Bioactivity-driven fungal metabologenomics identifies antiproliferative stemphone analogs and their biosynthetic gene cluster

Abstract

Introduction: Fungi biosynthesize chemically diverse secondary metabolites with a wide range of biological activities. Natural product scientists have increasingly turned towards bioinformatics approaches, combining metabolomics and genomics to target secondary metabolites and their biosynthetic machinery. We recently applied an integrated metabologenomics workflow to 110 fungi and identified more than 230 high-confidence linkages between metabolites and their biosynthetic pathways.

Objectives: To prioritize the discovery of bioactive natural products and their biosynthetic pathways from these hundreds of high-confidence linkages, we developed a bioactivity-driven metabologenomics workflow combining quantitative chemical information, antiproliferative bioactivity data, and genome sequences.

Methods: The 110 fungi from our metabologenomics study were tested against multiple cancer cell lines to identify which strains produced antiproliferative natural products. Three strains were selected for further study, fractionated using flash chromatography, and subjected to an additional round of bioactivity testing and mass spectral analysis. Data were overlaid using biochemometrics analysis to predict active constituents early in the fractionation process following which their biosynthetic pathways were identified using metabologenomics.

Results: We isolated three new-to-nature stemphone analogs, 19-acetylstemphones G (1), B (2) and E (3), that demonstrated antiproliferative activity ranging from 3 to 5 µM against human melanoma (MDA-MB-435) and ovarian cancer (OVACR3) cells. We proposed a rational biosynthetic pathway for these compounds, highlighting the potential of using bioactivity as a filter for the analysis of integrated-Omics datasets.

Conclusions: This work demonstrates how the incorporation of biochemometrics as a third dimension into the metabologenomics workflow can identify bioactive metabolites and link them to their biosynthetic machinery.

Keywords: Biochemometrics; Biosynthesis; Fungi; Metabologenomics; Natural products; Secondary metabolism.

Fig. 1

Bioactivity-guided metabologenomics workflow for prioritizing bioactive natural product-BGC pairs for targeted downstream analyses

Fig. 2

S-plots from PLS models of antiproliferative activity of A NRRL 5080 extract and fractions (2 component model, 93.99% independent, 99.42% dependent variation explained). B NRRL 5071 (2 component model, 92.20% independent, 95.51% dependent), and C NRRL 5074 (5 component model, 94.64% independent, 98.88% dependent). The upper right quadrants show the ions with the highest correlation to bioactivity. Compounds 1 and 2 were identified in all three S-plots among the top 10 contributors to biological activity. Points highlighted in blue were not purified, but clustered in the same molecular family as purified compounds. D Molecular family of 19-acetylstemphones detected by networking related MS² spectra. Nodes colored in pink correspond to compounds 1–3. Nodes highlighted in blue were not purified, but were identified through S-plot analysis as contributing to the observed antiproliferative activity and may represent additional bioactive stemphone analogs

Fig. 3

Structures of 19-acetylstemphone G (1), 19-acetylstemphone B (2), and 19-acetylstemphone E (3)

Fig. 4

Proposed biosynthesis of 19-acetylstemphones. In final structures, R=H for stemphones and R=COCH₃ for acetylstemphones