TY - JOUR T1 - Phylogenomic analysis of natural products biosynthetic gene clusters allows discovery of arseno-organic metabolites in model streptomycetes JF - bioRxiv DO - 10.1101/020503 SP - 020503 AU - Pablo Cruz-Morales AU - Johannes Florian Kopp AU - Christian Martinez-Guerrero AU - Luis Alfonso Yáñez-Guerra AU - Nelly Selem Mojica AU - Hilda Ramos-Aboites AU - Jörg Feldmann AU - Francisco Barona-Gómez Y1 - 2016/01/01 UR - http://biorxiv.org/content/early/2016/02/29/020503.abstract N2 - Natural products from microbes have provided humans with beneficial antibiotics for millennia. However, a decline in the pace of antibiotic discovery exerts pressure on human health as antibiotic resistance spreads, a challenge that may better faced by unveiling chemical diversity produced by microbes. Current microbial genome mining approaches have revitalized research into antibiotics, but the empirical nature of these methods limits the chemical space that is explored.Here, we address the problem of finding novel pathways by incorporating evolutionary principles into genome mining. We recapitulated the evolutionary history of twenty-three enzyme families previously uninvestigated in the context of natural product biosynthesis in Actinobacteria, the most proficient producers of natural products. Our genome evolutionary analyses where based on the assumption that expanded-repurposed enzyme families-from central metabolism, occur frequently and thus have the potential to catalyze new conversions in the context of natural products biosynthesis. Our analyses led to the discovery of biosynthetic gene clusters coding for hidden chemical diversity, as validated by comparing our predictions with those from state-of-the-art genome mining tools; as well as experimentally demonstrating the existence of a biosynthetic pathway for arseno-organic metabolites in Streptomyces coelicolor and Streptomyces lividans, using gene knockout and metabolite profile combined strategy. As our approach does not rely solely on sequence similarity searches of previously identified biosynthetic enzymes, these results establish the basis for the development of an evolutionary-driven genome mining tool that complements current platforms. We anticipate that by doing so real ‘chemical dark matter’ will be unveiled. ER -