Machine learning-based virtual screening and density functional theory characterisation of natural inhibitors targeting mutant PBP2x in Streptococcus pneumoniae

Abstract

Abstract Streptococcus pneumoniae ( S. pneumoniae ) has developed resistance to β-lactam antibiotics, largely due to mutations in penicillin-binding protein 2x (PBP2x), particularly within conserved motifs such as STMK and KSG. PBP2x mutations are frequently reported in multidrug-resistant pneumococcal strains associated with pneumonia, meningitis, and septicaemia. especially in serotypes 19A, 19F, and 23F, showing reduced susceptibility to β-lactam antibiotics. These mutations in the PBP2x disrupt antibiotic binding and enzymatic functions, highlighting the need for alternative therapeutic strategies. This study focused on five clinically relevant PBP2x mutations (T338A/G/P and K547G/T) within its active site. A library of phytocompounds was screened using a machine learning model trained to identify antibacterial compounds. Top candidates were filtered based on ADMET properties, and their electronic characteristics were assessed using HOMO–LUMO analysis and electrostatic potential mapping, through density functional theory (DFT). Glucozaluzanin C, a phytochemical derived from Elephantopus scaber , emerged as a potential candidate. Molecular docking and dynamics simulations revealed strong binding affinity and structural integrity with all PBP2x mutants, over a 100-ns timescale. RMSD, RMSF, and hydrogen bonding analysis confirmed stable interactions, suggesting Glucozaluzanin C may effectively interact with PBP2x mutants. Overall, the study highlights an effective strategy for identifying plant-derived inhibitors against β-lactam-resistant S. pneumoniae.