How population structure and nest membership shape pathogen patterns in bumble bees?

Host density, genetic diversity, and social groups are key factors influencing pathogen transmission in wildlife populations, but their interactions remain poorly understood in insects. Islands can provide natural laboratories with distinct populations that vary in density and genetic diversity, whereby dense, genetically homogenous populations are expected to facilitate pathogen transmission. We used bumble bees to test these predictions, assessing the population structure of the two common species Bombus pascuorum and B. terrestris across island and mainland sites in the British Isles and France, and testing bees for five micro-parasitic and four viral pathogens. B. pascuorum formed distinct genetic clusters on islands, with varying levels of heterozygosity, and only the Isle of Arran clustered with mainland populations. B. terrestris populations were less structured, but populations on the Isle of Man and the Scilly Isles were genetically separated from other island and mainland populations while showing low heterozygosity. Colony density was similar between species and not linked to genetic diversity, but had a positive effect on the prevalence of some pathogens. Contrary to expectations, there was no protective effect of high genetic diversity, suggesting that generalist bumble bee pathogens could be more affected by host species diversity and density. Yet, within B. terrestris populations, we found that nestmates showed more similar pathogen profiles than unrelated individuals, suggesting that genetic similarity and high contact rates within nests affect pathogen prevalence in wild bees.

Methods
216 Bombus pascuorum and 573 B. terrestris were collected from flowers from mid June to late August in 2021 across seven islands in the Irish Sea, the English Channel, and off the coast of Brittany in France, and four coastal sites in England and France in 2021. From late June to early August 2022, islands (except for the Scilly Isles) were sampled again using five sampling sites on large islands (Arran, Isle of Man, Guernsey, Belle-Ile) or collecting across the whole island on smaller islands (< 15 km2, Alderney and Ouessant). A total of 988 B. pascuorum and 1,163 B. terrestris were obtained in 2022.

DNA was extracted from bees using a single hind leg. Bees were genotyped at 12 loci, and fragments were run on a Thermo Fisher Scientific Genetic Analyzer 3130 Series. Alleles were scored in Geneious Prime® v2023.0.1. Sibship reconstruction was performed in COLONY v2.0.6,7, and population structure was assessed using DAPC (discriminant analysis of principal components) and the Bayesian clustering algorithm in Structure v2.3.4.

To study how population structure affects pathogen prevalence, we used island samples from 2022. We screened bumble bees for nine pathogens. Tissue from half bees (laterally bisected) was homogenized, and RNA was extracted and reverse transcribed to cDNA to test for RNA viruses and mRNA from parasite infections. PCR was used to screen for black queen cell virus (BQCV), deformed wing virus type A and B (DWV-A and DWV-B), slow bee paralysis virus (SBPV), Apicistis bombi, Crithidia bombi, Nosema apis, N. bombi, and N. ceranae.