Attaching a fatty acid to a peptide is one of the few modifications that changes a molecule's behaviour in the body more than it changes its sequence. It is also widely misdescribed as simply "adding a lipid". In the designs that work, the lipid is one of three parts, and the other two — where it attaches and what sits between the attachment point and the chain — do as much of the work. This page covers what lipidation achieves, how it is built on resin, and what changes downstream once the peptide carries a long hydrophobic chain.
The reason matters, because it determines which chain and which attachment point make sense.
The published design of semaglutide is a compact lesson in how these constructs are put together. The extended half-life comes from albumin binding, driven by a large fatty-acid-derived moiety on the lysine at position 26. The molecule carries three deliberate changes: alanine at position 8 replaced by 2-aminoisobutyric acid to resist dipeptidyl peptidase-4; the lysine at position 34 replaced by arginine, specifically to limit the options for acylation to the one remaining lysine; and a large hydrophilic spacer between that lysine and the γ-glutamate carrying a C18 fatty diacid with a terminal acidic group. Spacer and fatty acid both contribute to albumin binding, which lowers renal clearance; the reported half-life in humans is 165 hours, close to a week7. Liraglutide, the earlier compound in the same series, is acylated with a shorter chain through a simpler attachment and is dosed once daily1,7. The trade-off is explicit in the published work: semaglutide binds the GLP-1 receptor about threefold more weakly than liraglutide, and that loss of potency was accepted in exchange for the stronger albumin binding that gives the longer duration7.
The transferable point is the Lys34 substitution. Site-specific acylation was not achieved by finding selective coupling conditions. It was achieved by removing the competing nucleophile from the sequence. If a design has two lysines and only one should be modified, that is a sequence problem to solve at the design stage, or an orthogonal protecting group problem to solve on resin — not something a coupling protocol will fix.
The standard approach assembles the backbone by normal Fmoc chemistry, with the target lysine carrying a side-chain protecting group orthogonal to both Fmoc and the final TFA cleavage. That group is removed selectively at the end of the assembly, and the exposed ε-amine is then acylated with the spacer units and the fatty acid in sequence2.
| Group | Removed by | Stable to | Watch for |
|---|---|---|---|
| Mtt, Mmt | 1% TFA in DCM, often with 1–5% methanol or a silane scavenger3 | Piperidine, hydrazine, Pd(0)3 | 1–3% TFA needs tuning per sequence to avoid stripping Boc or cleaving the linker; 1% TFA has been shown to take an O-trityl off serine4 |
| Dde, ivDde | 2% hydrazine in DMF, repeated; up to 10% where ivDde proves stubborn3 | 20% piperidine and TFA3 | Hydrazine also removes Fmoc, so the backbone must be finished first3 |
| Alloc | Pd(0) with a silane or borane scavenger2 | Piperidine, TFA, hydrazine | Residual palladium; requires scrupulous washing |
The ordering constraint in row two is the one that most often governs the route. Because hydrazine strips Fmoc as readily as it strips Dde, a Dde or ivDde strategy commits you to completing the whole backbone before the side chain can be touched3. An Mtt strategy leaves the option of unmasking mid-synthesis, at the cost of working with dilute acid near acid-labile linkers.
Milder alternatives to dilute acid now exist for the trityl family. A head-to-head comparison found that 30% hexafluoroisopropanol or 30% perfluoro-tert-butanol in dichloromethane removes Mtt cleanly (three treatments of 5 minutes for hexafluoroisopropanol, three of 15 minutes for perfluoro-tert-butanol) while leaving Boc and tert-butyl groups untouched even after four hours — which is the real reason to reach for them, since the acid route is tuned not for speed but to avoid collateral damage4.
One procedural detail is worth stating plainly because it is easy to get wrong: keep the N-terminus protected until the side-chain work is finished. In a published comparison of Lys(Mmt), Lys(Alloc) and Lys(ivDde) routes to the same branched peptide, the N-terminal Fmoc was deliberately left in place through deprotection and side-chain coupling to prevent the free N-terminus competing for the acylation; the three routes gave the product at 79%, 82% and 93% purity respectively6. An unmasked N-terminus turns a site-specific modification into a mixture.
Process work on double-acylated GLP-1 analogues shows how much latitude exists even within one programme: the ε-amine was protected as Mtt, Mmt, Dde, ivDde or Boc depending on the route chosen; Dde and ivDde were removed with 2% hydrazine in NMP; in one variant Mtt was removed with neat hexafluoroisopropanol rather than acid; and the lipid arm could be installed either by acylating the resin-bound peptide or by acylating the fully unprotected peptide in solution5.
Acylating a purified peptide in solution avoids every orthogonal-protection question, but reintroduces the selectivity problem: the N-terminus and every unprotected lysine now compete. It is a reasonable route when the sequence contains a single amine, when the N-terminus is capped, or when a modest mixture can be separated chromatographically. It is also the pragmatic answer when a peptide already in hand needs a lipid added, rather than being resynthesised. On resin, by contrast, the protecting-group scheme is what delivers the selectivity, and the excess reagent is simply washed away — which is usually decisive at scale. Incomplete deprotection and reaction at unintended sites still have to be excluded analytically.
A lipidated peptide is a different compound to handle, and most of the surprises are physical rather than chemical.
For the orthogonal chemistry used to attach other things at a chosen residue, see bromoacetyl peptides; for how terminal modifications affect charge and stability, see acetylation and amidation.
| Attachment | Usually a lysine ε-amine, often through a γ-glutamate unit |
| Spacer | Hydrophilic units that set reach and restore solubility |
| Lipid | Chain length and terminal group set the strength of albumin binding |
| On-resin route | Orthogonal protection (Mtt/Mmt, Dde/ivDde, Alloc); selectivity guaranteed |
| Solution route | No special protection needed; selectivity must come from the sequence |
LifeTein synthesises lipidated and otherwise modified research peptides, including site-specific acylation through orthogonal lysine protection, γ-glutamate and PEG-type spacers, and a range of fatty acid chains. Send us the sequence, the attachment point and the chain you want, and we will confirm the route before synthesis starts. All material is supplied for research use.