S. meliloti-Trichoderma combinations differentially affect the quality of symbiosis
Since the most notable and ecologically relevant phenotype of rhizobia is their mutualistic interaction with leguminous plants, giving rise to nitrogen-fixing symbiosis, we evaluated if Trichoderma may modify the mutualistic interaction (symbiosis quality) between S. meliloti and the host legume plant M. sativa . Additionally, we aimed to inspect if the modulation of the mutualism possibly operated byTrichoderma could display evidence for genotype-by-genotype interaction, as did the previous results on S. melilotiphenotypes (i.e. growth, root biofilm formation, auxin, siderophore, EPS production). Results from in vitro symbiosis tests are reported in Figure 3 and Supplementary Figure S2 . Symbiosis quality was assessed in terms of root, shoot, and stem length, number of leaves, number of root nodules, and plant dry weight. Differences among combinations were found for all these parameters indicating that the level of mutualism is indeed affected by co-inoculation withTrichoderma and at different extent based on theTrichoderma species and the rhizobial strain. In particular, while all assessed phenotypes were modulated by the strain and/or the condition (Trichoderma ), plant dry weight and above ground plant length (shoot length and stem length) showed a statistically significant contribution of the interaction Trichoderma species x S. meliloti strain. (Supplementary Table S2 ). Going into the details of single combinations, (Figure 3, Supplementary Figure S2 ), plant dry weight was significantly higher when co-inoculated withT. velutinum -BL225C, T. harzianum -BL225C and T. tomentosum -BL225C compared to the other combinations. The highest value of plant dry weight was achieved in presence of the co-inoculationT. velutinum -BL225C (Figure 3d ). In terms of root length, the highest results were obtained with T. velutinum-S. meliloti and T. tomentosum-S. meliloti co-inoculations, although post-hoc Tukey Test showed no differences among the different groups (Supplementary Figure S2a ). This phenotype partially mirrors the auxins production profile (Figure 1 ), where these twoTrichoderma strains showed a less inhibiting power over. We may hypothesize that T. velutinum and T. tomentosum under the tested symbiotic conditions still allow S. meliloti to produce auxins as their spent media do under culture, giving rise to an increased root length . Regarding the number of nodules, values were significantly higher with the Trichoderma -AK83 combinations and the single S. meliloti AK83 strain. Indeed, this over-nodulating phenotype with AK83 has been already observed and interpreted in terms of higher capacity of competitiveness for nodule colonization , including possible cheating related to reduced nitrogen-fixation ability.
Summing up above-mentioned results, it is possible to distinguish combinations which were neutral (neither increasing nor decreasing the quality of symbiosis), positive (increase of the quality, indicating a synergistic interaction), and negative (decrease of the quality of symbiosis, viz. showing antagonistic interaction) with respect to M. sativa phenotypes.Figure 4 reports an evaluation of the modulation of mutualism by Trichoderma species with respect to the phenotypes with significant contribution of the interaction S. meliloti strain xTrichoderma species (Supplementary Table S2 ). Combinations displaying evidence for synergism, neutralism, and antagonism were present, again emphasizing a genotype-by-genotype effect on modulation of mutualism. Striking results were obtained for T. velutinum , which was antagonistic or nearly neutral for all strains but BL225C, which increased 2-fold the shoot length in combination withT. velutinum . Moreover, in general BL225C strain was the one showing for most of the combinations a synergistic effect ofTrichoderma co-inoculation.