Structural basis of insulin receptor antagonism by bivalent site 1-site 2 ligands S961 and Ins-AC-S2

Article https://doi.org/10.1038/s41467-026-73851-1 Structural basis of insulin receptor antagonism by bivalent site 1-site 2 ligands S961 and Ins-AC-S2 Received: 7 August 2025 Check for updates 1234567890():,; 1234567890():,; Accepted: 13 May 2026 Amber Vogel Danny Chou 1 2 , Alan Blakely1, Yuankun Dao & Christopher P. Hill 1 2 , Nai-Pin Lin 2 , Congenital hyperinsulinism is a rare genetic disease characterized by overproduction of insulin. One class of potential treatments is insulin receptor antagonists like S961 and Ins-AC-S2, which comprise segments for binding each of the two insulin-binding sites (site 1 and site 2) on the receptor. Notably, S597 – containing the same receptor binding segments as S961 but in the opposite order (site 2-site 1) – is an insulin receptor agonist rather than an antagonist. Using cryo-EM, we show how both S961 and Ins-AC-S2 bind an inactive conformation of the receptor, thereby explaining their antagonism. Furthermore, our structures reveal how agonist vs. antagonist activity is influenced by the order of site 1- and site 2-binding modules in bivalent ligands. Additionally, we show subtle differences between the receptor-binding mechanisms of S961 and Ins-AC-S2, which include displacement or engagement of αCT, and a binding interface between the Ins-AC-S2 insulin and the receptor FnIII-2/insert domains. These structural insights may inform development of next generation insulin receptor antagonists for treatment of congenital hyperinsulinism. Congenital hyperinsulinism (HI) is a rare genetic disease, occurring in roughly 1/25,000-1/50,000 infants, that is characterized by overproduction of insulin by pancreatic beta cells, leading to persistent hypoglycemia1,2. Left untreated, patients with HI can experience seizures, brain damage, learning disabilities, and death. HI can be caused by a vast number of genetic mutations and other transient causes and metabolic diseases, making treatment complicated and case-specific3. Most available treatments target endogenous insulin secretion pathways but are not universally effective, carry serious side effects, and drug resistance can develop3. This forces many patients to undergo surgical removal of part or most of the pancreas, which can result in type 1 diabetes. Thus, new treatment options for patients with HI are desirable therapeutic targets. Insulin signaling is mediated by binding of insulin to the insulin receptor (IR), a member of the receptor tyrosine kinase family that is critical for import of glucose into the cell4. In the absence of insulin, the fibronectin stalks that point to the cell membrane are separated and hold the intracellular tyrosine kinase (TK) domains apart, preventing trans-autophosphorylation of the TK domains required for downstream signaling (Fig. 1a). Binding of insulin to the ectodomain (ECD) results in dramatic conformational changes that bring the fibronectin stalks together, allowing for receptor activation (Fig. 1b)4,5. IR is a disulfide crosslinked homodimer, where each protomer contains α and β chains comprising multiple domains (Fig. 1c). Each α chain contains two binding sites for insulin: site 1, composed of the leucine-rich 1 (L1) domain and the alpha C-terminal helix (αCT) of the other protomer; and site 2, composed of the fibronectin type-III 1 (FnIII-1) domain. Differential splicing of the IR gene results in A and B isoforms, which differ by the absence or presence of a 12-residue extension at the C-terminus of the α chain6. A crystal structure shows how apo/inactive IR adopts an inverted-V conformation with the extracellular fibronectin stalks separated by ~120 Å (Fig. 1d)7. Due to its dimeric structure, 1 Department of Biochemistry, Spencer Fox Eccles School of Medicine, University of Utah, Salt Lake City, UT, USA. 2Division of Endocrinology and Diabetes, e-mail: ; Department of Pediatrics, School of Medicine, Stanford University, Palo Alto, CA, USA. Nature Communications | (2026)17:5068 1 Article https://doi.org/10.1038/s41467-026-73851-1 a c b d L1 ectodomain FnIII-1 CR αCT TK apo IR (inactive) site 1 insulin FnIII-2 L1 FnIII-3 apo IR (PDB: 4ZXB) C-tail TK domains e αCT ID β chain TM JM membrane CR ctin sta FnIII-1 FnIII-2 ID FnIII-2 FnIII-3 Fibron e insulin L2 lk α chain L2 bound IR (active) f site 1’ insulin g site 1 insulin h human insulin (agonist) A chain GIVEQCCTSICSLYQLENYCN L2 FnIII-1 CR L2 L2 B chain FVNQHLCGSHLVEALYLVCGERGFFYTPKT CR L1 S597 (agonist) αCT FnIII-1 CR S597 L1 site 2 insulin site 2’ insulin FnIII-2 ID FnIII-1 site 2 segment L1 FnIII-2 FnIII-2 site 1 segment S961 (antagonist/partial agonist) GSLDESFYDWFERQLGGGSGGSSLEEEWAQIQCEVWGRGCPSY ID FnIII-3 ID FnIII-3 S597 SLEEEWAQIECEVYGRGCPSESFYDWFERQL site 1 segment site 2 segment FnIII-3 Ins-AC-S2 (antagonist) GIVEQCCTSICSLYQLENYCGGSLPETGGGSLEEEWAQIQSEVWGRGSPSY 1:4 IR:insulin (PDB: 6PXV) 1:1 IR:insulin (PDB: 7STI) 1:2 IR:S597 (PDB: 8DTL) FVNQHLCGSHLVEALYLVCGERGFFYTPK site 2 segment insulin (site 1 segment) Fig. 1 | Insulin receptor and ligand structure. a Cartoon of the apo/inactive fulllength insulin receptor (FL-IR) illustrating the physical separation of the tyrosine kinase (TK) domains. b Cartoon of insulin-bound/active FL-IR showing dimerization of the TK domains. Cartoons of a and b generated using PDBs: 6PXV10, 4ZXB7, and Alphafold344. The ectodomain (ECD) is boxed. c Domain organization and disulfide linkages of FL-IR drawn using IBS45. The α and β chains are indicated and both protomers are shown. d Structure of the apo IR ECD in the inverted-V conformation (PDB: 4ZXB7). e Structure of the IR showing only the ECD bound to four insulins in the activated T-shape conformation (PDB: 6PXV10). f Structure of IR ECD bound to one site 1 insulin in an asymmetric, active conformation (PDB: 7STI12). g Structure of IR ECD bound to two S597 molecules (PDB: 8DTL21), where IR adopts a variation of the active T shape conformation. For structures shown in panels d-g, one IR protomer is shown in gray and the other color-coded by domain according to the color scheme shown in panel c. L1: leucine-rich 1 domain. CR: cysteine-rich domain. L2: leucine-rich 2 domain. FnIII: fibronectin type-3 domain. ID: insert domain. TM: transmembrane helices. JM: juxtamembrane domain. h Sequences of human insulin and IR ligands used in this study with site 1- and site 2-binding segments indicated. Human insulin A and B chains: magenta and gold, respectively. S597 and S961 site 1 and site 2 segments: purple and salmon, respectively. S961 linker, gray. Ins-AC-S2 A chain binding segment for site 2, salmon; linker, gray. Disulfide bonds are indicated. there are four binding sites on IR: site 1 and site 2 on one receptor half, and the corresponding site 1’ and site 2’ of the other receptor half (Fig. 1e). Multiple cryo-EM structures show activated complexes with 1:1, 1:2, and 1:3 IR:insulin stoichiometries8–13. Interestingly, the (...truncated)