Discovery of Mangifera indica-based natural inhibitors against TEM-1 β-lactamase from Escherichia coli using machine learning approaches.

Journal: Computational biology and chemistry
Published Date:

Abstract

Antimicrobial resistance driven by β-lactamase enzymes, particularly TEM-1 in Escherichia coli, compromises the efficacy of β-lactam antibiotics. This study applied a deep learning-guided pipeline to identify Mangifera indica phytochemicals as potential TEM-1 inhibitors. A neural network trained on 220 compounds achieved a ROC-AUC of 0.81, PR-AUC of 0.85, and Matthews correlation coefficient (MCC) of 0.64, with enrichment factors 1.55 at both 1 % and 5 % thresholds, demonstrating robust predictive ability. Screening of M. indica phytochemicals yielded 25 top-ranked compounds, of which seven satisfied drug-likeness and pharmacokinetic criteria. Docking analysis showed free binding energies ranging from -9.10 to -7.80 kcal·mol⁻¹ , outperforming Tazobactam (-8.07 kcal·mol⁻¹) and clavulanic acid (-6.20 kcal·mol⁻¹). The best-scoring ligands (LTS0215935, LTS0057337, LTS0207894, and LTS0086727) formed extensive hydrogen-bond networks with catalytic residues Ser70, Ser130, Asn132, Ser235, and Arg244. Molecular dynamics (500 ns) confirmed complex stability, with RMSD values between 2.0 and 3.5 Å, reduced residue fluctuations in the SDN and Ω loops, and compact global folds reflected by Rg (26-28 Å). Solvent-accessible surface area analysis revealed tighter packing for LTS0215935 (24.5-26.0 × 10 ³ Ų) compared to Tazobactam (26.5-27.5 × 10 ³ Ų). MM-GBSA calculations ranked LTS0215935 (-47.5 kcal·mol⁻¹) as the strongest binder, followed by LTS0057337 (-43.5 kcal·mol⁻¹), LTS0207894 (-42.6 kcal·mol⁻¹), and LTS0086727 (-43.1 kcal·mol⁻¹), all surpassing Tazobactam (-32.4 kcal·mol⁻¹). Density functional theory (DFT) analysis corroborated these findings, with HOMO-LUMO gaps indicating suitable electronic reactivity and molecular electrostatic potential maps highlighting complementarity with TEM-1 active-site residues. These results nominate M. indica scaffolds, particularly LTS0215935, as promising natural templates for next-generation β-lactamase inhibitors.

Authors

Keywords

No keywords available for this article.