Identification of ApPRXamide receptors.

A, Cloning of four putative ApPRXamide receptors and detection of the activation profiles induced by three classes of ApPRXamide peptides, as measured by the IP1 accumulation assay. B, The PCR products of putative ApPRXamide receptors (left to right): ApPRXa-R1 with a length of 1,599 bp, ApPRXa-R2 with a length of 1,704 bp, ApPRXa-R-like1 with a length of 1,767 bp, ApPRXa-R-like2 with a length of 1,581 bp. Lane 1: DNA marker (M); Lane 2: the target gene. C-D, Representative examples of dose-response curves showing activation of ApPRXa-R1 (C) and ApPRXa-R2 (D) in CHO-K1 cells by MMG2-DPs, pleurins and SCPs. E, Comparison of p[EC50] values for ApPRXa-R1 and ApPRXa-R2 activation by MMG2-pDPb and MMG2-DPb shown in (C-D), n ≥ 3. Two-way ANOVA, Fpeptide (1, 25) = 27.41, p < 0.0001, Freceptor (1, 25) = 695.1, p < 0.0001. F, Comparison of Emax of ApPRXa-R1/R2, n ≥ 3. Two-way ANOVA, Fpeptide (1, 24) = 0.1468, p > 0.05, Freceptor (1, 24) = 56.76, p < 0.0001. Bonferroni post-hoc test: n.s., not significant; **P < 0.01; ***P < 0.001; ****P < 0.0001. Error bar: SEM. G, Summary of the average p[EC50] and EC50 values shown in (C-D), with peptide sequences listed. IP1, inositol monophosphate; CHO-K1, Chinese hamster ovary-K1; p[EC50], -log10(EC50); Emax, the maximum IP1 concentration.

Binding modes of MMG2-pDPb and MMG2-DPb in ApPRXa-R1 from molecular docking and receptor mutagenesis.

A-B, The cutting face of the ligand binding pocket in the (A) MMG2-pDPb-and (B) MMG2-DPb-bound ApPRXa-R1 structure. MMG2-pDPb and MMG2-DPb are represented as green and blue sticks respectively, with their N-terminal pyroglutamic acid and glutamine marked by red lines. ApPRXa-R1 is depicted as light orange surfaces. C-D, Overall interactions between MMG2-pDPb (C) and MMG2-DPb (D) with ApPRXa-R1. ApPRXa-R1 is shown as light orange cartoons with its residues represented as orange sticks. Hydrogen bonds and salt bridges are indicated by (C) green and (D) blue dashed lines. All receptor binding sites and corresponding ligand residues are labeled. E-G, Comparison of detailed interactions between MMG2-pDPb/DPb and ApPRXa-R1. The merged maps of the three regions are indicated in different colored panels in (C-D). Differences in interactions are observed for residues Arg6 (E), Pro2 (F) and Leu4 (G). Hydrophobic interactions are shown as dash lines with arrows, with interaction differences highlighted by red dashed circles. H-I, Representative examples of dose-response curves showing the activation of ApPRXa-R1 and its mutants (based on predicted binding sites) by (H) MMG2-pDPb and (I) MMG2-DPb in CHO-K1 cells, as measured by the IP1 accumulation assay. Dose-response curves are normalized to 100% of the maximal response elicited by the ligands in ApPRXa-R1. J-K, Summary of mutant effects on ligand potency, wildtype ApPRXa-R1 was used as a control. Panel (J) corresponds to (H). n ≥ 3. One-way ANOVA, F (6, 17) = 115.9, p < 0.0001. Panel (K) corresponds to (I). n ≥ 3. One-way ANOVA, F (6, 16) = 107.8, p < 0.0001. Bonferroni post-hoc test: n.s., not significant; ***P < 0.001; ****P < 0.0001. Error bar: S.E.M.

Binding modes of MMG2-pDPb and MMG2-DPb in ApPRXa-R2 from molecular docking and receptor mutagenesis.

A-B, The cutting face of the ligand binding pocket in the (A) MMG2-pDPb-and (B) MMG2-DPb-bound ApPRXa-R2 structure. MMG2-pDPb and MMG2-DPb are represented as green and blue sticks respectively, with their N-terminal pyroglutamic acid and glutamine marked by red lines. ApPRXa-R2 is depicted as light pink surfaces. C-D, Overall interactions between MMG2-pDPb (C) and MMG2-DPb (D) with ApPRXa-R2. ApPRXa-R2 is shown as light pink cartoons with its residues represented as pink sticks. Hydrogen bonds and salt bridges are indicated by (C) green and (D) blue dashed lines. All receptor binding sites and corresponding ligand residues are labeled. E-G, Comparison of detailed interactions between MMG2-pDPb/DPb and ApPRXa-R2. The merged maps of the three regions are indicated in different colored panels in (C-D). Differences in interactions are observed for residues Arg6 (E), Pro2 (F), Pro3, Pro5, and Tyr7 (G). Hydrophobic interactions are shown as dash lines with arrows, with interaction differences highlighted by red dashed circles. H-I, Representative examples of dose-response curves showing the activation of ApPRXa-R2 and its mutants (based on predicted binding sites) by (H) MMG2-pDPb and (I) MMG2-DPb in CHO-K1 cells, as measured by the IP1 accumulation assay. Dose-response curves are normalized to 100% of the maximal response elicited by the ligands in ApPRXa-R2. J-K, Summary of mutant effects on ligand potency, wildtype ApPRXa-R2 was used as a control. Panel (J) corresponds to (H). n ≥ 3. One-way ANOVA, F (6, 17) = 46.46, p < 0.0001. Panel (K) corresponds to (I). n ≥ 3. One-way ANOVA, F (7, 17) = 29.82, p < 0.0001. Bonferroni post-hoc test: n.s., not significant; **P < 0.01; ***P < 0.001; ****P < 0.0001. Error bar: SEM.

Schematic illustrating how N-terminal pyroglutamylation (pQ) remotely modulates receptor subtype-specific activation without direct contacts.

In both ApPRXa-R1 and ApPRXa-R2, MMG2-pDPb (pQPPLPRYamide), which exhibits reduced conformational flexibility due to pQ, does not engage new receptor contacts but instead adopts a distinct bound conformation, resulting in differences in the number or strength of bonds with largely shared sets of interacting receptor residues. Compared with a more flexible MMG2-DPb, in ApPRXa-R2 (left, red), which may possess a loose binding pocket, the more rigid ligand with pQ strengthens key interactions with Y186ECL2, F210ECL2, S1032.64 and E1002.61, thereby potentiating receptor activation. In contrast, in ApPRXa-R1 (right, cyan), whose pocket is intrinsically compact, the same rigid pQ-modified ligand disrupts the interaction network formed by the more flexible MMG2-DPb, including contacts involving W432ECL3, Y237ECL1, E2292.61 and T4246.58, resulting in reduced receptor activity. This bidirectional, subtype-specific mechanism—enhancement in ApPRXa-R2 versus inhibition in ApPRXa-R1—accounts for the preferential activation of ApPRXa-R2 by MMG2-pDPb and of ApPRXa-R1 by the non-cyclized form MMG2-DPb.

Activation of ApPRXa-Rs by MMG2-pDPb analogs.

A, There were two types of MMG2-pDPb analogs: with one substituting the residues at positions 2-7 with Ala, and the other modifying the post-translational modifications (PTM) including pyroglutamic acid and amidation group. B-C, Comparison of conformations of [Ala1]MMG2-pDPb and MMG2-DPb in ApPRXa-R1 (B) and ApPRXa-R2 (C). [Ala1]MMG2-pDPb and MMG2-DPb are represented as green and blue sticks respectively. ApPRXa-Rs residues are shown as sticks with white surfaces. All receptor binding sites and corresponding ligand residues are labeled. D-E, Representative examples of dose-response curves showing the activation of ApPRXa-R1 (D) and ApPRXa-R2 (E) by PTM-related analogs. F, Summary of the average p[EC50] and EC50 values shown in (D-E), with peptide sequences listed. G-H, Representative examples of dose-response curves showing the activation of ApPRXa-R1 (G) and ApPRXa-R2 (H) by [Ala2-7]MMG2-pDPb. I, Summary of the average p[EC50] and EC50 values shown in (G-H), with peptide sequences listed. Dose-response curves are normalized to 100% of the maximal response elicited by MMG2-pDPb in receptors. J-K, Group data comparing the effects of all MMG2-pDPb analogs on ApPRXa-R1 (J) and ApPRXa-R2 (K). MMG2-pDPb was used as a control. Panel (J) corresponds to (D) and (G). n ≥ 3. One-way ANOVA, F (9, 22) = 153.2, p < 0.0001. Panel (K) corresponds to (E) and (H). n ≥ 3. One-way ANOVA, F (9, 34) = 151.3, p < 0.0001. Bonferroni post-hoc test: n.s., not significant; *P < 0.05; **P < 0.01; ****P < 0.0001. Error bar: SEM.

Human Neuromedin U receptor subtypes reveal pocket-dependent distal steering by N-terminal pyroglutamylation.

A-B, Representative examples of dose-response curves and summary data showing the activation of hNmU-R1 (A) and hNmU-R2 (B) by NmU-8 (YFLFRPRNamide), pQ-NmU-7 (pQFLFRPRNamide) and Q-NmU-7 (QFLFRPRNamide). NmU-8 exhibits distinct potencies at the two receptors as previously reported (17). For assays of pQ/Q-NmU-7 on hNmU-R1, n =3. One-way ANOVA, F (2, 6) = 11.75, p < 0.01; hNmU-R2, n =3. One-way ANOVA, F (2, 6) = 13.58, p < 0.01. Bonferroni post-hoc test: n.s., not significant; *P < 0.05; **P < 0.01. Error bar: SEM. C, Summary of the average p[EC50] and EC50 values from (A-B), with peptide sequences listed. NmU-8, Neuromedin U-8; hNmU-R1/2, human Neuromedin U receptors 1/2. D, Orthosteric binding pockets of ligand-bound active-state hNmU-R1(red) and hNmU-R2 (blue). Receptors are shown as cartoons and CASTp-defined molecular pockets are shown as white surfaces. E-F, Rg (E) and RMSF (F) profiles from 300-ns MD simulations of NmU-8-bound hNmU-R1 and hNmU-R2 complexes. G-H, Rg (G) and RMSF (H) profiles from 300-ns MD simulations of pQ-NmU-7-bound hNmU-R1 and hNmU-R2 models. I, Conceptual framework for PTM distal steering mechanism in relation to classical ligand-receptor interaction models. Left panel, classical paradigms and their limitations. The lock-key model assumes rigid, pre-complementary binding and neglects conformational dynamics of both ligand and receptor. The induced-fit model incorporates receptor rearrangement but does not account for non-contacting ligand features or divergent receptor-subtype architectures. Middle panel, structural diversity of ligands and receptors illustrated by PRXamide/NmU signaling. Ligands lacking N-terminal pQ sample multiple dynamic conformations (dashed lines, arrows), whereas N-terminal pQ rigidifies the peptide and pre-organizes its core into a constrained conformation. Receptor pockets range from compact environments (red; e.g., ApRPXa-R1, hNmU-R2) to looser pockets (green; e.g., ApPRXa-R2, hNmU-R1). Right panel, PTM distal steering mechanism. Top left: a compact receptor pocket paired with a flexible, non-pQ ligand permits baseline activation. Top right: the same compact pocket challenged by a pQ-rigidified ligand induces conformational frustration and reduced activation, defining an intermediate binding mode between the lock-key and induced-fit. Bottom left: a loose pocket interacting with a flexible, unmodified ligand enables baseline activation. Bottom right: the loose pocket productively accommodates the pQ-rigidified ligand, enhancing activation via canonical induced-fit.