Rhizobial bacteria are known for their capacity to fix nitrogen for legume hosts. However ineffective rhizobial genotypes exist and can trigger the formation of nodules but fix little if any nitrogen for hosts. Legumes must employ mechanisms to minimize exploitation by the ineffective rhizobial genotypes to limit fitness costs and stabilize the symbiosis. Here we address two key questions about these host mechanisms. What stages of the interaction are controlled by the host, and can hosts detect subtle differences in nitrogen fixation? We provide the first explicit evidence for adaptive host control in the interaction between Lotus japonicus and Mesorhizobium loti. In both single inoculation and co-inoculation experiments, less effective rhizobial strains exhibited reduced in planta fitness relative to the wildtype M. loti. We uncovered evidence of host control during nodule formation and during post-infection proliferation of symbionts within nodules. We found a linear relationship between rhizobial fitness and symbiotic effectiveness. Our results suggest that L. japonicus can adaptively modulate the fitness of symbionts as a continuous response to symbiotic nitrogen fixation.

Legumes preferentially associate with and reward beneficial rhizobia in root nodules, but the processes by which rhizobia evolve to provide benefits to novel hosts remain poorly understood. Using cycles of in planta and in vitro evolution, we experimentally simulated lifestyles where rhizobia repeatedly interact with novel plant genotypes with which they initially provide negligible benefits. Using a full-factorial replicated design, we independently evolved two rhizobia strains in associations with each of two Lotus japonicus genotypes that vary in regulation of nodule formation. We evaluated phenotypic evolution of rhizobia by quantifying fitness, growth effects and histological features on hosts, and molecular evolution via genome resequencing. Rhizobia evolved enhanced host benefits and caused changes in nodule development in one of the four host–symbiont combinations, that appeared to be driven by reduced costs during symbiosis, rather than increased nitrogen fixation. Descendant populations included genetic changes that could alter rhizobial infection or proliferation in host tissues, but lack of evidence for fixation of these mutations weakens the results. Evolution of enhanced rhizobial benefits occurred only in a subset of experiments, suggesting a role for host–symbiont genotype interactions in mediating the evolution of enhanced benefits from symbionts.