sign in sign up
<<
Home , Epistasis

PERSPECTIVES

EVOLUTION Laboratory experiments with shed light on how epistatic infl uence In , the Sum Is Less the pace of evolution. than Its Parts

Sergey Kryazhimskiy, 1,2 Jeremy A. Draghi, 1 Joshua B. Plotkin1

ropagating bacteria in a lab for thousands of genera- Ptions may seem tedious, or even irrelevant, to most evolution- Ancestor strain Adapted strain ary biologists. Nonetheless, such Evolves in lab experiments provide an opportunity to deduce quantitative principles of W evolution and directly test them in ab controlled environments. Combined with modern sequencing technolo- Downloaded from gies, as well as theory, recent micro- bial experiments have suggested a critical role for genetic interactions W Fitness a among , called , Diminishing returns Reconstruct intermediates in determining the pace of evolution.

Two papers in this issue, by Khan et http://science.sciencemag.org/ al. on page 1193 (1 ) and Chou et al. (2 ) on page 1190, present precise Fitness Antagonistic epistasis experimental measurements of these W < W • W Fitness W ab a b b epistatic interactions. 1st 2nd 3rd Microbial evolution experiments mutation mutation in a simple, constant environment reveal a characteristic pattern: At Antagonistic epistasis. Bacteria adapt to a laboratory environment by acquiring benefi cial mutations. Khan et al. fi rst, a population rapidly acquires and Chou et al. identifi ed the mutations that accrued in an adapted strain, and measured their fi tness benefi ts (growth beneficial mutations, but then advantage relative to the ancestor). The mutations conferred smaller marginal benefi ts in combination than they did

adaptation progressively slows so individually. This antagonistic epistasis causes progressively slower rates of adaptation over time. on March 13, 2019 that thousands of generations pass between subsequent benefi cial substitutions late predictive evolutionary models or to infer tions, individually and in combination, the ( 3). Unexpected outcomes, however, can and such interactions from empirical data. Nev- researchers were able to directly quantify the do occur even in these simple experimental ertheless, epistasis is at the heart of classi- extent and form of epistasis (see the fi gure). conditions. Populations evolve a dramatically cal theories, such as the evolution of sex (8 ), Both studies found a predominance of elevated mutation rate (4 ), discover rare phe- and also of modern concepts such as robust- antagonistic epistasis, which impeded the notypic innovations ( 5), or diverge into dis- ness and (a population’s ability rate of ongoing adaptation relative to a null tinct lineages that either coexist ( 6) or com- to evolve) ( 9). Moreover, recent theoretical model of independent mutational effects. pete vigorously as each strain races to acquire work ( 10) suggests that the overall dynami- Chou et al. further interpreted the prevalence more adaptive mutations ( 7). Recent theory cal pattern of adaptation observed in long- of antagonistic epistasis in terms of meta- suggests that a common cause underlies all term microbial experiments can be explained bolic costs and benefi ts. The concordance of these phenomena: the structure of epistatic by a prevalence of what is called antagonistic results from the two studies is noteworthy, interactions among mutations. epistasis, in which benefi cial mutations con- especially because Khan et al. analyzed Esch- Epistasis describes how the fi tness conse- fer less benefi t in combination than they do erichia coli populations [from the long-term quence of a mutation depends on the status of individually. experiments of Lenski (3 )], whereas Chou et the rest of the . In one extreme exam- To quantify epistasis among beneficial al. studied an engineered strain of Methylo- ple, called sign epistasis, a mutation may be mutations and to test these theoretical predic- bacterium extorquens. The remarkable preci- benefi cial if it arises on one genetic back- tions, both Khan et al. and Chou et al. exam- sion with which both studies quantifi ed epis- ground, but detrimental on another. Although ined the initial substitutions that occurred in tasis among benefi cial mutations was made interactions among may seem an obvi- populations of bacteria adapting in the labo- possible only by leveraging whole-genome SCIENCE ous fact of biology, the myriad possible forms ratory. The researchers identifi ed the hand- sequencing combined with the ability to of epistasis have made it diffi cult to formu- ful of mutations across the genome that had reconstruct mutational combinations and substituted in an evolved strain, and then con- assay them in the same environment in which 1Department of Biology, University of Pennsylvania, Phila- structed intermediate strains containing com- the mutations fi rst arose. delphia, PA 19103, USA. 2Department of Organismic and , Harvard University, Cambridge, MA binations of these mutations. By measuring The view of epistasis across a genome that

02138, USA. E-mail: [email protected] the fi tness benefi ts conferred by these muta- emerges from this work contrasts sharply HUEY/ ADAPTED BY P. CREDIT:

1160 3 JUNE 2011 VOL 332 SCIENCE www.sciencemag.org Published by AAAS PERSPECTIVES with the type of epistasis found among adap- whether experiments in simple environments, epistasis has been implicated in the evolu- tive mutations within a single protein ( 11). with only one or a few niches for coexisting tion of drug resistance in infl uenza viruses Notably, Weinreich et al. studied mutations strains, will refl ect the pattern of adaptation in (12 ) and in bacterial pathogens ( 13). Ulti- in an resistance , β-lactamase, more complex ecologies, such as Pseudomo- mately, populations of bacteria tediously and found a prevalence of sign epistasis, nas fl uorescens in structured environments propagated in the lab may be key to predict- which limits the number of genetic paths ( 6). Nonetheless, the compelling consistency ing the next moves of the most mutable and that evolution can follow ( 11). In contrast, between these two studies should inspire dangerous human pathogens. the epistasis documented by Khan et al. and efforts to test the generality of their fi ndings, Chou et al. exerts less constraint on the order by measuring epistasis in a wide range of References 1. A. I. Khan, D. M. Dinh, D. Schneider, R. E. Lenski, T. F. of substitutions that increase fi tness, so that experimental and even natural systems. Cooper, Science 332, 1193 (2011). the specifi c path that evolution will take is These studies, and the long-term labora- 2. H.-H. Chou, H.-C. Chiu, N. F. Delaney, D. Segrè, C. J. less predictable. At the same time, the preva- tory evolution experiments from which they Marx, Science 332, 1190 (2011). lence of antagonistic epistasis measured by derive, represent a resounding achievement 3. S. F. Elena, R. E. Lenski, Nat. Rev. Genet. 4, 457 (2003). 4. P. D. Sniegowski, P. J. Gerrish, R. E. Lenski, Nature 387, the two groups ensures a predictable tempo for the reductionist approach to studying 703 (1997). of adaptation characterized by diminishing biology. The mechanistic picture they paint 5. Z. D. Blount, C. Z. Borland, R. E. Lenski, Proc. Natl. Acad. marginal returns (10 ). of evolution is complex but not incompre- Sci. U.S.A. 105, 7899 (2008). 6. P. B. Rainey, M. Travisano, Nature 394, 69 (1998). Although these new experiments suggest hensible; although epistatic interactions lead 7. R. J. Woods et al., Science 331, 1433 (2011). a consistent principle of how epistasis shapes to surprising phenomena, the advantages 8. A. S. Kondrashov, Nature 336, 435 (1988). the pattern of adaptation, many questions of a frozen “fossil record” of laboratory- 9. G. P. Wagner, L. Altenberg, Evolution 50, 967 (1996). 10. S. Kryazhimskiy, G. Tkačik, J. B. Plotkin, Proc. Natl. Acad. Downloaded from must be answered before their results can raised isolates, and the ease of manipulat- Sci. U.S.A. 106, 18638 (2009). be extended to evolution outside the labora- ing—and, now, fully sequencing—evolved 11. D. M. Weinreich, N. F. Delaney, M. A. Depristo, D. L. tory. It remains unclear, for instance, whether strains enables researchers to tease apart and Hartl, Science 312, 111 (2006). 12. J. D. Bloom, L. I. Gong, D. Baltimore, Science 328, 1272 these results would be altered by changing examine the underlying causes of these phe- (2010). fundamental evolutionary parameters, such nomena. Moreover, the theory and concepts 13. S. Trindade et al., PLoS Genet. 5, e1000578 (2009).

as population size, rate of mutation, and rate developed to explain these simple experi- http://science.sciencemag.org/ of recombination. Likewise, it is unclear ments may have broad payoffs. Already, 10.1126/science.1208072

GENOMICS Does behavior evolve through gene expression Behavior and the Dynamic Genome changes in the in response to the environment? Alison M. Bell1, 3 and Gene E. Robinson 2,3

on March 13, 2019 hen circumstances change, an ples (3 ), the latter as a driver of behavioral over both physiological and evolutionary organism’s fi rst response is often evolution has never been widely accepted, time scales, provide a possible mechanism Wbehavioral. But how does adap- perhaps as a reaction against Lamarckian- for how behavioral plasticity might drive tive behavior evolve, given that it requires ism—the idea that characteristics acquired rapid behavioral evolution through changes constant and often instantaneous interac- by habit, use, or disuse can be passed on in gene regulation. In an environment with tions between an individual and its environ- across generations. However, behavioral more predators, colonies producing more ment? The dominant view emphasizes new and , especially for ani- bees with lower thresholds for responding random DNA mutation as the starting point. mals in natural populations, lend some plau- to alarm pheromone would have fared bet- This may lead to behavioral variation. If sibility to the view. ter, which would then result in a popula- the resulting variants have different fi tness The ability to analyze genome-wide gene tion with patterns of gene expression whose values, then could result expression through “transcriptomics” has output was an “aroused” behavior, even in in behavioral evolution through changes in shown that the genome responds dynami- the absence of alarm pheromone. Although frequencies across generations. An cally to stimuli (4 ). One illustrative exam- this view does not rule out the possibility alternative theory proposes environmentally ple is the honey bee. The African honey bee that these differences in aggression arose induced change in an organism’s behavior as (Apis mellifera scutellata) responds much through new mutation, the transcriptomics the starting point (1 ), and “phenotypic plas- more fi ercely when its hive is attacked than agrees with the idea of “genetic accommo- ticity” that is inherited across generations do other subspecies of honey bee. Evolu- dation” (3 ), the modern, more inclusive ver- through an unspecifi ed process of “genetic tionary changes in brain gene expression sion of , which could assimilation” ( 2). Despite numerous exam- may have resulted in an increase in respon- involve either evolutionary increases or siveness to alarm pheromone (the chemical decreases in plasticity. In certain environ- 1Department of Animal Biology, University of Illinois, bees use to alert each other to danger) for ments, plastic might be favored, Urbana-Champaign, IL 61801, USA. 2Department of Ento- African honey bees ( 5). About 10% of the but in other environments, nonplastic gen- mology, University of Illinois, Urbana-Champaign, IL 61801, same genes regulated in the brain by alarm otypes might be preferred instead. Future 3 USA. Neuroscience Program, Program in Ecology, Evolution- pheromone are also differentially expressed studies will determine whether differences ary Biology and Conservation, Institute for Genomic Biology, University of Illinois, Urbana-Champaign, IL 61801, USA. between African and the less aggressive in honey bee aggression can be explained E-mail: [email protected] European honey bees. These genes, acting by selection on regulatory regions of the

www.sciencemag.org SCIENCE VOL 332 3 JUNE 2011 1161 Published by AAAS In Evolution, the Sum Is Less than Its Parts Sergey Kryazhimskiy, Jeremy A. Draghi and Joshua B. Plotkin

Science 332 (6034), 1160-1161. DOI: 10.1126/science.1208072 Downloaded from

ARTICLE TOOLS http://science.sciencemag.org/content/332/6034/1160

SUPPLEMENTARY http://science.sciencemag.org/content/suppl/2011/06/02/332.6034.1160.DC1 MATERIALS http://science.sciencemag.org/

RELATED http://science.sciencemag.org/content/sci/332/6034/1193.full CONTENT http://science.sciencemag.org/content/sci/332/6034/1190.full

REFERENCES This article cites 13 articles, 7 of which you can access for free http://science.sciencemag.org/content/332/6034/1160#BIBL

on March 13, 2019 PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title Science is a registered trademark of AAAS.