A single amino acid in MRGPRX2 necessary for binding and activation by pruritogens

To the Editor

The Mas-related G-protein–coupled receptor family is implicated in allergy, inflammation, and itch.^1,2 Ligands of the human receptor MRGPRX2 that induce itch include substance P (SP),^2,3 which is increased in atopic skin, bovine adrenal medulla (BAM) peptide, and compound 48/80.^1,2 In contrast, LL-37, an antimicrobial peptide and MRGPRX2 ligand implicated in inflammation, is not known to mediate itch and is decreased in atopic skin.^5,6 We hypothesized that distinct receptor amino acids would be associated with Mrgpr activation by pruritogens versus non-pruritogens. To test this hypothesis, we used molecular modeling and site-specific mutagenesis to evaluate the interaction of Mrgprs with these ligands in the context of itch. The data indeed reveal that a specific amino acid in MRGPRX2 and the corresponding ones in the homologous mouse receptors MrgprA1^2,3 and MrgprB2¹ are critical for binding and activation by pruritogens. Furthermore, predictive alterations in BAM 8–22 rescued activity on mutated MRGPRX2. In contrast, LL-37, which does not induce itch, retained activity on the mutant receptors.

We used the publicly available RaptorX⁷ and Phyre2⁸ Web portals for predictive modeling of MRGPRX2, MrgprA1, and MrgprB2 in association with SP and BAM peptides (see Table E1 and this article’s Methods section in the Online Repository at www.jacionline.org). Each of the portals predicted an MRGPRX2 ligand-binding pocket to consist of a number of structurally conserved hydrophilic residues along with a buried glutamic acid residue, E164 (see Figs E1 and E2 in this article’s Online Repository at www.jacionline.org). The charged nature of E164 was consistent with a key role in ligand coordination and activation of MRGPRX2. As predicted, a single point mutation in which the glutamic acid was changed to arginine, MRGPRX2 E164R, was found to abolish activation of MRGPRX2 by SP (Fig 1, A), BAM13-22 (see Fig E3 in this article’s Online Repository at www.jacionline.org), and compound 48/80 (see Fig E4 in this article’s Online Repository at www.jacionline.org). Because positions 19 and 20 of the C-terminus of BAM13-22 are positively charged, these would be predicted to possibly interact directly with MRGPRX2 E164 depending on which way the backbone of BAM is rotated. “Rescue” mutations of BAM were generated and tested against both native MRGPRX2 and the mutant MRGPRX2 E164R (see Fig E3 and Table E2 in this article’s Online Repository at www.jacionline.org). The mutant peptide BAM13-22E20 failed to activate both the native and mutated MRGPRX2. In contrast, although mutant BAM K19D did not activate native MRGPRX2, it recovers function when tested against MRGPRX2 E164R. This confirmation of MRGPRX2 E164 led us to consider its importance in the function of homologous Mrgprs.

FIG 1.

SP activates MRGPRX2, MrgprA1, and MrgprB2 but not site-specific mutagenesis of Glu164 to Arg164 in MRGPRX2, and the equivalent residues in MrgprA1 and MrgprB2. HeLa cells were transfected with cDNAs encoding the mutant receptors. An HEK-293 cell line stably expressing the receptors was used for MRGPRX2, MrgprA1, and MrgprB2. Intracellular calcium [Ca²⁺]i was determined by ratiometric Fura-2 imaging after addition of SP as an indicator of receptor activation. Two other MRGPRX2 mutants, MRGPRX2T181A and MRGPRX2K251E, were generated as controls and SP activated these receptors in a manner similar to the native receptor (data not shown).

In comparison to the recently published article by Lansu et al,⁹ we had also initially considered residue D184 in our selection process. However, our ensemble of models demonstrated significant divergence of sequence registry, with some models placing this residue on helix 5 and others placing it on differing locations along loop 2. When comparing structures of GPCRs solved via x-ray crystallography, this loop has been trapped in what are called the “open” and “closed” states. The significant conformational change that is predicted to take place on ligand binding further confounds the level of difficulty in the predicted location, and significance, of residue D184. It is not clear whether this residue is important for the initial steering of the ligand into the binding pocket and/or in securing the ligand in place, as is predicted to be the case for residue E164.

We used the CLUSTALW¹⁰ sequence alignment program to predict that MrgprA1 N172 and MrgprB2 E171 are the residues likely to “sit” in the MRGPRX2 E164 position. We next used Multiple Alignment using Fast Fourier Transform to further examine the active sites prediction (see Fig E5 in this article’s Online Repository at www.jacionline.org). The mutated residues are flanked by cysteines in each of the 3 receptors. The predicted topology of MRGPRX2, MrgprA1, and MrgprB2 and their transmembrane domains (TMDs) was then considered. The mutated residue in MRGPRX2E164R is at the beginning of the third extracellular domain adjacent to the extracellular end of TMD4. Both MrgprA1N172R and MrgprB2E171R are at the end of TMD4 adjacent to the ECD3 (see Table E3 in this article’s Online Repository at www.jacionline.org).

We next extended the finding above that native MRGPRX2 responds to SP and compound 48/80 but MRGPRX2E164R does not, to the equivalent residues in MrgprA1 and MrgprB2. Neither MrgprA1N172R nor MrgprB2E171R responded to SP (Fig 1, B and C) or compound 48/80 (Fig E4). Two other MRGPRX2 mutants, MRGPRX2T181A and MRGPRX2K251E, were generated as controls and SP activated these receptors in a manner similar to the native receptor (data not shown). In contrast, LL-37, which is not a pruritogen, activates both wild-type and mutant Mrgprs (see Fig E6 in this article’s Online Repository at www.jacionline.org) and this activation is not inhibited by an Mrgpr antagonist (see Fig E7 in this article’s Online Repository at www.jacionline.org).

To further study the interaction of SP with wild-type and mutant Mrgprs, the binding of SP to these receptors was evaluated using ELISA. The wild-type receptors and the control MRGPRX2 mutants (MRGPRX2T181A and MRGPRX2K251E) bind SP as expected (Fig 2). The 3 Mrgpr mutants, MRGPRX2E164R, MrgprA1N172R, and MrgprB2E171R, do not bind SP (Fig 2).

FIG 2.

SP binds to wild-type but not mutant Mrgprs. ELISA was used to quantify the binding of SP to wild-type receptors (MRGPRX2, MrgprA1, and MrgprB2) and mutated receptors (MRGPRX2E164R, MRGPRX2T181A, MRGPRX2K251E, MrgprA1N172R, and MrgprB2E171R). P values for MRGPRX2 vs MRGPRX2E164R, MrgprA1 vs MrgprA1N172R, and MrgprB2 vs MrgprB2E171R, are .0021, .0003, and .0016, respectively. P values greater than .05 are considered nonsignificant (ns).

LL-37 activates MRGPRX2 to induce mast-cell degranulation.⁶ However, LL-37 has not been reported to have pruritogenic properties while atopic skin has decreased expression of antimicrobial peptides.⁵ Given the role of MRGPRX2 in itch and its interaction with pruritogens, including SP and compound 48/80, we evaluated LL-37–induced itch using the mouse cheek model. As opposed to pruritogens, which induce scratching in this model, LL-37 does not induce scratching (see Fig E8 in this article’s Online Repository at www.jacionline.org).

The data presented here demonstrate that SP and compound 48/80 activate native Mrgprs but fail to activate Mrgprs in which a single amino acid residue was mutated. In contrast, LL-37 activates both native receptor and mutant receptors. This finding indicates that the site(s) of Mrgpr activation for the pruritogens SP and compound 48/80 differs from the activation site(s) for LL-37. Furthermore, an Mrgpr antagonist, QWF, inhibits the interaction of SP and compound 48/80 with MRGPRX2 but not that between LL-37 and the receptor.

These findings have several implications. First, they reveal that the Glu and Asn residues at the end of the fourth TMD and beginning of the fourth ECL of Mrgprs are necessary for the interaction of SP and compound 48/80 with members of this receptor family. Second, not all compounds that activate Mrgprs or induce mast-cell degranulation result in itch. It is possible that LL-37 may interact with other cells such as keratinocytes and additional receptors to induce the release of molecules with antipruritic properties such as semaphorin 3A.¹¹ Another possibility is that activation of MRGPRX2 on mast cells by SP versus LL-37 may result in differential release of granules and their associated mediators from mast cells. Third, at least some Mrgprs appear to have more than 1 site for signaling followed by distinct behavioral effects. Fourth, although MrgprB2 has been proposed as the mouse ortholog of human MRGPRX2, modeling demonstrates that mouse MrgprA1 and human MRGPRX2 also exhibit topological similarities (see Fig E9 in this article’s Online Repository at www.jacionline.org). Taken together, these data provide further support for the possibility that antagonists of MRGPRX2 may be useful for the treatment of itch and inflammation while providing guidance with respect to the development of receptor antagonists.