Human insulin. Masses and charge calculated with the residue masses and pKa set used by the tools on this site, for the native form with three disulfide bonds.
Discovery
By the early twentieth century it was known that removing the pancreas caused diabetes in dogs, and that the pancreas must release something into the blood that controlled glucose. Isolating it proved difficult, because the digestive enzymes of the same gland destroyed any extract.
In 1921 and 1922, in Toronto, Frederick Banting and Charles Best, working in John Macleod's laboratory, prepared extracts that lowered blood glucose in diabetic dogs, and the biochemist James Collip developed a purification that made the extract safe enough for human use. In January 1922 a 14-year-old boy, Leonard Thompson, became the first person to be treated. Banting and Macleod received the Nobel Prize in Physiology or Medicine in 1923 and shared it with Best and Collip respectively. Within a few years insulin was being produced industrially from cattle and pig pancreas, and a diagnosis that had been fatal became a chronic condition.
Insulin went on to feature in a remarkable series of scientific firsts:
| Year | Milestone |
|---|---|
| 1955 | First complete amino acid sequence of any protein, by Frederick Sanger (Nobel Prize 1958) |
| 1963–1966 | First chemical synthesis of a protein, achieved independently by groups in Aachen, Pittsburgh and Shanghai |
| 1967 | Discovery of proinsulin, the single-chain precursor, by Donald Steiner |
| 1969 | Three-dimensional structure solved by X-ray crystallography, by Dorothy Hodgkin's group |
| 1982 | First medicine made by recombinant DNA technology to be approved: recombinant human insulin |
Structure
Insulin consists of two peptide chains. The A chain has 21 residues and the B chain 30. They are held together by two interchain disulfide bonds, A7–B7 and A20–B19, and the A chain carries a third, intrachain disulfide between A6 and A11. With 51 residues, insulin sits exactly on the conventional boundary between peptides and proteins, and is usually called a peptide hormone.
In three dimensions the A chain forms two short antiparallel helices, and the B chain a central α-helix followed by an extended strand at its C-terminus. The disulfides hold the two chains in a compact globule with a small hydrophobic core. The fact sheet above gives the mass of this native, disulfide-linked form. The MW calculator computes each chain separately and in reduced form, so the two chain masses must be added and 6.05 Da subtracted for the three disulfides, as explained in the guide to disulfide bonds.
Insulin's calculated pI of about 5.3, in close agreement with the measured value of 5.3–5.4, reflects its four glutamates against a single lysine and two arginines, plus two histidines. It carries a net negative charge at physiological pH and is least soluble around pH 5, a property exploited for crystallisation and, as described below, in the design of long-acting analogues.
Biosynthesis: from one chain to two
Two separate chains raise an obvious question: how does the cell ensure that the right chains pair and the right cysteines connect? The answer is that insulin is made as a single chain and cut apart afterwards.
- The gene encodes preproinsulin, 110 residues long. Its N-terminal signal peptide directs it into the endoplasmic reticulum and is removed there.
- The remaining proinsulin, 86 residues, folds as a single chain: B chain, then a connecting segment, then A chain. In this form the three disulfides form correctly, with the help of protein disulfide isomerase.
- In the secretory granules, the prohormone convertases PC1/3 and PC2 cut at pairs of basic residues on either side of the connecting segment, and carboxypeptidase E trims the remaining basic residues.
- The result is mature insulin plus the 31-residue C-peptide, which are released together in equal amounts.
Because C-peptide is co-secreted one-to-one with insulin but is not present in pharmaceutical insulin, its measurement distinguishes the body's own insulin production from injected insulin. This is a classic illustration of how processing of a precursor, described in the guide to post-translational modifications, produces the active peptide.
Dimers, hexamers and zinc
The active form of insulin, the one that binds the receptor, is the monomer. At the high concentrations found in the secretory granules of pancreatic β-cells, however, insulin associates into dimers, and three dimers assemble around two zinc ions into a hexamer. The hexamers pack into dense crystalline cores inside the granules, an efficient way to store a large amount of hormone in a small space.
Once released and diluted in the blood, hexamers dissociate into monomers within seconds to minutes. For injected insulin, however, the rate of this dissociation at the injection site controls how fast the hormone reaches the circulation. This single physical property became the main target of insulin engineering.
Engineering insulin analogues
Since the 1990s, most insulin used in medicine has been modified from the human sequence to change how quickly it is absorbed. The modifications illustrate the principles covered elsewhere on this site with unusual clarity: small changes in sequence alter self-association, charge and binding to other proteins.
| Analogue | Change from human insulin | Principle |
|---|---|---|
| Insulin lispro | B28 Pro and B29 Lys swapped | Weakens the dimer interface, so hexamers dissociate faster |
| Insulin aspart | B28 Pro → Asp | Charge repulsion at the dimer interface |
| Insulin glulisine | B3 Asn → Lys, B29 Lys → Glu | Reduced self-association |
| Insulin glargine | A21 Asn → Gly; two Arg added to the B-chain C-terminus | Raises the pI towards neutrality, so it precipitates after injection |
| Insulin detemir | B30 Thr removed; C14 fatty acid on B29 Lys | Binds serum albumin, which slows its release |
| Insulin degludec | B30 Thr removed; C16 diacid via γ-Glu on B29 Lys | Forms long multi-hexamer chains at the injection site |
Glargine is a textbook example of pI engineering. The two added arginines, together with the A21 substitution, shift the calculated pI from about 5.3 to about 7.0 by the method used on this site; reported experimental values are around 6.7. Glargine is formulated as a clear solution at acidic pH, where it is fully soluble. After injection, the neutral pH of the tissue is close to its isoelectric point, so it forms a fine precipitate that dissolves slowly. The isoelectric point guide explains the underlying principle. The A21 glycine replaces an asparagine that is prone to deamidation in acid, the reaction described in the guide to peptide stability.
Detemir and degludec use lipidation, the strategy of attaching a fatty acid so that the peptide binds albumin or forms larger assemblies. The same principle has since been applied to many other peptide drugs.
Insulin across species
Insulin is highly conserved among mammals. Pig insulin differs from human insulin at a single position, the last residue of the B chain (B30 alanine instead of threonine). Cattle insulin differs at three: A8, A10 and B30. These small differences mattered historically, because for sixty years patients were treated with animal insulins, and the closer similarity of the pig sequence made it less likely to provoke antibodies. Before recombinant production, human insulin was made semi-synthetically by exchanging the B30 residue of pig insulin enzymatically.
Insulin-like peptides are found throughout the animal kingdom, from nematodes to molluscs. A striking example comes from cone snails: some species release a small, fast-acting insulin into the water to induce hypoglycaemic shock in the fish they hunt.
The insulin receptor
Insulin acts through a receptor tyrosine kinase on the surface of cells in muscle, fat, liver and other tissues. The receptor is a disulfide-linked dimer of two α and two β subunits. Binding of insulin to the extracellular α subunits changes the arrangement of the dimer and activates the tyrosine kinase domains inside the cell, which phosphorylate each other and a set of substrate proteins. These in turn trigger the translocation of glucose transporters to the cell surface and a broad program of metabolic changes. The role of tyrosine phosphorylation in this signalling is described in the guide to post-translational modifications.
Working with insulin in the laboratory
- Solubility. Insulin is least soluble near its pI, around pH 5.3. It dissolves readily in dilute acid, such as 0.01 M HCl, and is stable in neutral buffers once dissolved.
- Fibrillation. Under agitation, at elevated temperature, or at low pH, insulin forms amyloid fibrils. It is one of the most widely used model proteins for studying amyloid formation.
- Quantification. With four tyrosines, no tryptophan and three cystines, insulin has a calculated extinction coefficient at 280 nm of about 6,335 M⁻¹cm⁻¹ by the method used in the MW calculator.
- Mass spectrometry. Intact insulin is readily detected by ESI as multiply charged ions; reduction with DTT separates the two chains, whose masses confirm the sequence.
Frequently asked questions
Why is insulin not taken as a tablet?
Like other peptides, insulin is digested by proteases in the stomach and intestine, and the small amount that survives is poorly absorbed across the gut wall. Oral delivery of peptides is an active area of research, and one oral peptide drug of a different class has been approved, but insulin itself is still administered by injection or other routes.
Is insulin a peptide or a protein?
Both descriptions are used. With 51 residues it lies on the conventional boundary, and it is generally called a peptide hormone.
What is the difference between insulin and C-peptide?
C-peptide is the connecting segment cut out of proinsulin during processing. It is released in equal amounts with insulin and is not needed for insulin's activity, but its measurement reveals how much insulin the body itself is making.
Why does the MW calculator give a different mass for insulin?
Because it calculates a single linear chain with free cysteines. For insulin you need to enter the A and B chains separately, add their masses, and subtract 2.016 Da for each of the three disulfide bonds.
References
- Banting FG, Best CH, Collip JB, Campbell WR, Fletcher AA (1922) Pancreatic extracts in the treatment of diabetes mellitus. Canadian Medical Association Journal 12:141–146.
- Ryle AP, Sanger F, Smith LF, Kitai R (1955) The disulphide bonds of insulin. Biochemical Journal 60:541–556.
- Adams MJ, Blundell TL, Dodson EJ, et al. (1969) Structure of rhombohedral 2 zinc insulin crystals. Nature 224:491–495.
- Steiner DF, Oyer PE (1967) The biosynthesis of insulin and a probable precursor of insulin by a human islet cell adenoma. PNAS 57:473–480.
- Mathieu C, Gillard P, Benhalima K (2017) Insulin analogues in type 1 diabetes mellitus: getting better all the time. Nature Reviews Endocrinology 13:385–399.