PaCS-perm: Insights into Membrane Permeation by Drugs using Parallel Cascade Selection Molecular Dynamics

Drug efficacy is often limited by low membrane permeability that restricts access to the site of action. However, improving drug permeation is limited by our understanding of the membrane permeation process, which is complex, and there is a need for improved, predictive models of model membrane permeation. In this study, we investigate the combination of parallel cascade molecular dynamics (PaCS-MD) simulations with umbrella sampling (US) to calculate the free energy barriers to permeation. We denote this protocol PaCS-perm. Using PaCS-perm, we have modelled the membrane permeation of 20 structurally diverse, charge-neutral drugs using a POPC/cholesterol membrane bilayer and calculated permeation barriers. Examination of the permeation trajectories reveals initial formation of a bridge between polar drugs and bulk water. As the drug reaches the membrane centre, the bridge breaks and the drug passes through the second membrane leaflet. The strong drug-water interaction is evident as hysteresis in the permeation potentials of mean force (PMF). Comparison of the predicted free energy barriers for drug entry into the membrane with experimental permeation measurements from parallel artificial membrane permeability assay (PAMPA) experiments (LogPeff) shows strong corelation (R = −0.76). The PaCS-perm protocol performs substantially better than using steered molecular dynamics with umbrella sampling (SMD-US) method (R = −0.62) using a 5 ns umbrella sampling time. This study validates the PaCS-perm protocol as a method for studying bilayer permeation and shows that the PaCS method effectively provides atomic-scale structure insights into permeation processes.