GPU-Accelerated All-Atom Particle-Mesh Ewald Continuous Constant pH Molecular Dynamics in Amber

Area of Science:

• Computational Chemistry
• Biophysics
• Molecular Dynamics

Background:

• Constant pH molecular dynamics (CpHMD) simulations are vital for studying biological processes influenced by pH.
• Current CpHMD methods have limitations in accuracy due to implicit-solvent models or hybrid schemes.

Purpose of the Study:

• To develop and validate a GPU-accelerated all-atom CpHMD method with Particle Mesh Ewald (PME) electrostatics.
• To improve the accuracy of molecular dynamics simulations for pH-dependent biological phenomena.

Main Methods:

• Implemented and parameterized an all-atom CpHMD method in the Amber22 pmemd.cuda engine.
• Utilized asynchronous pH replica-exchange titration simulations for validation across six benchmark proteins.
• Derived titration parameters for key amino acids (Asp, Glu, His, Cys, Lys) for CHARMM c22 and Amber ff14sb/ff19sb force fields.

Main Results:

• Achieved a root-mean-square deviation of 0.76 pH units from experimental pKa values for Asp, Glu, His, and Cys.
• Demonstrated strong correlation (Pearson's r=0.80) between calculated and experimental pKa shifts.
• Showcased the method's ability to capture biologically relevant protonation events in challenging protein systems.

Conclusions:

• The PME-CpHMD method significantly enhances the accuracy of molecular dynamics simulations under varying pH conditions.
• This advancement enables more reliable mechanistic studies of ubiquitous proton-coupled biological processes.
• Finite-size corrections or larger simulation boxes are recommended to further refine pKa calculations.