ADGRG2 was initially identified in a recombinant cDNA library constructed from a human epididymis (Osterhoff et al. 1997). It was highly represented in this library (approximately 0.01% of clones), indicating that the corresponding mRNA was enriched in this tissue. Differential screening employing polyA+RNA from various control tissues, i.e., human testis, liver, and brain, further suggested prevalent expression in the epididymis. Hence, it was named HE6 for Human Epididymis cDNA 6. The nearly 5 kb “full length” cDNA comprised one long open reading frame predicting an approximately 1000-residue polypeptide chain. Protein structure prediction revealed a multipass transmembrane protein with seven hydrophobic membrane-spanning (7TM) stretches, hallmark of the G protein-coupled receptors (GPCRs). Accordingly, HE6 was later renamed to GPR64 by the HUGO Gene Nomenclature Committee. A conspicuous cysteine-rich motif, initially named the Cys-box (Osterhoff et al. 1997), is located between the 7TM region and the long N-terminal part. This motif includes a HL*T tripeptide which is now known as the G protein-coupled receptor proteolysis site (GPS, see below). The GPS represents an integral part of a much larger ~320-residue domain termed the GPCR-Autoproteolysis INducing (GAIN) domain (Araç et al. 2012) which sets the family of Adhesion-GPCRs apart from the other GPCRs. Most recently, the Adhesion GPCR Consortium, together with the International Union of Basic and Clinical Pharmacology Committee on Receptor Nomenclature and Drug Classification, proposed ADGRG2 as the new name for HE6/GPR64 (Hamann et al. 2015), with ADGR as the common dominator for all Adhesion-GPCRs followed by a letter and a number to denote subfamily and subtype, respectively. ADGRG2 was grouped into Subfamily VIII of Adhesion-GPCRs which comprises members of ancient origin with only few or no known functional domains in their long N-terminal parts.
Epididymis-prevalent expression of ADGRG2 was confirmed by northern blot (Osterhoff et al. 1997) and microarray analyses (Gottwald et al. 2006). In situ hybridization showed that the mRNA was highly enriched in the duct epithelium with maximum levels in the proximal parts of the organ (Osterhoff et al. 1997), a region important for male fertility. Interfering with ADGRG2 hence might impair male fertility. To find a suitable animal model in which this hypothesis could be tested, a combined RT-PCR and database mining strategy was employed to clone the homologous cDNAs of various mammalian species. Cross-hybridizing genes and highly similar cDNAs (>80% identity) were detected in all species analyzed, including mouse (Obermann et al. 2003), suggesting that ADGRG2 was conserved among mammals. The encoding gene was assigned to the short arm of the human X chromosome (Xp22.13) which is almost completely syntenic between mammalian species. Multiple tissue northern analysis confirmed epididymis-prevalent expression also for the mouse counterpart, a cross-hybridizing mRNA of approximately 5 kb length being prominent in the proximal epididymis (Obermann et al. 2003). After prolonged exposure, however, faint signals were additionally seen in poly(A) + RNA extracts of fetal brain which had not been previously analyzed in the human (own unpublished results).
Knock-Out (KO) Phenotype
As the proximal epididymis is a major contributor to sperm maturation and sperm concentration, it was speculated that ADGRG2 signaling might be essential to male fertility. Targeted gene disruption indeed resulted in age-dependent male subfertility or even complete infertility (Davies et al. 2004). Female reproductive functions, for comparison, remained unaffected. The hemizygous mutant males revealed an abnormal accumulation of spermatozoa within distended efferent ductules, the sperm masses confining and eventually occluding the duct lumen. The very few sperm present in the more distal parts of the genital tract were morphologically abnormal (Davies et al. 2004). Dilation of the rete testis and lumen of seminiferous tubules was also frequently observed in the KO males, probably caused by backpressure build-up in the testis. The affected seminiferous tubules showed thinning and compression of the epithelium, occasionally resulting in complete tubular atrophy. Hormonal levels, on the other hand, were not significantly altered, indicating that ADGRG2 is not part of the endrocrine system controlling male reproductive functions. Rather, it seemed to operate immediately in the male genital tract.
In wild-type males, up to 90% of the fluid produced by the testis is moved across the duct epithelium and reabsorbed into the vasculature within the efferent ductules and proximal epididymal duct (for review, see Hess 2014). The conspicuous accumulation of spermatozoa in the Adgrg2 KO males is indicative of an increased fluid reabsorption in efferent ductules and/or proximal epididymal duct which led to increased luminal viscosity and finally sperm stasis (Davies et al. 2004). This chain of events apparently was followed by pressure build-up in the testis due to blockage of the excurrent duct system in rodents, where 3–6 ductules funnel into a single small duct prior to entering the epididymis (compare Fig. 1). Thus, dysregulation of fluid reabsorption seemed to be the primary effect of the Adgrg2 mutation while no apparent changes in epididymal gross morphology and/or epithelial structure were detected. As a caveat, it should be kept in mind that different anatomic designs are found among mammals concerning the connection of efferent ductules and epididymal duct proper (for review, see Hess 2014). In larger mammals, including human, multiple entries of efferent ductules into the epididymis are seen. Accordingly, duct occlusion seems less likely making cross-species inferences of the Adgrg2 KO phenotypic effects difficult.
The effects on global epididymal gene expression of the Adgrg2 mutation were studied to possibly find an explanation for the KO phenotype (Davies et al. 2007). Differential cDNA library and microarray screening comparing epididymides from wild-type and Adgrg2 KO adult males revealed an overlapping set of differentially expressed transcripts, the majority of which appeared to be downregulated in the KO epididymides. For example, claudin 10A, which has a role in tight junction formation and paracellular fluid movement and which is strongly expressed in the proximal epididymis in the wild-type, was found to be downregulated over seven-fold in the KO organs. The excitatory amino-acid transporter 3 (EAAT3), also known as Slc1a1, encodes a glutamate transporter which contributes to the regulation of osmolyte balance in the proximal epididymis and was found to be downregulated at least four-fold. Downregulation of particular biochemical pathways might indicate that ADGRG2 plays a role in these pathways. However, as many epididymal genes are regulated by testis-derived luminal factors and blockage of the excurrent duct system as seen in the KO mice would impede these signals, the significance of these expression changes demands further investigation.
Expression Profile and (Patho-)Physiological Functions
In the KO animals, the 7TM-encoding region of the Adgrg2 gene had been replaced by a beta-galactosidase reporter under the control of an internal ribosome entry signal (Davies et al. 2004). The resulting lacZ reporter expression faithfully recapitulated the endogenous expression pattern within the male genital tract (see above). However, despite the initial characterization of ADGRG2 as an epididymis-specific gene product, signs of beta-galactosidase staining were additionally seen within a subset of neurons in the trigeminal and dorsal root ganglia (DRG), within the synovial membranes of the developing joints, and also in the developing parathyroid (see Davies and Kirchhoff 2010, for review). These observations are compatible with the expression data documented in the Human Protein Atlas and also with our own Northern blot results (see above). Further, Adgrg2/Gpr64 was identified in a differential screen aimed at the identification of marker genes expressed by differentiating DRG neurons in mice. Consistently, the Adgrg2 protein colocalized with parvalbumin, a well-established marker of proprioceptive sensory neurons (Kramer 2005). Knowledge of the global ADGRG2 expression profile may provide additional functional information. Haitina et al. (2008) performed an extensive tissue localization analysis in rodents. In line with the above results, their findings indicated that Adgrg2 is expressed and possibly has a function in the central nervous system (CNS). Thus, although for the Adgrg2 KO animals no nonreproductive phenotype has been reported (Davies et al. 2004), the receptor may still have a function in the CNS and in other tissues.
Aberrant ADGRG2 expression occurred in human pathology, including cancer. GPR64 mRNA was expressed in synovial fibroblasts isolated from osteoarthritis patients, suggesting a role in joint (patho)-physiology (Galligan et al. 2007). Moreover, ADGRG2/GPR64 mRNA was found to be highly upregulated in Ewing sarcomas (Richter et al. 2013). The study suggested that the receptor was able to induce invasiveness and metastasis in ES by orchestrating placental growth factor (VEGF receptor 1 ligand) and MMP1 expression, recommending ADGRG2 as a candidate target for the development of novel antitumor therapies. Also, ADGRG2/GPR64 was significantly and uniquely overexpressed in a subgroup of medulloblastoma tumors characterized by overactive signaling in the WNT mitogenic pathway and was thus suggested as a candidate for the development of new imaging or radiotherapeutic agents (Whittier et al. 2013).
Genomic Structure, Splice Variants, and Single Nucleotide Polymorphisms (SNPs)
Human-rodent comparative sequence analyses were performed to possibly detect homologies and differences between species in exon-intron structures and alternative splicing (Obermann et al. 2003). Co-linear transcript variants were prevailing; however, exon E4 which seemed to be constitutive in the human (Fig. 2b) was alternatively spliced in mice and rats; the same may apply to E7. On the other hand, mini-exon E6, which is subject to alternative splicing in the human, was consistently lacking in rodents and thus could have originated from a more recent event of exon creation or loss during mammalian evolution. Additional rodent splice variants have been recorded which predicted an alternative translation start site and which, if functional, could affect the Adgrg2 expression pattern; corresponding ESTs, however, are lacking in the databases.
The GPCR Natural Variants database (NaVa, update of 2009) lists two nonsynonymous single nucleotide polymorphisms (SNPs) of missense type for ADGRG2/GPR64, which will lead to amino acid substitution, i.e., H290Q and N771S. Such variants may potentially affect receptor function and turnover, or the ability to recognize and respond to pharmacologic agents. They may also be important for cancer risk assessment and treatment modalities. However, any ADGRG2 genetic variants or mutations related to human disease are as yet unknown.
Like most other ADGRs, ADGRG2 still has the status of an orphan receptor, i.e., a ligand is unknown. The endogenous epididymal protein represents a highly glycosylated two-subunit apical membrane component which is expressed in a specialized polarized cell type (see above); hence, functional studies are challenging. Representing a highly abundant structural component of microvilli/stereocilia (Kirchhoff et al. 2008), it was initially questioned whether ADGRG2 would represent a veritable cell surface receptor which signals through heterotrimeric G-proteins. There was merely a brief note suggesting that when over-expressed in Xenopus melanophores, ADGRG2 will bind stimulatory subunits Gs/Gq (C. Jayawickreme, personal communication, cited in Foord et al. 2002) which will then activate adenyl cyclase and/or phospholipase C, respectively. Recently, however, substantial progress has been made in revealing the mechanisms of ADGR activation, including ADGRG2. A recombinant overexpression strategy was applied to measure ADGRG2 basal activity which resulted from an equilibrium between inactive and active receptor conformations (Demberg et al. 2015; Peeters et al. 2015). As in other ADGRs, the autoproteolysis-inducing GAIN domain seemed to be central to the mechanism of receptor activation, and truncated ADGRG2 mutants that mimic the postcleavage receptor up to the point of the GPS motif were found to exhibit enhanced constitutive activity. The new N-terminus, which is exposed after GPS cleavage, named the stalk or “Stachel” region, activates the receptor as a tethered cryptic agonist which then signals via downstream effectors of Gαs and Gαi (Demberg et al. 2015). This observation led to the proposal of a “disinhibition model” of activation, the N-terminal fragment acting as a negative regulator.
However, endogenous ADGRG2 largely existed as closely associated heterodimer in epididymal membrane preparations, and no evidence for significant amounts of soluble ectosubunits was found (Obermann et al. 2003; see above). This suggested that the cleaved receptor is trafficking to the plasma membrane as one and remains associated in a state poised for ligand engagement and receptor activation. Accordingly, it remains unresolved whether and how the ADGRG2 ectosubunit would dissociate from its cognate “Stachel” and 7TM regions in a regulated manner. Also, it is conceivable that the dissociating ectosubunits could take the “Stachel” with them after a second proteolytic cleavage step as it seems to be the case for ADGRL1/CIRL/Latrophilin-1 (Krasnoperov et al. 2009). There are further indications that the tethered cryptic agonist model may be an incomplete concept for ADGRG2 activation and signaling. Peeters et al. (2015) demonstrated that noncleavable versions of the ADGRG2 protein will still activate downstream signaling via RhoA-SRE and Gαq-NFκB pathways, presumably via coupling to Gα12/13 and Gαq. Thus, ADGRG2 signaling may include both stalk-dependent and stalk-independent mechanisms as was recently proposed for the closely related ADGRG1/GPR56 and ADGRB1/BAI1 receptors (Kishore et al. 2016).
ADGRG2 is a highly conserved orphan member of subfamily VIII of adhesion-GPCRs (ADGRs). In mammals, it is primarily expressed in the epithelium of the proximal male excurrent ducts. It shows a chimeric protein structure consisting of a hydrophobic endosubunit with seven hydrophobic membrane-spanning (7TM) stretches, and a hydrophilic ectosubunit with numerous N- and O-glycosylation sites, but no known structural adhesion motifs. ADGRG2 co-localizes with the apical microvilli and stereocilia of the principal epithelial cells which are implicated in testicular fluid reabsorption and sperm concentration. Targeted gene disruption in the mouse led to male infertility; additional physiological and pathological role(s) are suggested by its expression pattern. The receptor is cleaved at the G protein-coupled receptor Proteolysis site (GPS) within a GPCR-Autoproteolysis INducing (GAIN) domain, but both subunits remain noncovalently bound at the apical membrane. It appears that ADGRG2 signals via multiple G protein-signaling cascades depending on the cellular context and possibly involving context-dependent higher-order complexes. Further elucidation of the endogenous signaling pathway(s) will have to await ADGRG2 deorphanization. The role of the numerous splice variants in regulating protein interaction, ligand binding, and receptor function remains to be established.
- Kramer I. Molecular pathways of proprioceptive dorsal root ganglion (DRG) sensory neuron specification. Inaugural dissertation, University of Basel. 2005. edoc.unibas.ch/566/1/DissB_7369Google Scholar