Non-covalent engagement of the cGAMP-binding pocket by STING inhibitor H-151: Characterization by molecular docking, Gaussian accelerated molecular dynamics, and MM-GBSA free energy analysis
Computational Biology and Chemistry, vol.124, 2026 (SCI-Expanded, Scopus)
- Publication Type: Article / Article
- Volume: 124
- Publication Date: 2026
- Doi Number: 10.1016/j.compbiolchem.2026.109164
- Journal Name: Computational Biology and Chemistry
- Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Applied Science & Technology Source, BIOSIS, Chemical Abstracts Core, Chimica, Compendex, EMBASE, MEDLINE, zbMATH, Academic Search Ultimate (EBSCO)
- Keywords: CDN-binding pocket, Gaussian accelerated molecular dynamics, H-151, MM-GBSA, Molecular docking, STING inhibitor
- Hacettepe University Affiliated: Yes
Abstract
STING drives microglial neuroinflammation in Alzheimer's disease; its inhibitor H-151 suppresses tau-induced STING activation in human iPSC-derived microglia. H-151 acts covalently at transmembrane palmitoylation site C91, yet its indole-urea scaffold is consistent with non-covalent binding at the cytosolic CDN-binding pocket. We characterize this secondary binding mode using a multi-stage computational pipeline on the 7SII cryo-EM structure of human STING chain A. Blind docking identified the CDN pocket as the preferred non-covalent site: 12 of 14 top poses clustered there with a 2.97 kcal/mol affinity advantage over the TM region in non-covalent scoring. Focused high-exhaustiveness docking converged tightly (-8.278 kcal/mol; modes 1–2 RMSD 0.30 Å). A 95.2 ns dual-boost GaMD simulation confirmed pose stability: 83.8% of frames belonged to a single binding mode, with the pharmacophore anchored and the ethylphenyl tail sampling rotameric states (inter-cluster RMSD 2.32 Å). MM-GBSA analysis yielded ΔHbind = -41.4 ± 3.7 kcal/mol, dominated by van der Waals interactions. Per-residue decomposition and ProLIF fingerprinting across 4800 frames identified Y261 as the dominant contact (MM-GBSA −4.81 kcal/mol; H-bond donor occupancy 95.6%), with T263 and A262 as secondary hot spots. The cGAMP phosphate-anchoring residues R238 and Y240 were absent from the H-151 hot-spot map, and cGAMP redocking failed to converge (centroid deviation 9.35 Å), demonstrating pharmacophoric distinction. BBB permeability was concordantly predicted by three tools (pkCSM log BB = 0.33; DeepPK p = 0.998), supporting CNS access. These results identify a secondary CDN-pocket pharmacophore for H-151 and provide a structural template — the T263/Y261/A262 sub-pocket — for non-covalent STING antagonist design competitive with cGAMP.