Fully Flexible Docking via Reaction-Coordinate-Independent Molecular Dynamics Simulations

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Predicting protein-ligand binding geometry is crucial for structure-based drug discovery.
• Molecular dynamics (MD) offers a reliable computational approach, but simulations are often computationally expensive.
• Current MD methods for binding pose prediction are limited by simulation costs.

Purpose of the Study:

• To develop a general framework for accelerating the generation of protein-ligand binding modes.
• To reduce the computational expense of molecular dynamics simulations for drug discovery.
• To provide a method independent of predefined reaction coordinates.

Main Methods:

• Implementation of a potential-scaled molecular dynamics (MD) framework.
• Application of the dynamical protocol to GSK-3β and HSP90α systems.
• Validation of the approach for identifying correct ligand binding modes.

Main Results:

• The potential-scaled MD approach successfully generated putative protein-ligand binding modes.
• The method was applied to pharmaceutically relevant targets like GSK-3β and HSP90α.
• The protocol identified correct binding modes for multiple ligands without requiring predefined reaction coordinates.

Conclusions:

• Potential-scaled MD simulations offer a general and efficient framework for predicting protein-ligand binding modes.
• This approach significantly reduces computational time for dynamic docking simulations.
• The method enhances the practical applicability of MD in structure-based drug discovery.