Typically pathogens infect multiple host species. Such multi-host pathogens can show considerable variation in their degree of infection and transmission specificity, which has important implications for potential disease emergence. Transmission of multi-host pathogens can be driven by key host species and changes in such transmission networks can lead to disease emergence. We study two viruses that show contrasting patterns of prevalence and specificity in managed honeybees and wild bumblebees, black queen cell virus (BQCV) and slow bee paralysis virus (SBPV), in the context of the novel transmission route provided by the virus-vectoring Varroa destructor. Our key result is that viral communities and RNA virus genetic variation are structured by location, not host species or V. destructor presence. Interspecific transmission is pervasive with the same viral variants circulating between pollinator hosts in each location; yet, we find virus-specific host differences in prevalence and viral load. Importantly, V. destructor presence increases the prevalence in honeybees and, indirectly, in wild bumblebees, but in contrast to its impact on deformed wing virus (DWV), BQCV and SBPV viral loads are not increased by Varroa presence, and do not show genetic evidence of recent emergence. Effective control of Varroa in managed honeybee colonies is necessary to mitigate further disease emergence, and alleviate disease pressure on our vital wild bee populations. More generally, our results highlight the over-riding importance of geographical location to the epidemiological outcome despite the complexity of multi-host-parasite interactions.

We collected foraging bees (355 A. mellifera, 281 B. pascuorum, 640 B. terrestris and 38 B. lucorum individuals) within a 1x1 km area, from four Varroa-free islands; three Varroa-positive islands; and five Varroa-positive mainland sites. We extracted RNA from individual bees, and converted it to cDNA and carried out PCRs and qPCR assays for individual viruses (black queen cell virus and slow bee paralysis virus) to collect prevalence data and viral load data. We also carried out Sanger and Pacbio sequencing, these data are uploaded to Genbank and accession numbers are detailed in the paper.