CHARACTERIZATION OF STING INTRINSIC DYNAMICS: PAVING THE WAY FOR VARIANT INTERPRETATION AND RATIONAL PHARMACOLOGY

The Stimulator of Interferon Genes (STING) protein is a crucial mediator of type I interferon signaling. STING dysfunction due to genetic variants is related with autoinflammatory conditions and cancer. Given that missense variants can lead to disease by altering protein structure and function, rational interpretation of their effects can benefit patients with undiagnosed variants. We previously developed a tool for variant interpretation and prioritization using a structure-assisted approach. However, the pathogenicity of the analyzed variants remained unclear. Thus, a detailed characterization of STING's structural dynamics is essential to uncover the complexity of the conformational landscape, improving variant interpretation and pharmacologic rationalization. Here, we integrated a detailed description of intrinsic dynamics with Normal Modes Analysis and deep mutational scanning (DMS) calculations to obtain free-energy variations due to mutations.  The full-length STING model was built from the Cryo-EM apo structure (PDB: 6NT5). All missing loops were grafted into the structure, and side-chain flips were corrected using the Molprobity server with Reduce. Atomistic NMA calculations were performed in R using the Bio3D library. To identify potential functional movements, we compared the low-frequency normal modes with structural variations that lead to the experimentally observed activated state (PDB: 8IK3) and a disease-related mutant (PDB: 8GZS). Moreover, perturbation response scanning (PRS) was applied to identify residues involved in propagation of allosteric signals. DMS free-energy changes were calculated using pyRosetta for the apo and activated STING conformations. We identified a subset of 8 intrinsic movements that account for 90% of the total STING fluctuations. Overall, these modes capture interdomain transitions in which the ligand-binding domain moves simultaneously with the transmembrane domain (TMD) and the C-terminal region, indicating an intrinsic allosteric potential. Future PRS calculations will help identify key residues involved. While a single mode described STING activation, providing the structural basis for understanding its basal activity, other modes revealed rotations and stretching of TMD helices, likely related to the opening of an ion channel, as recently proposed. We are currently identifying conformational hotspots. Moreover, comparing DMS energy profiles in apo and active states will help evaluate the impact of mutations on protein stability across different structural contexts.