MDO: A Computational Protocol for Prediction of Flexible Enzyme-Ligand Binding Mode

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Predicting enzyme-ligand binding modes is crucial for understanding biological complex functions.
• The vast sampling space presents a significant challenge in accurate binding mode prediction.

Purpose of the Study:

• To develop a novel computational protocol for accurate enzyme-ligand binding mode prediction.
• To enhance computer-aided drug design methodologies.

Main Methods:

• The MDO protocol combines molecular dynamics simulations for enzyme sidechain sampling and molecular docking for ligand pose sampling.
• Clustering techniques are used for both enzyme and ligand configurations.
• Geometry optimization and the ONIOM method are employed for final pose selection and ranking.

Main Results:

• The MDO protocol achieved a 67% success rate (RMSD < 2 Å) on 15 known enzyme-ligand complexes.
• This success rate is substantially higher than the 40% achieved by conventional methods.
• Applying ONIOM calculations to initial poses within binding cavities further increased the success rate to 83%.

Conclusions:

• The MDO protocol offers high-quality enzyme-ligand binding mode prediction at a reasonable computational cost.
• MDO is a valuable tool recommended for structure-based drug design.