Screening of Machine Learning Predicted Inhibitors of Mur‐E Ligase

Abstract

Antibiotic resistance is an ever‐growing threat, which necessitates the development of novel antibiotics. Peptidoglycan (PTG) is an ideal target for antibiotic development as it is necessary for the structural integrity of the cell and is unique to bacterial cells. Mur‐E, a cytoplasmic ligase, is an integral component of PTG synthesis and shares many conserved structures with PTG ligases in other bacterial species. Thus, inhibitors of Mur‐E are likely to be effective broad‐spectrum antibiotics. Mur‐E's role in PTG synthesis is to add meso‐A2pm or L‐lysine to a nucleotide precursor of PTG, UMAG; this reaction is facilitated by ATP hydrolysis. Residues N449, R451, K157, E220, D392 were identified in a mutagenesis study as having significant impacts on enzyme kinetics. This project combines the knowledge of each residue’s known impact on kinetics with an analytical, dynamics‐based investigation of protein structure using Essential Sit Scanning Analysis (ESSA) and ClustENMD. These two methodologies implement elastic network models of proteins to predict allosteric pockets and alternate conformations of the protein, respectively. This will give insight into which pockets are likely to be effective drug targets based on their impact on enzyme kinetics and protein dynamics; these pockets will be evaluated using a library of predicted ligands for Mur‐E to identify potential inhibitors with AutoDock Vina. By evaluating ligand binding sites and affinities, probable antibiotics can be determined. Targeting residue pockets that facilitate key Mur‐E functions, like ATP binding and substrate addition, will produce the most effective antibiotic treatment by compromising PTG production.