WIDOCK: a reactive docking protocol for virtual screening of covalent inhibitors

Area of Science:

• Computational Chemistry
• Drug Discovery
• Biochemistry

Background:

• Covalent inhibitors offer unique therapeutic potential but require specialized screening methods.
• Accurately predicting ligand reactivity is crucial for identifying effective covalent drugs.

Purpose of the Study:

• To introduce WIDOCK, a virtual screening protocol for selecting diverse electrophiles as covalent inhibitors.
• To integrate ligand reactivity towards cysteine residues into the AutoDock4 docking protocol.
• To validate WIDOCK's performance in retrospective and prospective drug discovery efforts.

Main Methods:

• WIDOCK utilizes a reactive docking approach, extending existing methods with facile parametrization (experimental or computational).
• It incorporates a ligand-focused evaluation scheme and validates performance against therapeutically relevant targets.
• Reactivity parameters are derived from experimental kinetics or computed reaction barriers.

Main Results:

• WIDOCK demonstrated high sensitivity in identifying experimental actives against targets like KRAS^G12C, MurA, cathepsin B, OTUB2, and NUDT7.
• The protocol outperformed standard covalent docking tools in ROC curves, AUC values, and enrichment factors for diverse warheads.
• Prospective screening identified novel covalent human MAO-A inhibitors, with experimental confirmation and Cys323 labeling.

Conclusions:

• WIDOCK is a validated, warhead-sensitive virtual screening protocol for covalent inhibitor discovery.
• The protocol effectively integrates ligand reactivity into docking, improving the identification of cysteine-reactive drug candidates.
• WIDOCK shows significant utility in both retrospective analysis and prospective identification of novel covalent inhibitors.