How many peptide bonds are in a peptide?
The counting rule, its exceptions and the chemistry behind it
Matic Broz, PhD7 min
Count the amino acids and subtract one. A linear peptide with n amino acid residues has n − 1 peptide bonds, so a dipeptide has 1, a tripeptide has 2 peptide bonds, and a chain of 100 residues has 99. Each bond joins one residue to the next, and the first residue has nothing before it to join.
The rule holds for a single open chain. A protein built from several chains loses one bond per chain, which is why insulin's 51 residues carry 49 peptide bonds. A head-to-tail cyclic peptide gains one, so a ring of n residues has n bonds. Disulfide bridges never count, because they are not peptide bonds at all. The worked examples below cover each case, followed by why the bond is flat and why water takes centuries to break it.
What is a peptide bond?
A peptide bond is the amide bond that joins two amino acids: the carbonyl carbon of one bonds to the nitrogen of the next, with the formal loss of one water molecule, a reaction called condensation.[1] Repeated along a chain, these C(=O)NH groups form the backbone of every peptide and protein. What remains of each amino acid inside the chain is a residue.
The bond is not an ordinary single bond. Resonance spreads electrons across the oxygen, carbon and nitrogen, giving the C-N bond partial double-bond character. The six atoms of the peptide group, the two alpha carbons and the C, O, N and H between them, are therefore most stable in one plane, and the chain cannot rotate freely about the bond.[2]
A protein folds instead by turning the single bonds on either side of each alpha carbon, the phi and psi angles plotted in a Ramachandran plot. The angle across the peptide bond itself, omega, almost always sits near 180°, the trans form, with the neighbouring alpha carbons on opposite sides.[2] These are the three angles PeptideBuilder takes as input when it builds a model peptide.
Cis peptide bonds, with omega near 0°, are rare except where the next residue is proline.[3] In a set of 571 protein structures, Jabs and colleagues found only 43 cis bonds that did not precede a proline.[4]
In cells, peptide bonds are made on the ribosome. Its peptidyl transferase center, the site that joins each incoming amino acid to the growing chain, is built entirely from ribosomal RNA.[5] Every residue added during translation creates exactly one new peptide bond, and that is where the counting rule comes from.
How do you count peptide bonds in a peptide?
For one linear chain, subtract 1 from the number of residues. A connected chain of n residues contains n − 1 amide bonds, each linking a residue to the one before it.[6]
For a protein made of separate chains, count each chain on its own, or subtract the number of chains from the total number of residues:
Written as Gly-Ala-Ser, a tripeptide shows its two peptide bonds as the two hyphens between residue symbols. The count does not depend on which amino acids make up the chain.
| Molecule | Structure | Residues | Peptide bonds |
|---|---|---|---|
| Dipeptide | One linear chain | 2 | 1 |
| Tripeptide | One linear chain | 3 | 2 |
| Tetrapeptide | One linear chain | 4 | 3 |
| Pentapeptide | One linear chain | 5 | 4 |
| Hexapeptide | One linear chain | 6 | 5 |
| Polypeptide of 100 residues | One linear chain | 100 | 99 |
| Human insulin | Two chains of 21 and 30 residues | 51 | 49 |
| Cyclosporin A | Head-to-tail ring | 11 | 11 |
| Glutathione | One chain, first link through a side chain | 3 | 2, of which 1 is an isopeptide bond |
The last three rows are the cases where subtracting 1 from the total gives the wrong answer, or the right number for an unexpected reason. The next two sections explain them.
How many peptide bonds are in insulin?
Mature human insulin contains 49 peptide bonds: 20 in its 21-residue A chain and 29 in its 30-residue B chain.[7]
Subtracting 1 from the combined 51 residues gives 50, which treats insulin as one chain. It starts as one: the beta cell makes proinsulin, a single chain in which a connecting C-peptide joins the future B and A chains. Enzymes cut the C-peptide out before secretion, leaving the two chains of mature insulin.[7]
Three disulfide bonds hold those chains together, two between the A and B chains and one within the A chain.[7] They join the sulfur atoms of cysteine side chains, not a carbonyl carbon to a nitrogen, so they add nothing to the count.
Do cyclic peptides and side-chain links change the count?
A head-to-tail cyclic peptide has as many peptide bonds as residues, because one more peptide bond closes the ring. IUPAC calls a ring built only from backbone peptide bonds a homodetic cyclic peptide.[1] The immunosuppressant cyclosporin A is one example, with 11 residues joined into an 11-bond ring.[8] Predicting the shape of such a ring needs a model that knows the chain has no ends, which is what HighFold adds to AlphaFold2.
Not every ring closes through a peptide bond. A peptide can also cyclize through an ester, a disulfide or a new carbon-carbon bond, and then its backbone keeps the linear count of n − 1.[1]
Side-chain amides are the subtle case. IUPAC calls every amide bond in a peptide a peptide bond, then separates eupeptide bonds, which join the main carboxyl group of one amino acid to the alpha amino group of the next, from isopeptide bonds, which involve a side-chain group.[1][6] An isopeptide bond is still a peptide bond, just not a link between alpha carbons.
Glutathione shows why the label matters. Its glutamate is attached to cysteinylglycine through the side-chain carboxyl, so the tripeptide's two peptide bonds are one isopeptide bond and one eupeptide bond.[9] The n − 1 rule gives 2; counting only alpha peptide bonds gives 1. An isopeptide bond that cross-links two side chains instead of extending a chain adds a bond beyond the backbone count, as a ring closure does.
Why are peptide bonds so hard to break?
Hydrolysis reverses the condensation: water adds across the bond and splits it back into a carboxyl group and an amino group. Without a catalyst the reaction is very slow. Radzicka and Wolfenden measured it at high temperature and extrapolated to neutral water at 25 °C, finding half-lives of about 350 years for glycylglycine and 600 years for an internal peptide bond.[10]
The partial double-bond character that keeps the bond flat also helps it resist hydrolysis.[2] Cells and laboratories therefore break peptide bonds with proteases, enzymes that cut next to particular residues. PeptideCutter maps where common proteases and chemical reagents would cut a given sequence.
The water lost when each bond forms shows up in molecular weight. A linear peptide weighs the sum of its free amino acids minus about 18 Da for every peptide bond, the correction the Protein molecular weight calculator applies.
The count, the mass and the stability of a peptide all trace back to one reaction repeated along the chain. Each condensation adds a residue, one peptide bond and one lost water, and the bond it leaves is flat, mostly trans and stable for centuries in water. Counting bonds is counting those reactions, which is why chains, rings and side-chain links each change the answer in a predictable way.
