MM/PB(GB)SA benchmarks on soluble proteins and membrane proteins

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurate prediction of protein-ligand binding free energy is crucial for drug discovery.
• Molecular mechanics/Poisson-Boltzmann (Generalized Born) surface area (MM/PB(GB)SA) is a valuable computational tool for binding free energy calculations.
• Existing methods require optimization for diverse biological systems.

Purpose of the Study:

• To benchmark the accuracy of MM/PB(GB)SA for predicting binding free energy across various protein systems.
• To investigate the impact of different computational parameters on MM/PB(GB)SA performance.
• To compare MM/PB(GB)SA with other established methods like docking and free energy perturbation (FEP).

Main Methods:

• Systematic benchmarking of MM/PB(GB)SA across three membrane-bound and six soluble protein systems.
• Exploration of various parameters including ligand charges, force fields, GB models, and dielectric constants.
• Comparative analysis of MM/PB(GB)SA against docking and FEP methods.

Main Results:

• MM/PB(GB)SA demonstrated competitive accuracy compared to FEP for binding free energy prediction.
• Parameter optimization, particularly for GB models and membrane dielectric constants, is essential for system-specific accuracy.
• The study identified key parameters influencing MM/PB(GB)SA performance.

Conclusions:

• MM/PB(GB)SA is a powerful and efficient method for predicting protein-ligand binding free energy in drug discovery.
• System-specific parameterization is critical for maximizing the accuracy of MM/PB(GB)SA.
• This approach offers a valuable tool for accelerating the drug design process.