Rational Design and In Silico Evaluation of a Dual-Action Antimicrobial Peptide Targeting Efflux Pump and Membrane...

The increasing prevalence of multidrug-resistant (MDR) Pseudomonas aeruginosa, largely associated with overexpression of the MexAB-OprM efflux pump, highlights the need for alternative antimicrobial strategies. In this study, antimicrobial peptide (AMP) candidates were rationally designed using the Artificial Neural Network (ANN) and Random Forest (RF) models implemented in the CAMPR4 platform. Approximately 18,000 AMP sequences were screened, resulting in nine candidate peptides that were subsequently evaluated for physicochemical properties, predicted hemolytic activity, and antimicrobial potential using the dbAMP framework. Five non-hemolytic peptides (1ANN, 2ANN, 4RF, 5RF, and 6RF) were selected for further analysis based on their predicted physicochemical characteristics and antimicrobial profiles. Molecular docking against the three components of the MexAB-OprM efflux pump (MexA, MexB, and OprM) predicted binding affinities ranging from - 7.1 to - 14.1 kcal mol⁻¹. Among the evaluated peptides, 2ANN was the only peptide predicted to form a hydrogen bond between Lys15 and Asp61 of MexA. Molecular dynamics simulations demonstrated that the 2ANN-MexA complex remained structurally stable over 200 ns, while membrane simulations showed spontaneous adsorption of 2ANN onto a Gram-negative membrane and persistent membrane association throughout the simulation. Collectively, these computational analyses identified 2ANN as a promising AMP candidate for further investigation against MDR P. aeruginosa. However, as the findings are based exclusively on computational prediction and simulation, experimental validation, including peptide synthesis, antimicrobial susceptibility testing, hemolytic assays, and efflux pump inhibition studies, is required to confirm its biological activity and mechanism of action.