Semaglutide is a synthetic 31-residue analog of glucagon-like peptide-1 (GLP-1), built by making three deliberate changes to the native hormone's sequence and chemistry. It is one of the most thoroughly characterized peptides in the research supply chain — its receptor has been solved by cryo-electron microscopy with semaglutide bound in the pocket [5], the DPP-4 degradation chemistry it is engineered around was published in 1998, seventeen years before the paper describing its discovery [1, 2], and its impurity profile has been mapped to the level of single stereocenters [10]. That depth of characterization is exactly why it is a common reference ligand in GLP-1 receptor (GLP-1R) assay work.

This page describes the molecule, the receptor it binds, the assay systems it is used in, and the analytical questions that decide whether a vial of white powder is actually semaglutide. It is written for laboratory context. Echo Peptides supplies research chemicals for in vitro and preclinical work only.

Verified identity data

These values are published on the product page and come from our verified chemistry map, not from a supplier's label:

The average mass computed from that formula matches the quoted molecular weight; a build-time check on this site recomputes it from the formula and refuses to publish a mismatch. At roughly 4.1 kDa, semaglutide sits well above the mass range where a simple molecular ion tells you much on its own — identity confirmation for a peptide this size requires sequence-level evidence, not a single mass reading.

Three changes to the GLP-1 backbone

Native GLP-1 is a poor experimental tool because it disappears. The design work behind semaglutide addressed that in two independent ways, and a third change bears on the chemistry of the second.

Aib at position 8. Dipeptidyl peptidase IV (DPP-4) cleaves GLP-1 near its N-terminus, and the fix was published in 1998: analogs substituted at the N-terminal region with threonine, glycine, serine or α-aminoisobutyric acid (Aib) were all more resistant to DPP-4 in porcine plasma in vitro, with the Aib8 analog showing no detectable degradation after six hours against a half-life of 28 minutes for GLP-1(7-36) amide [1]. The Aib8 analog also retained receptor binding close to the native peptide — an IC₅₀ of 0.45 nmol/l against 0.78 nmol/l for GLP-1(7-36) amide in a cell line expressing the cloned receptor — which is the part that mattered, since only the Aib8 and Gly8 analogs held affinity similar to the native peptide, while the serine- and threonine-substituted analogs did not [1].

An acyl chain at lysine 26. Metabolic stability alone does not give a peptide a long circulating life. Semaglutide is derivatized at lysine 26 with a fatty acid moiety attached through a designed linker, and the stated aim of that program was to increase albumin affinity and secure full stability against metabolic degradation; the fatty acid moiety and the linking chemistry to GLP-1 are named as the key features for securing high albumin affinity and GLP-1R potency and for obtaining prolonged exposure and action [2]. The design review makes the same point about the chemistry: reversible albumin binding was the protraction strategy, with optimal fatty acid and linker combinations identified to maximize albumin binding while maintaining GLP-1R potency [3].

Arg at position 34. Semaglutide's two amino acid substitutions relative to human GLP-1 are Aib8 and Arg34 [2]. GLP-1 carries lysines at both position 26 and position 34; replacing the second one leaves a single lysine available to be acylated. The discovery paper states the substitution without spelling out that rationale, so treat the reasoning as chemistry, not as the authors' claim.

The measured result: GLP-1R affinity of 0.38 ± 0.06 nM — about three-fold lower than liraglutide's — with albumin affinity increased, and a plasma half-life of 46.1 hours in mini-pigs [2]. Note the trade the numbers record. Receptor affinity went down slightly and the molecule still became far longer-lived, because the limiting factor was never affinity.

Two numbering systems, and why they look contradictory

Positions in this molecule are quoted two different ways in the literature, and the mismatch trips people up. Design papers use GLP-1(7-37) numbering, which is why the substitutions are called Aib8 and Arg34. Analytical papers usually number semaglutide's own chain from 1 to 31, where residue 1 is histidine. The offset is six. So an analytical report describing a "D-Ser8" impurity is not contradicting "Aib8" — it is describing the serine at GLP-1 position 14. Any comparison of two documents about this peptide should start by establishing which convention each one uses.

The receptor

GLP-1R is a class B G-protein-coupled receptor that signals primarily through the stimulatory G protein Gs. The cryo-EM structure of the peptide-activated GLP-1R–Gs complex showed the peptide clasped between the receptor's N-terminal domain and its transmembrane core and further stabilized by extracellular loops, with a sharp kink in the middle of transmembrane helix 6 pivoting its intracellular half outward to accommodate the α5 helix of Gs [4]. That is the general activation mechanism for the class.

Semaglutide has been resolved in that pocket directly. Cryo-EM structures and 3D variability analysis of semaglutide- and taspoglutide-bound GLP-1R–Gs complexes showed peptide interactions similar to GLP-1 but different motions within the receptor and within the bound peptides [5]. The word doing the work there is motions: static contacts can look alike while the dynamics differ, and the authors present this as insight into the molecular determinants of peptide engagement rather than as an explanation of any downstream outcome.

Where the receptor is

Receptor distribution is a mapping question, and the map was redrawn once the antibody reagents improved. Using an extensively validated monoclonal antibody in monkey and human tissue, GLP-1R was localized predominantly to pancreatic β-cells, with markedly weaker expression in acinar cells and none in pancreatic ductal epithelial cells; in kidney and lung it was found exclusively in smooth muscle cells of arteries and arterioles; in heart, in myocytes of the sinoatrial node; and in the gastrointestinal tract, most strongly in Brunner's gland of the duodenum, with lower expression in parietal cells, in smooth muscle of the muscularis externa, and in myenteric plexus neurons throughout the gut [6]. No GLP-1R was seen in primate liver or thyroid, and the immunohistochemistry was backed by in situ ligand binding with ¹²⁵I-GLP-1 [6]. Reviews summarize GLP-1R expression across pancreas, gastrointestinal tract, heart, lungs, kidneys and brain [3].

The phrase "extensively validated" in that paper's own title is not decoration, and neither is the word "novel": the distribution described above is what emerged once the reagent itself was validated and cross-checked against ligand binding [6]. Receptor-expression claims that rest on an unvalidated antibody should be read as provisional until the reagent is characterized.

What semaglutide reaches in rodent brain

A rodent imaging and transcriptomic study found that semaglutide directly accessed the brainstem, septal nucleus and hypothalamus but did not cross the blood-brain barrier; access occurred through the circumventricular organs and several select sites adjacent to the ventricles [7]. Central c-Fos activity appeared in ten brain areas, including hindbrain regions the peptide reached directly and secondary areas without direct GLP-1R interaction, such as the lateral parabrachial nucleus, and microdissected areas from treated rats showed upregulation of prolactin-releasing hormone and tyrosine hydroxylase in the area postrema [7]. The distinction between direct access and downstream activation is the useful part for experimental design: detecting a response in a region does not establish that the compound was present there.

Signaling, bias and receptor trafficking

GLP-1R agonists are not interchangeable at the level of what the receptor does after binding. Beyond cAMP generation through Gs, the receptor recruits β-arrestin and undergoes agonist-mediated endocytosis, and different ligands drive those processes to different degrees.

In a study of biased GLP-1R agonists with varying propensities for internalization and recycling, compounds that retained GLP-1R at the plasma membrane produced greater long-term insulin release in comparison with a panel of approved GLP-1 mimetics, and that effect depended on reduced β-arrestin recruitment and faster agonist dissociation rates [8]. A later program building long-acting G protein-biased peptides used semaglutide explicitly as the benchmark their compounds had to match on acute signaling potency, and reported that very substantial reductions in β-arrestin recruitment efficacy — not moderate ones — were required to realize fully the benefits of GLP-1R agonism on blood glucose lowering in mice [9]. The same work found that its lead compound's cAMP signaling was differentially modulated by single and double GLP1R coding variants seen in human populations [9]. That last result was measured for their compound, not for semaglutide; whether semaglutide's cAMP output is similarly variant-sensitive is a separate question that study does not answer.

The practical implication for assay design: a single cAMP EC₅₀ is an incomplete description of a GLP-1R ligand. Bias between cAMP and β-arrestin, dissociation kinetics, and internalization versus membrane retention are separable readouts, and semaglutide's position on each is a measurement, not an assumption.

What a purity number does not cover

A certificate reading "99%" is a chromatographic statement about one method's ability to resolve one class of impurity. It is not a statement about stereochemistry. Characterization work on semaglutide using reverse-phase UPLC followed by high-resolution tandem mass spectrometry, with off-line collection, deuterated hydrochloric acid hydrolysis and chiral derivatization, identified low-level D-Ser8, D-His1 and D-Asp9 isomeric impurities — species with the same molecular weight as the target peptide [10]. Under the FDA guidance the authors were working to, each peptide-related impurity at 0.10% of the drug substance or greater has to be identified [10]. D-amino acid isomers are the hardest class to see precisely because mass spectrometry alone cannot distinguish them.

A separate analysis compared follow-on and compounded semaglutide and liraglutide samples against originator products using MHC-II-associated peptide proteomics, LC-MS, photostability testing and a fibrillation assay. The follow-on drug substance and the follow-on and compounded products carried distinct impurity profiles — amino acid deletions and additions, plus unidentified impurities — versus the originators, and potentially immunogenic peptides were presented on impurity-stimulated monocyte-derived dendritic cells from healthy donors [11]. The light-exposure result is narrower than that: the significant disparity in strength, impurity sum and high-molecular-weight protein level was reported between compounded semaglutide and originator products specifically [11]. Two caveats belong with that finding. The authors are employees and shareholders of the originator manufacturer, disclosed in the paper. And the assay measures immunogenic potential in a cell-based system; the paper's own conclusion calls for immunogenicity studies rather than declaring the question settled [11].

What the photostability testing supports is narrower than that: after light exposure, compounded semaglutide and the originator products differed measurably in strength, impurity sum and high-molecular-weight protein level [11]. Light exposure is therefore a handling variable worth controlling for this material, and nothing in that comparison licenses treating a stated purity figure as a prediction of how a given vial behaves under it.

What the market data show

The strongest reason to care about identity testing on this particular compound is empirical. A 2024 market surveillance study purchased semaglutide from illegal online pharmacies selling without prescription and ran visual inspection, sterility and endotoxin testing, and LC-MS quantitation. Of six attempted purchases, three vials were delivered and the other three orders were never sent at all — non-delivery e-commerce scams [12]. In the delivered vials, measured semaglutide purity ranged from 7.7% to 14.37% against 99% claimed on the labels; content exceeded the labeled amount by 28.56% to 38.69%; the lyophilized samples were free of viable microorganisms at testing, but endotoxin was detected in all of them, from 2.1645 to 8.9511 EU/mg [12].

Read those two findings together: the material was both overstated in purity by roughly an order of magnitude and overfilled relative to label. A quantitation assay alone would not have caught the first problem, and a purity assay alone would not have caught the second. This is the argument for batch-level documentation that reports both, from a laboratory that is not the seller.

What is not known

Several limits are worth stating plainly.

  • Species differences in receptor expression are real and have been measured. GLP-1R was localized to rodent thyroid C-cells, where GLP-1R agonists stimulated calcitonin release, upregulated calcitonin gene expression and produced C-cell hyperplasia in rats and to a lesser extent in mice; in contrast, humans and/or cynomolgus monkeys had low GLP-1R expression in thyroid C-cells, and GLP-1R agonists did not activate adenylate cyclase or generate calcitonin release in primates [13]. The authors' own conclusion is that species-specific differences in receptor expression and action exist in the thyroid and that long-term consequences of sustained GLP-1R activation in the human thyroid remain unknown [13]. A rodent result at this receptor does not transfer by default.
  • Much of the foundational pharmacology comes from one manufacturer's laboratories. The discovery paper, the design review, the receptor localization study, the rodent brain-access study and the impurity comparison above all carry originator-company affiliations or shareholdings in their disclosure statements [2, 3, 6, 7, 11]. The work is not thereby wrong — it is generally the most detailed available — but independent replication is thinner than the volume of literature suggests.
  • Structural data describe complexes, not consequences. The semaglutide-bound structures characterize receptor and peptide dynamics in a purified GLP-1R–Gs complex [5]. Mapping those motions onto cellular signaling outcomes is an open inferential step.
  • Sequence-level identity is rarely what a certificate proves. Unless a document reports the method — chiral analysis, peptide mapping, high-resolution MS/MS — a purity percentage tells you how much of the sample eluted in one peak under one gradient [10].

Bottom line

Semaglutide is a rationally designed GLP-1 analog. Two of its three changes address documented limitations of the native hormone: Aib at position 8 confers resistance to DPP-4 cleavage [1], and derivatization at lysine 26 supplies the reversible albumin binding the design program used to obtain prolonged exposure [2, 3]. The Arg34 substitution is reported without a stated rationale; what it does chemically is leave a single lysine available for acylation [2]. Its receptor is structurally solved with semaglutide bound [5], its tissue distribution is mapped with validated reagents [6], and its signaling is separable into cAMP, arrestin and trafficking readouts that behave differently across ligands [8, 9]. Against that, the analytical literature is unusually blunt about how easily material sold under this name fails to be this molecule [10, 11, 12]. For research use, the identity data on the vial and the batch documentation behind it are not paperwork around the science — for a 4.1 kDa peptide with same-mass stereoisomeric impurities, they are part of it.

Research use only. Not for human or veterinary use. Nothing on this page describes an application in people or animals.