Evolution and molecular basis of caste differentiation in bees

Lead Research Organisation: University of East Anglia
Department Name: Biological Sciences


A fundamental topic in biology concerns how the switching on and off (differential expression) of genes creates morphological and behavioural diversity. 'Polyphenism' (when different adult morphs arise from the same genome) provides systems in which to address how differential gene expression creates diversity at the level of whole organisms. A major example of polyphenism occurs in the ants, bees and wasps, in which female larvae develop into one of two castes, queens or workers. It is already known that the huge differences between the castes in size, shape, reproduction and behaviour stem from differential gene expression. Hence the process by which castes develop (caste differentiation) allows us to investigate how differential gene expression generates polyphenism. However, our understanding of the genetic basis of caste differentiation is in its early stages. The genes involved, their evolutionary background and the role of microRNAs (miRNAs, small RNA molecules important in regulating gene expression) are all poorly known.

In this project we will use bumble bees (Bombus) to elucidate the evolution and regulation of the genes influencing queen-worker caste differentiation. Bumble bees represent excellent models for this purpose. First, species vary in the number of critical stages of larval caste determination (when differentiation is initiated) and the evolutionary history of this variation is known. One species (B. terrestris) has two critical stages in female larvae (early- and late-instar), with the early critical stage being secondary (evolutionarily derived); another (B. hypnorum) has one such stage (late-instar larvae only); and 'cuckoo' bumble bees such as B. vestalis have no critical stages, as they have secondarily lost the worker caste and all females are queens that reproduce in nests of other Bombus species. Second, we have discovered that a miRNA, miR-6001-5p, is overexpressed in queen-destined late-instar B. terrestris larvae, suggesting that it regulates genes influencing caste.

These data allow us to create a predictive framework for expected patterns of differential expression in caste-associated genes and miR-6001-5p in Bombus. Specifically, if the critical stages are conserved in evolution via regulation of the same genes, and new critical stages are gained via additional regulatory pathways, B. terrestris and B. hypnorum should share common patterns of differential gene expression among late- but not among early-instar larvae. Likewise, patterns of gene expression specific to worker-destined larvae in these two species should be absent in the cuckoo species B. vestalis. Furthermore, if the role of miR-6001-5p is evolutionarily conserved, similar patterns of expression of miR-6001-5p and its target genes should occur across queen-destined late-instar larvae of all three Bombus species.

We will test these predictions by comparing profiles of differentially expressed genes across larval phenotypes, by using our expression data to predict targets of miR-6001-5p and by conducting assays to validate these targets. The proposed work will provide a comprehensive understanding of the evolutionary and genetic basis of changes in the number of critical stages of caste determination in eusocial insects and will elucidate the role of a miRNA in caste differentiation. As well as addressing the fundamental topic of the genetic regulation of polyphenism, the work is novel because no previous study has exploited a comparable predictive framework. It is timely because gene profiling by next-generation sequencing now provides the tools required and our existing data have created an exciting opportunity to make rapid headway in understanding how miRNAs affect caste differentiation. Finally the work fits BBSRC's strategic priorities because it could reveal means of manipulating worker production to maximize pollination services to horticulture provided by commercially-reared bumble bee colonies.

Technical Summary

We will investigate the molecular and functional basis of evolutionary changes in the number of critical stages in larval caste determination in three focal bumble bee (Bombus) species, including the role of miR-6001-5p, a miRNA that we have discovered to be upregulated in queen-destined late-instar larvae of B. terrestris. This will be done in a predictive framework based on our knowledge of interspecific caste differences and of miR-6001-5p.

B. terrestris queen- and worker-destined early- and late-instar larvae will be obtained by sampling larvae of appropriate size classes following and preceding queen removal, respectively, since queen removal takes away the queen's inhibition of queen differentiation. Control larvae will be left within colonies to verify caste fates. B. hypnorum queen- and worker-destined late-instar larvae will be obtained by sampling larvae from late-stage colonies, and early-instar larvae developing into queens or workers will be obtained by sampling small larvae and again verifying caste fates of control larvae. B. vestalis female larvae will be obtained by sampling from colonies hosted within B. terrestris colonies.

Larval transcriptomes of all three species will be profiled by RNA-Seq. In B. terrestris, we will use qRT-PCR to validate the most salient differentially expressed genes. Using our predictive framework, we will compare larval transcriptomes, and use qRT-PCR of selected genes, to test for expected similarities and differences in gene expression profiles across stages, castes and species.

We will use prediction software, genome annotations and our B. terrestris expression data to identify putative targets of miR-6001-5p and the luciferase reporter system to validate them. We will then use Northern Blots and qRT-PCR to test for corresponding patterns of expression of miR-6001-5p and its targets in queen-destined late-instar larvae of all three species, so elucidating its role in caste differentiation.

Planned Impact

As well as having a substantial academic impact, this research will potentially benefit four other groups, namely commercial bumble-bee rearing companies, bee-related NGOs, policy-makers and practitioners in government departments covering biodiversity and agriculture, and the general public.

Many companies worldwide mass-rear bumble bee colonies for the pollination of glasshouse crops. These companies would potentially benefit from the project because, given worker bees make the most efficient and frequent pollinators, an improved understanding of the molecular basis of worker production in bumble bee colonies might help maximize the effectiveness of pollination by commercially-reared colonies. The PI is committed to dialogue with the bee-rearing companies via existing projects and will seek further dialogue to promote relevant results of the new research.

Several organizations, including the Bumblebee Conservation Trust, the British Beekeepers Association, the Bee Guardian Foundation and Buglife, are active in conducting and/or promoting the understanding and conservation of bees. The project would benefit from their aid in disseminating results and impact and they would benefit from having, if they chose, a science-based bee-related 'story' to publicize. The project team will send team members to the Bumblebee Working Group (BWG) meeting, a forum for UK bumble bee conservation held every 1-2 years, which is regularly attended by representatives of relevant NGOs. The PI regularly attends the BWG and has previously hosted it. Existing contacts with the NGOs will be used for direct communication outside the BWG.

Policy-makers and practitioners in government departments covering biodiversity and agriculture have an interest in understanding the biology of insect pollinators in its broadest sense. This group might benefit from the evidence base that the project will provide as regards bumble bee biology. The PI will seek opportunities to present relevant results to policy-makers and practitioners, building on contacts made at two recent workshops (UK government's Science and Innovation Network workshop of experts on 'Insect Pollinators: From Research to Policy', London, February 2012; Insect Pollinators Initiative midterm workshop, York, November 2012). The forthcoming Insect Pollinators Initiative wrap-up stakeholder event in London in October 2014, which (as a PI on an IPI project) the PI will attend, will provide another prime opportunity for such interactions. The PI has a record of seeking to maximize the impacts of his research, for example through applying genetic census techniques that he and colleagues have developed for wild bumble bee populations to projects investigating the population biology of declining bumble bee species or testing the effectiveness of agri-environment schemes for bumble bees in the UK.

Bumble bees are immensely popular organisms with the public, and there is widespread affection for them and appreciation of their pollinating role. In addition, ongoing publicity over declines of honey bees and wild bees has increased awareness of the need to conserve bee populations. The public would benefit from learning the results of the project to satisfy the clear appetite that therefore exists for scientific information on the biology of bees. The project team will engage the public with the results of the research via press releases, dedicated project webpages and events such as exhibitions and public talks. The PI has wide experience of dealing with the media and a demonstrable commitment to interacting with the public in the manner proposed.


10 25 50
Description We used RNA-seq and qRT-PCR to elucidate the evolution and molecular basis of caste differentiation in queen- versus worker-destined female larvae in bumble bees (Bombus spp.), in order to advance our knowledge of the genetic regulation of polyphenism. Our main findings are as follows:

1. Genes underpinning larval caste differentiation in B. terrestris: We used RNA-seq to profile the transcriptomes of queen- and worker-destined larvae at different instars. Across 17 replicated libraries, we found 14,817 expressed genes. Consistent with caste determination occurring between the early- and mid instars, mid-instar larvae showed the greatest number of differentially expressed genes (DEGs) between caste pathways. In 17 comparisons across 14 selected genes in which RNA-seq predicted differential gene expression, the same pattern was found by qRT-PCR. We also isolated >6,500 orthologues between B. terrestris (our data) and two previous studies of caste-associated DEGs in the honey bee Apis mellifera. We found that, while the majority of caste-associated orthologues did not overlap in both species, some did, including the genes Hexamerin and Kruppel-homolog 1.

2. Transcriptomic basis of loss of the worker caste in B. vestalis: We used RNA-seq to compare expression profiles of female larvae of B. terrestris with those of the socially parasitic 'cuckoo' bumble bee B. vestalis, which has secondarily lost its worker caste. We predicted that B. vestalis females would most strongly resemble B. terrestris queens in expression profile. However, having created transcriptomes for each species with >6,400 shared genes, we found that, for genes changing expression during larval development, B. vestalis differed from both B. terrestris queen- and worker-destined larvae and did not more closely resemble either caste pathway.

This is the first study to have isolated and validated genes whose differential expression is associated with larval caste determination in bumble bees. It is also the second such study in any bee and indeed in any other eusocial Hymenopteran. We showed that overlap with corresponding genes in the honey bee is low but present. We also tested for the first time the hypothesis that a socially parasitic species has silenced the genes associated with the worker-destined larval caste pathway. However, we found that, unexpectedly, caste loss in B. vestalis is associated with completely divergent caste gene evolution. Overall, our findings, which are currently being prepared for publication, have added considerably to our fundamental understanding of the molecular basis of caste differentiation, caste evolution and polyphenism.
Exploitation Route Many companies worldwide mass-rear colonies of Bombus terrestris and, increasingly, other Bombus species, for sale as pollinators of commercial glasshouse crops. Given worker bees make the most efficient and frequent pollinators, an improved understanding of the molecular basis of worker production in bumble bee colonies might help maximize the effectiveness of pollination by commercially-reared colonies. By isolating genes associated with queen-worker caste determination in B. terrestris, our work provides a step along the path to manipulating colonies in order achieve this.
Sectors Agriculture, Food and Drink