Ligand-Binding-Site Structure Refinement Using Molecular Dynamics with Restraints Derived from Predicted Binding Site...

Area of Science:

• Computational Biology
• Structural Biology
• Drug Discovery

Background:

• Accurate modeling of ligand-binding sites is crucial for structure-based virtual screening.
• Predicted protein models often have low-quality ligand-binding site structures requiring refinement.

Purpose of the Study:

• To present a novel protocol for refining ligand-binding site structures.
• To enhance the accuracy of protein models for virtual screening applications.

Main Methods:

• Utilized molecular dynamics (MD) simulations.
• Employed restraints derived from predicted binding site templates for refinement.
• Validated the protocol on 40 diverse protein sets from the Astex list.

Main Results:

• Achieved consistent refinement of ligand-binding sites on modeled protein structures.
• Demonstrated an average Cα RMSD improvement of 0.90 Å for protein structures.
• Showed an average RMSD improvement of 1.97 Å in ligand binding modes after refinement.

Conclusions:

• The developed protocol offers a promising method for refining protein ligand-binding site structures.
• Improved structure accuracy facilitates more reliable virtual screening and drug design.