Abstract by Jianzhong Zhu

Natural products have long served as an important source of medicine for humanity, with notable examples including aspirin, penicillin, taxol, and artemisinin. However, natural product-based drug discovery faces numerous challenges. These include the presence of unculturable microorganisms in natural environments, the absence of secondary metabolite production in laboratory settings due to silent biosynthetic gene clusters, difficulties in isolation and purification, the frequent re-discovery of known compounds, challenges in identifying active compounds, and the complexity of structure elucidation. Against the global backdrop of bacterial and cancer multidrug resistance, this thesis aims to discover new natural products with antibacterial or anticancer potential through concerted effort to address several of the challenges above.

Firstly, the chemical repository of an Australian desert-adapted bush, Eremophila platycalyx subsp. woolgorong, was investigated. Chemical extraction was carried out on its root bark, followed by extensive isolation and purification, which yielded a suite of twelve new compounds. Among them, compound 2 (eremoplawone A) was found to possess an unprecedented chemical scaffold, while compound 11 or 14 (eremoplawone I or eremoplawone L) exhibited promising antibacterial activity against Staphylococcus aureus. The structure elucidation of compound 9 (eremolongine B) demonstrated the effectiveness of density functional theory calculations in determining the absolute configuration of organic compounds.

Next, we investigated the endophytes present in the roots of Eremophila spp. Nine endophytes were cultured under laboratory conditions with various stimulatory strategies, including the one strain many compounds approach, epigenetic manipulation, and bacterium–fungus co-cultivation. The fungus ILF-E001 co-cultured with S. aureus exhibited the strongest antibacterial activity, which was attributed to the production of bikaverin. This discovery was enabled by the combined use of biochemometric analysis—which differentiates active from inactive compounds—and high-resolution microfractionation. A workflow integrating the strengths of both techniques, along with dereplication, was established to facilitate the discovery of new bioactive natural products.

Building on the experience gained from the previous study involving the biochemometric analysis script, we took a further step by predicting the relative potency among active compounds using the concept of effect size in the context of anticancer activity from a cyclotide-rich crude extract of Viola odorata. The predicted potency was consistent with experimental observations for the active components, cycloviolacin O3 and O8. These cyclotides exerted their anticancer effects through cell membrane disruption, and their solution structures were elucidated by NMR spectroscopy.

Finally, we discovered a new cyclotide, cycloviolacin O37, from V. odorata. Its sequence was determined through liquid chromatography–mass spectrometry sequencing combined with transcriptome mining. The isolated form of cycloviolacin O37 appeared as a mixture of diastereomers, resulting from a methionine oxidation. The solution structures of both diastereomers were resolved using NMR. Unfortunately, cycloviolacin O37 did not exhibit antibacterial or anticancer activity. However, it is rich in lysine and possesses unique loop sequences, distinguishing it from all previously known cyclotides.