Peptide Structure, Classification & Scientific Terminology

By What Peptides Editorial Team · Updated 2026-09-14 · Pillar guide

A peptide is a chain of amino acid residues joined by peptide bonds, the amide linkage formed between the carboxyl group of one residue and the amino group of the next. Most working definitions also set a size limit: a chain of roughly two to fifty residues is commonly called a peptide, longer chains are called polypeptides or proteins, and the boundary is a convention rather than a law of nature. What actually unifies the group is the chemistry of the bond itself, because every residue in every peptide, from a dipeptide to a chain of 34,350 residues, is connected by the same amide link with the same geometry.

Two numbers make peptide chemistry practical. The average residue mass is about 110 Da, so a chain's molecular weight is close to 110 times the residue count minus the 18.01056 Da of water that each bond formation has already given up. Direction matters just as much: sequences are written from the free amino terminus to the free carboxyl terminus, and that convention governs how every database record, product label and specification sheet is read. This page sets out the vocabulary used across the rest of the guide, from bond geometry through classification to quaternary assembly, and it links to the individual term pages that sit beneath it, including alternative names for peptides and the pages on specific molecules.

Terminology in this field drifts. The same word can denote a chain length, a function, a receptor family or a marketing claim, and many search queries arrive truncated or misspelled. Where a term cannot be resolved to a structure, the honest answer is to say so rather than to guess, which is the approach taken for the imida peptide term and similar cases elsewhere in this guide. The sections below separate the structural facts, which are measurable, from the naming conventions, which are agreements between people, and they close with a practical walk through the documents used to describe a research material.

What Counts as a Peptide: Bonds and Residue Counts

A peptide is defined by its bond before it is defined by its size. The peptide bond is an amide formed when the carboxyl carbon of one amino acid reacts with the amino nitrogen of the next, releasing one molecule of water, 18.01056 Da, per linkage formed. Any chain held together this way is chemically a peptide, whether it contains two residues or two thousand, and the same bond recurs in every protein that has been sequenced. Size enters only as a naming convention, and the conventions differ between textbooks, databases and regulatory texts. One widely used cut places peptides at about fifty residues or less and reserves protein for longer chains, but that boundary is administrative rather than chemical, and plenty of molecules sit awkwardly on it.

Residue count supplies the working vocabulary. Two residues make a dipeptide, three a tripeptide, and short chains up to about twenty are often grouped as oligopeptides. Beyond that, polypeptide describes a longer single chain, and protein usually implies a chain long enough to fold into a stable tertiary structure. Real examples show why the borders are soft. Insulin has 51 residues held in two chains by disulfides, hepcidin has 25 residues with four disulfide bonds and is described in the hepcidin entry, and atrial natriuretic peptide has 28 residues with a 17-residue ring closed by a single disulfide, set out in the ANP page. Serum albumin has 585 residues, lysozyme 129, and titin about 34,350 in UniProt entry Q8WZ42.

Two cautions keep the vocabulary honest. First, function and size are independent axes: a short chain can act as a hormone, and a very long one can be a structural repeat with no signalling role at all. Second, not everything labelled a peptide is one. Bromantane is an adamantane derivative rather than an amino acid chain, and the bromantane page explains why the word peptide in that query is a misnomer. Where a term cannot be tied to a structure, saying so is the correct answer, and that is the position taken on similarly unresolvable names elsewhere in this guide. Both of those examples also show how a disulfide bridge changes the arithmetic: forming each bridge removes two hydrogen atoms, so an oxidised chain is slightly lighter than the same chain in its reduced form, and the difference has to be included before a measured mass will match a calculated one.

The Peptide Bond: Length, Planarity and Rotation

The peptide bond is shorter than a single bond and longer than a double bond, and that one measurement explains most of its behaviour. The C-N distance is about 1.32 A, against roughly 1.47 A for a C-N single bond and about 1.27 A for a C=N double bond. The shortening comes from resonance: the nitrogen lone pair delocalises into the carbonyl, giving the C-N link partial double-bond character, holding the carbonyl carbon, the oxygen, the nitrogen and its hydrogen in one plane, and blocking free rotation around the bond. Six atoms of the backbone therefore behave as a single rigid unit, and the flexibility of the chain has to come from the two bonds that flank it rather than from the peptide bond itself.

Planarity also fixes the relative position of the substituents. The trans arrangement, with the alpha carbons on opposite sides of the bond, is favoured by roughly 1000 to 1 over cis because it keeps the bulky side chains apart. Proline is the exception: its nitrogen sits inside a five-membered ring that already carries a substituent, so the energy gap narrows and the cis form becomes common enough to matter in structure determination. The torsion angle across the bond is called omega and sits near 180 degrees in the trans form and near 0 degrees in the cis form. Rotation is available at phi, about the N-C alpha bond, and psi, about the C alpha-C bond, giving two degrees of freedom per residue.

Those two angles are the reason a sequence alone does not fix a shape. Every residue except glycine and proline occupies a limited region of the phi-psi map, and steric clashes exclude most combinations, which is what makes secondary structure predictable at all. Cleavage of this bond carries its own vocabulary. A scissile bond is simply one that a given protease can cut, and the scissile bond page sets out the terms used for the bond, for the residues on either side of it, and for the numbering of the fragments produced. Hydrolysis returns the water that condensation removed and is the reaction every protease catalyses. Because the two states of a proline peptide bond interconvert slowly on the timescales that matter for folding, that isomerisation is often the slow step in a refolding experiment, and the enzymes that accelerate it are described as peptidyl-prolyl isomerases.

Direction, Residue Mass and Molecular Weight

Sequences are written from the free amino terminus to the free carboxyl terminus, and every database entry, catalogue listing and figure follows that direction unless it explicitly says otherwise. The convention is not cosmetic. Residue numbering starts at the N-terminus, terminal modifications are named by the end they occupy, and solid-phase synthesis runs in the opposite direction, from C to N, on a resin. Reading a sequence also means knowing which alphabet is in use: the one-letter code is compact and standard in databases, the three-letter code is clearer in structural work, and both are read left to right as N to C. Two terminal modifications are routinely left out of a written sequence and both shift the mass: N-terminal acetylation adds 42.01056 Da, and C-terminal amidation replaces a hydroxyl with an amine for a net change of minus 0.98402 Da. A measured mass that disagrees with a plain sequence calculation usually means a terminus carries one of them.

Mass follows from the same arithmetic. Each residue contributes its own mass, not the mass of the free amino acid, because every bond formation has already given up water. The average residue mass is about 110 Da, enough for a quick estimate: a chain of n residues weighs roughly 110n minus 18.01056 times the number of bonds, and the number of bonds is n minus 1. For a ten-residue chain that estimate gives about 1100 minus 162 Da, close to 938 Da. Exact work uses monoisotopic residue masses instead: glycine 57.02146, alanine 71.03711, cysteine 103.00919, carbamidomethylated cysteine 160.03065, and tryptophan 186.07931. One distinction must not be blurred in the middle of a calculation: average mass averages over the natural isotope abundances, whereas monoisotopic mass uses the lightest isotope of each element. The gap between the two widens with chain length, so the two conventions can never be mixed inside a single calculation.

Two related conventions often confuse readers. The genetic code is read in triplets, and the peptide codons page sets out the count behind it: 64 codons, 61 sense, 3 stops, with AUG as the usual start. Codons specify residues rather than bonds, so a codon count and the residue count of a mature chain rarely match once signal sequences and post-translational processing are taken into account. Glutamine appears frequently in both contexts and is easily confused with the free amino acid sold as a supplement, a distinction that the glutamina peptide page separates cleanly from the residue itself. Notation carries meaning as well: a lowercase letter in a written sequence usually marks a D-residue, and an X marks a position that could not be assigned, both of which are common in sequences determined from natural material rather than from a gene.

Classification by Shape: Linear, Cyclic and Branched

Shape classification describes topology, the way the chain is connected, independent of anything the molecule does. A linear peptide has one free N-terminus, one free C-terminus and a single path through the backbone. Cyclisation removes one or both termini from the picture by closing the chain, and it is among the most common modifications in both natural products and designed molecules, because a closed backbone constrains the phi and psi angles and lowers the entropic cost of adopting one conformation. Branching attaches a second chain to a side chain rather than to a terminus, producing a molecule with more than one path through its covalent structure. Topology also decides how many chains count as one molecule: insulin is usually described as a single peptide because its two chains are covalently linked by disulfides, whereas two chains held together only by non-covalent contacts are described as an assembly.

Cyclisation takes several forms. Head-to-tail closure joins the N-terminus to the C-terminus through a normal amide bond. Side-chain closure joins two side chains, most often by oxidising two cysteines to cystine, as in the 17-residue ring of atrial natriuretic peptide or the four disulfides of hepcidin. Side-chain-to-terminus closure joins a group such as the epsilon amino group of lysine, or the carboxyl of aspartate or glutamate, to one end of the chain. Stapling is a synthetic variant in which a hydrocarbon brace links two residues on the same face of a helix. Each form changes the molecular formula, which is why a stated formula is worth checking against the drawn structure. Ring size is normally reported as the number of residues enclosed by the loop, and the conformational constraint is most severe in rings of about five to ten residues, where few backbone arrangements are sterically available at all.

Branched and conjugated forms extend the same logic. Lysine, with its epsilon amino group, is the usual branch point, and multiple antigenic peptide constructs use a lysine core to present several copies of one sequence. Lipidation attaches a fatty acid to a side chain or a terminus; semaglutide is described in the literature as a GLP-1 analogue carrying an 18-carbon diacid side chain, and that appendage changes how the molecule behaves rather than the sequence itself. Mixtures complicate matters further, because a hydrolysate is defined by a distribution of fragments rather than by one sequence, which is the point made in the hydrolyzed vegetable protein entry. PEGylation follows the same pattern with a different appendage: the polymer is described either by the number of repeat units or by an average molecular weight, and that average is a distribution rather than a single value, which matters as soon as a mass spectrum is interpreted.

• Linear chain: one N-terminus, one C-terminus, one continuous path through the backbone.
• Head-to-tail cyclic: the two termini joined by an amide bond, leaving no free terminus.
• Disulfide-bridged cyclic: two cysteine side chains oxidised to cystine, as seen in ANP and hepcidin.
• Side-chain-to-terminus cyclic: a lysine, aspartate or glutamate side chain joined to one end of the chain.
• Branched: a second chain grown from a lysine epsilon amino group, giving two or more paths.
• Lipidated or PEGylated: a fatty acid or polymer attached to a side chain or terminus, changing mass and handling.
• Mixture: a hydrolysate such as HVP, defined by a fragment distribution rather than a single sequence.

Classification by Function: Hormones, Signals and Antimicrobials

Functional classification groups peptides by the role they are observed to play rather than by their covalent structure, so one molecule can sit in more than one category. Hormonal peptides are secreted and act through a receptor at a distance. Neuropeptides are released by neurons and act on neighbouring cells or across short distances. Antimicrobial peptides are typically cationic and amphipathic and are discussed in the literature in terms of membrane interaction. Signal peptides are N-terminal sorting sequences that route a nascent chain to a compartment and are usually removed once they have served that purpose. These labels are historical rather than logical. Many were assigned when a molecule was first extracted from a tissue, long before its sequence was known, and several survived the later discovery that the molecule is a fragment of a larger precursor. A class name is best treated as a starting point for searching the literature, not as a statement of mechanism.

Signal sequences are the clearest case of function residing in a region rather than in a whole molecule. A nuclear localisation signal is a short, basic stretch that marks a protein for import into the nucleus, and the nuclear localisation signal page describes how such motifs are recognised and how they are written in a sequence record. The natriuretic family shows the hormonal pattern: short circulating chains sharing a conserved ring, compared across members in the natriuretic peptide page. The family is defined by sequence similarity and by that shared ring rather than by a fixed residue count, so two members can differ by a handful of residues and still carry different prefixes. Both entries describe structure and nomenclature only, and neither should be read as a statement about any outcome.

Cosmetic labelling uses the language differently again. In an ingredient list a peptide appears under an INCI name, the position of that name reflects an ordering rule rather than a stated quantity in the way many readers assume, and a product name containing the word peptide may refer to a single sequence, a blend, or a purely promotional construction. The exlinea pro peptide serum page works through what can and cannot be read from such a label, and the companion page on before and after claims states plainly that this site publishes no photography of that kind and makes no efficacy claim. The checks that a label can actually support are narrow but real: does the INCI name correspond to a defined sequence, does the manufacturer state a concentration anywhere, and are the packaging and fragrance status declared. Those are questions about documentation, and only those are answered here.

From Primary Sequence to Quaternary Assembly

Four levels describe how a chain is organised. Primary structure is the residue sequence itself, written N to C, and it is the only level fully specified by a string of letters. Secondary structure is the local, repeating arrangement of the backbone held by hydrogen bonds: the alpha helix runs 3.6 residues per turn with a rise of 1.5 A per residue and a pitch of 5.4 A, stabilised by a hydrogen bond from residue i to residue i plus 4, while a beta strand in a sheet extends about 3.5 A per residue. Tertiary structure is the overall fold of one chain, and quaternary structure is the arrangement of several folded chains in one complex.

The word quaternary does not apply to a short peptide on its own, because there is only one chain to arrange. It enters the discussion when several chains assemble, when a peptide binds a larger protein, or when identical subunits oligomerize. Hemoglobin is the standard teaching example of a tetramer, four chains each carrying a heme group, and the assembly behaves differently from any single chain taken alone. The quaternary structure page and the tetrameric peptide page both unpack this vocabulary, because the terms are used loosely in descriptions of materials where no such assembly has been shown to exist. Oligomeric state is measured rather than assumed. Size-exclusion chromatography, analytical ultracentrifugation and native mass spectrometry each report on whether a sample holds one chain or several, and they often contradict a name that was assigned by analogy to a better studied molecule.

Laboratory courses usually introduce the same hierarchy from the bottom up, starting with detection of the bond, then hydrolysis, then separation of the residues released. That sequence of experiments is the framework behind the peptides and proteins lab page, which treats the number as a course label rather than as part of a structure. Folding is also why storage conditions appear on a specification sheet: a freeze-thaw cycle can unfold or aggregate a chain without breaking a single covalent bond, which is why lyophilised powder is commonly held near -20 C and reconstituted material near 2 to 8 C. The same caution applies to a purity figure. Reversed-phase HPLC reports how much of the sample is the intended sequence; it says nothing about whether the chains are folded, so a material can be highly pure by that measure and still contain a fraction that has aggregated or adopted the wrong conformation.

Reading a Sequence, a Label and a Specification Sheet

A specification sheet, often called a certificate of analysis, is the document that turns a name into a defined material. The fields worth reading in order are the sequence with its termini stated, the molecular formula, the calculated and observed molecular weight, the salt form and counter-ion, the purity figure together with the method used to obtain it, the identity method, the residual moisture and solvent figures, and the storage and handling statements. Purity without a method is not a number. Reversed-phase HPLC purity is reported with a gradient and a detection wavelength, and identity is usually established by LC-MS or MALDI-TOF mass measurement. The salt form deserves particular attention, because a peptide supplied as a trifluoroacetate salt weighs more than the free chain, and in a short sequence the counter-ion can account for a noticeable share of the mass of the material in the vial.

Three abbreviations carry most of the interpretive load. RUO means research use only, and a material marked that way is not intended for human or animal use of any kind; the research use only page explains what that marking does and does not tell a reader. Tryptic refers to cleavage by trypsin, which cuts C-terminal to lysine and arginine except before proline, and the tryptic peptide definition shows how fragment masses are predicted from that rule. Endotoxin figures are reported in endotoxin units from an LAL assay, sterility is tested to a named pharmacopoeial method, and residual water is usually measured by Karl Fischer titration. Lot numbers matter for the same reason. A certificate is issued for one lot, so any figure printed without a lot number cannot be traced to the material in hand, and a document without a date, a method and a lot identifier is a product description rather than a specification.

From here the guide splits by subject. Structural and definitional terms stay on this page's branch. The cosmetic peptide guide covers labelled ingredients and the vocabulary of product claims, the research and industry guide covers analytical testing and regulatory reporting, the vendor review guide explains how supplier documentation can be assessed without endorsing anyone, and the wellness guide covers collagen hydrolysate and supplement labelling. Each of those pages lists its own child articles, and none of them publishes dosing, medical advice or a recommendation of any kind. This page likewise states no amount for any person: every number above is a bond length, a mass, a residue count or a temperature, reported because those are the quantities a reader needs in order to check a sequence, a formula or a document against the structure it claims to describe.

• Sequence with N- and C-terminal modifications named explicitly, not implied only by a structure drawing.
• Molecular weight given as calculated from the formula and observed from the mass spectrum, with the salt form stated.
• Purity reported by RP-HPLC with the gradient and wavelength, since a bare percentage is not comparable between laboratories.
• Identity by LC-MS or MALDI-TOF, with the observed mass matching the calculated mass inside a stated tolerance.
• Residual moisture by Karl Fischer, endotoxin by LAL, sterility to a named pharmacopoeial method.
• Storage stated separately for lyophilised powder, commonly near -20 C, and for reconstituted material, commonly 2 to 8 C.