Assessing protein-ligand binding modes with computational tools: the case of PDE4B

Area of Science:

• Computational Chemistry
• Drug Discovery
• Molecular Modeling

Background:

• Phosphodiesterase 4B (PDE4B) inhibitors are crucial for treating various diseases.
• Developing potent PDE4B inhibitors with minimal side effects remains a challenge.
• Accurate prediction of inhibitor potency is essential for efficient drug discovery.

Purpose of the Study:

• To develop and validate a computational protocol for ranking novel phosphodiesterase 4B (PDE4B) inhibitor candidates.
• To establish a reliable method for estimating IC50 values of newly designed molecules.
• To improve the accuracy of predicting binding affinities for PDE4B inhibitors.

Main Methods:

• Utilized molecular docking (AutoDock) to generate initial poses of known PDE4B inhibitors.
• Employed Molecular Mechanics/Generalized Born Surface Area (MM-GB/SA) calculations on molecular dynamics (MD) trajectories.
• Integrated AutoDock results with averaged MM-GB/SA computations from multiple independent MD trajectories.

Main Results:

• AutoDock binding free energies alone were insufficient for accurately ranking inhibitors.
• MM-GB/SA post-processing of top AutoDock poses improved the correlation between calculated and experimental IC50 values.
• A robust linear relationship between experimental IC50 and calculated binding free energies was achieved by averaging MM-GB/SA results from multiple conformations and trajectories.

Conclusions:

• The developed protocol enhances the accuracy of predicting PDE4B inhibitor potency.
• This method provides a reliable computational tool for screening and ranking novel drug candidates.
• The findings facilitate the discovery of potent PDE4B inhibitors with improved therapeutic profiles.