Abstract
The recent availability of several ligand-bound hERG/Kv11.1 cryo-EM structures creates an opportunity for prospective structure-based modeling, but it also introduces a practical problem: a new ligand may be compatible with only a subset of the available receptor conformations, and docking may generate several plausible poses within each template. We therefore evaluated a template-aware pose-selection framework using the astemizole-, pimozide-, and E4031-bound hERG structures as experimental reference environments, with apo 8ZYN included as a receptor-state control. The ligand set separated into two study-defined chemical regions: a low-polarity, highly lipophilic Group I represented by astemizole and pimozide, and a more polar Group II represented by E4031. Induced-fit docking (IFD) improved self-docking relative to rigid Glide, whereas apo 8ZYN did not reproduce ligand-bound behavior during short molecular dynamics (MD) assessment. Across the retrospective ligand-bound analyses, six of seven group-matched systems met the dual MD-quality criterion, whereas none of eight group-mismatched systems did. Highest static shape similarity was the most consistent pose-selection criterion for Group I, while an interaction-fingerprint (IFP)-favored 30:70 shape:IFP composite was retained provisionally for the single retrospective Group II reference. Application to dofetilide, ibutilide, and terfenadine produced binding hypotheses with residue-contact patterns consistent with independent hERG literature. Together, these results support using ligand-bound hERG structures as chemically selective reference environments for pose prioritization rather than as interchangeable receptor templates.