Natural products remain a cornerstone in the search for new therapeutics, particularly in the context of antimicrobial drug discovery. The alarming rise of multidrug-resistant bacterial infections has created an urgent need for novel antibiotics. Fungi, with their vast and largely unexplored biosynthetic potential, represent a promising but underutilized resource for such compounds. However, traditional approaches to fungal extract generation and screening are labor-intensive and low-throughput, limiting the pace of discovery.
This thesis aimed to develop and implement a high-throughput platform for the efficient exploration of fungal chemical diversity, with the overarching goal of identifying novel bioactive metabolites. The strategy integrates miniaturized fungal cultivation, MS-based metabolomics, and innovative antimicrobial screening into a streamlined and scalable workflow. Public databases and in-house dereplication tools were leveraged to prioritize a subset of underexplored strains. These strains were cultured under standardized yet flexible conditions to rapidly produce crude extracts, which were subsequently evaluated for antimicrobial activity. Metabolomic analyses and application of an innovative data treatment, called Bioscore, enabled the identification of features associated with bioactivity, guiding the selection of promising strains for scale-up and targeted compound isolation. Active metabolites were purified and structurally elucidated by 1D and 2D NMR spectroscopy, confirming the discovery of novel chemical entities.
To further expand chemical diversity, an OSMAC (One Strain–Many Compounds) approach was applied to over 50 strains under varying culture conditions, generating approximately 1,600 extracts. These were systematically analyzed by LC-HRMS/MS and tested in a panel of biological assays, enabling a broad mapping of chemical and bioactivity landscapes. This large-scale effort validated the performance of the developed screening platform, FLECS-96, which dramatically outperformed conventional methods in both throughput and depth of analysis. The complexity and volume of data generated in this work necessitated the development of novel integration tools. A sample-centric Knowledge Graph, MycoMedix, was constructed to facilitate the interrogation and visualization of chemical, taxonomic, and biological information. This scalable and iterative database provides a valuable resource for fungal natural product discovery and prioritization.
Finally, the flexibility of the platform was demonstrated by extending its application beyond antibacterial discovery to include anti-RSV and anticancer screening. Altogether, this thesis presents an efficient, modular, and reproducible workflow that enables the systematic exploration of fungal biodiversity for drug discovery and lays the groundwork for future large-scale metabolomic studies.
