Abstract
Ligand-target dissociation rates (k(off)) strongly correlate with efficacy and safety profiles, as well as with the therapeutic effect of drugs. As a prototypical example, muscarinic receptor antagonists used as bronchodilators show similar affinity profiles toward the muscarinic M3 receptors (M3R) and M2 receptors (M2R), whereas their kinetic selectivity toward M3R avoids the adverse effects that a prolonged inhibition of M2R would induce at the cardiac level. Previous studies on the dissociation kinetics of human M3R showed that the residence time and binding affinity of muscarinic antagonists are deeply affected by the presence of specific mutations. The aim of our work was to reproduce the rankings of these experimental kinetic rates through an approach based on the application of adiabatic-bias molecular dynamics (ABMD) simulations using Path Collective Variables (PCVs), PCV-ABMD. Employing this methodology, we simulated the translocation of tiotropium, a long-acting bronchodilator targeting M3R, from the orthosteric site to the extracellular vestibule, without considering the whole unbinding process. The estimated times necessary for translocation displayed a strong correlation with the experimental pk(off) values. Moreover, a thorough analysis of protein-ligand contacts provided deeper insights into the mechanism of unbinding of muscarinic antagonists. The newly described PCV-ABMD protocol captured relevant metastable states and offered a reliable approach for the prediction of kinetic selectivity in sets of mutants.