Structure and dynamics of the pyroglutamylated RF-amide peptide QRFP receptor GPR103

Abstract

Pyroglutamylated RF-amide peptide (QRFP) is a peptide hormone with a C-terminal RF-amide motif. QRFP selectively activates a class A G-protein-coupled receptor (GPCR) GPR103 to exert various physiological functions such as energy metabolism and appetite regulation. Here, we report the cryo-electron microscopy structure of the QRFP26-GPR103-G_q complex at 3.19 Å resolution. QRFP26 adopts an extended structure bearing no secondary structure, with its N-terminal and C-terminal sides recognized by extracellular and transmembrane domains of GPR103 respectively. This movement, reminiscent of class B1 GPCRs except for orientation and structure of the ligand, is critical for the high-affinity binding and receptor specificity of QRFP26. Mutagenesis experiments validate the functional importance of the binding mode of QRFP26 by GPR103. Structural comparisons with closely related receptors, including RY-amide peptide-recognizing GPCRs, revealed conserved and diversified peptide recognition mechanisms, providing profound insights into the biological significance of RF-amide peptides. Collectively, this study not only advances our understanding of GPCR-ligand interactions, but also paves the way for the development of novel therapeutics targeting metabolic and appetite disorders and emergency medical care.

Fig. 1. Overall structure of the GPR103–mini-G_sqiβ₁γ₂–scFv16–Nb35 complex.

a Amino acid sequences of QRFP43 and QRFP26. b Cryo-EM density map of the GPR103–mini-G_sqiβ₁γ₂–scFv16–Nb35 individually colored. c Refined structures of the complex are shown as a ribbon representation. d Diagram of GPR103. N-terminal forms a helix-loop-helix motif. ECL2 forms a long β-sheet. e Ribbon representation of the QRFP26 and GPR103. Density focused on QRFP26 (pink). Two disulfide bonds are represented by stick models. The one is the highly conserved disulfide bond between C118^3.25 and C201^ECL2, and the other is atypical disulfide bond between C285^6.47 and C327^7.48. The C118^3.25A and C201^ECL2A mutations abolished the QRFP26 potency (Supplementary Fig. 1a–d). By contrast, the C285^6.47A and C327^7.48A mutations did not alter the potency, indicating their lesser importance for receptor function.

Fig. 2. QRFP26 binding site in the transmembrane region.

a, b Binding pocket for QRFP26 in the TMD. Residues involved in the QRFP-GPR103 interaction within 4.5 Å are shown as pink and blue sticks, respectively. Black dashed lines indicate hydrogen bonds. c Effects of mutations in the ligand-binding pocket of GPR103. QRFP26-induced activation of GPR103 was analyzed by the TGFα shedding assay. From the concentration–response curves (Supplementary Fig. 1b), ΔpEC₅₀ values relative to the wild-type were calculated. Colors in the mutant bars indicate an expression level matching that of titrated wild-type. NA, parameter not available because of lack of the ligand response. WT, wild-type. Statistical analyses were performed using the ordinary one-way ANOVA followed by Dunnett tests with the expression-matched (colored) wild-type response. ns, p > 0.05; *p < 0.05; **p < 0.01; ***p < 0.001. Data are presented as mean values ± SEM from at least three independent experiments performed in triplicate (n = 5 for the wild-type and n = 3 for the mutants). Source data are provided as a Source Data file. d, e Close-up views of F26 (d) and R25 (e) of QRFP26. f Surface representation of GPR103 viewed from the extracellular side. Positive and negative charges of the receptor are colored in blue and red, respectively.

Fig. 3. Architecture of the extracellular region.

a Ribbon representation of GPR103 focused on the extracellular side. b Hydrogen-bonding interactions of R17-P33^N-term and K18-Y36^N-term are indicated by black dashed lines. c Hydrophobic interaction of the N-terminal HLH with ECL2 and QRFP26. d–f Effects of mutations in the ECD of GPR103. QRFP26-induced activation of GPR103 was analyzed by the TGFα shedding assay. From the concentration–response curves (Supplementary Fig. 1a), ΔpEC₅₀ values relative to the wild-type were calculated (d). The concentration–response curves of the N-terminus-truncated mutants and the ECL2 deletion mutant were shown in e and f, respectively. Data for the wild-type response was obtained from the same experiment as Fig. 2c. Colors in the mutant bars indicate an expression level matching that of titrated wild type. NA, parameter not available because of lack of the ligand response. WT, wild-type. Statistical analyses were performed using the ordinary one-way ANOVA followed by Dunnett tests with the expression-matched (colored) wild-type response. ns, p > 0.05; *p < 0.05; **p < 0.01; ***p < 0.001. Data are presented as mean values ± SEM from at least three independent experiments performed in triplicate (n = 5 for the wild-type and n = 3 for the mutants). Source data are provided as a Source Data file.

Fig. 4. Structural polymorphism of the ECD.

a, b Superposition of the GPR103 structures in the tilted, upright, and refined states. c, d Superimposition of the PTH1R structures in class 1–3 (PDB 7VVM (PTH1R-G_s structure), 7VVL (PTH1R-G_s structure), and 7VVK (PTH1R-G_s structure)).

Fig. 5. Comparison of other related amide peptide receptors.

a Comparison of residues interacting with the RX-amide moiety in the RF- and RY-amide receptors. b Superimposition of GPR103 (blue), Y₁R (PDB 7X9A (NPY–Y₁R–G_i structure), green), CCK₁R (PDB 7MBY (CCK-8–CCK₁R–mini-G_sqi structure), magenta), and OX₂R (PDB 7L1U (OxB–OX₂R–mini-G_sqi structure), purple) structures in complex with QRFP26, NPY, CCK-8, and OxB, peptides, respectively. NPY, neuropeptide Y; CCK-8, cholecystokinin octapeptide; OxB, orexin-B. c, d Structural comparison of the binding mode of the RF-amide moiety in GPR103 (c) and the RY-amide moiety in Y₁R (d). The residues involved in ligand-receptor interactions are represented by stick models. Black dashed lines indicate hydrogen bonds. e Schematic representations of RX-amide recognition conserved in RF- or RY-amide receptors. The most conserved residues are indicated based on the structures of GPR103 and Y₁R.

Fig. 6. Model of receptor activation.

a–c Schematic representations of GPR103 activation model upon QRFP binding. From the apo state (a), the peptide binds (b) and changes to the active state (c).

Fig. 7. Structural comparison of GPCRs with N-terminal regions.

a–e Structural comparison of the N-terminal region in PTH1R (PDB 7VVK(PTH-PTH1R–G_s structure)) (a), TSHR (PDB 7XW5 (ML109–TSHR–G_s structure)) (b), CXCR2 (PDB 8XWN (CXCL8–CXCR2 structure)) (c), C5aR (PDB 8HK5 (C5a–C5aR1–G_i structure)) (d), and GPR103 (e).