• 100 1,327 8
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview



DOUGLAS J. FUTUYMA State University of New York at Stony Brook Chapter 19, "Evolution of Genes and Genomes" by Scott V. Edwnrds, Hnrvnrd University

Chapter 20, "Evolution and Development" by foltn R. Trlle, S/n/e UIl;versity of New York at Stony Brook

SINAUER ASSOCIATES, INC. • Publishers Sunderland, Massachusetts U.S.A.

Front cover

Back cover

The luxurious plumes of a male bird of pBradise (PnmdisnE'R rngginnn) ,'Ire the result of sexual selection (see Chapters 11 and 17). Photograph © Art \'Volfe/ Art Wolfe, Inc.

Modern birds are almost certainly descended from feathered dinosaurs, such as the famolls fossil

Archaeopteryx litltogmphict1. Long flight feathers were borne by clawed hands and a long tail, characteristic of theropod dinosaurs but not of modern birds (see Chapter 4). Photograph © Tom Stack! Painet, Inc.

Evolution Copyright © 2003 by Sinauer Associates, Inc. AU rights reserved. This book may not be reprinted in ''''hole or in part ,,,,,ithout permission from the publisher: Sinauer Associates, Inc., 23 Plumtree Road, Sunderland, MA 01375 U.s.A. FAX: 413-549-1118 W\'\"N.si nauer.com Sources of the scientists' photographs appearing ill Chapter 1 are gr TTT}

of polypeptide synthesis and the genetic code. An RNA ribozyme (green), an ancestor of mRNA, binds to a cofactor consisting of an amino acid (AA) and a short oligonucleotide, which have been joined by another ribozyme (R1 ) that joins specific oligonucleotides to amiIlO acids accordiIlg to a primitive code. This system evolves to one in \·vhich ribozyme R2, ancestor of the modern ribosome, links amino acids together. (After Mavnard Smith and SZ 200 Ill)

Figure 5.17

The distribution of land masses at several points ill geological time. (A) In the earliest Triassic, most land was aggregated into a single mass (Pangaea). (B) Eurasia and North America were fairly separate by the late Jurassic. (C) Cond\vanaland had become fragmented into most of the major sOllthern land masses by the late Cretaceous. (0) By the late Oligocene, the land masses were close to their present configurations. The outlines of the moderndav continents are visible in all of the m~ps; other black lines delineate important tectonic plate boundaries. (Maps © 2004 by C. R. Scotese/ PALEOMAP Project.)

(B) Late Jurassic

so called because of the evolution of crabs and bony fishes with the ability to crush mollusc shells and of molluscs with the kinds of protective mechanisms, such as thick shells and spines, that characterize many molluscs today (Vermeij 1987). The teleosts, today's dominant group of bony fishes, evolved and began to diversify. DLlIing the JUl'assic and Cretaceous, modern groups of gastropods (snails and relatives), bivalves, and bryozoans rose to dominance, gigantic sessile bivalves (rudists) fonned reefs, and the seas harbored several (A) groups of marine rep tiles. The end of the Cretaceous is marked by what is surely the best-known -mass extinction (often called the Kff extinction, using the abbreviations for Cretaceous and Tertiary). Ammonoids, rudists, marine reptiles, and nl,llly fa.m1l1es of invertebrates and planktonic protists became entirely extinct. The last of the 11011avian dinosaurs becarne extinct at this tin1e. Most paleontologists believe thls extinction was caused by the impact of an asteroid or SOlne other extraterrestrial body (see Chapter 7).



(C) late Cretaceous

(D) Lale Oligocene

Figure 5.18 Seed plants. (A) A living cycad, a member of a once diverse group.

(8) A leaf of the living ginkgo (Gillkgo Terrestrial plants and arthropods

For Inost of the Mesozoic, the flora was dominated by "gym.nosperms" (i.e., seed plants that lack flowers). The major groups were the cycads (Figure 5.18A) and the con.ifers and their relatives, including Ginkgo, a Triassic genus that has left one species that: survives as a "living fossil" (Figme 5.18B). The nllgiosperl1ls, orflower-


bi/aba) next to (C) a fossilized Ginkgo leaf from the Paleocene. (D) ProtomillJosairlen, a Paleocene/Eocene fossil member of the legume family that includes mimosas and acacias. (A © John Cancalosi/Peter Arnold, Inc./ Alamy Images; B photo by David Mcintyre; C © The Natural History ML1seum, London; D courtesy of W. L. Crepet.) (D)


CHAPTER 5 Mammals

Figure 5.19 Phylogenetic relationships and temporal duration

_ _••_T.he.....p.'id.'. ~¥1~

(thick bars) of major groups of amniote vertebrates. Some authors define "reptiles" as one of the h.. .'o major lineages of anmiotes, the

other being the synapsids, V\:hich includes 111i'lmmals. (After Lee et aJ.2004.)

Mesosaurs ~;;:::~

~r --. ,';-L if :;






,:? ~-_

.... -






,,• ,

.. .' '





•'" ••'"






= "0 3


~". ,





:, 2




.... • '"

Euryopsid, (marine rePti~

0: 0





Squamates (lizards and snakes)


Rhynchocephalians (Tuatara)

.g ,o




c 359




I_----'K'--'_ _

Tr 200





,QJ Present


Time (Mya)

il1g plants (Figure 5.18C),jirst appeared i/1 the early Cretaceolls. Many of the anatomical features of angiosperms, including flowerlike struchIres, had evolved individually in various Jurassic groups of "gymnosperms:' some of which were almost certainly pollinated by insects. Beginning about 130 Mya in the Cretaceous, the angiosperms began to~ rapidly in diversity and achieved ecological dominance over the "gymnosperms." The anatomically most "advanced" groups of insects made their appearance in the Mesozoic. By the late Cretaceous, most families of living insects, including ants and social bees, had evolved. TIm)llghollt the Cretaceous and thereafter, insects and angiospenns affected each other's evolution and may have augmented each other's diversity. As dif-



Figure 5.20 An ichthyosaur. The dorsal and tail fins of this extinct J11arine reptile are superficially similar to those of sharks and porpoises, although the taiJ fin of porpoises is horizontal. (The Field Museum, negative #GE084968c,


ferent groups of pollinating insects evolved, adaptive modification of flowers to suit different pollinators gave rise to the great floral diversity of modern plants. It is largely becalise of tlte spectacular iI/crease of al1giosperms alln ii/sects tllnt terrestrial diversity is greater to-

day tilan ever before. Vertebrates The major groups of amniotes are distinguished by the number of openings in the temporal region of the skull (at least in the stem members of each lineage; Figure 5.19). One such was a group of mariJle reptiles that flourished from the late Triassic to tlie end of the Cretaceolls and included the dolphinlike ichthyosaurs, which gave birth to live young (Figure 5.20). 11,e diapsids, with h"o temporal openings, becaJlle tile IIIOst diverse grollp of reptiles. One major diapsid lineage, the lepidosauronlorphs, includes the lizards, which became differentiated into modern suborders in the late Jurassic and i.nto modern families in the late Cretaceous. One group of lizards evolved into the snakes. They probably originated in the Jurassic, but their sparse fossil record begins only in the late Cretaceous. TIle archosauIomorph diapsids were the most spectacular and diverse of the Mesozoic reptiles. Most of the late Permian and Triassic archosaurs were fairly generalized predators a meter or so in length (Figure 5.21). From this generalized body plan, numerous specialized forms evolved. Among the most highly modified ardlosaurs are the pterosaUIs, one of the three vertebrate groups that evolved powered flight. The wing consisted of a membrane extending to the body from the rear edge of a greatly elongated fourth finger (Figure 5.22). One pterosaur "vas the largest flying vertebrate known; others were as small as sparrows. Dinosnurs evolved from nrcllOsnllrs related to the one pictured in Figure 5.21. Dinosaurs are not simply any old large, extinct reptiles, but members of the orders Saurischia and Ornithischia, which are distinguished from each other by the form of the pelvis. Both orders included bipedal forms and quadrupeds that were derived from bipedal ancestors. Both orders arose in the Triassic, but neither order became diverse Luttil the Jurassic. Dinosaurs became very diverse: more than 39 families are recognized (Figure 5.23). The Saurischia iJlcluded carnivorous, bipedal theropods and herbivorous, quadrupedal sauropods. Am.ong the noteworthy theropods are Deinollydllfs, with a huge, sharp claw that it probably used to disemboweJ prey; the renowned Tyrnnnost1l1rl1s rex (late Cretaceous), which stood 15 meters high and weighed about 7000 kilograms; and the small theropods from which birds evolved. The sauropods, herbivores with small heads and long necks, include the largest animals that have ever lived on land, sucl, as Apatosaul'lls (= Brontosaurus); Bmciliosmu'us, which weighed more than 80,000 kilograms; and DiplOdocus, which reached about 30 meters in length. The Ornithischia-herbivores with specialized, sometimes very numerous, teeth-included the well-known stegosams, with dorsal plates that probably served for thermoregulation, and the ceratopsians (horned dinosaurs), of whidl Triceratops is the best known. The extinction of the ceratopsians at the end of the Cretaceous left only one surviving lineage of dinosaurs, which radiated extensively in the late Cretaceous or early Tertiary and today includes about 10,000 surviving species. Aside from


Lngo5ucJllfS, a Triassic thecodont archosaur, shO\~'ing the generalized body form of the stem Figure 5.21

group from which dinosaurs

evolved. (After Bonaparte 1978.)



A Jurassic pterosaur, RlmmpllOrl'yllclllfs, shm·\'i.ng the wing membrane supported by the greatly elongated fourth finger, the large breastbone (stemum) to which flight muscles were attached, and the terminal tail membrane found in this genus, possibly used for steering. (After Williston 1925.)

Figure 5.22

Figure 5.23 The great diversity of dinosaurs. The root of this proposed phylogeny is central, near the top. The two great clades of dinosaurs, Omithischia (1; left) and Saurischia (2; right), are curved downward to fit the page. The Saurischia included the Sauropoda (3) and Therapoda (4), of which birds (Aves) are the only survivors (5). Ornithischia included stegosaurs (6) and ceratopsians (7). All ornithiseman lineages are now extinct. (From Sereno 1999, 5ci",ce 284: 2139. Copyright © AAA5.)






J'J""'J"': ...


EII/nIlSmIflIS ScelidoSllllrus Hlf(/)'lllIgoS(/flrrlS



DIICClllnlrtlS . .





SlllltlOSilllfliS UampaSf/UfUS Ollleisaums


KODOSAUIUNAE Gargor/cos(lllflls













• By transposition and Lmegual crossing ovel~ TEs can i.ncrease in number, and so increase the size of the genome. Most or all of these transposable element-induced effects have been observed within experimental populations of organisms such as maize (corn, Zea mays) and Drosophila melnllogaster. The transposition rate of various retroelements ranges from about 10-5 to 10-3 per copy in inbred lines of Drosophila, resulting in an appreciable rate of mutation (Nuzhdin and Mackay 1994). All the kinds of changes engendered by transposable elements can be found by comparing genes and genomes of organisms of the same or different species. For instance, L] retrotransposon insertions are associated with many disease-causing mutations in both mice and humans (Kazazian 2004), and a difference in flower color behveen hiVO species of Petllnia has been caused by the insertion and subsequent incomplete excision of a transposon that disabled a gene that controls anthocyanin pigment production (Quattrocdtio et al. 1999). Examples of mutations Geneticists have learned an enormous amOLmt about the nature and causes of mutations by studyiJlg model organisms such as Drosophila and E. coli. iV1oreover, many human mu~ tations have been dlaracterized because of their effects on health. Human mutations are usually rather rare variants that can be compared with normal forms of the gene; in some instances, newly arisen mutations have been found that are lacking in both of a patient's parents. Single base pair substitutions are responsible for conditions such as sickle-cell anem..ia, described earlier, and for precocious puberty, in which a single amino acid change in the receptor for luteinizing hormone causes a boy to sho'w sjgns of puberty when about 4 years old. Because many different alterations of a protein can diminish its function, the same phenotypi.c condition can be caused by many different mutations of a gene. For exall1ple, cystic fibrosis, a fatal condition afflicting olle in 2500 live births in northern Europe, is caused by mutations in the gene encoding a sodium channel protejn. The most common such mutation is a 3 bp deletion that deletes a single am.ino acid from the protein; another converts a codon for arginine into a "stop" codon; another alters splicing so that an exon is missing from the mRNA; and lHany of the more tIlan 500 other base pair substitutions recorded in this gene are also thought to cause the disease (Zielenski and Tsui ]995). Mutations in any of the many different genes that contribute to the normal development of some characteristics can also result in similar phenotypes. For example, retinitis pigmentosa, a degeneration of the retina, can be caused by mutations in genes on 8 of the 23 chromosomes in the (haploid) human genome (Avise 1998).



A mutated Im\'-density lipoprotein (LDL) gene in humans lacks exon 5. It is believed to have arisen by unequal crossing over betvveen two normal gene copies, due to out-oF-register pairing between tvvo of the repeated sequences (Alu, shown as blue boxes) in the introns. The numbered boxes are exons. (After Hobbs et aI. 1986.) Figure 8.8

Normal LDL gene Out-oF-register pairing





of two Alu repeat sequences leads to


unequal crossing over.

Normal LDL gene


x -"'-'---'----




crossing avec

Known mutant LDt gene

5' fI,'!utant gene (not found

in human population)






One of the resulting


genes lacks exon 5.


The other, with duplicated exon 5, has ,,,.?, not yet been found in ,""""",----,----,-----,3 I human populations.


Hemophilia can be caused by mutations in tvvo different genes that encode blood-clotting proteins. In both genes, nlany different base pair substitutions, as \-vell as small deletions and duplications that cause fralueshifts, are known to cause the disease, and about 20 percent of cases of hemophilia-A are caused by an inversion of a long sequence within one of the genes (Greenet aJ. ]995). Huntington disease, a fatal neurological disorder that strikes in mjdlife, is caused by an excessive number of repeats of the sequence CAG: the normal gene has 10 to 30 repeats, the mutant gene more than 75. Unequal crossing over between the hvo tandemly arranged genes for a-hemoglobin (see Figure 8.3) has given rise to variants with three tandem copies (duplication) and with one (deletion). The deletjon of one of the Joci causes a-thalassemia, a severe anemia. Another case of deletion, which results in high cholesterol levels, is tJle lack of exon 5 in a low-density lipoprotein gene. This deletion has been attributed to unequal crossing over, facilitated by a short, highly repeated sequence called AlII that is located in the introns of this gene and in many oU,er sites in the genome (Figure 8.8). TI1ese examples might make it seem as if mutations are nothing but bad news. While this is close to the truth-far more mutations are harmful than helpful-these mutations represent a biased sample. rVlany advantageous mutations have become incorporated into species' genomes (fixed) and thus represent the current wild-type, or normal, genes. For example, most genes that have arisen by reverse transcription from mRl\Aare nonhmetional pseudogenes, but at least one has been found that is a fully functional member of the hl.unan genome. Phosphoglycerate kinase is encoded by two genes. One, on the X ehrOlTIOSOme, has a norma] structure of 11 exons and 10 introns. The other, on an autosome, lacks introns, and clearly arose from the X-linked gene by reverse transcription. It is expressed only in the testes, a novel pattern of tissue expression that suggests that the gene plays a new functional role (see Li 1997). When biologists seek those genes that have been involved in the evolution of a specific characteristic, they often use rare deleterious mutations of the kind described here as indicators of CANDIDATE GENES, those that may be among the genes they seek. For example, a rare mutation in the human FOXP2 gene (jorkhead box 2, which encodes a transcription factor) causes severe speech and language disorders. Two research groups, led by Jianzhi Zhang (Zhang et al. 2002) and Svante Paabo (Enard et al. 2002), independently found that this gene has undergone two nonsynonymous (amino add-changing) substitutions in the human lineage since the divergence of the human and chimpanzee lineages less than 7 Mya. This is a l1"'ltiCh higher rate of protein evolution than would be expected, considering that only one other such substitution has ocew"red between these species and the mouse, which diverged almost 90 Mya (Figure 8.9). Both research groups propose that these substitutions occurred in the human Lineage less than 200,000 years ago and that they are among the important steps in the evolution of human language and speech.





;\'!ouse (DlltgroUp)

The two non synonymous mutations in the human lineage represent an unusually high rate of evolution for this protein.


Figure 8.9 A phylogeny of the Hominoidea and its divergence from an outgroup, the mouse. Each box shows the number of nonsynonymous (yellmv boxes) and synonymous (white boxes) substitutions in the FOXP2 gene. The two nonsynonymolts substitutions in the human lineage represent an unusually high rate of evolution of the FOXP2 protein, and may represent mutations that have been important in the evolution of language and speech. (After Zhang

et aJ. 2002.)

Rates of mutation Recurrent mutation refers to tJle repeated origill of a particular mutation/ and the rate at which a particular mutation occurs is typically measured in terms of recurrent mutation: the number of independent origins per gene copy (e.g., per gamete) per generation or per unit time (e.g., per year). rvlutation rates are estiJnates, not absolutes, and these estimates depend on the method used to detect mutations. Tn classical genetics, a mutation was detected by its phenotypic effects, such as white versus red eyes in Drosophila. Such a mutation, however, might be caused by the alteration of any of many sites ·within a locus; moreover, Illany base pair changes have no phenotypic effect. Thus phenotypically detected rates of mutation Lmderestirnate the rate at which all mutations occur at a locus. With modern Illolecular methods/ 111utated DNA sequences can be detected directly, so mutation rates can be expressed per base pair. ESTIMATING MUTATION RATES. Rates of mutation are estimated in se,·eral ways (Drake et al. 1998). A relatively direct method is to count the number of mutations arising in a laboratory stock (which is usually initially homozygous), scoring mutations either by their phenotypic effects or by molecular methods. An indirect method (Box A) is based on the nLUnber of base pair differences between homologous genes in different species, relative to the number of generations that have elapsed since they diverged from thelr common ancestor. This method depends on the neulTal theory of molecular evolution, which is described in Chapter 10. Mutation rates vary among genes and even among regions within genes, but on average, as measured by phenotypic effects, a locus mutates at a rate of about 10-6 to 10-5 mutations per gamete per generation (Table 8.2). TIle average mutation rate per base pair, based mostly on the indirect method of comparing DNA sequences of different species, has been estimated at about ]0- 11 to 10-10 per replication in prokaryotes (see Table 8.3), or about 10-9 per sexual generation in eukaryotes. The mutation rate in the human genome has been estimated at about 4.8 x 10-9 per base pair per generation (Lynch et aJ. 1999). Back mutation is mutation of a "mutant" allele back to the allele (usually the wild type) from which it arose. Back mutations are ordinarily detected by their phenotypic effects. They usuaUy OCCUI at a mudl Jower rate. than "forward" Jllutat-ions (from \.viJd type to mutant), presumably because many more substitutions can impair gene function than can restore it. At the molecular level, most phenotypically detected back mutations are not restorations of the original sequence, but instead result from a second amino

TABLE 8.2 Spontaneous mutation rates of specific genes, detected by phenotypic effects Species and locus

ESc/leric/lin coli Strcptomyci n resistance Resistance to TJ phage Arginine independence

Mutations per 100,000 cells or gametes

0.00004 0.003 00004

Salmone/ln lyphi11l1(rillm Tryptophan independence


NellrDSpOrn crassn Adenine independence

Drosophila lIIelmlOgasler Yellow body Brown eyes Eyeless

0.0008-D.029 12 3 6

Homo sapiens Retinoblastinoma Achondroplasia Hunt-ington's chorea Source: After Dobzhansky 1970.

1.2-2.3 4.2-14.3 0.5



BOX SA Estimating Mutation Rates from Comparisons among Species n Chapter 10, we will describe the neutral theory of moleclliar evoilltion.


This theon' describes the fate of purely neutr~1 mutations-that is, those mutations that neither enhance nor lower fitness. One possible fate is that a mutation will become fixed-that is, at-

tain a frequency of l.D-entil"ely by chance. The probability that ti,is event \-",ill occur equals Ii, the rate at which neutral mutations arise. In each genera-

tion, therefore, the probability is



a mutation that occurred at some time

in the past will become fixed. After tile passage of t generations, the fraction of mutations that \",ill ha\'e become fixed is therefore ut. If h"lO species diverged from a common ancestor t generations ago, the expected fraction of fixed mutations in both species is 0 ; 21ft, since various mutations have become fixed in both lineages. If the mutations in question are base pair changes, a fraction 0 ; 2/1t

of the base pairs of a gene should differ beh,\leen the species, assuming that all

base pairs are equally likely to mutate. Thus the average mutation rate per

base pair per generation is


= D/2t.

Thus 1Ne can estimate Ii if \·ve can measure the fraction of base pairs in a gene that differ between nvc species (D), and if we can estimate the number

of generations si.nce the two species diverged from their CDnUTIon ancestor (t). This requires an estimate of the length of a generation, information from the fossil record on the absolute time at which the common ancestor existed,

and an understanding of the phylogenetic relationships among the living and fossilized taxa.

In applying this method to DNA sequence data, it is necessary to assume that most base pair substitutions are neutral and to correct for the possibility that earber substitutions at some sites in the gene have been replaced by later

substitutions ("multiple hits"). Uncertainty about the time since divergence from the common ancestor is usually the greatest source of error in estimates obtClined by tltis method. The best estimates of mutat.ion

rates at the moleculaT level have been obtained from interspecific comparisons of pseudogenes, other non translated sequences, and fourfold-degen-

erate third-base positions (those in which all mutations are synonymous), since these are thought to be least subject to natural selection (although probably not entirely free of it). In comparisons among mammal species, ti,e average rate of nucleotide substitution has been about 3.3-3.5 per nucleotide site per 109 years, for a mutation rate of 3.3-3.5 x 10-9 per site per year (li and Graur 1991). If the average generation time were 2 years during ti,e history of the lineages studied, the a"erage rate of mutation per site would be about 1.7 x la-" per generation. Comparison of human and c11impanzee sequences yielded an estimate of 1.3 x 10-9 per site per year, asswlling divergence 7 Mya. If the average generation time in these lineages has been 1:>-20 years, then the mutation rate is about 2 x 10-8 per generation. The human diploid

genome has 6 x 109 nucleotide pairs, so this implies at least 120 new muta-

tions per genome per generation-an astonishingly high number (Crow 1993).

acid substitution, either in the same or a different protein, that restores tile function that had been altered by the fiTst substitution. Advantageous mutations arose and compensated for severely deleterious mutations within 200 generations in experimental populations of E. coli (Moore et al. 2000). EVOLUTIONARY IMPLICATIONS OF MUTATION RATES.

With such a low mutation rate per locus,

it might seem that mutations occur so rarely that they cannot be important. However,

summed over all genes, the input of variation by mutation is considerable. If the haploid human genome has 3.2 x 109 bp and the mutation rate is 4.8 x 10-9 per bp per generation, an average zygote wiIJ carry about 317 new mutations. [f only 2.5 percent of the genome consists of functional, transcribed sequences, 7 of these new mutations will be expressed

and will have the potential to affect phenotypic characters (lynch et al. 1999). Oti,er authors have estimated the number of new base pair changes per zygote in the functional part of the genome as 0.14 in Drosophila, 0.9 in mice, and 1.6 in humans (Table 8.3). So, in a population of 500,000 humans, at least 800,000 new mutations arise every generation. If even a tiny fraction of these mutations were advantageous, the amount of new "raltv

material" for adaptation would be substantial, especially over tile course of tilousands or millions of years. Experiments on Drosophila have confinned that the total mutation rate per gamete is

quite high. For example, Terumi Mukai and colleagues (1972), in a heroically large experiment, counted more than 1.7 million flies in order to estimate the rate at ,vhich the




Estimates of spontaneous mutation rates per base pair and per genome Mutation rate

Base pairs Organism

n,14 phage

in haploid genome

pair per

per replication per haploid



per base

in effective genomea

1.7 X 10 5





per replication per effective

genome a

per sexual generation per effective genomeb


Escherichia coli




X 10- 10

Sacc!wJ"Olllycfs cerevisine (yeast)

1.2 X 10'






X 10- 11









Neurospora crassn (bread mold)




CaeJlorhnbditis e1egalls







Drosophila II1elt1l1ogasier


1.6 x 10'


X 10- 10





2.7 x 10'

8.0 x 10'





















SOl/ree: After Drake et a1. 1998. "The effective genome is the number of base pairs in functional sequences that could potentially undergo mutCltions that reduce fitness. I, Calculated for multiccllul'hkh multiple DNA repliciltion events occur in development behveen zygote and gametogenesis.

chromosome 2 accumulates mutations that affect egg-to-adult survival (VLABILITY). They used crosses (see Figure 9.7) in which copies of the wild-type chromosome 2 were carried in a heterozygous condition so that deleterious recessive mutations could persist 'without being eliminated by natural selection. Every 10 generations, they performed crosses that made large nun1bers of these chromosomes homozygous and measured the proportion of those chromosomes that reduced viability. The mean viabiJjty declined, and the variation (variance) among chrolTIOSOmes increased steadily (Figure 8.10). From the changes in the mean and variance, Mukai et al. calculated a mutation rate of about 0.15 per chromosome 2 per galnete. This is the sum, over alllod on the chromosome, of mutations that affect viability. Because chromosome 2 carries about a third of the Drosophila genome, the total mutation rate is about 0.50 per gan1ete. Thus almost every zygote carries at least one Figure 8.10 Effects of the ClCCllnew mutation that reduces Viability. Subsequent studies have indicated that the mutation Jnulation of spontaneous mutations on the egg-to-adult survival of rate for Drosophila is at least this high, and that it reduces viability by] to 2 percent per Drosophila lIlelmlOgaster. The mean generation (Lynch et al. 1999). Indirect estimates indicate that humans likewise suffer viability of flies made homozygous about 1.6 new mutations per zygote that reduce survival or reproduction (Eyre-Walker for chromosome 2 carrying new and Keightley 1999). recessive mutations decreased, and the variation (variance) among those Mutation rates vary among genes and chromosome regions, and they chromosomes increased. TIle rate of are also affected by environn1ental factors. MUTAGENS (mutation-causing mutation was estimated from these agents) include ultraviolet light, X-rays, and a great array of chemicals, data. (After Mukai et aJ. ]972.) many of which are environmental pollutants. For example, mutation rates in birds and mice are elevated in industrial areas, and mice exposed to par1.0 20 ticulate air pollution in an urban-indusb"ial site showed higher rates of mu0.9 18 CO tation in repetitive elements than mice exposed only to filtered air at that ~ 0.8 16 :COMPLETE OOMINANCE, measured by the degree to ,""hich the heterozygote resembles one or the other homozygote, may occur. Inheritance is said to be additive i1 the heterozygote's phenotype is p,.ecisely intermediate between those of tl,e homozygotes. For example, A,A" (AI

22 bristles


The drastic phenotypic effect of homeotic mutCltions thClt switd1 development from one patlwlay to another. (A) Frontal view of the head of a wild-type

Figure 8.14

Drosophila IIIclallogaster,


normal antennae and mouthparts. (B) Head of a fly carrying the Alltcllllapedia mutation, which converts antennae into legs. The AllteHIInpedin gene is part of a large complex of Hox genes that confer identity on segments of the body (see Chapter 20). (Photographs courtesy of F. R. Turner.)



Figure 8.15

If the Aallele is dominant. a single

Two of the possible relationships between phenotype and geno-

type at a single locus with h'\'o alleles. If inheritance is additi\"e, replacing eadl A' allele \,,'ith an A allele steadily increases the amount of gene product, and

copy is enough for nearly full expression of the character.

• AA AA"~_---~

the phenotype changes accordingly. If A is dominant over A', the phenotype of AA' nearly equals that of AA because the single dose of A produces enough gene product for full expression of the character.


A dominant over A'",,-

AlA" and A,A, may have phenotypes 3, 2, and 1, respectively; the effects of replacing each A, with an Al simply add up. LossOF-FLNCTION mutations, in which the activity of a gene product is reducedr are often at least partly recessive, \vh.ile dominant mutations often have enhanced gene product activity (Figure 8.15).

If inheritance is additive, each replacement of an A' allele with an A allele increases the phenotype

'Additive inheritance (no dominance) A'A' Amount of gene product

by the same amount.

Effects of mutations on fitness The effects of new mutations on fitness may range from highly advantageous to highly disadvantageous. Undoubtedly, many mutations are neutral, or nearly so, having very sl.ight effects on fitness (see Chapter 10). The average, or net, effect of those that do affect fitness is deleterious. This was shown, for example, by the decline in mean viabiHty in Mukai's Drosophila experiment, described above (see Figure 8.10), and by the fitness effects of single mutations isolated in experimental populations of E. coli and yeast (Figure 8.16). A few mutations slightly enhanced fitness, some greatly decreased it, and the majority had small deleterious effects. Under some circumstances, slightly deleterious mutations may act as if they are nearly neutral and aCC1Il11uJate in populations. They may therefore have more harmful collective effects on a population than do strongly deleterious mutations, which are more rapidly expunged

by natural selection. The frequency distribution of mutational effects is not fixed, for the fib1€SS consequences of many mutations depend on the population's environment and even on its existing genetic constitution. For exan1ple, the decline of fitness due to new mutations in some experi.mental Drosophila populations 'was more than 10 times greater if the flies were assayed under crowded, competitive conditions than under noncompetitive conditions (Shabalina et aI. 1997). Most mutations are pleiotropic-that is, they affect more than one character. For example, the yellow mutation in Drosophila affects not only body color, but also several components of male courtship behavior. In some cases, tl,e basis of deleterious pleiotropic efA very few mutations fects is understood; for example, some mutations appear to be slightly beneficial. that affect Drosophila bristle number also disrupt 1.0 the development of the nervous system and re., duce the viability of larvae, which do not have u bristles (Mackay et al. 1992). ~ 'to.8 Evolution would not occur unless SOUle n1U-;; The long "tail" shows that tations were advantageous. Many experiments .2 a minority of mutations that demonstrate advantageous mutations have " have severe effects ~0.6 been done 'with microorganjsms such as phage, (reduce fitness by >5%). bacteria, and yeast because of their short generaThe steep rise in the curve at this point shows that most mutations have neutral or very slightly deleterious effects.

0.00 (neutral)


0.10 Selection coefficient




Figure 8.16 The cumulati\'e frequency distributions of the effects of ne,,\' mutations on fitness in the bacterium Escllericllia coli and the yeast Saccharomyces cerevisiae. The higher the selection coefficient, the more the mutation reduces fitness. Beneficial cffccts are indicated bv values to the left of 0.0

(neutral). (After Lynch et aJ. 1999.)



tion times and the ease with which huge populations can be cultured (Dykhuizen 1990; Elena and Lenski 2003). Because bacteria ca.n be frozen (during \'\lhich time they undergo no genetic change) and later revived, samples taken at different times from an evolving population can be stored, and their fitness can later be directly compared. The fitness of a bacterial genotype is defined as its rate of increase in numbers relative to that of another genotype with whid, it competes in the same culture, but \vhich bears a genetic marker and so can be distinguished from it. Suppose, for example, that a culture is begun with equal numbers of genotypes A and B, and that after 24 hours B is h·vice as abundant as A. If the bacteria have grown for x generations, each initial cell has produced 2" descendants. Thus, if genotypes A and B have grown at the respective rates of 25 and 26 (Le., B has produced one more generation per 24 hours), their relative numbers are 32:64, or 1:2. The relative fitnesses of the genotypes-that is, their relative growth rates-are measured by their rates of cell division per day: namely, 5:6, or 1.0:1.2. If the genotypes had the same growth rates-say, 25_ both would increase in number, but their fitnesses would be equal. Richard Lenski and colleagues used this metl'lOd to trace the increase of fitness in populations of E. coli for an astonishing 20,000 generations. Each population was initiated with a single individual and \-vas therefore genetically uniform at the start. Nevertheless, fitness increased substantially-rapidly at first, but at a decelerating rate later (Figure 8.17A). In a similar experiment (Bennett et a!. 1992), E. coli populations adapted rapidly to several different temperatures (Figme 8.17B,C,D). Bacteria can be screened for mutations that affect their biochemical capacities by placing them on a mediwTI on which that bacterial strain cannot grow, such as a medium that lacks an essential amino acid or other nutrient. Whatever colonies do appear on the medium must have grown from the fev.,.' cells in wruch mutations occurred that conferred a new biochemical ability. For example, Barry Hall (1982) studied a strain of E. coli that lacks the IneZ gene, which encodes ~-galactosidase, t1le enzyme that enables E. coli to metabolize the sugar lactose as a source of carbon and energy. Hall screened popu lations for the abil(A) ity to grow on lactose and recov(8) § 1.8 ered several rnutations. A mutau tion in a different gene (ebg) ~ 1.6 • altered an enzyme that normally E u 1.4 performs another function so ·?: ~ 0.9 a that it could break do\vn lactose. "E 1._? Another mutation altered regu~ 1.0