Accelerated flexible protein-ligand docking using Hamiltonian replica exchange with a repulsive biasing potential

Area of Science:

• Computational Biology
• Biophysics
• Molecular Modeling

Background:

• Identifying ligand binding sites on biomolecules is crucial for drug discovery and understanding biological processes.
• Traditional docking methods often require prior knowledge of binding sites and may oversimplify molecular flexibility.
• Limitations in current methods necessitate the development of more efficient and flexible approaches for binding site identification.

Purpose of the Study:

• To develop a novel molecular dynamics (MD) based method for the rapid identification of putative ligand binding sites on biomolecular surfaces.
• To enhance the exploration of potential binding regions by preventing ligand or protein trapping at non-specific sites.
• To offer a flexible approach that accommodates full flexibility of binding partners, unlike many rigid-body docking methods.

Main Methods:

• Utilized a molecular dynamics replica exchange (REMD) based approach with ambiguity restraints between receptor and ligand.
• Simulations employed a set of restraints that allowed close contacts in a reference replica while promoting transient dissociation in higher-energy replicas.
• The method was applied to peptide-protein, protein-protein, and drug-receptor systems, including explicit solvent simulations for a DNA-ligand complex.

Main Results:

• The developed REMD method rapidly identified putative ligand binding sites across various molecular systems.
• The approach demonstrated accelerated exploration of binding regions by avoiding non-specific trapping, outperforming continuous MD simulations.
• Near-native binding regions were sampled efficiently, even when starting from distant configurations.
• The method proved effective in explicit solvent simulations, showcasing its versatility.

Conclusions:

• The novel REMD method offers a powerful and efficient tool for rapid identification of ligand binding sites.
• This approach overcomes limitations of traditional docking by allowing full flexibility and avoiding non-specific interactions.
• The method shows broad applicability across different molecular systems and simulation conditions, advancing molecular modeling and drug discovery efforts.