The overall structure of G_q-coupled MRGPRX1 bound to compound 16

To improve receptor expression, we fused thermostabilized apocytochrome b₅₆₂ (BRIL) [22] at the N-terminus of MRGPRX1. NanoBit tethering strategy [23] was used for the complex formation, with the LgBit and HiBit fused to the C-terminus of the receptor and G_β subunit, respectively (S1A and S1B Fig). We used bioluminescence resonance energy transfer (BRET) assay to evaluate the impact of receptor modification on G protein coupling capability. The fusion of BRIL and LgBit to receptor only marginally affected receptor activity (S1C Fig and S1 Table). To further stabilize the complex, we used an engineered G_αq chimera in the complex assembly. The engineered G_αq chimera was designed based on the mini-G_αs/q71 [24,25] with several modifications (S1D Fig). Briefly, the N-terminal 1–18 residues of the mini-G_αs/q71 [24] were replaced by corresponding N-terminal sequences of the human G_αi1, while the α-helical domain of G_αi1 was subsequently inserted into the mini-G_αs/q71, thus providing possible binding sites for 2 antibody fragments scFv16 and Fab-G50 [26,27]. Additionally, 2 dominant-negative mutations (G203A and A326S) were introduced to decrease the affinity of nucleotide binding [28]. The same engineered G_αq chimera had been successfully used in the structure determination of several G_q-bound GPCRs, including the G_q-bound ghrelin receptor [29] and bradykinin receptors [30]. The engineered G_αq chimera used in the further structure study will be simplified as G_αq. We co-expressed BRIL-MRGPRX1-Lgbit, G_αq, and G_βγ-HiBit to obtain the MRGPRX1-G_αq complex. The complex was further stabilized by incubating with Nb35 [31] and scFv16 [32] in the presence of compound 16 (S2 Fig).

The compound 16-MRGPRX1-G_αq complex structure was determined by cryo-EM to yield a final map at an overall resolution of 3.0 Å (Figs 1A and S3A–S3F and S2 Table). In the map, the densities for the receptor, G_αq, G_βγ, Nb35, scFv16, and compound 16 could be well distinguished, and the interface residues between MRGPRX1 and G_αq (α5-helix) were clearly defined (S3G Fig). Thus, we built a reliable atomic model based on the well-traced α-helices and aromatic side chains (Fig 1B). Due to the flexibility, the N-terminus (M1-K25), part of the extracellular loops (I90) and long C-terminal residues (R279-Q322) of the receptor are invisible.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Cryo-EM structure of compound 16-MRGPRX1-G_αq complex.

(A) Cryo-EM density map of MRGPRX1-G_αq in complex with compound 16. Receptor, compound 16, G_αq, G_β, G_γ, Nb35, and scFv16 are colored slate, cyan, marine, forest, pink, salmon, and yellow-orange, respectively. The density of compound 16 is shown in mesh at the top-right corner. Compound 16 is fitted and shown as sticks. (B) Structural model for the compound 16-MRGPRX1-G_αq complex. Receptor, ligand, G_αq, G_β, G_γ, Nb35, and scFv16 are colored the same as (A). Compound 16 is shown as sticks. The 2D chemical structure of compound 16 is shown in the top-right corner. cryo-EM, cryo-electron microscopy.

https://doi.org/10.1371/journal.pbio.3001975.g001

The orthosteric pocket of MRGPRX1

The MRGPRX1 exhibits a shallow, broad, and wide-open ligand-binding pocket (Fig 2A). The distance between compound 16 and the critical toggle switch residue G229^6.48 is about 16.8 Å (Fig 2B), indicating that compound 16 is positioned near the extracellular surface but not buried deep in the receptor. The shallow pockets are also observed in the MRGPRX2-G_αq [14,15] and MRGPRX4-G_αq complex [14], suggesting the common pocket features among all MRGPRX receptors. Compound 16 occupies only about one-third of the pocket (Fig 2A) but is sufficient for the receptor activation. Similarly, the agonist (R)-ZINC-3573 and MS47134 take up only a small part of the pocket in MRGPRX2 and MRGPRX4, respectively (S4A and S4B Fig). The structure alignment shows that 3 agonists bind to different regions of the orthosteric pocket in 3 receptors, indicating distinct recognition mechanisms (S4C Fig). The electrostatic potential of the MRGPRX1 pocket is partially negative, partially positive, and partially hydrophobic (S4D Fig). In contrast, the electrostatic potential of the MRGPRX2 pocket is partially negative (sub-pocket 1) and partially hydrophobic (sub-pocket 2), and the electrostatic potential of MRGPRX4 pocket is positive (S4E and S4F Fig). These results suggest that MRGPRXs may prefer agonist scaffolds with distinct electro-properties.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The binding pocket of compound 16 on MRGPRX1.

(A) Top view of the compound 16-binding pocket from the extracellular side (surface mode). Pocket is colored gray, and compound 16 is shown as cyan sticks. (B) Cut-away view of the compound 16-binding pocket. The distance between the ligand and the traditional toggle switch position is shown as dashed lines. (C) Top view of the compound 16-binding pocket from the extracellular side (cartoon mode). TMs and ECLs are colored slate. Compound 16 is shown as cyan sticks. C161^4.64 and C173^5.32 are shown as sticks and colored slate. (D) BRET validation of residues C161^4.64 and C173^5.32. (E, F) Interaction between compound 16 and MRGPRX1 from 2 views. The key residues are shown as sticks and colored slate. The polar interactions are shown as yellow dashed lines. (G, H) BRET validation of residues in compound 16-binding pocket. Data are presented as mean ± SEM. n = 3; Emax, maximum effect; WT, wild type. The underlying data for Fig 2D, 2G and 2H can be found in S1 Data. BRET, bioluminescence resonance energy transfer.

https://doi.org/10.1371/journal.pbio.3001975.g002

The orthosteric pocket of MRGPRX1 accommodating compound 16 is composed of residues majorly located on TM3/4/5/6 (Fig 2C). C161^4.64 and C173^5.32 form a disulfide bond, which is conserved among MRGPRXs [14] (S5 Fig). Moreover, alanine substitutions of C161^4.64 and C173^5.32 nearly abolish the G_q coupling activity (Fig 2D). Interestingly, MRGPRX1 lacks the canonical disulfide bond between TM3 and ECL2 in other class A family GPCRs [33]. Taken together, the disulfide bond substitution in MRGPRX1 may help to reorganize the extracellular loops and maintain the wide-open orthosteric pocket.

The interaction between compound 16 and MRGPRX1

Compound 16 adopts a hairpin conformation in the pocket due to the intramolecular π-π interaction of the 1-aminoisoquinoline and phenylmethyl groups (Fig 1A). Notably, the amino group of 1-aminoisoquinoline forms strong salt bridges with E157^4.60 and D177^5.36 (Fig 2E and 2F), 2 conserved residues in all MRGPRXs (S5 Fig). Alanine substitutions of E157^4.60 and D177^5.36 dramatically affect receptor activation, suggesting that these 2 acidic residues play a crucial role in ligand recognition (Figs 2G and S6 Fig and S1 Table). Additionally, the variable residue Y99^3.29 recognizes compound 16 through 2 types of interactions, the π-π interaction with the middle aromatic ring of compound 16 and polar interaction with the amino group of 1-aminoisoquinoline (Fig 2E and 2F). Alanine substitution of Y99^3.29 nearly abolishes the G_q coupling activity (Figs 2G and S6 and S1 Table). Thus, the variable residue Y99^3.29 is essential for recognizing compound 16. Besides, the positive charge residue K96^3.26 in MRGPRX1 is conserved in MRGPRX4 but substituted by a serine in MRGPRX2 (S5 Fig). K96^3.26 in MRGPRX1 is close to compound 16 (Fig 2E and 2F), but S103^3.26 in MRGPRX2 is far away from (R)-ZINC-3573 (S7A Fig). However, the alanine substitution of K96^3.26 slightly affects the receptor activation (Figs 2G and S6 and S1 Table), suggesting that K96^3.26 does not directly participate in compound 16 recognition. In contrast, the K96^3.26 in MRGPRX4 is critical for MS47134 recognition [14] (S7B Fig). The variation of pocket residues may partly explain the agonist selectivity among MRGPRXs [12]. Apart from these closer residues, the farther hydrophobic residues around compound 16, such as W158^4.61, F236^6.55, F237^6.56, L240^6.59, and W241^6.60, also participate in the ligand-binding pocket formation (Fig 2E and 2F). These hydrophobic residues, except for F237^6.56, exhibit indirect and relatively weak interactions with compound 16 (Figs 2H and S6 and S1 Table). Alanine mutation of F237^6.56 shows a notable impact on compound 16-induced MRGPRX1 activation (Figs 2H and S6 and S1 Table), indicating a potentially crucial role of F237^6.56 in ligand recognition or receptor activation. However, the alanine substitution of F236^6.55 only partially affects the receptor activation, double confirmed by BRET assay and calcium imaging assay (Figs 2H and S6 and S1 Table). It suggests that F236^6.55 may help maintain the ligand-binding pocket instead of directly interacting with the ligand. Besides, the activity of compound 16 is partially reduced by substituting L240^6.59 with alanine in the BRET assay (Fig 2H and S1 Table). In contrast, it is not affected in the calcium imaging assay (S6 Fig). The discrepancy may indicate L240^6.59 is less important for the compound 16-binding pocket. Conversely, the corresponding residue L247^6.59 in MRGPRX2 was reported to be critical for its agonist (R)-ZINC-3573 binding (S7A Fig), and Y240 is located on ECL3 of MRGPRX4, which is also important for the recognition of the agonist MS47134 [14,15] (S7B Fig). The differences in hydrophobic residues surrounding the ligand pocket illustrate the preference of ligand recognition among MRGPRXs. Together, these detailed structural analyses of compound 16-binding pocket can provide important information for a better understanding of the ligand recognition mechanism in MRGPRX1.

Activation of MRGPRX1

The compound 16-MRGPRX1-G_αq complex structure exhibited TM rearrangement in the cytoplasmic half. The cytoplasmic ends of TM3 and TM6 are about 16 angstroms apart, consistent with other class A G protein-engaged GPCRs in an active conformation [33] (Fig 3A). Notably, the conserved toggle switch W^6.48 in other GPCRs is replaced by G229^6.48 in MRGPRX1. This vital substitution results in an inward movement of the extracellular half of TM6, narrowing the gap between TM3 and TM6 and initiating the formation of a shallow orthosteric pocket (Fig 3B). Briefly, Y106^3.36 in TM3 engages with G229^6.48 in TM6 to form a twist. This twist is then stabilized by the hydrophobic interactions network among Y106^3.36, F232^6.51, and F237^6.56, which prevents the ligands from entering the deeper location and exhibits a shallow pocket to accommodate ligands. Additionally, we conducted a structural comparison of MRGPRX1 complex to its functional closely related μ opioid receptor (μOR) in the active state (PDB 7U2L) [34] and inactive state (PDB 7UL4) [35] (Fig 3C). The structural comparison demonstrates that the MRGPRX1 complex shows a similar structure as the active μOR. Moreover, the structure superposition of G_αq-coupled MRGPRX1 with G_αq-coupled 5-HT_2AR (PDB 6WHA) [36] and G_αq-coupled B1R (PDB 7EIB) [30] by receptors also exhibits similar conformations (S8A and S8B Fig), suggesting a common activation mechanism among these receptors. Significantly, except for the extracellular half of TM6, G_αq-coupled MRGPRX1 shows nearly identical conformations of TM3, TM6, and TM7 with these receptors in active state (Figs 3C and S8). MRGPRX1 possesses several unique residues in its extracellular half of TM6, which are essential for receptor activation. Furthermore, consistent with our speculation, alanine mutations of these residues dramatically affected MRGPRX1 activation induced by compound 16 (Fig 3D). Hence, the initiation of MRGPRX1 activation is likely triggered by touching F237^6.56 at the bottom of the pocket upon agonist binding, pushing a series of residues in TM6 to move towards TM3 and resulting in the conformational change of G229^6.48. G229^6.48 shifts to get close to Y106^3.36, triggers the rotation of conserved F^6.44, and further facilitates the intracellular half of TM6 moving outward to accommodate the downstream G protein (Figs 3C and S8).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Activation mechanism of MRGPRX1 and conformational switches of MRGPRX1 in active state.

(A) Structural representation of TMs arrangement between TM3 and TM6. The distance between the C-terminus of TM3 and TM6 is shown as a dashed line. (B) Structural representation of the interactions near the substituted toggle switch G229^6.48. (C) Structural superposition of active MRGPRX1, active μOR (PDB 7U2L) [34] and inactive μOR (PDB 7UL4) [35] from the side, cytoplasmic, and magnified views. The movement directions of TM6, TM7, and residues in MRGPRX1 relative to inactive μOR are highlighted as red arrows. MRGPRX1, active μOR, and inactive μOR are colored in slate, light pink, and gray, respectively. (D) BRET validation of essential residues in the extracellular half of TM6. Data are presented as mean ± SEM. n = 3; Emax, maximum effect; WT, wild type. Magnified view of D/ERY motif (E), LLLF motif (F), and NPXXY motif (G). Polar interactions are shown as yellow dashed lines. The underlying data for Fig 3D can be found in S1 Data. BRET, bioluminescence resonance energy transfer; μOR, μ opioid receptor.

https://doi.org/10.1371/journal.pbio.3001975.g003

In addition to the unique twist structure in TM6, MRGPRX1 also shows significant differences in some classic motifs for class A GPCR activation. Firstly, the conserved D(E)^3.49R^3.50Y^3.51 motif on TM3 of most class A GPCRs forms an ionic lock in an inactive conformation and is broken upon activation [33]. In MRGPRX1, Y^3.51 is replaced by C121^3.51, E119^3.49 interacts with R134^ICL2 via a strong salt bridge, and R120^3.50 interacts with T217^6.36 to limit the movement of TM6 and Y359 on G_αq to stabilize the complex via polar interactions (Fig 3E). Secondly, the conserved motif P^5.50I^3.40F^6.44 [37] is substituted by an L187^5.46L191^5.50L110^3.40F225^6.44 motif, constraining the conformation of TM3/5/6 (Fig 3F). However, the conserved N^7.49P^7.50XXY^7.53 motif, which does not interact directly with G proteins but is essential for receptor activation [33], is conserved in MRGPRX1 (Fig 3G). The above structure features are highly conserved among MRGPRXs, suggesting the common activation mechanism despite distinct agonist recognition features [14,15].

The coupling of G_αq to MRGPRX1

The coupling of G_αq to MRGPRX1 is mainly maintained by interacting with residues on TM2, TM3, TM5, TM6, and ICL2 (S9A Fig). The interface between G_αq and TMs comprises a series of hydrophobic residues, including F61^2.39, V124^3.54, I202^5.61, L211^6.30, and L214^6.33 on the receptor and L352, L356, and L361 on the α5-helix of G_αq (S9B Fig). However, alanine substitutions of these hydrophobic residues have little effect on G_αq coupling activity (S9C Fig and S1 Table). Only I202^5.61A partially reduces G_αq activation (S9C Fig and S1 Table), suggesting that the TM bundles are less critical for G_αq activation in MRGPRX1. Moreover, in most class A GPCRs, ICL2 does not interact with G protein directly. However, extensive interactions of ICL2 with the αN-helix and α5-helix of G_αq are observed in the MRGPRX1 structure (S9D Fig). Alanine substitutions of I128^ICL2 and H133^ICL2 nearly impair G_αq coupling activity (S9E and S9F Fig, S1 Table), indicating that ICL2 plays a crucial role in G_αq coupling. Similar G protein coupling interfaces are also observed in the complex structures of MRGPRX2 and MRGPRX4 previously reported (S10A and S10B Fig).

Construct design

The full-length wild-type (WT) human MRGPRX1 (residues M1-S382) was cloned into pFastBac vector with an HA signal peptide, an N-terminal Flag tag, and a C-terminal 10×His tag. A BRIL [22] protein was fused at the N-terminus of MRGPRX1 to improve the expression of the receptor. LgBit [23] was fused at the C-terminus of MRGPRX1 to stabilize the whole MRGPRX1-Gαq complex. There was no more modification for the MRGPRX1 sequence. Prof. H. Eric Xu from Shanghai Institute of Materia Medica donated the engineered Gαq chimera plasmid. This engineered G_αq [29,30] was designed based on a mini-G_αs/q71 [24,25] skeleton with the replacement of G_i1 N-terminus and the insertion of G_αi1 α-helical domain, thus providing possible binding sites for 2 antibody fragments scFv16 and Fab-G50 [26,27]. Additionally, 2 dominant-negative mutations (G203A and A326S) were introduced to decrease the affinity of nucleotide binding [28]. The WT G_β1γ2 with HiBit [23] fused at the C-terminus of the β1 subunit was cloned into pFastBac-Dual vector. Ric8A is a molecular chaperone that has been reported to be essential for the biogenesis and signaling of G_αq subunits, and the involved functions include facilitating the proper folding of G_αq and promoting the formation of G_α guanine nucleotide-binding pocket [44–46]. The full-length Ric8A was cloned into pFastBac vector. For biochemical assay, the full-length WT MRGPRX1 and the MRGPRX1 point mutants were cloned into mammalian expression vector pCDNA 3.1 with an N-terminal Flag tag. All these constructs were generated with a standard PCR-based strategy and homologous recombination (CloneExpress One Step Cloning Kit, Vazyme).

Expression and purification of scFv16

The scFv16 with a C-terminal 8×His tag was expressed in Trichoplusia ni Hi5 insect cells and purified precisely as previously described [26,32,47]. Briefly, the Hi5 insect cells (3.0 to 4.0 million cells per mL) were infected with the scFv16 virus produced by the Bac-to-Bac system (Invitrogen) for 96 h. Then, the medium was collected and pH balanced to pH 8.0 by adding Tris-base powder. Chelating agents were quenched by adding 1 mM nickel chloride and 5 mM calcium chloride. After incubation at room temperature (25°C) for 1 h by stirring constantly, the supernatant was isolated by centrifugation (4,750 rpm, 30 min, 4°C) and incubated with Ni-sepharose (GE Healthcare) for 1 h at room temperature with constant stirring. The resin was collected and then washed by washing buffer (20 mM HEPES (pH7.5), 100 mM NaCl, and 20 mM Imidazole). The protein was eluted by elution buffer (20 mM HEPES (pH 7.5), 100 mM NaCl, and 250 mM Imidazole) and then incubated with HRV-3C protease at 4°C for 2 h to remove the C-terminal 8×His tag. The cleaved protein was further loaded onto Superdex 200 Increase 10/300 GL column (GE Healthcare) with running buffer (20 mM HEPES (pH 7.5), 100 mM NaCl). ScFv16 peak fraction was collected, flash-frozen, and stored at −80°C until use.

Expression and purification of Nb35

Nanobody-35 (Nb35) was expressed in the Escherichia coli BL21 (DE3) and purified as previously reported [30,31]. Nb35 was cultured in LB with 50 μg/mL ampicillin at 37°C, 220 rpm for about 3 h until OD₆₀₀ reached 1.0. Then, IPTG was added to induce the protein expression at 25°C, 220 rpm with a final concentration 1 mM. Then, the E. coli bacteria were collected by centrifugation (4,000 rpm, 20 min, 4°C) after 16 h. The pellets were resuspended in lysis buffer (25 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM PMSF) and disrupted by ultrasonication. The supernatant was isolated by centrifugation (14,000 rpm, 60 min, 4°C) and incubated with Ni-NTA resin (Qiagen) for 1 h. The resin was washed by washing buffer (25 mM HEPES (pH 7.5), 150 mM NaCl, and 25 mM Imidazole) and then eluted by elution buffer (25 mM HEPES (pH 7.5), 150 mM NaCl, and 250 mM Imidazole). The eluted protein was concentrated and loaded onto Superdex 200 Increase 10/300 GL column (GE Healthcare) with running buffer (25 mM HEPES (pH 7.5), 150 mM NaCl). The Nb35 monomeric fractions were pooled and stored at −80°C for further use.

Expression and purification of MRGPRX1-G_αq complex

The complex was expressed in Spodoptera frugiperda (SF9) insect cells. Baculovirus preparation was accomplished based on the baculovirus system manual (Invitrogen). For expression, the SF9 cells were cultured in serum-free medium (UK1000, UNION-BIOTECH) to 2.0 million cells per mL and co-infected with BRIL-MRGPRX1-LgBit, engineered G_αq, G_β1γ2-HiBit, and Ric8A at a virus ratio 1:1:1:1 for 48 h. The complex cell pellets were resuspended in lysis buffer (25 mM HEPES (pH 7.5), 150 mM NaCl, 5% Glycerol, 10 mM MgCl₂, 20 mM KCl, 5 mM CaCl₂) supplemented with Protease Inhibitors (1 mM PMSF, 2 μg/mL Aprotinin, 5 μg/mL Leupeptin, and 1 μg/mL Pepstatin). The suspension was homogenized and incubated with 10 μM compound 16 (MCE), scFv16 (20 μg/mL), Nb35 (15 μg/mL), and apyrase (25 mU/mL, NEB) for 1 h at room temperature. Then, 0.5% (w/v) lauryl maltose neopentylglycol (LMNG) (Anatrace) and 0.05% (w/v) CHS (Anatrace) was added into the suspension to solubilize the membrane. After extracting for 2 h at 4°C, the supernatant was isolated by centrifugation (18,000 rpm, 40 min, 4°C). The flag resin was washed with 10 column volumes washing buffer 1 (25 mM HEPES (pH 7.5), 150 mM NaCl, 5% Glycerol, 5 mM MgCl₂, 5 mM CaCl₂, 10 μM compound 16, 0.1% (w/v) LMNG, and 0.01% (w/v) CHS) and then washed with 10 column volumes washing buffer 2 (25 mM HEPES (pH 7.5), 150 mM NaCl, 5% Glycerol, 5 mM MgCl₂, 5mM CaCl₂, 10 μM compound 16, 0.01% (w/v) LMNG, and 0.001% (w/v) CHS). The complex was eluted with elution buffer (25 mM HEPES (pH 7.5), 150 mM NaCl, 5% Glycerol, 5 mM MgCl₂, 5mM CaCl₂, 10 μM compound 16, 200 μg/mL Flag peptide (GenScript), 0.01% (w/v) LMNG, 0.005% GDN (Anatrace), and 0.001% (w/v) CHS). The final complex elution was concentrated to less than 2 mL using an Amicon Ultra Centrifugal Filter (MWCO 100 kDa) and incubated with 50 μM compound 16 for 1 h at 4°C. Then, the complex was subjected to Superdex 200 Increase 10/300 GL column (GE Healthcare) with running buffer (25 mM HEPES (pH 7.5), 150 mM NaCl, 2 mM MgCl₂, 10 μM compound 16, 0.00075% (w/v) LMNG, 0.00025% (w/v) GDN, 0.000075% (w/v) CHS) to remove uncoupled components. The monomeric peak fractions were collected and concentrated to the final concentration 12 mg/mL for cryo-EM grid preparation.

Cryo-EM sample preparation and data acquisition

The concentrated sample (3.5 μL) at a concentration of 12 mg/mL was applied to glow-discharged holey carbon-coated grids (Quantifoil 200 mesh, Au R1.2/1.3). The grids were blotted for 3.5 s and flash-frozen in liquid ethane using a Vitrobot (Mark IV, Thermo Fisher Scientific). Images were recorded on a 300 kV Titan Krios G3i electron microscope (Thermo Fisher Scientific) equipped with Gatan K3 Summit direct detector and a GIF Quantum energy filter (slit width 20 eV). Movie stacks were collected using SerialEM [48] in counting mode at a magnification of 105,000× with the corresponding pixel size of 0.85 Å. Movies stacked with 50 frames were exposed for 2 s. Two data sets of the same sample were collected, including 3,447 and 2,799 movies separately. Two data sets were recorded at a total dose of about 56.15 e/Å2 and 58.32 e/Å2, respectively. The defocus range was set from −1.0 μM to −2.0 μM.

Data processing

Data processing was performed using cryoSPARC [49]. Movies frames were aligned using Patch motion. CTF estimation was performed using Patch CTF. Particles were first picked using a blob picker with partial micrographs. 2D templates were generated by 2D classification. Particle picking of all micrographs was performed by a template picker. A total of 6,687,388 particles from 6,246 micrographs were extracted using a box size of 288 pixels and cropped into 72 pixels. After 2 rounds of 2D classification and 1 round of ab initio reconstruction, 1,808,631 particles were selected and re-extracted using a box size of 288 pixels. Ab initio reconstruction using partial particles was performed, and 279,237 particles were removed, remaining 1,529,394 particles. After that, 1 round of nonuniform and local refinement was performed. One round of ab initio reconstruction was performed again, generating a new data set with 1,006,848 particles. Finally, 1 round of nonuniform and local refinement was performed, generating a 3.0 Å map.

Model building and refinement

The initial complex model was built using the structure of MRGPRX2 (PDB code: 7S8N) and Nb35 (PDB code: 7F4H) as templates. Models are then fitted into the density map and manually adjusted and rebuilt in COOT [50]. The restraint files of compound 16 and CHS were generated by Phenix.elbow package [51]. The complete model was finally refined in Phenix using real-space refinement with secondary structure and geometry restraints [52] and COOT. Overfitting of the model was checked by refining the model using one of the 2 independent maps from gold-standard refinement and calculating FSC against both half maps [53]. The final model was validated using Molprobity [54] (S2 Table). Structural figures were prepared in PyMOL (https://pymol.org/2/), UCSF Chimera [55], and UCSF ChimeraX [56].

Bioluminescence resonance energy transfer assay (BRET)

BRET assays [57] were performed as previously reported [47] to measure MRGPRX1-mediated G_q protein activation. Briefly, HEK293T cells (ATCCCRL-11268; mycoplasma free) were co-transfected at a 1:1:1:1 ratio of receptor: G_q-Rluc8: G_β: G_γ-GFP2. After at least 18 h, transfected cells were harvested and reseeded in opaque white bottom 96-well assay plates (Beyotime) at a density of 30,000 to 50,000 cells per well (DMEM plus 10% FBS). The medium was removed after 24 h. Cells were incubated with 40 μL 7.5 μM coelenterazine 400a (Goldbio) in drug buffer (1× Hank’s balanced salt solution (HBSS), 20 mM HEPES (pH 7.4), 0.3% BSA) for 2 min, and then treated with 20 μL compounds prepared in drug buffer at serial concentration gradient for additional 5 min. Plates were read in an LB940 Mithras plate reader (Berthold Technologies) with 395-nm and 510-nm emission filters. BRET ratios were calculated as the ratio of GFP2 emission (510 nm) to Rluc8 emission (395 nm) and analyzed by GraphPad prism 8.0.

Behavioral studies

WT C57BL/6J mice (8-week-old males) in acute itch behavioral tests were purchased from the Disease Control and Prevention Center of Hubei Province in China. Vehicle and MRGPRX1 agonists were subcutaneously injected into the nape of neck of mice after acclimatization. Scratching behavior in mice was observed for 30 min following injection. A bout of scratching was defined as a scratching movement with the hind paw directed at the area of the injection site. Then, the scratching bouts directing at the injected site were quantified.

Ethics statement

All acute itch behavioral tests in mice were performed with subcutaneous injection of compounds and approved by the Animal Care and Ethical Committee of Wuhan University under the International Association for the Study of Itch guidelines (Approval number: WDSKY0201707-2). After itch experiments, all mice were humanely killed by carbon dioxide asphyxiation.

Calcium imaging

HEK293T cells were loaded with 5 μM Fluo-4AM (Yeasen) for 30 to 45 min in the dark, supplemented with 0.01% Pluronic F-127 (wt/vol, Yeasen), in 1× HBSS containing 140 mM NaCl, 5 mM KCl, 10 mM HEPES, 2 mM CaCl₂, 2 mM MgCl₂, and 10 mM d-(+)-glucose (pH 7.4). After washing 3 times with HBSS, emission at 520 nm was detected from 488 nm excitation for Fluo-4AM. Data were analyzed from at least 3 repeated experiments at 80 to 200 cells.

Molecular docking

Molecular docking was performed using AutoDock Vina [58,59]. Briefly, the structure of MRGPRX1-compound 16 and the model of MrgprA3 using homology modeling based on the structure of MRGPRX1-compound 16 were used to do CQ docking. Homology modeling of MrgprA3 was performed using Swiss-model [60,61]. Open Babel was used to prepare the coordinates file of CQ and receptor with polar-hydrogens. A grid box with 25 × 25 × 25 grid points was used for searching. The results were checked, and the 2 top-ranking binding poses were selected. Figures were prepared in PyMOL (https://pymol.org/2/).