Skip to main content

Branched Peptides as Drug Candidates

Why Branching Helps a Peptide Drug Candidate

Peptides make attractive drug leads and poor drugs, for the same reason: they are built of the units proteases exist to cleave, and they are small enough for the kidney to clear in minutes. Branching addresses both problems without changing the sequence that does the binding, which is why so many marketed peptide drugs carry something on a lysine side chain.

The problem branching is solving

An unmodified peptide in plasma faces two independent clearance routes. Proteases cut it, working inward from the termini and at internal recognition sites. Separately, anything below roughly 50 kDa is filtered at the glomerulus, and a 3 to 4 kDa peptide passes straight through. A sequence can be made entirely protease-resistant and still be gone in an hour because of renal filtration alone.

Increasing the effective size of the molecule addresses the second route. PEGylation does it by attaching a large synthetic chain. Acylation does it differently and more elegantly: a fatty acid on a side chain binds reversibly to serum albumin, and the peptide travels as part of a 66 kDa complex that is too large to filter. Because the binding is reversible, free peptide is released continuously from a circulating reservoir.

The worked example: a side chain that buys a weekly dose

Semaglutide is the clearest illustration, because the design trade-off was stated explicitly by the group that made it. The molecule is a GLP-1 analogue carrying, on the lysine at position 26, a short spacer of two AEEA units and a C18 diacid. The diacid is what binds albumin.

What makes it instructive is that the branch cost potency. Semaglutide binds the GLP-1 receptor about threefold more weakly than liraglutide, and that loss was accepted deliberately in exchange for stronger albumin binding and the once-weekly dosing it allows.1 The branch is not a free improvement; it is a trade of receptor affinity for residence time, and it is worth making only when exposure rather than potency is what limits the molecule. The chemistry of building that side chain is covered on the lipidated peptides page.

The design question is which limit you are against. If a peptide is potent but cleared in minutes, a side-chain modification that extends exposure is the obvious move even at some cost in affinity. If it is already long-lived but weakly binding, adding a branch makes it worse on the axis that matters. The published GLP-1 work is useful precisely because it shows the trade being made on purpose rather than discovered afterwards.1

Avidity: several copies binding at once

The second reason to branch has nothing to do with pharmacokinetics. When a target presents several binding sites close together, a molecule carrying several copies of the binding sequence can engage more than one at a time. Dissociation then requires all the contacts to break at once, and because a released arm tends to rebind before the whole molecule diffuses away, the complex comes apart far more slowly than the single-site affinity would predict.

The practical consequence is that functional affinity can improve by orders of magnitude while the intrinsic affinity of each arm is unchanged. It also means the improvement is conditional: it appears only when the target's sites are spaced so the arms can reach them. A multivalent construct against a target presenting one isolated site gains nothing, and the geometry of the spacer becomes part of the design. See peptide versus protein binding affinity for how this shows up in measurements.

Protease resistance, and its limits

A branch point crowds the backbone around it, and exopeptidases working in from a terminus are slowed by a substituted lysine in their path. This is a real effect but a local one: it protects the region near the branch, not the whole molecule, and it does nothing about an endopeptidase site elsewhere in the sequence. Where proteolysis is the limiting problem, backbone modification, D-amino acid substitution or cyclisation address it more directly. See peptide backbone modifications and peptide stability and degradation.

What this means for a synthesis request

GoalArchitectureWhat to specify
Extend circulating half-lifeSingle side-chain acylationThe lysine position, the spacer, and the fatty acid or diacid
Raise functional affinity for a clustered targetTwo or four copies on a branched coreCopy number and the spacing between arms
Raise antibodies without a carrierMAP-4 or MAP-8See MAP peptides
Carry a payload alongside a targeting sequenceBi-epitopic branchWhich arm carries which sequence, and the attachment chemistry

In every case the position of the branch point matters as much as what is attached to it. A lysine inside the binding motif will usually cost activity; one on a solvent-exposed face away from the motif usually will not. Tell us what the peptide has to do, and we can advise on where the branch should sit.

Further reading

References

  1. Lau J, Bloch P, Schäffer L, Pettersson I, Spetzler J, Kofoed J, Madsen K, Knudsen LB, McGuire J, Steensgaard DB, Strauss HM, Gram DX, Knudsen SM, Nielsen FS, Thygesen P, Reedtz-Runge S, Kruse T. Discovery of the once-weekly glucagon-like peptide-1 (GLP-1) analogue semaglutide. Journal of Medicinal Chemistry 2015;58(18):7370–7380. doi:10.1021/acs.jmedchem.5b00726