Insulin A chain: the smaller half of the blood-sugar hormone
One of two protein chains that together form insulin, the hormone the pancreas releases to keep blood sugar in check; not a drug on its own, used as a lab research tool.
A researcher, an agent, or an algorithm wrote down the sequence and picked a target to hit.
An AI model like OpenFold3 or AlphaFold built a 3D structure and scored how well it fits the binding site.
A second contributor repeated the computation on their own hardware and the scores matched.
Endogenous peptide — produced naturally and routinely synthesized for research
Fork this card to add platform evidence →
Endogenous peptide — receptor binding and activity established in published literature
Fork this card to add platform evidence →
What this is
The insulin A chain is the smaller of the two peptide chains that make up mature human insulin — the hormone the pancreas releases to regulate blood sugar. It is 21 amino acids long (the sequence GIVEQCCTSICSLYQLENYCN), encoded by the INS gene on chromosome 11. Together with the 30-residue B chain (/card/pep-04453), it constitutes intact insulin, joined through three disulfide bonds: one intra-chain (A6–A11, linking Cys⁶ to Cys¹¹ within the A chain itself), and two inter-chain bridges (A7–B7 and A20–B19) that lock the two chains together. All three disulfide bonds are essential — reducing any one of them abolishes insulin receptor binding.
The A chain is not used as a therapeutic agent in isolation; its biological significance is entirely as part of the intact, disulfide-bonded two-chain hormone. Several A chain residues — A1(Gly), A2(Ile), A3(Val), A19(Tyr), and A21(Asn) — are critical elements of insulin's receptor-binding surface (Site 1), and mutations at these positions dramatically reduce or abolish insulin receptor affinity (De Meyts 2015). The stored sequence here is the canonical human A chain as determined by Frederick Sanger; the long-acting analog glargine carries one A chain modification — A21Asn→Gly — which shifts the isoelectric point for subcutaneous depot formation, while all other approved insulin analogs (lispro, aspart, glulisine, detemir, degludec) modify only the B chain and use this same canonical human A chain sequence.
History
Insulin was discovered in 1921 by Frederick Banting, Charles Best, John MacLeod, and James Collip at the University of Toronto. The first human patient, Leonard Thompson, received bovine insulin extract in January 1922; his recovery from diabetic coma established insulin as the first effective treatment for type 1 diabetes and transformed what had been a uniformly fatal disease.
The A chain's sequence was determined by Frederick Sanger as part of the complete sequence determination of bovine insulin, work for which Sanger received the 1958 Nobel Prize in Chemistry. The intra-chain disulfide (A6–A11) posed a particular analytical challenge because it crosslinks two segments of the same chain, requiring careful oxidative and reductive strategies to assign disulfide connectivity unambiguously. The complete bovine insulin sequence was published in 1955; the human insulin sequence was subsequently determined and found to differ from bovine at A8 (Thr in human, Ala in bovine), A10 (Ile in human, Val in bovine), and from porcine insulin at A8 only — porcine insulin has Ala at A8, making it one position different from human (Sanger 1959).
Dorothy Hodgkin's group solved the first insulin crystal structure at 2.8 Å resolution in 1969, revealing the A chain's characteristic two-helix fold (an A1–A8 helix and an A13–A20 helix) stabilized by the A6–A11 disulfide. The helical arrangement places A1, A2, A3, A19, and A21 on a common face — the receptor-binding surface, confirmed structurally decades later in crystal structures of insulin bound to its receptor (De Meyts 2015).
The transition from animal-source to recombinant human insulin was driven partly by supply constraints and partly by immunological concerns: anti-insulin antibodies were detected in patients on animal-source insulins, with higher titers against bovine than porcine insulin, correlating with the greater A chain sequence divergence. Recombinant human insulin (Humulin, Eli Lilly, 1982) — the first recombinant DNA pharmaceutical — used the exact human A chain sequence.
What it does
The A chain, as part of intact insulin, participates directly in insulin receptor (IR) recognition. Separated A and B chains have negligible IR binding affinity on their own; the receptor-binding surface depends on the combined three-dimensional scaffold that forms only after disulfide bonding.
Receptor contact residues: Mutagenesis and structural studies identify two clusters of A chain residues as critical for IR binding (De Meyts 2015). On the N-terminal helix (A1–A8), residues A1(Gly), A2(Ile), and A3(Val) are direct receptor contacts; alanine substitution at A1 or A2 reduces binding affinity by 50–90%, and deletion of A1 abolishes binding. On the C-terminal helix (A13–A20), A19(Tyr) and A21(Asn) contact the receptor's αCT segment; substitution of A19(Tyr) with Leu reduces binding by 50%, and crystal structures confirm A21(Asn) makes direct contact with αCT (De Meyts 2015). Position A8 (Thr in human) lies outside the binding surface and tolerates species variation without affecting receptor affinity.
Intra-chain disulfide (A6–A11): The Cys⁶–Cys¹¹ crosslink constrains the relative orientation of the two A chain helices, fixing the A1/A2/A3 cluster and A19/A21 in their receptor-contact geometry. Reduction of this disulfide abolishes receptor binding even when the two inter-chain disulfides remain intact (Belfiore et al. 2017).
Biosynthesis: The A chain is generated in beta cells by proteolytic processing of proinsulin, an 86-amino-acid single-chain precursor in which the B and A chains are connected by the C-peptide. In the trans-Golgi network and secretory granules, prohormone convertases PC1/3 and PC2 cleave at the B30–C-peptide and C-peptide–A1 junctions, releasing C-peptide and mature two-chain insulin. The three disulfide bonds form during proinsulin folding in the endoplasmic reticulum, before C-peptide cleavage. Proinsulin has roughly 100-fold lower IR affinity than mature insulin, because the C-peptide linkage blocks the A21 receptor-contact position (Belfiore et al. 2017).
Metabolic actions of insulin (A chain in context): Intact insulin acts on the IR to mediate GLUT4 translocation to the cell surface in skeletal muscle and adipose tissue, glycogen synthesis, suppression of hepatic glucose production, and protein synthesis. These effects are coordinated through IR→IRS-1/2→PI3K→Akt signaling and IR→Ras→MAPK signaling.
Evidence
- Human: The A chain itself has no approved clinical use in isolation; it functions only as part of intact insulin. Human pharmacology data come from decades of clinical use of intact insulin formulations. Hereditary mutations at A chain receptor-contact residues — such as the Wakayama insulin variant (A3Val→Leu), which has approximately 1% of normal receptor affinity — cause hyperinsulinemia with insulin resistance in affected patients, providing in-human validation of the mutagenesis data (Belfiore et al. 2017).
- In vitro: Alanine-scanning mutagenesis across the A chain established the quantitative contribution of each residue to receptor affinity, summarized in De Meyts (2015). Crystal structures of insulin bound to a truncated "micro-receptor" fragment containing the L1 domain and αCT peptide confirmed direct A21(Asn)–αCT contact (De Meyts 2015).
Myths and misconceptions
- "The insulin A chain is the active chain and the B chain is just structural" — Both chains contribute critical receptor-binding residues to insulin's Site 1 surface. The A chain contributes A1, A2, A3, A19, and A21; the B chain contributes B12, B16, B24, B25, and B26. Neither chain alone can bind the receptor with meaningful affinity (De Meyts 2015).
- "Porcine insulin differs from human insulin in both chains" — Only the A chain differs: porcine A8 is Ala, human A8 is Thr — one position. The B chains of porcine and human insulin are identical. Bovine insulin differs from human at two A chain positions (A8 and A10), while the B chains remain identical (Sanger 1959).
- "Recombinant human insulin was produced by expressing the A chain alone" — Early recombinant insulin production (Genentech/Eli Lilly, 1978–1982) expressed the A and B chains separately in E. coli, then cleaved, reduced, and recombined them under oxidizing conditions to form the native disulfide-bonded dimer. This two-chain approach was later supplanted by proinsulin expression followed by enzymatic C-peptide removal. Neither approach uses the isolated A chain therapeutically.
Common questions
What does the A chain's intra-chain disulfide (A6–A11) do structurally? The A6(Cys)–A11(Cys) disulfide crosslinks the loop between the two A chain helices, constraining their relative orientation. Without this crosslink, the A chain becomes flexible and the two helices lose their fixed angular relationship. This disrupts the co-planar presentation of the A1/A2/A3 cluster (helix 1) and A19/A21 (helix 2) that together form the receptor-contact surface — structurally, it acts as a conformational lock that holds the A chain in its receptor-competent geometry.
How does the A chain sequence differ between human, porcine, and bovine insulins? Human insulin A chain: GIVEQCCTSICSLYQLENYCN (A8=Thr, A9=Ser, A10=Ile). Porcine insulin A chain: identical except A8=Ala (one position different). Bovine insulin A chain: A8=Ala, A10=Val (two positions different from human). The B chains are identical across all three species. These A chain differences are outside the receptor-binding surface and do not affect receptor affinity, but they influence immunogenicity — bovine insulin, with two non-human A chain residues, elicits higher anti-insulin antibody titers in humans than porcine insulin (Sanger 1959).
Does the C-peptide interact with the A chain? In proinsulin, the C-peptide N-terminus is connected to A21(Asn) through a dipeptide linker, blocking A21 from receptor contact and explaining proinsulin's ~100-fold lower IR affinity compared to mature insulin (Belfiore et al. 2017). After secretory granule processing, C-peptide is released as a separate peptide and circulates in equimolar amounts with insulin — making it a useful clinical marker of beta-cell secretion — but does not re-associate with the A chain.
Mechanism
Insulin binds its receptor (a dimeric receptor tyrosine kinase) through a two-step, cross-linking mechanism involving two binding surfaces on insulin (Site 1 and Site 2) and two receptor-binding sites located on the two receptor α-subunits (De Meyts 2015). The A chain's contribution is primarily to Site 1: residues A1, A2, A3, A19, and A21 form a contiguous receptor-contact surface together with B chain residues B12, B16, B24, B25, and B26. Upon insulin binding, the receptor undergoes conformational change that activates its intracellular tyrosine kinase domains (β-subunits), triggering autophosphorylation and downstream IRS-1/2→PI3K→Akt and Ras→MAPK cascades.
The IR exists in two isoforms — IR-A (exon 11 excluded) and IR-B (exon 11 included) — that differ in the αCT domain contacting A21. The A chain receptor-contact residues are conserved across both isoform interactions; all insulin analogs intended to preserve high receptor affinity retain the canonical A chain sequence at positions A1–A3 and A19 (Belfiore et al. 2017).
Related peptides
- Insulin B chain — the 30-aa partner chain; B chain + A chain = mature insulin via three disulfide bonds; B chain residues B12, B16, B24, B25, B26 contribute to receptor Site 1 alongside the A chain residues
- Semaglutide — GLP-1 receptor agonist that stimulates insulin secretion from pancreatic beta cells; approved for type 2 diabetes and obesity
- Glucagon — counter-regulatory peptide from pancreatic alpha cells; opposes insulin's metabolic actions at hepatic and peripheral receptors
Research directions for this peptide, selected from the current sources — hypotheses you can explore and model. None of it is proven yet; tap any one to see the full thinking.
Does the short tail of the A chain (its last three amino acids) mainly determine whether insulin also activates a growth-factor receptor linked to cancer risk?
If yes, future long-acting insulins could be engineered at the A chain tip to be safer for people who need lifelong insulin therapy, reducing any added cancer risk from inadvertent growth-factor signaling.
Could the A chain alone, when its own internal shape is preserved, sit in the insulin receptor's docking site and partially block insulin from binding?
If true, fragments of degraded insulin released in fat tissue or blood could act as natural brakes on insulin signaling, which could help explain why insulin resistance worsens in some conditions. This would give researchers a new angle on type 2 diabetes biology.
If you swap the A chain's internal sulfur-sulfur bond for a bond that cells cannot cut as easily, would the resulting insulin last longer once taken up by cells?
A more cell-stable insulin could lower the doses needed to control blood sugar in diabetes and might help patients whose cells degrade insulin too quickly. This could also reduce the frequency of injections.
Does swapping one amino acid at the tip of the A chain (Asn to Gly in glargine) change how tightly the two insulin chains are held together by their connecting sulfur bond?
If confirmed, this could explain subtle differences in how long-acting insulins behave after injection and could guide smarter engineering of future insulin analogs that last longer or break down more predictably in the body.
▸full evidence table2 metrics
| metric | value | tool |
|---|---|---|
| ipTM | 0.28903138637542725 | boltz-2 |
| ranking score | 0.6247049570083618 | boltz-2 |
▸structural qualityopenfold3
| metric | value | note |
|---|---|---|
| gpde | 1.959 | global PDE — lower = better |
| disorder | NaN | fraction disordered |
▸3-letter notation
▸recipeboltz-2 1.0
| parameter | value |
|---|---|
| model | boltz-2 1.0 |
| weights | — |
| hardware | nvidia_nim_api |
| mlx version | — |
| python | — |
| random seed | — |
| msa strategy | none |
| diffusion samples | 1 |
| runtime | — |
| predicted by | mlx@peptide |
| predicted at | 2026-04-24 |
▸citationbibtex
@peptide{pep04454,
sequence = {GIVEQCCTSICSLYQLENYCN},
target = {insr},
author = {peptidemodel},
year = {2026},
status = {bioassayed}
}