Heavy metal hyperaccumulation is prevalent throughout plant evolution, particularly in the legume family (Fabaceae), and acts as a presumed chemical defense against herbivory. However, heavy metal hyperaccumulation can have non-target impacts on other biological interactors including plant-microbe interactions. This information is important given the interest in utilizing legume plants for phytoremediation of anthropogenically contaminated soils. Here we employ a greenhouse experiment manipulating selenium level along with 16S amplicon sequencing methods to explore the effect of selenium on the prokaryotic microbiome of the selenium hyperaccumulator Astragalus crotalariae and non-accumulator Astragalus lentiginosus var. borreganus. Regardless of hyperaccumulator status, both plants accumulated high levels of selenium in leaf tissue when grown on soils with high selenium levels. However, the effect of selenium on the prokaryotic communities was more drastic in A. lentiginosus than A. crotalariae, explaining 37.57% of the observed variation and significantly affected community diversity and evenness. Many individual microbes were affected by selenium addition; notably, Mesorhizobium the nodulation-inducing genera of Astragalus spp., appeared in significantly less abundance in roots of both plant species grown on highly seleniferous soils. This project highlights the potential significance of ecological partners in heavy metal accumulating plants and the necessity of their consideration when using these plants for phytoremediation.

Symbiosis often occurs between partners with distinct life history characteristics and dispersal mechanisms. Many bacterial symbionts have genomes comprised of multiple replicons with distinct rates of evolution and horizontal transmission. Such differences might drive differences in population structure between hosts and symbionts and among the elements of the divided genomes of bacterial symbionts. These differences might, in turn, shape the evolution of symbiotic interactions and bacterial evolution. Here we use whole genome resequencing of a hierarchically-structured sample of 191 strains of Sinorhizobium meliloti collected from 21 locations in southern Europe to characterize population structures of this bacterial symbiont and its host plant Medicago truncatula. Sinorhizobium meliloti genomes showed high local (within-site) variation and little isolation by distance. This was particularly true for the two symbiosis elements pSymA and pSymB, which have population structures that are similar to each other, but distinct from both the bacterial chromosome and the host plant. The differences in population structure may result from among-replicon differences in the extent of horizontal gene transfer, although given limited recombination of the chromosome, different levels of purifying or positive selection may also contribute to among-replicon differences. Discordant population structure between hosts and symbionts indicates that geographically and genetically distinct host populations in different parts of the range might interact with genetically similar symbionts, potentially minimizing local specialization.

Symbiosis often occurs between partners with distinct life history characteristics and dispersal mechanisms. Bacterial symbionts often have genomes comprised of multiple replicons with distinct rates of evolution and horizontal transmission. Such differences might drive differences in population genetic structure between hosts and symbionts and among the elements of the divided genomes of bacterial symbionts. These differences might, in turn, shape the evolution of symbiotic interactions and bacterial evolution. Here we use whole genome resequencing of a hierarchically-structured sample of 191 strains of Ensifer meliloti collected from 21 locations in the native range to characterize population structure of this bacterial symbiont and its host plant Medicago truncatula. E. meliloti genomes showed high local (within-site) variation and little isolation by distance. This was particularly true for the two symbiosis elements pSymA and pSymB, which have population structures that are similar to each other, but distinct from both the bacterial chromosome and the host plant. The differences in population structure may result from differences in mobility or selection driving bacterial adaptation to life in the soil versus in association with plants. Discordant population structure between hosts and symbionts indicates that geographically and genetically distinct host populations in different parts of the range might interact with genetically similar symbionts, potentially minimizing the potential for local specialization.

Coevolution is predicted to depend on how the genetic diversity of interacting species is geographically structured. Plant-microbe symbioses such as the legume-rhizobium mutualism are ecologically and economically important, but distinct life history and dispersal mechanisms for these host and microbial partners, plus dynamic genome composition in bacteria, present challenges for understanding spatial genetic processes in these systems. Here we study the model rhizobium Ensifer meliloti using a hierarchically-structured sample of 191 strains from 21 sites in the native range and compare its population structure to that of its host plant Medicago truncatula. We find high local genomic variation and minimal isolation by distance across the rhizobium genome, particularly at the two symbiosis elements pSymA and pSymB, which have evolutionary histories and population structures that are similar to each other but distinct from both the chromosome and the host. While the chromosome displays weak isolation by distance, it is uncorrelated with hosts. Patterns of discordant population structure among elements with the bacterial genome has implications for bacterial adaptation to life in the soil versus symbiosis, while discordant population genetic structure of hosts and microbes might restrict local adaptation of species to each other and give rise to phenotypic mismatches in coevolutionary traits.