Tackling antibiotic resistance: a dual-criterion computational filter for selective MexAB-OprM efflux pump inhibitors
- 11Institut Supérieur Mohammed VI de Biosciences et Biotechnologies, Université Mohammed VI des Sciences et de la Santé
Abstract
Drug repositioning against bacterial efflux pumps has classically relied on a single criterion: the predicted binding affinity for the target transporter. This approach overlooks a distinct failure modetranscriptional derepression of the pump's own expression, through inhibition of the associated transcriptional regulator. Here we propose a dual-criterion screening protocol, combining (i) a relative selectivity index between affinity for the MexB pump and affinity for its repressor MexR, and (ii) an absolute risk threshold calibrated on a documented derepression inducer (pyocyanin). The method was tested by molecular docking (AutoDock Vina) on 139 compounds spanning four independent chemical families (antidepressants, antipsychotics, H1 antihistamines, phytomolecules), against the MexB (PDB 3W9J) and MexR structures. The ΔG(MexB)–ΔG(MexR) correlation proved robust and generalizable across the four families (r = 0.834–0.924, all p < 10-7 after FDR correction), validated by bootstrap, permutation testing, and leave-one-out analysis. Controlling for molecular weight and LogP across the full dataset (n = 135) confirms a partial but non-exclusive contribution of molecular size to this correlation (partial r = 0.805 vs. raw r = 0.880). The dual-criterion filter revealed that 43.1% of the tested compounds, despite showing a good relative selectivity index, exceed the absolute risk thresholdamong them the historical reference compound PAβN and the previously proposed candidate vilazodone. Ensemble docking on three MexR structures (1LNW, 3ECH, 3MEX), performed on a subset of 14 priority compounds, quantified the associated structural uncertainty (mean standard deviation 0.512 kcal/mol): this test confirms the robustness of vilazodone's at-risk status, but reveals that a candidate initially classified as favorable (mianserin) becomes at-risk once this uncertainty is taken into account. These results indicate that filtering by relative index alone is insufficient to rule out derepression risk, and propose a reproducible methodological framework, including an uncertainty estimate, for drug repositioning in the context of combination therapy.
Related
- Heliocide H2 and cadinatriene glucopyranoside as promising dual β-lactamase inhibitors against antibiotic resistance: an integrated computational study
- Targeting the CDK4-Cyclin D Complex: A New Generation of Selective Kinase Inhibitors for Cancer Therapy
- Knowledge and misconceptions about antibiotic use and antibiotic resistance among secondary school students in Riyadh, Saudi Arabia: implications for school- based antibiotic resistance education
- Beliefs and Behaviours related to Antibiotic Use and Antimicrobial Resistance in Poland: General Population Survey
- Collateral effects of bacteriophage resistance on antibiotic sensitivity in Pseudomonas aeruginosa
BibTeX
@misc{mbanimi2026tackling,
title = {{Tackling antibiotic resistance: a dual-criterion computational filter for selective MexAB-OprM efflux pump inhibitors}},
author = {Divin Stoni MBANIMI},
howpublished = {ChemRxiv},
year = {2026},
month = sep,
doi = {10.26434/chemrxiv.15007681/v2},
url = {https://chemrxiv.org/doi/full/10.26434/chemrxiv.15007681/v2},
note = {CC BY},
}
RIS
TY - UNPB
TI - Tackling antibiotic resistance: a dual-criterion computational filter for selective MexAB-OprM efflux pump inhibitors
AU - Divin Stoni MBANIMI
PY - 2026
DA - 2026/09/02
PB - ChemRxiv
DO - 10.26434/chemrxiv.15007681/v2
UR - https://chemrxiv.org/doi/full/10.26434/chemrxiv.15007681/v2
AB - Drug repositioning against bacterial efflux pumps has classically relied on a single criterion: the predicted binding affinity for the target transporter. This approach overlooks a distinct failure modetranscriptional derepression of the pump's own expression, through inhibition of the associated transcriptional regulator. Here we propose a dual-criterion screening protocol, combining (i) a relative selectivity index between affinity for the MexB pump and affinity for its repressor MexR, and (ii) an absolute risk threshold calibrated on a documented derepression inducer (pyocyanin). The method was tested by molecular docking (AutoDock Vina) on 139 compounds spanning four independent chemical families (antidepressants, antipsychotics, H1 antihistamines, phytomolecules), against the MexB (PDB 3W9J) and MexR structures. The ΔG(MexB)–ΔG(MexR) correlation proved robust and generalizable across the four families (r = 0.834–0.924, all p < 10-7 after FDR correction), validated by bootstrap, permutation testing, and leave-one-out analysis. Controlling for molecular weight and LogP across the full dataset (n = 135) confirms a partial but non-exclusive contribution of molecular size to this correlation (partial r = 0.805 vs. raw r = 0.880). The dual-criterion filter revealed that 43.1% of the tested compounds, despite showing a good relative selectivity index, exceed the absolute risk thresholdamong them the historical reference compound PAβN and the previously proposed candidate vilazodone. Ensemble docking on three MexR structures (1LNW, 3ECH, 3MEX), performed on a subset of 14 priority compounds, quantified the associated structural uncertainty (mean standard deviation 0.512 kcal/mol): this test confirms the robustness of vilazodone's at-risk status, but reveals that a candidate initially classified as favorable (mianserin) becomes at-risk once this uncertainty is taken into account. These results indicate that filtering by relative index alone is insufficient to rule out derepression risk, and propose a reproducible methodological framework, including an uncertainty estimate, for drug repositioning in the context of combination therapy.
ER -