Short answer: The peptide backbone is the long, repeating –N–CH–CO– chain that links every residue in a peptide or protein [1]. Each residue contributes three backbone atoms (the amide nitrogen, the alpha carbon and the carbonyl carbon), and the shape of the chain is described by three torsion angles: φ (phi) about N–Cα, ψ (psi) about Cα–C and ω (omega) about the C–N peptide bond [3]. Because the C–N bond has partial double-bond character, ω sits near 180° (trans) or 0° (cis), so almost all of the chain's flexibility comes from φ and ψ [1][3]. Plotting φ against ψ gives the Ramachandran plot, a standard tool for describing and validating peptide and protein structure [8][9].
This article is educational chemistry for laboratory readers. It does not describe effects in people or animals and is not guidance for human or animal use.
What counts as the backbone
In organic chemistry terms, the long, repetitive sequence of –N–CH–CO– atoms that makes up a continuous chain is called the backbone [1]. The side chains (the R groups that make each amino acid different) sit on the alpha carbons and are not part of it [13].
The IUPAC-IUB conformation rules write a residue as -NH-CHR-CO- and number residues from the amino-terminal to the carboxyl-terminal end, with i as the residue number, so the alpha carbon of residue i is written Cαi [2]. The same rules also define an alternative "peptide unit," grouping the atoms -CHR-CO-NH- together; the two numbering schemes agree for every atom except the NH [2]. (See our N-terminus vs C-terminus guide for the direction convention and our dipeptide guide for the simplest backbone.)
Why the peptide bond is flat
An amide nitrogen is nonbasic because its unshared electron pair is delocalized by resonance with the carbonyl group [1]. That overlap gives the C–N bond some double-bond character and restricts rotation around it, so the amide bond is planar, with the N–H oriented 180° to the C=O [1]. (Our peptide bond guide covers how the bond forms.)
The IUPAC rules state the consequence in angle terms: owing to the partial double-bond character, ω can normally take values only near 0° or 180°, and ω ≈ 180° (trans) is the value generally found [3].
"Planar" is an approximation, though. A survey of peptide ω angles in the Cambridge Structural Database found that substantial deviations from planarity are tolerated, with a standard deviation of up to 6° about a trans mean that is less than 180° [4]. The distortion can come from pyramidalization at the amide nitrogen as well as simple twisting about the bond [4].
Cis peptide bonds: rare, and mostly before proline
A cis ω is possible but uncommon. In one survey of the Brookhaven Protein Data Bank, only 17 of 31,005 amide bonds (0.05%) were cis, compared with 99 of 1,534 X-Pro imide bonds (6.5%) [5]. Cis bonds were found mainly in bends and turns [5].
Non-proline cis bonds do occur. A later study of 571 non-redundant proteins identified 43 of them, and found they were significantly more frequent in structures solved at high resolution, suggesting they may be more abundant than previously thought [6]. All residues on the N-terminal side of these non-Pro cis bonds sat in the beta region of the φ/ψ plot [6].
The three backbone torsion angles
The IUPAC definitions are precise [3]:
- φ (phi): rotation about the N–Cα bond, defined by the atoms Ci−1, Ni, Cαi, Ci.
- ψ (psi): rotation about the Cα–C bond, defined by Ni, Cαi, Ci, Ni+1.
- ω (omega): rotation about the C–N bond between residues i and i+1, defined by Cαi, Ci, Ni+1, Cαi+1.
The IUPAC-IUB rules also list approximate angles for familiar regular structures [3]:
- Right-handed α-helix: φ −57°, ψ −47°, ω 180°.
- Parallel pleated sheet: φ −119°, ψ +113°.
- Antiparallel pleated sheet: φ −139°, ψ +135°, ω −178°.
- Polyproline II: φ −78°, ψ +149°, ω 180°.
- Polyproline I: φ −83°, ψ +158°, with ω 0° (cis).
The Ramachandran plot
Using torsion angles to describe polypeptide conformation was developed by Sasisekharan while studying collagen as a graduate student in G.N. Ramachandran's group [8]. The classic reference is the 1963 paper by Ramachandran, Ramakrishnan and Sasisekharan, "Stereochemistry of polypeptide chain configurations" [7]. The original plot treated atoms as simple impenetrable spheres [8]. For alanine-like residues (neither glycine nor proline), it gave three main "allowed" regions: a large alpha region, a large beta region and a much smaller αL region that mirrors the alpha region [8].
Real structures refined the picture. When ~60,000 well-defined residues from crystal structures solved at better than 1.2 Å were overlaid, most fell into clusters near the alpha, beta and αL regions but did not completely match the classical boundaries [8]. Validation tools such as PROCHECK, MOLEMAN2 and MolProbity therefore use Ramachandran regions based on empirical distributions rather than the original predictions [8].
One widely used empirical version came from 81,234 non-Gly, non-Pro and non-pre-Pro residues from 500 high-resolution proteins [9]. It showed sharp boundaries and separated large empty areas from regions that are allowed but disfavored [9]. One example is the γ-turn conformation near φ, ψ = (+75°, −60°): common validation programs counted it as forbidden, yet it occurs in well-ordered parts of good structures [9].
A finer map of backbone conformations
Hollingsworth and Karplus argued that the three-region picture (alpha, beta, αL) is inadequate and misleading, and that at least five regions are needed to describe the dominant areas: α, δ, β, PII and δ′ [8]. Their proposed labels include [8]:
- α: a fairly small area centered on the α-helix peak at φ, ψ = (−63°, −43°), containing roughly one-third of residues in diverse proteins.
- β: the natural grouping that holds most residues forming β-strands.
- PII: the right-hand part of the classical beta region. When repeated, it makes the polyproline-II conformation, which the authors suggest renaming "polypeptide-II." They say grouping it with β-structures was a historical mistake, caused by its lack of regular backbone hydrogen bonding.
- γ and γ′: the rare γ-turn region near (+80°, −80°) and its mirror image near (−80°, +80°).
Glycine, proline and pre-proline get their own plots
Residue type matters. There are four basic Ramachandran plot types: generic (the 18 non-glycine, non-proline amino acids), glycine, proline and pre-proline (residues that come just before a proline) [10]. Separate favored and allowed φ, ψ regions have been defined for Pro, pre-Pro and Gly [9].
- Glycine lacks a side chain, and in particular the Cβ atom that causes many steric clashes in the generic plot [10]. Its plot can therefore cross the −180°/180° borders, and the observed glycine map has five regions of density [10]. Glycine's ψ is typically clustered at 180° and 0° [10]. The sparsely populated ε region, a fairly extended conformation, is largely populated by glycine, which Hollingsworth and Karplus describe as the only amino acid that can readily adopt it [8].
- Pre-proline has a region unique to it, the ζ region, which is stabilized by a weak CO(i−1)···CδHδ(i+1) hydrogen bond to the following proline [10].
Changing the backbone on purpose
Because the backbone sets the conformational rules, chemists modify it directly:
- N-methylation of backbone amides is one such edit; cyclosporine is a natural, multiply N-methylated cyclic peptide [12]. One group synthesized a complete library of N-methylated cyclic alanine pentapeptides, which gave insight into how N-methylation modulates cyclic peptide conformation [12]. (See our cyclic peptides guide.)
- Peptoids go further. They are oligomers of N-substituted glycine units, with side chains on the amide nitrogens instead of the alpha carbons [13]. Their conformational flexibility comes from a lack of backbone chirality, the absence of backbone NH hydrogen-bond donors and easy cis-trans isomerization of their tertiary amide bonds [13].
- D-residues flip the stereochemistry at Cα. For D-amino acids, the IUPAC φ/ψ reference table swaps the Cα–H and Cα–R entries [3]. (See our D-amino acids guide.)
FAQ
What is the peptide backbone?
The continuous, repeating –N–CH–CO– chain of a peptide or protein, without the side chains [1].
What are phi and psi angles?
φ is the torsion angle about the N–Cα bond and ψ the torsion angle about the Cα–C bond of a residue [3].
Why is omega usually 180°?
The C–N peptide bond has partial double-bond character, which keeps ω near 0° or 180°. The trans value of about 180° is the one generally found [1][3].
How common are cis peptide bonds?
In one Protein Data Bank survey, 0.05% of ordinary amide bonds and 6.5% of X-Pro bonds were cis [5].
What does a Ramachandran plot show?
The φ and ψ angles of residues plotted against each other. Empirical versions are used to check the stereochemical quality of solved structures [8][9].
Bottom line
The backbone is the –N–CH–CO– repeat shared by every residue [1]. Its planar, mostly trans peptide bonds [1][3][5] leave φ and ψ as the main degrees of freedom [3]. The Ramachandran plot maps those angles [7][8], and modern empirical versions separate generic, glycine, proline and pre-proline residues [9][10]. Backbone edits such as N-methylation and peptoid N-substitution change those rules on purpose [12][13].
All products are for research use only — not for human or animal consumption.