A parameterized, continuum electrostatic model for predicting protein pKa values.

Area of Science:

• Biophysics
• Computational Chemistry
• Protein Biochemistry

Background:

• Accurate prediction of pK(a) values is crucial for understanding protein function and behavior.
• Purely electrostatic models have limitations in achieving chemical accuracy for pK(a) calculations.
• The Poisson-Boltzmann solver provides a framework for macroscopic electrostatics with atomic detail.

Purpose of the Study:

• To improve the reliability and accuracy of pK(a) calculations by introducing empirical corrections to the Poisson-Boltzmann solver.
• To develop a robust model with a minimal number of parameters for predicting pK(a) values of compounds outside the fitting dataset.
• To investigate the contributions of local functional group interactions, desolvation work, and site-to-site interactions to intrinsic pK(a).

Main Methods:

• Empirical corrections were added to the Poisson-Boltzmann solver.
• Parameters were derived from 286 residues across 30 proteins, based on electrostatic interactions and desolvation work.
• The linearized Poisson-Boltzmann equation was used to calculate electrostatic fields and their relation to intrinsic pK(a).

Main Results:

• The parameterized model achieved a root mean square error (RMSE) of 0.70 on the fitting set.
• An independent test set of 8 proteins yielded an RMSE of 1.08.
• The model demonstrated improved performance compared to other existing models.
• Residue burial depth correlated strongly with prediction error, a common limitation across models.

Conclusions:

• The developed empirical model significantly enhances the accuracy of pK(a) predictions in proteins.
• Local electrostatic effects, rather than long-range site-to-site interactions, are the primary drivers of pK(a) values.
• The model offers a robust and accurate approach for pK(a) calculations in biochemical and biophysical studies.