Before most people searched “retatrutide,” the molecule had a code on an Eli Lilly badge: LY3437943. In 2022, Tamer Coskun and colleagues at Lilly published that peptide in Cell Metabolism as a single-chain agonist built to engage three receptors at once: the glucagon receptor (GCGR), the glucose-dependent insulinotropic polypeptide receptor (GIPR), and the glucagon-like peptide-1 receptor (GLP-1R).
That is the discovery story that matters for chemistry. The later clinical program is a different literature. This article stays with the molecule: how it was described, what “triple agonist” means at the receptor, and what structural work has since shown about how one peptide fits three pockets.
Retatrutide is the nonproprietary name associated with LY3437943. Chemical registries list CAS 2381089-83-2. It is investigational as a medicine. It is not FDA-approved. A research reagent with this identity is not a Lilly drug product.
The 2022 paper: a GIP backbone with three jobs
Coskun, Urva, Roell, Qu, and co-authors titled the report in full: “LY3437943, a novel triple glucagon, GIP, and GLP-1 receptor agonist for glycemic control and weight loss: From discovery to clinical proof of concept.” Cell Metabolism, volume 34, issue 9, 6 September 2022, pages 1234-1247.e9 (doi:10.1016/j.cmet.2022.07.013).
The chemical design they state is specific enough to quote in substance. LY3437943 is a fatty-acid-acylated single peptide. The sequence is engineered from a GIP peptide backbone, then modified so the same chain also signals at GLP-1R and GCGR. In vitro, they report balanced GCGR and GLP-1R activity but more GIPR activity.
That last sentence is the opposite of a marketing slogan. It is a potency ranking versus the three human receptors, measured in assay systems in that paper. Later reviews restated the comparison to the endogenous hormones this way: more potent at GIPR (by a factor of 8.9) and less potent at GCGR and GLP-1R (factors of 0.3 and 0.4, respectively), citing Coskun. Those ratios are in vitro pharmacology of the ligand, not instructions for anything.
Fatty-acid acylation is the same family of trick used on other long-acting incretin peptides: a lipid chain that supports reversible binding to albumin, which is how a peptide that would otherwise be cleared quickly is given a longer circulating residence time. Coskun’s paper also includes mouse pharmacology and a phase 1 single-ascending-dose study. Those sections exist. They are not the subject here.
What “triple agonist” means, chemically
Unimolecular polypharmacology is the unlovely name for the idea. Instead of three drugs, one peptide is tuned so that it is an agonist at three class B GPCRs:
GIPR, the receptor for glucose-dependent insulinotropic polypeptide
GLP-1R, the receptor for glucagon-like peptide-1
GCGR, the receptor for glucagon
Semaglutide-class ligands are GLP-1R agonists. Tirzepatide (LY3298176 in the Lilly series) is a dual GIPR and GLP-1R agonist. Retatrutide adds GCGR as the third arm. That is a receptor-identity statement. It is not a comparison of outcomes, and it is not a shopping guide.
The three receptors are related class B GPCRs. GLP-1 and GIP are incretin hormones. Glucagon is the counter-regulatory hormone at GCGR. They are not the same protein, and they do not have identical extracellular loops. A unimolecular agonist has to satisfy three binding sites with one sequence. That is a harder chemistry problem than making three separate peptides, which is the point of the Lilly design.
The endogenous hormones already share related N-terminal chemistry. The engineering problem is to keep enough of that shared recognition to activate all three receptors, while the middle of the sequence absorbs the differences between the pockets. That is a sequence-design problem, which is why the Coskun paper bothers to say “GIP backbone” instead of “incretin-like peptide” and stopping.
Li and colleagues make the same point from structures. In their sequence comparison, the N-terminal stretch around residues 4-9 is the conserved part of this hormone family. Variation clusters in the middle of the chain (they highlight residues 10-21), which is also where ECL1, ECL2, and the top of TM1 read the peptide. They contrast retatrutide’s GIP-backbone strategy with another triple agonist (peptide 20 / MAR423) built from a glucagon backbone. Two ways to get to three receptors. Same class of idea, different starting sequence.
Identity table
Field
As listed in public chemistry sources
Nonproprietary name
Retatrutide
Developer code
LY3437943 (also written LY-3437943)
CAS
2381089-83-2
UNII
NOP2Y096GV
ChEMBL
CHEMBL5095485
DrugBank
DB18993
Class
Synthetic, lipidated peptide; agonist at GIPR, GLP-1R, and GCGR
Medicine status
Investigational. Not FDA-approved.
PubChem indexes the name, CAS, and synonyms (including LY3437943). Formula and exact mass can differ across database records depending on how the fatty-acid moiety and salt form are counted. For a reagent lot, the mass that matters is the one on that lot’s mass spectrum, matched to the sequence the lab claims, not a rounded number copied from a blog.
The code versus the name is ordinary pharmaceutical housekeeping. LY3437943 is the internal Lilly identifier from discovery. Retatrutide is the nonproprietary name that later attached to that code. Informal labels such as “triple G” show up in secondary write-ups because of the three receptors (GIP, GLP-1, glucagon). Informal labels are not CAS numbers. If a COA or a paper uses only a nickname, you still need LY3437943, retatrutide, or 2381089-83-2 to know you are in the same file.
2024: three cryo-EM complexes
In 2024, Wenzhuo Li, Qingtong Zhou, Zhaotong Cong, and colleagues reported cryo-EM structures of retatrutide bound to GLP-1R, GIPR, and GCGR, each in complex with Gs (Cell Discovery). The maps were determined at overall resolutions of 2.68 Å (GLP-1R), 3.26 Å (GIPR), and 2.84 Å (GCGR). Coordinates went to the Protein Data Bank as 8YW3, 8YW4, and 8YW5.
The structural picture is a single continuous helix. The N-terminal stretch (residues 1-13 in their numbering) goes into the transmembrane core. A more C-terminal stretch (residues 14-30) contacts the extracellular domain region, the top of TM1, and extracellular loop 1. The three complexes are similar overall. The interesting chemistry is where they are not: ECL1 in GIPR is a flexible loop (they note three prolines), while ECL1 in GLP-1R and GCGR forms a short helix. That is a physical reason a GIP-backbone peptide can still be taught to satisfy GLP-1R and GCGR. The middle of the sequence is where the receptors disagree.
They also place the fatty-diacid acylation at lysine 17 for retatrutide, compared with position 20 on semaglutide and tirzepatide in their sequence figure. Semaglutide and tirzepatide are included there as structural comparators, not as a clinical ranking. The covalent lipid on retatrutide is a map of where the chain is modified. It is not a schedule.
Non-natural residues show up in that same comparison figure: α-aminoisobutyric acid (Aib) and α-methyl leucine. Those substitutions are a standard peptide-engineering move against peptidases and for conformational control. They are part of why the chain is not simply “GIP with extra letters.” Coskun’s “engineered from a GIP peptide backbone” and Li’s residue-level contacts are the same object described at two resolutions: design intent, then a picture of the complex.
What the cryo-EM work does not do is turn a research lot into the Lilly clinical candidate. PDB entries 8YW3, 8YW4, and 8YW5 are structures of defined complexes. A lyophilized research vial is a chemical sample. Identity still has to be shown for that lot (mass, HPLC, the usual COA pair). For how those documents work, see how to read a peptide COA.
What the discovery record does not make this molecule
A published agonist is not an approved drug. Retatrutide remains investigational. Research-grade material offered for laboratory work is not interchangeable with a finished pharmaceutical, even when the amino-acid sequence is meant to match the literature compound.
This page does not recap trial percentages, does not rank retatrutide against tirzepatide for any human endpoint, and does not tell anyone what to do with a vial. For CAS, molecular profile, and the encyclopedia-style reference page, use the Retatrutide (LY3437943) research overview.
References
Coskun T, Urva S, Roell WC, et al. LY3437943, a novel triple glucagon, GIP, and GLP-1 receptor agonist for glycemic control and weight loss: from discovery to clinical proof of concept. Cell Metab. 2022;34(9):1234-1247.e9. doi:10.1016/j.cmet.2022.07.013.
Li W, Zhou Q, Cong Z, et al. Structural insights into the triple agonism at GLP-1R, GIPR and GCGR manifested by retatrutide. Cell Discov. 2024. doi:10.1038/s41421-024-00700-0. PDB 8YW3, 8YW4, 8YW5.
PubChem compound record for retatrutide (CAS 2381089-83-2; synonyms LY3437943, LY-3437943).
IUPHAR/BPS Guide to PHARMACOLOGY, ligand retatrutide (LY3437943).
For laboratory research only. Not for human or veterinary use. Not intended to diagnose, treat, cure, or prevent any disease.
The Discovery of Retatrutide (LY3437943)
Before most people searched “retatrutide,” the molecule had a code on an Eli Lilly badge: LY3437943. In 2022, Tamer Coskun and colleagues at Lilly published that peptide in Cell Metabolism as a single-chain agonist built to engage three receptors at once: the glucagon receptor (GCGR), the glucose-dependent insulinotropic polypeptide receptor (GIPR), and the glucagon-like peptide-1 receptor (GLP-1R).
That is the discovery story that matters for chemistry. The later clinical program is a different literature. This article stays with the molecule: how it was described, what “triple agonist” means at the receptor, and what structural work has since shown about how one peptide fits three pockets.
Retatrutide is the nonproprietary name associated with LY3437943. Chemical registries list CAS 2381089-83-2. It is investigational as a medicine. It is not FDA-approved. A research reagent with this identity is not a Lilly drug product.
The 2022 paper: a GIP backbone with three jobs
Coskun, Urva, Roell, Qu, and co-authors titled the report in full: “LY3437943, a novel triple glucagon, GIP, and GLP-1 receptor agonist for glycemic control and weight loss: From discovery to clinical proof of concept.” Cell Metabolism, volume 34, issue 9, 6 September 2022, pages 1234-1247.e9 (doi:10.1016/j.cmet.2022.07.013).
The chemical design they state is specific enough to quote in substance. LY3437943 is a fatty-acid-acylated single peptide. The sequence is engineered from a GIP peptide backbone, then modified so the same chain also signals at GLP-1R and GCGR. In vitro, they report balanced GCGR and GLP-1R activity but more GIPR activity.
That last sentence is the opposite of a marketing slogan. It is a potency ranking versus the three human receptors, measured in assay systems in that paper. Later reviews restated the comparison to the endogenous hormones this way: more potent at GIPR (by a factor of 8.9) and less potent at GCGR and GLP-1R (factors of 0.3 and 0.4, respectively), citing Coskun. Those ratios are in vitro pharmacology of the ligand, not instructions for anything.
Fatty-acid acylation is the same family of trick used on other long-acting incretin peptides: a lipid chain that supports reversible binding to albumin, which is how a peptide that would otherwise be cleared quickly is given a longer circulating residence time. Coskun’s paper also includes mouse pharmacology and a phase 1 single-ascending-dose study. Those sections exist. They are not the subject here.
What “triple agonist” means, chemically
Unimolecular polypharmacology is the unlovely name for the idea. Instead of three drugs, one peptide is tuned so that it is an agonist at three class B GPCRs:
Semaglutide-class ligands are GLP-1R agonists. Tirzepatide (LY3298176 in the Lilly series) is a dual GIPR and GLP-1R agonist. Retatrutide adds GCGR as the third arm. That is a receptor-identity statement. It is not a comparison of outcomes, and it is not a shopping guide.
The three receptors are related class B GPCRs. GLP-1 and GIP are incretin hormones. Glucagon is the counter-regulatory hormone at GCGR. They are not the same protein, and they do not have identical extracellular loops. A unimolecular agonist has to satisfy three binding sites with one sequence. That is a harder chemistry problem than making three separate peptides, which is the point of the Lilly design.
The endogenous hormones already share related N-terminal chemistry. The engineering problem is to keep enough of that shared recognition to activate all three receptors, while the middle of the sequence absorbs the differences between the pockets. That is a sequence-design problem, which is why the Coskun paper bothers to say “GIP backbone” instead of “incretin-like peptide” and stopping.
Li and colleagues make the same point from structures. In their sequence comparison, the N-terminal stretch around residues 4-9 is the conserved part of this hormone family. Variation clusters in the middle of the chain (they highlight residues 10-21), which is also where ECL1, ECL2, and the top of TM1 read the peptide. They contrast retatrutide’s GIP-backbone strategy with another triple agonist (peptide 20 / MAR423) built from a glucagon backbone. Two ways to get to three receptors. Same class of idea, different starting sequence.
Identity table
PubChem indexes the name, CAS, and synonyms (including LY3437943). Formula and exact mass can differ across database records depending on how the fatty-acid moiety and salt form are counted. For a reagent lot, the mass that matters is the one on that lot’s mass spectrum, matched to the sequence the lab claims, not a rounded number copied from a blog.
The code versus the name is ordinary pharmaceutical housekeeping. LY3437943 is the internal Lilly identifier from discovery. Retatrutide is the nonproprietary name that later attached to that code. Informal labels such as “triple G” show up in secondary write-ups because of the three receptors (GIP, GLP-1, glucagon). Informal labels are not CAS numbers. If a COA or a paper uses only a nickname, you still need LY3437943, retatrutide, or 2381089-83-2 to know you are in the same file.
2024: three cryo-EM complexes
In 2024, Wenzhuo Li, Qingtong Zhou, Zhaotong Cong, and colleagues reported cryo-EM structures of retatrutide bound to GLP-1R, GIPR, and GCGR, each in complex with Gs (Cell Discovery). The maps were determined at overall resolutions of 2.68 Å (GLP-1R), 3.26 Å (GIPR), and 2.84 Å (GCGR). Coordinates went to the Protein Data Bank as 8YW3, 8YW4, and 8YW5.
The structural picture is a single continuous helix. The N-terminal stretch (residues 1-13 in their numbering) goes into the transmembrane core. A more C-terminal stretch (residues 14-30) contacts the extracellular domain region, the top of TM1, and extracellular loop 1. The three complexes are similar overall. The interesting chemistry is where they are not: ECL1 in GIPR is a flexible loop (they note three prolines), while ECL1 in GLP-1R and GCGR forms a short helix. That is a physical reason a GIP-backbone peptide can still be taught to satisfy GLP-1R and GCGR. The middle of the sequence is where the receptors disagree.
They also place the fatty-diacid acylation at lysine 17 for retatrutide, compared with position 20 on semaglutide and tirzepatide in their sequence figure. Semaglutide and tirzepatide are included there as structural comparators, not as a clinical ranking. The covalent lipid on retatrutide is a map of where the chain is modified. It is not a schedule.
Non-natural residues show up in that same comparison figure: α-aminoisobutyric acid (Aib) and α-methyl leucine. Those substitutions are a standard peptide-engineering move against peptidases and for conformational control. They are part of why the chain is not simply “GIP with extra letters.” Coskun’s “engineered from a GIP peptide backbone” and Li’s residue-level contacts are the same object described at two resolutions: design intent, then a picture of the complex.
What the cryo-EM work does not do is turn a research lot into the Lilly clinical candidate. PDB entries 8YW3, 8YW4, and 8YW5 are structures of defined complexes. A lyophilized research vial is a chemical sample. Identity still has to be shown for that lot (mass, HPLC, the usual COA pair). For how those documents work, see how to read a peptide COA.
What the discovery record does not make this molecule
A published agonist is not an approved drug. Retatrutide remains investigational. Research-grade material offered for laboratory work is not interchangeable with a finished pharmaceutical, even when the amino-acid sequence is meant to match the literature compound.
This page does not recap trial percentages, does not rank retatrutide against tirzepatide for any human endpoint, and does not tell anyone what to do with a vial. For CAS, molecular profile, and the encyclopedia-style reference page, use the Retatrutide (LY3437943) research overview.
References
For laboratory research only. Not for human or veterinary use. Not intended to diagnose, treat, cure, or prevent any disease.
Last reviewed: 2026-10-14