pKa values in proteins determined by electrostatics applied to molecular dynamics trajectories

Area of Science:

• Computational Biology
• Biophysics
• Protein Chemistry

Background:

• Accurate prediction of protein pKa values is crucial for understanding protein function and interactions.
• Previous methods like Karlsberg(+) (KB(+)) relied on static crystal structures, limiting accuracy.
• Molecular dynamics (MD) simulations offer a more dynamic representation of protein conformations.

Purpose of the Study:

• To develop and validate an improved computational method for predicting protein pKa values.
• To assess the impact of using molecular dynamics (MD) conformations versus static crystal structures.
• To refine electrostatic energy calculations for pKa prediction.

Main Methods:

• Utilized four 10 ns molecular dynamics (MD) simulations for each protein to generate conformational ensembles.
• Employed a new approach (KB2(+)MD) solving the Poisson-Boltzmann equation with MD-derived conformations.
• Reformulated electrostatic energy expressions to simplify calculations and avoid intrinsic pKa dependencies.

Main Results:

• Reduced the pKa root-mean-square deviation (RMSD) from 1.17 pH units (KB(+)) to 0.96 pH units (KB2(+)MD).
• Further improved pKa RMSD to 0.79 pH units by energy-minimizing conformations with a dielectric constant of 4.
• Developed electrostatic energy terms that decouple protein and solvent contributions.

Conclusions:

• The KB2(+)MD approach significantly enhances the accuracy of computed protein pKa values.
• Incorporating protein dynamics through MD simulations is key to improving pKa prediction.
• The refined electrostatic energy calculations provide a more direct and interpretable method for pKa analysis.