Evaluating Force Matching as a Parametrization Strategy for the CHARMM36m Force Field Using Phosphorylation

Phosphorylation is a central regulatory post-translational modification whose accurate representation is essential for molecular simulations of biomolecular systems. Although CHARMM includes parameters for phosphorylated residues, their nonbonded interactions were largely estimated decades ago and have remained insufficiently validated due to the scarcity of experimental data for highly charged phosphate groups and the inherent chemical instability of the phosphoester linkage in model compounds. Here, we reparametrize the amino acid side-chain analog methylphosphate, a model for phosphorylated serine, in multiple charge states using force matching to density functional theory reference data obtained from our own quantum chemical calculations in an aqueous environment. The resulting parameters are validated against new experimental measurements, including osmotic pressure measured as osmotic concentration and nuclear magnetic resonance (NMR) relaxation data for phosphorylated dipeptides. This combined computational-experimental approach enables a systematic refinement of the nonbonded parameters for all relevant phosphate charge states, yielding physically accurate hydration and ion-interaction behavior while maintaining compatibility with the CHARMM36m force field.