What Peptides Are: Structure, Bonds and How Chains Are Built

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

A peptide is a chain of amino acid residues held together by amide linkages, each one formed between the alpha-carboxyl carbon of one residue and the alpha-amine nitrogen of the next. Every linkage is a condensation step, so the chain gives up one water molecule, about 18.02 Da, for each residue added beyond the first. The result has direction: a free amine at one end called the N-terminus and a free carboxyl at the other called the C-terminus. Two residues are enough to qualify, and nothing in the chemistry stops at any particular length.

Because the same covalent bond runs through a three-residue tripeptide and through a muscle protein of tens of thousands of residues, the vocabulary layered on top of that bond is convention rather than law. Chemists, structural biologists and regulators each cut the size range a little differently, which is why one molecule can be described as an oligopeptide in a formulation document and as a small protein in a structural paper. This hub starts with the bond, then the size conventions, then the direction and geometry of the chain, and finally how a sequence is read and checked. Two conventions are worth pinning down early: the bond-level test is set out in our page on how a peptide is defined at the bond level, and the fate of the chain under stress is handled in the note on what denaturation does and does not do to peptide bonds.

Four neighbouring hubs take the same subject in other directions: peptides as they appear in topical skincare formulations, the documentation and handling side of research peptides, neutral reviews of peptide vendors and peptides in wellness and supplement marketing. Each of those pages assumes the structural vocabulary set out here.

What Counts as a Peptide

The test for membership is chemical rather than statistical. A molecule belongs to the peptide family when amino acid residues are connected through amide linkages between the alpha-carboxyl group of one residue and the alpha-amine group of the next, which is the same rule applied in our page on the formal definition of a peptide. What changes with length is not the linkage, which stays an amide about 1.32 angstroms long between carbon and nitrogen, but the label used for the chain. A dipeptide has two residues and one such bond, a tripeptide three residues and two bonds, and a chain of n residues carries n minus one peptide bonds; those atoms are taken apart in the page on what a polypeptide chain is chemically made of.

The middle of the range is where the convention is loosest. Oligopeptide is generally used for short chains of roughly 2 to 20 residues, which covers most of what appears on cosmetic ingredient lists, while polypeptide starts somewhere around 20 residues and runs upward without a defined ceiling. Well known molecules sit on both sides of these lines. Glutathione is a tripeptide, but it uses an unusual bond: the glutamate is linked through its side-chain gamma carboxyl rather than its alpha carboxyl, so it is an isopeptide rather than a standard alpha-linked tripeptide. Oxytocin is a nonapeptide of nine residues whose cysteines at positions 1 and 6 form a disulfide that closes a six-residue ring and leaves a three-residue tail. Insulin has 51 residues in two chains, 21 in the A chain and 30 in the B chain, held together by two interchain disulfides plus one intrachain disulfide, and it is described variously as a peptide hormone and as a small protein. More of these borderline cases appear in the comparison of where the dipeptide to polypeptide line sits, and the insulin example raises the related question of whether every protein is also a polypeptide.

Above that, protein carries a functional implication rather than a numerical one: a chain, or an assembly of chains, that folds into a stable structure and does something with it. Roughly 50 residues is a common working threshold, but a folded 40-residue domain is often called a protein while a disordered 200-residue chain sometimes is not. Sequence space expands fast as well: with 20 common residues there are 20 to the power n sequences of length n, about 3.2 million pentapeptides and roughly 1.0 times 10 to the power 13 decapeptides. The repeating element behind every one of these labels is set out in the page on the structural unit shared by peptides and proteins, the top of the size scale in the note on the longest known human polypeptide chain, and the loose use of the fragment in the page on what pepti means inside ingredient names.

The Peptide Bond Itself

Formation of the bond is formally a condensation: the amine nitrogen attacks the carboxyl carbon, and water is eliminated. In water that reaction is thermodynamically uphill, so biology and chemistry both pay for it in different ways. The ribosome uses an activated ester, the aminoacyl-tRNA, and the peptidyl transferase centre of the large ribosomal subunit, which is RNA rather than protein, catalyses the transfer. Solid-phase synthesis instead uses coupling reagents such as HATU, HBTU or a carbodiimide to activate the incoming carboxyl group, then deprotects and repeats. In both cases the product is the same amide.

The amide is not an ordinary single bond. The nitrogen lone pair delocalises into the carbonyl, so the carbon to nitrogen bond measures about 1.32 angstroms, sitting between a typical C-N single bond near 1.47 angstroms and a C=N double bond near 1.27 angstroms, and the group is planar: the carbonyl carbon, oxygen, nitrogen, amide hydrogen and the two flanking alpha carbons all lie in one plane. Rotation around the bond costs roughly 20 kcal per mole, so the six atoms behave as a rigid unit. The practical consequence is that flexibility in the chain comes only from the two bonds to the alpha carbon, described as the phi and psi torsion angles, and not from the peptide bond itself.

That geometry also fixes the arrangement around the bond. The trans form, with the two alpha carbons on opposite sides, is favoured by roughly 1000 to 1 for most residue pairs because the cis form puts adjacent side chains into each other. Proline is the exception: its side chain loops back to the backbone nitrogen, making the imide-like bond, and the energy gap narrows enough that cis-proline can reach somewhere in the range of 5 to 30 percent depending on the local sequence and solvent. Kinetics matter as much as geometry. Uncatalysed hydrolysis at neutral pH and room temperature is extraordinarily slow, with a half-life on the order of hundreds of years, which is why backbones are stable enough to serve as structural material. Forcing hydrolysis needs strong acid, such as 6 M hydrochloric acid at 110 degrees C for 24 hours, strong base, sustained heat, or a protease. The amide also absorbs in the far ultraviolet, a strong transition near 190 nanometres and a weaker one near 210 to 230 nanometres, which is why peptide detection during chromatography is set at 214 or 220 nanometres. More measured values are collected in our list of peptide bond facts, and the consequences of that rigidity are traced in the page on how bond geometry shapes the whole chain.

• Carbon to nitrogen bond length about 1.32 angstroms, between a single bond (about 1.47) and a double bond (about 1.27).
• Rotation barrier around the peptide bond is roughly 20 kcal per mole, which is why the amide group stays planar.
• Only phi and psi, the bonds either side of the alpha carbon, rotate freely; the omega angle of the peptide bond stays near 180 degrees.
• Trans is favoured about 1000 to 1 for most pairs; before proline the cis form can reach roughly 5 to 30 percent.
• Uncatalysed hydrolysis at neutral pH has a half-life measured in hundreds of years; 6 M HCl at 110 degrees C for 24 hours cuts a chain for amino acid analysis.
• The amide absorbs near 190 to 230 nanometres, which is why HPLC detection for peptides commonly uses 214 or 220 nanometres.

Reading Direction: N-terminus to C-terminus

A chain has two distinguishable ends because the ends are chemically different. The N-terminus carries a free amine, or a modified one; the C-terminus carries a free carboxyl, or an amide or ester derived from it. By universal convention a sequence is written from the N-terminus on the left to the C-terminus on the right, and residue numbering follows the same direction. Real chains often carry caps at one or both ends: N-terminal acetylation is found on a large majority of human cytosolic proteins, and C-terminal amidation is common among secreted peptide signals, where it removes a negative charge. Both caps change the mass and the net charge of the molecule, so they belong in any written sequence.

Ribosomal synthesis runs in the same direction as the notation. Translation begins at an AUG start codon, with formylmethionine in bacteria and methionine in eukaryotes, and each new residue is added to the C-terminus of the growing chain, so the N-terminal end is finished first. Bacterial ribosomes incorporate on the order of 15 to 20 residues per second, mammalian ribosomes closer to 5. Each codon is a triplet of nucleotides in the messenger RNA read by an anticodon on a transfer RNA, and the codon table maps 61 sense codons onto 20 residues plus a stop signal, with three codons, UAA, UAG and UGA, specifying termination rather than an amino acid. A codon therefore specifies a single residue, never a chain; that distinction is set out in the page on how codons map onto peptide sequence, and the two kinds of monomer are compared directly in our note on the differences between nucleotides and amino acids.

Chemical synthesis runs the other way. In stepwise solid-phase synthesis the first residue is anchored through its carboxyl group to an insoluble resin, and the chain is extended one residue at a time toward the N-terminus, meaning the C-terminal residue is put in place first. Yield arithmetic explains why length matters so much here: at 99 percent coupling efficiency per step a 30-residue crude product is about 0.99 to the power 29, roughly 75 percent of the theoretical full-length material, while at 95 percent per step the same length falls to about 22 percent, and the remainder is a mixture of deletion sequences that purification has to remove. After cleavage from the resin, reversed-phase HPLC separates the target from those failures, and the purified fractions are frozen and dried. The chemistry and the notation meet at the end of the chain, which is why the page on what sits at the C-terminal end is worth reading alongside this one.

Side Chains and the Twenty Common Residues

Every residue in a chain shares the same backbone segment of nitrogen, alpha carbon and carbonyl carbon, and differs only in the group attached to the alpha carbon. Nineteen of the twenty common residues are chiral at that carbon; glycine, whose side chain is a single hydrogen, is not. Ribosomal translation uses the L configuration, while D residues appear in bacterial cell wall material and in a number of natural peptides made by non-ribosomal pathways. Proline is also structurally unusual because its side chain bonds back to the backbone nitrogen, giving a secondary amine whose geometry constrains the preceding bond and whose own phi angle is locked near minus 65 degrees.

Side chains are usually grouped by chemistry, and the groupings are quantitative rather than decorative. Ionisable groups have characteristic pKa values: aspartate around 3.9, glutamate around 4.1, histidine around 6.0, cysteine around 8.3, tyrosine around 10.1, lysine around 10.5 and arginine around 12.5, with terminal amine and carboxyl groups near 9 and 2 respectively. Net charge is the sum of those contributions at a given pH, and the pH at which the sum crosses zero is the isoelectric point. Hydropathy scales put numbers on the other axis: on the Kyte-Doolittle scale isoleucine scores about 4.5 and valine about 4.2 at the hydrophobic end, while lysine sits near minus 3.9 and arginine near minus 4.5. Aromatic residues matter analytically because tryptophan and tyrosine carry the absorbance at 280 nanometres used to estimate concentration, with molar extinction coefficients of roughly 5500 and 1490 per molar per centimetre respectively.

Sequence order then decides how those properties are presented in space. A helical chain whose hydrophobic residues fall every three or four positions puts them on one face of the helix and leaves the opposite face polar, giving an amphiphilic structure with a hydrophobic side and a hydrophilic side, a pattern common in membrane-active host-defence peptides such as magainin 2 at 23 residues and melittin at 26 residues. The same idea can be checked without a structure by plotting the sequence on a helical wheel, where 3.6 residues per turn means positions i, i plus 3 and i plus 4 land close together in space. That arrangement, and why it matters for solubility and for membrane association, is the subject of our explanation of what amphiphilic means for a peptide.

Folding: Secondary and Tertiary Structure

Structure is described in levels. Primary structure is the covalent sequence, including any disulfides or caps. Secondary structure is the local, repeated hydrogen-bonded arrangement of the backbone. Tertiary structure is the overall three-dimensional fold of one chain, and quaternary structure is the arrangement of several chains. Because the peptide bond is rigid, the fold is determined mainly by the phi and psi angles of each residue, and only a subset of angle combinations is sterically allowed; the allowed regions were mapped out in Ramachandran plots decades ago, and glycine alone can occupy the regions with a positive phi angle that other residues cannot reach.

The alpha helix is the most common repeating element: 3.6 residues per turn, a rise of about 1.5 angstroms per residue, a pitch of 5.4 angstroms, and a hydrogen bond from the carbonyl oxygen of residue i to the amide hydrogen of residue i plus 4, closing a 13-atom loop and giving the helix its formal name, 3.6-13. Helices in globular proteins average roughly 10 to 12 residues, about three turns, and carry a dipole, positive at the N-terminal end and negative at the C-terminal end, which is why anionic groups often bind near the N-terminal end. Proline breaks helices because its backbone nitrogen carries no amide hydrogen, while glycine is often excluded for the opposite reason. Beta structure is nearly the geometric mirror image: a strand is extended at about 3.5 angstroms per residue against 1.5 in a helix, so ten residues span roughly 33 angstroms as a strand and about 15 as a helix, and strands hydrogen bond to neighbours in parallel or antiparallel sheets. Reversals of direction are handled by beta turns spanning four residues, which favour glycine and proline; two rarer helices also occur, the 3-10 helix with three residues per turn and an i to i plus 3 hydrogen bond, and the left-handed polyproline II helix with three residues per turn and about 9 angstroms of pitch.

Tertiary structure is held together by interactions that are individually weak and collectively decisive: hydrophobic burial, hydrogen bonds, salt bridges, aromatic stacking, and sometimes metal coordination or covalent crosslinks. Disulfides are the covalent exception, a sulfur to sulfur bond about 2.05 angstroms long formed by two oxidised cysteines and broken again by reducing agents. Denaturation disrupts the noncovalent set: heat, 6 to 8 M urea, 6 M guanidinium chloride, detergents such as SDS, pH extremes and organic solvents all unfold chains, and many globular proteins lose their fold between 40 and 80 degrees C while proteins from thermophilic organisms stay folded above 90 degrees C. None of these treatments hydrolyses peptide bonds, so a denatured chain keeps the same sequence and mass, which is why re-folding is sometimes possible. Enzymatic cleavage is a separate matter and is highly specific: trypsin cuts on the C-terminal side of lysine and arginine except before proline, chymotrypsin after phenylalanine, tyrosine and tryptophan, and cyanogen bromide at methionine. The parameters above are collected in the page on peptide secondary structure, the packing of a complete fold is described in the note on tertiary structure and the bonds behind it, and the unfolding-versus-hydrolysis distinction is set out in the page on denaturation and peptide bonds.

• Alpha helix: 3.6 residues per turn, 1.5 angstroms rise per residue, 5.4 angstroms pitch, hydrogen bond from i to i plus 4.
• Beta strand: about 3.5 angstroms per residue extended, so 10 residues span roughly 33 angstroms versus about 15 in a helix.
• Beta turn: four residues reverse the chain direction; glycine at position 2 and proline at position 3 are statistically favoured.
• Disulfide: a covalent sulfur to sulfur bond of about 2.05 angstroms between two cysteines, broken by reduction rather than by heat alone.
• Denaturants: 6 to 8 M urea, 6 M guanidinium chloride, SDS, pH extremes and heat unfold chains without cutting the backbone.
• Protease specificity: trypsin after Lys and Arg unless followed by Pro, chymotrypsin after Phe, Tyr and Trp, cyanogen bromide at Met.

Reading a Sequence and a Specification Sheet

Sequences are written as one-letter codes read from N-terminus to C-terminus, with modifications shown as prefixes and suffixes. Acetylation of the N-terminus adds about 42.01 Da and removes a positive charge; amidation of the C-terminus replaces a hydroxyl with an amine, changing the mass by about minus 0.98 Da and removing a negative charge; a disulfide costs about 2.02 Da relative to two free thiols; pyroglutamate forms by cyclisation of an N-terminal glutamine and loses about 17.03 Da. Sequences that include non-standard or D residues, lipid tails, chelators or polymer conjugates usually state them explicitly, because none of them are visible in the one-letter string alone. Short functional motifs also appear inside longer sequences, for example the basic cluster recognised by nuclear import machinery, described further in our page on the nuclear localization signal motif.

Mass follows from the string. Add the residue masses, which are the free amino acid masses minus 18.02 Da of water, then add one water back for the intact chain with free termini, then apply any modification offsets. The monoisotopic residue masses of glycine, histidine and lysine are 57.02, 137.06 and 128.09 Da, so the tripeptide GHK comes to about 340.19 Da monoisotopic, a figure that any mass spectrometer should reproduce to within a few parts per million. Average masses, which use the natural isotope distribution rather than the lightest isotope, run roughly 0.05 to 0.1 percent higher and are the numbers usually quoted on product documentation. For rough work, about 110 Da per residue is a serviceable average, so a 30-residue peptide sits near 3.3 kDa. Chains are also described by the fragments they produce under enzymatic cleavage, which is what the term in our note on the tryptic peptide definition refers to.

Two orthogonal measurements make up the standard analytical pair. Reversed-phase HPLC, typically a C18 column with an acetonitrile gradient in 0.1 percent trifluoroacetic acid and detection at 214 to 220 nanometres, gives a purity figure expressed as the percentage of peak area at a stated wavelength; it says nothing about identity, because a co-eluting impurity is invisible and neither water nor counterion absorbs. Mass spectrometry supplies identity by comparing an observed ion, usually the protonated or multiply charged molecule, against the theoretical mass. Amino acid analysis after acid hydrolysis gives composition and estimates net peptide content, the fraction of material in a vial that is peptide once water and counterion are accounted for. Cold-chain conventions follow from the same chemistry: lyophilised powder is commonly stored near minus 20 degrees C, solutions made from it at about 2 to 8 degrees C, and repeated freeze-thaw cycles avoided, as laboratory practice rather than a stability guarantee. Those conventions are set out in the guides on storage practice for lyophilised peptides and how long a reconstituted vial keeps in the fridge, the measurement side in peptide purity testing methods and third-party peptide testing, the drying step in the note on the freeze-drying process, and the applications that make this documentation necessary in what peptides are used for.

Where Peptides Meet the Rest of the Site

Structure explains why the same chemistry shows up in such different places. A short chain is small enough to sit on the surface of a formulation and large enough to carry recognisable information in its side chains, which is why topical skincare has adopted fragments of matrix proteins and copper-complexed tripeptides as ingredient stories. That side of the subject is covered in the skin peptides hub, where the emphasis is on what an ingredient list tells you about sequence, concentration and evidence rather than on the chemistry set out here; the copper peptide pages there, including the discussion of copper tripeptide complexes in topical skincare, start from the coordination chemistry described above.

The documentation side is the other large branch. Once a chain is bought, dissolved, aliquoted, frozen, dried and tested, the questions become arithmetic and record-keeping: vial mass against diluent volume to give a concentration in mg per mL, unit conversions between mass and molar amount, chromatogram interpretation, and what a lot number does and does not cover. Those are the subjects of the practical guides hub, which includes reconstitution arithmetic worked through alongside the storage and testing material linked earlier. Safety questions, including regulatory status, sterility, endotoxin limits and counterfeit risk, are handled as regulatory and laboratory topics rather than as instructions, as in the page on regulatory status and sterility questions around injectable peptides; anything touching on human use belongs with a licensed clinician and a national regulator.

Two further hubs complete the map. Vendor documentation is reviewed on a neutral, evidence-of-quality basis in the peptide vendor reviews hub, which looks at what can be checked publicly rather than at marketing claims; no vendor is endorsed there. The wellness side is treated in the peptides and wellness hub, where hydrolyzed collagen fragments are discussed as composition rather than as outcomes, and where the physiological roles peptides are studied for are kept separate from product marketing.