Natural selection acting on the genetics of host response to commensal bacteria

BackgroundMicrobiota members collectively contribute to individual performance in humans, animals, and plants. This led to the quest for probiotics to improve host health and reproductive performance. Although the efficacy of probiotics is known to be strongly affected by environmental factors, microbe-microbe interactions, and microbial strain identity, the effects of host genotype and the underlying genetic architecture have been overlooked. In addition, the evolutionary causes of such genetic variation are typically not addressed. In this study, we aimed to describe the genetic architecture of the adaptation of the host plant Arabidopsis thaliana to commensal bacterial members of its native microbiota by identifying candidate genes associated with fitness proxies and presenting signatures of natural selection. ResultsA Genome-Wide Association study conducted under field conditions revealed extensive variation within a new mapping population of 162 genotypes of A. thaliana scored for total seed production and its two underlying components, namely fruit number and mean seed number per fruit, in response to 13 commensal strains. In agreement with the strong host genotype x commensal strain identity interactions observed for each reproductive trait, the polygenic genetic architecture was highly flexible among the 13 commensal strains. Candidate genes exhibited a significant enrichment in signatures of both local adaptation and balancing selection. In line with the phenotyped reproductive traits, we identified seven candidate genes with functions specifically and strongly linked to seed germination and fertility. ConclusionsOur findings reveal the importance of genotype-by-genotype interactions when measuring fitness proxies on a wild plant species inoculated with key members of its native microbiota. In addition, this study improves our understanding of the genetic signatures of natural selection acting on native host-microbiota adaptive interactions.