Drivers of alpha and beta-diversity in Xylocopa microbiomes
Xylocopa gut samples hosted ~222% more ASVs and had greater evenness than did crop samples (Figure 1; Chao1 lmer F1, 54= 148 p <0.001; Shannon lmer F1, 45 = 171 p <0.001) across bee species. Xylocopa sonorina hosted marginally greater species richness than X. tabaniformis (F1, 56 = 3.78 p = 0.057) and significantly greater diversity (F1,51 = 15.9, p <0.001), while the interaction of tissue and species was not significant (Chao1 F1,54 = 0.12, p = 0.73; Shannon F1, 45 = 0.056 p = 0.81). Within X. sonorina , female bees had significantly greater species richness and diversity within the gut than male bees (Chao1 p = 0.001, Shannon p = 0.04; Supplementary Figure S3), but no difference was detected inX. sonorina between sexes in the crop (Chao1 p>0.05; Shannon p>0.05), nor between sexes in either tissue forX. tabaniformis (p >0.2).
Bacterial species composition varied with sampling location, tissue type, species, their interactions, as well as the sex of bees (Figure 1D; full model Bray-Curtis PERMANOVA p <0.05 for all, Supplementary Table S1, Supplementary Figure S3). Paired analyses conducted on bees where both the crop and gut sequenced confirmed that crop and gut communities differed in both bee species (Supplementary Table S2; X. sonorina Bray F = 5.63, R2 = 0.098, p <0.001; X. tabaniformis Bray F = 3.87, R2 = 0.16, p <0.001). The gut and crop communities differed in dispersion. Using Bray-Curtis dissimilarities, the crop samples were more variable than the gut for both species (betadisper X. sonorina , p <0.05; X. tabaniformis p <0.01; Fig 1B; Supplementary Figure S4).
In our validation analysis, species differed in bacterial composition when samples from a single location where both species were sampled were considered (Davis), for both the crop (Supplementary Table S3, Bray R2 = 0.10 F=2.56, p=0.01) and the gut tissues (Bray R2 = 0.21 F=8.52, p<0.001).