Abstract
The genus Streptomyces is known to harbor numerous biosynthetic gene clusters (BGCs) of potential utility in synthetic biology applications. However, it is often difficult to link uncharacterized BGCs with the secondary metabolites they produce. Proteomining refers to the strategy of identifying active BGCs by correlating changes in protein expression with the production of secondary metabolites of interest. In this study, we devised a shotgun proteomics-based workflow to identify active BGCs during fermentation when a variety of compounds are being produced. Mycelia harvested during the non-producing growth phase served as the background. Proteins that were differentially expressed were clustered based on the proximity of the genes in the genome to highlight active BGCs systematically from label-free quantitative proteomics data. Our software tool is easy-to-use and requires only 1 point of comparison where natural product biosynthesis was significantly different. We tested our proteomining clustering method on three Streptomyces species producing different compounds. In Streptomyces coelicolor A3(2), we detected the BGCs of calcium-dependent antibiotic, actinorhodin, undecylprodigiosin, and coelimycin P1. In Streptomyces chrestomyceticus BCC24770, 7 BGCs were identified. Among them, we independently re-discovered the type II PKS for albofungin production previously identified by genome mining and tedious heterologous expression experiments. In Streptomyces tenebrarius, 5 BGCs were detected, including the known apramycin and tobramycin BGC as well as a newly discovered caerulomycin A BGC in this species. The production of caerulomycin A was confirmed by LC-MS and the inactivation of the caerulomycin A BGC surprisingly had a significant impact on the secondary metabolite regulation of S. tenebrarius. In conclusion, we developed an unbiased, high throughput proteomics-based method to complement genome mining methods for the identification of biosynthetic pathways in Streptomyces sp.