Artificially selected rhizosphere microbiota modify plant growth in a soil-independent and species-dependent way

Aims: Artificial selection of microbiota is an innovative approach to steer plant phenotype. Still, driving ecological determinants governing the success of this approach are yet to be characterized. We aimed to test the importance of the plant species and the soil type on the effects of selected microbiota. Methods: We previously selected rhizosphere microbiota associated to Brachypodium distachyon leaf greenness in a sandy cambisol. Three selection treatments were applied: low greenness, high greenness, and random (control). We tested the reproducibility of selected effects by inoculating the evolved microbiota on different Poaceae species (B. distachyon, maize, wheat and oat), either in the sandy cambisol used for selection, or in a clayey calcaric stagnosol. Results: Maize and wheat were not affected. B. distachyon greenness was altered in the intended way, while oat showed significant, but opposite, phenotypic changes. These effects were reproducible in the two different soils. The rhizosphere microbiota of B. distachyon and oat were significantly impacted by selected inoculants. We identified microbial sub-communities either positively or negatively associated with greenness, revealing the presence of a soil-independent microbial sub-community that is always associated with higher greenness in B. distachyon. Conclusions: The effect of selected inoculants was dependent on plant species, but not on the soil type. Thus, despite plant species dependency, plant phenotype changes induced by selected microbiota seemed transferable under different pedological contexts. This may have consequences for plant breeding programs that include microbiota selection, as we show that the plant species dependence is paramount for success across different soils.