Computational Design and Virtual Screening of Enumeration-Generated Benzothiazinone Derivatives as Cholinesterase Inhibitors Using Pharmacophore Modelling, Molecular Dynamics, DFT and ML-Based Predictive Modelling Analysis

Acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) are key cholinergic targets implicated in Alzheimer’s disease (AD), and simultaneous modulation of these enzymes represents a promising strategy for addressing cholinergic dysfunction. In the present study, a comprehensive computational approach was employed to investigate benzothiazinone derivatives as potential cholinesterase inhibitors. Molecular docking of a virtual library against AChE (PDB ID: 6O4W) identified several compounds with favorable binding affinities, among which compounds 1d, 1k, and 1a were prioritized based on their overall docking performance and binding interactions. Compound 1d exhibited a favorable AChE docking score of -9.441 kcal/mol, comparable to the reference drug donepezil. To assess their binding affinity to BChE, the selected compounds were docked with BChE (PDB ID: 5DYW). This resulted in docking scores of -8.796 kcal/mol for 1a, -8.859 kcal/mol for 1k, and -8.748 kcal/mol for 1d. Further MMGBSA binding free-energy calculations confirmed these compounds' strong binding, emphasizing the role of hydrophobic and van der Waals interactions in stabilizing the complexes. The stability of the selected ligand–protein complexes were subsequently evaluated through 100 ns molecular dynamics simulations, with compounds 1k and 1d exhibiting stable interactions within the respective binding sites. Pharmacophore modelling identified the AHRRR_1 hypothesis, comprising one hydrogen-bond acceptor, one hydrogen-bond donor, and three aromatic ring features, with a selectivity score of 1.985 and a survival score of 5.975. Density functional theory (DFT) calculations revealed favorable electronic characteristics, including suitable HOMO–LUMO distributions and charge-transfer properties. ADMET analysis indicated favourable pharmacokinetic profiles for several derivatives, while CatPredbased machine-learning analysis predicted inhibition constants (Ki) in the micromolar range, suggesting moderate to strong binding potential. Overall, the integrated AChE/BChE docking, MMGBSA, molecular dynamics, pharmacophore modelling, DFT, ADMET, and machinelearning analyses highlight benzothiazinone derivatives, particularly 1d, 1k, and 1a, as promising cholinesterase-targeting candidates and provide valuable structural insights for the rational development of potential therapeutic agents against Alzheimer’s disease.