Plant-mediated selection of endosymbiotic bacteria – SELECT

We studied the various mechanisms governing the ecological success of rhizobia associated with the plant Mimosa pudica, using a range of experimental approaches.

First, we exploited an evolutionary experiment—developed within our team over many years—aimed at evolving a plant-pathogenic bacterium into a symbiont of Mimosa. During this experiment, bacteria gradually acquired and improved their ability to form nodules and proliferate within them. By sequencing the genomes of these experimentally evolved bacteria, we were able to identify the mutations responsible for the observed symbiotic improvements and we analyzed which stage of the symbiotic interaction improved most rapidly.

We then studied in greater detail the extent to which the host plant can control the abundance of bacteria in its nodules based on their level of nitrogen fixation. Using a natural rhizobium from M. pudica, we constructed mutant strains with varying degrees of impairment in their ability to fix nitrogen, and we measured the proliferation of these strains within nodules.

To assess the effect of competition on the rhizobial populations of M. pudica, we then assembled a collection of natural rhizobia, which we inoculated onto their host plant either individually, in pairs, or in groups of 6 or 8 strains. In each case, we measured the abundance of each bacterial strain in the plant, calculated as the product of the number of nodules formed by each strain and of their proliferation within each nodule. When the presence of one strain altered the fitness of another one (indicating an interaction between two strains), we then analyzed at which stage of the life cycle this interaction occurred (life in the rhizosphere, root entry, or proliferation within the nodules).

Finally, we focused on the late stages of this symbiosis: nodule senescence. We investigated whether the proliferation and survival of rhizobia in mature and senescing nodules depended on the strains’ nitrogen-fixing capacity or on their ability to store carbon reserves.

The study of experimentally evolved bacterial strains showed that plants exerted strong selective pressure on bacterial entry into roots and that the bacteria’s ability to enter roots was partly linked to their ability to proliferate within root nodules. Indeed, among the identified mutations that enhance the symbiotic capabilities of bacteria, all had a positive effect on entry into the roots, and some also increased bacterial proliferation within the nodules.

Second, we observed that the relationship between the level of nitrogen fixation and bacterial survival in the nodules was not linear. Bacterial strains with a slightly reduced level of nitrogen fixation have a low fitness equivalent to that of strains that no longer fix nitrogen at all. Therefore, there seems to be a threshold-based mechanism in the control of nitrogen fixation, below which rhizobial fitness is strongly reduced by the host plant.

Third, we were able to show that interactions between rhizobia were very frequent and occurred mainly at the root entry level. This finding independently validates the results obtained with experimentally evolved strains. However, we also demonstrated that interactions could occur during later stages, during nodule colonization. In particular, we identified a strain that had a negative effect on the proliferation of another strain when these two strains colonized independent nodules of the same plant. We also observed that the interactions detected in inoculations involving pairs of strains were preserved in inoculations involving communities of 6 or 8 strains. This result indicates that the biological complexity of rhizobial communities has no significant effect on the interactions that occur between pairs of strains.

Finally, we demonstrated that neither nitrogen fixation nor the ability to store carbon reserves had any impact on bacterial survival in senescent nodules. However, analysis of mature nodules showed that bacteria that do not fix nitrogen accumulate more carbon reserves, whereas strains that do not store carbon reserves do not fix more nitrogen. This result indicates that these two important rhizobial properties are not subject to a bidirectional physiological trade-off, whereby an investment in one trait would automatically lead to a decreased investment in the other trait. It also suggests that certain intrinsic bacterial properties (the ability to store reserves) influence how the plant controls the fitness of non-nitrogen-fixing rhizobia.

This project has enabled us to describe and quantify the various mechanisms that control the fitness of rhizobia during their symbiotic interaction with the plant Mimosa pudica. We have identified a new type of interaction between two rhizobial strains, occurring within separate nodules and leading to a decrease in the fitness of one of the strains. We also demonstrated that the relationship between nitrogen fixation and rhizobial fitness within nodules is not linear. In both cases, it would be of interest to investigate the molecular mechanisms involved in these processes and to test whether they are also observed in other rhizobia-legume symbiotic associations.

Furthermore, our results, along with some preliminary observations, indicate that the physiological status of the symbiotic organisms (carbon reserve storage in the bacteria, level of plant exposure to light) impacts how legume plants control bacterial proliferation within their nodules. These observations pave the way for a more in-depth characterization of these effects, which could be conducted both in the laboratory and in natural or agricultural environments.

Finally, a study of our collection of natural rhizobial strains from Mimosa pudica showed that strains with equivalent nitrogen-fixing levels exhibit very different viability levels within the nodules. This observation is reminiscent of the phenomenon of terminal differentiation, well described in certain legume plants that impose a very specific physiological state to the bacteria present in their nodules, leading to a loss of viability. However, it is generally accepted that this phenomenon is controlled solely by the plant. It would therefore be interesting to investigate whether this differentiation phenomenon could also be controlled by the bacterial strain within the nodules of Mimosa pudica.

Eukaryotic organisms are exposed to complex microbial communities that affect their health, development and evolution. They have evolved multiple layers of control mechanisms to defend themselves from pathogens while, at the same time, allowing colonization by beneficial microbes. The filtering of microbes by eukaryotes imposes strong selective constraints on microbial populations and shape their eco-evolutionary dynamics. It is thus important to describe the effect of host-mediated selection on microbial populations to understand, and possibly predict, the evolution of pathogenic and beneficial microorganisms.

The rhizobium-legume symbiosis has a major impact on natural and agricultural ecosystems, contributing to a substantial part of the global flux of nitrogen. During this symbiosis, rhizobia are hosted within specific root organs (called nodules) where they fix atmospheric nitrogen and provide organic ammonium to the plant in exchange for carbon sources. Large phenotypic variations are observed among the different strains that can be found in soil rhizobial populations, some strains being highly beneficial while others provide little or no nitrogen to their host. During the interaction with the plant, rhizobial fitness is determined by the realization of three symbiotic steps: nodulation, leading to formation of new plant organs; infection, the entry and proliferation of bacteria within plant cells; and nitrogen fixation and the provision of ammonium to the plant. These three steps are jointly controlled by the genotypes of the two partners and influenced by ecological conditions, including the complexity of the rhizobial community in the rhizosphere.

SELECT will study how plant-mediated selection shapes bacterial fitness and the evolution of phenotypic traits involved in the different symbiotic steps. Three complementary experimental systems will be exploited to investigate different facets of this question. First, an evolution experiment that recapitulates in the laboratory the emergence of new rhizobia from a pathogenic ancestor will be analyzed to describe the dynamics of phenotypic and genetic evolution leading to the acquisition and improvement of the first two symbiotic steps, nodulation and infection. Next, I will genetically engineer rhizobia that fix different amounts of nitrogen to establish the quantitative relationship between nitrogen fixation and bacterial fitness within nodules. Last, I will use a collection natural rhizobial strains to compare the expression and evolution of symbiotic phenotypes after the inoculation of one or multiple bacterial strains on the host plant. The different experimental results will be integrated through mathematical modelling to analyze the relative contribution of the different symbiotic steps on bacterial fitness.

Overall, SELECT will thus couple several approaches (from bacterial genetics and genomics to experimental evolution and mathematical modelling) and experimental systems to provide a detailed view of the factors determining bacterial fitness during a complex mutualistic interaction. The fundamental insights on the eco-evolutionary dynamics of host-associated bacteria generated during SELECT may contribute, on the longer term, to improve agro-ecological practices by allowing the development of selection procedures to improve rhizobial performances (both nodulation competitiveness and nitrogen fixation efficiency) under different ecological contexts and on different host plants.