All twenty amino acids share the same core: a central alpha carbon bonded to a carboxyl group, an amino group, a hydrogen atom and a variable side chain. A peptide bond joins the carboxyl group of one amino acid to the amino group of the next. Side chains differ from residue to residue, but this link is the same along every peptide and protein chain, so its properties are shared by all of them.
Three of those properties matter most, and the rest of this page follows them. The bond is flat and rigid, which limits how a chain can fold. It is polar, which lets neighbouring bonds hold each other in helices and sheets. And it is extremely slow to break in water, yet easy for enzymes to cut.
| Bond type | Secondary amide, –C(O)–NH– |
|---|---|
| C–N length | About 1.32–1.33 Å, against 1.47 Å for an ordinary C–N single bond |
| Double-bond character | About 40%, from resonance with the carbonyl group |
| Usual form | Trans. In proteins about 5% of bonds before proline are cis, and only 0.03% of the rest |
| Half-life in water, pH 4–8 | About 350 to 600 years at room temperature without a catalyst |
How a peptide bond forms
A peptide bond forms when the carboxyl group of one amino acid reacts with the amino group of another. The product is an amide, and one molecule of water is released, which is why biologists call the reaction a dehydration synthesis.
H2N–CHR1–COOH + H2N–CHR2–COOH → H2N–CHR1–CO–NH–CHR2–COOH + H2O
The loss of water explains two practical facts. A peptide's mass is the sum of its residue masses, each an amino acid minus water, plus one water for the two ends. And a linear chain of n residues contains n − 1 peptide bonds. The chain also has a direction: its amino end is called the N-terminus and its carboxyl end the C-terminus, even when these groups are chemically modified, and sequences are always written from N to C.
Mixing two amino acids in water does not make a peptide. In water the equilibrium lies on the side of hydrolysis, and at neutral pH the two groups exist as an ammonium ion and a carboxylate, neither of which reacts in the right way. The main biological and laboratory routes therefore start by activating the carboxyl group. Ribosomes use aminoacyl-tRNAs, in which the amino acid is held as a reactive ester. Bacteria and fungi also make peptides without ribosomes, on large enzymes called nonribosomal peptide synthetases; this is how cyclosporin and antibiotics such as vancomycin and gramicidin S are built, often from unusual and D-amino acids. In the laboratory, chemists activate the carboxyl group with a coupling reagent and protect every other reactive group, as described under solid-phase synthesis.
Resonance and planarity
The first property, flatness, comes from the way electrons are shared. Although it is drawn as a single bond, the peptide bond behaves partly like a double bond. The lone pair of electrons on the nitrogen is shared with the neighbouring carbonyl group, so the real structure lies between two resonance forms.
Linus Pauling, Robert Corey and Herman Branson used this in 1951 to work out how protein chains fold. The C–N distance they measured in small amides, 1.32 Å, is much shorter than an ordinary single bond and corresponds, in Pauling's words, to "nearly 50 per cent double-bond character". They concluded that the amide group must be flat, and that twisting it by as little as 10 degrees out of the plane would already cost about 1 kcal/mol[1]. With the amide plane as a fixed constraint, they used it in that paper to predict the α-helix. Later estimates put the double-bond character a little lower, at about 40%[2].
Six atoms lie in the plane: the α-carbon of the first residue, the carbonyl carbon and oxygen, the amide nitrogen and its hydrogen, and the α-carbon of the next residue. Rotation about the C–N bond is blocked, so the backbone can only twist at the bonds on either side of each α-carbon.
The same resonance gives the second property: it makes the group polar. The carbonyl oxygen carries a partial negative charge and is a good hydrogen-bond acceptor; the N–H carries a partial positive charge and is a good donor. Hydrogen bonds between these groups hold together every α-helix and β-sheet. Because the nitrogen's lone pair is tied up in resonance, the amide nitrogen is not basic, and backbone amides do not add to a peptide's charge when its isoelectric point is calculated.
Cis and trans
A flat bond can have two arrangements. In the trans form the two α-carbons sit on opposite sides of the C–N bond; in the cis form they sit on the same side, where they crowd each other.
Almost all peptide bonds are trans. A survey of 571 protein structures found only 0.03% of ordinary peptide bonds in the cis form[3]. The exception is the bond in front of proline. Proline's side chain is joined back to its own nitrogen, so that nitrogen carries a carbon on both sides and the cis and trans forms are much closer in energy. In that survey 5.2% of these bonds were cis, and in short model peptides with uncharged ends, NMR measurements put the cis share before proline at about 10–15%[3].
The two forms interconvert slowly, and this shows up in the laboratory. NMR spectra of proline-containing peptides often show two sets of peaks for one compound, and HPLC peaks can broaden or split for the same reason; neither is a sign of impurity. In cells, slow proline isomerisation can hold up protein folding, and enzymes called peptidyl-prolyl isomerases speed it up.
Backbone flexibility: the angles φ and ψ
Since the peptide bond itself is rigid, most of the flexibility of the backbone comes from rotation about the two bonds on either side of each α-carbon. Rotation about N–Cα is described by the angle φ (phi) and rotation about Cα–C by ψ (psi), so the shape of the backbone at each residue comes down to one pair of numbers.
In 1963 G. N. Ramachandran, C. Ramakrishnan and V. Sasisekharan worked out which pairs are possible by checking where atoms of neighbouring residues would collide. Most combinations turned out to be forbidden, and the allowed ones fall into a few regions of what is now called the Ramachandran plot[4]. Familiar structures sit in these regions:
| Right-handed α-helix | φ about −60°, ψ about −45° |
|---|---|
| Antiparallel β-sheet | φ about −140°, ψ about +135° |
| Parallel β-sheet | φ about −120°, ψ about +115° |
| Polyproline II helix | φ about −75°, ψ about +145° |
Two residues break the rules. Glycine, with only a hydrogen for a side chain, can reach regions closed to all others and is common in tight turns. Proline's ring fixes its φ angle near −65° and it has no N–H to donate a hydrogen bond, so it often bends or interrupts the regular hydrogen-bonding pattern of a helix or sheet, and is most at home at their ends and in turns.
How long a peptide bond lasts
The third property is stability. Breaking a peptide bond with water releases energy, so in thermodynamic terms every peptide in water is unstable. In practice the bond is extraordinarily slow to break. Anna Radzicka and Richard Wolfenden heated simple peptides in sealed quartz tubes and extrapolated the rates down to room temperature. In neutral water at 25 °C, they found half-lives of about 350 years for glycylglycine and about 600 years for a peptide bond inside a chain[5]. Acid or base speeds the reaction considerably: in that study, glycylglycine at 25 °C lasts about 150 days in 1 M hydrochloric acid and about 2 days in 1 M sodium hydroxide.
For living things this is a useful combination. Proteins last as long as they are needed, yet an enzyme can cut a chosen bond in a fraction of a second. Proteases do this routinely, and industry relies on them too: about 60% of all industrial enzymes are proteases, used in food, leather, textile and paper processing[2]. One of them, subtilisin, is added to laundry detergents to digest protein stains. Removing water slows hydrolysis further, although how long a dry peptide keeps also depends on its sequence and on other reactions such as oxidation; see stability and storage.
Cutting peptide bonds on purpose
Because the bond is so stable on its own, breaking it on purpose takes either an enzyme or harsh conditions. Both are basic tools of peptide analysis, and they differ in how selective they are:
| Trypsin | After Lys or Arg, but not before Pro. The standard enzyme in proteomics |
|---|---|
| Chymotrypsin | After Phe, Tyr and Trp (more slowly after Leu and Met) |
| Glu-C | After Glu (and after Asp in phosphate buffer) |
| Cyanogen bromide | After Met. Splits large proteins into a few pieces |
| 6 M HCl, 110 °C, about 24 h | Every peptide bond. Used for amino acid analysis |
Complete acid hydrolysis is not perfectly clean. It destroys tryptophan, partly degrades serine and threonine, and turns asparagine and glutamine into aspartic and glutamic acid. That is why amino acid analysis reports Asx and Glx instead of the separate residues, and why tryptophan has to be measured another way. Enzyme digests are the usual first step before mass spectrometry.
Isopeptide bonds
Not every amide bond in a peptide runs along the backbone. An amide bond that involves a side-chain group instead of the main-chain amino or carboxyl group is called an isopeptide bond. Several important molecules depend on one:
- Glutathione (γ-Glu-Cys-Gly) links glutamate through its side-chain carboxyl group. Most peptidases recognise only ordinary α-linked bonds, so this protects it.
- Ubiquitin, the protein tag that marks other proteins for destruction, is attached through a bond between its C-terminal glycine and the side-chain amino group of a lysine.
- Transglutaminases join glutamine and lysine side chains. One of them, factor XIIIa, stabilises blood clots, and microbial transglutaminase is used in the food industry to bind proteins together.
An isopeptide bond has the same mass and formula as an ordinary peptide bond, so a mass measurement alone cannot tell them apart.
Detecting peptide bonds
Because the amide group is the same in every residue, the backbone itself can be detected, without knowing the sequence. Three classic methods do this, each with its own limits: a colour reaction, ultraviolet absorption and infrared spectroscopy.
Colour: the biuret test
One of the oldest ways to show that a solution contains peptides or proteins is the biuret reaction. An alkaline solution of copper sulfate, usually stabilised with tartrate, is pale blue; with protein it develops a reddish-violet copper complex that absorbs most strongly at about 545 nm. In 1949 Allan Gornall and colleagues in Toronto turned the reaction into a simple, reproducible method for measuring the protein content of blood serum[6], and biuret methods are still used for total protein in clinical laboratories. The colour generally needs at least two peptide bonds close together, so the test does not reliably detect dipeptides.
Ultraviolet light
The amide group absorbs ultraviolet light strongly below about 230 nm, with a strong band near 190 nm and a weaker one around 210–220 nm. HPLC detectors set to 214–220 nm therefore see every peptide, with or without aromatic residues. The familiar reading at 280 nm is different: it comes from tryptophan and tyrosine side chains, not from the peptide bond. Circular dichroism in the far UV, roughly 190–250 nm, uses the same absorption to report on secondary structure, because amide groups arranged in a helix or sheet absorb left- and right-circularly polarised light differently; an α-helix gives characteristic minima at 208 and 222 nm.
Infrared: the amide bands
In an infrared spectrum the peptide bond gives two strong signals. The amide I band, near 1650 cm⁻¹, comes mainly from stretching of the C=O bond; the amide II band, near 1550 cm⁻¹, from bending of the N–H bond combined with C–N stretching. Side chains hardly affect amide I, but the backbone structure does, which makes it the band of choice for estimating secondary structure: α-helices absorb close to 1655 cm⁻¹, while antiparallel β-sheets give a strong band near 1630 cm⁻¹ and a weaker one near 1685 cm⁻¹[7].
In short
- A peptide bond is the amide link formed between two amino acids with the loss of one water.
- Resonance makes it flat and rigid, so a chain can bend only at the bonds beside each α-carbon; these angles, φ and ψ, define its shape.
- Nearly all peptide bonds are trans; the bond before proline is the main exception.
- In neutral water, without a catalyst, a peptide bond lasts for centuries, yet enzymes cut it in a fraction of a second, which is what makes proteases so useful.
References
- Pauling L, Corey RB, Branson HR (1951). The structure of proteins: two hydrogen-bonded helical configurations of the polypeptide chain. Proceedings of the National Academy of Sciences USA 37:205–211. DOI: 10.1073/pnas.37.4.205
- Prabhakar R (2013). Peptide bond. In: Brenner's Encyclopedia of Genetics, 2nd edition, vol. 5, p. 262. Elsevier. DOI: 10.1016/B978-0-12-374984-0.01133-5
- Jabs A, Weiss MS, Hilgenfeld R (1999). Non-proline cis peptide bonds in proteins. Journal of Molecular Biology 286:291–304. DOI: 10.1006/jmbi.1998.2459
- Ramachandran GN, Ramakrishnan C, Sasisekharan V (1963). Stereochemistry of polypeptide chain configurations. Journal of Molecular Biology 7:95–99. DOI: 10.1016/S0022-2836(63)80023-6
- Radzicka A, Wolfenden R (1996). Rates of uncatalyzed peptide bond hydrolysis in neutral solution and the transition state affinities of proteases. Journal of the American Chemical Society 118:6105–6109. DOI: 10.1021/ja954077c
- Gornall AG, Bardawill CJ, David MM (1949). Determination of serum proteins by means of the biuret reaction. Journal of Biological Chemistry 177:751–766. DOI: 10.1016/S0021-9258(18)57021-6
- Barth A (2007). Infrared spectroscopy of proteins. Biochimica et Biophysica Acta 1767:1073–1101. DOI: 10.1016/j.bbabio.2007.06.004