Discovery of novel bioactive metabolites from marine bacteria is becoming increasingly challenging, and the development of novel approaches to improve the efficiency of early steps in the microbial drug discovery process is therefore of interest. For example, current protocols for the taxonomic dereplication of microbial strains generally use molecular tools which do not take into consideration the ability of these selected bacteria to produce secondary metabolites. As the identification of novel chemical entities is one of the key elements driving drug discovery programs, this study reports a novel methodology to dereplicate microbial strains by a metabolomics approach using liquid chromatography-high resolution mass spectrometry (LC-HRMS). In order to process large and complex three dimensional LC-HRMS datasets, the reported method uses a bucketing and presence-absence standardization strategy in addition to statistical analysis tools including principal component analysis (PCA) and cluster analysis. From a closely related group of Streptomyces isolated from geographically varied environments, we demonstrated that grouping bacteria according to the chemical diversity of produced metabolites is reproducible and provides greatly improved resolution for the discrimination of microbial strains compared to current molecular dereplication techniques. Importantly, this method provides the ability to identify putative novel chemical entities as natural product discovery leads.
Keywords: Actinobacteria; Chemical dereplication; Cluster analysis; Metabolomics; Natural products.
Copyright © 2013 Elsevier B.V. All rights reserved.