
This article covers receptor binding and activation only; the site’s existing cagrilintide mechanism article and the cagrilintide-versus-eloralintide comparison cover indications, clinical comparisons and efficacy rankings, none of which is repeated here. Cagrilintide engages the calcitonin receptor (CTR) core in essentially the same geometry at AMY1R, AMY2R, AMY3R and CTR: a mid-region loop arches over the transmembrane (TM) bundle, proline 37 seats in a hydrophobic pocket on the CTR extracellular domain (ECD), and limited density in the AMY1R structure places the lipidated N-terminal linker between ECL2 and TM6, with a short acyl segment merging into the detergent micelle[1]. The principal contacts described here involve the CTR protomer, and P37 makes no direct contact with RAMP1 or RAMP3. That structural observation does not by itself explain activation: the evidence that one peptide raises Gs/cAMP at all four complexes comes from separate functional assays, and a shared contact map is not proof of identical activation at each subtype.
The receptor cagrilintide actually contacts: CTR plus a RAMP
The three AMY receptors share the CTR protomer and differ by RAMP subunit — RAMP1 for AMY1R, RAMP2 for AMY2R, RAMP3 for AMY3R[1][2]. In AMY1R structures, F83 of RAMP1 packs over R126 of the CTR ECD and stabilises the RAMP1–CTR interface, whether the bound peptide is cagrilintide or rat amylin[1]. The peptide’s N-terminal helix sits in the class B pocket lined by TM3/TM5/ECL2 on one side and TM6/TM7/ECL3 on the other — the region mapped for rat amylin and CGRP at AMY1R[3] — and the cagrilintide backbone in the AMYR complexes closely matches the rat amylin-bound backbone[1].
The RAMP therefore shapes the receptor rather than forming most of the peptide’s footprint. It influences the ECD position and its dynamics, including how the interface responds to the peptide’s C-terminus. These structures locate the principal contacts on CTR; they do not quantify the binding energy contributed by each receptor component[1].
N-terminal lipidation: what the structures resolve
Cagrilintide carries an acyl chain on K1. In the AMYR cryo-EM maps that modification is poorly resolved: only limited density for the γGlu linker and a very short acyl segment appears, in the AMY1R complex. The linker exits the TM bundle between H296^ECL2 and W361^6.58, and the acyl segment merges into the detergent micelle beside a conserved lipid density that is also present in the rat amylin–AMY1R complex[1].
The non-lipidated backbone (CagriBB) shows what the lipid does and does not do. Once the peptide is stably engaged at AMYRs, removing the lipid changes neither peptide nor receptor conformation, apart from the K1 rotamer[1]. Functionally, CagriBB matched cagrilintide’s potency and maximal response in CRE-reporter cells co-expressing CTR and RAMP3, but was less potent in cells expressing CTR alone. The H296^ECL2 rotamer and the more open ECL3 adopted with the lipidated peptide are proposed explanations, so lipidation may help stabilise an active CTR conformation[1]. A second structural report also discusses N-terminal lipid modification as a route to extended half-life in this peptide class[4]. Neither observation establishes lipidation’s quantitative contribution to equilibrium affinity at each receptor.
One number needs care. The complexes were assembled with peptide at 10 µM[1]. That is a structural-biology condition, not a potency, affinity or exposure value.
The mid-region bypass motif and the E14–R17 salt bridge

Residues S19–P25 form the bypass motif, held by intramolecular hydrogen bonds between the L16 and F23 backbone and S20, and between the P25 backbone and N21. The F23 phenyl ring makes extensive van der Waals contact with L16, Y149^1.40 and Y146^1.37, anchoring the segment at the level of the CTR TM bundle and the micelle interface[1]. These conserved interactions support the bypass motif, a structural determinant associated with high potency of amylin analogues at AMYRs.
The E14–R17 pair distinguishes cagrilintide from rat amylin. N14E and V17R were introduced to counter the fibril formation seen with human amylin and were expected to stabilise the N-terminal helix through an intramolecular salt bridge. In the active-state AMYR structures, E14 adopts a downward rotamer that hydrogen-bonds and forms an electrostatic contact with R17 on the same helix face, close to ECL2. Rat amylin carries asparagine at the equivalent position, pointing up toward solvent and hydrogen-bonding the L291^ECL2 backbone. The larger E14 and R17 side chains push the helix away from ECL2 and slightly expand the peptide pocket on the extracellular side of the TM domain[1].
A conformational difference appears at CTR. Rat amylin bound to CTR-Gs splits between a bypass state (55% of particles) and a CT-like state (45%) in which the bypass motif becomes disordered and the ECD shifts. Cagrilintide holds 85% of particles in the bypass conformation[1]. This conformation accompanies opening of the extracellular TM bundle, with outward movement of TM1, ECL1/TM3 and TM4/ECL2 relative to rat amylin. These particle fractions describe the analysed structural preparations, not receptor-state populations measured in living tissue.
Proline 37 and the extracellular domain

At all three AMYRs, P37 occupies the same pocket above W79^ECD as Y37 of rat amylin, with backbone hydrogen bonds to S129^ECD and hydrophobic contacts to D77, G78, W79, Y131 and W128[1]. The difference is size. Proline’s small side chain is completely buried by CTR ECD residues and makes no direct contact with RAMP1 or RAMP3, consistent with cagrilintide’s ability to bind the CTR protomer on its own[1].
The missing RAMP contact has subtype consequences. At AMY1R, F83^RAMP1 still packs over R126^ECD and the interface holds. At AMY3R, rat amylin’s Y37 hydrogen-bonds E74^RAMP3 and holds the RAMP3 ECD in place; without an equivalent contact the RAMP3 ECD moves away from P37 and from the CTR ECD, the RAMP3–CTR interface appears weaker, and both ECDs show considerable motion — the maps for that interface were low resolution, so those details are indicative rather than settled[1]. ECD motion also separates peptides at AMY1R: CGRP produces 3.4 Å of ECD displacement in variance analysis against 0.8 Å for rat amylin[3], and the P37 change alters cagrilintide-bound ECD dynamics in a subtype-dependent way[1].
From binding to Gs/cAMP activation
All four cagrilintide structures were solved as active, Gs-coupled complexes, placing the binding contacts described above within an activated receptor–transducer assembly[1]. Functional evidence includes CRE-reporter responses in transfected cells and Gs activation assays[1][4]. A CRE reporter is a downstream readout, distinct from directly measuring cAMP accumulation. An earlier pharmacological comparison profiled the same peptide, as AM833, across 25 endpoints covering binding, activation and regulation, and classified it as a nonselective AMYR and CTR agonist[5]. An independent cryo-EM study reached the same set of contacts — bypass binding at AMY1R and CTR, F23 anchoring the peptide at the TM bundle, the E14–R17 salt bridge, P37 against the ECD — and likewise reported nonselective Gs activation through CTR and AMY1R[4].
Potency across AMY1R, AMY2R, AMY3R and CTR
In a 2024 comparison panel run in HeLa cells transduced with human calcitonin-family receptors, cagrilintide’s cAMP potency was nearly flat across the four targets and more than four orders of magnitude lower at CGRP and adrenomedullin 2 receptors. The table reports functional potency measured through cAMP accumulation, not binding affinity[6]:
| Receptor | Cagrilintide pEC50 (95% CI) | Pramlintide pEC50 (95% CI) |
|---|---|---|
| CTR | 10.64 (10.10–11.18) | 9.48 (8.16–10.80) |
| AMY1R | 10.73 (9.97–11.50) | 10.66 (9.53–11.80) |
| AMY2R | 10.37 (9.77–10.97) | 10.19 (8.65–11.74) |
| AMY3R | 10.33 (9.92–10.74) | 10.53 (9.18–11.89) |
| CGRPR | 6.25 (5.87–6.63) | 7.63 (6.78–8.49) |
| AM2R | 6.12 (5.85–6.39) | 6.83 (6.15–7.52) |
The source table additionally flags the cagrilintide lower confidence bounds of 5.87 at CGRPR and 5.85 at AM2R with footnote “a”; these weak-activity estimates should not carry the same interpretive weight as the four principal receptor results[6].
Two readings matter. First, the spread across AMY1R, AMY2R, AMY3R and CTR is under half a log unit — less than a threefold difference in EC50 — which is the pharmacological content of nonselective in this case. Second, the contrast with pramlintide is at CTR, not at the AMY receptors: point estimates differ by about 1.2 log units at CTR while the AMY1R values are almost identical, although the confidence intervals overlap. These are functional potency values, not radioligand binding constants (Ki), and they come from receptor expression in one cell background rather than native tissue.
Binding affinity is a separate measurement: the read sources provide no reportable cagrilintide Ki or Kd for CTR; none for AMY1R; none for AMY2R; and none for AMY3R. The JPET comparison states that binding was among the measures profiled for AM833/cagrilintide, but its abstract reports no numeric values[5]. The structural reports describe contacts and conformational states rather than equilibrium constants[1][4], and the 2024 cAMP panel reports potency, not Ki or Kd[6]. This is an evidence gap after retrieval of the sources supplied for this article, not a finding that affinities are equal across the four receptors. Functional pEC50 values should not be substituted for those constants.
Biased signaling: what the assays measure and what they do not
The evidence for signaling presented here centres on Gs: cAMP-related functional responses, Gs activation, and structures of Gs-bound complexes[1][4]. That supports nonselective activation across receptors for the Gs pathway and nothing wider. The 25-endpoint comparison reports that cagrilintide’s profile across binding, activation and regulation differs from selective and nonselective comparators[5], but the abstract gives no transducer-resolved potency ratios, and a Gs-coupled structure cannot show whether the same receptor also engages arrestin or other transducers. A claim that cagrilintide is biased for or against a particular transducer at a particular receptor is therefore not established by these sources.
The related idea that receptor dynamics explain the different metabolic effects of cagrilintide and salmon calcitonin is presented by the knockout-mouse authors as an inference from their data plus previously published in vitro work[2]; it is a hypothesis about mechanism, not a measured signaling bias.
Receptor distribution and the RAMP1/3 knockout mouse
The mouse experiments examined hindbrain-associated neuronal responses rather than providing a complete receptor-distribution map. In chow-fed male mice, cagrilintide raised cFos in the area postrema, nucleus of the solitary tract and lateral parabrachial nucleus. In RAMP1/3 double-knockout mice — which cannot assemble AMY1R or AMY3R while CTR and the RAMP2-containing AMY2R remain possible — cagrilintide’s area postrema activation fell by 57% relative to wild type, with no significant between-genotype difference in the nucleus of the solitary tract or lateral parabrachial nucleus. Within the knockout group, cagrilintide did not significantly increase cFos over vehicle in these regions. Salmon calcitonin kept activating all three areas in the knockouts at both experimental levels tested, possibly through the remaining AMY2R[2].
The result supports a substantial contribution from RAMP1/3-containing receptors to cagrilintide’s area postrema response in mice. Residual cFos does not establish a CTR or AMY2R component: the knockout experiment does not assign that signal to a specific remaining receptor. cFos counts are a proxy for neuronal activation, the acute cagrilintide experiment used one treatment level and each group had five animals, so this does not translate into a human dose, effect size or safety statement.
How much weight each evidence tier carries
Cryo-EM gives residue-level contacts and information about conformational variability in isolated complexes prepared in detergent micelles; it cannot by itself report affinity or efficacy. cAMP pEC50 values are functional potencies in a transfected cell line, useful for comparing ligands within one assay and unsuitable for predicting exposure. The knockout mouse is the causal test of RAMP dependence in the supplied evidence, and it is animal data with a proxy readout. Human evidence in these sources is limited to statements that cagrilintide has been tested in phase 1 and 2 trials[1][5]; nothing in this record measures receptor binding or occupancy in humans, and no clinical efficacy claim is made here.
Related research: Cagrilintide Mechanism of Action: AMY1R, AMY2R, AMY3R and CTR; What Is the Amylin Receptor? CTR and RAMP Proteins Explained; Cagrilintide vs Eloralintide: Receptor Profiles Compared; Cagrilintide.
Research Use Only. This information supports laboratory research on receptor pharmacology. It is not medical advice and does not support human or veterinary use, dosing or administration.
References
- Structural and dynamic features of cagrilintide binding to calcitonin and amylin receptors (Nature Communications, 2025)
- Cagrilintide lowers bodyweight through brain amylin receptors 1 and 3 (eBioMedicine, 2025)
- Cryo-EM structure of the human amylin 1 receptor in complex with CGRP and Gs protein (Biochemistry, 2024)
- Structural and mechanistic insights into dual activation of cagrilintide in amylin and calcitonin receptors (Acta Pharmacologica Sinica)
- AM833 is a novel agonist of calcitonin family G protein-coupled receptors: pharmacological comparison with six selective and nonselective agonists (JPET, 2021)
- NN1213, a potent, long-acting and selective analog of human amylin (J Med Chem, 2024)