Quantum mechanics-guided parameterization enhances molecular dynamics accuracy for flavonoid-membrane biophysical properties.

Understanding how small molecule parameterization influences the outcomes of molecular dynamics (MD) simulations is crucial for accurately modeling drug-membrane interactions. This study examines the impact of refined force-field parameterization on the ability of MD simulations to reproduce in vitro experimental biophysical outcomes for three flavonoids (baicalein, chrysin, and luteolin) interacting with model lipid membranes composed of dioleoylphosphocholine and dipalmitoylphosphocholine. We compared an analogy-based approach using CHARMM General Force Field (CGenFF ver. 4.0) with a quantum mechanics (QM)-based optimization protocol using Force Field Toolkit (ffTK). The results of in vitro experiments to assess changes in membrane dipole potential were compared with corresponding changes observed in MD simulations. Elastic properties of the membrane were assessed using differential scanning microcalorimetry and compared to MD-derived order parameters of lipid bilayers (SCD). Compared to the initial CGenFF parameters, the ffTK-optimized model demonstrated enhanced consistency with the experimental changes in membrane dipole potential and elastic properties for all three flavonoids. While the analogy-based CGenFF approach showed reasonable agreement for luteolin, the QM-guided ffTK refinement provided a more accurate description of the behavior of baicalein and chrysin, better aligning the computational results with in vitro observations. These findings underscore the necessity of careful QM-guided parameterization for small molecules in MD simulations aimed at elucidating membrane electrostatics and elasticity, and provide a framework for future computational studies of drug-membrane interactions.