A key characteristic of eusocial species is reproductive division of labour. Honey bee colonies typically have a single reproductive queen and thousands of sterile workers. Adult queens differ dramatically from workers in anatomy, physiology, behaviour and lifespan. Young female workers can activate their ovaries and initiate egg laying; these 'reproductive' workers differ from sterile workers in anatomy, physiology, and behaviour. These differences, however, are on a much smaller scale than those observed between the queen and worker castes. Here, we use microarrays to monitor expression patterns of several thousand genes in the brains of same-aged virgin queens, sterile workers, and reproductive workers. We found large differences in expression between queens and both worker groups (~2000 genes), and much smaller differences between sterile and reproductive workers (221 genes). The expression patterns of these 221 genes in reproductive workers are more queen-like, and may represent a core group of genes associated with reproductive physiology. Furthermore, queens and reproductive workers preferentially up-regulate genes associated with the nurse bee behavioural state, which supports the hypothesis of an evolutionary link between worker division of labour and molecular pathways related to reproduction. Finally, several functional groups of genes associated with longevity in other species are significantly up-regulated in queens. Identifying the genes that underlie the differences between queens, sterile workers, and reproductive workers will allow us to begin to characterize the molecular mechanisms underlying the evolution of social behaviour and large-scale remodelling of gene networks associated with polyphenisms.

We report results that address a long-standing controversy in honey bee biology, the identity of the queen-produced compounds that inhibit worker honey bee ovary development. As the honey bee is the only organism for which identities have been proposed for any pheromone that regulates reproduction, the resolution of its identity is of broad significance. We examined the effects of synthetic honey bee queen mandibular pheromone (QMP), four newly identified queen retinue pheromone components, and whole-queen extracts on the ovary development of caged worker bees. The newly identified compounds did not inhibit worker ovary development alone, nor did they improve the efficacy of QMP when applied in combination. QMP was as effective as queen extracts at ovary regulation. Caged workers in the QMP and queen extract treatments had better developed ovaries than did workers remaining in queenright colonies. We conclude that QMP is responsible for the ovary-regulating pheromonal capability of queens from European-derived Apis mellifera subspecies.

Environmental changes threaten plant-pollinator mutualisms and their critical ecosystem service. Drivers such as land use, invasions and climate change can affect pollinator diversity or species encounter rates. However, nitrogen deposition, climate warming and CO(2) enrichment could interact to disrupt this crucial mutualism by altering plant chemistry in ways that alter floral attractiveness or even nutritional rewards for pollinators. Using a pumpkin model system, we show that these drivers non-additively affect flower morphology, phenology, flower sex ratios and nectar chemistry (sugar and amino acids), thereby altering the attractiveness of nectar to bumble bee pollinators and reducing worker longevity. Alarmingly, bees were attracted to, and consumed more, nectar from a treatment that reduced their survival by 22%. Thus, three of the five major drivers of global environmental change have previously unknown interactive effects on plant-pollinator mutualisms that could not be predicted from studies of individual drivers in isolation.

We examined the effect of larval and adult nutrition on worker honey bee (Apis mellifera L.) ovary development. Workers were fed high or low-pollen diets as larvae, and high or low-protein diets as adults. Workers fed low-protein diets at both life stages had the lowest levels of ovary development, followed by those fed high-protein diets as larvae and low- quality diets as adults, and then those fed diets poor in protein as larvae but high as adults. Workers fed high-protein diets at both life stages had the highest levels of ovary development. The increases in ovary development due to improved dietary protein in the larval and adult life stages were additive. Adult diet also had an effect on body mass. The results demonstrate that both carry-over of larval reserves and nutrients acquired in the adult life stage are important to ovary development in worker honey bees. Carry-over from larval development, however, appears to be less important to adult fecundity than is adult nutrition. Seasonal trends in worker ovary development and mass were examined throughout the brood rearing season. Worker ovary development was lowest in spring, highest in mid-summer, and intermediate in fall.

We present a novel way to select for highly polygenic traits. For millennia, humans have used observable phenotypes to selectively breed stronger or more productive livestock and crops. Selection on genotype, using single-nucleotide polymorphisms (SNPs) and genome profiling, is also now applied broadly in livestock breeding programs; however, selection on protein/peptide or mRNA expression markers has not yet been proven useful. Here we demonstrate the utility of protein markers to select for disease-resistant hygienic behavior in the European honey bee (Apis mellifera L.). Robust, mechanistically-linked protein expression markers, by integrating cis- and trans- effects from many genomic loci, may overcome limitations of genomic markers to allow for selection. After three generations of selection, the resulting marker-selected stock outperformed an unselected benchmark stock in terms of hygienic behavior, and had improved survival when challenged with a bacterial disease or a parasitic mite, similar to bees selected using a phenotype–based assessment for this trait. This is the first demonstration of the efficacy of protein markers for industrial selective breeding in any agricultural species, plant or animal.