Higher Accuracy Achieved for Protein-Ligand Binding Pose Prediction by Elastic Network Model-Based Ensemble Docking.

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Molecular docking typically treats protein receptors as rigid, limiting accuracy due to inherent protein flexibility.
• Protein conformational changes upon ligand binding can significantly alter binding pockets, affecting docking predictions.

Purpose of the Study:

• To develop and validate an iterative anisotropic network model (iterANM)-based ensemble docking approach.
• To incorporate protein flexibility into molecular docking for improved prediction of receptor-ligand interactions.

Main Methods:

• Generated holo-like receptor structures using iterANM from apo receptor conformations.
• Employed ensemble docking with multiple flexible receptor structures.
• Validated the approach using datasets of cyclin-dependent kinase 2 (CDK2) inhibitors and diverse receptor-ligand complexes.

Main Results:

• iterANM-based ensemble docking demonstrated superior performance in reproducing native-like binding poses compared to rigid docking and molecular dynamics ensembles.
• Prediction accuracy was further enhanced by reranking with the molecular mechanics-Poisson-Boltzmann surface area (MM/PBSA) method.
• The approach showed higher success rates than AutoDockFR and other flexible receptor methods on CDK2 inhibitor datasets.

Conclusions:

• The iterANM-based ensemble docking provides an accurate, efficient, and practical framework for predicting ligand binding modes in flexible receptors.
• This method addresses limitations of rigid receptor docking and offers a robust alternative for drug discovery and structural biology research.