STUDY OF BINDING AFFINITIES OF FabZ INHIBITORS: NAS-21 AND NAS-91 ANALOGUES BASED ON RECEPTOR - CENTRIC COMPUTATIONAL METHODS

Abstract

NAS21, NAS91 and its derivatives belong to class FabZ inhibitors have been focused to develop better anti-malarial drugs. Library of 17 analogues was designed from NAS21, NAS91 scaffold structure, and NAS75, NAS79 was considered for computational study. Their molecular interactions, binding affinities with FabZ was studied using receptor-centric approaches: glide docking, molecular mechanics using generalized Born/surface area solvation model and multi-ligand bimolecular association with  energetic. Prediction models were developed between FabZ inhibition activity (pIC₅₀) of these compounds and molecular descriptors like glide score, binding energy and calculated free energy binding. The r² value varies from 0.66-0.77 indicating good data fit, and r²_cv was within range of 0.64-0.76, indicating acceptable predictive capabilities of models. A linear correlation was observed between predicted and experimented pIC₅₀ with r² = 0.66-0.78, suggesting the robustness of models. The ensemble-average free energy estimation including implicit solvation energy terms significantly improves the hit enrichment of virtual screening.