Item

Abstract

The phyllosphere of wild Arabidopsis thaliana was found to be dominated by the genus Pseudomonas. Though dramatic fitness effects on the host have been reported for infections with single Pseudomonas strains, A. thaliana is known to be colonized by many bacterial species, and even different Pseudomonas strains of the same species simultaneously. Little is known about the reciprocal fitness outcomes in a more realistic scenario of multiple coexisting Pseudomonas strains and Arabidopsis genotypes. The collective fitness outcomes derived from multiple interacting Pseudomonas strains is unlikely to be predictable from information about individual strains, neither for the bacteria nor for the host. Furthermore, the extent to which the plant genotype shapes its Pseudomonas community at the intra-species level has not been studied. Here, we present a novel method to accurately track single Pseudomonas strains in metagenomic data, and leverage it to test the effect of plant genotype, strain type and strain-strain interactions on the fitness of both host and bacteria. A panel of local Arabidopsis thaliana genotypes was infected with synthetic communities of Pseudomonas strains isolated from wild A. thaliana. In order to accurately evaluate the abundance of each Pseudomonas strain independently in metagenomic data, we genome-barcoded the strains using unique sequences integrated utilizing the Tn7 system. Plants were infected with mixtures of tagged strains, and 14 days post-infection, plant weight and strain abundance were estimated. The plant genotype had a statistically significant effect on which strain dominated. Moreover, colonization success of individual strains depended on the co-colonizing strains. Lastly, we observed that co-colonization of pathogenic and non-pathogenic strains resulted in disease protection, that may be partly mediated by the host genotype. Collectively, our results reflect the wide range of possible interactions and fitness outcomes between highly similar members of the same species.