Cyclic and Linear Peptides: What Cyclisation Changes
Few modifications change a peptide as thoroughly as closing it into a ring. Stability shifts, selectivity shifts, and – the part that gets least attention – so does the set of analytical questions the compound has to answer before you can trust what is in the vial.
The ring chemistries in use
Three routes account for most cyclic peptides, and the choice is not cosmetic: it decides what the finished compound will tolerate in a buffer.
Disulfide bridges
The mildest option, formed by oxidizing two cysteine thiols. Synthesis is normally run in dilute solution, since high dilution suppresses intermolecular reaction and pushes the equilibrium toward the intramolecular bridge, with air, DMSO or iodine serving as oxidant. When more than one cysteine pair is present, orthogonal protecting groups such as Acm, Trt or Mmt are removed in a controlled sequence so each bridge forms between its intended partners. Perform that sequence wrongly and you obtain a disulfide isomer: right mass, wrong three-dimensional structure, and invisible to mass spectrometry.
The bond is redox-sensitive by design. Dithiothreitol, TCEP, 2-mercaptoethanol and glutathione all reduce it, so any of them present in a buffer opens the ring, and a reducing intracellular environment may not leave the bridge intact either. That is a practical limitation rather than a theoretical one, and it is the usual reason a chemist reaches for a lactam instead when the constraint has to survive.
Lactam bridges
Here the bond is an amide between a side-chain amine, typically lysine or ornithine, and a side-chain carboxyl, typically aspartate or glutamate. It is built on resin using orthogonal side-chain protection and a conventional coupling reagent, and what results is an ordinary amide: unaffected by reduction, stable across the normal pH range, and cleavable only under conditions that would take the backbone apart anyway.
Spacing is a design variable. A bridge from i to i+4 locks a single alpha-helical turn, while i to i+7 spans two. Selecting that spacing is how the chemist chooses which conformation to fix, and it explains why two lactam analogs of one parent sequence can differ substantially in receptor selectivity.
Head-to-tail closure
Bonding the N-terminal amine directly to the C-terminal carboxyl eliminates both free ends and yields a fully cyclic backbone. It constrains the molecule most and is correspondingly the hardest to make, since the desired reaction competes against cyclodimerization and oligomerization. High dilution plus a pseudo-proline or turn-inducing residue to pre-organize the chain is the standard answer. Sequences of five residues or fewer are notoriously awkward for purely geometric reasons.
Less common alternatives
- Staples and thioethers. All-hydrocarbon staples installed by ring-closing metathesis across two alkene-bearing residues. They cannot be reduced, and they frequently improve membrane permeability along with helicity.
- Side-chain-to-tail and tail-to-side-chain. One terminus tied to a side chain, leaving the other terminus available for further modification.
- Click chemistry. An azide and an alkyne give a triazole: chemically inert, straightforward to install, but bulkier and more rigid as a linker than an amide.
What the constraint is worth
Take away the free termini and you take away the substrate exopeptidases need, which is why cyclic peptides are generally the more stable class. The deeper effect is conformational.
A linear peptide in solution samples an enormous ensemble of shapes, and only a small slice of that ensemble resembles the receptor-bound geometry. Binding therefore carries an entropic price: freedom has to be surrendered to reach the bound state. Pre-organizing the chain into something near that geometry pays part of the price up front, and affinity rises accordingly.
Selectivity is the same argument run backwards. A flexible peptide can mold itself to several related receptors; a constrained one presents one fixed shape, which suits the receptor it was designed against and suits its relatives less well. Hence the familiar pattern where cyclization improves selectivity while costing potency at the primary target – a locked conformation is seldom an exact match at the first attempt, and finding a bridge position and length that improves both takes iteration. Much of the selectivity engineering in our note on melanocortin receptor subtypes rests on exactly this.
Permeability is not automatic
Cyclic peptides can cross membranes better than their linear counterparts, but the benefit does not follow from the ring by itself. The mechanism is intramolecular hydrogen bonding: a cyclic backbone can satisfy its own amide donors internally and so present less polar surface to solvent during transit. N-methylating selected backbone amides pushes the same effect further by deleting donors entirely. Because the whole thing is conformation-dependent, permeability has to be measured rather than inferred from a structure drawing.
The analytical price of a ring
The core difficulty is that ring closure alters molecular mass by a small and entirely predictable amount – and several quite different outcomes produce that same change.
- Disulfide formation costs two hydrogens: minus 2.016 Da. A bridge between the wrong cysteine pair costs exactly the same, and so does an intermolecular bridge in a dimer, at twice the mass.
- Lactam and head-to-tail closure each lose a water: minus 18.011 Da. An aspartimide side product loses the same water, and a cyclodimer does too at twice the mass.
So a mass measurement establishes that some cyclization happened, not that the intended one did. Compounding this, the open-chain form shares the target’s amino acid composition and sits within two hydrogens of its mass, so it can co-elute or elute very near the product, while oligomeric forms match the monomer’s composition exactly. Retention time on its own is therefore weak evidence of identity for a cyclic compound, and a purity figure from one chromatographic run may simply not have resolved the isomers.
Separating them requires orthogonal evidence:
- RP-HPLC retention. A correctly folded cyclic peptide is usually more compact and less hydrophobic at the surface, and generally elutes ahead of its linear precursor. Isomers that mass cannot separate often separate here.
- Ellman’s assay, which quantifies free thiol. A fully oxidized disulfide peptide should register essentially none; anything left indicates incomplete oxidation.
- Size-exclusion chromatography or native MS, which pulls the intramolecular product apart from the cyclodimer where HPLC alone may not.
- Tandem MS following partial reduction. Selective reduction, alkylation and fragmentation maps which cysteines were genuinely paired – the definitive approach for multi-bridge peptides.
- CD or NMR. Circular dichroism gives a quick read on secondary structure; NOE-based NMR gives the actual constraint geometry when the question justifies the effort.
The reasoning behind these methods is set out in our notes on HPLC and mass spectrometry and why certificates disagree on purity.
Reading a certificate written for a ring
Documentation drafted for a linear peptide does not address what a cyclic one raises. Above all, confirm that identity rested on mass rather than retention time alone, and that oligomeric forms were considered at all. Beyond that:
- Observed mass stated next to a calculated mass for the cyclic form, with the lost water or hydrogens accounted for explicitly rather than left for the reader to infer.
- Which cyclization chemistry was used, because that fixes what the compound will survive in a buffer.
- For multi-bridge peptides, evidence of correct connectivity and not merely a matching mass.
- RP-HPLC purity with gradient and column named, since shallow gradients let isomers co-elute.
- Free-thiol content where disulfides are involved.
- Net peptide content, so concentration rests on something firmer than vial weight.
- Counterion identity, since residual trifluoroacetate carries its own activity in certain cell assays.
For what a certificate can and cannot establish, and why two laboratories can report different purity from one vial, see third-party versus in-house peptide testing.
Buffer and storage decisions that fall out of the chemistry
- Keep reducing agents away from disulfide-cyclized peptides. That covers DTT and TCEP introduced for unrelated purposes elsewhere in a protocol, and freshly made buffers carrying 2-mercaptoethanol.
- Minimize headspace oxygen where free cysteine is present, since slow air oxidation will build unintended bridges over weeks in storage.
- Keep an eye on pH. Disulfide exchange speeds up sharply above roughly pH 8, so near-neutral or slightly acidic buffers are the safer default.
- Exclude trace metals. Copper and iron catalyze thiol oxidation, and a chelator in the buffer is inexpensive insurance.
- Aliquot at reconstitution, because repeated freeze-thaw cycles drive both aggregation and disulfide scrambling.
- Reach for low-binding plasticware with cationic or strongly hydrophobic cyclic peptides, where adsorption at low working concentrations is a genuine loss.
General conditions are covered in storage and stability of lyophilized peptides, and record-keeping in keeping a peptide inventory.
How these compounds get written down
Notation across the literature is inconsistent: square brackets around the bridged residues, a cyclo(…) prefix, an explicit “Cys1-Cys6 disulfide” annotation, or no indication whatsoever. One compound can appear under several conventions, and a search built on one of them will miss the rest. The conventions worth learning are gathered in peptide nomenclature: analogs, fragments and salts.
Abbreviations used above
- Disulfide bridge – covalent S-S bond between two cysteine side chains; reversible under reducing conditions.
- Lactam bridge – amide bond between a side-chain amine and a side-chain carboxyl; not reducible.
- Head-to-tail cyclization – amide bond joining the N- and C-termini.
- Staple – hydrocarbon crosslink installed by ring-closing metathesis, usually to stabilize a helix.
- Disulfide isomer – correct mass, wrong cysteine pairing.
- Cyclodimer – two chains closed onto each other rather than each onto itself; twice the mass.
- Aspartimide – a cyclic side product at aspartate residues, also minus one water.
- Pre-organization – restricting conformational freedom so that less entropy is surrendered on binding.
ExoLabz supplies compounds for laboratory research use only. Nothing on this page constitutes medical advice or a suggestion of human or veterinary use. Analytical documentation is published where it is available.
