by the presence of a thalas- semic gene, the relative amount of Ilb-A:.. appeal's to be rloubled over normal values. In sickle-ccll/thalassemia anrl in other combinations of a thalassemic gene with one of the (S alleles, such as /3^, f3^, or (3^, the rate of pi'oduction of /)"'. f3^\ or /3" chains, resl^ecti^•ely. is liigher tlian that of the /3 chains i)roduced under the control of the jff''"'' gene. fS chain thalassemia is thus of an increased Hb-A^ type and interacts with ft alleles.

The denomination thalassemia includes a group of hei'editary anemias heterogeneous fi'om a genetic point of view; thalass(>mia minor may be caused by heterozygosity for several possible a"^'' oi- /?''''' pseudoalleles and thalassemia major by homozygosity for two a^^' or /?''"'' genes or pos- sibly by interaction of one a'^^' with one ft'^^ gene. The heterogeneity of thalassemia is in accordance with the variability of the hematological and clinical picture; the possibility exists that different thalassemic genes have different expressivity and penetrance. A form of thalassemia of intermediate severity, whicli had been designated thalassemia inter- media, may possibly result from the combination of thalassemic genes with low expressivity; other factors besides the hemoglobin genotype may also control the expression of the thalassemic genes. The pheno- typic effects of the thalassemic genes are consistent with the hypothesis of the amino acid substitution involving non-charged amino acids (Ingram and Stretton, 1959b). However, no chemical alteration of the hemoglobin of thalassemic patients has been reported so far. Few research groups have been involved in the search for the hidden amino acid substitution in tlialassemia, and tlie author is among tlunn. In spite of extensive examination of the "Hb-A" from patients with thalassemia major, no alteration of the fingerprinting pattern of this hemoglobin, compared to Hb-A from noinial subjects, has yet been observed (C. Baglioni, V. M. Ingram, and x\. O. W. Stretton, unpublished). The analysis of the peptides of the "Hb-A" isolaterl from a sickle-cell/ thalassemic patient has not revealed any difference in amino acid com- position with peptides isolated from authentic Ilb-A (O. duidotti, per- sonal communication). The evidence against the amino acid substitution hypothesis is, how^ever, not conclusi\-e at all; moic wo'k and more chemi- cal data are neces^ai-v before this possihilitv can be exclude(|. Tm-ersions

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 453

of the amino acid sequence, for instance, are j^robably not revealed by the fingerprinting analysis or by the analysis of the amino acid com- position of peptides; the only conclusive chemical evidence is the deter- mination of the entire primary sequence of the "Hb-A" from a thalas- semic. To reach this goal may require years of hard work.

Moreover, very little is known about the effect of a change in the genetic information of a structural gene on the primary sequence of the corresponding peptide chain. Some mutations may not result in an altera- tion of the primary structure of a peptide chain, but only in a difficult "reading" of the genetic code: i.e., ambiguity in the genetic information may cause the production of a large proportion of senseless peptide chains unable to fold, or a difficulty in translating the genetic code may slow down the synthesis of a peptide chain. An alternative hypothesis has been advanced by Ingram and Stretton (1959b), based on Freese's (1958) idea of "connecting units" alternating in linear arrangement with structural genes. If these connecting units serve to induce or repress structural genes, then thalassemia may be the consequence of a change of these regulators. How^ever, thalassemia is not likely to be a mutation of an ''i" type gene, in the teniiinology of Jacob and Monod (1961); mutations of these genes exhibit a ti'ans effect, which is not observed in thalassemia. Only the synthesis of the a or /? peptide chain in cis with the thalassemic mutation is repressed. The validity of a comparison between the function of the genes of a bacterial chromosome with the genes of higher organisms' chromosomes is, however, questionable. The complexity of chromosomal structures in higher organisms and the larger dimensions of the genetic material may limit the diffusion of repressor or inducer molecules to neighboring groups of genes and mimic in this way a cis effect.

H. ABNORMAL HEMOGLOBINS FROM ALTERED AGGREGATION OF PEPTIDE CHAINS

Under this denomination are classified abnormal hemoglobins which are not properly described as variants of the normal hemoglobins.

1. Hb-H

Rigas et al. (1955) and Gouttas et al. (1955) described this electro- phoretically fast moving hemoglobin, characterized by instability and denaturation on storage. Hb-H is precipitated in the erythrocytes as inclusion bodies on incubation with cresyl blue.

Jones et al. (1959a) have shown that Hb-H is made up of four /3-^ chains. Motulsky (1956) pointed out the close association between Hb-H and thalassemia; Hb-H carriers usually pi-esent the blood picture char-

454 CORRADO BAGLIONI

acteristic of thalassemia and one parent is found to be affected by thalassemia minor. The thalassemia in the Hb-H carrier and in the affected parent is of the non-increased Hb-Aa type (Silvestroni et al., in the infants, when Hb-A substitutes Hb-F. The presence of Hb-Bart's in cord blood has been attributed in these cases to heterozygosity for a chain thalassemic genes. These genes may manifest themselves with ever, of no use for the carriers; its oxygen affinity is far too high to permit release of oxygen to the tissues (Benesch et al., 1961).

Hb-H is not inherited in a Mendelian way (Motulsky, 1956). The parent affected by thalassemia does not show, in the majority of the cases reported in the literature, any Hb-H; transmission from father to son has been reported (Fessas, 1959). Only a small fraction of the individuals presumably affected by a thalassemia are Hb-H carriers. The unpredictable appearance of Hb-H may be related to other factors, which affect the eiythropoiesis or the rate of destruction of red cells in subjects affected by a thalassemia. The amount of Hb-H is quite vari- able in the carriers and undetected amounts of this hemoglobin may be present in individuals classified as normal. The presence of Hb-H may also be caused by the interaction of two abnormal genes, one of which has a very low expressivity and does not ajipreciably depress hemoglobin synthesis in heterozygotes.

2. Hb-Bart's

Fessas and Papaspyrou (1957) and Ager and Lehmann (1958) reported an abnormal hemoglobin in infants with a thalassemic blood picture. This hemoglobin has been designated Hb-Bart's. Hunt and Lehmann (1959) have fingerprinted Hb-Bart's and have shown that tliis hemoglobin is made up of y'" chains only; Hb-Bart's seems to have the structural fornnila 74'''. Like Hb-H, Hb-Bart's has been found to be associated with a thalassemia. Baglioni et al. (1961) have shown the presence of Hb-Bart's in a newborn, who inherited a thalassemic gene interacting in his mother witli an a chain abnormality (Hb-I = a2 ^2^^) and thought to be an a^^' gene. Recent studies on the hemoglobins of newborns have shown that Hb-Bart's occurs with high frequency among newborns in some populations. Vella (1959) observed a fast abnormal component, presumably identical with Hb-Bart's, in 4% of the cord bloods of Cliinese newborns. Hcndi'icksc et al. (1960) have found a 10% frequency of Hb-Bart's in cord bloods of Nigerian newborns. Sim- ilar findings for the Negro population of the United States have been reported by Minnich et al. (1962) and by Weatherall and Boyer (1961). Lie-Injo (1961) has reported the occurrence of Hb-Bart's in Indonesians, Malays, and Chinese. The high fi-criucnoe with wliich Hb-Bart's has been

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 455

observed in Negroes is quite surprising, Hb-Bart's disappears almost at the same time that Hb-F disappears and usually no Hb-H is observed in the infants, when Hb-A substitutes Hb-F. The presence of Hb- Bart's in cord blood has been attributed in these cases to heterozygosity for a chain thalassemic genes. These genes may manifest themselves with the presence of Hb-Bart's only in periods of stress for the erythropoietic system, when the production of red cells is accelerated by environmental or physiological factors around birth. The frequency with which Hb- Bart's is observed in newborns would correspond to the true frequency of such genes in population, unaffected by expressivity or penetrance factors.

If the frequency of a^^ genes is as high as estimated by the above- mentioned authors, one wonders why homozygous cases of a-thalassemia have not been observed and reported in the literature. An answer to this question may be found in the description by Lie-Injo (1961) of several cases of severe hydrops and erythroblastosis fetalis in stillborn babies or in babies who died shortly after birth. The hematological picture was that of a severe hemolytic disease and the red cells showed sickling although no Hb-S was present. The hemoglobin of these stillborn babies consisted mainly of Hb-Bart's with small amounts of Hb-F and Hb-A (Lie-Injo and Lie, 1961). In a cord blood sample examined by the author (obtained through the courtesy of Dr. Lie-Injo, Institute for Medical Research, Kuala Lumpur, Malaya) approximately 80% of the hemo- globin was found to be Hb-Bart's! Lie-Injo (personal communication) has suggested that these babies were affected by homozygous «-thalas- semia. The failure to observe cases of homozygous a-thalassemia in adults may thus be due to the fetal or stillborn lethality of this geno- type. Since a chains are common to Hb-F, Hb-A, and Hb-Aa, the synthesis of these hemoglobins will be equally affected by a deficiency of a chain production. Moreover, no compensatory effect is possible in homozygous a thalassemia, while in /? thalassemia the production of high quantities of Hb-F allows the affected individuals to live for a few to several years.

It is interesting that no abnormal aggregation of « chains to form a hypothetical 0:4 hemoglobin is observed in a thalassemia. The a^ mole- cule does not form either because of steric hindrance or because of a lack of chemical affinity of the a chains to polymerize. It has indeed been observed that upon prolonged exposure of Hb-A to acid pH and recombination, small quantities of (3^ chains polymerize to form Hb-H and that an equivalent amount of a^ chains are liberated as monomers and not as «4^ tetramers (Huehns et al., 1961b) ; it is not known whether these a chains arc altered or partially denaturatcd.

456 CORRADO BAG LION I

3. Ilb-Augustd I and Ilb-Augusta II

Huisman (19(30) has reported the occurrence of a fast minor com- ponent in the cord blood of Negro babies, together with Ilb-F, llb-A, and Hb-S. Dissociation and recombination analysis of this minor com- ponent, called Hb-Augusta I, showed the i)resence of (3^ chains only and the structural formula /3.,^ was assigned to this hemoglobin. Family studies showed the presence of minute amounts of Hb-H in one parent and in some of the brothers of the propositus; it was concluded that the formation of Hb-Augusta I (J3i^) was caused by a genotype.

Preliminary evidence has been reported by iluisman (19(30j that a different minor component observed in the cord blood of a baby, pre- sumably heterozygous for the Hb-C gene and for a thalassemic gene {a^^'/a^ l3^//3^), is made up of /?'' chains only; this ft/' liemoglobin has been designated Hb-Augusta II.

V. The Control of the Synthesis of Hemoglobin

A. THE CONTROL OF THE RATE OF SYNTHESIS OF THE PEPTIDE CHAINS

1. Rate of Synthesis of Normal Peptide Chains

Approximately 95% of the red cell protein is hemoglobin. Tlie pep- tide chains of hemoglobins are thus synthesized at a rate much higher than that of any other protein in the red cells and the hemoglobin syn- thesis is one of the main aspects of the differentiation of stem cells into erythrocytes. Veiy little is known of the mechanisms controlling this differentiation. One can guess that in the course of differentiation the hemoglobin genes become activated in such a way that their products become the main components of the I'ed cells; however, how the activa- tion of a grouj) of genes is brought about selectively in mannualian tissues is not known.

In the normal red cells only "complete" hemoglobin molecules are observed, i.e., Hb-A = a^.-^jS-^, Hb-F = ao/ y./, or Hb-A, = av^ 8,"^ These hemoglobin molecules have in common a/'' dimers. During hemoglobin synthesis the amount of jSo'^, y.''", and 8-^- dimers jjioduced is eriuivalent to the amovmt of «., dimers produced. This raises the question of how this equilibrium in the synthesis of ])roducts of independent genes is maintained. It is known that when a defective a gene is present (as in a chain thalassemia), abnormal aggregation of ^o or y... dimers may be observed. It may thus be sujiposed tliat the (3 and y genes have an autonomous regulation of (3 and y chain production. In pathological conditions in whicli an equivalent amount of a cliains is not produced,

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 457

the ^ genes during adult life or the y genes during fetal life produce (3.i and y-z dimers in excess over the available an dimers. A similar situa- tion is not observed in conditions involving a decrease in /? chains production (as in ft thalassemia). No abnormal aggregation of a.. dimers is observed in this disease and no evidence of an accumulation of these dimers has been reported. The a gene seem thus to have a non- autonomous regulation. The amount of a chains produced seems to be dependent on the amount of y or /? chains available; if not enough (3 or y chains are made, the production of a chains seems to be lowered in a parallel and corresponding way. It is not known how this control is achieved; one can guess that steric hindrance factors, prohibiting the polymerization of a chains, are of importance, if the solubility of the dimers or of the single chains is low in physiological conditions. One can visualize an equilibrium between a peptide chains sitting on or in the proximity of the ribosomal templates, dimerization of the a chains, and release of the a-^ dimers upon combination with /J-, ys, or 82 dimers. The answer to these questions may possibly come from experiments on hemoglobin synthesis in reticulocytes or in cell-free systems (Dintzis, 1961) (see Schweet and Bishop, Chapter VIII).

Within the hemoglobin peptide chains there are extreme variations in the rate of synthesis. Hb-Ao (uo''" 82"^=) represents approximately only 2.5% of the hemoglobin in normal adults. For eveiy 40 ft-'' chains only one 8-^- chain is synthesized. This ratio is quite constant, within the limits of random variability, and is only found to be altered in favor of the 8 chains when ft chains are produced with lower efficiency, as in ft thalassemia. One wonders how this constant ratio is maintained and why the 8 chains, which are extremely similar to the ft chains from a chemical point of view, are produced with so much lower an efficiency. The existence of controlling genetic elements has been postulated by sevei'al authors (see Neel, 1961); in their view the synthesis of the 8 chains is repressed by a specific regulator gene or, vice versa, the syn- thesis of ft chains is activated to a higher extent by a specific regulator gene. Thalassemia may then correspond to a mutation of a regulator gene, rather than of a stmctural gene (Ingram and Stretton, 1959b). No clear-cut genetic or chemical evidence has so far been presented in support of this hypothesis.

2. The Rate of Synthesis of Abnormal Peptide Chains

The finding of Pauling et al. (1949) that an abnormal hemoglobin is present in lower amounts than Hb-A has been confirmed for all the abnormal hemoglobins discovered. The only possible exception to this rule reported in the literature is Hb-J (Thorup et al., 1956).

458 coRRADo ba(;lk)NI

'riu' rate of synthesis of an aluioiiiial peptide chain is, in general, lower than thai of the eorresponding normal chain. The lib-A/abnornial hemoglobin ratio, which is assumed to be e(|uivalent to the ratio normal chain/abnormal chain, is ([uite variabh' for different abnormal hemo- globins and for different individuals. Itano (1953) has reported the distribution of lib-S and llb-(" in members of several families, pointing out tlie \aiiability of the latio lib-A/Hb-S or Hb-A/Hb-C, and intro- ducing the concept of relative rate.s of synthesis as tlie determining factor in hemoglobin ratios. Itano (1953) suggests that the variability in the ratio Hb-A/abnormal hemoglobin may be dependent on the rela- tive ability to synthesize Hb-A, rather than on the different rate of synthesis of the abnormal hemoglobin. Chernoff (1956) has reported the distribution of Hb-S, Hb-C, and Hb-D in heterozygotes, observing con- spicuous variations in the amount of these hemoglobins relative to Hb-A. Several genetic and environmental factors are likely to deter- mine this variability; no definite correlation has, however, been estab- lished between the Hb-A/abnormal hemoglobin ratio and any specific factor.

Different abnormal liemoglobins arc present in different amounts relative to Hb-A. On examining data reported in the literature one observes a fairly wide range of variation; it may thus be supposed that the abnomial peptide chains are synthesized at a characteristic rate, determined by the specific alteration of these peptide chains. How does a single amino acid substitution affect the rate of synthesis of peptide chains 141-146 amino acids in length? At which stage in the synthesis of a peptide chain does the amino acid substitution become a rate- limiting factor? It does not seem likely that an amino acid substitution may decrease the rate at which the amino acids are aligned along the ribosomal template and bound in peptide linkage for the synthesis of a peptide chain. If this were the case, abnormal hemoglobins having the same type of amino acid substitution should have an identical rate of synthesis; this docs not seem to happen for Hb-C, Hb-E, and Hb-OArahia, which have a glutamic acid substituted by a lysine in different positions of the fi chain. The ratio Hb-A/abnormal hemoglobin in heterozygotes for Hb-C, Hb-E, and Hb-OArabia has been found to be different.

The following steps in the synthesis of the hemoglobin molecule are the folding of the individual [leptide chains, the dimerization of identical chains, and the association of symmetric dimers. As far as we know, any of these processes can be disturbed by the presence of a wrong amino acid in a peptide chain; however, the most likely to be disturbed seems to be the folding of a peptide chain to assume its secondary or tertiarv configuration.

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 459

B. THE CONTROL OF THE SYNTHESIS OF FETAL AND ADULT HEMOGLOBIN

Hb-F is the principal component of human fetus hemoglobin. The synthesis of Hb-A starts, however, early in fetal life; Hb-A has been detected in fetuses of 13 weeks (Walker and Turnbull, 1955). The level of Hb-A raises to about 10% by 22 weeks and to an extremely variable 20% at birth. The Hb-F level is fairly constant from the twenty-fourth to the thirty-second week of fetal life and then falls at a rate of 2.5 to 4% per week, Hb-F being substituted by Hb-A (Cook et al, 1957).

The level of Hb-F at birth in infants of different gestational ages has been detemiined by Cook et al. (1957) ; the concentration of Hb-F decreases with increasing gestational age of the newborns (Fig. 13). Babies born prematurely have higher levels of Hb-F, while babies bom after term have lower levels of Hb-F; i.e., babies born 4 weeks after teiTn seem to have Hb-F levels comparable to those of 4-week-old babies born at the nonnal term. Betke (1958) has also studied the fall of Hb-F concentration in newborns and has come to the conclusion that the Hb-F level falls as a function of gestational age and that birth has a negligible effect upon the variation of Hb-F level.

Hb-F is present for a variable time after birth (Chernoff and Singer, 1952; Jonxis and Visser, 1956); the Hb-F level decreases to about 1% or less by the end of the first year of adult life. The disappearance of Hb-F takes significantly longer than the time predicted on the basis of a complete cessation of Hb-F synthesis (White and Beaven, 1959) ; the production of Hb-F decreases after birth and is very limited in normal adults.

In the course of fetal differentiation, eiythropoietic activity resides in different morphological sites. Hemoglobin containing cells appear first in the mesenchyme of the embryo; the liver and the spleen are the next active erythropoietic sites until birth (Maximow and Bloom, 1957). The bone marrow is the main center of red cell production around birth; after birth, active bone marrow disappears from the long bones and the erythropoietic activity is limited to the red mari'ow of the steiTium and vertebrae (Maximow and Bloom, 1957).

No correlation has been established between the morphological sites of erythropoiesis and the switch from Hb-F synthesis to Hb-A synthesis. Betke (1958) has shown that Hb-F and Hb-A are both present in red cells obtained from liver and bone marrow of human fetuses. Thomas et al. (1960) have determined the hemoglobin types synthesized by fetal spleen, liver, and bone marrow; these hematopoietic tissues synthesize both Hb-A and Hb-F, although in a different ratio. More Hb-A is syn- thesized in bone marrow preparations than in spleen or liver prepara-

460

CORRADO BAGLIONI

"I {""■

V)

X

I-

z

in o

2

1 -

a K

1 J I

QQ. t

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 461

tions; bone marrow is the latest hematopoietic site in fetuses and red cells produced in bone marrow are presumably more differentiated toward the Hb-A containing red cell type.

Kleihauer et al. (1957) have devised a method to demonstrate Hb-F within single cells, by washing blood smears with buffer solutions, which selectively dissolve Hb-A; the hemoglobin left in the red cells is assumed to be Hb-F. By this method of analysis Kleihauer et al. (1957) have shown that Hb-A and Hb-F are both present in variable ratios in the same red cells. Moreover, under suitable conditions, all the red cells of sickle-cell anemic patients show the sickling phenomenon. Since relatively large amounts of Hb-F may be present in sickle-cell anemia, this prove sthat Hb-S and Hb-F are present together in the same red cells.

1. The Occurrence of Hb-F in Adults

Hb-F is present in some adults affected by hereditary or acquired hematological disorders. Hereditary conditions in which Hb-F is con- stantly found in adults include sickle-cell anemia, thalassemia major, and the anemias caused by a thalassemic gene in combination with a gene for an abnomial hemoglobin. Acquired pathological conditions in which Hb-F has occasionally been reported include pernicious anemia, hypoplastic anemia, leukemia, and other types of anemia (Singer et al., 1951; Chernoff, 1953; White and Beaven, 1959).

The level of Hb-F in anemic patients varies over a wide range; the highest levels of Hb-F have been observed in thalassemia major, where Hb-F frequently approaches 90-95% (Rich, 1952). A slight amount of Hb-F has been reported to be present in pregnant women (Rucknagel and Chernoff, 1955) and in women with hydatid moles (Bromberg et al, 1957), conditions in which high amounts of chorionic gonadotropin are secreted. These observations suggested that gonadotropin may stimulate Hb-F production (Rucknagel and Chernoff, 1955).

The appearance of Hb-F in hematological disorders has been con- sidered to be the consequence of severe and chronic anemia or a com-

FiG. 13. The switch from Hb-F synthe.sis to Hb-A .synthesis. indicates

the mean relative amount of y peptide chain present during the last months of fetal

life and the first 10 months after birth; indicates the mean relative amount

of y8 peptide chain; and the relative amount of a peptide chain during

the same period. The a peptide chain is con.stantlv 100% since this chain is common

to Hb-F and Hb-A; = percentages of Hb-F present at birth in relation to

gestational age (data from Cook et al., 1957). x x x x x = percentages of Hb-F pre.sent in children from 1 to 10 months of age (data from Chernoff and Singer, 1952).

462 CORRADO BAG LION I

pensatory phenomenon when the syntliesis of Ilh-A is im]jaircd, as in thalassemia major (Rich, 1952).

Hb-F does not seem to be evenly distributed in the red cells of anemic jKitients. Bradley et al. (1961) have reported that Hb-F and ilb-S arc unevenly distributed in the red cells of sickle-cell anemic patients. Similar results have been obtained by the Kleihauer et al. (1957) method of analysis by Thompson et al. (1961) and by Witchener et al. (1961). Identical observations have been reported for the distribu- tion of Hb-F in red cells of subjects affected by thalassemia major (Thompson et a/.. 1961). Two distinct cell populations are observed in these patients by this method of analysis, one with high amounts of Hb-F and the other with rather small amounts of Hb-F (Thompson et al., 1961). The Hb-F in the red cells of severe anemic patients seem thus to have a clonal distribution, with clones of red cells containing Hb-F and clones of red cells apparently of the normal adult type.

C. HEREDITARY PERSISTENCE OF Hb-F

Edington and Lehniann (1955), and subsequently other investigators, have reported a condition in which high proportions of Hb-F are found in healthy Negro adults. This condition has been designated "hereditaiy persistence of Hb-F" by Jacob and Raper (1958), since Mendelian inheritance of this anomaly was shown by family studies. The gene responsible for the persistence of Hb-F during adult life has also been called "high F" gene. In carriers of this gene 20 to 30% of the hemo- globin is Hb-F; approximately the same amount of Hb-F is observed in individuals heterozygous for the high F gene and for the abnormal hemoglobin genes yS^ or fi^. It is quite remarkable that no Hb-A is present in these heterozygous individuals (Jacob and Raper, 1958; Went and Maclver, 1958).

Several marriages of heterozygotes for the high F and jS^ or fS^' genes with normal partners have been reported (Bradley et al., 1961 ; Thomp- son et al., 1961; Kraus et al., 1961; Maclver et al., 1961) ; children born from these marriages inherit either the high F gene or the abnormal hemoglobin gene.

The high F gene thus behaves as if allelic or closely linked to the /3 locus. It has been suggested that the high F abnormality is the result of a mutation at the /3 locus and that the production of Hb-F is a com- pensatory phenomenon (Bradley and Conley, 1960). The high F gene in this view is interpreted as a thalassemic-like gene, suppressing com- pletely the production of p peptide chains. However, the majority of the heterozygotes for a thalassemic gene do not show any increase of the Hb-F level over normal values. Individuals heterozygous for the high

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 463

F gene and a j8 thalassemic gene have been reported (Kraus et al., 1961). The high F gene seems to interact with the Z^"^'' gene; 70% Hb-F and 30% Hb-A are present in these individuals. Similar levels of Hb-F are nonnally found in thalassemia major; but the high F/thalassemia condition is of intermediate severity and quite different from thalas- semia major as a pathological entity (Kraus et al., 1961).

Individuals with the high F sickle-cell condition have levels of Hb-F comparable to some of the sickle-cell anemic patients (Bradley et al., 1961 ) ; the ratio Hb-S/Hb-F is, however, quite variable in sickle-cell anemia. Sickle-cell anemia is a severe disease, while the high F/sickle- cell condition is almost asymptomatic (Edington and Lehmann, 1955; Jacob and Raper, 1958; Went and Maclver, 1958). Thompson et al. (1961) and Bradley et al. (1961) have reported that Hb-F is evenly distributed in the red cells of high F and of high F/sickle-cell individuals, in contrast with the uneven distribution of Hb-F in thalassemia major and sickle-cell anemia. This obser^'ation provides a clue to explain the absence of pathological manifestations in the high F/sickle cell condition, as compared to sickle-cell anemia. In sickle-cell anemia the cells with high Hb-S content sickle and lyse at normal physiological oxygen tension, while in the high F/sickle-cell condition the unifomi distribu- tion of Hb-F among the red cells prevents sickling at the same oxygen tension. Hb-F does not participate in the sickling process, since it does not form mixed aggregates with Hb-S (Allison, 1957).

The uniform distribution of Hb-F in red cells of high F individuals is of more profound significance, since it shows that the synthesis of high amounts of Hb-F is common to all the cells and is not characteristic of particular cellular clones, as it seems to be the case in severe anemias. The synthesis of Hb-F in the high F condition is the primary manifesta- tion of the "high F" gene and does not seem to result indirectly from severe anemia (Thompson et al., 1961).

Wheeler and Krevans (1961) have reported a child apparently homozygous for the high F gene; the child is in healthy condition and does not possess any detectable Hb-A or Hb-As. The absence of Hb-Aa in this child constitutes the only evidence that the high F gene suppresses the synthesis of ^ and 8 chains at the same time. The Hb-F of this child appears to be identical by fingerprinting to authentic Hb-F from cord blood (C. Baglioni, unpublished).

The hereditary persistence of Hb-F seems to be a genetic entity com- pletely different from the hereditary anemias in which high levels of Hb-F have been reported. It has been suggested that the high F muta- tion may involve a regulator or operator gene, rather than a structural gene (P. S. Gerald, unpublished; Ceppellini, 1961; Motulsky, 1962; Neel,

464 CORRADO BACLIONI

1961). The t'uiu'tion of such a i-cgiilaloi- ii,vnc is to turn off the y chain production and to turn on the (3 and 8 chain production. The fact that indivi(hials heterozytj;ous for the hijih F gene and for a f3^ or a /3*' gene foi-in Ill)-S or llh-C in noiinal amount seems to indicate that tlie regula- tor gene exerts its control only on the structural genes in cis, located on the same chromosome (Neel, 1961). The suppression of S chains produc- tion in the child liomozygous for the high F gene (Wheeler and Krevans, 1961) is in agreement with the idea that the regulator gene extends its control over the [3 and S structural genes, which are known to be closely linked (Ceppellini, 1959b).

The mode of action of the high F gene is analogous in form with that of the operator gene of Jacob and Monod (1961). The operator gene has been defined by Jacob and Monod (1961) as a genetic element which controls the expression of several adjacent structural genes; the operator as defined in the Lac system of Jacob and Monod is the site of action of repressors produced by regulator genes. The interaction of regulators, operators, and structural genes results in the repressive genetic regulation of protein synthesis.

In Fig. 14 the high F mutation has been indicated as o~ and the normal allele as o^, by analogy with the nomenclature of Jacob and Monod (1961). The close linkage of the postulated o gene with the jS and 8 genes is in agreement with the structure of the coordinated unit of genetic expression defined by Jacob and Monod (1961) as operon. It must be realized that the existence of an operator among the hemoglobin genes is not substantiated by genetic evidence minimally comparable to that obtained by the fine genetic analysis of the Lac region of E. coli. It is extremely useful, however, to borrow Jacob and Monod's model of the operon to explain tentatively the chemical findings for the high F condi- tion. This model provides the investigator with a challenging hypothesis, and may be apt to stimulate further experimentation and criticism.

In the operator gene model it is hard to understand the relationships between concomitant events: the suppression of /? (and 8) chain produc- tion and the persistent synthesis of y chains. In order to explain these apparently linked events, one has to assume that the /3-8 operator, which is controlled via repression by a regulator gene, in its turn controls the y operator. In this way the synthesis of y chains may be shut off at the same time that the synthesis of /? and 8 chains is activated. However, only one y gene, presumably the one in cis if y is linked to /? and 8, should fall under this control, since the normal operator in the hetero- zygotes does not suppress y chain production. Nothing is known about the position of tlie y locus, since y alleles have not been found. The operator gene hypothesis is still highly speculative in the case of the

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY

465

human hemoglobin genes and needs to be substantiated by more direct and critical evidence.

It cannot be excluded on genetic grounds that the high F condition results from an inactivation or a deletion of the two linked genes (3 and 8, however unlikely a recurrent genetic event of this type may be; the concomitant elevation of Hb-F still remains to be explained. It has been

70 % S

I I I I I I

si

0"

S

I !

30 %F

(r/) / \r' -r:

Fig. 14. Schematic representation of the postulateil genotype of incUvicluals heterozygous for the "hereditary persistence of Hb-F" and for the f3^ gene. The normal "operator" or "regulator" gene (Neel, 1961) is indicated as o*; the abnormal allele responsible for the persistence of Hb-F is indicated as o~. The arrows illustrate the dimers sjTithesized under the control of the hemoglobin genes and the hemo- globins produced by random combination of the dimers (capital letters). The (i and 5 genes in <?(« with the o~ mutation are supposedly inactive (see Section V,C) ; it is not known whether y chains are synthesized under the control of the y locus in cis with the o~ mutation or whether both the y loci are active. The relative levels of Hb-S and Hb-F observed in these individuals are also .shown.

suggested (V. M. Ingram, unpublished) that the difference between thalassemia and high F condition may reside in the production of a defective RNA-template in the former and in no production at all of RNA-templates in the latter, if the /3 and 8 genes have been deleted. If messenger-RNA molecules compete for the ribosomal sites of protein synthesis, in thalassemia the ribosomes would be occupied by the defec- tive messengers, while in the high F condition the ribosomes may be free to accept y chain RNA-templates.

466 CORRADO BAG LION 1

D. MODELS OF RED CELL DU FERENTLVriON AND SYNTHESIS OF Hb-F

Red cells are formed through a process of differentiation from nucle- ated precursors. A continuous series of morphologically well-defined intermediate stages of differentiation between undifferentiated precursors (stem cells) and red cells has long been recognized (see Wintrobe, 1962). Several models of stem cell differentiation have been proposed recently to explain the kinetics of red cell production and its modifications in pathological conditions (Alpen and Cranmore, 1959; Stohlman, 1959; Lajtha and Oliver, 1960). Every model of red cell proliferation indicates the necessity for the perpetuation of the primitive stem cells in the absence of differentiation. The stem cells appear to be morphologically and biochemically undifferentiated; they act as a reservoir of potential blood cells. Only a small proportion of stem cells per unit of time under- goes differentiation normally, while replication of undifferentiated stem cells provides the somatic continuity of the cell line.

Barnes et al. (1959) have transplanted 10^ nucleated cells from the spleen of an infant T6+ mouse to an irradiated CBA mouse; from the bone marrow of the latter the same number of cells were serially trans- planted into irradiated CBA mice, succeeding each time in populating the bone marrow of the CBA mice with cytologically distinguishable T6+ cells. This experiment provides direct experimental evidence of the unlimited proliferative ability of the stem cells.

It has been proposed by Alpen and Cranmore (1959) that the normal supply of red cells is, in large part, provided by the division of already differentiated eiythroid precursors, called erythroblasts or pronormo- blasts. Since the synthesis and the accumulation of hemoglobin initiates soon after the stem cells differentiate into erythroid precursors, these cells can be specifically labeled by means of the radioactive isotope Fe^^. The erythroblast is the earliest cell type that incorporates meas- urable amounts of Fe^^ ; the label incorporated can be estimated by auto- radiographic techniques in which the grains produced by single cells are counted; division of the cells causes halving of the grain count. Alpen and Cranmore (1959) have observed that some of the erythroblasts divide without differentiating further and suggested that the erythro- blasts may divide into a nearly equal number of cells which will dif- ferentiate further, and of cells which do not differentiate but provide the continuity of the cell line.

In animals made anemic by bleeding, a large number of stem cells differentiate rapidly into erythroblasts (Lajtha and Oliver, 1960) to compensate for the increased demand in red cell output. Whereas 24 hours after Fe^^ injection 100% of the erythroblasts are labeled in

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 467

normal animals, only 10% are found to be labeled in bled animals (Alpen and Cranmore, 1959) ; the average generation time of the erythroblast does not appear to be different in bled or control animals. The erythro- blasts seem to be a partially self-maintaining cell population, supple- mented by the differentiation of the stem cells. This implies that the eiythroblasts are able to make a choice between dividing only or dividing and differentiating.

A definite correlation exists between higher erythropoietic activity in pathological conditions and the appearance of Hb-F in adults. Al- though w^e are far from a full understanding of the process of red cell differentiation, some of the above-mentioned facts suggest a possible explanation for the switch from Hb-F to Hb-A production at the end of fetal life and for the synthesis of Hb-F in anemic individuals.

It may be assumed that the hemoglobin type synthesized by a red cell precursor is determined by the genetic infonnation possessed by that cell at the moment of initiating hemoglobin synthesis and during the course of hemoglobin synthesis. By analogy with protein synthesis in bacteria, we can speculate that the hemoglobin genes are repressed in stem cells and that when these cells differentiate into erythroblasts, specific hemoglobin genes become derepressed or activated by genetically determined mechanisms. The factors determining derepression or activa- tion of (3 rather than y genes are unknowTi. It has been suggested that the change in oxygen tension from the uterine environment to the extra- uterine environment after birth may be the activating mechanism for the synthesis of ft chains (Allen and Jandl, 1960; Thomas et al., 1960) or that gonadotropic hormones may play a similar role (Rucknagel and Chernoff, 1955). None of these proposed mechanisms accounts satisfac- torily, in the author's opinion, for the synthesis of Hb-F in pathological conditions of severe anemia during adult life.

By analogy with the observations in bled animals, we can speculate that in anemic subjects the stem cells differentiate into red cells after a limited number of cell divisions, while in normal erythropoiesis a con- siderable number of cell divisions occur at the eiythroblast stage before maturation of the red cells. It has been reported that stem cells may sometimes differentiate into red cells without intervening division (Suit, 1957). Similarly, in the fetus the red cells are released in circula- tion after a number of divisions of the erythroid precursors, limited by the age of the fetus itself and possibly by the accelerated fetal ejythro- poietic activity. A correlation may thus exist either between the number of divisions that the progenitors of a red cell have undergone in the absence of morphological differentiation or the time spent in the bone marrow by these cells as erythroblasts and the hemoglobin type con-

468

CORRADO IJAtlLlONI

•©

y- ^ bti

■^ ^ 5

	o	3 u
	'Z^	T "3
<	-f	.-2 ^
s	~	>' -c
LU	--^
2	A	.2 'S
<	W
	o	cc fl
		CJ ^
		OJ
	f^	CK (V|
	Jj	^— ' I
	w
		_£2
		^^ ^■*-t
	3	.5f o
	_s

	oo
in	4
O	()-
E 0)	O^
W	o

!/ 1/ 1/

■o-*o •o

© — ►©

o =^ §

CO

3

LU Ll_

a Si

3 x ^

:: — "^ cJ

-!i; CD

5 ^

" > o o

-- 3 'S O

s '- c- g

■^ a -^

•o

" H aj

(u

3

IX. GENETICS AND HUMAN HEMOGLOBIN CliEMLSTRY 469

tained by the daughter red cell. Red cells produced after few cell divisions from stem cells may possibly contain more Hb-F than red cells produced after several divisions of the erythroblasts. In this view the information directed to the hemoglobin chain synthesis changes with the aging and/or the replication of the erythroid precursors (Fig. 15).

Sublethal irradiation of monkeys (Tuttle et al., 1961) or of humans for therapeutic purposes (R. Ceppellini, personal communication), which causes rapid differentiation of stem cells (Lajtha and Oliver, 1960), has been reported to be followed by the appearance of Hb-F in the blood of the irradiated subjects.

The present hypothesis regarding the relationship between red cell differentiation and the switch from Hb-F to Hb-A production accounts formally for most of the observations made in recent years: the lack of effect of birth on the switch from Hb-F to Hb-A production (Cook et al., 1957) and the appearance of Hb-F in conditions of abnormal eiythro- poiesis. It fails, however, to provide an explanation of the mechanisms involved in the control of the hemoglobin genes. This seems to be a point of more general interest since it deals with the control of differentiation, the challenging problem of today's biology.

The Hb-F/Hb-A system may be useful in the study of mechanisms of differentiation, because of the well-defined chemical nature of the proteins involved. In the switch from Hb-F to Hb-A production, only the synthesis of one type of peptide chain, the y chain, seems to be discontinued, while the synthesis of a new type of peptide chain, the p chain, is initiated. Only two genes, the p and the y genes, are apparently involved in switching their activity; this fact may provide a unique opportunity for the study of the differentiation of a simplified genetic and biochemical system.

Acknowledgments

The author wishes to thank Dr. Vernon M. Ingram for reading this manuscript, for suggesting several improvements and corrections, and for discussion and advice. Thanks are due to all who have made available unpublished data and who have contributed to this chapter with personal communications. The author is grateful to the Medical Foundation, Inc., Boston, and to the National Science Foundation, Washington, who have supported his work with grants.

References

Ager, J. A. M., and Lehmann, H. (1958). Brit. Med. J. I, 929.

Aksoy, M, and Lehmann, H. (1957). Nature 179, 1248.

Allen, D. W., and Jandl, J. H. (1960). J. Clin. Invest. 39, 1107.

Allen, D. W., Schroeder, W. A., and Balog, J. (1958). J. Am. Chem. Soc. 80, 1628.

Allison, A. C. (1955). Science 122, 640.

Allison, A. C. (1957). Biochem. J. 65, 212.

470 CORRADO liAGLIONI

Alpen. E. L., ami Cranmoro, D. (1!)59). /// "The Kinetics of Cellular Proliferation"

(F. Stohlman, Jr., cd.), p. 290. Grune & Stratton, New York and London. Atwater, J., Schwartz, I. R., and Tocantins, L. M. (19G0a). Blood 15, 901. Atwater, J., Schwartz, I. R., Erslev, A. J., Montgomery, T. L., and Tocantin.s, L. M.

(1960b). New Engl. J. Med. 263, 1215. BagHoni, C. (1961). Biochim. vt Biophijs. Acta 48, 392. Baghoni, C. (1962a). J. Biol. Clam. 237, 69. Baghoni, C. (19621)). Biochim. cL Biophys. Acta 59, 437. Baglioni, C, and Ingram, V. M. (1961). Biochim. et Bio/ihijs. Ada 48, 253. Baghoni, C, and Lehmann, II. (1902). Nature, in press. Baghoni, C, Ingram, V. M., and Sullivan, E. (1961). Nature 189, 167. Bailey, I. S., and Prankerd, T. A. J. (1958). Brit. J. Haematol. 4, 150. Barnes, D. W. H., Ford, C. E., and Loutit, J. F. (1959). Sang 30, 762. Benesch, R. E., Ranney, H. M., Benesch, R., and Smith, G. M. (1961). J. Biol.

Chem. 236, 2926. Benzer, S., Ingram, V. M., and Lehmann, H. (1958). Nature 182, 852. Betke, K. (1958). Schweiz. med. Wochschr. 88, 1005. Boulard, C, Co.sset, A., Dcstaing, F., Duzer, A., Jonxi.s, J. H. P., Mullcr, C. J., and

Portier, A. (1961). Blood 18, 750. Bradley, T. B., and Conley, C. L. (1960). Trans. Assoc. A7n. Physicians 73, 72. Bradley, T. B., Jr., Brawner, J. N., Ill, and Conley, C. L. (1961). Bull. Johns Hop- kins Hasp. 108, 242. Braunitzer, G., Rudloff, V., Hilse, K., Liebold, B., and Miiller, R. (1960). Z. phy.^iol.

Chem. Hoppe-Seyler's 320, 283. Braunitzer, G., Hillschmann, N., Rudloff, V., Hilse, K., Liebold, B., and Miiller, R.

(1961a). Nature 190, 480. Braunitzer, G., Gehring-Miiller, R., Hillschmann, N., Hilse, K., Hobam, G., Rudloff, v., and Wittman-Liebold, B. (1961b). Z. physiol. Chem. Hoppe-Seyler's 325, 283. Bridges, C. B. (1936). Science 83, 210. Bromberg, Y. M., Yehuda, M., Salzbergcr, M., and Abrahamov, A. (1957). Blood

12, 1122. Cabannes, R., and Portier, A. (1959). In "Abnormal Haemoglobins" (J. H. P. Jonxis

and J. F. Delafresnaye, eds.), p. 51. Blackwcll, Oxford. Carcassi, V., Ceppellini, R., and Siniscalco, M. (1957). Haematologica Pavia 42,

1635. Ceppellini, R. (1956). Sangue 1, 1.

Ceppellini, R. (1959a). Acta Genet. Med. et Gemellol. 8, 47.

Ceppellini, R. (1959b). Ciba Foundation Symposium, Biochemistry of Human Ge- netics, p. 133. Ceppellini, R. (1961). 2nd Intern. Con}. Human Genet. Rome. Charlwood, P. A., Gratser, W. B., and Beaven, G. H. (1960). Biochim. et Biophys.

Acta 40, 191. Chernoff, A. I. (1953). Blood 8, 413.

Chernof!, A. I. (1956). Nero Engl. J. Med. 253, 322 (1955). Chernoff, A. I., and Liu, J. C. (1961). Blood 17, 54. Chernoff. A. I., and Singer, R. (1952). Pediatrics 9, 469. Clegg, M. D., and Schroeder, W. A. (1959). J. Am. Chem. Soc. 81, 6065. Cook, C. D., Brodie, H. R., and Allen, D. W. (1957). Pediatrics 20, 272. Crick, F. H. C, Barnett, L., Brenner, S., and Watts-Tobin, R. J. (1961). Nature 192, 1227.

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 471

Dintzis, H. M. (1961). Proc. Natl. Acad. Sci. {U.S.) 47, 247.

Drabkin, D. L., and Austin, J. H. (1935). /. Biol. Chem. 112, 89.

Drescher, H., and Kiinzer, W. (1954). Kli7i. Wochschr. 32, 92.

Dreyer, W. J. (1960). Brookhaven Symposia in Biol. 13, 243.

Edington, G. M., and Lehmann, H. (1955). Brit. Med. J. II, 1328.

Fessas, P. (1959). In "Abnormal Haemoglobins" (J. H. P. Jonxis and J. F. Dela-

fresnaye, eds.), p. 153. Blackwell, Oxford. Fessas, P., and Papaspyrou, A. (1957). Science 126, 1119. Fessas, P., Karaklis, A., and Gnafakis, N. (1961). Acta Haeviatol. 25, 62. Fessas, P., Stamatoyannopoulos, G., and Karaklis, A. (1962). Blood 19, 1. Field, E. O., and O'Brien, J. A. P. (1955). Biochem. J. 60, 656. Freese, E. (1958). Cold Spring Harbor Symp. Quant. Biol. 23, 13. Gammack, D. B., Huehns, E. R., Shooter, E. M., and Gerald, P. S. (1960). /. Mol.

Biol. 2, 372. Gammack, D. B., Huehns, E. R., Lehmann, H., and Shooter, E. M. (1961). Acta

Genet, et Statist. Med. 11, 1. Garen, A., Levmthal, C., and Rothman, F. (1961). J. cJiim. phys. 58, 1068. Gerald, P. S., and Diamond, L. K. (1958). Blood 13, 835. Gerald, P. S., and Efron, M. L. (1961). Proc. Natl. Acad. Sci. U.S. 47, 1758. Gerald, P. S., and Ingram, V. M. (1961). /. Biol. Chem. 236, 2155. Gerald, P. S., Efron, M., and Diamond, L. K. (1961). A?n. J. Diseases Children 102,

514. Goldstein, J., Guidotti, G., Konigsberg, W., and Hill, R. J. (1961). J. Biol. Chem.

236, PC77. Gouttas, A., Fessas, P., Tsevremis, H., and Xefteri, E. (1955). Sang 26, 911. Gratzer, W. B., and Allison, A. C. (1960). Biol. Rev. 35, 459. Gratzer, W. B., and Beaven, G. H. (1961). /. Chromatog. 5, 315. Halbrecht, J., and Klibansky, C. (1956). Nature 178, 794. Hasserodt, V., and Vinograd, J. (1959). Proc. Natl. Acad. Sci. US. 45, 12. HeUnski, D. R., and Yanofsky, C. (1962). Proc. Natl. Acad. Sci. U.S. 48, 173. Hendrickse, R. G., Boyo, A. E., Fitzgerald, P. A., and Ransome, S. (1960). Brit.

Med. J., I 611. Hill, R. J., and Craig, L. C. (1959). J. Am. Chem. Soc. 81, 2272. Hill, R. L., and Schwartz, H. C. (1959). Nature 184, 642. Hill, R. J., and Konigsberg, W. (1960). J. Biol. Chem. 235, PC21. Hill, R. J., and Konigsberg, W. (1961). /. Biol. Chem. 236, PC7. Hill, R. L., Swenson, R. T., and Schwartz, H. C. (1960). /. Biol. Chem. 235, 3182. Hitzig, W. H., Frick, P. G., Betke, K., and Huisman, T. H. J. (1960). Helu. Paediat.

Acta 15, 499. Hoch, H. (1950). Biochem. J. 46, 199.

Horlein, H., and Weber, G. (1948). Deut. med. Wochschr. 39, 476. Horton, B., Payne, R. A., Bridges, M. T., and Huisman, T. H. J. (1961). Clin. Chim.

Acta 6, 246. Huehns, E. R., and Shooter, E. M. (1961). Nature 189, 918. Huehns, E. R., Flynn, F. V., Butler, E. A., and Beaven, G. H. (1961a). Nature 189,

496. Huehns, E. R., Shooter, E. M., Dance. X., Beaven, G. H., and Shooter, K. V.

(1961b). Nature 192, 1057. Huisman, T. H. J. (1960). Nature 188, 589. Huisman, T. H. J. (1961). Biochim. et Biophys. Acta 46, 384. Huisman, T. H. J., and Meyering, C. A. (1960). Clin. Chim. Acta 5, 103.

472 CORRADO BAG LION I

Huisman. T. H. J., Punt, K., and Schaad. J. I). C. (lilGl). lil,,o<l 17, 747.

Hunt, J. A. (1959). Nature 183, 1373.

Hunt, J. A., and Ingram, V. M. (1958a). Bioclnm. ,1 liinphtj.s. Avhi 28, 546.

Hunt, J. A., and Ingram. V. M. (1958b). Nature 181, 1(K)2.

Hunt, J. A., and Injiiam. \. M. (1959). Ciba Foinnhil mn Si/m /idsium. IVtorlu niisl nj

of Human Gcmlics, ]>. 114. Hunt, J. A., and Ingram, V. M. (1960). liiochim. el Hioplujs. Acta 42, 409. Hunt, J. A., and Lohmann, H. (1959). Nature 184, 872. Ingram, V. M. (1956). Nature 178, 792. Ingram, V. M. (1958). Biochim. et Biophys. Ada 28, 539. Ingram, V. M. (1959a). Nature 183, 1795. Ingram, V. M. (1959b). Biochim. ct Biophya. Ada :36, 402. Ingram, V. M. (1961u). Nature 189, 704. Ingram, V. M. (1961b). //( "Plcnioglobin and lis .Mmorinalitics." (^ C Thomas,

Springfield, Illinois. Ingram, V. M., and Strctton, A. O. W. (1959a). Federation Prur. 19, 343. Ingram, V. M., and Stretton, A. O. \\'. (1959b). Nature 184, 1903. Ingram, V. M., and Stretton, A. O. W. (1961). Nature 190, 1079. Ingram, V. M., and Stretton, A. O. W. (1962). Biochim. et Biophyi^. Acta 62, 456;

63, 20. Itano, H. A. (1951). Proc. Natl. Acad. Sci. U.S. 37, 775. Itano, H. A. (1953). Am. J. Human Genet. 5, 34. Itano, H. A. (1957). Advances in Protein Chem. 12, 215. Itano, H. A., and Noel, J. V. (1950). Proc. Natl. Acad. Sci. U.S. 36, 613. Itano, H. A., and Robinson, E. A. (1959). Nature 183, 1799. Itano, H. A., and Robinson, E. A. (1960). Proc. Natl. Acad. Sci. 46, 1492. Itano, H. A., and Singer, S. J. (1958). Proc. Natl. Acad. Sci. l\S. 44, 522. Itano, H. A., Bergren, W. R.. and Sturgeon, P. (1954). J. Am. Chem. Soc. 76, 2278. Itano, H. A., Singer, S. J., and Robinson, E. (1959). Ciba Foundation Symposium,

Biochemistry of Human Genetics, p. 96. Jacob, F., Monod, J. (1961). /. Mol. Biol. 3, 318. Jacob, G. F., and Rapor, A. B. (1958). Brit. J. Haematol. 4, 138. Jones, R. T., SchroodcT, \V. A., Balog. J. E., and Vinograd, J. R. (1959a). ./. Am.

Chem Soc. 81, 3161. Jones, R. T., Schrocder, W. A., and ^'iiu.grad, J. R. (1959b). ./. Am. Chem. Soc. 81,

4749. Jonxis, J. H. P. (1959). In "Alnioiinal Haemoglobins" (J. H. P. Jonxis and J. F.

Delafresnaye, eds.), p. 114. Blackwell, Oxford, England. Jonxis, J. H. P., and Visser. H. K. A. (1956). Am. J. Diseases Children 92, 588. Kendrew, G. C, Watson, H. C, Strandl)erg, B. E., Dickerson, R. E., Phillips, D. C,

and Shore, V. C. (1961). Nature 190, 666. Kleihauer, E., Braun, H., and Betke, K. (1957). A7//,. Wochschr. 35, 637. Konigsberg, W., Guidotti, G., and Hill, R. J. (1961). J. Biol. Chem.. 236, PC55. Kraiis, A. P., Koch, B., and Burckett, L. (1961). Brit. Med. ./. 1434. Kunkel, H. G., and Beam, A. G. (1957). Federation Proc. 16, 760. Kunkel, H. G., Ceppellini, R., Miiller-Eberhardt, V.. and Wolf, J. (1957). ./. Clin.

Invest. 36, 1615. Kunkel, H. G., and Wallenius, G. (1955). Science 122, 288. Lajtha, L. G., and Oliver, R. (1960). Ciba Foundation Symposium. Ilaeviopoiesis:

Cell Production and Its Regulation p. 289. Lehmann, H. (1960). In "The Metabolic Basis of Inherited Disea.se" (J. B. Stan-

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 473

bury, J. B. W^ngaardcn, and D. S. Fredrickson, eds.), p. 1086. McGraw-Hill,

New York. Lengyel, P., Speyer, J. F., and Ochoa, S. (1961). Proc. Natl. Acad. Sci. U.S. 47,

1936. Licbokl, B., and Brauuitzer, Ci. (1959). Z. physiol. Chem. Hoppe-Seyler's 315, 271. Lie-Injo, L. E. (1961). Nature 191, 1314.

Lic-Injo, L. E., and Lie, H. G. (1961). Acta Haematol. 25, 192. Maclver, J. E., Went, L. N., and Irvine, R. A. (1961). Brit. J. Haematol. 7, 373. Matsuda, G., Schroeder, W. A., Jones, R. T., and Weliky, N. (1960). Blood 16, 984. Matthaei, J. H., Jones, 0. W., Martin, R. G., and Nirenberg, M. W. (1962). Proc.

Natl. Acad. Sci. U.S. 48, 666. Maximow, A. A., and Bloom, W. (1957). "Textbook of Histology," 7th ed. Saun- ders, Philadelphia. Minnich, V., Cordonuicr, J. K., Williams. W. J., and Moore, C. V. (1962). Blood

19, 137. Mitchener, J. W., Tliomjison. R. B., and Huisman, T. H. J. (1961). Lancet i, 1169,

1349. Morrison, M., and Cook, J. L. (1955). Science 122, 920. Motulsky, A. G. (1956). Nature 178, 1055. Motulsky, A. G. (1962). Nature 194, 607. Muller, C. J. (1961). A comparative study of the structure of mammalian and a\-ian

haemoglobins. Doctoral Thesis. Groningen, Holland. Muller, C. J., and Jonxis, J. H. P. (1960). Nature 188, 949. Muller, C. J., and Kingma, S. (1961). Biochim. et Biophys. Acta 50, 595. Mm-ayama, M., and Ingram, V. M. (1959). Nature 183, 1798. Neeb, H., Beiboer, J. L., Jonxis, J. H. P., Kaars Sijpesteijn, J. A., and Muller, C. J.

(1961). Trop. geogr. Med. 13, 207. Neel, J. V. (1949). Science 110, 64. Neel, J. V. (1956). Ann. Human Genet. 21, 1. Neel, J. V. (1961). Blood 18, 769.

Nirenberg. M. W., and Matthaei, J. H. (1962). Proc. Natl. Acad. Sci. U.S. 48, 1588. Pauling, L.. Itano, H. A., Singer, S. J., and Wells, I. C. (1949). Science 110, 543. Perutz, M. F., Rossman, M. G., Cullis, A. F., Muirhead, H., Will, G., and North,

A. C. T. (1960). Nature 185, 416. Peterson, E. A., and Sober, H. A. (1956). /. .4m. Chem. Soc. 78, 751. Poulik, M. D. (1957). Nature 180, 1477.

Prins, H. K., and Huisman, T. H. J. (1956). Nature 111, 840. Ranney, H. M. (1954). J. Clin. Invest. 33, 1634. Raper, A. B., Gammack, D. B., Huehns, E. R., and Shooter, E. M. (1960). Brit.

Med. J. II, 1257. Rhine.smith, H. S., Schroeder, W. A., and Pauling, L. (1957a). ./. Am. Chem. Soc.

79, 609. Rhinesmith, H. S., Schroeder, W. A., and Pauhng, L. (1957b). J. Am. Chem. Soc.

79, 4682.

Rhinesmith, H. S., Schroeder, W. A., and Martin, N. J. (1958). J. Am. Chem. Soc.

80, 3358.

Rich, A. (1952). Proc. Natl. Acad. Sci. U.S. 38, 187. Rigas, D. A., Koler, R. D., and Osgood, E. E. (1955). Science 121, 372. Robinson, A. R., Zuelzer, W. W., Neel, J. V., Livingstone, F. B., and Miller, M. L. (1956). Blood 11, 902.

474 CORRADO BAGLIONI

Rmknagcl, D. L.. and Chcrnoff, A. I. (1955). Blood 10, 1092.

Rucknagol, D. L., ami Neel, J. V. (1961). In "Progress in Medical Genetics" (A. G.

Steinberg, ed.), Vol. I, p. 158. Grune & Stratton, New York and London. Schnek, A. G., and Schroeder, W. A. (1961). J. Am. Chem. Soc. 83, 1472. Schroeder, W. A., and Matsuda, G. (1958). J. Am. Chem. Soc. 80, 1521. Schroeder, W. A., Jones, R. T., Shelton, J. R., Shelton, J. B., Cormick, J., and M<-

Calla, K. (1961). Proc. Natl. Acad. Sci. l'.S. 47, 811. Schroeder, W. A., SheUon, J. R., Balog-Shelton, J., and Cormick, J. (1962). I'mc.

Nail. Acad. Sci. U.S. 48, 284. Scott, E. M., and Griffith, J. V. (1959). Biochim. et Biophys. Acta 34, 584. Shooter, E. M., Skinner, E. R., Garhck, J. P., and Barnicot, N. A. (1960). Brit. J.

Haematol. 6, 140. Silvestroni, E., and Bianco, I. (1952). Blood 7, 429.

Silvestroni, E., and Bianco, I. (1958). Policlinico (Rome), Sez. prat. 65, 203. Silvestroni, E., Bianco, I., Muzzolini, M., Modiano, G., and Vallisnori, E. (1957).

Progr. med. 13, 705. Singer, K., Chernoff, A. I., and Singer, L. (1951). Blood 6, 413, 429. Singer, S. J., and Itano, H. A. (1959). Proc. Natl. Acad. Sci. U.S. 45, 174. Smith, E. L. (1962a). Proc. Natl. Acad. Sci. U£. 48, 677. Smith, E. L. (1962b). Proc. Natl. Acad. Sci. U.S. 48, 859.

Smith, E. W., and Torbert, J. V. (1958). Bull. John Hopkins lIo.y>. 101, 38. Smithies, O. (1959). Advances in Protein Chemistry 41, 65. Smithies, 0., Dixon, G. H., and Connel, G. E. (1961). 27id Intern. Conf. Human

Genet. Rome. Speyer, J. F., Lengyel, P., Basiho, C., and Oclioa, S. (1962). Proc. Natl. Acad. Sci.

U.S. 48, 441. Stohlman, F., Jr. (1959). In "The Kinetics of Cellular Proliferation" (F. Stohlman,

Jr., ed.), p. 318. Grune & Stratton, New York and London. Suit, H. D. (1957). J. Clin. Pathol. 10, 267. Swenson, R. T., Hill, R. L., Lehmann, H., and Jim, R. T. S. (1962). J. Biol. Chem.

237, 1517. Thomas, E. D., Lochtc, H. L., Jr., Greenough, W. B., Ill, and Wales, M. (1960).

Nature 185, 396. Thompson, R. B., Mitchener, J. W., and Huisman, T. H. J. (1961). Blood 18, 267. Thorup, O. A., Itano, H. A., Wheby, M., and Lea veil, B. S. (1956). Science 123,

889. Tuttle, A. H., Newsome, F. E., Jackson, C. H., and Overman, R. R. (1961). Science

133, 578. Valentine, W. N., and Neel, J. V. (1944). A.M.A. Arrh. Intrrnal Med. 74, 185. Vella, F. (1959). Nature 184, 272. Vinograd, J. R., Hutchinson, W. D., and Schroeder, W. A. (1959). J. Am. Chem.

Soc. 81, 3168. Walker, J., and Turnlnill, E. P. N. (1955). Arch. Disease Childhood 30, 11, 114. Watson, H. C, and Kendrew, F. R. S. (1961). Nature 190, 670. Weatherall, D. J., and Baglioni, C. (1962). Blood, in press.

Weatherall, D. J., and Boyer, S. H. (1961). Trans. Assoc, of Am. Physicians 74, 89. Weatherall, D. J., and Boyer, S. H. (1962). Bull. Johns Hopkins Hosp. 110, 8. Wheeler, J. T., and Krevans, J. R. (1961). Clin. Research 9, 168. Went, L. N., and Maclver, J. E. (1958). Blood 13, 559. White, J. C, and Beaven, G. H. (1959). Brit. Med. Bull. 15, 33.

IX. GENETICS AND HUMAN HEMOGLOBIN CHEMISTRY 475

Wilson, S., and Smith, D. B. (1959). Canad. J. Biochem. and Physiol. 37, 405. Wintrobe, M. M. (1962). In "Clinical Hematology," p. 38. Lea & Febiger, Phila- delphia. Zuelzer, W. W., and Kaplan, E. (1954). Blood 9, 1047. Zuelzer, W. W., Xeel, J. V., and Robinson, A. R. (1956). Progr. in Hematol. 1, 91.

Chapter X

Contributions from TMV Studies

to the Problem of Genetic Information

Transfer and Codmg

A. TSUGITA' AND H. FRAENKEL-CONRAT

Virus Laboralory, University of Calijornia, Berkeley, California

I. Introduction

II. TMV-Protein

A. Rod Structure and Cliain Configuration

B. Amino Acid Sequence .

III. TMV-RNA

I\'. Viral Infectivity ....

.\. Chemical Basis of Infectivity

B. Mechani.sm of Infection

C. Cell-Free Biosynthesis of TMV-Protein .... V. Natural Strains of TMV

A. Symptomatology of TMV and Its Strains .

B. Composition of Natural Stra'ns

C. Mixed Reconstitution

D. Structural Differences in the RXA of Different Sti'a-ns \T. Chemically Produced Strains

A. Modification of the RXA

B. Protein Composition of Chemically Produced Mutants

C. Virulence and Viability of Strains

\TI. The Coding Problem

References

477 479 479 484 487 489 489 490 490 491 491 493 495 496 497 497 499 508 509 517

I. Introduction

The firtjt de.sci-iption of a virus as a new kind of biological entity we owe to Iwanowski (1892), and particularly to Beijerinck (1899). who demonstrated that the tobacco mosaic disease was due to a repli- cating substance which was able to pass through bacterial filters. In the following 40 years many plant and animal diseases were attributed to this novel class of agents, termed the filtrable viruses. That bacteria could also harbor viruses was shown by Twort and by d'Herelle about 1915. However, until about 1930 few chemists or bioclieniists concerned ' Pre.sent address: Research Institute for Protein, Osaka University, Osaka, Jajian.

477

478 A. TSrciTA AM) II. IHAKNKKL-CUNHAT

themselves with these hi()k)gic:il agents which apijarently were difficult to s(>parate from cellular niateriah

This situation changed when, in 1935, Stanley announced that he had isolated the tobacco mosaic virus in i)ure and ciystalline form. The term virus then acquired a rlistinctive meaning in physicochemical terms. Soon several otiicr plant viruses were isolated and crystallized, and all were proven to consist of RNA and protein (Bawden and Pirie, 1937; Stanley, 1938). Greater difficulties were encountered in the isolation of animal viruses, but several of them have also been obtained in crj'stal- line form (Schaffer and Schwerdt, 1955). These, as well as the bacterial viruses, proved to contain either RNA or DNA. Through these develop- ments viruses became the subject of intensive physical and chemical, as well as biological, research. With the application of new techniques for the study of macromolecules, such as electron microscopy, X-ray or light scattering, and ultracentrifugation, the structure of many viruses came to be well understood.

The general pattern emerged that most of the simjile viruses, consist- ing only of RNA and protein, were either of approximately spherical dimensions, or rod-shaped. In all viruses, the nucleic acid was covered by the protein. The amount of protein per jiarticle varied greatly, but the amount of RNA was approximately constant: 1-2 million "molecular weight" per virus particle ranging fi-oni about 5 million for the turnip yellows mosaic virus to 100 million for the influenza virus (Frisch- Niggemeyer, 1956).

In recent years it has become evident and generally accepted that in the case of many of the simpler viruses the RNA is by itself infec- tious even though of low efficiency (Fraenkel-Conrat, 1956; Gierer and Schramm, 1956). Furthermore, it has been clearly established that the RNA carries the entire genetic information required for virus production. Indications have been obtained that intracellular release of the nucleic acid is one of the first events ujion virus inf(H'tion with the RNA viruses, as was so well demonstrated to be the case with the bacteriophages (Hershey and Chase, 1952). Thus, virus particles are now generally regarded as representing relatively small-molecular genetic nucleic acids wrapped for transport in specifically designed and genetically determined protein overcoats of greater or lesser functional sophistication. In the simple RNA viruses, the role of the protein may be solely structural, but even here evolution has produced some surprising results. Thus, the complete protein coat may represent not only an adequate protective covering for the RNA, but may also give the jiarticle properties of thermal stability and enzyme resistance which are remarkable, consider- ing it.^ predominant protein nature.

X. TMV STUDIES IN GENETIC CODING 479

Tobacco mosaic \'iru.s (TMV) was the virus first to be discovered, first to be isolated, and fii-st to be degraded into protein and infective RNA, and to be reconstituted from these components. Because so much more is known about this than most other viruses, this article will be largely devoted to TMV. We shall first discuss the ])rinciple features of the structure and function of the tobacco mosaic virus, and we shall then attempt to show in what manner the study of this virus has con- tributed and may continue to contribute to our understanding of the molecular aspects of genetics. ]VIore detailed discussions of various aspects of this field can be found in recent reviews (Fracnkel-Conrat and Ramacliandran, 1959; Schuster, 1960; Klug and Caspar, 1960; Gierer, 1960).

II. TMV-Protein

A. ROD STRUCTURE AND CHAIN CONFIGURATION

Tobacco mosaic virus (TMV) is a rod-shaped particle (radius 150- 180 A, length 3000 A) composed of 5% ribonucleic acid (RNA) and 95% protein and having a particle weight of 40 million. The model presented in Fig. 1 is largely based on data obtained through a thorough X-ray diffraction analysis (Franklin and Klug, 1955, 1956; Franklin et al., 1957; Klug and Caspar, 1960), particularly of a methyl mercury derivative of the virus (Franklin and Holmes, 1958). As shown in the figure, the rod is believed to be hollow along its axis (40 A diameter) and surrounded by protein units in a gently pitched helical array (49 units in 3 turns, pitch 23 A). The nucleic acid is threaded through about 2130 protein units, thus also forming a helix of the same pitch with a diameter of 80 A. This concept of the architecture of the virus particle is well supported by recent improvements in electron microscopic tech- niques (Nixon and Woods, 1960). Figure 2 shows such an electron micrograph of TMV.

The virus rod is disaggregated by many agents which disrupt sec- ondary bonds occurring in proteins, such as heat, urea, detergents, alkali, acid, phenol and acetic acid. Some of these agents supply methods for the isolation of one or the other of the two components in native or biologically functional state while the other component is more or less degraded or denatured. Of these methods, acetic acid is the most con- venient and simple for the isolation of the protein (Fraenkel-Conrat, 1957), and phenol for that of the RNA (Gierer and Schramm, 1956; Haschemeyer et al., 1959).

Although the minimum molecular weight of the protein component is about 18,000, as shown later, the protein as usually obtaincMJ exhibits

480 A. TSUtllTA AN'l) H. FRAKN KKI.-CONKAT

a molecular ux-ight of about 100.(X)0 (4Sl, corrcsi)on(ling to an aggre- gation product of six molecules (Sclirannn ct (iL, 19ool. The protein as (lisaugi-ciiated carefully 1)V alkali at low ionic strength, shows a molec-

Fiti. 1. Model of tobacco mosaic structure (Kiug and Caspar, 1960). The model represents protein .subunits (white bodies) aggregated by helical stacking and held together by RNA (black tube). Some protein units (jjeptide chains) have been removed to show the location of the RXA. The segment shown corresponds to about one-tenth of tlic total length, since one virus particle consists of about 2130 protein chains (nioleculur weight about 18,000), arranged in 130 helical turns.

ular weight of 36,000 (Newmark and Alyers, 1957), and at high pH, or in quite dilute solution, or in 67^ acetic acid, the protein shows the molecular weight of 18.000 characteristic for the monomer (Ansevin and LaulTer, 1959; Wittmann, 1959b). The electrophoretic behavior of the

X. TMV STUDIES IN GENETIC CODING

481

Fig. 2. Electron micrograph of TMV obtaiued by new technique of staining with phosphotungstate. The longest straight rod in the picture is 0.45 /i long, 50% longer than the typical length of RNA containing virus rods. The helical stacking of subunits and the central channel, here outlined by a coat of contrasting phos- photungstate solution, are clearly visible.

482 A. TSUfilTA WD H. I- R AKX KKI,-r( )N HAT

protein is also greatly dependent on ionic strength, j)ll, and tempciature (Schranun and Zillig, IBoo; Kramer and Wittniann, 1958 j.

The diflieultii's in obtaining i)liysicocheniical parameters for the TMV protein niononici' are due to its marked tendency to aggregate, thus expressing its biological function. With increasing ionic strength and temperature, and witli a i)H optinnun of about 5, the protein aggregates to discs, rods, and finally to rods of ever-increasing lengths, indistin- guishable by all criteria, exce])t that of irregular and, under cei-tain con- ditions, excessive length, from TMV jjarticles. These observations prove that the general architecture of the virus particle is a function only of the shape of the protein monomer, wliile the length of the particle, as we shall see later, is determined by the length of the polynucleotide chain.

If the shape of the protein subunit determines that of the virus, the cjuestion arises: what determines the shape of the protein? Recent studies with ribonuclease and other proteins have yielded evidence that the native conformation of a peptide chain may be a consequence of its amino acid sequence alone (White and Anfinsen, 1959). A similar conclu- sion appears justified in regard to the 3-dimensional structure of the TMV protein chain, on the basis of the report by Anderer (1959) that the denatured and structureless protein can readily be renatured by dialysis from 6 M urea or 0.1 A" NaOH solution. These observations have been confimied and extended in our laboratory. It thus appears probable that the shape of the protein monomer represents the thermodynamically most stable conformation of its peptide chain. It has not yet been demonstrated in detail what bonds account for this structure, but the occurrence of several chain segments folded into «-helices, interspersed with non-helical areas and bends due to proline, or an accumulation of those side chains which interfere with stable helix formation appears probable (Blout et al., 1960). Quite possibly the folds are stabilized by hydrophobic interactions of groups of ali])hatic side chains, as well as by ionically linked and hydrogen-bonded polar residues. It must be noted that 12 of the 16 serines occur within 15 residues from both ends of the chain which are, in turn, quite low in basic and acidic residues; further, the majority of the serines, leucines, and isoleucines, as well as of the primary amide groups, occur in clusters along the chain.

Certain specific bonds have been suggested as ])articipating in either or both, chain conformation and interchain bonding. Thus, the — SH group can be detected and titrated only in the presence of dcnaturating reagents, and it is particularly unusual in its behavior in the intact virus. When the virus is treated with iodine the — SH group becomes substituted. The resultant sulfenyl iodide ( — -SI), usually quite a labile

X. TMV STUDIES IN GENETIC CODING 483

group, is permanently stable in the virus rod but decomposes concomi- tantly with disaggregation of the virus. In contrast, the — SH group of the isolated protein is oxidized by iodine (Fraenkel-Conrat, 1955a). The — SH group of the virus does not react with p-chloromercuribenzoate, but is able to bind small mercurials (e.g., methyl mercuric nitrate) and the methyl mercury group is not dissociated from the \-irus by cysteine, in contrast to that bound to the isolated protein (Fraenkel-Conrat, 1959). These observations suggest that the — SH is located near one of the protein surfaces which are exposed upon dissociation of tlie virus, and that the masked — SH carries an H atom which can be replaced by mercurials without change in the tertiary structure. Such an — SH group may be engaged in hydrogen bonding by acting as acceptor to another X — H donor group. This would account for the unusual iiroperties of the sulfur.

Another type of bond probably contributing to the tendency of the protein to aggregate to rod-shaped particles involves certain carboxyl groups. The disaggregation of the virus has been found to be accompa- nied by the release of one to two H" ions per protein unit, and, conversely, H"" ions are consumed in the aggregation of the protein with or without RNA. A further interesting observation supporting the existence of this type of hydrogen atom is that about 2000 lead atoms are bound by the virus and that the introduction of lead seems to stabilize the inter- subunit bonding. It has also been observed by X-ray scattering that this lead is located at two distinct cross-sectional sites in the virus particle (Caspar, 1956; Klug and Caspar, 1960). On the basis of the above experiments it has been suggested that there exists a particular H bonding involving undissociated — COOH groups which may play an important role in the build-up of the cylindrical virus structure (Fraen- kel-Conrat and Narita, 1958).

The attractive forces between the nucleic acid and the protein units may be largely ionic bonds between the basic residues of the protein and the phosphate groups of the nucleic acid, but hydrogen bonds and other forces probably contribute to the stability of the coaggregate of the two components. No definite experimental approach has yet been applied to the study of these forces.

Thus, to summarize, the architecture of the TAIV particle is the consequence of protein aggregation reactions at two levels: the first, the establishment of a specific chain conformation of the protein unit, and the second, the aggregation of these protein units, at first by side- to-side attachment and then by helical stacking into stable rods with a minimal protein surface area and maximal bonding of each unit to its six neighbors.

484

A. TSUGITA AND 11. !• H \K\ KKL-COXRAT

B. AMINO ACID SEQUENCE

'I'lif aiiiiiu) acid cDiiipo.-itioii of llic TAIN' protein has Ween (l('tci-iiiiiu'(l by several means. Recently the composition has been (h'linitdy estab- lished by ion exchange chromatogi'ai)hy, using the autnnialic amino acid analyzer (Tsugita and Fraenkel-Conrat, 1960). Almost tiie same (mjiii- position was deduced also by Wittmann and Braunitzcr (19o9) by a summation of the compositions of the separated peptides resulting from tryptic digestion of the virus protein. The two analyses agreed well except for one isoleucine residue. Nine residues of isolcucine were detected after sufficiently long hydrolysis of the protein, as well as in the

TABLE I Ami.xo Acid Comtosition of N.^vtural Strains (jf TMV"

	Ciroup .A			B	C	D
.■\miiio
acid	TMV	masked	J14D1	YA	GA	Y-TAMV	Dahlemens-e	G-TAMV	HR
Asp	18	18	17	19	19	18	17	22	17
Thr	16	16	16	17	17	17	17	19	14
Ser	1«	16	17	14	16	15	16	10	13
Glu	It)	1(5	16	16	16	19	19	16	22
Pro	8	8	8	8	S	8	8	10	8
Gly	6	6	6	6	6	6	6	4	4
Ala	14	14	14	14	14	11	11	IS	IS
Cys	1	1	1	1	I	1	1	I	1
Val	14	14	14	14	14	15	15	12	10
lieu	'J	9	9	8	8	7	7	s	s
Leu	12	12	12	12	12	13	13	11	11
Tyr	4	4	4	4	4	5	5	6	7
Phe	8	8	8	8	8	8	8	8	6
Try	3	3	3	3	3	3	3	2	2
Lys	2	2	3	2	2	2	2	1	2
Arg	11	11	11	12	12	9	9	8	11
Met	0	0	0	0	C	1	1	2	3
His	0	0	0	0	0	0	0	0	1
Total	158	158	158	158	158	158	158	158	158
± from
TMV		0	2	3	3	8	8	17	17
° Numbers in boldface indicate	±f	rem — ♦
differences from the prototype.	V	-TAMV	1	IS	17
	=t	from
								G-TAMV	16

sequential study at Berkeley, and that this number rather than eight was correct was later confirmed by Anderer (1962). The main problems in amino acid analysis reside now in the purity of the material, and in the rate of the liberation or decomposition of the amino acids during hydrolysis rather than in the accuracy of the analysis. Table I shows

X. TMV STUDIES IN GENETIC CODING 485

the amino acid composition of the protein. It should be noted that the protein hicks histidine, cystine, and methionine and has but one cysteine.

The sequential analysis was initiated at the Virus Laboratory at Berkeley in 1952. As usual in the protein field, the structural study began with end group analyses, and in particular with the finding that carboxypeptidase liberated about 2000 threonine residues per virus particle (Harris and Knight, 1952, 1955). This finding served as the first clear indication that the minimal molecular weight of the virus protein represented the subunit or peptide chain weight, a fact which was later supported by physical methods. The structure in the C-terminal region was shown by partial hydrazinolysis to be -Pro-Ala-Thr (Niu and Fraenkel-Conrat, 1955a) and the terminal hexapeptide, -Thr-Ser-Gly- Pro-Ala-Thr, was detected in chymotryptic digests soon thereafter (Niu and Fraenkel-Conrat, 1955b).

The general methods for amino terminal analysis either failed to yield any knowledge or yielded incorrect information (Braunitzer, 1956). The difficulty was explained when Narita (1958) found that an acetyl group blocked the amino-terminal end of the protein. This finding leads to interesting questions concerning the mechanism and the timing of the acetylation of the chain in relation to its biosynthesis and con- cerning the mode by which this feature is controlled by genetic infoiTna- tion. This is a problem of general significance since terminal acetyl groups have since been found to occur not only in all plant viruses investigated, but also in several other proteins, e.g., ovalbumin, cyto- chrome c, and certain hormones.

Starting from identification of the end groups, the results of a systematic study of the amino acid sequence of the protein were reported from the Virus Laboratory in Berkeley in a series of papers (Gish et al., 1958; Ramachandran and Gish, 1959; Gish, 1959, 1960, 1961; Tsugita, 1960, 1962a,b; Tsugita et al., 1960) . An independent study by Anderer and co-workers at the Max-Planck-Listitut in Tubingen began to yield valid data in 1960 (Anderer et al., 1960a). The first step in the complete analysis is the splitting of the protein into definite peptides and the separation of these peptides from one another. For this purpose trypsin was the main tool employed by both groups, since other enzymes lack the necessary specificity. Countercurrent distiibution methods were selected for separation of the digested peptides at our laboratory, and more recently basic ion exchange column chromatography (Dowex 1) has been employed. Further purification was necessaiy to obtain most peptides in pure form. The structural studies on some of the individual peptides with as many as 41 residues are rather complicated but almost complete agreement has been reached on the structures of all but this

486 A. TSUGITA AND II. FRAENKEL-CONRAT

longest one by the two research groujjs, and the few differences remaining between our sequences (Tsujrita et al., 1960, 1962a) and the recently corrected ones of Andcrcr (1962) will probably be soon resolved.

Tn order to obtain the sequential arrangement of the peptides, Witt- niann (1960a) investigated the tiypsin split products of several natural mutants characterized by fewer arginine and/or lysine residues than the common strain. From the comjiarison between the number and composi- tion of the peptides obtained from the wild tyj^e and those from such mutants the order of some of the jx'ptides has tentatively been estab- lisluMl. The results were confirmed and extended bv studies of the basic

Acetyl N-Scf^Tyr— ••Ser-»'lleu-

•Pro-	NH2 10 15 -*• Ser -♦ Glu -*■ Phe -»-Val-»Phe -►Leu -»Ser-».Ser
Leu'	'•-Glu'*	20 1 //etv^— ^=*rO^— >as	^•♦-^/o'*— Try^ — Alo
■Glu-	NHp -♦Glu-	-►Ala— •(Ara)-*Thr	/V/Vj, 45 — ►Vol— ♦Glu— ►Vol
■Lys-	*-Try.	50 /V//^ 1— Val'4— Glu-«— Ser	■*— Phe-«— Glu<— (Ar^

„„,.-.„„ .)

V Leu-^Gly — ►Asp— ►Glu— ►Phe— ►Glu — ►Ihr — ►ijiu — ►Ciiu — ►Ala— •<Arg)— ►! nr— ►vol — ►(jiu — ►«! ->^

Cvai'« — I nr'W — VQi^ UIU'^ — rro'* — oef^— rru'^ — Lyb^— i ry^ — vui^ — oiu'^ — aef^ — nnf^ — oiu'^— tMrq) m (Arq)-frPhe-^Pro — ►Asp-^Sef — ►Asp-^Phe— •{Us)— ►Vol — ►Tyr— •{Q)— ►Tyr— ►Asp— ►Ala— ►Vol ■>^

90 85 80 }

C(ArQ)«— Thr-«— Asp-*— Phe-«-Ala-« — Gly«— Leu-4— Leu-*— Alo-* — Thr*— Vol-*— Leu-«— Pro*— Asp-*— LeiJ« NHp - 95 NH2 NH2 NH2 100 NH2 105

Asp-^(Argi)-^lleu — ►lieu— ►Glu — ►Vol— ►Glu — ►Asp-^GIu — ►•Alo— ►Asp-^Pro— ►Thr— ►Thr— ».AIa -^^

120 H5 . ~ no . J

(Alo*— Vol-* — Thr-* — Ala-*— Asp-*— Asp-*-Val-* — Arg-*— (Arq)* — ^Thr-*— Ala*— Asp-*-Leu^— Thr-*— Glu M 125 NH^ 130 ^-^ 135

lieu— ♦^rgi-*Ser — ►Ala— ►Asp— ►lieu — ►-■asD-^/eo — ►*»l/— ►vo/— ►cs/u— »-Leu— ►lieu — •(Arg)— ►Gly ^

150 145 ,^ 140 NH2 J

CI pii*-rj«-* — Crff-* — ^er-*-Ser-* — Glu*— Phe-* — Ser-* — Ser-*— (Ara>*— Asp-*— Tvr* — Ser-* — Gly* — Thr 4i

Vol— ►Try— ►Thr — ► Ser -►Gly— ► Pro-frAio

Fig. 3. Amino acid scqiieiK-e of protein from common TMV. This sequence is based on all information available in February 1962. It now appears (June 1962) that one inversion (residues 25, 26) is necessary to correct the structure previously published from our laboratory (Tsugita et al., 1960; Tsugita et al., 1962a). About five changes have been made in the structure reported by Anderer ct al. (1960b; Anderer, 1962). All but the positions of a few amide groups appear to be agreed upon.

peptides obtained by chymotrypsin and subtilisin which supplied the overlapping sequences needed to align all tryptic peptides in a unique order (Anderer et al, 1960b; Tsugita et al, 1960). By these efforts the amino acid sequence of the TMV-protein has been determined almost completely in each of the two laboratories, and since the minor uncer- tainties do not overlap, the entire sequence can be regarded as known

X. TMV STUDIES IN GENETIC CODING 487

(see Fig. 3). The amino acid sequence of no other polypeptide of similar size has been established at two laboratories, and thus with a similar degree of certainty. Some of the particular features of this structure have previously been mentioned in connection with the functional aspects of the protein.

III. TMV-RNA

The amount of RNA in TMV and most other simple RNA viruses corresponds to a molecular weight of about 2 X 10". In the case of TMV the evidence from X-ray diffraction and electron microscopic analyses suggests strongly that each virus particle actually contains but one molecule of RNA of that molecular weight (Franklin et al., 1957; Hart 1958). The RNA occurs as a single unbranched chain of about 6400 nucleotides running through the rod at a diameter of 80 A and of the same helical pitch as the protein units to which it is probably bound by 3 phosphates per subunit. The length of the RNA thus limits the length of the stable virus rod to 6400/3 protein units, or a particle weight of (2130 X 17,800) + (2 X lO'') = 40 X lO''. It should be noted that the helical path through which passes the RNA within the rod does not allow for any contact or interaction of bases (Ginoza, 1958) as indicated by the hypcrchromed state of the RNA in situ (Bonhoeffer and Schach- man, 1960).

A number of different conditions which disaggregate the protein units have been employed for the separation and isolation of the RNA. Since the RNA is more sensitive to H^ and OH" ions than the protein, only those methods which proceed near neutrality yield biologically fully active RNA. Detergents (Fraenkel-Conrat and Williams, 1955; Fraenkel- Conrat et al., 1957) and phenol treatment (Gierer and Schramm, 1956) fulfill this requirement, and the phenol method has found particularly wide acceptance because of its simplicity and reliability.

TMV-RNA, as isolated in such manner, appears to contain less than 1% of protein, and no other known contaminants. If the acidic clay bentonite is added in the isolation of the RNA, the resultant prepara- tion contains less than 0.05% of protein, and from its markedly increased stability properties it is concluded that nucleases probably represent the main proteinaceous contaminant of non-bentonite-treated RNA prepa- rations (Fraenkel-Conrat et al., 1961; Singer and Fraenkel-Conrat, 1961).

Sedimentation analyses of the RNA show a main component of about 30S, which corresponds to the expected molecular weight of 2 X lO*' for the complete RNA moiety of one virus particle (Gierer, 1958a,b; Boedt- ker, 1959). However, from 20-50% of the material shows a broad

488 A. TSUGITA AND II. FRAEXKEL-COXRAT

distribution of lower sedimentation rates, as it might be expected from random frafjments of the predominant molecuhir species, which alone appears to be infectious. That the infectious molecules were single- stranded and of the length corresponding to tiie highest S-values was most clearly demonstrated by Giercr's study of the kinetics of enzyme inactivation of the viral RNA (19r)7, 19o8a,b). Light-scattering and viscosity data support these conclusions. As stated, TAIV-RNA, like <j!.X174 DNA and unlike most other DNA's, exists in single-stranded form in the virus. It is also single-stranded if isolated at low ionic strength. In the presence of salts, and even at low concentrations of divalent cations, RNA undergoes marked conformational changes (Haschemeyer et al., 1959; Doty et al., 1959). The drop of UV absorbance (hypochromism), and the dramatic rise in optical rotation renders RNA similar to DNA, and these effects, as well as alterations in chemical reac- tivity, have therefore been attributed, by analogy, to helical double- strandedness involving about 607^ of the bases (Haselkorn and Doty, 1961). However, these effects are freely reversible by dilution, heat, Versene, and other agents, and it thus seems certain that the primary biological role of the RNA as a carrier of genetic information is not dependent upon such conformational features. The same is prol)al)ly ti'ue also for DNA.

The chemical study of TMV-RNA began with the analysis of its component bases. The four typical bases, adenine, guanine, cytosine, and uracil, were found in the proportions of 1.13:1.00:0.77:1.10 (Knight, 1952), and no minor components such as methylpurines or pseudo-uracil have been detected (Littlefield and Dunn, 1958). Attempts to further characterize the RNA in chemical terms have followed two paths. The determination of end groups and terminal sequences may be regarded as the beginning of a systematic sequence analysis. Methods for the identification of terminal residues are available, although they have heretofore not been applied successfully to any RNA exceeding a molecular weight of 100,000 (S-RNA). Such application is dependent on the availability of RNA of the highest possible purity in terms of free- dom from chain fragments or shorter oligonucleotides. For those end group methods which rely on a certain type of enzymatic digestion the absence of other nucleases from both the RNA and the enzyme employed is also of critical importance. These difficulties have largely been over- come in recent studies of this laboratory. Through the use of virus carrying much radioactive label (4-40 X 10'' cpm'mg of C^* or P''-) it was possible to show that neither end of the chain was phosphorylated. Adenine was found to occur at both ends of the chain, and the 5'-linked terminal adenine was preceded by a pyrimidine, probably uracil (Sugi-

X. TMV STUDIES IN GENETIC CODING 489

yama and Fracnkel-Conrat, 1961; Sugiyama et al., 1962; Whitfeld, 1962; Fraenkcl-Conrat and Singer, 1962).

The other approach is a quantitative study of the cUgestion products obtained from RNA with specific enzymes. Structural work along those Hues may be facilitated by recent improvements in oligonucleotide frac- tionation methods, such as the 2-dimensional mapping procedure of Rushizky and Knight (1960), and the column chromatographic method of Staehelin et al. (1959).

IV. Viral Infectivity

A. CHEMICAL BASIS OF INFECTIVITY

It was reported in 1955 that TMV-RNA by itself had little biological activity, but that the coaggregation of native TMV protein with TMV- RNA prepared at neutrality yielded particles which were several hundred times more infective than either component (Fraenkel-Conrat and Williams, 1955). At that time it seemed that infectivity was a property only of the complete virus, be it isolated from a plant, or reconstituted in vitro from its components. Further studies, however, revealed that the low infectivity found in RNA preparations represented an intrinsic property of the RNA itself, and that this infectivity was affected by the protein coating only in a Cjuantitative sense (Fraenkel-Conrat, 1956; Gierer and Schramm, 1956). This conclusion has been confirmed by others and has been supported by independent observations with other viruses including animal viruses, as reviewed by Colter (1958) and Fraenkel-Conrat (1962).

The main pieces of evidence for the intrinsic infectivity of the RNA were (1) that traces of ribonuclease destroyed the infectivity of the RNA, while proteases did not, (3) that the infectivity was sedimentable together with the RNA, and not at the high rate characteristic for the virus and (3) that a variety of chemical and serological tests failed to detect in the RNA significant amounts of contamination by virus protein or peptide. In addition, electron microscopy and other physical methods have been employed and all observations support the conclusion that the RNA carries the infectivity.

We may now raise the question of why the RNA has only about 0.1% of the infectivity expected from the activity of TMV. Similarly low infectivity is observed in RNA preparations from different sources and prepared by various methods. This low infectivity may be presumed to be due to the RNA being damaged in the course of purification. However, the fact that the low activity of the RNA is greatly increased upon reconstitution and then approaches that of the original virus

490 A. TSUGITA AND H. FKAKN KP:L-C0NRAT

(Fraonkcl-Coiirat and Sinf2;cr. 1957, 19')9) ilis])rove^; this explanation. On the other liand, tlie ])resence of rihonucleasc in the leaves of the test plant suggests the possibility of digestion and inactivation of the RNA during the process of infection. The recent finding that the infectivity of the RNA is greatly potentiated by assay in presence of bentonite, an agent which absorbs and inactivates many nucleases, seems to support this interpretation (Singer and Fraenkel-Conrat, 1961).

B. MECHANISM OF INFECTION

Our knowledge concerning the inti-accl hilar jirocesses concerned with plant virus replication is as yet quite fragmentary. It seems that these viruses lack the sophisticated infection mechanisms of bacterial and animal viruses, and cellular entry of the vims appears to require a mechanical wounding, which is frequently achieved through an insect vector. Soon after entry into a viable cell the virus appears to shed its protein coat. This, in the case of TMV, requires several hours and accounts for the fact that the infection process is faster when free RNA rather than when the complete virus is used as the inoculum (Siegel et al., 1957; Engler and Schramm, 1959; Fraenkel-Conrat et al., 1959). Subsequently, an eclipse period of a few hours can be observed, during which the RNA is believed to enter the nucleus and initiate its replica- tion (Wildman, 1959; Zech, 1961). The progeny RNA then seems to reenter the cytoplasm and may there initiate virus protein synthesis. While thus infective RNA appears to be formed prior to any complete vims, excess viral protein is often observed at later stages in infected cells. Thus many virus isolates contain virus-like particles lacking nucleic acid and thus void of infectivity. In the case of TMV, this so- called X-protein, first isolated by Takahashi and Ishii (1952), was more I'ccently shown to reconstitute fully active virus when combined with TMV-RNA in vitro (Takahashi, 1959).

C. CELL-FREE BIOSYNTHESIS OF TMV-PROTEIN

Nirenberg and ]\Iatthaei recently reported on an amino acid-incorpo- rating system from Escherichia coli which was stimulated by the addition of macromolecular RNA (1961). The system consists of the ribosomal fraction of E. coli, the supei'iiatant (containing S-RNA), an ATP generating system, and amino acids one or all of which are labeled with C'^ When the extent of amino acid incorporation into hot TCA- insoluble material was found greatly stimulated by the addition of TMV-RNA to the reaction mixture, the question arose whether tlie RNA might act as messenger and evoke the syiitliesis of TMV-protein in this cell-free system. This concept schemed to be suppoi'ted when a

X. TMV STUDIES IN GENETIC CODING 491

serological relationship between the product and anti-TMV serum was detected. A detailed analysis of the labeled material produced under the influence of TMV-RNA was thereupon initiated. When unlabeled TMV- protein was added as carrier after the incubation and subsequently reisolated by column chromatography, it contained and retained a significant fraction of the counts, if TJMV-RNA had been present during the reaction, but much less if another RNA or no RNA had been present. Upon tryptic degradation of the protein and separation of the typical peptides, the counts were found distributed in characteristic manner; thus, with phenylalanine or tyrosine as labeled amino acids the radio- activity was predominantly associated with the peptides known to carry the respective amino acid. These data seemed to establish rather firmly that a protein resembling TMV-protein was made by the cell-free E. coli enzyme system under the influence of TMV-RNA. However, one test suggested that there was something wrong with much of the product. While the carrier protein reisolated from control reaction mixtures com- bined with TMV-RNA in the usual way under the conditions of re- constitution, the protein carrying what was presumed to be newly synthesized TMV-protein formed very little stable virus. This strong inter- ference with the reconstitution reaction by the small amount of "product" present may be tentatively interpreted as evidence that a good part of this protein carried a structural imperfection which prevented it from participating normally in the aggregation reaction which leads to rod formation. If one thinks of the typically aggregating protein units as analogous to bivalent antibody molecules, then the in vitro synthesized protein might be regarded as monovalent. Such a mental image would readily explain the marked interference of the latter with the reconstitu- tion reaction.

To conclude this review of a number of quite recent and preliminary results, it now appears that TMV-RNA can act as messenger and evoke in a completely foreign system the synthesis of a protein which shows very great resemblance to TMV-protein in serological, chromatographic, and amino acid sequential tenns, but probably differs in some as yet unknown feature necessary for the specific native conformation which enables TMV-protein to form rods (Tsugita et al. 1962b).

V. Natural Strains of TMV

A. SYMPTOMATOLOGY OF TMV AND ITS STRAINS

Plant virus diseases in general are either systemic, i.e., they spread throughout the plant, or they result in localized damage. In the case of TMV, different varieties of tobacco and other hosts respond typically in

492 A. TSL'GITA AND H. I- UAKX KKL-CONHAT

one or the othiT manner. The local lesion response observed in Xicotiana glutinosa and the artificial hybrid, -V. talxu-uin var. Xanfhl, nc, serves for the quantitative assay of the virus, whih' the spi'cadin^ of the \ini> throughout the growing plant in A', tabacuni is most favorable for i.-ohi- tion of the virus in jireparative scale. Apart from the pui'poses of (juanti- tation and isolation, the host response plays a crucial role in the detection and cliaracterization of strains or mutants of tlic viius. Thus, many strains ^ive characteristically small and non-spreading lesions and can in this manner be differentiated on the local lesion host. Many strains give local lesions in N. sylvestris and Java tobacco, varieties on which connnon TMV spreads systemically. Finally, the nature of the systemic response of A'', tabacum and other varieties can be markedly different. Virus sti-ains are known which do not show any detectable symptoms, even though they build uji to high concentrations throughout the plant. Others give mild symptoms (mottling), or the typical blistered effect of the mosaic disease, which may be accompanied by more or less yellow discoloration of leaves. Finally, necrotic areas may develop upon infec- tion with certain strains, and leaves or even the whole plant may die as a consequence of the infection. Thus, the observation of the response of several varieties of tobacco usually permits the identification of a known strain or the characterization of a new strain. The same tech- niques serve to test the biological homogeneity of a strain or to free it from a contaminant. Repeated passage through a local lesion host at limit dilution gives the best assurance that one is dealing with a geneti- cally homogeneous population of virus particles. Further, the use of a differential host i.e., one which responds systemically to one strain and locally to another can rule out the simultaneous presence of both types of strains.

An additional criterion has been proposed for the differentiation of strains which also supplies data concerning the viability of such strains. This method consists in homogenizing individual lesions (8-day- old, on A', tabacum var. Xanthi, nc) and testing these homogenates after appropriate dilution on new Xanthi plants. The number of lesions which can be evoked by one parental lesion has been termed productivity and was found to vaiy over tenfold for different natural strains (Veldee and Fraenkel-Conrat, 1962).

Since mechanical wounding is required to initiate infection, the virus solution is usually rubbed onto the surface of the leaves in presence of an abrasive (i.e., carborundum). Any trace of virus accidentally de- posited on the leaf surface at any time prior to this rubl)ing through transmittal by agents such as the nozzle of a water hose can upon

X. TMV STUDIES IN GENETIC CODING 493

rubbing with an abrasive evoke lesions which then are not due to the strain used in the inoculum, obviously a serious source of error.

For the quantitative testing for infective virus particles solutions containing on the order of 1 to 10 X lO"*^ mg virus/ml are rubbed onto half-leaves of the Xanthi, nc variety, the most sensitive test plant, and the resultant lesion number compared to that given by a known virus solution, preferably applied to the opposite half-leaves of the same plants. Over the range of 5 to 50 lesions per half-leaf, the response is approximately proportional to the virus concentration, but in absolute terms of multiplicity plant virus infection is veiy inefficient. The fact that it takes somewhere between 10^ and 10' particles to produce a lesion or to initiate a systemic disease has been cause for some concern as to whether the population of virus particles is uniformly infective. How- ever, most workers in the field tend to attribute this inefficiency of plant virus infection to the crude mechanics of the wounding and infecting process. The fact that most of the virus which is applied and rubbed onto the leaf can be washed off a few seconds later without decreasing the resultant lesion count seems to support this interpretation.

B. COMPOSITION OF NATURAL STRAINS

Analytical comparisons of natural strains of TMV were initiated by Knight and Stanley in 1941 and the first complete amino acid analyses appeared in 1947 (Knight). The strains originally investigated, the masked (M), J14D1 (JD), yellow aucuba (YA), green aucuba (GA), Holmes ribgrass (HR), and cucumber viruses 3 and 4 (CV3 and CV4) were obtained from field isolates from various diseased plants, and all but the last two were classified as strains of TMV on the basis of physicochemical, serological, and biological tests.- Later Black and Knight (1953) investigated two series of more closely related TMV strains, in part derived under such greenhouse conditions from one another, that single-step mutations may be presumed.

Analyses of the proteins of natural strains were also reported from the Max Planck Institut fiir Biologic (Aach, 1958; Wittmann, 1960b), and nucleotide analysis were obtained by Black and Knight (1953), as well as by Markham and Smith (1950).

The base analyses of the RNA of different strains have revealed no differences in composition, quite in contrast to the RNA of different

^ Cucumber viruses 3 and 4 were classed as TMV strains on the basis of physico- chemical relationship. When chemical .studies (RNA composition, end groups, cysteine content, etc.) .set them sharply apart from all bona fide strains of TMV, Knight suggested reclassification (Knight, 1955).

494 A. TSUGITA AND H. FRAENKEL-CONRAT

viruses, which was found to vaiy widely (Knight, 1952, 1959). It appears to be a general phenomenon that within a family of virus strains the RNA composition seems alike, but it must be remembered that each base occurs about 1500 times per molecule and that the error of the analytical method is at best ±30 residues.

In contrast to the uniformity in RNA composition, the amino acid make-up of most TMV strains was found to vary. Table I summarizes the results of recent analyses (Tsugita, 19f)2a; Knight et al., 1962) which in general bear out the earlier data. These analyses show that TjMV strains can conveniently be classified in the following analyti- cal groups: A. One group of many strains which are either identical with TMV, or quite similar to it, showing 1-3 amino acid residues replaced by others, and always lacking histidine and methionine. To this group belong most of the seemingly one-step mutants detected under green- house conditions. B. Another group, exemplified by Y-TAMV, showing a large change in composition including one methionine and C-terminal serine (Knight et al., 1962). Apparently the Dahlemense strain analyzed by Wittmann (1960b) belongs to this group. C. G-TAMV (again quite different, with two methionines). D. The Holmes ribgrass strain [again different, including three methionines and 1 histidine (Tsugita, 1962a)].

All strains, even those showing a very different composition, appear to contain the same number of residues, namely, 158 per peptide chain. Almost all strains studied in our laboratoiy which differ in composition from T]\IV give local lesions on A^ sylvestris, rather than the systemic disease, but no other correlations between composition and symptoma- tology were detected. Thus, strains of the same group gave no symptoms on A'', tabacum, or killed the plant; and some of the symptoms due to strains of all groups could resemble one another on one or the other host.

Recent studies of representatives of these classes of TMV strains have not confined themselves to amino acid composition. For the pur- pose of locating the amino acid exchanges on the peptide map, analyses of the amino acid composition of the peptides derived from trypsin digest of the protein, as well as preliminarj^ studies of their amino acid sequences, have been initiated. Figure 4 shows comparative aspects of the tentative structure of parts of these proteins (Tsugita, 1962a). These data suggest some interesting generalizations. Although the difference in net composition between two strains may appear relatively small (e.g., 8 exchanges for Y-TAMV as compared to TMV), the difference in peptide composition is much greater, involving in the pair used as example more than 30 exchanges, i.e., 20% of the total amino acids. On the other hand, some particular peptide sequences such as Arg.Asp-NH2. Arg (90-921. Val.Tyr.Arg (69-71), and Arg.Arg.Val.Asp.Asp.Ala.Thr.

X. TMV STUDIES IN GENETIC CODING 495

1 ►|2h- 3— h*

TMV ——J

Y-TAMV ■

*|-^6— 1*|— 8-*|-9-|-H0*|-M-|— 12—

G-TAMV H r-

HR

r

Fig. 4. Location of differences in protein structure of natural strains of TMV. The peptide chain of common TMV, subdivided at the sites of trypsin attack into 12 peptides is illustrated (A), as well as segments of this structure in which differ- ences have been observed (thin line) in strains of groups B, C, and D (see text). It may be noted that peptides 5, 7, and 9 are the same throughout.

Val.Ala.Ileu.Arg (112-122) are not changed throughout the four groups, which indicates that these sequences might play an important role in the biological activity of the protein molecule. These data are as yet cjuite fragmentary, but it is believed that they will ultimately supply the basis for important genetic and protein structural considerations.

C. MIXED RECONSTITUTION

Although DNA has long been regarded as the sole genetic material on the molecular level, it has been unequivocally proven in recent years that RNA fulfills this function for the RNA viruses. This was first clearly demonstrated by means of the reconstitution technique (Fraen- kel-Conrat and Singer, 1957). Three kinds of natural strains and the normal TMV strain were chosen for this purpose. These strains could be distinguished from the normal strain by either their symptoms on plants or their amino acid compositions, or both. Both the protein and the nucleic acid components were isolated from common TMV and from the natural strains, and the various possible combinations of pairs of these were reconstituted and tested. For instance, with the HR strain the following couples of reconstitution experiments were made:

HR-protein + HR-nucleic acid (a)

HR-protein + TMV-nucleic acid (b)

TMV-protein + HR-nucleic acid (c)

TMV-protein + TMV-nucleic acid (d)

Each mixed reconstituted virus was allowed to infect tobacco plants and the resultant symptoms were observed, and the amino acid composi- tion of the progeny protein determined. The results showed clearly that the nucleic acid determined both the symptoms and the amino acid composition, so that the progeny from (a) and (c) seemed to be HR and those from (b) and (d) seemed to be TMV both chemically and biologically. In other words, the nucleic acid of TMV was experimentally proved to be the sole genetic material carr>'ing the complete information

490

A. TSUGITA AND H. FRAKXKEL-CONRAT

for its own structure as well as for the structure of its homologous protein, undisturbed by the jircscncc of heterologous protein in the infecting particle (see Fig. 5).

ORIGINAL STRAINS

DEGRADATION

REPLICATION PROGENY

Fig. 5. Mi.xed rcconstitution. The identity of tlu^ progeny protein, as compared to the two parental strains, symbolized as A and B, has been established bj^ sero-

lofiiial tests and amino acid analyses.

D. STRUCTURAL DIFFERENCES IN THE RXA OF DIFFERENT STRAINS

While differences in the amino acid composition of viral strains have long been known to exist, attempts to detect differences in nucleotide composition have been unsuccessful. However, as previously pointed out, an analytical error as low as ztO.5% would give the composition of TMV-RNA with an accuracy of only ±6 to 9 residues for each base. Therefore, one cannot conclude from the available data that the strains of a given virus are identical in composition, but only that they resemble one another. Actually, jiresent knowledge of the genetic cai)a- bility of the RNA strongly suggests that there must exist structural differences between the RNA of viral strains. Such differences could conceivably be due to base interchange and thus become evident only from sequential analyses rather than from the composition of the RNA.

Experimental searches for structural difference of the RNA from strains have only recently been begun. End group analyses of four strains have shown these to have the same tenninal adenine residues as

X. TMV STUDIES IN GENETIC CODING 497

the common TMV (Sugiyama and P'raenkcl-Conrat, 1962), even though the proteins of two of these show differences in the N-terminal region of the peptide chain. Further information has been obtained by Rushizky and Knight (1960) through comparative studies of enzymatic digests of various strain nucleic acids. Pancreatic ribonuclease, which splits only next to pyrimidines, has been found useful for this purpose. By the 2-dimensional mapping procedure (electrophoresis and chromatography) many of the smaller digestion products can be separated and quantitated (cytidylic and uridylic acid, as well as di-, tri-, and tetranucleotides composed of purines and terminal 5'-linked pyrimidines). These studies have revealed definite differences in the frequency of recurrence of some oligonucleotide sequences in natural strains not closely related to TMV.

VI. Chemically Produced Strains

A. MODIFICATION OF THE RNA

Ever since the early days of their isolation the viruses have been subjected to various reagents or conditions which modified their struc- ture (Stanley, 1936). Only more recently has it become evident that most of those agents which inactivated vimses did so by force of their affinity for, and reaction with, the viral RNA. The three main classes of modify- ing agents are (1) physical agents, such as ultraviolet light. X-rays, heat, and sonic oscillations; (2) the biological approach of allowing the virus to replicate in the presence of base analogs which then tend to become incorporated into the virus structure; and (3) a host of chemical reagents which affect the amino groups or substitute the rings of the purines or pyrimidines.

This subject has been critically reviewed by Schuster (1960), and will not be systematically dealt with here. The modification reactions usually lead to inactivation of the viiTises and thus are of some interest as antiviral agents, and for the purposes of vaccine preparation. In connection with the present chapter, the main interest in these agents resides in their ability to produce mutations and they will be discussed primarily from this viewpoint.

The effects of the physical agents are varied, ranging from chain breakage to modification of the pyrimidine rings, but none of these reac- tions appears to lead to the frequent production of mutants. The same is true for the incorporation of base analogs. Azaguanine appears to replace some guanine residues (Matthews and Smith, 1955) and 2- thiouracil (.Teener and Rosseels, 1953; ]\Iandel et al., 1957) as well as 5-fluorouraeil (Gordon and Staehelin, 1958) replace uracil, the latter uji

498 A. TSUGITA AM) H. !• RAKXKKL-CONRAT

to 50%. These evcnls aw accoiiiijaiiicd l)y more or lesjs extensive inaeti- vation (no inactivation with fluorouraeil ), but no mutants ha\c hcen obsen'cd as a consequence of these reactions.

The chemical modification reactions are performed (Urcctly with the virus, or jircferably with the isohited RNA. Formaldeiiyde adds to or condenses with the amino groups of the RNA (Fraenkel-Conrat, 19o5bj, and is a relatively inefficient inactivating agent, probably because of the reversibility of the reaction (Staehelin, 1958). Glyoxal and similar com- pounds have a particular affinity for guanine residues, and are better inactivating agents (Staehelin, 1959a,b). No mutants have l)een obtained as the result of aldehyde treatment. However, it must be stressed that the only technique which renders the detection of relatively infrequent mutants a not too laborious task, the differential host method, has not been systematically nor frequently applied to RNA modified by the agents discussed in this and the preceding paragraph.

The next class of agents to be discussed are the alkylating agents. These vary greatly in their affinity for RNA, with mustard gas being the most reactive and iodoacetate the least reactive of the series that was investigated (Fraenkel-Conrat, 1961). It appears probable that the introduction of one or two alkyl groups per mole of RNA (6400 nucleo- tides) causes inactivation, regardless of the reagent used. The most readily alkylatable sites are the guanine and adenine residues. When various alkylated RNA preparations were surveyed for the presence of mutants the methylated dcrivates were found to contain significant numbers of mutated molecules among the survivors, but all other bigger substituents appeared to be not clearly mutagenic under the same test conditions.

One other agent, A'-bromosuccinimide, which seems to introduce bromine predominantly into the pyrimidine rings, was reported in pre- liminary manner to be mutagenic to a similar extent as mcthylation (Fraenkel-Conrat, 1961). Hydroxylamine and similar agents which at- tack the pyrimidine bases have also been reported to effect mutations (Schuster, 1961).

The reaction which has proven particularly fruitful in tlie field of vims mutagenesis has been the deamination reaction by means of nitrous acid. Schuster and Schramm (1958) and Vielmetter and Schuster (1960) established that the purine bases were deaminated at similar rates and slightly faster than cytosine in RNA, and that one hit per 6000 nucleo- tides had a 50% chance of inactivating the molecule. Oierer and INfundiy (1958) subsequently showed that many of the non-lethal hits were muta- genic. The single-hit nature of the curve relating mutation fi-efjuency to reaction time showed that a single deamination could result in a muta-

X. TMV STUDIES IN GENETIC CODING 499

tion. Further work from other hiboratorics confirmed these observations, and it now appears that with average hit numbers greater than 1 or 2, almost all deamination survivors are mutants, though only a fraction of these give the differential host test (i.e., local lesions in X. sijlvestris or Java tobacco). It has been established beyond doubt that deamination is per se truly mutagenic, and that selection does not play a significant role in this phenomenon (Mundry, 1959).

Having discussed the reactions which were found to effect mutations in RNA, a consideration of the presumed nature of the mutagenic event, in terms of RNA structure, appears advisable. The reaction most fre- quently employed is that with nitrous acid. In the case of DNA, evidence was adduced that the two reactions involving the 6-amino position (adenine -» hypoxanthine, and cytosine—> uracil) can be mutagenic, but that deamination of the guanine is only lethal (Vielmetter and Schuster, 1960). This is interpreted in terms of the current concepts of the replica- tion mechanism based on the Watson-Crick scheme of base pairing. In RNA, and only in RNA, is nitrous acid an astonishingly effective muta- gen, leading to mutation of almost all survivors. This has been attributed to the obligatory mutagenic effect of the deamination of cytosine since this leads to another genetically fully competent symbol in the RNA language but not in the DNA language, i.e., uracil (Fraenkel-Conrat, 1961; Tsugita, 1961). In contrast, the deamination of adenine leads to a base equally unnatural to RNA and to DNA. Thus, nitrous acid acting on RNA is the only direct mutagenic reaction known. Its effect contrasts with the various indirect mutagenic reactions, such as the deamination of the purines, or of cytosine in DNA, the methylation of the purines, the action of i¥-bromosuccinimide or hydroxylamine on the pyrimidines, the biological incorporation of base analogs into DNA, and others. All of the indirect mutagens are believed to increase the probability of a mis- reading of the coded information, but only the deamination of cytosine is believed to directly alter that information.

B. PROTEIN COMPOSITION OF CHEMICALLY PRODUCED MUTANTS

When mutants became available through chemical modification of the RNA, research programs were initiated, both at our laboratory and at the Max-Planck-Institut fiir Biologie, to determine the composition and the structure of the proteins of these mutant strains. Since the amino acid sequence of the wild type protein was known, and analytical methods for the detection and localization of single amino acid exchanges were available, there was reason for hope that this approach would reveal significant information concerning the relationship between the structure of the template RNA and tlie protein product.

500 A. TSUCMTA AND II. I- KAKX KF.!.-('< )\HAT

Tilt' iiu'tlit)(l> ciiiplox'cil l)y us wcic hiiclly as follows. The IvXA was sul)jt.'C'tc'(l to one of a \ai-i('t\- of cliciiiical agents, then freed of tlie reagent, and i-cconstiluled with untrt'atcd pi-otein. Only rarely was the conipU'tt' \-iiiis subjected to the cheinical treatment. As pre\-i()usly stated, only methylation and hroniination yielded significant numbers of nuitants while dcamination mutated almost all survivors. The extent of chemical modification was gaged from the inactivation, 37, 14, 5, and 2% residual infectivity being regarded as indicating 1, 2, 3, and 4 inactivating hits per average molecule (e~^ = 0.37). Previous studies had demonstrated that each chemical modification, as ascertained ana- lytically, had at least a 50% chance of being an inactivating event.

Two techniques were employed for the detection of mutants. Dif- ferential host mutants, as defined by Siegel (I960), are those which are detected as local lesions on .V. sijlvestns, a systemic host for the wild type virus. Such local lesions were individually excised and re- peatedly transferred to the same host until they produced local lesions unaccompanied by systemic symptoms and thus appeared to have been freed from contaminating wild type virus particles. Symptom mutants are detected by inoculating the reconstituted RNA onto a local lesion host, Xanthi, nc, and then transferring homogenates of randomly selected (or all) individual lesions to N. tabacum, N. sylvestris, and again to Xanthi. IMutants are distinguished by the appearance of unusual num- bers or shapes of lesions on Xanthi, nc, unusually mild or severe symp- toms on the systemic hosts, or other odd symptoms, singly or combined. To verify the genetically stable nature of any such symptoms, the virus was usually passed at least once through a local lesion and again onto that host which showed a distinctive reaction. Stable mutants were then transferred to 24-72 plants of .V. tabacum and virus harvested from these after 21 days. The symptomatology of th(> nuitant progeny was again verified by plant tests.

The mutant virus was split with acetic acid, and its protein was isolated. Samples of 4—5 mg of protein were hydrolyzed with 6 N HCl at 108=!=1°C for 24 and 72 hours. It is important that the acid be evaporated in a few minutes to avoid losses in serine and glutamic acid. One milliliter of the residue, redissolved to about 1 mg 'ml, was ai)plied to the amino acid analyzer (Tsugita and Fraenkel-Conrat, 1962).

The f|uestion arises whether direct analysis of hydrolyzates of a protein of molecular w'eight 18,000 can yield amino acid composition data sufficiently accurate to detect a single amino acid change. Since the most frequently occurring amino acid in TAIV is aspartic acid with 18 residues per mole, an analytical error of ±2.59^ would repi-esent the limit for the accurate detei-mination of this integral number. An accuracv

X. TMV STUDIES IN GENETIC CODING 501

better than this (less than ±2%) can generally be obtained with the analyzer and repeat analyses performed on each of two hydrolyzates can supply enough data to establish a reasonable degree of certainty.

In contrast to our primary use of direct analysis (Tsugita and Fraenkel-Conrat, 1960) Wittmann continues to determine the composi- tion of strains by the summation of the analyses of the peptide fragments obtained by trypsin digestion. That approach is less dependent on high accuracy in the analytical techniques, but the requirement that each peptide be obtained in a state of high purity makes this method laborious and expensive. The fact that the analytical results for common TMV obtained with the whole protein are in agreement with those obtained by peptide summation (see Section II,B) validates both methods; the direct method, being the simpler, thus appears preferable for the primary sur- vey of a large number of mutant proteins.

For mutants showing differences in composition from the wild type the question arises as to which particular amino acids are exchanged along the peptide sequence of the protein. Obviously it is important to establish the nature and the location of an exchange of amino acids, including an identification of glutamine and asparagine as contrasted to glutamic and aspartic acid.

For this purpose 20-300 mg of the proteins were digested by trypsin in the usual manner. From the digest the isoelectrically insoluble peptide was separated, and the soluble peptides were fractionated on Dowex 1X2, using a volatile solvent system similar to that used by Wittmann and Braunitzer (1959), as shown in the caption of Fig. 6. A typical pep- tide elution pattern is shown in that figure. A few of the peaks appeared to consist of a single peptide as shown by rechromatography and analyti- cal composition. Others were mixtures, but upon analysis of aliquots from the rising or falling edge of the curve the composition of the pure components was often closely approximated. For preparative isolation and definitive analyses such mixtures were rechromatographed. For the purpose of amino acid analysis of peptides the 50-cm and 5-cm columns of the analyzer were usually used rather than the 1 50-cm and 15-cm columns. This procedure has the advantage that the time for analysis is but one-third that of the normal procedure and that only one-tenth as large a sample is required (0.01-0.3 ^umoles), without much loss in accuracy (d=5%) (Tsugita, 1962a).

Any peptide showing a change in composition from that of the corresponding peptide of the normal strain was subjected to the usual variety of techniques of protein chemistry, i.e., digestion by proteinases or peptidases and chemical methods including dinitrophenylation, phen- ylisothiocyanation. and hydrazinolysis. Thus, the localization of the

502

A. TSUGITA AND H. I'RAENKEL-CONRAT

specific exchange was usually possible. Ccitain changes, involving resi- dues near the .V-acetyl end or the C-teiiuinus, could be ascertained by application of similar methods to the intact protein.

AVe shall now proceed to consitler the results of these studies (see Table II). The first important observation is that many mutants (about

Fig. 6. Chromatugnipliic separation of peptides in soluble fraction of tryptic digest of TMV protein. Buffer syslei7i:

1. 40 ml collidine + 40 ml pyridine + 1.5 ml acetic acid (pH 8.2) (180 ml).

2. Gradient (by Antegrade) (Technicon Chromatography Co., New York) Cham- bers 1 and 2: 40 ml collidine + 40 ml pyridine + 8.0 ml acetic acid (pH 7.3); 3: O.l N acetic acid; 4- 0.5 N acetic acid; 5: 0.5 N acetic acid; 6: IN acetic acid; 7: 2N acetic acid; 8 and 9: glacial acetic acid (each chamber contains 120 ml).

Column, 0.9 X 150 cm, Dovvex 1x2 (200-400 mesh) ; flow rate 30 ml/hour by Mini pump (Milton Roy Co.); fraction, 3.3 ml; 80 mg protein.

The numbers identify the main component of each peak with the secjuential peptide number on the polypeptide chain; 1 is not present because it is the N-terminal I-peptide removed by centrifugation prior to chromatograph3\

50% of those analyzed) did not differ from connnon TMV in their amino acid composition. Thus, the original report of Wittmann (1959a) con- cerning a mutant w'hich showed the same composition as common TMV has been confirmed.

Another class of mutants was observed which showed the replace- ment of 1-3 amino acids. In no case was the total number of i-esidues found to be changed. Certain amino acids were observed to be frequently substituted, and frequently by a particular one; others have not as yet been found to be replaced. Although our studies covered mutants ob- tained with three different types of reactions (deamination, alkylation, and bromination), the extent and the nature of the observed changes were similar. Thus, 1 deamination, 4 bromination, and 3 methylation mutants showed the replacement of 1 proline by a leucine, which in-

X. TMV STUDIES IN GENETIC CODING

503

TABLE II Amino Acid Exchanges in Chemical Mutants of Common TMV"

Mutagen	Mutant No.	Amino acid exchanges	Location
HNO2	273	Asp* -^ Ser	Peptide #1
HNO2	282	Asp* -^ Ala Thr -^ Ser	Peptide #10 Peptide #1
HNO2	171	Pro — > Leu Asp* — > Ala Thr ^Ser	Residue #156 Peptide #10 Peptide #1
HNO2	332	Pro -^ Leu	Peptide #1
HNO2	220	Ser ^Phe	Residue #138
HNO2	329	Ser ^Phe Ser ^Phe	Residue #138 Peptide #12
HNO2	237	Arg -^Gly Arg -^ Gly	Residue #61 Residue #134

HN02

NBSI

3216"=

326

Arg -*Gly Arg -^Gly Arg -. Gly

Residue #61 Residue #134 Residue #122

HX02	284	Glu Arg	^Gly — * La's		Residue #97
HN02 HN02	328 262	16 amino acids exchanged in many peptides; compo- sition identical with G-TAMV (see Table I)
NBSI	218	Pro	-^ Leu		Peptide #1
NBSI	233''	Pro	-^ Leu		Peptide #1
NBSI	207	Pro	-H> Leu		Peptide #1
NBSI	235	Pro	—> Leu		Peptide #1
NBSI	187 '^	Arg Asp*	-Gly ->Ser		Residue #46 Peptide #1

Asp* —* Ser Asp* — > Ser Asp* -^ Ser

Peptide #1 Peptide #1

Table continued

504

A. TSUGITA AND H. FRAENKEL-CONRAT

TABLK IT {Conliniied)

Mutagen	Mutant No.	AiiiiiKi acid exchanges		i>ocation
NBSI	830	lieu ^ Thr
NRSI NRSI	2()G	17 exchanges, similar to ( 1 Ala -^ Gly	:;-i	IWMV. l.ut for
NBSI	331	7 exchanges, similar to Y 1 Leu — > lieu	-T	AMY (Table I), but for
DMS	214	Pro — * Leu		Peptide #1
DMS	278	Pro — ► Leu		Peptide #1
DMS	176	Pro — > Leu		Peptide #1
DMS	215	Ser -Fhe		Residue #138
DMS	178	Arg -^ Gly
P.O/	249	16 amino acid exchanges sition identical with G-'	in many pejjtides; compo- TAMV

° Besides these 27 strains showing differences from common TMV, which in most instances were confirmed also bj' the analyses of progeny isolates, 29 other strains showed no detectable differences in composition from the parental strain.

* Not known whether aspartic acid or asparagine residue.

" Secondary change from strain 237.

'' Secondar}' change.

' Propylene oxide.

volved not the same near C-terminal pi'oline previously reported to be replaced in 1 nitrous acid mutant (Tsugita and Fraenkcl-Conrat, 1960) ; and 1 particular serine (#138) was found replaced by phenylalanine in 1 methylation and 2 deamination mutants. These data are summarized on Table II.

A survey of the location of changes on the peptide map indicates that these occur at sites scattered over the entire sequence. Where multiple changes were observed, these were not in vicinal j'jositions. While the total number of located changes is too small to justify any far-reaching conclusions, it would appear from the data illustrated in Fig. 7 that the distal parts of the peptide sequence are more prone to change than the central ones. It must be stressed, however, that this may have no genetic cause but be due to selection, in that mutation resulting in the exchange of specific amino acids might not give a functional or stable protein, and

X. TMV STUDIES IN GENETIC CODING

505

thus i)i'oducc either no virus progeny, or too little matci'ial for chemical analysis.

We have discussed one class of mutants showing the same amino acid composition as connnon T]\IV, and another group showing a limited number of alterations. Finally, another class was observed at our lab- oratory wliicii was veiy markedly different from the parental strain, differing by 8 or by 16-17 net amino acid exchanges, and probably in many more sites on the molecule. Representatives of this group closely resembled two natural strains previously described (groups B and C in

CH^CONH

273

282 -BS-

171 -BB-

332 -B-

DEAMINATION 220

329

237

32IB

_284

218 -B- 233 -fl- 207 -9- 235 -B-

187

326

D ODD

-B— e-

-B— &■

BROMINATION

D ODD

0 DDD

I 214 — B- METHYLATION ^Z? "f"

-B-

i

176 215

5 7

t^?^^4i(^6^W8-^ 9

10

ee-

fiB-

I1K-I2^C00H

HUBU

Fig. 7. Location of amino acid exchanges in chemicalb' evoked mutants. Each horizontal line represents the peptide chain of a mutant, subdivided into tryptic peptides as in Fig. 4. SoHd rectangles indicate definite location of a change, open rectangles possible alternate sites of a single change in a specific sequential peptide (#1-12). Only in strain 326 are 2 of the 4 potential aspartic acids of peptide #1 transformed to .serines.

Table I), and concern arose that such virus isolates were actually derived from contaminating lesions. The finding that three of the five members of this group clearly differed from the natural strains by the replacement of 1 or 2 amino acids, in one instance involving an increase in glycine content by 25% seemed to rule out this interpretation. Concerning the origin of these strains, they were derived from random lesions on iV. sylvestris resulting from the same reaction mixtures (deamination. bromination, and alkylation), which also yielded strains showing from 0 to 2 amino acid replacements. The average number of hits per particle, as judged from the level of inactivation was 3-6. Thus, no reason becomes evident for the appearance of a new family of mutants

506 A. TSUGITA AM) II. 1 KAKXKEL-COMiAT

showing a conii)08ition so grossly unlike the pai'cntal strain and most of the mutants resulting from it.

While no explanation can be ori'ei'cd for these observations, a review of what we know concerning the composition of various natural strains may supply some basis of understanding. As previously discussed, natural TMV strains are known to occur in groups: the M strain is indistinguish- able from the connnon strain, and YA, GA, and J14D1 differ by 2 to 3 residues. Then suddenly we are faced with several extremely different strains, bearing little relationship to TMV and in only a few instances resembling one another. Yet no strains of intermediate compositions are known, although this would be expected if the latter arose by a series of 20 or more mutational events from common TMV. Thus, the situation is similar for both series, that of the naturally observed and of the chemically evoked strains, even though this fact does not facilitate an understanding of the mechanism of either. The possibility must naturally be considered that any isolated mutant is actually due to a spontaneous mutation, rather than to the chemical treatment employed. This possi- bility cannot be excluded in any individual case, and therefore no con- clusions should be based on single isolates. Yet, obviously, it would be hardly less surprising to find spontaneous mutation to lead to such marked changes. Further the infrequency of detection of spontaneous mutation makes this appear unlikely on statistical grounds, since the unexpected observation of gross changes in protein composition upon treatment with three different mutagens were made repeatedly.

Finally, there exists the possibility that the various chemical agents through their inactivating action might select for a more stable strain or family of strains present in traces in common T]\1V preparations, This might seemingly account for the six observed strains of unusual composition, all of which resulted from ])reparations inactivated to 95% or more. Nevertheless, such an explanation appears improbably because common TMV generally appears more stable and more viable than its strains. Furthermore, the fact that the RNA rather than the whole virus was subjected to the inactivating reactions makes this interpreta- tion even more unlikely, for it appears veiy improbable that the RNA differs greatly in stability from one strain to another.

In the course of our chemical analyses of many mutants, these were repeatedly passed through singk^ lesion hosts and through large groups of A^. tabacum. Occasionally, and particularly in hot weather, the virus progeny so isolated was not identical in composition or symptomatology with the parental strain. At times, altered symptoms were detected on part of the inoculated plants and these plants were found to contain the chemically different strains. Of the six instances in wliich such phenom-

X. TMV STUDIES IN GENETIC CODING 507

ena were observed, four showed a single "secondary" amino acid ex- change. This could be due to some heterogeneity in the primary mutant, with a minor component (which then could be attributed to the original modification reaction) being favored by the test conditions. On the other hand, these observations may also signify that some of the chemically evoked mutants are rather unstable and readily undergo additional changes due to spontaneous mutations in the course of replication. The corollary of this hypothesis would be that the spontaneous mutation frequency is much higher for RNA than generally presumed, but that this fact is usually concealed by the lesser viability of the vast majority of the mutants as compared to the parental strain. If this were the case, then the observed mutants showing very different composition from the parental virus might be attributed to a series of spontaneous mutational events within a low- viability population, leading ultimately to a stable or rather a viable virus which then, not too surprisingly, would tend to resemble a known natural strain. In that case no direct relationship between the mutagenic reaction and the final stable strain would be expected.

These considerations probably do not apply to the strains showing only one or two exchanges, and yet among these also the same amino acid replacements were frequently observed to recur, regardless of the mutagenic reagent used.

In conclusion, then, no evident correlation was detected between the presence and extent of amino acid exchanges and either the nature of the modifying reaction or its intensity. The only clear correlation which was noted in this work was that amino acid changes were detected in almost all mutants giving local lesions on A^ sylvestris, and in only one mutant not showing this distinct biological property. It is this recog- nition which enabled us to select more promising mutants for chemical study, and thus to obtain relatively more data on amino acid exchanges than could be obtained by random analysis of mutants irrespective of the nature of their symptom alterations.

Parallel to our studies Wittmann analyzed a considerable number of nitrous acid mutants (1960a, 1961). His techniques differ from our in the following regards. He modified the virus directly, rather than the RNA, and has used only nitrous acid as the mutagen. As starting mate- rial he frequently employed a strain (A-14), rather than common TINIV. For the detection of mutants, Wittmann relied on tlie transfer of single primary lesions from Xanthi to N. tabaccum (symtom mutants) avoid- ing selection on A'', sylvestris. The reason for this was his desire to minimize the danger of secondary mutations in the course of repeated transfer. After isolation of the progeny protein, Wittmann directly

508 A. TSUCITA AND H. !• HAKXKEL-rOXHAT

proceeded to tryptic splittinp;, and aiialyzcfl the i-csultaiit peptides f^oparatc'ly. Ho has never reported in detail how he jmrified each peptide and it is our experience that some, and particuhirly the I-iiei)tide (41 residues), can he frcH'd oidy witli (Hfliculty from traces of adsoi'hecl p(^p- tides. The po.ssil)ility thus exists that a seeming (hfference in the composi- tion of the T-peptide is due to this technical difficulty. The results reported hy W'ittniami ai'e in part in line with our obser\'ations, hut show some unexpected differences. His observed i)roportion of mutants of unchanged comi)osition (70%) is higher than ours, as would he expected from the fact that he relied on random sampling only, while we fi'equently used a method which selects for protein changes.

A brief discussion of this group of mutants of unchanged pi'otein composition might here be interjected. These nmtants are of considerable interest since their presence clearly signifies that part of the RNA carries information which is not related to the stnicture of the virus coat protein. This is not a surprising fact, but had nevertheless not been previously demonstrated, and it is important to evaluate the validity of the evidence on which it is based. It would seem that the failure to detect a difference between two proteins might well be due to methodological causes. Thus, interchanges of two amino acids within one peptide, and translocation of amide groups have not been ruled out by the methods used, although the latter occurrence would probably be revealed upon chromatographic separation of the peptides. In view of these possibilities, the failure to detect an actual difference is a definite possibility for any one strain protein, even though it is not very probable. However, the prol)ability of this having occurred 7o times in one laboratoiy, and about 30 times in another is as good as nil, and thus the conclusion that different strains may share the same pro- tein appears on statistical grounds firmly established.

Besides this group, Wittmann also observed mutants showing 1 or 2 amino acid replacements, but none showing the many alterations observed six times in our laboratoiy. His exchanges were also often the same, but oddly enough his most frequent exchanges did not coincide with those of our laboratory. Yet in certain instances (e.g., serine #138 — > phenylalanine) the same exchanges were detected in both laboratories, an occurrence which would be statistically most improbable if it were due only to chance.

C. VIRULENCE AND ^■IAHILITY OF STRAINS

The isolation of a number of new strains of TM\' calls forth Cjues- tions about their comparative virulence and viability. As jireviously mentioned, manv of tlie new strains are more harmful to the host than

X. TMV STUDIES IN GENETIC CODING 509

the parental common TMV. Others are similarly, others less pathogenic, even to the point of giving no visible symptoms on two varieties of tobacco. Thus, virulence may be increased or decreased by mutation. However, viability, as frequently indicated by lesion size and measured by lesion productivity, seems to be invariably decreased by nmtation. All natural strains were less productive than common TMV, and this difference is more pronounced for many chemically produced strains (Veldee and Fraenkel-Conrat, 1962). Thus, as one might expect, com- mon TAIV, the result of extensive selection, is probably the specimen most perfectly adapted to its medium within the range of all possible variations of that theme. All newly produced mutants may well have appeared spontaneously at some previous time and may have fallen by the wayside in the course of natural selection.

The first mutant to be elucidated by us in structural terms, #171, has illustrated one possible mechanism of biological evaluation of a chemical structure of the kind that would be required for such selection (Tsugita and Fraenkel-Conrat, 1960) . Resistance to enzymes is probably a favorable attribute for viruses. The blocking of the N-terminus of TMV protein by an acetyl group, and the enzyme blocking action of proline near the C-terminus may contribute in this manner to the stabil- ity of the virus. The replacement of this proline by a leucine which was observed in strain #171 abolished its resistance toward the enzyme, carboxypeptidase A, and it appears very possible that the lesser biolog- ical viability of strain #171 can be largely attributed to this cause, and thus to the particular amino acid exchange in position 156 along the chain. Similarly simple explanations have not offered themselves in regard to other strains. It is important to remember, however, that we are able to analyze only the most viable mutants, and that we are thus forced to discriminate against all the more interesting protein altera- tions which render it nonfunctional.

VII. The Coding Problem

One of the most important problems for an understanding of the mechanism of protein synthesis is the functional relationship between the genetic material and the protein. Considerable evidence has accumu- lated which indicates that the bulk of the genetic infomiation in typical organisms is transmitted from parent to offspring through DNA. Accord- ing to present information, the messages of DNA are transmitted by Vv^ay of RNA (the so-called messenger RNA) and result in a specific protein structure, with soluble RNA probably acting as intermediary. In TMV as well as in other RNA viruses, the RNA was proven to be the direct bearer of the entire genetic information required to determine

510 A. TSUGITA AND H. FRAENKEL-CONRAT

the ainino acid ."sequence of its jirotoin coat and its symptoms in the host. This was best illustrated by the mixed reconstitution experiments described previously. This concept was further supi)orted by the study of chemical mutants, since one or very few changes in the nucleotide chain were able to effect changes in terms of biological symptoms and protein structure. Conclusive proof for the direct messenger activity of viral RNA has come from the recent experiments on protein synthesis in a cell-free E. coli system.

With a small virus like TMV, one should in principle be able to study both the structure of the RNA and that of the corresponding protein. If we assume that the linear nucleotide sequence in RNA is related to the amino acid sequence of the protein coat, and that the nature of the correspondence can be generalized, then TMV appears ideal for the study of the "coding problem" of protein synthesis, since we already possess experimental methods for the in vitro production of mutants and have knowledge of the complete amino acid sequence of this protein. Unfortunately, the chemical study of the RNA is only in its infancy, and the localization of mutagenic chemical events on the polynucleotide chain is not yet possible. Chemical studies of the protein of strains, on the other hand, has yielded definitive data, as reported in the preceding section. We shall now attempt to discuss and interpret these experimental results in the light of current theories and experi- mental facts pei'taining to the problem of coding and information transfer.

In 1954 Gamow publislicd the first theoretical approach to the problem how nucleotide sequences might code for amino acid sequences. The main assumptions of Gamow's scheme were the following. First, that a trinucleotide in DNA determined one amino acid. Second, that the trinucleotide fragment of DNA determining an amino acid and its neighbor "overlapped" by two-thirds of their length. Third, that the code was "degenerate," which means that more than one triplet may code for the same amino acid.

A triplet of nucleotides was assumed to code for one amino acid because pairs of the 4 nucleotides gave only 16 possibilities, while triplets gave 64 possibilities. The first number is not enough to code for 20 naturally occurring amino acids, while 64 possible combinations are more than necessary and permit degeneracy.

With the increase in our knowledge of amino acid sequences in many kinds of proteins, Gamow's overlajiping scheme has been experimentally ruled out. Brenner (1957) has given a proof of the theoretical impossi- bility of any univci'sal ()\-crlapping triplet code.

In 1957 Crick rf al.. introduced a non-overlapping triplet code. The

X. TMV STUDIES IN GENETIC CODING 511

authors suggested that of the 64 triplets some "made sense/' i.e., they represented an amino acid, while others did not. As an example, if A-G-U'* stood for leucine and C-A-U^ for valine, then G-U-C and U-C-A would be condemned to represent nothing ("nonsense"), so that . . . A-G-U-C-A-U . . . anywhere along the chain could be read only as Leu.\'al. What made this theoiy most appealing was the fact that the 64 possible triplets were thus automatically reduced to 20 sense-making ones, in agreement with the number of common building blocks in proteins. The resulting code was obviously non-overlapping, and it was comma-free, since each nucleotide along the chain pertained to only one sense-making and thus coding triplet. It should be noted that in a commaless code triplets composed of the same nucleotide (i.e., A-A-A) are excluded, since a sequence such as -A^-A--A^-A*-A^- does not give a unique sense; for it can be read starting at position 1, 2, or 3.

More recently. Crick et al. (1961) presented the first experimental evidence, based on genetic studies of T4 phage and acridine mutants derived from it, that the code for the DNA of bacterial genes was actually triplet in nature. They also arrived at the conclusion that the reading of the code must be from one end, and thus does not require nonsense triplets as commas. Their new data seemed to rule out the existence of meaningless triplets. Only between cistrons was there evi- dence for a break in the continuity of decipherable nucleotide sequences.

In addition to the theoretical and genetic approaches to the coding problem, an analytical approach was also attempted. Thus, Yeas pro- posed a code based on correlations between published analyses of the nucleotide and amino acid compositions of simple viruses. Woese (1961a,b) followed this path further, additionally using amino acid exchange data from several kinds of j)roteins in related species.

A great advance in our understanding of the coding jiroblem has recently come about when Nirenberg and ]\Iatthaei reported on their use of an excellent system to attack this problem in direct and experi- mental manner (1961). We have previously described the cell-free protein synthesizing system of E. coli (containing S-RNA, ribosomes, an ATP-generating system, and C"-amino acids) which these authors reported to be greatly stimulated by the addition of any RNA, and par- ticularly by TMV-RNA. Surprisingly, polyuridylic acid (poly-U) was found to be particularly effective in this regard, and Nirenberg and Matthaei were soon able to demonstrate that this additive led to the exclusive synthesis of polyphenylalanine.

This finding attracted the deserved attention, and othei- l;d)oratories

^A represents a nucleotide residue, the base of which is adenine; G represents guanine; U, uracil; and C, cystosine.

512 A. TSUGITA AND II. FRAENKEL-CONRAT

(juickly initiated research projects along similar lines and reported preliminary results. Thus Lengyel et al. (1961), using essentially the same system, coniirmed that poly-U led to the synthesis of polyjihenyl- alauine. These workers obtained a molar ratio of 3.25 between the amount of nucleotide used and the newly incor])orated amino acid, compared to 2.4 reported by Nirenberg and Alatthaei (1961j. This figure may represent a first a])proximation by biochemical nu'thods to the coding ratio, and it is heartening that it is so close to the tri])let recently postulated by Crick et al. (1961) on genetic grounds.

Further amino acid incorporation studies in the laboratories of both Nirenberg and Ochoa have centered on the use of various nucleotide copolymers of different composition (Martin et al., 1962; iSpeyer et al., 1962a,b; Lengyel et al., 1962). These favored in more or less specific manner the incorporation of other amino acids, but, as one might expect, these random copolymers were cjuantitatively much less effective than poly-U. All polymers that have been reported on contained uridylic acid and their effectiveness in regard to the incorporation of various amino acids was generally expressed in terms of the ratio to phenylalanine incorporation. Although the use of random copolymers did not enable the authors to determine the sequential structure of any coding unit, except for poly-U, nor whether the unit was definitely a triplet, they were able to suggest probable coding combinations for 17 out of the 20 naturally occurring amino acids, with only a few instances of disagree- ment in the results of the two laboratories participating in the race. In addition, Nirenberg et al. (1962) have shown that phenylalanine linked to soluble RNA was an intermediate in the process, using polyuridylic acid.

The finding that poly-U carries the message for phenylalanine poly- merization is in disagreement with Crick's "commaless triplet code" as described. As mentioned before, these facts and their own data have led Crick et al. (1961) to modify their concepts and to abandon the postulation of nonsense triplets. It thus becomes ever more likely that the code is a degenerate one.

Two years preceding these exciting demonstrations, we reported the amino acid exchanges in a chemically evoked nnitant of TJNIV (Tsugita and Fraenkel-Conrat, 1960), and Wittmann (1959b) observed no amino acid exchanges in another chemical mutant. Since then both groups have concentrated their efforts on analyzing many mutant proteins, and these observations have supplied important experimental data pertaining to the coding problem.

Wittmann analyzed 120 nitrous acid mutants (1961) and we analyzed 56 chemically induced mutants (Tsugita and Fraenkel-Conrat, 1960,

X. TMV STUDIES IN GENETIC CODING 513

1962; Tsugita, 1962a,b). In both laboratories at least half of the observed mutants showed no amino acid exchange. We have ])reviousIy discussed the grounds for accepting this negative finding as valid. This finding means that a good part of the RNA docs not find expression in the composition of the virus coat protein. Thus, coding schemes based on the direct com])arison between the composition of the RXA and the protein of the viruses, such as those proposed by Yeas and Woese, seem to have lost any basis in reality. The finding that not all RXA alterations lead to coat protein alterations is in line with expectation, if one considers the numbers of residues involved. Assuming a triplet code, the 158 amino acids would be expected to be governed by only 474 nucleotides, which corresponds to only 7.4% of the nucleotides in TMV-RNA. Most of the RNA would thus be assumed to carry information not pertaining to the coat protein, which alone has been analyzed. Actually one might expect that only about 1 out of 10 mutants should show protein changes, but it is ciuite possible that a high proportion of the particles altered in seg- ments pertaining to properties other than protein coat structure are not viable.

The chemical mutant (#171) first analyzed by us (Tsugita and Fraenkel-Conrat, 1960) had 3 amino acid exchanges and the location of these changes was scattered throughout the sequence. This observation, as well as the detection of mutants with single exchanges, disproves any overlapping code.

Investigations of the deamination reaction with DNA have led Viel- metter and Schuster (1960) to the conclusion that only the deamination of cytosine and adenine was mutagenic. The resulting uracil and hy- poxanthine would on chemical grounds be expected to pair preferentially as if they were thymine and guanine, respectively. In contrast, the deamination product of guanine, xanthine, still resembles guanine. Based on such consideration and an assumed triplet code, Gierer (1962) arranged all possible permutations of the 64 triplets in a scheme gov- erned by the result of their deamination (C^'U, A— > (hypoxanthine) — > G) in the form of 8 octets as shown in the diagram.

CAC

./

\

UAC CGC CAU

UGC UAU CGU

\ 1 ^ UGU

514 A. TSUGITA AND H. KRAHNKEL-CONRAT

It is evident that this scheme sets strict liiiiilalioiis on the possible results of deaniination. Thus, certain amino acids might be replaced by three others but none by more. Any amino acid ever observed to result from an exchange could only be exchanged to two or one other. Finally, no amino acid could result from more than three others, and those in this latter group would never be rejjlaced. Wittmann has relied on this scheme in the interpretations of his lesults, l)iit he finds some exchanges that do not fit, and we have observed others.

As previously discussed, we believe that in UNA, as contrasted to DNA, the deaniination of cytosine is the greatly predominating muta- genic event. Only much more rarely might the deaniination of A cause an A— >G shift, and other base exchanges may arise with even lower frequency during replication of the deaminated as well as the brominated or methylated KNA, and attributable to random mistakes in replication in a manner similar to natural mutation. Thus, the octet scheme ajijiears too rigid to allow for all possible events, and not truly expressive of the predominant role of the C-»U change in RNA deaniination.

Focusing our attention on this predominant result of deaniination, we have arranged our observed amino acid exchanges schematically in groups headed by triplets of decreasing cytosine content, allocating amino acids to triplets on the basis of the recent coding findings discussed above (see Fig. 8). The implication in such an evaluation of our exchange data is that they are the consequence of single step mutations, which is probably the case for most single exchanges, but can in no individual case be stated with assurance.

Further, one must consider whether the coding triplets as observed with E. coll would be of the same "language" as those in tobacco plant cells, since there exists the possibility that different codes might pertain in different organisms. This question appears to be answered by the results reported in a preceding section where it was show'n that TAIV- RNA acted as messenger for the production of material resembling the TMV protein by the same E. coli system, which has been used to obtain the coding data with synthetic polynucleotides. Thus, the coding language, whether it is universal or not, appears to be very similar for E. coli and tobacco, and the correlations shown in Fig. 8 are thus justified.

The data presented make it appear probable that proline, threonine and aspartic acid or its amide are coded as C2U; leucine, serine, and alanine as CU2; and phenylalanine as U.-,. These hypotheses agree in part with the results obtained by one and/or the other group studying the amino acid incorporation in the E. coli system. Cases which do not agree with these data (e.g., Tyr-^Pho, with tyrosine being coded by an

X. TMV STUDIES IN GENETIC CODING

515

c,,u CgX

cxu CX2

U3 XU2 XoU

Pro

2 (7)

-»- Leu

Tyr

AU, (O)

A<U (N)

3 (1)

t -Phe

U3(N,0)

(1)

Asp

lieu

(AspN) (Asp)

A2U (O) CAU GAU (O)

T2 Ala

AU2 (O) A,-U,(N)

rCGU(N,0)1

Glu

AGU (O) CGU (N)

Arg-

2^(2)

CGU(N,0)

Fig. 8. Correlation of observed amino acid replacements with coding symbols. All amino acid exchanges that have been observed in this laboratory are shown, with their frequency of occurrence. Numbers in parenthesis indicate exchanges resulting from reactions other than deamination, and thus unrelated to the arrangement scheme of this presentation which is based on the decreasing cytosine content from left to right. Solid lines and underlined code s>'mbols indicate agreement of amino acid exchange data with C -^ U transformations. "N" and "O" after the code symbols refers to the origin of the respective identification, "N" standing for Nirenberg and Matthaei (1961), Nirenberg et al. (1962), and Martin et al. (1962), and "O" standing for Lengyel et al., 1961, 1962. and Speyer et al. (1962). The symbol 2* in the bottom line refers to the fact that a progeny isolate of this strain showed three arginines to be replaced by glj^cines, rather than two.

alf) A. TSUGITA AND H. FRAENKEL-CONRAT

A-coiitaininji imclcotidc, according to Ijoth groups) must be attiil)utcd to mutational events other than C— > U and A — > G wliich, as stated, are definitely possil)le. Altei-nately, the degeneracy of the code might supply an explanation. It is encouraging, however, that at least 55% of the amino acid exchanges observed by us can be attributed to cytosine deaminations on the basis of present coding data. About 15% of the changes suggest an A -^ G exchange, but an ecjual fraction is indicative of C-^G and A^U exchanges. On the other hand, the fact that the allotment of code symbols to amino acids has been influenced by ex- change data obtained with nitrous acid mutants naturally renders the "agreement" less surprising. It also suggests that the limits in the useful exploitation of random nucleotide copolymers for coding information have now been reached, if not overstepped.

To conclude this short survey of the present concepts concerning the code that relates RNA to protein structure, one is now in the fortunate position that much experimental data has recently become available which brings this problem from the realm of speculation to that of factual information. It appears that three or possibly a slightly greater number of nucleotides code for one amino acid; it appears probable that several oligonucleotide units may represent the same amino acid. It is certain that a few uridylic acids represent one phenylalanine, and the approximate nucleotide composition of the words spelling 16 other amino acids are known with varying certainty. The code seems to be quite similar for very different organisms. It is probably deciphered starting from an end of the nucleotide chain and thus requires no markers (commas) between "words."

The main projects for the immediate future are (1) the preparation, by chemical or biosynthetic means, of polymers of known repeating nucleotide sequences and their systematic testing for messenger activity, (2) the establishment of the nature of many decidedly single-hit, one- step mutations in terms of resulting amino acid exchanges, and (3) the development of methods to determine the nucleotide exchange produced by the mutation, be it in the DNA or in the messenger RNA.

It would appear that within a few years it may become possil)le to produce a known modification in one of the simpler DNA or RNA genomes and find it to result in a predicted protein alteration. The natural mode of transfer of information from the genetic DNA to the messenger RNA is now reasonably well understood. There remains only one major obstacle to overcome before we can fruitfully exploit our new knowledge toward the manipulation of even the simplest genetic ma- terial, which is that of the small viruses. The remaining hurdle is the

X. TMV STUDIES IN GENETIC CODING 517

lacking methodology for the sequential analysis of long polynucleotides, a feat which appears to be the prerequisite for the exact superposition of the genetic map on the polynucleotide chain, the crowning achievement in molecular genetics.

References

Aach, H. G. (1958). Z. Naturforsch. 13b, 426.

Anderer, F. A. (1959). Z. Naturforsch. 14b, 642.

Anderer, F. A. (1962). 5th Intern. Congr. Biochem., Moscow, 1961.

Anderer, F. A., Weber, E., and Uhlig, H. (1960a). Z. Naturforsch. 15b, 79.

Anderer, F. A., Uhlig, H., Weber, E., and Schramm, G. (1960b). Nature 196, 922.

Ansevin, A. T., and Lauffer, M. A. (1959). Nature 183, 1601.

Bawden, F. C, and Pirie, N. W. (1937). Proc. Roy. Soc. B123, 274.

Beijerinck, M. W. (1899). Zentr. Bakteriol. Parasitenk. Abt. II 5, 27.

Black, F. L., and Knight, C. A. (1953). J. Biol. Chem. 202, 51.

Blout, E. R., DeLoze, C., Bloom, S. M., and Fasman, G. W. (1960). /. Am. Chem.

Soc. 82, 3787. Boedtker, H. (1959). Biochim. et Biophys. Acta 32, 519. Bonhoeffer, F., and Schachman, H. K. (1960). Biochem. Biophys. Research Com-

muns. 2, 366. Braunitzer, G. (1956). Biochim. et Biophys. Acta 19, 574. Brenner, S. (1957). Proc. Natl. Acad. Sci. U.S. 43, 416. Caspar, D. L. D. (1956). Nature 177, 928. Colter, J. S. (1958). Progr. iji Med. Virol. 1, 1. Crick, F. H. C, Griffith, J. S., and Orgel, L. E. (1957). Proc. Natl. Acad. Sci. U.S.

43, 687. Crick, F. H. C, Barnett, L., Brenner, S., and Watts-Tobin, R. J. (1961). Nature 192,

1227. d'Herelle, F. (1917). Compt. rer^d. acad. sci. 165, 373. Doty, P., Boedtker, H., Fresco, J. R., Haselkorn, R., and Litt. M. (1959). Proc. Natl.

Acad. Sci. UJi. 45, 482. Engler, R., and Schramm, G. (1959). Nature 183, 1277. Fraenkel-Conrat, H. (1955a). J. Biol. Chem. 217, 375. Fraenkel-Conrat, H. (1955b). Biochim. et Biophys. Acta 15, 307. Fraenkel-Conrat, H. (1956). /. Am. Chem. Soc. 78, 882. Fraenkel-Conrat, H. (1957). Virology 4, 1. Fraenkel-Conrat, H. (1959). In "Sulfur in Proteins" (R. Benesch et al, ed.), p. 339.

Academic Press, New York. Fraenkel-Conrat, H. (1961). Biochim. et Biophys. Acta 49, 169. Fraenkel-Conrat, H. (1962). In "Survey of Biological Progre.ss," (B. Glass, ed.).

Vol. IV, p. 59. Academic Pre.ss, New York. Fraenkel-Conrat, H., and Narita, K. (1958). Symposium on Protein Structure, Paris,

1957, p. 249. Fraenkel-Conrat, H., and Ramachandran, L. K. (1959). Advances in Protein Chem- istry 14, 175. Fraenkel-Conrat, H., and Singer, B. (1957). Biochim. et Biophys. Acta 24, 540. Fraenkel-Conrat, H., and Singer, B. (1959). Biochim. et Biophys. Ada 33, 370. Fraenkel-Conrat, H., and Singer, B. (1962). Biochemistry 1, 120.

518 A. TSUGITA AND H. FRAENKEL-COXRAT

Fracnkel-Conrat, H., and Williams. R. C. (1955). Proc. Nail. Acad. Sci. U.S. 41, 690. Fraenkel-Conrat, H., Singer, B., and Williams, R. C. (1957). Biochim. el Biophys.

Acta 25, 87. Fraenkcl-Conrat, H., Singer, B., and VcKlec, S. (1959). Biochim. el Biophys. Acta

29, 639. Fnunkcl-Conrat, H., Singer, B., and Tsugita, A. (1961). Virology 14, 59. Franklin. R. E.. and Holmes, K. C. (1958). Acta Crysl. 11, 213. Franklin. R. E., and Klug. A. (1955). Ada Crysl. 8, 777. Franklin, R. E., and Klug, A. (1956). Biochim. ct Biophys. Ada 19, 403. Franklin, R. E., Klug, A., and Holmes, K. C. (1957). "The Nature of Viruses"

(G. E. W. Wolstenholme and E. C. P. Millar, eds.), p. 39. Churchill, London. Frisch-Niggemeyer, W. (1956). Nature 178, 307. Gamow, G. (1954). Nature 173, 318. Gierer, A. (1957). Nature 179, 1297. Gierer, A. (1958a). Z. Nalurforsch. 13b, 477. Gierer, A. (1958b). Z. Nalurforsch. 13b, 485. Gierer, A. (1960). Progr. in Biophys. and Biophys. 10, 299. Gierer, A. (1962). 5lh Intern. Congr. Biochem., Moscow, 1961. Gierer, A., and IMundry, K. W. (1958). Nature 182, 1457. Gierer, A., and Schramm, G. (1956). Z. Nalurforsch. lib, 138 (1956). Ginoza, W. (1958). Nature 181, 958. Gish, D. T. (1959). Biochim. el Biophys. Acta 35, 557. Gish, D. T. (1960). J. Am. Chem. Soc. 82, 6329. Gish. D. T. (1961). J. Am. Chem. Soc. 83, 3303. Gish, D. T., Ramachandran, L. K., and Stanley, W. M. (1958). Arch. Biochem.

Biophys. 78, 433. Gordon, M. P., and Staehelin, M. (1958). J. Am. Chem. Soc. 80, 2340. Harris, J. I., and Knight, C. A. (1952). Nature 170, 613. Harris, J. I., and Knight, C. A. (1955). J. Biol. Chem. 214, 215. Hart, R. G. (1958). Biochem. et Biophys. Acta 28, 457. Haschemej-er, R., Singer, B., and Fraenkel-Conrat, H. (1959). Proc. Natl. Acad. Sci.

U.S. 45, 313. Haselkorn, R., and Doty, P. (1961). /. Biol. Chem. 236, 2738. Hershey, A. D., and Chase, M. (1952). J. Gen. Physiol. 36, 39. Iwanowski, D. (1892). St. Petersburg Acad. Imp. Sci. Bull. 35, 67. (Reprinted in

English (1942) : Phytopathol. Clas.^ics 7, 27.) Jeener, R., and Rosseels, J. (1953). Biochim. et Biophys. Ada 11, 438. Klug, A., and Caspar, D. L. D. (1960). Advances in Virus Research 7, 225. Knight, C. A. (1947). J. Biol. Chem. 171, 297. Knight, C. A. (1952). J. Biol. Chem. 197, 241. Knight, C. A. (1955). Virology 1, 261. Knight, C. A. (1959). In "The Viruses" (F. M. Burnet and W. M. Stanley, eds.),

Vol. 2, p. 127. Academic Press, New York. Knight, C. A., and Stanley, W. M. (1941). J. Biol. Chem. 141, 39. Knight, C. A., Silva, D. M., Dahl, D., and Tsugita, A. (1962). Virology 16, 236. Kramer, G., and Wittmann, H. G. (1958). Z. Nalurforsch. 13b, 30. Leng>-el, P., Speyer, J. F., and Ochoa, S. (1961). Proc. Natl. Acad. Sci. US. 47, 1936. Lengyel, P., Speyer, J. F., Basilio, C, and Ochoa, S. (1962). Proc. Natl. Acad. Sci.

U.S. 48, 282. Littlefield, J. W., and Dunn, D. B. (1958). Biochem. J. 70, 642.

X. TMV STUDIES IN GENETIC CODING 519

Mandel, H. G., Markham, R., and Matthews, R. E. F. (1957). Biochim. et Biophys.

Acta 24, 205. Markham, R., and Smith, J. D. (1950). Biochem. J. 46, 513. Martin, R. G., Matthaei, J. H., Jones, O. W., and Xirenberg, M. W. (1962). Biochem.

Biophys. Research Communs. 6, 410. Matthews, R. E. F., and Smith, J. D. (1955). Advances in Virus Research 3, 49. Mundry, K. W. (1959). Virology 9, 722. Narita, K. (1958). Biochim. et Biophys. Acta 28, 184. Newmark, P., and Meyers, R. W. (1957). Federation Proc. 16, 226. Nirenberg, M. W., and Matthaei, J. H. (1961). Proc. Xatl. Acad. Sci. lj\S. 47, 1588. Xirenberg, M. W., and Matthaei, J. H., and Jones, O. W. (1962). Proc. Xatl. Acad.

Sci. U.S. 48, 104. Niu, C. I., and Fraenkel-Conrat, H. (1955a). Biochim. et Biophys. Acta 16, 597. Niu, C. I., and Fraenkel-Conrat, H. (1955b). Arch. Biochem. Biophys. 59, 538. Nixon, H. L., and Woods, R. D. (1960). Virology 10, 157. Ramachandran, L. K., and Gish, D. T. (1959). J. Am. Chem. Soc. 81, 884. Rushizky, G. W., and Knight, C. A. (1960). Virology 11, 236; Proc. Xatl. Acad. Sci.

U.S. 46, 945. Schaffer, F. L., and Schwerdt, C. E. (1955). Proc. Xatl. Acad. Sci. U.S. 41, 1020. Schramm, G., and Zilhg, W. (1955). Z. Xaturjorsch. 9b, 494. Schramm, G., Schumacher, G., and Zillig, W. (1955). Z. Xaturjorsch. 10b, 481. Schuster, H. (1960). in "The Nucleic Acids" (E. Chargaff and J. N. Davidson, eds.),

Vol. 3, p. 245. Academic Press, New York. Schuster, H. (1961). J. Mol. Biol. 3, 447.

Schuster, H., and Scliramm, G. (1958). Z. Xaturjorsch. 13b, 697. Siegel, A. (1960). Virology 11, 156.

Siegel, A., Ginoza, W., and Wildman, S. G. (1957). Virology 3, 554. Singer, B., and Fraenkel-Conrat, H. (1961). Virology 14, 59. Speyer, J. F., Lengyel, P., BasiHo, C, and Ochoa, S. (1962a). Proc. Xatl. Acad. Sci.

UJS. 48, 63. Speyer, J. F., Lengyel, P., Basiho, C, and Ochoa, S. (1962b). Proc. Xatl. Acad. Sci.

U.S. 48, 441. Staehelin, M. (1958). Biochim. et Biophys. Acta 29, 410. Staehelin, M. (1959a). Biochim. et Biophys. Acta 31, 448. Staehelin, M. (1959b). Experientia 15, 413. Staehelin, M., Peterson, E. A., and Sober, H. A. (1959). Arch. Biochem. Biophys.

85, 289. Stanley, W. M. (1935). Science 81, 644. Stanley, W. M. (1936). Science 83, 626. Stanley, W. M. (1938). In "Handbuch der Virusforschung" (R. Doerr and C. Hal-

lauer, eds.). Vol. IV, p. 447. Springer, Berhn. Sugiyama, T., and Fraenkel-Conrat, H. (1961). Proc. Xatl. Acad. Sci. U.S. 47, 1393. Sugij'ama, T., and Fraenkel-Conrat, H. (1962). In preparation. Sugiyama, T., Singer, B., and Fraenkel-Com-at, H. (1962). In preparation. Takahashi, W. N. (1959). Virology 9, 437. Takahashi, W. N., and Ishii, M. (1952). Xature 169, 419. Tsugita, A. (1960). Biochim. et Biophys. Acta 38, 145. Tsugita, A. (1961). Protein Xucleic Acid, Enzyme (Tokyo) 6, 385. Tsugita, A. (1962a). /. Mol. Biol., in press. Tsugita, A. (1962b). J. Mol. Biol., in press. Tsugita, A., and Fraenkel-Conrat, H. (1960). Proc. Xatl. Acad. Sci. U.S. 46, 636.

520 A. TSUGITA AND H. FRAENKEL-CONRAT

Tsugita, A, and Fraenkel-Conrat, H. (1962). J. Mol. Biul 4, 73.

Tsugita, A., Gisli, D. T., Young, J., Fraonkel-Conrut, H., Knight, C. A., and

Stanley, W. M. (1960). Proc. Natl. Acad. Sci. UJS. 46, 1463. Tsugita, A., Funatsu, G., and Young, J. (1962a). In preparation. Tsugita, A., Fiaonkel-Conrat, H., Nirenberg, M. W., and Matthaei, J. H. (1962b).

Proc. Natl. Acad. Sci. U.S. 48, 846. Twort, F. W. (1915). Lancet ii, 1241.

Veldee, S., and Fraenkel-Conrat, H. (1962). Virology, in pre.ss. Vielmetter, W., and Schuster, H. (1960). Z. Nalurforsch. 15b, 304. Whitfcld, P. (1962). In preparation.

White, F. H., Jr., and Anfinsen, C. B. (1959). AnJi. N.Y. Acad. Sci. 81, 515. Wildman, S. G. (1959). In "The Viruses" (F. M. Burnet and W. M. Stanley, eds.),

Vol. 2, p. 1. Academic Press, New York. Wittmann, H. G. (1959a). Z. Vererbungslehre 90, 463. Wittmann, H. G. (1959b). Expcrientia 15, 174. Wittmann, H. G. (1960a). Virology 11, 505. Wittmann, H. G. (1960b). Virology 12, 613. Wittmann, H. G. (1960c). Virology 12, 609. Wittmann, H. G. (1961). Naturwiss. 48, 729. Wittmann, H. G., and Braunitzer, G. (1959). Virology 9, 762. Woese, C. R. (1961a). Nature 190, 697.

Woese, C. R. (1961b). Biochem. Biophys. Research Communs. 5, 88. Yeas, M. (1960). Nature 188, 209. Zech, H. (1961). Z. Naturforsch. 16b, 520.

AUTHOR INDEX

Numbers in italics indicate the pages on wluch the references are hsted at the end of the article.

Aach, H. G., 493, 517

Abbo, F. E, 155, 159, 200

Abelson, P. H., 178, 20^

Abrahamov, A., 461, 4'^0

Acs, G., 356, 357, 358, m, 40^

Adelberg, E. A., 146, 151, 218, 265

Adler, J., 13, 14, 23, 24, 29, 40, 41, 42, 43,

44, 45, 46, 48, 60, 61, 63 Ager, J. A. M. 454, 469 Aksoy, M., 436, 469 Alberts, B. M., 247, 266 Albertsson, P.-A., 310, 350 Alexander, P., 232, 233, 237, 238, 264, 268 Alfert, M., 92, lOS, 157, 190, 194, 200, 201 Allen, E. H., 355, 356, 358, 361, 362, 364,

365, 366. 367, 375. 399, 400, 401, 403 Allen, D. W., 408, 443, 459, 467, 469, 470 Allfrey, V. G., 395, 397, 400, 401, 404 Allison, A. C., 421, 463, 469, 471 Alpen, E. L., 466, 467, 470 Alper, H., 243, 264 Ames, B. N., 119, I48 Anderegg, J. W., 13, 61, 173, 200 Anderer, F. A.. 482, 484, 485, 486, 517 Anderson, B., 13, 17, 62 Anderson, E. S., 120, I48 Anderson, R. S., 242, 266 Anderson, T. F., 292, 351 Andrew, L. E., 218, 264 Anfinsen, C. B., 354, 373, 4OO, 404, 482,

520 Ansevin, A. T., 480, 517 Apgar, J., 361, 4OI Aposhian, H. V., 55, 56, 60 Arber, W., 133, I48 Arient, M., 165, 166, 204 Arlinghaus, R., 370, 379, 403 Armstrong, J. A., 120, I48 Arnstein, H., 393 Aronson, A. I., 179, 187, 200, 298, 305,

315, 329, 350, 351

Askonas, B., 373, 4OO

Astrachan, L., 179, 200, 273, 275, 276, 278,

279, 286, 289, 336, 337, 350, 352, 386,

400 Attardi, G., 384, 386, 402 Awater, J., 437, 438, 439, 440, 441, 450,

451, 470 Atwood, K. C, 245, 268 Aub, J. C., 191, 200

Auerbach, C., 210, 234, 235, 240, 264, 268 Aurisicchio, S., 119, I48 Austin, J. H., 416, 471 Avery, O. T., 72, 108

B

Baddiley, J, 166, 200

BagUoni, C., 410, 411, 412, 413, 416, 423, 424, 425, 427, 434, 438, 439, 440, 443, 446, 448, 449, 454, 463, 470, 474

Bailey, I. S., 451, 470

Baker, W. K., 243, 264

Baldesten, A., 3, 63

Baldwin, R. L., 50, 60

Balis, M .E., 185, 200

Balog, J. E., 408, 441, 443, 453, 469, 472

Balog-Shelton, J., 416, 417, 474

Barclay, R. K., 125, 149

Barker, T., 176, 202

Barlow, J. L., 119, 120, 150

Barner, H. D., 9, 60, 61, 143, I48, 156, 158, 170, 174, 194, 200, 201

Barnes, D. W. H., 466, 470

Barnes, H. B., 275, 287

Barnett, L., 225, 248, 249, 250, 253, 255. 260, 265, 285-286, 287, 377, 390, 4OI, 435, 436, 470, 511, 512, 517

Barnicot, N. A., 445, 474

Barnum, C. P., 166, 191, 200

Baron, F., 227, 264

Baron, L. S., 147, I48, 149, 150

Bartuska, D., 204

Basch, R. S., 165, 202

Basilio, C, 282, 283, 288, 390, 391, 392,

521

522

AUTHOR INDEX

400, 404, 434, 435, 436, 474, 512, 515,

51S, 619 Bass, L. W., 227, 265 Baucli, R., 225, 264 Bautz, E.. 222, 228, 229, 230, 234, 237,

238, 239, 249, 250, 251, 253, 255, 261,

263, 264, '266, 279, 2S6, 381, 389, 4OO Bawdcn, F. C, 478, 517 Bayley, C. R., 230, 264 Beam, A. G., 418, 420, 472 Beaven, G. H., 408, 416, 418, 428, 455,

459, 461, 470, 471, 474 Beer, M., 126, I48 Beers, R. ¥., 225, 264, 269 Beiboer, J. L, 447, 473 Beijerinck, M. W., 477, 517 Bell, D., 13, 62, 171, 172, 202, 205 Bello, L. J, 16, 60 Belser, N. D., 247, 269 Bendet, I., 114, 126, I48 Bendich, A., 107, 108 Benesch, R., 454, 470 Benesch, R. E., 454, 470 Ben-Ishai, R., 384, 404 Bentzon, M. W., 155, 195, 205 Benzer, S., 121, 150, 211, 212, 214, 219,

222, 225, 226, 238, 248, 249, 250, 252,

253, 255, 264, 265, 361, 4OO, 424, 426,

470 Berends, W., 245, 265 Berg, P., 17, 60, 280, 283, 284, 287, 289,

354, 355, 356, 358, 359, 360, 361, 380,

384, 388, 389, 394, 4OO, 4OI, 403, 404 Bergmann, F., 356, 358, 359, 361, 4OO, 403 Bergren, W. R., 422, Jir2 Bernal, J. D., 114, I48 Bernardi, G., 145, I4S Berns, K. I, 128, 129, 130, 138, I48, 151 Bertani, E., 3, 4, 5, 9, 50, 60 Bessman, M. J., 13, 14, 15, 16, 20, 21, 23,

24, 26, 27, 29, 31, 32, 33, 46, 47, 48, 50,

60, 61, 63, 160, 200, 203, 273, 280, 286,

287 Betke, K., 434, 459, 461, 462, 470, 472 Beukers, R., 245, 265 Bevelander, G., 165, 202 Bianchi, P. A., 13, 14, 61, 171, 172, 174,

200 Bianco, I., 418, 447, 449, 454, 474

Billon, D, 173, 174, 176, 178, 179, 181,

190, 191, 200, 201 Binkley, S. B., 170, 201 Binnio, B., 186, 202 Biringer, R., 280, 287 Birnic, G. D., 7, 61 Bishop, C. N., 165, 166, 201 Bishop, J., 363, 364, 365, 366, 367, 371,

372, 374, 375, 379, 386, 394, 398, 399,

400, 403 Black, F. L., 493, 517 Blakely, R. L, 7, 61 Bhikeslec, A. F., 248, 265 Bloch, D. P., 92, 96, 109, 157, 188, 200 Bloemcndul, H., 368, 369, 4OO, 4OI Bloom, S. M., 482, 517 Bloom, W., 459, 473 Blout, E. R., 482, 517 Blumson, N. L., 166, 200 Boedtker, H., 487, 488, 517 Boehner, L., 136, 137, 149 Boezi, J. A., 277, 287, 343, 350 Boivin, A., 72, 109 Bojarski, R. B, 14, 61, 62, 171, 172, 174,

200, 202 Bolle, A, 179, 180, 186, 187, 203 Bollum, F. J., 13, 50, 51, 52, 53, 54, 61,

171, 172, 173, 190, 200, 204 Bolton, E. T., 178, 204, 277, 287, 294,

298, 301, 302, 303, 305, 306, 309, 313,

315, 333, 336, 350, 352 Bolton, T. T., 305, 350 Boman, H., 356, 358, 402 Bond, V. P, 169, 201, 396, 4OI Bonhoeffer, F, 487, 517 Bonner, D. M., 288 Bonner, J., 186, 200, 279, 280, 287, 292,

294, 301, 352, 395, 396, 397, 4OI, 403 Borek, E., 10, 63 Borka, T., 157, 203 Borsook, H., 299, 301, 350, 351, 353, 354,

374, 401 Bosch, L., 368, 4OO, 4OI Boulanger, P., 277, 287 Boulard, C. 446, 470 Bovard, F., 358, 403 Bowman, H. G., 24, 61 Boycp, R., 184, 205

Boycr, S. H, 441, 442, 443, 446, 454, 474 Boyo, A. E., 454, 471

AUTHOR INDEX

523

Brachet, J., 272, 3S7, 299, 350, 395, 396,

401 Bradley, D. F., 225, S65 Bradley, T. B, Jr., 462, 463, 4''0 Brakke, M. K, 309, 350 Brammer, K. W., 230, 264 Brandt, P. W, 97, 104, 105, 110 Braiin, H, 461, 462, 472 Braun, W., 196, 201 Braunitzer, G., 413, 414. 415, 416, 417,

470, 473, 484, 485, 501, 517, 520 Brawner, J. N., Ill, 462, 463, 470 Brenner, S., 180, 186, 200, 205, 222, 225, 248, 249, 250, 253, 255, 260, 265, 268, 274, 285-286, 2S7, 310, 344, 350, 377, 383, 390, 401, 435, 436, 470, 510, 511, 512, 517 Bresler, A., 280, 2S7, 347, 348, 351, 382,

402 Bridges, C. B., 448, 470 Bridges, M. T, 444, 448, 471 Britten, R. J, 178, 179, 200, 204, 277, 287, 288, 296, 297, 299, 301, 305, 309-310, 312, 313, 315, 322, 323, 333, 344, 350, 351, 352 Brodie, H. R., 459, 470 Brody, S., 185, 200 Bromberg, Y. M., 461, 470 Brookes, J., 237, 265 Brown, A, 358, 401 Brown, B. J., 156, 174, 175, 184, 202, 375,

401 Brown, D. M., 227, 228, 264, 265 Brown, G., 358, 401 Brown, G. B, 243, 269 Brown, G. L., 125, 149 Brownell, L. W., 165, 166, 205 Bruce, V, 131, 132, 151 Buchanan, J. M., 5, 50, 55. 57. 62, 63 Bucher, N. R. L., 191, 197, 200 Buchwald, M, 313, 351 Buettner-Janusch, T., 165, 166, 204 Buettner-Janusch, V., 12, 63, 165, 166,

204 Bulivant, S., 107, 109 Bulman, H. N., 196, 201 Burckett, L., 462, 463. 472 Burgi, E., 122, 123, 125, 130, 139, 148, 149.

188, 201 Burma, D., 187, 201, 204, 280, 288

Burns, V. W., 159, 201 Bussard, A., 278, 287, 385, 401 Butler, E. A., 471 Butler, J. A. V., 13, 14, 61 Butros, J. M., 196, 201

Cabannes, R., 438, 470

Cairns, J., 126, 127, 148

Caldwell, P. C, 299, 351

Callan, H. G., 69, 107, 109

Cameron, I. L., 196, 202

Campbell, A., 146, US

Campbell, P. N., 355, 362, 363, 373, 400,

401 Canellakis, E. S., 13, 61, 160, 169, 171,

172, 185, 201, 203, 204, 357, 362, 402 Canellakis, Z. N., 160, 169, 204 Cantarow, A., 204

Cantoni, G. L., 357, 358, 360, 401, 404 Caputto, R., 57, 61 Carcassi, V., 449, 470 Cardini, C. E., 57, 61 Carey, W.. 147, 148, 149 Caro, L. G., 346, 351 Carrier, R. F., 190, 204 Carrol, E., 3, 4, 63 Cartledge, J. L., 248, 265 Case, M. E., 96, 109 Caspar, D. L. D., 121, 150, 479, 480. 483.

517, 518 Caspersson, T., 272, 287, 299, 351 Cassir, R., 176, 205 Catlin, B. W., 33, 61 Cavalieri, L. F., 66, 81, 82, 89, 97, 101,

107, 109, 144, 145, I48, 183, 184, 201,

203 Celano. A., 119, I48 Ceppellini, R., 418, 444, 445, 446, 448, 449,

463, 464, 469, 470. 472 Chamberlin, M.. 280, 284, 2S7 , 389, 4OI Champagne, M., 145, I48 Chandhuri, N. K., 9, 62 Chantrenne, H.. 384, 385. 395, 4OI Chao. F. C, 292, 294, 301, 351 Chargaff, E., 50, 63, 230, 269, 272, 288 Charlwood, P. A., 416, 418, 470 Chase, M. C, 72, 109, 136, 149, 162, 202,

210-211, 244, 265, 266, 478, 518 Cheong, L, 219, 265

524

AUTHOR INDEX

Chernoff, A. I., 423, 424, 458, 401, 407,

!,70, 474 Chiozzotto, A., 119, 14S Chonlikion. ¥., 176, 205 Cliou. S.. 156, 174, 175, 202 Clui. E.. 91, 109. 156. 158, 205 Chun, E. H. L, 80, 109 Clark. J. B., 244, 269 Claude, A., 292, 351 Clogg, M. D., 408, 470 Cloland, G. H., 240, 265 Clovoland, L. R., 67, 109 Clifton. C. E., 225,^65 Cohl). M.. 232, 233, 26S Cohen, A., 12, 63 Cohen, S. S., 8, 9, 10, 12, 46, 55, 56, 60,

61, 62, 64, 89, 109, 143, I48, 156, 158,

162, 170, 174, 194, 200, 201, 273, 275,

287, 382, 403 Cohn, W. E., 272, 287 Cole, L. J.. 166, 203 Colowick, S. P., 13, 61 Colter, J. S, 489, 517 Comly. L. T, 368, 369, 402 Conley, C. L., 462, 463, 470 Connel, G. E., 448, 474 Cook, C. D., 459, 470 Cook, J. L., 408, 473 Coppo, A., 119, 148 Cordero Funes, J. R., 204 Cordes, E., 356. 358, 4OO, 404 Cordes, S., 119. I48

Cordonnier. J. K., 416, 442, 443, 454, 473 Cormick, J., 416, 417. 443, 474 Cosset, A., 446, 470 Coughlin. C. A., 218, 265 Countryman, J. L., 273, 289 Cousens, S. F., 232. 233, 268 Cowie, D. B., 178, 204. 277, 287, 303, 305,

311, 343, 350. 351, 362, 378, 4OI Coy, U., 358, 402 Craig, D. L., 245, 267 Craig, L. C, 413, 471 Crane, H. R., 79, 110 Cranmore. D., 466. 467, 470 Crasemann, J. M.. 220, 222, 268 Cratliorn, A. R., 13. 14, 61. 171, 172, 174,

190, 200. 201 Croighton, H., 87. 109 Crick, F. H. C, 26, 38, 46, 49, 64, 72, 73,

111, 215, 225, 246, 250, 255, 260, 265, 269, 285-286, 287, 359, 368, 377, 380, 390, 401, 435, 436, 470, 510, 511, 512, 517

Crocker, T. T, 395, 396, 397, 40I

Crosbie, G. W., 6, 7, 61

Crouso, H. v., 92, 109

Culii.s, A. F., 409, 430, 473

Cummings, D. J., 142, 149, 156, 175, 202

Cunningham, L., 33, 61

Dahl, D, 494, 518

Dalgliesh, C. E., 4OI

D'Amato, F., 225, 265

Dance, N., 455, 471

Daniel, V., 368, 369, 4OI

Danneberg, P., 9, 62

Darwin, C, 208, 265

Das, N. K., 190, 194, 200, 201

Davidson, E. A., 17, 61

Davidson, J. N., 13, 62, 165, 166, 171,

172, 190, 201, 202, 205, 277, 288 Davison, P. F., 77, 109, 122, 123, 125, 126,

130, 139, 148, 149, 150 De, D. N., 92, 109, 188, 201 Deasy, C. L., 299, 350, 354, 4OI Deeley, E. M, 157, 158, 192, 204 de Garilhe, M. P., 20, 33, 61 De Giovanni, R., 245, 265 de Kloet, S., 397, 404 de la Haba, G. L,. 379, 404 Delamater, E. D., 193, 201 Delbriick, M., 73, 109, 208, 241, 267, 269,

292, 351 DeLoze, C, 482, 517 Dellweg, H., 219, 245, 269 Demailly, J., Jr., 281, 288 Do Mars, R. I., 225. 265 Demerec, M., 214, 223. 235, 241, 242. 243,

245, 247, 252, 265, 266 De Serres, F. J., 243, 252, 265. 266 Destaing, F., 446, 470 Deutsch, F., 144, 145, I48 Derreux, S., 385, 4OI De Vries, H., 209, 265 Dewey, D. L., 194, 202 De Wolf, S., 9, 62 Doysson. M.. 218, 265 d'Herelle, F., 477, 617

AUTHOR INDEX

525

Diamond, L. K., 447, 454, 1^11

Dickenson, R. E., 430, 47^

Dickey, F. H., 240, 265

Dickman, S., 362, m

Dieckmann, M., 356, 358, 359, 361, JfiO,

403 Dienstbier, Z, 165, 166, 204 Dinger, R., 347, 348, 351 Dintzis, H. M, 301, 351, 374, 375, 401,

457, 471 Diringer, R., 382, 402 Dische, Z, 27, 61 Dixon, G. H, 448, 474 Dixon, J., 162, 202

Djordjevic, B., 79, 109. 219, 222, 265, 269 Doctor, B., 361, 401 Doermann, A. H., 136, 137. 141), 211, 244,

265 Dohan. F. C, Jr.. 156, 191, 203 Donalson, J. T., 156, 201 Donini, P., 119, 148 Doty, P., 18, 44, 61, 63, 81, 109, 110, 144,

147, 150, 151, 230, 231, 265, 267, 488,

517, 518 Doudney, C. 0., 173, 176, 178, 179, 180,

187, 201, 245, 266 Dounce, A. L., 185, 188, 201, 205 Drabkin, D. L., 416, 471 Drauculic, M., 179, 185, 201, 203 Drasil, V., 196, 202 Drescher, H., 421, 471 Drew, R. M., 156, 157, 169, 194, 204 Dreyer, W. J., 428, 471 Dreyfus. J. C, 374, 398, 399, 402, 403 Dubin, D. T., 119, 14S Duerksen. J. D., 298, 303, 305, 311, 351,

352, 362. 378, 4OI Dulbecco, R., 244, 269 Dunn. D. B., 49, 61, 98, 109, 120, 149,

219, 265, 267, 357, 4OI, 488, 518 Duschinsky, R., 9, 62 Dutton, R. H., 196, 201 Dutton, R. W., 196, 201 Duzer, A., 446, 470

Eady, J. D., 196, 201 Eason, R.. 190. 205 Ebert, M.. 243. 264 Ebisuzaki, K., 16, 64

Edelman, I. S., 294, 351

Edgar, R. S, 96, 109, 132, 136, 149, 151

Edington, G. M., 415, 462, 463, 471

Edmunds, M., 3, 61

Edwards, J. L., 156, 201

Efron, M. L., 423, 431, 432, 433, 447, 471

Eidinoff, M. L., 219, 265

Eisenstadt, J. M., 187, 201, 281, 288, 389,

394, 403 El-Ani, A. S., 96, 109 Ellis, F., 156, 203 Ellis, M. E., 166, 203 Ellmauer, H., 244, 265 Elson, D., 298, 303, 305, 351 Elwyn, D., 7, 61 Endo, H. W., 245, 265 Engler, R., 490, 517 Ephrati-Elizur, E., 220, 269 Ephrussi-Taylor, H., 125, 149, 226, 267 Epstein, H. T., 119, I48 Erikson, R. O., 159, 163, 201 Errera, M., 179, 201, 396, 403 Erslev, A. J.. 450, 451, 470 Evans, E. A., Jr., 162. 165, 166, 204

Fahmy, M. J., 233, 235, 265

Fahmy, O. G.. 233. 235, 265

Falcone, G., 193, 201

Falkow, S., 147, 149, 150

Fankuchen, I. J., 114, I48

Fasman, G. W., 482, 517

Favelukes, G., 363. 369, 370, 371, 372,

379, 386, 397. 398, 399, 4OO Favelukes, S., 370, 403 Feinendegen, L. E., 169, 201, 396, 4OI Fe.s.sas. P., 442, 445, 447, 453, 454, 471 Fessenden, J., 367, 4OI Field, E. O., 440, 471 Fiers, W., 122, 149 Finamore, F. J., 164, 201, 279, 287 Firket, H., 156, 201 Firshein, W., 196, 201 Fischer, G. A., 170, 203 Fischer, R., 243, 265, 267 Fisher, H., 225, 265 Fisher, J., 380, 403 Fisher, T. X.. 276. 286. 337, 350 Fitzgerald, P. A., 454, 471 Fitz-James, P. C, 155, 158, 206

526

AUTHOR INDEX

Flaks, J. G., 8. 9. 10. 55. Ul, 62, 162, 201, 273. 2S7

Flynn, F. V., Jfll

Ford, C. E., 466, 470

Yoyo, F., Jr., 76, lO'J, 142, U9, 161, 183, 201, 346, 351

Fosse, R., 227, 265

Foster, T., 158. 163, 201

Fottrell, P.. 174, 201

Fox, E., 220, 222. 26S

Fox, M. S., 193, 201

Fraenkel-Conrat, H., 120. 121, l.'fO, 240, 266, 284, 285, 2SS, 382, .^0/. 478, 479, 483. 484, 485, 486. 487, 488, 489, 490, 491. 492. 495. 497, 498, 499, 500, 501, 504. 509. 512. 513, 517, 518, 519. 520

Franklin, R. E., 114, 1^9, 479, 487, 51S

Franzen, J. S., 170, 201

Eraser, D., 10, 62, 117, 120, 149, 150, 244, 266, 273, 2S7

Froderiksen, S., 243, 266

Freese. E.. 49. 62, 96, 97. 109, 142, 11,9, 217, 219, 220. 222, 223, 226. 227, 228, 229. 230, 231, 234. 237, 238, 239. 246. 247, 248. 249. 250, 251, 252, 253, 255, 257, 259. 261. 263, 264, ^05, 266, 453, 471

Freese, E. B., 222, 227, 228, 229, 230, 238, 239, 249, 250, 251. 253, 255, 257, 258, 259. 260. 261. 262. 263, 266

Freese, H. L., 156, 201

Freifelder. D.. 123, 126, 139, 149

French, E. F.. 312, 351

Frenster, A., 395, 401

Fresco, J. R., 247, 266, 48S, 517

Frick, P. G., 434, 471

Fricke, H., 242, 266

Friedewald, W. F., 242, 266

Friedkin, M., 5, 6, 7, 8, 10, 11, 50, 62, 64, 160. 202

Frisch-Xiggemeyer, W., 478, 518

Frontali. C., 119, I48

Fuerst, C. J., 119, 151

Fukai, H., 375, 403

Fukuhara, H., 281, 2S8

Funatsu, G., 486, 519

Gall, J. G., 69, 76, 92. 96, 107, 109. 159, 187, 202

Gallant, J., 179, 187, 202

Gammack, D. B., 423, 424, 438, 441, 440,

471, 473 Gamow, G., 285, 287, 510, 518 Gard, S., 243, 266 Garen. A., 428, 471 Garlick, J. P., 445, 474 Gaschlbauer, W., 172, 206 Gay, H., 97, 109 Gehatia, M., 142, 150 Gehring-Miiller, R., 413, 414, 415, 416,

417, 470 Geidu.schek, E. P., 79, 109, 185-186, 202,

227, 266, 280, 287 Gelbcr, B., 160, 188, 206 Gelfant, S., 191, 202 Gellert, M., 114, 115, 149 Genin, C., 281, 288 Gerald, P. S., 413, 423. 431, 432,, 433, 441,

446, 447, 454, 463, 471 Giauque, B. G., 157, 169, 194, 204 Gierer, A., 225. 251, 26S, 382. 4OI, 478,

479, 487, 488, 489, 498. 513, 518 Gilbert. W., 186, 202. 274, 276. 281, 287,

315, 326, 329, 344, 351, 383, 384. 386.

402 Giles, N. H., 91, 96, 109, 234, 235. 243.

252, 266, 267 Ginoza, W., 487, 490, 518, 519 Gish, D. T., 485, 486, 518, 519, 520 Gladstone, L.. 280. 289 Glaess, E., 245, 266 Glass, B., 241, 266 Glass, E. A., 218, 266 Glassman, E.. 356, 358, 4OO, 4OI, 403 Gnafakis, N., 442, 471 Goddard. J. W., 204 Godin, C., 373, 4OO Godman, C. G., 92, 109. 157, 200 Goldstein. A.. 156, 174. 175, 184, 202, 375,

377, 401 Goldstein, D. A.. 114, I48 Goldstein, D. B., 156, 174, 175, 202 Goldstein, J., 356. 402, 414, 415, 417, 471 Goldstein. L.. 395, 396. 397, 4OI Goldthwait, D., 311, 352 Gollub, E. G., 278, 287 Gordon, J., 358, 401 Gordon, M. P.. 497, 518 Gosling, R. G., 114, 151

AUTHOR INDEX

527

Gots, J. S., 50, 6i, 278, 2S7

Gottschling, H., 219, 266

Gouttas, A., 453, 471

Grant, P., 164, 202

Gratzer, W. B, 406, 408, 416, 418, 428,

470, 471

Gray, E. D., 13, 62, 171, 172, 202, 205

Graziosi, F., 119, I4S, 149

Gregg, J. R., 164, 202

Green, D., MacD., 239, 266

Green, M., 10, 62

Green, M. H., 297, 351

Greenberg, D. M., 7, 8, 62, 63

Greenberg, G. R., 9, 16, 50, 62, 64, 273, 288

Greengard, O., 362, 4OI

Greenough, W. B., Ill, 459, 467, 474

Greer, S. B., 218, 222, 231, 266, 269

Griboff, G., 219, 269

Griesbach, L., 9, 62

Griffith, J. S., 285, 287, 431, 474, 510, 517

Gros, F., 174, 176, 179, 186, 202, 203. 274, 276, 278, 281, 287, 288, 297, 299, 315, 326, 329, 344, 348, 351, 352, 363, 382, 383, 384, 385, 386, 394, 399, 4OI, 402, 403, 404

Gross, J., 299, 351

Grossi, L. G., 364, 366, 402, 403

Grossman, L., 120, 150, 189, 203

Gruelich, R. C, 196, 202

Grumbach, M. M., 88, 89, 91, 110

Grunberg-Manago, M., 280, 287, 348, 351

Guidotti, G., 413, 414, 415, 416, 417, 452,

471, 472 Gulyas, S., 193, 206 Guttes, E., 181, 190, 202 Guttes, S., 159, 181, 190, 202, 204 Glitz, H., 226, 243, 252, 266

H

Haagen-Smit, A. J., 299, 350, 354, 401 Haas, F. L,, 180, 187, 201, 244, 245, 266.

269 Hadorn, E., 235, 266 Hammerling, J., 395, 402 Hagerty, G., 24, 63 Haggmark, A., 3, 4, 5, 9, 50, 60 Hakala, M. T., 219, 266 Halberg, F., 166, 191, 200 Halbrecht, J., 421, 471

Hall, B. D., 186, 202, 274, 275, 279, 286,

287, 288, 310, 315, 347, 351, 352, 381,

383, 389, 400, 402, 403 Hall, C. E., 81, 97, 109, 295, 296, 307, 351 Hall, D., 186, 205

Halvorson, H., 362, 378, 384, 402, 404 Hamaguchi, K., 185-186, 202 Hamilton, L. D., 125, 126, 149, 150 Hamilton, M. G., 292, 301, 352 Hammarsten, E., 3, 62 Hanawalt, P. C, 89, 110, 142, 149, 150,

156, 174, 175, 202, 203 Handschumacher, R. E., 218, 219, 266 Hanson, J., 245, 265 Hanzon, V., 307, 351 Hardesty, B., 399 Harding, C. V., 191, 202 Harm, W., 227, 266 Harold, F. M., 179, 202 Harrar, G., 132, 151 Harris, J. I., 303, 352, 485, 518 Harris, J. K., 485, 518 Hart, R. G., 487, 518 Hartmann, G., 356, 358, 402 Hartmann, P. E., 214, 265 Haschemeyer, R., 479, 488, 518 Haselkorn, R., 488, 517, 518 Hasserodt, V., 416, 418, 440, 471 Haut, W. F., 93, 95, 111, 195, 206 Hayashi, M, 275, 276, 279, 287, 337, 351,

382, 383, 402 Hecht, L. I., 13, 62. 166, 171, 202. 357,

359, 363, 364, 365, 402 Hede, R., 123, 149 Heidelberger, C, 9, 62 Heiner, R. E.. 238, 266 Helinski, D. R.. 215, 266. 285, 287. 428,

471 Hell, A., 396, 403 Hems, R., 17, 63 Hendley, D. D., 225, 264 H(>ndrickse, R. G., 454, 471 Henning, U., 285, 287 Heppel, L. A, 13, 24, 33, 62, 63, 280, 288 Herbert, E., 13, 62, 171, 202, 357, 362, 402 Hermodsson, L. M., 307, 351 Hershey, A. D„ 72, 109. 119, 122, 123,

124, 125, 130, 136, 139, I48, 149. 150,

162, 188, 201, 202, 210-211, 266. 330,

352, 478, 518

528

ALTIIUR INDEX

Hiatt, C. W., 225, 266, 267

Hiatt, H. H, 14, 61. 62, 171, 172, 174, 186,

200, 202, 21 A, 276, 281, 2S7 , 315, 326,

329, 344, 351, 383, 384, 386, J,02 Hidaka, F., 245, 265 Hieulle, A., 227, 265 Hill, N. O.. 119, 150 Hill, R. J.. 413. 414. 415. 416, 417, 471.

472 Hill. R. L., 415, 424, 427, .'fll, 47.'t HiUschmann. N., 413, 414, 415, 416, 417,

470 Hilmoe, R. J., 24, 33, 62, 63 Hilse, K., 413, 414, 415, 416, 417, 470 Hinsholwood, C. N., 299, 351 Hirokawa, R., 362, 363, 364, 402, 403 Hirota, Y., 225, 266 Hiruk, C. 245, 265 Hitzis. W. H., 434, 471 Hoagland. M. B., 354, 355, 356, 357, 362,

363, 364, 365, 368, 369, 374, 402 Hobam, G., 413, 414, 415. 416, 417, 470 Hoch, H., 407, 471 Hodes, M. E., 230, 260 Horlein, H., 422, 471 Hoff-J0rgenson, E., 164, 202 Hogness, D. S., 133, 149 Hokin, L. E., 277, 287 Hokin, M. R., 277, 287 Holden, M., 243, 267 Hollaendcr, A., 243, 266 Holley. R. W.. 356, 358. 361, 401 Hollingworth. B. R.. 296, 301. 305, 307,

352 Holloway, B., 354, 402 Holmes, K. C, 479, 487, 518 Holzner, J. H., 176, 202 Hooper, C. W., 125, 150 Hopkins, J. W., 281, 288 Hopkins, J., 387, 404 Hornscy, S., 156, 202 Horton. B., 444, 448, 471 Hotchkiss, R. D.. 157-158, 193, 201, 202 Hotta, Y., 159, 163, 164, 171, 174, 202, 205 Howard, A., 155, 156, 194, 202 Hoyer, B. H., 294, 350 Hu, A., 362, 378, 402 Huang, R. C, 186, 200, 280, 287, 397, 4OI Huohns, E. R.. 421, 423, 424, 438, 441.

442, 446, 455, 471. 473

Hiilsniann, W. C, 356, .358, 364. 365. 367,

402, 403 Hughes, W. T., 66, 73, 74, 75, 104, ;;; Hughe.s, W. L., 138, 151, 156, 161, 183,

206 Hughes-Schrader, S., 67, 109 Huisman, T. H. J., 408, 418, 434, 444,

445, 448, 456, 462, 463, 4^1, 472, ^73,

474 Hultin, T., 354, 363, 364, 365, 366, 368,

379, 402, 404 Humphreys, G. K., 7, 8, 62 Hunt, J. A., 416, 424, 425, 436, 438, 443,

454, 472 Hunter, A. W. S., 90, 92, 109 Hurlbert, R. B., 3, 63 Hurwitz, J., 14, 62, 280, 287, 347, 348,

351, 382, 402 Huskins, C. L., 90, 92, 109 Hutchinson, W. D., 440, 474 Huxley, H. E., 296, 307, 351

I

Ingraham, L., 125, 149

Ingram, V. M., 406, 409, 410, 413, 416, 418, 419, 420, 423, 424, 425, 426, 427, 436, 438, 440, 442, 443, 444, 446, 447, 448, 449, 450, 451, 452, 453, 454, 457, 465, 470, 471, 472, 473

Inman, R. B., 50, 60

Irvine, R. A., 462, 473

Ishihama, A., 315, 352

Ishii, M., 490, 519

Itano, H. A., 407, 422, 423, 424, 426, 438, 439, 440, 441, 442, 451, 457, 458, 472, 473, 474

Iwamura, T., 158, 202

Iwanowsky, D., 477, 518

Iyer, V. N., 218, 234, 266

J

Jackson, C. H.. 469, 474

Jacob, F., 145, 146, 149. 186, 190, 200, 202, 244, 266, 272, 274, 277, 287, 288, 310, 344, 350, 351, 381, 383, 384, 389, 397, 401, 402, 403, 453, 464, 472

Jacob, G. F., 415, 462, 463, 472

Jaffe. J. J., 13, 61, 171, 172, 201

Jalin, T. L., 192, 205

AUTHOR INDEX

529

Jandl, J. H., 467, 469

Jardetzky, C. D., 166, 191, 200

Jecner, R., 497, 5 IS

Jensen, K. A., 234, 240, 266, 267

Jerne, N. K., 139, 151

Jesaitis, M. A., 56, 62

Jim, R. T. S., 424, 427, J,74

Jkeda, Y., 245, 267

Joklik, W. K., 17, 60

Jones, A. S., 230, 26^

Jones, O., 366, 390, 391, 392, 393, 403

Jones, O. W., 282, 283, 284, 2SS, 434, 435,

436, 473, 512, 515, 519 Jones, R. T., 416, 417, 418, 421, 441, 443,

453, 472, 473, 474 Jonxis, J. H. P., 374, 402, 415, 419, 446,

447, 459, 470, 472, 473 Josse, J., 16, 32, 33, 34, 35, 36, 37, 39, 46,

55, 56, 57, 58, 59, 62, 162, 202, 203,

273, 287, 337, 351

K

Kaars Sijpesteijn, J. A., 447, 473

Kaesberg, P., 284, 2S9

Kahn, P., 141, 149

Kaiser, A. D., 32, 33, 34, 35, 36, 37, 39,

46, 62, 133, 134, 135, 149, 337, 351 Kalckar, H. M., 5, 13, 61, 62 Kalle, G. P., 50, 62 Kameyama, T., 281, 288, 348, 351, 362,

387, 389, 394, 402, 403 Kanazir, D., 165, 166, 202, 218, 267 Kane, M. R, 12, 63 Kaplan, E., 450, 475 Kaplan, M. S., 222, 267 Kaplan, R. W., 244, 265 Kaplan, W. D., 240, 268 Karaklis, A., 442, 445, 447, 471 Karnofsky, D. A., 165, 202 Karpfel, Z., 196, 202 Kas, E., 185, 203 Kaudewitz, F., 225, 267 Kaufman, E., 225, 267 Kaufmann, B. P., 97 Kay, E. R, 185, 201 Kayeyama, T, 187, 201 Kayser, A, 281, 288 Keck, K., 10, 62, 273, 287 Keighly, G., 299, 350, 354, 4OI

Keir, H. M, 13, 62, 171, 172, 186, 190,

202, 205 KcUcnbergcr, E., 98, 109, 117, 118, 119, 142, 149, 179, 180, 186, 187, 188, 203 Kellenberger, G., 86, 87, 109, 121, 141,

149, 150 Keller, E, 354, 404 Keller, E. B., 299, 351 Kendrew, F. R. S., 430, 431, 432, 433, ^74 Kendrew, J. C., 285, 289, 430, 472 Kernot, B. A., 362, 4OI Khorana, H. G., 24, 62 Kielley, R. K., 13, 62 Kihara, H., 362, 378, 402 Kihlman, B, 218, 235, 267 Kilkson, R., 117, 150 Kimball, R. F, 157, 187, 203, 204, 243,

266 Kingma, S., 424, 434, 473 Kinosita, R., 156, 205 Kirby-Smith, J. S., 245, 267 Kirchner, C. W. J., 252, 264, 267 Kirk, J., 240, 267 Kirkland, R. J., 17, 62 Kirsclifeld, S., 219, 269 Kitani, Y., 96, 109 Kitazume, Y., 276, 287 Kjeldgaard, N. O., 192, 195, 203, 205 Kleinhauer, E., 461, 462, 472 Kleinschmidt, A., 142, 150 Klenow, H., 13, 17, 62, 170, 203, 204, 243,

266 Klibansky, C., 421, 471 Klug, A, 121, 150, 479, 480, 483, 487, 518 KIuss, B. C., 190, 204 Knight, C. A, 485, 486, 488, 489, 493, 494,

497, 517, 518, 519, 520 Koch, A. L., 156, 201, 218, 267 Koch, B., 462, 463, 472 Koch, G., 119, 150 Koerner, J. F., 50, 55, 57, 62 Koler, R. D., 453, 473 Kolmark, G, 234, 235, 240, 266, 267 Konigsberg, W., 413, 414, 415, 416, 417,

471, 472 Konrad, M., 245, 267 Konzak, C. F., 238, 266 Kornberg, A, 5, 6, 13, 14, 16, 20, 21,

22, 23, 24, 25, 26, 27, 29. 31, 32, 33, 34,

35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45,

530

AUTHOR INDEX

46, 47, 48, 49, 50, 52, 55, 56, 57, 58, 59,

60, 61, 62, 63, 64, 160, 161, 162, 200,

203, 253, 267, 273, 280, 2S7 , 337, 351 Kornbcrg, S. R., 16, 46, 55, 56, 57, 58, 59,

62, 64, 162, 203, 253, 267, 273, 287 Kozinski, A. W., 86, 110, 140, 150, 222,

267 Kramer, G., 482, 518 Krakow, J. S, 13, 61, 171, 172, 201 Kraus, A. P., 462, 463, 472 Krcbs, H. A., 17, 63 Krcvans. J. R., 445, 463, 464, 474 Krieg, D. R., 239, 244, 266, 267 Kroger, H., 187, 201, 204, 280, 288 Krone, W., 357, 404 Kruh, J, 374, 398, 399, 402, 403 Kiinzer, W., 421, 471 Kulad, C. G, 186, 202 Kunitz, M., 19, 63 Kimkel, H. G, 407, 408, 418, 420, 445,

449, 472 Kuriki, T, 165, 166, 204 Kuriki, Y., 164, 165, 166, 203, 204 Kurland, C. A., 315, 326, 329, 351 Kurland, C. G, 274, 276, 281, 287, 305,

344, 351, 383, 402

Labie, D, 374, 403

Lacks, S., 363, 402

LaCour, L. F., 76, 110

Lagerkvist, V., 360, 361, 400

Laird, C., 220, 222, 268

Laite, M. B., 156, 201

Lajtha, L. G., 156, 203, 466, 472

Lamarck, J. B., 208, 267

Lamborg, M, 355, 386, 402

Lamfrom, H., 355, 362, 375, 378, 398, 399,

402, 403 Lamond, W. A., 218, 2^7 Lampen, J. 0., 5, 63 Langridge, R., 125, 150 Lark, C, 195, 203 Lark, K. G., 156, 158, 165, 166, 167, 169,

170, 173, 178, 179, 180, 181, 182, 184,

186, 187, 190, 191, 192, 195, 196, 197,

203, 296, 351 Laskowski, M., 20, 24, 61, 63 Laskowski, M., Sr., 185, 203

Latarjet, R., 244, 267

Laudcrbach, A. L., 192, 205

LaulTer, M. A., 114, 126, I48, 480, 517

Laurila, U., 24, 63

Lawley, P. D., 237, 265

Lea, D. E., 242, 267

Leahy, J., 364, 365, 366, 375, 4OO

Leavell, B. S., 457, 4^4

Lcgault, R. R., 238, 266

Leliinan, J. R., 253, 267

Lehman, I. R., 13, 14, 18, 20, 21, 23, 24,

26, 27. 28, 29, 31, 32, 33, 40, 41, 42, 43,

44, 45, 46, 47, 48, 50, 56, 57, 60, 61, 63,

122, 150, 160, 200, 203, 280, ^67 Lchmann, H., 406, 415, 422, 423, 424, 425,

426, 427, 436, 441, 449, 450, 454, 462,

463, 469, 470, 471, 472, 474 Leloir, L. F, 57, 61 Lengyel, P., 282, 283, 288, 390, 391, 392,

394, 400, 400, 402, 404, 434, 435, 436,

473, 474, 512, 515, 518, 519 LePage, G. A., 165, 203 Lerman, L. S., 119, 150, 193, 203, 225, 267 Lett, J. T., 237, 238, 264 Leuppold, V, 245, 252, 267 Levan, A, 218, 267 Levine, L., 119, 120, 150 Levinthal, C., 77, 79, 109, 110, 122, 123,

125, 130, 136, 139, 149, 150, 161, 203,

225, 265, 428, 471 Lichtenstein, J, 9, 61, 62, 170, 201, 273,

275, 287 Lichtler, E., 13, 62 Lie, H. G., 455, 4^3 Lieberman, I, 5, 62, 176, 203 Liebold, B., 415, 470, 473 Lie-Injo, L. E., 454, 455, 473 Lima-de-Faria, A., 68, 88, 110 Lipmann, F, 188, 203. 356, 357, 358, 363,

364, 365, 366, 367, 368, 370, 371, 373,

379, 380, 394, 399, 402, 403, 404 Lipschitz, R., 230, 269 Lissouba, P., 96, 110 Litman, R. M., 219, 221, 222, 226, 267 Littauer. U. Z., 368, 369, 4OI Litt, M., 488, 517 Littlefield, J. W, 80, 109, 219, 267, 299.

351, 488, 518 Liu, J. C., 423, 424, 470

AUTHOR INDEX

531

Livingslono, F. B., 422, 4?3

Loc-hte, H. L., Jr., 459, 467, 474

Loeb, M. R., 9, 61

L0trup, S., 164, 202

Loftfield, R, 373, 402

Logan, R, 277, 2S8

Longer, L., 170, 203

Lorkiewicz, Z., 220, 222, 233, 235, 267, 269

Lotz, C, 240, 265

Loutit, J. F., 466, 470

Loveless, A., 234, 235, 238, 267

Lowy, P. H, 299, 350. 354, 401

Luippold, H. E., 91, 93, 111, 242, 269

Luria, S. E, 208, 225, 243, 265, 267, 292,

351 Luzzati, v., 126, 150, 305, 352 Lyons, M. E., 88, 110

M

Maal0e, O, 89, 110, 142, 143, 149, 150,

155, 156, 158, 170, 174, 175, 192, 195,

202, 203, 205, 243, 266 McCalla, K., 416, 417, 443, 474 McCarthy, B. J., 179, 200, 277, 287, 296,

297, 298, 305, 312, 313, 315, 322, 323,

329, 333, 344, 350, 351 McCarty, M., 72, 108 McClary, D. O, 157, 204 McClintock, B., 82, 87, 109, 110 McClure, F. T., 277, 288 McDonald, B. B., 157, 190, 203 McDonald, M. R., 97, 109 McElya, A. B., 13, 61, 173, 200 McFall, E., 155, 203, 384, 403 McGregor, H. C, 69, 107, 109 Maclver, J. E., 462, 463, 473, 474 Mackal, R. P., 162, 165, 166, 204 McKinley, E. B., 243, 265, 267 McLeish, J., 157, 203 MacLeod, C. M., 72, 108 McMaster, R. D., 163, 206 McNutt, W. S., 3, 4, 5, 63 McQuillen, K, 297, 299, 305, 312, 352 Maestre, M. F, 117, 150 Magrath. D. J., 243, 269 Maheshwari, N., 186, 200, 280, 287, 397,

401 Mahler, H. R., 10, 62, 120, 150, 273, 287 Main, R. K., 166, 203

Maley, F., 10, 15, 63

Maley, G. F., 10, 63

Mandel, H. G., 497, 519

Mandcl, L. R., 10, 63

Mandell, J. D., 122, 150, 330, 352

Manson, L. A., 5, 63

Mantsavinos, R., 13, 61

Marinelli, L. D., 241, 267

Markham, R., 24, 33, 62, 63, 493, 497, 519

Marmur, J, 44, 63, 81, 110, 119, 144, 145,

147, 148, 149, 150, 151, 189, 203, 230,

231, 265, 267 Marquardt, H, 241, 267 Martin, N. J., 409, 410, 473 Martin, R., 390, 391, 392, 393, 403 Martin, R. G., 282, 283, 288, 512, 515, 519,

434, 435, 436, 473 Maruyama, Y., 156, 158, 170, 173, 178,

190, 197, 203 Marvin, D. A, 125, 126, 149, 150 Mashima, S., 245, 267 Masson, F., 126, 150 Mather, K., 90, 92, 110 Matheson, A. T., 119, 150 Matsudo. G, 416, 418, 421, 473, 474 Matthaei, J. H, 282, 283, 284, 288, 345,

348, 352, 366, 387, 390, 391, 392, 393,

394, 403, 434, 435, 436, 473, 490, 491,

511, 512, 515, 519, 520 Matthews, R. E. F., 497, 519 Maximow, A. A., 459, 473 Mayor, H. D., 119, 150 Mazia, D., 73, 110, 197, 200, 203 Mendel, G., 209, 267 Mendelsohn, M. L., 156, 161, 191, 203 Merrill, S.. 358, 402 Meselson, M, 66, 77, 79, 80, 81, 86, 107,

110, 121, 133, 138, 139, 141, 142, 150,

151, 161, 183, 186, 189, 200, 203, 215,

267, 274, 287, 310, 344, 350, 383, 401 Meyering, C. A., 408, 444, 471 Meyers, R. W., 480, 519 Michelson, A. M., 20, 63 Micou, J, 395, 396, 397, 4OI Midgley, J. E., 277, 287, 306, 311, 330,

331, 337, 340, 352 Miescher, F., 272, 288 Miles, H. T, 215, 267 Miller, A., 398, 403

532

AUTHOR INDEX

Miller, I. H.. 7!). 110

Miller. M. L., 422. /,7S

Miller, O. L.. 69, 107, 110

Minagawa, T., 119, 120, 150

Mineklcr, S., 157, 20^

Minnich, V., 416, 442, 443, 454, 473

Mirsky, A. E., 72, 110. 395, 397. 400. .',01.

m

Mitchener. J. W., 462, 463, J,73, 474

Mitchi.son, J. M., 157, 206

Mitra, S., 92, 93, 110

Mitsui, N., 315, 352

Miura, K., 245, 265

Miyake, T., 219, 247, 267

Modiano, G, 418, 454, 474

Moldave, K., 364, 366, 367, 4OI, 402, 403

Monod, J., 190, 202. 272, 277, 278, 2S7 ,

28S, 344, 351, 381, 383, 384, 385, 389,

397, 401, 402, 403, 453, 464, 472 Monroe, R., 366, 370, 380, 394, 399, 403 Montgomery, T. L., 450, 451, 470 Montreuil, J., 277, 2S7 Montsavinos, R., 171, 172, 185, 201, 203 Moore, C. V., 416, 442, 443, 454, 473 Moore, E. C, 63 Moore, H. A., Jr., 156, 191, 203 Mooren, D., 9, 62 Mordah, J., 204 Morishima, A., 88, 89, 91, 110 Morris, A. J., 362, 364, 365, 366, 370, 371,

374, 375, 378, 379, 394, 399, 4OO, 403 Morris, N. R., 170, 203 Morris, R., 68, 110 Morrison, M., 408, 473 Mosor, H., 234, 240, 264 Moses, M. J., 72, 91, 107, 110 Motulsky, A. G., 453, 454, 473 Moussear, J., 96, 110 Mueller, G. C, 194, 205 Midler, R., 415, 470

Miiller-Ebcrhardt, V., 418, 445, 449, 472 Muirhead, H., 409, 430, 473 Muller, C. J., 419, 420, 421, 424, 434, 446,

447, 470, 473 Muller, H. J., 210, 26S Mundry, K. W., 225, 251, 26S, 498, 499,

618 Mun.sch-Peterson, A., 170, 203 Murakami, W. T., 120, 150 Murayama. M.. 423, 446, 450, 473

Muzzolini, M., 418, 545, 474

N

Nakada, D., 175, 180, 203, 248, 268 Nakamoto, T., 186, 202, 280, 287, 289,

347, 348, 352, 382, 404 Naono, S., 278, 287, 288, 385, 4OI, 402, 403 Xarita, K., 483, 485, 517, 518 Nasrat, G. E., 240, 268 Nathans, D., 188, 203, 364, 365, 366, 367,

368, 370, 373, 380, 394, 397, 399, 403 Neeb, H., 447, 473 Neel, J. v., 406, 422, 436, 449, 450, 457,

463-464, 465, 472, 473, 474, 475 Neidhardt, F. G., 176, 179, 204 Newmark, P., 480, 519 Newsome, F. E., 469, 474 Newton, A. A., 157, 158-159, 193, 195,

204, 205 Nicolaieff, A.. 126, 150 Niggli, H., 235, 266 Nilan, R. A.. 238, 266 Nirenberg, M. W., 282, 283, 284, 288,

345, 348, 352, 366, 387, 390, 391, 392,

393, 394. 403. 434, 435, 436, 473, 490,

491, 511. 512, 515, 519, 520 Nisman, B., 281, 288, 310, 352 Niu, C. I., 485, 519 Niven, J. S. F., 120, 148 Nixon, H. L., 479, 519 Nohara, H., 354, 403 Nomura, M., 121, 150, 179, 204, 274, 279,

288, 298, 310, 315, 352, 383, 403 Norcross, F., 296, 351 North, A. C. T., 114, 125, 150, 409, 430,

473 Novelli, G. D., 187, 201, 281, 288, 348,

351, 355, 356, 362, 387, 389, 394, 402,

403, 404 Novick, A., 218, 219, 266, 268 Nygaard, 0. F., 159, 204

O'Brien, J. A. P., 440, 47/

Ochoa, S., 13, 15, 63, 187, 201, 204. 280, 282, 283, 287, 288, 348, 351, 390, 391, 392, 394, 400, 4OO, 402, 404, 434, 435, 436, 473, 474, 512, 515, 518, 519

O'Donnell, J. F., 162. 165, 166, 204

Oehlkers, F.. 210, 218, 245, 268

AUTHOR INDEX

533

Ostergren, G., 92, UO

Ofengand, E,, 356. 358, 359, 361, m, 403

Ogata, K, 354, 362, 363, 364, 402, 403

Ogiir, M., 157, m

Okazaki, R., 164, 165, 166, 168, 169, 174,

176, 180, 203, 204, 205 Okazaki, T, 165, 166, 169, 174, 176, 180.

204, 205 Okubo, S, 187, 205. 233, 269 Okun, L., 220, 222, 268 Olive, L. S., 96, 109, 110 Oliver, R., 156, 203, 466, 472 Olson. M., 362, 374, 403 Omori, S., 362, 363, 364, 402, 403 Oota, Y., 315, 352 Opara-Kubinska, Z., 220, 269 Ord, M. G., 165, 166, 204 Orgel, A., 225, 255, 260, 265, 268, 285, 287,

510, 517 Ortiz, P. J., 348, 351 Osawa, S., 315, 352, 357, 403 Osgood, E. E., 453, 473 Otaka, E., 315, 352 Ove, P., 176, 203 Overgaard-Hansen, K., 170, 204 Overman, R. R., 469, 474

Padieu, P., 374, 402

Paigen, K., 121, 133, 151, 298, 352

Painter, R. B., 156, 157, 169, 194, 201,

204, 396, 401 Palade, G. E., 292, 352, 362, 378, 404 Paladini, A. C, 57, 61 Papaspyrou, A., 442, 454, 471 Pappas, G. D., 97, 104, 105, 110 Pardee, A. B., 155, 159, 160, 174, 200, 204,

219, 221, 222, 267, 111, 292, 298, 343,

352, 383, 384, 403 Parizek, J., 165, 166, 204 Parkins, G. M., 237, 238, 264 Partridge, C. W. H., 243, 266 Paul, J., 13, 17, 64, 172, 173-174, 176. 177.

201, 204, 206 Pauling, L., 407, 409, 422, 457, 473 Payne, R. A., 444, 448, 471 Pearce, J. D., 196, 201 Pelc, S. R., 76, 110, 155, 156, 202 Perry, R. P., 396, 403 Perutz, M. F., 409. 430, 473

Peterinann, M. L., 292, 301, 352

Peters, T., 362, 403

Peterson, E. A., 408, 473, 489, 519

Phear, E. A., 7, 63

Phillips, D. C, 430, 472

Pinkerton, T. C, 130, 139, 151

Pirie, N. W., 478, 517

Pizer, L., 382, 403

Plaut, W. S., 73, no, 396, 4OI

Plever, E., 9, 62

Plough, H. H., 231, 268

Pogo, A. O., 204

Pomper, S., 245, 268

Popper, H., 176, 202

Portier, A., 438, 446, 470

Poschkis, K. E., 204

Potter, R. L., 12, 63, 165, 166, 204

Potter, V. R., 12-13, 51, 61, 62, 63, 166,

171, 172, 173, 200, 202 Poulik, M. D., 407, 473 Pouyet, J., 18-19, 63 Prankerd, T. A. J., 451, 470 Pratt, D., 222, 253, 267, 268 Pratt, E. A., 56, 57, 63 Preiss, J., 356, 357, 359, 403 Prescott, D., 396, 403 Prescott, D. M., 156. 187, 190, 204 Prestidge, L. S., 174, 204, 298, 343, 352 Price, C. C, 232, 268 Prins, H. K., 408, 473 Prusoff. W. H., 170, 204 Punt, K., 444, 445, 472 Putnam, F. W., 139, 150, 240, 268

Q

Quastler, H.. 191, 193. 194, 204, 205

Raacke, I. D., 355, 357, 358, 362, 403 Rabinowitz, M., 362, 374, 380, 403 Racker, E.. 5. 63

Radding, C. M., 40. 41, 42, 43, 44, 45, 63 Ramachandran, L. K., 479, 485, 517, 518,

519 Ranney, H. M., 426, 436, 454, 470, 473 Ransome, S., 454, 471 Paper, A. B., 415, 438, 441, 462, 463, 472,

473 Rapoport, J. A., 210, 234, 235, 240, 268 Razell, W. E., 24, 62

534

AUTHOR INDEX

R(-a.l, J.. 242, 269

Keddi, K. K.. 347, 348, 353

Roicluinl, P., 3, 4, 5, 9, 50, 60, 62, 63,

160, 169, 204 Reiner, B., 233, 237, 238, 26S Remy, C. N., 5, 63 Remy, W. T., 5, 63 Rhinesmith, H. S, 409, 410, >^73 Rho, J. H, 395, 396, 403 Rich, A., 114, 125, 150, 449, 461, 462, 473 Rich, M. A., 219, 265 Richards, B. M, 157, 158, 192, 204 Richards, J., 13, 62, 171, 172, 202, 205 Rigas, D. A., 453, 473 Riley, M, 277, 2SS, 383, 403 Ripley, S, 354, 402 Ris, H, 72, 96, 97, 110 Rissebrough, R. W., 186, 202, 274, 276, 281, 2S7. 315, 326, 329, 344, 351, 383, 402 Ritter, D. B., 294, 350 Rixon, R. H., 192, 206 Rizet, G., 96, 110 Roberts, D., 5, 62

Roberts, R. B., 178, 179, 186, 200, 204, 277, 287, 2SS, 293, 297, 298, 299, 301, 303, 305, 309-310, 311, 312, 313, 315, 333, 344, 350, 351, 352, 362, 363, 378, 401, 403 Robinson, A. R, 422, 450, 473, 475 Robinson, E. A., 423, 424, 426, 438, 439,

441, 442, 472 Robson, J. M., 210, 234, 235, 264 Rogers, S., 219, 268 Rolfe, R., 143, 150 Roll, P. M., 3, 4, 63 Roller, A., 140, 150 Rosa, J., 398, 399, 402 Rose, I. A., 3, 63

Rosenberg, B. H., 66, 81, 82, 89, 97, 101, 107, 109, 144, 145, I4S. 183, 184, 201, 203 Roscnkranz, H. S., 107, 108 Rosenthal, S. M., 119, I48 Ross, W. C. J., 232, 268 Rosseels, J., 497, 518 Rossman, M. G, 409, 430, 473 Rotherham, J., 165, 166, 204, 205 Rothman, F., 471

Kownd, P., 117, 150

Rubenslein, I., 116, 117, 120, 123, 124, 150

Rucknagcl, D. L, 406, 422, 436, 449, 450,

461, 467, 474 Rudloff, v., 413, 414, 415, 416, 417, 470 Rudner, R., 189, 205, 219, 223, 252, 263,

268 Rueckert, R. R., 194, 205 Riisch, H. P., 159, 204 Ru.shizky, G. W., 489, 497, 519 Russell, E., 363, 371, 372, 379, 386, 398,

400 Pulhcrfi, L., 3, 63 Ryan, F. J., 175, 203, 210, 248, 268 Ryter, A, 117, 118, 119, 149

Sable, H. Z., 12, 63

Sadron, C, 145, I48

Saeki, R., 375, 403

Saluste, E., 3, 62

Salzberger, M., 461, 470

Sams, J, 241, 243, 265

Sarkar, N. K., 185, 201, 205

Sarvella, P., 68, 110

Saska, J., 196, 202

Sato, C., 395, 404

Sato, G. H., 139, 151

Sato, J., 375, 403, 404

Sauerbier, W., 240, 268

Sauerland, H., 92, 110

Sax, K., 91, 92, 93, 110. 231, 241, 242, 268

Scarano, E., 10, 63

Schaad, J. D. G., 444, 445, 1/72

Schachman, H. K., 29, 40, 41, 42, 43, 44,

45, 63, 292, 351, 352, 487, 517 Schachter, E., 126, I48 Schachtschabel, D., 357, 404 Schaeehter, M., 142, 149, 155, 156, 175.

192, 195, 202, 203, 205 Rchaoffer. P., 146, 149 Schaffcr, F. L., 478, 519 Schairer, M. V., 76, 111 Schapira, G., 374, 398, 399, 402, 403 Scheimer, J., 9, 62 Schell, P, 228, 265 Scherbaum, O. H, 192, 206, 206 Schildkraut, C. L., 81, 110, 144, 147, 150,

151. 230, 267

AUTHOR INDEX

535

Schlessinger, D., 274, 281, 2SS, 296, 297, 299, 301, 307, 347, 332. 383, 386, 394,

399, m

Schlesinger, S, 12, 6',^. 165, 166, m Schneider, J. H, 176, 203, 395, 396, 403 Schneider, M. J., 248, 268 Schneider, W. C, 165, 166, 204, 205 Schnek, A. G., 408, 420, 474 Schramm, G.. 225. 2GS, 382, 40L 478, 479,

480, 482, 485, 486, 487, 489, 490, 498,

517, 518, 519 Schroeder, W. A., 408, 409, 410, 416, 417,

418, 420, 421, 440, 441, 443, 453, 469,

470, 472, 473, 474 Schuitzer, R. J., 9, 62 Schulman, H. M., 279, 2SS Schumacher, G., 480, 519 Schuster, H., 225, 227, 228, 268, 269, 479,

497, 498, 499, 513, 519, 520 Schwartz, D., 82, 96, 110 Schwartz, H. C, 424, 427, 471 Schwartz, I. R., 437, 438, 439, 441. 450,

451, 470 Schweet, R. S., 355, 356, 358, 361, 362,

363, 364, 365, 366. 367, 369, 370, 371,

372, 374, 375, 378, 379, 386, 394, 397,

398, 399, 400, 401. 403. 404 Schweigert, B. S., 3, 63 Schwerdt, C. E., 478, 519 Scopes, A. W., 159, 206 Scott, D. B. M., 156, 158, 205 Scott, E. M., 431, 474 Scott, J. E., 191, 200 Scott, J. F., 363, 364, 365, 402 Scott, S. S., 166, 203 Sechaud, J., 117, 118, 119, 149 Seed, J., 193, 205 Seeds, W. E, 114, 125, 150. 131 Setlow, R. B., 142, 151. 184, 205 Setlow, S. K., 184, 205 Shapiro, H. S., 50. 63. 230, 269 Shelton, J. B., 416, 417, 443, 474 Shehon, J. R.. 416, 417, 443, 474 Sherman, F. G., 191, 194, 204, 205 Shimura, K., 362, 375, 403, 404 Shooter, E. M., 423, 424, 438, 441, 442,

445, 446, 455, 471. 473, 474 Shooter, K. V., 13, 14, 61. 171, 172, 174,

190, 200, 201, 455, 471

Shore, V. C., 430, 472

Shugar, D., 20, 63, 245, 268

Sibatani, A., 397, 404

Siegel, A., 490, 500, 319

Siekevitz, P., 12-13, 63, 354, 362, 378, 404

Silva, D. M., 494, 518

Silvestroni, E., 418, 447, 449, 454, 474

Simkin, J., 378, 404

Simmons, N. S, 125, 149

Simms, E. S., 5, 13, 14, 20, 21, 23, 24, 26,

27, 29, 31, 32, 33, 46, 47, 48, 50, 60, 61,

62, 63, 160. 200, 203, 280. 287 Simon, E. H.. 79, 110, 161, 205 Singer, B., 120, 149, 479, 487, 488, 489,

490, 495, 497, 517, 518. 519 Singer, K., 461, 474 Singer, L.. 461, 474 Singer, M., 357. 404 Singer, M. F., 24, 63 Singer, R., 461, 470 Singer, S. J., 407, 422, 439, 440, 442, 457,

472, 473, 474 Siniscaico, M., 449, 470 Sinsheimer, R. L., 20, 26, 27, 38, 54, 64,

119, 121. 122, 139. 149. 151, 253, 268 Sisken, J. E., 156, 187, 205 Skinner, E. R., 445, 474 Skoda, J., 165, 166, 204 Slayter, H. S., 295, 296, 307, 331 Skiyser, M., 368, 4OO. 4OI Smelhe. R. M. S., 13, 17. 62. 64. 171, 172,

173-174, 186, 190, 202. 203, 206 Smith, C. L., 157. 193, 195. 205 Smith, D. B.. 409, 413, 475 Smith, E., 392, 404 Smith, E. L., 435, 436, 474 Smith, E. W., 438, 439, 474 Smith, G. M., 454, 470 Smith, H. W., 243, 269 Smith, J. D , 49, 61. 63. 98, 109. 120, 149,

219, 222, 265. 357, 401. 493, 497, 519 Smith, K., 358, 404 Smith, M. S., 50, 55, 57, 62 Smith, R., 170, 205 Smithies, 0., 407, 448, 474 Sneoke, N., 231, 265 Sobels, F. H., 240, 268 Sober, H. A., 408, 473. 489, 319

536

AUTHOR INDEX

Somervillo. R. L, 9, 16, 50, 62, 6/,, 273,

■JSS Spahr, F., 384. 386. 402 Spahr, P. F., 305. 306. 352, 357, 401 Spencer, M.. 126. 150, 215, 26S Speyer, J. F., 282, 283, 2SS, 390, 391, 392,

394. 400, 400, 402, 404, 434, 435, 436,

473, 474, 512. 515, 518, 519 Spiegelman. S., 179, 186, 187, 200, 202,

205, 274, 275. 276, 279, 280, 281, 2S7,

28S, 303, 305, 310, 311. 315. 337, 347.

351, 352, 362, 378, 382, 383, 384, 4OI,

402, 403, 404 Spilman, W., 147, I48, I40 Spinolli, v., 247, 269 Sprinson, D. B., 7, 61 Srinivason, B. D., 191, 202 Stacey, K. A., 232, 233, 264, 268 Stadler, L. J., 210, 245, 268, 269 Staehelin, M., 240, 268, 497, 498, 518, 519 Stahl, F. W., 66, 77, 79, 80, 81, 96, 107,

110, 133, 138, 142, 150, 151, 161, 183,

189, 203, 215, 220, 222, 243, 244, 265.

267, 268 Stamatoyannopoulos, G., 445, 447, 4^1 Stanier, R. Y., 292, 352 Stanley, W. M., 478, 485. 486, 493, 497,

518, 519, 520 Stanners, C. P., 155, 156, 193, 205, 206 Steffensen, D., 96, 110 Steinberg, D., 354. 373, 4OO. 404 Steiner, R. F., 225, 264, 269 Stent, G. S., 73, 109, 119, 139. 151, 155.

180, 203, 205, 222, 244, 268, 269, 384.

403 Stephenson. M. L., 354, 355, 357, 358, 359,

363, 364, 365, 402, 404 Stern, C, 87, 111 Stern, H., 158, 159, 163. 164. 171. 174. 201.

202, 205 Stevens, A., 280, 288, 382, 404 Stevens, M. A., 243. 269 Stoc-ken. L. A., 165, 166, 204 Stohlman, F., Jr., 466, 4'^4 Stone, L. H. A., 90, 110 Stone, W. S., 243, 244, 269 Storck, R., 186, 205, 275, 288, 310, 347.

352 Strandberg, B. E., 430, 472 Strauss, B., 187, 205

StraiLss, B. S., 233, 234. 239. 240, 269

Stroisinger, G., 131, 132. 151

Strolzoff, E.. 96. 111. 221. 223, 252, 263,

269 Stretton, A. O. W.. 410. 418, 419, 444,

447, 449, 450, 451, 452, 453, 457, 472 Strominger, J. L., 166, 205 Studior. F. W., 151 Stulborg. M. P., 355, 356, 404 Sturgeon. P., 422, 1^72 Suooka, X., 44, 64. 80. ///, 161, 183, 189,

205 Sugino, N., 165, 205 Sugino, Y., 165, 205 Sugiyama, T., 488-489, 497, 519 Suit, H. D., 467, 474 Sullivan, E., 416, 454, 470 Sunderland, N., 157, 203 Suskind, S. A., 179, 187, 202 Suzuka, I., 362, 404 Swanson, C. P., 245, 269 Swartz, M. N., 38, 39, 45, 47, 49, 50, 64 Swenson, R. T., 424, 427, 471, 474 Swift, H. H., 157, 205 Swift. J.. 160, 188, 206 Szybalski. W., 79, 109, 193, 201, 218, 219,

220, 222, 233, 234, 235, 257, 265, 266,

267, 269

Takahashi, T., 363, 364, 402

Takahashi, W. N., 490, 519

Takanami. M., 364, 366, 367, 404

Takats, S., 164, 205

Takemura, S., 227, 269

Tal, M., 303, 351

Tamm, C, 230, 269

Taylor, A. L., 146. 151

Taylor, J. H., 66, 72, 73, 74. 75. 76, 80, 82,

83. 85. 88. 89, 91, 92. 93. 95, 96, 97, 98,

99, 104. 107. 110. 111. 138. 151, 156.

157. 161, 163, 169, 183, 195, 206 Taylor, L. S., 241, 267 Tener, G. M., 24, 62 Tessman, E. S., 119, 151, 244, 269 Tessman, I., 119, 151, 227, 269 Terasima. T.. 159, 192, 194, 206 Thoday, J. M., 91, 111, 242, 269 Thomas, C. A.. Jr., 119, 122, 123, 124.

128, 129. 130. 138, 139, I48, 150, 151

AUTHOR INDEX

537

Thomas, E. D, 459, 467, 4^4

Thompson, L., 12, 63

Thompson, L., 165, 166, 204

Thompson, R. B, 462, 463, 473, 474

Thorup, O. A., 457, 474

Thrasher, J. D., 196, S02

Tijima, T., 225, 266

Till, J. E, 155, 156, 193, 205, 206

Timasheff, S. N, 305, 352

Timofeeff-Ressovsky, M. W., 241, 269

Tomlin. P. A, 222, 267

Tissieres. A., 274, 281, 2SS, 296, 297, 299,

301, 306, 307, 348, 352, 383, 386, 387,

394, 399, 404 Tobita, T., 373, 404 Tocantins, L. M., 438, 439, 441, 450, 451,

470 Tolarico, M., 10, 63 Tolmach, L. J, 159, 192, 193, 194, 203,

206 Toschi, G, 307, 351 Trager, L., 219, 269 Trautner, T. A, 38, 39, 45, 47, 49, 50, 64,

136, 151 Treffers, H. P, 247, 269 Trobert, J. V, 438, 439, 474 Tsevremis, H., 453, 471 T'so, P., 151, 292, 294, 301, 351, 352, 395,

404 Tsujita, A., 284, 285, 288, 484, 485, 486,

487, 491, 494, 499, 500, 501, 504, 509,

512. 513, 518, 519, 520 Tung, J., 93, 95, 111, 195, 206 Tiirnbull, E. P. N, 415, 459, 4^4 Turner, J. F., 17, 62 Tuttle, A. H, 469, 474 Twort, F. W, 477, 520

u

Uhlig, H., 485, 486, 517 Utsumi, K. R.. 157, 206

Valentine, W. N., 449, 474 Vallisneri, E., 418, 454, 474 Van Bibber, M. J., 16, 60, 61 Van Tubergen, R. P., 142, 151 Van Vunakis, H, 119, 120, 150 Vaughan, M, 373, 404 Veldee, S., 490, 492, 509, 518, 520

Vella, F, 442, 454, 474

Vendrely, C, 72, 109, 157, 206

Vencirely, R., 72, 109, 157, 206

Verly, W. G., 156, 201

Vielmetter, W., 225, 227, 251, 269, 498,

499, 513, 520 Vincent, W. S, 276, 287, 289, 336, 337,

352 Vinograd, J., 77, 110, 292, 294, 301, 351,

352, 374, 401, 416, 418, 440, 441, 442,

453, 471, 472, 474 Vischer, E, 272, 288 Visser, H. K. A., 459, 4^2 Vogt, M., 218, 269 Volkin, E., 56, 64> 164, 179, 200, 201, 273,

275, 276, 278, 279, 286, 287, 288, 299,

336, 337, 352, 386, 4OO von der Decken, A, 364, 365, 366, 368,

404 von Ehrenstein, G., 363, 364, 365, 366,

370, 371, 379, 380, 394, 399, 403, 404

w

Wacker, A, 219, 245, 269

Wahba, A. J., 7, 64, 390, 4OO

Wake, R. G., 50, 60

Wakonig, T., 92, 110

Wales, M., 459, 467, 474

Walker, J., 415, 459. 474

Walker, P. M. B., 157, 158, 192, 204, 206

Wallenius, G., 407, 408, 418, 420, 449,

472 Waller, J.-P.. 303, 352 Warner, R., 187, 201 Warner, R. C., 19, 64 Warren, W., 311, 352 Watson, H. C., 276, 285, ^55, 430, 431,

432, 433, 472, 474 Watson, J. D., 26, 38, 46, 49, 64, 72, 73,

111, 179, 186, 202, 204, 215, 246, 269,

274, 281, 287, 296, 298, 301, 307, 315,

326, 329, 344, 351, 352, 383, 384, 386.

402 Watts-Tobin, R. J., 250, 255, 265, 285-

286, 287, 377, 390, 4OI, 435, 436, 470,

511, 512, 517 Weatherall, D. J., 441, 442, 443, 446, 454,

474 Weber, E., 485, 486, 517 Weber, G., 422, 471

538

AUTHOR INDEX

Webster, G. C, 361, 365. 360, 370, 404

Weed, L. L., 10, 01

Weigle, J, 80, 86, 87, 10!), 110, 121, 133,

in. 14.9. 150. 151, 244, 269 W.mII, .1., 187, 201, 204 Weill. J. D., 280, 2SS AV.Mnhlum, D.. 245, 269 WcmlVM, H., 3. 4, 63 Weislilum, B., 361, 4OO Wois.s, S. B., 186, 202. 280, 2S7, 2S9, 347,

348. 352. 382, 397, 404 Weis.sman. J. M., 13. 17. 64 Weissman, S. M., 02, 171. 172, 173-174.

202, 206 Welch, A. D., 218, 219, 266 Weliky. N., 418, 421, 473 WelLs, I. C, 407, 422, 457, 473 Went, L. N., 462, 463, 473, 474 Wertheimer, S. A., 76, 109, 142, I49, 161,

183, 201 Westergaard, M., 234, 235, 240, 252, 266,

267, 269 Weygand, F., 219, 269 Wheby, M., 457, 474 Wheeier, J. T., 445, 463, 464, 474 White, F. H., Jr., 482, 520 White, J. C, 416, 459, 461, 474 Whitfeld, P., 489, 520 Whitfield, J. F., 192, 206 Whitmore, G. F., 193, 206 Wieder, C. M, 225, 251, 269 Wilber, P. B., 12, 63 Wildman, S. G., 490, 519, 520 Wildy, P., 157, 158-159, 193, 195, 204, 205 Wilkins, M. H. F., 114, 125, 126, 140, 150,

151, 215, 260, 308, 352 Will, G., 409, 430, 473 Williams, R. C., 117, 121, 149. 487, 489.

518 Wilhams, W. J., 416, 442, 443, 454, 473 Williamson, D. H., 159, 206 Williamson, M. B., 172, 206 Wilson, H. R.. 125. 149. 150 Wilson, S., 409, 413, 475 Wimber, D. R., 191. 205 Wintrobe, M. M., 466, 475 Witkin, E. M., 187, 206, 225, 244, 245,

269 Wittman-Liebold, B., 413, 414, 415, 416,

417, 470

\\ iiiiuanii. H. G., 480, 482, 484, 486, 493, 494, 501, 502, 507, 512, 518, 520

Witz, J., 305, 352

Woese, C. R., 511, 520

Wolf, J., 418, 445, 449, 472

Wolf. M. K., 225, 265

Wolff, S., 91, 93, 111, 242, 269

Wollman, E. L., 145, 146, I49

Woodard, J., 160, 188, 206

Wood, W., 388, 394, 404

Wood, W. B., 283, 284, 289

Woods. P. S.. 66, 73, 74, 75, 76, 104, /;/, 138, 151. 156. 161, 183, 206

Woods, R. D., 479, 519

Work, T., 373, 4OO

Wyatt, G. R., 10, 46, 56, 64

WVss, 0., 244, 269

Xefteri, E, 453, 471

Yamamoto, N., 225, 269

Yamazaki, H., 284, 289

Yanofsky, C, 215, 266. 285, 287, 428, 471

Yarmolinsky, M. B., 379, 404

Yates, H. B., 157, 206

Yeas, M., 276, 287, 289, 336, 337, 352, 511,

520 Yehuda, M., 461, 470 Yoshida, A, 373, 404 Youcis, P, 156, 201 Young, I. E., 155, 158, 206 Young, J, 485, 486, 520

Zarhau. H., 357, 404 Zalin. R. K.. 142, 150 Zalokar, M., 396, 404 Zamecnik, P. C., 272, 289, 299, 351, 354,

355, 357, 358, 359. 363, 364, 365, 369,

386, 402, 404 Zamenhof, S., 219, 220, 231, 233, 237, 238,

266, 268, 269 Zech, H.. 490, 520 Zetterberg, G., 235, 269 Zeuthen, E., 164, 192, 202, 206 Zicliichi. M. L., 86, 87, 109, 121, 141, 149,

150 Zillig, W., 357, 404, 480, 482, 519

AUTHOR INDEX 539

Zimmer, K. G., 241, 269 Zobel, R, 126, 1^8

Zimmerman, S. B., 13, 14, 16, 46, 48, 50, Zogerman, J., 20^

55, 56, 57, 58, 59, 61, 62, 63, 64, 162, Zubay, G., 296, 307, 308, 351, 352

203, 253, 267, 273, 2S7 Zuelzer, W. W., 422, 450, 473, 475

Ziporin, Z. Z., 179, 202

SUBJECT INDEX

Acridine dyos as mutagens, 224-225 Adenosine triphosphale-generating sj's-

tem, 14 Adenjiate kinase, 12-13 Alkylating agents, 231-237 Amino acyl-RNA, 355-356 2-Aminoi)urine, 222-225 Analogs, base, 46-50

substitution in DNA synthesis, 46-50, 219-227 Autoradiography, 73-77, 122-128

molecular, 123-124, 127

B

Bacteria,

cytology of, 142-147 Bacteriophage, 113-141

distribution of parental DNA to prog- eny, 138-140

0X174, 119-122

ghosts, 115

HETS, 136

maturation, 117-119

organization of DNA, 114-117

recombination in T2 and T4, 132-133

structure of heterozygotes, 136-138

transfer of parental DNA, 139-141

T-2, 114-119, 123-125

turnover of RNA in cell, 278-279 Base analogs (see analogs) Bromodeoxyuridine,

substitution for thymidine in healing chromosomal lesions, 94 5-Bromouracil, 48-50, 220-222

incorporation in copolymer, 49

Carbon-14 as label for DNA, 18, 73 Chloramphenicol, effects on ribosome

synthesis, 333 Chloromercuribenzoate, 55 Chromatid, 67-68

sister exchangers, 82-85 Chromatography, 315-316

separation of hemoglol)ins, 406-407

Chromonema, 68

Chromosome re])lication and jjrotein

synthesis, 89-90 Chromosomes, 65-108 axial element, 72 du])lication, 73-77 exchanges, 82-86 giant, 68-72 lampbrush, 69-72, 107 labeling witli tritium, 73-77 mechanism of healing. 93-90 models of, 96-108 replication of, 72-82 reproduction of, 89-92 reunion of breaks, 92-96 sequences in duplication, 88-89 structure, 67-72

of bacterial, 143-144 types of changes produced by radia- tion, 90-92 Cistron, 214

Coding, genetic, 282-286, 390-395 evidence for triplet, 284-285 from mutant hemoglobins, 434-435 from TMV, 509-517 Colchicine and chromatid exchanges, 83 Complementation, 214 Copolymers, 42

Deamination by nitrous acid, 225-227

Deoxyriboaldolase, 5

Deoxycytidine, 3

Deoxycytidine triphosphatase, 50

Deoxycy tidy late hydroxymethylase, 11-

12 Deoxyribonucleic acid (DNA), base alterations, 216 breakage b3' shear forces, 122, 130-131 change in state during replication, 184-

185 changes in amount over the cell cj^le,

154-158 control mechanisms and nucleotide

pools, 159-160, 162-165 d-AT polymer, 40-45

540

SUBJECT INDEX

541

d-GC polj-mer, 45-46 density-gradient centrifugation of, 77-

79 "melting temperature," 231 molecular weights of bacterial DNA,

145 nearest neighbor analysis, 32-40 number of chains in molecule, 81, 125-

128, 144 organization in phages T2 and T4,

114-117, 120 organization into chromosomes, 96-108,

113-148 phage 0X174, 121-122 precursors of, see Deoxyribonucleo-

tides primer, 25-27, 161, 180 polymerase, see Polj'merase replication, 77-82 semiconservative, 74-80, 161 sequence of, 188-190 single chain regions or weak points,

119, 128 synthesis, 17-60

base composition of product, 29-32 control mechanisms, 160-200 dissociation from cell division, 197 effects of injury, 191

of nutrient environment, 194-196 of protein synthesis, 173-180 of radiation, 192-194 of temperature, 192-194 limited reaction, 23-24 nature of primer, 25-27 of product, 27-29 nearest neighbor analysis of prod- uct, 32-40 physical state of DNA in control- ling, 180-183 relation to the cell cycle, 155, 175-

180, 190-192, 197 replacement of natural bases bj' ana- logs, 48 requirements for cell free system, 20 simultaneous RNA synthesis, 186-

187 site in the cell, 190 structural stabilization by protein,

188 variations in the cvcle, 191-197

transfer and recombination in lambda

phage, 141 transfer experiments with BUDR, 140 with E. coli, 142 Deoxyribonuclease, 28

effects on chromosomes, 69 Deoxyribonucleoside,

monophosphate kinase, 12, 16 Deoxyribonucleotide pools, in anthers of Lilium and Trillium anthers, 163- 164 in bacteria, 166-169 Deoxyribonucleotides, 1-17 phosphor3'lation of, 12-17 synthesis of, 3-12, 163-172 Deoxyuridine triphosphatase, 50 Deoxyuridylate, synthesis of thy midy late

from, 6-8 Depolymerization of DNA, 238 Depurination of DNA, 230, 237-240 Diesterase, splenic, 34 Dyes, mutagenic effects of, 224, 225

Electrophoresis, 408

Enzyme induction, 277, 343-344

Enzymes,

amino acid activating, 354-356

phosphorylating in control of DNA synthesis, 171

ribosomal, 303-305 Episomes, bacterial, 147 Exchanges,

chromatid, induced by radiation, 90- 92

half-chromatid, 92

single and twin in chromosomes, 83

sister-chromatid (see Chromatid, ex- changes)

Fingerprints of hemoglobin, 410-412 5-Fluorodeoxyuridylate, 9, 93-94 Folic acid, &-12 Formaldehyde as a mutagen, 240

j8-Galactosidase, gene for, 281 induction of, 277, 281

542

SUBJECT INDEX

Gene for beta-galactosidase, 281 Gene mutations, 212-214 Glueosyl transferases, 57-58 Glucosj'lation of 5-hydroxymetliylcyto-

sine, 56-59 Gradient centrifugation,

cell components, 273-274, 316-321

DNA, 77-80

H

Half-chromatids, 67, 99 Hcmoglobin-A2, 418-419 Hemoglobin-A3, 420-421 Hemoglobin-F, 415, 418, 421 Hemoglobins, 405-469 abnormal, 421-434 amino acid substitution in variants,

42&-434 control mechanism in sj'nthesis, 456-

461 dissociation and recombination, 440-

442 genetics of, 436-438, 444-447, 462-465 mode of assembly of molecule, 439-440 sequence of amino acids in the chains,

419-420 structure of, 409-415 Homopolymers, 54

of DNA, 54 Hj'drazine, reaction with DNA, 227 Hydrogen bonds in DNA, 81 Hydrogen ion concentration (pH),

mutagenic effect, 230 Hydroxylaminc, mutagenic effect of, 227- 230 reaction with bases of DNA, 227-228 with bromouracil, 229 with cytosine, 228-229 5-Hydroxymethyldeoxycytidylic acifl, 10- 12 synthesis of, 10-12 Hypochromicity as a test for DNA syn- thesis, 19

K

Kinase, deoxycytidylate, 15 Kinases, 12-17

dGMP, 16

deoxj'nucleoside diphosphate, 15-17

thymidylale, in bacteria and HeLa cells, 173-174 in liver, 171-172

Linkers, chromosome, 98, 108

M

Maps, genetic, E. coli, 146 mutations, 248-249 T-even phages, 131-135, 212 in rll region, 213 Messenger RNA, see RNA Microsomes, 292 Mitotic cycle, 155-156

measurement of length, 156-157 Mitosis, synchronization, 181 Mutagenic specificit}^ and the reaction

of certain bases, 254 Mutagens, 217-218 specificity of, 252 Mutations, 207-264 classification of, 20&-209 forward, specificity of, 252 induction, in TMV, 497-499

kinetics of, 250 phages for stud}' of, 211-215 reverse, 249

specificity of, 253 spontaneous, 245-248, 263-264 suppressor, 254

N

Nitrogen-15 and DNA labeling, 78-79 Nitrogen mustard, see Alkylating agents Nitrous acid, mutagenic effect of, 225-227 Nucleosides in relation to mitotic cycle, 159

Peptide bond

formation of, 374-378 Peptide chain

formation of, 373-374 Phage, see Bacteriophage Phosphodiesterase, 23-24, 54 Phosphorus-32 as label for DNA, 18

SUBJECT INDEX

543

Phosphorylase, polynucleotide, 280 Phosphorylation of deoxATibonucleotides,

12-16 Photod\'namic effect of dyes, 225 Point mutation, 214 Polymerase, DXA,

bacteriophage-specific, 54-56

calf thymus, 50-54

E. coli. 17-21. 40-45 Polymerase, RXA, 280-281 PohTiucleotides, synthetic, 280

use in coding studies, 282-285 Pools, metabolic, 311-314 Primer as a factor in controlling DXA

replication, 181-183 Proflavin, mutagenic effect of, 224, 239 Protein s^mthesis, 353-400

final stages of, 378-381

kinetics of incorporation of label, 331- 333

requirement for DXA replication, 178- 179

sequence of bond formation, 374-378 P^Tophosphorolysis, 21

Radiation, indirect effects in mutation

studies. 242 Radiations, ionizing, mutagenic effect of, 240-243 ultraviolet light, mutagenic effect of. 243-245 Recombination, genetic. 82-88

mechanism of. 87, 96. 107, 132, 137 X phage DXA, 86-87 Replication, see DNA replication Revertants, 254 Ribonucleic acid (RXA) bios^-nthesis of, 272-286, 326-331. 335-

341 DX\A.-like, 273

RXA, homologous to DXA, 275 messenger, 275-276, 344, 382-386

isolation of, 279 soluble (see also transfer RXA), 368-

369 sj-nthesis of, 271-286

cell free s^^stems, 397-400 DXA-directed, 281, 386-390, 39&-397

relation to DXA SA-nthesis, 186-188 T2-specific, 274 turnover of, 275-276

upon phage infection, 278 Ribosomes, 292-350 assembly of, 333 effects of chloramphenicol, 298

of magnesium on, 293-296 fractionation of, 308-310 nucleic acid components of, 305-306 protein-RX'A ratio, 300-302 protein components, 302-303 purification of, 299-302 site of protein s^Tithesis, 299 stnicture of, 295, 306-307 s>-nthesis of, 321-350, 395-396

site of, 346, 363-367 Ribotides, conversion to deoxjTibotides,

169-170

Sedimentation analysis, see Gradient

centrifugation Serine as methyl donor, 7 Sulfur mustard, see Alky'lating agents

Target theor>-, 241 Tetrahydrofolic acid, 7-9 Thalassemia. 450-453

genetics of, 450 Thymidine-H", 74-76 Thymidine kinase, 13 Thymidylate kinase, 13 Thymidylic acid, 6-10

SA-nthesis of, 6-10 Tobacco mosaic virus (TMV) 477-517

t:mv,

amino acid composition, 493-495 infectivity. 489-491 protein, 484-489

amino acid sequence. 486 protein composition of mutants, 499-

508 ribonucleic acid. 487-489 structure of. 479-483 SATnptomatology. 491-493 virulence and viability, 508-509

544

SUBJECT INDEX

Tran.s(l(M)xyril)0.si(iasp. 4-5

Transfer RNA. 272. 356-362 structure of, 351)-360

Transformation, genetic, 133-134 with lambda-dg DXA, 133-134 with phage DNA broken by shear, 134-135

Transitions, 217, 256-257, 260-263

Transvorsions, 217

Triethylenemelamine as a mutagen, 233

u

I'll ra violet light, effects on pyrimidinos, 245 mutagenic effect of, 243

\'ariations, heritable, see Mutations Viscosity measurements and tests for DNA synthesis, 18-19

iu., i LiuuARY )"f