Abstract
BACKGROUND: Working toward a general framework to understand the role of microbiota in animal biology requires the characterisation of animal-associated microbial communities and identification of the evolutionary and ecological factors shaping their variation. In this study, we described the microbiota in the cloaca, brood patch skin and feathers of two species of birds and the microbial communities in their nest environment. We compared patterns of resemblance between these microbial communities at different levels of biological organisation (species, individual, body part) and investigated the phylogenetic structure to deduce potential microbial community assembly processes. RESULTS: Using 16S rRNA gene amplicon data of woodlarks (Lullula arborea) and skylarks (Alauda arvensis), we demonstrated that bird- and nest-associated microbiota showed substantial OTU co-occurrences and shared dominant taxonomic groups, despite variation in OTU richness, diversity and composition. Comparing host species, we uncovered that sympatric woodlarks and skylarks harboured similar microbiota, dominated by Proteobacteria, Firmicutes, Actinobacteria, Bacteroidetes and Acidobacteria. Yet, compared with the nest microbiota that showed little variation, each species' bird-associated microbiota displayed substantial variation. The latter could be partly (~ 20%) explained by significant inter-individual differences. The various communities of the bird's body (cloaca, brood patch skin and feathers) appeared connected with each other and with the nest microbiota (nest lining material and surface soil). Communities were more similar when the contact between niches was frequent or intense. Finally, bird microbiota showed significant phylogenetic clustering at the tips, but not at deeper branches of the phylogeny. CONCLUSIONS: Our interspecific comparison suggested that the environment is more important than phylogeny in shaping the bird-associated microbiotas. In addition, variation among individuals and among body parts suggested that intrinsic or behavioural differences among females and spatial heterogeneity among territories contributed to the microbiome variation of larks. Modest but significant phylogenetic clustering of cloacal, skin and feather microbiotas suggested weak habitat filtering in these niches. We propose that lark microbiota may be primarily, but not exclusively, shaped by horizontal acquisition from the regional bacterial pool at the breeding site. More generally, we hypothesise that the extent of ecological niche-sharing by avian (or other vertebrate) hosts may predict the convergence of their microbiota.