Abstract
Chronic, dysregulated inflammation contributes to colitis-associated colorectal cancer (CRC), and the cGAS-STING pathway represents a central but therapeutically challenging node because both insufficient and excessive STING activity can be pathogenic. Here, we integrate AlphaFold3 (AF3) receptor modeling, diffusion-based docking, and explicit-solvent molecular dynamics (MD) simulations to characterize how the marine briarane diterpenoid excavatolide B (ExcB) engages the human STING (hSTING) cyclic dinucleotide (CDN)-binding cleft. The structural integrity of the AF3 hSTING model was validated through both intrinsic confidence scores (pLDDT, PAE) and comparative benchmarking against experimental CTD structures (PDB: 4EF5, 6A05). Notably, the local geometries of key pocket-defining residues-including His157, Tyr167, and Thr263-remained consistent with established crystallographic data. Across three independent 100 ns MD replicas, ExcB exhibits a consistent spatial progression from an entrance-proximal pose at the solvent-accessible rim of the cleft (Site-2) to a more embedded, non-canonical corridor on the Cys148-adjacent side (Site-2'). Distance and contact analyses support a predominantly non-covalent within-cleft mechanism and do not indicate a persistent approach to the literature-reported covalent regime near Cys91. Residue-level profiling over the stabilized sampling window defines a reproducible corridor "contact signature" and reveals within-cavity sub-pose diversity rather than a single rigid bound pose. Mechanistically, competitive docking of the native agonist cGAMP to ExcB-conditioned receptor snapshots yields consistently less favorable docking outcomes in ExcB-conditioned conformations than docking to the native/open receptor; retaining ExcB coordinates does not further penalize cGAMP, supporting a receptor-reshaping (conformational conditioning) component rather than persistent static steric clash. Our findings characterize ExcB as a non-covalent modulator targeting a cryptic pocket within the STING CDN-binding cleft, establishing a structural basis for targeted mutagenesis and structure-activity relationship (SAR) studies.