Ligand Dissociation Pathways from Membrane Receptors Revealed by Weighted Ensemble Simulations

G-protein-coupled receptors (GPCRs) are pivotal in cellular signal transduction and serve as key drug targets. Among them, the β-adrenergic receptors (β₁AR and β₂AR) regulate cardiovascular function and are activated by endogenous catecholamines, norepinephrine (NE), and epinephrine (EP). While ligand interaction kinetics, especially the dissociation rate constants (<i>k</i>_off), have been recognized as a key parameter for drug development, accurate and atomistic modeling of these processes remains a major challenge due to the rarity and slow time scales of unbinding events. Hence, we applied the weighted ensemble method (WE), an enhanced sampling method, in molecular dynamics (MD) simulations to sample the dissociation pathways of NE and EP from β₁AR and β₂AR. To analyze and visualize the results, we used machine learning-based techniques, such as principal component analysis (PCA), <i>t</i>-distributed stochastic neighbor embedding (<i>t</i>-SNE), and density-based spatial clustering of applications with noise (DBSCAN). These methods enabled unbiased sampling of ligand dissociation from membrane-embedded GPCRs while preserving kinetics. We found that NE and EP dissociate via front-side pathways in β₁AR but favor a back-side pathway in β₂AR. We also found that WE simulations could yield reasonable <i>k</i>_off values.