<i>In silico</i> discovery of novel quinazolin-4-one derivatives as potent 3CLpro inhibitors using 2D-QSAR, molecular docking, and molecular dynamics

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) main protease (3CLpro) is a pivotal enzyme in viral replication. 3CLpro serves as a highly conserved and mutation-resistant target, rendering it a promising focus for the development of effective antiviral agents. To develop and assess possible 3CLpro inhibitors, we used a multilayered in silico approach. Quinazolin-4-one derivatives were designed using a fragment-based drug design (FBDD) technique. The promising candidates were chosen based on their predicted activities using a genetic algorithm-multiple linear regression quantitative structure-activity relationship (2D-QSAR) model. The 2D-QSAR model used satisfied both internal and external validation criteria (R 2 = 0.8659, Q 2 loo = 0.8054, and R 2 ext = 0.7385). Following molecular docking to the 3CLpro active site, ADMET (absorption, distribution, metabolism, excretion, toxicity) profiling was performed on the top-ranked compounds to evaluate their pharmacokinetic and toxicity characteristics. The most favorable candidates, A-100, A-117, and D-119, exhibited improved predicted activities (7.459, 7.703, and 7.137, respectively), strong binding affinities of (-8.5, -8.3, and -7.8 kcal mol -1 , respectively), and drug-like profiles. Molecular dynamics (MD) simulations over 150 ns further confirmed the stability and favorable interaction patterns of these ligands within the binding site. MD analysis showed that 3CLpro-D-119 complex exhibited the lowest RMSD (0.192 ± 0.032 nm), indicating the highest structural stability compared with the apo, Baicalein, and the Lead compound. Overall, this integrative computational workflow identified three promising 3CLpro inhibitors with high therapeutic potential, offering an improved foundation for future in vitro and in vivo validation and drug development against SARS-CoV-2.