Abstract
The escalating global crisis of antimicrobial resistance (AMR) necessitates the discovery of novel therapeutic targets that bypass conventional orthosteric resistance mechanisms. Allosteric sites—distinct from the primary active site—offer a promising avenue for drug design, providing high selectivity and the potential to modulate protein function through non-competitive mechanisms. This study presents a comprehensive integrated framework for the identification and validation of allosteric sites in critical drug-resistant bacterial enzymes, specifically focusing on multidrug resistance (MDR) efflux pumps and transcriptional machinery. Using a combination of dynamic profile analysis, machine learning-based ranking via PASSerRank, and molecular dynamics (MD) simulations, we identified three novel putative allosteric pockets in the Porphyromonas gingivalis efflux pump and Mycobacterium tuberculosis RNA polymerase. Computational results were validated through site-directed mutagenesis and kinetic assays using light-inducible allosteric control and amperometric biosensing. Our findings demonstrate that these allosteric sites are highly conserved across resistant strains and provide a structural basis for overcoming active-site mutations. The identified sites showed high druggability scores (D-score > 0.75) and were successfully modulated by small-molecule fragments. This research underscores the utility of combining advanced in silico predictions with robust experimental validation to expand the druggable landscape of the bacterial proteome, offering a strategic blueprint for the next generation of allosteric antibiotics to combat MDR pathogens.