Most mechanism explainers on this site describe how a peptide turns something on. This one describes the system that is simultaneously taking it apart. Every half-life figure in the research literature — the two minutes for native GLP-1, the week for semaglutide, the thirty minutes for CJC-1295 No-DAC — is a number produced by the competition between a molecule and a set of enzymes and a kidney.
That matters for reading this catalog because a surprising fraction of the structural oddities on the shelf are not receptor-related at all. Aib at position 8. An acetyl cap on the N-terminus. A C-terminal amide. A single D-amino acid at position 2. An all-D reversed sequence. None of those changes were made to improve receptor binding. They were made to survive the peptidase system long enough to reach one.
Peptides Are Food, and the Body Treats Them That Way
A peptide bond is a chemically ordinary amide. Human physiology contains hundreds of enzymes that hydrolyse it, distributed across plasma, cell surfaces, gut lumen, brush border, lysosomes and the kidney. Their day job is nutrition and signal termination — endogenous signalling peptides need to be switched off quickly, and hydrolysis is how that is done.
Peptidases split into two families by where they cut:
- Exopeptidases trim from an end. Aminopeptidases work inward from the N-terminus, carboxypeptidases from the C-terminus. Dipeptidyl peptidases remove two residues at a time.
- Endopeptidases cut internally at a recognised sequence motif, generating fragments rather than shortening the chain residue by residue.
The practical consequence: a peptide's ends are its most exposed positions. A great deal of peptide medicinal chemistry is the defence of those two ends.
DPP-4: The Best-Characterised Off-Switch on This Shelf
Dipeptidyl peptidase-4 (DPP-4, also CD26) is a serine exopeptidase existing both as a membrane-anchored ectoenzyme on endothelium and epithelium and as a soluble circulating form. It removes an N-terminal dipeptide when the second residue is proline or alanine.
Native GLP-1 has alanine at position 8 — the second residue of GLP-1(7-36)amide. DPP-4 cleaves there, producing GLP-1(9-36)amide, which does not activate the receptor as the intact peptide does. This single cleavage is the dominant reason native GLP-1 has a plasma half-life on the order of one to two minutes. GIP carries an alanine in the equivalent position and meets the same fate.
Two independent lines of evidence confirm DPP-4 is the rate-limiting step rather than one of many. First, the gliptin drug class — inhibiting the enzyme raises intact endogenous incretin levels, an approved pharmacological validation. Second, the substitution route: replacing Ala8 with α-aminoisobutyric acid (Aib), a non-proteinogenic residue with a quaternary α-carbon that DPP-4 cannot process, removes the cleavage. Published comparisons report native GLP-1(7-36)amide degrading with a roughly 28-minute in-vitro half-life while the Aib8 analogue was undetectably degraded over six hours.
That is the Aib8 in semaglutide — and the reason the design conversation about lipidated incretins is really two separate conversations. Aib8 defeats the enzyme; the C18 fatty diacid defeats renal clearance. Neither alone gives a week-long half-life. Work on aza-substituted N-terminal analogues published in 2025 suggests the enzyme-resistance half of that problem may not be fully solved even now, with azapeptide backbones reported to block DPP-4 cleavage more completely than Aib8 does.
Neprilysin, and Why One Enzyme Is Never the Whole Story
Neprilysin (NEP, neutral endopeptidase 24.11, CD10) is a zinc metallo-endopeptidase with famously broad substrate tolerance: natriuretic peptides, substance P, enkephalins, and incretins among them. Its clinical validation runs through sacubitril, an inhibitor deployed specifically to stop the enzyme degrading natriuretic peptides.
The instructive point for peptide research is what happens when you block only DPP-4. Rodent work published in 2025 found that measuring GLP-1 secretion accurately required inhibiting both DPP-4 and neprilysin in vivo — blocking the headline enzyme still left substantial degradation running underneath it. Insulin-degrading enzyme handles insulin, glucagon and amylin. Prolyl oligopeptidase acts on short proline-containing peptides. Angiotensin-converting enzyme has its own substrate list.
So "protease-resistant" is almost always shorthand for resistant to the enzyme somebody measured. A modification validated against one peptidase says nothing about the others, and a clean in-vitro stability result in buffer with one purified enzyme is a much narrower claim than a plasma stability curve.
Reading the Modifications on This Shelf as Defences
Once you know what the enzymes recognise, several catalog structures decode immediately.
C-terminal amidation. Converting the terminal carboxylate to an amide removes the anionic handle carboxypeptidases require. On oxytocin, ipamorelin and SS-31, the amide is doing double duty — in many neuropeptides it is also required for receptor activity, so it is not purely a stability feature.
N-terminal acetylation. Capping the free α-amino group blocks aminopeptidase entry. Thymosin Alpha-1 is acetylated, as is the Ac-LKKTETQ heptapeptide sold as TB-500 and the acetyl-capped cosmetic peptides including SNAP-8.
Proline content. Proline's ring constrains the backbone and most peptidases handle it poorly. The Pro-Gly-Pro tail shared by Selank and Semax was designed as exactly this kind of stabilising chassis — with the caveat covered in our Selank vs Semax comparison that the PGP fragment is not inert once released. BPC-157 is likewise proline-rich, which is part of why its parent sequence survives a gastric environment at all.
D-amino acids. Peptidase active sites are chiral. Substituting a single D-residue at a vulnerable position blocks cleavage there; building the entire peptide from D-residues makes it broadly resistant. Dermorphin carries D-Ala at position 2 — the classic exopeptidase-blocking placement. SS-31 opens with D-Arg. FOXO4-DRI is all-D throughout, which is the point of the D-retro-inverso design.
That last category carries a quality-control consequence worth restating: D and L residues are isobaric, so mass spectrometry confirms composition and is blind to chirality. A peptide whose entire stability strategy is stereochemical cannot have that strategy verified by the standard identity line on a COA. See reading a chromatogram and mass spectrum and our /quality/ overview.
The Kidney Is the Other Clearance Route — and Often the Bigger One
Suppose a peptide is fully protease-resistant. It can still disappear in minutes, because the second clearance system is physical rather than enzymatic.
Glomerular filtration passes circulating molecules below roughly 60 kDa, with charge selectivity layered on top. Almost every compound on this shelf is one to ten kilodaltons — well under the cutoff and freely filtered. Filtered peptides are then captured in the proximal tubule by the endocytic receptors megalin and cubilin, internalised via clathrin-dependent uptake, and delivered to lysosomes for hydrolysis. The peptide is not excreted intact so much as reabsorbed and digested.
This reframes the half-life-extension strategies from our PK primer. Albumin binding — the DAC linker on CJC-1295 DAC, the fatty diacid on the incretins — works largely because albumin at ~66 kDa sits at the filtration threshold, so a bound peptide travels with a carrier too large to filter efficiently. PEGylation, as on PEG-MGF, increases hydrodynamic radius toward the same end. Both are anti-filtration strategies more than anti-protease strategies.
The mirror image appears in IGF-1 LR3, where engineering away from a binding-protein reservoir produced measurably faster clearance in rats — the same physics running in reverse.
Why Oral Routes Fail, in One Paragraph
The gut is the peptidase system at maximum intensity: gastric pepsin, pancreatic trypsin, chymotrypsin and elastase in the lumen, then a dense brush-border layer of aminopeptidases and dipeptidyl peptidases on the enterocyte surface — with DPP-4 itself abundantly expressed there. A peptide surviving all of that still faces a tight epithelial barrier it is too large and too polar to cross. This is why oral peptide bioavailability is conventionally reported in fractions of a percent, why oral semaglutide requires a co-formulated absorption enhancer, and why the orally active compounds discussed in the growth-hormone axis roundup tend to be small molecules rather than peptides.
FAQ
Does protease resistance mean a long half-life? No. It removes one clearance route. Renal filtration, receptor-mediated internalisation and hepatic uptake are all still running. A stereochemically bulletproof peptide of two kilodaltons with no carrier binding can still be cleared in minutes.
Can a stability claim be checked on a COA? Not directly. A COA characterises the vial's contents at the time of testing — identity, purity, mass accounting, moisture, sometimes endotoxin. Metabolic stability is a property of the molecule in plasma, established in the literature rather than on a certificate. What a COA can confirm is that the stabilising modification is actually present — that an acetyl group adds its +42 Da to the theoretical mass, that an amide reads one dalton below the free acid. See reading a peptide COA.
Why do published half-lives for the same peptide differ so much? Because they measure different things: in-vitro incubation with a purified enzyme, ex-vivo plasma stability, and in-vivo terminal half-life are three distinct experiments, and species differ in peptidase expression. Numbers are only comparable within the same method — the same rule that applies to potency values and HPLC purity percentages.
Browse compound profiles in the research library.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.