Routes to inhibitors of colistin antibiotic resistance mediated by bacterial phosphoethanolamine transferases

: computational and microbiological investigations

Abstract

As one of the few remaining reliable antibiotics, colistin retains effectiveness against most multi-drug resistant (MDR) Gram-negative strains and is considered as the last resort to treat infections by these organisms. The mobile colistin resistance gene mcr-1 was identified in 2015 in Escherichia coli and encodes a zinc-dependent phosphoethanolamine (PEtN) transferase that modifies lipid A of the bacterial outer membrane to reduce colistin affinity. One option to overcome colistin resistance is to combine colistin with inhibitors of colistin resistance. This project aims to use a combination of computational and experimental approaches to identify MCR-1 inhibitors to restore sensitivity of producer strains to colistin and related agents.
Virtual screening against the MCR-1 catalytic domain was carried out using the Bristol University Docking Engine (BUDE). Biochemical and biophysical assays were developed to investigate protein:ligand interactions and demonstrated interaction of the small thiol, thioglycolic acid (TGA) with the MCR-1 catalytic domain. 35 candidate compounds including top ranked ligands from BUDE and reported inhibitors of other zinc metalloproteins were tested in these and colistin susceptibility assays. Although in vitro assays did not give clear results, compound C4 (4-iodoisoxazole) was identified as reducing the colistin minimum inhibitory concentration (MIC) of MCR-1 producing E. coli and Klebsiella pneumoniae strains. Compound C4 also potentiated colistin activity against ArnT mediated colistin-resistant K. pneumoniae and meropenem activity against carbapenem-resistant E.coli producing the metallo-β-lactamase IMP-1. Compound C4 did not affect the cell membrane integrity/permeability of E.coli cells, but induced a colistin heteroresistance phenotype in K. pneumoniae strains with ArnT mediated colistin resistance.
A multiscale workflow to model zinc metalloprotein:ligand complex structures, including molecular mechanics (MM) molecular dynamics (MD) and quantum mechanics (QM)/MM MD simulations, was established. The workflow was extensively tested and accurately modelled a variety of tested complexes (Sfh-I, HDAC2, IMP-1, etc.) and applied to MCR-like complex systems including the MCR-1:TGA complex. Modelling results were consistent with current experimental findings, indicating low affinity of TGA for the mono-zinc form of MCR-1.
This project developed an experimental and computational pipeline to screen and validate inhibitors and study their interactions with MCR-1. An inhibitor of MCR-1-mediated colistin resistance, compound C4, was identified and investigated, and a modelling workflow developed with potential general application to zinc metalloproteins. These findings provide tools and knowledge useful to understanding of the MCR-1 zinc site and discovery of inhibitors for MCR-1 like PEtN transferases able to preserve colistin efficacy against MDR Gram-negative bacterial infections.

Date of Award	21 Jun 2022
Original language	English
Awarding Institution	University of Bristol
Supervisor	Jim Spencer (Supervisor) & Adrian J Mulholland (Supervisor)