Research on guppies provides evidence that phenotypic plasticity — an organism's ability to alter its characteristics in response to changes in the environment — can both constrain and facilitate adaptive evolution. See Letter p.372
Altered or new environmental conditions, such as those brought about by climate change, are important sources of selection pressures that drive organismal adaptation and evolution. But alongside genetic adaptation, organisms can respond to environmental challenges through adaptive phenotypic plasticity, which refers to a non-genetic shift in the average characteristics (phenotype) of a population towards an evolutionary optimum. Whether phenotypic plasticity generally facilitates or constrains adaptive (genetic) evolution remains a contentious issue1,2,3,4. On page 372 of this issue, Ghalambor et al.5 provide experimental evidence from guppies suggesting that adaptive phenotypic plasticity in gene-expression patterns constrains evolution. But they also find that non-adaptive plasticity — phenotypic changes that do not directly contribute to increased fitness under the changed conditions — may facilitate adaptive genetic change by increasing the strength of natural selection.
The authors' experiments involved transplanting wild Trinidadian guppies (Poecilia reticulata; Fig. 1) from a stream that also hosted predatory cichlid fish into two replicate streams without cichlids. They then compared patterns of brain gene expression between the introduced and original (ancestral) populations after three or four generations. Parallel changes in gene expression had occurred for 135 genes in the two introduced populations, and these new levels of gene expression were similar to those exhibited by a native cichlid-free population. This suggested rapid adaptive evolution in the introduced populations.
However, the evolved differences were mostly (89% of the genes) in the opposite direction to that of phenotypic plasticity in expression patterns in the ancestral population. This was inferred by comparing the gene expression in ancestral fish reared in either the presence or absence of chemical cues from predatory cichlids. Thus, the phenotypic plasticity in these genes can be considered non-adaptive. The remaining 11% of genes exhibited adaptive plasticity — the evolved differences in gene expression in the experimentally introduced populations were concordant with the direction of change of expression levels in ancestral fish raised in the absence of predatory-fish cues. The authors also observed that there was little or no population divergence in the expression of these genes in either of the introduced populations.
The latter findings support evolutionary models predicting that adaptive phenotypic plasticity should weaken the strength of directional selection and thereby slow the rate of evolution (see refs 6 and 7 for examples). However, the real stunner of the study was the discovery that most of the evolved (genetic) differences in gene-expression patterns in the introduced guppy populations had taken place in the opposite direction to the direction of plasticity in the ancestral population. This inverse relationship between the direction of plasticity and the direction of adaptive evolution suggests that non-adaptive plasticity may facilitate (in the authors' words, potentiate) evolution by increasing the strength of directional selection required to create the observed divergence in gene-expression patterns.
The authors obtained support for the hypothesized increase in directional selection against non-adaptive plasticity by examining evolutionary changes in the magnitude of plasticity (quantified as the mean difference in expression levels of gene transcripts in the predator-cue-treated groups) between ancestral and introduced populations. Although phenotypic plasticity is, by definition, a non-genetic response to environmental cues, the capacity to express it, and its magnitude, can be genetically variable1,4. Consequently, if directional selection had acted most strongly on gene transcripts exhibiting non-adaptive plasticity, then the magnitude of plasticity in introduced populations in response to this selection should be reduced. This was just what Ghalambor et al. observed. Moreover, the decline in the magnitude of plasticity in the introduced populations was inversely proportional to plasticity in the ancestral population. This also aligns with the expectation that transcripts exhibiting the greatest non-adaptive plasticity should be the ones that are most strongly selected against.
Although the findings that phenotypic plasticity can both constrain and facilitate evolutionary (genetic) adaptation are not unprecedented, several features of Ghalambor and colleagues' study set it apart from earlier work on this topic. For instance, instead of focusing on a limited number of traits, the authors assessed the plasticity of a large number of traits (expressed genes), which allowed them to draw robust quantitative conclusions. Nevertheless, a question to be addressed is whether results from gene-expression analyses can be extended and generalized to macroscopic traits that have more-direct ecological relevance. Similarly, most previous empirical studies that focused on the direction of plastic responses and the direction of subsequent evolutionary divergence in wild populations have been limited to comparisons between ancestral and derived populations long after they diverged. The new study's focus on initial patterns of plasticity and subsequent rapid adaptive divergence in the wild provides a thought-provoking complement to laboratory experiments that have provided evidence supporting both positive (adaptive)8,9 and negative (non-adaptive)10 relationships between the directions of plastic responses and evolution.
Ghalambor and colleagues' results are also intriguing because most (but not all) attempts to model the effects of plasticity on subsequent evolution have assumed it to be adaptive. Thus, the observed negative relationship between the direction of plasticity and the direction of evolution in guppies may guide future theoretical work in the field. Furthermore, although increased strength of selection caused by non-adaptive plasticity may contribute to rapid adaptation and increase the likelihood of population persistence, it may also lead to reduced population size and an increased risk of demographic collapse3. By reducing population size, selection stemming from non-adaptive plasticity may expose a population to an increased rate of random genetic changes owing to a process known as genetic drift. This would in turn propagate loss of genetic variation and reduced efficiency of selection, counteracting the proposed benefit from non-adaptive plasticity.
As fascinating as it is to suggest that maladaptive plasticity may be a strong driver of evolution, sceptics may require further experimental studies from the wild with more population replicates and with a focus on traits with established ecological relevance (such as behaviours and morphology) to be convinced. Such studies would also be helpful, if not essential, in developing parameters for models that aim to understand how the interplay between phenotypic plasticity, natural selection and random genetic drift influences evolutionary changes. Footnote 1
West-Eberhard, M. J. Developmental Plasticity and Evolution (Oxford Univ. Press, 2003).
Price, T. D., Qvarnström, A. & Irwin, D. E. Proc. R. Soc. B 270, 1433–1440 (2003).
Chevin, L.-M., Lande, R. & Mace, G. M. PLoS Biol. 8, e1000357 (2010).
Pfennig, D. W. et al. Trends Ecol. Evol. 25, 459–467 (2010).
Ghalambor, C. K. et al. Nature 525, 372–375 (2015).
Ancel, L. W. Theor. Popul. Biol. 58, 307–319 (2000).
Paenke, I., Sendhoff, B. & Kawecki, T. J. Am. Nat. 170, E47–E58 (2007).
Waddington, C. H. Adv. Genet. 10, 257–293 (1961).
Suzuki, Y. & Nijhout, H. F. Science 311, 650–652 (2006).
Schaum, C. E. & Collins, S. Proc. R. Soc. B 281, 20141486 (2014).
Related links in Nature Research
About this article
Behavioral Ecology (2020)
Science of The Total Environment (2019)
A comparison of life-history traits in calcifying Spirorbinae polychaetes living along natural pH gradients
Marine Ecology Progress Series (2018)
De novo transcriptome assembly and annotation for the desert rainbowfish ( Melanotaenia splendida tatei ) with comparison with candidate genes for future climates
Marine Genomics (2017)