Experimental disruption of social structure reveals totipotency in the orchid bee, Euglossa dilemma

Eusociality has evolved multiple times across the insect phylogeny. Social insects with greater levels of social complexity tend to exhibit specialized castes with low levels of individual phenotypic plasticity. In contrast, species with simple social groups may consist of totipotent individuals that transition among behavioral and reproductive states. However, recent work has shown that in simple social groups, there can still be constraint on individual plasticity, caused by differences in maternal nourishment or social interaction. It is not well understood how these constraints arise, ultimately leading to the evolution of nonreproductive workers. Some species of orchid bees form social groups of a dominant and-one to two subordinate helpers where all individuals are reproductive. Females can also disperse to start their own nest as a solitary foundress, which includes a nonreproductive phase characterized by ovary inactivation, not typically expressed by subordinates. Little is known about individual flexibility across these trajectories. Here, using the orchid bee Euglossa dilemma, we assess the plasticity of subordinate helpers, finding that they are capable of the same behavioral, physiological, transcriptomic, and chemical changes seen in foundresses. Our results suggest that the lack of nonreproductive workers in E. dilemma is not due to a lack of subordinate plasticity.

Gene-regulatory context of honey bee worker sterility

The highly organized societies of the Western honey bee Apis mellifera feature a highly reproductive queen at the center of attention and a large cohort of daughters that suppress their own reproduction to help rear more sisters, some of whom become queens themselves. This reproductive altruism is peculiar because in theory it evolves via indirect selection on genes for altruism that are expressed in the sterile workers but not in the reproductive queens. In this study we attempt to situate lists of genes previously implicated in queenright worker sterility into a broader regulatory framework. To do so we use a model bee brain transcriptional regulatory network as a template to infer how sets of genes responsive to ovary-suppressing queen pheromone are functionally interconnected over the model's topology. We predict that genes jointly involved in the regulation of worker sterility should be tightly networked, relative to genes whose functions are unrelated to each other. We find that sets of mapped genes - ranging in size from 17 to 250 - are well dispersed across the network's substructural scaffolds, suggesting that ovary de-activation involves genes that reside within more than one transcriptional regulatory module. For some sets, however, this dispersion is biased into certain areas of the network's substructure. Our analysis identifies the regions enriched for sterility genes and likewise identifies local hub genes that are presumably critical to subnetwork function. Our work offers a glimpse into the gene regulatory context of honey bee worker sterility and uses this context to identify new candidate gene targets for functional analysis. Finally, to the extent that any sterility-related modules identified here have evolved via selection for worker altruism, we can assume that this selection was indirect and of the type specifically invoked by inclusive fitness theory.

Division of labor in honey bees is associated with transcriptional regulatory plasticity in the brain

Studies in evolutionary and developmental biology show that relationships between transcription factors (TFs) and their target genes can be altered to result in novel regulatory relationships that generate phenotypic plasticity. We hypothesized that context-dependent shifts in the nervous system associated with behavior may also be linked to changes in TF-target relationships over physiological time scales. We tested this hypothesis using honey bee (Apis mellifera) division of labor as a model system by performing bioinformatic analyses of previously published brain transcriptomic profiles together with new RNAi and behavioral experiments. The bioinformatic analyses identified five TFs that exhibited strong signatures of regulatory plasticity as a function of division of labor. RNAi targeting of one of these TFs (broad complex) and a related TF that did not exhibit plasticity (fushi tarazu transcription factor 1) was administered in conjunction with automated analyses of foraging behavior in the field, laboratory assays of aggression and brood care behavior, and endocrine treatments. The results showed that changes in the regulatory relationships of these TFs were associated with behavioral state, social context and endocrine state. These findings provide the first empirical evidence that TF-target relationships in the brain are altered in conjunction with behavior and social context. They also suggest that one mechanism for this plasticity involves pleiotropic TFs high up in regulatory hierarchies producing behavior-specific transcriptional responses by activating different downstream TFs to induce discrete context-dependent transcriptional cascades. These findings provide new insights into the dynamic nature of the transcriptional regulatory architecture underlying behavior in the brain.