Abstract
Understanding plant-microbe interactions with the possibility to modulate the plant's microbiome is essential to design new strategies for a more productive and sustainable agriculture and to maintain natural ecosystems. Therefore, a key question is how to design bacterial consortia that will yield the desired host phenotype. This work was designed to identify the potential genomic features involved in the interaction between Micromonospora and known host plants. Seventy-four Micromonospora genomes representing diverse environments were used to generate a database of all potentially plant-related genes using a novel bioinformatic pipeline that combined screening for microbial-plant related features and comparison with available plant host proteomes. The strains were recovered in three clusters, highly correlated with several environments: plant-associated, soil/rhizosphere, and marine/mangrove. Irrespective of their isolation source, most strains shared genes coding for commonly screened plant growth promotion features, while differences in plant colonization related traits were observed. When Arabidopsis thaliana plants were inoculated with representative Micromonospora strains selected from the three environments, significant differences were in found in the corresponding plant phenotypes. Our results indicate that the identified genomic signatures help select those strains with the highest probability to successfully colonize the plant and contribute to its wellbeing. These results also suggest that plant growth promotion markers alone are not good indicators for the selection of beneficial bacteria to improve crop production and the recovery of ecosystems.