Conserved residues in the extracellular loop 2 regulate Stachel-mediated activation of ADGRG2

Cleavage and dissociation of a large N-terminal fragment and the consequent unmasking of a short sequence (Stachel) remaining on the N-terminus have been proposed as mechanisms of activation of some members of the adhesion G protein-coupled receptor (aGPCR) family. However, the identity of residues that play a role in the activation of aGPCRs by the cognate Stachel remains largely unknown. Protein sequence alignments revealed a conserved stretch of residues in the extracellular loop 2 (ECL2) of all 33 members of the aGPCR family. ADGRG2, an orphan aGPCR, plays a major role in male fertility, Ewing sarcoma cell proliferation, and parathyroid cell function. We used ADGRG2 as a model aGPCR and generated mutants of the conserved residues in the ECL2 via site-directed mutagenesis. We show that tryptophan and isoleucine in the ECL2 are essential for receptor stability and surface expression in the HEK293 cells. By adjusting the receptor surface expression levels, we show that mutation of these residues of ECL2 ablates the Stachel-mediated activation of multiple signaling pathways of ADGRG2. This study provides a novel understanding of the role of the ECL2 in Stachel-mediated signaling and degradation of ADGRG2, which may lay the foundation for the rational design of therapeutics to target aGPCRs.


Introduction
G protein-coupled receptors (GPCRs) play fundamental roles in various cellular processes such as proliferation, metabolism, hormone secretion, contraction, and neurotransmission 1 . The signaling pathways of GPCRs via G proteins and -arrestins, and their trafficking trajectories have been studied for several decades 2,3 . This in combination with recent advances in our understanding of GPCR structures has formed a strong foundation for rational drug design to target this largest superfamily of surface receptors 4,5 . Recent genomic and model organism studies have revealed the versatile roles that the adhesion GPCRs (aGPCRs), the second-largest family of GPCRs, play in endocrine, nervous and immune systems 6 . aGPCRs have structural differences with other classes of GPCRs, namely autoproteolysis at a GPCR proteolysis site (GPS) and formation of a two-segmented receptor with a large N-terminal fragment (NTF) and a C-terminal fragment (CTF). The CTF includes the Cterminus, seven-transmembrane helical domains (TM1-7), extracellular loops (ECL1-3), intracellular loops (ICL1-3), and a short extracellular sequence on the N-terminus just before the TM1. Many studies 7-10 , including our recent reports 11, 12 , have shown that this short remaining peptide, known as Stachel sequence, can activate aGPCRs and might be the mechanism of activation in physiological and pathological states. NTF-dissociating molecular partners can potentially unmask the Stachel sequence for its binding to the cognate aGPCR. These studies exploited two main tools to reveal this mechanism: (a) NTF-truncated aGPCRs that show constitutive activation of G proteins; (b) synthetic peptides resembling the Stachel sequence that activate aGPCRs 13,14 .
We recently showed that human ADGRG2 (GPR64), an orphan aGPCR, is expressed in human parathyroid glands and regulates the signaling and function of calcium-sensing receptor 12 .
We discovered that similar to other aGPCRs (ADGRG1 15 , ADGRD1 8 , ADGRG6 8 ), ADGRG2 is activated by either the endogenous 15-amino acid long Stachel (P-15) on the N-terminus of its NTF-truncated mutant (ADGRG2-∆NTF) or the synthetic P-15 12 . We showed that the deletion of P-15, in addition to the NTF, ablates constitutive activation of Gs and cAMP production in a mutant that starts with the proline 622 (ADGRG2-P622) 11 and elevates receptor response to synthetic P-15. Nevertheless, the binding site of Stachel remains unknown among aGPCRs.
Previous studies in other families of GPCRs revealed that some residues in the ECL2 play major roles in ligand access, receptor subtype selectivity, and activity [16][17][18] . Despite high degrees of diversity in the structure of ECL2 among GPCRs 19 , there is a conserved disulfide bond between the cysteines of ECL2 and TM3, which ensures receptor structural integrity. In the Secretin family (class B1), this conserved ECL2 cysteine is followed by a tryptophan residue, forming the CW motif, which is further followed by an acidic residue (aspartic or glutamic acid). The aGPCR family has the highest homology to the Secretin family. By aligning the ECL2 residues of all 33 human aGPCRs (predicted by either GPCRdb or Uniprot), we show that most of the aGPCRs have an aliphatic residue (leucine or isoleucine) after the CW motif ( Here, we use ADGRG2-P622 as a model aGPCR to investigate the role of tryptophan and isoleucine of ECL2 (Figure 1b) in receptor activation by the Stachel. Our study shows that these residues not only regulate receptor activation by P-15 but also control receptor surface expression and degradation.

Mutations in ECL2 alter the expression of ADGRG2
We transiently expressed the ADGRG2-P622 (WT) and its ECL2 mutants, CWA, CAI, and CAA in HEK cells and used a cell-based ELISA assay to determine their surface expression via their Nterminal HA-tag. We noticed that all ECL2-mutants show significantly reduced expression on the surface in comparison to the WT receptor ( Figure 2a). This lower surface expression was confirmed by immunofluorescence imaging of receptors on the cell surface ( Figure 2b). Measuring the total expression via the V5-tag at the C-terminus by ELISA revealed that all three mutants are expressed at reduced levels compared to the WT receptor ( Figure 2c). Western blotting of the whole-cell lysates confirmed the reduced expression of mutants ( Figure 2d). Together, these data suggest that the expression of ADGRG2 is controlled by tryptophan and isoleucine in the ECL2.

ECL2-mutations accelerate ADGRG2 degradation
To determine if the reduced expression is due to possible degradation of ECL2 mutants, we isolated cell lysates at several time points post-incubation with a protein translation inhibitor, cycloheximide. The expression of the WT receptor remained substantially unchanged up to 8 hours post-inhibition of translation. However, the expression of all ECL2 mutants was significantly reduced starting at 2 hours post-inhibition and continued to diminish continuously up to 8 hours ( Figure 3). This set of data reveals that mutation of conserved residues in the ECL2 of ADGRG2 increases receptor degradation.

Mutation of tryptophan and isoleucine in ECL2 ablates ADGRG2-mediated induction of transcription factors
We have previously shown that the synthetic P-15 induces cAMP-response element binding protein (CRE) transcription factor in a luciferase reporter assay. To test whether ECL2 mutations affect receptor activation by the Stachel peptide despite their reduced surface expression, we used two approaches. First, we normalized the CRE induction downstream of P-15-activated WT and CAA mutants to that of their surface expression (Supplementary Figure 2). This approach revealed a lack of activation of CAA mutant by P-15. Second, we adjusted the dose of the transfected plasmids so that the mutants and WT receptors reach similar surface expression levels ( Figure 4a).
Thereafter, for all signaling assays, we adjusted the doses of mutant and WT plasmids and supplemented the total plasmid dose with empty backbone pcDNA3.1 plasmid.
Using the CRE and SRE reporter assays, we stimulated cells expressing comparable surface levels of WT and mutant receptors with increasing concentrations of P-15 or vehicle for 5 hours. The basal CRE and SRE activity was similar among WT and mutant receptors.
Concentration-response curves revealed that while the WT receptor responds in a concentrationdependent manner to P-15, none of the mutants are activated by this synthetic peptide (Figure 4b and c). These data suggest that activation of some pathways downstream of ADGRG2 is regulated by the conserved residues in the ECL2.

Gs pathway activation by the Stachel is controlled by the ECL2
We have previously used an endpoint homogenous time-resolved FRET (HTRF) assay to measure cAMP production downstream of ADGRG2-P622-Gs pathway 11 . To monitor the production of cAMP in kinetic mode, we developed a new cAMP production assay by using a geneticallyencoded biosensor 20 . We induced cAMP production by forskolin, an activator of adenylyl cyclase, and optimized the amount of sensor needed to reach a robust adenylyl cyclase response (Supplementary Figure 3). A 20-second baseline measurement was followed by stimulation with increasing concentrations of P-15 in cells that were transfected with adjusted doses of WT and ECL2-mutated receptors (Figure 5a-d). While the WT receptor responded to P-15 in a concentration-dependent manner, the mutants showed no response, except for the CWA mutant that responded marginally to the highest P-15 concentration. To ensure that increased doses of transfected plasmids did not alter the responsiveness of the adenylyl cyclase-cAMP axis, we compared the forskolin-induced cAMP production. Our data show that despite increased doses of transfected mutant plasmids, cells respond similarly to the forskolin (Figure 5e).
Using the HTRF assay in the presence of IBMX, an inhibitor of phosphodiesterases, to accumulate cAMP production overnight, we found that the basal levels of cAMP are comparable among receptors (Figure 5f). A 2-hour stimulation with P-15 results in a concentration-dependent production of cAMP by the WT receptor ( Figure 5g) and a modest response by CWA mutant.
Together, these data show that induction of the Gs-adenylyl cyclase-cAMP pathway by Stachel is dependent on the presence of an intact ECL2.

ECL2 mutations derail the whole-cell response to P-15
We and others have shown that ADGRG2 couples to multiple signaling pathways including Gq, Gs, and G13 7,11, 21,22 . Considering that measuring the production of either a single second messenger (e.g. cAMP) or induction of a couple of transcription factors (e.g. CRE and SRE) cannot detect the 'overall' response of cells to GPCR ligands, whole-cell label-free assays have been developed [23][24][25][26] . To investigate if the overall response of cells to synthetic P-15 is regulated by the ECL2 conserved residues, we developed a new non-invasive whole-cell label-free impedance assay. Activation of WT-transfected cells with P-15 resulted in a rapid decrease in cell monolayer impedance to 1kHZ frequency of electrical field voltage ( Figure 6). However, neither empty plasmid nor ECL2-mutated receptors responded to the P-15 stimulation. This data confirms the lack of receptor activation by Stachel peptide in the absence of ECL2 conserved tryptophan and isoleucine residues.

Discussion
Our understanding of the mechanisms of activation of aGPCRs has substantially increased over the last decade. This progress was made possible due to basic pharmacology and cell biology studies in various cellular backgrounds and model organisms 8,9,[27][28][29] . Now we know that binding of an extracellular molecular partner to the NTF of ADGRG1 (GPR56) dissociates it from the CTF and unmasks a Stachel sequence, which subsequently activates ADGRG1 14,15,30 . Similar mechanisms of activation exist for a few other aGPCRs 7,31 . However, the binding site of the Stachel on the cognate aGPCR remains largely unknown. Structural similarities of class B1 GPCRs to aGPCRs (formerly class B2) and the fact that class B1 receptors bind to peptide hormones provides hints on the possible binding site of the Stachel. Specifically, the ECL2 forms an orthosteric binding pocket for peptide hormones in PTH 1 receptor (PTH1R) 32 and Corticotropin-releasing factor 1 receptor (CRFR1) 33 . In the current study, we tested the role of conserved residues in and around the CW motif of the ECL2 in the activation of an NTF/Stacheltruncated ADGRG2 (ADGRG2-P622) by its Stachel, P-15. Our results provide evidence that the lack of isoleucine and/or tryptophan in the ECL2 not only abolishes the signaling pathways of ADGRG2 but also dysregulates receptor stability in HEK293 cells.
We chose the ADGRG2-P622 receptor for this study because we had previously shown that ADGRG2-P622 responds significantly larger to the synthetic P-15 in comparison to the fulllength (ADGRG2-FL) or truncated ADGRG2-∆NTF in a battery of assays 11 . This maximal response by ADGRG2-P622 is presumably due to the availability of the binding site for P-15, which might be either masked by the NTF in ADGRG2-FL or occupied by the endogenous P-15 in ADGRG2-∆NTF. The lack of constitutive activity in cells that express ADGRG2-P622 in the current study confirms the role of endogenous Stachel in the activation of ADGRG2.
Here, we used several readouts to assess receptor response to the synthetic peptide. The use of kinetic and endpoint cAMP production assays allowed us to measure the Gs signaling of ADGRG2. Genetically encoded biosensor and HTRF assays showed a lack of activation of ECL2 mutants either in the short-term (minutes) or long-term (2 hours). We supplemented these assays with CRE and SRE reporter assays, in which cells are activated for 5 hours. Both assays showed a lack of ECL2 activation by P-15, suggesting that the coupling of both Gs and G13 proteins to ADGRG2 is ablated in the absence of either tryptophan or isoleucine or both. To assess other possible but unknown signaling pathways that may originate from ADGRG2, we developed a whole-cell label-free impedance assay. We have previously used similar label-free assays to study GPR55 and cannabinoid 2 receptor 34,35 . This method allowed us to measure the overall response, rather than a singular signaling pathway, from a monolayer of cells upon activation with P-15 in kinetic mode. Using this readout, we observed a strong fast Stachel-specific reduction of impedance from WT receptor-expressing cells, indicating possible movement of cells and consequent disruption of cell-cell contact in the monolayer. This response was transient, and impedance was mostly restored in 15 min. Nevertheless, we did not detect a response from cells expressing any of the ECL2 mutated receptors, further confirming the lack of activation by P-15.
It is noteworthy that mutant receptors were degraded in the absence of either endogenous or synthetic Stachel peptide. Do the ECL2 mutations induce constitutive activity followed by sustained internalization and degradation? This does not seem to be plausible because we did not detect any altered basal activity either in CRE reporter assay or cAMP production assays. Do the ECL2 mutations hinder receptor transport and incorporation into the plasma membrane? Our western blotting, ELISA, and immunofluorescence imaging data suggest that the diminished surface expression of mutants is likely to be due to the fast degradation of receptors upon translation. We have previously shown that truncated ADGRG2-∆NTF gets ubiquitinated but is expressed on the plasma membrane at levels comparable to that of the full-length ADGRG2 11,12 .
Further studies are required to unravel 1) the mechanisms by which ADGRG2 gets degraded (lysosomal or proteasomal 36 ) and 2) how the isoleucine and tryptophan of ECL2 impact these mechanisms. Irrespective of the underlying reason for diminished total expression and degradation, we adjusted the surface expression levels by increasing the dose of the transfected plasmids and conducted our signaling assays in this condition. Nevertheless, comparable surface expression of ECL2-mutated and WT receptors did not rescue their response to P-15, supporting the role of ECL2 in Stachel-mediated activation of ADGRG2. Whether the ECL2 of other aGPCRs play similar roles in receptor stability and activation by Stachel peptides warrants further studies.
While finalizing this manuscript, we noticed that a separate group reported that several residues, including the tryptophan and isoleucine of the ECL2, participate in the binding of Stachel peptide to ADGRG2 37 . Although this study corroborates some of our findings, differences exist between the two studies. We used 'human' ADGRG2-P622 that lacks the NTF and the Stachel sequence to prevent interaction of the endogenous P-15 on the N-terminus of the receptor with the receptor. Sun et. al., however, determined the binding of a 'mouse' P-15 that was 'modified' at several residues to the 'mouse' ADGRG2 ortholog. This modified P-15 showed higher potency compared to natural P-15 and is potentially a new pharmacological tool for future ADGRG2 studies. However, its application remains to be validated by testing whether it shows affinity or activity towards other aGPCRs that share similar Stachel sequence and ECL2 motifs as ADGRG2.
Another major difference between these two studies is that while we observed a significant alteration of ECL2-mutated receptor stability and cell surface expression, Sun et. al. did not report such changes. These differences make a direct comparison of the two studies challenging.
ADGRG2 plays various physiological and pathological roles. For instance, mutations or truncations in ADGRG2 are associated with different forms of male infertility 38,39

and male
Adgrg2-null mice are infertile 40 . ADGRG2 also promotes proliferation and metastasis of Ewing sarcoma cells both in vitro and in vivo 41 . We have previously shown that ADGRG2 is enriched in parathyroid glands, is upregulated in parathyroid adenomas from patients with primary hyperparathyroidism, and its activation by P-15 elevates PTH secretion 12 . Although ADGRG2 is still an orphan aGPCR, our findings on the role of ECL2 in receptor stability and Stachel-mediated activation provide a foundation for future targeting of this receptor via small molecules or biologics.

Receptor mutagenesis
We used the Q5 site-directed mutagenesis kit (NEB, Ipswich, MA #E0552S) to generate mutations in the ECL2 of human ADGRG2. Our template for mutations was pcDNA3.1-3xHA-P622-V5 (WT) 11 , a plasmid that expresses an NTF/Stachel-truncated ADGRG2 with an N-terminal 3xHA tag and a C-terminal V5 tag. The following primer pairs were used to construct mutants: P622- Ti-E microscope equipped with a 16.2 MegaPixels DS-Ri2 camera and images were analyzed with Nikon NIS-Elements Basic Research software.

Luciferase reporter assays
Luciferase reporter assay was performed as described previously 12

Statistical analysis
Statistical analyses were conducted using two-tailed Student's t-test for comparisons between two groups, One-Way or two-Way ANOVA with Dunnett's multiple test or Holm-Sidak multiple comparison test in GraphPad Prism 9.0 software; P < 0.05 was considered significant.