Landscape structure affects temporal dynamics in the bumble bee virome: Landscape heterogeneity supports colony...

Virus spillovers from managed honey bees, Apis mellifera, are thought to contribute to the decline of wild pollinators, including bumble bees. However, data on the impact of such viruses on wild pollinators remain scarce, and the influence of landscape structure on virus dynamics is poorly understood. In this study, we deployed bumble bee colonies in an agricultural landscape and studied changes in the bumble bee virome during field placement under varying habitat composition and configuration using a multiscale analytical framework. We estimated prevalence of viruses and viral loads (i.e. number of viral genomic equivalent copies) in bumble bees before and after placing them in the field using next generation sequencing and quantitative PCR. The results show that viral loads and number of different viruses present increased during placement in the field and that the virus composition of the colonies shifted from an initial dominance of honey bee associated viruses to a higher number (in both viral loads and number of viruses present) of bumble bee associated viruses. Especially DWV-B, typical for honey bees, drastically decreased after the time in the field. Viral loads prior to placing colonies in the field showed no effect on colony development, suggesting low impacts of these viruses in field settings. Notably, we further demonstrate that increased habitat diversity results in a lower number of different viruses present in Bombus colonies, while colonies in areas with well-connected farmland patches decreased in their total viral load after field placement. Our results emphasize the importance of landscape heterogeneity and connectivity for wild pollinator health and that these influences predominate at fine spatial scales.