8/8/2019 Geneticla Theory of Adaptation

1/3

R440 Dispatch

Evolution: Towards a genetical theory of adaptationMichael Travisano

The population genetic basis for adaptation hasremained obscure despite a longstanding body oftheory. Microbial selection experiments are beginningto provide some answers.

Address: Department of Biology & Biochemistry, University of Houston,Houston, Texas 77204-5513, USA.E-mail: mtrav@uh.edu

Current Biology 2001, 11:R440R442

0960-9822/01/$ see front matter 2001 Elsevier Science Ltd. All rights reserved.

Knowledge of the genetic basis of adaptation should providethe centerpiece of a unified theory of evolution. Such atheory would greatly aid in the determination of past evo-lutionary changes, such as those leading to the evolution of

Homo sapien s. Perhaps more importantly, forward-lookingevolutionary predictions would have a much stronger foun-dation. Exactly how adaptation occurs at the genetic levelis still uncertain. Recent studies have focused on two broadquestions: are adaptive mutations of large effect, and whatfactors restrict rates of adaptation?

Current thinking on the genetic basis of adaptation has itsroots in the great debate that occurred at the end of the

19th century on the genetics of speciation and the apparentcontradiction between Darwinian natural selection andMendelian genetics [1]. One view was that speciationresulted from the gradual fixation of a large number of changes, each of small effect. The opposing view held thatspeciation required mutations of large effect, and wasbased in part on the often large effects of Mendeliangenetic factors. After several decades of vitriol, a numberof geneticists pointed out that there is actually no conflict,and that the rules of Mendelian genetics apply equallywell to genes of small effect as to those of large effect.

The reconciliation of Mendelian genetics and natural selec-tion provided a mathematical basis for the study of evolu-tion and adaptation, upon which hypotheses for the geneticsof adaptation have been made. One hypothesis states thatmost adaptive mutations fixed in a population are of smalleffect [2]. Mutations having large phenotypic effect wouldhave a greater likelihood of having wide-ranging deleteri-ous pleiotropic effects, negating any adaptive benefit. Asecond hypothesis states that the rate of adaptation shoulddecline over time as a result of the limited number of pos-sible adaptive mutations [35].

Assessment of these hypotheses has often relied upon

indirect or static measurements of adaptation. Paleontology

has provided important information on the tempo of evo-lutionary change, but it is much less useful in the assess-ment of selective benefits of phenotypic change or thegenetic basis for adaptation. Surveys of extant diversity,either phenotypic or genetic, have identified substantialdiversity, but have been less able to quantify selectiveeffects. An ideal approach would be to document adapta-tion as it occurs. Recent work has taken just this approach,using rapidly growing microbial populations to directlyobserve evolution in action.

In a series of experiments, Lenski and colleagues [4,6,7]propagated twelve replicate populations of Escherichia coli for over 10,000 generations. The goal was to investigate thedynamics of adaptation, using replicate populations propa-gated under the same environmental conditions and havingthe same immediate ancestor. Under these conditions, theimprovements in Darwinian fitness could be directly asso-ciated with the fixation of novel adaptive mutations arisingwithin each of the replicate populations. Using step-modelregression to detect adaptive events, three step-like fitnessincreases were found on average within each populationover the first 2,000 generations, each step of roughly a 10%fitness increase. However, the rate of adaptation declinedover the course of propagation: between 1000 and 2000

generations, the average rate of improvement was approxi-mately 0.108 per 1000 generations, but the rate fell toapproximately 0.008 between 5,000 and 10,000 generations.

Subsequently, over far fewer generations but with greaterprecision, Burch and Chao [8] measured the fitness benefitdue to adaptive mutations in populations of the RNA virus6. A low fitness ancestral genotype was generated by thefixation of a single deleterious mutation, reducing fitness byabout 90%. Phage populations, of different sizes, were thenallowed to recover and the fitness step size of the firstadaptive mutation was measured. Fitness was measuredevery five generations for 100 generations and at least onebeneficial mutation was always observed prior to 100 gen-erations of selection. The fitness benefit conferred by theadaptive mutations were generally smaller than the fitnessreduction due to the deleterious mutation, but populationsize was an important determinant of fitness benefit: thesmaller the effective population size, the smaller thebenefit. Further investigation demonstrated that the adap-tation of 6 was limited by the accessibility of advanta-geous genotypes within the mutational neighbourhood (theset of mutants one or a few mutational steps away) [9].

A difficulty for most selection studies, including those

mentioned above, has been the identification of adaptive

8/8/2019 Geneticla Theory of Adaptation

2/3

Dispatch R441

events. If an adaptive mutation yields only a 10% fitnessimprovement, identifying the time frame for the appear-ance of the mutation is technically difficult. Initially, anadaptive mutation is at a low frequency (1/population size)and has little effect on the fitness of a population. Inaddition, multiple adaptive mutations may occur simulta-neously, obscuring the fitness effect of any single mutation[10]. Burch and Chao [8] overcame these difficulties, butbegan with an ancestral genotype of very low initial fitnessand focused primarily on the very first adaptive mutation.

Imhof and Schlotterer [11] have recently been able toperform a detailed examination of step size over 1,000 gen-erations of selection, by an ingenious approach using thehigh rate of mutation inherent in microsatellite DNA. Amicrosatellite DNA locus (GA)n was cloned into a rel-atively low copy plasmid (pBR322), which was then trans-formed into a single E. coli genotype. The plasmid-bearinggenotype was used to initiate ten replicate populations thatwere then subsequently propagated for 1,000 generations.Carriage of the plasmid was ensured by periodic treatmentof the cultures with ampicillin, resistance to which wasencoded on the plasmid. During propagation of the repli-cate populations, the rapid evolution and diversification of the microsatellite DNA provided a marker for differentclonal lineages within each population. The populationswere assayed every 90 generations by restriction analysisof the plasmid DNA, and biologically significant changesin allele frequency were determined relative to the signal

intensity of the restriction bands.

The ability to track clonal lineages within populationsallowed Imhof and Schlotterer [11] to make several impor-tant observations. Firstly, numerous adaptive events wereidentified, 66 across all ten populations over the course of the propagation. Each event was identified by a shift inmicrosatellite allele frequency within a population, result-ing from the evolution of new microsatellite alleles and acombination of genetic drift and the selection of adaptivemutations associated with bacterial lineages having differ-ent microsatellite alleles. It was shown that selection, andnot drift, was the primary cause of the frequency shifts byrestarting four populations, with five-fold replication, atleast 90 generations prior to the observation of a signif-icant microsatellite allelic shift in the original fourpopulations. In 19 of the 20 populations evolution wasrecapitulated and the same allelic shift was observed as inthe original populations.

Secondly, the rate of adaptation was not observed todecline over the course of the selection (Table 1). Theoccurrence and persistence of selective sweeps did notdecline over time, nor did the selective benefit of adaptivemutations. This result stands in contrast with most other

selection experiments, either with E. coli [6] or with other

species [12]. Imhof and Schlotterer [11] suggest that thediscrepancy may simply have been due to the relativelyshort-term length of the experiment, but declines in therate of adaptation have been observed in other 1000-generation selection experiments with E. coli [13]. Thedifference is more likely to have alternative explanations,two of which are the particular nutrient environment thatwas used in the experiment, and the manner in whichfitness was determined.

Imhof and Schlotterer [10] chose for their experiments anutrient-rich environment, containing abundant peptides

and vitamins, in contrast with most other bacterial selectionexperiments, which were performed in minimal environ-ments containing a single carbon and energy source. Richnutrient conditions have been shown to greatly increase therange of adaptive possibilities for bacterial populations, pri-marily because of increases in ecological interactions [14]. Itis probable that there were a greater number of highly ben-eficial adaptive mutations in the nutrient-rich environmentthan has been observed in the nutrient-poor environmentsof other selection studies. The manner in which fitnessassays were performed may have been important, as fitnesswas determined by the changes in allele frequency withinthe evolving populations, and not by separate competitionexperiments against the common ancestor, as has been typ-ically performed. If frequency-dependence is important forfitness, as has been observed in microbial populations[1517], then the rate of fitness improvement will varydepending upon the competing genotypes.

Finally, and most importantly, Imhof and Schlotterer [11]were able to make an empirical measure of the distribu-tion of fitness benefits for adaptive mutations. The fitnessmeasures ranged from 0.0060.059, with a median fitnesseffect between 0.01 and 0.02. The distribution of fitnesseffects appears to be exponentially distributed, in agree-

ment with the hypothesis that most adaptive mutations

Table 1

Time course of the number of alleles that significantly increasein frequency for each population.

Generation90 180 270 360 450 540 630 720 810 900 990

Population1 0 2 0 2 1 1 0 1 0 0 02 0 1 2 0 0 1 0 0 1 0 13 0 0 0 0 2 0 2 1 1 5 04 0 1 0 1 0 1 0 0 0 0 05 0 2 0 1 1 5 0 0 2 0 06 0 2 1 0 1 1 0 3 1 1 17 0 1 1 2 1 0 0 0 0 0 08 0 0 2 0 0 0 1 0 0 0 09 0 2 0 0 0 0 1 2 1 0 1

10 0 0 0 0 0 1 0 0 0 1 0

8/8/2019 Geneticla Theory of Adaptation

3/3

are of relatively modest effect. As the authors note,however, the estimates obtained are likely to be some-what downward biassed, given the manner in which fitnesswas measured and the somewhat imprecise nature of themicrosatellite marker.

Understanding the genetics of adaptation is one of the mostvexing and complex problems in Biology. Adaptation is aprocess common to all life and plays a central role in shapingbiological diversity. However, quantitative hypotheses onthe genetics of adaptation, devised over half a century ago,have remained essentially untested because of the lack of appropriate data. The primary difficulty has been identify-ing those mutations which are important during adaptationand then quantifying their selective effect. The combina-tion of appropriate molecular techniques and microbialselection experiments shows great promise for the deriva-tion of a genetical theory of adaptation.

References1. Provine WB: The Origins of Theoretical Population Genetics.

Chicago: University of Chicago Press; 1971.2. Fisher RA: The Genetical Theory of Natural Selection. Oxford:

Clarendon Press; 1930.3. Wright S: Evolution and the Genetics of Populations. Volume 3:

Experimental Results and Evolutionary Deductions. Chicago:University of Chicago Press; 1977.

4. Lenski RE, Rose MR, Simpson SC, Tadler SC: Long-termexperimental evolution in Escherichia col i. I. Adaptation anddivergence during 2000 generations. Am Nat 1991,138: 1315-1341.

5. Kauffman SA: The Origins of Order: Self-Organization and Selection in Evolution. Oxford: Oxford University Press; 1993.

6. Lenski RE, Travisano M: Dynamics of adaptation and diversification:a 10,000 generation experiment with bacterial populations. Proc Natl Acad Sci USA 1994, 91: 6808-6814.

7. Elena SF, Cooper VS, Lenski RE: Punctuated evolution caused byselection of rare beneficial mutations. Science 1996,272: 1802-1804.

8. Burch CL, Chao L: Evolution by small steps and ruggedlandscapes in the RNA virus f6. Genetics 1999, 151: 921-927.

9. Burch CL, Chao L: Evolvability of an RNA virus is determined by itsmutational neighbourhood. Nature 2000, 406: 625-628.

10. Gerrish PJ, Lenski RE: The fate of competing beneficial mutationsin an asexual population. Genetica 1998, 102/ 103: 127-144

11. Imhof M, Schlotterer C: Fitness effects of advantageous mutationsin evolving Escherichia coli populations. Proc Natl Acad Sci USA2001, 98: 1113-1117.

12. Bell G: Selection The Mechanism of Evolution. London: Chapmanand Hall; 1996.

13. Travisano M: Long-term experimental evolution inEscherichia coli

.VI. Environmental constrains on adaptation and divergence.Genetics 1997, 146: 471-479.

14. Travisano M, Rainey PB: Adaptive radiation in microbes. Am Nat 2000, 156: S35-S44

15. Paquin C, Adams J: Relative fitness can decrease in evolvingpopulations of S. cerevisiae. Nature 1983, 306: 368-371.

16. Rainey PB, Travisano M: Adaptive radiation in a heterogeneousenvironment. Nature 1998, 394: 69-72.

17. Rozen DE, Lenski RE: Long-term experimental evolution inEscherichia coli VIII . Dynamics of a balanced polymorphism. Am Nat 2000, 155: 24-35.

R442 Current Biology Vol 11 No 11