D-MBIS Nonbonded Force Field Parameters Improve Specificity and Selectivity Prediction in Bromodomains.

Area of Science:

• Computational chemistry
• Drug discovery and development
• Molecular modeling

Background:

• Free energy calculations are essential for predicting ligand binding affinities in drug discovery.
• Accurate force fields are critical for reliable prediction of ligand specificity and selectivity.
• Current methods face challenges in modeling conformational changes, binding poses, and molecular interactions.

Purpose of the Study:

• To evaluate ab initio derived nonbonded force field parameters for predicting specificity and selectivity of drug candidates.
• To assess the impact of improved force field parameters on absolute binding free energy calculations.
• To compare computational predictions with experimental data for BRD4 inhibitors and bromodomains.

Main Methods:

• Utilized ab initio derived nonbonded force field parameters from Minimal Basis Iterative Stockholder (D-MBIS) atom partitioning.
• Replaced Open Force Field Sage 2.0.0 parameters with new atomic charges, van der Waals radii, and dispersion coefficients.
• Incorporated ligand polarization energies and focused on experimentally resolved protein structures.

Main Results:

• Achieved a mean unsigned error of 0.48 kcal/mol in predicting absolute binding free energy for nine BRD4 inhibitors.
• Demonstrated a strong correlation between calculated and experimental binding free energies.
• Accurately replicated experimental selectivity rankings for seven out of eight bromodomains when using experimental structures.

Conclusions:

• Ab initio derived force field parameters significantly improve the accuracy of binding free energy predictions.
• Using experimentally resolved structures is crucial for accurate selectivity predictions.
• The developed parameters show promise for enhancing specificity and selectivity assessments in drug discovery.