Function is a better predictor of plant rhizosphere community membership than 16S phylogeny.

Abstract

Rhizobacterial communities are important for plant health but we still have limited understanding of how they are constructed or how they can be manipulated. High-throughput 16S rRNA sequencing provides good information on taxonomic composition but remains an unreliable proxy for phenotypes. In this study, we tested the hypothesis that experimentally observed functional traits would be better predictors of community membership than phylogenetic origin. To test this hypothesis, we sampled communities on four plant species grown in two soil types and characterized 593 bacterial isolates in terms of antibiotic susceptibility, carbon metabolism, resource use and plant growth-promoting traits. In support of our hypothesis we found that three of the four plant species had phylogenetically diverse, but functionally constrained communities. Notably, communities did not grow best on complex media mimicking their host of origin but were distinguished by variation in overall growth characteristics (copiotrophy/oligotrophy) and antibiotic susceptibility. These data, combined with variation in phylogenetic structure, suggest that different classes of traits (antagonistic competition or resource-based) are more important in different communities. This culture-based approach supports and complements the findings of a previous high-throughput 16S rRNA analysis of this experiment and provides functional insights into the patterns observed with culture-independent methods.