From Water Networks to Binding Affinities: Resolving Solvation Dynamics for Accurate Protein-Ligand Predictions

Water molecules play a critical role in mediating protein–ligand interactions by forming hydrogen-bonding networks and contributing to ligand solvation. However, their behavior—ranging from rapid solvent exchange to persistent occupancy of buried sites—makes accurate binding free-energy estimation with molecular dynamics challenging. Inadequate sampling of water reorganization can bias computed affinities and obscure key interactions by preventing representative hydration states during calculations. To address this, we employ the polarizable AMOEBA force field together with Lambda-ABF-OPES, an integrated enhanced-sampling framework combining lambda-dynamics, multiple-walker adaptive biasing force, and on-the-fly probability enhanced sampling. AMOEBA provides a polarizable description of protein–ligand, ligand–water, and protein–water interactions, while the dynamic alchemical coordinate and adaptive biasing facilitate exploration of coupled ligand–pocket–solvent configurations during ligand decoupling. This construction allows water exchange, binding-site rehydration and solvent reorganization to emerge along the alchemical pathway without explicitly biasing water-based collective variables. Applied to five protein–ligand complexes spanning buried and semi-buried environments, the approach yields binding affinities in good agreement with experiment and captures relevant water-mediated reorganization to provide reproducible absolute binding free energies.