Pharmacophore guided deep learning approach to identify novel inhibitors targeting mycobacterial polyketide synthase Pks13-TE domain

Abstract

Tuberculosis (TB), an enduring global health challenge, persists due to the rise of drug-resistant Mycobacterium tuberculosis (MTB) strains. Among potential therapeutic targets, Pks13, a protein crucial for mycolic acid biosynthesis, is key for Mtb's virulence and survival. This study employed a pharmacophore-based drug design approach, employing the Pharmacophore-Guided Molecular Generation (PGMG) tool to target Pks13 inhibitors. Aromatic, hydrophobic, positive ion and hydrogen bond acceptors pharmacophoric features were identified from co-crystal ligands. Candidate compounds underwent evaluation of pharmacokinetic properties with the ADMET_AI tool. Further refinement involved molecular docking with PLANTS software, absolute binding free energy calculations via KDeep, and toxicity assessments using eToxPred. MM-GBSA, PCA, DCCM, and FEL were incorporated to validate and refine inhibitors accurately. From this analysis, we discovered five novel hit molecules. These includes, ethyl (S)-4-(4-(2-(2-(cyclohexylamino)-2-oxoethoxy)-2-oxoethoxy)phenyl)-6-methyl-2-oxo-1,2,3,4-tetrahydropyrimidine-5-carboxylate (Pk1), 2-(4-((2‑butyl‑6-pivalamido-1H-benzo [d]imidazol-1-yl)methyl)phenoxy)acetic acid (Pk2), (S)-2-(4-(4-cyclopentyl-5-(1-((6-methyl-2-morpholinopyrimidin-4-yl)amino)ethyl)-4H-1,2,4-triazol-3-yl)phenoxy)acetic acid (Pk3), (E)-2-(4-oxo-5-(4-((6-phenylpyridin-2-yl)methoxy)benzylidene)-2-thioxothiazolidin-3-yl)acetic acid (Pk4), and N-((1R,4r)-4-(((1R,2r,3S,5S,7S)-5‑hydroxy adamantan-2-yl)oxy)cyclohexyl)-4-morpholinobenzamide (Pk5). We conclude that the screened hit compounds may act as potential inhibitors targeting Pks13 and further preclinical and clinical studies may pave the way for developing them as effective therapeutic agents for the treatment of MTB.