Rhizobia are ecologically important, facultative plant symbiotic microbes. In nature there exists large variability in the association of rhizobial strains and host plants of the same species. Here, we evaluated whether plant and rhizobial genotypes influence the initial transcriptional response of rhizobium following perception of host plant. RNA-sequencing of the model rhizobium Sinorhizobium meliloti exposed to root exudates or luteolin was performed in a combination of three S. meliloti strains and three Medicago sativa varieties. The response to root exudates involved hundreds of changes in the rhizobium transcriptome. Of the differentially expressed genes, expression of 35% were influenced by strain genotype, 16% by the plant genotype, and 29% by strain x host plant genotype interactions. We also examined the response of a hybrid S. meliloti strain, in which the symbiotic megaplasmid (~ 20% of the genome) was mobilized between two of the above-mentioned strains. Dozens of genes were up-regulated in the hybrid strain, indicative of nonadditive variation in the transcriptome. In conclusion, this study demonstrated that transcriptional responses of rhizobia upon perception of legumes is influenced by the genotypes of both symbiotic partners, and their interaction, a wide spectrum of genetic determinants involved in the phenotypic variation of plant-rhizobium symbiosis.

Genes differentially expressed (log₂[fold change] ≥ 1, p-value < 0.01)  relative to the control condition are reported, and all fold change values for these genes are shown. To compare expression of genes conserved between Rm1021, AK83, and BL225C, the pangenome of the three strains was calculated using Roary version 3.13.0 with an identity threshold of 90%, and the genes found in all three strains (the core genes) were recorded. For each condition, core genes differentially expressed in at least one strain relative to the control condition were identified, and the fold change values for the gene and its orthologs in the other strains were extracted and reported in tables.

Heatmaps in the .zip folder show the clustering of genes differentially expressed and are using the ComplexHeatmap and Heatmaply packages of R