Inclusion of High-Field Target Data in AMOEBA's Calibration Improves Predictions of Protein-Ion Interactions

Area of Science:

• Computational chemistry and biophysics
• Molecular modeling and simulation
• Protein-ion interactions

Background:

• Molecular mechanics simulations require accurate descriptions of ion-protein, ion-water, and protein-water interactions for reliable predictions.
• Existing force fields (FFs) and nonbonded-fix (NB-fix) methods show limitations in accurately modeling ion-protein interactions, even with polarizable FFs.
• Previous work suggested recalibrating polarization models in FFs could improve ion-protein interaction predictions.

Purpose of the Study:

• To recalibrate polarization models for carbonyls, carboxylates, and hydroxyls within the AMOEBA protein force field.
• To assess the impact of these recalibrations on the accuracy of predicting ion-protein interaction energies.
• To evaluate the combined effect of recalibration and many-body NB-fix corrections on simulation accuracy.

Main Methods:

• Recalibration of polarization models for specific functional groups in the AMOEBA protein FF.
• Calculation of Na+ and K+ -protein interaction energies using the modified FF.
• Benchmarking results against van der Waals-inclusive density functional theory and experimental data.
• Application of many-body NB-fix corrections to Lennard-Jones components.

Main Results:

• Recalibration substantially improved predictions, decreasing mean absolute errors for Na+-protein and K+-protein interaction energies from 8.7 to 5.3 and 9.6 to 6.3 kcal/mol, respectively.
• Further reduction in errors to 2.7 kcal/mol for Na+ and 2.6 kcal/mol for K+ was achieved by applying many-body NB-fix corrections.
• The recalibrated AMOEBA FF maintained its accuracy in predicting protein structure and dynamics.

Conclusions:

• Recalibrating polarization models in the AMOEBA protein FF significantly enhances the accuracy of ion-protein interaction energy predictions.
• Combined with NB-fix corrections, these improvements lead to highly reliable molecular mechanics simulations of ion binding.
• The modified FF offers a more robust tool for studying protein-ion interactions in biological systems.