Grid-Based Surface Generalized Born Model for Calculation of Electrostatic Binding Free Energies

Area of Science:

• Computational Chemistry
• Molecular Modeling
• Biophysics

Background:

• Accurate calculation of solvation free energies is essential for rational drug design.
• Existing methods may face challenges in balancing computational speed and accuracy.

Purpose of the Study:

• To introduce and evaluate a grid-based molecular surface implementation of the "R6" generalized Born (GB) implicit solvent model, named GBNSR6.
• To assess the speed, accuracy, and parameter sensitivity of GBNSR6 compared to numerical Poisson-Boltzmann (PB) methods.

Main Methods:

• Implementation of a grid-based molecular surface for the GBNSR6 model.
• Testing on 15 protein-ligand complexes and biomolecules (268-25099 atoms).
• Comparison of GBNSR6 results with numerical PB calculations.

Main Results:

• GBNSR6 achieves a good balance between speed and accuracy for polar solvation free energies (ΔGpol) and binding free energies (ΔΔGpol).
• The model is robust to coarse grid sizes (h = 0.5 Å) with minimal grid artifact errors (∼0.6 kcal/mol).
• Estimated ΔΔGpol values show high correlation (r² = 0.97) with numerical PB, with low RMSE (1.43 kcal/mol) and no systematic bias.

Conclusions:

• The grid-based GBNSR6 model provides a computationally efficient and accurate method for calculating solvation free energies.
• This model is suitable for applications like drug design, offering reliable results comparable to more computationally intensive methods.
• GBNSR6 is integrated into the Amber (AmberTools) package for broader accessibility in molecular simulations.