Key points
- Cyclisation removes the flexible ends of a peptide. The result is usually more rigid, more resistant to proteases, and often binds its target more tightly.
- Head-to-tail cyclisation joins the two termini with an ordinary peptide bond and loses one water: −18.011 Da. Side-chain bridges have their own mass changes.
- Nature makes cyclic peptides on a large scale, mostly through nonribosomal synthetases and through the RiPP pathways that produce lantibiotics and cyclotides.
- A cyclic peptide has no free termini, so sequencing it by Edman degradation or by standard tandem MS is much harder than for a linear peptide.
Why cyclise a peptide
A linear peptide in solution is a flexible chain that samples an enormous number of conformations. Only a small fraction of them resemble the shape that binds a target, so most of the binding energy is spent on ordering the peptide rather than on the interaction itself. Linear peptides are also easy prey for exopeptidases, which chew inwards from a free N- or C-terminus, and their typical half-life in plasma is measured in minutes.
Cyclisation addresses both problems at once:
- Reduced entropy loss on binding. A constrained peptide is already close to its bound conformation, so affinity often improves by one to two orders of magnitude when the constraint is chosen well.
- Protease resistance. With no free termini, exopeptidases have nothing to grip. Endopeptidases can still cut, but the rigid conformation often prevents the peptide from fitting their active site.
- Better membrane permeability in some cases. Cyclic peptides can shield their backbone amides from water by folding and by forming internal hydrogen bonds, which helps them cross membranes. Cyclosporin A, an orally available cyclic undecapeptide, is the classic demonstration that peptides can break the usual rules of drug-like properties.
Ways to close the ring
| Type | Link | Monoisotopic Δ | Example |
|---|---|---|---|
| Head-to-tail (backbone) | N-terminus to C-terminus amide | −18.011 | Gramicidin S, cyclotides |
| Side-chain to side-chain, disulfide | Cys–Cys | −2.016 | Oxytocin, somatostatin |
| Side-chain to side-chain, lactam | Lys amine to Asp or Glu carboxyl | −18.011 | Constrained analogues of many hormones |
| Side-chain to terminus | Lys or Asp side chain to the other end | −18.011 | Daptomycin, polymyxin B |
| Thioether | Cys to a dehydro residue | varies | Nisin and other lantibiotics |
| Hydrocarbon staple | Alkene between two unnatural residues | varies | Stapled α-helices |
Many natural products combine several of these. Daptomycin has a ten-residue ring closed through a side chain plus a lipid tail on an exocyclic segment; polymyxin B has a similar architecture. Peptides in which only part of the chain is in the ring are called partially cyclic or depsipeptide-like, depending on whether the ring contains an ester.
Replacing an amide in the ring with an ester gives a depsipeptide. Esters are more easily hydrolysed and remove one hydrogen-bond donor, which can improve membrane permeability. Valinomycin and the didemnins are natural examples.
Mass and analysis of cyclic peptides
Head-to-tail cyclisation is a condensation like any other peptide bond formation, so the cyclic peptide is exactly one water lighter than its linear counterpart. For the pentapeptide RGDFV, the linear form has a monoisotopic mass of 592.297 Da and the cyclic form 574.286 Da. Note that the MW calculator always adds the terminal water, so for a head-to-tail cyclic peptide you subtract 18.011 Da from its answer.
Cyclisation also changes the charge. A head-to-tail cyclic peptide has no free α-amine and no free α-carboxyl, so two ionisable groups disappear and the isoelectric point is determined entirely by the side chains.
Sequencing is the real difficulty. Edman degradation needs a free N-terminal amine and simply fails. Tandem mass spectrometry works but produces confusing spectra: the ring first has to open at a random amide, so a population of linear acylium ions with different starting points is produced, each fragmenting further. The resulting spectrum contains overlapping series rather than the clean b and y ladders described in the mass spectrometry guide. The usual solution is to open the ring chemically or enzymatically at a defined position first, then sequence the linear product.
Making cyclic peptides
The chemistry itself is an ordinary amide coupling, but it has to happen between two ends of the same molecule rather than between two different ones. The competing reaction is intermolecular: two peptides joining to form dimers and higher oligomers.
The standard remedy is high dilution, typically 0.1–1 mM, so that a given chain is more likely to meet its own other end than another molecule. Slow addition of the peptide into the reaction and the use of pseudo-dilution on solid support achieve the same effect.
Ring size matters. Cyclising peptides of five or more residues is usually straightforward; tetrapeptides and especially tripeptides are difficult because the ring is strained, and the competing formation of diketopiperazines from the first two residues becomes significant. Residues that favour turns help a great deal: proline, glycine and D-amino acids at the right positions can raise cyclisation yields dramatically by pre-organising the chain into a turn that brings the ends together. This is one reason so many natural cyclic peptides contain D-residues.
Practical approaches in Fmoc chemistry include:
- Cyclisation in solution after cleaving a fully side-chain-protected linear peptide from a very acid-labile resin such as 2-chlorotrityl, followed by a final deprotection.
- On-resin cyclisation, where the peptide is anchored through a side chain and the termini are joined while still attached, which benefits from pseudo-dilution.
- Native chemical ligation between an N-terminal cysteine and a C-terminal thioester in the same molecule, used for larger rings.
- Click chemistry and ring-closing metathesis, which give triazole and hydrocarbon links rather than amides.
Cyclic peptides in nature
Most natural cyclic peptides come from two kinds of machinery. Nonribosomal peptide synthetases are large modular enzymes found in bacteria and fungi; they assemble peptides without a ribosome and routinely incorporate D-amino acids, N-methylated residues and other building blocks that the ribosome cannot use. Cyclosporin, gramicidin S, vancomycin and daptomycin all come from this route.
The other route is RiPPs, ribosomally synthesised and post-translationally modified peptides. Here a normal gene encodes a precursor that is then heavily modified and cyclised by dedicated enzymes. Lantibiotics such as nisin, used as a food preservative, and the plant cyclotides are made this way.
| Peptide | Source | Ring | Role |
|---|---|---|---|
| Cyclosporin A | Fungus Tolypocladium inflatum | Head-to-tail, 11 residues, several N-methylated | Immunosuppressant; orally available |
| Gramicidin S | Aneurinibacillus migulanus | Head-to-tail cyclic decapeptide with two D-Phe | Topical antibiotic |
| Vancomycin | Amycolatopsis orientalis | Cross-linked glycopeptide | Antibiotic of last resort for some infections |
| Daptomycin | Streptomyces roseosporus | Side-chain lactone ring plus lipid tail | Antibiotic against Gram-positive bacteria |
| Nisin | Lactococcus lactis | Five thioether rings | Food preservative |
| Kalata B1 | Plant Oldenlandia affinis | Head-to-tail plus cystine knot | Plant defence; a scaffold for drug design |
| α-Amanitin | Death cap mushroom | Bicyclic octapeptide | Inhibits RNA polymerase II; highly toxic |
The recurring theme is defence. Cyclic peptides are chemically tough, and organisms use them where a molecule must survive outside the cell.
Cyclic peptides as drugs
Cyclic peptides occupy a space between small molecules and biologics. They are large enough to cover the flat, extended surfaces of protein–protein interfaces, which small molecules struggle with, yet small enough to be made by chemical synthesis.
Two design approaches are widely used. Stapled peptides lock an α-helix by joining two residues on the same face, at positions i and i+4 (one helical turn) or i and i+7 (two turns), usually with a hydrocarbon bridge made by ring-closing metathesis. The staple enforces helicity, improves protease resistance, and in some cases improves cell uptake.
Macrocycle libraries take the opposite approach: instead of designing a single molecule, enormous numbers of cyclic peptides are made and screened. mRNA display methods such as the RaPID system can search libraries of 10¹² or more cyclic peptides for binders to a chosen target, and can include non-standard amino acids through reprogrammed genetic codes.
The main remaining obstacle is oral bioavailability. Cyclosporin manages it by folding so that its polar amide NH groups are turned inward and hydrogen-bonded to each other, presenting a greasy exterior in a membrane and a more polar one in water. N-methylation of selected amides, which removes hydrogen-bond donors entirely, is the most common way of pursuing the same effect by design.
Frequently asked questions
Does a cyclic peptide have an N-terminus?
Not in the head-to-tail case: every amine and carboxyl is part of a peptide bond. Sequences of such peptides are written starting from an arbitrary residue, so the same molecule can appear under several rotations of its sequence in different papers.
How many peptide bonds does a cyclic peptide of n residues have?
n, one more than the linear peptide of the same length, because the ends are joined as well.
Is a disulfide-closed peptide such as oxytocin a cyclic peptide?
It contains a ring, so it is cyclic in the broad sense, but its backbone is not cyclic: it still has free N- and C-termini, and the tail beyond the ring is linear. The term is usually reserved for peptides whose backbone forms a closed loop, with disulfide-bridged peptides described separately.
Why do so many natural cyclic peptides contain D-amino acids?
Two reasons. The nonribosomal synthetases that make them can epimerise residues, so D-amino acids are available to that machinery. And a D-residue at the right position favours the turn geometry that brings the two ends together, which makes the ring easier to close and more stable once closed.
References
- Craik DJ, Fairlie DP, Liras S, Price D (2013) The future of peptide-based drugs. Chemical Biology & Drug Design 81:136–147.
- Walsh CT (2004) Polyketide and nonribosomal peptide antibiotics: modularity and versatility. Science 303:1805–1810.
- Schafmeister CE, Po J, Verdine GL (2000) An all-hydrocarbon cross-linking system for enhancing the helicity and metabolic resistance of peptides. Journal of the American Chemical Society 122:5891–5892.
- Passioura T, Katoh T, Goto Y, Suga H (2014) Selection-based discovery of druglike macrocyclic peptides. Annual Review of Biochemistry 83:727–752.