Every compound in this library that is a peptide is a peptide because of this one link, repeated. It is worth understanding directly, because with one number you can check a published molecular weight for yourself — and find out when a sequence is not telling you everything.
What the bond is
Every amino acid carries an acid group (–COOH) at one end and an amino group (–NH₂) at the other. A peptide bond forms when the acid group of one meets the amino group of the next: they join, and a molecule of water leaves.
H₂N–CHR¹–COOH + H₂N–CHR²–COOH
↓ (condensation, −H₂O)
H₂N–CHR¹–CO–NH–CHR²–COOH
That is the whole mechanism. Do it repeatedly and the result is a chain: a peptide if it is short, a protein if it is long. The words describe length and role, not different chemistry — the backbone is the same bond either way.
Two consequences follow immediately, and both are useful.
A chain of n residues has n − 1 bonds. Fifteen residues, fourteen joints. The same arithmetic as fence posts and panels.
Each bond costs one water — 18.015 daltons. So a finished peptide weighs measurably less than its ingredients did.
The arithmetic, and how to check it
Those two facts give a formula that needs nothing but a table of amino acid weights:
Molecular weight = (sum of the free amino acid weights) − (n − 1) × 18.015
This library's amino acid chart carries the weights. Here is the formula run against three compounds that have their own entries, each checked against the figure PubChem publishes independently.
| Compound | Residues | Bonds | Free amino acids | − waters | Calculated | Published | Difference |
|---|---|---|---|---|---|---|---|
| BPC-157 | 15 | 14 | 1,671.75 | −252.21 | 1,419.54 | 1,419.5 | −0.04 |
| KPV | 3 | 2 | 378.47 | −36.03 | 342.44 | 342.43 | −0.01 |
| GHK | 3 | 2 | 376.41 | −36.03 | 340.38 | 340.38 | 0.00 |
(Published weights: PubChem CIDs 9941957, 125672 and 73587, read 2026-09-18. Differences are rounding in the published amino acid weights.)
The largest of the three is fifteen residues long and agrees to within four hundredths of a dalton. That is the point of doing it: the rule is not approximate.
When the arithmetic fails, it tells you something
Run the same calculation on TB-500, whose sequence is written LKKTETQ:
Free amino acids L+K+K+T+E+T+Q = 955.07
Six bonds, six waters lost − 108.09
────────
Calculated = 846.98
PubChem CID 62707662 = 889.00
────────
Unexplained = +42.02
Forty-two daltons is not rounding. 42.01 is the mass of an acetyl group (C₂H₂O added to the molecule), and TB-500 carries one on the front of its chain. The seven letters of the sequence do not record it.
This is the practical value of being able to do the sum. Gaps of this kind are recognisable:
| Gap | What it usually is |
|---|---|
| +14.03 | a methyl group |
| +42.01 | an acetyl group |
| +79.98 | a phosphate |
| −2.02 | a disulfide bridge (two hydrogens lost) |
| +18.02 | an amidated C-terminus, relative to the free acid |
A sequence is not a complete description of a molecule. Two products can carry the same letters and be different substances. Checking the mass is the cheapest way to notice.
Why the bond is flat, and why that matters
A single bond between two atoms normally allows them to spin freely. The carbon-to-nitrogen link in a peptide bond does not, because its electrons are shared across the whole group rather than staying put — it behaves as though it were partly a double bond. The result is that six atoms around each peptide bond lie in one plane, and the unit is rigid.
Linus Pauling, Robert Corey and Herman Branson set this out in the Proceedings of the National Academy of Sciences in April 1951 (volume 37, pages 205–211), in the paper that proposed the alpha helix. Their argument only worked because the peptide unit is flat: once the chain can bend only at the two single bonds on either side of each rigid plate, the number of shapes it can take collapses from unimaginable to enumerable.
That is why proteins have structures at all. A chain that could rotate anywhere would be a tangle; one built from flat, rigid segments with hinges between them folds reproducibly into the same shape every time.
The bond and the body
The same rigidity that makes the bond useful makes it durable. Peptide bonds do not fall apart on their own at any useful rate — breaking one requires water to be put back, and in practice an enzyme to do it.
The digestive system is full of exactly those enzymes, which is the single most important practical consequence in this whole library: a peptide taken by mouth meets machinery built specifically to cut peptide bonds. It is why nearly every peptide drug is injected, and why the exceptions need either an absorption enhancer or a chemistry that is not peptide at all. Orforglipron is the clearest example of the second route — a tablet that acts on a peptide hormone's receptor without containing a single peptide bond.
For what the words peptide and protein actually divide, see what peptides are; for the twenty building blocks and their weights, see the amino acid chart.
How much has been written about it
PubMed held 27,539 records naming the peptide bond in title or abstract on 2026-09-18 — a reminder that this is settled, century-old chemistry rather than a contested area. Nothing on this page is new. What is worth having is the ability to check a number yourself, which is the difference between reading a molecular weight and knowing it.
Related entries
- The amino acid chart — the twenty residues, with the weights this page uses
- What peptides are — where peptide ends and protein begins
- BPC-157, KPV, TB-500 — the three worked examples