Multiplexed activity metabolomics for isolation of filipin macrolides from a hypogean actinomycete

The Journal of Antibiotics https://doi.org/10.1038/s41429-024-00792-6 ARTICLE Multiplexed activity metabolomics for isolation of filipin macrolides from a hypogean actinomycete Jordan T. Froese1 Joseph A. Balsamo2 Benjamin J. Reisman3 Sierra M. Barone4 Jonathan M. Irish Brian O. Bachmann 2,3,7 ● ● ● ● 4,5,6 ● 1234567890();,: 1234567890();,: Received: 16 August 2024 / Revised: 11 October 2024 / Accepted: 18 November 2024 © The Author(s) 2024. This article is published with open access Abstract Chemical and biological stimulus screening in a hypogean actinomycete was used to elicit secondary metabolism. Optimal biosynthesis of bioactive natural products was identified using Multiplexed Activity Profiling for determining dosedependent activity via six single-cell biological readouts. Bioactive extracts were fractioned to establish candidate compounds for isolation using Multiplexed Activity Metabolomics by correlating microtiter well-isolated phenotypes and extracted ion current peaks. This guided the isolation of four filipin polyene macrolides including a new metabolite filipin XV, an alkyl side-chain hydroxylated congener of the filipin chainin, with substantially attenuated cytotoxicity. Filipinspecific cytotoxicity was confirmed using flow cytometry and fluorescence microscopy. Introduction Microorganisms across taxa are prolific for their potential to produce multiple unique secondary metabolites per organism, a potential that can now be quantified by automated These authors contributed equally: Jordan T. Froese, Joseph A. Balsamo Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41429024-00792-6. * Brian O. Bachmann 1 Department of Chemistry, Ball State University, Muncie, IN, USA 2 Department of Pharmacology, Vanderbilt University, Vanderbilt University School of Medicine, Nashville, TN, USA 3 Program in Chemical & Physical Biology, Vanderbilt University School of Medicine, Nashville, TN, USA 4 Vanderbilt-Ingram Cancer Center, Vanderbilt University Medical Center, Nashville, TN, USA 5 Department of Cell & Developmental Biology, Vanderbilt University, Nashville, TN, USA 6 Department of Pathology, Microbiology, & Immunology, Vanderbilt University Medical Center, Nashville, TN, USA 7 Department of Chemistry, Vanderbilt University, Nashville, TN, USA analysis of microbial genomic and metagenomic sequences [1], and the majority of which remain to be isolated [2]. However, activating the production of predicted secondary metabolite gene clusters in microbes, isolating metabolites, and determining their structures and biological functions remain rate-limiting steps in fully capitalizing on the genome-evidenced potential of microbes. An array of techniques substantially increases cryptic and relatively silent secondary metabolite biosynthesis in microorganisms and aid in their identification. Genetic refactoring in native organisms [3, 4] and/or heterologous expression in chassis organisms [5] is increasingly attempted, though this approach is sometimes limited by the availability of practicable genetic techniques for genomic manipulation in producing strains, potential complications in regulatory element compatibility, precursor supply, and metabolite toxicity in a given heterologous expressing chassis organism. For many years, simply varying production media compositions has yielded improvements in production of microbial secondary metabolites [6]. More recently, introducing discrete chemical and biological stimuli into a single medium can be used to identify conditions to upregulate production of secondary metabolites within metabolomes and to facilitate their identification via comparative metabolomics. These stimuli include culturing actinomycetes with sub-inhibitory concentrations of antibiotics [7, 8], toxic rare earth elements, mixed culture with competing/interacting organisms, and de-repression via selection of J. T. Froese et al. antibiotic resistance associated with transcription and translation apparatus. Development of high throughput metabolomic methodologies enabled screening of hundreds of stimulus conditions for activation of secondary metabolism in a given producer and comparative metabolomics has facilitated the rapid assessment of differences among perturbed systems to aid in identification of novel secondary metabolites in an activity-independent fashion [4, 9–11]. Through these methods, and providing sufficient effort, the products of the majority of previously cryptic gene clusters identified by genomic inference can be isolated. In these workflows, determination of activity or biological function often occurs after the isolation of upregulated, sufficiently abundant, and chromatographically tractable metabolites. However, it is not uncommon that hard-won new metabolites lack identifiable activity, representing an inherent inefficiency in such workflows. To address the gap between abundance-guided discovery of variously activated secondary metabolism and determination of potential biological relevance before isolation, this work provides a methodology for discovery of bioactive secondary metabolites by measuring biological responses in single cells within the presence of chemically complex microbial metabolomes. The quantitative single cell chemical biological activity metabolomics [12] tools described herein leverage fluorescent flow cytometry to efficiently identify bioactive microbial extracts, metabolomic subfractions, and compounds therein that impact cell injury and cell death in human-derived cell lines. These assays, termed Multiplexed Activity Profiling (MAP) [13] and Multiplexed Activity Metabolomics (MAM) [14], evaluate single cell responses to metabolite challenge using fluorescent cell barcoding [15, 16] and fluorescently labeled antibody functional readouts. The MAP assay is used to screen for dosedependent bioactivity in crude extracts while the MAM assay is used to fractionate bioactive extracts into a comprehensive metabolomic array to reveal the bioactive component(s) within the complex chemical mixture. Herein, this MAM platform performs 228 individual assays per extract metabolome comprising 40 fractionated wells, eight control wells, five cell functional readouts, and overall cytotoxicity. Chemically dependent responses in single cells are acquired using flow cytometry and the biological data are deconvoluted with the DebarcodeR algorithm to enhance and automate computation of biological activity [17]. This workflow can be used for diverse applications in the discovery of bioeffectors from virtually any source ranging from compound libraries to the products of primary and secondary metabolism. To demonstrate this workflow herein, an actinomycete sourced from Blue Spring cave (Sparta, TN) was isolated by culturing with selective media [18, 19] and cultivated in the presence of multiple chemical and biological stimulus to optimize conditions for secondary metabolit (...truncated)