Antimicrobial resistance continues to pose a threat to health and wellbeing. Unmitigated, it is predicted to be the leading cause of death by 2050. Hence, the sustained development of novel antibiotics is crucial. As over 60% of licensed antibiotics are based on scaffolds derived from less than 1% of all known bacterial species, bacterial secondary metabolites constitute an untapped source of novel antibiotics. The aim of this project therefore was to expand the chemical space of bacteria-derived antibiotic lead compounds, using genomics approach. To that end, a topsoil sample was collected from the rhizosphere in which antibiosis occurs naturally. Using starvation stress, sixty-five isolates were recovered from the sample, out of which four were selected based on morphology and designated A13BB, A23BA, A13AA and A23AA. A13BB was identified by 16S rRNA gene sequence comparison as a Pseudomonas spp. and the other three isolates as Hafnia/Obesumbacterium spp. A database search showed that species belonging to these genera have genomes larger than the 3 Mb size above which an increasing proportion of a bacterial genome is dedicated to secondary metabolism. Given their ecological origin, expected genome size and ability to withstand starvation stress, these four isolates were presumed to harbour antibiotic-encoding gene clusters. Isolates A13BB and A23BA were therefore selected for genome mining in the first instance. Illumina and GridION/MinION sequencing data were obtained for both isolates and assembled into high-quality genomes. Isolates' identities were confirmed by FastANI analysis as strains of P. fragi and H. alvei, with 4.94 and 4.77 Mb genomes, respectively. Assembled genomes were mined with antiSMASH. Amongst other secondary metabolite biosynthetic gene clusters (smBGCs) detected, the β-lactone smBGCs in both genomes were selected for activation as their end products bear the hallmarks of an 'ideal antibiotic' that can inhibit several bacteria-specific enzymes simultaneously. Analysis of these smBGCs revealed genes encoding two core enzymes: 2-isopropylmalate synthase (2-IPMS) and acyl CoA ligase homologues. In the biosynthetic pathway, 2-IPMS catalyses the condensation of acetyl CoA with the degradation product of valine or isoleucine to form 2-IPM. 2-IPM is isomerised to 3-IPM which then forms the β-lactone warhead through reactions catalysed by acyl CoA ligase. It was speculated that the β-lactone compound is biosynthesised to efficiently rid the organism of potentially harmful metabolic intermediates as it grows on poor carbon and nitrogen sources. Strain fermentation was therefore performed with 10.8 mM acetate as the main carbon source, and 5 mM L-valine or L-isoleucine as the nitrogen source. Fermentation extracts were analysed by LC-MS with at least thirty-seven metabolite ions detected. Many of these ions have masses in the range m/z 230-750, which is an ideal mass range for antibiotic molecules. As β-lactone compounds are difficult to identify in crude extracts, especially when utilising single-stage mass spectrometry, reactivity-guided screening of extracts with cysteine thiol probe was performed as the probe forms UV- and MS-visible adducts with β-lactone compounds. However, complete dimerization of probe at a faster-than-expected rate in extract matrices hindered successful screening. This meant that it was not possible to determine if any crude extract components were β-lactone compounds without further analysis. Measures to limit or eliminate probe dimerization are proposed, together with molecular networking strategies that can afford global visualisation and rapid dereplication of extract components, using tandem mass spectrometry fragmentation patterns of parent ions. This project provides an original and robust workflow that serves as a strong starting point in the isolation of novel β-lactone compounds from crude extracts, followed by structural optimisation and bioactivity profiling. The hitherto unrecognised potential of β-lactone natural compounds as 'ideal antibiotics' is highlighted, and several structural optimisation strategies required to harness this potential are proposed. The genomes assembled here, and associated data have been deposited in the repositories of the International Nucleotide Sequence Database Collaboration for repurposing by other researchers. Likewise, the hidden metabolic and biosynthetic potentials of P. fragi and H. alvei species uncovered by RASTtk and antiSMASH analyses have been catalogued and placed in the public domain, with many of these attributes reported for the first time.
AWOLOPE, O.K. 2023. Genome-guided bioprospecting for novel antibiotic lead compounds. Robert Gordon University, PhD thesis. Hosted on OpenAIR [online]. Available from: https://doi.org/10.48526/rgu-wt-1987833