Resumen de Genomics of specificity in the symbiotic interaction between rhizobium leguminosarum and legumes
For millenia legumes have been used in crop rotations to incorporate nitrogen into agricultural systems, thus avoiding the need for fertilization. This is possible because of their symbiotic interaction with specific diazotrophic soil bacteria, collectively known as rhizobia, that infect legume roots, form root nodules and fix atmospheric nitrogen into ammonia using the nitrogenase enzyme. The establishment and maintenance of this symbiotic partnership results from a molecular conversation that guarantees specificity and guides the co-development of both organisms during root nodule formation. Rhizobium leguminosarum bv. viciae is a member of the -Proteobacteria that can establish effective symbioses with members of the Fabeae legume tribe (Pisum, Lathryrus, Lens and Vicia). Previous studies have suggested that, although all R. leguminosarum bv. viciae isolates can effectively nodulate all Fabeae, different Fabeae select specific genotypes of rhizobia from those available in soil.
The aim of this thesis was to characterize the genomic, genetic and molecular bases of the selection of specific R. leguminosarum bv. viciae genotypes by different host legumes (Pisum sativum, Lens culinaris, Vicia faba and V. sativa) from a well-characterized agricultural soil.
We established a population genomics methodology based on pooled DNA samples (Pool-Seq) from R. leguminosarum isolates obtained from different sources: legume plant hosts used as rhizobial traps (P. sativum, L. culinaris, V. sativa and V. faba), as well as the isolation of R. leguminosarum directly from soil. This approach allowed us to confirm the hypothesis that different plant hosts select specific subpopulations of rhizobia from the available population present in the soil.
We also set out to characterize the indigenous R. leguminosarum soil population avoiding plant selection, in order to compare it with previously characterized host-selected subpopulations from this soil. As a side result, we uncovered a large number of previously uncharacterized, non-symbiotic rhizobia that contribute to the population pangenome. Host preference for specific genotypes was especially relevant in the case of pea plants. They selected a R. leguminosarum population significantly different from that present in the soil. Quite on the contrary, lentil and fava-bean plants did not show a significant genotype selection, and their nodule rhizobial populations reflected that present in soil (after one life cycle). Vetch plants revealed a certain genotypic preference, but not as substantial or as important as that from pea plants.
Given that plants can differentially select rhizobial genotypes and that viable rhizobia of those genotypes are released into soil after nodule senescence, we hypothesized that, in natural conditions after numerous cycles of selection-release, most nodules would be occupied by the preferred genotype/s in each plant. We experimentally tested this hypothesis in a mesocosm experiment aimed at mimicking these field conditions. We were able to demonstrate that the different plant hosts employed (P. sativum, L. culinaris, V. faba and V. sativa) selected different genotypes from those available in the P1 soil, and that the basis of selection was different for different plants. Pea and fava-bean plants strongly selected for specific genotypes, but in opposite ways. Pea nodules were colonized by strains endowed with a large set of genes probably implicated in rhizospheric fitness, irrespective of the symbiotic genotype they harboured. This suggestion should be confirmed by in situ transcriptomic studies. Fava-bean plants restricted their selection to a specific symbiotic genotype that was not always localized in the same symbiotic plasmid, or within the same chromosomal background. No hard conclusions could be obtained for vetch, although we suggest that this plant might behave dually, either as a selective host if a given genotype results in a rhizospheric advantage (such as in vetchB subpopulation), or as a non-selective host, with its nodules reflecting the genotypic diversity present in soil (such as in vetchA subpopulation). This last case was the situation found for lentils; none of the three subpopulations isolated from lentil nodules (initial lentil, lentilA and lentilB) differed significantly, either among themselves, regardless of the number of plant selection cycles, or with respect to the initial soil population.
