BPS2026 – Polarizable constant pH molecular dynamics for protein-cofactor and protein-DNA complexes

Constant pH molecular dynamics (CpHMD) algorithms are valuable for investigating biological mechanisms in which molecular conformation and protonation are tightly coupled. CpHMD methods have been limited to fixed-charge representations of electrostatics, impeding their capability to investigate highly polar environments such as those surrounding DNA-binding domains. To overcome this limitation, the first CpHMD algorithm for the atomic multipole optimized energetics for biomolecular applications (AMOEBA) force field was developed and validated on a set of crystal peptides under particle-mesh Ewald (PME). Here, we apply AMOEBA CpHMD to attain the most accurate CpHMD pKa predictions to date for three well-studied proteins—hen egg white lysozyme (HEWL), the chicken villin headpiece subdomain (HP36), and BBL. The root-mean-square-deviation from experiment was 0.59 with a Pearson’s correlation coefficient of 0.90. This marks an improvement over the fixed-charge methods, which is highlighted by significant improvement for the prediction of the catalytic dyad in HEWL (GLU-35/ASP-52) with near agreement to experiment. The fixed-charge models were shown to overestimate the ASP-52 pKa by as much as 2 pH units. We also apply AMOEBA CpHMD simulations to the highly polarizable cysteine-bound Zn 2+ ion and protein-DNA complexes. Fixed-charge models struggle to accurately predict cysteine pKa shifts, and we show that inclusion of polarizability improves upon those predictions. We perform CpHMD simulations of zinc fingers in which four cysteines are involved in binding the Zn 2+ ion, which is a highly conserved motif in transcription factors. We investigate titration state changes in the absence and presence of DNA. Lastly, we assess the impact of cysteine protonation states changes on DNA binding with free-energy calculations. Application of polarizable CpHMD will improve our understanding of transcription factor binding affinity and specificity.