RT Journal Article SR Electronic T1 Recapitulation of the evolution of biosynthetic gene clusters reveals hidden chemical diversity on bacterial genomes JF bioRxiv FD Cold Spring Harbor Laboratory SP 020503 DO 10.1101/020503 A1 Pablo Cruz-Morales A1 Christian E. Martínez-Guerrero A1 Marco A. Morales-Escalante A1 Luis Yáñez-Guerra A1 Johannes Florian Kopp A1 Jörg Feldmann A1 Hilda E. Ramos-Aboites A1 Francisco Barona-Gómez YR 2015 UL http://biorxiv.org/content/early/2015/06/08/020503.abstract AB Natural products, which result from secondary or specialized metabolism, have provided with molecules to human welfare for millennia. However, decline of chemical discovery pace has imposed a pressure upon human health, such as in antibiotic resistance. Current genome mining approaches have revitalized research into natural products, but the empirical nature of these methods limits the chemical space that is explored. By means of integrating evolutionary concepts related to emergence of specialized metabolism, we have gained fundamental insights that are translated into the discovery of hidden chemical diversity through a unique and unbiased genome mining approach. This method, termed EvoMining, can be defined as a functional phylogenomics platform for identification of expanded, repurposed enzyme families, with the potential to catalyze new conversions. A bioinformatics pipeline is proposed and validated by comparing its performance with the state-of-the-art genome mining approach antiSMASH. Moreover, as the founding assumption of EvoMining relates to the evolution of enzyme function, our approach was experimentally validated after solving two milestone problems that include unprecedented enzyme conversions. First, we report the discovery of a biosynthetic gene cluster for an orphan metabolite, which could not be unveiled with current methods, i.e. the biosynthesis of the protease inhibitor leupeptin by Streptomyces roseus ATCC 31245. Second, we characterized a novel enzyme, catalyzing the formation of an arsenic-carbon bond, in model organisms that have been thoroughly mined, i.e. Streptomyces coelicolor and Streptomyces lividans. This work provides evidence that bacterial chemical repertoire is still underexploited, as well as an alternative approach that promises to speed up the discovery of novel enzymes and biosynthetic logics that can feedback into current genome mining methods.