What Is a Peptide?

A peptide is two or more amino acids linked end to end by peptide bonds. By the convention most biochemists use, a chain stays a peptide up to somewhere between 50 and 100 amino acids; past that, it's usually called a protein. The line is a matter of custom, not a hard rule of chemistry — which is why the same molecule gets called different things in different textbooks.

What makes something a peptide

Every amino acid has an amino group on one end and a carboxyl group on the other. When two amino acids react, the carboxyl group of one and the amino group of the next combine and release a water molecule, forming a covalent amide linkage called a peptide bond. String enough of these bonds together and you get a peptide: a defined sequence of amino acids with a free amino end (the N-terminus) and a free carboxyl end (the C-terminus).

Short peptides get their own names by length — a dipeptide has two amino acids, a tripeptide three, a tetrapeptide four — and chemists sometimes use "oligopeptide" loosely for anything under about 20 residues and "polypeptide" for longer chains. None of that terminology is standardized by a single governing body; it is just how the literature has settled.

Peptide, protein, or something in between?

According to a researcher explainer from the University of Queensland's Institute for Molecular Bioscience, most of the peptides occurring naturally in the human body run about 20 amino acids long, and the conventional cutoff into "protein" territory sits around 50 to 100 amino acids — though the line is explicitly described as fuzzy rather than fixed. The same explainer uses hemoglobin, the oxygen-carrying molecule in red blood cells, as an unambiguous protein: it's built from four separate chains, two of 141 amino acids and two of 146, folded together into one functional unit.

Size isn't the only distinction researchers lean on. Proteins typically fold into a specific, stable three-dimensional shape that does much of the work of their function; many peptides are shorter and floppier, without a single folded structure to rely on. There are exceptions on both sides, which is part of why "peptide" and "protein" overlap rather than sitting in two clean boxes.

A single amino acid, by contrast, is neither — it's the monomer, the individual unit a peptide bond links into a chain. There are 20 amino acids that the genetic code routinely specifies for building the peptides and proteins organisms make; peptide synthesis in a lab can also incorporate amino acids outside that standard set.

Peptides your body is already making

Some of the best-known peptides were worked out before anyone had a word for "peptide drug."

Insulin is 51 amino acids, arranged in two chains held together by sulfur-to-sulfur (disulfide) bonds. Working it out took Frederick Sanger most of the 1950s: he broke the molecule into fragments, separated them, and read the sequence back together, piece by piece — the first time anyone had fully sequenced a protein. The work won him the 1958 Nobel Prize in Chemistry, and it stood as the first conclusive evidence that a protein has one defined, orderly amino-acid sequence rather than some looser arrangement (Nobel Prize, 1958 facts).

Oxytocin, the hormone released by the pituitary gland, is a nonapeptide — nine amino acids, full stop. Vincent du Vigneaud determined that sequence and then, in 1953, became the first person to synthesize a peptide hormone from scratch in a lab, work recognized with the 1955 Nobel Prize in Chemistry (Nobel Prize, 1955 facts; background via Wikipedia).

Glutathione is smaller still: a tripeptide of glutamate, cysteine and glycine. It's an unusual one — the bond between glutamate and cysteine is a "gamma" peptide bond rather than the standard kind, which is part of why the enzymes that normally chew through peptides in cells leave glutathione alone. The sulfur-containing side chain on cysteine is what lets glutathione participate in the cell's redox chemistry, donating electrons to help neutralize reactive byproducts of metabolism (Potęga, Molecules, 2022).

Gut hormones add more examples. Glucagon-like peptide-1 (GLP-1), named for its resemblance to glucagon, is cut from a much longer 160-amino-acid precursor protein in intestinal L-cells, which release it once nutrients from a meal reach them — one active form is written GLP-1(7-36)amide, meaning only residues 7 through 36 of that precursor, 30 amino acids in all, do the work (Willard & Sloop, Experimental Diabetes Research, 2012). Exenatide, a 39-amino-acid synthetic peptide approved for type 2 diabetes, was itself modeled on a peptide first isolated from Gila monster venom (UQ IMB).

How peptides get their names

Once a peptide becomes a drug, its nonproprietary name usually isn't arbitrary. The World Health Organization assigns a common "stem" to drugs that share a pharmacological family, so a reader can tell something about a molecule from its ending. In the WHO's own reference list, "-tide" is the stem reserved for peptides and glycopeptides generally, with narrower substems layered on top for specific families — "-lintide" for amylin-derivative peptides, the "-melanotide" group for melanocortin-receptor agonists, and more besides (WHO, Use of Stems in INN Selection). The ending on a peptide drug's generic name is doing real classification work, in other words, not just supplying a plausible-sounding syllable.

What's known, and what's still being mapped

The peptides described above were identified one at a time, often over years, because each required isolating a specific molecule and determining its exact sequence by hand. Modern mass-spectrometry-based proteomics can now search a single biological sample for thousands of peptides at once, and that has turned up far more of them than classical biochemistry ever worked through individually.

One 2024 study in Proteomes ran this kind of search on urine from 19 healthy volunteers and turned up 29,372 distinct endogenous peptides, representing more than 7,200 precursor proteins, in an initial pass. After the researchers tightened the criteria to peptides detected in at least 10 of the 19 people, that number fell to 1,503 peptides from 436 proteins — and only 69 peptides, from 26 proteins, showed up in every single volunteer (Proteomes, 2024). That gap between the first pass and the consistent core is itself the finding: most of what shows up in one person's peptidome is not reliably present in the next person's, and sorting signal from biological noise in that data is still ongoing work. Nobody has a settled count of "how many peptides are in the human body," and that is likely to stay true for a while — new ones keep turning up as instruments get more sensitive.

What this means for researchers

The word "peptide" is doing less taxonomic work than it looks like it's doing. It names a chemical bond and a rough size range, not a category with sharp edges or a single biological role — a peptide can be a hormone the body makes by the hour, like insulin, or a three-amino-acid molecule like glutathione that is more cofactor than messenger, or a drug built to copy or outlast a natural signal.

That looseness matters when reading anything about a specific peptide, including material published here. A sequence, a molecular weight and a manufacturing lot number are concrete, checkable facts; "peptide" by itself tells a reader almost nothing about what a given molecule does, where it came from, or how well-studied it is. For a researcher evaluating a specific compound, the more useful questions are the next ones: what sequence is actually in the sample, how was it verified, and what has been published about that exact molecule — not just its category. The Journal covers specific compounds at /learn, and the practical side of verifying what's actually in a given sample — not just trusting the label — is covered in our guide to reading a peptide lab report. For a look at the chemistry of the peptide bond itself from the other direction — enzymes engineered to cut it apart at a chosen site in disease-linked proteins — see our earlier piece on designed protein-cutting enzymes.