Skip to main content

Branched Peptide Synthesis

How a Branched Peptide Is Built

Building a branch means freeing one amino acid side chain while every other reactive group in the molecule stays protected. Standard Fmoc chemistry has only two levels of protection, and neither of them can do that, so the branch point needs a third protecting group chosen to be removable by chemistry that touches nothing else.

Why a third protecting group is needed

In ordinary Fmoc synthesis, the α-amine carries Fmoc and comes free with piperidine at every cycle, while the side chains carry acid-labile groups — tBu, Boc, Pbf — that come off together in the final TFA cleavage. Those two levels are enough to assemble a linear chain. They are not enough to open one specific side chain partway through, because removing the acid-labile set removes all of it and takes the peptide off the resin at the same time.

A branch therefore needs a lysine whose side chain carries a group that survives piperidine at every cycle, survives until it is wanted, and then comes off under conditions that leave both of the other sets intact.

Scheme of three protection sets on a peptide: Fmoc on the alpha amine removed by 20 percent piperidine at every cycle; a branch-point lysine carrying Dde, ivDde, Alloc or Mtt removed once by hydrazine, palladium or dilute acid; and permanent tBu, Boc and Pbf side-chain groups removed only by the final TFA cleavage. A caution describes Dde migration to a free lysine side chain during piperidine treatment.
The three sets and what removes each. The branch group is chosen for what it survives, not only for how easily it comes off.

The groups used, and what each costs

The approach dates from work that introduced Dde specifically to make branched peptides: a lysine protected as the branching motif allowed a 34-residue di-epitopic peptide to be assembled by continuous-flow solid phase synthesis, using a group orthogonal to both Fmoc and Boc.1

GroupRemoved byStrengthsWhat to watch
Dde2% hydrazine in DMFMild, quick, no metal catalystCan migrate to a free lysine side chain during piperidine treatment3
ivDde2% hydrazine in DMF, more slowlyMore hindered, so markedly more resistant to migration2More resistant, not immune; migration has still been reported
AllocPd(0) with a silane or other scavengerChemistry unrelated to amines, so no migration routeNeeds a palladium catalyst and careful scavenging; residual metal must be considered
Mtt1% TFA in dichloromethaneNo metal, no hydrazineAcid-labile, so the margin against the permanent groups is narrower; needs care on acid-sensitive sequences

The failure that is worth designing against

Dde does not always stay where it was put. An unprotected lysine ε-amine elsewhere in the peptide can acquire the Dde group from a protected one, and the transfer happens during the piperidine treatment used for Fmoc removal and in the washes that follow.3 The published investigation also found that an unprotected α-amine cannot take Dde from an ε-amine, so the migration is specifically between side chains, and that it can occur in DMF alone by direct nucleophilic attack.3

The consequence for a branched synthesis is serious and quiet: the branch ends up on the wrong lysine. The product has the right mass, because nothing has been gained or lost, and the error is visible only by a method that locates the attachment. That is why branched peptides with more than one lysine are built with ivDde or Alloc rather than Dde, and why a mass spectrum alone is not sufficient evidence that a branched peptide is what was ordered.

What this means when ordering. Tell us every lysine in the sequence, not only the branch point. If the peptide contains other lysines, the protecting-group strategy has to change, and an ε-amine that will be free in the final product is a particular constraint. This is usually decided before the first coupling, not after.

Assembling the arms

Once the side-chain amine is free, the second arm is built on it by ordinary Fmoc chemistry. Two practical points follow from the geometry rather than the chemistry:

  • Couplings on a branch are slower. The branch point is crowded, and the arms interfere with each other as they grow. Longer coupling times, repeated couplings or more aggressive activation are routine rather than exceptional.
  • A failure on one arm is a large failure. On a linear peptide an incomplete coupling gives a deletion of one residue. On a branched peptide a failure early on one arm gives a product missing most of a sequence, which is both a bigger impurity and a harder one to separate.
  • A spacer is often worth adding. Inserting a short spacer such as Ahx or β-alanine between the core and the arm moves the sequence away from the crowded centre, which improves both the coupling yields and, for an immunogen, the accessibility of the epitope.

Confirming that the branch is where it should be

Mass spectrometry confirms the composition but not the connectivity, and for a branched peptide those are different questions — a correctly branched product and a mis-branched one are isomers. Useful evidence comes from combining methods: HPLC to show a single product, amino acid analysis to confirm the overall composition and the peptide content, and where the attachment point itself is in question, enzymatic or chemical fragmentation followed by analysis of the pieces. See reading a peptide mass spectrum for what the mass data can and cannot settle.

Further reading

References

  1. Bycroft BW, Chan WC, Chhabra SR, Hone ND. A novel lysine-protecting procedure for continuous flow solid phase synthesis of branched peptides. Journal of the Chemical Society, Chemical Communications 1993;(9):778–779. doi:10.1039/C39930000778
  2. Chhabra SR, Hothi B, Evans DJ, White PD, Bycroft BW, Chan WC. An appraisal of new variants of Dde amine protecting group for solid phase peptide synthesis. Tetrahedron Letters 1998;39(12):1603–1606.
  3. Augustyns K, Kraas W, Jung G. Investigation on the stability of the Dde protecting group used in peptide synthesis: migration to an unprotected lysine. Journal of Peptide Research 1998;51(2):127–133.