Structure and function of gene regulatory networks associated with worker sterility in honeybees

Abstract

A characteristic of eusocial bees is a reproductive division of labor in which one or a few queens monopolize reproduction, while her worker daughters take on reproductively altruistic roles within the colony. The evolution of worker reproductive altruism involves indirect selection for the coordinated expression of genes that regulate personal reproduction, but evidence for this type of selection remains elusive. In this study, we tested whether genes coexpressed under queen-induced worker sterility show evidence of adaptive organization within a model brain transcriptional regulatory network (TRN). If so, this structured pattern would imply that indirect selection on nonreproductive workers has influenced the functional organization of genes within the network, specifically to regulate the expression of sterility. We found that literature-curated sets of candidate genes for sterility, ranging in size from 18 to 267, show strong evidence of clustering within the three-dimensional space of the TRN. This finding suggests that our candidate sets of genes for sterility form functional modules within the living bee brain's TRN. Moreover, these same gene sets colocate to a single, albeit large, region of the TRN's topology. This spatially organized and convergent pattern contrasts with a null expectation for functionally unrelated genes to be haphazardly distributed throughout the network. Our meta-genomic analysis therefore provides first evidence for a truly "social transcriptome" that may regulate the conditional expression of honeybee worker sterility.

Keywords: Gene regulation; RNA profiling; indirect selection; insect sociobiology; meta‐analysis; microarrays; social behavior.

Figure 1

Workers on pupae and pollen. In the presence of their queen mother, members of this all‐female caste deactivate their ovaries and adopt alloparental roles within the colony. In the queenright condition, workers are essentially sterile. Kin theory predicts that “genes for sterility” evolve via reproducing relatives who carry, but clearly do not express, these genes. (Photograph: Emma K Mullen, Cornell University).

Figure 2

The reconstructed honeybee transcriptional regulatory network (TRN). The network contains 2382 nodes representing transcription factors and their putative target genes, and 6757 edges representing regulatory interactions. Colors denote the eight functional clusters that we infer through model fitting.

Figure 3

Sterility genes mapped onto the eight clusters that best describe substructure of the honeybee brain transcriptional regulatory network. The clusters are arranged from largest to smallest: Cluster 1, 433 genes; Cluster 2, 384 genes; Cluster 3, 361 genes; Cluster 4, 291 genes; Cluster 5, 281 genes; Cluster 6, 234 genes; Cluster 7, 199 genes; Cluster 8, 197 genes (Gene lists associated with each cluster are available as Appendix S3). For demonstration purposes we show “sterility genes” mapped as blue nodes from (A) Cardoen et al. (2011; n = 267 genes) and (B) Mullen et al. (2014, n = 18 genes). Of the four gene sets tested, these two represent the largest and smallest that are statistically biased toward Cluster 3, which is shown by the * symbol.