Réexaminer le rôle de la plasticité phénotypique dans le processus évolutif / Revisiting the role of phenotypic plasticity in evolution

Séminaire FSER organisé par Christian Braendle (CNRS & University of Nice, France) et Henrique Teotónio (ENS Paris, France) du 25 au 30 août 2014

Participants

Erik Andersen, Charles F. Baer, Christian Braendle, Ivo Chelo, Asher Cutter, Marie Delattre, Ronald E. Ellis, Marie-Anne Félix, Eric Haag, Simon C. Harvey, Karin C. Kiontke, Patrick McGrath, Stephen Proulx, Scott Rifkin, Mattew Rockman, Henrique Teotónio, Mark Viney

—

Résumé

L’un des défis majeurs de la biologie actuelle est de comprendre comment les variations de l’environnement interagissent avec le génome et les processus de développement pour produire de la diversité phénotypique, et de comprendre également comment ces interactions évoluent. Notre but est de mieux définir ces problèmes et de considérer les différentes approches permettant de les résoudre, en nous consacrant à la biologie de Caenorhabditis elegans et d’autres espèces voisines. Récemment, C. elegans est devenu un organisme modèle de choix pour la génomique comparative des populations et l’évolution expérimentale. Ces études, associées à l’utilisation des outils classiques de biologie cellulaire et du développement disponibles pour ce nématode, font de cet organisme l’un des mieux adaptés à l’étude de la plasticité phénotypique. Ainsi, des processus cellulaires et développementaux bien caractérisés peuvent être soumis à des variations de l’environnement ou à des manipulations génétiques, tandis que des protocoles d’évolution expérimentale à long terme, à partir de variation naturelle existante ou de mutation, peuvent être couplés à la cartographie à haute résolution de phénotypes complexes. Les réseaux de gènes et leur dynamique temporelle peuvent être étudiés par la création de modèles informatiques. Cette conférence aux Treilles a rassemblé des spécialistes reconnus de la plasticité phénotypique, de l’évolution des réseaux de gènes, de la structure des génomes, de la génétique des populations, de la cartographie QTL et GWAS, de l’expression génique, de la détermination du sexe et du développement de la lignée germinale, mais aussi du comportement et de l’écologie. Des intervenants venant de différents pays d’Europe ou des Etats-Unis, avec une expertise à la fois théorique et pratique étaient présents. Parmi eux se trouvaient de nombreux jeunes chercheurs ayant récemment créé leurs groupes indépendants.

—

Meeting report

Introduction

Environmental variation modulates organismal development and substantially contributes to phenotypic variation within populations. Understanding how environmental and genotypic variation interact to generate phenotypic variation is thus a central challenge in biology.
At the individual level, any phenotype results through the interactions between genotype and environment, so that attempting to measure relative contributions of genes versus environment to individual phenotypes is meaningless. In contrast, at the population level, given that phenotypic differences can be measured, the relative contributions of genotypic versus environmental variation to phenotypic variation can be measured, albeit not without problems (Lewontin 1974; Fox Keller 2010). Partitioning the phenotypic variance found among individuals into that caused by genetics or environmental variation and their interactions allow estimates of the heritability of the phenotypes, for a given population in a particular environment (Falconer 1989). This quantitative genetics approach can further describe the genetic covariance across environments and therefore the heritability of phenotypic plasticity (Via and Lande 1985; Gomulkiewicz and Kirkpatrick 1992). However, a comprehensive understanding of phenotypic variation also requires the study of development and physiology, since quantitative genetics implicitly assumes linear genotype-phenotype-environment maps. Studies integrating statistical and causal analyses of environmental sources of phenotypic variation are still rare because they can only be adequately performed in organisms where both environment and genotype are amenable to manipulation and control.
During the summer of 2014, the workshop in Les Treilles focused on how the multidisciplinary study of the model nematodes Caenorhabditis elegans and related species can contribute to addressing these fundamental questions in developmental and evolutionary genetics. Henrique Teotonio (Ecole Normale Superieure Paris, France) introduced the workshop with key concepts of phenotypic plasticity and the challenges in measuring the environmental contributions to phenotypic variation; then exemplifying some of the challenges of understanding C. elegans transcriptome experimental evolution under changing environmental conditions. Christian Braendle (Institute of Biology Valrose, France) introduced the workshop with the key advantages of the Caenorhabditis model to study causal interactions between genotype and environment. In the past decade, C. elegans and related nematode species have emerged as powerful model organisms in evolutionary biology, which are particularly apt to address outstanding questions in the field of phenotypic plasticity (Carvalho et al. 2006; Gaertner and Phillips 2010; Gray and Cutter 2013).

C. elegans as a model to study phenotypic plasticity and genotype-by-environment interactions

C. elegans displays a large spectrum of developmental, physiological and behavioural responses to variable environmental conditions and such responses also show ample natural genetic variation (Braendle et al. 2008). From a practical perspective, (naturally) isogenic populations of C. elegans allow easy amplification and control of genetic background and thus facilitate analysis of phenotypic plasticity. C. elegans QTL mapping based on naturally occurring genetic variation has been successfully applied to the study of phenotypic plasticity underlying diverse traits. For example, Kammenga and colleagues have identified quantitative trait loci (QTLs) impacting the thermal plasticity of fertility, age at maturity, growth rate, lifespan and body size using Recombinant Inbred Lines (RILs) between the reference strain N2 and the divergent Hawaiian strain CB4856 (Gutteling et al. 2006; Li et al. 2006; Gutteling et al. 2007a; Gutteling et al. 2007b; Rodriguez et al. 2013). The molecular polymorphisms underlying certain QTLs could be identified, for example, a coding polymorphism in the tra-3 gene modulates body size response to temperature (Kammenga et al. 2007).
Novel standardized high-throughput approaches to identify natural variants of complex traits, such as plastic phenotypes, have recently been developed (Erik Andersen, Northwestern University, USA), which will greatly improve the power for trait mapping based on linkage and GWAS (Andersen et al. 2012). Similarly, the feasibility to characterize epistatic interactions of complex traits was illustrated by the presentation of Patrick McGrath (Georgia Institute of Technology, USA) who is taking advantage of related C. elegans laboratory strains, which have adapted to contrasting lab environments within a few decades (McGrath et al. 2009; Mcgrath et al. 2011). Recently, QTL mapping approaches have been extended to C. briggsae using recombinants derived from crosses between the isolates AF16 (India) and HK104 (Japan) (Koboldt et al. 2010; Ross et al. 2011). In contrast to C. elegans, C. briggsae presents higher levels of genetic diversity and wild isolates cluster into distinct phylogeographic groups, such as temperate and tropical clades (Prasad et al. 2011; Félix et al. 2013). Asher Cutter (University of Toronto, Canada) outlined recent analyses on C. briggsae genetic diversity using whole-genome sequencing and QTL mapping of thermal plasticity of life history and behavioural traits using high-throughput assays. This research indicates that C. briggsae emerges as an ideal model species to dissect the genetic architecture underlying thermal adaptation and plasticity in response to temperature variation. The meeting presentations, together with large body of past work, illustrate the feasibility of high-resolution complex trait mapping in Caenorhabditis nematodes, thus bridging interests in evolutionary and molecular genetics.
A remarkable example of C. elegans phenotypic plasticity is the formation of the alternative dauer stage in response to environmental cues (Cassada and Russell 1975). The propensity to form dauer larvae further shows considerable variation among C. elegans isolates (Viney et al. 2003; Harvey et al. 2008), and quantitative genetic analyses have yielded first insights into the complex polygenic architecture underlying intraspecific variation in dauer formation (Green et al. 2013; Green et al. 2014). Simon Harvey (Canterbury Christ Church University, UK) presented on-going research extending the analysis of natural genetic variation of dauer formation in C. elegans and C. briggsae isolates using mapping based on RILs, NILs and MA lines. Mark Viney (University of Bristol, UK) discussed recent assays quantifying natural genetic variation in dauer formation in response to specific ascarosides (Diaz et al. 2014) – key components of the pheromone controlling dauer induction (Golden and Riddle 1982; Srinivasan et al. 2008; Braendle 2012). Dauer induction in distinct C. elegans wild isolates may greatly differ in response to specific ascarosides and their mixtures; moreover, production of dauer-inducing ascarosides may also vary among isolates (Choe et al. 2012; Diaz et al. 2014). The extensive knowledge of molecular control mechanisms underlying C. elegans dauer formation, now increasingly being analysed in an evolutionary context, will make this system a key model for future studies on adaptive developmental plasticity in animals.
In addition to the well-known examples of plastic phenotypes, such as dauer formation, Caenorhabditis nematodes provide a virtually unlimited source of phenotypes that display both environmental sensitivity and evolutionary variation, both among and within species. Marie Delattre (Ecole Normale Superieure Lyon, France) presented research focusing on the first asymmetric cell division in the embryo, a robust major patterning event, which nevertheless underlies a surprising degree of evolutionary variability in various system parameters (Riche et al. 2013) which are further likely plastic, i.e. environmentally sensitive. Ron Ellis (Rowan University SOM, USA) discussed the potential roles of developmental plasticity in the evolution of sperm activation pathways, which show substantial divergence among the three hermaphroditic species, C. elegans, C. briggsae and C. tropicalis. These analyses involved the use of recently developed genomeediting methods (CRISPR-Cas9, TALENs), which allowed unprecedented targeted gene manipulation across species, thus providing exciting novel insights into the convergent evolution of hermaphroditism in the genus Caenorhabditis. Eric Haag (University of Maryland, USA) presented novel analyses on Caenorhabditis genome structure in the context of mating system evolution. Combining comparisons of gene expression (Thomas et al. 2012) and genome sequence between selfing and outcrossing species, the Haag lab identified candidates regulating mating-induced behavioural plasticity. These results thus shed light on the evolutionary changes in mating behaviour associated with the evolution of Caenorhabditis hermaphroditism.

Stochasticity in developmental systems

A single genotype may express phenotypic variation even in a single, homogenous environment. Such phenotypic variation is usually termed stochastic phenotypic variation, which can be considered as a particular form of environmental variation (given the absence of any underlying genetic variation), the regulation of which may be subject to evolutionary change. The mechanisms and consequences of stochasticity in the expression of phenotypic variation is a key topic in current systems biology and developmental genetics. The recent development of molecular techniques allowing the precise quantification of gene expression, such as single molecule fluorescent in situ hybridization (smFISH) (Raj et al. 2008) have significantly advanced our understanding of stochasticity at the molecular level. Scott Rifkin (University of California San Diego, USA) and collaborators applied the novel smFISH method in C. elegans to study the variability of a well-defined gene expression network underlying intestinal specification (Raj et al. 2010). This study is among the first to link how gene expression stochasticity relates to incomplete penetrance of mutations, and thus provides unprecedented insights into the phenomenon of developmental robustness. C. elegans further displays examples of stochastic cellular events, most prominently, the division of a particular hypodermal cell, P3.p (Sulston and Horvitz 1977). The propensity of this vulval precursor cell to either divide or fuse with the hypodermis is stochastic, but also strongly modified by mutation, genetic background and environmental conditions (Delattre and Félix 2001; Braendle and Félix 2008; Braendle et al. 2010). Marie-Anne Félix (Ecole Normale Superieure Paris, France) presented novel studies aimed at identifying genetic variation modulating the stochasticity of this cell fate patterning event. These examples illustrate the feasibility to connect molecular and evolutionary study of variation in developmental stochasticity.

Experimental evolution to study phenotypic plasticity

In the last 10 years or so, Caenorhabditis species have become major models in evolutionary biology through their use in evolution experiments (Gray and Cutter 2013). For example, such experiments have addressed the evolution of life-histories (Anderson et al. 2011; Carvalho et al. 2014), coevolution with pathogens (Morran et al. 2009; Schulte et al. 2010; Morran et al. 2011; Masri et al. 2013; Morran et al. 2014), the origin of self-fertilization (Theologidis et al. 2014), sexual selection (LaMunyon et al. 2007; Theologidis et al. 2014) or the population genetic consequences of mating system variation (Chelo et al. 2013a; Chelo et al. 2013b).
The use of evolution experiments to study phenotypic plasticity remains relatively unexplored. One exception is the recent report of genetic assimilation of survival to an acute heat shock in populations of C. remanei selected for larval resistance to high temperatures by shifting the temperature reaction norms (Sikkink et al. 2014b); a result that demonstrates rapid evolution of phenotypic plasticity and that vindicates some of the classic results by C.H. Waddington in Drosophila (Waddington 1953). Another form of phenotypic plasticity, transgenerational effects, was also shown to evolve in the same C. remanei populations: increased robustness to temperature variation in adults resulted from selection on larval heat shock survival (Sikkink et al. 2014a).
At the workshop, Steve Proulx (University of California Santa Barbara, USA) reviewed the population genetics theory about adaptation to fluctuating environments, emphasizing that for adaptation to occur the geometric mean fitness of populations across environments should increase. He then showed preliminary results consistent with this idea. Evolution experiments in populations of C. elegans with standing genetic diversity adaptation to a fluctuating anoxic larval environment appears to have occurred through the evolution of transgenerational effects only when parent-offspring environmental change was predictable. Using the same basic set of C. elegans populations, Ivo Chelo (Instituto Gulbenkian de Ciencia, Portugal) showed results on experimental evolution to high NaCl concentrations and discussed how to best describe individual fertility distributions across environments and the effects of selfing and outcrossing on those distributions.
Evolution experiments in Caenorhabditis have also been used to study the properties of mutational distributions across environments. Charlie Baer (Florida State University, USA) questioned the notion that increased environmental stress elevates mutation rates. In one mutation accumulation experiment temperature lead to increased deleterious mutation rates in a C. elegans but not in C. briggsae (Matsuba et al. 2012). In another C. elegans mutation accumulation experiment, deleterious mutation rates were indistinguishable under elevated endogenous oxidative stress from natural oxidative stress (Joyner-Matos et al. 2011). The environment thus conditions the appearance of novel genetic diversity by mutation, but this conditioning is dependent on the particular genotype studied.

The dark side: life cycle and ecology of natural Caenorhabditis populations

Our current understanding of Caenorhabditis biology is greatly hampered by the poor knowledge of ecology and life history of these nematodes in the wild. Although the study of natural Caenorhabditis populations has become increasingly studied over the past decade (Barrière and Félix 2005; Haber et al. 2005; Barrière and Félix 2007; Félix and Braendle 2010; Andersen et al. 2012; Félix and Duveau 2012; Petersen et al. 2014), we still lack basic information on life cycle, generation time, dispersal, population dynamics and structure as well as on key environmental factors underlying population and species distributions. Identifying ecologically relevant environmental conditions to study phenotypic plasticity and its relevance for fitness thus remains difficult. Prioritizing efforts to study elementary Caenorhabditis ecology in the wild – an arduous and time-consuming endeavour – emerged as a major task for the future. Integrating the ecological and population-genetic study of Caenorhabditis species other than C. elegans should provide particularly useful by allowing comparative and quantitative analyses of species differences in niche specialization and species-specific traits, such as molecular diversity and reproductive mode.
Matt Rockman (New York University, USA) presented new field data on metapopulation structure of tropical Caenorhabditis nematodes and discussed how to estimate fitness of gonochoristic species, emphasizing the need to take into account demographic parameters. Karin Kiontke (New York University, USA) provided an in-depth analysis of phenotypic diversity and character evolution in the rapidly expanding Caenorhabditis phylogeny with over 25 cultured species (Kiontke et al. 2011; Félix et al. 2014), which are available at the Caenorhabditis Genetics Center (http://www.cgc.cbs.umn.edu). In contrast to C. elegans, certain species display highly specialized interactions with invertebrate carriers, e.g. in C. japonica, which lives in close association with the shield bug, Parastrachia japonensis (Kiontke 2002)(Tanaka et al. 2010; Tanaka et al. 2012; Okumura et al. 2013; Yoshiga et al. 2013), thus greatly facilitating ecological and life cycle analysis. Integrating available information on niche specialists versus generalists with genetic diversity estimates across Caenorhabditis species (Li et al. 2014) is one particularly promising avenue of research to better understand Caenorhabditis ecology.
C. elegans and its relatives are microbivorous, feeding on a wide range of bacteria and fungi, which greatly affect growth and development and further represent potential pathogens (Félix and Braendle 2010). Analysis of the microbial fauna associated with Caenorhabditis is thus undoubtedly key to obtain a basic ecological understanding of this nematode group. The evolutionary ecological study of natural host-pathogen interactions in C. elegans has recently allowed the first identification of natural viruses in C. elegans and C. briggsae (Félix et al. 2011). Marie-Anne Félix discussed follow-up research leading to identification of molecular variants underlying evolutionary variation in viral sensitivity (Ashe et al. 2013). Given the strong selective pressures exerted by microbial pathogens and
parasites, future study of their interactions with Caenorhabditis will greatly benefit the evolutionary ecological understanding of these nematodes.

Perspectives

Presentations and discussions during the workshop at Les Treilles allowed for an integrative overview of how Caenorhabditis nematodes can be used to address specific conceptual and empirical questions in the field of phenotypic plasticity and its evolutionary significance. Without doubt, C. elegans and its relatives will continue to provide exceptionally powerful model organisms to integrate fundamental questions of genotypephenotype mapping at the interface of molecular and evolutionary biology.

Acknowledgements

This conference was made possible by the generous support of the Fondation des Treilles, Tourtour, France. We would like to thank all staff at Les Treilles for their great efforts in making this workshop an exceptional experience.