Modulating Complex Secondary Metabolism in Streptomyces rimosus by Targeted Genome Engineering

Numerous biosynthetic gene clusters (BGCs) encoding unknown structures have been discovered in the genomes of diverse microorganisms, representing a potentially rich source of novel natural products. However, most of the identified BGCs do not seem to be active, since we cannot detect any corresponding metabolites. Therefore, a better understanding of the regulation and biosynthesis of secondary metabolites encoded by these so-called 'silent' BGCs is of great importance.

We conducted a bioinformatic analysis of the Streptomyces rimosus ATCC 10970 strain, a producer of the antibiotic oxytetracycline, focusing on the expression of identified BGCs. We then reviewed experimentally identified compounds and putative structures predicted from genome data and similarity to known metabolites. We analysed available data on the regulation of two major metabolites - oxytetracycline and rimocidin, and experimentally evaluated the effect of the deletion of two oxytetracycline-competing pathways. Finally, we evaluated the effect of overexpressing BGC encoding the biosynthesis of the carotenoid isorenieratene, which cannot be detected in the culture of the native strain.

We identified 48 BGCs in the genome of Streptomyces rimosus ATCC 10970. However, only about 15 structures were predicted or identified in the culture of this strain. Transcriptional analysis of identified BGCs demonstrated a very high variability in expression strength. Interestingly, around 30 % of BGCs were 'silent'. In trans overexpression of one such silent BGC, encoding the biosynthesis of the carotenoid isorenieratene resulted in strong production of this metabolite, suggesting that silent BGCs are likely still functional. We also demonstrated that BGCs encoding two major metabolites, oxytetracycline and rimocidin, both derived from malonyl-coenzyme A (malonyl-CoA), are not competitive pathways. Surprisingly, deletion of one silent BGC, also derived from malonyl-CoA, has a very strong effect on the biosynthesis of oxytetracycline.

We observed that the expression strength of genes from BGCs identified in Streptomyces rimosus does not correspond to the experimental data obtained from the engineered strains, suggesting much more complex regulatory mechanisms than previously thought. Engineered Streptomyces rimosus host strains thus represent a very good model system to study the expression of 'silent' BGCs.