CIFDock: A novel CHARMM-based flexible receptor-flexible ligand docking protocol.

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Traditional docking methods face limitations in ligand sampling, protein flexibility, scoring accuracy, and explicit water retention.
• These limitations can hinder the accurate prediction of molecular interactions crucial for drug discovery.

Purpose of the Study:

• To introduce a novel CHARMM-based induced fit docking (CIFDock) workflow designed to address existing limitations in molecular docking.
• To enhance the sampling of protein, ligand, and explicit water molecule dynamics for improved accuracy in drug discovery.

Main Methods:

• Employed a CHARMM-based induced fit docking (CIFDock) workflow integrating all-atom force fields with enhanced sampling molecular dynamics.
• Utilized Self-Guided Langevin Dynamics (SGLD) for comprehensive sampling of ligand conformations, side-chain orientations, water positions, and protein residue motion.
• Implemented a modular, automated procedure using CHARMM scripting and SLURM array processing for efficient parallelization and induced fit effect capture.

Main Results:

• The CIFDock workflow successfully integrated protein, ligand, and water flexibility through simultaneous, unrestrained SGLD sampling.
• Cross-docking studies on 21 pharmaceutically relevant proteins demonstrated that CIFDock's results were comparable or superior to commercial docking programs.
• Developed and utilized five variants of CHARMM-based SWISSDOCK scoring functions to evaluate the generated binding poses.

Conclusions:

• The novel CIFDock workflow effectively overcomes key limitations in traditional docking, offering a more robust approach for drug discovery.
• This method provides enhanced sampling of molecular dynamics, leading to more accurate prediction of binding poses and induced fit effects.
• CIFDock presents a validated, automated, and parallelizable tool for improving the efficiency and accuracy of structure-based drug design.