Full text of "Biochemical systematics"

BIOCHEMICAL
SYSTEMATICS

PRENTICE-HALL BIOLOGICAL SCIENCE SERIES
William D. McElroy and Carl P. Swanson, Editors

BIOCHEMICAL SYSTEMATICS,* by Ralph E. Alston and B. L. Turner
CLASSIC PAPERS IN GENETICS, by James A. Peters
EXPERIMENTAL BIOLOGY, by Richard W. Van Norman
MECHANISMS OF BODY FUNCTIONS, by Dexter M. Easton
MILESTONES IN MICROBIOLOGY, by Thomas D. Brock
PRINCIPLES OF BIOLOGY, by Neal D. Buffaloe
SELECTED BOTANICAL PAPERS, by Irving W. Knobloch
A SYNTHESIS OF EVOLUTIONARY THEORY, by Herbert H. Ross

Concepts of Modern Biology Series

BEHAVIORAL ASPECTS OF ECOLOGY,* by Peter H. Klopfer

Foundations of Modern Biology Series

ADAPTATION, by Bruce Wallace and A. M. Srb

ANIMAL BEHAVIOR, by Vincent Dethier and Eliot Stellar

ANIMAL DIVERSITY, by Earl D. Hanson

ANIMAL GROWTH AND DEVELOPMENT, by Maurice Sussman

ANIMAL PHYSIOLOGY, by Knut Schmidt-Nielsen

THE CELL, by Carl P. Swanson

CELLULAR PHYSIOLOGY AND BIOCHEMISTRY, by William D. McEhoy

HEREDITY, by David M. Bonner

THE LIFE OF THE GREEN PLANT, by Arthur W. Galston

MAN IN NATURE, by Marston Bates

THE PLANT KINGDOM, by Harold C. Bold

* These titles are also in the Prentice- Hall International Series in Biological Science.
Prentice-Hall, Inc.; Prentice-Hall International, United Kingdom and Eire; Prentice-Hall
of Canada, Ltd., Canada; BerUner Union, West Germany and Austria.

Current Revisions for BIOCHEmCAL SYSTEMATICS
p. 110 Lath5rrine heterocyclic ring is aromatic.
p. 158 Anhalonidin N-containing ring is saturated,
p. 159 Berberine cation should appear as follows:

p. 161 .\lstonine skeleton; hydrolysis product of physostigmine; for rauwolfine
substitute established ajmaline structure;

N/ -^^^^

CHo

p. 162, p. 257 Gentianin should appear as follows:

^^C2H5

p. 163 Aconitine contains a C^ methoxyl group.

p. 176 Ceveratrum and jerveratrum skeletons contain OH at Ci-,- CHo °touds at
^10- *^13' ^20 andC25

p. 197 Rotenone C2-C3 bond is saturated,
p. 200 Apigenin lacks 3 ' -OH.

p. 216 Structure 2 has 3-OH. 5-OCH3 substituents; pinostrobin has 7-OCH.
structure 5 should read flavanonols. "*'

p. 217 Aromadendrin has C5 and C, hydroxy] and C4 keto groups; 4- substituent
in conidendrm should be OH.

p. 226 Embelin ring is unsaturated.

p. 249 Should read hecogenin skeleton.

p. 257 Cardenolid 5-membered rings are separated by a bond. Gentiopicrin
structure has now been proposed: ^„

Loganin structure has now been proposed:

p. 265 a-Santonin should appear as dienone.

p. 277 Betanidin carboxyl groups are unmethylated.

BIOCHEMICAL
SYSTEMATICS

Ralph E. Alston

Associate Professor of Botany
University of Texas

B. L. Turner

Professor of Botany
Director of the Herbarium
University of Texas

Prentice-Hall, Inc.

Englewood CUffs, N.J.

Dedicated to the memory of

Professor Donald Walton Davis

late Professor of Biology at the College of
William and Mary. His devotion to aca-
demic and scientific principles and his human
qualities represented a continuous inspiration

(R. E. ALSTON).

PRENTICE-HALL INTERNATIONAL, INC., London
PRENTICE-HALL OF AUSTRALIA, PTY., LTD., Sydney
PRENTICE-HALL OF CANADA, LTD., Toronto
PRENTICE-HALL FRANCE, S.A.R.L., Paris
PRENTICE-HALL OF JAPAN, INC., TokyO
PRENTICE-HALL DE MEXICO, S.A., Mexico City

All rights reserved. No part of this book may be reproduced in any form, by mim-
eograph or any other means, without permission in writing from the publisher.

Library of Congress Catalog Card Number: 63-14907

Printed in the United States of America
C

PREFACE

Although hundreds of thousands of words have
already been written about biochemical systematics
its actual impact upon formal systematics is still
trivial. So far, no significant taxonomic dispositions
of higher plants rest primarily upon biochemical
criteria. We consider that an important objective of
this book is to develop, more fully an appreciation
of the diversity of applications of biochemistry to
systematics.

The present treatment is oriented towards
botanical systematics. Many of the readers of this
book will be: (1) plant taxonomists with only shght
background in biochemistry and (2) chemists with
Httle background in classical plant systematics,
possibly unacquainted with certain concepts on

vi PREFACE

which the field is founded and with Hmited knowledge of modem
work in systematics. Chapters II through IV are written primarily
for the nonsystematist. We ask the indulgence of the well-informed
if this introductory matter reiterates much that has already been
written on the subject.

At the present stage of development, plant biochemical sys-
tematics is a difficult field to survey. It will be noted that nowhere in
the book is there a phylogenetic tree constructed out of chemical
correlations. Perhaps contrary to the expectation of some readers, we
do not see that even the beginnings of such a system are justified. Thus
the decision to organize the chapters about major groups of chemical
constituents rather than to focus upon taxonomic systems of catego-
ries is based upon our firm belief that it is more useful to consider
various "natural" chemical groups somewhat critically relevant to
their present and potential systematic value than to draw a series of
taxonomic judgments out of the usually fragmentary biochemical
data at hand. The latter approach has been used, at least eclectically,
by others, to no great advantage.

The writers cannot regard present limited biochemical data
as favoring one or another of the systems such as those of Engler and
Prantl, Bessey, Hutchinson, etc. Much of the literature in biochemi-
cal systematics includes references by the authors to competing sys-
tems when the data bear upon the systematic relationships of higher
categories, but in general the individual issue concerns only a small
part of the taxonomic whole, and the chemical data now available are
often quite limited.

Some readers may be puzzled by the fact that we speak else-
where of taxonomists who have no interest in phylogeny. The non-
taxonomist may be least capable of understanding this situation.
Nevertheless, professional taxonomists exist who favor the exclusion
of phylogeny from taxonomy. Similarly, although authoritative docu-
mentation from the literature is not available we have heard promi-
nent biologists express the belief that biochemistry could never make
a contribution to systematics since, e.g., nicotine and certain other
substances occur in obviously unrelated plant groups. Such argu-
ments as the latter may be transparent, but they are not fictitious,
and therefore some attention is given to answering them in the text.

We believe that the intellectual, technical and perhaps even
psychological gap (not intended to be construed as hierarchical in
nature) between systematics and chemistry has been the main factor
in delaying the maturity of biochemical systematics as a natural
discipline. Biochemical systematic studies of the present are often not
markedly different from those of 30 years ago. Modern statements

PREFACE VII

(Constance, 1955; Gibbs, 1958) are hardly distinguishable from those
of a generation ago (Redfield, 1936) or nearly a century ago (Abbott,
1886).

Classical cytogenetic methods, which offer far less, poten-
tially, than does comparative biochemistry in over-all application to
plant systematics, were quickly assimilated into the discipline, and as
a result some of the highest intellectual achievements are represented
by classical cytogenetical investigations (e.g. Cleland, 1949, 1954;
Clausen, 1953). Therefore, the conspicuous retardation of real prog-
ress in the development of sound principles of biochemical system-
atics is considered to reflect, in part, the wide technical and intellec-
tual separation of taxonomy and chemistry. Partly because of the
emergence of new research tools, and partly because a relentless
and natural trend toward molecular biology will otherwise turn the
field of biochemical systematics over to biochemists by default, the
writers believe that a reappraisal of biochemical systematics and the
development of a strongly positive attitude toward the field by tax-
onomists is desirable.

In our judgment the chief weakness of biochemical system-
atics has been and remains the threat of superficiality. If the present
book serves merely to foster a host of superficial shotgun chromato-
graphic comparisons miscellaneous irresponsible correlations and
naive interpretations, we will have failed completely in our purpose.
We hope that it will encourage an approach to biochemical system-
atics which is reflective and cautiously optimistic.

The book is offered with humility in recognition of our indi-
vidual and collective limitations. We have tried to avoid both pedantry
and oversimplification. In numerous instances we have taken the
liberty of professing a personal evaluation or criticism, always with
the objective of establishing a better perspective for viewing bio-
chemistry in its relation to systematics.

To our knowledge, there is no precedent for this book. Conse-
quently, it is based almost entirely upon research contributions from
technical journals. Because of the breadth of subject matter encom-
passed it is virtually impossible to cover the literature completely,
and it is likely that some work of major significance was not detected.
The words of Sir Francis Galton* come to mind:

I trust the reader will pardon a small percentage of error and inaccu-
racy, if it be so small as not to affect the general value of my results.
No one can hate inaccuracy more than myself, or can have a higher

* Galton, Francis. 1869. Hereditary Genius. 1st Edition. Reprinted, 1952. Horizon
Press, New York.

viii PREFACE

idea of what an author owes to his readers, in respect to precision; but
in a subject Hke this, it is exceedingly difficult to correct every mis-
take, and still more to avoid omissions.

Perhaps most importantly we regard this book as an effort to
consider perspectives in biochemical systematics. In this sense it is
written for the future. An encyclopedic compendium of biochemical
data organized in a taxonomic framework is badly needed. However,
the writers see no relationship between such a work and our present
endeavor.

Ralph E. Alston
B. L. Turner

TABLE OF CONTENTS

INTRODUCTION

■mm

TAXONOMIC PRINCIPLES

The categories 7

PhYLOGENETIC concepts and TAXONOMIC SYSTEMS l4

Systems of classification 16
Two-dimensional phylogenetic diagrams 18
Three-dimensional phylogenetic diagrams 23
Classification of vascular plants 23
Parallelism as a factor in classification 27
The fallacy of the "fundamental" character 31

PLANT TAXONOMY

TABLE OF CONTENTS

A BRIEF HISTORY OF MAJOR DEVELOPMENTS IN THE
FIELD 37

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS 41

Some preliminary considerations of the applica-
tions OF biochemistry to SYSTEMATICS 48

SEROLOGY AND SYSTEMATICS 67

AMINO ACIDS

Systematic studies involving amino acids 96

FATTY ACIDS

Fatty acid biosynthesis 123

CARBOHYDRATES

119

135

Simple sugars (a partial list only) 137

Sugar alcohols (acyclic polyhydric alcohols) 139

Inositol and related cyclic alcohols l4l

Oligosaccharides l43

Polysaccharides 150

TABLE OF CONTENTS xi

ALKALOIDS 1 55

Some major classes of alkaloids 157
Some general considerations of alkaloid distri-
bution AND PHYSIOLOGY 164

General considerations of the systematic value
of alkaloids 168

Specific examples of alkaloids of systematic
significance 170

CYANOGENETIC SUBSTANCES 181

PHENOLIC SUBSTANCES 191

Some basic considerations of biosynthetic path-
ways INVOLVED IN THE PRODUCTION OF PHENOLICS 194

Chemical structures of classes of flavonoid

compounds 198

Genetic studies concerning the flavonoid

compounds 202

Systematic aspects of the distribution of

phenolic compounds 209

QUINONES 223

TERPENOIDS 231

xii TABLE OF CONTENTS

MISCELLANEOUS COMPOUNDS 269

BIOCHEMICAL STUDIES OF HYBRIDS 295

Inheritance of oil characters in hybrids of
Eucalyptus macarthuri X £• cinera 311

GENERAL EVALUATION 327

Physiological or chemical races 332
Variation in the course of development and
within the mature plant 335
Methods of presenting comparative biochemical

DATA for systematic PURPOSES 336

Evaluation of specific biochemical data 342

APPENDIX I 345

bibliography

INDEX

347

387

INTRODUCTION

The great advances in biochemistry which have
come in a few decades have impressed both the in-
formed layman and scientist. The scientist who has
made an effort to acquire more than a passing
acquaintance with the subject is appreciative of not
only the elegance of method and the intellectual
challenge of the field but in addition the implica-
tions, sometimes of even a philosophical nature, of
these discoveries to other subdivisions of biology.
For instance, the biochemical unity disclosed
incidentally along with the elucidation of basic path-
ways of metabolism is as effective support for Dar-
winian evolution as is comparative anatomy. With-
out a fossil record, and assuming that evidence from
comparative anatomy were in some way unavailable,

2 BIOCHEMICAL SYSTEMATICS

comparative biochemistry would have already established unequivo-
cally the same concepts of evolution which now exist.

Four levels of biochemical unity may be recognized which,
collectively, provide a framework for evolutionary theory. Starting
with the most fundamental they are: (1) biochemical unity as ex-
pressed in the basic similarity of the hereditary material of all organ-
isms; (2) biochemical unity as expressed in the group of co-enzymes
which are essential to many of the basic biochemical processes; (3)
biochemical unity as expressed in the similarity of metabolic path-
ways, particularly those involved in energy exchange, of different
organisms; and (4) biochemical unity as expressed within major
taxonomic groups in the common presence of certain structural com-
ponents such as chitin, cellulose, and so on. At all of the levels there
is also some degree of diversity. For example, while deoxyribonucleic
acid is present in the chromosomes of diverse species, the same
sequence of nucleotide subunits is unlikely to be expressed even in
two individuals of a single species. All of this knowledge has a direct
bearing upon phylogeny in its broadest meaning. At least, all of the
facts have potential phylogenetic significance; those which emphasize
unity, to relate species, and those which emphasize diversity, to
separate species.

In recent years a number of books have been written about
various aspects of the broad subject of biochemistry in relation to
evolution. The Molecular Basis of Evolution by Anfinsen (1959), and
the six volume work in preparation edited by Florkin and Mason
(1960) are especially noteworthy. There are also numerous individual
articles on the subject of biochemical evolution, treating various as-
pects of the subject. Speculation upon the origin of life itself is now
centered almost entirely upon questions relating to molecular
evolution (Oparin, 1959).

Dating back many years before the beginnings of enzyme
chemistry and studies of metabolic pathways are numerous investiga-
tions of the distributions of various substances, initially in higher
plants and now including fungi and bacteria as well. Such investiga-
tions often had pharmacological and other economic objectives, but
some of the earliest workers were interested in correlations between
the distributions of substances and the taxonomic treatments of the
species investigated. Subsequent workers have continued to note
such correlations or even to make a tentative taxonomic judgment
based on their chemical results. Periodically, belief in the utility of
biochemical data for systematic purposes has been reiterated. Bio-
chemistry has not yet been responsible for any major advances in our
knowledge of phylogenetic relationships. Yet, inexorable progress in

INTRODUCTION 3

the accumulation of biochemical data, many of which are already seen
to be of phylogenetic importance, points to an obligatory integration
of these data in systematics. The systematist does not have the pre-
rogative of evaluating the purely chemical aspects of data, but he has
a responsibility to be alert to progress in biochemistry, particularly
when discoveries bear potentially upon phylogenetic considerations.
Biochemistry relates to phylogeny at several levels, only one of which
involves the taxonomic distribution of specific compounds. Certain
approaches discussed in Chapter 4 may seem to be remote or even
irrelevant, but the writers believe that no approach should be dis-
couraged provided it is theoretically sound though its practical value
may eventually prove to be slight.

It is not the purpose of this book to develop a case for
the use of biochemical data in systematics but rather to establish a
better perspective concerning the place of biochemistry in systematics.
There is a need for an exploration of some theoretical and intellectual
aspects of the subject, the development of a basic rationale, an inte-
gration of certain chemical and biological aspects, an analysis of the
advantages and hmitations of the biochemical approach, a broad and
essentially critical analysis of existing work. We have attempted to
accomplish this series of objectives.

We do not beheve that biochemistry represents a panacea for
all systematic problems. If anything, the writing of this book has
modulated our initial enthusiasm which even in the beginning did not
lead us to conceive of present biochemical data as providing more
than supplementary data for phylogenetic considerations. However,
profound and far-reaching new insight into phylogenetic relationships
is potentially available in biochemistry, ultimately, we predict, from
intensive study of the comparative chemistry of macromolecules.

Nowadays, much is spoken and written about what is pop-
ularly known as molecular biology and its relationship to descriptive
or classical biology. It is possible that some individuals regard these
two categories as mutually exclusive. It is true that in this age one per-
son rarely acquires eminence in both areas. However, there are many
who can excel in performance in one area and be intellectually in con-
tact with the other. It is the purpose of this book to contribute to an
integration of these disciplines by providing the groundwork for a more
effective utilization of biochemical data in systematics than has
previously existed.

TAXONOMIC
PRINCIPLES

Taxonomy is one of the oldest fields of biological
science. Organisms, and their relationships to other
organisms, have occupied man's thinking for hun-
dreds, if not thousands, of years. In order to classify,
even at the most elementary level, man had to
recognize (or identify) organisms. To do this he was
prone to observe, make comparisons, and to some
extent, integrate data, and develop generalizations
therefrom. It can be argued that taxonomy was
almost synonymous with biology in its beginning as
a science. The identity of organisms occupied the
thinking of early biologists. To derive order out of
the multitude of forms in existence, these biologists
were primarily concerned with writing descriptions
and giving names.

^ BIOCHEMICAL SYSTEMATICS

Many non-taxonomists, including biologists and other scien-
tists, believe that the sole function of the taxonomist is to describe
and name species. While this is still an important function of taxon-
omy, it is not its beginning or end. Taxonomy, like other areas of
biology, has kept pace with the mainstream of biological progress.

A well-trained worker in taxonomy today must have a broad
background in the fundamental concepts and basic working techniques
of a number of disciplines. He not only has to be familiar with the
special disciplines of his own field, but also should have some famiharity
with cytology, genetics, statistics, anatomy, and, it is hoped, bio-
chemistry. Without such breadth the worker is often confined to a
rather narrow avenue with much diminished perspective. If he is to
synthesize and integrate the data provided by classical methods and
augment this knowledge with new kinds of evidence he must be, as he
was in the beginning of the natural sciences, one of the better in-
formed and widest-read of all biologists.

Taxonomic thought, as indicated in more detail below, changed
radically with the advent of Darwinism. Taxonomists not only have
incorporated various new morphological approaches (for example,
embryology and palynology), but also have accepted enthusiastically
the contributions from genetics and cytology. In the present text we
are attempting to inform the interested taxonomic worker of some
present trends and developments in biological thinking which are or
may become relevant to taxonomy.

Certain biologists attempt to discredit taxonomy as a "clas-
sical" or dead field. This is unfortunate since taxonomy offers a con-
ceptual approach to biology at the organismal level such as chemistry
offers at the molecular level. Both taxonomy and chemistry are uni-
fying fields. The former, based on evolutionary principles, provides a
framework to account for morphological variation and its mecha-
nisms at the organismal and populational level, while classical and
theoretical chemistry provide a systematic framework to describe and
in part comprehend variations in the organization of elementary
particles.

While the term taxonomy has long been used to cover sys-
tematic work in the inclusive sense, more recently a number of new
approaches has occasioned the advent of new names, such as sys-
tematics,! biosystematics, neosystematics, and so on. Regardless of

1 Simpson (1961) defines systematics as "the scientific study of the kinds and
diversity of organisms and of any and all relationships among them," while taxonomy is
defined as "the theoretical study of classification, including its bases, principles, procedures,
and rules." In the present text we have used the terms interchangeably and in the inclusive
sense.

TAXONOMIC PRINCIPLES 7

their appellation, all such workers are, in fact, taxonomists; perhaps
a bit more modern by employing experimental procedures but other-
wise attempting to solve the same problems, namely, to show
relationships and to classify accordingly.

Constance (1960) in reviewing the book of Takhtajian (1959)
was impressed enough with certain statements made by this author
to quote in his review the following section:

Among many biologists of experimental aim the notion is widespread
that Systematics is a branch of knowledge that is absolutely out-
moded. This conception of Systematics is profoundly false and the
result of a certain narrow-mindedness of thought associated with one-
sided specialization. . . . The fundamental general-biological signif-
icance of Systematics consists in that millions of facts that have no
sort of scientific value in themselves find their place in the construc-
tion of Systematics. Systematics is consequently not only the basis of
biology, but also its coronation.

Placed in its proper perspective then, taxonomy becomes the
framework or the ordered arrangement of innumerable observations
and bits of information. This order is as useful for biochemical data
as it is for morphological features. Indeed, it would seem almost indis-
pensable for the former since the seemingly unlimited number of
molecular configurations might lose much of their interpretative
significance without such a foundation.

Taxonomists generally fall into one of two sorts: (1) those
who are primarily interested in the biological units, particularly with
respect to their identification, distribution, and proper description,
£md (2) those who are less concerned with the names and descriptions
of categories and more concerned with evolutionary histories and the
mechanisms of speciation. In taxonomy, as in most other fields, there
are specialists, some who are involved with floristic work, some with
identification, some with phylogeny, and some with evolutionary
mechanisms. There is room for all, in spite of the fact that different
approaches might seem to be more significant at different periods of
time. Ultimately all of the information must be consohdated into any
unified system of classification.

The categories

formal categories

There has been much misunderstanding about the nature of
biological categories. Such terms as species, genus, tribe, family,

BIOCHEMICAL SYSTEMATICS

"Lord, what a day!"

Fig. 2-1. From the systematic point of view, the original caption
might have read, with equal humor, "You mean they're not all
dogs?" (Drawing by George Price, 1954, The New Yorker Maga-
zine, Inc.)

order, and division have no specific meaning to most non-biologists
and frequently disputed meaning among biologists. The categories
may be regarded as highly arbitrary. Any attempt by man to
categorize natural variation must be arbitrary with respect to a
terminological system. This does not mean that the natural entities
which are being classified are, in themselves, arbitrary or subjective.
If Darwin's theory of evolution is accepted as the general mechanism
for the origin of extant taxa, it necessarily follows that the hierarchy
of formal categories erected by man do stand in certain positions
relative to each other.

It is often argued that the biological categories, in that they
are classified by man, are completely subjective in nature. What is
often overlooked here is that the subjectiveness is in applying the
terminology; the objectiveness of the category under consideration,
from a biological point of view, is real. If the biological entity were
completely subjective, then, to use a far-fetched analogy, one might
well expect the dog-catcher to bring into the pound occasionally lions,
orang-outangs, pelicans, and on rare occasions, snakes (cf. Fig. 2-1).

TAXONOMIC PRINCIPLES 9

Fortunately, however, the dog-catcher is not concerned with semantic
problems, and, though not trained in taxonomy, he finds no difficulty
in recognizing Canis familiaris despite its modern polymorphism.

The professional biological classifier has been said to arrive at
his classification through a process popularly known as the taxonomic
method. Several attempts have been made to define or otherwise ex-
plain the taxonomic method, but most definitions or descriptive
attempts fall short of their mark. While most taxonomists have a
fairly good idea what is meant by this method, they find it difficult to
express. Essentially, it can be defined as an attempt to make taxo-
nomic interpretations using pattern data from any source. Rogers and
Tanimoto (1960) among many others have clearly recognized the in-
herent complexities of this multiple correlate method and hence have
suggested the use of computer programs,^ using punch cards, for
classifying plants, since in making comparisons of many variables
when he studies his specimens the taxonomist is often unable to con-
vert his mental picture of these variables into a system which can be
communicated readily.

Anderson (1957) attempted to evaluate the objectivity of the
"taxonomic method" (he used the term "taxonomic intuition") by
sending pressed plant material to several specialists in different parts
of the world and asking these workers to classify the material as to
the number of taxa involved, particularly as concerned their designa-
tion as genus, species, and variety. The results of the study are
significant in that most of the workers were in essential agreement
as regards the degree of relationships expressed, and, in particular,
there was remarkable extent of agreement as to the generic status of
the material considered. To most taxonomists the nature of this ex-
periment would appear rather trivial. We think it can be fairly stated
that most taxonomists working today who might be working with the
same biological entities and using basically the same data will come
to essentially the same conclusions with respect to the recognition
and relative reink of the biological entities considered. The differences
that one might expect are the actual hierarchies assigned to the
categories recognized. For example, one worker might recognize ten or
fifteen genera in a given family, while another might designate only
a single genus for the same group, but recognize, instead, ten or
fifteen species within this major taxon. They both agree as to the
number of biological entities involved. The difference is one of rank

2 Grant (1959) has expressed little hope "at the prospects of purely mechanical
methods in systematics, such as the punching of cards and their classification by IBM ma-
chines. ... If the more obvious characters are selected for scoring . . . [then] . . . Conven-
ience is apt to go hand in hand with artificiahty in the classification of complex groups."

10 BIOCHEMICAL SYSTEMATICS

which involves a subjective judgment. The biological status of these
taxa would not be changed if they were called families or, for that mat-
ter, orders. However, one should understand that any changes in the
nomenclature of the categories of a portion of a taxonomic system or
arrangement should be followed consistently throughout that portion
of the system under consideration.

It is evident that the taxon which lends itself most readily to
experimental techniques, that is, the species, is also the taxon which
is most likely to intergrade morphologically and genetically with
some closely related taxon. Thus the species is the most difficult
taxon for which to discern discontinuities and to estabUsh parameters
for recognition purposes. As one proceeds from the species to the
genus, family, order, and so on, though the discontinuities between
these various taxa becomes increasingly large, and consequently easier
to circumscribe and identify, nonetheless the subjectiveness of these
categories increases.

Or, stated another way, it is easier for the taxonomist to
circumscribe and hence recognize the major taxonomic categories in
spite of the fact that the lesser specific and infraspecific categories are
better defined biologically and lend themselves to experimental
genetical and populational studies.

EXPERIMENTAL CATEGORIES

The development of cytogenetics and its application to taxon-
omy made possible a quasi-experimental approach to plant classifica-
tion. It was natural that early workers in this area of systematics felt
that a panacea was in the making and that with detailed (cytogene-
tical) study much of the difficulty in defining or circumscribing for-
mal categories would soon become a matter of the mere accumulation
and application of such data. Unfortunately, this has not proven to
be the case. It soon became apparent that sometimes obviously
closely related taxa would not hybridize while morphologically more
distinct taxa hybridized with ease, often both in the experimental
garden and in nature. Many studies which were conceived to establish
genetic affinities between taxa of given groups more often succeeded
in showing degrees of reproductive success or failure rather than
demonstrating comparative genomic differences.

Such reproductive data are often difficult to obtain, and even
where assembled the data may contribute little to the solution of the
species problem since, at least in the higher plant groups, taxa show
all degrees of reproductive affinity, depending on the time and cir-
cumstances under which hybridization occurs (either artificial or
natural).

TAXONOMIC PRINCIPLES ] }

Even such a promising criterion as chromosome number was
often found to be a poor guide for the identification or circumscrip-
tion of certain plant taxa. For example populations, and even in-
dividuals within populations, of Claytonia virginiana (Roth well, 1959;
W. Lewis, 1962) and Cardamine pratensis (Banach, 1950; and others)
tolerate a wide range of chromosome numbers. While polyploids of a
normally diploid entity are often ecologically, if not morphologically,
distinct, they are sometimes interspersed within populations which
appear to be fairly uniform from an ecological and morphological
point of view. Examples of diploid and tetraploid populations or in-
dividuals which can be distinguished in no other way than by their
chromosome number are becoming increasingly common in the taxo-
nomic literature, and this fact has understandably diminished the
hopes of many workers who would wish to use cytogenetical data as
the final criterion for categorical disposition.

Fortunately, most workers, while recognizing the value of
cytogenetical data for systematic purposes, have been aware of the
taxonomic chaos that might ensue at the specific and infraspecific
levels if any attempt were made to define rigidly the formal categories
in terms of reproductive affinity or chromosome number. The formal
categories, which are established by international agreement under an
appropriate code, have been erected and modified subsequently by
several generations of taxonomists. The taxa are usually circum-
scribed by discontinuities, and more often than not they are natural
biological entities classified according to their relative morphological
similarities or differences (which presumably is a reflection of their
genetical similarities or differences).

The "experimental categories" (see below) are in reality no
better defined than the formal categories and, as indicated above,
they suffer an inherent classificatory deficiency in that they may or
may not reflect relative genetic differences between and among
taxa. Lewis (1957) has clearly set forward the value of experimental
systematics from the standpoint of taxonomy by pointing out that
while such approaches do not permit an objective definition of the
species, they do provide an orientation for the concept. Hecht and
Tandon (1953) have appropriately stated that:

The delimitation of two species upon the basis of their failure to form
a hybrid is untenable wherever single or few gene differences or simple
structural heterozygosity leads to the formation of nonviable combina-
tions. Incipient species may owe their origins to differences such as
these, but the accumulation of further differences must follow before
what was once a single species may be considered as two.

BIOCHEMICAL SYSTEMATICS

Lewis (1957), in a brief and excellent paper dealing with the
relation of genetics and cytology to taxonomy, has stated.

Highly interfertile geographical races of a species may be genetically
far more different and phylogenetically much more distant than
morphologically comparable but, intersterile populations. . . . Con-
sequently, we should not attempt to reflect in our formal taxonomy
evidence of discontinuity in the genetic system unaccompanied by cor-
responding genetic differentiation.

Unfortunately too few of the early experimental workers
recognized the limitations of their approaches, and, instead of accept-
ing a modicum of rationale in the classical approaches, they were
often overanxious to submerge or erect a species on the basis of rather
limited or questionable cytogenetical data.

The most widely used series of experimental categories are
the ecotype, ecospecies, and cenospecies which are based on an

Table 2-1. Analytical key to the experimental categories. (After Clausen, 1951.)

ECOLOGY

GENETIC RELATIONSHIPS

MORPHOLOGY

Hybrids

Fertile.
Second Gener-
ation Vigorous

Hybrids Par-
tially Sterile.
Second Gener-
ation Weak

Hybrids
Sterile,
or None

Distinct

distinct
environ-
ments

Distinct

subspecies

(or ecotypes)

of one species

Distinct

species

(ecospecies)

Distinct
species complexes

In the
same
environ-
ment

Local

variations

of one species

Species

overlapping

in common

territory

(cenospecies)

Similar

distinct
environ-
ments

Distinct
ecotypes

of one species

Genetic species
only

In the
same
environ-
ment

Taxonomically

the same

entity

(autoploidy o
repat!

r chromosome
terning)

TAXONOMIC PRINCIPLES ]3

ecological-genetic classification (Grant, 1960). Table 2-1 shows the
characteristics and relationships of these informal groups. These and
similar categories are becoming increasingly common in the sys-
tematic literature. They are useful additions to the vocabulary in that
they enable the experimental worker to describe more accurately the
kinds of biological entities with which he is concerned. Information
conveyed in this form avoids any cumbersome explanatory extra-
polations to the formal categories. In addition to the experimental
categories shown in Table 2-1, many additional informal descriptive
terms have been proposed by numerous workers (Camp and Gilly,
1943; Grant, 1960; and others).

BIOCHEMICAL CATEGORIES

With the accumulation of chemical data from various plant
groups it seems likely that some serious attempt will be made to erect
a special nomenclature to deal with those categories so delimited.
Tetenyi (1958) has already proposed a series of infraspecific categories
such as chemouar, chemoforma, and chemocultivar, and so on to
designate appropriate races or forms of chemically defined taxa. We
are inclined to agree with Lanjouw (1958) "that chemical strains or
varieties formed in the wild should be treated as ordinary infraspecific
units"; however, we doubt that these groups, unless accompanied by
sufficient morphological divergence, should bear formal names accord-
ing to the International Code of Botanical Nomenclature. It is al-
ready apparent that chemical components may show variation just as
do morphological features, and any effort to encourage a formalized
nomenclature would only invite a deluge of names which would
further extend the lists of synonymy and in other ways increase the
nomenclatural burden. For the present, it appears wiser to develop
informal descriptive categories, much as has been done by the cyto-
genetical workers. As an example, one could speak of the chemical
races of a given taxon using the distinguishing constituents as adjec-
tives—thus, cyanogenetic race or acyanogenetic race, and so forth.
There seems to be little merit in a formal system along the line sug-
gested by Tetenyi (1958) and Mansfeld (1958). If we are to believe in
the biochemical individuality within Homo sapiens (Williams, 1956),
there would be nearly as many formal "varieties" or forms as there
are people.

The field of biochemical systematics is too poorly developed
to predict accurately its long-term effect on plant taxonomy. We are
certain that it will add greatly to the data with which to develop
further our system of classification. However, any changes in the
nomenclatural system will surely be incidental to its more important

14 BIOCHEMICAL SYSTEMATICS

contribution, that of providing a biochemical basis for showing
relationships and ultimately the recognition and incorporation of
molecular evolution into the over-all, synthetic concepts of taxonomy.

Phylogenetic concepts and taxonomic systems

Crow (1926) has presented an excellent argument in defense
of phylogenetic approaches to taxonomy, the following exerpt being
typical:

The relationships of organisms with one another are not theoretical
interpretations at all, but descriptions of the actual facts of the relation-
ship of parts of one organism to another. Phylogeny consists of theories
and hypotheses formed from these facts. . . . Phylogeny can give little
satisfaction to those who desire absolute truth, but those who hold a
partial view to be better than none at all may find it an interesting
study.

Theories and hypotheses, essential to analytical science,
while often rejected ultimately in the light of unfavorable subsequent
evidence, are symbolic of progress, and failure of a new theory or a
new hypothesis to emerge is perhaps indicative not of vitality but
rather stagnation in that instance. No scientific discipline, unless it is
purely descriptive, can afford to discourage or impugn the erection of
rational hypotheses from available knowledge. Nevertheless, in sys-
tematic biology, which is an analytical science, those attempting to
erect phylogenetic systems of classification, particularly those treat-
ing groups at higher taxonomic levels, often must defend not only
their particular hypothesis, but even the utility of hypotheses per se.
Doubtlessly many of those who object to phylogenetic classifications
(Gilmour, 1961; Russell, 1962; and others) have, in part, acquired
such an attitude as a result of the multiplicity of differing systems
which have been proposed for particular groups. All of the systems
are stoutly defended by their proponents, and, among the compre-
hensive systems, all are constructed from more or less the same
available data. Even some taxonomists have argued that systematists
should not strive to arrange and classify plants on an evolutionary
basis but rather should classify only on the basis of total similarities
(such a system may be referred to as "natural" even though not im-
plicitly phylogenetic). However, such a position cannot possibly be
defended on philosophical or even pragmatic grounds, and the writers
consider it axiomatic that phylogeny is the intellectual forte of
systematics.

TAXONOMIC PRINCIPLES 15

Any hypothetical arrangement purporting to show phyletic
relationships, whether based on cytogenetical, biochemical, morph-
ological, or a combination of such data, although of limited value in
itself, may be catalytic in the sense that it elicits further speculation
and wider associations or suggests preferred additional investigation.
In fact, it has succeeded if it has merely received sufficient attention
to persuade its declaimers to crystallize their own position and re-
appraise the total evidence. Of course, a parade of tenuous and
vacuous theories of trivial nature is to be discouraged, but most
of this type are rather easily perceived by the competent systematist.

Prior to Darwin's publications there were few, if any, pur-
portedly phylogenetic systems of classification proposed by the
serious plant taxonomist, for so long as taxonomists accepted the idea
of special creation, they were not likely to be concerned with phy-
logeny. While several outstanding taxonomists during the 1800's
classified plants by a "natural" system, they often made no serious
or conscious attempt to place the major taxa together according to
their evolutionary relationships. For example, such outstanding
workers as Bentham and Hooker, in their classic Genera Plantarum,
placed the gymnosperms between the dicots and monocots instead of
placing the latter two together as most phyletic workers have done
since that time. Nonetheless Bentham and Hooker's work remains to
this day a useful system, mostly "natural," but not phylogenetic.

Much has been written about the speculation involved in
numerous attempts by taxonomists to show phylogenetic relation-
ships at various taxonomic levels. While most workers concede that
it is possible to hypothesize with considerable assurance at the generic
and specific level, mainly because these lower categories are suited to
experimental, cytogenetical, and populational study, they also rec-
ognize that attempts to construct phylogenetic classifications at the
higher taxonomic levels often involve highly subjective judgments.
The fact that it becomes more difficult to position taxonomic groups
with respect to each other at the higher taxonomic levels in no way
invalidates the objectives sought, and the admission that this can
be done at the lower levels, in principle at least, assures the worker
that attempts to do this with the higher categories are fundamen-
tally sound.

Some workers have despaired of ever achieving any stable^
or useful phylogenetic classification and have argued for a system
that is both reasonably "natural" and useful but without phyletic

3 Many persons concerned with taxonomic problems, not necessarily taxonomists
themselves, deplore the repeated rearrangements of taxa that appear necessary as new in-
formation accumulates. Let us suppose a species, long placed in a particular genus, after
careful study is found to belong to some other genus. Under a set of international rules,

16 BIOCHEMICAL SYSTEMATICS

overtones. Whatever the argument against the incorporation of phy-
logeny in classificatory systems, it seems obvious that if plants are
arranged in as close a phylogenetic order as possible, along somewhat
practical lines,^ the taxonomist has performed a service, however
small, to the biochemist interested in natural plant products, to the
geneticist interested in making realistic crosses, or to the pharm-
aceutical worker in his efforts to locate new sources of drugs. In
addition to these factors, as noted previously, phylogeny provides in-
tellectual vitality to taxonomy.

Actually, most phylogenetic workers are cognizant of the
speculative nature of their various systems, but many outside of the
field are not fully aware of the tentative nature of differing and often
contending systems. The fact that evidence is not available to prove
or disprove one of two contending hypotheses concerning a particular
relationship does not invalidate the system as a framework for future
investigations. As new evidence accumulates, one of two competing
systems may increase in favor. Indeed, the two may be replaced by a
third which, while perhaps incorporating parts of both previous sys-
tems, may be substantiated with new evidence and information which
were not available to previous workers.

Systems of classification

Lawrence (1951) in an excellent treatment of the history of
classification stated that:

Many different classifications of plants have been proposed. They are
recognizable as being or approaching one of three types: artificial,
natural, and phylogenetic. An artificial system classifies organisms for
convenience, primarily as an aid to identification, and usually by
means of one or a few characters. A natural system reflects the situa-
tion as it is believed to exist in nature and utilizes all information
available at the time. A phylogenetic system classifies organisms
according to their evolutionary sequence, it reflects genetic relation-
ships, and it enables one to determine at a glance the ancestors or

the taxonomist now must make a formal name change, replacing the generic name, and,
if necessary, the specific name. But, why make the change? Changing the name doesn't
change the plant. What is gained by such action?

The answers should be obvious. The previous position of the species was un-
natural and phylogenetically unsound. The scientist must recognize natural relationship
or phylogenetic position by making the appropriate transfer and resulting name change.
The latter is only incidental to the primary purpose in this redisposition.

* Other systems might be easier to erect, maintain, and use for identification pur-
poses, but the utility often ends there.

TAXONOMIC PRINCIPLES ]7

derivatives (when present) of any taxon. The present state of man's
knowledge of nature is too scant to enable one to construct a phylo-
genetic classification, and the so-called phylogenetic systems represent
approaches toward an objective and in reality are mixed and are
formed by the combination of natural and phylogenetic evidence.

In the discussion below we vdll confine our attention primarily
to those systems of a phylogenetic nature. Artificial systems are no
longer used by the professional taxonomist, and since as indicated
earlier truly natural and truly phylogenetic systems are theoretically
synonymous, there is little need to prolong a distinction between
the two except in a historical-philosophical sense such as Lawrence
has done.

After Darwin's work there began to appear numerous and
varying systems of classification, nearly all of which were based on
phylogenetic considerations. Turrill (1942) has perhaps justly criti-
cized much of this speculation as has Lam (1959). The latter author,
in particular, emphasized the necessity of fossil evidence before any
substantial phylogenetic classification might be achieved, and he dis-
tinguishes between systems erected on the basis of "static taxonomy"
(proposed without paleobotanical data) and systems based on
"dynamic taxonomy" (utilizing fossil data).

Since, in the case of most flowering plants, nothing resembling
a progressive fossil sequence exists equivalent to the classic zoological
examples (for example, horse, ammonites, and so on), nearly all sys-
tems of classification for the group are based frequently on arbi-
trary principles as to what constitutes primitiveness or, in turn,
advancement.

Over fifty such principles have been advanced, some of a
contradictory nature depending on the point of view of the systema-
tist (Just, 1948; Constance, 1955). For example, Engler and Diels
(1936) considered that the majority of plants with simple unisexual
flowers were primitive, while Bessey, Hutchinson and others have
considered these same floral types indicative of advancement, the
condition having developed by reduction processes from complete, bi-
sexual flowers. The bases for some of the principles are well docu-
mented by extensive, detailed correlative studies on both living and
extinct groups (for example, the derivation of vessels from tracheids;
Bailey, 1944), while other principles are based more or less on a priori
judgment (for example, the assumption that free petals are more
primitive than connate petals, and so on). It should also be empha-
sized that any evaluation of the various principles must be considered
with respect to the group under examination. Thus Hutchinson

18 BIOCHEMICAL SYSTEMATICS

(1959), in setting forth his views on the phylogeny of angiosperms,
adopts the principle that "the spiral arrangement of leaves on the stem
and of the floral leaves precedes that of the opposite and whorled type."
However, Cronquist (1955), in considering the phylogeny of the
family Compositae, considered opposite leaves to be the primitive
condition for the family, but this need not mean that he considers
this to be a primitive character for the angiosperms generally.
Similarly, Hutchinson's view that the herbaceous habit is primitive
in the Ranunculaceae does not conflict with his supposition that
woodiness is a primitive condition for the angiosperms generally.

Practically all of these principles concern morphological
features, but it is not unlikely that as studies of "molecular evolution"
(Anfinsen, 1959) develop there will be as many, if not more, principles
formulated from purely chemical data. At least one worker (McNair,
1945) has ventured, though prematurely, into this field of conjecture,
and others are sure to follow.

Many of the more recently proposed classificatory systems
are accompanied by schematic diagrams showing the relative taxo-
nomic positions of the taxa treated. Lam (1936) has written an excel-
lent summary of such presentations, some of which are rather bizarre.
Little advance in this type of symbolization has occurred since Lam's
review of the subject. Most workers have presented their diagrams in
a two-dimensional framework, mainly because fossil data are lacking
to substantiate speculations in time. However, some workers, on the
basis of several other kinds of evidence, have sought to reconstruct
the chronological phyletic history of a given group and thus have
added a third dimension, time, to their scheme. Diagrams of the sort
mentioned have been constructed for taxonomic groups at all levels
from the species to the kingdom (Fig. 2-2 to Fig. 2-7). Most two-
dimensional schemes are presented merely to show relative similarities
and differences between taxa, although attempts are sometimes made
to include the "lines of evolution" for the taxa concerned, usually with-
out time connotations.

Two-dimensional phylogenetic diagrams

The two-dimensional presentation is popular because it is
simple to construct and need not reflect phylogeny, though it would
usually imply that the presentation was the best approximation from
the data at hand. One popular form of the two-dimensional scheme
is that shown for the genus Dicentra (Fig. 2-4). While phylogenetic
lines are shown in this scheme and the relative positions of the

TAXONOMIC PRINCIPLES

Miocene

Oligocene

Fig. 2-2. Schematic representation of the suggested origin and
evolution of present day Hymenopappus species (Turner, 1956).

various taxa, as determined from degrees of specialization, are in-
dicated, the factor of time for the assumed branching is not indicated.
The diagram indicates that D. torulosa is morphologically the
most speciahzed, or advanced, and that D. chrysantha is the most
primitive. In terms of position, as determined from the morphology of
the characters selected, D. torulosa is closer to D. scandens than it

BIOCHEMICAL SYSTEMATICS

Fig. 2-3. Diagram showing relationships between Araliaceae and
Unbelliferae. Horizontal lines mark beginning and end of Tertiary;
dotted lines connect morphological levels; solid lines indicate typo-
logical relationships; long double arrows signify actual relation-
ship; short double arrows indicate homologous (parallel) evolu-
tionary lines; oo, 5, 4, 3, 2, 1 = number of carpels; O = genus or
tribe; x = basic chromosome numbers; A = Araliaceae, Mc =
Myodocarpus, U = Umbelliferae, Ap = Apioideae, H = Hydro-
cotyloideae, S = Saniculoideae, M = Myrtales plexus, Ro = Rosales
plexus, Ra = Ranales plexus. Adapted from Baumann (Just, 1948);
copyright (1948) by the University of Chicago.

is to D. chrysantha, but its actual phyletic relationship might be
closer to D. chrysantha, its extreme specialization being a result of
more rapid evolution from the phyletic line culminating in D. chry-
santha. D. scandens possibly diverged earlier from the chrysantha
stock, but diverged at a much slower rate (Fig. 2-5a and 2-5b). As in-
dicated by Stern (1961), "the angles of divergencies, etc. are strictly
diagrammatic and are not designed to denote constant rates of
divergencies of evolution."

The Dicentra diagram was constructed primarily from inter-
pretations of exomorphic features. It is sometimes possible to con-
struct two-dimensional phyletic diagrams with assurance, often with
experimental support, when working with species groups where
hybridization, autoploidy, and amphiploidy have been major imme-
diate factors in the speciation process. The diagram for the genus
Clarkia by Lewis and Lewis (1955) is one of the better documented

TAXONOMIC PRINCIPLES

D. tofulosa
I
D. lichlanqe'nsis

» /'- •

D. macrocapnos

' 'A

J. paucinerva

15 14 13 12 11 10 9

1 12 13 14 15

Fig. 2-4. Graph based on specialization indices indicating the
probable phylogeny of Dicentra. Upper left: Subgenus Dactyli-
capnos. Lower left: Subgenus Chrysocapnos. Right: Subgenus
Dicentra. Higher values indicate a greater degree of specialization
(Stern, 1961).

MORPHOLOGICAL SPECIALIZATION

chrysantha scandens

Fig. 2-5. Two possible phyletic interpretations of portions of the
diagram shown in Fig. 2.4.

3 «

?- c

01 .S
ISO — '

, >i

s >

-^^ CD

»■=

■ii

a. -o

:a s

cc cC

C o

.a 43

CO •■— '

_UJ

m 0)

fl)

^ S

" M

•+3 a

0) -a

l-l

o .2

t-g

'o 'S

(Ti

fi 2

r-H

« t;;

•2 i
s

c«

* ^4

-:s

o
c

ti^

TAXONOMIC PRINCIPLES 23

cases utilizing such information in conjunction with exomorphic fea-
tures for phyletic evaluation. As is obvious from this diagram (Fig.
2-6), putative diploid species must have preceded the derived poly-
ploids, but again the relative time of such divergences is not shown in
the diagram.

Three-dimensional phylogenetic diagrams

For higher plant groups where fossil data are mostly lacking,
three-dimensional schemes are usually purely speculative. Nonethe-
less, some monographers have ventured to reconstruct the phyletic
past using geomorphological, phytogeographical, ecological and other
lines of subtle evidence. If, for example, a North American genus with
five species is critically studied and it develops that two of the species
occupy mesic habitats which are believed to be floristically old (such
as extant remnants of the Arctotertiary flora; Chaney, 1938), while
the remaining species occur in grassland and desert habitats (which,
on paleobotanical grounds, are believed to be more recently derived
vegetational types; Axelrod, 1950, and others), then this information
can be used to give relative time dimensions to any appropriate
phylogenetic diagram. Phylogenetic schemes constructed from such
data are often severely criticized, but, as indicated elsewhere, as a
framework for future investigation they are often of definite value.

Time-dimensional phyletic diagrams have been proposed for
the evolution of organic matter and organisms for the planet Earth
(Fig. 2-7), for the relationships between and within several families
(Fig. 2-3), for species within a genus (Fig. 2-2), and so on.

Classification of vascular plants

Because of the complex morphological variation of the vas-
cular plants, this group has been the most extensively and successfully
studied from a phylogenetic standpoint. This is particularly true of
the flowering plants, and a number of systems of classifications, usually
to the level of family, have been proposed for this group (Lawrence,
1951, for review; Cronquist, 1957, 1960; Benson, 1957; Hutchinson,
1959; Takhtajian, 1959; and others). However, only a few phylogenetic
systems have gained wide acceptance or attention, the more important
being the systems of Engler, Bessey, and Hutchinson. Certain aspects
of these three systems are discussed briefly below, mainly to acquaint
the nontaxonomist with their nature and objectives.

BIOCHEMICAL SYSTEMATICS

Basid
Myc. 16.0

Flagell. 0.4
Ascomyc. 16.0 (^^1 Bact. 1.2

Cyan. 1.8

Perid.
1.0

Fie 2-7. Spherical system of the microcosm, consistmg of an
infinite number of concentric "Time spheres." Adapted from
H. J. Lam, 1936, with the authors permission.

1. ENGLER SYSTEM.

As indicated by Lawrence (1951, pp. 118-120), Engler
"attempted to devise a system that had the utiUty and practicahty of
a natural system based on form relationships and one that was com-
patible with evolutionary principles." However, Engler considered the
anriosperms to be polyphyletic, and his arrangements are more an
attempt to show progressive complexity in structure rather than a
^hyTgenetic sequence. This system has gained w,de acceptance pr.
mi^ily because of its broad and detailed coverage, and the plante m
many of the world's major herbaria are arranged accordmg to this sys-

TAXONOMIC PRINCIPLES 25

tern as are the treatments in numerous floras and texts. Engler's system
is not ordinarily displayed in schematic form, mainly because its author
did not claim his treatment to be phylogenetic (Turrill, 1942), and the
system is recognized by most taxonomic workers as a useful but
partly artificial arrangement,

2. BESSEY SYSTEM

Bessey was one of the most astute and prolific American
taxonomists to put forward a system of classification for the higher
plants. His system (Fig. 2.8) differed considerably from that of Engler
in that, instead of emphasizing progressive specialization from the
superficially simple flowers of both monocots and dicots, such as Engler
proposed, Bessey felt that progressive diff"erentiation has proceeded
along a number of lines, one of these being the loss of parts from a
relatively simple but multicarpellate perfect flower such as is found
in the families Ranunculaceae and Magnoliaceae. This system was
not elaborated nearly to the degree that Engler's system was, and, in
addition, it suffered certain shortcomings resulting from the fact that
Bessey had only fragmentary knowledge of the families indigenous to
other parts of the world. In any case, Bessey's system did not receive
wide acceptance outside of the United States, although, as is apparent
from the Hutchinson system (discussed below), the principles on
which Bessey's system was erected have received wide approval
elsewhere.

3. HUTCHINSON SYSTEM

Hutchinson's system of classification for the flowering plants
was formulated on about the same principles as Bessey's system with
one important exception: Hutchinson thought that there occurred
early in the evolutionary history of the group a major phyletic di-
chotomy, resulting in an herbaceous offshoot which produced both the
herbaceous dicots and the predominantly herbaceous monocots of
today. The ancestral woody plexus was believed to have given rise to
those dicot families with mainly woody species. When the herbaceous
habit is found in otherwise essentially woody families such as the
Leguminosae, it is assumed by Hutchinson to have an independent
origin. The same is believed to be true for those semi-woody groups
which occur in essentially herbaceous families (for example. Clematis
in the Ranunculaceae).

Hutchinson's scheme allows for the wide separation of what
heretofore have been looked upon as fairly closely related taxa (for
example, the Umbelliferae and Araliaceae; see Baumann's phyletic
diagram for these groups. Fig. 2-3). Hutchinson ascribes much of this

BIOCHEMICAL SYSTEMATICS

o
o

I —

•et
O

o
O

'in

a
o
O

S-.

I-H

s:;

r-t

-S

0)
0)

■*-.

o
o

r/1

h-H

bf)

C
CO

o
C

<i>

-C

-tJ

ffi

•

;-i

tlH

^■1

IM)

TAXONOMIC PRINCIPLES 27

similarity to convergent evolution (discussed below). Hutchinson con-
tends that there is a considerable and fundamental phylogenetic gap
between a buttercup and a magnoha tree and that, although the
herbaceous habit has developed independently in several woody
famihes, the preponderance of morphological evidence supports his
arrangement.

Figures 2-8 and 2-9 do not show the arrangement of all the
famihes within the orders recognized by Bessey and Hutchinson, but
this information is included in their original presentations. It is im-
portant to remember that these systems, while agreeing in parts, are
contending hypotheses. The authors recognized this, for, as Hutchin-
son (1959) stated in the preface of his latest work.

Botanical systems can never remain static for long, because new facts
and methods of approach are liable at any time to modify them. Like
other things in this changing world, that which seems to be a prob-
ability or even a certainty one day may quite weU prove to be a fallacy
the next.

Diagrams of the type mentioned above enable the interested
worker to tell at a glance the presumed phyletic relationships within
the groups concerned; however, it cannot be overemphasized that
these are, at the most, hypothetical in nature and only in the rarest
instances are they free of gross oversimplifications. For the experienced
taxonomists such schemes may prove more irksome than instructive,
but to the systematically inclined organic chemist (possibly even for
speciahsts such as palynologists, embryologists, floristic cataloguers,
and so forth) they might provide some insight not apparent from the
more formalized monographic treatments.

Parallelism as a factor in classification

Grant (1959) attributes to two principal factors the main
responsibility for the differing generic treatments accorded the phlox
family (Polemoniaceae) by several workers on the basis of facts
available. These are: reticulate relationships following ancestral
hybridization and parallelism in evolution. As indicated in Fig. 2-10,
the two phenomena are often concomitant. Several workers have
felt that convergence and parallelism per se make it diflicult, if not
impossible, to erect meaningful phylogenetic classificatory schemes,
and some discussion of these phenomena will be included here.

Parallelism may occur as a result of hybridization and sub-
sequent backcrossing (Fig. 2-lOb). This type of parallelism, whether

BIOCHEMICAL SYSTEMATICS

' i ^

a
o

to
g

c
o

CO
O

2 i^

o s:

C o

.2 «3

cd -^
CO <i)

bD CO
,2 «

a >.

-D Sh

-^ .5

bj) ><

•| '^
o of

u bC

.2 >.

Q &

TAXONOMIC PRINCIPLES

detected or not, would hardly affect classificatory systems since, in
both the phylogenetic and typological approaches, the taxa con-
cerned would be grouped in about the same relative systematic posi-
tions. The type of parallelism shown in Fig. 2- 10a poses a more
difficult problem, but, except where one or only several criteria are
selected for emphasis over other kinds of data, such cases are
apparently uncommon. When autonomous paralleHsm^ following
"convergence" has been a factor in the evolution of a plant group, its
discovery is more likely to reflect the soundness of a broad, synthetic
(albeit predominately morphological) approach to higher plant classi-
fication.

The case for convergence in most closely related taxonomic
groups usually rests upon the quantitative features in one or at most
only a few characters. If these characters are important "key" charac-
ters (discussed below), then any systematic treatment based on such
features is hkely to be more artificial than natural, and to cite such
examples as instances of erroneous phylogeny resulting from conver-
gence and parallelism is to stretch the case. If two taxa have diverged
sufficiently to be recognized by their phenotypic differences, reflecting
multiple gene differences, then, on a priori grounds, the chance for
absolute genetical convergence seems most unlikely in view of our
present knowledge of mutational rates and the subtleties involved in
the selective forces having to do with character fixation.

Several workers have mentioned examples of what appear to
be autonomous convergence and parallelism for certain characters of
different taxa of higher plants (Bailey, 1944). An even more striking
parallelism has been described for some chemical components of
otherwise widely differentiated taxa. One rather striking example is
the occurrence of the hemoglobin molecule in cells of fungi and in the
root nodules of legumes (White, et al, 1959). Several additional
examples of chemical convergence and parallelism will be discussed
elsewhere in the present text more fully.

The argument that convergence and parallelism make it im-
possible to achieve a meaningful phylogenetic system can be appro-
priately countered with the following remark from Crow (1926):

The problem of the cause of convergence and parallel development is,
of course, an extremely important one. But inasmuch as convergence
itself was discovered by systematic and morphological investigations,
and is itself a phylogenetic conclusion from the systematic and
anatomical facts, the necessity of making more detailed study of
phylogeny is all the more necessary. ... To use the polyphyletic origin

5 As distinguished from parallelism due to hybridization and subsequent back-
crossing.

BIOCHEMICAL SYSTEMATICS

Fig. 2-10. Two ways by which parallelisms can develop in evolu-
tion Parallel selection is assumed in each case. (A) Independent
parallel mutations at homologous loci. (B) Hybridization foUowed
by segregation in the direction of one or both parental species (Bi),
or followed by backcrossing, viz. introgression (B2), V. Grant,
Natural History of the Phlox Family. Systematic Botany. Inter-
national Scholars Forum A Series of Books by American Scholars,
Sciences 1. The Hague: Martinus Nijhoff, 1959.

TAXONOMIC PRINCIPLES ^l

of a group formerly supposed to be a natural (monophyletic) one as an
argument against the possibility of constructing a natural system is
nothing more nor less than to use the conclusion of phylogeny to dis-
prove phylogeny.

The fallacy of the "fundamental" character

Most workers today are aware that any ultimate system of
classification must be based upon the available data from all fields.
To assemble these data is difficult enough, but to assess their phyletic
significance often appears impossible. This is particularly true with
respect to morphological features (as opposed to chromosomal or
genetic data). For example, what genetic (or phylogenetic) signif-
icance does an inferior ovary versus a superior ovary have? How does
one evaluate the genetic significance of separate carpels as opposed
to fused carpels? Of course the answer is sometimes obvious when one
is considering the mere presence or absence of a given character
(other characters being similar), but when two taxa are separated by
a combination of morphological features, all of which vary, both
quantitatively and qualitatively, there is no simple solution.

Because of the complexities involved, many workers set
arbitrarily certain "fundamental" or technical characters to mark
given groups. Consider the largest angiosperm family, the Compositae,
with over 30,000 species. All of the species are more or less ahke
m that most contain an involucrate head, four or five united petals, a
modified calyx-like structure (the pappus), an inferior ovary and two
carpels, a single style with two branches, and so on. In spite of
the extraordinary similarity of all of the species in this family, most
workers have grouped the species into twelve or thirteen tribes. The
tribal groupings are mostly natural, but occasionally certain taxa are
misplaced as to tribe, mainly because of the too rigid adherence to the
so-called "fundamental" features used to delimit the tribes initially.
For example, the genus Hymenopappus had long been placed in the
tribe Helenieae because of the absence of chaff. However, more recent
investigation has shown this genus to be unnaturally placed in the
Helenieae, since its most closely related taxon, Leucampyx, an
obvious prototype for the chaffless Hymenopappus, is apparently cor-
rectly placed in the tribe Anthemideae. The presence or absence of
chaff in this case appeared to be sufficiently "fundamental" to some
workers to separate two very closely related taxa, not only into
separate genera, but even into separate tribes. However, Turner
(1956), on the basis of total data, united the groups in a single genus

-- BIOCHEMICAL SYSTEMATICS

and suggested that their proper tribal disposition should be in the
Anthemideae. Numerous similar cases could be cited.

Of course, the term "fundamental" as applied to such charac-
ters is misleading. They are more appropriately called "key" charac-
ters in that they usually furnish an easily observed, mostly constant
feature by which to recognize the affinities of a given taxon. It often
takes the beginning student many years to appreciate this distinction,
and even today some otherwise well-informed professional taxono-
mists still think of certain single characters as "making" a given speci-
men and/or population a member of this or that species, tribe, family,

and so forth.

Cronquist (1957) has appropriately emphasized this pomt m

stating:

Every taxonomic character is potentially important, and no character
has an inherent, fixed importance; each character is only as important
as it proves to be in any particular instance in defining a group which
has been perceived on the basis of all of the available evidence.

Stated otherwise, there is no inherent value in a selected
single character. As will be indicated in more detail elsewhere, this is
as true for chemical characters as it is for megamorphic features.

Most systematic work of a biochemical nature has been
directed towards the evaluation and construction of phyletic schemes
for the higher taxonomic categories. For example, the detailed
serological work of Mez eventually resulted in the creation of
his now famous "Stammbaum" (Fig. 2-11). It seems apparent that
Mez' diagram was influenced by previous work which was based
essentially on exomorphic features. A critical discussion of the objec-
tives and hmitations of serology will be presented in Chapter 5 of the

present text.

With the development of rather rapid chromatographic tech-
niques which allow rapid detection of numerous chemical constituents
of organisms, it is now possible to make considerable new use of the
many phyletic diagrams which have been prepared by various mono-
graphers. Most chemotaxonomic studies of a correlative nature have
dealt with presumed phyletic relationships at the family level or
higher, reflecting, no doubt, the textbook famiharity of such systems
to many non-biologists. Interpretations of relationships at this level
are perhaps no better than the data on which they are based, and at
the present time these data are still quite hmited.

With present knowledge and techniques, a more meaningful
application of biochemical data towards classificatory schemes may be

Aquifoli
168

Loasa

.167

Begon

im-

Erica

.Poiygala

)66

Vitaceae

Umbelliferae
144 /

y143 LmS Balsamina

165

Symploca Composifae

/ Sapoto
190^ ^

-<vEbena
Opuntia X^ 189

Datisca
Oxalida ^

169 ^171
163^^ '70

Solano '''X''° 'T'""'

Nolana^:\186~\ 185 ^^

'^°^^ 18A ^183
ScrophulariaA w*^

193 BignoniOv ^*\ L^181 184j

192 177 i'/8 ^-^^^

188 ^191 1 / ^^'^'Y"

y VA-,. 179 ] m-^^^ ^

^ l^^^'° y ^ J^^"""^ Polemonio

b 187. ^ ^^^0.^

160

Corna '39 f 162 ^^^ i ^^^

'42^ ^ i.n (/'^^ ^\ Legumes . il34

140 \ 137V \ nM

132,
131

13T

129

128

133 141

Vellozio

127l

126

Palms
M12

125 ^Bennettites
Dioon 74

CycGS V°"^'"'<90

135^

115

123
Aristo lochig

Anona

66 101

95 i

Davallia 84
' Pteridium

89 .^r^7;
^^_rj 86 85

91 -jf ^^ Dicksonia

Gleichenia

Frullanio

Microlepis ^°

Florideae
Azolla
^g Salvinia

69Y5
70^

57>

Psilotunr_

54 52/10 50
55^ \48 ^

56 53 ^''
Lophocolea

Ricciocarpur
Bangiales

99\

100

^ Morsil

Ectocarpa-

Fucaceae

Heterotrophic
Bacteria

-176

Casuarina-

Fagaceae

Viola
.174

^146

Cista

153 Pipero'

Bixa Batida-

Cruciferae Balanopsida ^

147 >

150^
Caryophylla

149

^156

rll7

^j^g Calycantha

111

:116

Taxus-

Cedrus

pi 10 'ffL'^n 120-

PinUSyi'lU"

^148
Juniperus-
Gnetum
'Ephedra

'121

Thuja

122

H07

Lycopodium

.103
h02

^T Polytrichum

Amanita
Phallus 43^42
25i/
Musci f24

40) /4]

37>

^46

Cha vales
12

'29

,44

Tilleri

sria-

^10

Ustilago

-Oedogonium
JJIothrix

4 5 Desmidium

'Crenothrix

-Euglenoidales
-Mycetozoa

-Amoebeae Animal
Kingdom

Autotrophic ■

' Bacteria

Fig. 2-11. The "Stammbaum" of Mez, a phylogenetic tree pur-
portedly constructed, in part, from serological data; adapted from
Gortner, 1929, Outlines of Biochemistry, John Wiley & Sons, Inc.
The ending "aceae" is omitted from many of the famihes. For the
names of taxa that correspond to the numbers indicated in the
diagram, see Appendix I, p. 345.

^ - BIOCHEMICAL SYSTEMATICS

made at the family level and lower. Carefully constructed phylogenetic
systems have been prepared for numerous generic groups, but only in a
few instances (for example, Baker and Smith, 1920, on Eucalyptus) has
there been any concentrated effort to evaluate such systems with
purely biochemical data. For example, detailed chemotaxonomic work
of this nature on the phylogenetic groupings proposed within the
genus Crepis (Babcock, 1947) should prove exceedingly rewarding,
and might provide new data for relationships yet undetected. The
hypothetical phyletic diagram for the genus Hymenopappus (Fig. 2-2)
could be used profitably for the orientation of a purely chemotax-
onomic study; for example, will biochemical data further support the
basic dichotomy indicated by the Series Biennis and Perennis, or
will new data come to Hght that might indicate a much more reticulate
relationship between the species of these two series than is indicated?
It might even be possible to test by chemical data the vahdity of some
of the time speculation indicated in the Hymenopappus diagram. For
example, it has been demonstrated in numerous instances that certain
molecular configurations must occur before some more "advanced"
reaction is possible (cf. Ch. 11, p. 197, rotenones). If the latter molecular
configuration was found only in the morphologically more advanced
desert species, then this would correlate with the evidence from both
morphology and paleobotany as to the time of origin of desert habitats
and the plant types which must have become adapted to such regions
after or concomitant with their development. By the same reasoning,
species which have retained certain hypothetical ancestral morpholog-
ical features and ecological associations might be shown to have one
or several of the metabolic precursors necessary for the molecular
advancement indicated.

The approach to systematics of genera and lower categories
using biochemical patterns has not been vigorously pursued, but
as indicated by Alston and Turner (1962) it is capable of sufficient
refinement that not only are species detectable but also degrees of
hybridity for individuals from hybridized populations. Furthermore,
it appears likely that with appropriate controls biochemical patterns
can be constructed which permit rather objectively determined visual
presentation of numerous chemical features for inter- and intra-
populational comparisons. Data obtained chromatographically can
also be expressed mathematically with a minimum of interpretative
effort so that considerable exactness in the presentation of relationship
data can be achieved. Limitations involved in this type of comparison
are obvious, of course, and further discussion will be devoted to eval-
uation of biochemical data in a later chapter.

The present categorization of vascular plants was developed

TAXONOMIC PRINCIPLES 35

by several generations of taxonomists, each generation adding obser-
vations and concepts to the preceding. Descriptive data were compiled
for the lower taxa first and their significance and limitations deter-
mined before meaningful interpretations and circumscription of the
higher categories could be made. Many errors were forthcoming in the
extrapolations and interpretations incidental to its construction, but,
over-all, the resulting taxonomic structure rests on a solid foundation
of observational fact as opposed to mere conjecture.

Phylogenetic knowledge of both the major and minor catego-
ries of classification is certain to advance as our knowledge of
biochemistry advances. To be sure, the ultimate proofs of the system
must depend on the evidence from all fields, mainly paleobotany, but
we can no longer tacitly assume that "... a natural classification must
in the main be based on external characters, simply on account of the
much larger number of these and their much more restricted inci-
dence" (Sprague, 1940). There is a wealth of biochemical data awaiting
exploration, and, while the gross examination of leaves and floral parts
might be the most practical method for the classification of most
plants today, the chemical approach is certain to add significantly to
any ultimate phylogenetic system. Even at the level of identification
there is a significant advantage to the biochemical approach, for if an
exomorphic taxonomist were asked to identify a plant from a leaf or
petal fragment he might despair, but given chemical data he might be
able to identify the fragment to species. This can be done with
certainty in the case of the species of Baptisia so far examined
chemically.

PLANT TAXONOMY

a brief history of

Major developments in the field

The history of civihzation, or indeed all time-
dependent phenomena, can be divided into a num-
ber of major chronological periods according to the
intellectual imagination or disposition of the
recorder. Thus one might partition historic time
into one-hundred-year periods and graphically treat
each unit with equal systematic coverage as if
history were a straight line whose ascending time-
event path was devoid of significant event fluctua-
tions. Fortunately for students of history, most
historians have found it more appropriate to divide

-o BIOCHEMICAL SYSTEMATICS

recorded history into large or small time periods according to the
importance or significance of the events surveyed.

Botanical historians have also recognized the special signifi-
cance of certain contributions in making possible the development of
new vistas in botany. Greene (1909) in his Landmarks of Botanical
History emphasized the major early descriptive developments in
taxonomic botany, particularly as related to specific individuals and
their contributions to systematics. Beginning with prehistoric time, he
recognized as foremost (1) the descriptive contributions of Aristotle
and Theophrastus (followed by a long quiescence up to the fifteenth
century), (2) the significance of the observations of the herbalists
Tragus, Brunfels, Bauhin, et al of the sixteenth century, (3) the first
distinction of the monocots and dicots by John Ray in 1703, (4)
recognition of sexual characters and their significance by Linnaeus and
others in the mid-eighteenth century, and so on. Greene purposely
selected the word "Landmarks" in his published title since he recog-
nized "the impossibility of any such thing as a complete and faithful
history of any period when once that period is past."

While such a treatment of botanical history might be sufficient
to show the major descriptive phases, it seems that from a dynamic-
developmental point of view (in the historical sense) taxonomic history,
beginning with Aristotle, can be logically divided into four or five
major periods, each of which is terminated (or initiated as the case may
be) by some major "breakthrough" in scientific thought or through the
development of techniques which have permitted the acquisition
of new data (Table 3.1).

Different writers might recognize yet other "breakthroughs
than those hsted below, but we beheve that few readers will argue
about the impact of each on taxonomic practice and thought.

It should be obvious that the present treatment of taxonomic
history in no way supposes that the vahd techniques or methods
of any prior period give way to those of another. Rather the methods
and ideas of succeeding periods are usually superimposed on the
pre-existing framework; and all are necessary (or at least have so far
been found necessary) in our efforts to obtain an "ultimate" phyloge-
netic system of classification.

These periods of botanical history have been treated exten-
sively by a number of writers. Greene (1909) treated essentially the
Megamorphic Period; Sachs (1890) treated, among others, the Micro-
morphic Period; a number of workers have recently reviewed the
Evolutionary Period (Constance, 1955; Tax, et al, 1960; among
others); certain aspects of the Cytogenetical Period have been
adequately reviewed by several workers (Stebbins, 1950; Clausen,
1951; Heslop-Harrison, 1953; Constance, 1955; Darlington, 1956;

PLANT TAXONOMY

Table 3-1. The major historical or developmental periods of systematic biology.

Period

Time

1. Megamorphic

2. Micromorphic

3. Evolutionary

ca. 400 B.C. to ca. 1700 a.d.
(Beginning with Aristotle's time
and continuing to Leeuwen-
hoek's invention of the micro-
scope.)

ca. 1700 to ca. 1860.
(Beginning with Leeuwenhoek
and continuing to Darwin's pub-
lished views on evolution.)

Characterization
of the period

A terminological-descriptive
period characterized by the
development of formal group
concepts (e.g., families, genera,
species, etc.) and the establish-
ment of a descriptive language
to define these groups better.

4. Cytogenetical

5. Biochemical

ca. 1860 to ca. 1900.
(Beginning with Darwin's evo-
lutionary theory and extending
to the rediscovery of Mendel's
laws of inheritance. )

ca. 1900 to ca. 1960(7).
(Beginning with the rediscovery
of Mendel's laws and extending
to the present time.)

Leeuwenhoek's microscope and
lens systems made possible the
recognition of hitherto unknown
microorganisms, the recognition
of sexual features, and their sig-
nificance and made possible the
acquisition of new morphologi-
cal data (viz., anatomical em-
bryological, palynological, etc.).

Darwin's theory profoundly
affected systematic thinking.
Hereafter most classification
systems were constructed on a
phylogenetic basis.

ca. 1950(?) to

-(?).

(Beginning with the biochemi-
cal approach, made possible by
the development of rapid and
relatively simple techniques
such as chromatography, and
possibly extending to the deter-
mination of the sequences of sub-
units of polynucleotides such as
DNA and RNA and of proteins.
Techniques are already avail-
able whereby nucleotide and
amino acid sequences can be
analyzed.)

This period is characterized by
the detailed application of cyto-
genetical data and populational
statistics to plant taxa, mostly
at the generic, specific, and
infraspecific levels. These tech-
niques permitted the first truly
experimental approach to sys-
tematics.

Characterized in its early stages
by the establishment of "bio-
chemical profiles" for various
plant taxa and their compara-
tive use in solving taxonomic
problems; in later stages by a
comparative biochemical ap-
proach that takes into consider-
ation metabolic pathways, pro-
tein evolution, and comparative
enzymology.

.Q BIOCHEMICAL SYSTEAAATICS

Lewis, 1957; Hedberg, 1958; and others), and it is probable that this
period has not yet made its total contribution (i.e. in terms of broad
principles and ultimate potential).

The following questions may be raised: Are we really at the
beginning of a new period of taxonomic history? Will taxonomically
oriented biochemical investigations yield data that make possible
a better phylogenetic scheme? Will they give answers to taxonomic
questions that previous methods did not permit? Will chemotaxonomy
become as significant in the next half-century as cytotaxonomy
has during the last? Is the time at hand for this molecular approach?

We beheve that plant taxonomy is now entering this new
phase of biochemical investigation. The purpose of the chapters that
follow is to document (though selectively) the present state of our
knowledge in this field, to give our interpretations of the significance of
certain approaches already in use, to evaluate critically the limitations
as well as potential of the field, and, finally, to develop philosophical
concepts that might lead to increased activity and more important
contributions in the future.

INTRODUCTION TO

BIOCHEMICAL

SYSTEMATICS

If there are any biologists who deny the evolution
of metabolic pathways in plants, they must certainly
constitute a small minority. At our present state of
knowledge the evolution of metabolic pathways
may be considered axiomatic. It is self-evident that
certain fundamental pathways such as those in-
volved in energy transfer and the synthesis of basic
protoplasmic constituents appeared before the seed
plants evolved, probably even before the origin of
cellular organisms. However, there are numerous
plant components, broadly classified as secondary
substances, which have undoubtedly evolved late
in the evolutionary progression. Pertinent questions
concerning each of these substances are related to
when and in what group of plants they first occurred

42 BIOCHEMICAL SYSTEMATICS

and how often they have arisen independently. Not only are the
secondary substances proper subjects for such considerations, but
important structural components, such as lignin, which are of
relatively restricted distribution, also have an evolutionary history
which may be informative. Finally, it is highly probable that innova-
tions have appeared even in the fundamental pathways, from time to
time, which have been preserved in the descendants of the organisms
in which the change occurred. Thus it is not gross exaggeration
or mere wishful thinking to assert that a natural system of classifica-
tion is potentially available based on comparative biochemistry.
Actually, comparative biochemistry, itself, may be studied at several
levels. At one level emphasis is upon the distribution of certain
classes of substances, such as, for example, the isoquinoline type
alkaloids. Ultimately, comparative biochemistry will likely be repre-
sented by comparative enzymology or perhaps even the comparative
chemistry of RNA and DNA. It may well be that such studies
will yield the most accurate image of phylogeny, but the first level
approach must precede these more technically exacting ones or at least
be pursued concomitantly.

The distribution of a substance will not necessarily have
positive phylogenetic significance in all cases. Sometimes the com-
pounds may have clearly evolved independently in several plant
groups and will thus be phylogenetically useless at major taxonomic
levels. Nevertheless, those compounds may be valuable in pointing out
relationships within a given taxonomic group where they are found.
The authors have heard a prominent biologist state that biochemistry
can never make any contribution to systematics because certain
substances, such as nicotine, are found in such obviously unrelated
plant groups as Equisetum and Nicotiana. It is tempting to dismiss
this type of argument summarily as not worthy of rebuttal. It follows
from such reasoning that the person making such a statement believes
that the vast majority of compounds have evolved again and again
throughout the plant kingdom or that chemical substances appear,
somewhat capriciously, via a mechanism that transcends the usual
order so that their appearance has no real phylogenetic meaning. Since
the latter argument has been decimated through biochemical genetics,
it need not be taken seriously. As for the first, it is probably that the
mode of evolution of biochemical characters roughly parallels that of
morphological characters in that certain characters evolve repeatedly
(for example, pubescence) and are thus inconsequential at major
category levels, or they may arise once (or appear to have arisen once)
as in the case of double fertilization in the angiosperms. Frequently,
even after intensive study, one does not know whether a given mor-

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS

phological character represents convergent evolution or phylogenetic
affinity. Should it come as a surprise or disillusionment to find that
the same problem may confront one who is attempting to evaluate a
biochemical character? Hansel (1956) has discussed some of the
problems raised in this paragraph. He illustrates clearly the point that
the same basic problems are involved in the phyletic interpretation of
biochemical as well as morphological data. The "percentage of
frequency rule," illustrated with an example from indole alkaloid
distribution (Fig. 4-1) is often useful in the interpretation of the
systematic significance of members of a related series of substances.

At the present time there is no phylogenetic system based on
the distribution of biochemical constituents, nor is there likely to be
one, at least one derived out of the first biochemical level referred to
previously. What comparative biochemistry has to offer is supple-
mentary evidence which, when added to other systematic knowledge,
may clarify or help to clarify a given situation. If comparative
biochemistry seriously contradicted any part of the major structure
of plant systematics, it would be equally as disturbing to the pro-
ponents of comparative biochemistry as to other biosystematists.

The matter of weighing equitably biochemical data, of evalu-
ating it, and comparing it with a given unit of morphological, cytolog-
ical, physiological, or anatomical data is so important that a separate
chapter will be devoted to this topic. In the final analysis one would
hke to translate all differences into gene differences. It is difficult to
do this in the case of most biochemical or morphological data, and
unless hybridization is successful, it is impossible to analyze directly
the genetic basis for a particular difference.

There is reason to believe that in special situations bio-
chemical characters provide advantages if one is considering the
question of the genetic basis for a particular difference. In work to be
more fully described in Chapter 15, Turner and Alston (1959) have
demonstrated recombination of species-specific characters in individ-
uals from natural hybrid swarms of Baptisia species. Most species-
specific substances of the parents were present together in the hy-
brids though often in reduced amounts. In order to translate these
species-specific chemical characters into genetic units of differences
one must produce an F2 generation. Theoretically, if a particular sub-
stance required only one gene from one parent not present in the
other parent, three-fourths of the F2 generation should produce it,
and if n genes were required to form the substance, (|)« of the F2
generation should contain that compound. Therefore, a moderately
large F2 generation should suffice to translate units of biochemical
data into unlinked gene differences, assuming that pairing relation-

BIOCHEMICAL SYSTEMATICS

13 ^

■q ^

JS ITS

C« <T2

^ --H

1— H

a ^

ts :=*

.S K

^-1

O CO

a s

3 .2

b 3

be _o

cfl -n

c +^

T3 -3

^ ^

CC 1)

Sh -^^

o -"

S 2i

a> o

ca oj

CD J=

F-H -^

3 QJ

^ >

>-. «

-o

« cu

■ 03

o ^

OJ t*H

6E O

ca CO

-<^ S-(

C (D

ID ^

c2 g

^ ^

^ "a

ti S

:" o

fc o

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS 45

ships are regular. The advantage of the biochemical characters as
opposed to morphological characters is presumed to lie in the fact that
the biochemical characters are affected in general only in a quantita-
tive way by modifiers while many morphological characters (for
example, leaf form) may be influenced quaHtatively by numerous
modifiers, many of which exert their effect in a cryptic way. Perhaps
this generaHzation may prove invahd, but it is offered tentatively on
the basis of our personal experience with Baptisia hybrids to date and
the much larger background of evidence from biochemical genetics
in general.

Historically, interest in the application of chemistry to sys-
tematics goes back almost 150 years. In some of the writings of
A. P. de Candolle, as Hegnauer (1958) has noted, considerable atten-
tion was given to the chemical properties of plants as correlated with
their morphological characters. Examples from de Candolle cited by
Hegnauer were the observations that all Cinchona species aided fever,
all Pinus species produced terpenes, all Amentifera had astringent
bark and all Convolvulaceae were laxative. However, since it was not
possible before Darwin's time to accumulate any large amount of
chemical data, and since the theoretical implications from the later
fields of genetics, evolution, and comparative biochemistry were lack-
ing, it is understandable why little interest was displayed. In fact,
more often than not, chemical characters seemed to complicate the
existing taxonomic systems. An example of what may have been the
prevailing pre-Darwinian attitude is the statement by John Lindley
in his preface to Vegetable Kingdom, quoted by Gibbs (1958):

In the first place such matters belong to Chemistry, and not to Botany;
secondly, it does not appear possible to connect them with any known
principle of botanical classification; and, moreover, the extremely un-
steady conditions of the opinions of chemists themselves upon the re-
sults of their own researches, would render the introduction of the
supposed results of chemists embarrassing rather than advantageous.

Yet, in 1886, twenty-eight years after the appearance of the
Darwin- Wallace papers, Helen C. De S. Abbotfi published a paper
entitled. Certain Chemical Constituents of Plants Considered in Re-
lation to Their Morphology and Evolution. After noting that Haeckel
had divided the flowering plants into three groups: those with sim-
plicity of floral elements, those with multiplicity of floral elements,
and those with condensation of floral elements; she stated that

1 Helen Abbott Michael's scientific and philosophical writings, including the
reference cited, may be found in Studies on Plant Chemistry and Literary Papers by
Helen Abbott Michael. The Riverside Press, Cambridge, 1907.

46 BIOCHEMICAL SYSTEMATICS

saponin-containing groups all belonged to the middle group of Haeckel
and that saponin was a "constructive element in developing the plant
from the multiplicity of floral elements to the cephalisation of these
organs." She considered that saponin was "an indispensable principle"
in those plants in which it occurred. Later, she stated that saponin
was a "factor in the great middle realm of plant life when the
elements of the individual are striving to condense and thus increase
their physiological action and the economy of parts."

Such dogmatic assertions concerning the role of saponins as
an "indispensable principle" in "cephalisation," are, at best, exceed-
ingly tenuous, and she has resorted to an anathema to some botanists,
namely, a teleological statement, but in the following remarks she ex-
presses an idea that, in some circles, would be regarded as somewhat
radical even today.

The evolution of chemical constituents in which they follow parallel
Hnes with the evolutionary course of plant forms, the one being
intimately connected with the other, and consequently that chemical
components are indicative of the height of the scale of progression and
are essentially appropriate for a basis of botanical classification. In
other words that the theory of evolution in plant life is best illustrated
by the chemical constituents of vegetable form. (Sic.)

Further, in support of her proposal to utilize plant chemistry in the
pursuit of phylogenetic relationships she called attention to the fact
that disagreement among botanists themselves pointed to an inade-
quacy of morphological criteria. Also she noted that plant chemistry
represented internal influences controlling function and modifying
form rather than external forces. In addition to the preceding ideas
which were basically sound, she concluded, rather naively (not, perhaps
for the period) that "the percentage of any given compound in a plant
would gauge the progress or retrogression of the plant, species or
genus. . . ."

Abbott also pointed out that "albuminous compounds" and
chlorophyll were not likely to be of much use in classification because
they were necessary for the maintenance of life and presumably
occurred in all species. A similar idea has been expressed, in substance,
more recently by Erdtman (1956) and others who noted that secondary
compounds are probably more useful in systematics than are basic
metabolites, so that the idea which has been equated with modern
thought, in reahty, goes back to the previous century.

In the early twentieth century, some remarkably modern or
progressive statements appear. For example, Greshoff (1909), in the

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS 4/

Kew Bulletin, used the term "comparative phytochemistry" which
he defined as "the knowledge of the connection between the natural
relationship of plants and their chemical composition." Greshoff ad-
vocated the use of a short chemical description as a part of the for-
mal description of a new genus or species. This is none other than
the "biochemical profile" which Alston and Turner suggested re-
cently (1959).

As early as 1925 Munkner indicated full appreciation but pre-
mature optimism concerning the use of chemical data in the solution
of phylogenetic problems when he noted that in the "older" sys-
tematics morphological characters alone, and later anatomical charac-
ters, served to relate plants, while in "recent" time comparative chem-
istry was being utilized in the determination of phylogenetic
relationships.

It is reasonable to ask why, in the hght of these early
recognitions of the potential role of biochemistry, so little progress
has been made. The question, of course, does not have a single answer,
but the lack of progress in biochemical systematics may be explained
partly through the developing interest in genetics and enzymology
around the turn of the century. These fields may have lured some
investigators who possibly would have turned to biochemical sys-
tematics.

Also, many of the early surveys of natural products were in-
stigated from the more practical pharmacological approach. Biologists
of fifty years ago were not generally cognizant of the relationship of
chemistry to biology, and the biologist was, therefore, not likely to be
trained in chemistry. If present circumstances reflect persistent view-
points, the systematic botanists must have been even repelled by
chemistry.

Until recently, techniques have not been available to yield
the refined chemical information necessary for biochemical systematics
to contribute greatly to systematics. A long and distinguished period
of survey has provided vast amounts of information concerning the
distribution of chemical substances among the plant species, and in
some instances this has proved helpful, usually in a corroborative way
to morphology, in systematics. Only in the last decade have techniques
such as chromatography allowed the study of microquantities of sub-
stances from individual plants. The application of chemistry to in-
dividual plants has provided almost unlimited opportunity. Of para-
mount importance is the fact that it allows the study of populations
—natural and otherwise. It allows biochemical systematics to become
experimental. Among lower taxonomic categories it often may allow
the natural affinity of phylogeny and genetics to be expressed through

._ BIOCHEMICAL SYSTEMATICS

analysis in a new way-the genetic basis of the expression of chemical
characters. It is not expected that knowledge derived from such tech-
niques in higher plants will equal that gained from studies of certam
microorganisms (in which the focus has been upon biochemical path-
ways rather than phylogeny), but important advances will undoubtedly

be forthcoming. . .

Among some groups of organisms, whose simple organization
limits a morphological basis of systematics, chemical criteria have long
been utihzed. Unfortunately for the present argument it seems that
bacterial systematics is about as far away from "naturalness" as that
of any group of organisms, and chemical criteria have failed to produce
a natural system. According to Van Niel (1946):

Now the fact that the bacteria also have gradually been assigned to
famiUes, orders, and classes does not imply that our understanding of
their phylogeny is on approximately the same level as our understand-
ing of the plants and animals, in spite of the close resemblance of the
structure of the systems of classification. Bacterial taxonomy is far
more similar to Linnaeus' original system of the plants,

However there are special reasons why this situation is to be expected

in bacterial taxonomy. , , i , u-

Among certain groups of lower plants, notably the algae, bio-
chemical criteria, especially the pigment complement and the principal
photosynthetic products, have been given a considerable amount ot
weight, and it is probably correct to say that such criteria were im-
portant factors in the recent revision of algal taxonomy at the highest
level Even in this situation the biochemical information was usually
applied negatively, that is, not to show relationship but to support
non-relationship. It exposed problems too, for example, m the case ot
the siphonaceous alga, Vaucheria, now placed, somewhat conspicu-
ously, in the Chrysophyta (Smith, 1950).

Some preliminary considerations of the

APPLICATIONS OF BIOCHEMISTRY TO SYSTEMATICS

Moritz (1958) in a review of plant serology called attention to
the fact that serology may make contributions to both major and
minor systematic categories; that is, at the family, order, or higher
taxon level (major); or to the systematics of genera, species, and
infraspecific categories (minor). What is true of serology, itself essen-
tially a biochemical method, is true of biochemical contributions m

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS 49

general. Although these two taxonomic levels do not impose any
absolute restrictions upon the particular biochemical approach, and
there is no mutual exclusion, it is important to emphasize some funda-
mental differences.

Biochemical systematics at the major taxonomic category
level involves the use of classical studies of such substances as alka-
loids, and so on. For example, certain plant families tend to produce
alkaloids, while others do not. Within those famihes which do, one is
likely to find a certain class of alkaloid, and related genera are apt to
form a particular example of this type of alkaloid. The basic rationale
is that of associating specific secondary products of restricted occur-
rence with specific groups of plants. Some groups of secondary prod-
ucts, such as anthocyanins, are rather too widespread to be of great
value although we shall find that, even here, the distribution of un-
usual types is meaningful in systematic terms.

Biochemical systematics as applied to minor categories may
be approached in diverse ways. It is theoretically capable of the ut-
most refinement, as will be discussed later in this section. Experimen-
tal chemical systematics is most likely to make a contribution at this
level. One form is the work by Turner and Alston on Baptisia which
has been referred to earlier in this chapter. It may be assumed as a
valid generalization that emphasis is shifting from the major to in-
clude the minor category level. It has only been within the last few
years that certain ultimate goals have even been conceived. A few
examples here will serve to illustrate that definite progress is being
made in directions undreamed of ten years ago.

From the area of serology, an exceedingly interesting situa-
tion has been reported by Suskind (1957). In Neurospora crassa
a number of tryptophan-deficient mutants (td series) have been studied.
Evidence from serological studies indicates that a protein closely related
to tryptophan synthetase (the functioning enzyme) is present in a tryp-
tophan-requiring mutant. In fact, several td mutants have been
studied serologically, and those which exhibit serological cross reactivity
are referred to as CRM (cross reaction mutant). Some mutants (for ex-
ample, td) show no serological difference from the wild type allele
while others, although exhibiting a cross reactivity, demonstrate a de-
gree of reactivity indicating a serological difference. It is particularly
interesting to note that most CRM mutants can be suppressed while
CRM-less mutants cannot be suppressed (Suskind, 1961).

The basic method is to obtain rabbit antibody (using partially
purified preparations of tryptophan synthetase) which neutralizes
enzyme activity. Tests, using td mutants, were conducted to deter-
mine whether or not they could yield a substance capable of combin-

50 BIOCHEMICAL SYSTEMATICS

ing with the antibodies to normal tryptophan synthetase, and it was
discovered that certain td mutants did contain serologically active
though enzymatically inactive material. This presumably represented
an altered protein, formed in the presence of the mutant. It was
sufficiently close to the normal enzyme to behave as a similar antigen,
but the protein had not retained its catalytic property. Furthermore,
the td mutant could elicit antitryptophan synthetase when injected
into rabbits. If the inferences drawn from these studies are correct,
classical genetics will be served at least to the extent of an elegant ex-
periment suggesting indirectly the idea of mutation as a change, not
a loss. Suskind and other workers are primarily interested in the
study of an allelic series as applied to questions of intragenic structure.
Adams (1942) and Markert and Owens (1954) have prepared antisera
against a tyrosinase preparation from the fungi, Psalliota campestris
and Glomerella, respectively. Antiserum for the tyrosinase of Psalliota
was inactive against tyrosinase from a related genus, Lactarius pi-
paratus, and antiserum for the tyrosinase of Glomerella was inactive
against tyrosinase preparations from Neurospora, Psalliota, Tenebrio,
and the vascular plant genus, Solanum. Therefore, the same enzyme
from different species, by serological criteria, may be somewhat differ-
ent. Novel applications of such serological methods are theoretically
possible, though perhaps impractical at this time. For example. Bird-
song, Alston, and Turner (1960), noting the absence of canavanine in
seeds of certain species in a genus in which canavanine occurs, suggest
that interspecific crosses of canavanine-less forms might yield a
canavanine-producing hybrid, disclosing latent pathways in much the
same way that complementary mutants in Neurospora are indicative
of metabolic blocks affecting different steps. But what if the species
are incompatible or even compatible but yield a canavanine-less
hybrid? Serological tests of the type described above might be applied
to disclose an enzymatically inert, but related protein. This result
would imply that canavanine synthesis was a lost property, and such
information would have definite taxonomic value.

Even now, it is apparent that the disclosure of homologous
genes by serological tests is becoming feasible (Nisselbaum, et al.,
1961). Stimpfling and Irwin (1960a) have recently reported a study
of gene homologies in species of the Columbidae (including doves and
pigeons). Through extensive previous genetic and serological inves-
tigations, it has been possible to demonstrate in these species a series
of species-specific antigens which segregate in backcross generations
as simple Mendelian characters. If one thus produces an antiserum
and adsorbs with appropriate mixtures of test sera, it is possible to
produce a single-antibody-containing antiserum which can then be

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS

used to test a number of related species. Gene homology is implied in
each case where such antiserum is active against the serum furnished
by the test species. One series of species-specific antigens, through
appropriate matings, has been defined as the product of genes occupy-
ing homologous loci in four different species (Stimpfling and Irwin,
1960b). The antigens behave as contrasting characters in backcross
hybrids and are considered to be products of genes that had a com-
mon origin but underwent subsequent change. They then constitute
a series of multiple alleles considered in this situation at the generic
level. Additional complexity of the locus, at least in the serological
expression, is indicated by the fact that, within a species, variants of
the species-specific allele occur. Furthermore, another antigen, in this
case from a different genus, has been shown to have some serological
affinities with the series discussed above, and it may represent another
allelic variant which, if confirmed, would elevate the character to the
family level. In this last case, the serological affinities of the allelic
series are greater among species of the same genus than with the
extrageneric related antigen.

We hardly know where enzyme studies will have arrived by
ten more years. The following discussion is indicative of the course of
future progress, and discloses the potential refinements of biochemical
systematics.

In recent years much progress has been made in the analysis
of the amino acid sequences within certain protein molecules. The
classic example is the work of Sanger on the insulin molecule. Applica-
tion of these techniques to genetics appears in the work of Ingram on
sickle cell anemia and altered hemoglobin. In a recent book by Anfin-
sen (1959), The Molecular Basis of Evolution, some of this work has
been summarized. A few of the pertinent facts follow:

(1) Insulin from five different species has been studied (beef,
pig, sheep, horse, and whale), and only insulins of pig and
whale were found to be identical.

(2) Adrenocorticotropic hormone (ACTH) of sheep, beef,
and pig has been examined, and that of pig diff'ers from
the other two.

(3) Sheep and beef ribonucleases differ.

(4) Vasopressin (with only eight amino acids) of beef has
arginine while that of hog has lysine.^

(5) Ferriporphyrin peptides from cytochrome C of pig, horse,
beef, and salmon are alike, but in chicken, serine replaces
alanine.

2 Addendum from A. C. Allison, 1959.

52 BIOCHEMICAL SYSTEMATICS

Even more remarkable than the facts hsted above are the
imphcations of recent work by Zuckerkandl et al. (1960) who have
utihzed trypsin lysis of hemoglobins of various animals and have then
examined the patterns of the derived peptide mixture by means of
combined electrophoresis and paper chromatography. Although it is
admitted that comparison of individual spots is limited by the
methods, the authors suggest that when two complex peptide pat-
terns are generally similar the probability is high that most of the
spots represent identical or highly similar sequences. Among several
primates studied the basic patterns were very similar; other mammals
showed less similarity to the primates than did primates to each
other; three fish patterns (bony fish, lungfish, and shark) showed few
similarities, and a cyclostome and Echiurid "worm" showed none.
The three fish patterns differed among themselves more than did the
mammals observed. Apparently most of the hemoglobin molecule has
been subject to the effects of gene mutations which have been retained
in the course of vertebrate evolution,^ and probably mutations affect-
ing the same peptide region have occurred repeatedly. The hetero-
geneity of these hemoglobins is remarkable in itself, but even beyond
this, it suggests that this type of comparative biochemical study may
be expected to make a major contribution in the not too distant future.
Enzyme heterogeneity is now well established, and a major conference
has already been devoted to the question of multiple molecular forms
of enzymes (Wroblewski, 1961). Introductory remarks at this con-
ference, held by the New York Academy of Sciences, by Gregory (1961)
reflect the current lively interest and realistic possibilities of studies
of the comparative biochemistry of enzymes:

It is apparent that enzyme heterogeneity is a common phenomenon.
More than 30 enzymes have been shown to exist in multiple forms
within individual organisms. They have been observed in both plants
and animals, in unicellular microorganisms as well as multicellular
species. They have been distinguished on the basis of a variety of
characteristics including electrophoretic and chromatographic be-
havior, serological specificity, differential solubility, and differential
response with coenzyme analogues. . . . The importance of the study
of multiple forms of enzymes stems in part from their frequent but by
no means universal occurrence. Their study promises to expand our
knowledge in a variety of fields ranging from embryology and the
study of evolution to physiology and pathology.*

3 For an interesting discussion of the evolution of hemoglobin and myoglobin, see
V. M. Ingram, "Gene evolution and the haemoglobins." Nature, 189: 704-708 (1961).

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS 53

Although the exact wording of the quoted paragraph empha-
sizes variation within individual organisms, variations among individ-
uals and among species occur. In the beginning, it is to be expected
that enzyme differences within individuals will comphcate taxonomic
appraisals of interspecific differences, but as the bases for such differ-
ences are better understood, the problem should be simpHfied. An
excellent illustration of the extent of variations in similar enzymes is
to be found in Fig. 4-3.

Directly related to the remarks made above are implications
from studies of the effect of partial degradation of enzymes upon
their activity. Selected examples from those summarized by Anfinsen
are the following:

(1) ACTH consists of 39 amino acid residues.

(a) With carboxy peptidase, three C- terminal residues may
be removed without loss of activity.

(b) With hmited pepsin digestion, eleven C-terminal resi-
dues may be removed without loss of activity.

(d) Loss of even one or two A'^-terminal residues results in
loss of activity.

(2) Papain consists of 180 amino acids.

(a) About eighty residues from the A/^-terminal end may
be removed without loss of activity.

(3) Ribonuclease consists of 124 residues. (Fig. 4-2.)

Figure 4-2 may be consulted to show the extensive modifica-
tions of ribonuclease which may be tolerated without loss of activity
of the enzyme. Note that it is at the extremes of the protein chain
wherein modification is permitted without loss of activity.

The discovery that enzymes may have a rather large number
of nonessential amino acids (that is, nonessential with respect to the
overt action of the enzyme) associated with them and also substitu-
tions within the essential parts at some points without loss of activity
supports the earlier implications of serological findings: that the same
enzyme from two species may differ. This fact, again, provides for a

^ Although enzyme heterogeneity is becoming recognized as commonplace, there
are also examples of enzymes from different sources which appear to be identical, at least
by serological criteria. For example, Fredrick (1961) has reported that a purified phos-
phorylase preparation from the bluegreen alga, Oscillatoria princeps, was serologically
active against other blue green algae. Yet, as noted, serological activity may not indicate
total similarity.

54 BIOCHEMICAL SYSTEMATICS

_ Activity retained on oxi-

, . u .'.„" 'ouc^ (Tm iun 'r,N mc jt^^^^ dation by performic acid

Ser ,n sheep ^lA LYS PHE GLU ARG GN HIS „eT ^ ,o ,he sulfone

^ ALA ^ A^P

TIlV^*^" /^ Initial split by subtilisin sER

A »• > , • A ^^ , ^FP 5^" ^^^ *^* ALA SER THR ••

Activity retained on guanidination StK

of amino groups AN

30 TYR 80 -,, ^-,

''' "" MET ,, CYS ^„ THR ILEU SER MET THR SER TYR SER ,,

^.,*«^; ^SE^^^^ ARG ^0 ^^

v*"; (LYS 90 ,SERP ^Pepsin inactivates ,*'^ LYS aLA^

^^YR \l20 ^™*^^"/^''^

^ GLY -/ iVALj

THR PRO VAL,^^^*^^"'' ''"^ "'^ VAL PRO VAL TYR PRO AN -Ar

fir X ryj '\'™J

Glu ^-^ l-YS AN CXPose removes to here A^'i^'^Y '"'t o" reoction 1^^-;

in sheep ASP ALA without inactivation with bro,.oacetate; activity ,,0 fGLU .J

^ , . ,wr 'ost on photooxidotion (?) ^ ""

^"%f TYR LYS THR THR AN ALA GN LYS,, HIS JLEU ILEU VAL ALA \^^^^ 60
40 LYS 100 ^ ^— - ^^^^^.^

> ALA

DInitrophenylation / °R0 g^

Inactivates ^ VAL „ yy^L

A"*" THR AIA ASP

^"'"'< PHE VAL HIS GLU SER LEU ALA

Fig. 4-2. Amino acid sequence of ribonuclease. Note the effect of
specific modifications upon activity. Courtesy C. B. Anfinsen; the
figure is a composite of results from the following: C. H. W. Hirs,
S. Moore and W. H. Stein, J. Biol. Chem., 235:633 (1960).
R. R. Redfield and C. B. Anfinsen, J. Biol. Chem., 385 (1956).
D. H. Spackman, W. H. Stein, and S. Moore, J. Biol. Chem.,
235:648 (1960). J. T. Potts, A. Berger, J. Cooke and C. B. Anfinsen,
J. Biol. Chem. (in press). C. Smyth, W. H. Stein and S. Moore,
J. Biol. Chem. (in press).

comparative biochemistry of enzymes at the molecular level. This
basis for potential enzyme heterogeneity, the extent of which is illus-
trated in Fig. 4-3, has been pointed out by Paul and Fottrell (1961).
Mutations that affect the amino acid sequence in nonessential
parts of the enzyme can in all likelihood be preserved and will thus
represent extremely subtle indices of relationship somewhat analogous
to the system of reciprocal translocations of Oenothera chromosomes
which have provided valuable insight into the phylogeny of this
genus. Progress in comparative enzymology is accelerating now, and
such investigations may play an increasingly important role in the
study of evolution. Recently, Esser et al (1960) studied twenty-five
reverted mutants, presumably back mutations, of the tryptophanless
(td2) mutant of Neurospora crassa. Reaction rates for the specific
reaction system governed by the gene differed among the reverted

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS

mutants, while reaction rates for six different wild strains were
similar. These results suggest that the reverted mutants are not
qualitatively identical and represent further evidence for enzyme
heterogeneity, although simultaneously they emphasize the com-
plexity of the problem of establishing the phylogenetic meaning of
those differences which are disclosed.

Another important question which is more apropos now than
ever before in the hght of advances in biochemical techniques is
raised by Anfinsen:

One of the major questions to be answered in arriving at a clear
understanding of the phylogenetic relationships between different
forms of life is whether there exist identical, or closely homologous,
genes in widely separated species, or whether similarities in pheno-
types are due to analogous genes which determine equivalent appear-
ance or function by different pathways.

Horse

Chicken

Lamb

Cow

Rabbit

Human

Guinea pig

Rat

Mouse

Fig. 4-3. Esterase zymograms of serum from different species.
From Paul and Fottrell, 1961, "Molecular Variation in Similar
Enzymes from Different Species." Ann. N. Y. Acad. Sci. 94:671

56 BIOCHEMICAL SYSTEMATICS

Actually, Anfinsen's question may be modified to apply to
enzymes, which are generally assumed to be direct or indirect template
products of genes, and further subdivided:

(1) Do widely separated organisms, which possess the ability
to synthesize a certain substance, employ the same
sequential order and precursor series, implying enzymatic
homology, or do they travel different roads to the same
destination?

(2) If two biochemical sequences are identical stepwise, to
what extent are the enzymes involved homologous and
thus identical or nearly so? The minor differences in, say,
ribonuclease, of different organisms would not seem to
suggest non-homology.

With respect to the second question, if only a small active
site has critical spatial arrangement, then independent evolution of an
enzyme might be expected to yield chemically different enzymes— un-
less new enzymes evolve by minor variations in a member of a pair of
repeats. This last idea is expressed in detail by Demerec and Hartman
(1956) following studies of non-random distribution of genes in-
volved in histidine and tryptophan synthesis in Salmonella. Non-
random gene distribution appears to be characteristic of amino acid
synthesis in Salmonella. More recently it has been reported that
4-threonine and 5-isoleucine-valine loci are clustered in an order
corresponding to the sequence of biochemical reactions they control
(Glanville and Demerec, 1960). Nonetheless the examples of non-
random gene distribution of which we are aware at present are found
in only a few organisms and may not represent a widespread phenom-
enon. In Horowitz' words (1950):

Biochemical mutants of Neurospora should provide excellent material
for study of the possibility of non-random gene distribution. At present,
all that can be said is that if such a distribution exists, it does not leap
to the eye.

The situation, at present, in Neurospora is essentially unchanged al-
though Wagner et al. (1958) have described a system involving valine
and isoleucine synthesis wherein the gene sequence seems to be cor-
related with a sequence of metabolic steps.

In connection with the question of comparative enzyme
chemistry, the following excerpt from a discussion at a recent sym-
posium (Haslewood, 1959) illustrates the trend of thought among bio-
chemists at present.

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS

Bloch: "On the other hand you see no difficulty in assuming entirely
separate pathways in the evolution of the specific bile acids?"

Haslewood: "No difficulty. If you were to tell me, as the result of
researches on protein, that the enzymes making cholic acid in the cod
are quite different substances from the enzymes making cholic acid in
man, I would not be at all surprised."

When the time arrives at which amino acid sequences of in-
dividual proteins can be efficiently analyzed, this procedure will
doubtlessly provide some of the answers to questions of enzyme
homology versus analogy. At present the procedure is complex and
tedious, and only a few laboratories are involved. Serological inves-
tigations may provide considerable circumstantial evidence, as has
been discussed, for if two similar enzymes behave as a single antigen,
they are best considered to be homologous. It is possible, of course,
but not highly probable, that the enzymes have evolved independ-
ently as serologically identical molecules. There are some interesting
possibilities for the study of enzyme systems which appear to be
definitely non-homologous. For example, in the squid eye the pros-
thetic group of the visual pigment is described as neo-b-retinine
(Hubbard and St. George, 1958), similar to the pigment of the ver-
tebrate eye. Since the squid eye and vertebrate eye are generally re-
garded as one of the classic examples of convergent evolution in
structure, the precise molecular configuration of the "enzymes" in-
volved in the remarkable correlated biochemical parallelism is of
interest. In another similar case Johnson et al. (1960) reported an
interacting luciferin-luciferase system between a crustacean {Cypri-
dina) and a fish {Apogon). There is reason to believe that the
biochemical mechanisms of phosphorescence are similar in the two
species and represent another example of convergent biochemical
evolution. Surprisingly, the authors seem to interpret the discovery
somewhat differently, however, "although the similarities in the
luminescent systems of a fish and crustacean could represent a rare,
evolutionary coincidence, they as likely indicate that more of a thread
of unity exists in the comparative biochemistry of luminescence
among diverse types of organisms than has been hitherto supposed."
Anyway, serological comparisons of luciferase from the two sources
would be of interest as they represent potentially analogous enzymes.

Dessauer et al. (1962) have compared certain iron-binding
proteins (transferrins) of 150 reptiles and amphibians by electro-
phoresis. Large differences in migration rates were observed; in some
instances there was considerable intraspecific variation, and also
multiple transferrins were often found. In some cases the transferrins

BIOCHEMICAL SYSTEMATICS

were quite constant and similar among related species. Dessauer
speculated that variation in the transferrin pattern might be greater
in species in a more active phase of evolution. The results in general,
while they raised a number of questions, indicated that the iron-
binding proteins might be of considerable value in direct systematic
comparisons or in population studies when intraspecific variation is
encountered.

Another method of studying comparative enzymology which
appears to be very promising has been described recently by Kaplan
et al. (1960), and Kaplan and Ciotti (1961). This technique involves a
comparison of the catalytic properties of selected enzymes. Several
related methods have been utilized by Kaplan's group. For example,
they have shown that certain diphospho-pyridine nucleotidases of
ruminants (for example, goat, beef, lamb, deer) are inhibited strongly
by isonicotinic acid hydrazide while those of a number of other
mammalian groups, as well as the frog, are relatively insensitive. This
implies a distinctiveness in these enzymes in one related group of
mammals which is systematically significant.

In the work reported by Kaplan et al. (1960) reaction rates
were compared at high and low substrate concentration using the
"same" enzyme from a relatively wide assortment of vertebrates. The
specific enzyme reported on was lactic acid dehydrogenase, using both
lactate and pyruvate as substrates. In addition to the normal diphos-
phopyridine nucleotide (DPN) cof actor they prepared specific analogs
of the pyridine ring of DPN such as acetyl pyridine and thionicotina-
mide. These cofactors participated in the reaction either as electron
donors (with pyruvate) or acceptors (with lactate). Table 4-1 presents
some of Kaplan's results. The values reported could be duplicated,
according to the authors, within a few per cent when a number of dif-
ferent individuals of the same species were analyzed.

One notable feature of the data from Table 4-1 is the fact
that ratios for heart muscle and for skeletal muscle of the same
species consistently differed. Also there were outstanding differences
in the ratios of flounder, sole, and halibut (all flatfishes) as opposed to
the other animals, including a number of other fishes. Differences,
though somewhat less marked, were typical between the enzymes of
most of the species examined, giving the impression that, if these re-
action rate differences truly reflected enzyme structural differences,
all of the animals possessed different enzymes. The authors were
conservative, however, and did not stress the smaller differences in
ratio. A few additional analyses were carried out with invertebrates,
and extremely wide differences were observed, notably a greatly en-
hanced affinity of the enzyme with the pyridine analog, acetyl
pyridine, in crustaceans.

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS

Table 4-1. The ratio of the reaction rates for lactic dehydrogenases from heart
and muscle of different animals, with high and with low concentrations of
pyruvate or lactate. Kaplan et al. (1960). Reprinted from Science by permission.

DPNH//,

DPNHi

APDPNz.

TNDPN,.

Animal

Heart

Skeletal
muscle

Heart

Skeletal
muscle

Man

0.4

0.7

1.8

Mouse

0.5

0.8

1.2

3.0

Rat

0.4

0.8

0.7

2.4

Guinea pig

0.4

0.8

0.9

2.8

Rabbit

0.3

1.1

0.4

5.1

Beef

0.5

1.1

2.9

Pig

0.5

0.8

0.7

2.4

Lamb

0.4

1.2

0.7

3.5

Pigeon

0.2

0.7

0.8

3.0

Chicken

0.5

0.9

1.1

4.4

Bullfrog

0.7

0.8

0.7

4.9

Grass frog

0.4

0.7

0.5

4.5

Salamander

0.9

1.3

1.6

4.1

Box turtle

0.6

0.7

2.3

4.8

Painted turtle

0.8

1.0

2.8

9.0

Herring

1.2

1.9

2.9

11.6

Mackerel

0.9

3.2

0.8

11.5

Flounder

2.0

1.9

Sole

2.1

3.2

Halibut

1.9

Sea bass

0.9

1.9

0.9

5.6

Butterfish

0.8

1.7

0.8

10.6

Scup

0.6

1.4

0.6

4.9

Sea robin

1.1

1.4

1.9

8.5

Puffer

1.2

1.3

7.4

14.0

Toadfish

0.9

1.3

1.4

9.5

Suckerfish

0.5

1.0

0.9

6.0

Dogfish

0.4

1.1

1.2

8.0

' Not available.

Kaplan et al. state, in summary that:

The data presented indicate that it is possible to classify animals not
only by their physiological and morphological characteristics but also
by their enzymatic properties, and they also suggest that change in
enzyme structure may have been of significance in the establishment
of new species.

Boser and Pawelke (1961) have discovered that there are two
mahc dehydrogenases in potato. This finding is pertinent to Kaplan's

^0 BIOCHEMICAL SYSTEMATICS

work in that variations in the relative amounts of two enzymes could
yield differences such as reported by Kaplan, among different species
or within an individual. In this case, no qualitative difference in the
enzymes is required and the implications of the results would differ.

Somewhat related to the work of Kaplan is that of Blagove-
shchenskii (1955) who has emphasized in his writings the fact that the
same enzymes from different organisms exhibit different activation
energy thresholds and particularly that more advanced organisms
have reduced in general the activation energy required for a partic-
ular enzymatic process (for example, legume catalase has lower
activation energy requirement than does bacterial catalase) implying
greater enzyme efficiency in the more advanced species.

So far the great achievements in biochemistry have been in-
tegrative and unifying in their influence. The metabohc similarities
of all organisms, from bacteriophage (in their hmited metabolic
abilities), to higher plants, to man are emphasized. Examples of this
are so well known that it is no longer necessary to cite them. There is
aheady evident a turn of the tide, a focus upon minor category dif-
ferences in biochemistry. This thinking is expressed succinctly by
H. C. Crick (1958):

Biologists should realize that before long we shall have a subject which
might be called "protein taxonomy"— the study of the amino acid
sequences of the proteins of an organism and the comparison of them
between species. It can be argued that these sequences are the most
dehcate expression possible of the phenotype of an organism and that
vast amounts of evolutionary information may be hidden away within
them.

It is unfortunate that anyone should equate biochemical sys-
tematics merely with a survey of the distribution of a given chemical
entity. Novel situations providing challenge and reward for ingenuity
and perception abound in the area of biochemical systematics as in all
areas of biology. An example of an ingenious use of biochemical data
in Drosophila systematics is the work of Hubby and Throckmorton
(1960). It has been shown that more primitive Drosophila species
produce red pteridine pigments in various parts of the body while the
more advanced forms have the distribution of such pigments re-
stricted to the eyes. In the primitive forms, the red pigments are
present in the testes of males. The red pteridine pigments of the
testes are identical with the eye pigments, and the relationship sug-
gests an evolutionary change in pteridine metabolism in Drosophila
so as to restrict pteridine accumulation to the eyes where presumably
functional significance may be attributed to the pigment. Hubby and

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS

Throckmorton studied the pteridines of 156 species representing five
sub-genera. Estimations of pteridine content were made visually from
paper chromatograms of extracts. Their results indicated that primi-
tive species from each sub-genus contained greater amounts of
several pteridines in the body than did the more highly evolved
forms. Even some of the colorless pteridines were reduced in amount
in certain highly evolved forms. Notably, drosopterine and sepiapteri-
dine have been almost eliminated in the testes of advanced forms of
each evolutionary line, and the authors suggest that there is evidence
that different mechanisms have arisen to bring about this decrease
among the various evolutionary lines. It is obvious that this system
offers an additional valuable key to Drosophila phylogeny with no
indication presented that it contradicts or challenges the system
established from genetic and cytological or morphological criteria.

Another interesting approach to comparative biochemistry
with an essentially phylogenetic focus is represented by the work
of Vogel (1959a, 1959b, 1960, 1961) on lysine synthesis. It has been
known for some time (see Wagner and Mitchell, 1955, p. 203) that the
biosynthesis of lysine proceeds via two different pathways involving
either a-aminoadipic acid (in Neurospora) or diaminopimelic acid (in
E. coli).

COOH

CH2
CH2

Neurospora

CH2

CHNH2

COOH

oc -aminoadipic acid

CH2NH2

V CH2

X CH2
CH2

COOH

y^ CHNH2

CHNH2

^ COOH

CH2
CH2

E. coli y^

lysine

CH2

CHNH2

COOH

diaminopimelic acid

Vogel extended this knowledge to numerous plants represent-
ing various major taxonomic groups including bacteria, algae, fungi,
and vascular plants. The results of his investigations are summarized
in the table below (Table 4-2). Vogel utilized a technique involving
radioactive tracers in which the labelling pattern of lysine was indica-
tive of the pathway by which it was formed.

Conclusions from these data are that, by the criterion of com-

BIOCHEMICAL SYSTEMATICS

Table 4-2. Taxonomic differences in the synthesis of lysine.

Pathway Suggested

DIAMINOPIMELIC ACID

a-AMINOADIPIC ACID

Bacteria

Algae

Bacillus subtilis

Euglena gracilis

E. coli

Algae

Chlorella vulgaris

Fungi

Saprolegnia ferax

Allomyces macrogynus

S. parasitica

Rhizopus stolonifer

Achlya bisexualis

Mucor hiemalis

A. americana

Cunninghamella blakesleeana

Hypochytrium catenoides

Candida subtilis

Neurospora sp.

All other ascomycetes and basidiomy-

cetes studied.

3 species of chytrids

Higher plants

fern (Azolla Carolina)

duckweed (Lemna minor)

pollen tissue (Ginkgo biloba)

leaf parenchyma (Agave toumeyana)

habituated root tissue (Melilotus

officinalis)

petiole crown gall (Helianthus annuus)

parative lysine synthesis alone, bacteria, some algae, and higher plants
show a closer relationship to each other than to the majority of fungi
and Euglena. However, among the fungi, the Saprolegniales and also
Hypochytrium catenoides are atypical in that they utilize the di-
aminopimelic acid pathway. Such data are clearly of phylogenetic
interest, though it is obvious that at present it is uncertain as to how
much weight must be given this evidence.

That higher plants do synthesize a-aminoadipic acid is evi-
dent from work by Grobbelaar and Steward (1955) who found that
this acid became radioactive after C^* lysine was fed to Phaseolus.
Fowden also (personal communication) has noted that a-aminoadipic
acid is frequently encountered as a minor or trace component of
many plants and that C^^ lysine and tritiated pipecolic acid give rise
to radioactive a-aminoadipic acid in Acacia. How these facts will in-
fluence, ultimately, the assessment of the phylogenetic implications of

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS 63

Vogel's work remains to be seen. Obviously absolute metabolic distinc-
tions represent the most satisfactory types of criteria because they
imply the emergence of a new synthetic ability or independent origins
of alternative biosynthetic routes. When both systems are potentially
available (as the formation of a-aminoadipic acid from lysine implies),
the demonstration of a selection favoring one pathway over the other
becomes the systematic criterion.

A final example illustrating the varied approaches to bio-
chemical systematics is taken from the field of zoology. It is derived
from a Harvey Lecture by Wald (1947) entitled "The Chemical
Evolution of Vision." Some of the major points are summarized below:

It had been noted that extracts of fish retinas were purple-
colored while those of mammals, birds, and frogs rose-colored. Sub-
sequently, marine fish were found to yield extracts colored Hke those
of mammals, birds, and frogs. The purple pigment (characteristic of
freshwater fish) was found to be closely related but distinct from
the rhodopsin of the second group and was named porphyropsin. The
essential biochemistry of rod vision was strictly analogous, in the por-
phyropsin system, to the rhodopsin-retinene-vitamin A system, but
biochemical studies suggested minor differences in the carotenoid
moiety of the chromoprotein. Subsequently this difference has been
found to reside in the ring of the retinene portion wherein, in the
porphyropsin system, one additional double bond is present (Wald,
1960).

It is noteworthy that such a division between marine and
freshwater fish receives no support from their taxonomy. Indeed,
investigation of euryhaline fish showed that the visual system was
related to their spawning environment.

(1) Anadromous types (spawning in freshwater) have the
porphyropsin system.

(2) Catadromous types (spawning in salt water) have the
rhodopsin system. Some yield mixtures of the two pigments.

Wald noted that it is with evolutionary migrations between
freshwater and the sea that the patterns are associated. Since it is
commonly believed that freshwater fishes provided ancestors of the
amphibia, Wald investigated other pertinent species.

(1) The sea lamprey (spawning in freshwater, however), a
primitive vertebrate type, has the porphyropsin system.

(2) The newt has porphyropsin.

(3) The bullfrog {Rana catesbiana) provides the most signif-

BIOCHEMICAL SYSTEMATICS

Fig. 4-4. Transfer from the porphyropsin to the rhodopsin system
during metamorphosis of the bullfrog. Antimony chloride tests
with extracts of bleached retinas from tadpoles approaching meta-
morphosis, and from newly emerged frogs. The tadpole retina
contains a high concentration of vitamin A2 with only a trace of
Ai, the frog retina just the reverse pattern (Wald, 1945-46).
Reprinted from the Harvey Lectures by permission of the Academic
Press, Inc.

icant information, for tadpoles have mostly porphyropsin
with only a trace of rhodopsin, while the frogs have mostly
rhodopsin. Intermediate stages exhibit mixtures (Fig. 4-4).

Wald says, "It is difficult to view this [latter] phenomenon
otherwise than as a recapitulation." This work is of unusual interest
to biochemical systematics. In the first place, at the major systematic
category level it provides modest further support for the argument
that freshwater fishes are the most primitive and that they were the
progenitors of amphibians.

However, at the lower systematic category level it provides

INTRODUCTION TO BIOCHEMICAL SYSTEMATICS 65

spurious biochemical evidence. That is, all marine fishes are linked to
the rhodopsin system while freshwater fishes are linked to the por-
phyropsin system. Yet, certain groups of freshwater and saltwater
fishes are obviously closely related by every other criterion.

The interpretation thus derived might better be that, for
some unknown reason, there is rather strong selection pressure for the
rhodopsin system in marine or terrestrial habitats, so that it evolved
independently with each evolutionary emergence of a group from
freshwater. There is no obvious reason why there should be strong
selection for the rhodopsin system under such circumstances. In
Wald's words, "there is an order here that goes with the ecology, but
with the genetically determined rather than the causal ecology."

Subsequent work has supported the original data except that
a few marine fishes are now known to utilize the porphyropsin sys-
tem and not all frogs exhibit the conversion of the porphyropsin to
the rhodopsin system associated with metamorphosis (Wald, 1960).

The lesson which may be learned from this is that data which
provide valid support for a systematic interpretation at one level may
be simultaneously misleading at another level. In this work one gains
the impression that strong selection pressure may be present, when
unexpected on a priori grounds. One must, therefore, be cognizant of
cryptic selection pressure which could produce a biochemical correla-
tion which might be deceptive in its implications.

The foregoing discussion has touched briefly upon several
facets of biochemical systematics. The authors hope that some of the
ideas expressed serve to indicate the need for an enlarged perspective
from which to view the field. The scope of the subject greatly exceeds
the somewhat sterile cataloging of compounds and their host species.
There is scarcely any doubt that this broad field offers a tremendous
potential to systematics. Its past near-neglect has stemmed almost
certainly from limitations of technique, but instrumentation is ad-
vancing at an incredible pace, and techniques are now commonplace
that were totally unavailable even ten years ago.

SEROLOGY

AND SYSTEMATICS

Intensive serological investigations preceded sus-
tained or general interest in other biochemical
approaches to systematics. Since serology is, in
apphcation and methodology, fundamentally differ-
ent from those approaches (for example, studies of
specifically known chemical entities), treatment of
serology precedes that of particular "natural"
classes of compounds in subsequent chapters.

Although some interest in biochemical
applications to systematics presumably developed
as a result of the work of Abbott more than seventy-
five years ago (1886, 1887a, 1887b), and indeed
sporadically the subject was introduced even earlier,
as indicated in the preceding chapter, only a hand-
ful of important contributions appeared prior to the

68 BIOCHEMICAL SYSTEMATICS

classic work of Baker and Smith (1920) on the terpenoids of Eucalyp-
tus. However, as early as 1901 Nuttall published his significant work
on the use of essentially serological methods in establishing species
relationships. These serological methods were, in turn, adopted by
numerous workers and extended to include a wide variety of organ-
isms, both plant and animal, over the succeeding several decades. In
its period of greatest emphasis (that is, during the period of 1920-
1930), the serological approach received mixed reactions. Some inves-
tigators embraced this development as a panacea which, almost alone,
would provide a completely objective approach to systematics gen-
erally. One prominent group of plant serologists emerged at Konigs-
berg, Germany, following the initial investigations of Gohlke in 1913,
Later, Mez was the dominant figure in the Konigsberg group. The
Konigsberg work culminated in the development of the much de-
bated, but now often overlooked, "Serodiagnostiche Stammbaum"
(Fig. 2-12) purporting to show a phylogenetic tree derived almost
entirely from comparative serological investigations (Mez and Ziegen-
speck, 1926). The serological data evoked in some quarters a con-
siderable amount of skepticism and in fact some severe criticism.
Most skepticism, as might have been expected, came from the
classical morphological systematists while violent emotional criticism
of Mez's contributions came surprisingly from other serological
workers, such as the Berlin group represented by Gilg and Schurhoff
(1927) who stated, "the serodiagnostic method is, for investigation of
plant relationships, completely useless."

The controversy between the Berlin serologists and the Mez
group at Konigsberg was discussed by Chester (1937) in a series
of three general reviews of plant serology. These papers were master-
fully written, and they represent a classic summary of the early period
of plant serological investigations. The present authors are indebted
to Chester's review for much of the information on basic methodology
presented in the succeeding pages. It is ironic that at about the time
the Chester review appeared, interest and activity in plant serology
waned. Plant serological investigations have subsequently revived
somewhat, in Germany in the work of Moritz and in America by
Johnson and Fairbrothers. The revival of interest in America in
plant serology represents an offshoot from the animal serological
systematic studies of Boy den and co-workers, begun in 1925 and
continuing at Rutgers University. In the following paragraphs selected
examples will be drawn from zoological studies when they illustrate,
particularly well, a certain principle. In general, however, botanical
studies will be emphasized.

It is doubtlessly recognized, by even the general reader, that

SEROLOGY AND SYSTEMATICS 69

serology concerns essentially antigen-antibody responses. That is,
certain foreign substances (called antigens and formerly regarded
as proteins though now it is recognized that other substances than
protein may be antigenic), when injected into a host, may elicit
the formation, in the host, of other substances (called antibodies,
likewise generally regarded as proteinaceous) which may agglutinate
or otherwise affect the foreign substance. Various species of domes-
tic animals may serve as the host although rabbits are most fre-
quently used.

Chester has listed the types of reactions which were utilized
up to 1937. Since his review, the first method to be discussed
below, the precipitin reaction, has become the most widely used.
The precipitin reaction is probably the simplest of the various
methods of evaluating antigen-antibody reactions. In this method,
one mixes aliquots of the antigen in varying dilutions with the anti-
body preparation (antiserum); this mixture produces an amount of
precipitate corresponding to the "strength" of the reaction, and
the precipitate is appropriately measured. In addition to the precipitin
reaction various reactions classified as anaphylaxis reactions have
been utilized. In principle these last methods involve sensitizing a
host, then later injecting into the host a second dose of antigen
preparation. The second injection may induce some physiological
response such as inflammation or spasms. An interesting modification
of the anaphylaxis reaction is known as the Schultz-Dale technique.
A sensitized virgin female guinea pig is killed, and the uterus removed,
placed in Ringer's solution, and attached to a kymograph. The antigen
preparation is added directly to the uterus, and the degree of uterine
contraction is measured on the kymograph. A third type of reaction
involves the destruction (or agglutination) of particulate antigen
carriers such as bacteria, blood cells, pollen or other unicellular bodies
by antisera from sensitized hosts. Complement fixation, a fourth type
of reaction, utilizes the knowledge that a non-specific, heat labile
substance (complement) which participates, essentially, in the antigen-
antibody interaction, is used up in the process. Therefore, in principle
one measures the presence or absence of residual complement, follow-
ing the exposure of the antiserum to an unknown antigen preparation.
Residual complement is measured by comparing the efficiency of a
second reaction to a standard antigen preparation, for example sheep
blood cells. Complement fixation would be suitable as an indirect
indicator of an interaction which could not be followed visually.
In another type of reaction, the Aberhalden reaction, the serum and
the antiserum used to test it are mixed in a dialysis membrane.
Subsequently one tests the external medium with ninhydrin for

70 BIOCHEMICAL SYSTEMATICS

dialyzable cleavage products. Some other rather uncommon reactions
have been utihzed but so infrequently as to render them insignificant
for present purposes.

As in any other biochemical approach, the validity of serologi-
cal data depends directly upon the reliability of the techniques
utilized. From the earliest investigations strong support for particular
innovations of technique has been the rule, and often the attitude has
been taken that other techniques, usually equally vigorously sup-
ported by their adherents, are, nevertheless, almost completely worth-
less. Controversy over technical procedure was particularly rife be-
tween the Konigsberg and Berlin investigators, and it seems that they
hardly agreed on anything. Subsequent improvements in technique,
to be discussed later, indicate that the controversy could only disclose
which group's technique represented the greater imperfection. Since
the question of technique in serology is exceptionally relevant to
a reasonably objective appraisal of the method itself, some details will
be included below.

In botanical serology, seeds are most frequently used as a source
of antigen. These may be ground in a mortar and pre-extracted
with some non-polar solvent such as petroleum ether to remove lipids.
The ground material may also be extracted with ethanol. The protein
is finally extracted, most often with physiological saline, in proportions
of about 100 ml per 10 gms of tissue. Extraction time is controlled, of
course, and may represent several hours, or overnight. Sometimes
expressed sap is used directly.

Considerable disagreement arose among earlier investigators
as to whether individual plants were serologically homogeneous or
whether different organs or even tissues from the same plant had
different antigenic complements. Mez believed that plants were homo-
geneous, but the Berlin group disagreed. Chester noted, however, that
comparisons between seed proteins and other plant parts were particu-
larly distinctive and supported the Berlin viewpoint. Quite recently,
strong evidence for antigenic heterogeneity has been adduced by
Kloz, et al. (1960); this evidence will be presented in detail later,
following some additional discussion of present methodology. If there
is significant adaptive enzyme formation during development of higher
plants, as a priori considerations and precedent from microorganisms
suggest, then antigenic heterogeneity may be expected. Furthermore,
present work on multiple enzymes, discussed earlier (Chapter 4),
suggests strongly that large differences in antigenic composition may
be expected within an organism.

Another important question of technique relates to the pro-
tein concentration of a particular plant extract. For example, if a

SEROLOGY AND SYSTEMATICS J]

protein extract from one plant is twice as concentrated as that of
another, should they be adjusted to a standard concentration for
vahd comparison? Dissenters would note that it has not been estab-
lished that there is a necessary correlation between total protein and
antigenic activity. It has been suggested that a constant ratio of
tissue to solvent is preferable.

Injection of the extract into the host is intraperitoneally,
intracutaneously, or intravenously. A typical inoculation schedule
might be 5 cc doses administered at three to four day intervals with a
total of five to eight injections followed by a nine to ten day rest
before bleeding (Chester, 1937). It may be noted that individual
differences in the reactivity of different host animals, while reduced
by careful breeding, can never be entirely ehminated. Consequently,
some of the differences in serological reactions must represent
variations in host reactivity. This factor is undoubtedly taken into
account by workers in serology but is not often expressed. The com-
plications stemming from the requirement of a supply of host animals
have probably deterred many botanists otherwise receptive to
serological investigations. If the "Kunstsera" (artificial serum from
beef) reported by Mez had proven as reliable in the hands of other
investigators as claimed by its developer, we might have witnessed a
dramatic adoption of the serological approach.

In the earlier serological investigations, there were two differ-
ent methods of reading the precipitin reactions. As usual, one was
favored in Berlin and one in Konigsberg. The first of these, the
"flocculation test" was utilized by the Mez group in Konigsberg. In this
technique a carefully diluted antigen solution was mixed in a standard
sized test tube with an aliquot of undiluted antiserum. The mixture
was shaken, incubated for a standard time, and the height of the
precipitate which had, in the interim, flocculated, was read. The
observer, by design, did not know the identity of the serum being
tested. The second method was called the "ring test." In this test the
denser liquid was added to a test tube and the less dense liquid
pipetted carefully onto its surface. Without disturbing the layers, the
tube was incubated under standard conditions and the width of the
ring of precipitate measured. The ring test is not used frequently
at present, but Lewis (1952) has used this test in studies of the sero-
logical manifestations of pollen incompatibility factors.

A final commentary on the rather unfortunate controversy
between the Konigsberg and Berlin groups may be appropriate at this
point before passing to the post-Chester period in plant serology.
Chester felt that the controversy was to a considerable degree respon-
sible for the failure of systematists to become receptive to the

72 BIOCHEMICAL SYSTEMATICS

serological approach. In any event the serological data were in general
ignored by the majority of systematists, though von Wettstein is said
to have regarded serological data as useful, within limits, in phylogeny.
Chester offers a quotation from the Swedish systematist Heintze as a
reflection of the opinion of many systematists:

Serodiagnostic investigations have hardly contributed to a clearing up
of the relationships within the Cormophytes. By and large they only
"confirm" the errors of Engler and Prantl.

With the passage of time details of the Mez "Stammbaum"
have faded from the memory of most of the relatively few people who
saw it. The illustration was copyrighted, and aside from its appearance
in the original article, has been published only rarely. It is difficult to
believe that the "Stammbaum" possesses much validity in view of the
current recognition of certain limitations of the early serological
methods. Yet, when higher taxonomic categories are compared sero-
logically, correspondingly, the sensitivity of the method may not need
to be as great to provide clues to relationships.

Boyden (1942), in an important general review of serology
and systematics, discussed some of the methodological innovations in
use at that time, particularly the "photronrefiectometer," which is
essentially a modified densitometer. The Rutgers serologists (botanical
as well as zoological) are now using mostly densitometric measurements
of the precipitin reaction (Boyden and De Falco, 1943), but Moritz, in
Germany, is using a micromethod called micronephelometry in which
a beam of light passes through a microscope slide, containing the test
solution, mounted on a microscope. A photocell is attached to the
ocular position, and light reduction resulting from turbidity is recorded
through the photocell and an ammeter (Moritz, 1960). An interesting
point brought out by Boyden concerns the phenomenon of optimal
proportions. Briefly, it has been established that the amount of pre-
cipitate obtained with constant amounts of antiserum and increasing
dilutions of antigen rises from zero to a maximum then falls off again
to zero with considerable excess of antibody. As a result of this
phenomenon (for which several hypothetical explanations exist), one
must compare interactions over a series of dilutions. The optimal
proportion for different antigen preparations may vary significantly,
as indicated in Fig. 5-1. Since in most of the early serological work only
one proportion of antigen and antibody preparations was utilized, it
is obvious that the reliability of the method was accordingly lessened.
This disclosure cannot help but reduce the value of much of the early
serological work including, of course, that of Mez.

SEROLOGY AND SYSTEMATICS

LEGEND:

1. Woodland caribou

100.0

2. Woodland deer

71.3

3. Elk

50.7

4. Bison

43.1

5. Pig

5.4

5 6 7 8 9 10 11 12
Antigen Dilution

Fig. 5-1. Illustration of the concept of optimal proportions in the precipitin re-
action. Note that the greatest amount of interaction occurs at different antigen
dilutions for different species. Thus the total area under the curve represents,
more accurately, the degree of interaction (Gemeroy, Boyden & De Falco, 1955).

The technique recommended by Boyden requires a measure-
ment of the total area under the curve obtained by measuring the re-
action with various antigen dilutions. Relationships are expressed as
the per cent area of a particular heterologous reaction compared to
the homologous reaction (in a homologous reaction, the antiserum is
matched with its original inducing antigens, and the amount of re-
action, or curve area, is denoted as 100 per cent). The higher the per-
centage of reaction obtained with a particular heterologous reaction
the closer would be the presumed serological affinity.

Despite the improvements in technique such as described
above, some workers question the validity of methods based on
strictly quantitative reactions. Gell et al. (1960) have pointed out, for
example, that when species A gives more precipitate in a heterologous
reaction with B than it does with C, this may reflect varying amounts
of a single protein which is abundant in B and not in C. How-
ever, species A and C may contain several common, non-cross-reacting
substances and might reasonably be regarded as more closely related
species, although the serological method utilized would obscure this
relationship. The actual extent to which such theoretical objec-
tions, in practice, detract from the validity of serological data cannot
easily be ascertained, but it does appear that such complications are
sufficiently probable that every effort should be made to come to
grips with some of the more fundamental aspects of the method.

74 BIOCHEMICAL SYSTEMATICS

Systematic applications of serology require more insight into the
molecular basis of the phenomenon than now exists, yet there is no
clear indication that this facet of the problem is being aggressively
explored except by biochemists. No vigorous analytical treatment of
the problem of the qualitative aspects of the reaction appears from
the literature of systematic serology. The critical question of precisely
what is being measured has not been faced. For example, what are the
relative contributions of structural proteins, storage proteins, enzy-
matic proteins and non-proteins to the total reaction? Most proteins
which have been tested are antigenic, but their effectiveness varies.
What bearing, for example, does the disclosure that one may convert
gelatin from a weak antigen into a potent antigen by the attachment
of tyrosine, tryptophan or phenylalanine peptides (Sela and Arnon,
1960) have upon the question?

Granted that it is perhaps not required that systematic
serologists establish, themselves, the precise molecular dynamics of
the reaction, there remains nevertheless an obligation to attempt to
establish some parameters with respect to the presumptive validity
of the method through experimentation. Several possibilities are
apparent. For example, it should not be difficult to obtain genetic
stocks of different highly homozygous lines of intensively investigated
species, such as maize. Hybrids could then be produced and back-
crossed, in successive generations, to each parental type. By this
method one could obtain, empirically, a graded series of genetic types
from one parental extreme to the other. If a parental "standard" is
utilized to prepare an antiserum, then the remaining lines would, if
the technique is valid, be predicted to yield heterologous reactions of
decreasing amount. Should a linear trend in the serological results
appear, it would provide convincing corroboration of the method.
Unfortunately, this type of experiment has not apparently been done.
In hybrid populations, where morphological hybrid indices (and even
biochemical indices in the case of Baptisia) are available, the serolog-
ical data could be correlated with data of these other types. Unless
experiments of these types are conducted, one cannot accept the
taxonomic implications of serological data without considerable reser-
vations. Actually, Moritz has undertaken a number of serological in-
vestigations of hybrids (for example, Moritz and von Berg, 1931), in
which pre-adsorption with appropriate parental serum was utilized,
among other devices, in establishing the hybrid nature of a putative
hybrid. That is, if a putative hybrid is used to produce an antiserum,
and following pre-adsorption with a serum from one and then the
other parent, no reaction occurs with the homologous serum, it is
concluded that the hybrid has no antigens not present in one or the

SEROLOGY AND SYSTEMATICS 75

other parent and is truly a hybrid. This work will be described more
fully in Chapter 15 (Hybrid Studies).

In the older literature, there are some serological investiga-
tions which have been correlated with other criteria but not in a man-
ner which is wholly capable of resolving questions such as those
raised above. For example, Baldwin et al. (1927) studied the serolog-
ical interrelationships of a number of cross-inoculation groups of
legumes. Cross-inoculation groups are groups of species within which
certain nodule-inhabiting strains of bacteria may be cross-inoculated.
Generally, it may be assumed that the species belonging to a partic-
ular cross-inoculation group are rather closely related to each other.
In some cases a particular cross-inoculation group may also be affected
similarly by some pathogen; for example, in the cowpea group several
genera are susceptible to the bacterial spot disease. Using a variety of
reactions (including precipitin reactions and the Schultz-Dale tech-
nique) these investigators found that in general the serological re-
sponses were in agreement with cross-inoculation grouping. In the same
paper a summary of previous serological investigations of some genera
indicated agreement with the cross-inoculation group disposition of
the genera, and these investigations offered some support for the
validity of the serological method. A final precaution was taken to
establish that the reaction involved host-plant antigens rather than
bacterial antigens. Legume-seed antisera did not agglutinate the
corresponding nodule inhabiting bacteria.

Recent botanical investigations in systematic serology at
Rutgers University begin with the work of Baum (1954) on the
Cucurbitaceae and by Johnson (1953, 1954) on the Magnoliaceae.
Subsequently, Hammond (1955a, 1955b) reported on serological in-
vestigations in the Solanaceae and Ranunculaceae, and Fairbrothers
(Fairbrothers and Johnson, 1959; Fairbrothers and Bouletta, 1960)
has investigated some grasses and certain species of the Umbel-
liferae. Since these investigations are quite similar in methodology
and approach, and in general do not introduce highly controversial
interpretations, only the Johnson work and the Hammond work on
the Ranunculaceae will be discussed.

In Johnson's serological investigation of the Magnoliaceae
several genera were compared with Magnolia and then several species
of Magnolia were compared to establish intrageneric serological re-
lationships. The serological data are shown in Table 5-1. It has been
noted that the precipitin reaction was, in the Rutgers laboratory,
derived from calculating the area under the curve of photronreflec-
tometer readings at various antigen dilutions. At the generic level
the contention that Magnolia, Michelia, and Talauma form a natural

BIOCHEMICAL SYSTEMATICS

Table 5-1. Results of titrating, with the photronreflectometer, antisera against
four species of Magnolia with antigens from several species of the genus (Johnson,
1953). By permission of the Serological Museum.

Het.

Homolo-

Heter-

Area

Antiserum

Species

gous

ologous

Homol.

Area

Area
per cent

M-14-la(l + 1)

Magnolia kobus DC.

353

M. acuminata L.

279

79.2

M. tripetala L.

244

69.2

M. virginiana L.

240

68.1

M. portoricensis Bello

129

36.6

#3(1 + 1)

Magnolia tripetala L.

332

M. obovata Thunb.

308

92.7

M. kobus DC.

250

75.3

M. virginiana L.

240

72.3

M. acuminata L.

214

64.5

M. portoricensis Bello

178

53.6

9-1(1 + 2)

Magnolia portoricensis Bello

406

M. obovata Thunb.

274

67.5

M. tripetala L.

267

65.8

R-l(l + 1)

Magnolia obovata Thunb.

327

M. tripetala L.

304

M. portoricensis Bello

202.5

61.9

group is supported by the data, while Liriodendron is relatively
distant, serologically, and Illicium, which has been removed from
the Magnoliaceae by some investigators on anatomical grounds,
gave no reaction. Illicium was shown to produce a highly reactive
antiserum, when tested against homologous serum, so that the lack
of reaction with Magnolia virginiana antiserum is not due to generally
low antigen content. It is notable that McLaughlin (1933) placed the
genus in the Hamamelidaceae. However, Disanthus, in this latter
family, gave no reaction with Illicium, perhaps supporting its treat-
ment by some taxonomic workers as a monotypic family, Illiciaceae.
At the species level the data suggest a closer serological re-
lationship between the Asiatic species. Magnolia obovata and the
American species, M. tripetala than between the latter and two other
American species, M. acuminata and M. virginiana. Magnolia por-
toricensis (Table 5.1) is farthest removed from all species, and, accord-
ing to Johnson, there is some morphological evidence to support its
separation as a single species of a separate sub-genus.

SEROLOGY AND SYSTEMATICS

Other data presented by Johnson indicate that inter-specific
differences in Magnolia surpass the inter-generic differences in certain
cases. For example, the heterologous reaction between Magnolia
tripetala and M. portoricensis is 53.6 while the heterologous reaction
between Magnolia virginiana and Michelia champaca is 83.0. In fact,
the latter heterologous reaction is greater than many heterologous re-
action among species of Magnolia tested against M. tripetala antisera.
Since, apparently, similar procedure was used in all cases, it is difficult
to account for this apparent paradox, even when it is recognized that
different host animals were used which may differ in their antibody
responsiveness. The same type of situation is noted in Baum's work.
Unfortunately no explanation of this is presented in the original pub-
lications. In this connection an interesting statement in a discussion
of serological work on songbirds seems pertinent:

An additional point to consider in the interpretation of these [serolog-
ical] tests is that the techniques used tend to separate more sharply
species that are closely related, while species distantly related are not
so easily separated. In other words, comparative serological studies
with the photronreflectometer tend to minimize the differences be-
tween distant relatives and to exaggerate the differences between close
relatives. (Stallcup, 1961.)

This remarkable statement provides for a somewhat confus-
ing situation wherein, in the interpretation of data, one doesn't know
whether to consider two species farther apart or closer together than
the data indicate. If Stallcup's generalization is supportable, then cer-
tain taxa, of problematical familial alliance, such as Hydrastis (to be
discussed below), would be almost incapable of placement by ser-
ological results.

Hammond (1955b) compared a number of genera in the
Ranunculaceae on the basis of their serological interactions, and this
criterion, together with cytological and morphological data, was used
to produce a new systematic treatment of the genera. According to
Hammond the family is "serologically close-knit," and he considers
this observation to be in contradistinction to the generally held view-
point that certain genera of the family are relicts of ancient evolu-
tionary lines and thus genetically quite distinctive. The basis of Ham-
mond's statement is, however, not clear, since only a few families of
flowering plants had been studied at that time, and furthermore a
number of genera which he tested gave no reaction to the antiserum.
Hammond produced a three-dimensional model to depict the serological
relationships within the Ranunculaceae. Notable among his conclu-
sions is the placement of Hydrastis into the Ranunculaceae on the
basis of a positive reaction with Aquilegia antiserum. Hydrastis has

BIOCHEMICAL SYSTEMATICS

I W

« '%

(a)

(b)

Fig. 5-2. Immunoelectrophoretic patterns: (a) unadsorbed anti-
rye-wheat serum (in trough) against rye (right), rye-wheat hybrid
(center), and wheat (left); (b) adsorbed antirye-wheat serum
against rye (right), rye-wheat hybrid (center), and wheat (left)
(Hall, 1959). By permission of Hereditas.

been placed by some systematists into the Berberidaceae. Alkaloid
chemistry presents another line of biochemical evidence relevant to
the placement of Hydrastis, but in this case the biochemical affinities
are with the Berberidaceae. This question will be considered further
in a subsequent chapter devoted to the alkaloids.

The immunogenetic studies of M. R. Irwin and his colleagues
and students at Wisconsin are well-known, and these have been
alluded to briefly in the preceding chapter. Further consideration of
this significant work is included in the later chapter on biochemical
studies of hybrids.

A serological method which, in contrast to the straight pre-
cipitin reaction, is qualitative in nature has been utilized to advantage
in animal systematic investigations and to some extent in plant
studies. In principle this technique, known as Immunoelectrophoresis,
is similar to the other serological methods. Extracts of seeds or other
plant material are prepared and then subjected to agar-gel electro-
phoresis. As described by Hall (1959), in one modification, parallel

SEROLOGY AND SYSTEMATICS

troughs are cut into the agar along the electrophoretic track between
each sample, and antiserum is poured in (Fig. 5-2). The antigens and
antibodies diffuse into the agar, and when they meet, corresponding
antigens and antibodies form stabilized precipitates in the shape of
arcs which may be detected by appropriate methods.

A recent botanical study involving the immunoelectrophoretic
technique is that of Gell, Hawkes, and Wright (1960) on the genus
Solanum. They studied the gel diffusion patterns of fifteen Mexican
and twenty-two South American species of this genus. Since the anti-
sera were relatively ineffective in distinguishing the South American
species, only the fifteen Mexican species will be discussed. The latter
species are arranged into seven series according to Hawkes' sys-
tematic treatment. (He divided the tuberous solanums into seventeen
series, some of which are listed below.)

HI. Morelliformia

IV. Bulbocastana

V. Cardiophylla

VI. Pinnatisecta

XII. Demissa

XIII. Longipedicellata

XIV. Polyadenia

S. morelliforme

S. bulbocastanum

S. cardiophyllum
S. ehrenbergii
S. sambucinum

S. pinnatisectum
S. Jamesii

S. demissum
S. guerreroense
S. semidemissum
S. spectabile
S. verrucosum

S. polytrichon
S. stoloniferum

S. polyadenium

In the preparation of the extracts from the tubers the crude
juices were adjusted to yield a protein concentration of 0.5 per cent.
Antisera were prepared from rabbits.

80 . BIOCHEMICAL SYSTEMATICS

With an antiserum prepared from S. tuberosum the Mexican
species could be divided into three groups: one gave four precipitin
lines, another two lines, and the third one line. (All South American
species give reactions similar to S. tuberosum.) One aberrant epiphy-
tic species, S. morelliforme, yielded only one line, but in addition this
line could not be further resolved into two lines as was the case with
a precipitin line in a similar position in the other fourteen species.
Therefore, serologically, S. morelliforme appeared farthest removed
from S. tuberosum. In some comparisons, notably in strains of S.
polytrichon, differences within a species proved greater than those be-
tween species.

Species of Series V and VI without exception yielded two lines
with S. tuberosum, antiserum and further showed a similar pattern
against antiserum of one of the Series V species (namely S. ehren-
bergii) and even a moderately close relationship to species of Series
IV and XIII. Thus the authors consider Series V and VI to be a link-
ing group between the two pairs of series mentioned.

All other species were grouped together when tested against
S. tuberosum, but against S. ehrenbergii (preadsorbed with S. tubero-
sum) only S. tuberosum, S. verrucosum and S. semidemissum were
placed together (showing no precipitin lines).

Series XII, XIII, and XIV, which gave four-Hne patterns
against S. tuberosum, have in common an important morphological
feature, the rotate or wheel-shaped corolla, while the other series
have a stellate corolla type. (South American species have, also, the
rotate corolla.) Furthermore, crosses within Series XII, XIII, and
Series Tuberosa can be made as well as between Tuberosa and various
South American species. Series Bulbocastana, which also gave a four-
line pattern with Tuberosa, is exceptional in that it does not hybridize
with species of the other Series readily. Finally, Series V and VI,
giving the two-line spectrum, are fairly interfertile. Although the
results should prove to be interesting, reaction patterns to specific
antisera of all species concerned were not reported.

In general, the serological data from Solarium follow rather
closely the patterns of morphology and hybridization. The fact that
the type of serological methods used by these authors provides a
pattern in accord with other lines of evidence lends validity to the use
of the method in systematic studies. It is interesting that immuno-
electrophoretic studies generally yield only a relatively few arcs of
interaction, even though no preadsorption is carried out. From pre-
cipitin reactions one gains the impression that a large number of
antigen-antibody interactions are involved in a single precipitin reac-
tion. It is probably that in Immunoelectrophoresis, when crude extracts

SEROLOGY AND SYSTEMATICS

are used, only the major constituents in the serum are in sufficient
quantity to yield a visible reaction, and antigens of an enzyme nature
are not detected.

Somewhat similar work on the legume genus Phaseolus has
been done by Kloz (1962), but the number of species investigated at
this time (four) is too few to allow significant conclusions. However,
apparently significant serological differences exist, and it is likely that
the more extensive analysis in progress will provide further insight
into the relationships of the species in this genus.

Earlier in this section it was stated that even in the 1920's
differences of opinion existed between the Konigsberg and Berlin ser-
ologists as to whether there were serological differences within a
plant, that is, whether different organs were serologically distinct.
Recently Kloz et al. (1960) have demonstrated unequivocally that
such differences exist and indeed often exceed inter-specific serological
differences. These workers compared the antigenic substances from
cotyledons, "subcotyledonous" parts (roots) of seedlings, and mature
leaves in Phaseolus vulgaris, P. coccineus, Glycine soja and Vicia faba.
They employed essentially the technique of the Rutgers group. Some
of their precipitin results (in per cent) are given below.

Phaseolus vulgaris (antiserum of cotyledons against sera from
cotyledons of the following species):

P. vulgaris 100

P. coccineus 88.2

Glycine soja 3.4

Vicia faba 1.8

Phaseolus vulgaris (antiserum of leaves against sera from
leaves of the following species):

P. vulgaris 100

P. coccineus 89.7

Glycine soja 41.7

Vicia faba 19.9

Comparison between serological properties of individual or-
gans of the same species, data taken from P. vulgaris.

Antiserum against cotyledons tested against sera from the
following sources:

cotyledons 100

subcotyledons 8.9

leaf 5.2

BIOCHEMICAL SYSTEMATICS

Antiserum against subcotyledonous tissue of seedlings against
sera from the following sources:

subcotyledons 100

leaf 23.7

Antiserum against leaf tissue tested against sera from the
following sources:

leaf 100

subcotyledons 53.4

cotyledons 8

There is no question but that there are serological differences
among the organs investigated. Kloz et al. stated that the protein
characters of cotyledons (reserve proteins) showed weakest cross re-
actions between species of different genera, indicating that generic dif-
ferences were more pronounced in these organs. These authors
presented the hypothesis that the protein characters of the sub-
cotyledonous and leaf tissues are phylogenetically older than storage
protein of the cotyledons and therefore emphasize the common
origin of taxa more than do characters which have undergone
differentiation at later stages of evolution. Although the hypothesis is
interesting and, if valid, of theoretical importance, it may be an over-
simplification. Presumably, in this instance what one is measuring are
differences which parallel and reflect the evolution of several genera
of a single tribe at a time when the cotyledons had already made
their evolutionary appearance. It is therefore possible that in some
cases more subsequent specialization appeared in organs such as leaves
than appeared in cotyledonous proteins. In any event, the major
point, that serological differences exist among different tissues within
a plant, should not be obscured by further attention to the second
question.

Wright (1960) has refined, further, investigation of organ
specific antigens. By combining ultracentrifugation and immuno-
diffusion he was able to demonstrate an antigen in the microsome
fraction of three-day old coleoptile tissue of wheat. In order to ex-
clude non-microsomal antigens, the antiserum was first absorbed with
the supernatant of the microsome fraction. The precipitin band
associated with the microsome fraction of three-day coleoptile tissue
was absent from coleoptile tissue of a younger age and from root and
leaf tissue. These data imply, in the words of the author, "that a non-
organ specific meristematic pattern of antigens has superimposed upon
it, during differentiation, a combination of proteins characteristic of
differentiated cells."

SEROLOGY AND SYSTEMATICS 33

In recent years there has been considerable interest in a modi-
fied serological method using relatively crude plant extracts to dis-
tinguish human blood types. Although an important motivating
factor in this work is a rather practical consideration, namely, the
commercial application of specific plant agglutinins in blood typing,
there are in addition a number of intriguing problems of a fundamen-
tal nature involved (Boyd, 1960). Although this work is not widely
known, even now, the first report of the existence of a plant agglutinin
manifesting some degree of selectivity was made as early as 1888
when Stillmark noted that an extract from the seeds of the castor
bean {Ricinus communis) agglutinated the red blood cells of animals
selectively. Although occasionally some minor work was devoted to
plant agglutinins, for the most part the subject was ignored until
1948 when Renkonnen at Helsinki revived interest in plant agglutinins
with a survey of ninety-nine legume species, six of which showed
definite affinity for either A or blood types. Subsequently, nu-
merous investigations have disclosed a large number of legume species
which agglutinate red blood cells, sometimes with no antigenic speci-
ficity but frequently with definite specificity. No knowledge is
available concerning the botanical function of the agglutinins which
are usually, but not always, obtained from the seeds. By 1955 per-
haps a thousand species of plants had been screened, and an over-
whelming proportion of the species disclosed to be producers of
"specific" agglutinins were in the family Leguminosae. The relevance
of this work to biochemical systematics lies in the question of whether
such investigations can disclose any meaningful patterns of distribu-
tion of agglutinins among the plant species.

Following the work by Renkonnen, selected examples of some
important early surveys are those of Boyd and Reguera (1949) who
studied 262 species from sixty-three families and Krupe (1953) who
studied 167 species in the Leguminosae and, in addition, ninety-four
different varieties of lima bean {Phaseolus lunatus). Krupe, for
example, found some genera such as Lathyrus and Phaseolus which
showed quite consistent agglutinin activity; that is, a large propor-
tion of the species were active. Other genera, for example, Caragana,
showed wide species differences. Species which specifically favored
certain blood groups were recognized; for example, Lotus tetragono-
lobus favored blood type O; Vicia cracca and Phaseolus lunatus
favored blood type A; and Sophora japonica and Coronilla varia
favored blood type B. By far the most comprehensive survey and
general study of plant agglutinins, however, has been that of Makela
(1957), a student of Renkonnen, who studied 743 species of the
family Leguminosae, including 165 genera. Thirty-seven per cent of

84 BIOCHEMICAL SYSTEAAATICS

the species studied contained agglutinins in their seeds, and a number
of the agglutinins showed some degree of specificity. This work will
be discussed in more detail later.

The technique used in testing for plant agglutinins is rather
simple. Seeds are ground and extracted in saline, appropriately
diluted, and then incubated at room temperature with a suspension
of erythrocytes. The mixture is then examined for e\idence of agglu-
tination. Apparently the extracts are quite stable and may be re-
tained for months vsithout significant loss of acti\'ity. Makela, who
studied such pai-ameters as temperatui-e, salt content, and pH, re-
ported a surpiising tolerance for such an appai'ently specific reaction.
For example, agglutination occurred over a fairly wide concentration
of NaCl in the medium (though optimum results were obtained near
the "physiological" range) and over a pH range from 5 to 11. Despite
the simphcity of the technique there appears to be fairly good
reproducibihty, and some of the results are actually, in themselves,
validation of the method. For example. Schertz et al. (1960)
reported that a specific hemagglutinating substance. '"anti-A," from
the lima bean is inherited as a simple Mendelian dominant. The Fi of
a cross between one high-activity parent and one inactive parent
yielded seventy-two plants showing high activity, none showing no
activity. The F2 segregated essentially three active to one inactive.
Furthermore, Morgan and Watkins (1956) have utihzed specific plant
agglutinins to show that the blood group antigen of type AB individ-
uals is a unique molecule rather than a mixture of A and B substances.

There is no certain knowledge of the chemical nature of the
plant agglutinins. Some investigators consider them to be muco-
proteins. Rigas et al. (1955) obtained a highly active mucoprotein
fraction which, when hydrolyzed, yielded an inactive polysaccharide
and a very active euglobulin. Presumably ehmination of the poly-
saccharide enhanced activity of the protein. However, he does
not beheve that the term "antibody" is entirely appropriate and in-
stead refers to the agglutinins as "lectins." Part of his objection to
the use of "antibody" Ues in the fact that their formation is not
ehcited as in the case of most animal antibodies. Some investigators
consider that the plant agglutinins do not react with the same re-
ceptors as do the typical antibodies. The question is not completely
settled, however. One argument in favor of a different mode of
specificity for the plant agglutinins is that the plant agglutinins are
neutralized in many cases by certain simple sugars while animal
agglutinins are not. Makela beheves that the plant agglutinins
accidently possess a configuration that is complementary to the
chemical grouping of the blood group substances.

SEROLOGY AND SYSTEMATICS

One point raised previously is of particular interest. It was
stated that simple sugars may sometimes inhibit specific plant
agglutinins. The first report of inhibition of this type was that of
Watkins and Morgan (1952) who, in this case using anti-H agglutinin
of eel serum, discovered that agglutination was inhibited by L-fucose.
In addition to certain simple sugars, the sugar derivation N-acetyl-
galactosamme inhibits anti-A, and anti-B agglutinins from a number
of sources (Makela, 1957). This observation acquires added signif-
icance following the disclosure that X-acetyl-galactosamine is present
in hydroly sates of blood (particularly high \-ields are derived from
blood gi-oup A, lower fields in 0, and practically none in B). It is
suggestive of the presence of a carbohydrate-like terminal group on
the antigen. A general theory to account for the inhibition of agglu-
tinins by simple sugars is that the sugars resemble the reactive end
group of a red blood ceU receptor. The sugars then attach to the
agglutinin complementary site blocking it and thereby preventing
agglutinin-receptor contact. On the basis of the structural relation-
ships of groups of sugars which are effective inhibitors as opposed to
the ineffective sugars it has been suggested by Krupe (1956) that the
configuration of carbon 3 and 4 is important in determining the
abihty of the sugar to inhibit agglutinins.

Four patterns available in placement of hydroxyl
groups in carbons 3 and 4 of aldopyranoses.

h6\oh

The more strongly inhibiting sugars fall into groups 1 and 2
above. For example inhibitors of anti-B extracts, such as L-arabinose
and D-galactose, represent group 2. Inhibitors of anti-H extracts, such
as D-ai'abinose, D-digitoxose. L-fucose, and L-galactose ai'e of group

86 BIOCHEMICAL SYSTEMATICS

1. However, Makela reported that some group 3 sugars such as
D-glucose, D-mannose, and the ketose, D-fructose, strongly inhibited
some other types of plant agglutinins.

In connection with the question of the systematic value of the
plant agglutinins it is premature to attempt a final evaluation. Plant
agglutinins have been found in a number of families of flowering
plants, but it has already been noted that the Leguminosae have
special proclivity toward the production of agglutinins. Yet the
agglutinins are not common within the sub-families Mimosoideae and
Caesalpinioideae of the family. The Mimosoideae are especially
deficient (only a few species of the genus Parkia are positive). The
extensive survey of the Leguminosae by Makela has provided enough
data to permit some generalizations to be made concerning the distri-
bution of agglutinins within the legume family. It is obvious that
only tendencies are disclosed by the data. That is, at most taxonomic
levels the character tends not to be constant. For example, within the
sub-family Papilionoideae only the tribes Podalyrieae and Trifolieae
have not proved to have any agglutinins. The tribes Dalbergieae and
Hedysareae are somewhat poor in agglutinins, but the rest of the
tribes contain numerous agglutinin producers. In the tribes Phaseoleae
and Galegeae, there is a great variation, but in the tribe Vicieae, which
is particularly consistent, the large majority of species studied produce
agglutinins. There is some apparent regularity of a qualitative nature
with respect to the distribution of the agglutinins. For example, anti-H
agglutinins are quite rare except in the tribe Genisteae where they are
frequently encountered in the genera Cytisus, Genista, Laburnum,
Petteria, and Ulex, though absent from some other genera in the tribe.
Elsewhere, only the genus Lotus, of the tribe Loteae and Virgilia of
the Sophoreae are known to produce anti-H agglutinins.

Makela does not emphasize strongly the systematic aspect.
He makes only a few general comments such as the following:

The occurrence can be said to conform to the taxonomic plant system
to some extent though by no means absolutely. Proofs of this are,
in particular, the total absence of agglutinins in the seeds of certain
tribes, e.g. Trifolieae, and the almost regular presence in the seeds of
Viceae.

When data concerning the specificities of plant agglutinins
are supplemented by further knowledge of their responses to various
inhibitors, they may involve sufficient qualitative refinement to dis-
close a more meaningful pattern to the distribution of plant agglu-
tinins than we now have. Some progress in this direction has been
made, but the results so far have not given much cause for optimism.

SEROLOGY AND SYSTEMATICS 87

For example, Makela has combined specificities and inhibitor charac-
teristics to distinguish two groups of genera as follows. Certain genera
produce agglutinin which act upon rabbit cells but not guinea pig
cells and are inhibited by certain sugars of group 2 (for example,
D-galactose). Most of them contain also anti-A + B or anti-B agglu-
tinins. Another group of genera produce agglutinins which act on
rabbit cells but also on guinea pig cells. These agglutinins are inhibited
by group 3 sugars (for example, D-glucose). Genera falling into the
former class are Bandeiraea, Sophora, Crotalaria, Cytisus, Caragana,
and Coronilla. Genera of the second group are Lathyrus, Lens,
Pisum, Vicia, and possibly Parkia. It is evident to taxonomists that
the two groups do not fall neatly into any systematic order. In fact,
the groups individually overlap even the sub-family level, and each
group includes a number of tribes, some of which are represented by
species in both groups. There is reason, however, in the opinion of the
writers, to expect the plant agglutinins to be systematically important,
if not at the tribal level then at least at the genus level. Since it has
been shown that the presence of at least one agglutinin is genetically
controlled, by that fact alone there is established a rational basis for
their distribution which is phylogenetic in principle.

In the preceding chapter an example from zoological studies
and one from botanical studies were utilized to illustrate biochemical
systematics approaches which provided significant information but
which did not represent, entirely, correlations of the distribution of a
compound with a taxonomic system (for example, Vogel's investiga-
tion of lysine synthesis and Wald's investigation of the visual pig-
ments). Although it was not intended that this establish a precedent
for later chapters, in a general discussion of the possible role (includ-
ing the future role) of serology in phylogenetic studies it is useful to
review briefly one or two special applications of serology which,
although not directly relevant to systematics, nevertheless indicate
some of the possibilities of the method. In general, some of the limita-
tions of the gross quantitative serological method are obviated when
refined genetic stocks are available and appropriate preadsorption is
utilized. Since the genetic knowledge of the materials to be discussed
below is more complete than that of earlier studies, the implications
of the work seem to have greater validity. Again, one example is drawn
from plant studies and one from animal studies.

The first example treats serological investigations of a series of
four pollen incompatibility alleles (S2, S3, S4, and Sq) of Oenothera
organensis carried out by Lewis (1952). Preadsorption was utihzed
not only to precipitate common protein antigens not connected with
the S-factors, but also to provide, artificially, what are referred to as

BIOCHEMICAL SYSTEMATICS

"half-homologous" extracts. For example, if one wishes to form a half-
homologous antiserum to a plant of the genotype S2-3, an antiserum
is prepared against S2-3 serum and pre-adsorbed with S2-4 serum. In
the first reaction the common proteins, including the S2 antigen are
precipitated, leaving in the antiserum, presumably, only the S3 anti-
bodies. Another S2-3 pollen extract is then said to be "half-
homologous" with the antiserum, and an 82-6 extract is said to be
"heterologous." In the latter case no reaction, or at most a very slight
reaction, may be expected. The results of a series of cross reactions
are shown in Fig. 5-3. The figures within the squares represent the
time, in minutes, required to form a precipitin ring; therefore the
lower values indicate the stronger reaction. A blank indicates that the
reaction was not recorded. Certain inconsistencies are presumed to
result from extraneous genie differences in the material since the
stocks were not isogenic. However, these gene differences were obviously
not sufficient to prevent generally good correlations.

TA S

2.6

»3.4

>3.6

J3.6

S4.6

^2.4

>2.3

AS AA

S3.6 S2.4

S3.4 S2.6

S3.4 S3.6

S3.4 S4.6

S2.4 S36

S2.4 S3.6

S2.6 S3.4

S2.6 S4.6

Fig. 5-3. Time in minutes to form a ring precipitate at ^ dilution of test antigen (TA).
Antiserum is designated AS, and the adsorbing antigen is designated AA.

B = Heterologous antigen, that is, no S alleles common to test antigen and adsorbed serum.

n = Completely homologous, that is, both S alleles common to test antigen and adsorbed serum.

D = Half-homologous, that is, one S allele common to test antigen and adsorbed serum.

I I = Untested combinations.

. - 00 JB

WrillWilHM»li IP iiMBitff »f^

SEROLOGY AND SYSTEMATICS

The original purpose of the investigation was to discover, if
possible, the nature of the mechanism by which diploid pollen (from
tetraploids) carrying two different S alleles was not inhibited. If the
interaction fails to produce the "S-substances" or if a new product with
new specificity was produced under such circumstances, this might be
disclosed by the serological method. However, Lewis did not consider
the tests sensitive enough to yield vahd results. Linskins (1960)
reported results similar to those of Lewis in a study of three S alleles
in Petunia hybrids. It is obvious that methods analogous to those of
Lewis, and to some extent Irwin (that is, the gene homologies in
certain birds), might be apphed successfully to a study of gene
homology in related species. The critical point is to be able to pre-
adsorb in such a fashion as to leave in the antiserum a designated single
antibody which can then be used to screen other species.

In the second example. Fox (1949) employed a serological
method to study specific eye color mutants of Drosophila. Using
isogenic stocks, he analyzed serologically various combinations of the
mutants, ruby and vermilion, along with the wild type. By means of
selective adsorption such as described above Fox showed that vermilion
antiserum and the double mutant, ruby-vermilion, antiserum were ser-
ologically equivalent. He inferred that the normal allele at the ruby
locus further modifies an "antigen" dependent upon the wild type
allele of vermilion. Therefore, in the presence of the vermilion allele
the normal allele at the ruby locus cannot effect the modification, and
consequently in the vermilion phenotype no serological difference be-
tween Rb (normal allele of ruby) and rb could be detected. If the
antigens are indeed enzymes and if the conclusions are valid, this
is yet another example of gene interaction to produce a single enzyme,
an exception to the classic one- gene, one-enzyme hypothesis. (Of course,
in these days, when the gene is becoming almost as difficult to define
as a species and virologists are threatening to reduce the "unit of
crossing-over" to as little as two nucleotide pairs (Benzer, 1957) such
aphoristic generalizations are inviting targets anyhow.)

The examples just discussed do not by any means represent
all of the instances in which serological methods have been applied to
the study of specific enzymes or genetic factors. Several other studies
of this nature have been noted by Moritz (1958).

In summary, there is reason to believe that serological tech-
niques, especially those utilizing immunogenetic methods (in general,
restricted to the lower taxonomic categories), will make important
contributions to systematics. Immunoelectrophoresis, however it may
be applied, presents in addition a distinct advance over earlier tech-
niques in that a qualitative element is introduced, and this technique

90 BIOCHEMICAL SYSTEMATICS

may be expected to become increasingly significant. Modifications of
the strictly quantitative precipitin tests, even with increased sensi-
tivity, seem to possess some inherent limitations. As a general criticism
it appears that proponents of the precipitin methods have expended
tremendous effort in the development of techniques without exerting
equivalent efforts to set up critical "test" experiments or to pursue the
theoretical aspects of antibody-antigen reactions at the biochemical
level.

AMINO ACIDS

Amino acids are generally recognized primarily in
their role as structural units of proteins. The fact
that amino acids may have additional important
roles may not always be fully appreciated. There are
already more non-protein than protein amino acids
known, for example, and the ratio of non-protein to
protein amino acids will increase as new amino acids
are discovered (Fig. 6-1). Numerous ninhydrin posi-
tive substances, yet uncharacterized, are known.
Most, if not all, of the protein amino acids are of
little taxonomic value by virtue of their cosmopoli-
tan distribution! while the non-protein acids are

1 Eventually, however, as indicated previously, the struc-
tural sequence of protein amino acids may prove to be among the
ultimate phylogenetic criteria.

BIOCHEMICAL SYSTEMATICS

*S 20

Advent of

paper

uomatograph

__</

r^^'"^

k_-<5r

1910

1920

1930

1940

1950

1960

Date

Fig. 6-1. The effect of paper chromatography on the rate of
identification of new non-protein amino acids (Fowden, 1962).

often of somewhat more limited distribution and therefore may be
effectively used as systematic characters.

Since the vast majority of the non-protein amino acids are
probably not basic metabolites in the strict usage of the term, there is
some question as to the appropriateness of including amino acids along
with carbohydrates and lipids among basic metabolites. Whatever
group of substances to which the amino acids may be more appropri-
ately related, there are at least some historical reasons for beginning
the treatment of specific classes of compounds with a chapter on amino
acids.

The technique of paper chromatography, which has been
primarily responsible for renewed interest in biochemical systematics,
had its origin in studies of amino acids (Consden et al, 1944). Partly
because development of new techniques in chromatography and re-
finements of older methods proceeded most rapidly in amino acid

AMINO ACIDS 93

chemistry, chromatographic investigations of amino acids have greatly
outnumbered similar investigations of other classes of compounds.
Consequently, more workers are aware of the possible application of
chromatography to the study of amino acids than perhaps any other
group of compounds. Therefore, it is not surprising that some of the
earliest investigations into the application of chromatographic tech-
niques to systematics involved amino acid patterns. There is no
indication in these early studies that there was careful consideration
of the question of whether amino acids were, on apriori grounds,
likely to be of greater systematic value than other classes of substances.
As has already been noted amino acids are among the least useful
classes of substances if one concentrates upon the approximately
twenty amino acids of protein. Not only are these protein amino acids
nearly always present in tissues but, in addition their absolute and even
their relative concentrations are so closely dependent upon the
physiological state of the moment and so sensitive to metabolic dis-
turbances that their quantitative as well as qualitative relationships
are Kkely to be of little systematic value. This last point will be dis-
cussed further in a later paragraph.

Before the advent of paper chromatography, the study of
amino acids contributed very little data of taxonomic importance.
Chromatographic techniques, however, not only provided new dimen-
sions of study of the common amino acids (for example, comparisons
of amino acids of individuals and accurate measurements of the con-
centrations of various free amino acids in a single root apex), but in-
advertantly disclosed the presence of a variety of "new" amino acids.
Fowden (1959) describes these latter compounds as "products of
the chromatographic revolution." It should not be inferred that all of
the non-protein amino acids owe their discovery to paper chromatog-
raphy. In lists compiled by Vickery (1941) and Dunn (1943), prior to
the development of chromatography, a number of suspected non-
protein amino acids were included (though they were reported simply
as not known to be constituents of protein). Each of these lists con-
tained approximately fifty compounds, about twenty of which were
the ubiquitous protein amino acids.

Table 6-1 illustrates some of the non-protein amino acids of
plants and relates them structurally to protein amino acids when
possible. At least one acid of column two is found in protein
(a-aminoadipic acid is found in the protein of corn seeds) though it is
more typically associated with non-protein amino acids.

New amino acids continue to be reported, and since 1958
some twenty or more additional amino acids have been characterized.
Several new amino acids have been discovered in the Mimosaceae,

BIOCHEMICAL SYSTEMATICS

£2

•S

+J
O

o
a

-a
a

o
a

o
<v

• I— I

o
C/2

j::

-2

Ci,

•S

1^)

8 c

■4^

3
u
o

o «

l^)

n >,

13 g

Sob

o
3

■^

__o ^^

o d

-a

a
(/}

CS
Ol

o
03

3
o
3
O

01
Ol

o
e

S
a

c
o
c

o
o
o ^

3 «

o ;3

(w) 0)

o
o
o

2 ^
O §

^^ C

o
o

X OJ

X I
o -

i-Ih 01
U Ol

U
O

O
O

X S

3-1 tH

o «

o w

X
O

^ 03
c-i

2 >,

- — ' -M

CJ I

2 ^
o

o
o
o

X 1^

O o

" S

o X

X
O

X
o
o
o

r— 1

2—2,

X ffi/"K

o— o

o
o
o

CM S

o
o
o

o -=

o
o

fc CM

^ X

o
o
o

X aj

2 .c

E 03

CM a

O^
O

2
X

X
o
o
o

X
2

X S

X ^
u

o
o
o

QJ a;

e X

cq o

X
O
o
o

X
2

.3
-a

k^ X

AMINO ACIDS

cc'

1^
3

§^

o
CO

3
be

T3
C
CS

,1^

-fit

O cu

^ c

K ^
2 ^

■^ o

o
o

X >.

X 5

3
"-2

o
o
o

X X

X
o
o
o

X X

Ct3

'S.
>.

-3

\ /
Q

X
O

o
p

X
u

X 2

o .s

^ he

X
O

X
o

>>
a
o
;-i
a
_o
o

+-»

<3C^

•t-i HiH OJ

ffi

'1^2

800

!-i
CL,

3 X

■3 ^

ri f "^ X
X X/

CO X

►^

0—0

96 BIOCHEMICAL SYSTEMATICS

particularly in the genus Acacia (Gmelin, 1959; Virtanen and Gmelin,
1959; Gmelin et al, 1959; Gmelin and Hietala, 1960). New amino
acids have been identified in Lathyrus species (Bell, 1961); in Reseda
(Larsen and Kjaer, 1962); in Allium (Virtanen and Matikkala, 1960);
in Ecballium (Gray and Fowden, 1961); in crown gall tissue (Biemann
et al, 1960); and in a red alga (Kuriyama et al, 1960. Even a
selenium-containing amino acid has been identified in Astragalus
(Trelease et al, 1960). These amino acids fall into several different
chemical sub-types.

Although at present, the majority of these newly discovered
non-protein amino acids are known to occur in only a few plants,
Fowden (1962) has noted that one still cannot detect trace amounts,
and it may be that their distributions are far more extensive than now
suspected. Steward et al (1955) stated that eighty-one ninhydrin
reactive substances found in the non-protein portion of various plant
extracts did not correspond to any of the known amino acids.

Systematic studies involving amino acids

One of the first taxonomic studies employing chromatography
of amino acids was that of Buzzati-Traverso and Rechnitzer (1953).
In this brief paper the authors compared the chromatographic pat-
terns of fish muscle protein hydrolysates from different species. The
amino acids themselves were not identified, and the chromatograms
showed few spots, but it was evident that differences in the patterns
occurred. It is strange, in view of the general occurrence of twenty
amino acids in protein that, in protein hydrolysates, many of the
twenty amino acids were missing. According to the authors the
amino acid patterns of species regarded as more closely related
by other criteria were more alike chromatographically, and they
further maintained that stocks from geographical races of the
same species could sometimes be distinguished. Although the authors
forecast wide use of chromatography in population and genetic studies,
little work of this type on fish has appeared subsequently. Vismanathan
and Pillai (1956) repeated, essentially, the work of Buzzati-Traverso
and Rechnitzer, in a study of sardines, but the results contributed
nothing to the systematics of the group.

Another paper by Buzzati-Traverso (1953) has gained con-
siderable attention. The work is not primarily systematic and does
not even treat exclusively the amino acids. However, imphcations of
the work, if substantiated, bear directly upon the sensitivity of the
chromatographic method and indirectly upon apphcations of similar

AMINO ACIDS 97

types of investigations to systematics. Although Buzzati-Traverso's
work included examination of fluorescent substances in addition to
amino acids, the former class was probably a rather heterogeneous
assemblage of undefined substances. These may be considered here
since the principles concerned are independent of the nature of the
compounds compared. Fruit flies (Drosophila) were studied inten-
sively, but Buzzati-Traverso also included some plant studies, as will
be disclosed. The flies were fed a standard diet and chromatographed
by mashing the individual flies directly on paper. It was not stated
that the Drosophila compared were isogenic. The ninhydrin patterns
(revealing amino acids) of different Drosophila strains were said to be
similar, but with respect to fluorescent spots, males and females ex-
hibited distinct differences. Since later work by Fox (1956) on sex dif-
ferences in Drosophila elaborates this point somewhat, and will in
turn be discussed, no description of these sex differences is necessary
here. Of more importance for this discussion are the results of
Buzzati-Traverso's comparisons of a series of mutants with corre-
sponding wild type flies. As a background, it should be noted that
Hadorn and Mitchell (1951) had undertaken a chromatographic
study of fluorescent patterns of Drosophila mutants in eye color and
body color. Although conspicuous changes in the fluorescent patterns
occurred in different developmental stages, Hadorn and Mitchell, in
their early work reported no significant differences in the fluorescent
patterns of either eye color or body color mutants, as opposed to wild
type, at any stage examined. More recently, however, with improved
chromatographic techniques for the separation of pteridines (which
include the Drosophila eye pigments), distinctive chromatographic
differences are now correlated with a number of eye color mutants
(Hadorn, 1962). Buzzati-Traverso used a group of m.orphologicgd
mutants rather than biochemical mutants (that is, the overt
phenotypic expression was morphological rather than biochemical).
By Buzzati-Traverso's interpretation of his results each of the strains
tested gave a distinctive fluorescent pattern, and each genotype had
a characteristic biochemical pattern. According to the author the
heterozygotes could always be detected, though the morphological ex-
pression of the gene indicated dominance. The present writers, after
examining the illustrations in the Buzzati-Traverso paper, have some
reservations concerning his interpretation. It appears possible, if not
probable, that the pattern differences were in part artifact. In some
cases, for example, two patterns may appear to be different, but by
our interpretation the only difference that is apparent is a shifting of
the Rf values of the same series of spots upward in certain cases.
Perhaps, in the photographic reproduction the detafls were lost, but

9g BIOCHEMICAL SYSTEMATICS

on the basis of the evidence presented, certainly no clear-cut dif-
ferences can be detected.

On the basis of previous results and on theoretical grounds,
differences such as those reported are indeed surprising, though
Buzzati-Traverso did not indicate this to be so. For example, Hadorn
and Mitchell did not detect differences in the heterozygote, and those
authors were studying biochemical mutants.2 in the Buzzati-Traverso
work in which a series of morphological mutants were compared, it is
remarkable that of a small number of unidentified fluorescent sub-
stances, one or more of them are invariably affected, quantitatively or
qualitatively by each mutation. It is even more surprising that the
heterozygotes could be detected chromatographically. In general when
a biochemical mutant is dominant (for example, a flower color factor)
one can scarcely detect the heterozygote even by sensitive quantitative
methods. Dominance may be above 90 per cent in the majority of

such cases.

Similarly Buzzati-Traverso reported that a recessive mutant
of tomato had two fluorescing spots not present in the wild type strain.
However, in this case the fluorescent pattern of the heterozygote was
similar to that of the double recessive, but its ninhydrin pattern was
intermediate. Finally, in a yellow-green mutant of muskmelon the
chlorophyll content of the heterozygote could not be distinguished
from the homozygous dominant, but the ninhydrin pattern of the
heterozygote was indistinguishable from that of the double recessive.
The writers consider that the interpretation given by Buzzati-Traverso
to his results is not necessarily the only interpretation which is
plausible, since the photographs resemble closely anomalies we have
sometimes observed in our experiences with paper chromatography.

Interesting work on the comparative fluorescent patterns of
male and female Drosophila has been reported by Fox (1956). He was
principally concerned with whether any biochemical differences could
be attributed to the presence of the Y chromosome. Therefore, he
compared males and females chromatographically and then compared
normal (XX) females and females carrying, in addition, a Y chromo-
some (XXY). These two types of females showed similar patterns,
suggesting that the Y chromosome itself was not responsible for
any overt biochemical effect. Of more significance to systematic in-
vestigations was the disclosure by Fox that, in isogenic stocks, strik-
ing differences in the fluorescent patterns of males and females
occurred. Ten spots were common to males and females, but in all

2 In the case of white eye, which is morphologically recessive, the chromatographic
pattern of the heterozygote is distinguishable from the homozygote of either class, but
this appears to be rather exceptional (Hadorn, 1962).

AMINO ACIDS 99

but one of these spots there were detectable quantitative differences.
In addition there were seven spots pecuhar to males and two spots
peculiar to females. Some of these spots were probably the pteridines
mentioned in Chapter 4.

Although Fox was not immediately concerned with potential
systematic applications stemming from his work, it is possible that
the comparative biochemistry of sex, could be extended profitably to
other species and genera, or even higher taxa, of insects. In the
present case a number of absolute sex-linked differences were re-
corded in one species, and a systematic extension of this comparison
could hardly fail to provide valuable insight into relationships.

Micks (1954) applied amino acid chromatography to a study
of certain mosquito species which are difficult to separate on morpho-
logical bases, and his illustrations of chromatographic differences are
convincing. Later, Micks (1956) studied several different groups of in-
sects, and again his illustrations of comparative ninhydrin patterns
show distinctive differences at the order level (that is, in a compar-
ison of certain Hemiptera, Diptera, and Orthoptera). Even three
genera of cockroaches could be distinguished chromatographically.
Within a single genus, however, any differences which were apparent
were quantitative. Intrageneric qualitative differences in mosquito
(Culex) had previously been reported by Ball and Clark (1953). These
investigators found aspartic acid in Culex quinquefasciatus and C
stigmatosoma, though an extract five times as concentrated was used
in the last named species. They also reported the unusual sulfonic
amino acid, cysteic acid, in C tarsalis and C. stigmatosoma but not
in C. quinquefasciatus. Cysteic acid may possibly have arisen as an
artifact by oxidation of cysteine. It is noteworthy that specimens of
C quinquefasciatus as widely separated as California and Texas were
qualitatively identical, and Ball and Clark concluded that the inter-
specific differences were intrinsic, not environmental.

In other systematic zoological studies involving chromatog-
raphy, Kirk et al. (1954) found that seven species of land snails could
be distinguished by their fluorescent patterns. The pattern for a
given species was the same regardless of diet or geographical location.
A few other reports are scattered throughout the literature such as
those of Mohlmann (1958) who studied fluorescent patterns of butter-
flies and Wright (1959) who studied mollusks of the genus Lymnea by
similar methods. In summary, however, in the judgment of the
present writers none of the papers in this series extends beyond the
point of suggesting that chromatographic studies might be valuable
in future taxonomic investigations. None was addressed to any
specific problem or shed any light upon an actual systematic problem.

]00 BIOCHEMICAL SYSTEMATICS

Botanical investigations of amino acid patterns, while
initially lagging somewhat behind zoological studies, now seem to be
providing even more information of direct taxonomic utility, probably
because unusual amino acids are more often involved.

Bell (1962a) has recently examined forty-nine species of the
legume genus, Lathyrus, and has presented data which appear to be
potentially quite valuable in interpreting species affinities within the
genus (compare the work of Pecket on the phenolics of Lathyrus,
Chapter 11). A new guanidine amino acid (homoarginine) is present
in seeds of thirty-six species; seven unidentified ninhydrin-reacting
compounds in concentrations of the order of 1 per cent have been de-
tected in the seeds of one or more species. Some of the substances are
probably those responsible for the toxic condition known as neuro-
lathyrism (Chapter 10), and others may be related to the lathyrus
factors. Bell believes that these non-protein amino acids may con-
stitute a highly concentrated form of nitrogen storage in leguminous
seeds, and many of these amino acids, in fact, contain additional
nitrogen. Although free amino acids are not typically found in the
seeds of most plants in high concentration, the content of free amino
acids in seeds of the Leguminosae is often high.

The most important immediately significant taxonomic con-
clusion from the work of Bell is contained in the following statement
by the author:

Within the genus there existed well defined groups of species that were
characterized, not by the presence of an arbitrary concentration of
one specific ninhydrin-reacting compound, but rather by the presence
of associated groups of such compounds. These groups of associated
compounds appeared as characteristic patterns after the seed extracts
had been chromatographed or subjected to ionophoresis on paper. In
the extracts of most, but not all, of the species examined the spots
forming the characteristic patterns were of comparable size and
intensity.

(Table 6-2 illustrates the grouping of Lathyrus species on the basis of
the patterns described by Bell.)

Another recent systematic study of plant amino acids is that
of Renter (1957) who studied the principal forms of soluble nitrogen
in various parts of sixty-six species representing forty-eight families
(Fig. 6-2). Renter did not exaggerate the systematic imphcations of
the work. He noted that since the substances considered were fre-
quently common metabolites of plants and animals, their relative
quantities rather than strict presence or absence were of most signif-
icance. In some species the principal amino acids in various parts of

Cotoneaster multiflora

• '

Carya amara

Dicentra eximia

®V®0

Robinia pseudo-acacia

Bowieo volubilis

(;^

^
i_M

#J6

Nymphaea hybrida

Fig. 6-2. Patterns of amino acicis of storage organs of several plant species not considered
to be closely related:

1. Citrullin 6. Serine 39. Piperidine-2-carboxylic acid

2. Glutamine 9. Alanine 44. Azetidine-2-carboxylic acid

3. Asparagine 11. Gamma-aminobutyric acid 45. Delta-acetylornithine

4. Glutamic acid 12. Proline

5. Aspartic acid 17. Arginine (Reuter, 1957)

102

BIOCHEMICAL SYSTEMATICS

^
g

<Xi
o

T3
O

"o
&.

C<3

'T3
O

o
o

CO S

c
'2

+ + +

+ ^

p.LHL^tHPH^Hh'H ■ ■ ■

^_^.-|__|- + + + + + • • ■

+ + ^ + + ^^HHHH + + + + + + + +H • ■ ■

•4

+ + + + + + + + + + + +

+ + + + + + + + + + +

iliiil It i nil! llllif

kj k4 ,^' Kj Kj ^ k4 ^' Kj i^' Kj ,^' kj ^' Kj Kj 1^ Kj 1^ 1^ k4 k4 ^4

d
2

^(Mco^iocot-oooio-H ^^;2Jt25S^223Scsc^?3

p
O

,-1 (M

AMINO ACIDS

103

HHHHHH^^

+ + +

H B-

H H H P H ^ H

0)
03

H +

+ +

+ + + +

+ + + + +

+ +
+ +

+
+

H ^

^ ir^

+ + + + + + + + + +
+ + + ^J + ^ + ^^

^: +

+ + +

+
+ + +

+
+

+ + + + + + + + + + +

+ + + + + + +

+ + + + + + +
^ + + ^ + ^

+ + ^ H H

■^

a
.1

3
•;3

•2

03
3

s^
3
5J
C

3
C

§ « S

03 -CJ

3 p

03
03 5

ft) 03
03 C

k -C o ;i ~

c
Si. a

i^^q Kj^Kci^i^Kq^Kq^qK^^ K^K^Kq^K^Kj^

S^ Kj ^ ^ ^ 1^ ►^

3
O

(M CM

(M(M(M<MCOCOCOCOCO00CO

i> 00 03 o ^ oa CO

CO CO CO T}< Tf -5)" -^

■^ lO CC I> 00 03 o
T)> Tt rr -^ Tf -<* lO

104 BIOCHEMICAL SYSTEMATICS

the same individual differed. Thus arginine was found to predominate
in the lower stem in ash (Fraxinus), but moving toward the stem
apex glutamic acid, asparagine, and glutamine in the order listed rep-
resented the main amino acids of corresponding positions.

The three acids found in high concentration in the stem apex
(namely, glutamic acid, asparagine, and glutamine), plus asparatic
acid, are consistently among the predominant amino acids of nu-
merous plant species. These acids are probably among the first com-
pounds into which amino nitrogen is incorporated. Asparagine is the
amide of aspartic acid which, in turn is derived via transamination
from the Krebs' cycle acid, oxaloacetic acid. Similarly, glutamine is
an amide of glutamic acid, also derived via reductive amination of
another Krebs' cycle acid, a-ketoglutaric acid.

Because of the metabolic position of these acids no special
significance is attached to their prominant occurrence in a large num-
ber of species. Therefore, statements such as that by Korohoda et al.
(1958) that glutamine, glutamic acid, and alanine are most character-
istic of the genus Brassica have very little systematic significance.
However, in certain cases less common amino acids appear to be con-
sistently prominent in a family and thus are characteristic of the
metabolism of that family. Some examples taken from Renter (1957)
are discussed below. It should be borne in mind, however, that rapid
changes in amino acid content may accompany development, and the
concept of "principal amino acid" should not be applied too vigorously.

Arginine, illustrated below, possesses the guanidine group:

R— NHCNH2
NH

CH2NHCNH2

CH2

CHNH2

COOH

Arginine is often present as a principal amino acid in the
families Saxifragaceae, Hamamelidaceae, and Rosaceae, but its occur-
rence elsewhere as a principal amino acid is sporadic. It is noteworthy
that other guanidine compounds are conspicuous in the family
Rosaceae (Reuter also reported the presence of certain guanidine
compounds) and in the one member of the family Hamamelidaceae
which was tested, namely, Parrotia persica. Only a few members of
the family Saxifragaceae were tested, and these were negative for
guanidines other than arginine.

AMINO ACIDS 105

Citrulline, as indicated by its formula, is closely related to
arginine. It was originally reported in the family Cucurbitaceae and
is present in several members of the family. It rarely appears in
Reuter's table except in the families Juglandaceae and Betulaceae
wherein it is the chief acid in all eight genera tested. The former
family is usually placed alone in the order Juglandales while the lat-
ter family is included with the Fagaceae in the order Fagales. Since
citrulline is absent in the three genera of the Fagaceae tested, the dis-
tribution of this compound (as a principal amino acid) is significant.

CH2NHCNH2
CH2 II
CH2
CHNH2
COOH

citrulline

A third acid, 6-acetylornithine, is probably the outstanding
example cited by Renter since it is restricted in the Papaverales to
the sub-family Fumarioideae of the Papaveraceae where it was found
to be present as the chief amino acid among all of the nineteen
species (representing four genera) examined. It was not found
as a principal amino acid in any of fifteen genera of the sub-family
Papaveroideae tested although a small amount is present in Hylome-
con, Chelidonium majus, and Glaucium flavum. In a later section,
additional biochemical data bearing on the relationship between the
Papaveroideae and the Fumarioideae will be presented and discussed
(Chapter 9). 5-Acetylornithine also occurs in ferns (Asplenium) and
grasses. Unusual amino acids which occur in widely separated taxa
provide potential opportunities to study analogous enzymes or anal-
ogous biosynthetic routes as defined in a previous chapter, but
they are not likely to have any direct systematic use at higher taxo-
nomic levels.

CH2NHCCH3
CH2 II
CH2
CHNH2
COOH

d-acetylornithine

A fourth acid, proline, which was once considered to be some-
what rare, is now reported from a number of different species in the
Leguminosae. Additional genera of other families, which have a high
proline concentration, are Taraxacum (Compositae), Mahonia (Ber-
beridaceae), Eleagnus (Eleagnaceae), Tamarix (Tamaricaceae), Phello-
dendron (Rutaceae), Ailanthus (Simarubaceae), and Moras (Mora-
ceae).

106

BIOCHEMICAL SYSTEMATICS

CH2 — CH2

CH2 CHCOOH
H

proline

Finally, azetidine-2-carboxylic acid which is a lower homolog
of prohne, seems to be typical of the Lihaceae where it is of rather
widespread occurrence. This fact is not evident from Reuter's lists,
but Fowden and Steward (1957a) reported the presence of this acid
in seventeen of fifty-six genera of Liliaceae tested. Renter also shows
it as the principal amino acid in roots of Convallaria majalis (some-
times treated as a separate tribe [Hutchinson, 1959] or family [Gates,
1918] of the Liliaceae).

/^.

CH2 CH2COOH

\ V
N
H

azetidine-2-carboxylic acid

Renter has described a scheme of probable inter-conversion
for the acids just considered which, if correct, suggests a rather close
biochemical affinity for all. It is not hkely that the distributions of
these acids are of great taxonomic importance in themselves. The
scheme of probable inter-conversion (shghtly modified from Renter)
appears below:

glutamine y, citrulline -^ arginine

glutamic -^ glutamic semialdehyde -^ ornithine^

^ proline 6 acetylornithine

aspartic -^ aspartic semialdehyde ^ azetidine-2-carboxylic acid
asparagine

When two acids such as prohne and azetidine-2-carboxylic
acid probably arise by analogous reactions of precursors differing by
a single carbon, comparison of the enzymes involved should prove
interesting. There is some hkehhood that two enzymes responsible
for such equivalent reactions are structurally related, perhaps even
phylogenetically related (homologous). The same enzyme may catalyze
both reactions, of course. Such a phenomenon exists in valine-iso-
leucine synthesis in Neurospora.

A particularly interesting study of plant amino acids is that
of Fowden and Steward (1957) who studied the amino acids of
eighty-nine species representing fifty-six genera of the Family

AMINO ACIDS 107

Liliaceae (actually six genera were in the family Agavaceae and three
in Amaryllidaceae, but all of these had at some time been placed in
the family Liliaceae). Eighteen amino acids were of relatively com-
mon occurrence and need not be discussed beyond mention of the fact
that methionine and histidine were notable by their absence. A total
of fifty-four ninhydrin-positive but unidentified spots were detected.
Most of these spots were restricted to one or at most only a few
species, but little can be said concerning systematic implications of
these distributions beyond the fact that they probably represent a
reservoir of biochemical information to be utilized in the future.

Eight amino acids, representing a series of recently dis-
covered types, all of which were identified, exhibit rather interesting
patterns. Five of these represent an apparently related group, all of
which may be conceived as derivatives of glutamic acid.^

CH2
Y-Methyleneglutamic acid HOOCCCH2CHCOOH

NH2
CH2
Y-Methyleneglutamine H2NOCCCH2CHCOOH

NH2

CH3
Y-Methylglutamic acid HOOCCHCH2CHCOOH

NH2

CH3
y-Hydroxy-Y-methylglutamic acid HOOCCCH2CHCOOH

OH NH2

Y-Hydroxyglutamic acid HOOCCHCH2CHCOOH

OH NH2

y-Methyleneglutamic acid was reported from seven genera:
Tulipa, Erythronium, Haworthia, Lilium, Notholirion, Fritillaria,
and Calochortus. The acid is apparently characteristic of Tulipa
wherein all species tested contained it. y-Methylglutamic acid was
found in six genera: Tulipa (most species), Erythronium, Lilium,
Notholirion, Calochortus, and Puschkinia.

3 Y-hydroxyglutamic acid, having a different carbon skeleton, may be excluded.
The fact that its distribution pattern is also distinctive as opposed to the others of the
group is then significant.

^ Qg BIOCHEMICAL SYSTEMATICS

y-Methyleneglutamine was reported only in the two genera,
Tulipa and Erythronium but was present in all species of Tulipa

examined.

y-Hydroxy-y-methylglutamic acid was found as traces in six
genera including Tulipa (many species), Erythronium, Littonia,
Lilium, Calochortus, and Puschkinia. y-Hydroxyglutamic acid was
found in only two genera: Hemerocallis and Gasteria.

The association of these unusual amino acids among certain
related genera is of taxonomic significance. It is especially interesting
that Calochortus contains two of the acids since Ownbey (1940) has
stated that "the relationship of the genus [Calochortus] as a whole,
although remote, is probably rather with the genus Tulipa.'' Recently,
Buxbaum (1958) established this genus as the single member of the
new tribe, Calochorteae. However, Hutchinson (1959) retains Calo-
chortus and related genera {Erythronium, Fritillaria, Tulipa, Lloydia,
Gagea, Notholirion, Lilium, Nomocharis, and Giraldiella) in the
more inclusive tribe Tulipeae. Using the amino acid criteria alone it
would appear that Ownbey's comments are especially significant, and
intensive chemical studies should contribute significant data to estab-
lish the phylogenetic affinities of the genera.

Two other acids, not derivatives of glutamic acid and hence
not included in the natural group above, had a rather restricted dis-
tribution within the Liliaceae. One of these, hydroxyproline, was de-
tected only in Dracaena (which is now placed in the family Agava-
ceae by many workers, Hutchinson, 1959). Another, azetidine-2-
carboxylic acid, has already been discussed. Finally, pipecohc acid,
the next higher analog of proline, appeared in nine genera, including
Hosta, Haworthia, Fritillaria, Chionodoxa, Hyacinthus, Muscari,
Smilacina, Convalleria, and Maianthemum.

CH2CH2CH2CH2CHCOOH

N '

pipecolic acid

These distinctive glutamic acid derivatives seem to occur
sporadically among widely separated taxa (for example, peanut, ferns,
phlox). Fowden and Steward state that this distribution implies the
genetic factors responsible do not operate at the generic or specific
level; that is, "in short, the accumulation of any of these compounds
may be determined by relatively few of the genes that characterize
the organism." In other words, since the compounds are not re-
stricted to a genus or species, their synthesis could not therefore de-
pend upon the specific association of a large species-dehmiting gene

AMINO ACIDS 109

pool. A generation of studies of genetic control by biosynthesis has in
fact aheady established this principle.

Fowden and Steward concluded from their study that numer-
ous metabolic pathways, previously unexpected, existed. This conclu-
sion certainly appears to be valid, since many yet unidentified
compounds exist, and these offer further promise for comparative
biochemical studies.

Birdsong et al. (1960) have reported on the distribution of the
guanidine, canavanine, an amino acid found thus far only in the
family Leguminosae. Within the family its appearance seems to be of
definite taxonomic significance.

NH
ONHCNH2

H2
CH2
CHNH2
COOH

canavanine

Prior to this study a total of sixty-eight species representing
thirty-one genera had been analyzed by various workers, and
Tschiersch (1959) was of the opinion that since canavanine appeared
somewhat randomly in the family, its distribution had no taxonomic
significance. Extension of the number of investigated species disclosed,
however, that canavanine occurs only in the sub-family Papilionoi-
deae, and of the tribes of that sub-family, it does not occur in Pody-
larieae and Sophoreae and is apparently rare in the tribe Genisteae.
Canavanine is particularly common in the tribes Trifoheae and
Loteae; all of seventeen species in these two tribes analyzed by Bird-
song et al. contained canavanine. Przybylska and Hurich (1960) have
reported the canavanine distribution in a few additional species, but
the pattern of distribution is maintained. The lack of canavanine in
the tribes Podylarieae and Sophoreae is interesting because inde-
pendent chromosomal evidence suggests that these tribes are offshoots
from the main Papilionoid stock (Turner and Fearing, 1959).

There is considerable circumstantial evidence that canavanine
is an important metabolite in those plants in which it occurs; for
example, it may be important in the storage and transport of nitrogen.
If this is true, the distribution of canavanine should be more vigorously
controlled by selection pressure, and therefore its distribution should
have greater significance (that is, subsequent loss of ability to form
canavanine would have negative survival value). The Birdsong et al.

1 1 BIOCHEMICAL SYSTEMATICS

study illustrates clearly the fact that in some cases a small sampling
may not disclose a pattern of the distribution of a substance.

Another point of interest in the canavanine work is that cer-
tain large and diverse genera such as Vicia, Astragalus, and Glycine
contain some species with canavanine present and some without. As
noted in an earlier section, it is possible that those species lacking
canavanine have, in certain cases, lost only one enzyme, specific for a
single step. Therefore, two different species could have complemen-
tary deficiencies. Hybridization could produce individuals capable of
forming canavanine. Canavanine would then appear as a hybrid sub-
stance in such a case. More will be said later about the formation of
"new" substances in hybrids.

It now appears that another newly discovered amino acid,
like canavanine, is restricted to the Papilionoideae of the family
Leguminosae. This unusual acid, which gives a brilliant scarlet color
with ninhydrin, was identified from Lathyrus tingitanus seeds by
Bell (1962). It was reported to occur in only a few of many species of
Lathyrus examined. We have now examined nearly 300 species in the
family Leguminosae for the presence of lathyrine, and the acid has
been detected only in several species of Lathyrus. Within the genus
Lathyrus this amino acid is likely to be of definite systematic utility.

CH,CH

\00H

lathyrine

Alston and Irwin (1961) have reported on the relative variation
in free amino acids and secondary substances in five species of Cassia.
They noted that, although definite differences did appear in the
amino acid chromatograms of different species, the extent of variation
was far less than that of fluorescent substances. For example, from
ten to twelve amino acid spots appeared, no more than nine of which
were present in a single species. Superimposed upon the relative
limitation in numbers of amino acids which are readily disclosed from
crude extracts is the fact that the quantities of free amino acids
present at a particular time tend to be quite sensitive to numerous
external and, presumably, internal influences, for example, light and
temperature, nutritional conditions, stage of development, and so on.
Fowden (1959) has observed that certain amino acids such as histi-
dine, tyrosine, cystine, and methionine are rarely detectable unless

AMINO ACIDS

•v^

(*•

#0*

0-'

V-

3 LU

\5'

f i

^-T

_aj

_>^

-w

'+J

S-i

CL,

-w

-a

r-"^

-M

^4-1

°C

"a,

-M

£_,

-M

-*^

-M

b£

-2

'-t-i

S
0)
73

-Q

-a
c

Si
M

o
CO

O
c«

X
0)

1 1 2 BIOCHEMICAL SYSTEMATICS

the plant is in circumstances which encourage protein breakdown.
Additionally, the amides, glutamine and asparagine, are extremely
sensitive to modifications of plant growth. Pleshkov et al. (1959)
compared the free amino acid content of corn leaves and roots grown
in a complete nutrient medium and grown in media minus nitrogen,
phosphorus, or calcium, respectively. With prolonged deficiency in
each case the amino acids decreased sharply, the greatest decrease
being evident in aspartic and glutamic acids, alanine, serine, and gly-
cine. In contrast to the nitrogen deficient plants, which responded by
a drop in free amino acid content within twenty-four hours after re-
moval from the complete medium, the phosphorus and potassium
deficient sets showed a slight increase in the amino acid content
during the first week, followed by a rapid decrease with prolonged
deficiency. Possingham (1956) found that in tomato plants cultured
in media deficient in copper, zinc, manganese, or iron, the free amino
acid fraction actually increased while in molybdenum deficient plants
the free amino acid fraction decreased. When molybdenum was
added, there was a rapid upswing in free amino acid content inter-
preted to reflect the role of molybdenum in nitrate reduction and
nitrogen uptake (Possingham, 1957). Although qualitative differences
were not great, an example such as the appearance of pipecolic acid
in iron and manganese deficient plants, while absent in controls, is
notable. A similar situation reported by Coleman (1957) occurs in flax.
In this plant citrulline, not previously found in flax, occurs in mod-
erate concentration in sulfur deficient plants. Possingham (1956)
noted that the relative amounts of amino acids changed in a different
pattern with each type of deficiency. The systematic implications of
these last observations are that, if one is interested in discovering
whether the enzyme system leading to the production of a substance
is present and not merely whether the plant normally accumulates
the substance, exposure to various types of physiological stress may
provide the opportunity in some cases.

There are other more recent studies, similar in principle to
those of Possingham cited above; for example, Tso and McMurtrey
(1960) found that, in general, mineral deficiencies other than N caused
an increase in the free amino acids of tobacco plants and variations
in the relative concentrations of amino acids. ^ Such evidence serves
to support the concept that apparently metabolically labile sub-
stances such as amino acids provide less reliable data than do meta-
bolic end products which accumulate. ^

4 Mineral deficiencies also affect the accumulation of other groups of substances,
for example, alkaloids, anthocyanins, and so on, but generally not so directly, hence as
quickly, as the common amino acids.

AMINO ACIDS

Trifolium hybridum

T. pratense

113

T. repens

200

100

sc80

.VN

sc 10

sc50

^^^y

60 6-5 70

Vicia Faba

V. sativa

Glycine Soja

sc 15

sc 10

sc 40

Fig. 6-4. Electrophoretic mobilities of major seed globulins
(Daniellson, 1949). Reproduced from The Biochemical Journal,
with permission.

In our comparison of the free amino acids of the seeds of
Baptisia species we have noted very consistent results within a
species and in fact quite similar patterns among all of the species
examined to date. Thus, if free amino acids vary greatly during de-
velopment as indicated and furthermore are easily affected by the
environment, as indicated, the seeds at least provide a rather stable
base for the analysis of developmental changes. We have found that
even the patterns of the free amino acids of the stem, leaves, and
flowers of Baptisia species are predictable and reliable although
quantitative differences certainly occur. The patterns of free amino
acids of the stem, leaves, and flowers are generally quite similar in
B. leucophaea (Fig. 6-3).

Although most work has been devoted to single amino acids
it is now evident that a variety of peptides may exist, and these may
prove, eventually, of considerable taxonomic importance (Virtanen
and Matikkala, 1960; Wiewiorowski and Augustyniak, 1960; Carnegie,
1961). Aside from the tripeptide, glutathione, and a recently dis-

5 It should be noted, however, that appropriate populational sampling for chroma-
tographic study should reduce the disadvantage of much of the individual variation which
might occur in nature.

114

BIOCHEMICAL SYSTEMATICS

covered tetrapeptide, called malformin (Takahashi and Curtis, 1961),
reported from Aspergillus niger, very little work on specific higher
peptides is available.

Haas (1950) reported on the peptides from four species of
marine algae. The breakdown of amino acids derived from the hydrol-
ysis of these algal peptides follows:

Algal Species

glycine

alanine

arginine

histidine

aspartic

glutamine

Griffithsia

—

ftosculosa

Pelvetia

—

canaliculata

P. canaliculata

—

f. libera

Corallina

—

officinalis

Since these peptides do not occur in detectable amounts in
summer months, Haas proposed that lack of light in winter interferes
with normal protein synthesis leading to the formation of mixtures of
peptides (intermediates?).

It is possible that at some future time alteration of normal
metabolism by exposure to stress will disclose abnormal but sys-
tematically enlightening metabolic pathways; that is, the accumula-
tion of substances normally found only in small amounts, such as
citrulline in flax described earlier, if enzymatically controlled, would
provide clues to relationships. Hoffman (1961) observed that several
species of the green alga, Oedogonium, could not be distinguished
chromatographically until cultures were allowed to remain in stale
media. Under these suboptimal conditions, a number of additional
compounds then appeared, some of which were species-specific.

In the older botanical literature there are a number of studies
of seed proteins. Some of the results are suggestive of taxonomic
affinities. In general the studies are of rather slight value because the
proteins are characterized somewhat crudely. One of the more recent
comparisons of seed proteins (of grasses and legumes) is that by
Danielsson (1949). His work consisted essentially of studying the
globuHn fractions by ultracentrifugation. The information yielded by
such techniques relates to the number of major globulin types, their
relative molecular weights and their relative abundance. The data
may be expressed in the form of a graph (Fig. 6-4), each peak
representing a component and the height of the peak its amount. To

AMINO ACIDS 1 1 5

the right is the higher molecular weight. From the standpoint of
systematics the legume data appear to be more interesting. Two
major globuHn components, vicilin and legumin (mol. wts. 186,000
and 331,000, respectively), are of widespread occurrence and two
others were detected in certain species of Leguminosae.

Some of the curves obtained by Danielsson have been repro-
duced to illustrate the nature of the information in these tests. In the
figure of Vicia faba (Fig. 6-4, lower left) two peaks are distinct. The
peak at left represents vicilin, that at the right legumin. In contrast
most species oi Acacia show legumin either weakly or not at all. With
the possible exception of Trifolium repens, Acacia is the only genus
showing so little legumin.

In general, the curves of related species tend to show similar
relative proportions of vicilin and legumin in their seeds. In the closely
related genera Lathyrus and Vicia legumin always predominated,
while in the less closely related Phaseolus, vicilin predominated. The
distributions of the minor components seem not to be amenable to
any systematic interpretation.

While this work is of interest, it is doubtful whether, in its
present form, a large diversity of critical data may be acquired.
Although the patterns definitely seem to bear resemblance at the
generic level, they are not likely to succeed in clarifying taxonomic
points in question or solving problems of phylogeny. Gerritsen (1956),
however, has obtained amino acid analyses of five highly purified seed
globulins of lupines, three globulins from Lupinus angustifolius, and
two from L. luteus. The highest molecular weight protein of each
species appeared to be identical; the next in size showed similarities
but also definite differences. The smallest (mol. wt. ca. 25,000) had no
counterpart in L. luteus. This type of investigation would seem to
offer much promise particularly at the intrageneric level.

Blagoveshchenskii (1960), who also studied the seed proteins
of various legumes, inferred from his results that in the "primitive"
species alkali-soluble proteins predominated while the contents of
albuminus and vicilin were low. In more "advanced" species vicilin
predominated over legumin, and the content of alkali-soluble proteins
was low. His illustrations were not very clear, however, and are there-
fore difficult to evaluate.

Sibley (1960) has utihzed electrophoretic patterns of egg-
white proteins in an extensive study of over 650 avian species. He has
assumed that the electrophoretic patterns are representative, in part,
of the genie complement of the species. It is indeed notable that
serologically related substances are found in the embryonic and adult
blood sera. The curves obtained from the egg-white proteins of

116

BIOCHEMICAL SYSTEMATICS

o >

Q -

1 1 — r 1 r — r 1 1 1 1 r"

C>4 O CO >0 TT CN

1 1 — I 1 1 1 1 1 1 1 r-

C-M ^ CO so 'J- CN

O <->

T 1 1 1 1 1 1 1 1 1 T"

Ol ^ eo -O TT CM

— I — I — I — I — I — 1 — I — I — I — I — r-
CM ^ oo >o ^1- og

1 1 1 1 r 1 1 1 1 1 T"

CM ^ OO so ^:}- CVI

CM a>

+ p

— I — r — I 1 — I — I — I — I I I

CM C3 OO -O "3- CM

AMINO ACIDS

117

r I I I I \ — I — I — I — I — r-

t^ O OO vo 'S- CVI

p >-

"1 — I — I — I — I — I — I — I — I — f — r

^ OO >0 T^ CNI

-"a- ^5

— I 1 1 1 1 1 — 1 r-

OO *0 ^T CM

O .S
O ^

— I — I —

CSI ^

— I — I — r-

CO NO

T 1 1 1—

"^ CSI

—I — 1 — I — I — I — I — r"

OO nO •^- CSI

s s

118

BIOCHEMICAL SYSTEMATICS

birds are considerably more complex than those obtained from
seed proteins (Fig. 6-4). It would be interesting to compare the results
of this approach with the serological data obtained by Boyden and
others (Chapter 4). There is, however, some danger of circular reason-
ing whenever one engages in apriori deductive speculations concerning
the systematic value of a character which is studied only superficially.
This statement, which is not intended as a condemnation of Sibley's
methods, needs some clarification. Sibley points out that the proteins
are of particular significance since they are more or less direct gene
products. While no one would dispute the general principle, the fact
remains that electrophoretic data yield only patterns. They are
accordingly a cumulative expression of the protein complement and
do not provide evidence as to the particular structure of particular
proteins. Therefore, it seems that the force of the argument is lost,
and it is a mistake to assume that the electrophoretic patterns
are more incisive indicators of phyletic affinity than the morpho-
logical pattern also evoked by those same agents (the genes) perhaps
more indirectly. It is unnecessary to reaffirm the conviction that
intimate knowledge of protein molecular structure is of profound
phylogenetic importance and the statements above are not intended
to refute this hypothesis.

FATTY ACIDS

The naturally occurring fatty acids, at least those
to be found among the higher plants , provide an
apparent mild paradox, insofar as their systematic
implications are concerned. For example, Hilditch
(1956), in his comprehensive treatment of the
chemistry and distribution of natural fats, makes
the following statement:

The fatty (glyceride) compounds of seeds are
specific and closely related to the families in which
the parent plants have been grouped by botanists.
It is, indeed, not an exaggeration to say that the
component acids of seed fats could themselves be
made the basis of a system of classification of
plants.

119

120

BIOCHEMICAL SYSTEMATICS

Despite the preceding statement, there are no cases known to
the writers in which data concerning the fat composition of species
were appHed to the solution of a taxonomic, or more specifically,
a phylogenetic, problem. There are not even any contributions which
attempt to infer relationships from such data, in contrast to, for
example, Hegnauer's treatment of the isoquinoline alkaloids (Chapter
9). It is apparent that Hilditch was not necessarily implying that a
natural system of classification could be constructed out of the distri-
bution of seed fats. This may be ascertained from inspection of
Table 7-1. For example, in Group A, whose major component acids are
linoleic, linolenic, and oleic, families in the gymnosperm order
Coniferae and the angiosperm families Juglandaceae, Labiatae, and
Oenotheraceae among others are included; in Group D, whose major
component acids are palmitic, oleic and linoleic, families such as
Gramineae, Magnoliaceae, Solanaceae, and others are included; and in
Group K, whose major components are stearic, palmitic, and oleic

Table 7-1. Distribution by family of some fatty acids (Meara, 1958).

Major Component Acids

Family

(A) Linoleic

Celastraceae

Moraceae

Linolenic

Coniferae

Oenotheraceae

and/or

Elaeagnaceae

Passifloraceae

Oleic

Juglandaceae

Rhamnaceae

Labiatae

Valerianaceae

Linaceae

(B) Linoleic

Amaranthaceae

Oleaceae

Oleic

Asclepiadaceae

Papaveraceae

Betulaceae

Pedaliaceae

Capparidaceae

Plantaginaceae

Compositae

Scrophulariaceae

Dipsacaceae

Staphyleaceae

Fagaceae

Theaceae

Hippocastanaceae

Typhaceae

Myrtaceae

Ulmaceae

Olacaceae

Vitaceae

Cucurbitaceae

Oleic

Euphorbiaceae

Rosaceae

Linolenic

Elaeostearic

Licanic

Ricinoleic

FAHY ACIDS

121

Table 7-1. (Continued)

Major Component Acids

Family

(D) Palmitic

Acanthaceae

Gramineae

Oleic

Anacardiaceae

Lecythidaceae

Linoleic

Anonaceae

MagnoUaceae

Apocynaceae

Malvaceae

Berberidaceae

Martyniaceae

Bombacaceae

Menispermaceae

Caprifoliaceae

Rubiaceae

Caricaceae

Rutaceae

Caryocaraceae

Solanaceae

Combretaceae

Tiliaceae

Families Elaborating Seed Fats Containing Characteristic Fatty Acids

Major Component Acids

Family

(E) Petroselinic

Araliaceae

Oleic

Umbelliferae

and

Linoleic

(F) Acetylenic

Simarubaceae

Tariric

[Picramnia sp.)

Octadecenynoic

Olacaceae

(G) Eicosenoic

Olacaceae

Buxaceae

(Oleic, Linoleic)

(Ximenia sp.)
Sapindaceae

(Simmondsia sp.)

(H) Erucic

Cruciferae

Oleic

Tropaeolaceae

Linoleic

(I) Cyclic unsaturated

Flacourtiaceae

acids

(J) Arachidic

Leguminosae

Lignoceric

Moringaceae

Oleic

Ochnaceae

Linoleic

Sapindaceae

(K) Stearic

Gnetaceae

Guttiferae

Palmitic

Burseraceae

Meliaceae

Oleic

Convolvulaceae

Sapotaceae

Dipterocarpaceae

Verbenaceae

(L) Laurie

Lauraceae

Simarubaceae

Myristic

Myristicaceae

Ulmaceae

Palmitic

Palmae

Vochysiaceae

Salvadoraceae

^22 BIOCHEMICAL SYSTEMATICS

acids, the families Gnetaceae, Verbenaceae, and Dipterocarpaceae
are among those included. Throughout the twelve groups listed,
families with few or no affinities are placed together, though in some
instances two or more families which are related on the basis of their
classical treatment occur together.

Obviously, attempts to read phylogenetic implications from
this pattern of distribution of fatty acids will avail nothing. Yet, fatty
acids may prove useful to systematics since a fairly large number
of fatty acids are known to have limited distribution (for example, the
cyclic unsaturated acids of the Flacourtiaceae). However, two factors
must be considered to bear upon the assessment of the systematic
significance of the distribution of fats. There are about the same num-
ber of fatty acids known as there are amino acids, and many of the fatty
acids, like many amino acids, are widely distributed. Secondly, as
Hilditch has pointed out, in the formation of typical triglycerides the
glyceride structure tends to be dependent upon the proportions of the
various component acids. In other words the enzymatic esterification
of the fatty acids with glycerol appears to be of low specificity so that
the distribution of fatty acids in glycerides tends towards the maxi-
mum degree of heterogeneity. The effect of this is that a full range of
variation, for a given fatty acid complement, is permitted. If two
species have a similar fatty acid complement, they will both produce
a similar fat complement. Doubtlessly, there are exceptions to this
generalization, and more needs to be known about the precise control
of fat synthesis.

One of the difficulties of the classification of Hilditch
illustrated above is that it is based upon major component acids
which in the majority of instances are of exceedingly wide occurrence.
For example, oleic acid is included in nine of the twelve groups. It is
not hkely that the distribution of this acid offers much to chemo-
systematics. Another difficulty hes in the fact that since the groups
are based on major components, not absolute distinctions, one is not
measuring the presence of a given metabolic pathway but rather cer-
tain favored pathways. In effect it is not always clear what is actually
being measured. Concentration on the unusual or the rare fatty acids
is likely to prove more profitable.

Excellent treatments of the chemistry and distribution of
fatty acids are available in the recent hterature (Hilditch, 1956;
Meara, 1958). No such comprehensive account is included here and in
fact much of the information in these works is not strictly relevant to
systematics. However, certain basic considerations of the chemistry
of fatty acids, specifically the major variations, and a brief description
of certain features of their biosynthesis may be useful in providing

FATTY ACIDS ] 23

heightened perspective from which to evaluate their systematic
significance. For example, it is pertinent to consider whether all fatty
acids are synthesized via one or via several basic biosynthetic routes.
Very little is known about the genetics of fatty acid synthesis, so one
potentially valuable aspect of the subject is temporarily obscured.
Finally, certain examples will be selected to illustrate the association
of specific fatty acids with particular genera or families. Table 7-2 in-
cludes a list of fatty acids.

Fatty acid biosynthesis

In recent years the pathways involved in fatty acid metabo-
lism have been rather well established, and it is now evident that
fatty acid metabolism is linked directly to the oxidative breakdown
of carbohydrate at the point of the formation of acetyl coenzyme A.
Presumably all aerobic organisms possess the ability to form acetyl
CoA. Since all fatty acids appear to be constructed from two-carbon
units supplied from acetyl CoA, this mechanism accounts for the
overwhelming predominance of even-numbered fatty acids found in
nature. The first step in the building up of fatty acids is assumed to
be a condensation between two molecules of acetyl CoA:

2 CH3C— S CoA — > CH3CCH2C S CoA + CoA SH

Within the last few years a slight modification of this step has
been recognized as a result of the demonstration that CO2 is essential
for fatty acid build up, yet CO2 does not become incorporated into the
fatty acid. To satisfy these requirements CO2 is postulated to combine,
with acetyl CoA to form malonyl CoA as an intermediate:

O
C— OH

(1) CH3CSC0A + CO2 — . CH2CSC0A

acetyl CoA malonyl CoA

O COH

CHo

(2) CH3CS CoA + CH2CS CoA * CH3CCH2CS CoA + CoA SH + CO2

acetyl CoA malonyl CoA acetoacetyl CoA

:z;
o

<;
o

c
_o

>> o

O 05

02 &

-a

T3
CO

O
O
O

s
O

CO
P

I-H

Q
<

Q
Pi

^oiSioooooooo

o^' §§§00000000
oF. ooooqqooqoq

3
S

■^

CI ■ TP O X r-l ^^ i-t iH r-^ (N C^ ^^

ci PC a CI CI CI cj CI C) CI ""c? cj ci

KOKKEKEKKKKKK

o^ooooooooo oo

CO ""T"

CO HH cococococorococofococo

rt
Z

G
O

6
o

o G

o o

5 fe

o^ooooooooooo

^ o _w

'o -^ o

C >. G

-M -1^ ix!

3 a> 01

e CO c e

-y §1
o «
C o o

« « s

=3 ^ §■§

2 2 -a

O OJ o

1-1 S-. s^ Co o ?^

o 8

G c

O o

2 CO

OQQFhKOHQHK

I I I I I I I I I I

O 01

3 O

a &, a 3

>, .a .y s -a -^ JG

S .2

cc " _2 c s

G O CO CO CO co^'coi^ SoJ.Sfo

PQ-2 000h-lScLHa}<mH:i:

cjcjocicicacjcicj

^^^j:'000000000

OCl^OClClCICOCO'^'tTj'iO

KHHHHHHpHpCPnhlHpHpHpHpnpH
MHHHHH OClTt'OXOCa-^CD

'fm':D00rH,-*rH--HrHC«]C^]OlC4

ooooooooooooo

124

o
o
o

Q s

^ o

c
0)

c
o

CO ^

3 3

M §

o o

a:
o
o
o

X X

o o
o o
o o

I — I X X

r^O

M . •

X X

o q

X X

o o

CI CO

X X

o o

X X

q q

t- CO

1—1 r— I

Cl o

X X

o o

CO CO

X X

.. o o

oic
enoic
-5-en
-9-en

<? 01 CO CO

^ "^ -^ -^

Q Q P P

CJ O O CJ

o o o o

X CJ ^ ;d

.-H C) CJ CJ

K tn III K

O Cl f rf

d o d CJ

-J—

-f—

-i—

"— r'

-4—

-i-

4—

■i—

-1—

:H:CH.[i
.CH(OH
H:CH.[C

-J—

.COOHt
.COOHt
.COOH

I.COOH
3. COOH
5.COOH
9.COOH

r— 1

1—1

X X

i-r-l

n c) CJ M CJ CJ ci

K K E E K X K

K K K

C^4

CJ U
K K K E X K ffi

X X
q

X
q

ffi ii K \
q /

^^ ji^ 5; 32 ^IH Jj. 5. JjH

>u q q q q q q q

ffi

1—1

t—

■*

0^1

f 1

r— 1

(— (

C^J

«
^

E K E K K ffi

X X X X X X X

•— 7J

1 — 1

'—7'

l— 7*

' — ^

^-^

'—7'

I— 7*

K ffi K E K K
CJ

hJh hh hh hh h^ hh hh

0000000

'0
c

■3

-a

1—1

»— 1

2
a

-3

1— t

ca
S

Hexadec-9-enoic
Octadec-9-enoic
Octadec-6-enoic

12-Hydroxy-octa
9-Hydroxy-octad

a.
W

CO
1-H

1— (

T— 1

1— f

Docos-U-enoic
Docos-13-enoic
Tetracos-15-enoi
Hexacos-17-enoi

'0
c
>>

-C
+J

'o
c
>>

a)
-a

"2
'o
c

£1

JS
+J

1
!>

0)
13

<S
.M

6
«

c
o

s
a

o
O

3 3

0) E
o o

-U

S-.

Ji>

CL,

CLl

ffi K K K ffi ffi

:o CO GO 00 00 00

c5 do

125

3
o

Si
O)

c
o

-3-^5.

o " Jrt c S
0) s-H oj --^ 3
O H W X hJ

CJ CJ CM CM CJ CM CJ

0000000

CO » CJ CJ to O C30
CO CO ■T ■^ ■^ lO lO

X X X X X X X

O O C^ CJ "^ f^O o
CJ CM CJ CJ CM CJ CO

0000000

ffi

i-H

2:;

u
a

I
s

C
^3

o
o

II ^

o— o

o
u

3)
T3

I
I

-(J

o
o

II (N

o
v

•13

X
O

o
o

II H,

X X

O— o

o
c
o
I

CD
I

<D
T3

X
O
O
O

O
o

o
c

X
O

o
o

o
q

X
u

X
o

-S
of

cB
O

-a

CO
o
O

o
o
o

s
o

o
o
o

o
o

CI -t— E

K X O
o o ^
■ 'O o

o o

X o

X iS'

X X
o o

K X

q q

X X
U U

<J<' X

O^.
h-i X

X X
o o

^J^ '^ <~ -r

X X

X
o

ii
o

X X

o o

X X

o o

X X

o o

K K K

XXX

y y y

K K K

o o o

XXX
o o o

o o

EKE
o o o

.s w

o
c

2 S

'3
c

o
c
a;

'. CO

2" (>f ^. -^'

CD of

c«

-O

-a

-M

CJ
Ol

-o

o
o

o
c

CL,

o
c

o
ed

00
o

■■e

1— c

C
O

o
O

a; C
I — I ^
o o

£4

■B

-n

o o

126

steari
ic

bO ^

8 §

is C

■— « -rt

X ;j

a J

C^ CJ

CJ M CO

00000

csl ro

ro ro ca

ffi K K K ffi

CD X

CO X CO

000

E E

o o

!^ ^n

1—^ c^

^ G

K ^

4—

X
o

-2

-t

-r

5-1

01
Ol

o o .a

O
o

O O H Q

-w

■-^

• 1— 1

o
c

-M

a
o

2 o

c c

S IS

- "^ X

cc S qj

0) CO ca
03

H Q E

O S 03

cj H X

o 15 <u

Q e- K

~ — '

03 _a

"3 2

J g

o o

S ^

^ « 2

03 -tj >H

P-i cc <

W (M (N

o o o

CO X w

OJ (TJ CO

XXX u

00 00 o

o o o t::

•a

■*

iTO

XXX

C^l

CO P

CM CM CM

o o o

CM to o

CO CO ■*

ffi HI ffi

CM ■* ^

C'J CM CM

o o o

127

] 28 BIOCHEMICAL SYSTEMATICS

Following this initial condensation, reduction occurs, and the
ketone group is eliminated. These reactions are accomplished in three
steps, two of which involve electron transfer from the pyridine
nucleotides (DPNH and TPNH). These steps are illustrated below.

?" II

(3) CHaCCHsC^S CoA p^'^" + "^ > CH3CHCH2CS CoA + DPN

OHO O

(4) CH3CHCS CoA — » CH3CH=CHCS CoA + H2O

O O

(5) CHaCH^CHCS CoA '^^''" ^ "^ > CH3CH2CH2CS CoA + TPN

Degradation of fatty acids proceeds by the stepwise removal
of two-carbon units via a pathway which is essentially the reverse of
that described above. There are some differences however. For ex-
ample, the initial oxidative step, corresponding by analogy to step
5 above, in reverse, is mediated by the coenzyme flavine adenine
dinucleotide (FAD). A second point of difference is that it does not
appear that the temporary binding of CO2 to form the malonyl
derivative is involved.

The significance of the mode of fatty acid synthesis described
above is that it represents an almost universal basic metabohc path-
way. Therefore, all of the various fatty acids are metabolically related,
and the variations in chain length, in degree of unsaturation, and even
those involving terminal cyclization, are secondary.

A question of some theoretical importance is that of whether
a single enzyme of low specificity is involved in the condensation of
malonyl CoA with the preformed carbon chain or whether several
enzymes, each with affinities for a carbon of particular length, may
cooperate in the build up of a sixteen carbon fatty acid. In fact, it is
possible that a different, specific enzyme exists for coupling of
malonyl CoA to C2, C4, Ce . . . C„ residues. There is insufficient
evidence on this point, to the writers' knowledge, to provide a general
statement. Likewise, it is not yet known whether any specific coen-
zymes participate in the oxidation-reduction steps involving different
carbon chain lengths, or even whether similar coenzymes but different
apoenzymes participate. Crane et al. (1955) have shown that three
enzymes are active in the first oxidative step in the degradation of
fatty acids in pig hver, and their specificities differ for different car-
bon chain lengths. All three of these are fiavoproteins. As indicated in
Fig. 7-1, enzyme Yi is most active on C8-C12 fatty acids, Y2 on Cg-

FATTY ACIDS

129

Fig. 7-1. Specificity of fatty acyl CoA dehydrogenases for sub-
states of different chain length (Crane et al. 1955).

Ci6 fatty acids, and G (which is probably equivalent to butyryl CoA
dehydrogenase) is most active on C4 acids. The three enzymes can
thus effect the degradation of fatty acids up to sixteen carbons. More-
over, these workers discovered a fourth enzyme which is specific for
oxidation of the reduced forms of G, Yi, and Y2. The last, also a
flavoprotein, was designated an "electron transferring flavoprotein."
It presumably gives up electrons to some intermediate in the basic
electron transport system.

In connection with the problem just posed above, a statement
of Hilditch (1952) is pertinent. Hilditch noted that the fatty acids of
more primitive plant and animal forms tended to represent a more
complex mixture, with a simpler mixture characteristic of more ad-
vanced organisms. This could represent the evolution of enzymes with
more specificity effecting more vigorous control over chain length in
fatty acids. Some authors, notably McNair (1941), have considered
that there is an increase in the molecular weight and the complexity
of the fatty acids during the course of evolution, a viewpoint which
is in part opposed to the idea of Hilditch, cited above. Goldovskii
(1960) has criticized both of McNair's premises:

Owing to the great diversity of chemical reactions, the process of fatty
acid formation from its very inception must have led to a muhitude of
acids (polycondensation always leads to a number of polymer homol-
ogues). In fact, the simplest lower plants, in particular the algae and
fungi, already possess a complex equipment of fatty acids, including
high molecular ones. Nor can we agree with the idea of a rise in the
degree of unsaturation in the course of evolutionary development as a
whole, since even in the algae acids of a high degree of unsaturation
are formed. And, on the other hand, seed fats in the Compositae, the
members of which are generally taken to be at the summit of the

] 30 BIOCHEMICAL SYSTEMATICS

evolutionary development of plants, are by no means distinguished by
having the highest iodine values, as might be expected if it is thought
that the degree of unsaturation increases during evolution.

Another problem which must be solved in the metaboKsm of
fatty acids is the enzymatic coupling of the acid to form an ester
linkage with glycerol. If this reaction is rather unspecific, then what
prevents the incorporation of shorter chain "intermediates" into
triglycerides? A related question is that of the mechanism for ter-
minating the extension of the carbon chain. One reaction has been
reported which provides a partial answer to both questions. This re-
action involves the acyl attachment from an acyl CoA group to
phosphoglyceric acid with the liberation of CoA. The reaction
proceeds more efficiently with sixteen- and eighteen-carbon acyl CoA
compounds and is probably an intermediate in phosphatide (for ex-
ample, lecithin) synthesis. In any case, it is clear that the presence of
such enzymes could result in the capture of fatty acids of appropriate
chain length as they are synthesized. Frequently, it appears that a
given species tends to synthesize saturated fatty acids possessing two-
carbon atoms more or less than that of the major component (Hil-
ditch, 1952), again suggesting the possibility of less than absolute
specificity for certain enzymes governing fatty acid synthesis. Perhaps
this area would be particularly fruitful for comparative enzyme
studies-both from the standpoint of their catalytic properties and
their absolute chemical constitution (that is, amino acid sequence).

It is evident from the review by Meara (1958) that, except for
the seed fats, relatively little systematic significance can be gleaned
from analysis of plant fats. For example, fats of roots usually rep-
resent minor components. In a few cases, notably the sedge Cyperus
esculentus, the oil content may be high. The fat content of bark is
usually on the order of 3 per cent and in the few species examined the
major fatty acid was the common oleic acid. More frequently, fruit
coat fats may acquire a relatively high concentration of oil. Yet, ac-
cording to Meara, "the characteristics and component fatty acids of
most of the fruit coat fats, irrespective of their botanical family, are
very similar."

It has generally been believed that leaf lipids and the con-
stituents of leaf surface waxes are rather similar among different plants
(Hilditch, 1956). However, since the advent of chromatographic tech-
niques including gas chromatography, much evidence has been ac-
quired indicating that the leaf waxes contain a quite diverse
assemblage of hydrocarbons of different lengths including branched
chains, alcohols, aldehydes, ketones, acids and esters. Even a new

FATTY ACIDS

131

carbon-methyl flavonoid component of wax is suspected (Price, 1962).
For separation of alkanes, the techniques of gas chromatography and
mass spectrography are utihzed together.

Purdy and Truter (1961) have compared the surface hpids of
leaves of sixty three species, using thin layer chromatography to sep-
arate mixtures. Characteristic patterns were obtained for each species.
It was also demonstrated that the patterns did not change with the
age of the plant.

The work of Eglinton et al. (1962) is of unusual interest and in-
dicates that wax constituents may be of exceptional taxonomic value.
These workers chromatographed unfractionated extracts of alkanes,
using only 40 g. of dried samples, and obtained a complete analysis of
straight chain and branched chain alkanes of 23 to 35 carbons. Some
of their results are shown in Fig. 7-2. The variety of patterns dis-
closed by their data suggests wide application of these methods in
taxonomy.

The patterns of alkanes derived from individual species are
apparently quite constant. A study of Aeonium urbicum (Crassulaceae)
collected from various places and including immature and even dead

Crassulaceae

-il4

Scrophularicceae Euphorbiaceae

Gramineae

Liliaceae

Solonaceae

Cruciferoe

t 1 1 1 1

1 1 I 1 r— 1 1 1

1 ^''

— n t

|-|

-1 1

-| r-

n 1

limn

n r

n 1

22 24 26 28 30 32 34

22 24 26 28 30 32 34
Carbon No.

22 24 26 28 30 32 34

22 24 26 28 30 32 34 22 24 26 28 30 32 34

n n— alkane

■ Branched chain alkane

Fig. 7-2. Distribution in mole percentage of n- and branched
chain alkanes C22-35 in the hydrocarbon fraction of the waxes
from individual plant species. Alkane percentages less than 2 mole
per cent have been omitted (Courtesy G. Eglinton).

132 BIOCHEMICAL SYSTEMATICS

leaves showed the pattern to be quite stable and characteristic for the
species (Eglinton, 1962). However, the techniques also served to dis-
tinguish species of the same genus. Four species of the genus Hebe, in
which hybridization frequently occurs, were compared and according
to Eglinton et al. the four species can be immediately distinguished
chemotaxonomically, e.g. Hebe odora has a major constituent of C29;
H. parviflora and H. diosmifolia, C31; and H. striata, C33. An entirely
unsuspected source of major chemical variation immediately acces-
sible to analysis is therefore disclosed by these investigations.

The remainder of this section will be concerned with a few
selected examples of fatty acids whose distributions are restricted to
or characteristic of certain plant families.

FLACOURTIACEAE

Chaulmoogric (Cie) and hydnocarpic (Cig) acids occur in this
family together with lower homologues in trace amounts. The distinc-
tive characteristic of this group of acids is the presence of a cyclo-
pentene ring. Although present in a number of genera of the Flacour-
tiaceae they do not occur outside the family. It is therefore evident
that the distribution of these cyclic acids is of taxonomic interest.

Fatty acids with cyclic groups occur also in the genus Sterculia
(Sterculiaceae). Thus sterculic acid (C19) possesses a three-membered
ring inside the chain— produced, possibly, by addition of a carbon and
yielding an uneven number of carbons. Subsequently, Shenstone and
Vickery (1961) have reported that, in addition to Sterculia and
Brachychiton of the Sterculiaceae, certain Malva and Gossypium
species (Malvaceae) produce the cyclopropene acids, sterculic acid and
malvalic acid. Both families are placed in the order Malvales.

CRUCIFERAE

In this family the unsaturated acid, erucic acid (C22), is quite
prominent. According to Meara (1958) it is probable that many, if not
most Cruciferae, contain erucic acid. Only a small proportion of the
total species has been subjected to detailed analysis, however, and
the statement is based partly on inferences derived from the low
saponification number of the fats from a larger number of crucifer-
ous species.

Outside the Cruciferae, the nasturtium {Tropaeolum minus),
of the monogeneric family Tropaeolaceae, contains a large quantity of
erucic acid. Since the Tropaeolaceae are usually placed in the Ger-
aniales and show no obvious phylogenetic affinity to the Cruciferae,
it is likely that the high erucic acid content in the two families is
coincidental. It is interesting that certain relatively uncommon iso-
thiocyanates occur in the two families (Chapter 14).

FAHY ACIDS 1 33

UMBELLIFERAE

A structural isomer of oleic acid, petroselinic acid, is appar-
ently confined to a few families including the Umbelliferae where it
represents a major component (for example, 75 per cent of the total
fatty acids in Petroselinum sativum) in most of the species tested. In
addition to the Umbelliferae, petroselinic acid has been found in ivy
{Hedera helix) of the related family, Argdiaceae. However, one other
plant known to produce the acid as a major component is Picrasma
guassioides of the family Simarubaceae. This family is not closely
allied with the UmbeUiferae though Meara (1958) states that petrose-
linic acid is confined to the members of the Umbelliferae and one or
two members of isolated but related species. It is significant that
tariric acid has been reported only from the genus Picramnia of the
family Simarubaceae. Petroselinic acid is an octadec-6-enoic acid
while tariric is an octadec-6-ynoic acid.

PALMAE

According to Hilditch (1952), "In both constancy and com-
plexity of the mixture— the seed fats of the palm family form the out-
standing instance of specificity of fatty acid composition within a
single botanical family."

Within this family lauric and myristic acids (C12 and C14)
usually make up 50 per cent or more of the total fatty acid
complement.

In addition shorter chain acids, such as caproic, caprylic, and
capric, are generally present along with palmitic, stearic, and the un-
saturated acids, oleic and linoleic. This group of acids represents re-
markable diversity with respect to carbon chain length, and in addi-
tion it is indicated that the relative proportions are rather constant
among different species and even among various genera within
the family.

The fatty acids of the Palmae demonstrate the fact that
chemical characters may be somewhat constant within a taxonomic
group dehmited on other grounds. This has been shown in a number
of other studies involving other compounds, of course, and the fatty
acids of the Palmae are distinctive only in that they involve the oc-
currence of rather constant proportions of a series of compounds. Un-
fortunately, the fatty acid complement of the Palmae does not
provide much insight into relationships within the family or with
other families.

Hilditch, a chemist, has obviously recognized that fatty acid
metabolism and morphological characters are often correlated, and
the following statement (Hilditch, 1952) illustrates at once the promise
and the pitfalls of phytochemical systematics:

1 34 BIOCHEMICAL SYSTEMATICS

Most interesting and least understood of all is the manner in and by
which species with common morphological relationships produce
qualitatively the same mixtures of fatty acids in their seeds so that
classification of species according to the constituent acids in their seed
fats leads to much the same results as that developed by the botanical
classifications of Linnaeus and his successors.

The statement quoted might elicit some justifiable criticism
from systematists, most of whom envision the goal of systematics to
extend beyond the mere convenient categorizing of species. A natural
classification of species based solely on the constituent acids in their
seed fats would bear no relationship to any present taxonomic system.
While the palms would be placed into a single group, other categories
would include gymnosperms, monocots and dicots together.

CARBOHYDRATES

Judging from the relatively small body of available
literature it seems that the potential contribution
of carbohydrates to biochemical systematics is
sKghtly regarded, although some early work on the
biochemistry of carbohydrates from the systematic
aspect exists. For example, Blackman (1921) dis-
cussed in rather general terms the use of carbo-
hydrates as phylogenetic criteria, citing specifically
the accumulation of pentosan mucilages in succulent
families such as the Cactaceae and Crassulaceae.
Blackman cited an older work by Meyer, who
arranged the flowering plants into five "classes" on
the basis of their propensity to form starch. Gen-
eralizations such as a tendency for most monocots
to fall into the low starch-producing classes or that

135

1 36 BIOCHEMICAL SYSTEMATICS

the families Solanaceae and Leguminosae form starch in large amounts
are not of great value in themselves. Blackman also discussed the
monumental work on starch grains by Reichert (1919) which had
appeared shortly before and which will be discussed in a later chapter.
In his comparative study of starch grains, Reichert exposed numerous
possibilities for the use of starch characters in systematics.

Carbohydrates are so diverse that one cannot relegate them
in toto to the general categories of either basic metabolites or sec-
ondary constituents. Not only simple hexose sugars such as glucose
and fructose but also the more recently familiar sugars such as ribose
and sedoheptulose must now be considered as basic metabolites. In
contrast, the sugars that represent the glycosidic portion of such
secondary substances as pteridines, steroids, flavonoids, cyanogenetic
principles, some oligo- and polysaccharides, and sugar derivatives
such as the sugar alcohols, are collectively more properly considered
as secondary products of metabolism.

Previously, it was stated that secondary constituents
promise generally to be more useful than basic metabolites in sys-
tematic investigations. However, a basic metabolite, sedoheptulose,
once considered to be restricted in occurrence to the family Crassul-
aceae, is now found in a number of other families (in the free, non-
phosphorylated state). Since free sedoheptulose has a restricted
distribution, it is potentially systematically useful. The previous gen-
eralization concerning basic metabolites is supported, however, as
most of the simple sugars which are involved principally in energy
relationships, for example, glucose, fructose, maltose, sucrose, and
others are so widespread that they can hardly be expected to have a
phyletically meaningful distribution. Some sugars and sugar deriva-
tives which appear to be quite restricted in their occurrence may
actually have a much broader distribution than expected. As in the
case of sedoheptulose, screening of large numbers of species is neces-
sary to expose meaningful patterns.

Probably many more simple sugars (up to the oligosaccharide
level of complexity) and sugar derivatives exist than is generally
appreciated. Because of the lack of any integrated treatment and the
encyclopedic effort required to bring together data on all the different
substances described to date, the present treatment is necessarily in-
complete. An excellent review of the carbohydrate literature is
available (Shafizadeh and Wolfrom, 1958). In the following brief
account, the main purpose is to identify some of the lesser known
sugars and sugar derivatives to illustrate further opportunities for
phytochemical systematic investigations. In appropriate situations.

CARBOHYDRATES 1 37

phyletically meaningful distribution of a substance is noted, but in
general at this time carbohydrates have not contributed nearly as
much to the field of biochemical systematics as have certain other
classes of compounds.

Simple sugars (a partial list only)

The five-carbon sugar, xylose, is frequently encountered in
higher plants and is a constituent of polysaccharides such as xylan
and the hemicelluloses. Other pentoses, such as arabinose and ribose
are frequently found in higher plants. Ribose is associated with co-
enzymes and nucleoproteins, and it is an intermediate in the dark re-
actions of photosynthesis. It would be indeed remarkable if ribose
were absent from a higher plant. Free pentoses are often present in
detectable quantities in plant tissues. Wilhams et al. (1952), who
examined thirty-one different food plants, and also leaves from
twenty-three trees, found free pentoses in the fruits of only the lemon
and strawberry, but in the leaf material examined, free pentoses were
present in about half of the species.

Of the hexoses, most are of little or no systematic significance.
Although it is true that the abihty to accumulate large amounts of
one or another of the common hexoses may conceivably be systemat-
ically meaningful. There is no evidence that such accumulation often
occurs. A chromatographic study of free sugars of twenty-seven
famihes of seed plants and ten species of algae representing three
phyla (Bidwell et al, 1952) showed some differences in the relative
concentrations of glucose and sucrose. Since sugar concentration may
be as sensitive to conditions of growth as amino acids, quantitative
differences must be evaluated conservatively. These authors reported
extremely low free sugar content in the algae tested, and a similar
statement appears in a report on carbohydrate accumulation by lower
plants by Young (1958).

Free galactose, though uncommon, has been found in a variety
of plants, and it also occurs in the sugar component of certain complex
glycosides. Mannose apparently does not often occur free, yet the
sugar is present in polysaccharide form (mannan) in certain palms
and orchids.

Although free fructose is of slight systematic importance, the
distribution of the fructose polysaccharide, inulin, appears to have some
taxonomic significance. It is quite widely known that certain Compo-
sitae (for example, dandelion, Jerusalem artichoke, and Dahlia) store

1 38 BIOCHEMICAL SYSTEMATICS

inulin. Related, but different fructose polysaccharides are apparently
found in some grasses. According to Hegnauer (1958) inulin has been
found in the family Boraginaceae, and Bacon (1959) found inulin in
species of the Campanulaceae. Hegnauer suggests a further study of the
distribution of fructosans in the plant kingdom, and Bacon infers from
the presence of inulin in the Campanulaceae a close evolutionary con-
nection between the family and the Compositae. A comprehensive
survey of the Compositae for inulin would be valuable. Sporadic
occurrences of inulin in two families is not necessarily indicative of
a close phylogenetic relationship between the families.

Heptoses have been known to occur for many years, but it
was probably Calvin's work on the dark reactions of photosynthesis
that led to recognition of the biological role of certain seven-carbon
sugars. Mannoketoheptose is found as a free sugar in the avocado
pear. Free sedoheptulose was first reported from Sedum spectabile by
LaForge and Hudson (1917). Subsequently, the presence of free
sedoheptulose has been associated with the family Crassulaceae, and
originally the sugar was considered to occur in the free state only in
this family. For example, Nordal and Klevstrand (1951a; 1951b)
found free sedoheptulose in all five sub-families of the Crassulaceae.
However, Nordal and Oiseth (1951, 1952) examined Primula elatior,
P. vulgaris and P. veris (Primulaceae) and found sedoheptulose and
probably mannoheptulose. These were the first reports of sedohep-
tulose outside of the family Crassulaceae. The search for sedohep-
tulose in P. elatior was prompted by the identification of the seven-
carbon sugar-alcohol, volemitol, from the species. Nordal and Oiseth
reported the detection of sedoheptulose in several species of Saxifraga-
ceae; Chrysosplenium alternifolium, Parnassia palustris, and several
species of Saxifraga were positive for sedoheptulose. Williams et al.
(1952) also reported heptuloses in sixteen of thirty-one plant tissues ex-
amined, and twenty-one of twenty-three species of trees had one or
more heptuloses in the leaves. Recently, Brown and Hunt (1961)
examined 200 species of plants representing seventy-eight families and
found free sedoheptulose in at least one species of each of sixteen fami-
lies. It was prevalent in the families Corylaceae and Oleaceae, but in
general no clear systematic implications outside of this observation
were apparent. Thus, while association of the sugar with the family
Crassulaceae alone is no longer valid, and this particular phytochemi-
cal systematic correlation is negated, perhaps in a broader framework
the distribution of sedoheptulose will be phylogenetically significant.
In view of the basic metabolic significance of phosphorylated sedohep-
tulose, it is not surprising to find that free sedoheptulose occurs
outside of a single family.

CARBOHYDRATES ] 39

Structurally modified sugars

Rhamnose, a methylpentose, is frequently present as a part
of the glycosidic component of certain flavonoids. It does occur as the
free sugar, for example, in leaves and flowers of Rhus toxicodendron.

Fucose, deoxygalactose, is found in the brown alga, Laminaria
digitata. Although botanical examples of the systematic utility of
fucose are lacking, there is an interesting example of genus specificity
involving fucose in the sea-urchin-egg jelly coat (Vasseur and Immer,
1949). The polysaccharide of Echinus esculentus jelly consists en-
tirely of galactose residues, that of Echinocerdium cordatum solely of
fucose residues, Strongylocentrotus draebachiensis, fucose with galac-
tose, and Paracentrotus lividus, fucose plus some glucose.

Numerous other deoxy sugars, occurring often as steroid
glycosides, are known (Shafizadeh and Wolfrom, 1958; Reichstein,
1958). Deoxyribose is of ubiquitous occurrence in nucleotides of DNA.

Branched chain sugars of higher plants are somewhat rare.
Apiose, illustrated below with hamamelose, are among authenticated
branched pentose and hexoses respectively:

HC=0

HCOH

HOH2CCOH

HOCCH2OH

HCOH

apiose

hamamelose

Hamamelose is found as a constituent of a tannin of Ham-
amelis virginica, but apiose has been identified as a free sugar in
leaves of the monocot, Posidonia australis (Potamogetonaceae). At
least it is the main sugar component following mild acid hydrolysis of
leaves, so it is to be considered as only conditionally a free sugar
(Bell et al., 1954). Previously apiose was known only as the parsley
flavone glycoside (the apioside of apigenin).

Sugar alcohols

(acyclic polyhydric alcohols)

Relatively few sugar alcohols have been identified as naturally
occurring compounds. Usually the compounds are reported from a
rather diverse group of plants. For example, galactitol (dulcitol) is
found in certain red algae, fungi, and higher plants; sorbitol is found
in algae, monocots, and dicots; mannitol, which is quite widespread,

140 BIOCHEMICAL SYSTEMATICS

occurs in some algae, fungi, lichens, Gnetaceae, and numerous mono-
cot and dicot species. Others, such as volemitol, polygalitol, and
styracitol are known in only one or a very few species.

In spite of their rather broad distribution, mentioned above,
within certain delimited taxonomic groups the sugar alcohols may
have a meaningful distribution. For example, Plouvier (1948) has
shown that galactitol occurs in branches, bark, and leaves of numerous
species of the family Celastraceae, while it is absent from the related
families, Rutaceae, Simarubaceae, Meliaceae, Rhamnaceae, and
Vitaceae. It also occurs in the Lauraceae {Cassytha filiformis),
Scrophulariaceae {Melampyrum species), and Hippocrateaceae {Pris-
timera indica). The last example is interesting in view of the fact
that Bentham and Hooker, and others, have included members of this
family in the Celastraceae.

Among algae galactitol seems to be restricted to the red algae.
In contrast, mannitol is often found in brown algae but not, appar-
ently, in the reds. Quillet (1957) found mannitol in seventeen species
of brown algae, sometimes comprising up to 50 per cent of the dry
weight. Volemitol was present in one species, Pelvetia canaliculata.
Cmelik and Marowic (1950) found mannitol in Adriatic species of
Cystosura, Sargassum, Laminaria, Dictyopteris, Fucus, and Padina
with a maximum accumulation at the beginning of winter. However
at no season was mannitol obtained from red or green algae. Actually,
the authors say that "practically" no mannitol was obtained from
red or green algae at any season. It is not clear whether or not
small quantities were actually detected in some. Large seasonal
variation in mannitol content sometimes occurs. Black (1948) noted
over fourfold differences in mannitol content of some Laminaria
species in Scotland, with the maximum concentration coming in mid
or late summer. In L. cloustonii the dry weight mannitol content in-
creased from 18 per cent at one-half fathom to 36 per cent at four
fathoms.

Among flowering plants, mannitol occurs rather widely, among
so many families that a significant familial distribution is unlikely.
Mannitol is exceptionally common, however, in the family Oleaceae.

Sorbitol has a more limited distribution among angiosperm
families, e.g., the Rosaceae. According to Barker (1955) if detached
leaves of certain Rosaceae are kept in the dark to eliminate starch and
then floated in a solution of sorbitol, starch is synthesized. The leaves
cannot utilize mannitol or galactitol to form starch. Similarly, leaves of
Adonis vernalis (Ranunculaceae) and certain species of the Oleaceae
will utilize dulcitol and mannitol respectively. It would be interesting
to repeat certain of these experiments using methods of modern tissue

CARBOHYDRATES 141

culture and with C^^ labelled sugar alcohols. One could utilize liquid
root-culture techniques or measure callus development from cotyledon
explants upon agar with sugar alcohols as the carbohydrate source.

Inositol and related cyclic alcohols

This group of substances is arbitrarily discussed along with
the other sugar alcohols. Empirically, they are close to the sugar
alcohols but biosynthetic pathways leading to the synthesis of cyclic
alcohols may be quite different.

The best known of the cyclic alcohols is the widely distributed
inositol. Meso-inositol occurs in numerous higher plants while other
isomers are more hmited in distribution.

H /OH H0\ OH
Ho\oH hAh

meso-inositol

Two monomethyl ethers of inositol, bornesitol and sequoyitol,
occur in several families. Another monomethyl ether, pinitol, is found
among a variety of conifers, in Ephedra, and also among a number of
angiosperm families (for example, Leguminosae). Dambonitol, a
dimethyl ether, occurs in a number of species of angiosperms, and the
deoxyinositol, quercitol, is found in a number of angiosperm families.

The cyclitols, pinitol and quercitol, have been investigated
from a systematic orientation (Dangschat, 1958; Plouvier, 1955).
Despite the fact that they occur among numerous unrelated groups,
within certain taxonomic groups the distribution of these compounds
is definitely meaningful.

Quercitol is found in a number of families of both monocots
and dicots. However Plouvier, who studied the distribution of quercitol,
found that quercitol generally occurred only in one or a few represen-
tatives of a family; that is, it did not appear to be particularly charac-
teristic of the family. In the family, Fagaceae, the compound was
described from the genus Quercus, but Plouvier found that several
species of Fagus and Castanea, of this family lacked quercitol. How-
ever, all of thirty-three species of Quercus investigated contained
quercitol.

Thus, quercitol appears as a genus-specific character in

142 BIOCHEMICAL SYSTEMATICS

Quercus, but not as a family character. In contrast, pinitol is signif-
icant at the family level. While pinitol is common in the Pinaceae, it
is not the pines that provide the example to be cited, for pinitol occurs
in other families of conifers. Even in the Leguminosae, where pinitol
apparently occurs throughout most of the tribes, it is not widespread
among the tribes Vicieae and Phaseoleae (Dangschat, 1958). However,
in the Caryophyllaceae pinitol approaches a family diagnostic charac-
ter. Plouvier (1954) examined forty-five species of this family rep-
resenting all sub-families (Paronychioideae: four genera, five species;
Alsinoideae: six genera, thirteen species; and Silenoideae: six genera,
twenty-seven species) and found pinitol to be present in forty-three of
the forty-five species. Repeated attemps to detect pinitol in Stellaria
media and Silene schafta were negative, so one could not consider
pinitol to be infallibly diagnostic. Yet, Plouvier states:

It appears as a constant chemical character of the Caryophyllaceae
providing biochemical homogeneity in this morphologically hetero-
geneous family; it establishes a connection between the three sub-
families particularly between the Apetaly and Dialypetaly which many
authors consider to be separate families.

Plouvier then extended his investigation to include twenty-
seven species belonging to related families, including Chenopodiaceae,
six genera; Amaranthaceae, four genera; Nyctaginaceae, two genera;
Aizoaceae, three genera; Phytolaccaceae, one genus; and Portulacaceae,
one genus. All of the above-mentioned families were included in the
Centrospermae of Engler and Diels (1936). In addition selected
famihes of the order Geraniales, which, according to some workers, are
phylogenetically close to the Caryophyllaceae were examined. These
included Oxalidaceae, one genus; Geraniaceae, three genera; Tropaeo-
laceae, one genus; Linaceae, one genus; and Zygophyllaceae, two
genera. Although the sample in each family was small, it is note-
worthy that only six of the twenty-seven species contained pinitol,
and it was absent from the families Amaranthaceae and Chenopodi-
aceae. The six positive species were: Nyctaginaceae, Mirabilis jalapa,
M. longiflora, and Bougainvillea glabra; Phytolaccaceae, Phytolacca
americana; Aizoaceae, Tetragonia expansa; and Zygophyllaceae,
Zygophyllum fabago (in low yield). Only the last-named famiily is
placed in the order Geraniales. Evidence with respect to the presumed
affinity between the two orders concerned is too limited to be
significant.

Since pinitol is shown to be present in several families of the
Centrospermae however, it appears that the utilization of pinitol as a

CARBOHYDRATES 1 43

link between the sub-families of the Caryophyllaceae is inappropriate
on the same objection as noted by Mothes and Romeike (1958) con-
cerning the isoquinoline alkaloids (to be discussed in Chapter 9);
namely, that it is not proper to relate a group of families on the basis
of the presence of a particular character and then, using that same
character, establish a hnk between various sub-families. In the case of
pinitol, no strong position was taken by Plouvier.

It is interesting to compare the systematic distribution of
pinitol with the systematic distribution of the betacyanins discussed
in Chapter 14. It may be noted that Hutchinson placed the Nyctagi-
naceae in the order Thymeleales which he derived from the Flacourti-
aceae of the Lignosae, while deriving the Chenopodiales from the
Herbaceae. The Nyctaginaceae, however, include a number of pro-
ducers of betacyanins which Reznik (1957) considers to represent a
diagnostic character significant at the ordinal level (Chapter 14). It is,
therefore, particularly interesting to note that all of the species of
Nyctaginaceae investigated by Plouvier contained pinitol. Curiously,
alone among the famihes of the Centrospermae, the Caryophyllaceae
apparently lack betacyanins yet it is this family in which pinitol is
most typical. The writers do not consider that the situation described
undermines in any way the validity of these biochemical data as
phylogenetic criteria because the distributions of pinitol and unusual
anthocyanin-like pigments in the Centrospermae, while exhibiting dif-
ferent patterns, at no point are in conflict, and they complement each
other with respect to the placement of the somewhat disputed family,
Nyctaginaceae.

Acid derivatives of inositol (not in a biosynthetic sequence
relationship) include quinic and shikimic acids. Quinic acid is of quite
general distribution. Shikimic acid until recently was thought to be
exceedingly rare (for example, Bonner, 1950, stated that at that time
it was reported only from species of Illicium). However, this acid, now
shown to be an intermediate in the synthesis of certain amino acids
as well as numerous secondary constituents (Chapter 11), is of general
occurrence. Hasegawa et al. (1954) detected shikimic acid in a num-
ber of species of angiosperms and gymnosperms.

Oligosaccharides

Tables 8-1 and 8-2 (Shafizadeh and Wolfrom, 1958) list the
typical disaccharides, the common oligosaccharides, and some of their
sources. Most disaccharides occur as glycosides and these often appear

Table 8-1. Constitution and natural origin of typical disaccharides (Shafizadeh
and Wolfrom, 1958).

Common name

Constitution

Origin

Trehalose

Sucrose

Inulobiose

Galactinol

Sophorose

Laminaribiose

Nigerose

Turanose

a-D-Glucopyranosyl
a-D-glucopyranoside

a-D-Glucopyranosyl
y8-D-fructofuranoside

a-D-Glucopyranosyl
a-L-sorbofuranoside

a-D-Glucopyranosyl
)8-D-^/ireo-pentuloside

1 -0-/3-D-Fructofuranosyl-
D-fructose

1-0-a-D-Galactopyranosyl-

D-mjo-inositol
2- 0-a-D-Glucopy ranosyl-

D-glucose

2-0-;8-D-Glucopyranosyl-

D-glucose
2-0-(a-D-Galactopyranosyl-

uronic acid)-L-rhamnose

2-0-a-D-Xylopyranosyl-

L-arabinose
2-0-(a-D-Glucopyranosyl-

uronic acid)-D-xylose
2-0-(4-0-Methyl-a-D-

glucopyranosyluronic acid)-

D-xylose

3-0-^-D-Glucopyranosyl-
D-glucose

3-0-a-D-Glucopyranosyl-
D-glucose

3-0-a-D-Glucopyranosyl-

D-fructose
3-0-;S-D-Galactopyranosyl-

D- galactose
3-0-a-D-Glucopyranosyl-

L-arabinose
3-0-a-D-Galactopyranosyl-

L-arabinose
3- O-yS-L- Arabinopyranosyl-

L-arabinose
3-O-a-D-Xylopyranosyl-

L-arabinose

Fungi, mushrooms, yeast, sea-
weeds, trehala manna

Most abundant sugar in plant

saps
Enzymic synthesis

Enzymic synthesis

Hydrolysis of inulin

Sugar beet
Enzymic synthesis

Sophoraflavonoloside from

Sophora japonica
Hydrolysis of mucilages from

slippery elm, flaxseed (Plan-

tago ovata) and okra
Hydrolysis of corn cob hemi-

cellulose
Hydrolysis of corn cob hemi-

ceUulose
Hydrolysis of corn cob hemi-

cellulose

Hydrolysis of Laminaria poly-
saccharide

Hydrolysis of amylopectin and
the polysaccharide from
Aspergillus niger

Hydrolysis of melezitose

Hydrolysis of arabic acid and
the gum of Acacia pycnantha
Enzymic synthesis

Autohydrolysis of arabic acid

Hydrolysis of e-galactan of

larch, and gums
Autohydrolysis of golden apple

gum

144

Table 8-1. (Continued)

Common name

Constitution

Origin

Hyalobiuronic
acid

Maltose

Cellobiose

Lactose

3-0-(/?-D-Glucopyranosyl-
uronic acid)-2-amino-2-
deoxy-D-glucose

4-0-a-D-Glucopyranosyl-
D-glucopyranose

4-0-/3-D-Glucopyranosyl-
D-glucopyranose

4-0-/?-D-Galactopyranosyl-

D-glucopyranose
4-0-a-D-Galactopyranosyl-

D-galactopyranose
4-0-j8-D-Xylopyranosyl-

D-xylopyranose
4-0-(a-D-Glucopyranosyl-

uronic acid)-D-xylopyranose
4-0-;8-D-Mannopyranosyl-

D-mannopyranose
4-0-(a-D-Galactopyranosyl-

uronic acid)-D-galacto-

pyranuronic acid
4.0-(4-0-Methyl-a-D-

glucopyranosyluronic acid)-

L-arabinopyranose
4-0-(^-D-Glucopyranosyl-

uronic acid)-D-glucopyranose

Isomaltose

6-0-a-D-Glucopyranosyl-
D-glucose

saccharides
Hydrolysis of amylopectin

Gentiobiose

6-O-yS-D-Glucopyranosyl-
D-glucose

Hydrolysis of gentianose and
glycosides of amygdalin and
crocin

Melibiose

6-0-a-D-Galactopyranosyl-
D-glucose

Hydrolysis of raffinose and in
plant exudates

Epimelibiose

6-0-a-D-Galactopyranosyl-
D-mannose

Hydrolysis of guaran polysac-
charide and epimerization of
melibiose

Galactobiose

6-0-a-D-Galactopyranosyl-
D-galactose

Hydrolysis of stachyose and
enzymic synthesis

Planteobiose

6-0-a-D-Galactopyranosyl-
D-fructofuranose

Hydrolysis of planteose

Hydrolysis of hyaluronic acid

Hydrolysis of starch

Hydrolysis of cellulose

Milk of mammals
Hydrolysis of okra mucilage
Hydrolysis of xylan

Hydrolysis of soluble hemi-
cellulose of com cob

Hydrolysis of guaran poly-
saccharide

Enzymic hydrolysis of pectic
acid

Hydrolysis of lemon gum

Hydrolysis of Types HI and
VIII Pneumococcus poly-

145

146

BIOCHEMICAL SYSTEMATICS

Table 8-1. (Continued)

Common name

Constitution

Origin

Rutinose

6-0-|8-L-Rhamnopyranosyl-
D-glucose

The glycoside rutin from Rata
graveolens

Vicianose

6-0-y8-L-Arabinopyranosyl-
D-glucose

The glycoside vicianin from
Vicia angustifolia

Primverose

6-0-y3-D-Xylopyranosyl-

D-glucose
6-0-(/?-D-Glucopyranosyl-

uronic acid)-D-galactose

The glycoside ruberythric acid
from madder root

Hydrolysis of the gums ob-
tained from many species of
Acacia

to have a restricted distribution taxonomically, sometimes in a sug-
gestive pattern (for example, gentiobiose is found among species of
the Rosaceae). More intensive investigation of the distribution of
specific disaccharides must occur before any evaluation of the sys-
tematic imphcations of their distributions can be made. The higher
ohgosaccharides are mostly products of partial hydrolysis of poly-
saccharides. However, several non-reducing ohgosaccharides are
known to occur in various parts of a number of different plant species.
The important non-reducing ohgosaccharides are rafiinose, planteose,
gentianose, stachyose, mehzitose, and verbascose.

In general there is little data available on the systematic
imphcations of the distribution of ohgosaccharides. MacLeod and
McCorquodale (1958) compared water-soluble carbohydrates of the
Gramineae and evaluated these substances as phylogenetic criteria.
Ohgosaccharides were among the sugars identified. These authors
were primarily concerned with the tribal disposition of certain genera.
Twenty-two species, representing eleven tribes, were analysed. Since
all of the species contained glucose, fructose, and sucrose, only the
more complex sugars provided any useful information. The authors
arranged the genera into six groups based on the presence of certain
types of oligosaccharides (Table 8-2).

Raffinose is a trisaccharide (galactose-glucose-fructose) while
stachyose, a tetrasaccharide, contains two galactose residues attached
to glucose of a glucose-fructose unit. This difference between the two
oligosaccharides may be regarded as minor. In fact, the authors note
that barley embryos infiltrated with concentrated raffinose solution
win form some stachyose though normally the sugar is absent.

CARBOHYDRATES

147

Table 8-2. Patterns of oligosides of the Gramineae.

Group Characteristic

Genera Represented

1. Hexoses only

Spartina

2. Homologous series of fructosans

Bromus

3. Fructosans, raffinose

Elymus
Agropyron

4. Raffinose

Glyceria
Phalaris
Nardus
Molinia

5. Raffinose and Stachyose

Brachypodium

Poa

Dactylis

Cynosurus

Arrhenatherum

Avena

Holcus

Anthoxanthum

Ammophila

Agrostis

Phleum

6. Isomer of raffinose

Festuca
Lolium

Nevertheless, some genera which contain much raffinose do not form
stachyose.

MacLeod and McCorquodale also compared hydrolysis prod-
ucts of the polysaccharides. These always yielded mostly glucose but
in addition, the pentoses, xylose and arabinose, were often present,
and the relative per cent of xylose and arabinose differed among the
genera. Finally, Nardus yielded 19 per cent mannose, and Molinia
yielded 23 per cent galactose.

General appraisal by MacLeod and McCorquodale of the
taxonomic significance of their findings was as follows:

(1) Nardus was distinctive in its high content of water-
soluble mannan.

(2) The Hordeae formed a natural group, but on the basis of
its content of soluble polysaccharides Hordeum itself is
rather distinctive from the other genera examined.

148

BIOCHEMICAL SYSTEMATICS

Table 8-3. Constitution and origin of typical higher ohgosaccharides (Shafizadeh
and Wolfrom, 1958)«.

Common name

Constitution

Origin

Maltotriose

Panose

Maltose
homologs

Cellobiose
homologs

Xylobiose
homologs

Raffmose

a-D-Gp-(l -^ 4)-a-D-Gp-(l ^ 4)-
D-Gp

a-D-Gp-(l -^ 6)-a-D-Gp-(l -^ 4)-
D-Gp

a-D-Gp-(l -^ 4)-[a-D-Gp-
(1 ^ 4)]2_3-D-Gp

p-D-Gp-il -^ 4)-[/8-D-Gp-

(1 ^ 4)]i_5-D-Gp

/S-D-Xylp-(l-^4H/?-D-Xylp-
(1^4)]i_5-y8-D-Xylp

a-D-Galp-(l -^ 6)-a-D-Gp-
(1 ^ 2) /J-D-Fru/

Hydrolysis of starch

Degradation of amylopectin
and enzymic synthesis from
maltose

Hydrolysis of starch

Acetolysis of cellulose

Hydrolysis of corn cob xylan

Sugar beet, cotton seed hull,
other plants

Planteose

a-D-Galp-(l -^ 6)-/3-D-Fru/"-

Seeds of various Plantago

(2^1) a-D-Gp

species and tobacco

Melezitose

a-D-Gp-(1^3)-i8-D-Fru/-

Mannas, honeydews, and

(2^1) a-D-Gp

exudations of several widely
different plants

Gentianose

j8-D-Gp-(l-^6)-a-D-Gp-

Rhizomes of several species of

(1^2)^-D-Fru/

Gentiana

a-D-Gp-(l -^ 4)-a-D-Gp-

Action of invertase on sucrose

(1^2)/S-D-Fru/

1-Kestose

a-D-Gp-(l ^ 2)-/3-D-Fru/-

Action of invertase and Asper-

(1^2);8-D-Fru/

gillus niger on sucrose

6-Kestose

a-D-Gp-(l-*2)-y3-D-Fru/-

Action of yeast invertase on

(6 -^ 2) ^-D-Fru/

sucrose

Neokestose

;S-D-Fru/-(2 -^ 6)-a-D-Gp-

Action of yeast invertase on

(1 ^ 2) p-v-Fruf

sucrose

Stachyose

a-D-Galp-(l -^ 6)-a-D-Galp-

Stachys tuberifera, soybeans,

(1^6)-a-D-Gp-(1^2)

ash manna, and various

j8-D-Fru/

plants

CARBOHYDRATES

149

Table 8-3. {Continued)

Common name

Constitution

Origin

Verbascose

a-D-Galjo-(l ^6)-a-D-Galp-

Roots of the mullein,

(1^6)-a-D-Galp-(1^6)-

Verbascum thapsus

a-D-Gp-(l-^2)/3-D-Fru/

Manninotriose

a-D-Galp-(l -^ 6)-a-D-Galp-

Hydrolysis of stachyose; in ash

(1^6)-D-Gp

manna

a-D-Galp-(l -^ 6)-/3-D-Man/j-

Hydrolysis of guaran

(1 ^4)-D-Manp

^-D-ManvD-(l ^ 4)-^-D-Manp-

Hydrolysis of guaran

(1 ^4)-D-Manp

a-D-GpA-(l -^ 4)-/3-D-Xylp-

HemiceUulose-B of corn cob

(1_^4)-D-Xylp

» The following standard abbreviations are used in this table. The monosaccharide radicals are represented by
the first three letters of their name, with the exception of the glucose radical, which is denoted by G. Furanose and
pyranose rings are indicated by / and p, respectively. The uronic acids are shown by the suffix A as in D-GpA which
indicates D-glucopyranuronic acid.

(3) The Bromeae formed a natural tribe quite distinct from the
Brachypodieae, Festuceae, and Hordeae.

(4) Two genera of the Festuceae, Lolium and Festuca, were
distinctive in containing an unusual trisaccharide.

(5) Two genera of the Aveneae, Anthoxanthum and Holcus,
differed from the other two tested {Avena and Arrhena-
therum) in that they lacked yS-glucosan. According to the
authors the taxonomic positions Anthoxanthum and
Holcus are slightly suspect on morphological grounds, and
their contents of soluble carbohydrate show affinities with
the tribe Agrostideae and to a lesser extent with the
Phalarideae.

With respect to the last statement, the biochemical data
applicable to the question of the relationship within the Aveneae lack
conviction. While over-all, the data appear to be suggestive, though
not conclusive, the rather small number of representatives of the
tribes sampled and the difficulty in appraising the biochemical
significance of the data raise doubts about their true phylo-
genetic significance. It is difficult to evaluate the significance of an
extra galactose unit on a trisaccharide with a terminal galactose or
the presence or absence of a soluble yS-glucosan. One is inclined to
suspect that the differences alone do not represent sound biochemical
criteria for adducing relationships at the tribal level and to predict

150 BIOCHEMICAL SYSTEMATICS

that such criteria would break down in an extensive survey of species.
There is no doubt that a basic problem in biochemical systematics in-
volves the need for greater insight in the appraisal of biochemical
data. The obvious fact is that some differences are apt to be more im-
portant than others.

There is an interesting paper, again involving the grass family,
by Belval and de Cugnac (1941) concerning "glucides" of Bromus and
Festuca. These glucides appear to be oligo- or polysaccharide in
nature. The type which is characteristic of Festuca is phlein, found
in several other grasses including Phleum pratense. Phlein is a fructo-
san, hence related to inulin. However, unlike the inulin fructosans
which represent 2, 1' glycosidic linkages, the phlein type possesses 2,
6' linkages. Fructosans are laevorotatory. Belval and de Cugnac
found that the specific optical rotation, (an), of the fructosan from
Festuca species, before and after acid hydrolysis, was —49 and —96°
while that of Bromus species was —37 and —84°. These values were
said to be quite consistent within the genus.

One questionable taxon, Bromus (or Festuca) gigantea, was
particularly interesting. This species has a glucide with the optical
rotation characteristic of Festuca. Lolium perenne and L. multi-
florum, which contain phlein, cross with Festuca, and the authors
suggest that the questionable taxon should be expected to cross with
certain Festuca species although it was not possible to cross it with
the two morphologically similar species, Bromus asper and B. erectus.
The authors also imply that it might also cross with Lolium. It is
pertinent to note that practically the entire argument in this case
rests upon the biochemical data. In Hubbard (1954) it is noted that
Festuca gigantea hybridizes with F. pratense and F. arundinacea,
and sterile hybrids may be obtained with Lolium perenne.

Natural hybrids of Festuca gigantea and other species of
Festuca as well as Lolium were known well before the Belval and de
Cugnac paper (for example, Jenkins, 1933). However, there was no
indication that the authors were aware of the work, and in principle
it does not detract from the significance of the biochemical data.

Polysaccharides

This class of substances is referred to generally as the glyco-
sans. Glycosans may be composed of pentoses (pentosans) or hexoses
(hexosans) or even mixtures of these. In general, if a single sugar
predominates in the glycosan the name is derived from the sugar in-
volved. Although a rather large number of plant gums and mucilages

CARBOHYDRATES 151

of polysaccharide character (or related to polysaccharides) are known,
and these may be obtained from leaves, stems, roots, and even flowers,
beyond a few generalizations (for example, their association with the
Leguminosae and specific genera therein such as Acacia and Astra-
galus), very little systematic importance is indicated for them at the
present time. It is still considered likely that certain gums are
synthesized by fungal or bacterial enzymes rather than via a metab-
olism strictly that of the host. This point is discussed briefly in the
comprehensive treatment of plant gums and mucilages by Smith and
Montgomery (1959).

Araban is composed of L-arabinose units. It is a common
constituent of pectic materials and is very widely distributed. Xylans,
whose chief constituent is xylose, occur in several forms, frequently of
1 : 4 y8-linked D-xylopyranose units in unbranched or branched chains.
It is highly probable that the specific xylans would prove to be sys-
tematically valuable, but as yet there is inadequate knowledge from
an insufficient number of species.

Galactans, which also comprise part of the pectin complex are
quite common. Arabogalactans are associated with the woods of
conifers, particularly various species of larch (for example, Larix
occidentalis). Mannans are widely distributed among higher and
lower organisms. Galactomannans are also known from a number of
species, and according to Neumuller (1958), they are associated
particularly with the family Leguminosae (for example, Medicago
sativa produces a galactomannan with a ratio of galactose to mannose
of 2 : 1). Glucomannans are known from several species of Amorpho-
phallus (Araceae) and seeds of certain Iris species.

Polyglucosides other than starch include such substances as
floridean starch and laminarin. The former is formed as the reserve
carbohydrate in red algae; the latter as the reserve carbohydrate in
brown algae. Floridean starch has recently been studied in detail by
Meeuse et al. (1960), who conclude that there is no basic distinction
between this starch and other polysaccharides of the starch family.
Floridean starch appears to have a branching pattern similar to that
of glycogen, that is with somewhat more frequent, shorter 1 : 6 side
chains than the amylopectin component of typical starch.

Algae also produce several other unusual types of polysac-
charides, some of which are httle known chemically. Recently, Stoloff
and Silva (1957) attempted to apply the distribution of particular
water-soluble polysaccharides to the phylogenetic treatment of sixty
species of red algae. The classification of the polysaccharides is based
mostly on physical properties. Three types, all of which occur esteri-
fied with sulfate residues attached to galactose units, were described:

]52 BIOCHEMICAL SYSTEMATICS

Agars These consist of two components: agarose, a

linear polymer of galactose and anhydro-
galactose, and agaropectin, a sulfated poly-
saccharide. Agars set to thermally reversible
gels.

Carrageenans These are also hexose-sulfate derivatives;
lambda carrageenan is a galactose sulfate and
kappa carrageenan is a mixture of anhydro-
galactose and galactose sulfate.

Gelans Strong gel formers similar in structure to

kappa carrageenan, but with a hexose-sulfate
ratio of about 0.5.

Stoloff and Silva found that all species of the same genus pro-
duced the same type of soluble polysaccharide. In general, as indicated
in their paper, there are few cases of more than one type of poly-
saccharide occurring in the same family. Exceptions are as follows:

Gelidiaceae Sulina produces gelan; four other genera,

agar.
Endocladiaceae Glocopeltis produces carrageenan; Endo-

cladia, agar.
Phyllophoraceae Gymnogongrus produces carrageenan;

Phyllophora and Ahnfeltia, agar.

Stoloff (1962) has reviewed the distribution of these poly-
saccharides and constructed a revised classification of the Florideae
on the basis of polysaccharide type alone. According to Stoloff the
taxonomist should "look up from his mounts and his microscope and
make fuller use of the technological advances in related disciplines. It
is not the tools but the viewpoints and objectives that should dis-
tinguish the botanist from the chemist or physicist." Later, Stoloff
says, "... at the familial level, and certainly at the generic level
of breakdown, the limits of usefulness of the evolutionary viewpoint
and the value to evolutionary theory seems to have been reached."
The present writers take a different view with respect to the lower
taxonomic categories, for it is in these that experimental methods,
cytogenetic, genetic and other macromolecular data are more appli-
cable. In any event it is not likely that the evolutionary viewpoint
will ever outlive its usefulness.

A type of polysaccharide which is somewhat difficult to
classify, namely, amyloid, has been studied intensively from the sys-
tematic point of view by Kooiman (1960a) who has examined the
seeds of many species of higher plants. Amyloids are complex poly-

CARBOHYDRATES 1 53

saccharides which yield glucose, galactose, and xylose. Partial acid
hydrolysis yields a product whose X-ray diffraction pattern resembles
cellulose. The main chain of the amyloids of different species is then
composed of 1-4 linked glucoses (Kooiman and Kreger, 1957). Enzy-
mological experiments with cellulase suggest that xylose and galactose
residues are attached as side chains in an undisclosed pattern. The
oligosaccharides derived by cellulase hydrolysis of a number of species
were the same, but their relative quantities differed (Kooiman, 1957).

The test for amyloid is a blue coloration of the amyloid solu-
tion upon exposure to I2-KI and sodium sulphate (Kooiman, 1960b).
Kooiman has tested over 2,500 species and finds certain families
which are general amyloid producers. For example, in the Legumino-
sae, the sub-family Caesalpinioideae are a particularly rich source of
amyloid. However, the positive species belong only to the tribes
Cynometreae, Sclerolobieae, and Amherstieae. Numerous genera in
these three tribes are amyloid containing. Galactomannan is frequently
encountered as a constituent of the endosperm in the tribes Cassieae
and Eucaesalpinieae.

Outside the Leguminosae amyloid was detected in sixteen
dicotyledonous famiUes, but no amyloid was found in the twenty-five
monocotyledonous families examined. Some noteworthy distribu-
tions follow. In the family Acanthaceae, of the ten species known to
produce amyloid all are in the tribe Justicieae. All Paeonia species in-
vestigated produced amyloid, but thirty other species of the family
Ranunculaceae were negative. All of the investigated taxa of the
order Primulales (including, according to the system of Engler, only
three families: Primulaceae, Myrsinaceae, and Theophrastaceae) were
found to produce amyloid. In connection with the latter observation,
it seems pertinent to mention that Hutchinson's arrangement of these
famihes differs considerably from that of Engler. Hutchinson places
the predominantly herbaceous Primulaceae and Plumbaginaceae in
the order Primulales and includes the woody Theophrastaceae, Myr-
sinaceae, and Aegicerataceae in the order Myrsinales; according to
Hutchinson these orders are in different phyletic groups. It is perhaps
premature to draw conclusions from the limited data available, but
it is difficult to ignore the striking amyloid distribution unless one
wants to assume convergence of both morphological and biochemical
characteristics.

Whatever the ultimate disposition of the famihes in question,
it appears likely that the carbohydrate chemistry of the groups will
play some contributory role, but much additional exploratory work
will be necessary before meaningful conclusions can be drawn from
the amyloid data.

ALKALOIDS

The alkaloids include a particularly heterogeneous
group of nitrogenous compounds, upwards of 1,000
in number, mostly from vascular plants (Willaman
and Schubert, 1961). A few non-vascular plants and
some animals synthesize alkaloids, but the com-
pounds are rare in both of these groups. Unlike
many classes of naturally occurring substances which
may be defined rather precisely in chemical terms,
no entirely adequate chemical definition of an alkaloid
seems possible because of the variety of alkaloid types
in existence. By a general operational definition an
alkaloid is considered to be a pharmacologically
active compound usually containing a basic group
and with a heterocychc nitrogen-containing ring.i It
is evident, from such a definition, that alkaloids are

155

156 BIOCHEMICAL SYSTEMATICS

not a chemically natural group, and it likewise follows that they do
not constitute a natural biological group, functionally, phylogeneti-
cally, or with respect to their biosynthesis. Therefore, few generaliza-
tions relevant to any of these above considerations are warranted.
Among nitrogenous substances of plants there is almost a continuum
from the universal products of metabolism to alkaloids in the strict
sense, and of course nitrogen-containing secondary compounds exist
which are not classified as alkaloids. Purine and pyrimidine bases and
the amino acid, histidine, are alkaloids except by the physiological
criterion. Betacyanins (formerly regarded as nitrogenous anthocyanins,
discussed in Chapter 14), except for the absence of any obvious
physiological effects, are clearly model alkaloids.

Generalizations concerning the stability of alkaloids in the
plant, factors affecting their synthesis, origin within the plant, and
histological distribution must also be treated conservatively because
alkaloids comprise such a heterogeneous group.

Since, by definition, alkaloids are physiologically active upon
animals, and many alkaloids are important drugs, the compounds are
best known to the pharmacologist. Much of the voluminous literature
on alkaloids is the direct or indirect result of their great economic im-
portance. It is probable that alkaloids are less well known to most
botanists than are certain compounds or classes of compounds that
serve some structural or obvious functional role in the plant (for ex-
ample, lignin, and plastid pigments). Consequently, a brief general
discussion, including a limited treatment of the chemical affinities of
the major classes of alkaloids will precede the section directly treat-
ing their sytematic significance. In this latter section no attempt is
made to give a comprehensive account of alkaloid distribution or to
develop any unified system of phylogenetic interpretation. Each of a
number of more natural classes of alkaloids could be given such a
treatment, and in fact some investigators have already done so. Cer-
tain of these latter types of studies will be described, but they have
been selected mainly to provide further insight into general principles
applicable to the evaluation of the systematic worth of alkaloids.

In general, discussion of biosynthetic mechanisms past the
point required to clarify some point of phylogenetic interpretation is
beyond the scope of this book, especially in the case of the alkaloids,
wherein many classes of compounds exist, each of which may be
formed by almost completely independent biosynthetic routes. The

1 According to Elderfield (1960): "No completely satisfactory all-inclusive defini-
tion of these compounds is possible. It will be sufficient to define an alkaloid as a nitro-
genous substance usually of plant origin, usually possessing basic properties, usually
optically active, and usually possessing some characteristic physiological action. Such a
definition is not perfect, and exceptions to all of the above criteria can be cited."

ALKALOIDS 1 57

subject of alkaloid biosynthesis is rarely given comprehensive formal
review, and then one is impressed with the incompleteness of knowl-
edge and the prevalence of hypotheses supported by circumstantial
evidence alone (Mothes and Romeike, 1958; Marion, 1958; and
Poisson, 1958). It should be noted that numerous alkaloids show
structurally a potential relationship to one or more amino acids. Con-
sequently, it is generally regarded that alkaloid synthesis is related
to amino acid synthesis. This generalization has proved helpful in
seeking relationships between alkaloids otherwise difficult to interpret
(discussed by Schhttler, 1956). Hegnauer (1958) has placed the major
alkaloid types into amino acid "families" for purposes of disclosing
useful systematic correlations. The arbitrary basis of such schemes
should be remembered, however. The directness of the relationship of
the biosynthesis of a particular alkaloid to a corresponding amino
acid may vary greatly in different cases. Wenkert (1959) has recently
suggested that we may be overemphasizing the relationship between
alkaloid and amino acid biosynthesis and thereby losing sight of a
potential relationship between alkaloid and carbohydrate metabolism,
particularly reaction sequences leading toward or derived from
aromatic synthesis.

Some major classes of alkaloids

protoalkaloi ds

These comprise a group of simple alkaloids lacking a hetero-
cyclic nitrogen-containing ring. Their structure suggests a relationship
to the aromatic amino acids (for example, tyrosine), and evidence
exists that hordenine is formed by decarboxylation of tyrosine
followed by N-methylation (Marion, 1958). Mescaline is also formed
from tyrosine by decarboxylation followed by hydroxylation of the
ring and methylation (Leete, 1959). These alkaloids are found in such
widely separated plant families as the Gnetaceae, Gramineae, Cacta-
ceae and Leguminosae, and their systematic value is therefore limited
to considerations of intrafamihal phylogeny.

Representative types of protoalkaloids are the following:

CH3

/ \cHCHNHCH3
~ OH

ephedrin

ho/;choCH2n;;^
^=^ CH3

hordenine

158

BIOCHEMICAL SYSTEMATICS

H,CO

H2CH2NH2

mescaline

ISOQUINOLINE ALKALOIDS

This large group of alkaloids may be considered to be deriva-
tives of a parent substance, isoquinoline. Like the protoalkaloids they
may also be regarded as, potentially, derivatives of pathways con-
nected with aromatic amino acid synthesis.

isoquinoline

The isoquinoline alkaloids range from simple derivatives with
a reduced heterocyclic ring and minor substitutions of the benzene
ring to very complex alkaloids of the bis-benzylisoquinohne type.
The distribution of this more or less natural class of alkaloids is of con-
siderable interest and will be discussed in a following section. The
isoquinohne alkaloids are quite characteristic of the Papaveraceae
and certain other families. Representatives of the classes of iso-
quinoline derivatives are illustrated below:

(a) Simple isoquinoline derivatives.

NH
HO CH3

anhalonidin

(b) Benzylisoquinoline derivatives.
H3CO.

CH2

OCH,

H3CO

OCH3

papoverin

ALKALOIDS

159

(c) Protoberberine derivatives. The basic berberine ring con-
figuration may be viewed as a benzylisoquinoline derivative in which
N-methylation is followed by condensation with the free phenyl
group.

H.C

OCH.

OCH,

bei'berine

(d) Protopine derivatives. The protopine basic configuration
may be considered to arise by opening of the original heterocyclic
ring of isoquinoline in a protoberberine configuration.

H2C

protopine

(e) Aporphine derivatives. These may also be regarded as
derivatives of the benzylisoquinoline type in which ring closure be-
tween the free-phenyl and the isoquinoline-phenyl ring occurs.

NCH3

H,CO

(f) Phthalideisoquinoline derivatives. These constitute a
relatively small group which may be regarded as derivatives of
benzyhsoquinoline formed by a secondary ring closure to produce a
five-membered, oxygen-containing ring. This group of alkaloids is
found in the Papaveraceae, with the exception of hydrastine (below)
which is found only in Berberidaceae and Ranunculaceae.

160

BIOCHEMICAL SYSTEMATICS

H2C

OCH3
OCH3
hydrastine

(g) Bisbenzylisoquinoline derivatives. These alkaloids consist
of two benzylisoquinoline groups joined by one or more ether linkages.
Thev' are found in several families including the Magnoliaceae and
Berheridaceae,

H3CN.

magnoline

INDOLE ALKALOIDS

This class is probably far more heterogeneous than the two
previous classes. Although other mechanisms of formation of the
indole nucleus doubtlessly exist, in some cases at least indole is
derived from tryptophane metabolism which, in turn, is derived as an
early offshoot from the pathway to the aromatic amino acids.
Shikimic acid is a parent substance for both groups, and at least in
the early stages of their formation all alkaloids discussed so far are
likely to have metabolic connections with each other. Some simple
indole alkaloids are known (for example, gramine, illustrated below),
and it is not surprising to find that they too occur in a number
of widely separated families.

.CHa

-CH2N

CH,

gramme

ALKALOIDS

161

In contrast, many of the indole alkaloids such as strychnine
(see below) are quite complex. Sometimes it is impossible to deter-
mine whether certain of these alkaloids are reduced indole or reduced
isoquinoline derivatives (for example, some Erythrina alkaloids such
as erysopin). Therefore, the systematic distribution may provide clues
to the interpretation of the alkaloid's biosynthetic affinities. Rep-
resentative indole alkaloids are the following:

H3CO

CH,

CH3 CH3

harmine

physostigmine

H3COOC

yohimbine

rauwolfme

H3COOC

alstonine

PYRIDINE ALKALOIDS

This group includes alkaloids in which the pyridine nucleus
itself is preserved as well as those with a reduced pyridine, or piperi-
dine, nucleus. The alkaloids may be quite simple (pyridine itself is
found in Haplopappus hartwegii) or moderately complex. Simple
pyridine derivatives such as coniin are found in numerous families
among gymnosperms, monocots, and dicots. Since pyridine is a part
of the fundamental coenzyme complex involved in oxidative phos-
phorylation, repeated evolution of simple pyridine derivatives is not
surprising. Although the pyridine alkaloids are widely distributed

162

BIOCHEMICAL SYSTEMATICS

some of them are of considerable systematic importance. Representa-
tive types are illustrated below.

CH2CH2CH3

OCH3

N-^^^O

ncinin

nicotine

anabasin

/
H2C

CH=CH-CH=CH-CO-N

piperine

r~\

C0H2C

1^^

N" "CH2CHOH-

CH3 ^^^

lobelin

CHCH.

CH=CH2

Several minor categories are illustrated below without further com-
ment. The term minor is used here strictly to indicate that these
groups do not exhibit, in general, the diversity of subtypes encountered
among the previous groups. Some of the most important and best
known alkaloids will be recognized among this group however.

Quinolizine derivatives.

CH2OH

'N^

lupinine

sparteine

ALKALOIDS

163

Pyrrolidine derivatives.

Tropane derivatives.

N-
CHs

CH2COCH3

^CHoOH
-0— COCH

C6H5

hygrin

hyoscyamin

Pyrrolizidine derivatives.

^^% /^\ 2 /C-CH3
H3C C ^CH C=0

C^ CHa O

o o .063

Imidazole derivatives.

CHc

N-

-CHo-

iCaH

2^5

O^^O

pilocarpine

senecionme

Sterol derivatives.

CH3 CH3
H3C

H3C

solanidine

Terpene derivatives

P^/OCOCsH

6^15
OCH3

OCOCH3

CH2 OCH3

OCH3

aconitine'^

Purine derivatives.
O

H3CN

O^^N-^N^
CH3

caffeine

NCH3

2 After Wiesner et al, 1959 and Bachelor et al, 1960.

164 BIOCHEMICAL SYSTEMATICS

Mothes and Romeike (1958) have summarized the major dis-
tribution of alkaloid types among orders and families of higher plants.
Table 9-1 is adapted from their data. In a few places suggestive cor-
relations in alkaloid content exist between certain families (or even
groups of families as in the isoquinohne types). While certain of these
correlations will be discussed in this chapter, in general the meaning
of such data is not yet sufficiently clear, and the known biosynthetic
relationships are too inadequate to allow meaningful speculation. The
present writers believe that the diversity of alkaloid types, their com-
plexity, and their wide distribution allow much optimism regarding
their systematic importance. The taxonomic value of alkaloids is not
necessarily restricted to simple correlations of distribution, but later
on the basis of studies of comparative biosynthesis, enzymology, and
genetic mechanisms, these compounds may yield even more substantial
insight into phylogenetic problems.

Some general considerations of
alkaloid distribution and physiology

It was noted above that alkaloids are rare in animals and in
lower plants. Among the former organisms alkaloids are found in such
widely separate groups as sea snails (Echinodermata), sand worms
(Annelida), toads^, and sharks (Chordata). Alkaloids are apparently
absent from algae, mosses, and liverworts, most fungi, and are rare
among the simpler vascular plants. For example, only protoalkaloids
or other relatively simple alkaloids occur in the divisions Sphenopsida
and Lycopsida; alkaloids are unreported from ferns, unreported from
cycads, and rare in gymnosperms in general. Thus, other than among
flowering plants alkaloids are not widely distributed. Willaman and
Schubert (1952) reported that about ninety-seven of the approximately
300 angiosperm families were known to have alkaloid-containing
genera. Cromwell (1955) stated that forty families of flowering plants
contained alkaloids. The reason for this discrepancy is not clear.
Surveys of the flora of various regions provide some information
about the actual percentage of alkaloid-containing species. A survey
of Russian species (Orechov, 1955) yielded 10 per cent, and Australian
species (Webb, 1949), yielded 20 per cent alkaloid-containing species.

3 Toads (Bufo) are notable in that secretions from their parotoid glands contain
not only alkaloids [bufotenines, also found in the plant, Piptadenia falcata (Giesbrecht,
I960)], but represent the only known vertebrate source of the plant sterol, phytosterol,
and also contain bufagins which are similar to the cardiac aglycones found in certain
plants. To our knowledge no proposal to include Bufo in the plant order Ranales has yet
appeared, even in a chemical journal.

ALKALOIDS

165

Table 9-1. Distribution by family of the major alkaloids of higher plants
(adapted from Mothes and Romeike, 1958).

Family

Alkaloid types present (among those illustrated)

Lycopodiaceae

pyridine

Equisetaceae

protoalkaloid and pyridine

Taxaceae

protoalkaloid

Gnetaceae

protoalkaloid

Magnoliaceae

isoquinobne

Lauraceae

isoquinoline

Anonaceae

isoquinoline

Menispermaceae

isoquinoline

Ranunculaceae

isoquinoline, quinolizine, terpene

Berberidaceae

isoquinoline, quinolizine

Nymphaeaceae

terpene

Papaveraceae

isoquinoUne, quinolizine, pyrrolidine

Crassulaceae

pyridine

Leguminosae

protoalkaloid, pyridine, pyrrolizidine,

quinolizine, isoquinoline, indole

Theaceae

purine

Sterculiaceae

purine

Rutaceae

isoquinoline, imidazole

Celastraceae

protoalkaloid

Aquifoliaceae

purine

Malvaceae

UmbelUferae

Piperaceae

Moraceae

Euphorbiaceae

Santalaceae

Chenopodiaceae

Cactaceae

Phytolaccaceae

Nyctaginaceae

Loganiaceae

Apocynaceae

Asclepiadaceae

Gentianaceae

Convolvulaceae

Boraginaceae

Labiatae

Solanaceae

Rubiaceae

Lobeliaceae

Compositae

Liliaceae

Amaryllidaceae

Dioscoriaceae

Gramineae

Palmae

protoalkaloid

pyridine, pyrrolidine

pyridine

pyrrolizidine

protoalkaloid, pyridine, quinolizine, indole

protoalkaloid, isoquinoline, pyridine

purine

isoquinoline, indole

indole, steroid

pyridine

pyrrolidine, tropane

pyrrolizidine

pyridine

protoalkaloid, pyrrolidine, pyridine,

pyrrolizidine, tropane, steroid

isoquinoline, indole, purine

pyridine

pyridine, pyrroUzidine

protoalkaloid, isoquinoline, steroid, purine

isoquinoline

tropane

protoalkaloid, indole

pyridine

166

BIOCHEMICAL SYSTEMATICS

Webb found that from 753 species in 110 families tested, 145 species
from forty-one families contained alkaloids.

The distribution of alkaloids within the plant and in the ceU
has been discussed by James (1950). Usually, the alkaloids, which are
water soluble, accumulate in the vacuoles and are rarely found in
dead tissues (even the quinine of Cinchona bark is said to be confined
to living cells). Alkaloids may be present in any part of a particular
plant and very often occur in meristematic tissue. In Baptisia leuco-
phaea, alkaloids have been found to be present in roots, stems, leaves,
flowers, fruits, and seeds, the highest concentration occurring in the
seeds. The absolute amounts and relative concentrations of the
various alkaloids in B. leucophaea differ from one part of the plant
to another (Brehm, 1962).

Synthesis of alkaloids exhibits a number of interesting varia-
tions. For example, nicotine synthesis in Nicotiana is initiated in the
root and completed in the leaves. The lupine alkaloids, however, are
produced in rapidly growing shoots. Doubtlessly, many variations are
to be expected in the pattern of synthesis of such a heterogeneous
group of substances.

Concerning the stability of alkaloids, James ( 1953) has said that

... a given species always forms the same group of related alkaloids,
in more or less fixed proportions and within fairly narrow limits of con-
centration. It has proved very difficult to modify these relations, even
quantitatively, by simple experimental means.

Mothes (1955) and other workers, however, suggest that
alkaloids are definitely affected by various external factors. In our
analyses of individual plants of Baptisia leucophaea from several
populations we have found a wide range in the absolute and relative
concentrations of the leaf alkaloids (Fig. 9-1). We do not know yet
whether or not these differences are genetic. However, even if the dif-
ferences were assumed to be genetic in origin, such extensive variation
would suggest a multiple gene system expressing the effects indirectly.
In such circumstance (for example, wherein alkaloid synthesis is
influenced by diverse internal factors) it seems likely that certain exter-
nal factors would also exert some influence. A relatively small propor-
tion of alkaloids has been studied with respect to questions of varia-
tion under experimental conditions. It is likely that the role of some
alkaloids in the plant is not critical, the factor of natural selection is
correspondingly presumed to be low, and therefore regulatory mecha-
nisms controlhng their synthesis would not be expected to be highly
refined. Furthermore, alkaloid synthesis is somewhat closely connected

ALKALOIDS 1 67

/ 1

Fig. 9-1. Circular chromatograms showing alkaloid variation in
individual plants of Baptisia leucophaea from a single population.
Leaf samples were from plants of similar stage of development col-
lected at the same time. (Courtesy of B. R. Brehm)

with amino acid metabolism, and environmental factors may have a
powerful effect upon free amino acid concentrations. It is therefore
understandable that alkaloids may in certain cases be quite sensitive
to environmental factors.

If it were established that alkaloids served some important
role in the plants in which they occur, additional systematic signif-
icance might underlie their presence; for example, the alkaloids
would, in turn, be related to other special physiological attributes of
the plant. However, in general, the role of alkaloids in the plant is un-
known. They have been regarded as sources of protection against in-
sects, organic waste products (detoxification products), regulatory de-
vices, or even energy sources, but Httle or no direct support of any of
these hypothetical functions is available. It is beyond the scope of
this book to explore the possible roles of alkaloids in detail. The
possibility that alkaloids serve as detoxification mechanisms in which
the products are collected in vacuoles is interesting however. Alkaloids

168 BIOCHEMICAL SYSTEAAATICS

are absent from algae and aquatic plants in general. In aquatic higher
plants water-soluble toxic products may be eliminated directly into
the environment without the requirement of detoxification mecha-
nisms. The aquatics which produce alkaloids are usually those with
floating leaves. For example, the submerged Ceratophyllaceae lack
alkaloids while the closely related Nymphaeaceae with floating leaves
are alkaloid containing. [It is interesting to note that in the genus
Cabomba (Nymphaeaceae), in which the leaves are mostly sub-
merged, alkaloids have not to our knowledge been reported.]

Thus far there has been little or no evidence of physiological
effects of alkaloids upon the plants in which they occur. Dawson
(1948) has expressed skepticism that negative results effectively settle
the question. Alkaloids of one species may apparently affect other
species— even close relatives. Mothes (1960) noted that when bella-
donna or tomato was grafted onto Nicotiana stock, nicotine migrated
into the scion and browning occurred. The browning was assumed to
result from the presence of nicotine.

General considerations of the
systematic value of alkaloids

Previously, it was noted that the protoalkaloids and nicotine
had a distribution which suggested that parallel evolution accounted
for their presence in certain widely separated plant groups. Rowson
(1958) has noted that the distributions of anabasine, berberine, and
caffeine did not closely correlate with the systematic position of the
plants in which they occur. The same is true of 3-methoxypyridine
(found in Equisetum and Thermopsis), the barman alkaloid types,
and others. Since parallel evolution of morphological attributes is
also regularly encountered, similar parallelisms among biochemical
components should not be cause for excessive pessimism concerning
their use. Parallel evolution is likely to be responsible for many
possible misinterpretations of biochemical data.

Within an alkaloid series it is probable that alkaloid com-
plexity is correlated generally with systematic advancement. However,
McNair (1935) somewhat naively correlated the molecular weights of
alkaloids with the Engler and Prantl family index number. The
"percentage of frequency rule" (Chapter 4) supports the previous
generalization but only within a closely knit group wherein parallel
evolution for the character is minimized. It is of course important to
estabhsh better criteria of complexity than merely molecular weight.
For example, the genetical basis for the synthesis of a bis-benzyliso-

ALKALOIDS 1 69

quinoline may be no more complex than that of the smaller berberine
type of isoquinoline. The former is a dimer of a simple benzyliso-
quinoline; the latter probably involves an N-methyl phenyl condensa-
tion of the simple benzylisoquinoline.

There seem to be an unusually large number of highly specu-
lative statements relevant to the systematic implications of the
alkaloids. Arguments based on criteria of simplicity versus complexity
of the alkaloids may be quite subtle in nature. Often it is difficult to
evaluate an argument fully because the logic, as applied in a chemical
reference-framework, may be sound, but not in accord with the bio-
logical facts. The following discussion by Wenkert (1959) provides an
example:

On the basis of the rapidly emerging patterns of the biosynthesis of
plant products, both theoretical and experimental, it is possible to
categorize, albeit yet crudely, natural substances into two classes, one
based to a large extent on acetate and, hence, on genetically and
enzymatically easy routes, and the other founded to a major degree on
non-acetate material, i.e. substances farther along in the tricarboxylic
acid cycles and hence, enzymatically difficult, circuitous routes. If it be
assumed that the evolution of life processes, i.e. the structure and
mechanism of enzymes, through geologic time proceeded from simple
to more complex patterns, a correlation of paleobotany with the
chemistry of natural products would be on hand. Substances originat-
ing from acetate would be expected present in the oldest plants. On
this basis the structure of Lycopodium alkaloid annotinine is no sur-
prise, nor is the discovery of triterpenes from petroleum and coal
deposits.

Despite specific reference to the tricarboxylic acid intermediates to
illustrate the "non-acetate" pathways, we infer from the main body
of the paper that Wenkert is considering the acetate-mevalonate
family (Chapter 13) of compounds on the one hand and the shikimate-
prephenate family (Chapter 11) on the other. The former lead to such
compounds as the carotenoids, terpenes, essential oils, and sterols; the
latter lead to indoles, aromatic amino acids, hgnins, and tannins. In
certain water-soluble plant pigments (flavonoids) there is a partial
contribution from each pathway (Chapter 11). Alkaloids of both
affinities are known as well as some unrelated to either (for example,
the purine derivatives). Perhaps purines, by the criterion above,
should be expected in the more ancient plants since it is generally be-
lieved that the earhest hving organisms formed polynucleotides con-
taining purines. However, alkaloids of the purine type are found
in coffee.

1 70 BIOCHEMICAL SYSTEMATICS

While the theoretical position of Wenkert is no doubt sound,
the fact is that the most "primitive" known alkaloid-containing plants
are vascular plants. These plants also produce carotenoids and lignin.
The mevalonate and prephenate pathways were undoubtedly well
represented among lower forms possibly hundreds of millions of years
before alkaloids appeared. The genetic complexity of a given alkaloid
is perhaps best represented by the extent of deviation of the alkaloid
from an already established basic metabolic pathway. The reasons for
assuming that non-acetate pathways are more likely to involve
enzymatically difficult and circuitous routes may be valid, but they
are not obvious to the present writers.

Specific examples of alkaloids of
systematic significance

Alkaloids of the isoquinoline class are probably among the
best examples to illustrate the application of biochemical criteria to
phylogeny of the higher categories. In this case the disposition of
families and perhaps even the proper delimitation of orders are in-
volved. Nevertheless, so far the alkaloid chemistry has failed to clarify
the taxonomic problems among the groups of plants concerned.

The isoquinolines, as noted above, are likely offshoots of
aromatic amino acid metabolism. Hegnauer (1952, 1954, 1958) has
discussed the taxonomic distribution of the entire group of related
alkaloids, and our treatment is derived principally from his compre-
hensive account. A partial list of families containing isoquinolines
follows. For later reference purposes the list is divided into the
categories of Lignosae and Herbaceae (Hutchinson 1959).

Lignosae

Herbaceae

Magnoliaceae

Ranunculaceae

Anonaceae

Berberidaceae

Monimiaceae

Menispermaceae

Hernandiaceae

Papaveraceae

Rutaceae

Aristolochiaceae

The families hsted above are recognizable as representing in
general rather "primitive" families by Hutchinson's criteria, and
except for the Aristolochiaceae, Rutaceae, and Papaveraceae, they
represent the Ranales of Engler and Diels. Other Ranalian famihes

ALKALOIDS

171

(for example, Nymphaeaceae) do not contain isoquinoline alkaloids
but contain other types of alkaloids. Certain families (for example,
Ranunculaceae) may contain isoquinoline derivatives and in addition
other types of alkaloids. Most of the families hsted contain several
types of isoquinolines. The Magnoliaceae contain, in addition, proto-
alkaloids, the simplest group of alkaloids, which are also derived from
aromatic amino acid metabolism.

General conclusions from the over-all distribution of iso-
quinoline alkaloids are that their wide occurrence among the Ranales
indicate phylogenetic interrelationship. Outside this group (for ex-
ample, in the Amaryllidaceae) they occur infrequently, and then the
specific mode of secondary ring closure differs, indicative of parallel
evolution. These Amaryllidaceae alkaloids, the lycorine types, have a
different type of linkage between the benzyl and heterocyclic
N-containing ring of the basic benzylisoquinoline. In the Leguminosae
the Erythrina alkaloids, regarded as possible isoquinoline derivatives,
may also be interpreted as indole derivatives. An exception is the
family Rutaceae wherein almost a full array of isoquinoline alkaloids
of the same types as occur in the Ranales are to be found.

Gibbs (1954) has noted the striking parallelisms to be found
in the alkaloids of the sub-families Papaveroideae and Fumarioideae
of the Papaveraceae. (However, see Chapter 6 for the distribution of
6-acetylornithine in these sub-families.) These sub-families are con-
sidered to be separate families by some systematists. It is more than
a question of the common presence of isoquinolines which relates these
sub-families. The extent of alkaloid parallelism is striking. Proto-
berberines, aporphines, phthalideisoquinolines, and protopines occur
widely throughout both sub-families. Phthalideisoquinolines are rare
outside the Papaveraceae while protopine is found elsewhere only in
Nandina (Berberidaceae). Gibbs states:

We must not let the finding of protopine in a plant outside the
Papaveraceae blind us to the very strong evidence from the work of
Manske and others that the Papaveroideae, Hypocoideae and Fumar-
ioideae are indeed very closely allied chemically. This work is one of
the best examples of the worth of comparative chemistry applied to
taxonomy.

Actually, demonstration of alkaloid similarities in the sub-
families of Papaveraceae does not necessarily bear upon the question
of whether the two taxonomic groups should be considered as sub-
families or as families. The significant point is that they have been
closely linked. The taxonomic position, as long as such a link is
emphasized, is a matter of descretion.

172

BIOCHEMICAL SYSTEMATICS

Mothes and Romeike (1955) have questioned the use of iso-
quinoHne alkaloids to relate the Papaveroideae and Fumarioideae to
each other, yet by the same data to relate the Papaveroideae to other
Ranahan famihes. They consider that such conclusions represent
circular reasoning. However, there are two levels of similarity in-
volved, and the evidence should be applied independently at different
levels. For example, various types of isoquinohne alkaloids occur in
the different famihes of the Ranalian complex, and it is merely the
presence of the general isoquinohne type that ties the groups together,
while, in addition, in the Papaveraceae it is the common presence
of a series of specific isoquinohne derivatives, some rare, which is con-
sidered to be especially significant in adducing the relationship of the
sub-families.

Comparison of certain alkaloids of the Ranunculaceae and
Berberidaceae proves to be interesting. In the Ranunculaceae,
Xanthorhiza, Coptis, Thalictrum, and Hydrastis produce isoquinohne
alkaloids. Except in Hydrastis the alkaloids are relatively simple
protoberberines. Species of Hydrastis contain hydrastine, a more com-
plex phthahdeisoquinoline, found only in Berberis laurina of the
Berberidaceae. In the Berberidaceae, Berberis, Mahonia, and Nandina
are alkaloid producers. Protopine, otherwise restricted to the Papaver-
aceae, is found in Nandina, and Nandina lacks the bis-benzyliso-
quinolines found in other Berberidaceae. It is interesting that Hutch-
inson (1959) and other workers have placed Nandina in a monotypic
family, Nandinaceae. Generally, the alkaloid distribution in Berberi-
daceae and Ranunculaceae does not suggest any unusually close re-
lationship between the two families. However, an interesting proposal
was published in this connection by McFadden (1950). McFadden
recognized a "small chromosome group" of five genera in the Berberi-
daceae: Nandina, Berberis, Jeffersonia, Hydrastis, and Glaucidium.
Basic chromosome numbers in this group vary from x = 6 to x = 14.
In the Ranunculaceae six genera also form a "small chromosome"
group (Gregory, 1941): Isopyrum, Aquilegia, Anemonella, Thalictrum,
Coptis, and Xanthorhiza. Basic chromosome numbers for these genera
range from x = 1 to x = IS. The isoquinohne alkaloids are found, in
the two famihes, only in the small chromosome groups. According to
McFadden:

From a morphological standpoint, treatment of this group of genera as
a systematic unit is at least as tenable as their present classification.
However, in grouping these genera as a taxonomic unit morphological
characters would be stressed that are different from those now em-
ployed by classification of these.

ALKALOIDS 1 73

Certainly, the cytological and biochemical data considered separately
would not constitute strong evidence. "^ The conclusions of McFadden
are not in accord with the serological data provided by Hammond
(Chapter 5) who placed Hydrastis in the Ranunculaceae but closer to
Ranunculus than to Thalictrum. Finally, the six small-chromosome
genera assigned to the Ranunculaceae do not form bis-benzylisoquin-
olines which are typical of the Berberidaceae. Other biochemical evi-
dence cited by McFadden was relatively meager and inconclusive, but
it is quite possible that intensive biochemical studies would clarify
this interesting situation.

The family Rutaceae (containing the orange) is the last to
be discussed in connection with the isoquinoline alkaloids. Hegnauer
(1958) noted that protoberherine, aporphine, protopine, and rare
chelidonine alkaloids were all present in Rutaceae as well as in
certain Ranalean families. He believes that the affinities of Rutaceae
and these families are much closer than most systematic treatments
imply. In Hegnauer's words:

The exactness, not similarity, of the complex phenylisoquinoline
alkaloids in both groups appear so surprising and convincing that a
new investigation of the systematic position of the Rutaceae may be
urgent.

In this connection it is interesting to note that Hallier presumably
derived the Rutaceae from "stocks ancestral to the Berberidaceae"
(Lawrence, 1951).

Another alkaloid of the Rutaceae, rutaecarpine, is a complex,
indole-containing substance. Chemically related alkaloids are present
in the families Apocynaceae, Loganiaceae, and Rubiaceae. No sys-
tematic bridge between these families and the Rutaceae is necessarily
implied, but it is interesting that Hutchinson (1959) has proposed a
relationship between Rubiaceae and the families Apocynaceae and
Loganiaceae.

Other indole-alkaloids of restricted systematic distribution
are those of the Amaryllidaceae. In this family over seventy alkaloids
are known. ^ While there are a number of rare alkaloids in this sub-

^Kumazawa (1938) on morphological grounds provisionally included the genera
Hydrastis and Glaucidium in the Ranunculaceae as the sole members of the sub-family
Glaucidioideae; however, he retained Jeffersonia in the Berberidaceae as have nearly all
subsequent workers.

5 To indicate the increased interest in phytochemical research, it seems worth
noting that only fifteen alkaloids were known in the Amaryllidaceae in 1954, the addi-
tional compounds having been acquired over a six-year period (Wildman, 1960).

174 BIOCHEMICAL SYSTEMATICS

family, some such as lycorine, have been found in all twenty-six of the
genera of Amaryllidoideae which have been examined to date (Wild-
man, 1960). Pax and Hoffmann (1930), in their treatment of the
Amaryllidaceae, recognized four sub-families: Agavoideae, Hypoxidoid-
eae, Campynematoideae, and Amarylhdoideae. Most workers have
treated the family similarly, but Hutchinson (1959) excluded all the
sub-families, other than the Amaryllidoideae, and simultaneously trans-
ferred three tribes of the classically constituted Liliaceae to the
Amaryllidaceae [including the tribe Alheae, which contains the genus
Allium (onion)]. It is interesting to note that Hutchinson's treatment,
except for the transfer of the three Liliaceous tribes (Agapantheae,
Allieae, and Gilliesieae), would be compatible with the alkaloid data.
However, alkaloids of the Amaryllidaceous type, while found in
nearly all of the tribes of the Amaryllidoideae as classically con-
stituted, are not found in the three transferred tribes, and therefore
the family, as reconstituted by Hutchinson, is perhaps as anomalous
from the standpoint of alkaloids as by the treatment of Pax and
Hoffmann.

Unrelated alkaloids, of the colchicine type, have been found
in five genera of three tribes of the Liliaceae. Two of the tribes
(Colchiceae and Iphigenieae) appear to be fairly closely related, but
the third (Uvularieae) is somewhat more distant. However, the
Lihaceae is a large and varied family, and as indicated by Hutchinson
(1959) it is still somewhat artificially classified, even with the removal
of several of its more distinct elements. A more inclusive biochemical-
morphological study might yield a better phylogenetic arrangement
than exists at present. Correlations between the comparative chem-
istry of the alkaloids and that of other chemical groups (for example,
the substituted glutamic acids. Chapter 6 and the saponins. Chapter
13) in the families Lihaceae and Amarylhdaceae should be informative.

Hegnauer (1958) considers that the occurrence of the Senecio
alkaloids outside the family Compositae, in one instance in the
Boraginaceae and in another the Leguminosae, is of phylogenetic
significance. He notes the presence of inulin in both the Boraginaceae
and Compositae and the flavonoid chalkone in both the Compositae
and Leguminosae and concludes that,

. . . the extensive structural resemblances of the Crotalaria, Borago
and Senecio alkaloids is altogether not understandable if no genetic
connection can be recognized between the families, and the alkaloids
are metabolic wastes.

Perhaps this is true, but parallel evolution, at least in the Legumino-
sae, seems to be the more likely explanation.

ALKALOIDS 1 75

An interesting group of alkaloids found in the legume genus
Lupinus and related groups may prove, eventually, to be of consider-
able systematic significance within the family Leguminosae. These
quinolizine derivatives, represented by a relatively small number of
specific types, are also known to occur in the families Berberidaceae,
Chenopodiaceae, Papaveraceae, and Solanaceae, but it is the Legumi-
nosae in which the alkaloids have been most intensively studied. In the
latter family, only the tribes Genisteae, Podalyrieae, and Sophoreae
of the sub-family Papilionoideae include genera which produce these
alkaloids; for example, Lupinus, Thermopsis, Baptisia, Cytisus,
Sarothamnus, Genista, Sophora, and Podalyria. Certain of the lupine
alkaloids, such as cytisine and spartein are of quite widespread
occurrence within these tribes. Biogenetic evidence plus correlated
genetic studies should provide information that will yield further in-
sight into phylogenetic problems. Intensive investigation of the lupine
alkaloids using several different approaches is currently underway by
Nowacki and colleagues (Kazimierski and Nowacki, 1961). Schutte
and Nowacki (1959) have presented evidence that sparteine is synthe-
sized from the amino acid, lysine, and Nowacki (1958) and Birecka
et al. (1959) have circumstantial evidence that sparteine is converted
into lupanine and then into hydroxylupanine. It is Hkely that some
parallelism and convergence have occurred in the origin of bio-
synthetic mechanisms involving lupine alkaloid synthesis. To what
extent these complexities can be explained in a phylogenetic sense re-
mains to be seen. Intensive investigations of large, natural genera
such as Lupinus should prove important in clarifying inter-generic
relationships by exposing the amount and nature of variation at the
infra-generic level.

Hegnauer (1958) has utihzed the presence of lupinine alkaloids
in the Leguminosae and Chenopodiaceae as evidence of a relationship
between the orders Rosales and Ranales. In support of this Hegnauer
cited Hutchinson's placement of the presumed parental stock of the
Chenopodiaceae and Caryophyllaceae, adjacent to the Ranales. How-
ever, in Hutchinson's scheme the Leguminosae are in the Lignosae,
alhed with the woody Magnohales, while the species producing lupine
alkaloids are to be found in the herbaceous Ranunculaceae and
Berberidaceae, both in Hutchinson's Herbaceae. In this instance,
then, it does not seem valid to imply that the argument derives
further support from Hutchinson's system.

Recently, the taxonomic significance of the steroid alkaloids
of the veratrum group has been evaluated (Kupchan et al, 1961).
This study is representative of other similar studies which involve a
group of alkaloids whose distributions within either a sub-family or

176

BIOCHEMICAL SYSTEMATICS

genus indicate definite taxonomic significance. This study will there-
fore serve as an example of the type. The veratrum alkaloids are
known, so far, from the tribe Veratreae of the Liliaceae. Numerous
individual alkaloids occur representing variations in the bsisic nucleus,
substitutions of the basic nucleus, and ester derivatives. Kupchan
et al. recognized two major groups, the jerveratrum group and the
ce veratrum group as follows:

Jerveratrum group. Veratramine, rubijervine, isorubijervine,

jervme.

Ceveratrum group. Zygadenine, veracevine, germine, proto-
verine.

The two groups possess the C27 ring structure (other vera-
trum alkaloid types are incompletely identified). The jerveratrum
types, with few hydroxyl groups, occur as free bases or as simple
glucosides. The ceveratrum types, with seven to nine hydroxyl sub-
stitutions, usually occur esterified with various acids or ester alkaloids,
never as glycosides. Among the ceveratrum types, zygadenine and
veracevine occur as monoesters; germine and protoverine occur as tri
or tetra esters.

The genera concerned are Veratrum, Zygadenus, Stenanthium,
Schoenocaulon, Amianthium, and Melanthium. Schoenocaulon is re-
garded as rather distinctive and homogeneus, Veratrum as relatively
homogeneous, Stenanthium as small and diverse, and Zygadenus as
quite heterogeneous possibly including several genera (as subgenera).

The distribution of veratrum alkaloids is given in Table 9-2.

ALKALOIDS

177

Table 9-2. Distribution of Veratrum alkaloids (adapted from Kupchan et al,
1961).

Species

Type of alkaloids

jerveratrum

cerveratrum

unclassified

Veratrum album album

V. album oxysepalum

V. album grandiflorum

V. vivide

V. eschscholtzii

7 •

V. stamineum

V. fimbriatum

V. nigrum

Amianthium muscaetoxicum

Zygadenus venenosus venenosus

Z. venenosus gramineus

Z. paniculatus

Schoenocaulon officinale

Kupchan et al. consider that alkaloid distribution supports in general
the classification on morphological grounds. Zygadenus and Schoe-
nocaulon contain only the ceveratrum alkaloids. The chemical
evidence postulated as the basis for considering Zygadenus inter-
mediate between Veratrum and Schoenocaulon is not particularly
convincing. Zygadenus has a higher proportion of zygadenine esters
than does Veratrum, and zygadenine is considered an alkaloid
"hybrid" between the two ceveratrum sub-types in that it occurs as
a monoester but possesses some structural similarities to the germine
and protoverine types together with which it frequently occurs. In a
phylogenetic sense the term "intermediate" has a connotation that
Zygadenus was derived from Veratrum and gave rise to Schoeno-
caulon. However, the chemical evidence does not exclude the equally
likely hypothesis that both Schoenocaulon and Zygadenus evolved
either from Veratrum or a Veratrum-\\\ie ancestor. Veratrum as in-
dicated appears to have a more primitive alkaloid chemistry.

Another point of taxonomic interest concerns certain Vera-
trum species. V. album var. album, V. vivide, and V. nigrum (Atlantic
coast taxa) contain alkaloids which yield mono- or dihydroxymethyl-
butyrate residues upon hydrolysis. V. album var. grandiflorum,
V. eschscholtzii, V. stamineum, and V. fimbriatum (Pacific coast taxa)
yield angelate and tiglate upon hydrolysis.

In Table 9-2 one may note that certain taxa, such as Vera-
trum album var. album or V. vivide contain numerous alkaloids while

178 BIOCHEMICAL SYSTEMATICS

others, for example V. nigrum, contain relatively few alkaloids. The
apparent difference may, however, reflect merely a more intensive
examination of one species. One should give more consideration to the
presence of a given alkaloid then to its apparent absence. In the
alkaloids, in particular, as a result of the fact that many alkaloids are
drugs, designation of a species as alkaloid-containing is based upon
arbitrarily designated minimum quantities. Hegnauer (1958) has
recommended that an alkaloid content of 0.01 per cent dry weight
represents the minimum in order for a plant to be considered alkaloid-
containing. Yet, in considering the taxonomic implications of alkaloid
distribution the more relevant data may be the presence of an
enzymatic mechanism for synthesis of even a small amount of a
particular type of alkaloid. Ability to accumulate the alkaloid in rela-
tively large amounts may also be genetic and therefore relevant, but
not necessarily as fundamental as the existence of the enzymes in-
volved in the primary pathway. For example, in the Solanaceae only
traces of nicotine occur in tomato and other species, but from a
phylogenetic, if not physiological, point of view the trace is quite
important.

In the literature of alkaloids, particularly, there are examples
of rather arbitrary taxonomic revisions by chemists, based principally
on chemical evidence. Thus Manske (1954) transferred Dicranostigme
franchetianum to the genus Stylophorum "because its alkaloids are
the same as those of S. diphyllum." Dicranostigme lactucoides was
retained "because the contained alkaloids, namely protopine, isocory-
dine, sanguinarine, and chelerythrine present a combination hitherto
encountered only in a Glaucium.'" Nowhere was there any discussion of
the basis for the previous taxonomic dispositions of the species. The
chemical evidence may be important, but it is possible that equally
significant morphological or cytological evidence was ignored.

Another example of arbitrary taxonomic "revision" prim£irily
on chemical grounds is that of Manske and Marion (1947) in Lycopo-
dium. This paper appeared in the Journal of the American Chemical
Society. Lycopodium annotinum var. acrifolium contained five alka-
loids absent from typical L. annotinum (they apparently replaced a
group of five alkaloids of the latter). Accordingly, L. annotinum var.
acrifolium was raised to specific rank, L. acrifolium, with the addi-
tional comment that the newly elevated species was more different,
morphologically, from L. annotinum, than the two species, L. flabelli-
forme and L. complanatum were from each other. This comparison
was presumed to lend additional validity to the taxonomic disposition
which otherwise was based solely on chemical data. However, Wilce,
a student of Lycopodium has stated (personal correspondence):

ALKALOIDS 1 79

So far as I know, L. annotinum has never been subjected to a critical
study using modern taxonomic methods. Before answers can be given
to the questions you ask about this species and its variety acrifolium,
such a study should be made. I feel that it is essential to study this
and other species of Lycopodium from a world-wide standpoint if one
hopes to avoid considerable error in the interpretation and evaluation
of the various characters. If after such a study were made, no distin-
guishing characters other than shape and texture of leaves had been
found, then I would certainly hesitate to recognize var. acrifolium at
the species level, regardless of the information given by Manske and
Marion. In fact, if it were not for their biochemical evidence to sup-
port the minor morphological difference, I should be reluctant to give
acrifolium even variatal status.

Since there is no indication that various populations of plants
were examined by Manske and Marion to discover the nature of
variation in alkaloid content even the chemical evidence is not
estabhshed satisfactorily by the taxonomists' criteria. From our ob-
servation of plant to plant variation in Baptisia alkaloids, unless one
has good reason to expect that variations will not occur, sampling of
populations and individual plants is of critical importance.

CYANOGENETIC
SUBSTANCES

The cyanogenetic substances of higher plants com-
prise a relatively small and somewhat heterogeneous
group of glycosides of the cell sap. The parent sub-
stances liberate cyanide apparently enzymatically
when the cells are damaged.

In recent years cyanogenetic compounds
have been rather neglected. Probably the most
significant recent advances have been in the elucida-
tion of the structures of certain cyanogenetic glyco-
sides, which have been known to exist for a number
of years. Only a few new cyanogens have been
disclosed since Robinson (1930) reported the exist-
ence of ten glycosides. The systematic importance
of the cyanogenetic compounds cannot be denied
since, although their distribution is somewhat

181

182

BIOCHEMICAL SYSTEMATICS

limited, the compounds are prevalent in certain families such as the
Rosaceae. However, Hegnauer (1958) concluded that at present "the
taxonomic significance of the character of cyanogenesis is very
limited. Its value may be more important once the cyanogenetic com-
pounds of most of the known cyanophoric species are known."
Hegnauer has alluded to a major Hmitation of many broad surveys of
the distribution of cyanophoretic species. The tests generally utihzed
merely disclose whether prussic acid (HCN) is liberated by the
species. The tests do not indicate the chemical nature of the parent
substance. The specific type of cyanogen is known in a number of
cases, but in the survey work such as that of Gibbs (1954) and others,
only presence or absence of HCN is noted by use of emulsin and
sodium picrate. Even with this limitation the distribution of cyano-
gen is often of taxonomic interest.

The most recent view of cyanogenetic compounds is that of
Dillemann (1958). Cyanogenetic substances do not include numerous
chemical structural analogs or modifications of the basic parent sub-
stance as do the alkaloids or flavonoids, and the hmited number of the
compounds reduces further their systematic significance.

According to Dillemann all the cyanogenetic substances which
have been fully characterized consist of a sugar, a cyanhydric acid,
and a third substance whose nature is variable. Since the number of
classes of these compounds is Hmited, a rather complete chemical
listing is possible. The following structural formulae are obtained from
Dillemann, using his classification:

True cyanogenetic heterosides.

In the first group the nitrile (C=N) group is attached to the
aglycone group.

(1) Amygdaloside occurs in many species of Rosaceae.

CHoOH

glucose

(2) Vicianoside is from Vicia angustifolia (Leguminosae)
OH ^ ^„ H /-^

N=C

c<3

CYANOGENETIC SUBSTANCES

183

(3)

A group of several closely related substances, also related
to the two types illustrated above, have the basic struc-
ture phenyglycolonitrile D-glucoside-/S.

(a) Prunasine (L-phenyl) from many Rosaceae, some
Myoporaceae {Eremophila maculata), Myrtaceae
{Eucalyptus corynocalix) and Scrophulariaceae
{Linaria striata and Choenorrhinum minus).

(b) Prulauroside (L-D-phenyl) from species of Rosaceae.

(c) Sambunigroside (D-phenyl) from Sambucus nigra
(Caprifoliaceae), Acacia glaucescens (Leguminosae)
and Ximenia americana (Oleaceae).

(d) Dhurroside, a para-OH phenyl analog of prunasine,
found in Sorghum vulgare (Gramineae).

(e) Phyllanthoside (may be same as dhurroside) found in
Euphorbiaceae {Phyllanthus gastroemii).

(f) Zierioside, a meta-OH phenyl analog of prunasine
found in Rutaceae (Zieria laevigata).

(4) Linamaroside, from several legumes and others, including
several species of Dimorphotheca (Compositae).

CHoOH
O

oh\

-o-

CH3
-C-CH3

teN

(5) Lotaustraloside is from species of Lotus and Trifolium
repens (Leguminosae).

CH3

CH.,OH

C — CH2 — CH3

(6) Acacipetaloside is found in the legume genus, Acacia.

CH.OH

1 84 BIOCHEMICAL SYSTEMATICS

(7) Gynocardoside (the structure has not been completely
established, in particular the positioning of three OH
groups attached to the aglycone) is from Gynocardium
and Pangium (family Flacourtiaceae).

CH2OH

C< {C5H5(OH)3

OH
Positions of (OH) groups are doubtful.

In the second series, the nitrile group is attached to the glycosidic
group.

(8) Lotusoside (lotusin) is from Lotus arabicus.

C11H21O10 — CH — O

C=N

Although the formula above had been accepted for many
years, work by Doporto et al. (1955) has established rather conclu-
sively that the "flavone" portion is incorrectly identified. These
investigators obtained some of the original samples and identified
the components as a mixture of quercetin and kaempferol (flavonols).

The authors did not discuss lotusin itself, only the flavonol
degradation products. If both these flavonols are derived from lotusin
itself, then there must be two different lotusins present. Apparently
more work is needed on the intact cyanogen.

The list of cyanogenetic compounds above agrees essentially
with that of Gibbs (1954) except for two compounds, hiptagin and
karakin, which Dillemann described as "pseudo-cyanogenetic"
heterosides, and a third, macrozamoside, derived from several
cycads including Macrozamia spiralis. Another substance, cycasin
found in Cycas revoluta, consists of the same parent substance as
macrozamin but is esterified with glucose (Nishidi et al., 1960). These
substances do not liberate HON in hydrolysis with dilute HCl unless
first treated with sodium hydroxide solution, then acidified. They are
called pseudo-cyanogenetic substances by Lythgoe and Riggs (1949).
According to these authors (see also Langley et al., 1951, who were
studying macrozamin, the substance responsible for the condition
known as "wobbles" or "staggers" in Australian livestock) it is a

CYANOGENETIC SUBSTANCES 1 85

glycoside containing the carbohydrate component, primeverose
[6-(B-D-xylosido)-D-glucose].

^-0

IfOH^

r—f ^9t-o-ch2N=n+ch3

macrozamin

Macrozamin represents one of the few known occurrences of
linked nitrogen atoms in a natural product (see Chapter 6). Systemat-
ically it is interesting in that a "new" synthetic ability is associated
with a primitive plant. Although such a situation should not be un-
expected, it may appear to be paradoxical if one should suppose that
primitiveness necessarily implies evolutionary quiescence. It is not
likely that the ability of cycad species to form linked N is of recent
origin, however, since the compound is found in at kast three cycad
genera. The important theoretical point is that macrozamin could
represent a recently acquired synthetic ability even though the cycads
themselves are phylogenetically old. The other pseudo-cyanogenetic
compounds, hiptagin and karakin, differ from each other in the posi-
tion of fusion of the nitrogen-containing group (B-nitropropionic acid:
NO2— CH2— CH2— COOH).

karakoside = 1:4:6 tri-(B-nitropropionyl)-D-glucopyranose
hiptaside = 1:5:6 tri-(B-nitropropionyl)-D-glucopyranose

Relatively httle is known concerning the origin of cyano-
genetic glycosides. Butler and Butler (1960) reported that when
white clover was supplied with C^* labeled isoleucine and valine,
radioactivity appeared in the aglycone portion of lotaustralin and
linamarin respectively. However, radioactivity failed to appear when
C14 labeled glycine and vahne, labeled only in the 1-C, were supplied.
These results suggested that isoleucine and vahne are involved in the
metabolism of these cyanogenetic glycosides, and that the formation
of the cyanide grouping includes decarboxylation.

The role of cyanogenetic substances is unknown. Ideas that
they represent protective agents, wastes, or reserve energy sources are
distinctive neither by virtue of originahty nor their susceptibihty to
direct experimentation.

Distribution of cyanogenetic substances in the plant is rather
widespread although apparently the leaves are particularly rich.
Green fruits in some cases are richer in cyanogenetic compounds than
are the mature fruits (for example, in Nandina domestica). In some
genera (for example, Vicia) only the seeds are cyanogenetic, and in

1 86 BIOCHEMICAL SYSTEMATICS

others (for example, Isopyrum) roots as well as aerial parts are
cyanogenetic.

More pertinent is the distribution of cyanogenetic compounds
within the plant kingdom. According to Dillemann, except for a few
isolated examples such as Bacillus pyocyaneus, and certain fungi,
cyanogenetic substances are restricted to advanced vascular plants:
about thirty species of ferns and nearly 900 species of angiosperms
representing ninety-five families. Families notable for the production
of cyanogenetic substances are the Rosaceae (150 species), Legumino-
sae (100), Gramineae (100), Araceae (50), Compositae (50), Euphor-
biaceae, Passifloraceae, Ranunculaceae, and Saxifragaceae. In some
families only one species is known to be cyanogenetic. Hegnauer
(1959b) lists about 750 species representing sixty-two families and 250
genera of seed plants. An indication of the frequency of cyanogenesis
among a broad sample of species may be obtained from results of an
Australian phytochemical survey (Webb 1949). Eleven cyanogenetic
species were found among 306 species representing sixty-seven
families. The positive species were scattered among several families.

At the generic level, in some genera all species studied were
cyanogenetic (for example, Passiflora, Prunus, Cotoneaster, Dimor-
photheca) while in others some species were cyanogenetic and others
were not. In some genera only a single species may be cyanogenetic.

There are several reports of the existence of physiological or
biochemical races within a species. Thus both cyanogenetic and
acyanogenetic individuals have been reported for Trifolium repens,
and Lotus corniculatus (Armstrong et al, 1912, 1913), Sorghum
vulgare (Petrie, 1913), Eucalyptus viminalis (Finnemore et al., 1938),
Euphorbia drummondii (Seddon, 1928) Trema aspera (Smith and
White, 1920), and other species. The subject of chemical races will be
considered in more detail in Chapter 16.

In Lotus corniculatus a rather complex situation is encoun-
tered. In an intensive investigation in 1911 of populations of L.
corniculatus (Armstrong et al, 1912), cyanide was rarely detected.
However, in the following year, in which the weather was unusually
warm and dry, cyanide was rarely absent in the same populations of
these perennial plants. There were populations of the species growing
near each other which were markedly different in the amount of
cyanide present. Futhermore, the variety major was always free of the
cyanogen and, hkewise, free of the enzyme which, in the typical L.
corniculatus, was present.

Trione (1960), who studied the cyanogen content of flax
seedlings in controlled environment, found that not only did the
HCN content increase with more light but even a diurnal variation in

CYANOGENETIC SUBSTANCES 1 87

HCN occurred. Ermakov (1960) likewise noted that linamarin content
of flax was higher under controlled conditions of lower soil moisture,
low temperature, after mechanical injury, and in young growing
organs, so apparently in this species cyanogen content is quite sensi-
tive to environmental factors.

It is evident from these data that both genetic and ecological
factors affect the production of the cyanogenetic compound. It would
be interesting to know whether the enzyme concentration was affected
similarly. The marked influence of ecological factors upon the occur-
rence of a biochemical component, while it may be exceptional in this
instance, needs to be taken into account in the studies of physiological
races either in population studies or classical genetic studies. In con-
trast to the situation in Lotus corniculatus, separate reports by
Williams (1939) and Atwood and Sullivan (1943) indicate that Tri-
folium repens produces similar quantities of cyanogen under differing
conditions. These authors, studying the inheritance of cyanogen
production, observed plants over a period of several years and re-
ported that individual plants always tested about the same for
cyanogen.

With rare exceptions, cyanogenetic glucosides are accom-
panied by enzymes which catalyze their hydrolysis with liberation of
HCN and sugar. A complex of enzymes is involved in the breakdown
of amygdalin, but the system of enzymes is called emulsin.
According to Robinson (1930) the emulsin system will liberate HCN
from sambunigrin, dhurrin, vicianin, prunasin, and prulaurasin but
not linamarin. Since, in all but the last-named, the linkage is quite
similar, this fact is not surprising. However, reports that linase will
liberate HCN from amygdalin are surprising and perhaps should be
treated with some conservatism. Certain plants which do not them-
selves form cyanogenetic substances contain enzymes which break
down amygdalin (Robinson, 1930).

Gibbs (1954) and Hegnauer (1958, 1959b) have reviewed the
distribution of cyanogenetic compounds from a taxonomic viewpoint.
Hegnauer investigated over 400 species and reported a number as
cyanogenetic. Included were first records of the conifers Taxus
cuspidata and T. media. It is unnecessary to reproduce their data in
detail, for no clear-cut systematic implications are evident. The most
interesting data are those in which subfamilies rich in cyanogenetic
species are compared. For example:

Rosaceae: Cyanogenesis is pronounced in the Pomoideae

and Prunoideae, less frequent in the Rosoideae
and Spiraeoideae.

188

BIOCHEMICAL SYSTEMATICS

Leguminosae: Although the sub-families Mimosoideae and
Caesalpinioideae contain a few cyanogenetic
species, the character is best expressed in the
Papilionoideae wherein most tribes contain
cyanogenetic species. Trifolium repens contains
two different types of cyanogens, apparently
the only such example.

An unusual cyanogen occurs in the legume genus Indigofera.
Morris et al. (1954) studying a toxic substance from the leaves of
Indigofera endecaphylla, found it to be ^-nitropropionic acid;
NO2 — CH2CH2COOH, the aglycone of hiptagin and karakin. Accord-
ing to these authors they isolated the compound from several species.
In a subsequent paper, Cooke (1955) studied several species of
Indigofera with the following results:

Concentration of j8-nitropropionic acid
(mg/g Fresh Weight)

Indigofera tetensis

I. suffruticosa

I. trita

I. dimorphophylla

I. subulata

7.6 (leaf)

I. endecaphylla

6.1-14.8 (leaf)
9.8 (immature leaves)
8.8 (mature leaves)
2.4 (stem)

Other workers (Schilling and Strong, 1955; Dupuy and Lees,
1956; Bell, 1962) have obtained yet another unusual nitrogenous
derivative from Lathyrus odoratus and Lathyrus pusillus. This sub-
stance, )S-N-(Y-L-glutamyl)-aminopropionitrile, is one of the agents
producing the condition known as lathyrism (Selye, 1957), in

CYANOGENETIC SUBSTANCES 1 89

particular the skeletal form of the disease. The active principle is
^-aminopropionitrile. The compound is absent from L. sativus, L.
cicera, L. latifolius, L. strictus, L. splendens, and others (Strong,
1956). However, most of these latter species are positive for the form
known as "neurolathyrism." This fact merely confirms what has been
suspected, namely, that the two forms of lathyrism result from two
different agents. Among the species reported by Selye hardly any
(these exceptions were also doubtful) were positive for both forms of
lathyrism. More recently, Ressler et al. (1961) have identified a neuro-
lathyrus factor from L. latifolius as L-a, y-diamino butyric acid. This
finding has led to a very interesting speculation that the two types of
lathyrus factors are derived from a common precursor. Apparently,
they do not occur together in a plant. The hypothetical scheme in
which the lathyrus factors stem from a parent substance, asparagine,
is shown in Fig. 10-1. Further support for the pathway illustrated in
Fig. 10-1 was provided by the subsequent discovery of the hypothet-
ical intermediate, /3-cyano-L-alanine, in related species, Vicia saliva
and V. angustifolia (Ressler, 1962).

CH2NH2
CH2
COOH — (a)/-C-N

CHNH2

CH2

CONH3 teN ^COOH

asparagine /^-cyano-L-alanine CHNH2

CH2
CH2NH2

Fig. 10-1. Hypothetical pathway in (a) Lathyrus odoratus and
(b) L. latifolius (after Ressler, Redstone and Erenberg, 1961;
reprinted from SCIENCE by permission).

Hegnauer (1959a) has investigated the distribution of cyano-
genetic substances among species of Taxus and certain related genera
{Cephalotaxus and Torreya). The other genera were acyanogenetic as
were certain species of Taxus. In certain cyanogenetic species,
varieties were found to be either negative (T. baccata var. aurea),
weakly cyanogenetic (var. dovastoniana), or strongly cyanogenetic
(var. baccata). According to Hegnauer:

It is interesting, in chemotaxonomic relationship, that the genera
Taxus and Cephalotaxus are clearly phytochemically different. Both
contain alkaloid but the bases are different. Cyanogenesis is found
only in Taxus.

190

BIOCHEMICAL SYSTEMATICS

However, the relationship of Cephalotaxus to Taxus on mor-
phological grounds is not considered to be close. Although formerly
included in the Taxaceae, Cephalotaxus is now considered a separate
family, Cephalotaxaceae (Buchholz, 1951). In view of this pre-
sumed lack of close relationship between Taxus and Cephalotaxus the
statement by Hegnauer has less significance. Cyanogenesis in Taxus
is not to be considered of systematic significance beyond perhaps
additional support for the recognition of varieties. Hegnauer says that
since separation of some of the cultivated forms of Taxus may be
difficult, cyanogen content may serve as a useful character. This
suggestion may be received with some reservation, since the character,
unsupported by correlated morphological differences, is of dubious
value in delimitation of anything more than a single or perhaps a few
genie differences; for example, in Trifolium repens it has been estab-
lished that a single dominant gene governs production of the cyano-
genetic compound (Wilhams, 1939) and another gene the enzyme re-
quired to hydrolyse the cyanogen (Atwood and Sullivan, 1943). If we
knew sufficiently well, the biochemistry of the species and its individ-
uals we might regard plants as biochemical individuals just as
R. J. Wilhams and Reichert before him regard individuals as bio-
chemically unique. 1

1 Reichert, 1919. "Recently data have been rapidly accumulating along many and
diverse lines of investigation which collectively indicate that every individual is a chemical
entity that differs in characteristic particulars from each other."

PHENOLIC
SUBSTANCES

This large and diversified group of compounds con-
tains a number of classes of substances which are
well known. They have been extensively investigated
in spite of a relative lack of economic value within
the group. Except for a few physiologically active
compounds, such as phloridzin and, according
to some reports, rutin, the phenohcs are of little
pharmacological interest, i Even within the plant in
which they occur no physiological function is readily
apparent for most phenolics, though some have been
found to be effective inhibitors of seed germination

iSome isoflavones, particularly those which form a 4th
ring and are therefore rather sterol-like in general configuration,
exhibit estrogenic activity. A potent estrogen of this type is ob-
tainable from Butea superba (Bickoff, 1961).

191

192

BIOCHEMICAL SYSTEMATICS

(de Roubaix and Lazer, 1960), and these may also be self-inhibitors in
the seeds and fruits in which they occur. There are also numerous
scattered reports of phenohc inhibitors of certain fungi and plant
viruses (Uritani, 1961).

Although the amino acids tyrosine and dihydroxyphenyl-
alanine, certain alkaloids, and other substances are phenolic in nature,
customarily the term, phenohc compound, is not extended to include
nitrogenous derivatives. Also, certain phenols are demonstrably
related to a parent substance belonging to a different chemical group,
for example terpenes, as in the case of thymol. Major categories
of phenolic substances include the following: simple phenols without
side chains; simple phenols with one, two, or three-carbon side
chains (occurring as acids, aldehydes, ketones, or alcohols); depsides
of simple phenols (for example chlorogenic acids: see Fig. 11-1);
and higher polymers of simple phenols such as the important
structural component of vascular plants, lignin. Another large and
important group of phenolic substances is the flavonoids, which in-
clude the vacuolar pigments such as the anthocyanins and antho-
xanthins, in addition to other classes. Finally, coumarins, which are
unsaturated lactone derivatives (for example, coumarin and scopoletin),
and which may be derived from the same biosynthetic pathway as that
leadmg to simple phenohc compounds, are also included with the
phenohc compounds. Phenohcs are usually present in the plant as
glycosides or esters.

Certain of the phenohcs have been the objects of a large num-
ber of productive biochemical genetic studies, and also recently there
have been important new advances in knowledge of the biosynthesis
of these compounds. Knowledge of the genetics and biosynthesis of
phenols should contribute to a clearer understanding of the meaning
of some of the results of biochemical systematics studies. For this
reason, brief discussions of the mode of biosynthesis and certain as-
pects of the genetics of flavonoids are included in this chapter.

Present knowledge of the comparative biochemistry of sec-
ondary compounds and particularly their mode of inheritance is often
inadequate to provide much important insight into their systematic
significance in a given instance beyond strict correlations of systematic
distribution. Consequently, many biochemical systematic studies rep-
resent a rather empu-ical search for patterns of distribution of partic-
ular substances or groups of substances. It should be recognized,
however, that for the vast majority of morphological characters used
as systematic criteria, the genetic mechanisms responsible for the
characters have not been revealed either. Therefore, in those cases

PHENOLIC SUBSTANCES

193

phloroglucinol

(Sequoia sempervirens)

/^%^n^

p-hydroxybenzaldehyde

COOH

gallic acid

H0-^^^^C-CH3
~" O

p-hydroxyacetophenone

(Populus trichocarpa)

H3CO

HO-/ V-CH=CH— CH2OH

coniferyl alcohol

mociurin

[Madura tinctoria)

CH=CH— CO— O

HO'

chlorogenlc acid

,COOH

{Coreopsis spp.)
Fig. 11-1. Structural formulas of some simple phenols.

wherein genetic and biochemical mechanisms governing the synthesis
of certain chemical substances are yet undisclosed, it does not neces-
sarily follow that the compounds are accordingly of little value as
taxonomic criteria.

BIOCHEMICAL SYSTEMATICS

194

Some basic considerations of
biosynthetic pathways involved in the
production of phenolics.

The biosynthesis of phenohc compounds has been reviewed
by Neish (1960) and others. A central problem, that of the initial
aromatization, appears to have been solved through investigations
into the biosynthesis of aromatic amino acids in microorganisms. The
presently accepted biochemical pathway to tyrosine and phenyl-
alanine is that elucidated by Davis and coworkers (Davis, 1956; Levin
and Sprinson, 1960) through studies of E. coli biochemical mutants.
The essential features of this scheme are illustrated and the path-
way extended to include several classes of phenols in Fig. 11-2.
Although not all evidence that these pathways are operative in
higher plants is direct, isotope studies from several laboratories pro-
vide independently strongly favorable circumstantial evidence for
such pathways.

There is, now, equally strong evidence from isotope studies
that in the flavonoids one benzenoid portion of the molecule comes
from a quite different pathway, namely head to tail condensation of
three acetyl groups (Rickards, 1961). Confirmation of the theory of
acetate condensation suggested by Birch and Donovan (1953) has
come from studies in four different laboratories in four different
countries, for example, Watkin, et al. in Canada (1957), Grisebach in
Germany (1957), Geissman and Swain in the United States (1957),
and Shibata and Yamazaki in Japan (1957). The acetate conden-
sations are involved in the formation of the benzene ring of the
flavonoid molecule customarily referred to as the "A ring" while the
general pathway to phenols provides the B ring and the three carbons
adjacent to the B ring (see below, formula of quercetin).

quercetin

In quercetin (and its anthocyanidin analog, cyanidin) ring B
is derived from the shikimic acid pathway and ring A from the acetate
pathway. This mechanism is probably generally representative of
flavonoid synthesis, possibly involving a chalkone (see below) inter-
mediate (Grisebach and Patschke, 1961). Hutchinson, et al. (1959)

£ £ 33

OJ 0) p

£ £ 3

3 C

o
-a

p
O

sx
d)

>i
X

p
'■5

-a

"> .,

-o

9 -S

O (/!

3
a"

CO 0^

5 "^^

-^"O

^ ^3

^:2

1 o

c^ cs

■o

-a

M +

■' CD

-»->

-a

3
u
o
o

C
_o

u
CB
O)
!h

a>
m

•+3

S-.

(1)

■^

l«

<*-

-|J

^-A

■^-1

o I

.2 o

bl)

-T!

■^

fTi

C/}

W) «

S .S

195

196 BIOCHEMICAL SYSTEMATICS

have shown, for example, that acetate is preferentially incorporated
into the A ring of phloretin by apple leaf tissue.

phloretin

The Ce — C3 compounds which apparently are synthesized by
way of the shikimic acid pathway are important units in the forma-
tion of lignin in addition to their roles in amino acid and flavonoid bio-
synthesis. While it is not appropriate to include herein a detailed dis-
cussion of lignin biosynthesis and lignin chemistry, it is pertinent to
note that, although the exact structure of hgnin is not known, it
is believed to be a phenylpropane polymer. There is reason to be-
lieve that many different kinds of lignin exist, and a specific lignin
may characterize a particular taxonomic group. The systematic im-
plications of Hgnin chemistry are discussed in Chapter 14.

Coumarins comprise a particularly interesting group of
phenohc compounds. There are a number of different coumarin
derivatives of widespread occurrence, and some, such as scopoletin,
affect plant growth. From the formula of coumarin itself, one un-
famihar with phenol chemistry may conclude that the coumarin
structure is homologous with the A ring plus the heterocycHc ring of
the flavonoid nucleus.

0-^0

coumarin

In such case the benzene ring of coumarin would be expected
to come directly from acetate. Coumarin, however, is a lactone of
o-hydroxycinnamic acid, and tracer studies support the view that
coumarin synthesis follows the shikimic acid pathway (for example,
labelled caffeic acid and labelled scopoletin are formed when
labelled phenylalanine is provided to Nicotiana, Reid, 1958).

The disclosure that isoflavones are formed from phenylalanine
by a mechanism which includes an aryl migration (Grisebach, 1961),
in addition to relating this flavonoid group to the shikimic acid path-
way, suggests that the rotenoids may also be included since rotenoids
bear a structural resemblance to isoflavonoids.

PHENOLIC SUBSTANCES

197

H,C

Grisebach and Ollis (1961) have noted a high frequency of
co-occurrence of isoflavonoids and rotenoids (see Table 11-1), and
furthermore there is a rather close correlation in their group sub-
stitution patterns.

It appears that a single pathway, in reality one that is possibly
as phylogenetically old as the first cellular organisms, leads to the
phenolic amino acids and hence to other phenolics. Secondly, another
pathway, originating also from an important basic metabolite (acetate),
cooperates to yield the complex flavonoids.

Table 11-1. Distribution of flavonoids, isoflavonoids and rotenoids in selected
species (Experientia 17: by permission of Grisebach and Ollis, 1961).

Plant

Flavonoid

Isoflavonoid

Rotenoid

Ferreirea spectabilis
(Leguminosae)

Naringenin

Biochanin-A

Ferreirin

Homoferreirin

—

Prunus puddum

(Rosaceae)

Sakuranin
Sakuranetin
Genkwanin
Taxifolin

Prunetin

Padmakastin

Padmakastein

Mundulea sericea
(Leguminosae)

Sericetin

Mundulone
Munetone

Munduserone

Pachyrrhizus erosus
(Leguminosae)

—

Pachyrrhizin
Erosnin

Rotenone
Pachyrrhizone

Pterocarpus angolensis
(Leguminosae)

Muningin
Angolensin

—

Derris malaccensis
(Leguminosae)

—

Toxicarol isoflavone

Rotenone
Sumatrol
Deguelin
Toxicarol
EUiptone
Malaccol

198

BIOCHEMICAL SYSTEMATICS

In vascular plants, wherein the production of phenylpropane
derivatives in lignin synthesis has been continuously (in evolutionary
time) a major metabohc activity, flavonoid compounds are of general
occurrence, though it is true that phenolics are far less prominent
among vascular cryptogams. In thallophytes, where hgnin does not
occur, flavonoid pigments are practically unknown although other
phenohcs may be numerous. According to Blank (1947) a report of
the occurrence of flavonoids in mosses is probably valid, but no algae
or fungi are known to produce flavonoids. The report of flavonoids in
Chlamydomonas (Moewus, 1950) was invalid (Kuhn and Low, 1960).
Alston (1958) showed that the purple pigment of the green alga,
Zygogonium ericetorum, was not an anthocyanin and that early re-
ports of the occurrence of anthocyanins in filamentous algae were prob-
ably erroneous. It seems therefore that the presence of an enzyme
system leading to lignin synthesis has provided an opportunity for
the appearance of phenylpropane derivatives to couple with the
acetate pathway to form the basic flavonoid nucleus. This step in
biochemical evolution may have been acquired quite early in view of
the wide distribution of flavonoids among pteridophytes, but it is
also possible that it evolved repeatedly.

The next section will be devoted to biochemical genetical
studies of certain classes of complex phenols, the flavonoid pigments.
In order to appreciate fully the implications of such studies a brief sur-
vey of the chemistry of these compounds might prove helpful.

Chemical structures of classes
of flavonoid compounds

This group of compounds contains a Cs-Ca-Ce carbon
skeleton in which the C3 unit links two aromatic groups. The C3 chain
is essentially the key to the different major classes of flavonoids since
these classes are recognized on the basis of the oxidation state of the
C3 unit in addition to the mode of ring closure to form a heterocyclic
middle ring (if ring closure ensues). Flavonoids usually occur as glyco-
sides and sometimes also as acylated compounds, the acyl group being
in many cases a phenoHc acid. Glycosides are mostly formed as esters
at carbons 3, 5, or 7 but some carbon glycosides at position 8 are known
(Horhammer and Wagner, 1961). Flavonoid glycosides are usually
water soluble and are located in the vacuole of the cell. Classes of
flavonoid compounds are discussed below:

PHENOLIC SUBSTANCES

199

ANTHOCYANINS

The basic aromatic unit is referred to as a phenylbenzo-
pyrilium salt with the configuration shown below. This class includes
most of the red and blue plant pigments.

6' 5'

All naturally occurring anthocyanins have the 4', 3, 5, and 7
positions occupied by an hydroxyl or some substituted group. They
usually occur as glycosides with the sugars attached at positions 3 or
3 and 5. Sugars commonly reported are glucose, galactose, rhamnose,
and arabinose. Disaccharides (for example, rhamnoglucose) may
occur as well as 3,5-dimonosides and even trisaccharides.

In acylated anthocyanins the organic acid is frequently p-
hydroxybenzoic, protocatechuic, p-hydroxycinnamic or other
phenolic acid. The attachment of the acyl group is apparently at
a free hydroxyl in the ring or an hydroxyl group of the sugar.

The aglycone of the anthocyanin, which may be obtained by
acid hydrolysis, is referred to as an anthocyanidin. Although dozens
of anthocyanins have been described, only a few anthocyanidins are
known, and some of these are rather rare. These compounds differ in
the substitution pattern involving positions 3', 4', 5', 3, 5, and 7.

Representative anthocyanidins.

,0H

B VOH

pelargonidin

cyanidin

pCHa

delphinidin

OCH3
malvidin

Hirsutidin is a 7-methoxy analog of malvidin. Capensinidin, a 5-
methoxy analog of malvidin, has been obtained from Plumbago
capensis (Harborne, 1962).

QCH3

peonidin

Rosinidin, a 7-methoxy analog of peonidin, has been reported to occur
in Primula rosea (Harborne, 1958).

200

BIOCHEMICAL SYSTEMATICS

FLAVONOLS

These, like most of the flavonoids other than anthocyanins,
are colored yellow or cream or have hardly any color. The basic
flavonol nucleus is illustrated below.

kaempferol

Substitutions, similar to some of those illustrated in the anthocyanins,
also occur in the flavonol class. For example, other well known
flavonols are quercetin (analogous to the anthocyanin cyanidin), and
myricetin (analogous to the anthocyanin delphinidin). Flavonols are
common flavonoid constituents and, like the anthocyanins, widely
distributed.

FLAVONES

These are similar to flavonols but lack the 3-hydroxyl group.

FLAVANONES

HO O

FLAVANONOLS

HO O

PHENOLIC SUBSTANCES

201

ISOFLAVONES

l^r\

HO O

CHALKONES

■0H^_/~\

HO O
AURONES (BENZALCOUMARANONES)

This group is distinguished by the presence of a five-
membered heterocychc ring.

HO. ^ ,0^

CATECHINS

LEUCO-ANTHOCYANINS

BIFLAVONYLS
HO

o-^^ /

HO HO

In addition to the many types of flavonoids already described
there are a few flavonoids substituted with isoprene units. The
substances artocarpin and isoartocarpin, found in the wood of Arto-
carpus integrifolia, are particularly interesting representatives since

202 BIOCHEMICAL SYSTEMATICS

they also possess the rare ortho-hydroxy substitution of the B-ring
(Dave etal, 1962).

H3CO

artocarpin isoartocarpin

The isoprenoid side chains which distinguish the two compounds
are hnked differently, that is, C-C in artocarpin and 0-C in isoarto-
carpin. Yet, spatially the ortho-OH of the B-ring is close to the 3
position at which the isoprene substitution in artocarpin occurs.

Flavonoids are of special interest in that they represent a
molecular composite formed via several basic pathways each of which
leads to other secondary compounds: the shikimic acid pathway,
mevalonic acid pathway, and acetate condensation. Other examples of
phenolic-isoprenoid derivatives are known, including other flavonoids,
rotenoids, coumarins, and quinones (Ollis and Sutherland, 1961).

When all the known derivatives of the classes of flavonoids
including glycosides, are totaled, they number into the hundreds.
Geissman and Hinreiner (1952) listed almost 200 different flavonoids
already known to occur in nature, and many new types have since
been described (for a recent comprehensive list see Geissman, 1962).

Genetic studies concerning the
flavonoid compounds

As noted previously the inheritance of certain flavonoid pig-
ments has been studied more intensively than perhaps any other group
of chemical substances in flowering plants (Alston, 1963). The antho-
cyanins, particularly, have been the objects of numerous investigations
extending back almost to the nineteenth century. Onslow (1916) called
attention to the possibility of biochemical genetic studies of anthocy-
anins shortly after Willstatter had established their chemical nature.
Apparently the first actual biochemical genetic investigation was that
of Scott-Moncrieff (1931) who showed that in Pelargonium zonale a
dominant gene, producing a rose-pink flower, governed the formation
of a cyanidin glycoside. The double recessive, in contrast, contained a
pelargonidin glycoside and was salmon-pink in color.

By 1936 a number of biochemical-genetic studies of flower

PHENOLIC SUBSTANCES 203

color had been completed, and Scott- Moncrieff in reviewing this work
outlined several generalizations concerning the inheritance of antho-
cyanins, such as the fact that the more oxidized form was usually
dominant to the less oxidized, and that 3-5 diglycosidic and acylated
anthocyanins were dominant to the 3 monoglycosidic and non-
acylated forms, respectively (Scott-Moncrieff, 1936).

Beale et al. (1941) in another important review of the subject
concurred in general with the findings of Scott-Moncrieff. The num-
ber of species which had at that time been investigated was surpris-
ingly large though most of the work suffered from limitations of the
techniques then available. Between 1941 and the early 1950's rela-
tively little additional work on the inheritance of flower color was re-
ported. Haldane (1954), who apparently had interested Scott-Moncrieff
in the subject, outlined some of the problems which remained un-
solved at the time and deplored the declining interest in the study of
the biochemical genetics of flower color. Yet, even then a number of
important studies along these lines were in progress. Apparently, re-
newed interest stemmed in part from the introduction of paper
chromatographic techniques. Before such techniques appeared, it was
almost impossible to resolve the anthocyanins, yet complex mixtures
of pigments were frequently encountered. The first report of the use
of paper chromatography in the study of anthocyanins was that of
Bate-Smith (1948), and most, if not all, of the major biochemical-
genetic work on anthocyanins since has been facilitated by paper
chromatographic investigations. Several significant publications on
the inheritance of flower color have appeared in recent years, yet
these have not answered some of the basic questions of flavonoid bio-
synthesis which now center on interconversions of classes of flavo-
noids, the point at which substitutions in the A and B rings occur, and
the exact mode of union of the A and B units of the flavonoid nucleus.
Some consideration will be given to these points later.

The extent of genetic investigations of flower color is empha-
sized by the work of Paris et al. (1960) who surveyed publications
treating the inheritance of flower color in seventy-five different species.
These workers attempted the formulation of a general inheritance
scheme governing flower color. They recognized six major analogous
genes on the basis of the frequency of appearance of the correspond-
ing phenotypic effect. While it is unquestionably desirable to attempt
to develop an integrated system of genetic notation in which factors
known to have equivalent biochemical expression are assigned the
same symbol, it is doubtful that the arbitrary recognition by these
authors of six types of analogous genes based entirely on the pheno-
typic expression of color alone is a positive contribution. Rather, it

204 BIOCHEMICAL SYSTEMATICS

oversimplifies the situation and conveys to the casual reader the idea
that the gene categories are possibly biochemically as well as pheno-
typically analogous when, in fact, it is demonstrable that in numerous
cases they are not.

As a result of a series of biochemical genetic studies involving
numerous plants several types of biochemical differences attributable
to single gene differences have been reported. (Since many of these
have been confirmed several times by different workers, only cases of
some special interest will be identified by citation.) In a number of
instances genes are known to govern the substitution pattern of the
B ring, that is, the number of hydroxyl groups present. Sometimes,
for example in Streptocarpus (Lawrence et al., 1939), a dominant gene
governs the formation of malvidin instead of pelargonidin. In this
case, it is possible that the gene permits the addition of one or more
OH groups in the B ring (or a precursor thereof ) and thus provides a
site for methylation so that a single gene may appear to govern a
more complex biochemical process than is actually the case. A similar
situation probably occurs in Impatiens (Alston and Hagen, 1958). Of
course, it is possible that the gene governs methoxylation, but present
evidence does not permit a choice between these alternatives. It is in-
teresting that, to the writers' knowledge, there is no report in the
hterature of a gene which governs substitutions in the A ring other
than the glycosidic pattern. It is highly probable that such genes
exist, since hirsutidin (a 7-methoxy malvidin), gossypetin ( a flavonol
with an 8-hydroxy substitution) and other compounds with atypical
A ring substitution patterns exist.

Numerous instances of the occurrence of single genes which
affect the glycosidic pattern are known, and as noted previously the
diglycoside is dominant to the monoglycoside.

There are several instances known of single genes which
govern acylation. Abe and Gotoh (1956) reported a dominant gene
governing acylation with p-hydroxy cinnamic acid in the eggplant,
and Harborne (1956) reported an interesting situation in Solarium
in which a single gene appeared to govern three biochemical differ-
ences in the same anthocyanin: a change in substitution of the B ring,
a change in the glycosidic pattern, and acylation of the glycoside
with p-coumaric acid.

There are numerous examples of genetic mechanisms which
involve interactions between anthocyanins and other classes of
flavonoids. In Dahlia, the classic example of such interaction
(Lawrence and Scott-Moncrieff, 1935), one factor, I, governs flavone
synthesis at (apparently) the expense of anthocyanin. The authors
concluded from these results that a precursor, hmited in amount, was

PHENOLIC SUBSTANCES 205

common to all the pigments. A similar type of competition is reported
in Primula (de Winton and Haldane, 1933). In Impatiens balsamina,
the gene L allows production of malvidin type anthocyanins and also
the related flavonol, myricetin (Clevenger, 1958), yet there does not
appear to be competition between anthocyanins and flavonols in this
plant since a fifty-fold increase in pelargonidin content of flowers does
not reduce appreciably the amount of its flavonol analog, kaempferol
(Hagen, 1959). A dominant gene which effects production of two
different classes of 3':4':5'-trihydroxylated pigments is known in
Solanum phureja and also in Primula sinensis. In Dianthus a dominant
gene, R, introduces cyanidin and its flavonol analog, quercetin, while
pelargonidin and its analog, kaempferol, occur in the absence of R
(Geissman et al., 1956).

One of the most informative examples of interaction between
several classes of flavonoids is that of Antirrhinum (Sherratt, 1958)
(Fig. 11-3). The types of flavonoids which occur in Antirrhinum are
anthocyanins, flavonols, flavones, and aurones. In most of these classes
more than one representative aglycone type is present, though not
necessarily together in a single plant. Genetic control of flower color
in Antirrhinum has been investigated by several groups independently,
and the present discussion is taken from Sherratt (1958) using the
genetic symbols of Dayton (1956). A factor, Y, is necessary for the
formation of flavonoids. Unless certain other dominant genes are
present, however, only pigments of the flavone and aurone types are
found (namely, apigenin and aureusidin). In the presence of the
double recessive, IaIa, aureusidin content is increased with no
apparent reduction in apigenin content. In fact no other factors under
consideration appear to affect apigenin. The gene, R, governs simul-
taneously the appearance of anthocyanins and flavonols (both classes
have a 3-OH in the heterocyclic ring). Gene B governs the substitu-
tion pattern of the B ring, introducing dihydroxy rather than mono-
hydroxy derivatives in the anthocyanins and flavonols present. Gene B
does not affect the other two classes of flavonoids. The interpretation
of these data is implicit in figure 11.3. Notably, it appears that the
pathway to aurone synthesis is determined rather early. Jorgensen and
Geissman (1955) have shown that increased anthocyanin synthesis re-
sults in some lowering of the aureusidin content however.

In Phaseolus a series of alleles, C", C and C' influence relative
quantities of flavonols and anthocyanins as well as the substitution
pattern of the B ring (Feenstra, 1960). None of these substances is
formed in the presence of the recessive, C"; C along with the factor
viae results in the formation of flavonols of the kaempferol type plus
a small amount of the quercetin type and no anthocyanins; C^^ with

'c
o

^3
1=1

o '^

o .

c S

4) O

-M -a
It

22 '^

s • —

<u >

3 +^

"o C

■2 S

<A bX)

s «

• b >.

O ^

o "

> T3

^ -I

^ X

o) o

" a

+^ &

s
s

o ^

O 00

t/j *^

0) cS

be _4, ..

bc ^ O

3 T3 "5

C» C CO

eo CO '-iS

206

PHENOLIC SUBSTANCES 207

yiae results in the formation of both flavonols and anthocyanins of
monohydroxy or dihydroxy types. The allele C in the presence of the
dominant allele V, governs flavonols and anthocyanins mainly of the
type, 3'-4'-5'-trihydroxy, and C^ with V yields anthocyanins only,
these having the trihydroxy substitution in the B ring. This case rep-
resents an unusually complex form of interaction which Feenstra
interprets as indicating that a shift in the hydroxyl pattern of the B
ring to the trihydroxy configuration favors anthocyanin synthesis over
flavonol synthesis.

Since it has already been established that the acetate and
shikimic acid pathways are involved in both anthocyanin and flavonol
synthesis, it is hardly surprising to find a number of instances of inter-
actions— in fact a number of expressions of this interaction— between
classes of flavonoids. Similarly, however, the total absence of a definite
instance of gene-controlled direct interconversion suggests that actual
interconversion of the classes of flavonoid pigments is not the rule. In
this connection it now appears that leucoanthocyanins, once thought
of as likely precursors to anthocyanins, do not function in this way.
Evidence is not unequivocal on this point, however. Genes affecting
leucoanthocyanins are known. In Impatiens a gene governs the
presence of a pelargonidin-type anthocyanin and, in addition, leuco-
pelargonidin (Alston and Hagen, 1955). Feenstra (1959) reported a
gene in Phaseolus governing the appearance of leucoanthocyanin.
The leucoanthocyanins, incidentally, provide some circumstantial
evidence favoring the position that methylation occurs at a late stage
in anthocyanin synthesis. Methylated leucoanthocyanins are prac-
tically unknown yet leucoanthocyanins are commonly found along
with methylated anthocyanins.

The quantitative inheritance of flavonoids is provided with
some interesting illustrations. In some plants a rather large number
of genes may influence the amount of anthocyanin. In Primula, for
instance, at least four different loci contain dominant intensifiers for
anthocyanin, and five loci contain dominant inhibitors (de Winton
and Haldane, 1933). In Dahlia (Lawrence and Scott-Moncrieff, 1935)
two loci affect the amount of anthocyanin and two others affect the
amount of yellow flavonoid pigments. From studies of the physiology
of anthocyanin synthesis it is clear that a host of extrinsic factors can
modify anthocyanin content, in fact a number of generaUy harmful
influences actually bring about increased anthocyanin synthesis. It is
therefore to be expected that a large number of different genes would
achieve a similar effect through diverse means. From a systematic
viewpoint, gene homology between two factors which exert quantita-
tive effects on anthocyanin synthesis in two different species has a

208 BIOCHEMICAL SYSTEMATICS

low probability. The same interpretation may be expected to hold for
the surprisingly large number of complex loci which affect, quantita-
tively, anthocyanin synthesis. Alston (1959) has discussed certain im-
plications of the existence of such loci in a large proportion of plant
species studied, and it is pertinent to note that homologies among such
complex loci are considered likely to be rare.

The foregoing discussion serves to provide a perspective from
which to view certain systematic investigations involving the flavo-
noid pigments or simpler phenols. Several illustrations have been
selected which disclose that a single gene may alter several bio-
chemical components of a plant (in one further case, the P*" allele of
Impatiens balsamina governs not only the amount of anthocyanin in
the stem, sepals, and petals, but in addition has a different qualitative
expression in each plant part). Despite such examples, it seems im-
proper to conclude, in the absence of genetic criteria, that when re-
lated anthocyanins and flavonols occur together or when similar
glycosides of anthocyanins and flavonols occur together, the same
enzyme (or gene) is necessarily implicated. In one such situation cited,
involving Lathyrus odoratus (Harborne, 1960a), this assumption was
made after examining a number of varieties but without benefit of
genetic studies. In a previous genetic study Beale (1939) reported
that the genes affecting anthocyanidin type did not influence flavonol
composition in Lathyrus odoratus.

Just as cases are known in which one gene governs several bio-
chemical differences, there are instances in which several different
genes may affect the same biochemical character. It seems to be
established that numerous gene effects are highly indirect, the
primEiry gene effect remaining completely unsuspected.

Gene mutations affecting relatively late stages in the flavo-
noid biosynthetic pathway appear to be far more frequently detected
than those affecting an early step. This assumption is based on the
rarity of cases in which a gene is known to inhibit the total synthesis
of aU flavonoids (either as a dominant or recessive). Most "white"
mutants involve the anthocyanins and in such mutants other types
of flavonoids may still be produced. Roller (1956) who studied the
flavonoids of certain white-flowered varieties of over forty species,
found other types of flavonoids present in practically every instance.
From two to six flavonoids were present, as a rule, with flavonols
most frequent. The infrequency of cases involving mutations inhibit-
ing total flavonoid synthesis was also noted in the discussion follow-
ing a recent paper on anthocyanin genetics by Harborne (1960b).

On a priori grounds, one may predict that the earlier

PHENOLIC SUBSTANCES 209

in the pathway the block to the synthesis of a secondary substance
occurs, the more hkehhood that the metaboHsm of a basic metabohte
is affected adversely. It follows that the early stage mutants would be
ehminated more often. Modification of a terminal step is also less
likely to provide the opportunity for the appearance of a new series
of compounds. The question of biochemical selection is also pertinent.
If selection becomes more critical, then the earlier in a sequence of re-
actions the change represented by the mutant occurs, the more likely is
its preservation to become dependent upon the total gene pool. Thus, in
general, the preservation of such a mutation rests upon a broader
underlying genomic constitution than that of a mutation affecting a
terminal step. Perhaps such considerations are purely academic at the
moment with respect to the systematic implications of biochemical
data, but they are nontheless potentially significant. Such considera-
tions bear upon the question of whether each newly acquired synthetic
ability should be given the same weight of systematic significance. In
the writers' opinion they should not. Even without consideration of
the actual systematic distribution of the compounds involved, the
appearance of an aurone (with a five-membered heterocyclic ring)
may be more significant than the appearance of a different glycosidic
pattern, although in the former case the empirical chemical formula
remains the same while in the latter it may be radically altered.

In summary it is evident that knowledge of the major bio-
synthetic route and some familiarity with the mode of inheritance of
a group of related chemical constituents should allow more critical
analysis and a more precise evaluation of the systematic implications
of a given distributional pattern.

Systematic aspects of the distribution of
phenolic compounds

The use of phenolic substances in systematic investigations
does not extend back as far as that of certain other groups of plant
constituents such as alkaloids and essential oils. In the past decade
a number of investigators have considered the phenolics, particularly
anthocyanins, leucoanthocyanins, flavonols, and phenolic acids. Bate-
Smith has stressed especially the leucoanthocyanins. In his first treat-
ment of the systematic distribution of leucoanthocyanins (Bate-Smith
and Lerner, 1954), over 500 species were surveyed for leucoantho-
cyanins in leaves. In general these compounds are more abundant
in woody families, especially in certain groups regarded by some

210 BIOCHEMICAL SYSTEMATICS

workers as primitive. In the predominantly herbaceous famiHes fewer
species contain leucoanthocyanins. In the family Leguminosae, which
is predominantly woody, both positive and negative results were
obtained, though most herbaceous members of the sub-family
Papilionoideae tested negatively. Within the Papilionoideae, one
tribe, Hedysareae, contains many positive species (ten out of eleven
tested). Bate-Smith reported that in many legumes the seeds tested
positively for leucoanthocyanins even when the leaves of the same
species were negative, a fact which complicates the interpretation of
distributional data.

In view of the above, Bate-Smith believes that families of
Hutchinson's Herbaceae should be examined for the presence of
leucoanthocyanins. It should be apparent that this application of
leucoanthocyanins as systematic criteria is at the higher taxonomic
levels (family, in this case). The character is not constant within
a family, necessarily, or even within a genus, and its systematic value
at the level indicated is somewhat questionable. Alston (unpublished)
has shown that leucoanthocyanins do not, apparently, serve to clarify
relationships within the genus Prosopis. In specific cases, however,
there is no reason to doubt that leucoanthocyanins may be valuable
as systematic characters particularly at lower taxonomic levels. For
example, in Iris leucoanthocyanins are virtually restricted to the
section Apogon wherein they are found in four out of seven species
examined (Bate-Smith, 1958).

There is some correlation between a morphological and
chemical character in the Papilionatae in that the groups having
a pulvinus at the base of the leaf (the pulvinate condition is
regarded as the primitive condition), including the tribes Sophoreae,
Dalbergieae, Phaseoleae, and parts of Galegeae and Hedysareae, tend
to be positive for leucoanthocyanins. The groups lacking a pulvinus
tend to be negative for leucoanthocyanins. The pulvinate condition is
also distinctly correlated with woodiness as opposed to herbaceous
habit, a factor which Bate-Smith also believes to be significant.
Among the herbaceous monocots, about equal numbers of positive
and negative species are known. Positive species are common among
gymnosperms.

With respect to qualitative aspects of leucoanthocyanins
Bate-Smith (1957) reported that the leucodelphinidin was common
among certain orders (for example, Rosales), but in other orders only
leucocyanidin occurred (as in Ranales). In Myrtales, leucodelphinidin
was quite common as was the analogous flavonol, myricetin, and the
tri-hydroxy derivative ellagic acid:

PHENOLIC SUBSTANCES 21 1

Thus, in Myrtales, the tri-hydroxy configuration is expressed within
several phenohc classes and is emphasized to the extent that it
assumes some systematic significance. Bate-Smith noted that
among the six sub-families of the Rosaceae proposed by various
authors, four (Spiraeoideae, Pomoideae, Prunoideae and Rosoideae)
contain only leucocyanidin while one (Chrysobalanoideae) contains
both leucocyanidin and leucodelphinidin. Bate-Smith noted that the
last-named sub-family had been treated as a family by at least
one worker.

The situation described above emphasizes one of the most
vexing problems facing the systematist, regardless of whether or not
he is concerned with biochemical data, namely, the proper systematic
evaluation of a particular correlation which has been established. In
this case the Chrysobalanoideae as well as the other sub-families of
Rosaceae have already been recognized as distinctive on morpho-
logical grounds, and in fact most have been treated as separate
families on occasion (Lawrence, 1951). The question is how much
additional distinctiveness is implied by the presence of leucodelphini-
din in this sub-family alone (information was not available for the
Neuradoideae). Unfortunately, we are not yet in a position to give un-
qualified opinions in many cases such as this, but conservatism with
respect to the systematic evaluation of leucoanthocyanins is justifiable.

In his general article on the taxonomic aspects of phenolics
Bate-Smith (1958) noted that three classes of these compounds are
widespread in their distribution in leaves of higher plants. These sub-
stances are leucoanthocyanins, flavonols and hydroxycinnamic acids.
He concluded that perhaps the absence of certain of these common
substances might be more significant than their presence. For in-
stance, leucoanthocyanins are for the most part absent from the orders
Centrospermae, Umbelliferae, and Contortae; entirely absent from
the Rhoeadales, Tubiflorae, Plantaginales and Cucurbitales; and al-
most entirely absent from the Campanulatae. They are also absent
from many families of the Ranales. Many of these same orders do not
produce flavonols. In connection with his discussion of flavonols Bate-

212

BIOCHEMICAL SYSTEMATICS

Smith states that among sympetalous families, those with zygomor-
phic flowers often lack flavonols in the leaves, and this fact may pro-
vide a clue to the morphological character with which the absence of
flavonols might be linked. It is important to establish any correlation
between metabohsm and form since the biochemical basis of develop-
ment in all its stages is so little known. However, the present writers
beheve that many, if not most, secondary compounds have no critical
role in morphogenesis, and consequently correlations between a given
chemical and anatomical character may be merely coincidental.

Many biochemical studies of phenoUcs have emphasized the
distribution of some of the commonest phenolics, for example, fla-
vonols and certain cinnamic acid analogs such as ferulic and caffeic
acids. This trend is natural, particularly when investigations are con-
ducted by biologists who must rely upon relatively simple chemical
procedures. However, in many cases the actual systematic value of the
studies may not be great, particularly when minor systematic cate-
gories are being considered. As illustrations of the broad distribution
of certain phenolics, Tomaszewski (1960) surveyed 122 species of 86
famihes for p-hydroxybenzoic acid and other simple phenols and
found that p-hydroxybenzoic acid was present in 120 of the 122
species tested. Caffeic acid was present in all but about a dozen
species. [The author stated that caffeic acid was absent from all
gymnosperms and legumes, but none of the species investigated in
the study was named, and since, of 122 species, a total of 86 different
families was included, not many gymnosperms or legumes could have
been examined. Pecket (1959) found caffeic acid in most species of the
legume, Lathyrus.] Ferulic acid occurred in 63 per cent of the species.
p-Coumaric acid occurred in all but three genera, but catechol was
found only in the Salicaceae wherein it occurred in all species examined.
Takahashi et al. (1960) found quercetin and kaempferol to be widely
distributed in the order Coniferae, and no significant pattern was
estabhshed. The chief significance of these studies is the additional
evidence adduced for a very wide distribution of certain phenohcs.
This fact does not exclude them from systematic utility within a
particular taxonomic group, since in combination with other chemical
constituents they might prove significant in individual cases.

An extension of such investigations to the more complex or
the more restricted phenoHc types should yield data of more obvious
meaning, in some cases even at the infra-specific level. One excellent
example has been described by Wilhams (1960). Apple and pear
are significantly different in their phenohc chemistry, the distinc-
tions remaining consistent even though individual varieties (cultivars)
of apple fruit (but not leaves) vary greatly in this respect. Apple

PHENOLIC SUBSTANCES 213

contains the rather uncommon dihydrochalkone, phloridzin, as
its principal phenolic, while pear contains another uncommon
phenolic, arbutin, a glycoside of hydroquinone. Phloridzin is absent
from all pear species while arbutin is absent from apple. Even more
significant is the disclosure that, among the twenty-five species of
apple other than the cultivated apple (Malus pumila), in most species
phloridzin is the dominant phenolic, but in some species phloridzin is
reduced greatly in amount, and another dihydrochalkone glucoside
occurs, the second containing one more phenolic hydroxyl group and
with glucose attached at a different position than in phloridzin. The
second compound is found, with the exception of one variant of one
species, only in the four species from eastern Asia comprising the
series Sieboldianae. It is difficult to ignore the phyletic significance of
such data.

Other examples of the potential value of rather unusual
phenolics which have a restricted distribution are the isoflavones and
the rotenoids. The former are reported only from the Rosaceae,
Leguminosae, Moraceae, and Iridaceae. In the first two instances,
since these families are closely related and often placed in the same
order, phyletic significance may be inferred while the other cases
doubtlessly represent convergent evolution. The rotenoids are, to the
writers' knowledge, restricted to the Leguminosae. The presence of
both isoflavones and rotenoids together is further circumstantial
evidence of a biosynthetic relationship between the two chemical
classes as suggested on chemical grounds earlier in this chapter. It
would be interesting to know if any species of Rosaceae produce
rotenoids. In general the phenolic chemistry of the Rosaceae and
Leguminosae are not similar (Bate-Smith, 1961).

In the genus Iris the distribution of isoflavones appears to be
correlated with the morphological species groups delimited by taxon-
omists. Isoflavones are found only in the sections Evansia and
Pogoniris, considered as equivalent eastern and western groups. As
noted earlier, section Apogon contains most of the leucoanthocyanin-
positive species. Bate-Smith noted that leucoanthocyanins are gen-
erally found in the mesic species of Iris and that this generaliza-
tion seemed to apply to other monocots as weU. Recently, Reznik and
Neuhausel (1959) reported on the occurrence of colorless anthocyanins
in submerged aquatics. They found that a large number of such
aquatic species contained a high concentration of colorless antho-
cyanins, but these were not leucoanthocyanins. Rather, they were
presumed to be the pseudobase form of the anthocyanin which turns
red in HCl in the cold. True leucoanthocyanins must be heated in
rather concentrated HCl to produce corresponding anthocyanidins.

214 BIOCHEMICAL SYSTEMATICS

The existence of colorless anthocyanins in numerous monocot and
dicot groups was noted, though they were not detected in certain
families which are predominantly aquatic, for example, Potamogeto-
naceae. In most cases the pigments are cyanidin derivatives, the most
commonly encountered types of anthocyanins in vegetative tissue.
The formation of these colorless anthocyanins may involve some type
of selection which results in the appearance of a physiological state
permitting the anthocyanins to exist in the pseudobase form. Possibly,
some of these may have been misidentified as leucoanthocyanins.

Some recent phenolic studies having systematic implications
are those of Pecket (1959, 1960a, 1960b) on Lathyrus; Griffiths (1960)
on Theobroma and Herrania; Reznik and Egger (1960) and Egger and
Reznik (1961) on Hamamelidaceae and Anacardiaceae; Bate-Smith
and Whitmore (1959) on the Dipterocarpaceae; Bate-Smith (1961) on
Prunus and Potentilla; Riley and Bryant (1961) on Iridaceae and
Billek and Kindl (1962) on the Saxifragaceae. In each of these studies
variations in patterns were observed, though the authors did not in
all cases consider the systematic significance of the patterns. Perhaps
more important than the establishment of taxonomic affinities at this
stage of such work is the fact that species can be distinguished from
other related species by the phenolic characters compared.

In the Lathyrus study several systematic judgments were made
on the basis of the various phenolic patterns established for certain
species, but the present writers, after examining the data offered, and
in consideration of the general characteristics of the compounds, would
be more conservative. In a genetic study of Lathyrus odoratus (Beale,
1939), flavonoid inheritance was shown to be quite complex with a
number of chemical phenotypes represented within a single species.
However, it is true that wild species, such as those studied by Pecket,
tend to have fewer variations than cultivated species. If the results of
studies of the comparative chemistry of the non-protein amino acids
of Lathyrus (Chapter 6) and the toxic nitriles (Chapter 10) are
integrated with the phenolic data, interesting taxonomic conclusions
may be possible.

Some of these phenolic studies tend to exaggerate the
systematic implications of the data. It is natural that enthusiasm
will sometimes exercise a subtle influence to magnify the positive as-
pects of interpretation, but the occasional direct assertion that the
particular biochemical data do not provide any clues to systematic
relationships should be anticipated. When only a few compounds are
being considered and when only a few individuals of a selected group
of species are screened, such results would not be cause for repudia-
tion of the methods, nor would they even be surprising. Studies such
as that of Stoutamire (1960), though preliminary in nature, show a

PHENOLIC SUBSTANCES 215

clear-cut rationale, and the data are evaluated conservatively. In this
work several species of Gaillardia were analyzed for anthocyanins.
Inter-specific differences involved particular cyanidin glycosides, and
the patterns conformed somewhat to the sub-generic disposition. Color
variations of geographic races of G. pulchella were found to involve
only quantitative differences in the three anthocyanins present. A
quite similar study of Papaver species was reported by Acheson et al.
(1956) with similar conclusions. Griffiths (1960), who compared the
seed polyphenols of various species of Theobroma and the related
genus, Herrania, could discover no general taxonomic implications
aside from the fact that he concluded that it is reasonable to suppose
that the genus Herrania is closely related to Theobroma. This con-
clusion, based solely on the chemical data, is questionable, however,
because again, only the common polyphenols were considered.

The phytochemical systematic studies of Erdtman (1956, 1958)
are especially interesting. He found a distinctive combination of
phenolic substances in the heartwood of the genus Pinus where
the compounds accumulate as inert deposits. Erdtman believes
that secondary constituents are generally far more useful in sys-
tematic studies than the basic metabolites such as sugars, certain
common fatty acids and amino acids. This same position has been
taken by others as noted elsewhere. Erdtman favors the bark and
wood constituents. He states (1956):

It is clear that compounds which occur in phylogenetically young,
highly specialized organs will possess a lesser taxonomic interest
especially when they take part in some of the biochemical processes
specific to the organ.

If the statement given above is intended to refer to flower
parts, it is more applicable to phylogenetic problems involving the
higher taxonomic categories, perhaps not at all applicable to problems
of systematics of the lower categories.

Phenolic compounds from the genus Pinus include the
following:

(1) Stilbenes

CH ^ ,

pinosylvin

pinosylvin monomethyl ether
pinosylvin dimethyl ether

216

(2) Dibenzyls

BIOCHEMICAL SYSTEMATICS

CH2

\h2

dihydropinosylvin monomethyl ether

(3) Flavanones

HO HO O

pinocembrin pinostrobin

cryptostrobin, either 6 or 8-methyl pinocembrin

(4) Flavones

tectochrysin, 7 methoxychrysin
strobochrysin, 6 methylchrysin

(5) Flavonols

pinobanksin (2:2-dihydrogalangin)
strobobanksin 6-(8?), methyl
2 : 3-dihydrogalangin

(6) Flavonols: none

(7) Cyclitols: pinitol (d-inositol, sequoyitol and myoinositol
reported from P. lambertiana)

According to Erdtman the generic sub-groups Haploxylon and
Diploxylon differ chemically, but Mirov, as noted elsewhere (Chapter
13), did not find this to be a feature of the terpene chemistry of the
genus. The tables in Erdtman however show clearly that the Diploxy-
lon group has the simpler heart wood chemistry; for example, most of
the compounds listed above are to be found in Haploxylon species

PHENOLIC SUBSTANCES 217

whereas, in Diploxylon only pinosylvin, its monemthyl ether, pino-
cembrin and pinobanksin occur. Erdtman states:

The observed differences between Haploxylon and Diploxylon are of
such nature that one is led to conclude that the Haploxylon pines have
an oxidation-reduction system at their disposal which has disappeared
or is defective in the case of the Diploxylon pines. Since "loss" muta-
tions are more common than progressive mutations, it is probable that
Haploxylon is more primitive than Diploxylon. Alternatively the
separation has taken place already at an earlier phylogenetic stage.2

Furthermore, Erdtman states that "more powerful methylating
systems" are characteristic of the Haploxylon pines, species of which
contain carbon methylated flavones and flavanones.

Outside the genus Pinus some other interesting situations are
discussed by Erdtman. For instance, of fourteen Tsuga species known,
five were investigated, and all contained the lignan conidendrin, an
unusual substance characteristic of Picea.

H3CO
HO

OCH3
H3CO

conidendrin

All Larix species investigated contained aromadendrin (2:3
dihydrokaempferol) and taxifolin (2:3 dihydroquercetin).

aromadendrin

Taxifolin has also been reported in Pseudotsuga taxifolia.

2 The argument that Haploxylon is more primitive than Diploxylon may be valid,
even on the chemical grounds, but not upon the logic that loss mutations are more fre-
quent than progressive mutations, a statement which appears to be a non-sequitur. There
are examples, in biochemical systematics in which a "loss" is postulated, and, accordingly,
the simpler compound is regarded as phylogenetically more advanced. Thus, Gottlieb
et al. (1959) reported that in certain Aniba (Lauraceae) species (for example,
rosewood) four methoxylated a-pyrones occur, while in others (for example, coto) only the
unsubstituted «-pyrones occur. These authors consider the plain a-pyrones of more recent
phylogenetic origin, but their argument rests on the observation that current theories of
the biogenesis of a-pyrones involve an expected oxygen fimction at position 4.

218 BIOCHEMICAL SYSTEMATICS

Roberts et al. (1958) investigated the phenolic constituents of
tea varieties as well as other species of the genus Camellia, and their
results are of considerable interest because of the type of problem in-
volved. It is rather Ukely that prolonged cultivation of the tea plant
may have almost obliterated the recent natural species history. In
fact, according to Kingdon-Ward (1950) "wild tea" as such, no longer
occurs, and despite the fact that tea taxa are recognized, the large
number of cultivated "varieties" must be subjectively assigned to one
of several major cultivar types or else they are classified as putative
hybrids. Roberts et al. initiated their study on the premise that, "If
the chemical compound could be shown to be a feature of one or the
other of the taxa conceived by botanists, then the chemical definition
could be accepted as relevant to a natural system of classification and
need not be regarded as a special or artificial classification restricted
to the circumstances of cultivation."

Tea plants are usually considered derived from Camellia
sinensis (China tea) or C. sinensis var. assamica (Assam tea). A
rather extensive phenolic complex is typical of the vegetative shoots
of the species, including several catechins, depsides such as galloyl-
quinic and chlorogenic acids, flavonols, anthocyanins, and leucoantho-
cyanins. Trihydroxy derivatives (for example, galhc acid, gallocate-
chin, myricetin, and leucodelphinidin) of these classes are prominent
in the species. Anthocyanin is more characteristic of shoots of the
China variety. In general, Assam tea lacks anthocyanin. The so-
called "southern" form of Assam, in the opinion of Roberts et al., has
been crossed with the China variety, and this accounts for the appear-
ance of anthocyanin in the form. It is interesting that these authors
reported that leucoanthocyanins were of sporadic occurrence, some-
times absent, sometimes abundant, and it was not possible to
associate them with a particular kind of tea.

These investigators further found that triglycosidic flavonols
were common in the China variety but not in the Assam tea variety
(except in trace amounts in some instances). An independently iso-
lated southern form contained a substance known as IC, which gave
an orange color with aluminum chloride, but which unfortunately was
not further characterized. The substance was absent from all other
tea varieties tested but was present in two other species of the section
Thea, namely Camellia taliensis and C. irrawadiensis. The authors
considered that this evidence opened the possibility that some popu-
lations of cultivated tea were derived as species hybrids. While such
statements are conjectural at present, the work illustrates another
possible apphcation of biochemical data to systematics. It is noted
that otherwise the three species which comprise the section Thea are

Table 11-2. Distribution of biflavonyls (Baker and Ollis, 1961; from the
Chemistry of Natural Phenolic Compounds by permission of Pergamon Press).

Orders

Families, Genera, and Species

Sc. K. So.

Cycadales

Cycadaceae

Cycas revoluta Thunb.

Ginkgoales

Ginkgoaceae

Ginkgo biloba L.

Coniferales

Taxaceae

Taxus cuspidata Sieb. and Zucc.
T. cuspidata var. nana Hort.
T. floriana Chap.
Torreya nucifera Sieb. and Zucc.

Cephalotaxaceae

Cephalotaxus drupacea Sieb. and Zucc.
C. nana Nakai

Podocarpaceae

Podocarpus macrophylla D. Don

P. chinensis Sweet

P. nagi Zoll. and Moritz

Pinaceae
Abies firma Sieb. and Zucc.
A. homolepsis Sieb. and Zucc.
A. mariessii Mast.
A. veitchii Lindley
A. sachalinensis var. Schmidtii

Tatewaki
Keteleeria davidiana Beissner
Pseudotsuga japonica Carriere
Tsuga sieboldii Carriere
T. diversifolia Mast.
Picea polita Carriere
P. glehnii Mast.
P. maximowiczii Regel
P. koyamai Shirasawa
P. bicolor Mayer

P. jezoensis var. hondoensis Rehder
Pseudolarix kaempferi Gordon.
Larix kaempferi Sargent
Cedrus deodar a Loud.
Pinus densiflora Sieb. and Zucc.
P. koraiensis

(and 22 other spp. of Pinus

not named)

Sciadopityaceae

Sciadopitys verticillata Sieb. and Zucc.

Taxodiaceae

Taxodium distichum Rich.
Sequoia sempervirens E.

219

220

BIOCHEMICAL SYSTEMATICS

Table 11-2. {Continued)

a;
w

en
O

I— I

Orders

Coniferales

Gnetales

Casuarinales

Families, Genera, and Species

Metasequoia glyptostroboides Hu and

Cheng
Glyptostrobus pensilis K. Koch
Cunninghamia lanceolata Hooker
C. lanceolata var. konishii Fujita
Taiwania cryptomerioides Hayata
Cryptomeria japonica D. Don
C. japonica var. araucarioides Hort.

Cupressaceae

Callitris glauca R. Brown

Thujopsis dolobrata Sieb. and Zucc.

Thuja standishii C.

T. occidentalis L.

Biota orientalis Endl.

Libocedrus formosana Frolin

L. decurrens Torrey

Cupressus funebris Endl.

C arizonica

Chamaecyparis obtusa Endl.

C. obtusa var. breviana Mast.

C. pisifera Mast.

C pisifera var. filifera Mast.

C. pisifera var. squarrosa Mast.

Sabina chinensis Antoine

S. virginiana Antoine

S. procumbens Sieb. and Zucc.

S. sargentii Nakai

S. sargentii var. kaizuka Hort.

Juniperis utilis Koidz.

J. conferta Pari.

Sc. K. So. H.

Ephedraceae

Ephedra gerardiana Wall.

Casuarinaceae

Casuarina stricta Ait.

G. = ginkgetin; I. = isoginkgetin; Sc. = sciadopitvsin; K. = kayaflavone; So. = sotetsuflavone; H. = hinoki-

flavone.

chemically similar, although C. taliensis more closely parallels
C. sinensis. A number of species from other sections of the genus were
examined chromatographically, but their patterns did not closely
resemble the Thea pattern.

PHENOLIC SUBSTANCES 221

In summary, phenolics may be regarded as potentially of great
systematic importance because of the existence of hundreds of different
types, many of which are of restricted distribution. It is probable that
a comprehensive review of the chemical, biochemical, and pharmaco-
logical literature would establish a number of interesting correlations
not already recognized. In most of these cases, additional work would
be necessary to substantiate a systematic evaluation. Hegnauer (1956)
has reviewed the comparative chemistry of an individual family, the
Leguminosae and considered among other groups of compounds,
the phenolics. Comprehensive chemical reviews such as that of
Karrer (1958), which lists the constitution and occurrence of or-
ganic plant constituents, provide insight into attractive possibihties,
for example, the distribution of the flavanone, naringenin and other
flavanones and their glycosides in the genus Acacia and other mem-
bers of the Mimosoideae. Among the commoner phenolics, specific
glycoside types are likely to be more significant than the aglycone
which have been more often studied.

Some of the more recently discovered flavonoids, such as the
biflavonyls discussed earlier, offer opportunities for phylogenetic in-
vestigations. The biflavonyls, for example, are known to occur only
in gymnosperms with the exception of Casuarina (Table 11-2). It is
notable that biflavonyls are not yet known from Pinaceae, although
numerous species have been examined. Baker and Ollis (1961),
in noting the presence of biflavonyls in Casuarina stricta, add
"This is particularly interesting because of all the angiosperms,
Casuarina is the most closely related to the gymnosperms." That this
viewpoint is far from unanimous may be quickly ascertained from
Lawrence (1951). The more important question is that of how much
weight ought to be given to the presence of hinokiflavone, the bi-
flavonyl of Casuarina, in linking the group to gymnosperms.

In contrast to the biflavonyls, another group of recently dis-
covered flavonoids, the C-glycosides, thus far have been reported from
such widely separated plant groups as the Gramineae, Lemnaceae,
Caryophyllaceae, Rosaceae, and Verbenaceae (Horhammer and
Wagner, 1961).

QUINONES

Three major classes of naturally occurring quinones
are recognized: benzoquinones, naphthoquinones,
and anthraquinones (see below). In addition a few
complex substances of quinone structure occur (for
example, tripterine, in Tripterygium wilfordii, family
Celastraceae), but these last are too little known to
allow much consideration of their systematic im-
portance at this time.

The most recent treatment of the chemistry
and distribution of quinones is that of Thomson
(1957). He emphasized mainly quinone chemistry,
but sources of all naturally occurring quinones were
given.

Quinones occur in plants, animals, and
micro-organisms. However, in the animal kingdom,

223

224 BIOCHEMICAL SYSTEMATICS

O O

benzoquinone naphthoquinone

quinones are known to occur only in certain echinoderms and insects.
They are rare in algae but common in fungi. In the vascular plants,
with the exception of certain quinones which are believed to function
as important coenzymes, quinones are of restricted occurrence.
Although not widespread among plant families, in those families in
which quinones do occur, the compounds may be characteristic for
the family.

The chemical properties of quinones include relative ease of
oxidation and reduction. In this connection 2-methyl,3-phytyl,
1-4-naphthoquinone (Vitamin K), found in high concentration in
chloroplasts, has been proposed as a coenzyme involved in electron
transport, particularly in the processes following the primary photo-
chemical event of photosynthesis.

Recently, another group of coenzymes of a quinoid nature
referred to collectively as coenzyme Q (or ubiquinone) has been
described. This group is thought to participate in electron transport
between cytochromes b and c. The basic structure of coenzyme
Q involves a 2,3,5-tri-methyl benzoquinone substituted with isoprenoid
side chains at the 6-position.

CH3 \

CH2CH=C — CH2/nH

At least five naturally occurring homologues of coenzyme Q
have been described, the differences involving the number of isoprene
units attached. Lester and Crane (1959) studied the distribution
of the coenzyme Q series in animals, plants, and microorganisms.
Coenzyme Q was found in all higher plants examined (six genera) and
among red, brown, and green algae. One bluegreen alga, Anacystis
nidulans, did not yield any coenzyme Q. The higher animals and
plants were found to contain, usually, coenzyme Qio, with Q9 appear-
ing in a few cases. (The subscript refers to the number of isoprene
units in the side chain.) Among microorganisms, there was considera-
bly greater variation in the types present; for example, in ascomycetes
coenzymes Qio, Q9, Qs, Q7, and Qs were present. A particular quinone

QUINONES 225

found in chloroplasts having an absorption maximum at 254 m/x has
been called "plastoquinone" (Crane, 1959).

The presence of a quinone coenzyme involved in an important
electron transport system in plants suggests that the ability to syn-
thesize the basic naphthoquinone nucleus is not limited but is charac-
teristic of plants in general. Therefore, those groups of plants which
accumulate naphthoquinones otherwise substituted than in the
vitamin K pattern may not possess a uniquely new enzyme system
for the formation of the naphthoquinone ring structure, but rather
may possess a metabolic system which permits the accumulation of
naphthoquinones, which, when coupled with appropriate enzymes,
provide for secondary structural modifications. It is well established
that some quinones are fungicidal. If then, there is some positive
selective value correlated with quinone accumulation, and the basic
quinone pathway pre-exists (even though production is hmited) among
green plants in general, it is not surprising to find distantly related
plants producing the same compound. Considered in this light there
is no reason to suspect cryptic phylogenetic association between taxa
possessing such compounds. For example, the quinones lawsone, and
its methyl ether (the latter is fungicidal) are found in Lawsonia alba
(Lythraceae) and in Impatiens balsamina (Balsaminaceae) respec-
tively. Thomson says that "it is noteworthy that such closely related
quinones occur in distantly related plant families." Actually, other,
even more complex quinones occur in equally distantly related families,
for example, lapachol (Bignoniaceae, Verbenaceae, Sapotaceae).

"oh ^ X /OCH3 ^^^A.OH

CH3
CH2CH=CCH3

lawsone lawsone methyl ether lapachol

Some quinones are physiologically active (as purgatives), and
others are valued as dyes. Despite their economic significance, how-
ever, relatively little is known of quinone biosynthesis, and practically
no genetic studies on quinones have been reported. The favored
hypothesis to account for the important anthraquinone group involves
the same mechanism as that producing the A ring of flavonoid com-
pounds, namely, the condensation of acetate units. Acetate-2-Ci4 has
been used to investigate the biosynthesis of emodin by Penicillium
islandicum, and the results suggest that head to tail condensation of
eight acetate groups was involved (Friedrich, 1959). Hegnauer (1959),
in contrast, emphasized the fact that compounds such as xanthones,

226 BIOCHEMICAL SYSTEMATICS

stilbenes, chalkones, asperulosides, and so on, which may be regarded
as variants of the basic Ce-Ca-Ce flavonoid nucleus, occur in the plant
families which are also notable for the production of anthraquinones.
Furthermore, he notes that Trim (1955) found that asperuligenin
accumulated in Rubiaceae during development, but only until the
synthesis of anthraquinone began. Thus, Hegnauer believes that com-
parative phytochemistry points to a relationship between the Ce-C.s
and Ce-Cs-Ce groups on the one hand and anthraquinones on the
other, so that the acetate theory alone could not satisfactorily
account for the facts; possibly phenol-related pathways are involved.
The simplest group of quinones, benzoquinones, are rarely
found among higher plants, being better known among fungi. Although
no attempt is made in this section to give a comprehensive list of the
quinones and their sources, there are so few benzoquinones from
higher plants, that it is practical to list them all. The following
benzoquinones from higher plants are included in Thomson (1957):

H3CO

2 : 6-dimethoxybenzoquinone

Adonis vernalis, Ranunculaceae

HO-Y^%-CnH

embelin

Myrsine, Embelia and Rapanea, Myrsinaceae

O
HO-

Rapanea maximowiczii; Myrsinaceae; Oxalis purpurata war. Jacquinii,
Oxalidaceae

QUINONES 227

HO^J-v^(CH2)i3CH=CH(CH2)3CH3

H3C J OH

maesaquinone

Maesa javonica, Myrsinaceae

O 9^3 p,TT

H0\^-V/CHCH2CHoCH=C

HaC^
O

perezone

Perezia adnata, Trixis calcalioides, tribe Mutisieae of the Compositae

The Myrsinaceae seem to be particularly rich in benzo-
quinones, and these compounds should prove to be useful as system-
atic criteria. It is significant that the plants which produce benzo-
quinones bear no particular taxonomic affinities to those species pro-
ducing the other classes of quinones.

Since these compounds are, for the most part, relatively
simple derivatives of naphthoquinone, it is probable that they have
arisen independently in many, if not all, of the families known to
produce them.

Extensive surveys for the presence of naphthoquinones have
not been made, and many naphthoquinone-containing species may re-
main undetected. Naphthoquinones, by present knowledge, are rather
rare, and a given type usually is restricted to one or two famihes. The
simple naphthoquinone, juglone, approaches a familial character in
the Juglandaceae. Although naphthoquinones seem to be of fittle
systematic significance above the family level, it is possible that at
the lower taxonomic levels the compounds may be of systematic value.

The most complex group of quinones, the anthraquinones, is
also the most widely distributed. In fact, if the three sub-types of
quinones were selected to illustrate the principle of the "percentage
of frequence" rule (see Fig. 4-1) the results would contradict the
principle, since the least complex have the most fimited taxonomic dis-
tribution. Coupled with the fact that there is little simultaneous occur-
rence of two or three sub-types of quinones, their general pattern of
distribution implies that there is no close biosynthetic relationship
between the types of quinones, and therefore this chemical class, in a
biosynthetic sense, appears to be artificial. (In contrast, in the fungal

228 BIOCHEMICAL SYSTEMATICS

genera Penicillium and Aspergillus, two and three groups, respec-
tively, of quinones are encountered, and within a single group, a num-
ber of different quinones occur.)

The association of anthraquinones with particular families of
higher plants is striking. The Rubiaceae, Polygonaceae, and Rhamna-
ceae are notable in this respect, with the family Rubiaceae the out-
standing example (Hegnauer 1959). Anthraquinones are rare among
monocots, having been reported only in the Liliaceae. Schnarf (1944)
investigated the presence of aloin in tribes of the Liliaceae. In the
tribe Asphodeleae, he found aloin in specialized cells in the genera
Asphodelus, Evenurus, Bulbine, Bulbinella, Bulbinopsis, and Alec-
torurus. "Aloin cells" are otherwise found only in the tribe Aloineae
(except for the presence of chrysophanol in Xanthorrhoea of a third
tribe, Lomandreae). Moreover, the above-named genera differed in
foliar anatomy and embryology from others of the Asphodeleae but
resembled the Aloineae. Thus, according to Hegnauer, the biochemical
evidence correlates nicely with embryological and anatomical evidence.

In a previous study, Munkner (1928) investigated extensively
the tribe Aloineae and particularly the genus Aloe. The older tech-
nique for the detection of anthraquinones was a color test, the
Borntrager test. (A slightly acidified benzene extract is shaken in a
test tube with ammonia. A rose red to raspberry color indicates the
presence of anthraquinone.) Since it now appears that negative tests
with the Borntrager reagent are not always reliable (Hegnauer, 1959),
some conclusions based on the presence or absence of anthraquinones
by this test might be proven spurious. Of genera related to Aloe, the
following results were obtained:

Gasteria (seven species tested; all positive)

Lomatophyllum (two species tested; both positive)

Apicra (four species tested; all negative)

Kniphofia (ten species tested; all negative)

Haworthia (seven species tested; two positive and five

negative)

One hundred and seventy eight species oi Aloe were examined,
and a large majority of the species gave a positive Borntrager re-
action. However, there is little indication of a definite pattern of the
distribution. For example, although nineteen species of the section
Leptoaloe were negative, there were two questionable exceptions
{A. kraussii and A. parvula); six of the seven remaining sections had
both positive and negative species, as did all but one of the five sub-
sections of section Eualoe.

In the family Polygonaceae, Jaretzky (1926) reported that of
the two sub-families, Eriogonoideae and Polygonoideae, only the latter
produced anthraquinones. Many species of the genera Ernex, Rumex,

QUINONES 229

Rheum, as well as species of Atraphaxis, Oxygonum, Polygonum, and
Muhlenbeckia produce anthraquinones. More recently, Tsukida
(1957) reported on the distribution of anthraquinones in the Polygon-
aceae and added several other genera to those known to produce these
compounds. Jaretzky (1926, 1928) believed the presence of anthra-
quinones to be a primitive character since it was inversely correlated
with morphological progression within the genus Rumex (for example,
dioecious species such as Rumex acetosa are anthraquinone free) as
well as within the sub-family as a whole (for example, Fagopyrum
with heterostyly, is anthraquinone free). This is one of the few cases
where the presence of a particular class of chemical substances is
believed to be a primitive character. Tsukida was principally con-
cerned with the localization of specific anthraquinones in plant organs
as well as the specific anthraquinone types produced by these species,
and he did not emphasize particularly the systematic imphcations of
his data.

Heppeler (1928) studied the distribution of emodin in the
genus Rhamnus (Rhamnaceae) and attempted a systematic arrange-
ment of the genus based on the presence or absence of this anthra-
quinone in dried plants. However, Maurin (1928) in the same year
reported a number of species positive which had been considered
negative by Heppeler. Furthermore, Hegnauer (1959) has noted that
the application of the Borntrager test to herbarium material is
unreliable since a number of species judged by Heppeler to be negative
have since been shown to contain anthraquinone. Hegnauer has
summarized present knowledge of the occurrence of anthraquinones
in the genus. Unfortunately (for purposes of phylogenetic implications),
a number of the series in both the sub-genera, Frangula and
Eurhamnus, which formerly had been considered negative, are now
shown to be positive for anthraquinones. It appears that anthra-
quinones are widespread in the genus Rhamnus and also in a number
of other genera of the Rhamnaceae. Since no distributional pattern is
now recognizable, further investigation on this family is recommended
by Hegnauer.

In the Leguminosae only a few genera are known to produce
anthraquinones, but in one genus, Cassia, the compounds are wide-
spread. The classic anthraquinone work on Cassia is that of Gilg and
Heinemann (1926). These authors assumed that oxymethylanthra-
quinones were to be found only in the section Chamaesenna of the
sub-genus Senna.

Within Chamaesenna the various series were analyzed for
presence of emodin-like anthraquinones, yielding some interesting
results. Probably the most noteworthy systematic conclusion stem-
ming from this survey was the redisposition of the taxa belonging
to the series Aphyllae, which includes only the two species,

230

BIOCHEMICAL SYSTEMATICS

Series

Species Examined

Results

Pachycarpae

( + ) Emodin present

Crassirameae

( — ) Emodin absent

Rostra tae

( — ) Emodin absent

Auriculatae

( — ) Emodin absent

Floridae

( - ) Emodin absent

Aculeatae

( — ) Emodin absent

Pictae

( + ) Emodin present

Brachycarpae

( + ) Emodin present

C. aphylla and C. crossiramea. According to Gilg and Heinemann
these two species are placed in Bentham's series Aphylla on a super-
ficial character (namely, absence of leaves). Gilg and Heinemann
treated C. aphylla as a leafless member of the series Pachycarpae,
while C. crassiamea was placed in a newly proposed series, Crassiraea.
Although these authors based their conclusions, in part, on certain
morphological and geographical evidence, considerable weight was
apparently given to the fact that C. aphylla tested positively for
emodin (as did the fourteen other species tested in the series Pachy-
carpae) while C. crassiramea was negative. A similar observation was
perhaps also responsible for the author's establishment of the series
Aculeatae, its only species, C. aculeatae, which was negative for
emodin, having previously been placed in the series Pictae (Bentham,
1871); the latter testing positive for those nine species examined.

However, Hegnauer (1959) has summarized more recent
literature on the anthraquinones of Cassia, and has noted reports
of a much broader distribution of anthraquinone in the genus. For
example, several species of the sub-genus Fistula have been reported
to contain anthraquinones: C. fistula, C. leptophylla, C. carnaval, and
C javanica. The majority of sections in the sub-genus Senna now
are known to have at least one representative which produces anthra-
quinones, and within the section Chamaesenna, two series other than
those noted by Gilg and Heinemann are included among anthra-
quinone producers. Finally, C. mimosoides, of the sub-genus Lasio-
rhegma (section Chamaecrista) has been found to produce anthra-
quinones. The situation in Cassia is, then, similar to that in Rhamnus.
It is quite hkely that an intensive study of the distribution of
quinones in the large genus Cassia would disclose a pattern. Such a
study would have to include a characterization of the more common
quinones as well as analysis of various plant organs, for it has been
demonstrated that related species may differ radically in the distribu-
tion of quinones within the plant.

TERPENOIDS

A rather heterogeneous group of substances is
actually included under terpenoids, yet with few
exceptions the compounds may be conceived as
structural derivatives of the five-carbon compound,

isoprene.

CH3

CH2=C-CH=CH2

Isoprene

Recently, the six-carbon compound, meval-
onic acid, has been found to be an important
precursor in cholesterol synthesis and is suspected
to be involved also in the synthesis of several other
classes of isoprenoid compounds such as terpenes^

231

232 BIOCHEMICAL SYSTEMATICS

and carotenoids (Wagner and Folkers, 1961). Mevalonic acid itself
apparently originates through acetate condensation. As was indicated
in a previous section (Chapter 11) mevalonic acid occupies a focal
position in isoprenoid synthesis somewhat analogous to that of
shikimic acid in aromatic synthesis (Fig. 13-1).

CH3

HOCH2CH2C— CH2COOH

mevalonic acid

Among the simplest terpenes, the relationship to isoprene is
evident at once. For example, note the monoterpene, myrcene, illus-
trated below:

CH2

H2C CH

H2C CH2
CH

H3C CH3

myrcene

The formula above, also of myrcene, represents the type
customarily used to represent terpenoid compounds.

In addition to differences in the position of double bonds and
degree of hydrogenation of a given basic terpene structure, alcohol,
aldehyde, ketone, and acid derivatives of simple aliphatic terpenes
exist. Geraniol, for example, is a widely distributed alcohol of this
type. Furthermore, ring closure provides for simple cyclic structures.
A common example of such a compound is phellandrene:

/^-phellandrene

1 Although Stanley (1958) has reported incorporation of C^* labeled mevalonic
acid into a-pinene of Pinus attenuata, Battaile and Loomis (1961) have evidence that
mevalonic acid is not incorporated into mint terpenes. These latter investigators found
C^^ from mevalonic acid in carotenoids and other compounds in the plant. Therefore, one
cannot readily discount their evidence concerning terpenes on the grounds that it is negative.

TERPENOIDS 233

It is evident that a very large number of simple terpene types are
theoretically possible (and, in fact, exist). More complex terpenoid
compounds, and other types of isoprenoid derivatives, also exist
in abundance throughout the plant kingdom and to a more limited
extent among animals. Even in simple monoterpenes internal rings may
form with the elimination of a double bond, for example, as in pinene.

«-pinene

Also, additional isoprene units may be incorporated. Ses-
quiterpenes, for instance, represent three isoprene units (C15 com-
pounds), and diterpenes represent four isoprene units. The latter are
relatively uncommon. Ti'iterpenes, with six isoprene units, are but
rarely encountered in higher plants. Plant steroids are best considered
allied with the terpenoid substances. Isoprene derivatives, or com-
pounds that may be derived theoretically from isoprene, of even
higher molecular weight include such compounds as the carotenoids,
and high polymers such as rubber or gutta percha. Although the
carotenoids may prove to be a valuable biochemical category for
chemosystematic purposes, as yet Httle work along such hues has
been done with such compounds except among the algae (Chapter
14). The phytyl group of chlorophyll is essentially a poly isoprene, and
the group also occurs as part of the napthoquinone derivative,
Vitamin K, discussed in Chapter 12. Such substances as these last are
examples of important basic metabolic pathways and are therefore
probably less useful in phytosystematic investigations.

A classic example of the apphcation of phytochemistry to
problems of phylogeny is the work of Baker and Smith (1920) on the
terpenes of Eucalyptus oils. One might suppose that the impetus from
this classic work would have encouraged considerable interest in the
systematic distribution of terpenes, yet this has not occurred on a large
scale. The Baker and Smith work was a remarkable achievement, but
on reading the book it is nontheless evident that the immediate
systematic imphcations of the study were quite limited. There is no
doubt however that the work disclosed clearly the possibihties of
phytochemical systematics. Some of the important aspects of the
Baker and Smith work will be considered at this time.

One important goal of their early investigations was to deter-
mine whether or not chemical characters are dependable (or constant)
enough to warrant their consideration as taxonomic characters. The
examination of large numbers of individual trees over an extensive

234

BIOCHEMICAL SYSTEMATICS

>»

0)>

o— o

C) >sS

S ^5 ^^

O k5

/ P 5

■5 "

>»

o=u

- <

5 *

) E

2 Jr

O— C

:> 8

H >•

^ 2

o=c

; S

ro O

- 8

; S

CO
1

ffi

3 1

C N

1 c

o— c

J 8

C' -s

J S

o OS

o
o

TERPENOIDS

235

CO
CD

O— O,

X
u

o X

c
a

n,
o

^ °-

9. J

T 2 s

>.x

o f-"'

I >*

I u>
M 4)

ffi £
Q 5

m
o

u— o.

IS «
|i

X
p

X ^'

\ CO

X
o

236 BIOCHEMICAL SYSTEMATICS

part of the range of the species was reqmred to answer this question.
As a result of such investigations Baker and Smith were impressed
with the constancy of oil characters within a species:

The theory has often been advanced that the chemical constituents of
the same species vary in different localities, but this idea is not verified
by our experiences as regards the Eucalypts, as they do not show these
differences in chemical constituents that might perhaps be expected
from differences of soils or localities. The reverse may possibly be
accounted for by the natural selective, ecological peculiarities shown in
many instances by the species themselves, as it is remarkable how a
certain species will flourish on a particular geological formation and
become singular to Hke formations, while at the same time objecting
to those entirely different. However that may be, those influences do
not appear to act detrimentally, or to interfere in any way with the
practical constancy of results.

It is apparent now that more variation in oil character existed
than was recognized by Baker and Smith, especially among the in-
dividuals of a population. For example, in discussing the oil characters
of Eucalyptus dives they noted that the constancy of oil characters
exhibited by Eucalyptus species generally also apphed to E. dives.
Components of E. dives oil are crude oil, largely phellandrene with 5
to 8 per cent cineole and some piperitone. Another fraction consisted
largely of the peppermint ketone, piperitone. But, Penfold and
Morrison (1927) also described major variations in the oil character
of E. dives. Ordinarily this species yields oils with 45 to 50 per cent
piperitone. Yet, some plants identical to typical E. dives yielded as
low as 8 per cent piperitone. In fact, there had been some complaints
that differences in the piperitone concentration were the fault of the
distillation techniques if not the result of adulteration. The following
picturesque statements from Penfold and Morrison indicate that
significantly different oil characters did indeed occur in two plants
which were morphologically indistinguishable:

Then again whilst engaged in field service during the end of December,
1924, repairing a punctured tyre of the car by which we were travelling
led us to examine a patch of trees of this species growing close to the
Main Southern Road about 18 miles on the Sydney side of Goulburn.
The observation was made of two trees growing together, indistin-
guishable from one another by both botanist and bushman, but each
containing a different essential oil. On crushing the leaves between the
fingers, one yielded the typical phellandrene-piperitone odour, whilst
in the other the odour of cineol-phellandrene-terpinol was most
pronounced.

TERPENOIDS

237

A communication from another worker concerning the mor-
phology of these plants is quoted by Penfield and Morrison:

I tried every point to see if there is any morphological difference be-
tween these two forms, but failed to find one single character to distin-
guish these two trees. . . . Seedlings, young and matured foliage, buds
and fruits, all agree with the other. I spent many hours over this
examination that I might not miss any point."

According to Penfield and Morrison, E. dives is thought to be
a hybrid. Thus they feel that such oil variation "seems only reason-
able." Yet, why is there so httle evidence in a hybrid of morphological
variations? The fact that so much effort and attention was devoted
to the question of whether the chemical forms of E. dives could
be otherwise distinguished is indicative of the influence that the earlier
phytochemical work had with respect to the question of chemical
constancy within a species.

Baker and Smith did not apparently consider hybridization
to be an important factor in the evolution of Eucalyptus.

It may be now shown that most of these supposed aberrant forms are
really distinct species, and in our opinion cross fertilization in the
Eucalypts under natural conditions is quite exceptional, especially
when we know that numerous species are growing intermixed, often
flowering at the same time, and so under supposed favourable condi-
tions for hybridization, yet preserving throughout extensive areas their
specific characters with remarkable constancy.

Four types of E. dives were subsequently recognized, all based
on oil character differences:

E. dives, type

piperitone, 40 to 50 per cent; phellandrene, 40 per cent
E. dives, var. A

piperitone, 5 to 15 per cent; phellandrene, 60 to 80 per

cent; piperitol (small amount).
E. dives, var. B.

piperitone, 10 to 20 per cent; cineole, 25 to 50 per cent;

together with phellandrene
E. dives, var. C.

Cineole, 45 to 75 per cent; piperitone, under 5 per cent;

phellandrene, absent, or present in small quantity only.

Although Baker and Smith considered that oil characters did
not usually vary greatly within a species, they found examples

238 BIOCHEMICAL SYSTEMATICS

wherein oil characters differed among morphologically similar plants.
A notable example is the species Eucalyptus phellandra discussed
prominently by Read (1944). According to Read, E. phellandra had
been included previously under E. amygdalina, and in the first
edition of the Baker and Smith work (1902) it was recorded under
amygdalina. According to Baker and Smith, "It is one of the few
species of this research that has been founded almost entirely on the
chemical constituents of the oil." There is then some circular reason-
ing if one creates a species on the basis of a difference in oil character
alone, while simultaneously maintaining that constancy of oil character
within a species is typical within species of the genus. As a generali-
zation the species constancy of oil character in Eucalyptus is doubt-
lessly accurate. Physiological races are regularly encountered, and
their appearance does not normally affect the integrity of the species.
In Pinus, Mirov (1961) has found that some species vary but little in
turpentine composition throughout their range while other species
are quite variable in this respect.

In connection with problems of phylogeny within the genus
and among other related genera. Baker and Smith noted the similari-
ties in oil constituents of Eucalyptus and Angophora (for example,
the presence of the sesquiterpene, aromadendrene, in both genera) as
opposed to a third allied genus, Tristania. They proceeded to develop
a postulated line of descent showing the supposed origin of each sub-
group, the pattern stemming from correlated chemical and morpho-
logical characters. Baker and Smith recognized four major sub-divisions
of the genus with distinctive chemical attributes:

(1) Those yielding oils consisting largely of the terpene pinene,
either dextro-rotatory or laevo-rotatory.

(2) Those yielding oils containing varying amounts of pinene
and cineole, but in which phellandrene is absent.

(3) Those yielding oils in which aromadendral is a character-
istic constituent and phellandrene is usually absent.

(4) Those yielding oils in which phellandrene is a pronounced
constituent with piperitone mostly present.

Since the majority of eucalypts yield oil largely of pinene
and cineole without phellandrene, the authors believed that phellan-
drene and thus piperitone appeared later, in fact even later than
aromadendral.

An interesting correlative morphological character is found
among the Eucalypts. The character involves the pattern of leaf
venation which seems generally to be correlated with the oil constitu-

TERPENOIDS 239

ents. The "primitive" {sejisii Baker and Smith) leaf venation pattern,
associated with cineole and pinine oils, exhibits the following featm-es:

(1) Angle with midrib is less acute (approaching a right
angle).

(2) Marginal vein is close to edge.

(3) Reticulations between veins are prominent.

The "advanced" leaf types, associated with phellandrene and
piperidine, possess the following attributes:

(1) Angles of veins with midrib acute.

(2) Marginal vein withdrawn from edge, a second marginal
vein withdrawn from edge, and a third marginal vein may
be in evidence (for example, E. dives).

(3) Reticulations between major veins are reduced, and thus
more space for oil glands is present.

(4) Looping arrangement of major veins particularly notice-
able in the bending of the marginal vein at positions of
major lateral veins.

Elsewhere Baker and Smith stated:

In other parts of this work we show that this alteration in leaf vena-
tion and chemical constituents is not local in its incidence, and that
the specific characters of each species are practically constant over the
whole range of its distribution, and numerous instances are given of
this constancy.

That the constituents of the oil have been fixed and constant
for a long period of time must be evident by the fact that, to whatever
extent or range any particular species has reached, it contains the
same characteristic constituents, and has its botanical characters in
agreement.

Baker and Smith prepared a phylogenetic tree designed to
show the evolutionary relationships of over 150 species of Eucalypts.
They also illustrated the general distribution of specific chemical
constituents of the oil. This distribution was purported to reflect the
major movement of the genus during its evolution in Austraha.

The foregoing discussion may have given the impression that
only a few oil constituents had been detected. In fact, even in 1920
Baker and Smith listed forty oil constituents, and at the present time
it is almost a certainty that many more are known. Many of the com-

240 BIOCHEMICAL SYSTEMATICS

ponents, however, occur in small quantities and often in only a few
species and these substances may provide further taxonomically useful
information.

Exceptional species such as Eucalyptus macarthuri and E.
citriodora, in which the chief constituents are geranyl acetate and
citronellal, were regarded as end members of sequences in which the
ancestral intermediate forms have disappeared in the course of evolu-
tion (Read, 1944).

McNair (1942) attempted to correlate the morphological and
chemical characteristics as reported by Baker and Smith and con-
cluded that sometimes "primitive" morphology and "advanced" oil
characters occurred together, or the opposite relationship occurred.
The extent to which this is borne out is difficult to determine from
the data in McNair's paper, since he presents no morphological data
to compare with the chemistry. Of course, some instances of more
rapid evolution in either morphology or oil chemistry are to be
expected. One point made by McNair which is noteworthy is that oil
constituents of the "advanced" type may appear independently in
groups which otherwise show no close genetic relationships.

As noted earlier, in spite of the classic work by Baker and
Smith, very little work on the biological aspects of terpene chemistry
has been carried out. This is noted by Mirov in 1948 emphatically:

. . . the chemistry of essential oils to the problems of biology has been
utterly neglected and very little organized work has been done in this
direction. A notable exception is, of course, the classical research on
the Eucalypts and their essential oils by Baker and Smith. . . .

Mirov (1948) reported on the terpenes of the genus Pinus. He
included extensive tables of data arranged according to species and
following the classification of Shaw (1914). In the sub-groups
Haploxylon (having a single vascular bundle in each needle with
usually five needles per dwarf shoot) and Diploxylon (having a double
bundle with two to three needles) there did not appear to be any
significant general differences in their terpenes. For example, both
groups contained dl-a-pinene as a major constituent, and other, more
complex substances occurred sporadically throughout both groups.^
However, Erdtman has shown distinctive differences in the heartwood
chemistry of the two groups. (For discussion of Erdtman's work see
Chapter 11.)

2 However, in a discussion following presentation of a paper on the distribution of
turpentine components (1958) Mirov stated that the Haploxylon group "have decidedly
more sesquiterpenes" and more new substances were found in that group. Mirov beUeves
that the two sub-groups split very etu-ly and underwent parallel evolution.

TERPENOIDS 241

Mirov cites several instances in which closely related species
have similar terpenes (for example, Pinus muricata, P. attenuata,
and P. radiata), and other cases in which closely related species have
quite different terpenes.

One extremely interesting situation involving pure species was
reported by Mirov (1948). Pinus ponderosa contains ^-pinene and
limonene (however, the variety scopulorum consists mostly of
a-pinene instead of /5-pinene). P. jejfreyi, which some botanists con-
sider a variety of P. ponderosa, contains no terpenes but rather hep-
tanes. To complicate the matter further, P. jeffreyi in its chemical
attributes approaches more closely the group Macrocarpa than the
group Australia to which P. ponderosa belongs (heptane is found in
all three species comprising the Macrocarpa group). Also, similar
aldehydes are found in P. jeffreyi and the pines of the Macrocarpa
group. Furthermore, P. jeffreyi crosses in nature with both P. ponder-
osa and P. coulteri, the latter a member of Macrocarpa. According to
Mirov, P. jeffreyi possibly crosses more readily with P. coulteri.

In the genus Mentha rather extensive chemical investigations
of the important flavoring substances have been conducted by nu-
merous investigators. Recently, genetic studies have advanced evidence
that a single pair of genes controls, directly or indirectly, the major
monoterpenic chemical constituents of mint oils (Murray, 1960a,
1960b). The action of the dominant gene apparently is upon a cyclic
intermediate to convert it to a spearmint (2-oxygenated-p-menthane),
while in the presence of the recessive only, the cyclic intermediate is
converted to the peppermint type (3-oxygenated-p-menthane)
(Reitsema, 1958a, 1958b). Except for the position of the oxygen a
corresponding series of compounds exists in both the peppermint and
spearmint lines. No authenticated instance of the coexistence of
spearmint and peppermint oils in a single plant exists. In general the
spearmint oils contain more unsaturated compounds and much more
saturated alcohols while odd side reaction products such as found in
some peppermint oils are lacking. A third group of species, the so-called
"lemon mints," do not produce cyclic derivatives but rather acychcs
such as citral and linalool. Reitsema has constructed a correlative
biochemical-phylogenetic sequence in which the progression is toward
increasingly more reduced compounds (Fig. 13-2).

Some very interesting work on higher terpenes of the
Cucurbitaceae has been reported by Enslin and Rehm (1960). These
substances, not fully characterized, appear to be related to the
tetracyclic triterpenes. They are bitter tasting, have a purgative
action, and are referred to as "cucurbitacins." So far, eleven different
cucurbitacins are known, ten of which have been crystallized and an
empirical formula assigned to them. All contain two or more

242

BIOCHEMICAL SYSTEMATICS

Linear Intei'mediate

i
Cyclic Intermediate

I I

Spearmint Types Peppermint Types

dihydrocarvone

M. spicata, etc.

pulegone

M. pulegium

menthone

Acyclic Types
C— C=C— C— C— C=C— C=0

C citral C

M. citrata

I diosphenolene

I M. rotundifolia

.OH

M. sylvestris

diosphenol

menthofuran

M. aquatica

menthol

M. piperita, etc.

Fig. 13-2. Hypothetical biochemical-phylogenetic sequence of
peppermint type oils (Reitsema in Jour. Amer. Pharm. Assoc, Sci.
Ed. 47: 268. 1958-by permission).

hydroxyl groups and several keto groups. They may be found as
glycosides or agly cones in various parts of the plant, and many species
contain an active glucosidase capable of hydrolysing the glycosides to
aglycones. The glycosidase is apparently of somewhat low specificity
since it is capable of hydrolysing steroidal saponins, the diterpene

TERPENOIDS

243

/8-D-glucoside, darutoside, and certain cardiac glycosides. Surprisingly,
in one species, Acanthosicyos horrida, cucurbitacins occur as glycosides
in the roots and as aglycones in the fruit. According to Meeuse (1954)
most, if not all, genera producing the cucurbitacins are in the sub-
family Cucurbitaceae (for example, Momordica, Bryonia, Ecballium,
Citrullus, Cucumis, Lagenaria, Cucurbita, and Sphaerosicyos).

Two cucurbitacins, designated B and E, are thought to be the
primary cucurbitacins, since other than cucurbitacin C these two are
the only ones which sometimes occur alone in mature plants, and
seedlings of all twenty-one species studied contained mainly B and/or
E, even in species containing up to eight different cucurbitacins. The
empirical formulas of B and E are given below:

Cucurbitacin B C32H48O8
Cucurbitacin E C32H44O8

Apparently the cucurbitacin content within a species may
vary greatly since several genera {Citrullus, Cucumis, and Lagenaria)
occur in bitter and non-bitter forms. In the case of Cucurbita pepo
var. ovifera from one to eight different cucurbitacins may occur,
though certain combinations are favored. Enslin and Rehm found
that genetic, environmental, and developmental factors influence the
cucurbitacin content.

The value of the cucurbitacin studies is further enhanced by

Table 13-1. Relative amounts of constituents in peppermint type oils" (Reitsema
in Jour. Amer. Pharm. Assoc, Sci. Ed. 47: 268. 1958— by permission).

arvensis

var.

rutundi-

piper-

folia

sylvestris

pulegium

aquatica

piperita

ascens

Piperitenone

Piperitenone oxide

50%

Diosphenolene

Piperitone

Piperitone Oxide

45%

Diosphenol

. .

Pulegone

80%

Menthofuran

40%

2-15%

Menthone

.v"

25%

Menthol

50%

80%

" Absence of quantitative data indicates lack of data rather than an implied absence of the compound in the oil.
' Identified by chromatography and ultraviolet absorption.
' Indicates presence without quantitative data.

244 BIOCHEMICAL SYSTEMATICS

investigations pertaining to enzymatic interconversion. In the fruit
juice of Lagenaria siceraria an enzyme catalyzes efficiently the con-
version of E to B while the reverse reaction occurs more slowly.
The authors conclude that A is formed from B, and C from B.
Enzymes occur which convert E to I, B to D, and C to F. All of these
conversions involve loss of a two-carbon group. Surprisingly, the
highest activity for this type of conversion is found in the fruit juice
of a non-bitter Golden Hubbard squash. An alternate pathway to
cucurbitacin D, from F, occurs apparently in leaves and fruits of
Cucumis angolensis.

Emslin and Rehm summarize their evaluation of the taxo-
nomic significance of the cucurbitacins in a brief paragraph, as
follows:

The main conclusion emerging from this study of the biogenetic inter-
conversions is that there are only two primary bitter principles, which
are chemically very labile, and easily transformed to other related sub-
stances by enzyme systems present both in bitter and non-bitter
plants. It is therefore not surprising that a knowledge of the bitter
principle composition of species appears to be of little value to the
taxonomist.

The authors then go on to note distinctions between Cucumis
and Citrullus as follows. All species of Cucumus investigated con-
tained cucurbitacin B while Citrullus species contained only E in their
seedling roots. In Cucumis the cucurbitacins occur mainly as agly-
cones while in Citrullus they occur as glycosides. Possibly the authors
are unduly pessimistic regarding the systematic significance of these
substances. Since they state that an effective paper chromatographic
method is available for their study, it is likely that intensive studies
of populations, particularly natural populations, would prove useful.
It is not likely that cultivated varieties would offer as much promise,
considering the labihty of the group, as would wild species.

Another phytosystematic investigation of higher terpenes is
that of Hollo way (1958) who studied the diterpenes of the phyllo-
cladene and podocarprene types. Among the former group several
diterpenes, phyllocladene, rimuene, mirene, and kaurene are closely
related and possibly isomers. The podocarprene group is chemically
similar, and representatives of both types occur together in at least one
genus, Sciadopitys. With the exception of Sciadopitys these diter-
penes are confined to the tribes Araucarineae or Podocarpineae.
Diterpenes of other types occur in other conifers, for example, in
Pinus. Since, in older classifications the order Coniferales was divided
into two families, the Pinaceae (including Araucaria and Sciadopitys)

TERPENOIDS

245

and the Taxaceae (including Podocarpus), Holloway considered that
the distribution of the diterpenes was opposed to this older taxonomic
disposition. Sciadopitys, by more recent treatments, is placed in the
Taxaceae, but the genus may be somewhat closer to Araucaria,
according to Holloway, if one uses certain criteria related to embryo
development, gametophyte structure, and fertilization. He does not
deny, however, the similarities between Sciadopitys and other genera
in the Taxodineae, but he still considers it possible that the genus
diverged from the main Araucarian stock at an early time. It would
be interesting to know the total distribution of these diterpenes.

Holloway constructed a diagram to illustrate the relation of
the occurrence of the phyllocladene and podocarprene diterpenes and
conifer phylogeny (Fig. 13-3). All genera to the right of the Taxineae
either have these diterpenes or are postulated to have them.

As noted in Chapter 14 the genus Podocarpus is biochemically
distinctive in that its lignin contains some syringyl derivatives typi-
cally absent from the lignins of other conifers.

A small group o