Conclusions
Interaction between soil fungi and bacteria in the plant root microbiota (rhizomicrobiota) is becoming a central topic in understanding the ecology of the rhizosphere and exploiting the potential of combined manipulation of the fungal and bacterial component in order to improve plant nutrition and growth and resistance to biotic and abiotic stress (for examples see
This interaction can span from almost random microbial assemblies to specific symbiotic associations of fungal hyphae and bacterial cells Concerning Trichoderma spp., most of the studies have focused on the antagonistic interaction toward pathogenic bacteria, while positive effect on mutualistic bacteria have been poorly investigated. A recent work on mycorrhizal fungi , showed the alignment of fungal/rhizobial mutualism toward eliciting plant gene expression, but no direct observation has been made on the modulation of rhizobial mutualism.
Evidence presented in this work indicates that Trichodermastrains (belonging to four species) have an impact on S. melilotiphenotypes, at either physiological and host plant growth promotion levels, and alter gene expression of a large fraction of the genome. Additionally, such effects are different for the differentTrichoderma species and for the single S. melilotistrains, clearly showing a strong component of genotype-by-genotype interaction. In particular, we found that the various combinations ofS. meliloti strains and Trichoderma species elicit the differential expression of a variable fraction of genes, spanning from 6% to 25% of total S. meliloti genes. The same fungal species can determine up to 3-fold differences in the extent of the stimulon among strains (e.g. T. velutinum from 6.6% to 25.0%), as well as the same S. meliloti strain can vary on a similar extent its stimulon among different fungal species treatments (e.g. 1021 from 6.0% to 25.0%). Under a condition simulating the first step of symbiosis, i.e. the recognition of the host plant by the perception of root exudates, the same S. meliloti strains displayed stimulons not higher than 19% of total genome and average values around 1%-8% in most strains and conditions . The larger stimulons observed here, with fungal spent media, led to speculate that several functions encoded by the S. meliloti genome could involve interaction with members of the soil and root microbiota rather than simply the symbiosis with plant. A metabolic model reconstruction done on S. melilotishowed the importance of many genes residing on the chromosome and the pSymB chromid (accounting roughly for 2/3 of the genome) for metabolic adaptation to soil and rhizosphere, while relatively few were relevant for symbiosis .
Within these large stimulons, ca. 23% of shared DEGs display a differential expression which is significantly modelled by rhizobium x fungus interaction, again supporting the hypothesis that a relatively important fraction of the S. meliloti genome and of its variation could be associated to the interaction with fungi in the rhizomicrobiota. Concerning the functional groups of the overall DEGs contributing to differences among stimulons, relevant contributions of COG categories K (transcription) and S (unknown function) were found. From one side this result highlights that at least some of the genes with still unknown functions may be related to functions normally not assessed in laboratory conditions, and involved in social interaction in the natural environment, from another side it emphasizes the role of transcriptional regulators (e.g. two-component systems) and gene x gene (i.e. epistatic) interactions in social interactions. The relevance of epistatic interaction is further confirmed by comparing the results from the cis -hybrid strain to the two parental strains 1021 and BL225C. Unexpectedly, these results indicate that the swapping of the symbiotic megaplasmid pSymA affects global stimulons in conditions (fungal spent media) apparently unrelated to symbiosis with plant. Indeed, stimulons of the cis -hybrid strain do not overlap with those of the two parental strains. Epistatic interactions between genes residing on the megaplasmid pSymA and the rest of genome were already shown, though very few in terms of number of genes involved, by genome reduction experiments and Tn-seq analyses . On the samecis -hybrid strain we showed that stimulons from root exudates are, as expected, displaying nonadditive (epistatic) effects . The finding that the transcriptome of the same strain reports similar evidence strongly suggests the presence on pSymA of genes related to social interaction in the rhizomicrobiota.
Moreover, the reported data on the symbiotic quality of the S. meliloti - Trichoderma combinations disclose the possibility that the rhizomicrobiota may modulate rhizobium-plant mutualistic interactions (synergistically or antagonistically, depending on the genotypic combinations, again emphasizing a genotype x genotype effect). Indeed, also results obtained in culture with spent media indicated effects, including the genotype-species specific response of plant-growth promoting related phenotypes of S. meliloti , which could modulate symbiotic quality, as the production of indole-3-acetic acid (IAA). IAA production in S. meliloti has been in fact demonstrated to positively modulate symbiotic quality . Moreover, within some stimulons (e.g. the cis -hybrid strain treated with T. velutinum ) we found the presence of some nod genes, which are required to produce the lipochito-oligosaccharide molecule (Nod Factor), the first molecular signal directed toward the host plant by rhizobia . Modulation, e.g. synergistic effect, by fungi on rhizobium symbiotic quality has been observed for mycorrhizal fungi but until now not for non-mycorrhizal fungi as Trichoderma and for the effect of fungi on rhizobium physiology and genome-wide transcriptional patterns. Data obtained on symbiotic tests with S. meliloti - Trichodermacombinations demonstrated that alfalfa growth performances are indeed affected by S. meliloti - Trichoderma mix. However, a plant variety dependency over the effect of the different combinations was observed, which points out to the importance of the tripartite interaction between host plant genotype and the genotypes of the rhizomicrobiota, here oversimplified with with S. meliloti -Trichoderma combinations.
We may conclude hypothesizing that rhizobium fitness in symbiosis, and consequently rhizobial genome evolution and genomic diversity, could be modulated by rhizospheric microbes, which may have favoured or antagonized the overall differentiation pathway leading to nitrogen-fixing root nodules. Under this hypothesis we propose novel models for studying rhizobium-plant interaction, which should include other components of the rhizomicrobiota, in the present proof-of-concept exemplified by Trichoderma . Such novel models should clarify the role of genes, many of them still with unknown function, in the tripartite social interaction between rhizobium, host plant, and microbiota, leading to understanding the multifaceted selective pressures over rhizobium in soil, rhizosphere and the benefits of symbiosis.
In perspective, such models would allow to translate laboratory evidence into the application of novel bioinoculant consortia due to high relevance for agriculture of the nitrogen-fixation operated by rhizobia and the biocontrol by Trichoderma , considering that for both formulations and regulation on their use already exist , as well application to alfalfa cultivation .