Structural Modifications: Acylation, PEGylation, and Aib Substitution

The short version

A peptide drug candidate that works beautifully in a test tube often fails in the body because it is destroyed before it reaches its target. Two culprits dominate: proteases that clip the peptide chain and the kidneys that filter small molecules out of circulation. Drug chemists have developed three main engineering approaches to solve these problems: blocking protease attack sites with non-standard amino acids (Aib substitution), hitching the peptide to the large serum protein albumin via a fatty acid chain (acylation), and attaching a bulky inert polymer to push the molecule above the renal filtration threshold (PEGylation).

Why native peptides make poor drugs

Glucagon-like peptide-1 (GLP-1) illustrates the pharmacokinetic problem clearly. In its native form GLP-1 is a ~30-amino acid peptide with potent insulin-stimulating and appetite-suppressing activity. But after intravenous injection, its plasma half-life is roughly 1.5 to 2 minutes.^[1] The enzyme responsible for this rapid clearance is dipeptidyl peptidase-4 (DPP-4), a serine protease present in endothelial cells, plasma, and the small intestinal brush border. DPP-4 cleaves dipeptides from the N-terminus of peptides that carry a penultimate proline or alanine, which is precisely the structure at position 2 of GLP-1 (Ala²).^[7]

Even if proteolysis were solved, peptides below roughly 65 kDa are freely filtered by the glomerulus and lost in urine. GLP-1 is only about 3.3 kDa. Without modification, any GLP-1-based drug would require continuous intravenous infusion to maintain therapeutic concentrations, making it clinically impractical.

Strategy 1: Aib substitution to block DPP-4

Alpha-aminoisobutyric acid (Aib) is a non-standard amino acid with two methyl groups on its alpha carbon instead of the single hydrogen found in standard amino acids. This gem-dimethyl substitution creates steric bulk that physically blocks the DPP-4 active site from accommodating the peptide. When the alanine at position 2 of an incretin is replaced by Aib, DPP-4 can no longer cleave the N-terminal dipeptide, and the peptide survives in circulation far longer.

Tirzepatide uses this strategy at two positions: Aib at position 2 (blocking DPP-4) and Aib at position 13 (improving overall helical stability of the peptide backbone and resistance to additional proteases).^[4] The structural change is subtle from a molecular weight standpoint but decisive for pharmacokinetics.

D-amino acid substitution is a related approach: replacing an L-amino acid with its mirror-image D-form makes the peptide bond at that position unrecognizable to most mammalian proteases, which are stereospecific.^[8] D-substitution is used in some shorter peptides but less commonly in long-chain therapeutic peptides where maintaining secondary structure matters.

Strategy 2: Fatty acid acylation and albumin binding

Serum albumin is the most abundant protein in plasma (~40 g/L) and has a half-life of roughly 19 days because it is too large to be filtered by the kidneys and escapes lysosomal degradation via the FcRn recycling pathway. Fatty acid acylation exploits albumin by attaching a long-chain fatty acid to a lysine side chain in the peptide, allowing the peptide to bind non-covalently to one of albumin's fatty acid binding sites. The peptide-albumin complex is now effectively ~67 kDa, above the renal filtration threshold, and is cleared with albumin's kinetics rather than the peptide's.

• Liraglutide: a C16 (palmitic) fatty acid attached directly to Lys²⁶ via a glutamate spacer. Terminal half-life ~13 hours, enabling once-daily dosing.^[3]
• Semaglutide: a C18 fatty diacid attached at Lys³⁴via a longer linker incorporating two gamma-glutamate units and a mini-PEG spacer. The longer linker reduces steric interference with the receptor while maintaining strong albumin binding. Combined with Aib⁸ to block DPP-4, semaglutide achieves a terminal half-life of approximately 165 hours (~7 days), enabling once-weekly dosing.^[2]
• Tirzepatide: a C20 fatty diacid (eicosanedioic acid) attached at Lys³⁴ via a linker incorporating a gamma-glutamate and two mini-PEG units. The C20 chain produces tighter albumin binding than the C18 chain in semaglutide, contributing to tirzepatide's mean terminal half-life of approximately 116 hours (~5 days) and its once-weekly dosing schedule.^[4]

The acylation chemistry must be carefully tuned. A fatty acid chain that is too short binds albumin weakly. One that is too long or attached too close to the receptor-binding region reduces pharmacological potency. Linker chemistry (spacers, mini-PEGs, glutamate units) is used to position the fatty acid away from the active pharmacophore.

Strategy 3: PEGylation

Polyethylene glycol (PEG) is a water-soluble synthetic polymer. Attaching PEG chains to a therapeutic protein or peptide (PEGylation) increases its hydrodynamic radius, the effective size the molecule presents in solution.^[5] Once the hydrodynamic radius corresponds to a molecular weight above the renal filtration threshold of roughly 65 kDa, the molecule is no longer freely filtered.

PEGylation also reduces immunogenicity (the PEG shell shields antigenic epitopes from immune recognition) and can increase aqueous solubility for otherwise hydrophobic peptides. The approach is well-established in biologics: pegfilgrastim (PEGylated G-CSF), pegylated interferons (PEG-IFN-alpha-2a, PEG-IFN-alpha-2b), and pegylated erythropoietins use this strategy to convert short-lived cytokines into drugs that can be dosed weekly or less frequently.

In modern GLP-1 receptor agonist design, fatty acid acylation with albumin binding has largely supplanted direct PEGylation for the peptide hormone class, because acylation does not occlude the receptor-binding surface as readily and because the albumin-recycling mechanism provides more predictable pharmacokinetics than PEG chain length alone. PEGylation remains relevant in other peptide drug classes and as a strategy for increasing oral bioavailability of large peptides.

Other strategies: cyclization and D-amino acids

Cyclization connects the N- and C-termini of a peptide (or two side chains) to form a ring. Cyclic peptides are more conformationally rigid than linear ones, which can improve receptor binding affinity (by reducing the entropic cost of adopting the bound conformation) and increase protease resistance (because exopeptidases cannot attack the now-absent free termini). Cyclosporin A is a long-used example. In the peptide hormone space, cyclization is used in smaller research molecules and is under investigation for oral peptide delivery.

D-amino acid substitution at a protease-sensitive cleavage site introduces a local stereochemical change that most mammalian proteases cannot accommodate.^[8]This is most useful for short, well-characterized sequences where a single cleavage site dominates degradation.

A worked comparison: native GLP-1 to semaglutide

The progression from native GLP-1 to semaglutide is the clearest worked example in the incretin drug class:

The ~5,000-fold difference in half-life between native GLP-1 and semaglutide comes from the combination of Aib⁸ blocking DPP-4 cleavage and the C18 fatty diacid promoting tight albumin binding. Neither modification alone achieves a once-weekly profile; the combination is required.^[1][2]

Putting it together

When you see a peptide drug with a long half-life, it almost always carries at least one of these modifications, often two in combination. Aib substitution addresses the enzymatic problem at a specific known cleavage site. Fatty acid acylation addresses renal clearance by borrowing albumin's size and recycling mechanism. PEGylation solves both problems by brute-force size augmentation. Each approach has trade-offs in manufacturing complexity, receptor binding interference, immunogenicity, and cost, which explains why different drugs in the same class use different combinations.