The evolutionary genomics of adaptation to stress in wild rhizobium bacteria

Microbiota comprise the bulk of life's diversity, yet we know little about how populations of microbes accumulate adaptive diversity across natural landscapes. Adaptation to stressful soil conditions in plants provides seminal examples of adaptation in response to natural selection via allelic substitution. For microbes symbiotic with plants however, horizontal gene transfer allows for adaptation via gene gain and loss, which could generate fundamentally different evolutionary dynamics. We use comparative genomics and genetics to elucidate the evolutionary mechanisms of adaptation to physiologically stressful serpentine soils in rhizobial bacteria in western North American grasslands. In vitro experiments demonstrate that the presence of a locus of major effect, the nre operon, is necessary and sufficient to confer adaptation to nickel, a heavy metal enriched to toxic levels in serpentine soil, and a major axis of environmental soil chemistry variation. We find discordance between inferred evolutionary histories of the core genome and nreAXY genes, which often reside in putative genomic islands. This suggests that the evolutionary history of this adaptive variant is marked by frequent losses, and/or gains via horizontal acquisition across divergent rhizobium clades. However, different nre alleles confer distinct levels of nickel resistance, suggesting allelic substitution could also play a role in rhizobium adaptation to serpentine soil. These results illustrate that the interplay between evolution via gene gain and loss and evolution via allelic substitution may underlie adaptation in wild soil microbiota. Both processes are important to consider for understanding adaptive diversity in microbes and improving stress-adapted microbial inocula for human use.

Methods
Strain isolation and soil analysis. Acmispon wrangelianus and A. brachycarpus were gathered during March 2019 from natural reserves in California and in May 2019 from areas in Oregon. Within reserves, both serpentine and non-serpentine sites were sampled for: 1) plant root nodules, and 2) the soil directly below the plant, if available. Details on strain isolation and soil analysis procedures are in SI Appendix.

Phenotyping Ni tolerance. The minimum inhibitory concentration (MIC) of Ni was determined for 681 field-collected strains of Mesorhizobium by growing the strains on NiCl2 enriched media plates that ranged from 0 mM to 5 mM increasing by 1 mM for each treatment with three replicates. Ni tolerance was also determined by the growth in liquid media containing 0 or 1 mM NiCl2 (OD600). Details of these procedures are in SI Appendix.