Discussion
Our results demonstrate that winter bees and nurses have similar transcriptional profiles in fat body tissues relative to foragers, while winter bees and foragers have similar transcriptional profiles relative to nurses. Thus, there are clearly tissue-specific expression patterns associated with the winter bee phenotype. These patterns are consistent with the differential functions of these tissues in bees. Winter bees thus represent a “mix and match” phenotype between summer nurse bees and summer forager bees, with winter fat bodies serving to store nutrition (as is the case of nurse bees) and winter flight muscles remaining active (as in forager bees) to generate heat. Our analysis of biological processes demonstrated that genes differentially expressed within both tissues corresponded to categories associated with the diapause phenotype in other insect species.
We found that there were twice as many differentially expressed genes in flight muscle tissue than in fat body tissues (Table 1). While many studies have evaluated differences between nurses and foragers in fat body tissue (Ament, S. A. et al., 2012; Khamis, A. M. et al., 2015; Seehuus, S-C. et al., 2013), few have evaluated flight muscle tissue. Interestingly, studies of flight muscle tissue report large changes in muscle performance or physiology occurring in young, nurse-age, worker bees, which are more related to maturation processes than the transition from nursing to forager (Harrison, J., 1986; Herold, R., Borei, H., 1963; Schippers, M-P. et al., 2010). This would suggest that few genes would vary between older nurses and foragers, which were used in our study. However, we found hundreds of differentially expressed genes in the flight muscle of nurses, foragers and winter bees. In our samples, bees were actively performing the behaviors (foragers were collected as they returned to the hive, and winter bees were collected from thermoregulating clusters) and thus expression patterns may be showing differences between active versus quiescent muscle tissue. However, overall transcriptional profiles of muscle tissue have not been broadly investigated, and future studies should evaluate to what extent there are “baseline” differences in expression patterns in flight muscle tissue in these different phenotypes, which may prime the bees for different levels of activity.
Based on total numbers of DEGs and clustering analyses, our results suggest that winter bees’ transcriptional profiles in the fat bodies that support the long-lived winter physiological state, while transcriptional profiles in the flight tissues support (or are reflective of) high levels of activity. Thus, winter bees appear to be “mixing and matching” the gene expression profiles and underlying physiological processes of nurse and forager bees, in a tissue-specific way. These results highlight the importance of considering tissue-specific expression patterns when evaluating processes involved in diapause. It will be valuable to determine how this tissue-specific phenotypic plasticity is controlled, both at the molecular and behavioral level. Nurse-like transcription profiles in the fat body tissue may be a result of reduced exposure to brood pheromone (since the there are no developing larvae in the colony) and potentially the presence of older forager-age bees which release ethyl oleate. Brood pheromone can accelerate the transition from nursing to foraging, reduce Vitellogenin levels and reduce longevity in worker bees (Amdam, G. V. et al., 2009). Exposure to ethyl oleate can slow the behavioral maturation from nursing to forager (Leoncini, I. et al., 2004). Forager-like transcriptional profiles in the muscle tissue may be the result of muscle activity, but it remains to be determined what neurophysiological mechanism triggers the generation of muscle activity in these winter bees.
When comparing the GO categories identified in our study with those from previous studies, we found differential regulation of genes associated with dendrite morphogenesis (GO:0048813) neurogenesis (GO:0022008) and morphogenesis of an epithelium (GO:0002009). Several of the differentially regulated dendrite morphogenesis-related genes participate in mTOR and TGF-beta signaling. Interestingly, dendrite morphological restructuring, driven in part by TGF-beta signaling, underlies a related occurrence of phenotypic plasticity in C. elegans : the Dauer phenotype. The Dauer larva is a stress-induced developmental stage wherein several tissues (particularly nervous tissue) undergo remodeling, producing a stress-resistant and long-lived alternative developmental phenotype, akin to the diapause state (Androwski, A. et al., 2017). Molecular analyses have established the Dauer stage to be driven by genes encoding components of an insulin-related pathway, a cyclic nucleotide pathway, and a TGF-beta-related pathway. DAF-7 – Dauer larva development regulatory growth factor 7 – is the TGF-beta-related ligand for the Dauer pathway. The DAF-7 signal is transduced by the DAF-1 and DAF-4 TGF-beta-family type 1 receptors as well as several SMAD-family transcription factors, including DAF-8. The honey bee orthologs of genes encoding DAF-1, DAF-4, and DAF-8 are TGF-beta receptor type 1 [LOC550930 ], activin receptor type 2A [LOC412471 ], and mothers against decapentaplegic homolog 3 [LOC412601 ], which were differentially regulated between winter bees and foragers in fat bodies. Similarly, the Dauer phenotype has been shown to promote neurogenesis of mechanosensory neurons that were experimentally damaged in C. elegans (Caneo, M. et al., 2019). Interestingly, studies in the vertebrate African turquoise killifish (Nothobranchius furzeri ) (Chi-Kuo H. et al., 2020) and invertebrate cotton bollworm (Helicoverpa armigera ) (Yu-Xuan L. et al., 2013), species which both exhibit diapause phenotypes, reveal a role for neurogenesis-associated Polycomb repressive complex members such as histone-lysine N-methyltransferase E(z) [LOC552235 ] (which was upregulated in winter bees and in foragers compared to nurses in the flight muscles) in diapause maintenance. Thus, conserved signaling and regulatory pathways might underly stress-induced long-lived developmental phenotypes across taxa.
Overall, our results indicate that honey bees exhibit tissue-specific transcriptional profiles associated with diapause as an adaptation to different seasonal conditions. Differential regulation of genes associated with dendrite morphogenesis and neurogenesis, including members of the TGF-beta signaling pathway and the Polycomb repressive complex suggest that conserved molecular pathways may underlie stress-induced long-lived developmental phenotypes across taxa. These studies lay the groundwork for future evaluations of the mechanisms, evolution, and consequences of these interrelated phenomena.