PREDICTING BINDING AFFINITIES OF FOSMIDOMYCIN ANALOGUES AS DOXP-REDUCTOISOMERASE INHIBITORS BASED ON STRUCTURE-CENTRIC APPROACHES

Abstract

Fosmidomycin and its derivatives belong to class DOX-reductoisomerase (DXR) inhibitors. A fosmidomycin analogues library was designed with 43 analogues, their molecular interactions and binding affinities with DXR (PDB ID: 1ONP) have been studied using Glide docking, QM-polarized ligand docking (QPLD), molecular mechanics based on generalized Born/surface area (MM-GB/SA) and multi-ligand bimolecular association with energetics (eMBrAcE). Prediction models were developed between DXR inhibition activity (pIC₅₀) of these compounds and molecular descriptors like Glide score, QPLD_score, binding energy and calculated free energy of binding. The r² value ranges from 0.608-0.807 indicating good data fit, and r²_cv ranges from 0.592-0.799 indicating acceptable predictive capabilities of models. Linear correlation between predicted and experimented pIC₅₀ was observed (r² = 0.603-0.781). Low root mean square error (0.39-0.84) of inhibitors was established in all structure-centric approaches, an efficient tool for generating potential and specific inhibitors of DXR by testing rationally designed lead compounds based on fosmidomycin derivatization.