A key question in biology is how novel biological complexities evolve, and an important example of this is the development of societies, when individuals within a species cooperate to function as a whole. Wasps are social insects and represent an attractive system for studying the genetic basis of such social behaviors. Seirian Sumner and colleagues in her lab at the University of Bristol investigate how the interactions between genes, the environment and behavior have led to the evolution of sociality in wasps. In a recent study in Genome Biology, they describe an RNA-seq analysis of the genetic mechanisms controlling alternative phenotypes in the eusocial wasp Polistes canadensis. We asked her to explain how the study came about and what were the main findings.
Could you give a brief overview of the background to your research?
I am interested in the evolution of social behavior. The success of animal societies lies in the emergence of division of labour, where individual group members specialize in different tasks. The epitome of social behavior is in the social insects (ants, some bees and wasps) where some group members specialize in reproduction – the queens, and others in maternal care – the workers. In many social insects queens and workers differ so much in appearance and behaviour that they could be mistaken as belonging to different species. Yet, these two phenotypes arise from the same genetic material – a shared genome. Queens and workers are examples of how the same genome can give rise to highly contrasting, specialized phenotypes.
Mechanistically, shared genes expressed at different levels in response to specific environmental cues at specific times in an individual’s life, give rise to distinct alternative phenotypes. On an evolutionary level, sociality is all about the emergence of commitment (or specialization) into reproductive or non-reproductive roles. During social evolution, the level of commitment – the nature and limits of phenotypic plasticity – changes. This makes social insects, as a group, excellent models for studying phenotypic plasticity as well as social behavior.
Why is studying alternative phenotypes is such an important question?
How and why social phenotypes arise, the nature of their phenotypes and their roles in societies are fundamental questions about one of the major transitions in evolution – the emergence and proliferation of societies. Whilst we understand a great deal about the complex lives of a few well studied highly eusocial species, like the honeybee and fire ant, our knowledge of the processes and machinery underlying the origins of social phenotypes is relatively unexplored. I am interested in the origin of sociality, and I use a combination of classical behavioural ecology field methods with gene-level analyses to examine how and why alternative phenotypes first arise.
How has your work changed our understanding of the genetic mechanisms controlling alternative phenotypes?
A long-standing hypothesis on the evolution of alternative phenotypes is that evolution makes do with what it inherited from its ancestor. In the case of social animals, the ancestor was a solitary insect, whose phenotypic traits associated with reproduction and maternal care become uncoupled into queen and worker specialists, respectively. Until now, assessment of this mechanism has been restricted to behavioural assays – which can be subjective. For the first time, we were able to quantify the process by which the genes underlying the early stage of queen and worker evolution are uncoupled, by looking across the whole genome at the genes expressed in each caste.
We sequenced everything expressed in the brains of queen and worker paper wasps, meaning we made no a priori selection of genes to examine. In this respect, our study differed from the many excellent studies on gene expression in the honeybee and other eusocial insects, where thousands of pre-selected genes were examined through microarray analyses. We stumbled across a rather unexpected result. We found a consistent asymmetry in the numbers of genes up-regulated in workers relative to queens: of the 2442 differentially expressed genes between these two social phenotypes, almost all of these (91%) were upregulated in workers!
How do you think the asymmetric expression of novel genes relates to alternative phenotypes?
Future studies will need to address what this means, but one idea is that this asymmetry reflects the relative specialization of the castes. Specifically, worker paper wasps perform almost all behaviours except for egg-laying: they forage for and feed the brood, they build and defend the nest, they mate, even when they have little chance of becoming a queen. Conversely, queens are specialized egg-layers who rarely leave the nest. The vast up-regulation of the genes putatively important in making phenotypes different may reflect the differences in the behavioural repertoires of queen and worker paper wasps. In hindsight, this hypothesis makes a lot of sense. But without ‘transcriptome fishing’, we may never have discovered this.
Could you explain how your work has informed the debate about the evolution of alternative phenotypes?
Genomic studies in social insects to date, and indeed phenotypic plasticity, has focused on the evo-devo hypothesis that a conserved suite of specific genes are responsible for generating similar phenotypes across independent origins (e.g. ants versus wasps or bees), and different levels of sociality (e.g. different degrees of ‘commitment’ by queens and workers). Indeed, there is strong support for a small number of ‘toolkit’ genes being differentially expressed with respect to castes in social insects, and the ‘candidate-gene’ approach to studying the molecular basis of alternative phenotypes has been highly successful in revealing these patterns.
A less popular idea is that genomic novelty is important in the origin and proliferation of alternative phenotypes. Genomic novelty includes genes that are restricted to a specific lineage, genus, or even species. How such novelty arises is still very much in debate (and not the topic of our study). We stumbled across an unexpected discovery – two thirds of the genes that putatively make queens different from workers were novel. These genes had not been sequenced ever before. They were entirely absent from the genomes of key insect model species. Even more surprising, was that 90% of these novel genes were upregulated in workers! After a bit of head scratching and delving back into the literature, we realised that genomic novelty and phenotypic novelty appear to go hand in hand in other organisms, like reptiles, and even the humble Hydra. Our findings therefore add to the emerging importance of novel genes in phenotypic evolution.
Why is Polistes canadensis such a good organism for studying phenotypic plasticity?
Polistes have been a model genus for understanding phenotypic plasticity for over a century. They display the same fascinating behaviours that the honeybee does, whereby individuals specialize in reproductive (queen) or maternal care (worker) behaviours. But, unlike the honeybee and most ants, they retain the ability to switch castes throughout adulthood. Such plasticity is a hallmark of the early stages of social evolution, where reproductive and maternal care behaviours show the first signs of uncoupling the phenotype of their solitary ancestor into two distinct (but flexible) phenotypes.
Polistes are also an excellent species for experimental work because each individual can be identified, monitored and observed throughout their life. For generations, scientists have been gripped by the mini soap-opera that unfolds in a Polistes colony: there is disagreement, conflict, coercion as well as cooperation, when individuals negotiate the terms by which they pass on their genes to the next generation. We can play ‘god’ with their societies, removing some individuals to instigate behavioural changes in other group members: promotions and demotions in the social hierarchy. Combined with genomic analyses, such in vivo manipulations are a powerful tool for understanding the mechanisms underlying phenotypic plasticity and its role in social evolution.
Has using next generation sequencing technologies allowed you to address questions that conventional genetics would not have?
Absolutely. Next Generation Sequencing (NGS) is causing a major revolution in the way we approach ecology and evolution. Conducting any gene-level analyses on species which are not a typical genetic model organism has been next to impossible until the last couple of years. Hypotheses on the molecular basis of alternative phenotypes, and social evolution, have been kicking around for a while. We are now for the first time able to address these questions. NGS provides a quick, affordable, unbiased and accessible way of looking at the genes, across entire genomes of non-model organisms for ecologically meaningful sample sizes.
In fact, I had tried to address these same questions in P. canadensis using ‘conventional’ methods in 2006. After many painful hours in the lab, my co-authors and I managed to identify a measly 42 genes that were differentially expressed between castes in P. canadensis (Proc B 2006, 273:19-26). There is no doubt that NGS is changing the scale at which we can address these questions and the range of organisms we can apply them to. Moreover the ease of acquiring the data is shifting the focus from hypothesis-driven to data-driven research. These are exciting times for biologists, and we need to be open-minded about what lies around the corner.
What will the future directions of your work be?
Our study is the tip of the ice burg. Next, we will be trying to tease out the extent of ‘within caste’ variation, to understand the limits on plasticity for queens and workers, the environmental factors responsible for them, and the mechanisms underlying them (e.g. epigenetics).
Our findings posit a number of new hypotheses about the origins of alternative phenotypes. For example, is upregulation of novel genes in workers typical of primitively eusocial species, or indeed in the primitive state of other organisms with diverse phenotypic proliferation? The next step is to determine how generally applicable they are. Since so little data is available for non-model organisms, especially for wasps and primitively eusocial insects, we have a lot of work to do to answer this question.