Abstract
High‑affinity host–guest systems, such as Cucurbit[n]uril (CBn) macrocycles, are vital across various scientific and technological fields – such as targeted drug delivery, smart (self‑healing) materials, sensitive biosensors, and molecular diagnostics – due to their exceptional molecular recognition capabilities. Molecular simulation (MS)-based predictions of ligands’ binding poses and affinities on the macrocycles would greatly help optimize such host-guest systems. Yet, the poor accuracy of force fields (FFs) for these synthetic receptors has limited the applicability of MS thus far. Here, we demonstrate that incorporating electron density-derived Lennard-Jones parameters and charges into FFs can drastically improve the accuracy of free-energy calculations for these systems. As a test case, we focus on five adamantane derivatives and two diamantane derivatives in complex with one of the macrocycles, CB7. Our free energies of binding, calculated via multiple-walker well-tempered funnel metadynamics, turned out to be in fair agreement with experiment for all the adamantane molecules. For the larger diamantane molecules, we still observe a discrepancy with experiments, which calls for deeper investigation. Overall, the calculations also provide insight into the binding mechanism and the role of the solvent. In particular, the chemical structures of the ligands and the ion strength play an important role in the binding process.
Supplementary materials
Title
Supporting Information
Description
Additional information on time evolution of CVs, concentrations effects on binding free energy, and convergence plots.
Actions