Computational Docking of Bioactive Ligands

Docking experiments were carried out using the full-length predicted structures of the target proteins, with the most optimal quality parameters. The simulations were performed using the AutoDock Vina v1.1.2 algorithm [42 (link)]_, integrated within the YASARA Structure. In the case of SERCA2b, the docking grid box included a potential binding pocket situated at the interface between the transmembrane and cytosolic domains, overlapping with the binding site of SERCA inhibitor BHQ (2,5-di-tert-butylbenzene-1,4-diol) [25 (link)]. The binding pocket was chosen based on previous studies, which suggested that the binding site of BHQ might overlap with the potential binding pocket of hypericin, a photosensitizing agent with a large molecule, similar to porphyrinic derivatives [43 (link)]. For potassium channels KCa1.1 and K_ATP, the grid box was set around the residues involved in heme binding, which were confirmed through previous mutation studies [21 (link),22 (link),44 (link)]. BHQ and heme were used as positive controls.
Both protein and ligand structures were protonated according to the physiological pH. The docking experiments were performed with flexible residues, and 12 docking runs were executed for each ligand. The predicted protein-ligand complexes were refined by minimization with AMBER14 force field. The results were retrieved as the binding energy (ΔG, kcal/mol) and predicted dissociation constant (Kd, µM). The predicted binding poses and molecular interactions were analyzed using ChimeraX v1.4 [45 (link)] and BIOVIA Discovery Studio Visualizer (BIOVIA, Discovery Studio Visualizer, Version 17.2.0, Dassault Systèmes, 2016, San Diego, CA, USA).