Dr. Susumu Ohno (Auth.) - Evolution by Gene Duplication-Springer-Verlag Berlin Heidelberg (1970)

Susumu Ohno
Evolution by
Gene Duplication
With 28 Figures
Springer Science+Business Media, LLC 1970

Dr. SusuMu HNO
Department of Biology
City of Hope Medical Center
Duarte, Calif.JUSA
@Springer Science+Business Media New York 1970

Originally published by Springer-Verlag New York Inc. in 1970
Softcover reprint of tbe bardeover Ist edition 1970
ISBN 978-3-642-86661-6 ISBN 978-3-642-86659-3 (eBook)

DOI 10.1007/978-3-642-86659-3
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concemed,
specifically those of translation, reprinting, re-use of illustrations, broadcasting, reproduction by photocopying
machine or similar means, and Storage in data banks.
Under 54 of the German Copyright Law were copies are made for other than private use, a fee is payable to
the publisher, the amount of the fee to be determined by agreement with the publisher.
The use of general descriptive names, trade names, trade marks etc. in this publication, even if the former are not
especially identified, is not to be taken as a sign that such names, as understood by the Trade Marks and Merchan-
dise Marks Act, may accordingly be used freely by anyone.
Library of Congress Catalog Card Number 78-112882.

Typesetting and printing: Carl Ritter & Co., Wiesbaden. Bockbinding: Kar! Hanke, Dsseldorf.
Title No. 1677

This book is dedicated to my deceased father, Kenichi Ohno,
and to my father-in-law, Keiji Aoyama, also deceased
Preface
It is said that "necessity is the mother of invention". To be sure, wheels and

pulleys were invented out of necessity by the tenacious minds of upright citi-
zens. Looking at the history of mankind, however, one has to add that "Ieisure
is the mother of cultural improvement". Man's creative genius flourished only when
his mind, freed from the worry of daily toils, was permitted to entertain
apparently useless thoughts.
In the same manner, one might say with regard to evolution that "natural
selection mere(y tnodifted, while redundanry created". Natural selection has been
extremely effective in policing alleHe mutations which arise in already existing
gene loci. Because of natural selection, organisms have been able to adapt to
changing environments, and by adaptive radiation many new species were
created from a common ancestral form. Y et, being an effective policeman,
natural selection is extremely conservative by nature. Had evolution been
entirely dependent upon natural selection, from a bacterium only numerous
forms of bacteria would have emerged. The creation of metazoans, vertebrates
and finally mammals from unicellular organisms would have been quite impos-
sible, for such big leaps in evolution required the creation of new gene loci
with previously nonexistent functions. Only the cistron which became redun-
dant was able to escape from the relentless pressure of natural selection, and by
escaping, it accumulated formerly forbidden mutations to emerge as a new gene
locus.
May 1970 SusuMu HNO

Acknowledgments
First of all, my gratitude is to my colleagues, especially to NIELS B. ATKIN,

WILLY and MARIA LuiSA BECAK, ALPRED GROPP and ULRICH WoLF. Although they
may not totally agree with what I have written, much of the evidence on which
my arguments are based has been collected as a result of our cooperative ven-
tures.
I am also grateful to Mrs. SHARYL BALES for her patient help in preparing
this manuscript. My esteemed colleagues, MELVIN CoHN and ERNEST BEUTLER,
were kind enough to go through a rough draft of this manuscript. Their advice
and comments proved invaluable.
Writing a book in one's spare timeisnot an easy task; I must have irritated
the people in my laboratory, my colleagues at the institute and my family often
during this period. I offer my sincere apologies.
This work was supported in patt by a grant (CA 05138) from the National
Cancer Institute, U.S. Public Health Service.
Contents
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Part 1
The Creation of Life Based on the lnherent Complementality
between Purine and Pyrimidine Bases
Chapter I. Perpetuation of the Germ Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Chapter II. Replication ofNucleic Acids on the Basis of A-T, G-C Complement-
ality and the Origin of Life. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1. The Presence of Self-replicating Nucleic Acids in "Prebiotic" Condition 6
2. Emergence of Transfer RNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3. Division of Labor between DNA and RNA . . . . . . . . . . . . . . . . . . . . . . . 12
4. Emergence of Ribosomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Chapter III. Chromosomes of Eukaryotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

1. Centromere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2. Nucleolar Organizer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3. Heterochromatic Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4. Further Need for Untranscribable Base Sequences . . . . . . . . . . . . . . . . . . 18
5. Histones as Nonspecific Repressors of Transcripdon . . . . . . . . . . . . . . . . 18
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
XII Contents
Part 2
Mutation and the Conservative Nature of Natural Selection
Chapter IV. Mutation as a Change in the Base Sequence of a DNA Cistron 21

1. Mutations Affecting Structural Cistrons . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
a) Frame-shift Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
b) Nonsense Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
c) Missense Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
d) Samesense Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2. Mutations Affecting Transfer RNA Cistrons . . . . . . . . . . . . . . . . . . . . . . . 24
a) Suppressor Mutation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
b) Mutations which Result in Ambiguous Coding . . . . . . . . . . . . . . . . . . 25
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Chapter V. Forbidden Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

1. Forbidden Mutations Affecting Tran{fer RNA Cistrons . . . . . . . . . . . . . . . . 26
2. Forbidden Mutations of Structural Cistrons . . . . . . . . . . . . . . . . . . . . . . . . . 27
3. Forbidden Mutations Favored . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Chapter VI. Tolerable Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

1. Neutral Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2. Favored Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3. Convergent Evolution and Recurrent Mutations . . . . . . . . . . . . . . . . . . . . 37
4. Atavism and Revertant Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Chapter VII. The Conservative Nature of Chromosomal Evolution . . . . . . . . 41

1. The Absence of a Close Linkage Requirement for Functionally Related
Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2. Inversion as an Interna! Rearrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
3. Robertsonian Fusion: The Creation of One Metacentric by Fusion of
Two Acrocentrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4. The Creation of a Sterility Bartier by Chromosomal Changes . . . . . . . . 44
5. Conservation of the Original Linkage Groups . . . . . . . . . . . . . . . . . . . . . . 45
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Chapter VIII. The Spontaneaus Mutation Rate . . . . . . . . . . . . . . . . . . . . . . . . . 48

1. Forbidden Mutations V ersus Tolerable Mutations . . . . . . . . . . . . . . . . . . . . 48
2. The Mutation Rate and Cistron Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3. Intragenie Recombination and the Principle of Polymorphism Genera-
ting More Polymorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4. On the So-called Living Fossils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Contents XIII
Chapter IX. The Rate of Evolution and the Importance of Isolation . 0 0 0 0 55

1. Isolation as a Prerequisite for Speciation . o o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 56
20Population Size and the Price of Success .. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 57
3. Generation Time and the Rate of Evolution . 0 0 0 0 0 0 0 0 0 0 0 0 57
References ......... 0 o o 0 0 o 0 0 0 o 0 0 o 0 0 0 o 0 0 58
Part 3
Why Gene Duplication?
Chapter X. Duplication for the Sake of Producing More of the Same 59

1. Genes for Ribosomal RNA .......... 0 0 0 0 0 0 0 0 0 0 0 0 0 60
2. Genes for Transfer RNA o 0 61
3. Inherent Disadvantage of Having Multiple Copies of the Same Gene . . 62
References .... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 65
Chapter XI. The Attainment of a Permanent Heterozygous Advantage by the

Incorporation of Two Former Alleles into the Genome . . . . . . . . . . . . . . . . 65
References .................. 0 0 0 0 0 0 0 0 67
Chapter XII. The Differential Regulation of Former Alleles and Their Trans-
formation to Isozyme Genes ............. 0 o 0 0 o 67
References .... 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 71
Chapter XIII. The Creation of a New Gene from a Redundant Duplicate of

an Old Gene ............................................ o 71
1. The Case of Trypsin and Chymotrypsin ..... 0 0 0 0 0 0 0 0 72
2. The Protein of Microtubules and Actin of the Skeletal Muscle ... 0... 74
3. Myoglobin and Hemoglobin ..... 0 0 0. 0................ 76
4. L-and H-chains ofimmunoglobulin .... o o o........ 77
5. The Emergence of a New Gene by a Frame-shift Mutation . o 0 80
References .................. 0 0 0 80
Chapter XIV. Duplication of Regulatory Genesand Receptors ..... 0..... 82

1. Hierarchy of Regulatory Mechanisms ..... 0 0 ................... 0 0 83
2. The Requirements to be a Regulator of Gene Activation . . . . . . . . . . . . 84
3. Concordant Duplication of a Primary Regulatory Gene and a Structural
Gene ........................... 0 0 0........... 85
4. Morphological Changes Due to Functional Diversification of a Dupli-
cated Regulatory Gene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
5. Duplication of Receptor Sites Adjacent to Structural Genes . . . . . . . . . . . 87
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
XIV Contents
Part 4
Mechanisms of Gene Duplication
Chapter XV. Tandem Duplication Involving Part ofOne Linkage Group at a
Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
1. Unequal Exchange between Two Chromatids of the Same Chromosome 89
2. Unequal Crossing-over between Two Homologaus Chromosomes during
Meiosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
3. Regional Redundant Replication of DNA . . . . . . . . . . . . . . . . . . . . . . . . . 92
4. Merits and Shortcomings of Regional Duplication. . . . . . . . . . . . . . . . . . . 94
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Chapter XVI. Polyploidy: Duplication of the Entire Genome . . . . . . . . . . . . . 98

1. Incompatibility between Polyploidy and the Well Established Chromo-
somal Sex Determining Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
2. Autopolyploidy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3. Allopolyploidy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4. Diploidization of the Tetraplaid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
5. Elimination of Certain Chromosomes during Diploidization . . . . . . . . . . 104
6. Dosage Effect of Regulatory Genes in the Tetraplaid . . . . . . . . . . . . . . . . 104
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Chapter XVII. Other Mechanisms for Achieving Gene Duplication . . . . . . . 107

1. Polysomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
2. Incorporation of Supernumerary Chromosomes . . . . . . . . . . . . . . . . . . . . 107
3. Lysogeny: Incorporation ofthe Viral Genome . . . . . . . . . . . . . . . . . . . . . 108
4. Viral Transducdon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Part 5
Evolution of V ertebrate Genomes
Chapter XVIII. Primitives Inherit the Earth . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
1. Emergence of the First V ertebrates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
2. The Type of Fish which were Prepared to Become Land Living Animals 113
3. Evolution from Fish to Amphibians . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4. The Invention of the Amniote Egg and the Emergence of Reptiles and
Birds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
5. Synapsida and the Emergence of Mammals . . . . . . . . . . . . . . . . . . . . . . . . . 118
6. Mammals in General............................................ 120
7. Primatesand Man . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Contents XV
Chapter XIX. Nature's Great Experiment with Gene Dupllcation during

Evolution from Tunicate-llke Creatures to Fish . . . . . . . . . . . . . . . . . . . . . . . 124
1. The Genome Size of Tunicate-llke Creatures from which Vertebrates
Emerged . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
2. Extreme Diversity of Genome Size Exhibited by Fish . . . . . . . . . . . . . . . 125
a) Changes in Genome Size which are Due Exclusively to Tandem
Dupllcation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
b) Meaninglessness ofExclusive Dependence upon Tandem Dupllcation 128
c) Efficacy of Tetraploidy as a Means of Acquiring Functionally
Diversified Dupllcated Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
d) Possible Polyphyletic Origin of Terrestrial Vertebrates . . . . . . . . . . . . 130
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Chapter XX. Evolution from Amphibians to Birds and Mammals and the
Abrupt Cessation of Nature's Experiment at the Reptillau Stage 132
1. Frogs Verstts Salamanders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
2. Diapsidian Reptiles and Birds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
3. Synapsida Line of Reptiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
4. Mammals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Chapter XXI. Whence Comes Man? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

1. Uniformity of the Genome Size and the Number ofDupllcated Gene Loci 139
2. Was Diversification of Placental Mammals Due Strictly to Allelle
Mutations of the Already Existing Gene Loci? . . . . . . . . . . . . . . . . . . . . . 141
3. An Evolutional Mechanism which Anticipated Future Needs . . . . . . . . . 142
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Plates I-VIII . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

Introducclon
When CHARLES DARWIN (1872) proposed the then revolutionary idea of evolution
by natural selection, he had no clear understanding of genetics. This is all the more
remarkable because evolution is nothing but the consequence of continuous accumula-
tion of genetic changes within the genome, and natural selection operates only
because individuals which comprise a population invariably display some degree of
genetic diversity.
The science of genetics, originally espoused by MENDEL (1885), did not come to
full bloom until the beginning of this century when MoRGAN and his group began a
systematic study on the fruit fly (Drosophila). They estahlished that genes (hereditary
units) are lineally aligned on individual chromosomes which are contained within the
cell nucleus (MoRGAN, 1911). At that time, only morphological traits were used as
genetic markers. With inherited morphological traits, it was not often certain
whether or not one trait seen in one species was genetically homologaus to a similar
trait in another species. Because of this limitation, until recently, the genetic studies
of evolution were confined to either natural selection within the species or the rela-
tionship between closely related spec1es, such as various fly species belonging to the
genus Drosophila (DoBZHANSKY and PAVOLOVSKY, 1958).
In the 1950's, the molecular structure ofDNA (Desoxyribose Nucleic Acid) was
finally resolved (WATSON and CRICK, 1953). As a result, we came to understand
individual genes as DNA cistrons and a mutation as a base substitution within a
cistron. Further, it became possible to translate the base sequence of a DNA cistron
to the amino acid sequence of the polypeptide chain it produces. This enabled us to
determine the homologaus gene loci in diverse organisms, including man and bac-
teria. The comparison of the amino acid sequences of homologaus peptide chains
from various vertebrate species immediately revealed the extremely conservative
nature of the structural genes (MARGOLIASH, 1963). We came to the realization that
alleHe mutations of already existing gene loci cannot account for major changes in
evolution.
1 Ohno, Evolution hy Gene Duplication
2 Introduction
As long as a particular function of an organism is under the control of a single

gene locus, natural selection does not permit perpetuation of mutations which
result in affecting the functionally critical site of a peptide chain specified by that
locus. Hence, allelic mutations are incapable of changing the assigned function of
genes.
Gene duplication emerged as the major force of evolution. Only when a redundant
gene locus is created by duplication is it permitted to accumulate formerly forbidden
mutations and emerge as a new gene locus with a hitherto unknown function.
Wehave written two short reviews (OHNo et al., 1968; HNO, 1969) stressing the
role gene duplication played in vertebrate evolution. This theme is expanded herein
by reconstructing the process of evolution which produced man and other mammals
from primitive fish of 300 million years ago.
In this golden age of biology, a book faces the danger of becoming obsolete
before its publication. It is my beliefthat in order to avoid early obsolescence, the
author, judging on the basis of the scant evidence available, is obliged to anticipate
future developments and paint a picture with broad strokes of his brush. This I have
clone rather freely in this book.
References
DARWIN, C. R.: The origin of species. 6th Ed., The world's classics, Oxford: University Press
1872 (reprinted 1956).
DoBZHANSKY, T., PAVOLOVSKY, 0.: Interraeial hybridization and breakdown of co-adapted
gene complexes in Drosophila paulistorum and Drosophila willistoni. Proc. Natl. Acad. Sei.
us 44, 622-629 (1958).
MARGOLIASH, E.: Primary structure and evolution of cytochrome C. Proc. Natl. Acad. Sei.
us 50, 672-679 (1963).
MENDEL, G.: Versuche ber Pflanzen Hybriden. Verhandl. Naturforsch. Verein Brnn,
4, 3--47 (1865).
MoRGAN, T. H.: An attempt to analyze the constitution of the chromosomes on the basis of
sex-limited inheritance in Drosophila. J. Exptl. Zool. 11, 365--413 (1911).
HNO, S.: The role of gene duplication in vertebrate evolution. In: The biological basis of
medicine, (BmAR, E. D., BmAR, N., Eds.) Vol. 4, Chapter 4, pp. 109-132. London:
Academic Press 1969.
-,WoLF, U., ATKIN, N. B.: Evolution from fish to mammals by gene duplication. Hereditas
59, 169-187 (1968).
WATSON, J. D., CRrCK, F. H. C.: Genetical implications of the structure of desoxyribose
nucleic acid. Nature 17, 964-966 (1953).
Part 1
The Creation of Life Based on the lnherent

Complementality between Purine and Pyrimidine Bases
Chapteri
Perpetnation of the Germ Line

A multicellular organism is a mortal. As such, a two-year-old mouse is a very old
mouse indeed. Even with the best medical care, no one would expect a mouse to live
5 years. In the case of man, gradual deterioration which accompanies aging appears to
begin at about 30; a 100-year-old man is as rare as a 2-year-old mouse. Yet, in a strict
sense, only the somatic cells of the body are mortals.
HAYFLICK and MooRHEAD (1961) have shown that fibroblasts removed from
human fetuses have a finite life span. They will divide 50 times and no more. Fibro-
blasts removed from an older person divide fewer times before the inevitable sene-
scence sets in.
Somatic cells of the body are, in a sense, slaves which are forced to perform
gratuitous functions for the good of the body as a whole. For instance, the production
of hemoglobin is essential to the sustenance of the body, but this act is a heavy
burden for erythroblasts of bone marrow which make the hemoglobin molecules.
Such a slave system functions only if individual somatic cells are given a finite life
span; otherwise, mutants which have ceased to perform an assigned gratuitous
function would enjoy an immediate selective advantage over still obedient fellow
slaves. Such rebels would then take over the system. If normal somatic cells were
endowed with immortality, the incidence of neoplastic growths would approach
proportians which would endanger the survival of a species. Neoplastic cells are
an example of such rebellious mutants which succeeded in their attempt to become
immortal (CoHN, 1968).
On the contrary, our germ cells as well as germ cells of all other creatures which
inhabit the earth today have lived for hundreds of millians of years, and each has the
potential of immortality.
Going back the time scale, we can see that the number of ancestors doubles at
each generation. It took only two parents to produce each of us, but there were four
grandparents and eight great-grandparents. A truly staggering number of more than
500,000 (219) different persans living twenty generations ago could have contributed
their germ cells to the formation of one individual living today. During the 15th
1*
4 The Creation of Life Based on the Inherent Complementality
century, however, with the exception of warriors of various conquering tribes, our
ancestors tended to stay in the same locality. It is doubtful that any interbreeding
unit of the 15th century approached the size of 500,000. One comes to the realization
that each of us is the product of some degree of inbreeding. In evolution, rather
intensive inbreeding forced by isolation has been a si11e qua 11011 of speciation.
Man is essentially a Pleistocene animal which emerged during the last great ice
age. Thus, little more than 2 million years ago, our germ cells were no doubt contained
in ape-like primates resembling Australopithecines. During the Eocene epoch which
began about 60 million years ago, a lemuroid creature was the sole representative of
the order Primate, and approximately 250 million years ago, a certain reptile derived
from the ancestral cotylosaurs was destined to be the ancestor of ail mammals.
Through these intermediaries, the ancestry of our germ ceils can be traced to a
particular crossopterygian fish which slithered out of the watet some 280 million years
ago to become the first amphibian.
Quite clearly, ailliving organisms on this earth are interrelated at sametime in the
past; even man and bacteria must have a common ancestry. Purine and pyrimidine
bases which are building blocks of nucleic acids as weil as some simpler amino acids
were already in abundance in primeval waters of this earth a billion or so years before
the emergence of the first life form. lt appears that self-replication of polynucleotides,
although imprecise, also preceded the creation of life. There is no a priori reason
why only the 3' to 5' linkage should be employed for polymerization of mononucleo-
tides, since the 5' to 5' linkage or even the 2' to 5' linkage in the case of RNA would
do just as weil. The very fact that every living creature of today utilizes the 3' to 5'
linkage indicates that they are the descendants of the first living form which happened
to have utilized polynucleotides of the 3' to 5' linkage type. Evolution, then, is the
history of an immortal germ line which was created eons ago and which has been
diverging ever since.
References
CoHN, M~: What can Escherichia coli and the plasmacytoma contribute to understanding
differentiation and immunology? Symp. int. Soc. Cell. Biol., Vol. 7, pp. 1-28. WARREN,
K. B., Ed. New York: Academic Press 1968.
HAYFLlCK, L., MooRHEAD, P. S.: The serial cultivation of human diploid cellline. Exptl.
Cell Research 25, 585-621 (1961).
Chapter!I
Replication of Nucleic Acids on the Basis of A-T, G-C

Complementality and the Origin of Le
The immortality of the germ line defines the exact molecular requirement to be a
gene (a master of inheritance). Unlcss affected by a mutation, each inherited trait
perpetuates itself through many, many generations of individuals. The molecule
which is a gene must have an inherent property, so that the exact replica of itself can
be made before each ceil division.
Replication of Nucleic Acids on the Basis of A-T, G-C Complementality 5
Desoxyribonucleic acid (DNA) uniquely quallfies for this role. During the
19th century, MrnsCHER (1871), working on pus removed from infected wounds of
patients, bad already identified desoxyribonucleic acid as a biological constituent.
Subsequently, FEULGEN (1928) found that Schiff's reagent for the identification of
aldehyde-containing compounds could be modified to localize the DNA-containing
structure within the fixed cell. The application of Feulgen's stain to cells of various
higher animals and plants revealed that DNA occurs exclusively in the chromosomes.
Thus, during the 1930's, it became almost certain that the so-called genes must be
DNA molecules.
DNA is merely a polymer of mononucleotides; each nucleotide being made of
one purine or pyrimidine base, one desoxyribose and one phosphate. There are only
four different kinds of bases; adenine and guanine representing purines, and thymine
and cytosine representing pyrimidines. On the surface, DNA appears far less compli-
cated than protein. How can suchapparently simple molecules self-replicate and specify
the exact amino acid sequences of hundreds of thousands of different kinds of
enzymes and non-enzymatic proteins contained in the body of an individual?
The answers to these questions began to come in the 1950's, at which time the
most exciting era in the history of biology began.
"DNA" and "double helix" are now well known words, yet the cardinal fact
which is not as well appreciated is that not only the replication of DNA but also the
transcription of genetic messages encoded within DNA are entirely dependent upon
the complementality that exists between adenine (A) and thymine (T) as well as be-
tween guanine (G) and cytosine (C). CHARGAFF (1951) discovered that while DNA
from different sources, such as from calf thymus and from salmon sperm, contained
adenine and guanine in varying ratios, in every DNA, the number of adenine mole-
cules was equal to the number of thymine molecules, and the number of guanine
molecules equaled the number of cytosine molecules. Thus, the base ratio, A + G
= T + C, holds true in all DNA.
The structure of DNA molecules elucidated by WATSON and CRicK (1953) is as
follows: DNA consists of two complementary strands. Bach strand is made of
nucleotides in 3' to 5' linkage. Mononucleotides are held tagether by bonds between
the phosphate molecule and carbon 3 and carbon 5 of adjacent desoxyribose. From
this poly-sugar-phosphate backbone, the purines and pyrimidines attached to carbon 1
of each sugar project. Two complementary strands form a double helix which is
held tagether by hydrogen bonds between the pairs of bases; adenine on one strand
facing thymine on the other, and guanine facing cytosine. The antiparallel strands
form a right-handed helix which undergoes one complete revolution with each ten
nucleotide pairs.
For replication, the two strands of the double helix are separated. Mter un-
coupling of the paired bases, each strand then serves as a template for recreating a
missing strand complementary to itself. Thus, if the bases projecting from one of the
two separated strands follow the sequence A,G,G,C,A,T, the newly synthesized
strand gets the sequence T,C,C,G,T,A. The other strand which bad the sequence T,C,
C,G,T,A, couples with the new strand which has the sequence A,G,G,C,A,T. This
mechanism based on the complementality between A and T, as well as between G
and C, enables the cell to make exact copies of the DNA molecules however many
times the cell may divide. Indeed, the self-replication of DNA molecules quallfies
them to be the masters of inheritance (Fig. 1). The actual work of synthesizing a
new strand of DNA is mediated by an enzyme, DNA polymerase, and the subunits
used are 5'-nucleoside monophosphate previously activated by becoming 5'-nucleo-
side triphosphatesuch as desoxy ATP (desoxyadenosine triphosphate) (KoRNBERG,
1961).
Fig. 1. The schematic illustration of DNA replication. In order to replicate, two strands of a
double helix have to separate from each other. Pentagons represent desoxyribose and circles
represent phosphate molecules. The old strands are painted solid black, while newly synthe-
sized strands are outlined
1. The Presence of Self-replicating Nucleic Acid in "Prebiotic" Condition

Synthesis of simple amino acids such as glycine, alanine and serine, as weil as
purines and pyrimidines, from ammonia, hydrocyanic acid, methane, and carbon
dioxide, probably occurred in primeval atmosphere and in the oceans, while being
catalyzed by ultraviolet light from the sun, cosmic radiation, and radioactive minerals
on the surface of the earth a billion or so years before the emergence of the first
living creature (CALVIN and CALVIN, 1964).
Since the complementary replication of a nucleic acid makes use of certain
structural characteristics inherent in the bases themselves, it is almost certain that
reasonably accurate self-replication of a polynucleotide is possible in the absence of
enzymes. This means that self-replication of polynucleotides also began to occur in
primeval soup long before the emergence of the first life on this earth (ORGEL, 1968).
The evidence for this belief comes from a variety of test tube studies. When solutions
containing simple derivatives of adenine and uracil are mixed, crystals containing
hydrogen-bonded mixed dimers often form. The same is true for a guanosine and
cytosine mixture. However, mixed crystals containing derivatives of other pairs of
bases are unknown (KATZ et al., 1965). In another experiment, it was found that a
stable helix could be formed in dilute aqueous solution for the pair polyuridylic acid-
Emergence of Transfer RNA 7
adenosine-5' -phosphate and for the pair polycytidylic acid-guanosine-5' -phosphate

which showed that once a polynucleotide chain is formed it can act as a template to
orient mononucleotides (HowARD et al., 1966). Furthermore, it has been shown
that adenosine-5' -phosphorimidazolide reacts with remarkable efficiency on a
polyuridylic acid template to give internucleotide bonds (WErMANN et al., 1968).
Thus, the demonstration is complete, showing that base pairing specificity resides in
the bases themselves, and that self-replication of nucleic acid could have occurred in
the "prebiotic" condition before the advent of enzymes.
It would appear that life on this earth owes its creation to the inherent comple-
mentality that exists between a couple of purine-pyrimidine base pairs. In this
connection, it is of interest to note that template-directed reaction in the "prebiotic"
condition does not lead to the preferential formation of 3' to 5' linkages. When
ribonucleotides are condensed with ribonucleosides, the 2' to 5' isomer is always the
most abundant and the 5' to 5' isomer is the next most abundant. Similarly, in the
desoxy series, the 5' to 5' isomer is the main product. Perhaps the earllest double
stranded polymers, if they were RNA, contained both 2' to 5' and 3' to 5' linkages.
The exclusion of the 2' to 5' linkages probably occurred along with the evolution of
polymerases after the creation of life (ORGEL, 1968).
2. Emergence of Transfer RNA

Transition from the self-replicating free polynucleotides to the first life form
would have required the creation of a mechanism which could direct the ordered
synthesis of polypeptide chains from available free amino acids, since polypeptide
chains, although incapable of self-replication, are infinitely more versatile than
nucleic acids as catalysts of biochemical reactions. Once the mechanism which
translated the base sequence of a self-replicating nucleic acid to the amino acid
sequence of a polypeptide chain came into being, the eventual emergence of life
forms was assured. Thus, it appears that the evolution of transfer RNA harked the
creation of life on this earth.
On one hand, each kind of transfer RNA had to have a specificity to recognize a
particular base sequence of nucleic acid, and on the other hand, it had to have a
preference to bind with a specific amino acid. Despite such specialized requirements,
not only is all transfer RNA of modern organisms made of about 80 bases, but they
also share the following common characteristics:
1. The 3'-end invariably has the base sequence CCA. When charged, an amino
acid is bound in an amino acyl linkage to carbon 3 of the ribose of the terminal
adenosine.
2. As first transcribed, transfer RNA is made of the four ordinary bases; A, G, U
and C. But, some of them are later modified to various derivatives. For example,
adenine may be modified to hypoxanthine, and uracil to dihydrouracil or pseudo-
uracil. Thymine, normally used only by DNA, is also found in this dass of RNA.
3. A transfer RNA assumes a "cloverleaf" configuration since the base sequences
in certain segments of a polynucleotide are complementary to those in other parts of
the same molecule, and these complementary sequences engage in base pairing by
hydrogen bonds (HoLLEY et al., 1965; MADISON et al., 1966).
s
8
~ i i
y-u-u-cr..-u
:i
~
s
8
~ i ~
u-u-cr..
I
zi
s8 s0 ~ u
~ :i
u-u
~ i .. i
t..(-u-0-u-Z--Z
7~ :i 'bn:
4'1 ~
I ' s<G
zi ~
s
0
u
~ i
u-u-0-u-u-z
.. :i :i :i
~
The smallness of their size and the common chru:acterlstics shru:ed by all Jransfer
RNA apparently indicate their ancient origin. At the beginning of life, how many
sequential bases did individual Jransfer RNA recognize? Nucleic acids (messenger RNA
oflatter times) were and ru:e madeoffout different bases; adenine (A) and guanine (G)
which ru:e purines, and thymine (T) or uracil (U) and cytosine (C), which are pyrimid-
ines. If each of the ancient Jransfer RNA recognized two successive bases in the base
Emergence of Tranrfer RNA 9
5
8
I
1 ::r:
u-Z>
5-tS-~-z u
~
::r:0
5
0
u 8 ::r: ::r:
4u-u-u-0-Z
:l:lll :l
0
4u-u-u(
:l J_J )u-o::r:
1.. I U=U
:l ::r: ::r:
~ z
::r:0
8 ::r: ::r:
~ :l
u-u-u(
J_J )u-::r:
L U=U
~ ~ ~
5 ::r:0
8I ..
::r:
.. o
8I ::r:o . .
5-5-5-8 5-0-5
~z L4
2
5
8
~u-u-0
:l ::r:
~z
sequence of a nucleic acid, these nucleic acids could have contained 16 (42) different
kinds of genetic messages, and 16 different kinds of transfer RNA could have specified
16 different amino acids. In the primeval soup of this earth many eons ago, there
probably were only 10 or so simple amino acids which were available to ernerging
life forms. At that time, the doublet coding system would no doubt have worked.
Y et, the very fact that allliving organisms universally use the triplet coding system
Table 2. The genetic code. Bach amino acid is specifted by a coding triplet in messenger RNA.
Termin. indicates a known terminating codon in E. coli
5'-terminal Middle nucleotide 3'-terminal
u c A G
u Phe Ser Tyr Cys u

Phe Ser Tyr Cys c
Leu Ser Termin. Termin. A
Leu Ser Termin. Trp G
c Leu Pro His Arg u

Leu Pro His Arg c
Leu Pro Gin Arg A
Leu Pro Gin Arg G
A Ile Thr Asn Ser u

Ile Thr Asn Ser c
Ile Thr Lys Arg A
Met Thr Lys Arg G
G Val Ala Asp Gly u

Val Ala Asp Gly c
Val Ala Glu Gly A
Val Ala Glu Gly G
reveals that from the very beginning transfer RNA evolved to recognize a set of three
consecutive bases of nucleic acids. Once life was created, a change in the codon size
would necessarily have made a mockery of all previous messages. Thus, such a
change would have exterminated allliving forms which existed at that time (CRICK,
1968). It appears that tran{fer RNA has not changed substantially since the time of
creation.
Inasmuch as transfer RNA read three consecutive bases (coding triplet) of nucleic
acids as a message, these nucleic acids generated 64 (43) different kinds of messages
and this resulted in a great redundancy of messages. A transjer RNA which recognized
a specific amino acid must have recognized not one particular codon but a number of
codons. Since subsequent increase in the number of amino acids available for
synthesis ofpolypeptide chains was merely from 10 or so to 20 (Table 1), redundancy
of codons persists to this date.
A method for actually deciphering the genetic messages contained in messenger
RNA was introduced by NrRENBERG and MATTHAEI (1961). Using the enzyme
polyribonucleotide phosphorylase, they synthesized a polyuridylic acid; an artificial
messenger RNA containing no base except uracil. When this RNA was mixed with
free amino acids and necessary ingredients such as ribosomes and transfer RNA
isolated from bacteria, a polypeptide made exclusively of phenylalaninewas synthesized
de novo. Thus, it was established that a codon, UUU, is recognized by a transjer RNA
which specifies an aromatic amino acid, phenylalanine. Subsequendy, the nature of
message contained in each of the 64 possible codons was clarified (Table 2). Observing
Table 2, it should be noted that redundancy most often resides on the third base
of the coding triplet. For instance, alanine is specified by any of the four codons
Emergence of Transfer RNA 11
having GC as the first two bases. The third base can be G, C, A or U. Similarly, any
of the four codons having GU as the first two bases can specify valine.
The recognition of a codon by a specific transfer RNA is also based on the comple-
mentality that exists between two pairs of bases, in that a set of three consecutive
bases (anticodon) in the middle of a transfer RNA engages in base pairing with a
codon in the nucleic acid which is to be translated. Unless a degree of infidelity in
the base pairing between a codon and an anticodon was introduced at the beginning,
64 different anticodons, and, therefore, that number of transfer RNA, were needed for
translation of the base sequence of nucleic acid to the amino acid sequence of a
polypeptide chain. For example, a codon GCA could only be recognized by an
alanine transfer RNA having the anticodon CGU. Three other kinds of transfer RNA
would be needed to translate all four codons for alanine.
It appears that this needed infidelity in the mannet of base pairing between
codons and anticodons was provided by the introduction of unusual bases to anti-
codons usually at the third position. Hypoxanthine (HyX) isaderivative of adenine
(A), but unlike A, HyX can pair with not only U, but also with C and A. Thus, an
alanine transjer RNA having the anticodon CGA would only have recognized a codon
GCU, but an alanine transfer RNA having the anticodon CGHyX can recognize three
of the four codons for alanine (GCU, GCC and GCA).
Again observing Table 2, it should be noted that three codons (UAA, UAG and
UGA) are marked as terminating codons. They are also known as nonsense codons
(KAPLAN et al., 1965; WEIGERTet al., 1966; GAREN, 1968). Apparently, the genome of
Escherichia coli (a colon bacillus) does not contain a set of genes which specify
functional transfer RNA to match the codons UAA, UAG and UGA. Thus, at least
in this bacterium, translation of the base sequence of messenger RNA to the amino
acid sequence of a polypeptide chain stops at the nonsense ot chain terminating codon,
so that an amino acid specified by the preceding triplet codon becomes the carboxyl
end of a peptide chain. The chain terminating signals are obviously useful in that a
long strand of nucleic acid can be translated to two or more independent polypeptide
chains rather than a single long peptide chain. Such nucleic acids of modern organisms
are known as polycistronic messenger RNA (ATTARDI et al., 1963).
Of the 20 amino acids used by modern organisms for the synthesis of polypeptide
chains, methionine (Table 1) is unique in that it is specified by a single codon (AUG)
(Table 2). In the in vitro system of E. coli, translation of an artificial messenger to a
polypeptide chain becomes markedly efficient if AUG, the codon for methionine, is
placed at the 5'-end. In fact, most, if not all, of the peptide chains synthesized by
E. coli normally have formylmethionine (methionine with a blocked amino group)
for the amino end (MARCKER and SANGER, 1964; ADAMS and CAPECCHI, 1966). It
follows that in E. coli, AUG in messenger RNA serves as the initiating codon. A methio-
nine transjer RNA might have had this unique property from the very beginning of
life forms. Conversely, an initiating codon might have evolved later as a sophisticated
device.
From the above, one can get an inkling of what the intermediate form which
bridged the gap between self-replicating "prebiotic" nucleic acid and the first living
organism was like. It must have contained a number of self-replicating nucleic acids
(more likely RNA than DNA), some of which functioned as primordial transfer
RNA. Because of the presence of transjer RNA, the directed synthesis of polypeptide
chains from available amino acids must have occurred. Some of these peptide
chains may have served as enzymatic catalysts, while others formed a crude cell
membrane together with prebiotically synthesized peptide chains which were present
in the primeval soup (Fig. 2).
Fig. 2. My imagination of the intermediate form between the "prebiotic" self-replicating

nucleic acid and the first living form. The sphere is a creature. Self-replicating nucleic acid is
depicted at the upper right comer of the sphere. In the middle, the base sequence of a nucleic
acid is translated to the amino acid sequence of a polypeptide chain (black beads on a string)
mediated by a primordial transfer RNA (clover-leaf). Such a directed polypeptide together
with a prebiotically formed polypeptide (white beads on a string) form a crude cell membrane
which serves as the boundary between a creature and the environment
3. Division of Labor between DNA and RNA

Needless to say, the gene of modem organisms has a dual role. Contained in a
fertilized egg it has to be transmitted to all descendant cells which constitute the
body of an individual and through germ cells it has to be transmitted to all progeny
which constitute the next generation. The exact copying mechanism ofDNA replica-
tion based on the complementality between bases enables the gene to play this first
role. In addition, the genetic information encoded within the DNA molecule has to be
deciphered and expressed in the form of polypeptide chains during ontogenic
development, so that an individual is formed in accordance with the genetic program
contained within the nucleus. The base sequence of DNA, however, is not directly
translated to the amino acid sequence of a polypeptide chain. Again utilizing the
inherent complementality between bases, one of the two strands of DNA is first
transcribed to RNA (messenger RNA, transfor RNA and ribosomal RNA), and the base
Emergence of Ribosomes 13
sequence of messenger RNA is read by transfer RNA. Thus, there is a distinct division
of labor between DNA and RNA. Self-replicating DNA is for the preservation and
transmission of genetic messages to the progeny, while RNA is used as an inter-
mediary for the materialization of genetic messages contained in DNA.
At the very beginning, however, each nucleic acid of a crude living form must
have served a dual role, in that it self-replicated on one hand, and on the other, its
base sequence was translated to the amino acid sequence of a polypeptide chain or
chains. This is an inherently wasteful process, since self-replication perpetuates
the production of two base sequences which are complementary to each other. When
both sequences are read by transfer RNA, two polypeptide chains with totally
unrelated amino acid sequences would be produced. If one of the polypeptide
chains is useful for a particular function, the other is likely to be totally useless,
although, there is a remote chance that the other might be useful for quite an un-
related function. Such a system cannot be efficiently modulated by natural selection.
Thus, division of labor between DNA and RNA was the most logical progrcss to be
followed by the crude living form. Once this divisionwas established, only one base
sequence of DNA could be materialized in the form of a polypeptide chain. In this
way, the DNA ring of prokaryotes and subsequently the chromosomes of eukaryotes
must have evolved.
4. Emergence of Ribosomes
Increased sophistication of the machinery for translation of genetic messages
apparently led to the creation of subcellular particles which are ribosomes. In all
modern organisms, reading of a messenger codon by transfer RNA can only take
place inside a ribosome. Each ribosome is a ribonucleoprotein complex of 10 to
20 m1-1- in diameter. Ribosomes consist of two unequal subunits which are bound
together. The large subunits are 50 S in size, and the small subunits 30 S. The 30 S
subunit receives the messenger RNA, while the 50 S subunit provides a cavity for one
or more transftr RNA and anchors itself to the membranaus component of the endo-
plastic reticulum in the cytoplasm.
Only after the 5'-end of a messenger RNA is inserted into a ribosome, does the
synthesis of a polypeptide chain begin. A tran{fer RNA charged with an amino acid
fits into the cavity in the 50 S subunit and recognizes the messenger RNA codon
occupying the roof of the cavity. As the messenger RNA moves through the ribosome,
much like a tape through the head of a tape recorder, translation is effectuated and the
peptide chain grows. Once the 5'-end of a ntessenger RNA emerges from the first
ribosome, it can attach to a second ribosome, and the second synthesis of a poly-
peptide chain starts from the amino end. When the 3' -end of a messenger RNA has
moved through the first ribosome, the first copy of a complete polypeptide chain is
released. Inasmuch as many copies of the polypeptide chain are made continuously
from a single messenger RNA, at any given moment, a single messenger is attached to
several ribosomes. This unit of beads (ribosomes) on a string (messenger RNA) is
known as a polysome unit. For example, a messenger RNA which is 600 bases long,
thus coding for a polypeptide chain made of 200 amino acid residues, is attached to
about eight ribosomes.
The three different kinds of RNA are integral parts of the ribosomes. In the
case of vertebrate species, their sizes are 5 S, 18 S and 28 S. 5 S ribosomal RNA is
made of only 120 or so bases, and it shows certain similarities to transfer RNA, in
that thls dass of ribosomal RNA uses unusual bases such as pseudouracil and thymine,
and that there exists long sequences of complementary nucleotides. Presumably,
a 5 S ribosomal RNA molecule folds upon itself much in the same manner as does
the transjer RNA molecule (FoRGE'I' and WEISSMAN, 1967). 18 S and 28 S ribosomal
RNA molecules are too large to permit the complete base sequence analysis.
It appears that a pair of closely linked genes for 18 S and 28 S ribosomal RNA
transcribes a single dicistronic RNA. When transcribed, the precursor RNA is 45 S
in size. Subsequently, this strand is split into three pieces; one is the 18 S, the other
28 S, and the third piece appears tobe degraded (PERRY, 1962; PENMAN, 1967).
It was suspected for some time that the gene which specifies ribosomal RNA
(18 Sand 28 S, but not necessarily 5 S) resides in the nucleolar organizing region of
the chromosome. At mitotic metaphase of many plant and animal species, the
nucleolus organizing region is conspicuously visualized as the constricted region in
a heavily condensed chromosome. This constricted region hardly takes FEULGEN
stain for DNA, hence, the name SAT-region (Sine Acido Thimonucleico). The critical
evidence which substantiated the above suspidon was furnished by a deletion
mutation in the African water frog, Xenopus laevis. This species, with the diploid
chromosome nurober of 36, normally carries the nucleolus-organizer on a single
pair of chromosomes. Thus, diploid nuclei of the wild-type frog contain two nucleoli.
The diploid nucleus of frogs heterozygous for the deleted nucleolus organizer
contains only one nucleolus, and the homozygous mutants contain none (ELSDALE
et al., 1968). The homozygote which dies during embryonie development is not
only totally incapable of de novo synthesis of ribosomal RNA, but also DNA extracted
from the homozygous mutant does not hybridize with the wild-type ribosomal RNA
(WALLACE and BrRNSTIEL, 1966).
When DNA is melted, the two complementary strands which form a stable
double helix dissociate. If the solution is allowed to cool slowly, the two complement-
ary strands seek each other out and reform the double helix. However, single strands
of DNA can be made immobile by absorbing them on membranous filters. Single
strands of DNA so prepared from the \Vild-type Xenopus laevis include at least two
copies of the gene for ribosomal RNA. When such filters are dipped into a solution of
ribosomal RNA, the ribosomal RNA gene and ribosomal RNA transcribed from it
seek each other out and form a firmly bound pair on the basis of more or less exact
base-for-base complementality. This, then, is the technique ofRNA-DNA hybridiza-
tion.
The fact that DNA from a frog homozygous for the deleted nucleolar organizer
fails to hybridize with ribosomal RNA reveals that the mutant genome does not
contain the gene which specifies this dass of RNA.
References
ADAMS, J. M., CAPECCHr, M. R.: N-formylmethionyl-sRNA as the initiator of protein
synthesis. Proc. Nad. Acad. Sei. US 55, 147-155 (1966).
ArrARDI, G., NAONO, S., Rouvr:ERE, J., ]ACOB, F., GRos, F.: Producdon ofmessenger RNA
and regulation of protein synthesis. Cold Spring Harbor Symposia Quant. Biol. 28,
363-372 (1963).
CALVIN, M., CALVIN, G. ]. :Atom to Adam. Am. Scientist 52, 163-186 (1964).
Chromosomes of Eukaryotes 15
CHARGAFF, E.: Structure and function of nucleic acids as cell constituents. Federation Proc.
10, 654-659 (1951).
CRICK, F. H. C.: The origin of the genetic code. J. Mol. Biol. 38, 367-379 (1968).
ELSDALE, T. R., FrsCHBERG, M., SMITH, S.: A mutation that reduces nucleolar number in
Xenopus laevis. Exptl. Cell Research 14, 642-643 (1958).
FEuLGEN, R.: Histochemischer Nachweis von Aldehyden. V erhandl. deut. pathol. Ges.
28, 159-200 (1928).
FoRGET, B. D., WErSSMAN, S. M.: Nucleotide sequence of KB cell 5 S RNA. Science 158,
1695-1699 (1967).
GAREN, A.: Sense and nonsense in the genetic code. Science 160, 149-159 (1968).
HOLLEY, R. w., APGAR, J., EVERETT, G. A., MARQHISEE, M., MERRILL, s. H., PENSWICK,
J. R., ZAMIR, A.: Structure of ribonucleic acid. Science 147, 1462-1465 (1965).
HowARD, F. B., FRAZIER, J., SrNGER, M. F., MrLES, H. T.: Helixformation between poly-
ribonucleotides and purines, purine nucleosides and nucleotides. J. Mol. Biol. 16,
415-439 (1966).
KAPLAN, S., STRETTON, A. 0. W., BRENNER, S.: Amber suppressors: Efficiency of chain
propagation and suppressor specific amino acids. J. Mol. Biol. 14, 528-533 (1965).
KATZ, L., ToMITA, K., RrcH, A.: The molecular structure of the crystalline complex
ethyladenine: Methyl-Bromouracil. J. Mol. Biol. 13, 340-350 (1965).
KoRNBERG, A.: Enzymatic rynthesis of DNA. New York: John Wiley and Sons, Inc. 1961.
MADISON, J. T., EVERETT, G. A., KuNG, H.: Nucleotide sequence of yeast tyrosine transfer
RNA. Science 153, 531-534 (1966).
MARCKER, K., SANGER, F.: N-formyl-methionyl-S-RNA. J. Mol. Biol. 8, 835-840 (1964).
MmscHER, F.: ber die chemische Zusammensetzung der Eiterzellen. Hoppe-Seyler's Med.
Chem. Unters. 4, 441-460 (1871).
NIRENBERG, M. W., MATTHAEI, J. H.: The dependence of cell-free protein synthesis in
E. coli upon naturally occurring or synthetic polyribonucleotides. Proc. Natl. Acad.
Sei. US 47, 1588-1602 (1961).
ORGEL, L. E.: Evolution of the genetic apparatus. J. Mol. Biol. 38, 381-393 (1968).
PENMAN, S.: Ribonucleic acid metabolism in mammalian cells. New Engl. J. Med. 276,
502-511 (1967).
PERRY, R. P.: Cellular sites of synthesis of ribosomal and 4 S RNA. Proc. Natl. Acad. Sei.
us 48, 2179-2186 (1962).
WALLACE, H., BrRNSTIEL, M. L.: Ribosomal cistrons and the nucleolar organizer. Biochim.
et Biophys. Acta 114, 296-310 (1966).
WATSON, J. D., CRrcK, F. H. C.: Genetical implications of the structure of desoxyribose
nucleic acid. Nature 17, 964-966 (1953).
WEIGERT, M. G., GALLUCCI, E., LANKA, E., GAREN, A.: Characteristics ofthe genetic code
in vivo. Cold Spring Barbor Symposia Quant. Biol. 36, 145-150 (1966).
WmMANN, B. J ., LoHRMANN, R., ORGEL, L. E., ScHNEIDER-BERNLOEHR, H., SuLsToN, J. E.:
Template-directed synthesis with adenosine-5'-phosphorimidazolide. Science 161, 387
(1968).
Chapter III
Chromosomes of Eukaryotes
In bacteria and certain other unicellular organisms, the entire genetic material
exists in the form of a DNA ring. Such organisms are typically haploids and within
the cell there exists no sharp demarcation between the nucleus and the cytoplasm.
These organisms are called prokaryotes.
In sharp contrast, eukaryotes are typically diploids and most of them are multi-
cellular organisms (metazoans). The union of two gametes (haploid cells) initiates
the development of an individual. Within the genome (haploid nucleus), the genetic
material is partitioned into a fixed number of distinct structural entities which are
called chromosomes. Chromosomes stay within the boundary of the nuclear membrane
except during cell division. Thus, there is a sharp demarcation between the nucleus
and the cytoplasm within the cell.
Since all vertebrates are eukaryotes, the subject of chromosomes should be intro-
duced in this chapter. Fig. 3 (Plate I) shows the karyotype of a normal human male as
originally eluddated by T JIO and LEVAN (1956). It should be noted that the 46 chromo-
somes of male Homo sapiens are made of 22 pairs of homologaus autosomes (ordinary
chromosomes) and a pair of sex chromosomes; the large X-chromosome and the
small Y-chromosome. In the case of females, the sex chromosomes also constitute
a homologaus pair, for there are two X-chromosomes and no Y-chromosome. These
chromosomes in the pietute were fixed and stained at metaphase of mitosis; thus,
each is made of two sister chromatids. At the end of mitotic anaphase, sister chromat-
ids of each chromosome move toward opposite poles of the mitotic spindie. It is
assured that each of the two daughter cells receives an identical diploid set of genetic
material. Although one sister chromarid of a larger chromosome may measure more
than 1 [.L in thickness, on the basis of available genetic data, one has to assume that
if one chromatid is stretched to the extreme it would become a single continuous
strand of a DNA double helix. During the mitotic phase of the cell cycle, a single
strand is very tightly packed to form a chromatid which fadlitates easy transportation
by spindie fibers to a daughter cell.
1. Centromere
Each metaphase chromosome is marked by a primary constriction where the
boundary between two sister chromatids is not very clear. Such a constriction can be
seen in the middle or near the middle in some chromosomes, while in others, it is
situated extremely close to one terminal end. The position of a primary constriction
is a usefullandmark and helps to distinguish chromosomes of similar size as different
entities. Each chromosome has one primary constriction for it contains a centromere
where the chromosome attaches itself to the spindie fibers. Those chromosomes
having the centromere in the middle or near the middle are defined as metacentric
chromosomes, and those which have the centromere rather near to one end of the
chromosome are called subterminal chromosomes. The term "acrocentric chromo-
somes" refers to those having an almost terminally located centromere.
Each chromosome can afford to have only a single centromere or a single functional
unit of centromeres. When a chromosome possesses two centromeres spaced apart
(dicentric), at anaphase, there is a fifty-fifty chance that a single chromatid would be
pulled to opposite poles of the mitotic spindie. This would result in a bridge forma-
tion and subsequent chromosome breakage. Needless to say, natural selection does
not talerate the persistence of dicentric chromosomes. It must be that even if located
in different chromosomes all centromeres arehomologaus entities, for the centromere
of one chromosome when separated and attached to another chromosome can serve
that chromosome equally well. We have no idea as to what DNA base sequence
defines a segment of the DNA strand as a centromere.
Heterochromatic Region 17
2. Nucleolar Organizer
Observing Fig. 3 (Plate I), it should be noted that the short arm of the three acro-
centric autosomes of the 13 to 15 group, as weil as the short arm of the two small acro-
centric autosomes of the 21 to 22 group, are marked by a secondary constriction. The
secondary constriction usually, but not always, indicates the site of a nucleolar
organizer (HEITZ, 1933; KAUFMANN, 1934; DEARING, 1934). In the case of man, all
secondary constrictions on the acrocentric autosomes appear to indicate the sites of
nucleolar organizers (FERGUSON-SMITH and HANDMAKER, 1961; OHNo et al., 1961).
The fact that the nucleolar organizer carries the gene for 18 S and 28 S ribosomal
RNA has already been mentioned. It then follows that the secondary constrictions
carried by five different autosomes of man must be homologaus to each other.
3. Heterochromatic Region
The primary and secondary constrictions are only noticeable landmarks on the
metaphase chromosomes. Y et, during mitotic prophase, certain other regions of the
chromosomes stand out from the rest by virtue of premature condensation. These
prematurely condensed regions of prophase chromosomes are conspicuous in the
preceding interphase nuclei as chromocenters (condensed masses of DNA). The
regions of the chromosomes which are prone to premature condensation are said
to be made of heterochromatin (HEITZ, 1933), and, during S (synthetic) phase of
interphase, the heterochromatic regions characteristically replicate their DNA later
than other chromosomal regions which are said to be made of euchromatin (TAYLOR,
1960).
In certain instances, the heterochromatic condition is merely a temporary state of
inactivity assumed by a chromosome or a chromosomal region. The best example of
this is seen in the mammalian X-chromosome. While the male interphase nucleus of
mammals is free of a chromocenter, each interphase nucleus of female somatic cells
is characterized by the presence of a single conspicuous chromocenter (BARR and
BERTRAM, 1949) which represents one of the two X-chromosomes in the precociously
condensed state (OHNo et al., 1959). Through this means, mammalian species ap-
parently equalize the disparity that exists between the male and the female with
regard to the dosage of the X-linked genes (LYON, 1961; BEUTLER et al., 1962).
Because one of the two X-chromosomes ofthefemale is rendered inert by hetero-
chromatinization, individual somatic cells of the male as well as of the female are
effectively endowed with a single dose of each X-linked gene. The heterochromatic
condition in this case is clearly a temporary functional state assumed by the X, since
in female germ cells both X-chromosomes remain euchromatic (OHNo et a!., 1962).
In other instances, the heterochromatic condition reflects the inherent property of
that chromosomal segment. Such heterochromatic regions are devoid of structural
genes, therefore, it is a genetically empty region so to speak. This region must be
made of a segment of useless, untranscribable DNA base sequences. On the surface,
it would appear that natural selection would not permit the apparently useless
chromosomal segments to persist. On the contrary, upon careful consideration, one
realizes that in order to preserve the structural integrity of individual chromosomes,
certain segments of the chromosome should be made of useless base sequences. The
2 Ohno, Evolution by Gene Duplication
regionaraund the centromere is characteristically made of heterochromatin (centro-

meric heterochromatin). According to DARLINGTON (1935), there is a good evolutional
reason for the existence of centromeric heterochromatin blocks. During speciation,
the centromeric regions are most often involved in translocation between chromo-
somes. As a result of translocation, the chromosomal segment near the centromere
is lost. While the loss of dispensable heterochromatin is permissable, the presence
of important structural genes near the centromere would have placed a severe restric-
tion on the occurrence of chromosomal changes accompanying speciation.
Similarly, both ends of a chromosome tend to be made of heterochromatin
(telomeric heterochromatin). MuLLER (1932) believed that the telomere is essential
to the maintenance of the structural integrity of individual chromosomes. When a
chromosome is broken into two pieces by irradiation and other means, the broken
ends are never satisfied until they are fused with other broken ends, and this is the
basis of translocation, inversion, insertion and the formation of dicentrics which
follow chromosome breaks. Without the protection of telomeric heterochromatin,
the end of a chromosome too would never be satisfied until it fused with the end of
another chromosome. Sooner or later, all the chromosomes in the haploid comple-
ment would fuse with each other end-to-end until the formation of one huge ring
chromosome occurred.
4. Further Need for Untranscribable Base Sequences

Is the euchromatic region of the chromosome made only of useful DNA base
sequences which are transcribed and translated to various gene products? One
should think not. Each gene does not exist as a separate molecule, rather it represents
a part of a continuous DNA strand. There apparently is no physical interruption
between one gene and an adjacent one on either side of it. How is it then that each
gene as a rule transcribes aseparate messenger RNA? It must be that the space between
adjacent cistrons is occupied by a stretch of nonsense base sequence. Perhaps, RNA
polymerase cannot use such nonsense base sequences for a template .U sing the technique
ofDNA-DNA self-hybridization, BRITTEN and KoRNE (1968) have shown that nearly
10% of the mammalian genome is occupied by multiple copies of either one peculiar
base sequence or a small number of similarly peculiar base sequences (satellite DNA).
It might be that this dass of DNA represents untranscribable base sequences which
are used not only for the spacing of structural genes within the euchromatic region,
but also for the entire inherently heterochromatic regions.
YASMINBH and YUNrs (1969) have recently shown that the bulk of satellite DNA
is indeed concentrated in the autosomal heterochromatin fraction of the mause
genome. When nucleolar organizing regions of the chromosome of a salamander
(Triturus viridescens) were isolated and examined under the electron microscope by
MrLLER and BBATTY (1969), it was found that between the genes which were actively
engaged in transcription of a precursor molecule for 18 S and 28 S ribosomal RNA,
there exists a stretch of untranscribable DNA.
5. Histones as Nonspecific Repressors of Transcripdon

Prokaryotes, such as E. coli, are unicellular organisms. As one cell represents a
whole organism, it follows that most of the genes contained in a ring DNA are
Histones as Nonspedfic Repressors of Transcripdon 19
expressed most of the time. Perhaps for this reason, a ring DNA of prokaryotes
stays exposed. Thus, nearly all of its structural genes engage in transcriptional
activity, unless individually repressed by a special genetic regulatory mechanism.
Eukaryotes, on the other hand, are typically multicellular organisms. The genome
(haploid chromosome set) of each mammalian species contains roughly 3.5 x 10-9mg
of DNA. There is room for thousands and thousands of structural genes. If all
these genes in the nucleus are fully transcribed and translated simultaneously, the
cell would literally hurst open from congestion of overproduced RNA and protein
molecules. In eukaryotes, the process of somatic cell differentiation during embryonie
development insures that each cell type be specialized and use only some of the genes
in the nucleus. Tobe sure, various enzymes for basic metabolic pathways and proteins
for cell multiplication are needed by every cell regardless of its somatic cell type,
but these genes for hausehold chores comprise only a small fraction of the vertebrate
genome. A majority of structural genes specify products which are gratuitous for the
cells making them, but needed for the body; for example, insulin and other peptide
hormones, hemoglobin and immunoglobin. With regard to these genes for gratuitous
products, there is a distinct division of labor among somatic cell types. The gene for
a precursor of the hormone insulin is active only in Langerhans' islet cells of the
pancreas, and the genes for hemoglo bin peptide chains are active only in erythropoietic
cells ofbone marrow. Among all the somatic cell types of the body, only plasma cells
are the producers of immunoglobulin.
Quite clearly, in the case of multicellular organisms, it is more desirable to keep
most of the genes in the repressed state.It must be that the structural genes in the
metazoan genome remain dormant unless individually de-repressed by the special
genetic regulatory mechanism.
The activating regulatory mechanism can function only if the eukaryotes are
endowed with indiscriminate repressor molecules which bind with DNA and prevent
transcriptional activity. In this manner, every cistron would remain dormant until
repressor molecules are specifically removed. Indeed, in the chromosomes of
eukaryotes, DNA is intimately associated with the group of basic proteins known as
histones. STEDMAN and STEDMAN (1950) suspected that histones serve as repressors
of gene activity. HuANG and BoNNER (1962) and ALLFREY et al. (1963) have sub-
sequently shown that DNA cistrons which are bound with histones cannot engage in
transcriptional activity.
Histones are rather small molecules being made of 110 or 220 amino acid residues.
Most vertebrates appear to produce five or six different kinds of histones. They are:
a very lysine-rich (fl) histone, two slightly lysine-rich (f2a2 and f2b) histones, and
two arginine-rieb (f2al and f3) histones (JoHNs, 1966). A sixth kind, identified as a
serine-rich (f2c) histone, has been found exclusively in the mature nucleated erythro-
cytes of both avian and non-avian species (HNILICA, 1966; NEELIN et al., 1964).
Free amino groups of basic amino acids, lysine and arginine, which are abundant in
the carboxyl half of histones, indiscriminately associate with phosphate groups of
any DNA cistron (BusTIN et al., 1969). In addition, the -OH group of serines in
histones also participates in binding with phosphate groups of DNA.
In order to activate a particular structural gene locus in the genome, an activator
molecule specified by a regulatory gene of the eukaryote must selectively recognize
that particular structural gene and remove histune from it.
2*
20 The Creation of Life Basedon the Inherent Complementality
References
ALLFREY, V. G., LrTAU, V. C., MrRSKY, A. E.: On the role of histones in regulating RNA
synthesis in the cell nucleus. Proc. Natl. Acad. Sei. US 49, 414-421 (1963).
BARR, M. L., BERTRAM, L. F.: A morphological distinction between neurones of the male
and female and the behavior of the nucleolar satellite during accelerated nucleoprotein
synthesis. Nature 163, 676-677 (1949).
BEUTLER, E., YEH, M., FAIRBANKS, V. F.: The normal human female as a mosaic of X
chromosome activity: Studies using the gene for G-6-PD deficiency as a marker. Proc.
Natl. Acad. Sei. US 48, 9-16 (1962).
BRITTEN, R. J ., KoHNE, D. E.: Repeated sequences in DNA. Science 161, 529-540 (1968).
BusnN, M., RALL, S. C., STELLWAGEN, R. H., CoLE, R. D.: Histone structure: Asymmetrie
distribution of Iysine residues in lysine-rich histone. Science 163, 391-393 (1969).
DARLINGTON, C. D.: Recent advances in cyto!ogy. London: J. and A. Churchill, Ltd. 1935.
DEARING, W. H., J R.: The material continuity and individuality of the somatic chromosomes
of Ambystoma tigrinum, with special reference to the nucleolus as a chromosomal compo-
nent. J. Morphol. 56,157-179 (1934).
FERGUSON-SMITH, M. A., HANDMAKER, S. D.: Observations on the satellited human chromo-
somes. Lancet 1961 I, 638-640.
HErTZ, E.: Die somatische Heteropyknose bei Drosophila me!anogaster und ihre genetische
Bedeutung. Z. Zellforsch. Abt. Histochem. 20, 237-287 (1933).
HNrLicA, L. S.: Studies on nuclear proteins. I. Observations on the tissue and species
specificity of the moderately lysine-rich histone fraction 2b. Biochim. et Biophys. Acta
117, 163-175 (1966).
HUANG, R. C., BoNNER, J.: Histone, a suppressor of chromosomal RNA synthesis. Proc.
Natl. Acad. Sei. US 48, 1216-1222 (1962).
JoHNS, E. W.: Metabolism and radiosensitivity. In: The ce/1 nuc!eus, p. 116. London: Taylor
and Francis, Ltd. 1966.
KAUFMANN, B. P.: Somatic mitoses of Drosophila me!anogaster. ]. Morphol. 56, 125-156
(1934).
LYON, M. F.: Gene action in the X-chromosome of the mouse (Mus muscu!us L.). Nature
190, 372-373 (1961).
MrLLER, 0. L., J R., BEATTY, B. R.: Visualization of nucleolar genes. Science 164, 955-957
(1969).
MuLLER, H. ]. : Further studies on the nature and causesof gene mutations. Proc. VIth Int' 1
Congr. Genet. Ithaca, N.Y., 1, 213-255 (1932).
NEELIN, J. M., CALLAHAN, P. X., LAMB, D. C., MuRRAY, K.: The histones of chicken
erythrocyte nuclei. Can. J. Biochem. and Physiol. 42, 1743-1752 (1964).
0HNO, S., KAPLAN, W. D., KINOSITA, R.: Formation of the sex chromatin by a single
X-chromosome in liver cells of Rattus norvegicus. Exptl. Cell Research 18, 415--418
(1959).
- , TRUJILLO, J. M., KAPLAN, W. D., KINOSITA, R.: Nucleolus-organizers in the causation of
chromosomal anomalies in man. Lancet 1961 II, 123-125.
- , KLINGER, H. P., ATKIN, N. B.: Human ogenesis. Cytogenetics 1, 42-51 (1962).
STEDMAN, E., STEDMAN, E.: Cell specificity of histones. Nature 166, 780-781 (1950).
TAYLOR, J. H.: Asynchronaus duplication of chromosomes in cultured cells of Chinese
hamsters. J. Biophys. Biochem. Cytol. 7, 455--464 (1960).
TJIO, ]. H., LEVAN, A.: The chromosome nurober ofman. Hereditas 42, 1-6 (1956).
YASMINEH, W. D., YuNis, J. ].: Satellite DNA in mouse autosomal heterochromatin.
Biochem. Biophys. Res. Commun. 35,779-782 (1969).
Part 2
Mutation and the Conservative Nature ofNatural Selection
Chapter IV
Mutation as a Change in the Base Sequence of a DNA Cistron

Due to an inherent complementality which exists between the two base pairs,
adenine-thymine and guanine-cytosine, DNA is endowed with the unique property
of being able to make an exact replica of itself before each cell division. Yet, if the
mechanism of DNA replication were perfect and there was no room for mistakes, the
creation of divergent living creatures from a common ancestor could not have
happened.
Changes in the base sequence of individual cistrons do occur and such changes
are the cause of individual variability within a population. Natural selection exploits
these individual differences and evolution occurs. A heritable change in the base
sequence of a cistron is defined as a mutation. The observation that one mutation
usually affects only a single base pair within a cistron attests to the fact that the
mechanism of DNA replication is nearly perfect and that mistakes do not occur very
often. Because of this stable replication mechanism, however, a new mutation which
is favored by natural selection can perpetuate itself as a new inherited trait.
Different types of mutations shall be defined in this chapter. Mutations that
affect structural cistrons which specify polypeptide chains shall be contrasted with
those which affect cistrons for transfer RNA.
1. Mutations Affecting Structural Cistrons

a) Frame-shift Mutations
A rare type of mutation which most profoundly affects the assigned function of a
structural cistron is a frame shift mutation. This type of mutation is due to either the
deletion or the insertion of a single or two consecutive base pairs. When a messenger
RNA transcribed from an affected cistron is translated, a growing peptide chain
would receive the proper amino acid sequence until the point of deletion or insertion.
However, because of the coding mechanism based on base triplets, from then on to
the carboxyl end, the amino acid sequence would be completely altered, so that very
little homology would remain between the original wild-type peptide chain and a
mutant polypeptide.
Aframe-shift mutation due to the deletion of a singlebasepair has actually been
found at the gene locus for lysozyme of a bacteriophage. A portion of a messenger RNA
22 Mutation and the Conservative Nature of Natural Selection
transcribed by the wild-type lysozyme dstron had the base sequence- AGU.CCA.
UCA.CUU.AAU.- which was translated to - Ser-Pro-Ser-Leu-Asn -. As the
result of a deletion in the cistron, the first A was lost from the corresponding portion
of a mutant tnessenger RNA. Thus, the message now read as- GUC.CAU.CAC.
UUA.- which was translated to -Val-His-His-Leu - (TERZAGHI et al., 1966). It
is of interest to note that so far as the deletion is concerned, the loss of a single or two
consecutive base pairs results in a more drastic consequence than the loss of three
successive bases from a structural dstron. The loss of triplets merely results in the
deletion of a single amino acid from the otherwise intact amino add sequence of the
gene product. The same can be said of an insertion of one or two consecutive base
pairs as compared to the insertion of triplets.
b) Nonsense Mutations
Mutations are more often base Substitutions than insertions or deletions. Of the
base Substitutions, nonsense mutations are the most drastic kind because such muta-
tions result in premature chain termination. It should be remernbered that three of the
64 codons are set aside as the chain terminating codons (Table 2, Chapter II). The
codons UGA, UAG and UAA cannot be recognized by any of the existing transfer
RNA specified by the genome of an organism. Thus, the translation of a messenger
RNA ends at the position occupied by a chain terminating codon. A nonsense mutation
refers to a base substitution which changes an amino add specifying codon to a
chain terminating codon.
Let us imagine a structural cistron designed to specify a polypeptide chain made of
145 amino acid residues with the 20th base triplets of its messenger RNA originally
reading AAG. Thus, the 20th position from the amino end ofthiswild-type poly-
peptide chain is occupied by lysine. When a single base substitution changes this
AAG codon to the nonsense codon UAG, the translation of a mutant messenger RNA
ends at the 19th base triplets, so that a mutant cistron can only specify a polypeptide
chain made of 19 amino acids.
In the case of two closely linked structural genes which tagether transcribe a
single polycistronic messenger, a base substitution which changes a chain terminating
codon to an amino acid specifying codon would also have a dramatic effect. When a
polycistronic messenger RNA is translated, only a single, long polypeptide chain
would be formed, instead of two independent polypeptide chains, each having a
specific function.
c) Missense Mutations
Most often, a base substitution which affects the structural cistron results in an
amino acid Substitution at a specific position in a polypeptide chain. For example, a
change from the codon GUU to GCU replaces valine with alanine. Such mutations
are defined as missense mutations. Certain types of amino acid substitutions affect the
assigned functions of a cistron less drastically than other types. For instance, the
replacement of glycine with alanine, valine with leucine, or phenylalanine with
tyrosine, is of little consequence, since the pairs of amino acids involved in the ex-
changes are of the same kind (Tables 1 and 2, Chapter II). These are defined as
conservative substitutions. In sharp contrast, such replacements as leueine with
arginine or aspartic acid with valine are more drastic. This type of replacement
Mutations Affecting Structural Cistrons 23
changes the net molecular charge of a polypeptide chain, so that a mutant chain
becomes distinguishable from the original wild-type chain by electrophoresis.
The amino acid cysteine is unique in that it has an -SH group. An -SH group is
often essential to the activity of enzymes; e. g., dehydrogenases. Furthermore, -SH
groups of a pair of cysteines are capable of forming a disulfide bridge between them.
Not only do intrachain disulfide bridges determine the molecular shape of a poly-
peptide, but also the formation of a polymerk molecule sometimes depends on the
disulfide bridge formation between two different polypeptide chains. Thus, the
replacement of cysteine with any other amino acid usually causes a drastic alteration
in the functional performance of a polypeptide chain.
The -OH group of serine also serves as an active radical of certain enzymes, and
it can bind with a phosphate group of nucleic acids. Therefore, the replacement of
serine with any other amino acid can also be a quite drastic change. The same can
be said of the replacement of histidine with other amino acids as will become clear in
later chapters.
d) Samesense Mutations
As shown in Table 2 (Chapter TI), all 20 amino acids except methionine and
tryptophan are specified by more than one codon. In the case of glycine, alanine,
valine, threonine and proline, the third base of the codon is completely redundant or
synonymous. For example, as long as the first two bases of the codon for glycine
read GG, the third codon can be any of the four bases. In isoleucine, phenylalanine,
tyrosine, aspartic acid, asparagine, glutamic acid, glutamine, cysteine and histidine,
on the other hand, the third base of the codon is only partially redundant. For example,
as long as the first two bases of the codon for tyrosine read UA, the third base can
be either U or C, but not A or G. As far as these codons are concerned, a substitution
of the third base does not necessarily alter the amino acid sequence of a polypeptide
chain. This type of base substitution is known as a samesense mutation.
On the surface, it may appear that samesense mutations are of no evolutional
significance. The following considerations, however; should make it clear that this
type of mutation can be of some importance. 1. A samesense mutation can serve as an
intermediate step for the missense mutation. For example, if isoleueine which occupies
a certain position of the wild-type polypeptide chain is specified by the codon AUA,
replacement of the isoleueine with phenylalanine (codons UUU and UU C) cannot be
accomplished by a single base substitution. But, if the codon AUA was previously
changed to AUU by a samesense mutation, such amissense mutation becomes possible.
2. The samesense mutation can conceivably affect the rate of translation of a messenger
RN 1\.. The rate at which a messenger RNA is translated to a polypeptide chain must
depend, in part, upon the nurober of transfer RNA molecules which are available for
translation ofthat messenger RNA. As already mentioned, of the four codons specify-
ing alanine, the codons GCU, GCC and GCA can be recognized by the same alanine
transfer RNA having the anticodon CGHyX. The codon GCG, however, has tobe
recognized by a different alanine transfer RNA having the anticodon CGC.
It might be that the anticodon CGHyX-type is a major species and the anticodon
CGC-type is a minor species of alanine transfer RNA, because the cell, at any given
time, contains 100 times more molecules of the formertype than of the latter type.
If this is the case, a samesense mutation which changed the codon GCU to GCG should
result in a marked slow-down of the rate of synthesis for a mutant polypeptide

chain, although the wild-type polypeptide chain and a mutant polypeptide chain
maintain the identical amino acid sequence.
As concluding remarks to this section on mutations of structural cistrons, the
probability of occurrence of various types of base substitutions shall be mentioned.
Since we know the nature and number of codons for each amino acid, and since we
know that only three of the 64 codons are chain terminating codons, we can calculate
that if a base substitution randomly affects any base pair in the cistron, for every
nonsense mutation, there should be 17 missense mutations and 6 samesense mutations.
The relative frequencies of these three types of base substitutions are expected to be
1:17:6 (WmTFmLD et al., 1966). As the method of electrophoresis is increasingly
used in distinguishing allelic products of a given structural gene locus, it should also
be helpful to know that 40% of all possible missense mutations result in changing the
net molecular charge of a polypeptide chain specified by that locus (FITCH, 1966).
Afran;e-shift mutation is due to the deletion or insertion of base pairs. Forthis
reason, there is no way to calculate the expected frequency ofjrame-shift mutations in
relation to the mutations due to base substitutions.
2. Mutations Affecting Transfer RNA Cistrans

The genome of vertebrates contains literally hundreds of thousands of structural
genes; each specifying a unique polypeptide chain with a particular function.
A mutation which affects one of these cistrons results in changing the amino acid
sequence of only one of the hundreds of thousands of polypeptide chains. In sharp
contrast, transjer RNA cistrons are short in variety; an organism requires only 30 or
so different kinds of transfer RNA. However, a mutation affecting any of the small
number of transjer RNA cistrons has a far reaching effect, since the presence ofthat
transfer RNA is required for the translation of nearly every messenger RNA.
a) Suppressor Mutation
In E. coli and in other unicellular organisms on which a detailed study has been
performed, UAG, UAA and UGA serve as the chain terminating codons, probably
because the genome does not contain genes for the three kinds of functional
transfer RNA with the anticodons to match these three codons. What happens if a base
substitution in the DNA cistron which transcribes tyrosine transfer RNA results in
changing the anticodon from AUG to AUC? A mutant tyrosine transfer RNA now
recognizes UAG and adds tyrosine to the growing polypeptide chain. UAG no
Ionger serves as a nonsense codon. A structural cistron which had been suffering from a
nonsense mutation would suddenly be repaired, for the fulllength of a mutant messenger
RNA could again be translated to a polypeptide chain. In E. coli, such a mutation
affecting the tyrosine transfer RNA locus is known as the suppressor 3+ (GAREN,
1968). Such a mutation, however, is a double edged sword. While it apparently
repairs the darnage to a particular structural cistron which was inflicted by a nonsense
mutation, if UAG is also used as the normal chain terminating signal by a poly-
cistronic messenger RNA, in exchange, two adjacent cistrons transcribing this poly-
cistronic messenger have to suffer the consequence of this very same mutation. A
Mutations Affecting Transfer RNA Cistrons 25
single, long polypeptide chain rather than two independent polypeptide chains would
be translated from a polycistronic messenger.
b) Mutations which Result in Ambiguous Coding

Normally, the codon AAG on a messenger RNA is recognized only by the lysine
transfer RNA presumably having the anticodon UUC. What happens if a mutation
in the glutamine transfer RNA locus results in changing its anticodon from GUC to
UUC? The codon AAG, residing on a variety of messenger RNA, is now recognized
not only by the lysine transfer RN.A, but also by a mutant glutamine transfer RNA.
The result is an ambiguous coding in that, if its messenger RNA contains the codon
AAG, a single structural cistron in the genome now produces two or more different
polypeptide chains; one differing from the other only by having glutaminein place
of lysine. Such a mutation which affects the anticodon of a transfer RNA is likely
to cause the ambiguous coding of not one particular messenger RNA, but rather of a
variety of messenger RNA transcribed by a number of different structural genes.
Every horse ( Equus caballus) produces two different kinds of hemoglobin
1X-chains; 1Xf and 1X'. The only difference between 1Xf and 1X' is at position 60; the
form er has glutamine w hile the latter has lysine. I t may appear that the genome of the
horse contains two separate gene loci for 1Xf and 1X', until it is realized that there is
an allelic substitution involving tyrosine and phenylalanine at position 24 (Fig. 6,
Chapter VI). Some horses are homozygous, having either tyrosine or phenylalanine
at position 24, while the heterozygous horses produce two different kinds of 1X-chains
with regard to position 24. A most interesting fact is that every heterozygous horse
producesnot only two kinds of 1Xf, but also two kinds of 1X'. They are :-Phe- Gln-,
- Tyr-Gln - , - Phe-Lys- and -Tyr-Lys- (KILMARTIN and CLEGG, 1967).
One possible explanation of the above finding is that both 1X' and 1X.f are specified
by a single gene locus in the horse genome (haploid set), and that ambiguous coding
of the codon AAG in its messenger RNA is responsible for the placement of either
glutamine or lysine at position 60. Although lysines at other positions such as 7 and
11 of the horse 1X-chain are not replaced by glutamine, this can be explained on the
assumption that these lysines are specified by the other codon (AAA) in the messenger
RNA. If, during speciation, the modern horse has indeed become homozygous to a
mutant glutamine transfer RNA having the anticodon UUC, the codon AAG
contained in every messenger RNA produced by the horse should show ambiguous
coding. The occurrence of either glutamine or lysine should be noted not only in
hemoglobin chains, but also in other polypeptide chains which are constituents of
enzymatic and non-enzymatic proteins. Such a mutation of the tran{fer RNA cistron
which causes widespread havoc among a variety of polypeptide chains is probably
incompatible with the normal development of an organism. An alternative explana-
tion to the findings on the horse hemoglobin 1X-chain shall be affered in a later
chapter.
On the basis of the above discussion on mutations which affect transfer RNA
cistrons, it should be realized that our cherished belief on the universality of codons
is based on one tacitly agreed assumption. One has to assume that natural selection
has been extremely effective in eliminating the types of mutations mentioned above
which must have affected transfer RNA genes time and again throughout the
history ofliving organisms. If a mutational change in the anticodon of a transfer RNA

had been permitted to accompany the process of speciation, the codon AAG which
originally specified lysine might have become a codon for glutamine in certain types
of organisms. Conversely, what would have happened if a mutation in a tyrosine
transfer RNA which resulted in the loss of the ability to bind with a spedfic amino
add at the 3'-end was permitted to perpetuate itself? The codon UAC, which used to
specify tyrosine, would now be serving as a chain terminating nonsense codon in
certain organisms.
References
FrrcH, W. M.: An improved method of testing for evolutional homology. J. Mol. Biol. 16,
9-16 (1966).
KrLMARTIN, J. V., CLEGG, J. B.: Amino-acid replacements in horse hemoglobin. Nature
213, 269-271 (1967).
TERZAGHr, E., KADA, Y., STRElSINGER, G., EMRICH, J., INOUYE, M., TsuGrrA, A.: Change
of a sequence of amino acids in phage T4 lysozyme by acridine induced mutations. Proc.
Nad. Acad. Sei. US 56, 500-507 (1966).
WHrrFIELD, H. J., MARTIN, R. G., AMES, B. N.: Classification of aminotransferase (C gene)
mutants in the histidine operon. J. Mol. Biol. 21, 335-355 (1966).
Chapter V
Forbidden Mutations
Only the amino acid sequence of a polypeptide chain defines its function. A
polypeptide chain having a certain amino acid sequence serves as a subunit of immuno-
globulin, while another polypeptide chain having a different amino add sequence
functions as a subunit of an enzyme, such as lactate dehydrogenase. It then follows
that a change in the amino add sequence of a polypeptide chain can deprive that
polypeptide chain from the performance of its assigned function.
As long as the genome (haploid set) contains only a single structural gene locus
for one vital function, natural selection has not permitted the perpetuation of such
mutations which resulted in the loss of the function assigned to that locus. Since
these mutations have been forbidden to accompany the process of spedation, they
shall be defined as forbidden mutations. For instance, an enzyme, dihydro-orotase,
catalyzes the middle step of de novo synthesis of pyrimidine bases. If a new polypeptide
chain spedfied by a mutant gene no langer functions as dihydro-orotase, it is indeed
a forbidden mutation, for individuals homozygous for this mutation would surely die
without extensive medical care. In fact, all known mutations which cause inherited
diseases of man are forbidden mutations.
1. Forbidden Mutations Affecting Transfer RNA Cistrons

As already mentioned, changes in either the anticodon or the amino add binding
site of a transjer RNA cause widespread havoc, for the effect is manifested through the
amino add sequence of not only one polypeptide chain, but a great variety of poly-
peptide chains.
Forbidden Mutations of Structural Cistrons 27
The indication that all living organisms utilize an identical set of codons to
specify each amino acid suggests that such mutations affecting transjer RNA cistrons
have been forbidden since the time of the first living creature. In addition, the base
sequences in certain segments of transjer RNA have to remain complementary to
those in other parts of the same molecule, so that the characteristic "cloverleaf"
configuration can be maintained ..
It must be that almost any change in the base sequence of the transjer RNA
cistron hinders the performance of the function assigned to its product. Indeed, no
matter what species they are derived from, whether from E. coli or from man, all
transfer RNA have the same characteristics. Because of the effective elimination of
forbidden mutations by natural selection, the base sequence of each transjer RNA
cistron has changed only slightly despite a billion or more years of existence. No
clearer case can be made to point out the extremely conservative nature of natural
selection.
2. Forbidden Mutations of Structural Cistrons

What kind of changes in the base sequence of a structural cistron are forbidden
mutations? The most deleterious types one can think of are jrame-shift mutations and
nonsense mutations. Unless such changes occur very near to the tail end of the structural
cistron so that only the carboxyl end of a polypeptide chain is affected, a modified
polypeptide chain specified by the mutant structural gene is not expected to retain
the ability to perform its assigned function. In short, both frame-shift mutations and
nonsense mutations are almost invariably deleterious, and, therefore, forbidden.
Missense mutations, on the other hand, can be forbidden or tolerated. By and large,
conservative amino acid substitutions, such as an exchange between alanine and
glycine or that between phenylalanine and tyrosine, are not forbidden by natural
selection. But whether or not the particular amino acid substitution can be tolerated by
natural selection depends upon the site at which the substitution occurs. For instance,
an exchange between histidine and tyrosine can be tolerated in certain sites of many
polypeptide chains. But, the same exchange affecting either the 58th or 87th positions
of the mammalian hemoglobin o.:-chain is forbidden. As shown in Fig. 4, in the case
of myoglobin and hemoglobin peptide chains, two histidine residues from the opposite
direction hold a heme. In the case of the mammalian o.:-chain, of the 141 sites, histidine
at the 58th and 87th positions represents the points of attachment to the heme group.
In man, an allelic substitution of one or the other histidine residues by tyrosine
causes an inherited disease; methemoglobinemia. Heroaglobin is oxidized to meth-
emoglobin at all times, and, normally, methemoglobin is easily reduced back to
hemoglobin. However, a tyrosine residue mutationally introduced in place of a
histidine resists reduction, since it forms too stable a complex with the ferric iron
of the heme group (GERALD and ScoTT, 1966). A mutation which causes methemo-
globinemia is apparently incompatible with successful speciation. Accordingly, a
pair ofhistidine residues which represent the points of attachment to the heme group
have been preserved in the hemoglobin as well as the myoglobin chains of all verte-
brates.
The pair of histidines mentioned above serves to introduce the concept of the
active site or sites within a polypeptide chain. Any functional polypeptide chain
contains an active site within. In the case of heme-containing peptide chains, the sites
of attachment to the heme group represent the most critical active sites. They may be
represented by a pair of cysteines, as in Cytochrome C, or by a pair of histidines, as in
hemoglobin and myoglobin. For a peptide chain which is a subunit of an enzyme
molecule, the active site represents the part which recognizes a substrate and binds
with a coenzyme. Lactate dehydrogenase (LDH) is an NAD dependent enzyme
catalyzing the interconversion of lactate and pyruvate, and having a molecular
weight of 135,000. Since it is a tetramerk molecule, each subunit specified by the
LDH cistron is made of340 or so amino acids. When LDH from members of different
classes of vertebrates were compared, wide differences in the amino acid composition
were noted. Yet, the active site of 12 amino acids remained inviolate (KAPLAN, 1965).
Fig. 4. A three dimensional configuration of the myoglobin and hemoglobin polypeptide

chains. Two histidine residues (H) from opposite directions hold a heme
These 12 amino acids are: - Val-Ile-Ser-Gly-Gly-Cys-Asn-Leu-Asp-Thr-Ala-

Arg -. Regardless of the vertebrate species they are derived from, whether from
fish or from man, it is the sequence of these 12 amino acids including the cysteine in
the center which defines a particular polypeptide chain as the LDH subunit. Quite
clearly, any disturbance, including a conservative amino acid exchange in the sequence
of the active site, results in the formation of either a functionless polypeptide chain or a
functionally hindered polypeptide chain. Thus, a mutation affecting the active site
has been forbidden throughout the history of vertebrate evolution; the length of
300 million years.
Two proteolytic enzymes, trypsin and chymotrypsin, shall be given as examples of
polypeptide chains having two separate active sites. The amino acid sequences of
these two active sites have been preserved throughout the history of mammalian
evolution. As schematically illustrated in Fig. 5, they have one active site around a
histidine and the other active site around a serine, which are separated from each
other by a stretch of more than 100 amino acid residues. In the actual three dimensional
configuration which both molecules assume, however, the active histidine and the
Forbidden Mutations of Structural Cistrons 29
a.tive serine are believed to be facing each other. The three dimensional configuration
of these two molecules is determined largely by five or six disulfide bridges formed
between cysteine residues (KEILet al., 1963; KAUFFMAN, 1965). Thus, the sequences of
amino acids which contribute to the three dimensional configuration to be assumed
by a polypeptide chain are also very important in the maintenance by a polypeptide
chain of its assigned function. The replacement of a cysteine involved in a disulfide
bridge formation with another amino acid certainly represents a forbidden mutation.
X-ray diffraction studies have shown that a myoglobin peptide chain and a hemo-
globin polypeptide chain of vertebrates are folded around the heme group in a nearly
identical way. Thus, the three dimensional configuration schematically illustrated in
NH2
I
\
Fig. 5. The two dimensional molecular shapes defined by the positions of disulfide bridges of
chymotrypsinogen (Ieft) and trypsinogen (right) are schematically illustrated. In chymo-
trypsinogen, five disulfide bridges between cysteines link positions 1-122,42-58, 136-201,
168-182 and 191-220. Of these, the 1-122 link is lost from chymotrypsin as the 15 resi-
dues from the amino end are discarded by activation. The 42-58 link forms the histidine
(H) loop, and the loop made by the 191-220 link contains the active serine (S). In trypsinogen,
six disulfide bridges link positions 13-143, 31-47, 115-216, 122-189, 154--168 and
179-203. The 31-47 link forms the histidine (H) loop, and the loop made by the 179-203
link contains the active serine (S)
Fig. 4 applies to either the myoglobin or hemoglobin polypeptide chains of any

vertebrate (PBRUTZ et al., 1960; K:ENnREw et al., 1960). The six bends shown by
these molecules are not caused by disulfide bridges, but by a more subtle affinity
that exists between stretches of amino acid sequences on both sides of each bend. As
far as these stretches of amino acid sequences in a myoglobin or hemoglobin poly-
peptide chain are concerned, only conservative amino acid substitutions have been
permitted to accompany the process of speciation. Drastic amino acid substitutions
have been forbidden.
The Darwinian concept in its original form, as well as the Neo-Darwinian
concept, is so wdl accepted that it is difficult to think of evolution except in terms of
natural selection for desirable characteristics and advantageaus genes. Thus, SIMPSON
(1964) stated that "natural selection is the composer of the genetic messages, and
DNA, RNA, enzymes, and other molecules in the system are successively its mes-
sengers."
The true character of natural selection revealed by discussions presented above is
contrary to the cherished belief of evolutionists. As long as one vital function is
assigned to a single gene locus in the genome, natural selection acts as an extremely
efficient policeman which preserves the base sequence ofthat locus. Natural selection
does not permit the basic character of a gene to change.
3. Forbidden Mutations Favored

Although the types of mutations cited above are by nature deleterious, there are
certain exceptional circumstances where apparently forbidden mutations are favored
by natural selection. Perhaps because of these exceptions, natural selection, which in
reality is extremely conservative, has been misunderstood as an advocator and
mediatot of genetic changes.
The coat color of mammals immensely in.fluences the survival value of a species,
for it can be used as camou.flage or as a warning. The coloring of hair is due to the
presence of eumelanin (black) and pheomelanin (yellow). These two forms of
melanin pigments are polymers of indole 5: 6 quinone, which is derived from the
aromatic amino acid tyrosine. An enzyme, tyrosinase, catalyzes the above reaction,
andin mammals this locus is known as the C-locus or the albino-locus (WoLFE and
CoLhMAN, 1966). A mutation at the C-locus (tyrosinase locus), which results in the
production of ineffective tyrosinase, serves to make the coat color lighter; chinchilla
and hi1llalqya are examples of such mutations, and these mutations were recurrently
utilized by various mammalian species (SEARLE, 1968). As an example, a mutant
tyrosinase which is temperature sensitive is produced by a recessive allele, eh
( hi1llalqya), of this locus. This mutant tyrosinase cannot catalyze the synthesis of
eumelanin as well as pheomelanin at the usual mammalian body temperature (37 C).
but at a slightly lower temperature, it functions as well as the wild-type tyrosinase,
A homozygous mutant (chfd) shows the Himalayan phenotype having a light
colared body and dark extremities. This phenotype is seen among mice and rabbits
as well as cats (Siamese).
The reason for this apparent tolerance of deleterious forbidden mutations at the
C-locus is found in the fact that the C-locus shows no pleiotropic effect. If tyrosinase
specified by the C-locus is also concerned with a more vital function of the body, such
as synthesis of a hormone, epinephrine, such mutations as chinchilla and hi!Jialqya would
never have been permitted to perpetuate themsel ves. This point can be made clear
when a deficient mutation at the phenylalanine hydroxylase locus is contrasted with
that at the C-locus. The deficiency of phenylalanine hydroxylase also results in
diluting the hair color because an excessive amount of phenylalanine inhibits tyrosin-
ase. But, this mutation has been forbidden to accompany the process of speciation,
simply because it causes a serious illness; phenylketonuria (]ERvrs, 1953).
An alkaloid, such as colchicine, is a universal poison of cell division, for it
inhibits the formation of the mitotic spindle. This inhibition is due to the specific
binding between one molecule of colchicine and one dimeric molecule of micro-
tubule protein which is a constituent of the mitotic spindie fibers. The microtubule
protein shall be discussed in greater detail in Part 3. The very fact that microtubule
proteins of all eukaryotes universally maintain an affinity to colchicine reveals that a
particular stretch of amino acid sequence in a microtubule polypeptide chain which
recognizes and binds with colchicine represents the functionally critical part of the
molecule. Therefore, the amino acid sequence of this stretch has been conserved by
References 31
natural selection throughout the evolutional history oE eukaryotes. Y et, a remarkable

exception exists in the Syrian hamster ( Mesocricetus auratus) which shows a marked
resistance to colchicine. This species oE rodents in their native habitat consume
grasses rich in colchicine. Obviously Eor this reason, this species appears to have
become homozygous Eor a mutation which changed the amino acid sequence oE a
Eunctionally vital part oE a microtubule polypeptide chain. It would not be surprising
if the Syrian hamster paid dearly Eor the acquisition oE this resistance. The micro-
tubule protein oE this species may be somewhat delinquent in the perEormance of its
assigned Eunction.
One can conceive of a rare situation where the deleterious effect oE a forbidden
mutation at one gene locus could be cancelled out by an equally forbidden mutation at
another gene locus. Galactosemia is a serious inherited disease oE man, resulting Erom
the inability oE an affected homozygote to convert galactose to glucose. The defect
is in the gene locus Eor an enzyme, galactose-1-P uridyl transEerase (IssELBACHER
et al., 1958). In mammals, the main source of galactose is the lactose in the mother's
milk, and in order to absorb lactose through the intestine, the presence oE another
enzyme (lactase) is required. If the species becomes homozygous deficient for both
lactase and galactose-1-P uridyl transferase simultaneously, members of the species
would no Ionger suffer an ill effect which is due to the accumulation oE galactose in
the body, Eor no galactose would enter the body. Such a pair ofjorbidden mutations
can be tolerated by natural selection only iE the other set of mutations has sub-
stituted galactose with another hexose as the carbohydrate source oE the milk.
Milk oE certain marine mammals, such as the California sea lion ( Zalophus
californianus), contains no trace oflactose and the genome oE this species has apparently
eliminated a Eunctional gene locus Eor lactase. However, there still exist a pair oE gene
loci Eor galactokinase and galactose-1-P uridyl transEerase (MATHAI et al., 1966).
References
GERALD, P. S., Scorr, E. M.: The hereditary methemoglobinemias. In: The metabolic basis of
inherited disease, 2nd ed., pp. 1090-1099. STANBURY, J. B., WYNGAARDEN, J. B., FREDERICK-
SON, D. S. Eds. New York: McGraw-Hill Book Co. 1966.
IssELBACHER, K. ]., ANDERSON, E. P., KuRAHASHI, K., KALCKAR, H. M.: Congenital
galactosemia, a single enzymatic block in galactose metabolism. Science 123, 635-636
(1956).
]ERVrs, G. A.: Phenylpyruvic oligophrenia: Deficiency ofphenylalanine oxydizing system.
Proc. Soc. Exptl. Biol. Med. 82, 514 (1953).
KAPLAN, N. 0.: Evolution of dehydrogenases. In: Evolving genu and proteins, BRYSON, V.
VoGEL, J. H., Eds. New York: Academic Press 1965.
KAUFFMAN, D. L.: The disulphide bridges of trypsin. J. Mol. Biol. 12, 929-932 (1965).
KEIL, B., PRusfK, Z., SoRM, F.: Disulphide bridges and a suggested structure of chymo-
trypsinogen. Biochim. et Biophys. Acta 78, 559-561 (1963).
KENDREW, J. C., DrcKERSON, R. E., STRANDBERG, B. E., HART, R. G., DAVIES, D. R.,
PHILLIPS, D. C., SHORE, U. C.: Structure of myoglobin: A three-dimensional fourier
synthesis at 2 Aresolution obtained by X-ray analysis. Nature 185, 422--427 (1960).
MATHAI, C. K., PrLSON, M. E. Q., BEUTLER, E.: Galactose metabolism in the sea Iion. Proc.
Soc. Exptl. Biol. Med. 123, 4--5 (1966).
PERUTZ, M. F., RossMANN, M. B., CuLLIS, A. F., MmRHEAD, H., WrLL, G., NORTH, A. C. T.:
Structure of hemoglobin: A three dimensional fourier synthesis at 5.5 A resolution,
obtained by X-ray analysis. Nature 185, 416--422 (1960).
SrMPSON, G. G.: Organisms and molecules in evolution. Science 146, 1535-1538 (1964).
SEARLE, A. G.: Comparative genelies of coat colour in mamma!s. London: Logos Press, Ltd.
1968.
WoLFE, H. G., CoLEMAN, D. L.: Pigmentation, In: Biology of the Iabaratory mouse, 2nd ed.
GREEN, E. L., Ed. New York: McGraw-Hill1966.
ChapterVI
Tolerable Mutations
The previous chapter pointed out that the functionally critical parts of a molecule
have not changed much throughout the history of evolution. This is because muta-
tions affecting these parts of a molecule have efficiently been eliminated by natural
selection. Nevertheless, when the amino acid sequence of homologaus polypeptide
chains from diverse species are compared, amino acid substitutions at a varying
nurober of sites are noted. Those mutations which have been permitted to accompany
the successive processes of speciation shall be called tolerable mutations. What is the
nature of these mutations which have been tolerated by natural selection? Some
mutations accompanied the process of speciation merely because they were harmless.
They are then neutral mutations. Others have been chosen actively by natural selection
because they affered definite advantage over their wild-type counterparts. They shall
be defined as favored mc.tations.
1. Neutral Mutations
Apparently, the idea of a neutral mutation is repugnant to most evolutionists.
SrMPSON (1964) stated that "The consensusisthat completely neutral genes or alldes
must be very rare if they exist at all. To an evolutionary biologist, it therefore seems
highly improbable that proteins, supposedly fully determined by genes, should have
non-functional parts, that dormant genes should exist over periods of generations,
or that molecules should change in a regular, but non-adaptive way." The fact is
that, if the process of speciation requires selection for an advantageaus mutant allele
at every gene locus within the genome, evolution becomes a mathematical impro-
bability. Neutralmutations must have occurred time and again, and tbe guite in-
cidental fixation of theseneutral mutations appears to have accompanied the process of
speciation. As shall be pointed out in a later chapter, the creation of a new species
requires intense inbreeding by a relatively small isolated population. Thus, the in-
cidental fixation of neutral mutations should frequently occur.
The only difference between human and gorilla hemoglobin ~X-chains is the
Substitution of aspartic acid in the gorilla for glutamic acid in humans at position 23
(ZucKERKANDL and ScHROEDER, 1961). Because this substitution is between two
aliphatic, dicarboxyl amino acids, a noticeable difference in the kinetic property of the
~X-chains of the two species is not expected. It is almost certain that this represents a
neutral mutation.
Fig. 6 compares the amino acid seguences of the wild-type hemoglobin 1X-chains
of man and the horse. It can be seen that both chains are 141 residues long, and that
the two differ only at 17 of the 141 sites (BRAUNITZER and MATSUDA, 1963). Fifteen
known allelic substitutions, mostly deleterious, are known to uccur in human
1 Asp 10 ~s
Humancx. Val-Leu-Ser~Ala-Asp-Lys-Thr-Asn-Val-Lys-Ala-Ala-Tyr Gl
Horse.cx. V al-Leu-Ser~Ala-Asp-Lys-Thr-Asn-Val-Lys-Ala-Ala-Tyr Ser

30
Asp 20 Asp Gin Gin
Humancx. Lys-Val-Gly~His-Ala-Gly-Glu-Tyr-Gly-Ala-Glu-Ala-Leu-Glu
Horse cx. Lys-Val-Gly-ltHis-Ala-Gly-Glu-Tyr-Gly-Ala-Glu-Ala-Leu-Glu

Phe
40
Humancx. Arg-Met-Phe-LeuMPhe-Pro-Thr-Thr-Lys-Thr-Tyr-Phe-Pro-His-
Horse cx. Arg-Met-Phe-Leu~Phe-Pro-Thr-Thr-Lys-Thr-Tyr-Phe-Pro-His-

Arg
G{y 50 Glu Asp Tyr 60
Humancx. Phe-Asp-Leu.,-Ser-His-Gly-Ser-Ala-Gln-Val-Lys~His:_Gly-Lys-
Horse cx. Phe-Asp-Leu-Ser-His-Gly-Ser-Ala-Gln-Val-Lys@His-Gly-Lys-

Gin
Lys 70
Human cx. Lys-Val-Ala-;'As~Leu-ThrRAla-Val~His-MAsp-Asp-
Horse cx. Lys-Val-Ala-AspLeu-Thr~Ala-Val~HisHAsp-Asp-

80 Tyr 90
Humancx. MPr~Ala-Leu.:.Ser~Leu-Ser-Asp-Leu-His-Ala-His-Lys
Horse cx. ~Pro~Ala-Leu-SerQLeu-Ser-Asp-Leu-His-Ala-His-Lys

100
Humancx. Leu-Arg-Val-Asp-Pro-Val-Asn-Phe-Lys-Leu-Leu-Ser-His-Cys-Leu-
Horse cx. Leu-Arg-Val-Asp-Pro-Val-Asn-Phe-Lys-Leu-Leu-Ser-His-Cys-Leu-
110 Lys 120

Humancx. Leu~Thr-Leu-Ala~His-Leu-Pro~Phe-Tht-Pro-Ala-
Horse cx. Leu~Thr-Leu-Ala~His-Leu-Pro~Phe-Thr-Pro-Ala-

130
Human cx. V al-His-Ala-Ser-Leu-Asp-Lys-Phe-Leu~Ser-Val-Ser-Thr-Val-
Horse cx. Val-His-Ala-Set-Leu-Asp-Lys-Phe-Leu~Ser-Val-Ser-Thr-Val-

140
Human cx. Leu-Thr-Ser-Lys-Lys-Tyr-Arg
Horse cx. Leu-Thr-Ser-Lys-Lys-Tyr-Arg
Fig. 6. The complete amino acid sequence of the human hemoglobin cx.-chain is compared to
that of the horse hemoglobin cx.-chain. Known allelic Substitutions of the human cx.-chain are
indicated by italics above the wild-type sequence. One allelic and the other non-alleHe
substitutions of the horse are indicated by italics below the wild-type sequence
populations. These mutational Substitutions in man are also listed in Fig. 6. It is

clear that no forbidden mutations, which are represented as deleterious allelic substitu-
tions in man, accompanied the successive processes of speciation which separated man
from the horse. No difference can be detected between man and the horse at the
functionally critical sites, such as histidine at the 58th and 87th positions. Furthermore,
many of the 17 differences that separate two species represent conservative sub-
stitutions; an example being an exchange between glycine and alanine occurring at
positions 19, 57, 65 and 71. They may be regarded as representing neutral mutations
which have been accumulated since the common ancestors of both species started to
follow separate paths of evolution nearly 70 million years ago.
At the functionally non-critical sites of a polypeptide chain, even the drastic type
of amino acid exchanges may represent neutral mutations. Such examples can be
found in fibrinopeptide A and B which are the amino-end portians of vertebrate
fibrinogen molecules. In the clotting of blood, these portions of fibrinogen molecules
are proteolytically removed by the trypsin-like action of thrombin. Their removal
permits spontaneaus polymerization of the resulting fibrin molecule to form an
insoluble fibrinogel. Since the function of fibrinopeptides is non-specific, many
missense mutations as weil as deletion of triplets affecting these parts of the cistron
have apparently been ignored by natural selection and accompanied the process of
speciation as neutral mutations. Although every mammalian fibrinopeptide has
arginine for the carboxyl terminal because it has been split off by the trypsin-like
action of thrombin, fibrinopeptide A can be only 17 amino acid residues long as in
man, or 19 residues long as in sheep and goats, and fibrinopeptide B can be only 15
residues long as in man, or 21 residues long as in reindeer. The amino acid sequence of
fibrinopeptides of two rather closely related species can show a remarkable difference.
For example, fibrinopeptide A of the horse and the donkey differ by 2 substitutions
and 2 insertions as shown below (BLOMBACK et al., 1965):
Hor.re:
Donkey:
The comparison between the above mentioned situation found in fibrinopeptides

and that found in histone IV helps us understand the nature of natural selection. As
already mentioned, the role of histones is to bind with phosphate groups of DNA
through free amino groups of its Iysine and arginine residues as weil as through
-OH groups of its serine residues. This role requires the maintenance of a fixed amino
acid sequence. It appears that histone IV of both cattle and the garden pea are 101
amino acid residues long and the two differ by only two amino acid substitutions
(DELANGE and FAMBROUGH, 1968). This conservation is quite remarkable in view
of the evidence that animals and plants appear to have followed separate paths of
evolution for at least one billion years. Indeed, natural selection permits the accumula-
tion of mutations only if they affect functionally non-erideal sites of the cistron. It
then follows that many of the mutations which accompanied the process of speciation
have been ignored by natural selection merely because they represented neutral
mutations.
Of the different types of mutations, samesense mutations, as a group, have to be
regarded as the most neutral of all, since the accumulation of samesense mutations by
Favored Mutations 35
structural cistrons does not result in alteration of the amino acid sequences of their
products. The Treffers mutator (mut T) gene of E. coli is believed to specify a
defective DNA polymerase which is prone to place mismatched bases during DNA
replication, so that many structural genes of the "mut T" strain are affected by base
substitutions. Cox and Y ANOFSKY (1967) selected the "mut T" strain for continued
viability by repeated subcultures, and observed a trend of change toward DNA of
higher guanine-cytosine content. Numerous substitutions of A and T in DNA with
C and G were compatible with continued viability of E. coli, since most of these
changes affected the synonymaus third bases of redundant codons (samesense muta-
tions). Fot example, a change from CGA to CGG in DNA changes the codon GCU to
GCC, but both codons specify the same amino acid (alanine).
In the evolutional time scale, the rat and the mause are close relatives. Yet,
WALKER (1968) estimated that as much as 13% of the nucleotide positions are
occupied by different bases in the DNA's of the rat and the mause. This appears tobe
a much greater difference than that observed on the amino acid sequences of homo-
logous polypeptide chains of two spedes. The conclusion to be draw is that most of
the evolutionary changes in DNA represent samesense mutations.
2. F avored Mutations
As lang as one vital function is assigned to a single gene locus within the genome,
natural selection does not permit the perpetuation of mutations which change the
fundamental character ofthat locus. Thus, the dihydro-orotase locus would forever
remain the dihydro-orotase locus until the extermination of allliving organisms on
this earth. On the other band, if all tolerable mutations were effectively neutral, there
would have been no evolution.
Taking the enzyme locus as an example, it is granted that tolerable mutations
which affected the functionally less critical sites of the enzyme polypeptide do not
alter the substrate spedfidty and other basic characteristics of the enzyme, but some of
these mutations do change the kinetic property ofthat enzyme with regard to its pH,
temperature optimum and Km (Michaelis constant). As far as thesetolerable mutations
are concerned, natural selection is afforded with an opportunity to actively favor a
particular mutant allele of that locus, if this alleHe product best fits the requirement
imposed by a particular environment in which a population of the organism is
placed.
Let us imagine a spedes of fish which inhabits a main body of water where there
is wide fluctuation of temperature between day and night as weil as summer and
winter. At a nurober of enzyme Iod, this spedes has become polymorphic; maintain-
ing multiple alleles which specify variant enzymes with different temperature optima.
Subsequently, a small nurober of them migrated into a small body of watet where
temperature is always warm because it is fed by a bot spring. In this subpopulation,
natural selection no doubt favors a particular allele which spedfies a variant with a
high temperature optimum. Eventually, a subpopulation becomes homozygous for
this type of allele at a nurober of enzyme Iod and emerges as a new spedes. This, then,
is the process of spedation by spedalization. Here, it should be realized that the
creation of a spedalized spedes by selection for Javared mutations at the already
existing gene Iod leads to an evaludanal cul-de-sac. The new fish spedes mentioned
3*
36 Mutation and the Conservative Nature ofNatural Selection
above would be exterminated as soon as the hot spring which fed that small body of
water dried up.
More often, certain mutant alleles are favored by natural selection because they
offer heterozygous advantages to individuals within a population. Although members
of the same species share many common characteristics, with the exception of
monozygotic twins, no two members of a randomly breeding population are identical.
Such individual diversity is due to allelic differences at a number of gene loci. With
regard to the gene locus for the ABO erythrocyte antigen system of man, some of
us type as A, others as either 0 or B, and yet others as AB. The traditional thought on
individual diversity has been that a population maintains multiple alleles at a given
gene locus only if it is advantageaus to be heterozygous at that gene locus. While this
line of thought can be challenged in many instances where the existing multiple
alleles differ from each other merely by neutral mutations, there have been well
proven examples of the heterozygous advantage.
The mutant human hemoglobin '-chain differs from the wild-type -chain by a
single substitution; glutamic acid at the 6th position of the normal -chain is replaced
by valinein the '-chain (PAULING et al., 1949; INGRAM, 1956). tx2 2 molecules within
the erythrocyte tend to pile upon each other, and this tendency causes erythrocytes
to assume a sickle-shape. It is believed that this apparently deleterious mutant allele
for an abnormal '-chain persists in Mrican populations with a remarkably high
frequency (as much as 40% in some areas), because relative resistance to falciparum
malaria is passed on to /'-heterozygotes (ALLISON, 1954). Yet, as long as the normal
-chain and the abnormal '-chain genes exist as two allelic alternatives of the same
gene locus, the production of desirable heterozygotes is invariably accompanied by
the production of deleterious '/'-homozygotes suffering from severe sickle-cell
anemia. At the most, only 50% of a population can enjoy this heterozygous advantage,
while 25% of the same population would be afflicted with sickle-cell anemia and
would die without extensive medical care. Thus, a mutant allele which is favored by
natural selection because of the heterozygous advantage can never be the cause of
speciation.
Accordingly, a frustrating situation may develop with regard to the hetero-
zygous advantage. The result isthat allelic polymorphism often persists in transcend-
ing the process of speciation. For example, Gm(a+) and Gm(a-) are allelic alter-
natives at the gene locus for y Gl-class of immunoglobulin heavy-chains in man.
This allelic difference is detected by the use of proper antisera. Man also maintains
alleles Gm(b +) and Gm(b-) at the other gene locus for y G3-class of heavy-chains
(KUNKEL et al., 1964). It is of extreme interest to note that with regard to these two
closely linked gene loci on the same chromosome, apparently the same allelic alterna-
tives are also maintained by the chimpanzee (BoYER and YoUNG, 1961). Although
the heavy-chain specified by the y Gl-locus of man is expected to differ by several
amino acid substitutions from that specified by the corresponding gene locus of the
chimpanzee, and the samc is expected to hold true for the y G3-locus, the fact remains
that one of these differences seen in each dass of heavy-chains is an allelic difference
rather than a species difference. At the fixed site or sites on the y Gl-class of heavy-
chains, the Gm(a +) polypeptide of man and the chimpanzee must share the same
amino acids. Similarly, the corresponding site or sites on the Gm(a-) polypeptide,
regardless of whether it is derived from man or the chimpanzee, must be occupied by
Convergent Evolution and Recurrent Mutations 37
partieular amino acids which are different from those found in the Gm(a +) peptide
chain. It appears that allelie polymorphism involving Gm(a+) and (a-) as well as
Gm(b +) and (b-) was already present in the common ancestor to both man and the
chimpanzee (Dt:_yopithecus of the Pliocene epoch, which existed 7 or 8 million years
ago). These allelic alternatives have apparently been maintained by both man and the
chimpanzee, despite the separate routes of evolution they subsequently followed.
The X-chromosome of man carries a gene locus which specifies the antigen on
erythrocyte surface. The allelic polymorphism of this locus can be ascertained by the
use of anti X~ sera. Erythrocytes of a person carrying an Xg (a +) allele can be
agglutinated by this antiserum, while the red cells of a person homozygous or
hemizygous for an Xg (a-) allele cannot (MANN et al., 1962). Gibbons ( H_ylobates
lar lar) apparently maintain essentially the same allelic alternatives at the correspond-
ing X-linked gene locus, for they type either as Xg (a+) or as Xg (a-) in the same
manneras humans (GAVIN et al., 1964). Allelic polymorphism can indeed transcend a
series of successive speciations, as man and gibbons shared a common ancestor in the
rather remote past (about 26 million years ago).
While Javared mutations which offer a heterozygous advantage can and do
accompany the process of speciation, such mutations cannot be the cause of speciation,
for the same reason that neutral mutations which accompany the process cannot be
the cause of speciation. Only those javared mutations which, under a given circum-
stance, offer selective advantage in the homozygous state can be the cause of specia-
tion.
3. Convergent Evolution and Recurrent Mutations
Convergence has long been regarded as the bane of students of fossils. It involves
resemblances whieh are sometimes extensive and detailed, although they are not
evidence of propinquity of ancestry. For instance, an iehthyosaur (an extinct marine
reptile of the Jurassie period) was remarkably similar in shape to a porpoise (a
living marine mammal of today).
The Triassie period marked the first time during the evolution of vertebrates
that truly land-living tetrapods turned in any appreciable nurober to a life in the sea.
At least three different lines of reptiles began to evolve toward the aquatic life in
early Triassie times. The iehthyosaurs, in many respects the most highly specialized of
the marine reptiles, appeared suddenly and dramatically in middle Triassie times, and
during the Jurassie period (130 to 165 million years ago) attained a body shape
remarkably similar to living porpoises. They finally became extinct before the dawn
of the Cenozoie era (the age of mammals).
Of all the placental mammals, whales and porpoises, or cetaceans, are certainly
the most atypieal. The first cetaceans were !arge whales which appeared in the middle
Eocene epoch. During the late Oligocene times (approximately 30 million years ago),
there appeared some small odontocetes (toothed whales) which were ancestral to the
modern porpoises and dolphins. Thus, extinct iehthyosaurs and living porpoises are
separated by at least 30 million years, yet, their resemblance goes beyond their fish-
like body shapes.
Ichthyosaurs were ovoviviparous. They hatched eggs inside their body and bore
living young. Some fossils of Ichthyosaurus from Germany showed unborn embryos
within the body cavity of the adult, and one specimen was found in the act of giving
birth when death overtook the mother. The young were being born tail first; the
manner of delivery used by existing marine mammals.
On the surface, such examples of convergent evolution may appear as a great
puzzle. On the genetic term, however, they are not the bane; rather they are very
instructive in understanding the extremely conservative nature of natural selection.
They also show us that at a homologaus gene locus natural selection may independ-
ently favor similar types of tolerable mutations to cope with particular demands im-
posed by similar environments.
Although adult body shapes of divergent vertebrate species tend tobe remarkably
different, the basic body design as revealed by the process of morphogenesis during
embryonie development has remained essentially the same (ontogeny recapitulates
phylogeny). This is in keeping with the already discussed fact of evolution that
natural selection by and large has conserved base sequences at the functionally
critical sites of each structural cistron. Thus, divergent vertebrate species do maintain
homologaus gene loci. As long as there are homologaus gene loci, there can be
recurrence of homologaus mutations at each of these gene loci; hence, convergent
evolution.
Among mammals, the best examples of recurrent mutations can be found at those
gene loci concerned with coat color. The coat color patterns are determined by the
concerted action of a number of independent gene loci; one specifying an enzyme
(tyrosinase) which catalyzes the synthesis of melanin pigments from an aromatic,
amino acid (tyrosine), others controlling the distribution pattern of eumelanin and
pheomelanin among individual hairs, and yet others controlling the migration of
melanoblasts from neural crests to hair follicle primordia during embryonie develop-
ment. Many examples of convergent evolution due to the recurrence of homologaus
mutations at one of these gene loci can be found in mammals. For example, at the
tyrosinase loeus (C-locus), a temperature sensitive mutant allele (eh) whieh gives the
Himalayan phenotype has recurred in the mouse, the rabbit and the cat as already
mentioned (SEARLE, 1968). The amino acid compositions of the wild-type tyrosin-
ases of these three species must be considerably different, yet natural seleetion must
have conserved nearly identical amino acid sequences at the functionally critical parts
of these polypeptide ehains. Because of this conservation, similar amino acid sub-
stitutions mutationally introduced at a corresponding functionally critical site can
change the tyrosinase of each of these three species to a temperature sensitive mutant.
4. Atavism and Revertant Mutations

Although genes do change the appearance of individuals, such obvious con-
sequences of evolution are no doubt determined in a precise but very indirect way by
interaction between the structural gene products which are invariably polypeptide
chains. Unforrunately, none of the gene products which ultimately determine the
morphological appearance have been identified to this date, probably beeause such
gene products are recognizable only during embryonie morphogenesis. Frustrated by
our own ignorance, it has often been assumed that those genes which determine
morphological appearance must have a mystical quality which is not shared by
known structural genes specifying functionally well-defined polypeptide chains.
Atavism and Revertant Mutations 39
At tbis point, it is rather instructive to ask a simple question - How many

independent gene loci underwent mutational changes to accomplish a known drastic
evolutional change in morphological appearance? A partial answer to the above
question can be found in atavistic or "throw back" mutations wbich have been
observed in a variety of vertebrate species.
The Jurassie bird, Archaeopteryx lithographia, was little more than an archosaurian
reptile. The winged forelimbs ended in a hand wbich was composed of the clawed
first three digits. Modern birds, with certain exceptions such as hoatzims (Opistho-
comus) of South America, have eliminated tbis grasping hand from the winged
forelimb. Y et, it takes a mere dominant allelic mutation at a single autosomally
inherited gene locus to put back these clawed digits on the wing of a domestic
cbicken (CoLE, 1967).
Man (Homo sapiens) is said to be a naked ape (MoRRIS, 1967). Indeed, the hairless-
ness of the body is a mostprominent characteristic of man wbich sets bim apart from
other mammals. Y et, it appears that all it takes for a "naked ape" to regain a full set
of fur is a dominant allelic mutation at a single gene locus. In Vienna, there exists a
family portrait painted by HoFNAGEL (an early 17th century artist). The father of
tbis family, Petrus Gonzalus, was born on the island of Teneriffa (in the Canary
Islands group). His entire face, as weil as bis hands, were covered with hair. While bis
wife was a normally hairless person, both daughters resembled their father in tbis
aspect.
It is possible that each of the two landmark morphological changes mentioned
above were accomplished merely by a forbidden mutation to a null allele at a single
gene locus. Even if the hairlessness of our body is due to the homozygous state for a
null allele at a single gene locus, the hairlessness would remain a stable characteristic
of Homo sapiens, because revertant mutations are the rarest of all mutations. Frame-
shift as weil as nonsense mutations affecting almost any base pair in the cistron change
the functional wild-type allele to a functionless null allele. But the reversion of a null
allele to a functional allele requires precise repair of the affected base pair. For
example, if the mutation to a nullallelewas due to a deletion of the 101st basepair
of the gene, only an insertion of a basepair precisely at the 101st position can serve as
a revertant mutation.
Nevertheless, the bigher primates appear to have regained trichromatic vision
by atavistic mutations wbich are revertant in nature. While many fish, ampbibians,
reptiles as well as birds are equipped with excellent trichromatic color vision (WARNER,
1931; WoJTUSIAK, 1933; HAMILTON and CoLEMAN, 1933), numerous experiments on
color vision of mammals other than bigher primates have indicated that, in tbis
entire group of animals, color vision is very rudimentary at best (PARSONS, 1924).
Protoinsectivores, a common ancestor of placental mammals probably bad nocturnal
habits, and for tbis reason, they might have become homozygous for null alleles at a
pair of gene loci wbich controlled color vision. Tbis accounts for the widespread
occurrence of achromatic (black and wbite) vision among mammals of today.
Of the primates, lower species belanging to the suborder Lemuroidea apparently
maintain achromatic vision (BrnRENS DE HAAN and FRIMA, 1930), while GRETHBR
(1939) found that the ringtail monkey representing the moreprimitive New World
monkey ( P laryrrhina) possessed dichromatic vision of the protanopic type (poor
discrimination in the red and yellow region). The more advanced Old World monkeys
(Catarrhina}, such as baboons and rhesus, on the other hand, were endowed with
trichromatic vision comparable to that of normal man. Indeed, it appears that higher
primates regained trichromatic color vision by revertant mutations which restored
function to null alleles.
References
ALLISON, A. C.: The distribution of the siekle-eeil trait in East Africa and elsewhere, and its
apparent relationship to the incidence of subtertian malaria. Trans. Roy. Soc. Trop. Med.
Hyg. 48, 312-318 (1954).
BrERENS DE HAAN, J. A., FRIMA, M. J.: Versuche ber den Farbensinn der Lemuren. Z.
vergleich. Physiol. 12, 603-631 (1930).
BLOMBAGK, B., BLOMBAGK, M., GRONDAHL, N. ]., GuTHRIE, C., HrNToN, M.: Studies on
fibrinopeptides from primates. Acta Chem. Scand. 19, 1789-1791 (1965).
BOYER, S. H., YouNG, W. ]. :Gammaglobulin (Gm group) heterogeniety in chimpanzees.
Science 133, 583-584 (1961).
BRAUNITZER, G., MATSUDA, G.: Primary structure of the a-chain from horse hemoglobin
]. Biochem. (Tokyo) 53, 262-263 (1963).
CoLE, R. K.: Ametapodia, a dominant mutation in the fowl. J. Heredity 58, 141-146 (1967).
Cox, E. C., Y ANOFSKY, C.: Altered base rados in the DNA of an Escherichia coli mutator
strain. Proc. Natl. Acad. Sei. US 58, 1895-1902 (1967).
DE LANGE, R. ]., FAMBROUGH, D. M.: Idendeal COOH-terminal sequences of an arginine-
rieb histone from calf and pea. Federation Proc. 27, 392 (1968).
GAVIN, ]., NoADES, ]., TrPPETT, P., SANGER, R., RAGE, R. R.: Blood group antigen Xg in
Gibbons. Nature 204, 1322-1323 (1964).
GRETHER, W. F.: Color vision and color blindness in monkeys. Comp. Psychol. Mono-
graphs 15, 76 (1939).
HAMILTON, W. F., CoLEMAN, T. B.: Trichromade vision in the pigeon as illustrated by the
spectral hue discrimination curve. J. Comp. Psychol. 15, 183-191 (1933).
INGRAM, V. M.: A specific chemical difference between the globins of normal human and
siekle-eeil anemia hemoglobin. Nature 178, 792-794 (1956).
KuNKEL, H. G., ALLEN, J. C., GREY, H. M.: Genedc characters and the polypeptide chains
of various types of gamma-globulin. Cold Spring Rarbor Symposia Quant. Biol. 29,
443-447 (1964).
MANN, ]. D., CAHAN, A., GELB, A. G., FrsHER, N., HAMPER, ]., TrPPETT, P., SANGER, R.,
RAGE, R. R.: A sex-linked blood group. Lancet 1962 I, 8-10.
MoRRIS, D.: The naked ape. London: The Trinity Press 1967.
P ARSONS, J. H.: An introduction to the study of color vision. Cambridge (England): Cam-
bridge University Press 1924.
P AULING, L., lTANO, H. A., SINGER, S. J., WELLS, I. C.: Sickle ceil anemia: A molecular
disease. Science 109, 443 (1949).
SEARLE, A. G.: Comparative genedes of coat colour in mammals. London: Logos Press,
Ltd. 1968.
SrMPSON, G. G.: Organisms and molecules in evolution. Science 146, 1535-1538 (1964).
WALKER, P. M. B.: How different are the DNA's from related animals? Nature 219, 228
(1968).
WARNER, L. H.: The problern of color vision in fishes. Quart. Rev. Biol. 6, 329-348 (1931).
WoJTUSIAK, R. J.: ber den Farbensinn der Schildkrten. Z. vergleich. Physiol. 18, 393--436
(1933).
ZuGKERKANDL, E., SGHROEDER, W. A.: Amino acid composition of the polypeptide chains
of gorilla hemoglobin, Nature 192, 984--985 (1961).
The Absence of a Close Linkage Requirement for Functionally Related Genes 41
ChapterVII
The Conservative Nature of Chromosomal Evolution

The foregoing discussions revealed the extremely conservative nature of natural
selection, that the important parts of the gene have never changed in evolution, and
that natural selection permits only trifling changes.
Viewed in this light, the fact that among placental mammals the diploid chromo-
some nurober ranges from a high of 84 in the black rhinoceros (HUNGERFORD and
SNYDER, 1967) to a low of 17 in two species of rodents (MATTHEY, 1953; HNO et al.,
1963) may be taken as evidence that chromosomal changes have not been subjected
to a strict surveillance by natural selection. Therefore, random reassortment of
chromosomes has safely accompanied a series of successive speciations. These
chromosomal changes, however, are more apparent than real. As will be shown, the
original linkage groups appear to have been conserved to a considerable extent
despite apparent changes in the karyotype (the morphological appearance of di-
ploid complements). This conservation is not due to the necessity of functionally
related genes remaining in close linkage, but merely due to the fact that heterozygous
carriers of most types of chromosomal rearrangements suffer from semisterility.
Because of this semisterility, many types of chromosomal changes cannot be fixedas a
new species characteristic.
1. The Absence of a Close Linkage Requirement for Functionally Related Genes

In prokaryotes, the genes specifying a series of catalytic enzymes of the same
metabolic pathway are often dustered and coordinated as a group, because they
transcribe a single polycistronic messenger RNA. Such a group is defined as an operon
(]AcOB and MoNOD, 1961). The best known example is the /ac-operon of E. coli
which contains the genes for transacetylase, permease and -galactosidase. These
enzymes are used for the metabolism of lactose.
In the case of vertebrates, the coordination of activities of functionally related
genes does not appear to depend upon their close linkage. In all known cases, two or
more genes which specify a series of enzymes for the same metabolic pathway have
been found to be unlinked (carried by different chromosomes). For example, glucose-6-
phosphate dehydrogenase (G-6-PD) and 6-phosphogluconate dehydrogenase
(6-PGD) catalyze two successive steps of the pentose phosphate shunt of carbo-
hydrate metabolism. Yet, in mammals, the former is X-linked (CHILDS et al., 1958;
KIRKMAN and HENDRICKSON, 1963; TRUJILLO et al., 1965; HNO etal., 1965; MATHAI
et al., 1966), while the latter is autosomally inherited (PARR, 1966; SHAW, 1966;
THULINE et al., 1967). The reason for this non-linkage of functionally related genes in
vertebrates can be found in gene duplication (tobe discussed in Part 3). As a result of
gene duplication in vertebrates, each of the enzymes concerned with the same
metabolic pathway is more often than not specified by two or more isozyme genes. In
such a situation as that where isocitrate dehydrogenase of the liver and heart is speci-
fied by two separate gene loci (isozyme genes), while the next step enzyme (oxalo-
succinate dehydrogenase) of both organs is specified by the same gene locus, the
coordination of activities of these two kinds of enzymes in the liver as weil as in the
heart has to depend on means other than gene clustering. In fact, isozyme genes
themselves do not appear to be closely linked to each other. ln man, three unlinked
gene loci specify the isozyme for phosphoglucomutase (HARRIS et al., 1967).
This non-linkage can be extended to a pair of genes which specify a single poly-
merized molecule. Two ~X-chains and two -chains make up a single hemoglobin
molecule which is ~X2 2 Y et, the ~X-chain gene is not linked to the -chain gene in man
(CEPPELLINI, 1959). Similarly, two light-chains and two heavy-chains make up a
single 7S immunoglobulin molecule. Yet, not only in man, but also in rabbits, the
gene loci for light-chains are located on one chromosome, and a group of gene loci
for heavy-chains are located on another chromosome (KuNKEL et al., 1964; OumN,
1966).
Paradoxically, when functionally related genes are dustered tagether in the verte-
brate genome, their functional activities are not coordinated. For example, the genes
which specify various classes of immunoglobulin heavy-chains are very closely
linked to each other in man (NATVIG et al., 1967) and the mouse (HERZENBERG et al.,
1967), yet, each clone of plasma cells makes use of only one particular heavy-chain
locus (PoTTBR and LIEBBRMAN, 1967; CoHN, 1967).
Inasmuch as vertebrates are perfectly capable of coordinating the activities of
unlinked genes, there is no a priori restriction against random reassortment of the
linkage groups. The restriction is purely of a mechanical nature.
2. Inversion as an Internal Rearrangement

A chromosome fragment produced by two breaks within the same chromosome
can be reinserted after making a 180 o turn; the result is an inversion. An inversion
does not change the gross appearance of a metaphase chromosome unless an inverted
segment contains a centromere (pericentric im ersion). By a pericentric inversion, a
two-armed chromosome (metacentric or subterminal) can change to a one-armed
chromosome (acrocentric) and vice versa.
Inversions are very useful means of creating a sterility barriet between a new
species and an old species. Accordingly, this type of chromosomal change often
accompanied the process of speciation. Without the effective sterility barrier, the
integrity of a newly emerged species is threatened when it again comes in contact
with the parental species. When two homologues which differ by an inversion pair
during meiosis of interspecific hybrids, a crossing-over between an original and an
inverted segment produces a dicentric chromosome and an acentric fragment, thus
effectively reducing the fertility of interspecific hybrids.
Deer mice of the genus Peron;yscus belong to the hamster subfamily Cricetinae,
rather than to the mouse subfamily Murinae. Numerous species and subspecies of
Peromyscus which inhabit the North American continent uniformly possess the
diploid chromosome nurober of 48, but the ratio of two-armed chromosomes to one-
armed chromosomes is markedly different. For example, in the cactus mouse
( P. eremicus), all48 chromosomes are two-armed, while in the bush mouse ( P. bqylii),
40 of the 48 are one-armed (Hsu and ARRIGHI, 1966). There is little doubt that in
this genus the visible chromosomal changes which accompanied speciation were
exclusively pericentric inversions.
It should be realized that while an inversion disturbs an internallinkage relation-
ship between genes at two points, the entire linkage group remains intact.
Robertsonian Fusion: The Creation of one Metacentric 43
3. Robertsonian Fusion: The Creation of One Metacentric

by Fusion of Two Acrocentrics
While inversion is an internal change, the exchange of chromosomal material
can take place between two non-homologaus chromosomes. Such interchanges are
known as translocations. Of the various types of translocations, a particular type
known as a Robertsonian fusion has most often accompanied the process of specia-
tion. This is essentially the creation of one metacentric chromosome by the centric
fusion of two acrocentrics (RoBERTSON, 1916).
The situation found in the family Bovidae illustrates the great contribution
Robertsonian fusions have made to the evolutional changes of karyotypes. Among
members of the family Bovidae, the highest diploid chromosome number is 60 as in
goats, cattle and bison, the intermediate number of 54 is represented in Congo
buffalo and sheep, and the lowest diploid chromosome number of 48 is found in the
musk ox. In the species with 60 chromosomes, all 58 autosomes are acrocentrics,
while in the musk ox, 12 metacentrics and 34 acrocentrics make up 46 autosomes
(HECKet al., 1968). If each arm of a metacentric is counted as one, every member of
this family thus far studied has 58 autosome arms. Quite clearly, Robertsonian
fusion has been the exclusive contributor to the visible karyological evolution of the
family Bovidae.
In germ cells of a heterozygote for a Robertsonian fusion, one metacentric
pairs with two acrocentrics, and, at the end of 1st meiosis, the metacentric moves
toward one divisionpole and the two acrocentrics to the other pole. No unbalanced
gametes are produced, and, consequently, heterozygotes do not suffer from semi-
sterility. This appears tobe the reason that natural selection has permitted this parti-
cular type of chromosomal interchange to accompany the process of speciation.
Robertsonian fusions can be regarded as the equivalent of neutral mutations.
The tolerance of natural selection to Robertsonian fusions can be illustrated by the
following two examples:
1. In domestic cattle (Bos taurus, 2n = 60), all 58 autosomes are normally
acrocentrics as already mentioned, but a single Robertsonian fusion which occurred
to one stud bull of the SRB breed of Sweden has been spread far and wide within
this breed. In fact, four stud bulls descended from the first were found to be homo-
zygous for this fusion having 58 chromosomes including a homologaus pair of
metacentric autosomes (GusTAVSSON, 1966).
2. In the Poschiavo V alley in Switzerland, there exists a small population of the
tobacco mouse (Mus poschiavinus, 2n = 26) which is distinguishable from the ordinary
house mouse (Mus musculus, 2n = 40) by its karyotype. All 40 chromosomes of
Mus musculus are acrocentrics, while Mus poschiavinus with 26 chromosomes is homo-
zygous for 7 Robertsonian fusions as illustrated in Fig. 7 (Plateii); 14 metacentdes and
12 acrocentrics make up the diploid complement (GROPP and VON LEHMANN, 1969).
However, at a number of coat color as well as enzyme loci, we found M. poschiavinus
to possess known alleles of M. musculus. A series of remarkable chromosomal
changes apparently have occurred in a rather short span of time, so that genetically
the tobacco mouse is still a member of the ordinary house mouse species.
4. The Creation of a Sterility Barrier by Chromosomal Changes
As shall be made clear in Chapter IX, isolation is a conditio sine qua non of speciation.
The geographical isolation however, does not last forever; sooner or later, a newly
created species again comes in contact with its parental species. In the absence of
geographical isolation, other means are needed to preserve the integrity of a newly
arisen species. Otherwise, a new species which is no doubt a minority would be
absorbed by the majority as a result of interbreeding. Differences that developed
during a period of geographical isolation with regard to body odor, courting pattern
and other aspects may result in the loss of sexual attraction between a minority and
majority species. This, then, is the behavioral isolation mechanism. In the presence of
sexual promiscuity, a chromosomal difference can serve as a very effective reproduc-
tive isolation mechanism by creating a sterility barrier. For instance, the horse
(Equus caballus, 2n = 64) and the donkey (Equus asinus, 2n = 62) mate with each
other without compunction, yet the integrity of either species is not endangered
simply because Fcinterspecific hybrids (mules and hinnies, 2n = 63) are uniformly
sterile.
In fact, the creation of a sterility barrier appears to be the only obvious contribu-
tion gross chromosomal rearrangements have made to evolution. Robertsonian
fusions often accompanied the process of speciation simply because heterozygous
carriers of this type of interchange do not suffer from semisterility. For this very
reason, Robertsonian fusion contributed very little to the creation of sterility barriers.
If a new species becomes homozygous for a reciprocal translocation, on the
other hand, Fchybrids between a new species and a parent species would suffer a 50%
reduction in fertility, thereby creating an effective sterility barrier. For this reason,
however, reciprocal translocation played only a minor role in evolutional change of
the diploid complement. When any chromosomal interchange first occurs to an
individual, it invariably occurs in the heterozygous state. For an interchange to
accompany the process of speciation, the heterozygous faunder has to leave enough
offspring, some of which are again heterozygous, to initiate the first group of homo-
zygotes (a firsthomozygote is produced only from a mating between heterozygotes).
A reciprocal translocation exchanges broken halves of two nonhomologaus chromo-
somes. When germ cells of a heterozygote enter meiosis, a quadrivalent is formed
between two chromosomes involved in an interchange and their intact homologues.
As a result, four types of gametes are produced. Of these, two are grossly unbalanced,
for they are simultaneously deficient for one chromosomal segment and duplicated
for the other. Thus, only 50% of the gametes produced by a heterozygous founder
give rise to viable zygotes, and only half of these are again heterozygous for a trans-
location. There is slight chance that a heterozygote suffering from semisterility can
serve as the faunder of a new species.
It is inversion which appears to have played a very significant role in creating a
sterility barrier. The genetic constitution of individuals appears to determine the
frequency of meiotic crossing-over at a given chromosomal segment. Thus, a
population can talerate inversions of the segments where crossing over rarely
occurs, and a new species may become homozygous for a nurober of inversions
during the period of geographical isolation, as evidenced in the already mentioned
deer mice of the genus Peromyscus. In hybrids between a new species and a parent
Conservation of the Original Linkage Groups 45
species, however, crossing-over may occur at the segments involved in inversions.

The result is the complete sterility of interspecific hybrids due to the formation of
dicentrics and acentric fragments (if the heterozygosity for one inversion reduces the
fertility by 50%, three or four inversions would produce complete sterility).
Starting from an ancestral karyotype, the combination of Robertsonian fusions
and pericentric inversions (visible inversions) can give rise to a remarkably wide
range of karyotypes. As an example, let us imagine a hypothetical ancestral karyotype
made of 96 acrocentric chromosomes. Twenty-four successive Robertsonian fusions
can reduce the diploid chromosome number to 48. This karyotype made of 48 meta-
centdes can in turn be changed to that made of 48 acrocentrics by 24 successive
pericentric inversions. Subsequent Robertsonian fusions can further reduce the
diploid chromosome number to 24, etc. In fact, the combination of Robertsonian
fusions and pericentric inversions can explain most of the differences in karyotypes
which we observe among diverse species of placental mammals.
5. Conservation of the Original Linkage Groups

As already mentioned, an inversion merely results in an internal rearrangement
of an originallinkage group, and a Robertsonian fusion simply links together two
intact linkage groups. If the combination of Robertsonian fusions and inversions
has been mainly responsible for evolutional changes of the karyotypes, it then follows
that despite apparently drastic changes in the appearance of diploid chromosome
complements the originallinkage relationship between various structural genes must
have been preserved to a surprising extent by diverse descendants of a common
ancestor. Serious disturbances in linkage relationship would have occurred only
when a Robertsonian fusion was followed by a pericentric inversion. The available
information on linkage relationships appears to verify the above assumption.
The most dramatic example of this conservation can be seen in the X-linkage
group of placental mammals. As discussed in Chapter III, the unique dosage com-
pensation mechanism for X-linked genes based on random inactivation of one or the
other X-chromosome of female somatic cells appears to have evolved more than
100 million years ago in the common ancestor to both marsupial and placental
mammals. Once this mechanism was established, natural selection no doubt favored
the conservation of the entire X-linkage group.
The X-chromosome of a great majority of placental mammals is nearly identical
in absolute size; comprising about 5% of the genome (OHNo et al., 1964). In the
case of unusually large X-chromosomes which are seen in exceptional species of
rodents and ungulates, the part in excess of the standard 5% has been rendered
permanently inert by heterochromatinization (WoLF et af., 1965; FRACCARO et al.,
1968; WuRSTER et al., 1968). There is growing evidence of the homology ofX-linked
genes in placental mammals. At present, seven separate gene loci are known to be
X-linked in two or morediverse species (OHNO, 1969). For example, the gene locus
for an enzyme, glucose-6-phosphate dehydrogenase, has been shown tobe X-linked
in man (CHILDS etal., 1958; BoYER et al., 1962; KrRKMAN and HENDRICKSON,1963),
the horse and the donkey (TRUJILLO et al., 1965; MATHAI et al., 1966), the hare
(HNO et al., 1965) and the house mouse (EPSTEIN, 1969). The two gene loci which
specify antihemophillc factors VIII and IX are X-linked in man as well as in the dog
(GRAHAM etal., 1947; HvTTetal., 1948; MusTARn etal., 1960).
As to the autosomal genes, the genes which were originally dustered tagether
appear to have remained so despite extensive speciation. For example, the genes for
various classes of immunoglobulin heavy-chains have remained in very close linkage
to each other not only in man (NATVlG et al., 1967), but also in the mouse (HERZEN-
BERGet af., 1967).
Both mammalian and avian species possess a third gene locus for an enzyme,
lactate dehydrogenase (GoLDBERG, 1962; BLANCO et al., 1964). Unlike A- and B-
subunits of this enzyme which are seen in a variety of tissues, the C-subunit specified
by this third locus is seen only in the sexually mature testis. In avian species, it has
been shown that the locus for the C-subunit is closely linked to that for the B-subunit.
These two gene Jod for B and C LDH subunits have apparently remained in close
linkage since the time of early reptiles; the length of more than 200 million years
(ZINKHAM and lSENSEE, 1969).
There is also evidence for the conservation of rather !arge segments of autosomes.
For example, a pair of coat color gene loci, pink-eyed dilution (a melanosome stroma
protein locus) and albino (tyrosinase locus), are about 15 crossing-over units apart
not only in the house mouse, but also in the rat (RoBINSON, 1960), and the deer
mouse (ROBINSON, 1964).
References
BLANCO, A., ZrNKHAM, W. H., KuPCHYK, L.: Genetic control and ontogeny of lactate
dehydrogenase in pigeon testes. J. Exptl. Zool. 156, 137-152 (1964).
BaYER, S. H., PoRTER, I. H., WErLBOECHER, R.: Electrophqretic heterogeneity of glucose-6-
phosphate dehydrogenase and its relationship to enzyme deficiency in man. Proc. Natl.
Acad. Sei. US 48, 1868-1876 (1962).
CEPPELLINI, R.: Biochemistry of human genedes (WoLSTENHOLME, G. E. W., O'CoNNOR, C.
M., Eds.), pp. 133-138. London: J. and A. Churchill, Ltd. 1959.
CHILDS, B., ZrNKHAM, W. H., BROWNE, E. A., KrMBRO, E. L., ToRBERT, J. V.: A genetic
study of a defect in glutathione metabolism of the erythrocyte. Bull. Johns Hopkins
Hosp. 102, 21-37 (1958).
CoHN, M.: Natural history of the myeloma. Cold Spring Barbor Symposia Quant. Biol. 32,
211-222 (1967).
EPSTErN, C. J.: Mammalian oocytes: X chromosome activity. Science 163, 1078-1079 (1969).
FRACCARO, M., GusTAVSSON, I., BuLTEN, M., LrNDSTEN, J., TrEPOLo, L.: Chronology of
DNA replication in the sex chromosomes of the reindeer ( Rangifer tarandus L.). Cyto-
genetics 7, 196-211 (1968).
GOLDBERG, E.: Lactic and malic dehydrogenase in human spermatozoa. Science 139, 602 to
603 (1962).
GRAHAM, J. B., BucKWALTER, J. A., HARTLEY, L. ]., BRINKHAUS, K. M.: Canine hemo-
philia: Observations on the course, the clotting anomaly, and the effect of blood trans-
fusion. J. Exptl. Med. 90, 97-111 (1947).
GROPP, A., voN LEHMANN, E.: Chromosomenvariation vom Robertsonschen Typus bei der
Tabakmaus, lvf. poschiavinus, und ihren Hybriden mit der Laboratoriumsmaus. Cyto-
genetics (in press).
GusTAVSSON, I.: Chromosome abnormality in cattle. Nature 209, 865-866 (1966).
HARRIS, H., HoPKINSON, D. A., LuFFMAN, J. E., RAPLEY, S.: Electrophoretic variation in
erythrocyte enzymes. In: Hereditary disorders of erythrocyte metabolism (BEUTLER, E.,
Ed.), City ofHope Sym. Series, Vol. 1, pp. 1-20. New York: Grune & Stratton 1967.
HECK, H., WuRSTER, D., BENIRSCHKE, K.: Chromosome study of members of the sub-
families Caprinae and Bovinae, family Bovidae; the musk ox, ibex, aoudad, Conga buffalo
and gaur. Z. Sugetierkunde 33, 172-179 (1968).
HERZENBERG, L. A., MINNA, J. D., HERZENBERG, L. A.: The chromosome region for
immunoglobulin heavy-chains in the mause. Cold Spring Barbor Symposia Quant.
Biol. 32, 181-186 (1967).
References 47
HUNGERFORD, D. A., SNYDER, R. L.: Somatic chromosomes of a black rhinoceros ( Dieeros

bicornis GRAY, 1821). Amer. Nat. 101, 357-358 (1967).
Hsu, T. C., ARRIGHr, F. E.: Chromosomal evolution in the genus Peromyscus (Cricetidae,
Rodentia). Cytogenetics 5, 355-359 (1966).
Hurr, F. B., RrcKARD, C. G., FrELD, R. A.: Sex-linked hemophilia in dogs. J. Heredity 39,
2-9 (1948).
]ACOB, F., MoNOD, J.: Genetic regulatory mechanism in the synthesis ofproteins. J. Mol.
Biol. 3, 318-356 (1961).
KrRKMAN, H. N., HENDRICKSON, E. M.: Sex-linked electrophoretic difference in glucose-6-
phosphate dehydrogenase. Am.]. Human Genet. 15, 241-258 (1963).
KuNKEL, H. G., ALLEN, J. C., GREY, H. M.: Genetic characters and the polypeptide chains
of various types of gamma-globulin. Cold Spring Barbor Symposia Quant. Biol. 29,
443-447 (1964).
MATHAI, C. K., HNO, S., BEuTLER, E.: Sex-linkage of the glucose-6-phosphate dehydro-
genase genein the family Equidae. Nature 210, 115-116 (1966).
MArrHEY, R.: La formule chromosomique et le problerne de la determination sexuelle chez
Ellobius lutescens THOMAS. Rodentia-Muridae-Microtinae. Arch. Klaus-Stift. Vererb.-
Forsch. 28, 65-73 (1953).
MusTARD, J. F., RosWELL, H. C., RoBINSON, G. A., HoEKSEMA, T. D., DowNrE, H. G.:
Canine hemophilia B. (Christmas disease). Brit. J. Haemat. 6, 259-266 (1960).
NATVIG, J. B., KuNKEL, H. G., LITWIN, S. P.: Genetic markers of the heavy-chain subgroups
ofhuman gamma G globulin. Cold Spring Barbor Symposia Quant. Biol. 32, 173-180
(1967).
HNO, S.: Evolution of sex chromosomes in mammals. In: Annual review of genetics, Vol.
III (RoMAN, H. L., Ed.) pp. 495-524. Palo Alto: Annual Reviews, Inc. 1969.
- JAINCHILL, J., STENIUS, C.: The creeping vole ( Microtus oregoni) as a gonosomic mosaic.
I. The OY/XY constitution of the male. Cytogenetics 2, 232-239 (1963).
- BECAK, W., BECAK, M. L.: X-autosome ratio and the behavior pattern of individual
X-chromosomes in placental mammals. Chromesoma 15, 14--30 (1964).
- PooLE, J., GusTAVSSON, I.: Sex-linkage of erythrocyte glucose-6-phosphate dehydro-
genase in two species ofwild hares. Science 150, 1737-1738 (1965).
OunrN, J.: Genetic regulation of immunoglobulin synthesis. J. Cell Physiol. 67, 77-108
(1966).
PARR, C. W.: Erythrocyte phosphogluconate dehydrogenase polymorphism. Nature 210,
487-489 (1966).
PorrER, M., LrEBERMAN, R.: Genetic studies of immunoglobulins in mice. Cold Spring
Barbor Symposia Quant. Biol. 32, 203-209 (1967).
RoBERTSON, W. R. B.: Taxonomie relationship in the chromosomes of Tettigidae and
Agrididae: V-shaped chromosomes and their significance in Agrididae, Locustidae and
Grylidae: Chromosomes and variation. J. Morphol. 27, 179-332 (1916).
RoBINSON, R.: A review of independent and linked segregation in the Norway rat. J. Genet.
57, 173-192 (1960).
- Linkage in Peromyscus. Heredity 19, 701-709 (1964).
SHAW, C. R.: Electrophoretic vadation in enzymes. Science 149, 936-943 (1965).
THULINE, H. C., MoRRow, A. C., NoRBY, D. E., MoTULSKY, A. G.: Autosomal phospho-
gluconic dehydrogenase polymorphism in the cat (Felis cattus L..) Science 157, 431-432
(1967).
TRUJILLO, J. M., WALDEN, B., O'NEIL, P., ANsTALL, H. B.: Sex-linkage of glucose-6-
phosphate dehydrogenase in the horse and donkey. Science 148, 1603-1604 (1965).
WoLF, U., FLrNSPACH, G., BHM, R., HNO, S.: DNS-Reduplikationsmuster bei den Riesen-
Geschlechtschromosomen von Microtus agrestis. Chromesoma 16, 609-617 (1965).
WuRSTER, D. H., BENIRSCHKE, K., NoELKE, H.: Unusually !arge sex chromosomes in the
sitatunga (Tragelaphus spekei) and the black buck (Antilope cervicapra). Chromesoma 23,
317-323 (1968).
ZrNKHAM, W. H., IsENSEE, H.: Linkage of Iactate dehydrogenase B and C Iod in pigeons.
Science 164, 185-187 (1969).
Chapter VIII
The Spontaneous Mutation Rate

Inasmuch as heritable changes are primarily caused by mutations affecting
individual cistrons within the genome, the accurate estimation of the spontaneaus
mutationrate is of the utmost importance in understanding evolution. Furthermore, in
view of the fact that a mutation randomly affects a single base pair of any cistron and
that only some of the mutations are permitted by natural selection to accompany the
process of speciation, we should also have some idea about relative proportians with
regard to forbidden versus tolerable mutations. In practice, it is convenient to consider
the spontaneaus mutation rate either in terms of per locus per generation or per base
pair per generation.
1. Forbidden Mutations Versus Tolerable Mutations

In the past, diverse methods have been used to estimate the spontaneaus mutation
rate at a variety of structural gene loci in mammals, as weil as in fruit flies and bacteria.
A surprisingly uniform mutation rate in the order of magnitude of 10-s per locus
per generation has been obtained. The following references for this mutation rate
are only those dealing with man and other mammals: RussELL, 1951; SLATIS, 1955;
STEVENSON, 1957; LYON, 1959. .
The mutation rate quoted above, however, is rather irrelevant to evolution,
since the rate estimated applies only to the type of mutations which deprive an
affected gene from the performance of its assigned function. Fot instance, the muta-
tion rate to null alleles is measured at the thymidine kinase locus of bacteria selected
for their resistance to a pyrimidine analog (BdUR). Similarly, from the incidence of
phenylketonuria among human newborns, one calculates the spontaneaus mutation
rate to null alleles at the phenylalanine hydroxylase locus.
Each of such forbidden mutations represents a change in the base sequence of a
cistron which resulted in disturbing the acti ve site amino acid sequence of a poly-
peptide chain (Chapter V). On tht: contrary, a tolerable mutation which is compatible
with successful speciation represents a missense mutation resulting in an amino acid
exchange at a functionally less-critical site of a polypeptide chain (Chapter VI). While
it is true that the spontaneaus mutation rate for forbidden mutations is of the order of
magnitude of 10-s, what we wish to know in analyzing evolution is the spontaneaus
rate for tolerable mutations. As far as some of the larger eistraus are concerned, there
is little doubt that the tolerable mutation rate per locus is considerably higher than the
generally accepted figure of 1 x 10-s (one in one hundred thousand), which is
applicable only to forbidden mutations.
How can we calculate the spontaneaus rate for tolerable mutations at a given gene
locus? The following approach has been explored by AMES and bis colleagues
(WHITFIELD et al., 1966). As already mentioned, the theoretically expected ratio
between nonsense mutations and missense mutations is 1:17. Unless it affects the very
end of a cistron, every nonsense mutation which results in the premature termination
of a polypeptide chain growth should represent a forbidden mutation. Missense muta-
tions, on the other band, can be either forbidden or tolerable depending upon the nature
of the amino acid exchange (conservative or drastic), as weil as the site at which the
The Mutation Rate and Cistran Size 49
exchange takes place. Using this information, a tolerablemutationrate per given gene
locus can be estimated from the knownforbidden mutation rate, provided one knows
the fraction of forbidden mutations contributed by nonsense base substitutions. AMES
and his colleagues applied this principle to the spontaneously occurting functionless
mutants in the amino transferase locus (C gene) of the histidine operon in Salmonella
ryphimurium. The functionless mutants (forbidden mutants) included a nearly equal
number of nonsense and missense mutants.
The conclusion to be drawn from the above observation is that for every for-
bidden mutation which occurred at the aminotransferase locus of this bacterial species,
there must have been 8 tolerable mutations which substituted amino acids at various
functionally less-critical sites of the polypeptide chain. It then follows that in the case
of larger cistrons specifying a polypeptide chain which contains many functionally
less-critical sites, the spontaneaus rate for tolerable mutations is in the order of
magnitude of 10-4 (one order of magnitude higher than that forforbidden mutations).
Here, the samesense mutations are excluded from consideration.
2. The Mutation Rate and Cistron Size

The magnitude of difference between the tolerable mutation rate and the forbidden
mutation rate should differ from cistron to cistron, partly because of size difference.
Since different cistrons are of different sizes, the ideal way to express the mutation
rate is not per locus, but per base pair. Let us arbitrarily assume the universal mutation
rate of 1 x 10-7 perbasepair per generation (KIMURA, 1968). Some cisti:ons are ex-
tremely short, being made of only 80 or so base pairs; e.g., those which specify
transfer RNA. The spontaneaus mutation rate for all possible mutations per transjer
RNA locus should be merely 8 x 10-6. As discussed in Chapter II, the molecular
requirement to be a transfer RNA is very exact. Almost any change in the base se-
quence of a transjer RNA cistron would represent a forbidden mutation. It appears that
the stringent conservation of the base sequence of transfer RNA throughout the
history of living organisms is due not only to the ruthless elimination of forbidden
mutations by natural selection, but also to the smallness of the cistron size; the
consequence of which is the extremely low spontaneaus mutation rate per locus.
As a rule, cytochrome C of vertebrates is made of 104 amino acid residues. As
each amino acid residue of a polypeptide chain is specified by base triplets in a
messenger RNA, the minimum length of the cistron for cytochrome Cis 312 base pairs.
It then follows that the true mutationrate per cytochrome C locus is merely 3 X 10-5
This is within the range of the generally accepted figure of 1 X 10-5 which is appli-
cable only to forbidden mutations. It is a rule that the shorter the polypeptide chain,
the greater the part of its sequence which represents the functionally critical sites
where amino acid substitutions of any kind areforbidden. It is rather expected that so
far as a short cistron is concerned, the true mutation rate for all possible mutations
becomes nearly the same as the conventional mutation rate applicable exclusively to
forbidden mutations. Under the circumstances, the low over-all mutation rate is a
protective shield agairrst the relentless pressure of natural selection. Indeed, the
amino acid sequence of cytochrome C has been conserved to a remarkable extent
throughout the evolution of vertebrates which began nearly 300 million years ago.
Cytochrome C of such diverse vertebrates as man and tuna fish differ at only 21 of the
104 sites (MARGOLIASH, 1966).
Polypeptide subunits of some enzymatic and non-enzymatic proteins, on the

other hand, may be made of as many as 600 amino acid residues; thus, the cistron for
such a polypeptide chain is 1,800 base pairs long. The over-allmutationrate per locus
in this case becomes 2 x 10-4, and one order of magnitude difference is seen between
the over-all mutation rate and the rate for forbidden mutations; the difference re-
presenting tolerable mutations. This great difference is expected, because in the Ionger
polypeptide chain there will be more sites where amino acid substitutions are toler-
able. Amino acid substitutions at the functionally less-critical sites do change the
kinetic property such as pH optimum and Km in the case of an enzymatic subunit.
Thus, natural selection is afforded with an opportunity to favor a particular tolerable
mutation.
As long as each vital function is assigned to a single gene locus within the genome,
only those long cistrons containing a !arge number of functionally less-critical sites
undergo rapid and meaningful evolutional changes. It is weil worth realizing that so
far as these long cistrons are concerned, the spontaneaus rate for tolerable mutations
is ab out 10 times higher than the generally acceptedjorbidden mutationrate of 1 X 10-5
3. Intragenie Recombination and the Principle of Polymorphism Generating

More Polymorphism
With the advent of electrophoretic studies, it became increasingly clear that the
maintenance of a multiple allelic system by an interbreeding population is not peculiar
to blood group and histocompatibility gene loci. Many of the enzyme loci studied
on man and other vertebrates have been found to maintain a number of alleles
(HARR!S, 1969; SALTHE, 1969). These alleles which specify electrophoretic variants
of the enzyme represent the drastic type of missense mutation, each of w hich replaced
an amino acid with one of a different charge; the example being an exchange between
alanine and glutamic acid. Forty per cent of all possible missense mutations would
result in changing the net charge of a polypeptide chain specified by that locus
(FrTCH, 1966). It appears that in most instances they represent tolerable mutations
which are effectively neutral, since the heterozygous advantage cannot be shown
and the allelic frequency in a population tends to conform to the Hardy-Weinberg
expectation.
With regard to the effectively neutral alleles, however, KrMURA and CRow (1964)
have shown that there are strict limits to the number of alldes which can be main-
tained by a population. Their formula is n = 4N[J. + 1, where n is the effective
number of neutral alleles, N is the effective population size as expressed by the number
of breeding adults, and (L is the spontaneaus mutation rate per locus per generation.
Accepting the spontaneaus tolerablemutationrate of 1 x 10-4, a population of 7,500
breeding adults is required in order to maintain four neutral alldes at a single gene
locus. When the population is defined as an interbreeding unit, individual popula-
tions of most species seldom approach that size.
Could it be that the spontaneaus occurrence of tolerable mutations is not the sole
contributor to the establishment and maintenance of a multiple allelic system? Could
there be another mechanism which generates and maintains polymorphism?
When one examines the same enzyme locus in divergent vertebrate species, one is
struck by the fact that while some species maintain a number of alldes at that locus,
other species show no sign of polymorphism at the corresponding locus. For example,
Intragenie Recombination and Polymorphism 51
at the autosomally inherited gene locus for an enzyme; 6-phosphogluconate dehydro-

genase, the existence of a multiple alleHe system specifying electrophoretic variants
has been found in man (PARR, 1966), rats (PARR, 1966), deer mice (SHAW, 1965),
cats (THULINE et al., 1967), quail (BAKER and MANWELL, 1967; HNO, et al., 1968)
andin goldfish and other fish (BENDER and HNO, 1968). But in the case of the house
mouse and the rainbow trout, although we have already examined nearly 1,000 indi-
viduals of each species, we have yet to find a single individual showing an electro-
phoretic variant of this enzyme. Yet, spontaneaus mutations should affect homologaus
cistrons of the same size with the same rate regardless of species.
It may be that there is a principle which can be stated as "polymorphism generates
more polymorphism" in that mutation-like events are more likely to occur in germ
cells of heterozygotes than in those of the homozygote. If there exists such a principle,
the base sequence of a gene locus would remain rather stable within a population,
unless the frequency of new alleles produced by tolerable mutations increase to
significant proportians by drift and other means. Once polymorphism is established,
however, frequent mutation-like events occurring in heterozygotes will perpetuate
polymorphism even in a small population. Such mutation-like events which occur
only in heterozygotes must be the consequence of intragenic recombination .. A re-
combination between homologaus cistrons is of no consequence if it occurs in homo-
zygotes, or in common heterozygotes which have two alleles which differ from each
other by a single base Substitution. Only when two alleles involved are different
from each other by two or more non-consecutive base substitutions does a recombi-
nation generate a new composite allele which is different from either of the two
already existing alleles.
How frequent is the rate of intragenie recombination? Some of the blood group
gene loci are characterized by having a large nurober of alleles, and, consequently,
individuals are more often heterozygous than homozygous. At the gene locus con-
trolling the blood group B-system of cattle, STORMONT (1965) has estimated the
spontaneaus incidence of recombination-like events detected by the combination of
antisera to be 2 x 10-3 or one in five hundred. This is an astonishingly high figure
when compared to the spontaneaus tolerable mutation rate of one in ten thousand
(1 X 10-4).
WRIGHT and ATHERTON maintain three alleles at the gene locus for LDH B-sub-
unit in their stock of brook trout. They are a wild-type allele (B) and two mutant
alleles (B' and B"). According to their abstract (WRrGHT and ATHERTON, 1968),
recombination between B' and B" alleles which produced an apparently revertant
wild-type B allele occurred at an astanishing frequency, so that 2 of the 100 progeny of
the BJB ~ x B'JB" cl cross typed and bred as BJB.
F 1 hybrids between two inbred strains of mice should be heterozygous for parental
alleles at all of the nearly 20 histocompatibility gene loci. Accordingly, reciprocal
skin grafts exchanged between F 1 mice should be accepted. Any rejection has to
reflect a mutation-like event which occurred either in the somatic cells of the F 1
mouse itself, or in a germ cell of its patent. BArLEY (1966) encountered skin graft
rejection at the rate of 13.5 x 10-3, or roughlyonein one hundred. Inasmuch as the
histocompatibility gene loci remain quite stable within each inbred strain, the most
reasonable explanation appears to be that observed mutation-like events reflect
intragenic recombinations which occurred in somatic cells of F 1 hybrids.
4*
An NADP-dependent enzyme, 6-phosphogluconate dehydrogenase, is a dimer

having a molecular weight of about 120,000. A polypeptide subunit of this enzyme
should be made of about 600 amino acid residues. Thus, the 6-PGD locus is a
rather large one being made of 1,800 or so base pairs. In our stock of quail (Coturnix
c. japonica), four alleles which specify electrophoretic variants A, B, C and D are
maintained at this autosomally inherited gene locus (HNO et al., 1968). This enabled
us to set up a nurober of matings between a homozygote for one allele and a hetero-
zygote for the other two alleles; the example being the AfD x B/B cross. The pro-
geny of such matings should have all been heterozygotes, with both A/B and B/D
-----
--
- - --
!!_
--
- -!!
-
--;;
0--------------------------------------
2 3 4 5 6 7 8 9 10 11 12
Q
Fig. 8. A tracing of a photograph of a starch gel plate stained for 6-PGD phenotype of quail.
The transmission of a new recombinant allele specifying D' (mu) subunit which arose in
a B/C heterozygous mother is shown. The anodal direction is upward. A, B, C and D sub-
units are specified by the known alleles of our population. The A 2-autodimer band stays
closest to the starting point, while the D 2-autodimer band moves farthest toward the anode.
Of the three bands shown by each heterozygote, the middle band represents a hybrid dimer
band. Slots 1 to 7: The phenotypes ofthefather and the mothet and five of their progeny;
four expected and one which received a new recombinant allele. A/D father (slot 1), B/C
mother (slot 2). A/B (slot 3), CfD (slot 4), A/C (slot 5), and B/D (slot 6) are phenotypes
expected ofthe progeny. While one progeny (slot 7) received D from his father, he received
an unexpected recombinant D'(mu) instead of Bor C from his mother. Slots 8 to 12: This
D/D'(mu) male was mated to an A/B female (slot 8) for the progeny test. Four expected
phenotypes recovered among the progeny were: A/D (slot 9), A/D'(mu) (slot 10), B/D
(slot 11) and B/D'(mu) (slot 12)
showing three bands of 6-PGD on the stained gel plate. Any other phenotype ex-
hibited by the progeny would reflect a mutation-like event which occurred in the
germ cells of one of the parents. When a total of 1,011 progeny from 26 such matings
were analyzed, six recombination-like events which occurred in the heterozygous
parents were detected. Three recombinations produced revertant alleles (e.g., an
A/D parent giving B-like electrophoretic variants to its progeny), while the other
three recombinations produced two types of new electrophoretic variants which had
not previously existed in our stock (Fig. 8) (OHNO et al., 1969).
The conclusion to be drawn from the above examples appears to be that in the
relatively large loci, intragenic recombination can occur at a rate between 10-2
and 10-3. Spontaneous tolerable mutations and recombinations between mutant
alleles supplement each other in generaring and maintaining a multiple allelic system.
On the So-called Living Fossils 53
Once independent Plissense mutations affecting different sites of the wild-type cistron
give rise to two mutant alleles, the recombination between the two can quickly
generate the new third allele as illustrated below:
Wild-type: -Pro-Arg-
Mutant 1: -His-Arg-
Mutant 2: -Pro-Met-
Recombinant between 1 and 2: -His-Met-
Furthermore, reversions by recombinations would also help maintain a multiple
allelic system within a small population. Mutations being random events, it should be
that the more tolerable allelic alternatives a population places at the disposal of natural
selection, the faster the rate of evolution ofthat population.
4. On the So-called Living Fossils

When the paleontologist speaks of different rates of evolution, he is stating the
observed fact that while certain groups of vertebrates appear to have diverged to
numerous types in a relatively short time, others seem to have remained essentially
unchanged for millions of years. Those living species whose almost exact replicas
are found in the fossil records of ancient times are commonly referred to as "living
fossils".
Of marsupial mammals, the North American opossum ( Didelphis virginiana) is a
good example of a living fossil. It appears that during the Cretaceous period, which
began 130 million years ago and ended at the advent of the Cenozoic era, the marsu-
pials were widely distributed throughout the world. Although a wide range of
adaptive radiation was enjoyed by later marsupials in Australia and South America
where they were sheltered from competitive placental mammals, the Cretaceous
marsupials were apparently of one generalized type, and the American opossum
appears to have maintained all the essential features of these ancestral Cretaceous
marsupials. One might say that the American opossum has not changed in 100 million
years.
The modern Australlau lungfish ( Neoceratodus) is another example of a living
fossil. The ancestry of the living lungfish, survivors of one line of the lobe-finned
fish, can be traced to the middle Devonian period. From this start, the centralline of
dipnoan evolution led to Ceratodus; a genus that became widely distributed during
the Triassie and subsequent periods of the Mesozoic era. Neoceratodus, a direct
descendant of Ceratodtts, has changed very little from its Mesozoic progenitor despite
the span of almost 200 million years. On the other hand, the South American lung-
fish ( Lepidosiren) and the African lungfish ( Protopterus) have modernized.
The rate of evolution is greatly influenced by population size, generation time
and other factors as shall be described in the next chapter. Nevertheless, the fact that
certain groups of vertebrates remained essentially unchanged over 100 million years,
while others in the same span of time evolved by successive waves of adaptive
radiation appears incompatible with our assumption that spontaneaus mutations
randomly affect the homologaus gene loci of all animals at about the same rate. In the
case of short cistrons where the amino acid sequence of their products had to be
stringently conserved, a living fossil and its more evolved relatives may indeed have
accumulated about tbe same nurober of neutral mutations in a given span of time.
In fact, I expect tbat if tbe amino add sequences of cytocbrome C of tbe modern
opossum and tbe kangaroo could be compared not only with eacb otber but witb
cytocbrome C of a Cretaceous marsupial ancestor, tbe two living species would be
sbown to bave accumulated about tbe same nurober of neutral amino acid sub-
stitutions since the time of a common ancestor.
Tbe possible explanation of living fossils migbt be found in tbe prindple of
polymorphism generaring more polymorphism. Natural selection primarily acts upon
tbe larger cistrons which bave tbe potential of generaring a large variety of functional
alleles. At tbese gene Iod, once tbe tbresbold is crossed by tbe accumulation of
tolerable mutations, a larger variety of alleles can be created in rapid succession by
intragenic recombination. Tbose spedes which bad crossed tbe tbreshold migbt bave
evolved by rapid adaptive radiation. In comparison, tbe rate of evolution bad to
remain very slow for tbose which bad not crossed the tbresbold; bence, a living
fossil.
References
BAILEY, D. W.: Heritable histocompatibility changes: Lysogeny in mice? Transplantation
4, 482-488 (1966).
BAKER, C. M. A., MANWELL, C.: Molecular genetics of avian proteins. VIII. Egg white
proteins of the migratory quail, Coturnix coturnix. New concepts of "hybrid vigour".
Comp. Biochem. Physiol. 23, 21-42 (1967).
BENDER, K., HNO, S.: Duplication of the autosomally inherited 6-phosphogluconate
dehydrogenase gene locus in tetraploid species of Cyprinid fish. Biochem. Genet. 2,
101-107 (1968).
FrTCH, W. M.: An improved method of testing for evolutional homology. J. Mol. Biol. 16,
9-16 (1966).
HARRIS, H.: Enzyme and protein polymorphism. Brit. Med. Bull. 25, 5-13 (1969).
KrMURA, M.: Evolutionary rate at the molecular level. Nature 217, 624--626 (1968).
- CRow, J. F.: The number of alleles that can be maintained in a finite population. Genetics
49, 725-738 (1964).
LYON, M. F. : Some evidence concerning the "mutational load" in inbred strains of mice.
Heredity 13, 334--352 (1959).
MARGOLIASH, E.: Sequence and structure of Cytochrome C. Advances in Protein Chem. 21,
113-286 (1966).
HNO, S., STENIUS, C., CHRISTIAN, L. C., HARRrs, C.: Synchronaus activation of both
parental alleles at the 6-PGD locus of Japanese quail embryos. Biochem. Genet. 2,
197-204 (1968).
- - - ScHIPMANN, G.: De novo mutation-like events observed at the 6-PG D locus of the
Japanese quail, and the principle of polymorphism breeding more polymorphism. Bio-
ehern. Genet. 3, 417-428 (1969).
PARR, C. W.: Erythrocyte phosphogluconate dehydrogenase polymorphism. Nature 210,
487-489 (1966).
RusSELL, W. L.: X-ray induced mutations in mice. Cold Spring Rarbor Symposia Quant.
Biol. 16, 327-336 (1951).
SALTHE, S. N.: Geographievariation of the lactate dehydrogenases of Rana pipiens and Rana
palustris. Biochem. Genet. 2, 271-304 (1969).
SHAW, C. R.: Electrophoretic variation in enzymes. Science 149, 936-943 (1965).
SLATrs, H. M.: Comments on the rate of mutation to chondrodystrophy in man. Am. J.
Human Genet. 7, 76-79 (1955).
STEVENSON, A. C.: Camparisans of mutation rates at single Iod in man. In: Effect of radia-
tion on human heredity, pp. 125-137. Geneva: World Health Organization 1957.
STORMONT, C.: Mammalian immunogenetics. In: Genetics today (GEERTS, S. J., Ed.), Vol. 3,
Chapter 19, pp. 716-722. New York: Pergarnon Press 1965.
The Rate of Evolution and the Importance of Isolation 55
THULINE, H. C., MoRROW, A. C., NoRBY, D. E., MoTULSKY, A. G.: Autosornat phospho-
gluconic dehydrogenase polymorphism in the cat. (Felis cattus L.). Science 157, 431-432
(1967).
WHITFIELD, H. J., MARTIN, R. G., AMES, B. N.: Classification of aminotransferase (C gene)
mutants in the histidine operon. J. Mol. Biol. 21, 335-355 (1966).
WRIGHT, J. E., ATHERTON, L.: Genetic control of interallelic recombination at the LDH B
locus in brook traut. Genedes 60, 240 (1968).
Chapter IX
The Rate of Evolution and the Importance of Isolation

It is a curious fact that when discussing evolution on a large scale even well-
disciplined Mendelian geneticists often succumb to a Lamarckian illusion. Thus, to
the question of why man does not have body hair, the answer is given that when our
brachiating ape ancestors of the forest came out to open ground they were forced to
travellong distances carrying heavy burdens on their backs. This tended to overheat
their hoclies which were covered with hair; hence, the loss of the body hair. This
version vividly paints in our mind the picture of a group of traveling man-apes,
panring and perspiring, suddenly making a community decision to comply with the
dictate of natural selection. The illusory implication is that every member of the
group of hairy apes transformed to a hairless ape. Such an unwittingly cultivated
illusion1tempts one to see the process of evolution as members of a successful species
advancing shoulder-to-shoulder to a higher and higher state of being and eventually
transforming to a new species. Thus, the process of speciation is viewed as a mere
extension of the accumulation of genetic changes which occurred in the previous
species, and the mutation rate is freely translated to the rate of evolution.
Uniform transformation from an old species to a new species can occur only if the
heritable traits responsible for the speciation are carried by a viral genome. Only then
can widespread infection and subsequent incorporation of the viral genome into the
hast genometransform a majority of the previous species to members of a new species.
In fact, an evolutionist who entertains the Lamarckian illusion is advocating evo-
lution by viral infection.
As lang as heritable changes in the chromosomal genes are the cause of evolution,
a new species can arise only as a result of intense inbreeding practiced by a small
minority of the parent species. The hairless trait no doubt arose as a mutation, most
likely recessive (Chapter VI), in either Australopithecus or Pithecanthropus (Homo
erectus) only one million years or so ago. Mutations being random events, a particular
mutation as a rule affects only one of the two homologaus genes of one particular
individual in a population. Thus, only one individual in a population of hairy apes
must have originally acquired a hairless trait. A species of hairless apes must have
arisen from the descendants of this particular mutant as a result of intense inbreeding
which established the homozygous state for hairlessness. Thus, the process of specia-
tion usually requires the reproductive "isolation" of a minority from the majority.
1. Isolation as a Prerequisite for Speciation

It appears that CHARLES DARWIN (DARWIN, 1888) clid not fully appreciate the
importance of "isolation" as a prerequisite to speciation. WAGNER (1889) appears to
be the first who appreciated the necessity of "isolation". Thus, he stated "The for-
mation of a genuine variety, which DARWIN considers an 'incipient species' will
succeed in nature only when a few inclividuals can spatially segregate themselves for
a long time from other members of the species by transgressing the confining barriers
of their range."
In man, Robertsonian fusions between five pairs of acrocentric autosomes [Fig. 3
(Plate I), Chapter III] occur sporaclically, but with recognizable frequency (HAMER-
'l'ON, 1968). Yet, in the absence of isolation, the chance of one heterozygote mating
with another heterozygote is very slim. Consequently, a homozygote for a Robert-
sonian translocation having 44 chromosomes instead of 46 has never been found.
In the case of mice, however, a small population which apparently isolated itself
in an abandoned tobacco factory in the Poschiave Valley of Switzerland has gone
through intensive inbreecling to become homozygous for seven independent Robert-
sonian translocations and emerged as an incipient species (Chapter VII).
In theory, speciation by isolation can occur without geographical isolation. Indeed,
man practices inbreecling to obtain new breeds of domesticated animals. Fot instance,
short-legged breeds of the dog, such as the dachshund, have been made homozygous
for a mutant gene which causes chondrodystrophy. In well established breeds of
cattle, only a few males with exemplary phenotypes are kept as stud bulls and each
stud bull sires hundreds or thousands of progeny. This practice results in a consider-
able degree of inbreecling, and, as a result, a Robertsonian translocation which
occurred in one stud bull of Sweden spread far and wide within the breed and a
number of homozygotes having 58 chromosomes instead of 60 were found (Chapter
VII).
Whether or not a minority can isolate itself from the majority in nature without
the benefit of a geographical barriet has been the subject of cliscussion for a con-
siderable period of time. MAYR (1963) has accumulated a great deal of evidence
which shows that the process of speciation invariably requires geographic isolation;
allopatric model of speciation. The hypothesis of {)'mpatric speciation (the idea that
biological or ecological races of a species can co-exist geographically in an area and
gradually cliverge genetically until they constitute clistinct species) has now been
largely cliscreclited. When the examples which were earlier alleged to establish the
existence of .rympatric speciation were examined more closely, most of these proved
to be examples of forms which had already diverged to the level of full species by
allopatric speciation. Recently, a stasipatric model of spedation was proposed(WHI'l'E,
1968). However, I fail to see a fundamental distinction between the stasipatric and
.rympatric models.
A rninority can isolate itself from the co-existing majority only if there exists an
effective barriet which prevents interbreecling between the two groups. While
clifferences in the courting behavior, body odor and the chromosome constitution
can serve as effective barriers as mentioned in Chapter VII, such clifferences cannot
be created overnight. Only during long periods of geographic isolation can a minority
accumulate enough clistinctive characteristics of their own. The reproductive barriet
Generation Time and the Rate of Evolution 57
that exists between an incipient species and the patent species sharing the same geo-
graphical area should not be mistaken as evidence of sympatric speciation, since such
a barriet owes its very existence to prior geographic isolation.
2. Population Size and the Price of Success

Mutations being rather rare events, the fixation (homozygosity) of a set of newly
acquired hereditary traits can occur only in a very small population in which in-
breeding occurs. Since a new species invariably arises from a very small population,
incidental fixation of a nurober of neutral mutations must also occur during the pro-
cess. Indeed, as pointed out in Chapter VI, many of the amino acid substitutions
which accompanied successful speciation appear to represent neutral mutations.
Geographie isolation is the prerequisite of evolution only because the rate of
evolution is inversely related to population size. Therefore, the extermination of a
majority also creates a situation equivalent to geographic isolation. Mter a sudden
drastic change in environment has taken a heavy toll from a successful species, the
species as a whole is reduced to a very small population. Inasmuch as the genotype of
a majority has already been proven unfit to cope with the changed environment, the
remnant of the now decimated species is confronted with only two alternatives:
either to perish or to emerge as a new species by becoming homozygous for a new set
of hereditary traits. Thus, maximum opportunity for speciation exists during the time
of drastic environmental change. Indeed, evolution of our own genus (Homo) took
place essentially within the last one million years of the Pleistocene epoch which
experienced the four successive periods of glaciation. Our immediate ancestor Homo
erectus ( Pithecanthropus) as well as our immediate relative Homo neanderthalensis
flourished during early interglacial periods and perished during glacial periods. It
appears that we owe the quick creation of our own species to periodical decimations
of our immediate ancestors.
The nurober of individuals which comprise a species is a good criterion for success
in evolution. Our own species, numbering in the hundreds of millians and occupying
allland areas of the world, has no doubt achieved a pinnacle of success. Y et, for this
reason, we have, for the moment at least, forfeited the chance for further evolution.
Although huge populations would become increasingly polymorphic by continuous
accumulation of tolerable mutations, a chance of any of the newly acquired hereditary
traits becoming fixed as a new species characteristic is practically nil. Such is the
price of success.
3. Generation Time and the Rate of Evolution

Other things being equal, the rate of evolution is also inversely related to the
generation time of a species. Here, the generation time is defined as the length of
time needed by individual members of a species to attain reproductive capability.
Thus, it differs from the life span of individuals.
Small rodents become sexually mature within a month or so after birth. In the
span of one million years, they would go through as many as 12 million generations.
Large, hoofed animals, on the other hand, may take 5 years or more of postnatallife
to attain sexual maturity. In one million years, they would go through only two
hundred thousand generations. Thus, during the same span of time, rodents would
encounter far more numerous opportunities for adaptative radiation than would
large hoofed animals.
The criteria of evolutionary success is various. We like to think of ourselves as
the most successful of all mammals, yet, we are but a single species, and the great
dominance we now enjoy has been a development of only the last several thousand
years. Primates as an order has been none too successful. In sharp contrast, rodents
were supremely successful during most of the Cenozoic era (the age of mammals). If
the range of adaptive radiation, the number of species, and the number of individuals
within a species are criteria for success in evolution, then rodents far outshine all
other mammals. There is little doubt that rodents owe their success to a great extent
to their short generation time. In contrast, animals with rather long generation time
often appear as though they have been standing still for one or two million years.
The effect of generation time on the rate of evolution can be illustrated by the
comparison between the rodent subfamily Microtinae (voles) and the ungulate family
Camelidae. The earllest known fossil record of the Microtinae is Mimomys from the
Pleistocene epoch found in Europe and Asia. The diversification of this subfamily
appears to have taken place within the Pleistocene epoch during the time span of
one million years. The geographical range of the Microtinae covers most of North
America southward to Guatemala and the northern two-thirds of the Eurasian con-
tinent. There are about 50 living species in this subfamily.
Living members of the family Camelidae originated from a common ancestor that
inhabited North America about one million years ago. Yet, there are only six living
species: the two-humped (Camelus bactrianus) and one-humped (C. dromedarius)
camels of the Old W orld and three species of llamas and one species of vicugna in
South America. Furthermore, the degree of diversification seen within the Microtinae
is much greater than that seen in the Camelidae. For example, the diploid chromosome
number of the Microtinae ranges from a high of 60 in Microtus chrotorrhinus (MEYLAN,
1967) to a low of 17 in Microtus oregoni (OHNO et al., 1963). In sharp contrast, all six
species of camelids, regardless of whether they are camels or llamas, have the appa-
rently indistinguishable diploid complement of 74 chromosomes (BENIRSCHKE, 1967;
1'AYLOR et a/., 1968.).
References
BENIRSCHKE, K.: Sterility and fertility of interspecific mammalian hybrids. In: Comparative
aspects of reproductive failure (BENIRSCHKE, K., Ed.). Berlin-Heidelberg-New York:
Springer 1967.
DARWIN, F.: The life and letters of Charles Darwin. London: John Murray 1888.
HAMERTON, ]. L.: Robertsonian translocations in man: Evidence for pre-zygotic selection.
Cytogenetics 7, 260-276 (1968).
MAYR, E.: Animal species and evolution. Cambridge (Massachusetts): Harvard Univ. Press
1963.
MEYLAN, A.: A1icrottt.r chrotorrhintt.r, another species with giant sex chromosomes. l'V[ammalian
Chromosome Newsletter 8, 280-281 (1967).
HNO, S., J AINCHILL, J ., STENIUS, C.: The creeping vole ( Microttt.r oregoni) as a gonosomic
mosaic. I. The OY/XY constitution of the male. Cytogenetics 2, 232-239 (1963).
TAYLOR, K. M., HuNGERFORD, D. A., SNYDER, R. L., ULMER, F. A., ]R.: Uniforrnity of
karyotypes in the Camelidae. Cytogenetics 7, 8-15 (1968).
WAGNER, M.: Die Entstehung der Arten durch rumliche Sonderung. Gesammelte Auf-
stze. Basel: Benno Schwabe 1889.
WHITE, M. ]. D.: Models of speciation. Science 159, 1065-1070 (1968).
Part 3
Why Gene Duplication?
Chapter X
Duplication for the Sake of Producing More of the Same
The discussions presented in Part 2 revealed the true character of natural selection.
It is not so much an advocator or mediator of heritable changes, but rather it is an
extremely efficient policeman which conserves the vital base sequence of each gene
contained in the genome. As long as one vital function is assigned to a single gene
locus within the genome, natural selection effectively forbids the perpetuation of
mutations affecting the active sites of a molecule. In the case of the enzyme locus,
tolerable mutations might change the kinetic property such as pH optimum and
Michaelis constant of the enzyme, but never the basic character. Therefore, the
dihydro-orotase locus would forever remain the dihydro-orotase locus, and the
-galactosidase locus would remain the -galactosidase locus.
lt becomes quite clear that while allelic changes at already existing gene loci
suffice for racial differentiation within species as weil as for adaptive radiation from
an immediate ancestor, they cannot account for large changes in evolution, because
large changes are made possible by the acquisition of new gene loci with previously
non-existent functions. Only by the accumulation ofjorbidden mutations at the active
sites can the gene locus change its basic character and become a new gene locus. An
escape from the ruthless pressure of natural selection is provided by the mechanism
of gene duplication. By duplication, a redundant copy of a locus is created. Natural
selection often ignores such a redundant copy, and, while being ignored, it accumu-
lates formerly Jorbidden mutations and is reborn as a new gene locus with a hitherto
non-existent function. Thus, gene duplication emerges as the major force of evolu-
tion.
Even before the advent of molecular biology, a number of geneticists with fore-
sight, such as l-IALDANE (1932), realized the important role gene duplication played
in evolution. However, full appreciation of the magnitude of importance was not
possible until the elucidation of the coding mechanism enabled us to interpret evo-
lutional changes refiected in the direct gene products.
Although the creation of new gene loci by supplying redundancy is the most
important role, there are other benefits the mechanism of gene duplication confers to
organisms. When the metabolic requirement of an organism dictates the presence of
60 Why Gene Duplication?
an enormaus amount of a particular gene product, the incorporation of multiple

copies of the gene locus by the genome often fulfills that requirement. This, then, is
the type of gene duplication which serves to produce more of the same gene product.
1. Genes for Ribosomal RNA

As stated in Chapter II, an organism only requires three different kinds of ribo-
somal RNA; 5S, 18S and 28S. While ribosomal RNA is short in variety, it has tobe
made in great quantity, for the translation of a single messenger RNA requires the
attachment of several ribosomes, and a single cell is likely to be producing many
copies each of hundreds of different kinds of messenger RNA. Thus, as much as 85%
of the total RNA extracted from ordinary somatic cells is ribosomal RNA.
Quite clearly, if the genome contains only a single DNA cistron for each of the
three different kinds of ribosomal RNA, individual somatic cells cannot synthesize
enough ribosomal RNA to sustain the ontogenic development. Using the technique
of DNA-RNA hybridization already mentioned in Chapter II, RITossA and SPIEGEL-
MAN (1965) have shown in the fruit fly (Drosophila melanogaster) that each nucleolar
organizer of this insect species contains 100 tandemly duplicated copies of a pair of
genes which transcribe for a dicistronic RNA which is later split into 18S and 28S
ribosomal RNA. In the fruit fly, the nucleolar organizer is carried by the X as well as
the Y-chromosome.
In the case of the Mrican water frog ( Xenopus laevis), the nucleolar organizer is
carried by a single pair of homologaus autosomes. The latest estimate indicates that
in this vertebrate species each nucleolar organizer contains 450 tandemly duplicated
copies of a pair of genes for 18S and 28S ribosomal RNA (BRoWN and DAwm, 1968).
However, it should be remernbered that the genome size (the haploid DNA content)
of Xenopus laevis is 30 or 40 times greater than the genome size of Drosophila melano-
gaster; therefore, it appears that in proportion to the genome size, Drosophila has a
greater number of genes for two classes of ribosomal RNA. The genome size of
Xenopus is only slightly smaller than that of mammals. But, in the case of mammalian
species, several pairs rather than a single pair of chromosomes tend to carry nucleolar
organizers. For instance, of the 46 chromosomes of man, five different pairs of acro-
centric autosomes carry nucleolar organizers [Fig. 3 (Plate I), Chapter III]. Could
it be that evolution from cold-blooded to warm-blooded vertebrates was accompanied
by an increase in the degree of duplication of a ribosomal RNA cistron? It is likely
that a higher rate of metabolism requires greater concentration of ribosomes in the cell.
It has been shown that a gene for the third dass of ribosomal RNA (5S) is not
contained in the nucleolar organizing region of the chromosome either in Drosophila
or in Xenoptts. However, there appears to be extreme redundancy of 5S DNA in the
genome. The latest estimate is that the genome of Xenopus contains 20,000 duplicated
copies of a DNA cistron for 5S ribosomal RNA (BROWN and DAwm, 1968).
Aside from the three classes of RNA mentioned above, ribosomes also contain
proteins. There is little doubt that for the continuous formation of ribosomes, the
cell has to synthesize as much ribosomal protein as ribosomal RNA. It is of extreme
interest to find out whether or not natural selection also favored the amplification by
tandem duplication of each structural gene for ribosomal protein. In the case of Escher-
ichia coli, at least 16 different kinds of ribosomal proteins, rauging in molecular weight
from 9,000 to 41,000, have been identified (KALTSCHMIDT et al., 1967).
Genes for Transfer RNA 61
Even further amplification of the genes for 18S and 28S ribosomal RNA appears
to occur during ogenesis of amphlbians and echinoderms. As mentioned earlier, an
individual Xenopus, which is homozygous for a deletion of the nucleolar organizer
is totally incapable of synthesizing 18S and 28S ribosomal RNA. Yet such a homo-
zygous deficient zygote derived from the mating of heterozygotes grows to the swim-
ming tadpole stage (ELSDALE et al., 1958). The amount of ribosomal RNA stored in
the egg cytoplasm by a heterozygous mother is sufficient to sustain the growth of
homozygous embryos to this advanced stage of development. It is clear that even the
nucleolar organizer, with 450 copies of a ribosomal RNA gene, cannot, by itself,
produce such enormous amounts of 18S and 28S RNA during ogenesis. It is now
shown that as the ocyte suspended in the diplotene stage of first meiotic prophase
begins to grow in size, the nucleolar organizer region of the chromosome dissemi-
nates its free copies to the nuclear plasm, so that the ocyte nucleus finally contains
1000 or so free copies of the nucleolar organizers; each of which appears to organize
an individual nucleolus. Since each nucleolar organizer already contains 450 tandemly
duplicated copies of a pair of genes for 18S as well as 28S ribosomal RNA, the number
of genes for two classes of ribosomal RNA which become available to the growing
ocyte is truly staggering; 450 x 1002. In sharp contrast to 18S and 28S, free copies
of a gene for 5S ribosomal RNA do not appear to be disseminated during ogenesis of
Xenopus. As the chromosomes already contain 20,000 duplicates of this gene, further
amplification appears unnecessary (BRoWN and DAwm, 1968).
In the case of amniote eggs of reptiles, birds, and mammals, such dissemination of
free copies of the nucleolar organizer during ogenesis probably occurs on a much
smaller scale, if it occurs at all. Nevertheless, the fact that a segment of the chromo-
some can engage in repeated DNA replication and disseminate its free copies, while
the rest of the chromosomes are not involved in DNA replication, has far reaching
implications.
2. Genes for Transfer RNA

The transjer RNA is also short in variety, for the genome of an organism needs to
contain only 30 or so different kinds of genes which transcribe transjer RNA (Chapter
II). However, individual cells need to produce each species of transfer RNA in rather
large amounts, for a single translation of a messenger RNA of avetage length to a
polypeptide chain requires a few hundred transjer RNA. Indeed, as much as 15% of
the total RNA extracted from the growing embryo is transfer RNA. One might
expect that in the case of transfer RNA too, natural selection favored the tandem
duplication of a cistron for each species of transjer RNA. Again utilizing the tech-
nique of DNA-RNA hybridization, RITOSSA, ATWOOD and SPIEGELMAN (1966 (1)]
arrived at the conclusion that if there are 60 different kinds of genes for transfer RNA,
the genome of Drosophila melanogaster contains 13 duplicated copies of each gene.
In the case of transfer RNA genes, however, there remains some doubt as to
whether or not the apparent redundancy revealed by the DNA-RNA hybridization
can really be taken as evidence of the genome containing multiple replicas of each
transfer RNA gene. The genome of E. coli contains two rather widely separated gene
loci (Su2 and Su3) or gene clusters for two subspecies of tyrosine transfer RNA. One
subspecies specified by Su3 has the anticodon which can recognize both codons
(UAU, UAC) for tyrosine. Thus, it would appear that the other subspecies specified
by Su2 must also have the same anticodon. Yet these two tyrosine transfer RNA's are
not identical to each other (GAREN, 1968). Su2 and Su3 of E. coli should be regarded
as two closely related but separate gene loci diverged from a common ancestral gene
after duplication, rather than exact replicas of each other. However, the fact that a
subspecies specified by Su3 amounts to only 10% of the total tyrosine transfer RNA
can be explained on the basis that while Su2 represents a duster of 10 tandem dupli-
cates, Su3 is a singleton.
3. Inherent Disadvantage of Having Multiple Copies of the Same Gene

On the surface, it would appear that whenever the need arises for an organism to
have an enormaus amount of one particular gene product, this need can easily be
satisfied by incorporating the multiple copies of the same gene into the genome. In
fact, the nature of natural selection and chromosomes are such that the incorporation
of the multiple copies entails inherent disadvantages.
The fact that ribosomal RNA isolated from Xenopus laevis, of the tailless amphibian
order Anura, hybridize very weil with the nucleolar organizing DNA isolated from
Salamanders, such as Axolotl mexicanum and Necturus maculosus, reveals that natural
selection has stringently conserved the base sequence of a pair of genes for 18S as
well as 28S ribosomal RNA (BRoWN and DAwm, 1968). Anurans and salamanders
followed separate paths of evolution for as long as 280 million years (since the first
amphibians emerged on this earth at the beginning of the Carboniferous period). If
base substitutions are permissible at many of the sites of this cistron, nearly 300 million
years of separation would have resulted in a marked difference in the base sequences
between the anuran nucleolar organizing DNA and the salamander nucleolar orga-
nizing DNA. Thus, anuran ribosomal RNA would not effectively hybridize with the
salamander nucleolar organizing DNA.
Natural selection can eliminate forbidden mutations and effectively police the base
sequence of a DNA cistron only if the genome contains a single copy of each gene.
Policing by natural selection becomes very ineffective when multiple copies of the
gene are present. For the sake of simplicity, let us assume that the genome contains
three exact replicas of the same gene. A forbidden mutation which rendered one of the
three copies functionless would be tolerated, since even a deficient homozygote still
has four doses of the good gene. The second jorbidden mutation which renders the
second copy useless would also be tolerated, for even a deficient double homozygote
still has two doses of a good gene. Thus, in a relatively short time, two of the three
duplicates would join the ranks of "garbage DNA", and finally only one func-
tional gene remains in the genome. Consequently, having hundreds of tandemly
duplicated copies of a single gene in the nucleolar organizer is not as ideal a situation
as it appears on the surface, for slowly but surely more and more duplicates would
become useless genes by mutation. Ideally, the gamete should contain only a single
gene each for 18S and 28S ribosomal RNA, and the tandem duplication of it should
occur only after fertilization. This way, all multiple copies of a ribosomal gene con-
tained in an individual are either uniformly defective or uniformly functional. Natural
selection can now eliminate unfit individuals which inherited a defective ribosomal gene.
It is surprising that no known organism employs this ideal solution for the policing
of its ribosomal genes.
Inherent Disadvantage of Having Multiple Copies of the Same Gene 63
CALI,AN (1967) has proposed a very ingenious mechanism by which the organism
might escape the hazard of containing multiple copies of the gene in the genome
He postulates that there is a hierarchy among the tandem duplicates in that the one at
the end is the master, while all others are slaves; the master-slave theory. When
chromosomes duplicate before each cell division, not the slaves, but only the master
serves as the template for DNA replication. The net effect of the master-slave system
is the same as the gamete having only a single dose of the ribosomal gene, since all the
ribosomal genes contained in an individual are either uniformly defective or uniformly
R R R
R R R
R R R R R
I
R R R
R R R
A
Fig. 9. The consequence of unequal crossing-over between duplicated segments is illustrated

on the nucleolar organizer which normally contains three tandemly duplicated copies of a
pair of genes (R) for 18S and 28S ribo.romal RNA. Middle column: First meiotic prophase.
The homologaus pairing between the duplicated segments is inexact. As a result, a chiasma
is exchanged between the third gene of the chtomosome at the left and the first gene of its
homologue at the right. Left and right columns: Two daughter cells in 2nd meiosis. At the
left, one of the two chromatids received the deleted nucleolar organizer (one R). At the
right, one of the two received the further duplicated nucleolar organizer (five R's). Aseach
crossing-over involves two of the four chromatids of the two homologues in pairing, two
of the four gametes produced are affected.
functional. If the master suffers a forbidden mutation, all the slaves of the next cell
generation would inherit the same defect.
One wonders if it is this master-slave system which enabled anurans and sala-
manders to stringently conserve the base sequence of the ribosomal genes despite the
presence in their genome of 450 or so tandemly arranged copies.
Another serious difficulty an organism encounters by having multiple copies of
the same gene is constant deletion and further duplication which afflicts the chromo-
somal region made of tandemly duplicated copies. Crossing-overs that normally
occur between homologaus chromosomes during meiosis are, as a rule, no problem,
for exchanges are preceded by exact gene-for-gene pairing between two homologaus
chromosomes. In the case of a duplicated region, however, homologaus pairing
becomes very inexact. For instance, No. 1 ribosomal gene at the head of the nucleolar
organizing region of one chromosome rnight pair with No. 250 ribosomalgenein the
middle of the nudeolar organizer of its homologue. The result of such shifted
pairing is "unequal crossing-over". Where both chromosomes had the nudeolar
organizer made of 450 copies of a ribosoJJJal gene, one would now receive only
200 copies (deletion), while the other receives 700 copies (further duplication) as
shown in Fig. 9. If homologaus pairing is truly based on the attraction that exists
between the DNA of nearly identical base sequences, such shifted pairing and sub-
sequent unequal exchange should also occur between two chromatids of the same
chromosome in somatic cells. Such unequal crossing-over and unequal exchange
between the nudeolar organizers on the X and the Y are constantly occurring in the
fruit fly (Drosophila JJJelanogaster). Those which received the considerably deleted
nudeolar organizers from both parents finally become recognizable because of their
markedly retarded growth [RrTossA et al., 1966 (2)]. Those affected flies have been
known as bobbed mutants (STERN, 1927). Because further unequal crossing-over
between the deleted nudeolar organizers occasionally result in restoration of the
normal nudeolar organizer, normal flies frequently emerge from a stock of bobbed
mutants. The reciprocal product of unequal crossing-over is the extraordinarily
large nudeolar organizer containing a greater than normal nurober of duplicates of a
ribosoJJJal gene. Contrary to what one might expect, a fly which inherited such a great
nudeolar organizer does not become a superfly.
There is yet another dass of mutations in Drosophila which result in generalized
growth retardation. They are known as Minutes, for they are homologaus lethal,
dominant traits. Although the Minutes form a phenotypically homogeneaus dass, any
of the over 50 independent gene loci widely scattered in the genome can mutate to
become a Minute. ATwoon [in RrTOSSA et al., 1966 (1)] postulates that each Minute is
also a deletion due to unequal crossing over affecting one duster of 13 duplicates of a
particular transfer RNA gene.
Such deleterious consequence of unequal crossing-over is the fate which has to be
endured by the chromosomal segment carrying the tandem duplicates of the same
gene. Y et, in the absence of either of the two ideal systems, one where the gamete
contains only one dose of the gene with duplication occurring after fertilization
and the other, the master-slave system, apparently deleterious deletions might be
beneficial to the species in the long-run. As a result of deletion, the nudeolar organizer
can deanse itself of degenerate duplicates which became functionless due to accumula-
tion of mutations. Subsequent unequal crossing-over between the partially deleted
nudeolar organizers can restote the original degree of duplication this time made
mostly of functional copies.
For mammalian species which carry the nudeolar organizers on several different
chromosomes, the additional problern of maintaining homology between the regions
of non-homologaus chromosomes is imposed. Unless all these nucleolar organizing
regions involve themselves in mutual exchange of genetic materials, some would
become a useless collection of degenerate copies not contributing to the production of
18S and 28S ribosoJJJal RNA.
Of 46 chromosomes in the diploid nucleus of man, the nucleolar organizers are
carried by the three pairs of medium-sized acrocentric autosomes (13th, 14th and
15th pairs) as well as by the two smallest pairs of acrocentric autosomes (21st and
22nd pairs) [Fig. 3 (Plate 1), Chapter III]. In human somatic cells, all these acrocentric
autosomes are often in very dose association with each other at their nudeolar organ-
The Attainment of a Permanent Heterozygous Advantage 65
izers (FERGUSON-SMITH and HANDMAKER, 1961; HNO et al., 1961). This appears to
be a mean employed by mammals to maintain the homology between the nucleolar
organizers carried by non-homologaus chromosomes.
References
BROWN, D. D., DAwm, I. B.: Specific gene amplification in ocytes. Science 160, 272-280
(1968).
CALLAN, H. G.: The organizationof genetic units in chromosomes. J. Cell Sei. 2, 1-7 (1967).
ELSDALE, T. R., FISCHBERG, M., SMITH, S.: A mutation that reduces nucleolar number in
Xenopus laevis. Exptl. Cell Res. 14, 642-643 (1958).
FERGUSON-SMITH, M. A., HANDMAKER, S. D.: Observations on the satellited human chromo-
somes. Lancet 1961 I, 638-640.
HALDANE, J. B. S.: The causes of evolution. New York: Rarper and Eros. 1932.
KALTSCHMIDT, I., DZIONARA, M., DoNNER, D ., WITTMANN, H. G.: Ribosomal proteins. Mol.
Gener. Genet. 100, 364-373 (1967).
0HNO, S., TRUJILLO, J. M., KAPLAN, W. D., KINOSITA, R.: Nucleolus-organizers in the
causation of chromosomal anomalies in man. Lancet 1961 li, 123-125.
RITOSSA, F. M., SPIEGELMAN, S.: Localization of DNA complementary to ribosomal RNA
in the nucleolus organizer region of Drosophila melanogaster. Proc. Natl. Acad. Sei. US
53, 737-745 (1965).
- ATwooo, K. C., SPIEGELMAN, S.: (1) On the redundancy of DNA complementary to
amino acid transfer RNA and its absence from the nucleolar organizing region of Droso-
phila melano,f!,aster. Genetics 54, 663-676 (1966).
- - - (2) A molecular explanation of the bobbed mutants of Drosophila as partial defi-
ciencies of "ribosomal" DNA. Genetics 54, 819-834 (1966).
STERN, C.: Ein genetischer und zytologischer Beweis fr Vererbung in Y -chromosome von
Drosophila melanogaster. Z. Induktive Abstammungs- u. Vererbungslehre 44, 187-231
(1927).
Chapter XI
The Attainment of a Permanent Heterozygous Advantage

by the lncorporation of Two Former Alleles into the Genome
The heterozygous advantage, which benefits only certain members of a population,
can be fixed as a new species characteristic if two alleles involved are incorporated
into the genome as two separate gene loci. In this way, every member of a species
would come to enjoy the heterozygous advantage without ever having to produce
homozygotes which may be unfit.
Let us imagine a hypothetical species of fish which inhabits a long stretch of a
river. This river originates in a high Northern mountain and runs through a Southern
desert before it pours out to the ocean. Further, let us assume that this species is
endowed with two allelic alternatives at the gene locus for an enzyme; esterase (Es).
The Es A-variant specified by one allele has the temperature optimum of 5 C, while
the Es B-variant specified by the other allele functions best at 20 oc. It is expected
that for a subpopulation inhabiting the high Northern mountain part, natural
selection unconditionally favored the A-variant, so that the Subpopulation as a whole
has become A/A homozygous. Conversely, the B-variant has been favored in the
other subpopulation inhabiting the low Southern desert part, and the B should
have become the wild-type allele ofthat subpopulation. The water temperature in the
intermediate part of the river fluctuates widely with seasons; very cold in winter and
quite hot in summer. There is an unquestionable heterozygous advantage for
members of a subpopulation occupying the intermediate area. The AfB heterozygote
can cope with both the cold temperature of winter and the hot temperature of summer.
Yet as long as the A- and B-variants are specified by two alleles of the same gene
locus, only 50% of the subpopulation in the intermediate area can become hetero-
zygotes. Twenty-five per cent of the zygotes would be A/A which have difficult
summers, and the other 25%, which are BfB, would encounter serious problems of
survival during the winter months. Under these conditions, one would expect that
natural selection has favored the duplication of the Es-locus in the intermediate
subpopulation. When two alleles for the A- and B-variants are incorporated into the
genome as two separate gene loci, every member of the Subpopulation would become
ABJAB and enjoy a permanent heterozygous advantage without ever having to
produce undesirable homozygotes. A close approximation of this hypothetical
situation has apparently occurred in populations of the catostomid fish (Catostomus
clarki) which inhabit tributaries of the Colorado River system (KoEHN and RAs-
MUSSEN, 1967).
Whenever natural selection strongly selects against homozygotes, a duplication
which confers the heterozygous advantage to every member of a population must be
favored. Incorporation of two former alleles into the genome, however, contains the
germ of disaster. No gene functions alone; rather a group of genes perform inter-
related functions. For example, glucose-6-phosphate dehydrogenase (G-6-PD) and
6-phosphogluconate dehydrogenase (6-PGD) catalyze two successive steps of the
pentose phosphate shunt of carbohydrate metabolism. Once the activities of these
two enzymes of the species have been coordinated on the basis of a one-to-one gene
dosage ratio (two-to-two in diploid somatic cells), duplication of the 6-PGD locus
without concordant duplication of the G-6-PD locus might be disastrous. Thus, even
if there is strong natural selection against homozygotes at the 6-PGD locus, in-
corporation of two 6-PGD alleles into the genome might not be permitted, for the
disadvantage of disrupting the established gene dosage relationship with G-6-PD
might outweigh the advantage to be gained by the duplication. Hemoglobin cx- and
-chains are specified by two unlinked genes. Yet two alpha's and two beta's tagether
make up a single hemoglobin molecule. The fact that a duplication which incorporates
the normal -chain and mutant '-chain genes into the genome has not occurred in
African populations despite a strong heterozygous advantage might indicate that
having two doses of the -chain gene while maintaining only a single dose of the
cx-chain gene is incompatible with proper ontogenic development. In diploid organ-
isms, the gene dosage appears to be of prime importance. Otherwise, mammals
would not have developed the elaborate dosage compensation mechanism for X-
linked genes as discussed in Chapter III.
It becomes clear that despite the obvious benefit of attaining a permanent hetero-
zygous advantage, this type of duplication cannot always be favored. The concordant
duplication of all genes with interrelated functions which are scattered over different
chromosomes in the genome is accomplished only by becoming tetraploid. This is
The Differential Regulation of Former Alleles 67
the very reason we believe that polyploidy played just as important a role in the
evolution of vertebrates as it did in the evolution of higher plants. This point shall
be discussed in detail in Part 4.
References
KoEHN, R. K., RAsMUSSEN, D. I.: Polymorphie and monomorphic serum esterase hetero-
geniety in Catostomid iish populations. Biochem. Genet. 1, 131-144 (1967).
Chapter XII
The Differential Regulation of Former Alleles

and Their Transformation to Isozyme Genes
Nonconcordant duplication involving only one of a group of functionally
interrelated gene loci becomes permissible if the incorporation of two former alleles
ofthat locus into the genome is quickly followed by the development of the differen-
tial genetic regulatory mechanism. As this genetic regulatory mechanism permits only
one or the other former allele to engage in transcriptional activity in any given somatic
cell type of an individual, the original one-to-one gene dosage relationship is effec-
tively restored among all functionally related genes in spite of discordant duplication
which had affected one locus.
Once it is possible for an organism to discriminate between duplicated genes for
the same enzyme and use them differentially during ontogenic development, the way
is open for an organism to derive ultimate benefit from this type of gene duplication.
Because of differential use, the duplicated genes are exposed to different pressures of
natural selection. As a result, the two would gradually diverge from each other by
accumulating different kinds of mutations. Finally, the products of the duplicated
genes, although they still act upon the same substrate and use the same coenzyme,
acquire kinetic properdes which are markedly different from each other. In such a
way, a group of duplicated genes for the so-called isozymes must have been born.
Most vertebrates are endowed with at least two separate gene loci for A and B-
subunits of an enzyme, lactate dehydrogenase (LDH). These different subunits can
recognize each other as weil as themselves. Accordingly, by polymerization, five
tetramerk isozymes are formed between two kinds of subunits. They are A 4, A 3B,
~B 2 , AB3 and B4 (MARKERT, 1964). The fact that the products of two separate gene
loci maintain infinite affinity for each other suggests that the two arose from a
common ancestral gene by duplication. While subsequent natural selection could
easily have maintained the already existing affinity between the two subunits, it
would have been much more difficult to create an affinity between the products of
two separate gene loci which had no initial affinity.
Similarly, mammals and other vertebrates are endowed with three separate gene
loci for A, Band C-subunits of an enzyme; fructose diphosphate aldolase. In the case
of this enzyme, two subunits are seldom produced in the same tissue. For instance,
muscle cells produce only A-subunits while liver cells produce only B-subunits.
However, A- and B-subunits mixed in vitro randomly polymerize with each other and
form five tetramerk isozymes in the same manner as do A- and B-subunits of LDH
(PENHOET et al., 1966). Since the high affinity between A and B subunits is seldom
utilized by an organism, in the case of aldolase, this affinity probably reflects the
fact that the three separate gene loci for aldolase are duplicates derived from a single
gene locus.
What is the nature of the differences in kinetic properdes of the products of
these duplicated gene loci and how are these differences exploited by an organism
during ontogenic development? In the case of LDH which catalyzes the inter-
conversion of lactate and pyruvate accompanied by the interconversion of NADH2
and NAD, it has been shown that LDH-5 which is made exclusively of A-subunits
(A4) has a low affinity for pyruvate. That is to say, it functions most efficiently when
the substrate (pyruvate) concentration is around 10-3 M. In sharp contrast, LDH-1
which is made exclusively of B-subunits (B 4) has a high affinity; it functions best
when the pyruvate concentration is around 2 x 10-4 Mol (PLAGEMANN et al., 1960).
Needless to say, LDH-2, 3 and 4, which are hybrid molecules made of A- and B-
subunits, have intermediate kinetic properties. While pyruvate occupies the key
position in carbohydrate metabolism, lactate appears to serve no useful purpose in
high organisms other than as a temporary electron acceptor or oxidant during periods
when oxygen is in short supply. Thus, the most important function of LDH may not
be the reduction of pyruvate or the oxidation of lactate, but rather the regulation of
the ratio of NAD to NADH2 , since this ratio affects the rates of many catalytic
reactions (MARKERT, 1964).
From the above, it is easy to see that LDH-5 (A 4) is most useful to the tissues
which are anaerobic because of a relatively poor blood supply. During the rapid
metabolism of glucose, pyruvate production is enhanced and NAD is rapidly reduced
to NADH2 In the absence of oxygen, NADH2 cannot be reoxidized to NAD, and,
unless something is clone, glycolysis soon comes to a grinding halt. The presence of
low affinity LDH-5 (A4) enables NADH2 tobe reoxidized to NAD by the conversion
of pyruvate to lactate. This metabolic arrangement permits continued energy pro-
duction under the anaerobic condition even to the point of toxic accumulation of
lactic acid. On the other band, in the case of the tissues which are weil oxygenated by
an abundant blood supply, LDH-1 (B 4), which has a high affinity for pyruvate, is no
doubt the preferred type. For the tissues which are periodically exposed to the anaer-
obic condition, the concommitant presence of both A- and B-subunits would be
preferable, for the majority of LDH molecules produced would then be the hybrid
types; LDH-2, 3 and 4.
In the case of mammals, LDH-5 (A 4) indeed predominates in allfetal tissues, for
mammalian fetuses lead a rather anaerobic existence. Mter birth, LDH-5 remains
predominant in skeletal muscle which has a poor blood supply and where lactic acid
accumulates to an alarming degree after strenuous exercise. In the well oxygenated
tissues, notably the heart, the production of LDH-5 is suppressed after birth and the
active transcription and translation of the gene for B-subunits begins, so that LDH-1
predominates in the postnatal heart (MARKERT, 1964).
In the case of aldolase too, the kinetic property of the A 4-molecule must accom-
modate the unique metabolic requirement of skeletal muscle, while that of the B 4-
molecule fulfills the different metabolic requirement of hepatic cells (PENHOET et al.,
The Differential Regulation of Former Alleles 69
1966). It is becoming increasingly clear that in higher organisms such as vertebrates

the task of specifying one particular enzyme type is more often assigned to a group
of duplicated gene loci rather than to a single gene locus. Thus, man and other mammals
are endowed with at least two separate gene loci for pyruvate kinase (PK) (KoLER
et al., 1964) and with three unlinked gene loci for phosphoglucomutase (PGM)
(HARRIS et af., 1967).
Starting from a single fertilized egg, the body of vertebrates becomes a complex
organization made of hundreds of different kinds of somatic cell types, and no two
somatic cell types are identical with regard to their assigned functions. There is
little doubt that the type of gene duplication discussed in this chapter contributed
greatly to the attainment of such a complex body organization. Cells having identical
genetic material can differentiate into different somatic cell types only because the
genome contains a group of duplicated gene loci for each of many key enzymes.
Although duplicated genes of the group specify the same enzyme so far as the choice
of substrate and coenzyme is concerned, each gene product is unique with regard
to its Km as weil as its pH and temperature optimum. Because the choice from each
group of duplicates is offered, different somatic cell types acquire different characteris-
tics even with regard to the process of basic carbohydrate metabolism.
The production of specialized non-enzymatic proteins is also benefited from this
type of gene duplication. Bach immunoglobulin molecule is made of light-chains
(L-chains) and heavy-chains (H-chains), and L- and H-chains are specified by un-
linked groups of duplicated genes. Man and other mammals apparently carry only
two separate gene loci for u- and A.-classes of L-chains on one chromosome, and as
many as 10 separate gene loci for various classes ofH-chains on another chromosome.
Natural selection has permitted such gross discordance with regard to the degree of
duplication of L- and H-chain genes only because the genetic regulatory mechanism
which insures that only a single L-chain locus and a single H-chain locus shall engage
in transcription in each clone of antibody-producing plasma cells had previously
been evolved (PuTNAM et al., 1967; NATVIG et al., 1967; HERZENBERG et al., 1967;
PorTERand LIEBERMAN, 1967; CoHN, 1967; Hoon et al., 1968). Having a group of
functionally diverged genes for H-chains enables mammals to cope with all sorts of
contingencies. The IgA-type of antibody made up of an o.:-class H-chain is secreted
into the milk by the mother, and the maternally derived IgA protects newborns
while they themselves are still incapable of antibody production. IgM-type of 19S
antibodies, which are polymeric molecules, use the [J.-class of H-chains. IgM fixes
complements and contributes greatly to the body's initial reponse to newly encoun-
tered antigens. IgG-type of antihoclies use the y-class of H-chains, and this is the
cornerstone of the body's defense against invading foreign organisms. Furthermore,
IgG of the mother passes through the placenta and protects the embryo.
Although there are numerous examples of functionally diverged duplicated genes
in vertebrates, most duplications appear to have occurred in ancient times, so that
even rather remotely related species usually show the same degree of gene duplication
and the same differential use of these duplicated genes. lt is not often that one is
afforded with an opportunity to catch the type of gene duplication described in this
and the previous chapter in the middle of the act so to speak. Fortunately, duplication
of the 6-PGD locus which occurred in certain members of the fish family Cyprinidae
afforded us with such an opportunity.
Minnows, carp and goldfish comprise thls fresh watet family. While most of the
minnows are diploid spedes having the mean 2n nurober of 52, Barbus barbus of the
Rhlne River in Europe as weil as the familiar goldfish and carp appear to be tetraploid
spedes for they have more than 100 chromosomes. The genome of many diploid species
contains a single autosomally inherited gene locus for thls enzyme, and each species
maintains a nurober of alleles at thls locus whlch specify electrophoretic variants of
6-PGD. Since each 6-PGD molecule is a dimer, the homozygote such as A/A of the
diploid species gives a singleband of thls enzyme (~) upon starch gel electrophoresis,
while the heterozygote such as A/B shows three enzyme bands (~, AB and B2)
in the expected 1:2:1 ratio (Fig. 10). As a result of duplication ofthe entire genome,
the tetraploid species have apparently incorporated two former alleles of the 6-PGD
----- __
-- --
1
0
-=- ---...,. ! ! : - -CI!!lllt-
2 3 4 5 6 7 8 9
Fig. 10. The stareh gel plate stained for 6-PGD. Verdeal stareh gel eleetrophoresis was
earried out at pH 8.6 using a borate buffer. The starting point is indicated by zero, and the
anodal direetion is upward. 1, 2 and 3 illustrate the allelic polymorphism of a single gene locus
in a diploid species, Barbus tetrazona. (1) A single C2 band of a CJC homozygote. (2) Three
bands of an A/C heterozygote. A hybrid dimer band in the middle is intensely stained. (3) A
single A 2 band of an A/A homozygote. 4, 5, and 6 illustrate the alleHe polymorphism of one
of the two gene loci in the goldfish which is a tetraploid species. (4) Three bands of an A/A,
B2 /B 2 homozygote. (5) Six bands of an A/A, B 3 /B 2 homozygote. (6) Three bands of an A{A,
B3/B 3 homozygote. 7, 8 and 9 illustrate differences in the isozyme patterns of different tissues
of the goldfish which is homozygous A{A, B 2 /B 2 (7) Gills where A2 band at the bottom is
most eonspicuous. (8) Liver where B~ band at the top is accentuated. (9) Kidney where A 2
band is again most conspicuous
locus as two separate gene loci. Accordingly, the double homozygote such as AfA,
B/B of the tetraploid species shows the three enzymeband pattern reminiscent of the
diploid heterozygote. The tetraploid individual whlch is heterozygous at one of the
two gene loci such as AfA, B3fB2 now shows six 6-PGD bands; ~' AB 2, AB 3, B~,
B2B3 and B~ (Fig. 10). Thus, the tetraploid cyprinid fish have indeed attained a perma-
nent heterozygous advantage by the incorporation of two formet alleles into the
genome as two separate gene loci. Furthermore, when different tissues of the tetra-
ploid spedes are compared, it is noted that in certain tissues such as gills and kidney
there is a predominance of the A-subunits, while in other tissues such as the liver,
more B-subunits than A-subunits were apparently produced (Fig. 10). It appears
that these tetraploid spedes are on the way to developing the differential genetic
regulatory mechanism whlch discriminates between the two formet alleles. The gene
loci for A- and B-subunits of 6-PGD are indeed becoming the isozyme genes as
described in thls chapter (BENDER and HNO, 1968).
The Creation of a New Gene from a Redundant Duplicate 71
References
101-107 (1968).
CoHN, M.: Natural history of the myeloma. Cold Spring Rarbor Symposia Quant. Biol. 32,
211-222 (1967).
RARRts, R., RoPKINSON, D. A., LUFFMAN, J. E., RAPLEY, S.: Electrophoretic variation in
erythrocyte enzymes. In: Rereditary disorders of erythrocyte metabolism (BEUTLER, E.,
Ed.). City ofRope Symposium Series, Vol. 1, pp. 1-20. New York: Grune & Stratton
1967.
RERZENBERG, L. A., MINNA, J. D., RERZENBERG, L. A.: The chromosome region for immuno-
globulin heavy-chains in the mouse. Cold Spring Rarbor Symposia Quant. Biol. 32,
181-186 (1967).
Roon, L., GRAY, W. R., SANDERS, B. G., DREYER, W. J.: Light chain evolution. Cold Spring
Rarbor Symposia Quant. Biol. 32, 133-146 (1967) ..
KoLER, R. D., BrGLEY, R. R., JONES, R. T., RrGAS, D. A., V ANBELLINGHEN, P., THOMPSON, P.:
Pyruvate kinase: Molecular differences between human red cell and leukocyte enzyme.
Cold Spring Rarbor Symposia Quant. Biol. 24, 213-221 (1964).
MARKERT, C. L.: Cellular differentiation-An expression of differential gene function. In:
Congenital malformations, pp. 163-174. New York: The International Medical Congress
1964.
NATVIG, J. B., KuNKEL, R. G., Lrrwrn, S. P.: Genetic markers ofthe heavy-chain subgroups
ofhuman gamma G globulin. Cold Spring Rarbor Symposia Quant. Biol. 32, 173-180
(1967).
PENHOET, E., RAJKUMAR, T., RurrER, W. J.: Multiple forms of fructose diphosphate
aldolase in mammalian tissues. Proc. Nad. Acad. Sei. US 56, 1275-1282 (1966).
PLAGEMANN, P. G., GREGORY, K. F., WROBLEWSKI, F.: The electrophoretically distinct
forms of mammalian lactic dehydrogenase. II. Properdes and interrelationships of rabbit
and human lactic dehydrogenase isozymes. J. Biol. Chem. 235, 2282-2293 (1960).
PoTTER, M., LrEBERMAN, R.: Genetic studies of immunoglobulins in mice. Cold Spring
Rarbor Symposia Quant. Biol. 32, 203-209 (1967).
PUTNAM, F. W., TITANI, K., WrKLER, M., SHINODA, T.: Structure and evolution of kappa
and lambda light chains. Cold Spring Rarbor Symposia Quant. Biol. 32, 9-29 (1967).
Chapter XIII
The Creation of a N ew Gene from a Redundant Duplicate

of an Old Gene
The type of gene duplication whlch produced the group of isozyme genes contri-
buted greatly to the evolution of increasingly complex organisms. These functionally
diverged duplicated genes, however, still specify the same enzyme in that their
products act upon the same substrate with the help of the same coenzyme. A, B and
C-subunits of LDH of any vertebrate must still maintain either the identical active
site sequence of 12 amino acids (- Val-Ile-Ser-Gly-Gly-Cys-Asn-Leu-Asp-Thr-
Ala-Arg -), or a sequence very similar to the above, for thls is the sequence whlch
binds with NAD and recognizes pyruvate or lactate as the substrate (KAPLAN, 1965).
Dillerences in the kinetic property of A, B and C-subunits must reflect amino acid
substitutions whlch affect sites other than the active site of the polypeptide chain.
The type of gene duplication to be discussed in this chapter is different because it

contributed to the creation of new gene loci which acquired previously non-existent
functions. The ability to produce specific antihoclies against each of the variety of
antigens an organism comes in contact with is a unique attribute of vertebrates. No
invertebrate metazoans, no matter how complex their body organization might be,
can respond to foreign invaders in a specific way. It must be that the gene loci for
L- and H-chains of immunoglobulin were created de novo in the genome of ancestral
vertebrates. Y et, in a strict sense, nothing in evolution is created de novo. Bach new
gene must have arisen from an already existing gene, and only the accumulation of
forbidden mutations which result in changing the active site of the molecule can alter
the basic character of the gene. How can the gene locus be permitted to accumulate
forbidden mutations? When the event of gene duplication was followed by the develop-
ment of the differential genetic regulatory mechanism, the duplicated genes eventually
became the isozyme genes. If the event of gene duplication were followed by acquisi-
tion of the first forbidden mutation by one of the duplicates, what then? As the genome
still contains one functional gene locus for a particular function, the production of a
useless polypeptide chain by a mutated duplicate is absolutely harmless to an organ-
ism. In fact, the mutated duplicate has now become the redundant gene locus.
Natural selection would ignore the redundant locus, and thus, it is free to accumulate
a series of forbidden mutations, aided by intragenie recombination. As a result, the
polypeptide chain specified by it might finally acquire a function which is quite
different from that assigned to the original gene. In such a way, a series of new genes
with previously non-existent functions must have emerged during evolution. The
creation of a new gene from a redundant copy of an old gene is the most important
role that gene duplication played in evolution.
1. The Case of Trypsin and Chymotrypsin

Digestion of protein brought to the intestinal tract as food is carried out by two
principle proteolytic enzymes; trypsin and chymotrypsin. They are produced in the
pancreas and secreted to the intestinal tract through a duct. As active trypsin and
chymotrypsin would lyse and kill the pancreatic cells which produced them from
within, the two genes for trypsin and chymotrypsin actually specify longer, inert
polypeptide chains which are trypsinogen and chymotrypsinogen.
The essential difference between trypsin and chymotrypsin is in the preferred
sites at which these enzymes split the peptide bonds of protein. Trypsin attacks the
peptide bond at the carboxyl side of the basic amino acids lysine and arginine, while
chymotrypsin attacks the peptide bond at the carboxyl side of the aromatic amino
acids phenylalanine and tyrosine. There is little doubt that it is better to be endowed
with both of these proteolytic enzymes. The digestion of food protein would be most
inefficient if an organism was endowed with only one or the other.
Since these two proteolytic enzymes have distinctly different functions, one
might assume that the gene loci specifying trypsinogen and chymotrypsinogen
evolved independently of each other. When the complete amino acid sequence of
one is compared with that of the other, however, the homology between the two,
and therefore their common ancestry, becomes evident (Fig. 11), although the bovine
chymotrypsinogen A, which is made of 245 amino acids (KErL et al., 1963), is con-
CH 1 Cys G!y Val Pro Ala litt Gin Pro Val Lett Ser G!y Lett Ser Arg Ilu Val Asn Gly Glu
TRP 1 Val Asp Asp Asp Asp Lys Ilu Val Gly
CH 21 Glu Ala Val Pro Gly Ser Trp Pro Trp Gin Val Ser Leu Gin Asp Lys Thr Gly Phe His
TRP 10 Gly Tyr Thr Cys Gly Ala Asn Thr Val Pro Tyr Gin Val Ser Leu Asn Ser Gly Tyr His
CH 41 Phe CYS GLY GLY SER LEU ILU
ASN GLU ASN TRP VAL VAL THR ALA ALA CYS Gly Val
1--
TRP 30 Phe CYS GLY GLY SER LEU ILU
ASN SER GLN TRP VAL VAL SER ALA ALA - - Lys
CH 61 Thr Thr Ser Asp Val Val Val Ala Gly Glu Phe Asp Gin Gly Ser Ser Ser Glu Lys Ilu
~~~~~~~~~~~~~~~~~~ Gly Asn Gin Gin
CH 81 Gin Lys Leu Lys Ilu Ala Lys Val Phe Lys Asn Ser Lys Tyr Asn Ser Leu Thr Ilu Asn
TRP 70 Phe Ilu Ser Ala Ser Lys Ser Ilu Val His Pro Ser Tyr Asn Ser Asn Thr Leu Asn ~
CH 101 Asn Asn Ilu Thr Leu Leu Lys Leu Ser Thr Ala Ala Ser Phe Ser Gin Thr Val Ser Ala QrJ>
(I)
TRP 89 Asn Asp Ilu Met Leu Ilu Lys Leu Lys Ser Ala Ala Ser Leu Asn Ser Arg Val Ala Ser 0
.....,
CH 121 Val Cys Leu Pro Ser Ala Ser Asp Asp Phe Ala Ala Gly Thr Thr Cys Val Thr Thr Gly
TRP 109 Ilu Ser Leu Pro Thr Ser Cys Ala Ser Ala Gly Thr Gin Cys Leu Ilu Ser Gly Trp Gly ~"'
CH 141 Trp Gly Leu Thr Arg Tyr Thr Asn Ala Asn Thr Pro Asp Arg Leu Gin Gin Ala Ser Leu s
TRP 129 Asn Thr Lys Ser Ser Gly Thr Ser Tyr Pro Asp Val Leu Lys Cys Leu Lys Ala Pro Ilu
CH 161 Pro Leu Leu Ser Asn Asn Lys Lys Tyr Trp Gly Thr Lys Ilu Lys Asp Ala Met
~
Thr Cys (")
TRP 149 Leu Ser Asn Ser Ser Cys Lys Ser Ala Tyr Pro Gly Gin Ilu Thr Ser Asn Met Phe Cys ~
CH 181 Ilu Cys Ala Gly Ala Ser Gly Val Ser Ser CYS MET GLY ASP ~GLY GLY PROLEUVAL
TRP 169 Ala Gly Tyr Leu Glu Gly Gly Lys Asn Ser CYS GLN GLY ASP ~GLY GLY PRO VAL VAL ~
CH 201 CYS Lys Lys Asn Gly Ala Trp Thr Leu Val Gly Ilu Val Ser Trp Gly Ser Ser Thr Cys "'s
TRP 189 CYS Ser Gly Lys Leu Gin Gly Ilu Val Ser Trp Gly Ser Gly Cys Ala Gin Lys Asn Lys
CH 221 Ser Thr Ser Thr Pro Gly Val Tyr Ala Arg Val Thr Ala Leu Val Asn Trp Val Gin Gin
TRP 209 Pro Gly Val Tyr Thr Lys Val Cys Asn Tyr Val Ser Trp Ilu Lys Gin Thr Ilu Ala Ser
CH 241 Thr Leu Ala Ala Asn
TRP 229 Asn
Fig. 11. The complete amino acid sequence of bovine chymotrypsinogen A is compared to that of bovine trypsinogen. The sequences in italics
from the amino ends indicate the portians which are split off when activated. The active site sequence in the histidine loop as weil as the active
--.]
site sequence areund the critical serine are capitalized t>J
siderably Ionger than the bovine trypsinogen, which is made of 229 amino acids
(KAuFFMAN, 1965). When activated to become chymotrypsin, the amino end peptide
made of 15 amino acids is split off from the chymotrypsinogen. Thus, the chymotryp-
sin itself is made of only 230 amino acids. In the case of the trypsinogen, on the other
hand, the activation splits off the amino end peptide made of only six amino acids.
Thus, the trypsin itself is made of 223 amino acids. The difference in length now
involves only seven amino acid residues. As already shown in Fig. 5 (Chapter V),
chymotrypsinogen has five disulfide bridges and the trypsinogen has six. Because of
these disulfide bridges, both molecules assume a rather similar shape. The center of the
first active site of chymotrypsinogen is histidine at position 57, while serine, at
position 195, is the center of the second active site. The corresponding centers are
found at position 46 (histidine) and position 183 (serine) of trypsinogen. In both
enzymes, the active histidine is part of a loop (16 amino acids long) made by a disulfide
bridge formed between cysteines at positions 42 and 58 in the case of chymotrypsino-
gen and at positions 31 and 47 in the case of trypsinogen. The active serine, too,
is part of another loop in both enzymes, formed by a disulfide bridge between
positions 191 and 220 in chymotrypsinogen and between positions 179 and 203 in
trypsinogen. In the actual three dimensional configuration which both molecules
assume, the active histidine and the active serine might be facing each other. When
the amino acid sequences of the two active sites of chymotrypsinogen and trypsinogen
are compared, it becomes clear that their active sites are similar, but distinctly
different from each other (NEURATH et al., 1967). Chymotrypsinogen and trypsinogen
differ at three of the sixteen sites of the histidine loop, and they differ at two of the
eleven sites which surround the active serine (Fig. 11).
There is little doubt that the gene locus for one enzyme evolved from a redundant
duplicate of the other enzyme gene by the accumulation of forbidden mutations
which affected the active sites. Inasmuch as the activation of not only trypsinogen
itself but also chymotrypsinogen is carried out by already activated trypsin, one is
inclined to believe that the cistron for trypsinogen was the ancestral gene and that
the gene for chymotrypsinogen was created from a redundant duplicate of the trypsin-
ogen gene. It should be noted in Fig. 11 that trypsinogen is split at the carboxyl
end of lysine at the 6th position, while chymotrypsinogen is split at the carboxyl end
of arginine which occupies the 15th position. The basic amino acids, lysine and
arginine, are preserved at these sites to offer the preferred site to trypsin.
Just as convergent evolution has occurred with regard to allelic mutations of the
homologous gene locus (Chapter VI), convergent evolution must have occurred
with regard to gene duplication. As long as divergent organisms have the homologous
gene locus which can serve as an ancestor, duplication can independently create a new
gene locus with the same basic characteristics. The fact that butterfly larvae apparently
possess both trypsin-like and chymotrypsin-like enzymes can be taken as an example
of convergent evolution.
2. The Protein of Microtubules and Actin of the Skeletal Museie

The protein subunit of microtubules must be one of the older proteins in an
evolutional sense, for microtubules are the components of bacterial flagella and
mitotic spindie of all the eukaryotes. In vertebrates, aside from mitotic spindle, the
The Protein of Microtubules and Actin of the Skeletal Museie 75
cilium of certain epithelial cells, the tail of spermatozooa, new:ofilaments of the nerve
axon and sarcotubules of the skeletal muscle utilize these protein subunits.
It has been shown that this negatively charged (acidic) protein is a dimer having
the molecular weight of 120,000. It assumes a globular shape and the subunit peptide
chain specified by the gene which should be about 600 amino acid residues long
contains 150 to 180 glutamic and aspartic acids. The universal characteristics of this
protein, regardless of the source it is derived from, are: 1. it binds specifically with
colchicine, and 2. it also binds with GTP (guanosine triphosphate) (SHELANSKI and
TAYLOR, 1968; WEISBNllERG et al., 1968). It is not surprising that colchicine, which is
best known as a disruptor of the mitotic
spindie (EmsTI and DusTIN, 1955), also
interferes with a variety of cellular func-
tions dependent upon microtubules (OK.A-
ZAKI and HOLTZER, 1965; TILNEY et al.,
1966). It is of extreme interest to find out
if the genome of vertebrates contains a
group of duplicated gene loci rather than
a single gene locus for this protein of
microtubules.
It appears that the gene for actin of
muscle developed from a redundant dupli- ~
cate of the gene for the protein of micro- ~
tubules. A striated skeletal muscle fiber
contains about 20% protein by weight. The
contractile portion of it consists almost
exclusively of two structural proteins;
myosin and actin. Myosin is extracted with Fig. 12. Tbc microtubule at thc top is
a 0.3 Mol KCI solution. It has the molec- compared to the sliding filarnent model of
ular weight of about 450,000 and it ag- the actomyosin complex at the bottom.
Both the colchicine-binding protein of the
gregates to form a spindle-shaped element microtubule and the actin of the actomyo-
with a considerable number of side pro- sin complex are indicated by circlcs. The
jections. Actin, on the other hand, is less diametcr of the tubule shown at the left is
soluble in KCl, as it takes 0.6 Mol KCl about 24<) A, while the unit of actomyosin
to extract actin. In the absence of salts, complex shown at the right is about 100 A
long
actin becomes globular (G-actin) with a
molecular weight of about 70,000. When
the isolated actin and myosin are put tagether in a test tube they form a complex
called actomyosin which contracts in the presence of A TP (adenosine triphos-
phate). The spatial arrangement of actin and myosin molecules in the actomyo-
sin complex is schematically illustrated in Fig. 12 (SZENT-GYRGYI, 1957).
Comparison between the protein of microtubules and actin reveals their similarity
and at the same time significant differences. The molecular weight of the monomeric
actin, which is 70,000, compares favorably with that of the subunit of the micro-
tubular protein, which is 60,000. The amino acid compositions and electrophoretic
mobilities of both proteins are very similar and each possesses a nucleotide-binding
site {RENAUD et a/., 1968). On the other band, the actin binds with ATP, whereas the
microtubular protein binds with GTP, and while the microtubular protein binds
with colchicine, the actin does not (BoRrSY and TAYLOR, 1967). The actin contains one
residue per mole of the rare amino acid 3-methylhistidine, whereas, the colchicine-
binding protein does not. There is little doubt that the gene for actin evolved from
the duplicate of the gene for the subunit of the mkrotubular protein by the accumula-
tion of forbidden mutations which resulted in changes in the active sites.
3. Myoglobin and Hemoglobin

While vertebrates from the most primitive hagfish to man are endowed with
hemoglobin molecules, urochordates and cephalochordates lack hemoglobin. It is
likely that so far as vertebrates are concerned the earllest gene for a hemoglobin poly-
peptide came into being at the onset of vertebrate evolution. On the other hand, the
existence of hemoglobin in diverse invertebrate species, which is a good example of
convergent evolution, suggests that a primordial gene capable of becoming a gene
for hemoglobin has long been present in the animal kingdom.
It is the view of INGRAM (1963) that this primordial gene specified a monomeric
heme-containing protein. Of the heme-containing proteins, the type exemplified by
cytochrome C can not be the ancestor of hemoglobin, for a heme in cytochrome C
is attached to the two cysteines closest to the amino end (MARGOLIASH, 1963). The
gene for the first vertebrate hemoglobin, on the other hand, could easily have
evolved from the gene for myoglobin. X-ray diffraction studies have shown that a
myoglobin peptide chain and a hemoglobin peptide chain are folded araund the
heme group in a nearly identical way as already illustrated in Fig. 4 (Chapter V)
(PERUTZ et al., 1960; KENDREW et al., 1960). In both polypeptides, a heme is attached
to two histidines spaced weil apart from each other. In myoglobin, histidines at
positions 64 and 93 hold a heme, while in hemoglobin IX-chain, the points of attach-
ment to a heme are represented by histidines at the 58th and 87th positions (see
Fig. 6, Chapter VI). The cardinal difference between myoglobin and hemoglobin is
that myoglobin remained a monomer, while hemoglobin became a tetramerk mole-
cule. The ability of hemoglobin peptide chains to recognize each other and form a
tetramerk molecule appears to have been acquired by a deletion or deletions. The
-like hemoglobin chains of vertebrates are made of 145 or 146 amino acids, while
myoglobin of the sperm whale is 153 residues long (EnMUNDSON, 1965). This deletion
appears to have occurred after the development of a jaw by vertebrates. Monomeric
hemoglobin is still found in the hagfish and the lamprey which represent the most
primitive jawless state in vertebrate evolution. An adult hemoglobin peptide chain
of the lamprey has, in fact, been shown tobe 156 residues long (RunLOFF et al., 1966).
Although the sequence of the lamprey hemoglobin shows 120 differences from that
of the sperm whale myoglobin, if the lamprey hemoglobin sequence were compared
to that of its own myoglobin, there may be only a few differences. In fact, so far as
the hagfish and the lamprey are concerned, the gene for myoglobin and that for
hemoglobin should be regarded as isozyme genes. In this connection, it is of interest
to note that the genome of the hagfish appears to contain several gene loci for
monomerk hemoglobin (OHNO and MoRRISON, 1966). Although there has been no
shortage of duplication in the hagfish, these duplicated genes remained as isozyme
genes of a group. None of these gene lod rose above the others by accumulations of
{orbidden mutations.
L- and H-chains ofimmunoglobulin 77
The next step in the evolution of vertebrate hemoglobin appears to have been the
emergence of -like chains whkh are 145 or 146 residues long. The y-chain or fetal
hemoglobin as weil as the c5 and other chains of minor adult hemoglobin belong to
this dass. A deletion or deletions occurred to the DNA cistron for monomerk
hemoglobin and this change gave rise to the genes which specify the -like chains.
This new chain acquired the ability to polymerize in a rather indiscriminate way, for it
was able to form autotetramers; such as 4 and y 4 When hemoglobin of many teleost
fish species was subjected to electrophoresis, it was not uncommon to observe five
evenly spaced hemoglobin bands on the gel plate. The five band pattern thus observed
was strikingly similar to that presented by the five tetramerk isozymes of LDH
which are made by indiscriminate association of A- and B-subunits specified by two
duplicated gene loci (Chapter XII). It appears that many teleost fish are endowed
with two or more duplicated gene loci for -like chains, and these -like chains can
recognize each other as well as themselves thereby forming tetramerk molecules of
all possible combinations.
The last step in the evolution of vertebrate hemoglobin appears to have been the
creation of the .x-chain. The gene whkh specifies .x-chains was no doubt derived from
one of the duplicated genes for the -like chains. This change has been accomplished
by further deletion of the DNA cistron, for .x-chain is only 141 amino acid residues
long.
The .x-chains cannot form tetramerk hemoglobin molecules by themselves, but
tagether with the -like chains, they form the most efficient hemoglobin molecules in
existence; they are .x2 2 , .x2y 2 and .x2c52 The sequence of the .x-chain and that of the
-chain of the same mammalian species differ in at least 48 sites as in sheep, and, at the
most, at 84 sites as in man and the horse. This great difference is in agreement with
the view that the creation of the .x-chain gene from a redundant duplicate of the
-like chain gene occurred long before the emergence of the first mammals on this
earth. The presence of hemoglobin .x-chain has been noted in certain teleost fish such
as the carp (HILSE et al., 1966).
4. L- and H-chains of Immunoglobulin

The ability of vertebrates to produce immunoglobulin molecules is a unique
property. The gene loci specifying L-and H-chains of these molecules are traceable to
ancient ostracoderm fish, since the survivors of the mostprimitive jawless vertebrates
include the lamprey and the hagfish, both of which produce IgM type of immuno-
globulin (PAPERMASTER et al., 1962; HILDEMANN and THOENES, 1969).
Higher vertebrates produce three major types of immunoglobulin subunits:
the L-chain which is 210 to 220 amino acid residues long, the H-chain whkh is
about 450 residues long, and a special [L-heavy-chain which is 650 or so residues long.
There aretetramerk and polymerk forms of immunoglobulins. IgG, IgD and IgE
dasses are represented by tetramerk molecules each consisting of two identkal
L-chains and two identkal H-chains (LH) 2 , as schematkally illustrated in Fig. 13.
The IgM dass, on the other hand, consists of polymerk molecules; a Ionger [L-heavy-
chain is used in place of the ordinary H-chain to combine with the L-chain (L2 [L2)n.
The cistrons which specify the subunits of immunoglobulin molecules have the
special ability of providing immensely variable amino acid sequences to the poly-
peptide chain they produce. Two antihoclies of distinct specificity produced by the
same individual differ by the contiguous, variable amino acid sequence of the L-chain
and H- or [L-heavy-chain. Mammals are capable of producing several thousand kinds
of antihoclies directed against the numerous antigens to whlch they are exposed.
These great variations in amino acid sequences whlch determine the antibody
specificity of individual immunoglobulin molecules appear to be brought about by
substitutions whlch occur in a unique region of the L- and H-chains. In the case of
L-chains, thls region is 105 amino acid residues long, starting from the amino end.
L (k) H H L(k) Ukl fV
ss ss ss ss
ss
Xn
Fig. 13. Sehemadeillustration of immunoglobulin molecules. Left: A tetrameric molecule

(7S) of IgG, IgD and IgE dass. Two L-chains are bound to two H-chains by disulfide
bridges. The length of the H-chain is twice that of the L-chain. The variable region, con-
sisting of 105 residues from the amino end, is painted solid black. Since this L-chain is
drawn to have cysteine at the carboxyl end, it belongs to the x-dass. Middle: A dimeric unit
of polymeric molecules of IgM dass. A special,u-heavy-chain used by IgM has three times
the length of the L-chain. Right: A more realistic representation of the L-chain. An obvious
symmetry that exists between the variable and constant regions is a reflection of the internal
homology
If this amino half region of the L-chain includes 40 variable sites, and if each site has
an alternative to choose one of three amino acids, the regionwill generate 340 different
variable sequences whlch are more than sufficient to account for the many different
kinds of antihoclies produced. The remaining region, nearer to the carboxyl end of an
L- or H-chain, does not have a variable site, and this region is defined as the constant
region.
The task of antibody production is assigned to a particular type of blood cell
known as a plasma cell. These plasma cells are transformed lymphocytes, and the
embryonie origin of lymphocytes can be traced to the thymus and other organs,
L-and H-chains oflmmunoglobulin 79
such as the bursa of Fahridus of avian species (GucK et al., 1956; MrLLER, 1961).
Each plasmablast and its descendants are committed to produce only one fixed
L-chain sequence and one fixed H-chain sequence. To put it another way, all the anti-
body molecules produced by a plasma cell clone are identical to each other. Therefore,
two separate plasma cell clones almost never produce an identical antibody. The
variable sequences are apparently generated by a special mechanism which affects the
40 or so sites already mentioned. This special mechanism shall be discussed in the
last chapter of this book. A particular antigen invading the body serves as the stimulus
to a lymphocyte or lymphocytes which happened to have acquired the ability to
generate the particular L-chain and H-chain sequences that can recognize an invading
antigen. Such a lymphocyte or lymphocytes respond to the stimulus by transforming
to plasmoblasts. Thus, by proliferation, the clones of antibody producing plasma
cells are formed. This is the clonal selection theory of antibody response (BuRNET,
1958).
Man, mice, rabbits, and presumably most other mammals, have at least two
separate gene loci which specify x- and A.-L-chains. Both chains are about 210 to
220 amino acid residues long; the difference in length is due to the fact that plasma
cell clones produce L-chains with variable regions of different length. The mammalian
x-chain has cysteine as the carboxyl terminal, while the A.-chain has one more amino
acid (serine in the case of man) after cysteine. Complete amino acid sequences have
been determined on several x- and A.-chains produced by monoclonal myelomas of
man and mice (HILscHMANN and CRAIG, 1965; MrLSTEIN, 1966; PoRTER, 1966;
BAGLIONI, 1967; GRAY et al., 1967). The most striking result suggested by a sequential
study of L-chains is that if substitutions at the variable sites identified with antibody
specificity are discounted, the L-chain itself shows internal homology. Various amino
acids occupying non-variable sites within the variable region (from the amino end to
the 105th position) are often found in the corresponding sites in the constant region
(from the 106th position to the carboxyl end which can be positions 212 to 218)
(LENNOX and CoHN, 1967). Cysteine occupies the 22nd and 87th positions in the
variable region and the 132nd (105 + 27) and 192nd (105 + 87) in the constant region
of a human x-chain. This indicates that an original cistron for L-chain arose by
tandem fusion of two duplicates as shown in Fig. 13 (MILSTEIN, 1966; BAGLIONI,
1967).
It is generally believed that mammals possess nearly 10 duplicated gene loci
(isozyme-like genes) for H-chains oflgG, IgA, IgD, IgE and IgM classes of immuno-
globulin. For instance, four closely linked gene loci specifyyt, yz, y3, and y4 H-chains of
IgG, and yet another two specify .xt and .x2 H-chains of IgA. The available data on
partial sequences of H-chains which are ab out 450 amino acid residues long indicates
that an original cistron for the H-chain arose by tandem fusion of two L-chain
genes (DooLITTLE et al., 1966). The !L-chain of IgM is about 200 amino acid residues
Ionger than other H-chains. For this reason, the gene locus for !L-chain can be con-
sidered a fused triplicate of the L-chain gene (WrKLER et al., 1969).
There remains little doubt that the three kinds of immunoglobulin genes for L-
and H-chains share the ultimate common ancestor in a redundant duplicate of the
gene which specified a polypeptide chain about 110 amino acid residues long. The
suggestion has been made that this locus might have been that which specified hapto-
globin .x-chain (BLACK and DrxoN, 1968). Haptoglobin, which isapolymer made of
two subunits (iX and -chains), is present in serum and its function is to combine with
free hemoglobin liberated by hemolysis and make it accessible to degradative enzy-
mes. The first duplication, which perhaps placed a redundant duplicate of the hapto-
globin lX-chain gene on another chromosome, and the second duplication which
placed two redundant duplicates adjacent to each other on the same chromosome
appear to have occurred in the ancient ostracoderm fish of 300 million years ago.
The subsequent fusion of adjacent duplicates created the first gene locus for L-chain,
while still preserving the gene locus for haptoglobin lX-chain. The second immuno-
globulin gene to emerge in vertebrate evolution appears to have been that for [1.-
heavy-chain. MARCHALONIS and EDELMAN [1966 (1)] have shown that the shark
produces only an IgM dass (L2 [L2)n of immunoglobulin. As the [1.-heavy-chain gene
could not have been created at the expense of the already existing L-chain gene, this
process would have required three duplications. The first must have placed a re-
dundant duplicate of the L-chain gene on a different chromosome; then, two succes-
sive duplications must have placed three redundant duplicates in tandem on the same
chromosome, and subsequent fusion of the three created the gene locus which
specifies [1.-heavy-chain. It appears that by the time vertebrates evolved to the am-
phibian stage, they were already endowed with all three kinds of immunoglobulin
genes for L-chain, [1.-heavy-chain and y-H-chain, for the frog produces IgG as well as
IgM dass of antihoclies [MARCHALONIS and EDELMAN, 1966 (2)].
5. The Emergence of a New Gene by a Frame-shift Mutation

In the case of the four examples cited above, the new gene preserved enough
homology with the ancestral gene so that despite a change in the active site the phylo-
genic relationship between the two genes became evident when various parameters
of their gene products were compared. What would have happened if the first
mutation suffered by a redundant duplicate was aframe-shift mutation (Chapter IV)?
When a messenger RNA transcribed from a mutant cistron was translated, a grow-
ing peptide chain would have received the proper amino acid sequence until the
point of deletion or insertion. However, from then on to the carboxyl end, the
arnino acid sequence would have been completely alteredas mentioned in Chapter IV.
Although a frame-shift mutation, as a rule, leads to premature chain termination
because of inevitable conversion of some codons to nonsense codons, nonsense codons
can be changed back to amino acid specifying codons by subsequent mutations. Thus,
starting with a frame-shift mutation, a duplicate might acquire a new function which is
totally different from that assigned to the original gene. Admittedly, this is a one in a
rnillion chance, but in evolution, events with odds of one in one million occurred time
and again. In such a case, the knowledge of neither their assigned functions nor their
complete amino acid sequences would readily reveal the common ancestry between
the original gene and a new gene derived from its deleted or inserted duplicate.
References
BAGLIONI, C.: Homologies in the position of cysteine residues of K and L type chains of
human immunoglobins. Biochem. Biophys. Res. Commun. 26, 82-89 (1967).
BLACK, J. A., DrxoN, G. H.: Amino-acid sequence of alpha chain of human haptoglobins.
Nature 218, 736-741 (1968).
References 81
BoRISY, G. G., TAYLOR, E. W.: The mechanism of action of colchicine: binding of col-
chicine-3H to cellular protein. J. Cell Biol. 34, 523-533 (1967).
BuRNET, F. M.: The clonal selection theory of acquired immunity. London: Cambridge
Univ. Press 1958.
DooLITTLE, R. F., SrNGER, S. J., METZGER, H.: Evolution of immunoglobulin polypeptide
chains: Carboxyterminal of an IgM heavy chain. Science 154, 1561-1562 (1966).
EnMUNDSON, A. B.: Amino-acid sequence of sperm whale myoglobin. Nature 205, 883-887
(1965).
EIGSTI, 0. ]., DusTIN, P., ]R.: Colchicinein agriculture, medicine, biology and chemistry.
Ames (Iowa): Iowa State College Press 1955.
GucK, B., CHANG, T. S., ]AAP, R. G.: The bursa of Fahridus and antibody production.
Poultry Sei. 35, 224-234 (1956).
GRAY, W., DREYER, W., Hoon, L.: Mechanism of antibody synthesis: Size difference be-
tween mouse kappa chains. Science 155, 465-467 (1967).
HILDEMANN, W. H., THOENES, G. H.: Immunological response of Pacific hagfish. I. Skin
transplantation immunity. Transplantation 7, 506-529 (1969).
HILSCHMANN, N., CRAIG, L. C.: Amino acid sequence studies with Bence-Jones proteins.
Proc. Natl. Acad. Sei. US 53, 1403-1409 (1965).
HILSE, K., SoRGER, U., BRAUNITZER, G.: Zur Phylogenie des Hmoglobinmolekls ber
den Polymorphismus und die N-terminalen Aminosuren des Karpfenhmoglobins.
Z. physiol. Chem., Hoppe Seyler's 344, 166-168 (1966).
INGRAM, V. M.: The hemoglobin in genedes and evolution. New York: Columbia University
Press 1963.
KAPLAN, N. 0.: Evolution of dehydrogenases. In: Evolving genes and proteins (BR YSON, W.,
VoGEL, H. ]., Eds.). New York: Academic Press 1965.
KAUFFMAN, D. L.: The disulphide bridges of trypsin. J. Mol. Biol. 12, 929-932 (1965).
KEIL, B., PRUSIK, Z., SoRM, F.: Disulphide bridges and a suggested structure of chymo-
trypsinogen. Biochim. et Biophys. Acta 78, 559-561 (1963).
KENDREW, J. C., DICKERSON, R. E., STRANDBERG, B. E., HART, R. G., DAVIES, D. R.,
PHILLIPS, D. C., SHORE, U. C.: Structure of myoglobin: A three-dimensional fourier
synthesis at 2A resolution obtained by X-ray analysis. Nature 185, 422-427 (1960).
LENNOX, E., CoHN, M.: Immunoglobin genetics. Ann. Rev. Biochem. 36, 365-406
(1967).
MARCHALONIS, ]., EDELMAN, G. M.: (1) Polypeptide chains of immunoglobulins from the
smooth dogfish (Muste/us canis). Science 154, 1567-1568 (1966).
- - (2) Phylogenetic origins of antibody structure. II. Immunoglobulins in the primary
immune response of the bullfrog, Rana catesbiana. ]. Exptl. Med. 124, 901-913
(1966).
MARGOLIASH, E.: Primary structure and evolution of cytochrome C. Proc. Natl. Acad. Sei.
us 50, 672-679 (1963).
MILLER, J. F. A. P.: Immunological function ofthe thymus. Lancet 1961 II, 748-749.
MrLSTEIN, C.: The disulphide bridges of immunoglobulin 11:-chains. Biochem. J. 101, 338-
351 (1966).
NEURATH, H., WALSH, K. A., WINTER, W. P.: Evolution of structure and function of
proteases. Science 158, 1638-1644 (1967).
0HNO, S., MoRRISON, M.: Multiple gene loci for the monomerk hemoglobin of the hagfish
( Eptatretus stoutii). Science 154, 1034-1035 (1966)
0KAZAKI, K., HoLTZER, H.: Aspects of myogenesis in vitro. J. Cell Biol. 27, 75A (1965).
PAPERMASTER, B. W., CoNDIE, R. M., FINSTAD, ]., Goon, R. A.: Immuneresponse in the
California hagfish. Nature 196, 355-356 (1962).
PERUTZ, M. F., RossMANN, M. B., CuLLIS, A. F., MurRHEAD, H., WILL, G., NoRTH, A. C. T.:
Structure of hemoglobin: A three dimensional fourier synthesis at 5.5 A resolution,
obtained by X-ray analysis. Nature 185, 416-422 (1960).
PoRTER, R. R.: A discussion of the chemistry and biology of immunoglobulins. Proc. Roy.
Soc. (London), B 166, 113-243 (1966).
RENAUD, F. L., RowE, A. ]., GIBBONS, I. R.: Some properdes of the protein forming the
outer fibers of cilia. J. Cell Biol. 36, 79-90 (1968).
RuDLOFF, V., ZELENIK, M., BRAUNITZER, G.: Zur Phylogenie des Hmoglobinmolekls,
Untersuchungen am Hmoglobin des Fluneunauges ( Lampefra jluviatilis). Z. physiol.
Chem., Hoppe Seyler's 344, 284-288 (1966).
SHELANSKY, M. L., TAYLOR, E. W.: Properries of the protein subunit of central-pair and
outer-doublet microtubulues of sea urehin flagella. J. Cell Biol. 38, 304-315 (1968).
SZENT-GYRGYI, A.: Chemistry ofmuscular contraction. New York: Academic Press 1957.
TILNEY, L. G., HrRAMOTO, Y., MARSLAND, G.: Studies on the microtubules in heliozoa.
III. A pressure analysis of the role of these structures in the formation and maintenance
of the axopidia of Actinosphaerium nucleoftlum (BARRETT). J. Cell Biol. 29, 77-95 (1966).
WEISENBERG, R. C., BoRISY, G. G., TAYLOR, E. W.: The colchicine-binding protein of
mammalian brain and its relation to microtubules. Biochem. J. 7, 4466-4479 (1968).
WrKLER, M., KHLER, H., SHINODA, T., PUTNAM, F. W.: Macroglobulin structute: Homology
ofMu and gamma heavy chains ofhuman immunoglobulin. Science 163,75-78 (1969).
ChapterXIV
Duplication of Regulatory Genes and Receptors

The manner in which an organism is going to utilize functionally diverged
duplicated genes is largely dependent upon whether or not duplicated structural
genes can be placed under separate regulatory mechanisms. Although our ignorance
on the genetic regulatory mechanism is profound, a discussion on the subject of
gene duplication should by necessity consider duplication of regulatory genes.
In embryology, the term "undifferentiated" has traditionally been used to de-
scribe the state of an early embryonie nuclei. This usage has an unfortunate eon-
notation because indiseriminate transeriptional aetivity by every structural gene in
the nucleus is implied. Aeeordingly, the subsequent differentiation ofvarious somatic
eell types has often been viewed as being due to the progressive inactivation of indi-
vidual struetural genes.
The fact is that the fertilized egg begins its development with only a fraetion of
the genome turned on. Undifferentiated eells of early embryos do not require many
gene produets for they merely multiply without performing specialized functions.
Not only are these gene products already stored in the egg eytoplasm by the mother,
but it also appears that the first set of genes in the embryonie nuclei to engage in
transeriptional aetivity are the same kind of genes which aetively eontributed to
matemal storage during ovogenesis. Utilizing the teehnique of DNA-RNA hybridi-
zation, it has been shown in the sea urehin that messenger RNA synthesized de novo by
embryonie nuclei of blastulae are of the same kind as those stored in the egg eyto-
plasm by the mother (WHITELEY et al., 1966; SLATER and SPIEGELMAN, 1966).
Quite clearly, embryonie differentiation is the proeess of seleetively aetivating
the silent genes rather than progressively silencing the aetive ones. Unless something
is clone, the DNA strand of the eukaryote ehromosome is indiseriminately blanketed
by histones (basic proteins) as mentioned in Chapter III. Itfollows that the genetic
regulatory mechanism of the eukaryotes has to exercise activating control, because
in order to permit transeription by eertain struetural cistrons, the regulatory gene
produet has to remove histones seleetively from these portions of the DNA strand,
Hierarchy of Regulatory Mechanisms 83
so that RNA polymerase can attach itself to DNA and begin transcription. This is a
fundamental difference between the eukaryote and the prokaryote. Inasmuch as the
prokaryote does not possess the indiscriminate repressor (histones), the genetic
regulatory mechanism, as a rule, exercises a repressive control as exemplified by the
/ac-operon system of E. coli (]AcoB and MoNon, 1961).
JACOB and MoNOD (1963) defined differentiation as " ... two cells are differen-
tiated with respect from one another if, while they harbor the same genome, the
pattern of proteins which they synthesize is different."
1. Hierarchy of Regulatory Mechanisms

Needless to say, there is a hierarchy of regulatory mechanisms. First of all, the
primary transcriptional control determines whether or not a given gene locus is to
be turned on in a given somatic cell type. Somatic differentiation in ontogenic develop-
ment is primarily under this type of control. It appears that the developmental deci-
sion made in this respect is more often than not irreversible, for once a certain group
of progenitor cells have committed themselves to be liver cells, their descendants
would forever remain liver cells.
Once a particular locus has been turned on in a specific somatic cell type, an
organism has the choice of either leaving it alone so that the gene product will be
made all the time in a constitutive manner or placing it under a secondary repressive
control so that the gene product is made only in the presence of an inducer. lt appears
that the gene locus for tyrosine aminotransferase is turned on in mammalian liver.
However, this enzyme is made only in the presence of an adrenal steroid hormone
(ToMKINS et al., 1966). The transcription and translation of the already turned on
locus is blocked until the repressor is removed by an inducer. Analogy to the modus
operandi of the /ac-operon system of E. coli is found not in the primary control, but
in the second rank repressive control of vertebrates. In both cases, the regulatory
protein specified by an independent gene locus is a two-headed monster in that it has
two separate binding sites. On one hand, it recognizes a particular base sequence of
nucleic acid, and, on the other hand, it recognizes an inducer molecule. In the case
of the /ac-operon, the regulatory protein, which is a large acidic protein having the
molecular weight of about 160,000, recognizes and binds with a stretch ofDNA base
sequence which comprises an operator portion of the /ac-operon. This binding effec-
tively inhibits the transcription, so that a polycistronic messenger RNA cannot be
copied from the three enzyme loci in the /ac-operon. When an inducer, which is a
metabolite of lactose such as IPTG, binds with this regulatory protein, however, its
allosteric configuration changes in such a way that it no Ionger recognizes the base
sequence of the operator (GrLBERT and MLLER-HrLL, 1966; RIGGSand BouRGEOIS,
1968; RrGGS et al., 1968). In this way, the structural genes concerned with lactose
metabolism are turned on only when E. coli encounters lactose.
Also in the case of the mammalian tyrosine aminotransferase system, until the
binding with an inducer changes its allosteric configuration, the regulatory gene
product must recognize and bind with a specific base sequence of nucleic acid to
prevent transcription and translation. The lac-repressor protein of E. coli is designed
to read a base pair sequence of double stranded DNA. Thus, it can only block tran-
scription. In order to block transcription and translation simultaneously, the mam-
6*
malian repressor must be able to recognize a base sequence which comprises a patt
of the DNA cistron as weil as messenger RNA transcribed from that DNA. This is
possible only if the repressor is designed to read a base sequence of a single strand.
lnasmuch as the half-life of messenger RNA is very short (a matter of minutes) in
prokaryotes such as E. coli, the repressive transcriptional control alone is very
effective. In the case of vettebrates and other metazoans, on the other hand, the
half-life of messenger RNA may become very prolonged (a matter of days) in non-
dividing adult somatic cells. It stands to reason that the secondary repressive regula-
tory mechanism of vertebrates exercises control over transcription and translation.
In the body of a higher animal, the environment to which individual somatic cells
are exposed is under the control of a whole organism. lt is not surprising that hor-
mones rather than derivatives of substrates often serve as inducers. While hormones
can be specific inducers of cettain enzymes and proteins, they can also have a systemic
stimulating effect. The latter effect of hormones appears to be mediated through
cyclic AMP (adenosine 3',5' monophosphate). For example, insulin stimulates adenyl
cyclase on the cell membrane which results in greater production of cyclic AMP.
It is this cyclic AMP which serves as a somewhat indiscriminate inducer of a variety
of enzymes concerned with glyconeogenesis (RoBINSON et al., 1968).
At the bottarn of the hierarchy, the function of already made polypeptide chains
is regulated by the inherent property of polypeptide chains themselves. For example,
the fotward reaction catalyzed by an enzyme may be slowed down after a certain
amount of substrate has been converted to the product. An enzyme molecule often
has a binding affinity to an inhibitor, and then the level of an inhibitor controls the
enzyme activity; an example being ccfeedback inhibition". A nurober of cases are
known where amino acid (or in some cases nucleotide) end-products act as specific
inhibitors for the first of the enzymes in the chain producing them, even though they
may differ considerably in structure from the substrates of these enzymes.
For example, the activity of one of the two asparate kinases of E. coli is subjected
to ccfeedback inhibition" by threonine which is the end-product of a pathway (STRADT-
MAN et af., 1961).
2. The Requirements to be a Regulator of Gene Activation

Going back to the primary genetic regulatory mechanism of development which
operates by exercising positive transcriptional control, on the basis of the above
discussion, I surmise the characteristics of this mechanism to be as follows:
a) In the diploid organism, two structural cistrons on a homologaus pair of
chromosomes are, as a rule, regulated together. Accordingly, the product of a
regulatory gene should be diffusable, and, the product being diffusable, a regulatory
gene locus need not be closely linked to the structural gene it controls.
b) A regulatory gene product must have an inherent propetty to recognize and
bind with a stretch of particular DNA base sequence which is situated adjacent to a
structural cistron it controls. A stretch of DNA which is recognized by a regulatory
gene product shall be defined as the receptor site of a structural cistron (analogous to
the operator of /ac-operon system). The specific affinity between a regulatory gene
product and the receptor cannot be dependent upon free amino groups of lysine and
arginine residues, therefore, a regulatory gene product is most likely to be a large
Concordant Duplication of a Primary Regulatory Gene 85
protein of complex amino acid composition in the same manner as the /ac-repressor
protein of E. coli.
c) In addition, a regulatory protein must also be endowed with the ability to
compete against non-specifie histones for binding sites on DNA and selectively
remove already present histones from the receptor and an adjacent structural cistron.
Knowing the charaeteristics of histones (Chapter III), the most likely mechanism of
histone removal appears to be transphosphorylation between serine residues of
regulatory protein and of histones. If a regulatory protein has phosphorylated serine
residues, it can exehange free phosphates for phosphate groups of DNA with serine
residues of histone. Recent evidenee indicates that once a single serine residue is
phosphorylated, lysine-rich histone becomes ineapable of preventing the attaehment
of RNA polymerase to DNA (LANGAN, 1969).
d) For the struetural cistron to be regulated, it has to be aceompanied by a
stretch of DNA base sequenee whieh serves as the receptor. In ontogenie develop-
ment, it is possible that a group of functionally divergent structural genes is plaeed
under the control of a single regulatory gene, beeause they are endowed with nearly
identical receptors. The findings of RuTTER and his colleagues on embryonie develop-
ment of the rat panereas may indieate that struetural genes for trypsinogen, ehymo-
trypsinogen, amylase, ribonuclease, earboxylpeptidase A and B, and lipase A and B
are under the control of a single regulatory gene. Their findings can be extrapolated
to mean that all these struetural genes simultaneously begin to engage in tran-
scriptional activities at about the 10th day of embryonie life when the panereas is
merely a poueh-like diverticulum made of a single layer of epithelium derived from
the primitive gut (RuTTER et al., 1968). Of all the somatic eell types of the body, a
number of enzymes are found only in liver and kidney. This appears to suggest that
some regulatory genes make their presenee felt in more than one somatie eell type.
If the above surmises on the nature of positive transeriptional control meehanisms
are not too far off target, duplieation of regulatory genes and their receptors must
have eontributed greatly to the evolution of vertebrates.
3. Coneordant Duplication of a Primary Regulatory Gene and a

Structural Gene
The evidenee presented in the previous ehapter suggests that the strueturalloci
for a variety ofhemoglobin ehains evolved from a eommon aneestral cistron by a
series of gene duplieations. This series of duplications of struetural genes appears to
have been aceompanied by a eoneordant series of regulatory gene duplications. In
human development, the .x-ehain gene is aetivated at the onset of embryonie hemo-
poiesis, and it eontinues to engage in transeriptional aetivity throughout the life of an
individual. A partner of the .x-ehain, however, changes. At first, s-ehain appears so
that embryonie hemoglobins are of two types: .x2s2 and s 4 Soon, s-chain production
is shut off and y-chain takes over until neo-natal life. Fetal hemoglobin is mainly
.x2 y2 and some y 4 Finally, - and <5-ehains make their appearance. A major component
of adult hemoglobin is .x22 ; .x2<52 comprising a mere 3% or so of the total.
Differential eontrol of duplieated struetural genes implies the funetional diversi-
fieation of duplieated regulatory genes. It would appear that the .x-chain cistron, the
s-ehain cistron, and the y-ehain cistron have different receptors for separate regulatory
86 Why Gene Duplication ?
genes. The - and ~-chain cistrons, on the other band, must still be under the control
of the same regulatory gene, and, therefore, they must still have nearly identical
receptors.
4. Morphological Changes Due to Functional Diversification of a Duplicated

Regulatory Gene
A great morphological change in evolution has also been accomplished by the
mod.ification of a redundant copy of the previous structure. The first major anatomical
improvement that occurred to early vertebrates was the development of jaws.
Palatoquadrate Hyomandibular
Maxilla
Fig. 14. The transformation of the third gill arch into the jaws (palatoquadrate and maxilla)
and of the fourth gill arch into the hyomandibular are illustrated. The third gill arch is cross-
hatched while the fourth gill arch is shaded. Gill slits are solid black. Bottom: The hypo-
thetical ostracoderm of the Silurian period. Top: The hypothetical placoderm of the early
Devonian period
The earllest known vertebrates of more than 300 million years ago were jawless
fish betonging to the dass Agnatha; collectively known as ostracoderms. The hagfish
and lampreys of today represent the extensively modified form of this dass. These
fish-like creatures had a mouth which was merely an opening that led to the digestive
tract with as many as 10 pairs of gills opening into it. Such a mouth could merely serve
as a vacuum cleaner and was only fit to grovel in the bottarn mud of streams, lakes
and, at times, estuaries. The development of jaws enabled the improved model to
grasp and chew foods which were too large to fit the vacuum cleaner type mouth of
the predecessor.
References 87
It seems that the jaws were originally derived from a gill arch. The ostracoderm
had a !arge number of gills with cartilaginous or bony supports. At least the first one
and probably two of the original anterior gill "arches" were eliminated, while an-
other arch, probably the third in the series, was changed from a gill support to a pair
of jaws. Each gill arch in the primitive vertebrate was formed by a series of several
bones arranged somewhat in the fashion of a V turned on its side, with the point
directed posteriorly. When such a V was supplied with teeth and hinged at the point
of the V, it became the primitive verrehrate jaws as shown in Fig. 14 (COLBERT, 1955).
Since both the gill arch and jaws were made ofbones, the change from one form to
the other does not appear to have required the creation of new structural gene loci.
But, the modification appears to have required duplication of a regulatory gene or a
group of regulatory genes which controlled gill arch development. First, the third
gill arch had to be placed under separate control of a duplicated group of regulatory
genes. Subsequent mutations may have changed the characteristics of these regulatory
genes, so that the jaws were organized in place of a gill arch.
5. Duplication of Receptor Sites Adjacent to Structural Genes

If the control by a single regulatory gene is responsible for the apparently simul-
taneaus activation in the same somatic cell type of a group of divergent structural
genes, it follows that each of these structural genes must be endowed with nearly
identical receptors. For example, the receptor adjacent to the chymotrypsinogen locus
must have essentially the same base sequence as that adjacent to the amylase locus,
since both loci are activated in the pancreas. Only duplication of a common ancestral
receptor can equip functionally divergent structural genes with the same or similar
receptors. A similar idea has recently been proposed by BRITTEN and DAVIDSON (1969).
References
BRITrEN, R. J., DAVIDSON, E. H.: Generegulation for higher cells: A theory. Science 165,
349-357 (1969).
CoLBERT, E. H.: Evolution of the vertebrates. Science Editions. New York: John Wiley &
Sons, Inc. 1955.
GILBERT, W., MLLER-HILL, B.: Isolation of the lac repressor. Proc. Natl. Acad. Sei. US 56,
1891-1898 (1966).
J ACOB, F ., MoNOD, J.: Genetic regulatory mechanism in the synthesis of proteins. J. Mol.
Biol. 3, 318-356 (1961).
- - Genetic repression, allosteric inhibition, and cellular differentiation. In: Cytodiffer-
entiation and macromolecular synthesis (LocKE, M., Ed.), pp. 30-64. London:
Academic Press 1963.
LANGAN, T. A.: Phosphorylation of liver histone following the administration of glucagon
and insulin. Federation Proc. 28, 600 (1969).
RIGGS, A. D ., BouRGEOIS, S. : On the assay, isolation and characterization of the lac repressor.
J. Mol. Biol. 34, 361-364 (1968).
- - NEWBY, R. F., CoHN, M.: DNA binding of the Ia& repressor. J. Mol. Biol. 34, 365-
368 (1968).
RoBINSON, G. A., BUTCHER, R. W., SUTHERLAND, E. W.: Cyclic AMP. Ann. Rev. Biochem.
37, 149-174 (1968).
RurrER, W. J., KEMP, J. D., BRADSHAW, W. S., CLARK, W. R., RoNziO, R. A., SANDERS,
T. G.: Protein synthesis in cytodifferentiation. J. Cell Physiol. 72, 1-18 (1968).
SLATER, D. W., SPIEGELMAN, S.: An estimation of genetic messages in the unfertilized
echinoid egg. Proc. Natl. Acad. Sei. US 56, 164-170 (1966).
STRADTMAN, E. R., CoHEN, G. N., LEB RAs, G., DE RoBICHON-SzuLMAJSTER, H.: Feed-back
inhibition and repression of aspartokinase activity in Escherichia coli and Saccharomyces
cerevisiae. J. Biol. Chem. 236, 2033-2038 (1961).
TOMKINS, G. M., THOMPSON, E. B., HAYASHI, s., GELEHRTER, T., GRANNER, D., PETER-
KOFSKY, B.: Tyrosine transaminase induction in mammalian cells in tissue culture. Cold
Spring Rarbor Symposia Quant. Biol. 31, 349-360 (1966).
WHlTELEY, A. H., McCARTHY, B. ]., WHITELEY, H. R.: Changingpopulations of messenger
RNA during sea urehin development. Proc. Nad. Acad. Sei. US 55, 519-525 (1966).
Part 4
Mechanisms of Gene Duplication
ChapterXV
Tandem Duplication lnvolving Part of One Linkage Group

at a Time
Just as mutacion is the consequence of amistakein DNA replication, gene dupli-
carion also arises as an irrfrequent mistake in mitoric as well as meiotic processes.
Duplicarion which involves a single cistron or a duster of cistrons on one linkage
group is dealt with in this chapter. This mode of gene duplicarion can be contrasted
with the simultaneaus duplication of all linkage groups due to polyploidy (to be
dealt with in a later chapter).
1. Unequal Exchange between Two Chromatids of the Same Chromosome

When, for the first time, TAYLOR et al. (1957) studied the DNA replicarion
pattern of individual chromosomes by utilizing tririated-thymidine, they unexpec-
tedly noticed frequent exchange between two chromatids of the same chromosome.
The arrangement of DNA strands in the chromosome is such that each metaphase
chromosome of a daughter cell whose mother cell has incorporated tritiated-thymi-
dine into its DNA includes a labelled chromatid and an unlabelled one. The ob-
served pattern, however, was unexpectedly complicated. Both chromatids of a
chromosome consisted of a labelled and an unlabelled part, and when a part of one
chromatid was labelled, the corresponding part of the other chromarid was not
labelled. Clearly, during the synthetic phase of the mitotic cycle, frequent exchanges
between two chromatids of the same chromosome occur. Two chromatids of the
same chromosome are assumed to be absolutely identical. On the surface, this type
of exchange seems to be of no consequence, but since it does occur frequently, it
could on occasion be unequal as indicated by Fig. 15. Even the slightest unequal
exchange would lead to the placement of two identical cistrons on one chromarid and
the deletion ofthat cistron from the other chromatid. The exchange would result in
one of the daughter cells becoming heterozygous for a gene duplication and the other
daughter cell hemizygous forthat locus. Were this to occur during mitosis of germ
cells and were duplication to confer an immediate selective advantage on the off-
spring, a small isolated popularion would in time become homozygous for this
duplication and this duplication might become a species characteristic.
90 Mechanisms of Gene Duplication
In most instances, whether an observed tandem gene duplication was produced

by this mechanism or by unequal crossing-over during meiosis cannot be ascertained.
However, only an unequal exchange between two sister chromatids of the same
chromosome can result in a very unique situation which is characterized by two or
more duplicated gene loci sharing a similar or identical set of alleles (Fig. 15). As
(Phe/Phe) Germ cells of {Tyr/Tyr)

homozygote ancestral horse homozygote
(one locus for
c:t - chain)
Unequal exchange
Gametes
(Phe. Phe) (Tyr . Tyr)
Gene duplication
Modern heterozygous
horse
r Furtherdifferentiation
(Phe. Phe/Tyr . Tyr)
(Lys. Gln/Lys. Gin)
(Loci for c:tsand c:t f)
Fig. 15. Gene duplication by unequal exchange between two sister chromatids of the same
chromosome can result in a unique situation in which two closely linked duplicated gene
Iod share the same allelic alternatives. This is schematically illustrated using the gene loci
for horse hemoglobin a<-chains as examples. A small open circle denotes the centromere of
the chromosome, while each large circle represents an a<-chain cistron. A cistron which
specifies phenylalanine at the 24th position is drawn as a circle, having an outlined upper
half, while the cistron which specifies tyrosine at the 24th position is drawn as a cirde having
a solidly black upper half. When the bottom half of the circle is outlined, it indicates that the
cistron specifies Iysine at the 60th position. The shaded bottom half, on the other hand,
indicates that the cistron specifies glutamine at the 60th position
already mentioned in Chapter IV and shown in Fig. 6 (Chapter VI), every horse
produces two different kinds of hemoglobin !X-chains; !X! and !X1 The two differ by a
single amino acid substitution at position 60; glutamine in oJ is substituted by lysine
in lX1 (KILMARTIN and CLEGG, 1967). The possibility that these two !X-chains are speci-
fied by the same gene locus and that the presence of a mutant glutamine transfer RNA
is responsible for an ambiguous coding was raised in Chapter IV. An alternative, and
Unequal Exchange between two Chromatids of the Same Chromosome 91
perhaps more reasonable, explanation is that 0(,! and 0(,1 are spedfied by an original and
very recently arisen duplicate.
It should be recalled that there is an allelic substitution from tyrosine to phenyl-
alanine at position 24 which simultaneously affects both 0(,-chains. An immediate
ancestor of the modern horse ( Equus caballus) may have produced only one kind of
0(,-chain, since it bad only a single 0(,-chain locus. This is not an unreasonable assump-
tion, as the donkey, the horse's cousin, makes only one kind of 0(,-chain. At the single
locus stage, the horse may have already maintained two alleles; one specifying
tyrosine and the other, phenylalanine at position 24. If duplication of this locus was
accomplished by two separate events involving an unequal exchange within the same
chromosome, an original and a duplicated loci on the same chromosome must have
received the same codon for the 24th position. Thus, the two different kinds of
homologous chromosomes came into being within the population; an original and a
duplicate on one chromosome that specify tyrosine at the 24th position, and both on
the other chromosome that specify phenylalanine at the corresponding position. The
dose linkage between an original and a duplicate might have reduced recombination
to a negligible level. A duplicate subsequently diverged from an original by re-
placing glutamine with lysine or vice versa at the 60th position. In this manner, the
modern horse came to have one gene locus for 0(,! and another locus for 0(.1, with both
loci on the same chromosome invariably sharing the same allelic change (Fig. 6,
Chapter VI).
The wild Przewalski horse ( Equus pzrewalskii) of Guter Mongolia has 66 chromo-
somes (BENIRSCHKE et al., 1965), while the horse of all breeds has 64 chromosomes. It
appears that either duplication of the 0(,-chain locus or a transfer RNA mutation which
resulted in the production of two different kinds of 0(,-chains occurred before a com-
mon ancestor diverged into two sibling species of horses, since the Przewalski horse
also has 0(,! and 0(, chains.
1
In Chapters VI, VII and XIII, it was pointed out that various heavy-chains
(H-chains) utilized by different dasses of mammalian immunoglobulin are spedfied
by 10 or so dosely linked gene Iod. In man (NATVIG et al., 1967) and mice (HERZEN-
BERG et al., 1967), each of these gene Iod maintains its own series of alleles. These
allelic alternatives of man and the mouse are believed to represent amino acid sub-
stitutions that occurred within the constant region near the carboxyl-terminal of each
class of H-chains. In rabbits, on the other hand, alleles known as Aal, Aa2 and Aa3
appear to represent amino acid substitutions that affected constant sites within the
variable region near the amino-terminal of H-chains (OuniN, 1966). Genetically,
Aal, 2 and 3 behave as alleles of a single gene locus, yet these alleles are not found in
one particular class of H-chain, but rather in all different classes of H-chains. For
instance, in the AalfAa2-heterozygous rabbit, allelic markers 1 and 2 are carried not
only by H-chains of the IgG dass, but also by those of the IgA dass and even by the
(J.-heavy-chain of the IgM class (Tonn, 1963). This situation is apparently analogous
to that found in hemoglobin 0(,-chains of the horse, since 10 or so dosely linked
H-chain gene Iod on the same chromosome of the rabbit appear to share an identical
or very similar amino add substitution. On the surface, it may appear that unequal
exchanges between sister chromatids of the same chromosome have been solely
responsible for the tandem duplication of rabbit H-chain genes. However, the evi-
dence which suggests that a dustering of H-chain genes occurred before the emer-
gence of the first mammals on this earth makes the above suggestion untenable
(Chapter XIII).
The composite picture which emerges from the above with regard to cistrons
Controlling immunoglobulin H-chains isthat each complete H-chain is specified not
by a single cistron but by two independent cistrons which were fused into one during
ontogenic development by a lysogeny-like process. It may be that each of the 10 or
so closely linked gene Iod specifies only a constant region (a part including the car-
boxyl-end) of H-chains, and man and the mouse maintain allelic polymorphism of
constant genes. Variable regions (parts including the amino end) of H-chains are
specified by another group of independent gene loci, and rabbits maintain allelic
polymorphism at one of these variable gene loci. Since one variable gene may fuse
with any of the 10 or so different kinds of constant genes, it then follows that allelic
markers Aal, 2 and 3 of a variable gene locus are found in every dass ofH-chain. The
dass specificity of each H-chain resides in the amino acid sequence of its constant
region.
2. Unequal Crossing-over between Two Homologous Chromosomes during

Meiosis
In order to produce haploid gametes (eggs and spermatozon), germ cells of a
diploid organism pass through meiosis. During first meiotic prophase, two homolo-
gaus chromosomes pair side-by-side and form chiasmata between the two. Con-
sequently, an exchange of genetic material takes place between the paternally derived
and maternally derived elements. This is the process of genetic recombination by
crossing-over. An exchange of this sort might occasionally be uneven; the result
being duplication on one chromarid and deletion of a gene locus from the other
chromarid (Fig. 16).
In most instances, it is impossible to discern whether it was an unequal exchange
within the same chromosome or an unequal crossing-over between two homologues
during meiosis which was responsible for the creation of two known tandemly
duplicated genes. However, only an unequal crossing-over occurring during meiosis
of a heterozygote can result in placing two form er alleles of the same gene locus on the
same chromosome. Such an example is known at the gene locus for haptoglobin
lX-chain of man. Three common alleles of this locus exist in human populations; Hp1F,
Hp 15, and Hp2. SMrTHms and bis colleagues (1964) have shown that the lX-Chain
specified by Hp2 allele has about twice the molecular weight of that specified by
either Hp 1F or Hp 15 allele, and that the amino terminal half of a certain type of Hp2
chain has the amino acid sequence of Hp1F, while the carboxyl-terminal half has that
of Hp1s. Undoubtedly, an unequal crossing-over which occurred in an Hp1FjHp 1s-
heterozygote produced this type of Hp2 chain. The two alleles placed in tandem
on the same chromosome fused and became one cistron. Needless to say, tandem
fusion of two identical alleles, either Hp 1F or Hp 15, also occurred in homozygotes,
and produced other types of Hp2 chains.
3. Regional Redundant Replication of DNA

Normally, replication of chromosomal DNA occurs only in preparation for
mitosis and meiosis, and newly synthesized DNA strands are divided evenly between
Regional Redundant Replication of DNA 93
two sister chromatids. If only a segment of the DNA strand undergoes replication
during the G-1 stage, taodem duplication ofthat entire segment can be accomplished
by a single break and reunion at each end of a duplicated portion of the DNA strand.
Heterozygote
Meiosis
Unequal Crossing-over
Gametes
Gene Oupl ication
Zygote
homozygous
for a duplication
Fig. 16. Gene duplication by unequal crossing-over between two homologaus chromosomes
during meiosis is schematically illustrated. lf thls occurs in a heterozygote, two alleles of the
same gene locus become two independent gene loci in extremely close linkage. A small
circle denotes the centromere of the chromosomes, while a !arge circle indicates a cistron.
Inirially, a solid black circle and an outlined circle exist as two aUeles of the same locus (Top).
Unequal crossing-over during meiosis of a heterozygote places two alleles on the same
chromarid (second row). One of the 4 gametes produced by such a germ cell carries thls
duplication (third row). If thls duplication means a selective advantage, in time, every
member of the population would become homozygous for this duplication. The hetero-
zygous advantage would thus become permanently conferred on a population (Bortom)
On the basis of the following findings, KEYL (1966) holds that gene duplication
by this mechanism has, in fact, played a part in the evolution of dipteran insects.
He observed that two subspecies of the midge (ChironomuJ thummt) showed a marked
difference in DNA content even though they had identical diploid complements. The
diploid nucleus of Ch. th. thummi contained 27% more DNA than that of Ch. th. piger.
When the DNA contents of individual homologaus bands of the two subspecies were
compared on giant salivary gland chromosomes, it was noted that certain bands
showed as much as a 16-fold difference, while certain other bands showed no difference
at all. He felt that a 16-fold difference in the DNA content of the homologaus
chromosomal segment can more easily be explained by assuming repeated duplication
of a whole chromosome segment rather than of individual gene loci within that
segment. Inasmuch as the coordination of activities of functionally related genes does
not appear to depend on their close linkage (Chapter VII), it appears that duplication
of a whole chromosomal segment does not confer more selective advantage than
duplication of individual gene Iod. Furthermore, a chromosomal segment containing
tandemly duplicated gene loci is subjected to increasingly frequent unequal exchange
andfor crossing-over as already mentioned. Hence, a 16-fold increase in the DNA
content of a particular chromosomal segment can easily be the consequence of repeated
unequal exchange or crossing-over.
4. Merits and Shortcomings of Regional Duplication

In vertebrates, a considerable number of well proven cases of obvious duplicates
in close linkage are known. There is little doubt that gene duplication by any one of
the three means discussed in this chapter repeatedly occurred during the course of
vertebrate evolution. Some of these known tandem duplications apparently occurred
in ancient times, so that the possession of a fixed set of tandemly duplicated genes has
been established as a firmly fixed dass specific characteristic. For example, all
mammals appear to be endowed with the same set of closely linked duplicated genes
specifying various classes of immunoglobulin H-chains. This fact has already been
mentioned on several occasions. Quite clearly, the common ancestor to all mammals
already had a full set of these duplicated genes. Repeated tandem duplication of an
ancestral H -chain gene must have occurred in the reptilian or even amphibian ancestor
of mammals. Pantotheres, which existed at the end of the Jurassie period (about
130 million years ago), must have already had the full set.
Other tandem duplications have occurred in more recent times. For example,
human hemoglobin -chain and Cl-chain differ at only 10 of the 141 amino acid sites
(INGRAM and STREl'TON, 1962), and, among mammals, only man's immediate relatives,
such as orangutans, have the Cl-chain gene (HrLL et al., 1963). Duplication of a -chain
gene and the creation of a Cl-chain gene by subsequent diversification of a duplicate
must have occurred within the superfamily Hominoidea. This duplication is less than
25 million years old.
Some tandem duplications have occurred in even more recent times. The creation
of the Hp2 cistron for haptoglobin 1X-chain by a fusion between tandemly duplicated
Hp 1F and Hp 1s cistrons apparently occurred within the human species, for none of the
great apes possess the Hp2 cistron. This unequal crossing-over must have occurred
less than one million years ago. Even within mankind, Hp 2 exists merely as an allelic
alternative to Hp 1F and Hp 1s.
The creation of a new and Ionger cistron by a fusion of tandem duplicates can
only be accomplished by either an unequal exchange within the same chromosome
or by an unequal crossing-over between the homologues during meiosis. There is no
Merits and Shortcomings of Regional Duplication 95
doubt that such fusions greatly contributed to the evolution of vertebrates, since not
only Hp2 of haptoglobin x-chain, but also the first irnmunoglobulin H-chain gene,
were created by a fusion as discussed in Chapter XIII. Without the utilization of at
least two of the three mechanisms of tandem gene duplication mentioned in this
chapter, evolution from fish to mammals would have been impossible. These tandem
duplications, however, contain their own germ of disaster. It should become clear
that if vertebrate evolutionbad depended solely on tandem duplication, such advanced
creatures as mammals would never have come into existence.
The first shortcoming of tandem duplication is that the presence of tandem dupli-
cates invites further unequal exchange and unequal crossing-over. This was realized
as early as 1925 by Sturtevant on the Bar locus of Drosophila. As discussed in Chapter X,
each nucleolar organizing region of the chromosome carries several hundred tandemly
duplicated copies of the gene for 18S and 28S ribosomal RNA. Consequently, nucleolar
organizing regions undergo repeated unequal exchange and unequal crossing-over
(see Fig. 9, Chapter X). When cumulative deletions finally depleted the intolerable
number of copies from the nucleolar organizer, the deleterious effect was manifested.
The bobbed mutation of Drosophila melanogaster is such an example (RI1'0SSA et al.,
1966).
Even the mere tandem duplicates for hemoglobin - and :5-chains of man appear to
engage in unequal exchange and unequal crossing-over with significant frequency, for
the hybrid Lepore chain, which has the amino terminal half of :5-chain and the carboxyl
terminal half of -chain, has been encountered as several independent mutation-like
events (BAGLIONI, 1965). The formation of a hybrid cistron is but one of many conse-
quences of unequal exchange and unequal crossing-over. There may be a number of
healthy persons walking the streets who carry two doses of either the -chain gene or
the o-chain gene on the same chromosome.
Under such unstable conditions, the functional diversification of a redundant
duplicate cannot advance too far. Indeed, the degree of functional diversification
exhibited by tandemly arranged duplicates is rather small. The duster of immuno-
globulin H-chain genes, despite their existence of over 130 million years, continues to
perform the same function in only a slightly different way.
The second shortcoming of tandem duplication is the disruption of the established
gene dosage ratio between functionally related genes. It is believed that an allele for
an abnormal '-chain of human hemoglobin persists in Mrican populations with
remarkably high frequency, because relative resistance to falciparum malaria is
passed on to /'-heterozygotes (Chapter VI). By a single unequal crossing-over
during meiosis of a heterozygote, the -allele and '-allele can be placed on the same
chromosome to become two separate gene loci in extremely close linkage. If thls
spreads, every member of a population would then have acquired a permanent
heterozygous advantage without ever having to produce a deleterious homozygous
type ('/ 1). The fact that thls type of duplication has never been observed among
Mrican people suggests that doubling of the -chain gene dosage without also
simultaneously increasing the x-chain gene dosage does more than offset the advantage
gained by placing - and 1-alleles on the same chromosome. The almost exclusive
formation of x 2 2-tetrameric hemoglobin molecules requires that x-chains and -
chains be produced in nearly equal amounts. The two gene loci specifying hemoglobin
x- and -chains are not linked.
The tandem duplication which produced the <5-chain gene from a redundant
duplicate of the -chain gene is tolerated probably because the o-chain gene is a very
inefficient producer; Hb ~ (<X 2o2) amounts to only 1.5 to 4.0% of the total adult
hemoglobin. Were the <5-chain gene as efficient a producer as the -chain gene, this
duplication would probably not have been tolerated. Since the hybrid Lepore chain
which has the amino terminal sequence of o-chain is produced in as small an amount as
the <5-chain itself, the receptor base sequence at the head of the o-chain gene appears to
be responsible for the poor productivity of this structural gene locus.
An unequal crossing-over which produced the Hp2-allele of haptoglobin <X-chain
did not disrupt the gene dosage relationship with the gene for haptoglobin -chain,
because two duplicates fused into a single cistron. Mammalian genomes contain
nearly 10 closely linked gene loci for immunoglobulin H-chains (Chapter VII) and
two unlinked gene loci for n- and .A.-classes of L-chains on the second and third
linkage groups (GrLMAN-SACHS et al., 1969). Natural selection has permitted such
gross discordance with regard to the degree of duplication of L- and H-chain genes
only because the genetic regulatory mechanism had previously been evolved to
ensure that only a single L-chain locus and a single H-chain locus would engage in
transcription in each clone of antibody-producing plasma cells.
The third and probably most serious shortcoming of tandem duplication is that
it tends to duplicate the structural cistron without duplicating the regulatory gene
which controls its activity. The regulatory gene product is diffusable, and, for this
reason, the regulatory gene locus need not be and is not closely linked with the struc-
tural gene locus it controls (Chapter XIV).
As long as tandem duplicates remain under the control of a single regulatory gene
locus, there is not much hope of duplicates acquiring divergent functions. The
differential use of functionally diverged duplicates during ontogeny is an eventual
possibility only if each duplicated structural gene is accompanied by its own dupli-
cated regulatory gene. Only then can the receptor site (operator) of the structural
gene and its regulatory gene evolve side-by-side as an independent unit.
As shall be described, in the case of lungfish and salamanders, the enormaus
increase of their genome size appears to have been accomplished exclusively by
regional duplication. The genome size of the South American lungfish is 35 times
greater than the size of the mammalian genome (HNO and ATKIN, 1966) and that
of the amphiuma (salamander) is 28 times greater (MIRSKY and Rrs, 1951). Yet,
there is no sign that they possess a proportionately greater nurober of functionally
diverged duplicated genes (CoMINGS and BERGER, 1969).
Lungfish and salamanders clearly show the tragic consequences of exclusive
dependence upon tandem duplication as the means of achieving gene duplication. No
matter how many copies of the structural cistron they might have acquired by
repeated tandem duplication, as long as tandem duplicates remain under the control
of a single regulatory gene locus, they can only serve to produce more of the same
gene product as do the multiple copies of the gene for 18S and 28S ribosomal RNA
which occupy the nucleolar organizing region (Chapter X). As their genome size
increased to a great extent, their cell size also increased enormously. There has been
an obvious need to produce more of each gene product. While tandem duplicates of
lungfish and salamanders apparently fulfilled this need, for this very reason, none of
their duplicates have been set aside to escape from the relentless pressure of natural
References 97
selection and to emerge as new gene loci with hitherto nonexistent functions. Un-
limited increase of the genome size exclusively by tandem duplication is a hopeless
proposition in that it is the best way to conserve and the worst way to change.
It is of interest to note that CALLAN (1967), who developed the Master-slave hypo-
thesis introduced in Chapter X, worked mainly with the genomes of salamanders.
Such a system might as weil exist not only for the ribosomal cistron, but also for
various structural cistrons of lungfish and Salamanders. The prototype of the modern
lungfish ( Dipterus) was already in existence in the middle Devonian period (300
million years ago) and the ancestor of salamanders ( Lepospondyls) branched out from
the main amphibian line as early as the beginning of the Carboniferous period
(260 million years ago).
References
BAGLIONI, C.: Abnormal human hemoglobins. X. A study of hemoglobin Lepore Boston.
Biochim. et Biophys. Acta 97, 37-46 (1965).
BENIRSCHKE, K., MALOUF, N., Low, R. J.: Chromosome complement: Difference between
Equus caba!!us and Equus przewa!skii, PoLIAKOFF. Science 148, 382-383 (1965).
CALLAN, H. G.: The organization of genetic units in chromosomes. J. Cell Sei. 2, 1-7 (1967).
CoMrNGS, D. E., BERGER, R. 0.: Gene products of amphiuma: An amphibian with an
excessive amount of DNA. Biochem. Genet. 2, 319-333 (1969).
GrLMAN-SACHs, A., MAGE, R. G., YouNG, G. 0., ALEXANDER, C., DRAY, S.: Identifica-
tion and genetic control of two rabbit immunoglobulin allotypes at a second light chain
locus, the c locus. J. Immunology, 103, 1159-1167 (1969).
HERZENBERG, L. A., MrNNA, J. D., HERZENBERG, L. A.: The chromosome region for im-
munoglobulin heavy-chains in the mouse. Cold Spring Barbor Symposia Quant. Biol.
32, 181-186 (1967).
BILL, R. L., BuETINER-]ANuscH, J., BuEITNER-]ANUSCH, V.: Evolution of hemoglobin in
primates. Proc. Nad. Acad. Sei. US 50, 885-893 (1963).
INGRAM, V. M., STRETI'ON, A. 0. W.: Human hemoglobin A 2 I. Comparison of hemo-
globins A 2 and A. Biochim. et Biophys. Acta 62, 456-474 (1962).
KEYL, H. G.: Increase ofDNA in chromosomes. In: Chromosomes today, Vol. 1, pp. 99-101.
(DARLINGTON, C. D., LEwrs, K. R., Eds.). London: Oliver & Boyd 1966.
KrLMARTrN, J. V., CLEGG, J. B.: Amino-acid replacements in horse hemoglobin. Nature 213,
269-271 (1967).
MrRSKY, A. E., Rrs, H.: The desoxyribonucleic acid content of animal cells and its evo-
lutionary significance. J. Gen. Physiol. 34, 451-462 (1951).
NATVrG, J. B., KuNKEL, H. G., LrrwiN, S. P.: Genedc markers of the heavy-chain sub-
groups of human gamma G globulin. Cold Spring Barbor Symposia Quant. Biol. 32,
173-180 (1967).
HNO, S., ATKIN, N. B.: Comparadve DNA values and chromosome complements of eight
species of fishes. Chromosoma 18, 455-466 (1966).
UDIN, J.: Genetic reguladon of immunoglobulin synthesis. J. Cell Physiol. 67, 77-108
(1966).
RrTossA, F. M., ATWOOD, K. M., SPIEGELMAN, S.: A molecular explanadon of the bobbed
mutants of Drosophila as partial deficiencies of "ribosomal" DNA. Genedes 54, 819-834
(1966).
SMITHIES, 0., CoRNELL, G. E., DrxoN, G. H.: Chromosomal rearrangements and protein
structure. Cold Spring Barbor Symposia Quant. Biol. 29, 309-319 (1964).
STURTEVANT, A. H.: The effects of unequal crossing-over at the Bar locus in Drosophila.
Genedes 10, 117-147 (1925).
TAYLOR, J. H., WooDs, P. S., HuGHES, W. L.: The organizadon and duplicadon of chromo-
somes as revealed by autoradiographic studies using tritium-labeled thymidine. Proc.
Nad. Acad. Sei. US 43, 122-128 (1957).
TODD, C. W.: Allotypy in rabbit 195 protein. Biochem. Biophys. Res. Commun. 11, 170 to
175 (1963).
Chapter XVI
Polyploidy: Duplication of the Entire Genome

By becoming tetraploid, what used to be the diploid chromosome complement
becomes the haploid set (genome). Thus, duplication of all gene loci, the structural
cistrons as well as the regulatory cistrons which control the structural cistrons, is
accomplished "in one fell swoop". The mechanism of gene duplication by poly-
ploidy has none of the shortcomings exhibited in tandem duplication. Concordant
duplication of all gene loci creates no problern with regard to the dosage relationship
of functionally related genes, and duplication of each structural gene is accompanied
by duplication of its own regulator. As the original gene and its duplicate are carried
by two separate chromosomes, no instability is induced by gene duplication. As will
be shown, however, polyploidy has its own shortcomings. Quite clearly, tandem
duplication and polyploidy complement each other as the means of achieving meaning-
ful gene duplication. It would not be surprising if during the course of vertebrate
evolution these two means were used alternatively.
1. Incompatibility between Polyploidy and the Well Established

Chromosomal Sex Determining Mechanism
In the plant kingdom, particularly among the flowering plants, examples of
polyploid evolution abound. For instance, of the cultivated cereal plant Sorghum,
three species, S. versicolor, S. sudanense and S. halepense, have chromosome numbers of
10, 20 and 40. The first is the diploid species (2n = 2 x 5), while the second and
third are respectively tetraploid (4n = 4 x 5) and octaploid (8n = 8 x 5) species. In
the maj ority of flowering plants, male argans (stamens) and female argans (carpels or
pistils) are present in the same flower. They are in fact hermaphroditic species.
Polyploid evolution and the hermaphroditic mode of reproduction are perfectly
compatible.
The scarcity of polyploids among not only vertebrates, but also invertebrates, is
due to the bisexuality that characterizes the majority of animal species. Polyploidy is
incompatible with the weil established chromosomal sex-determining mechanism.
When diploid organisms with the XYJXX scheme of sex-determining mechanism
become tetraploid, the male has to maintain the 4AXXYY-constitution and the
female the 4AXXXX. During meiosis of the 4AXXYY-male, the four sex elements
may pair off as the XX-bivalent and the YY-bivalent. If this occurs, every gamete
would be of the 2AXY-constitution. Consequently, all the offspring resulting from
a mating between a tetraploid male and a tetraploid female would receive the
4AXXXY-constitution. If the 4AXXXY gives the male phenotype, the species would
obliterate itself for lack of females. If the 4AXXXY gives the sterile intersex, the
situation is not much improved. Even if two XY-bivalents are formed in individual
spermatocytes of the tetraploid male, in 50% of the cases the X and the Y would move
to the samedivisionpole at 1st meiotic anaphase, again resulting in the production of
progeny with the 4AXXXY-constitution. There exists no mechanism which insures
the exclusive production of two classes of gametes, 2AXX and 2AYY, by the
4AXXYY-male. Thus, polyploidy invariably disturbs the chromosomal sex determin-
ing mechanism.
Autopolyploidy 99
Because of the well established chromosomal sex determining mechanism, the

possibility of polyploid evolution is denied to mammals, birds and reptiles. Although
triploid individuals can be viable in birds (OHNO et al., 1963) and reptiles, it appears
that the best they can do is become an all female gynogenic or parthenogenic species.
Triploid species of the Ted lizard (genus Cnemidophorus) contain the haploid chromo-
some complement of one species and the diploid set of the other species. Triploidy
is apparently the means employed by these interspecific hybrids to escape from
hybrid sterility (PENNOCK, 1965). Such an interesting oddity, however, is a side issue
of vertebrate evolution.
In amphibians and fish, on the other hand, the chromosomal determiners of the
opposite sexes, the X and the Y of the male heterogamety as well as the Z and the W
of the female heterogamety, are still in a rather initial state of differentiation, so that
the X and the Y or the Z and the W share many gene Iod and the Y can Substitute for
the X. In mammals, the presence of one X is essential to the viability of a zygote
regardless of its sex. The 2AOY or the 2AYY-constitution is lethal. In the Japanese
cyrinodont fish (Oryzias latipes), however, the 2AYY-constitution can not only give
rise to the normal male, but also by estradiol treatment, the 2AYY can be changed
into the fertile female (YAMAMOTO, 1961). It is not surprising ifpolyploid evolution
is still possible among living amphibians and fish of today. Indeed, the first bisexual
tetraploid species was found among frogs of South America (BECAK et al., 1966).
Inasmuch as polyploidy and tandem duplication complement each other as the
means of achieving meaningful gene duplication, and since polyploid evolution was
possible only at the initial stages of vertebrate evolution, it then follows that most of
nature's experiments with gene duplication must have been clone at the stages of
fish and amphibians. It appears that by alternatively employing tandem duplication
and polyploidy, the fish or amphibian ancestor of mammals had already attained the
characteristic genome size endowed with an adequate degree of genetic redundancy.
Subsequent big leaps in evolution, which culminated in the creation of mammals and
eventually man, appear to have been accomplished not by further extensive gene
duplication, but rather by differential use of the already existing genetic redundancy.
2. Autopolyploidy
When two daughter cells which were produced at the end of mitotic telophase
fuse into one, a tetraploid cell is produced in the body of a diploid organism. A
tetraploid cell may also arise from two successive DNA replications not interverred
by mitosis. It must be that during mitotic propagation of spermatogonia, as well as
ogonia, in the gonad, a nurober of tetraploid germ cells are normally produced.
At the completion of meiosis, such tetraploid germ cells should give rise to diploid
gametes. The tetraploid zygote, therefore, a tetraploid individual, is produced by the
union of two diploid gametes. It appears that in man and other mammals tetraploid
zygotes are normally produced with significant frequency, but tetraploidy is a lethal
condition. Of the 227 spontaneously aborted human fetuses studied by CARR (1967),
two were tetraploids; a 4AXXXX and a 4AXXYY. In the species where triploid
individuals are viable and fertile, tetraploid zygotes may also be produced as the result
of a mating between a triploid and a diploid. As already mentioned, the triploid
female tends to produce triploid eggs which is the basis for the creation of the all
7*
female parthenogenic triploid species. FANKRAUSER and HUMPHREY (1959) artifici-

ally produced triploid individuals of the salamander, Ambystoma mexicanum. Triplaids
of both sexes were of very poor fertility, and males were more sterile than females.
Therefore, the mating between triploids was not readily accomplished. When mated
to diploid males, however, triploid females produced many tetraploids. If tetraploid
individuals of both sexes are fertile, a bisexual tetraploid race and eventually a new
tetraploid species may be created. Such tetraploids, which have descended from a
single ancestral diploid species, are defined as autotetraploids, for their genome is
equivalent to the diploid chromosome complement of the ancestral species.
A newly arisen autotetraploid species can be readily identified by the presence of
quadrivalents in meiosis. In the diploid species there are two homologues of each kind
and during 1st meiosis only bivalents are seen. In a fresh autotetraploid, however,
there are four homologues of each kind, and these four associate with each other and
form one quadrivalent rather than two bivalents.
SAEZ and BRUM (1959) indicated that South American frog species belanging to the
family Ceratophrydidae might constitute an autopolyploid series. The lowest diploid
chromosome nurober of this series was 22 and the highest was 110. Indeed, MARIA
LurzA BECAK and her colleagues (1966) found that while 22 chromosomes of Odonto-
phrynus cultripes formed 11 bivalents in meiosis, 44 chromosomes of 0. americanus
tended to form 11 quadrivalents. Indeed, Odontophrynus americanus was the first
example of a newly arisen bisexual autotetraploid species ever found in vertebrates.
Fig. 17 (Plate III) shows the karyotype and a 1st meiotic metaphase figure of a male of
this species. It should be noted that four homologues each, of 11 different kinds, con-
stitute the somatic chromosome complement of 0. americanus. No difference has
been detected between the karyotype of the male and the female. This suggests that
the X and the Y or the Z and the W of this tetraploid species and its ancestral diploid
species are still morphologically identical. The rather initial stage of sex chromosome
differentiation which prevails in this family of frogs apparently permitted polyploid
evolution. A 1st meiotic metaphase figure [Fig. 17 (Plate III)] contains 10 quadrivalents
and two bivalents. In most instances, however, 11 quadrivalents are seen.
Subsequently, BECAK et al. (1967) have shown that Ceratophrys dorsata of this
family, having 104 chromosomes, is an auto-octaploid species evolved from such a
diploid species as Chacophrys pierotti with 26 chromosomes. The 104 chromosomes
of C. dorsata can be sorted out to eight homologues each of the 13 different kinds and
during male meiosis the presence of several octavalents, some quadrivalents, as weil as
a small number of bivalents were noted.
3. Allopolyploidy
The mule is an interspecific cross produced by mating the male donkey (Equus
asinus, 2n = 62) with the female horse (Equus caballus, 2n = 64). The mule shows a
conspicuous hybrid vigor, and because of this they have served mankind as a patient
beast of burden since time immemorial. The mule is mentioned in the Iliad by Homer
(850 to 800 B. C.) as well as in ancient Chinese literature. The mule, however, cannot
perpetuate itself, since mules of both sexes are unequivocally sterile. Their sterility
is due neither to the lack of steroid sex hormone production nor to the absence of
sexual desire. The sterility is due solely to the pairing difficulty during meiosis. The
Diploidization of the Tetraplaid 101
31 chromosomes in the donkey haploid set (genome) and the 32 chromosomes in the
horse haploid set clearly differ by a number of translocations and inversions (TRUJILLO
et al., 1962; BENIRSCHKE et al., 1962), so that homologous pairing between the horse
and donkey chromosomes is precluded during meiosis. The mule has no way of
producing the functional haploid gametes. If the mule is made a tetraploid, what then?
During meiosis, the donkey chromosome can pair with its homologous donkey
chromosome, and the horse chromosome can pair with its own kind; thus, the pro-
duction of balanced mule diploid gametes is insured. The only reason this trick
cannot be performed on the mule is that tetraploidy, as already mentioned, is a lethal
condition in mammals. In flowering plants, however, fertility was restored to inter-
specific hybrids time and again by polyploid evolution. Such polyploids which have
originated from interspecific hybrids are defined as allopolyploids. In almost total
absence of one-to-one pairing affinity between chromosomes of the two parental
haploid sets, a newly arisen allotetraploid would show no sign of quadrivalent for-
mation. Instead, the two allen diploid sets would form two independent sets of
bivalents. This, then, is a feature which distinguishes a new allotetraploid from a new
autotetraploid.
Among those belonging to the genus Barbus of cyprinid fish, Barbus barbus having
100 chromosomes appears tobe a tetraploid species, because others of the genus have
around 50 chromosomes, and the genome size (expressed as the amount of DNA
contained in the somatic nucleus) of Barbus barbus is about twice that of other species
(WoLF et al., 1969). Yet, only bivalents are seen during meiosis of this apparently
tetraploid species. Barbus barbus conceivably represents an allotetraploid species.
4. Diploidization of the Tetraploid

A newly arisen autotetraploid begins its evolution having four homologous
chromosomes of each linkage group. If this acquired state of tetraploidy is going to
contribute to the creation of new gene loci, the disomic state should eventually be
reestablished by functional diversification of four original homologues, so that one
original linkage group is split into two separate linkage groups. This, then, is the
process of diploidization which occurs in a tetraploid species.
As long as four homologues get together to form a quadrivalent during meiosis,
the four would be randomly sorted out into two each at the end of 1st meiosis.
There is no possibility of functional diversification. The preferential formation of two
separate bivalents instead of one quadrivalent is the prerequisite for diploidization.
Structural heterozygosity must be created among the four original homologues
(SHAVER, 1963). As an example, let us assume that a newly arisen autotetraploid
contained a set of four homologous metacentric chromosomes in the mitotic comple-
ment. For the time being, they invariably formed a single quadrivalent. Subsequently,
however, a pericentric inversion occurred in one of the four changing it to a sub-
terminal chromosome. Natural selection now favored the inclusion of one new
inverted subterminal and one original metacentric in the genome. This created a
situation where an inverted subterminal preferentially pairs with another of the same
kind leaving two original metacentdes to pair with each other. So far as these four
original homologues are concerned, the major step toward diploidization had indeed
been taken.
Ofthose fish belanging to the suborder Salmonoidea, trout, salmon, whitefish and
graylings appear to be autotetraploid species which have progressed toward diploidi-
zation in vadous degrees. Many of them probably odginated from a diploid ancestor
which had 48 acrocentric chromosomes. Thus, newly arisen autotetraploids may
have been endowed with 96 acrocentdc chromosomes. The creation of one new
metacentdc chromosome at the expense of two acrocentrics (Robertsonian fusion)
appears to have been the means chiefly employed to achieve structural hetero-
zygosity within four homologaus chromosomes of each set (OHNo et al., 1968).
Although further transformation of certain metacentdes by pericentric inversions to
acrocentrics occasionally confuses the picture, the closely related trout and salmon
species usually have the samenurober of chromosome arms (counting one acrocentric
as one and one metacentdc as two) despite apparent differences in the chromosome
number. The chromosome nurober is lower in those species with more metacentrics.
In certain species, such as the Pacific Coho salmon (Oncorf?ynchus kisutch), 1st meiotic
metaphase figures contain mostly bivalents and only a few quadrivalents. It appears
that they have almost completed the process of diploidization. Others, such as the
rainbow trout ( Salmo irideus), show extensive Robertsonian polymorphism within
the species. Although the nurober of chromosome arms remains constant at 104,
the chromosome nurober may vary from a low of 58 (50 metacentdes plus 8 acro-
centdcs) to a high of 65 (39 metacentdes and 26 acrocentrics) between individuals as
well as within an individual. As shown in Fig. 18 (Plate IV), 1st meiotic metaphase
figures of the rainbow trout contain a various nurober of quaddvalents and a few
multivalents in addition to bivalents (OHNO et al., 1965).
In the South American frog, Odontophrynus americanus, we have seen a newly
arisen autotetraploid, while in the trout and salmon we evidenced rather ancient
autotetraploids progressing toward diploidization and almest achieving the goal.
Even in the case of allotetraploids, if they arose from interspecific hybrids between
very closely related parental species, the initial formation of quaddvalents would be
unavoidable. Thus, they too must go through the process of diploidization described
above in order to regain the disornic state.
All mammalian and avian species which have been scrutinized to this date appear
as authentic diploid species; there is no hint that they evolved from immediate
tetraploid ancestors. Yet, it is our contention that either at the fish stage or at the
amphibian stage, the mammalian ancestor went through at least one tetraploid
evolution. Although avian species apparently beleng to a different genome lineage,
their ancestor must have also gone through tetraploid evolution. During the course of
more than 200 rnillion years, diploidization has been thoroughly accomplished by the
genomes of mammals and birds. Yet, the fact that in mammals the gene locus for
hemoglobin cx-chain is not linked tothat for Hb -chain, and the gene loci for immuno-
globulin L-chains arenot linked to those for immunoglobulin H-chains, appears to
indicate that these gene duplications were accomplished by tetraploid evolution in
mammal' s remote past.
As mentioned in the previous chapter, those species which have depended
exclusively on tandem duplication as the means of achieving gene duplication show no
sign that they have acquired a nurober of functionally diverged duplicated genes in
proportion to their genome size.
Diploidization of the Tetraplaid 103
What is the situation found in freshly diploidized autotetraploid species? At first

a newly arisen autotetraploid species should be endowed with four doses of every
gene; therefore, tetrasomic inheritance is expected. Indeed, in a tetraploid Odonto-
phrynus atJJericanus species, BECAK and his colleagues (1968) found that two alleles
which specify electrophoretically fast (F) and slow (S) variants of serum albumin are
inherited in a tetrasomic mannet. In addition to two homozygous types, F/F/F/F
and SfSfSfS, there were three types of heterozygotes, F/F/F/S, FfFfSfS and FfSfSfS.
The Hardy-Weinberg equilibrium in a population with regard to these two alleles is
expressed by the formula (p + q)4 = p4 + 4p3q + 6p2q2 + 4pq3 + q4 = 1, instead of
the customary formula for disomic inheritance which is (p + q) 2= p 2+ 2pq + qz = 1.
Next, four homologues of each original linkage group are transformed to two
independent pairs by the process of diploidization; therefore, at this stage, each original
gene locus has effectively been duplicated. A switch from tetrasomic inheritance of a
single gene locus to disomic inheritance of a pair of gene loci is expected. Thus, with
regard to almost any gene product, we expect a freshly diploidized autotetraploid
species to show possession of twice the number of gene loci as compared to its
diploid relatives. The closest diploid relatives of salmon, trout, whitefish and gray-
lings are smelts of the family OstJJeridae.
While various species of smelts have two and possibly three separate gene loci,
for subunits of an enzyme, lactate dehydrogenase (LDH) (HNO et al., 1968), trout,
salmon and other tetraploid members are endowed with at least five (MoRRISON and
WRIGHT, 1966; KLSE et al., 1968), and possibly eight separate gene loci for LDH
subunits (MASSARO and MARKERT, 1968).
Non-mitochondrial NAD-dependent malate dehydrogenase (MDH) of smelts is
specified by a single gene locus in their genome (QmRoz-GuTIERREZ and HNO,
unpublished data). In sharp contrast, trout and salmon are endowed with two
separate gene loci for MDH (BAILEY et al., 1969).
The situation found on NADP-dependent isocitrate dehydrogenase (lCDH) and
NAD-dependent sorbitol dehydrogenase (SDH) of tetraploid salmonoid fish is of
special interest. Diploid salmonoid fish such as herring and smelt are endowed with
one gene locus each for supernatant- and mitochondrial-form of ICDH. The rainbow
trout, on the other band, has two separate gene loci for mitochondrial-form lCDH,
but still a single gene locus for supernatant-form ICDH. This single locus, however,
is inherited in a tetrasomic mannet (WoLF et al., 1970). In the case of SDH, the brown
trout shows the disomic inheritance of two separate gene loci, while the rainbow
trout shows the tetrasomic inheritance of a single locus (ENGEL et al., 1970). The
tetrasomic inheritance is expected at certain gene loci of these freshly diploidized
tetraploid fish, since they still form a number of quadrivalents during meiosis.
Nevertheless, tetraploid members of salmonoid fish possess a greater number of
isozyme gene loci than either mammals or birds. Mammals and birds merely have
three gene loci for LDH and one gene locus each for non-mitochondrial MDH and
ICDH. After the initial experiment by freshly diploidized autotetraploids, many of the
duplicated gene loci are expected to be silenced during subsequent stages of evolution
for the lack of selective advantage. Therefore, the fact that mammals and birds have
a fewer number of duplicated genes compared to traut and salmon does not speak
against the idea that these two classes of warm-bloaded vertebrates evolved from
ancient tetraploid ancestors.
Mammals usually have three and at the most five separate gene loci for hemoglobin
chains; trout and salmon, on the other hand, appear to have nearly 10 gene loci for
hemoglobin chains (TsuruKr and GADD, 1963). It has also been reported that more
numerous isozymic forms of glyceraldehyde-3-phosphate dehydrogenase are found
in the rainbow trout than in mammals (LEBHERZ and RuTTER, 1967).
Mammals, birds and lungfish, as a rule, have only one gene locus each for trypsin
and chymotrypsin of the pancreas. Ungulates are exceptional in that they are endowed
with a pair of gene loci for chymotrypsin A and B (NEURATH et al., 1967). Salmon and
trout apparently have two gene loci for trypsin and two for chymotrypsin (CROSTON,
1965). Indeed, freshly diploidized autotetraploid species are undergoing an extensive
experiment to functionally diversify every pair of duplicated genes.
5. Elimination of Certain Chromosomes during Diploidization

The mechanism of mitosis normally insures that two sister chromatids of each
chromosome separate from each other at anaphase and move to opposite poles of the
mitotic spindle. This mechanism, however, falls every now and then. Instead of
separating from each other, both chromatids may move to the same division pole.
As a result, one daughter cell receives three homologous chromosomes (trisomy)
and the other only one (monosomy). If such occurs in germ cells, trisomic and mono-
somic individuals are produced. It appears that elimination of unneeded chromosomes
through mating between individuals each missing one chromosome may accompany
the process of diploidization of a tetraploid species. W e have examined the chromo-
some constitutions of nearly 500 tainbow trout. Among them were a few individuals
which possessed one more, as well as one less, than the normal nurober of chromo-
some arms which is 104. These aneuploids showed no apparent ill effect and they were
endowed with functional gonads. During the process of diploidization, certain of the
originallinkage group may fall to achieve functional diversification. As far as those
which failed are concerned, natural selection may favor the elimination of two of
the four original homologues from the genome.
6. Dosage Effect of Regulatory Genes in the Tetraplaid

Because the well established chromosomal sex determining mechanism makes the
bisexual propagation of polyploid individuals impossible, the possibility of polyploid
evolution is denied to reptiles, birds and mammals. If sexuality is the only reason,
however, we should expect to recover a nurober of viable polyploid individuals
among higher vertebrates even though they may be intersex. The fact is that tetra-
ploid zygotes of man (CARR, 1967) and other mammals are aborted early in fetallife.
The explanation for lethality of tetraploid zygotes cannot be found in disturbance of
the chromosomal sex determining mechanism.
The subject of the dosage effect of regulatory genes shall be introduced here to
clarify the issue. As discussed in some detail in Chapter XIV, many of the structural
genes in the vertebrate genome appear to be under double regulation, in that the
secondary repressive control is often superimposed on the gene loci which were
already committed by the primary activating control to be active in certain somatic
cell types. In the secondary repressive control, the repressor gene product binds with
a DNA cistron as well as its messenger RNA and blocks transcription and translation
References 105
of the already committed locus until it is removed by an inducer. Thus, the gene
product is made in appreciable quantity only in the presence of an inducer. The modus
operandi of the repressive secondary control of vertebrates is very similar to that of
the repressive control of the /ac-operon system of E. co!i.
In the /ac-operon system of E. coli, it has been shown that the rate of synthesis of
-galactosidase (a regulated gene product) varies reciprocally with the first power
level of the repressor (SADLER and NoVICK, 1965). To put it simply, in the normal
haploid state, the E. coli genome contains one dose of the repressor gene and one
dose of the /ac-operon. At this 1 : 1 dosage ratio the system works fine, but when the
ratio is changed to 2:2, as in diploid, superrepression results. In the absence of an
inducer, a diploid actually synthesizes less -galactosidase than a haploid and it
takes a higher level of inducer to achieve maximum synthesis by each of the two regu-
lated structural genes. It appears that in the normally diploid organism, such as
mammals, the 1:2 dosage ratio between the regulator and the regulated is preferable
to the expected 2:2 ratio. Indeed, the fact that in the development of extreme inter-
specific hybrids only the maternally derived allele of each structural gene locus tends
to be expressed (HrTZEROTH et al., 1968; CASTRo-SmRRA and HNO, 1968) can be
interpreted to mean that in diploid organisms only the maternally derived regulatory
genes function, while the paternally derived ones remain silent. Whatever may be
the actual dosage ratio between the regulator and the regulated in diploids, it is
likely that all the regulated structural genes are subjected to Superrepression in a
newly arisen tetraploid.
There is actual evidence of Superrepression in malignant plasma cells of the mouse.
When a myeloma producing a respectable immunoglobulin type was cloned accord-
ing to the ploidy, it was found that every diploid clone continued to produce immuno-
globulin, but only half of the tetraploid clones did so, and none of the octaploid
clones producedimmunoglobulin (CoHN, 1967). The synthesis of gratuitous products
was shut off merely by an increase in ploidy.
Perhaps the lethality of tetraploid zygotes in higher vertebrates is due to Super-
repression. Fish and amphibians, on the other hand, are apparently capable of
enduring the effect of Superrepression due to doubling of the regulatory gene dosage.
Nevertheless, newly arisen autotetraploid species must be under heavy pressure of
natural selection to escape from the strangle-hold of superrepression. The escape
can be accomplished by the simultaneaus functional diversification of a pair of
duplicated regulatory genes and a pair of regulated structural genes. This should
serve to speed up the process of diploidization. It is small wonder that polyploid
evolution seems so effective in achieving meaningful gene duplication.
References
BAILEY, G. S., CocK, G. T., WILSON, A. C.: Gene duplication in fishes: Malate dehydro-
genase of salmon and traut. Biochem. Biophys. Res. Commun. 34, 605-611 (1969).
BECAK, M. L., BECAK, W., RABELLO, M. N.: Cytological evidence of constant tetraploidy
in the bisexual SouthAmericanfrog, Odontophrynusamericanus. Chromosoma 19,188-193
(1966).
- - - Further studies on polyploid amphibians (Ceratophrydidae). I. Mitotic and meiotic
aspects. Chromosoma 22, 192-201 (1967).
- ScHWANTES, A. R., SCHWANTES, M. L. B.: Polymorphism of albumin-like proteins in
the South American tetraploid frog Odontophrynus americanus (Sa/ientia: Ceratophrydidae).
J. Exptl. Zool. 168, 473-476 (1968).
BENIRSCHKE, K., BROWNHILL, L. E., BEATH, M. M.: Somatic chromosomes of the horse,
the donkey and their hybrids, the mule and the hinny. J. Reprod. Fertil. 4, 319-326
(1962).
CARR, D. H.: Chromosome anomalies as a cause of spontaneous abortion. Am. J. Obstet.
Gynecol. 97, 283-293 (1967).
CASTRO-SIERRA, E., HNO, S.: AlleHe inhibition at the autosomally inhetited gene locus for
liver alcohol dehydrogenase in chicken-quail hybtids. Biochem. Genet. 1, 323-335
(1968).
CoHN, M.: Natural history of the myeloma. Cold Spring Barbor Symposia Quant. Biol. 32,
211-222 (1967).
CROSTON, C. B.: Endopeptidases of salmon ceca: Chromatographie separation of some pro-
perties. Arch. Biochem. Biophys. 112, 218-223 (1965).
ENGEL, W., P'T HoF, J., and WoLF, U.: Genduplikation durch polyploide Evolution: die
Isoenzyme der Sorbitdehydrogenase bei herings- und lachsartigen Fischen (Isospont{yli).
Humangenetik 9, 157-163 (1970).
FANKHAUSER, G., HuMPHREY, R. R.: The origin of spontaneous heteroploids in the progeny
of diploid, triploid and tetraploid axolod females. J. Exptl. Zool. 142, 379-422 (1959).
HrTZEROTH, H., KLSE, J., HNO, S., WoLF, U.: Asynchronaus activation of parental
alleles at the tissue specific gene loci observed on hybrid trouts during early embryonie
development. Biochem. Genet. 1, 287-300 (1968).
KLSE, J., WoLF, U., HrTZEROTH, H., RITTER, H., ATKIN, N. B., HNO, S.: Duplication of
LDH gene loci by polyploidization in the fish order Clupeiformes. Humangenetik 5,
190-196 (1968).
LEBHERZ, H. G., RurrER, W. J.: Glyceraldehyde-3-phosphate dehydrogenase variants in
phyletically diverse organisms. Science 151, 1198-1200 (1967).
MASSARO, E. J., MARKERT, C. L.: Isozyme patterns of salmonoid fishes: Evidences for
multiple cistrons for lactate dehydrogenase polypeptides. J. Expd. Zool. 168, 223-238
(1968).
MoRRISON, W. J., WRIGHT, J. E.: Genetic analysis of three lactate dehydrogenase isozyme
system in trout: Evidence for linkage of genes coding subunits A and B. J. Exptl. Zool.
163, 259-270 (1966).
HNO, S., KrTTRELL, W. A., CHRISTIAN, L. C., STENIUS, C., Wrrr, G. A.: An adult triploid
chicken (Ga!lus domesticus) with a left ovotestis. Cytogenetics 2, 42-49 (1963).
- STENIUS, C., FArssT, E., ZENZES, M. T.: Post-zygotic chromosomal rearrangements in
rainbow trout (Sa!mo irideus GIBBONS). Cytogenetics 4, 117-129 (1965).
- WoLF, U., ATKIN, N. B.: Evolution from fish to mammals by gene duplication. Here-
ditas 59, 169-187 (1968).
PENNOCK, L. A.: Triploidy in parthenogenetic species of the Teiid lizard, Genus Cnemi-
dophorus. Science 149, 539-540 (1965).
SADLER, J. R., NovrcK, A.: The properties of repressor and the kinetics ofits action. J. Mol.
Biol. 12, 305-327 (1965).
SAEZ, F. A., BRuM, N.: Chromosomes of South American amphibians. Nature 185, 945
(1960).
SHAVER, D. L.: The effect of structural heterozygosity on the degree of preferential pairing
in allotetraploids of Zea. Genetics 48, 515-524 (1963).
TRUJILLO, J. M., STENIUS, C., CHRISTIAN, L. C., HNO, S.: Chromosomes of the horse, the
donkey, and the mule. Chromesoma 13, 243-248 (1962).
TsuYuKr, H., GADD, R. E. A.: The multiple hemoglobins of some members of the Salmo-
nidae family. Biochim. et Biophys. Acta 71, 219-221 (1963).
WoLF, U., RITTER, H., ATKIN, N. B., HNO, S.: Polyploidization in the fish family Cyprinidae,
order Cypriniformes. Humangenetik 7, 240-244 (1969).
- ENGEL, W., FAusT, J.: Zum Mechanismus der Diploidisierung in der Wirbel-
tierevolution: Koexistenz von tetrasomen und disomen Genloci der Isozitrat-Dehydro-
genasen bei der Regenbogenforelle (Salmo irideus). Humangenetik 9, 150-156 (1970).
Y AMAMOTO, T.: Progenies of sex-reversal females mated with sex-reversal males in the
medaka, Oryzias latipes. J. Exptl. Zool. 146, 163-180 (1961).
Incorporation of Supernumerary Chromosomes 107
Chapter XVII
Other Mechanisms for Achieving Gene Duplication

The following are mechanisms other than those previously mentioned which
could conceivably contribute to the creation of genetic redundancy, although, to date,
there e:xists no evidence to suggest that these means were actually employed during
the evolution of vertebrates.
1. Polysomy
Nondisjunction, which is a mishap of mitosis, results in both chromatids of one
chromosome moving to the same division pole. In the case of diploid species, this
results in one of the daughter cells receiving three homologaus chromosomes (tri-
somy). If such occurs in germ cells, trisomic individuals may be produced. In the
Jimson weed ( Datura stramonium), the trisomic condition for each of the 12 chromo-
somes in the genome has indeed been o bserved (BLAKESLEE, 1930). A mating between
trisomic individuals of the same type should produce a tetrasomic progeny which
contains two homologaus chromosomes in the genome. In this way, duplication of
one entire linkage group can be accomplished.
Such duplication, however, appears to have been too deleterious to have con-
tributed to vertebrate evolution. The trisomies of larger chromosomes are lethal
conditions, and even those of smaller chromosomes result in sterility. For instance,
trisomic and monosomic conditions constitute a major cause of abortion in man
(CARR, 1967). Although the trisomies of three different kinds of small autosomes
are tolerated to the extent that affected individuals may be born alive, they are mal-
formed and seldom grow to reproductive age (LEJEUNE et al., 1959; PA-rAuet al.,
1960; EuwARDS et al., 1960). The deleterious effect of the trisomic condition effec-
tively precludes the possibility of gene duplication by tetrasomy in mammals and
other vertebrates.
In the tetraploid species which is undergoing the process of diploidization,
however, trisomic (actually pentasomic) as well as monosomic (actually trisomic)
conditions may be tolerated (as in the rainbow trout which was discussed in the
previous chapter). In these species, it appears that it is monosomy rather than trisomy
which is utilized for further evolution. Fot those chromosomes which have failed to
achieve functional diversification, during the process of diploidization the elimination
of two of the four original homologues may be favored by natural selection.
2. Incorporation of Supernumerary Chromosomes

Supernumerary chromosomes (B-chromosomes) were first recognized by LoNG-
LEY (1927) on corn, Zea mqys. Subsequently, their presence has been noted in a va-
riety ofplants (MUN"I"ZING, 1959) as well as in some invertebrate animals such as the
grasshoppet (JoHN and HEwru, 1964). Supernumerary chromosomes are illegitimate
carriers of genetic material, for the rules which govern the behavior of these super-
numeraries are not the rules which opetate for Standard chromosomes. There seems
to be very little genetic homology between standard chromosomes and super-
numeraries of the same species. Supernumeraries, which are usually smaller than the
standard chromosomes, pair only with each other and recombine among themselves
during meiosis. Individuals may accumulate a considerable number of supernume-

raries without any apparent i1l effect. Conversely, individuals do not suffer if they
do not have a supernumerary.
sTERGREN (1945) has compared the role of the supernumerary chromosome to
that of a parasite in the sense that such chromosomes might persist in a population
even when their effects were deleterious to the host. Could it be that the supernu-
merary chromosome is in fact a part of the genome of a parasitic organism? One or
more of the chromosomes of an intracellular parasite which were excluded from
their own cell might have attached themselves to the mitotic spindie fibers of the
host cell. If such is the origin of supernumerary chromosomes, the eventual incorpo-
ration of these illegitimate chromosomes into the host genome is analogaus to viral
lysogeny to be described next. However, it appears that such a mechanism of gene
duplication did not contribute to the evolution of vertebrates, for no vertebrate
species has been found to carry supernumeraries.
3. Lysogeny: Incorporation of the Viral Genome

Those viruses which use bacteria as their hosts are known as bacteriophages.
Normally, bacteriophages invade the bacterium and rapidly multiply within, finally
killing the host by lysis. Under certain conditions, however, the phage and the bac-
terium establish a state of synbiosis. DNA of the phage is inserted into the bacterial
genome, and both genomes replicate synchronously during exponential growth of
bacteria.
Lysogenization of Salmonella by bacteriophages causes the appearance of new
antigenic determinants and often produces weakened reactivity of already existing
determinants. Study of these lysogenic conversions has proved especially rewarding
in delineating the kinds of genetic information contained within the phage genome
which influence and modify biosynthesis of different antigenic determinants on the
cell wall ofthe host (RoBBINS and UcHIDA, 1962).
Oncogenic DNA viruses, such as simian papova-virus 40 (SV 40), affect mamma-
lian somatic cells in two different ways. These viruses multiply within permissive
cells and eventually kill them, but when the virus encounters non-permissive cells,
it establishes a state of synbiosis with the host. As a result, the host cells are often
transformed to malignant cells. While infective viruses can na Ionger be recovered
from transformed malignant cells, those transformed cells are marked by the persistent
presence af a newly arisen antigen (T-antigen). AsT-antigen is apparendy specified by
a cistron within the viral genome, the incorporation of the viral genome into the host
genome is suspected in transformed malignant cells. Indeed, WESTPHAL and DuLBEcco
(1968) utilized the technique ofDNA-DNA hybridization to show that the genome
of transformed malignant cells contains multiple copies of SV 40 DNA.
Can the origin of certain genes in the vertebrate genome be traced ta the infective
virus of the past which became permanently incorparated inta the hast gename? In
mammals, same of the blaod group genes such as that far B-system of cattle (STOR-
MONT, 1965), as well as same af the histocompatibility genessuch as the H-2locus of
the mause, are extremely camplex structures, each accupying a long stretch af the
chromosame. BAILEY and KoRN (1965) suggest that mutatian-like and recombination-
like events which they observed at histocompatibility gene loci of the mause can best
References 109
be explained by viral lysogeny. They found that most of the mutation-like events
observed on Fchybrid mice between two inbred strains represented the gains of
hitherto nonexistent histocompatibility antigens. For this and other reasons, they
proposed that the observed situation was analogaus to that found in lysogenized
Salmonella bacteria, in that new antigens were specified by the viral genome which
became incorporated into a region of the histocompatibility gene locus.
4. Viral Transducdon
When the phage genome which had existed in lysogenized bacteria breaks free
from the state of synbiosis and becomes infective again, it might carry an adjacent
piece of the host genome with it. When such a phage establishes a new state of syn-
biosis with another bacterium, the incorporation of the phage genome into the ge-
nome of the new host would be accompanied by that of a piece of the previous host's
genome. This, then, is the process of transduction.
A hypothetical situation is presented here to visualize the process of gene dupli-
cation by viral transduction in vertebrates. Let us assume that in both the hamster
and the mouse, the thymidine kinase locus is situated adjacent to the site on a certain
chromosome where the insertion of the viral genome preferentially occurs. When
this virus breaks free from the hamster host genome and becomes infective, it might
carry with it the hamster cistron for thymidine kinase. The next incorporation of the
viral genome by the mouse host results in taudem duplication of thymidine kinase
genes, as the hamster thymidine kinase cistron is inserted in front of the mouse
thymidine kinase locus.
No known case of gene duplication by viral transduction is available in verte-
brates, but such evidence may weil be forthcoming.
References
BAILEY, D. W., KaHN, H. I.: Inherited histocompatibility changes in progeny of irradiated

and unirradiated inbred mice. Genet. Res., Camb. 6, 330-340 (1965).
BLAKESLEE, A. F.: Extra chromosomes: A source of variation in the Jimson weed. Smith-
sonian Repts 1930, 431--450.
CARR, D. H.: Chromosome anomalies as a cause of spontaneaus abortion. Am. J. Obstet.
Gynecol. 97, 283-293 (1967).
EDWARDS, J. H., HARNDEN, D. G., CAMERON, A. H., CROSSE, M., WoLFF, 0. H.: A new
trisomic syndrome. Lancet 1960 I, 787.
joHN, B., HEwrrr, M.: The B-chromosome system of Myrmeleotettix maculatu.r (THUNB).
Chromesoma 16, 548-578 (1965).
LEJEUNE, J., GAUTIER, M., TuRPIN, R.: Etude des chromosomes somatiques de neuf enfants
mongoliens. Campt. rend. 248, 1721-1722 (1959).
LoNGLEY, A. E.: Supernumerary chromosomes in Zea may.r. ]. Agr. Res. 35, 769-784 (1927).
MuNTZING, A.: A new category of chromosomes. Proc. 10th int. Congr. of Genedes 1, 453
to 467 (1959).
sTERGREN, G.: Parasitic nature of extra fragment chromosomes. Botan. Notiser 2, 157 to
163 (1945).
PATAU, K., SMITH, D. W., THERMAN, E., INHORN, S. L., WAGNER, H. P.: Multiple con-
genital anomaly caused by an extra autosome. Lancet 1960 I, 790-793.
RoBBINs, P. W., UCHIDA, T.: Determinants of specificity in Salmonella: Changes in antigenic
structure mediated by bacteriophage. Federation Proc. 21, 702-710 (1962).
110 References
STORMONT, C.: Mammalian immunogenetics. In: Genedes today. Vol. 3, Chapter 19,
pp. 716-722 (GEERTS, S. J., Ed.). New York: Pergarnon Press 1965.
WESTPHAL, H., DuLBEcco, R.: Viral DNA in polyoma and SV 40-transformed cell lines.
Proc. Natl. Acad. Sei. US 59, 1158-1165 (1968).
Part 5
Evolution of V ertebrate Genomes
Chapter XVIII
Primitives Inherit the Earth

In addition to the Lamarckian illusion pointed out in Chapter IX, another illusion
persists with regard to the nature of evolution. We tend to view evolution as a
purposeful process by which simple creatures, which were primitive, changed step by
step to become more and more complex creatures, which are advanced. The commonly
held notion resulting from this misunderstanding is that amphibians evolved from a
type of fish which had achieved a pinnade of complexity as a fish, and that reptiles
evolved from the most advanced form of amphibians. The chronology of vertebrate
evolution briefly reconstructed in this chapter shows quite dearly that the contrary
has been the case. A big leap in evolution to a more advanced dass has always been
commenced by one of the more primitive members of the previous dass.
Why has it always been the more primitive of the previous dass which made a
big leap and evolved toward a higher dass? This apparent paradox disappears when
we substitute the words "uncommitted and generalized" for the word "primitive",
and "committed and specialized" for "advanced".
1. Emergence of the First V ertebrates

At the beginning of Cambrian times, some 500 million years ago, there first
appeared in the fossil record a considerable variety of invertebrate metazoan animals.
There were trilobites and crustaceans of the phylum Arthropoda, snails of Mo!!ttsca,
lamp shells of Brachiopoda, and some echinoderms. It is believed that the original
vertebrate ancestors were the most uncomplicated of the invertebrate creatures of
that time. They were simple sessile forms which resembled tiny, deep sea forms of the
dass Pterobranchia of today. Although this dass is placed in the phylum Chordata, the
body of pterobranchs, attached by a stalk to the ocean bottom, consists of little
except a digestive tract. Positioned around a mouth, arms with ciliated bands extend
to catch food partides drifting past in the water. Such a simple creature can be com-
pared to a sheet of white paper for only on white paper can a new style of manuscript
for evolution be written. More advanced forms, such as members of the phylum
Arthropoda, bad already committed themselves to remain invertebrates.
Further development from this humble beginning appears to have been accom-
plished by the creation of gill slits - paired openings on either side of the body leading
112 Evolution of Vertebrate Genomes
from the throat (pharynx) out to the surface. Although gill slits initially served as
food strainers, they were to be modified to breathing apparatus in the future. At this
stage of development, our ancestors must have resembled tunicates of today which
belong to the class A1cidiacea. As sessile filter-feeders, the forms represented by tuni-
cates have reached the evolutional dead end. Yet, what were the characteristics of the
larval forms of these tunicate-like creatures? In evolution, the process called paedo-
morphosis happened time and again. The adult stage was eliminated, and the larval
form became sexually mature and reproduced. It appears that from the larva of a
"'C
.g_
0>0 Living relatives
..
O.CJJ
c:N
'
c~
~!:?.
c
&
.,
~
Hagfish Lamprey
m
>
c:
~
.E
8N
Amphioxus
8
e
~
a:
w
~
0
N
0
w
...J
Pterobranchs
Fig. 19. A schematic drawing indicating the emergence of the first vertebrate by paedo-
morphosis from the larval form of tunicate-like creatures
tunicate-like form of ancient times there arose the body type from which the first
true vertebrates sprang. In the free-swimming tadpole-like larva of certain modern
tunicates, the gill apparatus is situated in its swollen "head" region. Behind, there is
a muscular swimming tail, strengthened by a stout, but flexible, longitudinal noto-
chord; predecessor of the vertebral column. There exists a longitudinal dorsal nerve
cord which, in the head region, receives sensory information from rudimentary
sense organs.
The first chordates of ancient times which emerged by paedomorphosis were
probably not too different from amphioxuses oftoday (YoUNG, 1950; RoMER, 1967).
Indeed, the fossil of a small, primitive, unarmored, jawless chordate was found in the
Silurian rock in England. This creature named jamoytim was constructed along the
line of the present day amphioxus, and by the dawn of the Devonian period (approxi-
The Type of Fish which were Prepared to Become Land Living Animals 113
mately 320 million years ago), early vertebrates in the form of jawless fish of the dass
Agnatha were found in abundance. These jawless forms of ancient times are collec-
tively designated as ostracoderms (Fig. 19).
2. The Type of Fish which were Prepared to Become Land Living Animals
As already mentioned in Chapter XIV and illustrated in Fig. 14, the first major
anatomical improvement that occurred to early vertebrates was the development of
jaws. Hagfish and lampreys, however, persist to this date as evolutional relics repre-
senting the dass Agnatha. Starting from the earllest of the jawed vertebrates of the
upper Silurian and lower Devonian times (dass, P lacodermz), all of the principallinef>
of piseine evolution proceeded to improve the jaw structures, to wear protective gill
covers (operculum) and to attain an advanced degree of ossification of the axial
skeletons. Thus, the dass Osteichthys (bony fish) was boro. Sharks and rays chose not
to perfect the improvement of jaws, not to wear gill covers and not to ossify the
axial skeletons at this early date and so established a dass of their own; Elasmo-
branchii.
In the middle Devonian period, the bony fish diverged into two distinct lineages.
An important difference between the two was that one had ray-fins, while the other
had muscular lobe-fins. Cheirolepis ofthat time characterized the ray-finned lineage
( Actinopterygii). The contemporary of Cheirolepis, which represented the lobe-finned
lineage (Choanichthyes), was Osteolepis. Cheirolepis and Osteolepis, however, had many
features in common as shown in Fig. 20. Their hoclies were covered by heavy dia-
mond-shaped rhombic scales, and their backbones turoed upward at the tail (hetero-
cercal tail). They also had well developed airbladders opening to the pharynx which
probably served as very functional air breathing apparatus (lungs). Among bony fish,
these features have been out of style for eons, and only the most primitive, such as
sturgeons, bowfins and gar-pikes of the subdasses Pa/eopterygii and Ho/ostei, preserve
such primitive features to this date. In addition, in Osteolepis, two pairs of muscular
fins were placed in appropriate positions on the sides of the body. The anterior pair
on the ehest could be easily modified to fore limbs, complete with pectoral girdle
(shoulders), of terrestrial vertebrates. The posterior pair was placed well behind the
middle of the body, so that they could be changed to hind limbs, complete with
pelvic girdle. It becomes dear that only the primitive bony fish of the Devonian
period which lived in shallow waters of continental drainage systems had enough
ambiguity in their features to evolve toward terrestrial vertebrates. Indeed, rhipi-
distian descendants of Osteoiepis induded venturesome species which came out of the
water to become the first amphibians at the end of the Devonian period. As fish, the
lobe-finned lineage has been singularly unsuccessful. Today only three genera of
lungfish (order Dipnoi) which still live in shallow fresh water and one species
(Latimeria) of Coe/ocanthini which lives in the ocean represent this lineage. In sharp
contrast, the ray-finned lineage has continu ~d to modernize and during the Cretaceous
period (170 million years after Cheirolepis) emerged as teleosts. Subsequent diversi-
fication of teleosts is a true success story ofevolutiTon.hey inhabit the deep ocean as
well as lakes, streams and estuaries. Consequently, they are at the present time the
most numerous of the vertebrates. The species of teleostean fishes outnumbers by far
all others species of recent vertebrates combined, and, in the number of individuals,
114 Evolution ofVertcbrate Genomes
some of the marine fishes reach population figures of almost astronomical magnitude.
Modem teleosts have evolved a long way from the primitive bony fish of the De-
vonian period (MoY-THOMAs, 1939; RoMER, 1946.)
PRESENT
I
I c:
l,g
I;; "'~
IE ~
IN >-
1a: END OF
DEVONIAN PERIOD
....
c:
0
c:
Cl.
0
THE RAY-FINNED LINEAGE THE LOBE-FINNED LINEAGE
Fig. 20. The early dichotomy of the Devonian fish to the ray-finned lineage as represented
by Cheirolepis and the lobe-finned lineage as reprcsented by Osleolepir. Orteolepir was the
rhipidistian crossopterygian which evolved to the first amphibian, lchthyorlega. Aside from
rhipidistians, two other branchcs of lobe-finned fish developed in the Devonian time.
Living lungfish are the descendants of the dipnoan branch, while living Latimeria belong to
the coelacanth branch
Evolution from Fish to Amphibians 115
3. Evolution from Fish to Amphibians

In upper Devonian sediments in Eastem Greenland, fossils of lchthyostega were
discovered. lchthyosfega was apparently the first amphibian directly descended from
t-
zw
U)
w
a:
0..
.."'
c:
.21
..C:(
g..
c: I
~
<!I
200 million
"'
e~
.,.-
~
..o"-
0 "'
~"C
.. c:
<t: u ..
a:
w
u
6N
0 280 million
w
...J
c:
"'
;:
1'~-~
0
&;
0
Crossopterygian fish
Fig. 21. Chronology of the evolution of amphibians
osteolepids. They were intermediate between crossopterygian fish of the suborder

Rhipidistia and characteristic early amphibians. lchthyostegids, however, were
already equipped with strong pectoral and pelvic girdles, with which were articulated,
completely developed limbs and feet quite capable of carrying the animal around on
land.
s
The direction of further development followed by amphibians during the Carboni-

ferous and Permian periods was: 1. to develop the tough skin frequently underlain by
bony plates to prevent evaporation of body fluid, 2. to strengthen the vertebral
column by producing interlocking vertebrae that, with the aid of musdes and Iiga-
ments, formed a strong horizontal column to support the body, and 3. to improve
air-breathing capability by forming well-developed nasal passages which opened to
the nostrils located on the dorsal surface of the skull. Throughout the Carboniferous
and Permian periods, the main groups of amphibians continued to evolve along the
above mentioned direction. However, one of the early descendants of the ancestral
ichthyostegids, collectively known as embolomeres, had already committed them-
selves to evolve toward the reptilian direction in early Carboniferous times, and the
first reptile, Cotylosaurus, emerged within 40 million years (just before the beginning
of the Permian period) after the emergence of the first amphibian. Meanwhile,
remaining members of the main labyrinthodont branch of amphibians continued to
modernize, and large, more advanced labyrinthodont amphibians competed on more
or less equal terms with their reptilian contemporaries until well after the beginning
of the age of dinosaurs (Fig. 21).
Of amphibians living today, frogs and toads of the order Salientia or Anura
appear to be the direct descendants of the labyrinthodont amphibians. The ancestry
of urodeles (salamanders, mudpuppies and newts), on the other hand, is traced back
to a quite different origin. From the very beginning of amphibian evolution, there
existed a minority lineage which was quite different from the main labyrinthodont
lineage. The amphibians of the minority lineage are collectively called Lepospom!Jls.
In the labyrinthodonts, the body elements of the vertebrae were pre-formed in
cartilage and the deposition of calcium phosphate occurred afterward; aspidospon-
dylous vertebrae. In the lepospondyls, on the other hand, the vertebrae were not
pre-formed in cartilage, but rather were formed directly as spool-like, bony cylinders
around the notochord, and were generally united with the neural arch. Urodeles are
modern lepospondyls (COLBERT, 1955).
4. The Invention of the Amniote Egg and the Emergence of

Reptiles and Birds
Although the amphibians which emerged at the end of the Devonian period were
no doubt the first land-living vertebrates, in a sense, they too have always been bound
to water. They returned to the water to reproduce and their larvae were equipped
with gills to lead an aquatic life. During Carboniferous times, a major innovation in
the evolution of land-living vertebrates occurred. This event was the development
of the amniote egg. The significance of this innovationwas comparable to one which
occurred earlier; the transformation of the third gill arch into a pair of jaws.
The first animals to have the amniote egg were reptiles. Once equipped with the
amniote egg, the larval stage was eliminated from ontogenic development. This egg,
which contained a large yolk that furnished nourishment for the development of an
embryo, was internally fertilized. A fertilized amniote egg developed two sacs; the
amnion, which was filled with a liquid, contained an embryo, and the allantois,
which served as a waste bag. The entire structure was encased in a shell that was
strong enough to protect the contents of the egg, yet porous enough to allow the
passage of oxygen into the egg and the passage of carbon dioxide out of the egg.
Invention of the Amniote Egg and the Emergence of Reptiles and Birds 117
With the advent of the amniote egg, it finally became possible for terrestrial
vertebrates to wander freely over the land. Although the oldest known amniote egg
was found in lower Permian sediments in North America, the actual transformation
.,
"'g
.,
(.)
~
u
<{
er:
IMAMMALsl ,I BIRDS
I 130
w million
u
0 .,
5"'
N ~
0
r.n
w .Bi
0
., I)! "'.,0
1'0
~
:E 5
....,
.s
.,
2 ~y 0
165
"'0
I)! million
.,
."0
I
I
V
I
t;"
-0
l"'
..............
"' ........
~
V .........
,( 200 million
~
~ ~.,.,,;,, ~
<(
er: "'c
w ~"'
u ~E
.,
., .-
,,,.,.;,, I~
0 0., Cotylosaurus (1st reptile)
N .Oe..
0 .,
~"0
c Amnioteegg
w u ..
...I
<(
c..
Embolomeres (Amphibian)
Fig. 22. Three subclasses of reptiles and their descendants. In the schematic drawings of
reptilian skulls at the bottom, each temporal opening, if present, is shown in solid black
from one of the early labyrinthodont amphibians (embolomeres) to reptiles (cotylo-

saurs) occurred considerably earlier.
At the end of the Permian period which closed a chapter on the Paleozoic era,
highly varied chondrostean fishes inhabited the streams and ponds of the continents
and the oceans that surrounded them, while numerous and diverse labyrinthodont
amphibians lived in the watery and swampy environments of the land area. At the
same time, reptiles, being the newest models of vertebrates, were busy establishing
their pattem of dominance that was to come at the dawn of the Mesozoie era.
The descendants of cotylosaurs first diverged into five subclasses, of which only
the following three are pertinent: 1. Diapsida, 2. Anapsida and 3. Synapsida. Of the
three, Diapsida was the most successfullineage as reptiles. A most persistent charac-
teristic of this subclass was that their skulls had two posterior temporal openings
separated by the postorbital and squamosal bones.
At the beginning of the Mesozoie era (Triassic period), thecodonts emerged to
establish the archosaurian branch of Diapsida. The success story of reptiles, which
culminated in the emergence of dinosaurs, is familiar to all of us. Dinosaurs comple-
tely dominated the land from the late Triassie to the end of the Cretaceous period
(a span of 100 million years). The Mesozoie era which ended barely 70 million years
ago was truly the age of reptiles. Some of the dinosaurs which weighed as much as
50 tons were indeed the largest and most terrifying animals that stalked the earth.
Y et, the triumph of reptiles was in a strict sense the triumph of only the archosaurian
branch of Diapsida. Members of Anapsida and Synapsida merely lived through the
Mesozoie era under the shadow of dinosaurs. In the same manner, the triumph of
mammals durlog the past 70 million years has been the triumph of placental mammals
and not of marsupials and monotremes.
In addition to dinosaurs, the archosaurian lineage gave rise to a new class of
vertebrates; birds. It appears that certain members of the archosaurian lineage main-
tained the original bipedal posture of thecodonts of the suborder Pseudost~Chia, and
their descendants evolved to two forms of flying vertebrates durlog the Jurassie
period (165-230 million years ago). There were flying reptiles collectively known as
pterosaurs which have long been extinct and birds which have enjoyed tremendous
success. As the class Aves was derived from the powerful archosaurian lineage at the
pinnacle of its success, the rule of primitive members of the previous dass always
malcing a big leap to a more advanced class appears to have been violated. However,
it should be realized that no major innovation was involved in creating birds from
reptiles. This is the very reason that birds are often referred to as "feathered reptiles"
(Fig. 22).
Of living reptiles, lizards and snakes of the order Squamata and crocodiles and alli-
gators of the order Croco4Jiia have descended from members of Diapsida. Turtles and
tortoises of the order Chelonia, on the other hand, are descendants of the Anapsida
line. A characteristic of Anapsida is that there is no temporal opening in the skull
behind the eye. The first reptile, Corylosaurus, also had no temporal opening in the
skull behind the eye (fiEILMANN, 1927; COLBERT, 1951).
5. Synapsida and the Emergence of Mammals

Members of the subdass Synapsida never gained eminence as reptiles. Instead
they proceeded directly from the primitive reptilian state to evolve toward mammals.
Clearly, synapsids bridged the gap between primitive reptiles and mammals, for its
early members were very close indeed to the first reptiles, Corylosaurus, whereas some
of its latest genera approached so closely to the mammalian state it is a debatable
Synaptida and the Emergence of Mammals 119
question whether the animals should be classified among reptiles or among mammals.
A persistent characteristic of this subclass was that the skull had a single temporal
opening bordered by the postorbital and squamosal bones. Unlike diapsids which
< <
:v;- :_
WO> <]" ~
::I:E
f-G>
Wo!!
:I:.!!! -"'
a::;; :;;
ob I-Cl. w .~ 1i ~ g>
<(
a:
t-0
o5 ...
<("
w"'
~ i= ~ = ; 2
w 0::::!!: ::!!:~ ~2 . ~ ~
D..-
u
0
N
0
z
w
t
u
- 130
<(
a:
I t
Multituberculates
'+'
u Symmetrodonts
0
N Pantotheres
T'loooodo~
0 115
(/)
w
::!!:
~~
0
Mammal-like reptiles
Fig. 23. A diagram illustrating the origin of monotremes, marsupials and placentals from
the Jurassie mammal-like reptiles
originally assumed the bipedal posture, synapsids maintained the quadrupedal

posture throughout their evolution.
The first reptile recognizable as a synapsid appeared in rocks of the late Carboni-
ferous period of the Paleozoic era, as the earllest descendant of Cotylosa~~rus. Among
those synapsids of the order Therapsida, the most mammal-like were carnivorous
reptiles of various sizes, which appeared during middle Permiau times of the Paleo-
zoie era and developed during the Triassie period of the Mesozoie era. They are
known collectively as theriodonts. Subsequently, a group of reptiles which appeared
to bridge the gap between advanced theriodonts and primitive mammals made their
appearance. They were the ietidosaurs of the Triassie and early to middle Jurassie
periods. The ietidosaurs, which were generally small animals, almost crossed the
threshold that separated reptiles from mammals, and in so evolving they brought
about the doom of the synapsids, and eventually of dominant reptiles. Although the
ietidosaurs established the foundation for the subsequent rise and triumph of mam-
mals in the early part of the Mesozoic era, the age of mammals was not to come for
many millions of years (Fig. 23). During a time lapse of perhaps 100 million years
after the rise of the first mammals from ietidosaurian ancestors, the dinosaurs and
other dominant reptiles of the Diapsida line were to reign supreme (BROOM, 1932;
CoLBERT, 1955).
6. Mammals in General
During the Jurassie period, the threshold from ietidosaurian reptiles to mammals
was crossed. As the name implies, the young of mammals are nourished on milk
supplied by the mother during the early stage of postnatallife. Mammals are poikilo-
thermic animals which use hair for body cover. In exchange for the attainment of a
high rate of metabolism which is correlated with the formation of the four-chambered
heart, mammals have acquired a definite aging process, in that the prime of life is
invariably followed by senescence. The following are other more notable mammalian
characteristics. In mammals, the quadrate and articular bones (the articulating ele-
ments between the skull and jaw in reptiles) have retreated into the middle ear to
become transformed into two of the three ear ossicles, the incus and malleus, that
together with stapes (inherited from the reptillau stapes) make a chain to transmit
vibrations from the eardrum to the inner ear. This no doubt improved the hearing of
mammals. There is a diaphragm which separates the body cavity into the thoracic
portion and the abdominal portion. The presence of this diaphragm aids inhalation
and exhalation of air. There is a notable change in dentition. Teeth are differentiated
into incisors, canines and cheek teeth. As to the skeletons, ribs which used to stick out
from cervieal as weil as lumbar vertebrae of reptiles are greatly reduced in size or have
disappeared. Three pelvie bones, the ilium, the ischium, and the pubis are fused to
form a single bony structure. Needless to say, of partieular importance is the com-
paratively large brain case found in all but the mostprimitive mammals. This enlarged
brain case is a reflection of the growth of the brain and the greatly increased intelli-
gence of mammals.
Living mammals belong to three different subclasses; Prototheria, Metatheria and
Eutheria. Of these, monotremes representing Prototheria are the most peculiar. The
platypus or duckbill, Ornithorl?J!nchus and the spiny anteaters, Echidna and Tacl?J!glosstts,
that inhabit Australia and New Guinea reproduce by laying hard-shelled eggs which
are hatched in burrows. The rectum and urogenital system open into a common
cloaca as in reptiles and birds, not separately as in other mammals. However, the
young are suckled on milk that is secreted by modified sweat glands of the mother.
Although monotremes are unknown in the fossil record previous to the Pleistocene
Primates and Man 121
epoch (only one million years ago), there is every reason to believe that other mam-
mals did not go through a Prototheria-like stage of evolution and that monotremes
represent a separate line of descent from the mammal-like reptiles of the Jurassie
period.
As to marsupials of Metatheria and placentals of Eutheria, the fossil record of
placentals is as ancient as that of marsupials. Thus, it appears that marsupials also do
not represent an intermediate step of evolution through which placentals finally
emerged. Rather the two groups appear to have risen simultaneously from Panto-
theres which existed at the end of the Jurassie period (Fig. 23). It is probable that
during Cretaceous times the marsupials were widely distributed throughout the
world. Through the duration of 60 million years (the Cretaceous period), they very
likely shared the land with the primitive placentals, and for a time the two groups
may have competed on more or less equal terms. But with the advent of the Cenozoie
era (70 million years ago), there was a great evolutionary hurst of the placentals that
brought about a stringent reduction, if not complete extermination, of the marsupials
in most continental regions. Needless to say, placental mammals are those in which the
young go through a considerable period of prenatal growth to be born as a miniature
replica of their parents. In ontogenie development of placentals, the allantoie mem-
brane inherited from the original amniote egg comes in direct contact with the
mother's uterus, and through this area of contact (the placenta) nourishment and
oxygen are carried from the mother to the developing embryo.
The earllest and most primitive eutherians were the insectivores, and it seems
obvious that they were probably the ancestors of all other placentals. From a proto-
insectivore stem, the placentals radiated into 28 diverse orders, 11 of which are now
extinct (SIMPSON, 1945).
7. Primatesand Man
There is little doubt that primates are direct descendants of the insectivores.
Indeed, primates retain many generalized eutherian characteristies, and, therefore,
are not highly specialized animals. The order Primates can be considered the second
most primitive order among placentals; Insectivores being first. The modern Odental
tree shrews (Tupaia), which are probably constructed similarly to the most ancestral
type of Primates, can be regarded as insectivores that have adapted themselves to
living in trees. In fact, many of the evolutional improvements made by primates are
the direct results of adapting themselves to living in trees. Such improvements in-
cluded binocular vision, a large brain, an increase of intelligence and grasping hands.
From a probable tree shrew ancestry, primates evolved along three principle lines.
There was an initial radiation of primitive primates in Paleocene times (70 to 55
million years ago), and a suborder Prosimiiwas established. Lemurs, loris and tarsiers
belong to this suborder. There was a second radiation of primates during the upper
Eocene epoch (approximately 40 million years ago) which established the second
suborder Simiae. New World as well as Old World monkeys belong to this suborder
(Fig. 24).
At last we come to the superfamily Hominoidea which includes our own species.
The origin of hominoids can be traced back to the lower Oligocene sediments of
Egypt (approximately 35 million years ago). Although Propliopithecus (the ancestral
hominoid) was very small in size, a more noticeable trend which characterized the
122 Evolution ofVertebrate Genomes
evolutional history of hominoids was a strong inclination toward a great increase in

the size of the body as well as the brain. As the body size increased, most of them
became brachiating animals, swinging by their arms from branch to branch. Con-
"'
c:
..8
..0
c,
Pithecanthropus
(Homo erectus)
- 1 million
australopithecus
I
"'c:
~
,Q
ii;
~
a:
w
-11 million
~ 1----+-- - l - -
""' /
N
0
zw
u Limnopithecus proconsul
Hominoidea
26 million
"'c:
"'
.~

36 million
..
., c:
c: .,
.. u
~Jl
..!!!-o
t'OC
o. ..
Fig. 24. The family tree of primates diagrarnmatically illustrated
sequently, there came the development of long arms equipped with long fingers.
The bind legs became short and the extemal tail disappeared. Mter Propliopithecus,
the next hominoids to appear were Limnopithecus and Proconsul from the lower Mio-
cene sediments of Mrica (approximately 25 million years ago). It appears that Limno-
pithecus represented a side branch which eventually led to the modern gibbons
References 123
( Hylobates), while the more generalized Proconsul served as the ancestor to all other
modern hominoids. The descendants of Proconsul were the various dryopithecines
of which the genus Dryopithecm was typical. These dryopithecines flourished in wide
areas of Eurasia during the Miocene and Pliocene epochs. This rather chimpanzee-
like ape was probably ancestral not only to modern chimpanzees but also to gorillas
and orangutans and possibly to the hominid line.
Let us look at the origin of our own species (Homo sapiens). It appears that our
hominoid ape ancestor made a very significant change in his way of living, perhaps
during early Pliocene times (10 million years or so ago). The forest-living hominoids
were peaceful, largely fruit and vegetable-eating animals. As the earth began to cool
off, lush tropical forests which covered many parts of the earth started to retreat, and
at that time our ancestors apparendy exchanged life in the forests for life on open
ground. In so doing they became omnivorous, and as hunters they came into direct
competition with the already successful carnivorous animals. Moving in packs, the
hominoids triumphedas hunters. During this adaptation to a new way of life, natural
selection appears to bave strongly favored an erect posture and an increase in intelli-
gence. The firstsuch hunting man-apes (pre-hominids) known to us were .Amlralopi-
thecNS and their allies. A skull of Amlralopithecm ( Zinjanthropm) boisei found in Olduvai
Gorge, Tanzania is believed tobe 1.75 million years old (ToBIAS, 1967). Here again,
we witness the instance of the primitive triumphing over its more advanced con-
temporaries. When the earth began to cool off at the beginning of the Pliocene epoch,
those apes which bad already attained a high degree of specialization as brachiating
animals must bave bad no choice but to retreat with the forests. It must be that our
ape ancestors of that time were the type which were sufficiendy primitive and un-
committed to a way of life in forests. For this very reason, they were able to remain in
the open ground to try a new way of life.
Further development of these hunting man-apes, which culminated in the emer-
gence of our own species, apparendy occurred within the last one million years
(Pleistocene epoch) which experienced the four successive periods of glaciation.
There were Pithecanthropm (Homo erectus) and Homo neanderthalensis. While Homo
neanderthalensis became extinct during the last glacial period, Homo sapiens with the
mean cranial capacity of 1,350 cc emerged after the last glacier (only 50,000 years or
so ago) as the only living species of the genus Homo (LE GRos CLARK, 1959).
References
BROOM, R.: Mammal-like reptiles of South America. London: H. F. & G. Witherby 1932.
CoLBERT, E. H.: The Dinosaur book, 2nd ed. New York: McGraw-Hill Book Co. 1951.
- Evolution ofthe vertebrates. Science Edition. New York: John Wiley & Sons, Inc. 1955.
HEILMANN, G.: The origin ofbirds. New York: D. Appleton & Co. 1927.
LE Gaos CLARK, W. E.: The crucial evidence for human evolution. Proc. Am. Phil. Soc. 103,
159-172 (1959).
MoY-THOMAS, J. A.: Paleozoic fishes. New York: Chemical Publishing Co. 1939.
RoMER, A. S.: The early evolution of fishes. Quart. Rev. Biol. 21, 33-69 (1946).
- Major steps in vertebrate evolution. Science 158, 1629 -1637(1967).
SIMPSON, G. G.: The principles of classification and a classification of mammals. Bull. Am.
Museum Nat. Hist. 85, 1-350 (1945).
ToBIAS, P. V.: Olduvai Gorge, Vol. II. New York: Cambridge University Press 1967.
YouNG, J. Z.: The life of vertebrates. Oxford: Clarendon 1950.
Chapter XIX
Nature's Great Experiment with Gene Duplication

during Evolution from Tunicate-like Creatures to Fish
1. The Genome Size of Tunicate-like Creatures from which

V ertebrates Emerged
Of the various invertebrates which existed nearly 500 million years ago in
Cambrian times, it was the simplest sessile form which gave rise to the first vertebrate
through tunicate-like creatures. What was the genome of these uncommitted creatures
like? I would venture a guess that the genome too had to be in an uncommitted state
containing the bare essential number of structural genes, and having the least degree
of redundancy. Only by starring in this way was it possible to utilize genetic redun-
dancy subsequently created to write a new chapter in the evolution of metazoan
animals.
Of the primitive chordates of today, the tunicate of the Pacific coast (Ciona
intestinalis) belongs to the subphylum Tunicata and produces free-swimming tadpole-
like larvae. Our measurement has shown the genome size of Ciona to be only 6% of
that of placental mammals; the haploid set of chromosomes containing about
0.21 mg X 10-9 DNA (ATKIN and HNO, 1967). This is indeed a very small genome
size, for some invertebrates, such as grasshoppers, are endowed with genomes which
are comparable in size to the genomes of mammals.
Since the time of BoVERI (1890), it has been known that primitive chordates of
today are characterized by having a small number of minute, slender chromosomes
less than one micron in width. The diploid complement of Ciona intestinalis (2n = 28)
contains 14 pairs of minute acrocentrics, while that of another tunicate (Styela plicata,
2n = 32) belanging to a different order has 16 pairs of equally minute acrocentrics
(MrnoucHr, 1936; T AYLOR, 1967). Inasmuch as the extreme smallness of the genome
size is a universal characteristic of living tunicates, it appears safe to assume that
tunicate-like creatures of nearly 400 million years ago whose tadpole-like larvae
furnished the vertebrate prototype also had similarly small genomes.
By studying amphioxuses of today, one rnight get some idea on the question of
what the genome of the first chordate which emerged by paedomorphosis was like.
Amphioxuses of the subphylum Cephalochordata have also been noted to have very
minnte chromosomes. The diploid chromosome number of32 was found in Amphioxus
belcheri of Japan (NoGusA, 1960). Our measurement on Amphioxus lanceolatus revealed
that its genome size is only 17% of that of placental mammals; the haploid set of
chromosomes containing about 0.60 mg x 10-9 DNA (ATKIN and HNO, 1967).
As will be shown, the smallest genome size to be found among fish and for that
matter among all vertebrates is about the same or a little more than that of the am-
phioxus. In other words, the rninimum genome size possessed by many diverse
species of fish is not too different from that of amphioxus. On this basis, I venture a
guess that paedomorphosis from tunicate-like creatures to amphioxus-like creatures
which took place either during the Ordovician period or the Silurian period was
accompanied by a two to threefold increase in the genome size; from about 6% to
Extreme Diversity of Genome Size Exhibited by Fish 125
about 17% of that of placental mammals. Whether this increase was accomplished
exclusively by taudem duplication or by a combination of taudem duplication and
tetraploidy cannot be resolved at the moment.
2. Extreme Diversity of Genome Size Exhibited by Fish

The genome sizes of representative species of fish belanging to 17 different
orders and representing 6 subclasses are shown in Table 3. Genome sizes of primitive
chordates which have already been discussed are also shown. In our series of studies
(HNO and ATKIN, 1966; ATKIN and HNO, 1967; HNO et al., 1967; MuRAMOTO
et al., 1968; HNo et al., 1969; WoLF et al., 1969), the nuclear Feulgen stain content
of each species was compared to that of man using the Deeley integrating micro-
densitometer, incorporating a crushing condenser (DEELEY, 1955). Thus, the genome
size was originally expressed as a percentage ofthat of man. Since the haploid DNA
content of man and other placental mammals is about 3.5 mg x 10-9, each per-
centage value was also translated to the haploid DNA content in Table 3. HrNE-
GARDNER (1968) has also clone a rather extensive study on genome sizes of diverse
teleost fish using a fluorometric method. A number of species were examined both
by HrNEGARDNER and ourselves, and, for each species, the values obtained were
nearly identical. Of the values shown in Table 3, those indicated by asterisks were
obtained by HrNEGARDNER.
Observing Table 3, it should be noted that the smallest genome to be found in
vertebrates is possessed by remotely related teleost fish belanging to diverse orders.
The genome of the surf smelt, the tetrazona barb, the swordtail, the hornyhead
turbot, the fantail sole, the sea horse and the puffer is slightly larger or about the
same as that of the amphioxus, which is 17% of the genome size of placental mammals.
Their close relatives, however, may be endowed with a genome of considerably
larger size. For example, both the surf smelt and the coho salmon belong to the same
suborder; Salmonoidea. Yet, the former has the minimum-sized genome, while the
latter has a genome which is almost as large as that of placental mammals. Such
diversity can be found even within the same genus. While Barbus tetrazona is the
possessor of a minimum-sized genome, Barbus barbus has a genome which is 50% of
that of placental mammals as shown in Table 4.
Possession of the minimum-sized genome does not indicate so-called "primitiv-
eness". None of the more ancient holocephalian, chondrostean or holostean fish are
endowed with such a small genome. Hagfish and lampreys, which represent the most
ancient jawless state ofvertebrateevolution, arealso possessors of rather large genom-
es. Of all the species listed in Table 3, the lungfish, which is one of the few survivors
of the crossopterygian lineage, sticks out like a sore thumb in that its genome is
incomparably enormaus; 35 times greater than that of placental mammals. Y et, the
lungfish has remained essentially unchanged since the middle Devonian period.
The state of apparent chaos which prevails with regard to genome sizes in fish is
merely a reflection of nature's great experiment with gene duplication. As discussed in
Part 4, taudem duplication as well as polyploidy must have occurred time and again in
fish not only during Devonian times (some 300 million years ago), but also in more
recent times.
Table 3. Genome size.r of primitive Chordates and diverse jish

Species Genome size Order Class Sub-
phylum
Commonand in% inmg or suborder or subdass

Latin name X 10-9
DNA
Tunkate 6 0.21 Enterogona Ascidiacea Tunicata
Ciona intestinalis
Amphioxus 17 0.60 Cephalo-
Amphioxus lanceolatus chordata
Hagfish 80 2.80 Hyperotreti Agnatha Vertebrata
Eptatretus stoutii
Lamprey 40 1.40 Hyperoartii
Lampetra planeri
Ratfish 43 1.50 Chimaeroidei Holocephali
Hydrolagus colliei
Shovelnose sturgeon 50 1.75 Chandrostei Paleopterygii
S caphirhynchus
platorhynchus
Bowfin 35 1.23 Protospondyli Holost~i
Amia calva
Spotted gar 40 1.40 Ginglymodi
Lepidosteus productus
South American 3,540 123.90 Dipnoi Crossopterygii
lungfish
Lepidosiren paradoxa
Pacific herrin~:; 28 1.00 Clupeoidea Teleostei
Clupea pallasai ( Clupeiformes)
California anchovy 43 1.50
Engraufis mordax
Surf smelt 20 0.70 Salmonoidea
Hypomesus pretiosus (Ciupeiformes)
Coho salmon 90 3.15
Oncorhynchus kisutch
Mud-minnow 75 2.70* Esocoidei
Umbralimi
Lizard-fish 35 1.20* S copeilformes
Synodus luciocepa
Tetrazona barb 20 0.70 Cyprinoidea
Barbus tetrazona (Ostariophysi)
Carp 50 1.70
Cyprymts carpio
Channel catfish 30 1.05 Siluroidea
lctarulus punctatus (Ostariophysi)
Smooth Armored 125 4.40*
catfish
Corydoras aneneus
Table 3 (continued)
Eel 40 1.40* Apodes

Anguilla rostrata
Swordtail 23 0.80 Microcyprini
Xyphophrus helleri
Killifish 43 1.50*
Fundulus heteroclitus
Bass 37 1.30* Percomorphi
Epinephelus morio
Green sunfish 31 1.10
Lepomis cyanellus
Goby 40 1.40*
Gobius sadanundio
Hornyhead turbot 18 0.63 Heterosomata
Pleuronichthys verticalis
FantaU sole 23 0.80
Xystreurys liolepis
Sea horse 19* 0.66 S olenichthyes
Hippocampus hudsonius
Puffer 14 0.50* Tetraodontiformes
Spheroides maculatus
Table 4
Spedes Ploidy Diploid Number of Genome Number of gene
chromosome chromosome size lod
number arms in % -:6--=P:--G=-D=--cL""'D=-=-H=-I=-c=o=-=H
Tetrazona barb Diploid 50 90 20 1 3 2

Barbus tetrazona
Fasciata barb 52 86 22 1 3 ?
Barbus fasciatus
Pltz 50 78 28 1 3 2
Rutilus rutilus
Schleie 48 80 30 2 3 2
Tinca tinca
Dbel 50 88 38 2 3 ?
Leuciscus cephalus
Black shark 50 82 40 1 3 1
Labeo chrysophekadion
Barbe Tetraploid 100 144 49 2 +5 ?

Barbus barbus
Carp 104 168 52 2 +5 2
Cyprinus carpio
Goldfish 104 166 53 2 +5 2
Carassius auratus
128 Evolution ofVertebrate Genomes
a) Changes in Genome Size whlch are Due Exclusively to Tandem Duplication

If changes in the genome size have been due exclusively to repeated tandem
duplications, such changes should not have affected the outward appearance of the
diploid chromosome complement. One of the peculiarities of fish is that a surprisingly
large number of remotely related species present very similar karyotypes made of
48 acrocentric chromosomes (N OGUSA, 1960; RoBERTS, 1964; PosT, 1965; HNO and
ATKIN, 1966; CHEN, 1967; T AYLOR, 1967). Ofthose listed in Table 3, the hagfish, the
California anchovy, the swordtail, the green sunfish, the hornyhead turbot and the
fantail sole are the possessors of 48 acrocentrics. One belongs to the dass Agnatha,
and the others to four different orders of Teleostei. Despite the apparently identical
karyotypes, their genome sizes vary from a low of araund 20%, as found in the
swordtail, hornyhead turbot and the fantail sole, to a high of 80% in the hagfish; the
30% value shown by the green sunfish and the 40% value shown by the California
anchovy represent intermediate values. Quite clearly, they comprise a series in which
changes in the genome size have been accomplished exclusively by tandem duplica-
tion [Fig. 25 (Plate V)].
Evidence of exclusive dependence upon tandem duplication can also be found in
closely related species belanging to the same family. Table 4 shows genome sizes of
nine species which belong to the family Cyprinidae. Of the nine, six have very similar
karyotypes made of 48 to 52 chromosomes; the number of chromosome arms range
between 80 and 90. Yet, their genome sizes vary from a low of 20% to a high of 40%
(MuRAMOTO etal., 1968; WoLF etal., 1969). As shown in Table 3, ofthe two species of
catfish (Suborder, Siluroidea), the genome value of 30% is shown by the channel
catfish, while the armored catfish shows a 125% value. This fourfold difference also
appears to be due exclusively to tandem duplication, for all of the catfish that have
been studied have very similar karyotypes; 2n =54 to 56, the chromosome arm
number = 90 to 94 (MuRAMOTO et al., 1968).
b) Meaninglessness of Exclusive Dependence upon Tandem Duplication

In Chapter XV, it was pointed out that tandem duplication has various short-
comings as the means of achieving useful gene duplication. Observing Table 4,
it should be noted that although the genome of Leuciscus cephalus and Labeo chryso-
phekadion is twice the size of the genomes of Barbus tetrazona and Barbus Jasciata, the
number of gene loci which specify LDH (lactate dehydrogenase), non-mitochondrial
NADP-dependent IDH (isocitrate dehydrogenase) and 6-PGD (6-phosphogluconate
dehydrogenase) have not increased in proportion to the increase in their genome size.
While the diploidization ofa tetraploid results in uniform doubling of every gene locus
in the genome, repeated tandem duplication should result in the genome containing
multiple copies of some genes, and only single copies of other genes. This is because
further duplication by unequal crossing-over involves mainly those chromosomal
segments which were already duplicated. Yet, only the doubling of the 6-PGD locus
was noted in Tinca and Leuciscus (BENDER and HNO, 1968; KLSE et al., 1969;
QurRoz-GuTIERREZ and HNO, 1970). This fact appears to confirm one of the short-
comings of tandem duplication which is that functional diversification of tandemly
duplicated copies of a gene is very difficult to achieve (Chapter XV). It is probable
that Tinca and Leucisctts do have many copies of each of a pair of 6-PGD loci, but
these copies merely serve to make more of the same gene product (Chapter X).
The genome containing many duplicated chromosomal segments is inherently
unstable, since duplicated segments invariably invite further unequal cros~ing-over
and unequal exchange. As a result, two types of gametes, those with still higher
degrees of genetic redundancy and those with diminished degrees of genetic redun-
dancy, would occasionally be produced. In view of the meaninglessness of exclusive
dependence upon tandem duplication, one wonders as to which of the reciprocal
products of unequal crossing-over or unequal exchange would be favored by natural
selection.
The trend of more ancient and presumably primitive fish having larger genomes
than those which are modern and highly specialized is evident in Table 3 (HrNE-
GARDNER, 1968). Flatfish of the order Heterosomata, sea horses of Solenichtf!Jes and
puffers of Tetraodontiformes are highly specialized modern fish. They uniformly
possess the minimum-sized genome which is either about the same or only slightly
larger than the genome of the amphioxus. Jawless fish as well as holocephalian,
chondrostean and holostean fish, in sharp contrast, have rather large genomes.
It appears probable that as a result of extensive experiments with gene duplication
which began as early as Devonian times, various lineages of ancient fish acquired
rather large genomes; perhaps between a 40 to 100% value. Sofarasthose which
depended exclusively upon tandem duplication were concerned, subsequent moderni-
zation and progressive specialization have been accompanied by progressive reduction
in the degree of genetic redundancy. This is expected since the possession of tandemly
arranged multiple copies of a cistron per se does not appear to greatly benefit an
organism. The trend for a diminished degree of genetic redundancy with progressive
modernization has also been noted in angiosperm species of plants belanging to the
genus Lat!ryrus. All 18 species of Lat!ryrus have nearly identical diploid chromosome
complements made of 14 metacentric chromosomes, but the genomes of more
ancient species contained twice the DNA ofthat of more modern spedes (REES and
HAZARIKA, 1969).
There is little doubt that the lungfish is an evolutional oddity. It is my view that
the ancestor of this particular lineage also increased its genome size exclusively by
tandem duplication. But, in this case, the increase went past the point of no return.
As their genome size increased beyond the threshold, their cell size had to increase to a
great extent. This situation created an obvious need to produce each gene product in
great quantity. Tandemly duplicated copies of each cistron had to be used to fulfill
this need, and it became impossible to eliminate excessive genetic redundancy.
By establishing such a system, the organism effectively forfeited an opportunity for
further evolution. In a manner of speaking, the genome became frozen, while con-
taining enormous genetic redundancy. It is clear that in so doing such a lineage
reached the evolutional dead end. It will be shown that what happened to the lung-
fish also happened to salamanders and newts of the amphibian order Urodeles. Indeed,
this side branch (Lepospondyls lineage mentioned in Chapter XVIII) stopped dead at
the amphibian stage.

c) Efficacy of Tetraploidy as a Means of Acquiring Functionally Diversified

Duplicated Genes
Of nine species of cyprinid fish listed in Table 4, three ( Barbus barbus, Cyprinus
carpio and Carassius auratus) are almost certainly tetraploids which have evolved from
diploid species with the least degree of genetic redundancy. Observing Table 4,
however, it should be noted that evidence for the doubling of every gene locus is not
as clear in these three spedes of cyprinids as it is in tetraploid salmonid fish (BENDER
and HNO, 1968; KLOSE et al., 1969). Our interpretation is that Barbus barbus, the
carp and the goldfish are older tetraploids which completed the process of diploidi-
zation long ago. Thus, some of the duplicated gene loci which failed to offer a
selective advantage may have been subsequently silenced. While a number of quadri-
valents are still found in meiosis of the tetraploid salmonids [Figure 18 (Plate IV)],
only bivalents are seen in these three tetraploid species of cyprinids.
Another instance of diploid-tetraploid relationship was found in fish of the order
Clupeiformes as shown in Table 3. While members of the suborder Clupeoidea, such as
the herring and anchovy, as well as smelts of the family Osmeridae (suborder, Sal-
monoidea) are diploid species, all other members of Salmonoidea are clearly tetraploids.
Tetraplaid members are trout and salmon of the family Salmonidae, whitefish of
Corregonidae and graylings of T/?_ymallidae. The genomes of diploids range from 22%
to 40%, while those of tetraploids range from 60% to 90%; the total chromosome
arms of diploids range from 48 to 60, those of tetraploids range from 100 to 160
(KLSE et al., 1968). The karyotype of the Pacific anchovy shown in Fig. 25 (Plate V)
should be compared tothat of the rainbow trout shown in Fig. 18 (Plate IV) (Chapter
XVI). As shown in Fig. 18 (Plate IV) a number of quadrivalents are still found during
meiosis of these tetraploid species. This indicates that these tetraploid salmonoid fish
are of relatively recent origin and that they have not quite completed the process of
diploidization.
As already mentioned in Chapter XVI, with regard to almost every gene product
which has been studied, these tetraploid species possess twice the number of gene
loci as compared to not only diploid members of the same order, but also to birds
and mammals. Tetraploidy is indeed a very efficient means of achieving meaningful
gene duplication.
Since tetraploid evolution has occurred in modern fish such as cyprinids and
salmonoids, it is clear that tetraploid evolution could also have occurred in ancient
fish of Devonian times. It is our contention that the ancestors of reptiles, birds, and
mammals have experienced at least one tetraploid evolution either at the stage of
fish or at the stage of amphibians. This belief is based on evidence presented in this
and previous chapters which indicates that tandem duplication is meaningless unless
supplemented periodically by simultaneaus duplication of all gene loci by tetraploidy.
d) Passihle Polyphyletic Origin of Terrestrial V ertebrates

Just as lungfish of the order Dipnoi of the crossopterygian lineage stand out from
all other fishin having genomes with enormous genetic redundancy, tailed amphibians
(salamanders, newts and mudpuppies) stand apart from all other terrestrial vertebrates
for the same reason as will be shown in the next chapter. This makes one suspect that
it was not one kind of rhipidistian lobe-finned fish which evolved to amphibians
References 131
through ichthyostegids, but rather that there were two distinct kinds. One gave rise to
the main labyrinthodont lineage from which frogs, reptiles, birds and mammals
were derived, while the other gave rise to a lepospondyls side branch which hit a dead
end after producing tailed amphibians. Furthermore, the following findings on some
of the more primitive bony fish make one wonder if, in fact, there were several
different kinds of rhipidistian fish which evolved to amphibians. The possibility of
polyphyletic origin of terrestrial vertebrates should be considered.
One prominent genome characteristic which is shared by avian species on one hand
and by members of the two orders of reptiles on the other is the possession of
microchromosomes. Dot-like microchromosomes are the size of small bacteria (less
than 1 [L in diameter) and a considerable nurober of these microchromosomes are the
constituents of diploid complements ofbirds, lizards, snakes and turtles. No mammals,
except possibly monotremes, have microchromosomes, and neither do amphibians
nor teleost fish. However, much to our surprise, we found an abundance of micro-
chromosomes in three species of ancient fish which are left-over relics. They were
the ratfish of the subdass Holocephali, a chondrostean sturgeon and a holostean
spotted gar. The ratfish is an evolutional curiosity. It has a cartilaginous skeleton and
fertilizes internally as do sharks and rays of the dass Elasmobranchii, yet it wears gill
covers (opercula) like bony fish (Osteichthyes). Chondrostean and holostean are the
most ancient of the bony fish, and, for this reason, they still preserve certain character-
istics shared by the Devonian crossopterygian fish which came out of the water to
become the first amphibians. Their body is covered with heavy rhombic scales, an air
bladder, like the lung of terrestrial vertebrates, connects with the throat and aids, or
once aided, respiration, and the backhone turns upward at the tail (heterocercal tail).
A diploid metaphase figure of the shovelnose sturgeon shown in Fig. 26 (Plate VI)
can easily pass for that of a reptilian species (especially a turtle). While it cannot be
denied that microchromosomes of reptiles and birds might have evolved de novo within
diapsid and anapsid lineages of reptiles, it seems probable that the particular rhipidi-
stian fish ancestor which eventually gave rise to diapsid and anapsid lineages of
reptiles already had microchromosomes.
References
ATKIN, N. B., HNO, S.: DNA values of four primitive chordates. Chromosoma 23, 10-13
(1967).
101-107 (1968).
BovERr, T.: Zellen-Studien. ber das Verhalten der chromatischen Kernsubstanz bei der
Bildung der Richtungskrper und bei der Befruchtung. Jena Z. Med. Naturw. 24, 314
bis 401 (1890).
CHEN, T.: Comparative karyology of selected deep sea and shallow water teleost fishes.
Ph. D. Dissertation, Yale University 1967.
DEELEY, E. M.: An integrating microdensitometer for biological cells. ]. Sei. Instr. 32,
263-267 (1955).
HrNEGARDNER, R.: Evolution of cellular DNA content in teleost fishes. Am. Naturalist 102,
517-523 (1968).
KLSE, ]., WoLF, U., HrrZEROTH, H., RITTER, H., ATKIN, N. B., HNO, S.: Duplication of
the LDH gene loci by polyploidization in the fish order Clupeiformes. Humangenetik 5,
190-196 (1968).
9*
KLosE, ]., WoLF, U., HITZEROTH, H., RITTER, H., OHNO, S.: Polyploidization in the fish
family Cyprinidae, order Cypriniformes II. Duplication of the gene loci coding for Iactate
dehydrogenase (E.C.: 1.1.1.27) and 6-phosphogluconate dehydrogenase (E.C.: 1.1.1.44)
in various species of Cyprinidae. Humangenetik 7, 245-250 (1969).
MINOUCHI, 0.: Notiz ber die Chromosomen von Tethyun1 plicatum LEs. (Ascidia). Z. Zell-
forsch. 23, 790-794 (1936).
MuRAMOTO, J., 0HNO, S., ATKIN, N. B.: On the diploid state of the fish order Ostariophysi.
Chromosoma 24, 59-66 (1968).
NoGUSA, S.: A comparative study of the chromosomes in fishes with particular considera-
tions on taxonomy and evolution. Mem. Hyogo Univ. Agr. 3, 1-62 (1960).
OHNO, S., ATKIN, N. B.: Comparative DNA values and chromosome complements of eight
species of fishes. Chromosoma 18, 455-466 (1966).
- MuRAMOTO, J., CHRISTIAN, L., ATKIN, N. B.: Diploid-tetraploid relationship among
Old-World members of the fish family Cyprinidae. Chromosoma 23, 1-9 (1967).
- - STENIUS, C., CHRISTIAN, L., KITTRELL, W. A., ATKIN, N. B.: Microchromosomes in
Holocephalian, Chondrostean and Holostean fishes. Chromosoma 26, 35-40 (1969).
QmRoz-GuTrERREZ, A., OHNO, S.: The evidence of gene duplication for S-form NADP-
linked isocitrate dehydrogenase in carp and goldfish. Biochem. Genet., 4, 93-99 (1970).
PosT, A.: V ergleichende Untersuchungen der Chromosomenzahlen bei Swasserteleosteern.
Z. Zool. Syst. Evolut.-forsch 3, 47-93 (1965)
REES, H., HAZARIKA, M. H.: Chromosome evolution in Lathyrus. In: Chromosomes today,
Vol. 2, pp. 158-165 (DARLINGTON, C. D., LEwrs, K. R., Eds.). Edinburgh: Oliver &
Boyd 1969.
RoBERTS, F. L.: A chromosome study of 20 species of Centrarchidae. J; Morphol. 115, 401 to
418 (1964).
TAYLOR, K. M.: The chromosomes of some lower chordates. Chromosoma 21, 181-188
(1967).
WOLF, U., RITTER, H., ATKIN, N. B., 0HNO, S.: Polyploidization in the fish family Cyprinidae,
order Cypriniformes. I. DNA-content and chromosome sets in various species of Cyprini-
dae. Humangenetik 7, 240-244 (1969).
ChapterXX
Evolution from Amphibians to Birds and Mammals and the

Abrupt Cessation of Nature's Experiment at the Reptilian Stage
1. Frogs Versus Salamanders
In Table 5, frogs and toads of the order Anura (descendants of the main laby-
rinthodont lineage of early amphibians) are compared to salamanders and newts of
Urodela (lepospondyls side branch). It immediately becomes clear that their genomes
reflect the fact that the two belong to different lineages.
It appears that what happened in the ancestral lungfish also happened in the
ancestor of urodeles. In both groups, progressive increase in the genome size ex-
clusively by taodem duplication crossed the point of no return, so that the genome
had to freeze while containing enormaus genetic redundancy. Urodeles are characteri-
zed by having a rather low diploid chromosome number; from a low of 22 to a high
of 38. Y et, each individual chromosome is enormaus in size (FANKHAUSER and
HuMPHREY, 1959; DoNNELLY and SPARROW, 1963; KEZER et al., 1965). Ofthe seven
Frogs Versus Salamanders 133
species of urodeles shown in Table 5, the smallest genome is still 7 times the size of the
genome of placental mammals, while the largest is 27 timesthat of mammals (ALLFREY
et al., 1955; JosEPH GALL, personal communication). Necturus maculosus (mudpuppy)
is rather similar to Lepidosiren paradoxa (South American lungfish) not only in the
genome size, but also in the diploid chromosome number. These two species, one an
amphibian and the other a fish, have the same diploid chromosome number, however,
the genome size of the amphibian is 2780%, while that of the fish is 3540% (Tables 3
and 5).
Two species of tritons belanging to the same genus (Triturus) have similar
diploid chromosome complements; 2n = 24 for T. cristatus and 2n = 22 for T.
viridescens. Yet, the genome of the latter is nearly twice the size of the genome of the
Table 5
Species 2n nurober Genome
Urode!a (Caudata) size%
Axolotl mexicanum 26 or 28 939

Plethodon cinereus 26 or 28 705
Triturus cristatus 24 830
Triturus viridescens 22 1300
Amphiuma means 28 2700
Necturus maculosus 38 2780
Species 2nnumber Genome

Anura ( S alientia) size%
Pipapipa 22 40
Xenopus laevis 36 85
Bufo americanus 22 140
Odontophrynus cultripes 22 70
Odontophrynus americanus 44 (4n) 125
Ceratophrys dorsata 104 (Sn) 250
Rana catesbiana 26 180
Rana pipiens 26 220
former (Table 5). There remains little doubt that, both in urodeles and in lungfish,
enormous increase in the genome size has been accomplished exclusively by tandem
duplication. The lepospondyls lineage of amphibians is clearly a side branch of
vertebrate evolution which hit a dead end.
In sharp contrast to urodeles, anurans, as a whole, have genomes of reasonable
size. Although the smallest genome listed in Table 5 is 40% of that of placental
mammals, KLAus RoTHFELS kindly informed me that the spadefoot toad ( Scaphiopus),
belanging to the suborder Anomocoela, has a genome which is only 20% of the genome
of man and other placental mammals.
lt should be recalled that the 20% value represents the minimum-sized genome of
all vertebrates, and that such genomes characterize many diverse species of modern
and specialized teleost fish (Table 3). The great difference in genome sizes which
we see in frogs and toads of today is due apparently to both taodem duplication and
polyploidy. As already mentioned in Chapter XVI, three species of frogs belanging
to the family Ceratophrydidae constitute a polyploid series. Starting with the 70%,
value of a diploid species, Odontophrynus cultripes, the genome size increased to 250%
in an octaploid species, Ceratophrys dorsata (BECAK et al., 1967). The Surinam toad
(Pipa pipa) and the South African water frog ( Xenopus laevis) belong to the same
family ( Pipidae) ; yet, the genome of the latter is about twice the size of the genome of
the former. Considering the difference in chromosome number, it is possible that
there exists a diploid-tetraploid relationship among members of the family Pipidae
(NrELS B. ATKIN, personal communication). A difference in the genome size observed
between two species of Ranidae, on the other band, is due exdusively to tandem
duplication, since Rana catesbiana and Rana pipiens have nearly identical diploid
chromosome complements (Dr BERARDINO, 1962).
Inasmuch as present day amphibians have apparently undergone both repeated
tandem duplication and polyploid evolution in recent times, it would be a reasonable
deduction that Nature's great experiment with gene duplication, beginning in ancient
fish of Devonian times, continued in the main labyrinthodont lineage of amphibians
during Carboniferous and Permian times.
2. Diapsidian Reptiles and Birds

It appears that coinciding with the invention of the amniote egg, the great experi-
ment with gene duplication came to an abrupt end. While the sex chromosomes of
fish and amphibians appear to have remained morphologically indistinct, revealing
that their chromosomal sex determining mechanism stayed at the initial stage of
differentiation, reptiles apparently acquired a well entrenched chromosomal sex
determining mechanism. This is reflected in the fact that a pair of opposing sex
determining chromosomes (the X and the Y in the case of the male heterogamety,
and the Z and the W in the case ofthefemale heterogamety) are no Ionger morphologi-
cally identical to each other in a number of present day reptilian species as shown in
Fig. 27 (Plate VII). The determiner ofthe heterogametic sex (the male Yin theXY/XX
system and the female W in the ZZJZW system) has become specialized, and, as a result,
it became considerably smaller than its original homologue; the X or the Z (BECAK
et al., 1962; KoBEL, 1962; BECAK et a!., 1964; GoRMAN and ATKINS, 1966). As the
well entrenched chromosomal sex determining mechanism was acquired, it became
impossible for reptiles to utilize polyploidization as a means of obtaining new gene
loci. From this stage of vertebrate evolution onward, gene duplication could only be
accomplished by unequal crossing-over, unequal exchange and other means of
tandem duplication. Inasmuch as exdusive dependence upon tandem duplication is
rather meaningless, it appears that Nature's great experiment with gene duplication
abruptly ceased at the reptilian stage.
The evidence presented below indicates that when embolomeres (a branch of
very early labyrinthodont amphibians) began to evolve toward reptiles in the be-
ginning of the Carboniferous period nearly 280 million years ago, only two dasses of
genome sizes were selected, and that each genomedass has remained stable ever since.
Of the living reptiles, snakes and lizards of the order Squamata and crocodiles and
alligators of the order Crococfy!ia are the descendants of the Diapsida line. It should be
recalled that birds of the dass Aves also evolved from the Jurassie reptile of the
Diapsida line.
Diapsidian Reptiles and Birds 135
Observing Table 6, it should be noted that snakes and lizards on one hand and
birds on the other constitute a rather uniform group with regard to their genome
sizes. The genome sizes of snakes and lizards range between 60 and 67% of that of
placental mammals, while the genome sizes of birds range between 44 and 59% of
that of mammals (ATKIN et al., 1965). The similarity between members of the reptillau
order Squamata and the dass Aves does not end in their genome sizes. Microchromo-
Table 6
Diapsidian reptiles
Order Species Zn number Genome
size%
Squamata Chameleon lizard 36 60

Ano!is caro!inensis (Z4 micro)
Alligator lizard 46 or 48 65
Gerrhonotus multicarinatus (Z6 micro)
Boa constrictor 36 60
Boa constrictor amarali (ZO micro)
South American Xenodon 30 67
Xenodon merremii (14 micro)
South American jararaca 36 60
Bothropsjararaca (ZO micro)
Birds
Order Species Zn number Genome
size%
Ga!!iformes Chicken 78 45
Gallus gallus domesticus (60 micro)
Columbiformes Pigeon 80 55
Columba livia domestica (60 micro)
P sittaciformes Parakeet 58 44
Melopsittacus undulatus (36 micro)
Passeriformes Canary 80 59
S erinus canarius (60 micro)
somes mentioned in the previous chapter are present in the diploid complements of
both this group of reptiles and birds.
In Fig. 27 (Plate VII), the female karyotype of the sidewinder rattlesnake (Crotalus
cerastes, 2n = 36) is compared to the female karyotype of the canary (Serinus canarius,
2n= 80). The rattlesnake has 10 pairs of microchromosomes, while the canary has
about 30 such pairs. There is yet another common characteristic. The female hetero-
gamety of the ZZjZW-type which operates in all avian species (YAMASHINA, 1951) also
prevails in poisonous snakes of the families Viperidae, Crotalidae and Elapidae (BECAK
et al., 1962; KoBEL, 1962; BECAK et al., 1964). Furthermore, the absolute size of
Z-chromosomes of poisonous snakes and birds are about the same; the Z comprising
about 10% of the genome. The W-chromosome of both groups has become a speci-
alized, smaller element (OHNO, 1967). There is little doubt that the genome of avian
species has not changed much from the time of their archosaurian reptilian ancestor.
Since the time of the first reptiles, there must have been dichotomy in reptilian
evolution. Some Carboniferous reptiles of the Diapsida line already had a genome
characterized by its size (around 50% ofthat of placental mammals) and by the
presence of microchromosomes. Furthermore, the sex determining mechanism of
this lineage probably operated on the principle of female heterogamety (ZZJZW-
system). This genome lineage has remained stable. Members of the reptilian order
SquatJlata, as well as birds, are living descendants of this lineage. As a result ofNature's
extensive experiment with gene duplication, many different kinds of genomes with
different degrees of genetic redundancy must have been created during fish and am-
phibian stages. Of those, why was this particular genome type selected in ancestral
reptiles?
It is my understanding that this genome type contained a greater potential than
most other types, because an increase in the genome size was accomplished by
tetraploidization rather than by repeated taodem duplication. When a diploid
species with a minimum-sized genome became a tetraploid, the genome size increased
to a near 50% level (see Table 4 of the previous chapter). Thus, what happened to
tetraploid cyprinid fish (carps and goldfish) may have also happened to either the
fish or amphibian ancestor of certain reptiles of the Diapsida lineage. The genomes of
lizards and snakes (60 to 67% values) are slightly larger than the genomes of birds
(44 to 59%) It appears that excessive genetic redundancy contained in the genomes of
snakes and lizards is due to repeated taodem duplication which occurred following
etraploidization.
3. Synapsida Line of Reptiles
Although the Synapsida line which gave rise to mammals left no reptilian survivors,
the study of turtles, alligators and crocodiles reveals that there was indeed dichotomy
in reptilian evolution and some of the reptiles, which surely included members of the
Synapsida line, were characterized from the very beginning by having a genome which
was about the same size as the mammalian genome.
Alligators and crocodiles of the order Crococ!Jlia belong to the Diapsida line which
gave rise to birds and which includes members of the order SquatJlata; yet, their
genomes are quite different from those of snakes, lizards and birds. The diploid
chromosome complements of alligators and crocodiles are very unreptilian and rather
mammal-like in that no microchromosome is present. The diploid chromosome
nurober ranges from a high of 42 in the South American spectacled caiman (Cai1!lan
sclerops) to a low of 32 in the Nile crocodile (Crococ!Jlus niloticus) and two other species
(CoHENand CLARK, 1967). By sheer coincidence, the diploid complement of the caiman
is nearlyidentical tothat of therat ( Rattusnorvegicus) [Fig. 28 (Plate VIII)]. Furthermore,
the genome of the caiman was found tobe nearly as large asthat of placental mammals;
an 82% value was obtained. Quite clearly, reptiles belanging to the Diapsida line were
of two kinds with regard to the genome characteristics. One kind gave rise to snakes
and lizards on one hand and to birds on the other. The other kind gave rise to alli-
gators and crocodiles. It appears probable that dinosaurs of the Jurassie period,
although they all belonged to the Diapsida line, also belonged to two distinct genome
Mammals 137
dasses; one snake and lizard-like or bird-like and the other alligator and crocodile-
like or placental mammal-like.
Turtles of the order Chelonia are descendants of the Anapsida line. They are of
particular interest in that their chromosome complements contain microchromosomes
as do the chromosome complements of lizards, snakes and birds, yet their genome
sizes are similar to that of mammals. The diploid chromosome number of turtles
tends tobe in the range of the 50's and 60's, and 20 or so of the chromosomes can be
considered as microchromosomes (BECAK et al., 1964). The mitotic metaphase figure
of the Desert tortoise (Gopherus agassizi, 2n =52) shown in Fig. 28 (Plate VIII) is also
strikingly similar in character tothat ofthe shovelnose sturgeon shown inFig. 26 (Plate
VI). The genome of this species, however, was rather large being 89% of that of
mammals. Another species of turtle belanging to quite a different family bad a genome
which was 80% ofthat of mammals (ATKIN et al., 1965).
On the basis of the above findings on certain descendants of the Diapsida line as
weil as on the descendants of the Anapsida line, we deduce that the genome size of the
Synapsida line from theriodonts of the late Carboniferous period of the Paleozoic era
has remained about the same as that of alligators, crocodiles and turtles.
Why was this particular genome size selected by certain early reptiles and why
has it remained stable ever since? It is my view that this genome dass also acquired a
great potential for future evolution, because an increase was accomplished by
tetraploidization. When a diploid species with the minimum-sized genome became
tetraploid either at the stage of fish or amphibian, it furnished the genome dass
which eventually led to the creation of birds. When a diploid species which bad
already increased its genome size to the 40 to 50% value by repeated tandem duplica-
tion became tetraploid, it furnished the genome dass which eventually led to the
creation of mammals. What happened in a fish or amphibian ancestor of the Synapsida
line of reptiles appears to have happened in tetraploid salmonoid fish in more recent
tim es.
At any rate, it is dear that only two of the diverse genome dasses created by
Nature's great experiments survived to the reptilian stage, and each of the two genome
dasses have remained essentially stable. While birds appear to have inherited the
ZZ/ZW system of sex determining mechanism from their reptilian predecessor, the
XY/XX system operates in mammals. It appears likely that there was also dichotomy
of evolution in reptiles with regard to the sex determining mechanism.
4. Mammals
Initial attempts to determine genome sizes of divergent placental mammals by
measuring DNA contents produced somewhat confusing results. Indications were
that a reproducible absolute DNA value for each species was hard to obtain. The
haploid DNA content per sperm head of cattle was found tobe 2.80 mg x 10-9 by
MIRSKY and R.J:s (1949), while LEUCHTENEERGER et a/. (1951) reported it tobe 3.49 rng
X 10-9
MANDEL and bis colleagues (1950) were among the first to show that each diploid
nudeus of man as well as cattle, sheep, pigs and dogs contains about 7.0 mg x 10-9
(haploid or genome value 3.5 mg x 10-9) ofDNA; no more, no less. More recently,
ATKIN et al. (1965) re-examined this matter of the constancy of genome size among
placental mammals by means of microspectrophotometry. Six species with diverse

chromosome numbers representing four different orders were chosen: man (Homo
sapiens, 2n = 46) of the order Primates, the dog ( Canis familiaris, 2n = 78) representing
the order Carnivora [Fig. 28 (Plate VIII)], the horse (Equtts caballus, 2n = 64) of the order
Perissodacryla, and the mouse (Mtts muscttlus, 2n = 40), the golden hamster (Mesocricetus
aurattts, 2n = 44) and the creepingvole (Microtusoregoni, 2n = 17 /18) [Fig. 28 (Plate VIII)]
representing the order Rodentia. There was no significant difference in DNA values be-
tween man, the horse, the dog, the golden hamster and the mouse. A single exception
was the creeping vole which had a DNA value 10% lower. This species (HNO et al.,
1963) shares with another member of the same subfamily, Ellobius ltttescens, 2n = 17
(MATTHEY, 1953), the distinction of having the lowest diploid chromosome number
known among placental mammals. Such a drastic reduction in the number of chromo-
somes appears to have been accompanied by substantial loss of heterochromatic
segments which represent non-transcribable DNA base sequences. Such loss of
dispensable heterochromatin would account for the 10% lower DNA value found in
the creeping vole. A small difference ( 10% or so) in genome sizes observable
among placental mammals can indeed be attributed to amounts of dispensable
heterochromatin. ScHMID and LEPPERT (1968) have shown that of 13 species of
rodents belonging to the subfamily Microtinae, those with genomes slightly larger
than that of man contained more heterochromatic segments in their chromosomes
than those with slightly smaller genomes.
The uniformity of genome size extends to members of two other subclasses of
mammals; Metatheria and Prototheria. BrcK and ]ACKSON (1967) studied the DNA
content of two monotremes; the duck-billed platypus (Ornithori?Jnchus anatinus) with
the diploid chromosome number of 70 and the spiny anteater ( Taci?Jglosstts aculeatus)
with 63 chromosomes (MATTHEY, 1949; VAN BRINK, 1959) as well as one species of
marsupial, the potoroo (Potorous tridacrylus, 2n = 13/12) (SHARMAN, 1952). Their DNA
values were 98.6%, 93.3% and 81.0% ofthat of the rat ( Rattus Jtorvegicus) which
represented placental mammals. Quite clearly, the genome size of the Synapsida line
has remained essentially unchanged in three branches of mammals; Prototheria
(monotremes), Metatheria (marsupials) and Eutheria (placentals).
References
ALLFREY, V. G., MrRSKY, A. E., STERN, H.: The chemistry of the cell nucleus. Advances in
Enzymol. 16, 411-500 (1955).
ATKIN, N. B., MATTINSON, G., BECAK, W., HNO, S.: The comparative DNA content of
19 species of placental mammals, reptiles and birds. Chromosoma 17, 1-10 (1965).
BECAK, W., BECAK, M. L., NAZARETH, H. R. S.: Karyotypic studies of two species of South
American snakes (Boa constrictor amara!i and Bothrops jararaca). Cytogenetics 1, 305-313
(1962).
- - - HNO, S.: Close karyological kinship between the reptilian suborder Serpentes
and the dass Aves. Chromesoma 15, 606-617 (1964).
- - LAV ALLE, D., SCHREIBER, G.: Further studies on polyploid amphibians (Cerato-
phrydidae). II. DNA content and nuclear volume. Chromosoma 23, 14-23 (1967).
BrcK, Y. A., ]ACKSON, W. D.: DNA content of monotremes. Nature 215, 192-193 (1967).
CoHEN, M. M., CLARK, H. F.: The somatic chromosomes of five crocodilian species.
Cytogenetics 6, 193-203 (1967).
nr BERARDINO, M. A.: The karyotype of Rana pipiens and investigation of its stability during
embryonie differentiation. Develop. Biol. 5, 101-126 (1962).
Uniformity of the Genome Size and the Number of Duplicated Gene Loci 139
DoNNELLY, G. M., SPARROW, A. H.: Karyotype and revised chromosome number of

Amphiuma. Nature 199, 1207 (1963).
FANKHAUSER, G., HuMPHREY, R. R.: The origin of spontaneaus heteroploids in the progeny
of diploid, triploid, and tetraploid axolotl females.]. Exptl. Zool. 142, 379-422 (1959).
GoRMAN, G. C., ATKINS, L.: Chromosomal heteromorphism in some male lizards of the
genus Anolis. Am. Naturalist 100, 579-583 (1966).
KEZER, J., SETO, T., PoMERAT, C. M.: Cytological evidence agairrst parallel evolution of
Necturus and Proteus. Am. Naturalist 99, 153-158 (1965).
KoBEL, H. R.: Heterochromosomen bei Vipera berus L. ( Viperidae, Serpentes). Experientia 18,
173-174 (1962).
LEUCHTENBERGER, C., VENDRELY, R., VENDRELY, C.: A comparison of the content of des-
oxyribose nucleic acid (DNA) in isolated animal nuclei by cytochemical and chemical
methods. Proc. Natl. Acad. Sei. US 37, 33-38 (1951).
MANDEL, P., M:ETAIS, P., CuNY, S.: Les quantites d'acide desoxypentose-nucleique par
leucocyte chez diverse especes de Mammiferes. Compt. rend. 231, 1172-1174 (1950).
MATTHEY, R.: Les chromosomes des vertebres. Lausanne (Switzerland) 1949.
- La formule chromosomique et le problerne de la determination sexuelle chez El!obius
lutescens THOMAS. Rodentia-Muridae-Microtinae. Arch. Julius Klaus-Stift. Vererbungs-
forsch. Sozialanthropol. u. Rassenhyg. 28, 65-73 (1953).
MrRSKY, A. E., Rrs, H.: Variable and constant components of chromosomes. Nature 163,
666-667 (1949).
HNO, S.: Sexchromosomes and sex-linked genes (LABHART, A., MANN, T., SAMUELS, L. T.,
Eds.), Vol. I. Monographs on Endocrinology. Berlin-Heidelberg-New York: Springer
1967.
- J AINCHILL, J ., STENIUS, C. : The creeping vole ( Microttts oregoni) as a gonosomic mosaic
I. The OY/XY constitution of the male. Cytogenetics 2, 232-239 (1963).
ScHMID, W., LEPPERT, M. F.: Karyotyp Heterochromatin und DNS-Werte bei 13 Arten von
Whlmusen ( Microtinae, Mammalia, Rodentia). Arch. Julius Klaus-Stift. Vererbungs-
forsch. Sozialanthropol. u. Rassenhyg. 43, 88-91 (1968).
VAN BRINK, J. M.: L'expression morphologique de la Digametie chez les Saumpsides et
les monotremes. Chromosoma 10, 1-72 (1959).
Y AMASHINA, Y.: Studies on the chromosomes in 25 species of birds. Genedes 2, 27-38
(1951).
ChapterXXI
Whence Comes Man?

1. Uniformity of the Genome Size and the Number of Duplicated Gene Loci
The uniforrnity of genome size within mammals implies that all mammals are
essentially endowed with the same genome in that most of the gene loci contained
in the genome of one species are included in that of the other species. For instance,
all placental mammals studied to this date contain three separate gene loci for lactate
dehydrogenase (LDH) subunits in their genomes. A and B subunits are produced in a
variety of tissues (MARKERT, 1964), while C subunits are produced only in the sexually
mature testis (GoLDBERG, 1962). The above situation found in mammals is in sharp
cantrast to the situation found in teleost fish. Some teleosts with rninimum-sized
genomes, have only a single gene locus for LDH (MARKERT and FAULHAUBER, 1965),
while tetraploid teleosts such as trout and salmon have as many as 8 separate gene
loci for LDH subunits (MAssARO and MARKERT, 1968). Sirnilarly, all placental
mammals appear to be characterized by having one gene locus each for the super-
nataut form and the mitochondrial form ofNAD-dependent malate dehydrogenase as
well as NADP-dependent isocitrate dehydrogenase. In the case of teleost fish, on the
other band, tetraploid species have two separate gene loci for the supernatant form
of each enzyme (BArLEY et al., 1969; QuiRoz-GuTIERREZ and HNO, 1970; WoLF
et al., 1970).
The very fact that amino acid sequence of human hemoglobin cx-chain more
closely resembles that of the mause cx-chain (17 differences) and that of the horse
cx-chain (18 differences) than the amino acid sequence ofhls own hemoglobin -chain
(84 differences) reveals that a duplication which produced the cx- and -chain genes
from a common ancestral hemoglobin gene occurred long before the emergence of
protoinsectivores (more than 70 million years ago). Indeed, it appears that Nature's
great experiments with gene duplication were clone when our ancestors were still fish
or amphlbians.
It is of interest to note that divergent mammalian species are about the same not
only with regard to the number of duplicated gene loci whlch are unlinked, but also
with regard to the number of tandemly duplicated gene loci. Both in man (NATVIG
et al., 1967) and the mause (HERZENBERG et al., 1967), a duster of nearly 10 closely
linked gene loci specify heavy-chains for different classes of immunoglobulin; IgG,
IgA, IgM, etc. It is generally believed that a pair of unlinked gene loci specify
two types of immunoglobulin light-chains; kappa and lambda. When the human
x-chain is compared to the human A.-chain, there is roughly 40% homology, but the
degree of homology increases to 60% when human kappa is compared to mause
kappa (LENNOX and CoHN, 1967). Apparently, a duplication whlch produced a pair
of immunoglobulin light-chain genes also occured at a stage prior to the emergence
of protoinsectivores.
Exclusive dependence upon taudem duplication as a means of acquiring new gene
loci is not very meaningful as repeatedly mentioned; however, when further increase
of the genome size by polyploidization became impossible (coinciding with the
invention of the amniote egg), subsequent creation of new gene loci bad to depend
exclusively on mechanisms of taudem duplication. It appears that as soon as polyploid
evolution became impossible, natural selection began to select against an excessive
degree of taudem duplication. Thls would explain why the genome size stabilized
at the reptilian stage and why divergent species of placental mammals have maintained
about the same number of closely linked duplicated genes. If there were no restriction,
it is expected that some placental mammals would contain vastly greater numbers of
tandemly duplicated genes in their genomes than others.
The genome of birds is only half as large as that of mammals; yet, for most
enzymes and non-enzymatic proteins, birds appear to have about the same number
of duplicated gene loci as mammals. For example, birds are also equipped with three
separate gene loci for LDH A, B and C subunits, and they too have a supernatant
form and a mitochondrial form of NAD-dependent malate dehydrogenase as well as
NADP-dependent isocitrate dehydrogenase. Thls is probably due to the fact that the
ancestor of birds also underwent tetraploid evolution as did the ancestor of mammals.
The twofold difference in their genome size appears to be due strictly to genetic
redundancy accumulated by taudem duplication.
Was Diversification of Placental Mammals Due Strictly to Allelic Mutations 141
It has been shown that LDH B and C subunits of birds are specified by a pair of
closely linked gene loci (ZrnKHAM and IsENSEE, 1969). Inasmuch as mammals are
also endowed with a pair of gene loci for LDH Band C subunits, it appears that the
tetraploid genome of the avian line and the tetraploid genome of the mammalian
line have been subjected to the same restrictions with regard to the accumulation of
tandemly duplicated gene loci.
2. Was Diversification of Placental Mammals Due Strictly to

Allelic Mutations of the Already Existing Gene Loci?
Even if the creation of new gene loci by tandem duplication has been strictly
forbidden during the past 70 million years, placental mammals need not have depended
exclusively upon allelic mutations of the already existing gene loci for their speciation.
One characteristic of evolution in many animal groups is the tendency to lose parts
of the body form and to show a specialization of the retained parts during adaptive
radiation (RENSCH, 1959). At the level of structural genes, either silencing by the
accumulation of forbidden mutations or deletion of unneeded gene loci may have
contributed to mammalian evolution.
Inasmuch as the mammalian genome started with enough sets of duplicated
genes, silencing of one or more of the duplicated genes in a set may have often accom-
panied the process of speciation. While man utilizes a special y-chain instead of the
-chain for fetal hemoglobin, many other placental mammals use the -chain for both
fetal and adult hemoglobin. On the surface, it may appear that the gene locus for
human hemoglo bin y-chain has arisen by a de novo tandem duplication in his immediate
ancestor. When the amino acid sequences of 1X, and y-chains are compared, however,
it becomes clear that the y-chain gene is a very ancient gene. Human and human y
differ by 39 amino acid substitutions, while known differences between human
-chain and -chains of non-primate placental mammals merely range from a minimum
of 5 substitutions as compared to camels to a maximum of 26 substitutions as com-
pared to cattle (EcK and DAYHOFF, 1966). It must be that the common ancestor of
placental mammals (protoinsectivores of more than 70 million years ago) had a pair
of gene loci for hemoglobin - and y-chains. During subsequent adaptive radiation
the y-chain locus has been rendered silent in most branches of placental mammals.
Chymotrypsinogen of the pancreas is specified by a single locus in many placental
mammals, but pigs, sheep and other members of the order Artiodac(Yla are exceptions
in that they are equipped with two separate gene loci for chymotrypsinogen A and B
(NEuRATH et al., 1967). It may be that the ancestral protoinsectivore originally had a
pair of chymotrypsinogen genes, but both have been retained in the functional state
only in the Artiodac(Yla line. It should be recalled that tetraploid fish, such as trout
and salmon, are also endowed with a pair of chymotrypsinogen genes (Chapter XVI).
Although natural selection has been imposing strict limitation to the creation of
new gene loci by tandem duplication, a small number of new gene loci have apparently
been created by de novo tandem duplication in placental mammals. Like man, fetal
hemoglobin is different from adult hemoglobin in cattle. In place of the -chain,
fetal cattle hemoglobin uses another -like chain. This fetal -like chain of cattle,
however, is not homologaus to the human y-chain, since the fetal cattle -like
chain is more similar to its own -chain (23 differences) than to human y-chain
(40 differences) (EcK and DAYHOFF, 1966). It appears that a tandem duplication
which produced a fetal -like chain gene from the original -chain gene occurred in
the ancestor of cattle in relatively recent times.
A tandem duplication involving the hemoglobin -chain gene locus gave rise to a
c5-chain locus of man. Human beta and human delta differ from each other by only
10 substitutions, indicating that this particular duplication occurred in man's imme-
diate past. Indeed, only hominoid apes appear to have the c5-chain gene (HrLL et al.,
1963).
It appears that the ancestral mammals began their evolution containing many
sets of duplicated genes which had not achieved high degrees of functional diversifica-
tion. Consequently, placental mammals have been endowed with continuous opport-
unities to specialize each of the duplicated gene loci in a set. In my opinion, progressive
specialization of the already existing duplicated genes contributed greatly to the
tremendous adaptive radiation enjoyed by placental mammals within the framework
of the stable genome.
As early as 1852, ScHROFF found that some rabbits do not respond to the admini-
stration of atropine (an alkaloid which dilates the pupil of the eye). It was sub-
sequently shown that this resistance is dependent upon the presence or absence of an
enzyme, atropinesterase, and that whether a rabbit produces this enzyme or not is
genetically determined (SAwrN and GucK, 1943). Placental mammals have a number
of gene loci which specify enzymes with esterase activities. Since there is a redundancy
of gene loci, a species can afford to change one of the gene loci to a specific esterase gene
when the need arises. In the rabbit, CLYDE STORMONT (personal communication)
has shown that only one of the five alleles of a particular esterase gene locus has ac-
quired the ability to specify an atropinesterase. In a similar manner, any of the several
esterase gene loci of placental mammals can change to become a specialized gene locus
for an esterase of high substrate specificity; atropinesterase, cocaine-esterase, etc.
3. An Evolutional Mechanism which Anticipated Future Needs

Was the creation of man from the hominoid ape an unprecedented advance in
evolution? This has to be the final question asked when considering the evolution of
vertebrates. Man's attainment of manipulating hands and of a bipedal postute which
is aided by powerful neck musdes to hold the head in an upright position does not
appear to be a specially novel achievement, for equally remarkable modifications of
body forms occurred time and again in evolution. For example, starting from the
ordinary foot with five digits, the horse acquired the ability to run swiftly by using
his hoofed middletoes. Equally remarkable is the elongated nose of elephants which
became an efficient instrument of manipulation.
The uniqueness of man unquestionably lies in his enormaus intelligence. Thus,
intelligence should first be defined. Natural selection often provides the genome of an
organism with the behavioral answers to cope with events experienced in the past.
Thus, in some birds and even in some fish, a certain sequence of events elicits very
complex mating and nest building behavior patterns. Such behavior patterns are
inherited, because, based on lessons learned in the evolutional past, the genome came
to contain a set of genes for a fixed behavior pattern. Although man also reacts quite
often to a certain sequence of events with a predictable, and, therefore, fixed be-
havior pattern, such instinctive behavior is not a reflection of his intelligence.
An Evolutional Mechanism Which Anticipated Future Needs 143
The process of learning comes into play when an organism encounters an event to
which its genome is not provided with a prepared behavioral answer. Clearly, all
vertebrates are capable of learning by experience. If given an electric shock every time
it sees a flashing red light, even a goldfish quickly learns to swim away the moment
it sees the red light. The amount of learning to be accumulated by an individual
should be proportional to its longevity. Man, in his 60 or more years, obviously
learns a great deal more than a mouse, which has a life expectancy of a few years.
Yet, if longevity is such an important factor, elephants should be exceptionally
intelligent.
Although all vertebrates have the capacity to learn, most are incapable of trans-
mitting what they have learned to their sibs and progeny. This is the problern of
communication. Expressive face, manipulating hands and the capacity to vocalize
using a variety of sounds are obvious prerequisites for developing a communication
system which is good enough for the exchange of acquired knowledge. At the receiv-
ing end of communication, the young should be endowed with infinite curiosity and an
unsatiable desire to irnitate the adult. The long nursing period of several years and
the formation of a family unit based on the permanent bond between a male and a fe-
male greatly improves the amount of learning which the young can receive from the
adult.
Observing the behavior of living great apes, it appears that the germ of these
prerequisites for advanced communication was already contained in our remote
ancestor. If man acquired his intelligence mainly as a result of an increased ability to
communicate with each other, the creation of man from the hominoid great ape was
no great leap in evolution, for such could have been accomplished simply by progres-
sive modification of anatornical features including an increase in the number of neurons
occupying brain hernispheres. The trend for an increased brain size was already
started at the earllest stage of mammalian evolution.
Man's intelligence enables him to discrirninate between the cause and the effect,
to associate apparently independent phenomena and find the underlying cause, and
to form the concept by assirnilating the observations.
It appears that development of rudimentary cultures as hunters and foodgatherers
began before the emergence of our own species. Ho111o neanderthalensis and even
Pithecanthropus ( HoiJJo erectus) had a culture of sorts, but many more thousand years
were to elapse before the practice of agriculture was started which subsequently
led to the establishment of permanent communities (10,000 years ago). Civilization,
as we understand it, began to develop in isolated parts of the world only 5,000 years
or so ago. Only then did the invention of hieroglyphics, phonetic letters, and numer-
ical signs begin to tax man's rnind to its full potential, and it was a mere few thousand
years ago when the spread of civilization from these isolated centers to surrounding
communities of barbadans occurred in any significant degree.
In a mere several thousand years, even the most severe natural selection could not
have improved man's native intelligence significantly. Furthermore, by the time man
began to live in a civilized way, human populations became too large and unyielding
for natural selection to act upon. There is but little evidence that selective breeding
for higher intelligence was ever practiced after the rise of civilization, or that the more
intellectually gifted left a greater number of progeny than their less gifted con-
temporaries. In short, one has to conclude that the degree of native intelligence or
inherited capacity for learning as a population or species characteristic of man has

not changed significantly since the rise of civilization. In history, surrounding tribes
ofbarbarians destroyed a civilized centertime and again, and, in so doing, they learned
the ways of the vanquished. It would appear that the uncivilized possessed no less the
mental faculty than the civilized.
Did the genome of our cave-dwelling predecessor contain a set or sets of genes
which enable modem man to compose music of infinite complexity and write novels
with profound meaning? One is compelled to give an affirmative answer. Further-
more, it appears that intelligence as such is not so much a species characteristic of
Homo sapiens, but rather a genus characteristic of Homo. Homo neanderthalensis
apparently had enough compassion to bury their dead and decorate their graves with
flowers. Herein lies the most fascinating aspect of the evolutional process which
created Homo.
It looks as though the early Homo was already provided with the intellectual
potential which was in great excess of what was needed to cope with the environment
of that time. What was the mechanism which originally endowed the genome of
Homo with a system which was not immediately needed, but in anticipation of future
needs? Fortunately, vertebrates have a known system which can anticipate future
needs. This is the immune system which responds specifically to an enormous array of
antigens, including artificial ones created in test tubes. Specific antibody molecules are
produced in response to an antigen which is unlikely to have been encountered as
selective influence in the organism's evolutional history (CoHN, 1969).
There is little doubt that such a system owes its existence to the genetic redun-
dancy created by gene duplication. Because the rabbit genome contained several
gene loci for esterases of low substtate specificity, random accumulation of tolerable
mutations accidentally resulted in giving a specific character to a certain allele of a
particular locus. In this mannet, an atropinesterase gene was bom. This could have
happened even if the rabbit had never encountered atropinein its history.
Here, we see the crude beginning of a system which can anticipate future needs.
If the genome came to contain thousands of copies of a gene which specified the
rather unspecific antibody molecule directed against the most common of the existing
antigens, by subsequent mutations, these copies in time could have diverged to
specify a variety of antibodies, and the organism would have acquired the ability to
cope with an enormaus array of antigens. Such a system, however, is delinquent in
its function to anticipate future needs, for natural selection effectively conserves the
base sequence of functionally critical sites of a gene only if its product is vital to the
well-being of the organism. A gene which became potentially useful but which is
presently of no use would be ignored by natural selection. While being ignored, its
character would change slowly but surely by random accumulation of mutations.
What would happen if the genome keeps only a small number of genes, each of
which is presently useful, and multiple copies of each are made after fertilization?
During development of an individual, multiple copies disttibuted in somatic cells
would acquire divergent base sequences by accumulation of mutations and by in-
tragenic recombination between each other. Generation after generation, certain
somatic clones of each individual would come to contain genes which are of no use
today, but which may prove to be useful in the future. In this way, the system which
References 145
anticipates future needs is established, although the capacity to anticipate is confined

by number and the base sequences of prototype genes to be contained in the genome.
In the case of genes for immunoglobulin subunits, it is becoming increasingly
indicative that the variable region and the constant region of individual light- as well
as heavy-chains are specified by separate gene loci as mentioned in Chapter XV. It
may well be that the genes for the variable regions which determine antibody specifi-
cities of light- or heavy-chains do undergo multiplication in the manner mentioned
above in somatic cells.
The establishment of an intricate network in the brain may depend upon the cell
surface recognition between individual neurons at their dendritic and axonal ends.
During ontogenic development, why did neuron A establish a connection with
neuron B, while avoiding a connection with neuron C? It appears feasible that this
selectivity was due to the complementary attraction that happened to have existed
between the cell surface protein of neurons A and B. It then follows that individual
neurons are clonally derived in that their progenitor cells during a phase of rapid
proliferation diverged to generate a great variety of amino acid sequences of the cell
surface protein. Thus, with regard to the cell surface protein, neurons became a
collection of divergent clones. Even within a single neuron, the cell surface protein at
its dendritic ends must, by necessity, have acquired the amino acid sequence which
is distinct from that at its axonal end.
The mechanism for generaring a great variety of amino acid sequences of the
neuronal cell surface protein must operate on the same principle as that by which
individual clones of plasma cells generate a variety of antibody molecules (CoHN,
1969). The genomes of not only all vertebrates but also many invertebrates which are
endowed with neurons must contain a varying number of progenitor genes for the
cell surface protein of neurons. The somatic production of multiple copies of each
progenitor gene, however, may be a more recent invention. Depending upon when
in evolution such an innovation came into being, the presence or absence of this
somatic amplification mechanism may separate vertebrates from invertebrates,
mammals from othervertebrates oreven hominoid apes and man from othermammals.
Progenitor genes contained in the genome are expected to be directly responsible
for instinctive (genetically programmed) behavioral response, whereas an organism
may acquire an increased capacity for learning by somatic amplification and modifica-
tion of the progenitor genes.
It may be that the genome of our cave-dwelling ancestors acquired an appropriate
number of progenitor genes, each with the base sequence optimized to generate an
unprecedented variety of amino acid sequences when its multiple copies underwent
random mutations and intragenic recombination in neurons. In this way, the first
Homo sapien was already endowed with a brain of infinite complexity in anticipation
of future needs which were to arise many thousand years hence.
References
BA.ILEY, G. S., CocKs, G. T., WrLSON, A. C.: Gene duplication in fishes: Malate dehydro-
genases of salmon and trout. Biochem. Biophys. Res. Commun. 34, 605-612 (1969).
CoHN, M.: Anticipatory mechanisms of individuals. In: Control processes in multicellular
organisms. WoLSTENHOLME, J. E. W., KNIGHT, A. J., Eds. CIBA Foundation Sym-
posium 1969.
EcK, R. V., DAYHOFF, M. 0.: Atlas of protein sequence and structure. National Biomed.
Res. Foundation, Silver Spring, Maryland, 1966.
GoLDBERG, E.: Lactic and malle dehydrogenases in human spermatozoa. Science 139, 602
to 603 (1962).
HERZENBERG, L. A., MINNA, J. D., HERZENBERG, L. A.: The chromosome region for immuno-
globulin heavy-chains in the mouse. Cold Spring Rarbor Symposia Quant. Bio!. 32, 181
to 186 (1967).
HrLL, R. L., BuEITNER-}ANUSCH, J., BuEITNER-}ANUSCH, V.: Evolution of hemoglobin in
primates. Proc. Natl. Acad. Sei. US 50, 885-893 (1963).
LENNOX, E., CoHN, M.: Immunoglobin genetics. Ann. Rev. Biochem. 36,365-406 (1967).
MARKERT, C. L.: Cellular differentiation-an expression of differential gene function. In:
Congenital malformations, pp. 163-174. New York: The International Medical Congress
1964.
- FAULHAUBER, I.: Lactate dehydrogenase isozyme patterns of fish. J. Exptl. Zoo!. 159, 319
to 332 (1965).
MASSARO, E. ]., MARKERT, C. L.: Isoenzymepatterns of salmonoid fishes: Evidence for
multiple cistrons for lactate dehydrogenase polypeptides. J. Exptl. Zoo!. 168, 223-238
(1968).
NATVIG, J. B., KuNKEL, H. G ., LrTWIN, S. P.: Genetic markers of the heavy-chain sub-
groups of human gamma G globulin. Cold Spring Rarbor Symposia Quant. Bio!. 32,
173-180 (1967).
Qumoz-GuTIERREZ, A., HNO, S.: The evidence of gene duplication for S-form NADP-
linked isocitrate dehydrogenase in carp and goldfish. Biochem. Genet., 4, 93-99 (1970).
RENSCH, B.: Evolution above the spedes level. New York: Columbia Univ. Press 1959.
SAWIN, P. B., GLICK, D.: Inheritance of "atropinesterase", a blood enzyme of the rabbit.
Genedes 28, 88 (1943).
WOLF, U., ENGEL, W., FAUST, J.: Zum Mechanismus der Diploidisierung in der Wirbeltier-
evolution: Koexistenz von tetrasomenund disomen Geniod der Isozitrat-Dehydrogenasen
bei der Regenbogenforelle (Salmo irideus). Humangenetik.
ZrNKHAM, W. H., IsENSEE, H.: Linkage of lactate dehydrogenase B and C Iod in pigeons.
Science 164, 185-187 (1969).
Subject Index
Actinopterygian (ray-finned) fish 113 Chromosomal structure 15-19

Active site of molecules 27, 28, 29, 59, 60, Chromosomes of man (Homo sapiens), Fig. 3
71, 72, 74 (Plate I)
Allopatric model of speciation 56 - - mouse ( mus 11/Usculus), Fig. 7 (Plate II)
Allotetraploid 100-101 - - tobacco mouse (mus poschiavinus),
Ambiguous coding 25-29 Fig. 7 (Plate II)
Amino acids 8-9 (Table 1) - - tetraploid frog (Odontophrynus
Amniote egg 61, 116-117 americanus), Fig. 17 (Plate III)
Amphioxus 112, 124, 126-127 {Table 3) - - rainbow trout ( Salmo irideus), Fig. 18
Anapsida line of reptiles 118, 131, 137 (Plate IV)
Anchovy ( Engraufis mordax) 128, 130 - - turbot ( Pleuronichthys verticalis),
Anticodons of lransfer RNA 11, 23, 24, 25 Fig. 25 (Plate V)
Anticipation, evolutional mechanism of - - sunfish ( Lepomis cyanellus), Fig. 25
142-145 (Plate V)
Archaeopteryx lithographia (Jurassic bird) - - anchovy ( Engraufis mordax), Fig. 25
39, 118 (Plate V)
Atavism and atavistic mutation 39-40 - - sturgeon (Scaphirhynchus platorhyn-
Australopithecus (man-ape) 4, 55, 123 chus), Fig. 26 (Plate VI)
Autotetraplaid 99-100 - - rattlesnake (Crotalus cerastes), Fig. 27
(Plate VII)
Bacteriophage 21, 108 - - canary (Serinus canarius), Fig. 27
Bar locus in Drosophila 95 (Plate VII)
bobbed mutation in Drosophila 64, 95 - - caiman alligator (Caiman sclerops),
Fig. 28 (Plate VIII)
Caiman alligator (Caiman sclerops) 136 - - tortoise ( Gopherus agassizi), Fig. 28
Camelidae (camels and llamas) 58 (Plate VIII)
Canary ( Serinus canarius) 135 - - dog (Canisfamiliaris), Fig. 28
Cattle (Bos Iaurus) 43, 51, 56, 72-74, 137, (Plate VIII)
141-142 - - vole ( Microtus oregoni), Fig. 28
Centromere 16 (Plate VIII)
Chelonia (turtles and tortoises) 118, 137 Chromosomal changes in evolution 41-42
Chimpanzee (Pan troglocytes) 36-37, 123 Chymotrypsin 28, 72, 73, 74, 85, 104, 141,
Chondrostean fish 117, 125, 129, 131 Figs. 5 and 11
10*
148 Subject Index
Clonal selection 79 Hemoglobin, Figs. 4, 6. 3, 19, 25, 27, 32,

- derivation of somatic cells 145 36, 42, 66, 76-77, 85, 90, 91, 94-96,
Codons, (Table 2) 10, 11, 23 104, 140, 141, 142
Colchicine 30, 31, 75, 76 Heterochromatin 17-18, 138
Color vision 39-40 Heterozygous advantage 36, 65, 95
Complementality between purine and pyri- Himalaya mutation of mammals 30, 38
midine bases 4--7, 21 Histocompatibility genes 51, 108-109
Convergent evolution 37, 38, 74, 76 Histones 18-19, 34, 82-83, 85
Crocodylia (alligators and crocodiles) 118, Holostean fish 113, 125, 129, 131
134, 136 Homo neanderthalensis 57, 123, 143-144
Crossopterygian (lobe-finned) fish 4, 115 Hormones 19, 30, 83, 84
Cyclic AMP (adenosine 3', 5' monophos- Horse ( Equus caba!lus), Fig. 6. 25, 32, 34,
phate) 84 44, 45, 77, 91, 100, 138, 140, 142
Cyprinidae (carp and allies) 128, 130, 136 Hybridization of nucleic acid 14
Cytochrome C 28, 49, 76
Ichthyostega (first amphibian) 115, 131
Deer mouse ( Peromysctts} 42, 44, 46, 51 Ictidosaurus (mammal-like reptiles) 120
Diapsida line of reptiles 118, 134--137 Immuneglobulin Figs. 13, 19. 36-37, 42,
Dinosaurs 118, 136 46, 69, 77-80, 91, 92, 94--96, 140, 145
Diploidization of tetraploids 101-104 Inbreeding, necessity of 32, 55
Disulfide bridge 23, 29, 74 Inducer (of transcription and translation)
Dog (Canis familiaris) 45, 56, 137, 138 83, 84
Donkey ( Equus asinus) 34, 44, 45, 91, 100 Intragenie recombination 50-53, 72, 144,
Dosage compensation for X-linked genes 145
17, 45, 66 Inversion of chromosomes 42, 44--45, 101
- effect of structural genes 66-67 Isozyme 41, 67-70, 71, 76, 77
- - - regulatory genes 104--105
Drosophila (fruit fly) 1, 60, 61, 64, 95 Jawless fish ( Agnatha) 86
- hagfish 76, 77, 86, 125, 128
Electrophoresis, Figs. 8, 10. 24, 50 - lamprey 76, 77, 86, 125
Escherichia co!i 11, 18, 27, 35, 41, 83-85, - ostracoderms 113
105
Esterase 65-66, 142, 144 Labyrintheclont amphibian 116, 131, 132,
Euchromatin 17 134
Eukaryote 13, 15, 16, 19, 82 LDH (Iactate dehydrogenase) 28, 46, 51,
67-68, 71, 103, 139-141
Favored mutation 30-31, 35-37 /ac-operon of E. coli 41, 83, 105
Fibrinepeptide 34 Lamarckian illusion 55
Forbidden mutation 2, 26-31, 34, 48-50, Lemuroid primates 39, 121
62-63, 72, 74, 76, 141 Lepospondyls amphibian 129, 131, 132, 133
Frame-shift mutation 21-22, 27, 39, 80 Lethality of tetraploid zygote 104, 105
Galactosemia 31 Life span of somatic cells 3
Generation time, effect of 57-58 Linkage of genes 41-42, 45-46, 91, 94,
Genome (haploid set) 16, 26, 59, 96, 97, 96, 141
98, 100, 107, 108, 139 - - nucleotides 2'-5' 4, 7
- size of vertebrates, Tables 3, 4, 5, 6 3'-5' 4, 7
124--131 5'-5' 4, 7
Germ line 3-4 Living fossil 53-54
- cells 17, 99 Lungfish (Dipnoa) 53, 96, 97, 113, 125,
Gibbon ( Hylobates lar !ar} 37, 122-123 129, 130, 133
G-6-PD (glucose-6-phosphate dehydro- Lysogeny 92, 108-109
genase) 41, 45, 66
Goldfish (Carassius auratus) 51, 70, 143 man (Homo sapiens) 16, 32, 34, 36-37, 39,
Gorilla (Gorilla gori/la) 32, 123 45, 46, 50, 55, 56, 64, 69, 76, 77, 79, 85,
91, 92, 94, 95, 99, 104, 123, 137, 138,
Hamster ( Mesocricetus auratus) 31, 138 140, 141, 142, 143-145
Haptoglobin 79-80, 92, 94 Marsupials 53, 121, 138
Subject Index 149
Master-slave theory 63, 64, 97 - - genes 59, 72, 74, 96, 124, 129, 136,
Melanin 38 142
Mierechromosomes 131, 135, 136, 137 Redundant replication of DNA 92-94
Microtinae (rodent family) 58 Regulatory gene 67, 82-87, 96, 104--105
Microtubule protein 30, 74--76 - protein 83, 84--85
Minimum-sized genome of vertebrates Revertant mutation 38-40
125, 128, 129, 133, 137 Ribosome 13-14
Minute mutation of Drosophila 64 Ribosoma/RNA 12,13-14,60-61,
Missense mutation 22-23, 27, 48, 50, 53 62-64, 95, 96
Monkey, Old World and New World - protein 60
39-40, 121 Robertsonian fusion or translocation
Monosomy 104, 107 43-45, 56, 102
Monotremes 120, 131, 138
Mouse (Mus musculus) 30, 35, 38, 43, 45, Salamander (Urodela) 18, 62, 96, 97, 100,
46, 51, 56, 79, 91, 140, 143 130
Mutation affecting structural genes 21-24, Salmonoidea (smelt, ttout, salmon, etc.) 102,
27-30 103, 104, 125, 130
- - transfer RNA 24--26, 26-27 samesense mutation 23-24, 49
Myoglobin, Fig. 4. 27, 28, 76-77 Self-replication of nucleie acid - DNA
4--6
Neoplastic cells 3, 79, 105 - in "prebiotic condition" 6-7
Neutralmutation 32, 34--35, 54, 57 Sex chromosomes and sex determining
Nonsense mutation 22, 27, 39, 80 mechanism 16, 98-99, 134, 135, 137
Nucleolar organizer of chromosomes 14, Siekle-eeil anemia 36
17, 18, 60-61, 64, 65, 96 Somatic amplification of gene dosage
60-61, 144--145
Octaploid frog 100, 134 Spontaneaus mutation rate 48-50
Odontophrynus americonus (South American Squamata (lizards and snakes) 118, 134,
frog) 100 135, 136
Operator base sequence 83 Sterility barrier 44-45, 56
Sturgeon ( Scaphirhynchus platorhynchus) 131
Paedomorphosis 112, 124 Sunfish ( Lepomis cyanellus) 128
Phenylketonuria 30 Supernumerary chromosome 107-108
Pithecanthropus (Homo erectus) 55, 57, 123, Suppressor mutation 24--25
143 Sympatric model of speciation 56
6-PGD (6-phosphogluconate dehydro- Synapsida line of reptiles 118, 119, 120,
genase 41, 51, 52, 66, 70, 128-129 136-137, 138
Pleiottopic effect of genes 30
Polycistronic messenger RNA 11, 25, 41, 83 Teleost fish 113, 125-131, 139-140
Polymotphism, alleHe 35, 37, 50 Tetraploid fish 70, 101-105, 130, 136,
Polyphyletic origin of terrestrial vertebrates 137, 141
130-131 - frog 100, 102
Polyribosome 13 Tetrasomie inheritance 103
Population size, effect of 50, 57 Tolerablemutation 32-40, 48-50, 57, 59,
Prokaryote 13, 15, 18, 41, 83, 84 144
Tortoise (Gopherus agassizi) 137
Quail (Cotttrnix c.japonica) 51, 52 Transduction, viral 109
Triploid bird 99
Rahbit (Oryctolagu.rcuniculus} 30, 38, 79, 91, - lizard 99
142, 144 - salamander 100
Rainbow trout (Salmo irideus} 51, 102, Trisomy 104, 107
104, 130 Tobacco mouse (Mus poschiavinus) 43, 56
Rat ( Rattus norvegims) 35, 46, 51, 85, 136 Transfer RNA 7-12, 13, 21, 24--27, 49,
Ratdesnake (Crotalus cerastes) 135 61-62, 64, 91
Recurrent mutation 37-38 Tunicate 112, 124
Receptor base sequence 84, 85, 86, 87, 96 Trypsin, Figs. 5, 11. 28, 72-74, 85, 104
Redundancy of codons 10, 23 Tyrosinase (C-locus) 30, 38
150 Subject Index
Tyrosine transamiaase 83 Untranscribable (non.ren.re) base sequence of

Turbot ( Pleuronichth)'S verticali.r) 128 DNA 17, 18, 138
Unequal exchange between chromatids Vole ( Microtu.r oregoni) 138
Fig. 15. 64, 89-92, 95, 134
- crossing-over during meiosis, Xenopu.r laevi.r (African water frog) 14,
Figs. 9, 16. 64, 95, 96, 134 60-61, 62, 134
Plates I-VIII
153
Plate I (Fig. 3). 46 chromosomes in the diploid complement of anormal human male. Top row
left: The 1-3 group of the three largest pairs of metacentric autosomes. Top row right: The
4- 5 group of two subterminal autosomal pairs. Second row: The 6-12 group is made of
7 pairs of metacentric autosomes. Third row left: The 13- 15 group of six acrocentric auto-
somes. Although all of them carry the nucleolar organizer on their short arms, in this
photograph, the nucleolar organizer is actually seen only on the 14th pair. Third roll' right:
The 16- 18 group of metacentric and subterminal autosomes. Bottom roll' left: The 19-20
group of four metacentrics. Bottom row middle : The 21- 22 group. Although both pairs of
acrocentrics carry the nucleolar organizer on their short arms, the secondary constriction
is actually seen only on the 21st pair. Bottom row right: The !arge metacentric X and the
small acrocentric Y
154
h AA d( nt Dn R~ at u na e1.
0I U 01
tf
lt ..
,.
u
s...,
H
PJate li (Fig. 7). The male karyotype of theordinary hause mause (Musmumtlus,2n ~ 40) shown
in the top two rows is compared with the male karyotype of the tobacco mause (Alus poschia-
vinus, 2n ~ 26) shown in the third and fourth rows. The X- and V-chromosomes of each
species are placed at the extreme right of the lower row. A male 1st meiotic metaphase
figure at the bottom is from an inrerspecific F 1-hybrid. Seven trivalents are seen, each being
made of one poschia11inus metacentric and two muscu/us acrocentrics (Courtesy of Professor
ALFRED GROPP, Bonn, Germany)
Plate III (Fig. 17). The karyotype of the tetraploid frog species, Odontophrynus americanus
(4n = 44), and a male meiotic figure. Boliom : In the karyotype, 44 chromosomes are
sorted out to four homologues each of the 11 different kinds. Top : In this first meiotic
metaphase figure, 10 quadrivalents and two bivalents (10 and 12 O 'clock) are seen.
An elongated body is a sperm head. These photomicrographs were taken from a
microscopic slide kindly given to us by Professor WrLLY BECAK and Dr. MARIA
LurzA BECAK, Sao Paulo, Brazil
S).L
~
ft )<..
~
S).L
j. ' A A,. lJ1
X ,. ~ "i" U1
-
""
Plate IV (Fig. 18). Two ka-
ryotypes and two male first
meiotic metaphase figures
ofthe rainbow tcout(Salmo
irideus), which is believed
l ,,
to be a tetraploid species in
the process of diploidiza-
tion. Due to extensive Ro-
bertsonian polymorphism,
somatic cells from different
tissues of the same fish
showdifferent chromosome
numbers, and meiotic figu-
res from the same fish con-
' )\r t
tain varying numbers of ~ f; ~ [I ~ 1 l !
multivalents. Topthreerows: '
The karyotype of a Iiver cell.
61 chromosomes. 43 meta- K t J ~ t l I ~
centrics and 18 acrocentrics
make up the total of 104
chromosome arms. F ourth 6 at
tosix throws: Thekaryotype f\ ' "
of a spleen cell. 59 chromo-
somes. 45 metacentrics and
14 acrocentrics make up
the total of 104 chromo-
some arms. First meiotic
metaphase figure at botiom
left : Four quadrivalents
(Z, 3 and 7 O'clock and
left center) are conspicuous.
First meiotic metaphasefigures
at botiom right : 28 bivalents
and only one quadeivalent
(11 O'clock)
n <1 n n
tlt\
" (\ {) t f'\ t")
" - "'
SJ.L
Plate V (Fig. 25). Apparently identical diploid complements made of 48 acrocentric chromosomes possessed by three teleost species belonging to
three different orders. Top two rows: The homyhead turbot representing the order Heterosomata. The genome is 18% ofthat of placental mammals.
ilfiddle two rows: The green sunfish representing Percomorphi. The genome is 31 o/o. Bottom two rows: The California anchovy representing Clupeoidea.
The genome is 43%. See Table 3. The hagfish also has 48 acrocentrics, but the genome size is 80% ofthat of placental mammals ~
-
158
.,;.:/.

..
t
1 '1_
1/}
- --l ,
1------,5::-)J.
Place VI (Fig. 26). A diploid mitotic metaphase figure of the shovelnose sturgeon (Zn= 112).
Of the 112 chromosomes, 48 are truly dot-like microchromosomes
' ..

. . . .. .

,. -.

:,,u ~,
..
Plate VII (Fig. 27). The karyotype of the female sidewinder rattle-snake (Crotalus ctrastes, 2n = 36) representing the Diapsida line of li ving reptiles is
compared to that of the female canary (Serinus canarius, 2n = 80 ) representing birds descended from the Jurassie reptile of the Diapsida line.
Top row: Crotalus cerastes: 10 pairs of microchromosomes. The 4th pair from the left is the heteromorphic ZW-pair. Boliom rows: Serinus canarius:
About 30 pairs of microchromosomes. In this species too, the 4th pair from the left is the heteromorphic ZW-pair. The Z-chromosomes of ,_
both species are about the same absolute size <J1
\Q
160
'
Plate VIII (Fig. 28). Mitotic figures of the spectacled Caiman (Caiman sclerops, 2n = 42) repre-
senting a side branch of the Diapsida line, and the desert tortoise ( Gopherus agassizi, 2n =52)
representing the A napsida line are compared to the mitotic figures of two placental mammals
descended from the Synapsida line. The genome sizes of these four species are about the same.
Top /eft: Caiman sclerops: 2n = 42. Top right: Gopherus agassizi: 2n = 52 (including micro-
chromosomes). Bottom left: Canis fami/iaris: 2n ~ 78 (the female dog). Only the two X-
chromosomes are metacentrics. Bottom right: .Microtus oregoni: 2n = 17 (the female creeping
vole). The female is normally XO

Dr. Susumu Ohno (Auth.) - Evolution by Gene Duplication-Springer-Verlag Berlin Heidelberg (1970)

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Dr. Susumu Ohno (Auth.) - Evolution by Gene Duplication-Springer-Verlag Berlin Heidelberg (1970)

Transféré par

Droits d'auteur :

Formats disponibles

Susumu Ohno

Springer Science+Business Media, LLC 1970

@Springer Science+Business Media New York 1970

ISBN 978-3-642-86661-6 ISBN 978-3-642-86659-3 (eBook)

Library of Congress Catalog Card Number 78-112882.

Title No. 1677

It is said that "necessity is the mother of invention". To be sure, wheels and

May 1970 SusuMu HNO

First of all, my gratitude is to my colleagues, especially to NIELS B. ATKIN,

Chapter I. Perpetuation of the Germ Line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Chapter III. Chromosomes of Eukaryotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Chapter IV. Mutation as a Change in the Base Sequence of a DNA Cistron 21

Chapter V. Forbidden Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Chapter VI. Tolerable Mutations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

Chapter VII. The Conservative Nature of Chromosomal Evolution . . . . . . . . 41

Chapter VIII. The Spontaneaus Mutation Rate . . . . . . . . . . . . . . . . . . . . . . . . . 48

Chapter IX. The Rate of Evolution and the Importance of Isolation . 0 0 0 0 55

Chapter X. Duplication for the Sake of Producing More of the Same 59

Chapter XI. The Attainment of a Permanent Heterozygous Advantage by the

Chapter XIII. The Creation of a New Gene from a Redundant Duplicate of

Chapter XIV. Duplication of Regulatory Genesand Receptors ..... 0..... 82

Chapter XVI. Polyploidy: Duplication of the Entire Genome . . . . . . . . . . . . . 98

Chapter XVII. Other Mechanisms for Achieving Gene Duplication . . . . . . . 107

Chapter XIX. Nature's Great Experiment with Gene Dupllcation during

Chapter XXI. Whence Comes Man? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147

Plates I-VIII . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151

As long as a particular function of an organism is under the control of a single

The Creation of Life Based on the lnherent

Perpetnation of the Germ Line

Replication of Nucleic Acids on the Basis of A-T, G-C

1. The Presence of Self-replicating Nucleic Acid in "Prebiotic" Condition

adenosine-5' -phosphate and for the pair polycytidylic acid-guanosine-5' -phosphate

2. Emergence of Transfer RNA

u Phe Ser Tyr Cys u

c Leu Pro His Arg u

A Ile Thr Asn Ser u

G Val Ala Asp Gly u

Fig. 2. My imagination of the intermediate form between the "prebiotic" self-replicating

3. Division of Labor between DNA and RNA

regionaraund the centromere is characteristically made of heterochromatin (centro-

4. Further Need for Untranscribable Base Sequences

5. Histones as Nonspecific Repressors of Transcripdon

Mutation and the Conservative Nature ofNatural Selection

Mutation as a Change in the Base Sequence of a DNA Cistron

1. Mutations Affecting Structural Cistrons

result in a marked slow-down of the rate of synthesis for a mutant polypeptide

2. Mutations Affecting Transfer RNA Cistrans

b) Mutations which Result in Ambiguous Coding

history ofliving organisms. If a mutational change in the anticodon of a transfer RNA

1. Forbidden Mutations Affecting Transfer RNA Cistrons

2. Forbidden Mutations of Structural Cistrons

Fig. 4. A three dimensional configuration of the myoglobin and hemoglobin polypeptide

These 12 amino acids are: - Val-Ile-Ser-Gly-Gly-Cys-Asn-Leu-Asp-Thr-Ala-

Fig. 4 applies to either the myoglobin or hemoglobin polypeptide chains of any

3. Forbidden Mutations Favored

natural selection throughout the evolutional history oE eukaryotes. Y et, a remarkable

Horse.cx. V al-Leu-Ser~Ala-Asp-Lys-Thr-Asn-Val-Lys-Ala-Ala-Tyr Ser

Horse cx. Lys-Val-Gly-ltHis-Ala-Gly-Glu-Tyr-Gly-Ala-Glu-Ala-Leu-Glu

Horse cx. Arg-Met-Phe-Leu~Phe-Pro-Thr-Thr-Lys-Thr-Tyr-Phe-Pro-His-

Horse cx. Phe-Asp-Leu-Ser-His-Gly-Ser-Ala-Gln-Val-Lys@His-Gly-Lys-

Horse cx. Lys-Val-Ala-AspLeu-Thr~Ala-Val~HisHAsp-Asp-

Horse cx. ~Pro~Ala-Leu-SerQLeu-Ser-Asp-Leu-His-Ala-His-Lys