Accurately describing protonation equilibria is critical for understanding biomolecular recognition and drug binding. Roughly 70% of small-molecule drugs contain ionizable groups, and changes in their protonation state strongly influence binding affinity, specificity, and pharmacokinetics. Yet, most molecular dynamics simulations still assume fixed protonation states, missing this essential piece of molecular behavior. Constant pH molecular dynamics (CpHMD) overcomes this limitation, but technical challenges have limited its widespread adoption. A particular gap exists in protein-ligand complexes, where experimental data on ligand pKa shifts remain scarce and scattered, leaving fundamental questions about how proteins reshape ligand ionization unanswered. We address this challenge with CpHMD enhanced by adaptive landscape flattening (ALF), a framework designed to lower sampling barriers and make ligands readily titratable within simulations. This methodology enables rigorous and efficient exploration of protonation states across protein-ligand systems, providing a path forward where experimental information is limited. As demonstrations, we studied riboflavin binding protein, and the folate receptor, capturing ligand-centered pKa shifts and coupled protonation events that underlie pH-dependent binding and release. We further extended the approach to drug interactions with human serum albumin, highlighting how ionization contributes to strong affinity. Together, these results show that CpHMD with ALF can illuminate new aspects of protein-ligand binding, expanding molecular dynamics toward more realistic, pH-sensitive biophysics.
Cherepanov et al. (Sun,) studied this question.