Nearly all multi-cellular organisms host microbial symbionts that can influence a range of critical life processes, from development to immunity and metabolism. However, little is known about how environmental factors shape host-associated microbial communities, and how this relates to symbiont function and host health. Pesticide exposure may influence the honey bee (Apis mellifera) gut microbiome because some known gut bacteria possess genes that are hypothesized to detoxify chemicals, and thus may offer symbiont and host protection. Here, we used a combined approach of field observations, field experiments, and laboratory experiments to evaluate the effects of pesticide exposure on the structure and function of the honey bee gut microbiome. First, using field-collected bees, we quantified pesticide concentration of six commonly-used agricultural and in-hive pesticides to test for a correlation between pesticide content in the bees and the structure of their gut microbiome using 16S rRNA gene amplicon sequencing. Second, we conducted an in-hive experiment testing the effects of three different pesticides and a probiotic on the structure of the gut microbiome, host immune function, and parasite levels. Lastly, we tested the growth of gut bacterial cultures in laboratory assays with varying pesticide concentrations.
Results/Conclusions
Four of the six tested pesticides were detected in our field-collected bees, and the levels of two of these pesticides varied significantly across sites (imidacloprid and t-fluvalinate). Bees from hives with different levels of the neonicotinoid pesticide, imidacloprid, had significantly different gut microbiomes, although we acknowledge that other variables, such as diet, could be confounding this observational correlation. In hive experiments, pesticide treatment shifted the relative abundances of certain bacterial groups (e.g. Lactobacillus sp.), as well as increased the ectoparasitic mite counts. Interestingly, probiotic bacteria mitigated the effects of the pesticides on parasite levels. There were no differences among treatments in the activity of two immune function markers (glucose oxidase and phenoloxidase). Lastly, in laboratory assays with imidacloprid, we found that the majority of bacterial isolates tested (91%) exhibited reduced growth compared to controls, even in low concentrations of imidacloprid. However, several isolates were robust to altered growth in the presence of imidacloprid and may contain detoxification genes. These in vitro effects of pesticides on bacterial growth could also exist in vivo. Our combined field and laboratory results suggest that pesticides can influence host-microbe-parasite interactions and can have important implications for bee pollinator health, and thus ecosystem health.