Address correspondence and reprint requests to Frédéric Simonin, Institut de Génétique et de Biologie Moléculaire et Cellulaire, CNRS/INSERM/ULP, 1 rue Laurent Fries, B.P. 10142 67404 Illkirch Cedex, C.U. de Strasbourg, France. E-mail: firstname.lastname@example.org
During the past few years several new interacting partners for G protein-coupled receptors (GPCRs) have been discovered, suggesting that the activity of these receptors is more complex than previously anticipated. Recently, candidate G protein-coupled receptor associated sorting protein (GASP-1) has been identified as a novel interacting partner for the delta opioid receptor and has been proposed to determine the degradative fate of this receptor. We show here that GASP-1 associates in vitro with other opioid receptors and that the interaction domain in these receptors is restricted to a small portion of the carboxyl-terminal tail, corresponding to helix 8 in the three-dimensional structure of rhodopsin. In addition, we show that GASP-1 interacts with COOH-terminus of several other GPCRs from subfamilies A and B and that two conserved residues within the putative helix 8 of these receptors are critical for the interaction with GASP-1. In situ hybridization and northern blot analysis indicate that GASP-1 mRNA is mainly distributed throughout the central nervous system, consistent with a potential interaction with numerous GPCRs in vivo. Finally, we show that GASP-1 is a member of a novel family comprising at least 10 members, whose genes are clustered on chromosome X. Another member of the family, GASP-2, also interacts with the carboxyl-terminal tail of several GPCRs. Therefore, GASP proteins may represent an important protein family regulating GPCR physiology.
candidate G protein-coupled receptor associated sorting protein
G protein-coupled receptor
histamine H2 receptor
5-hydroxytryptamine 7 receptor
m1 muscarinic receptor
m2 muscarinic receptor
opioid receptor like 1
thromboxane A2 α receptor
G protein-coupled receptors (GPCRs) form one of the largest gene family in the human genome (see Pierce et al. 2002). The most recent estimation indicates that, excepting the odorant-receptor subfamily that contains several hundred members, the repertoire of GPCRs responding to endogenous ligands consists of 367 receptors in human and 392 in mouse (Vassilatis et al. 2003). These receptors are involved in the regulation of numerous physiological processes and represent major targets for pharmaceutical drugs. Upon ligand binding, GPCRs trigger numerous cellular events, including modification of second messenger levels (see Gudermann et al. 1997), receptor desensitization and internalization (see Pierce et al. 2002) and modification of gene transcription (see Pierce and Lefkowitz 2001; West et al. 2002). The molecular mechanisms underlying these phenomena are not completely understood and are the subject of intensive research.
In the last few years, several novel proteins that physically interact with GPCRs have been identified, confirming that signal transduction associated with GPCRs is not restricted to heterotrimeric G protein activation (Hall et al. 1999; Brady and Limbird 2002). These associated proteins interact, either specifically with a small subset, or with a broad range of receptors, depending on the receptor region involved in the interaction. In most cases, this region is located in the COOH-terminal cytoplasmic domain, which is one of the most divergent domains within GPCR sequences. These observations indicate that each GPCR may activate a specific set of intracellular effectors and may be regulated or targeted differently depending on interacting partners.
Opioid receptors are members of subfamily A of GPCRs that represents almost 90% of GPCRs (Fredriksson et al. 2003). Persistent stimulation of these receptors in vivo is well known to produce tolerance and dependence, but the molecular and cellular mechanisms underlying these phenomena are not well understood (Nestler 2001; Taylor and Fleming 2001). Opioid receptor desensitization and trafficking, which represent key processes in the development of tolerance and dependence, seem highly regulated (Kieffer and Evans 2002), suggesting that new interacting partners still remain to be discovered.
Using the carboxyl-terminal tail of delta opioid receptor (DOR) as bait in a two-hybrid screen, we have identified GASP. This protein was shown recently to be a regulator of the intracellular sorting of DOR upon stimulation with agonists (Whistler et al. 2002) and will be hereafter called GASP-1 in this study. We show here that GASP-1 interacts with ORL1, delta- (DOR) mu- (MOR) and kappa- (KOR) opioid receptors via a small portion of the carboxyl-terminal tail, proximal to the seventh transmembrane domain. Interestingly, this protein interacts with numerous GPCRs from family A and B and is a member of a new family. Another member of this family, GASP-2, also interacts with the carboxyl-terminal tail of several GPCRs in vitro. Altogether, our results suggest that we have identified a novel gene family that potentially interacts with GPCRs in the central nervous system.
Yeast two-hybrid system
We constructed a cDNA library of approximately 2.5 × 106 recombinant clones from SHSY-5Y cells mRNA in pACT2 (Clontech, Palo Alto, CA, USA) using standard molecular biology methods (Sambrook and Russel 2001). The MATCHMAKER II two-hybrid system (Clontech) was used to screen this library with the carboxyl-terminal domain of human mu- and delta-opioid receptors as baits. The sequences encoding amino acids 334–400 and 326–380 of mu and delta receptors, respectively, were amplified by PCR and cloned into pAS2-1 (Clontech). Using the Clontech protocol, 6 × 106 transformants were screened.
GST pull-down assay
The cDNA encoding the carboxyl-terminal part of GASP-1 identified from the two-hybrid screen was subcloned into pcDNA3 expression vector (Invitrogen, Carlsbad, CA, USA). The full-length cDNA encoding GASP-1 was amplified by PCR from human genomic DNA and cloned into pcDNA3. The cDNA encoding the closest GASP-1 paralogue (GASP-2), subcloned into pCMV SPORT6 expression vector, was purchased from Invitrogen. Coupled in vitro transcription-translation of these clones was performed using the TNT quick-coupled transcription/translation system (Promega, Madison, WI, USA) according to manufacturer's instructions. Vectors for the bacterial expression of glutathione S-transferase (GST) fused to the carboxyl-terminal tail of various GPCRs were constructed in pGEX-2T (Amersham Biosciences, Sacley, France). The following pGEX constructs were engineered: pGEX-DOR encoding residues 314–372 (referring to number U10504); PGEX-DOR2 encoding residues 314–333; pGEX-DOR3 encoding residues 331–372; pGEX-KOR encoding residues 326–380 (referring to number U17298); pGEX-KOR2 encoding residues 341–380 (referring to number L11065); pGEX-MOR encoding residues 334–400 (referring to number L29301); pGEX-opioid receptor-like 1 (ORL1) encoding residues 315–370 (referring to number X77130); pGEX-m1 muscarinic (M1) encoding residues 414–460 (referring to number X52068); pGEX-m2 muscarinic (M2) encoding residues 436–466 (referring to number M16404; pGEX-histamine H2 (H2) encoding residues 284–349 (referring to number S57565); pGEX-5-hydroxytryptamine 7 (5-HT7) encoding residues 380–445 (referring to number L21195); pGEX-thromboxane A2α (TXA2α) encoding residues 304–343 (referring to number D38081); pGEX-β1 adrenergic (β1-AR) encoding residues 373–477 (referring to number J03019); pGEX-calcitonin receptor (CALCR) encoding residues 407–490 (referring to number L00587). Site-directed mutagenesis of DOR and β1-AR COOH-terminal tails was performed by PCR. GST-fusion proteins were immobilized on glutathione-agarose beads and incubated with [35S]-labelled in vitro translated GASP-1 protein or its close paralogue in binding buffer [20 mm Tris/HCl pH 8, 150 mm NaCl, 1 mm EDTA, 1 mm dichlorodiphenyltrichloroethane (DTT), 10% glycerol and 1% triton X-100] for 1 h at 4°C. Beads were then washed five times with the same buffer, re-suspended in sodium dodecyl sulfate – polyacrylamide gel electrophoresis (SDS–PAGE) sample buffer, incubated for 10 min at 65°C, pelleted for 30 s at 10 000 g and the supernatants were subjected to SDS–PAGE. Gels were then either stained with Coomassie blue or dried and exposed for autoradiography. The percentage of [35S]GASP-1 retained by DOR and β1-AR point mutants compared with wild-type receptor tails was calculated by scintillation counting of an aliquot of each supernatant.
A polyclonal rabbit antibody to GASP-1 was raised against glutathione S-transferase fusion protein containing residues 924–1395 of GASP-1. We have verified that, following immunoblot analysis (see below), we did not observe any immunoreactivity in COS-1 cells transfected with pcDNA3 vector (see Fig. 1b, lower panel) while four major immunoreactive bands were detected in cells transfected with the cDNAs encoding GASP-1, thus demonstrating the specificity of this antibody.
The Flag epitope-tagged DOR was generated as previously described (Cvejic et al. 1996). The cDNAs encoding Flag-DOR and GASP-1 were co-transfected in COS-1 cells by electroporation as described (Befort et al. 1996). After 48 h, cells were treated with 10 μm Deltorphin II (Neosystem, Strasbourg, France) for 24 h or 30 min. Cells were then washed two times and scraped off the plates with phosphate-buffered saline, pelleted by centrifugation at 400 g for 10 min at 4°C and frozen at −80°C. Pellets were thawed in lysis buffer containing 50 mm Tris/HCl pH 7.4, 300 mm NaCl, 1.5 mm MgCl2, 1 mm CaCl2, 1% triton X-100, 10% glycerol and a cocktail of protease inhibitors (Complete, Roche Molecular Biochemicals, Indianapolis, IN, USA) and incubated for 30 min at 4°C. The cell lysates were then cleared by centrifugation (10 000 g for 10 min at 4°C) and the protein concentration was measured using Bradford reagent (Bio-Rad Laboratories, Hercules, CA, USA). For immunoprecipitations, 500 μg of proteins were incubated with 2 μg of M1 anti-Flag antibody (Sigma, St Louis, MO, USA) for 2 h at 4°C, followed by incubation with protein A-Sepharose (Sigma) for an additional 2 h. The immunoprecipitation complexes were isolated by centrifugation at 2000 g for 1 min and washed three times with lysis buffer. The pellets were re-suspended in SDS–PAGE loading buffer, incubated 10 min at 65°C and centrifuged at 10 000 g for 2 min. Proteins from the supernatant were subjected to SDS–PAGE, transferred to immobilon-P membranes (Millipore Corporation, Bedford, MA, USA) in 50 mm Tris-boric acid and blocked with 5% non-fat powder milk in TBS-tween/CaCl2 (50 mm Tris/HCl pH 8, 150 mm NaCl, 0.4% tween 20, 1 mm CaCl2). Membranes were then incubated with M1 anti-Flag monoclonal antibody in the same buffer for 1 h and washed with TBS-tween/CaCl2. After incubation with horseradish peroxidase-conjugated anti-mouse IgG (Vector Laboratories, Burlingame, CA, USA) in 5% milk in TBS-tween/CaCl2 for 1 h, membranes were washed in TBS-tween and subjected to chemiluminescence detection method (ECL plus, Amersham-Biosciences). For the detection of GASP-1, anti-GASP-1 polyclonal serum and horseradish peroxidase-conjugated anti-rabbit IgG (Vector Laboratories) were used and CaCl2 was omitted from buffers. For immunodetection of GASP-1 directly from cell lysates, 50 μg of proteins were loaded on SDS gels and analyzed following the same procedure.
Northern blot analysis
Multiple tissue blots of poly(A)+RNA from human tissues (Clontech) were hybridized to probes labelled by random priming (Roche). A human GASP-1 cDNA probe (320 bp encoding residues 924–1031) and a human β-actin cDNA control probe (Clontech) were 32P-labelled to a specific activity of approximately 109 cpm/μg. Hybridization was performed at 68°C using ExpressHyb hybridization solution (Clontech) and washes were carried out in 0.1 × saline sodium citrate buffer (SSC)/0.1% SDS at 65°C. Hybridization signals were visualized and analyzed on a phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA).
In situ hybridization
Cryostat sections (10 μm) were prepared from frozen brains and hybridized with [35S]-labelled antisense riboprobes as described (Décimo et al. 1995). The GASP-1 probe (362 bp encoding residues 745–871 of mouse GASP-1) was amplified by PCR from mouse genomic DNA and subcloned into pBluescript for riboprobe synthesis. Brain sections were treated and hybridized in the same experimental series and exposed under Kodak NTB-2 emulsion for 2 weeks.
In order to identify opioid receptor interacting proteins, we constructed a two-hybrid cDNA library from SHSY-5Y cells, which endogenously express mu and delta-opioid receptors (Kazmi and Mishra 1987). We then performed two-hybrid screens using as baits the COOH-terminal tail of human delta and mu receptors starting at and including the NPXXY conserved motif. Twenty clones were isolated from the screening with the delta receptor carboxyl-terminal tail and confirmed as positive after yeast retransformation control experiments. Subsequent sequencing of these clones revealed that we isolated five times the same clone encoding the carboxyl-terminal part (residues 924–1395), of a large predicted protein of 1395 amino acids (KIAA0443, gene bank accession number: AB007903) that was identified very recently by Whistler and colleagues (Whistler et al. 2002) as a key player in the lysosomal sorting of DOR.
To further confirm that GASP-1 directly interacts with the carboxyl-terminal tail of delta opioid receptor, we produced the [35S]methionine labelled COOH-terminal portion of GASP-1 (residues 924–1395) by in vitro translation and expressed the COOH-tail of DOR as GST-fusion protein in Escherichia coli. When incubated with DOR/GST-fusion immobilized on glutathione-sepharose beads, the carboxyl-terminal region of GASP-1 strongly bound to DOR and not to GST alone (Fig. 1a, middle panel, lanes 1 and 2) confirming the result obtained in two-hybrid yeast experiments. Noticeably, analysis of in vitro translation products revealed the presence of labelled polypeptides of lower molecular weight that also interact with DOR GST-fusion (not shown). These lower molecular weight species most likely arose from alternate start codon usage as protease inhibitors did not affect their presence. This observation suggests that a limited portion of the GASP-1 COOH-terminus is critical for the interaction.
We then cloned the full-length cDNA encoding GASP-1 and observed the same capacity of in vitro translated GASP-1 to bind to purified DOR GST-fusions and not to GST alone (Fig. 1a, upper panel, lane 1 and 2). Finally, we examined the ability of GASP-1 to bind full-length DOR in vivo by co-immunoprecipitation from transfected COS cells (Fig. 1b). To facilitate specific immunoprecipitation and detection of receptors, we transfected a cDNA encoding a Flag-tagged version of DOR. To detect GASP-1 protein, we raised a rabbit polyclonal antibody against the COOH-terminus (residues 924–1395, see Experimental procedures). This antibody specifically recognized four major immunoreactive bands, on immunoblot of COS cells lysates transfected with the cDNA encoding full-length GASP-1. The upper band displayed an apparent molecular weight of 190 kDa (see Fig. 1b, lower panel), which is slightly higher than expected (157 kDa). Flag-DOR was recognized as a 70-kDa fuzzy immunoreactive band following immunoprecipitation and immunodetection with anti-FLAG antibody (Fig. 1b, middle panel). GASP-1 was co-immunoprecipitated specifically from cell lysates in association with Flag-DOR (Fig. 1b, upper panel). GASP-1 co-immunoprecipitation was not significantly affected by short-term (30-min) or long-term (24-h) treatments with a DOR selective agonist (Fig. 1b, upper panel, lanes 3 and 4). These data further demonstrate that GASP-1 displays binding activity for full-length DOR in a cellular context, as it does for purified DOR/GST carboxyl-terminal tail in vitro.
GASP-1 interacts with COOH-termini of opioid receptors
To assess selectivity of the interaction between GASP-1 and DOR, we performed GST pull-down experiments using the COOH tail of the other opioid receptors (MOR and KOR) as well as the ORL-1 receptor, a highly structurally related protein. Each carboxyl-terminal tail, starting from the NPXXY motif, was expressed as a GST-fusion protein in E. coli and tested with in vitro translated GASP-1 COOH-terminal portion or full-length protein (see Fig. 1a, middle and upper panels, respectively). We also observed a strong interaction of both labelled proteins with KOR, ORL-1 and, to a lesser extent, MOR carboxyl-terminal tails. This result suggested that these receptors might interact with GASP-1 via a conserved motif.
As shown in Fig. 1(c) (lower panel), opioid receptor COOH-terminal sequences share a conserved motif corresponding to a short stretch of 17 amino-acids spanning from the carboxyl-terminal end of the seventh transmembrane domain to an Arg residue. We therefore tested the GASP-1 binding activity of this conserved region, as well as that of non-conserved portion of DOR and KOR carboxyl-termini (Fig. 1c upper and middle panels). As expected, the non-conserved part of DOR and KOR COOH-termini (lanes 3 and 5, respectively) did not interact with GASP-1, while the conserved COOH-terminal region of DOR (lane 2) still retained substantial binding activity as compared with DOR and KOR full-length COOH-terminal tails (lanes 1 and 4, respectively). This result points to this sequence as a critical determinant for the binding of GASP-1 to opioid and ORL1 receptors.
GASP-1 interacts with the cytoplasmic tail of other GPCRs
The results obtained with opioid and ORL1 receptors prompted us to investigate GASP-1 binding to other GPCR carboxyl-terminal tails from the same subfamily A. Within this subfamily, we selected GPCRs with the most divergent COOH-terminal sequences and with different coupling to G-proteins. We also tested a GPCR from subfamily B, calcitonin receptor (CALCR), showing no sequence homology with GPCRs from subfamily A in its carboxyl-terminal region. Results of GST pull-down experiments (Fig. 2a) showed that, with the exception of histamine H2 (H2) COOH-terminus, for which the interaction was weaker, a substantial association with GASP-1 was observed for all of the COOH-terminal domains, including the short tails of m2 muscarinic (M2) and thromboxane A2α (TXA2α) receptors composed of 31 and 40 amino acids, respectively. Again, this result indicated a critical role for the region proximal to the seventh transmembrane domain in promoting the interaction with GASP-1. We did not observe any correlation between the strength of the interaction and the mode of G-protein coupling for the tested receptors. Indeed, the highest levels of binding were observed with DOR and β1-adrenergic (β1-AR) cytoplasmic tails, which are coupled to Gi and Gs, respectively, while low levels of binding were observed with MOR (Gi; see Fig. 2b) and H2 (Gs) carboxyl-termini. M1 (Gq), M2 (Gi), TXA2α (Gq) and 5-HT7 (Gs) cytoplasmic tails displayed intermediate levels of binding. Surprisingly, CALCR displayed a strong binding activity toward GASP-1 suggesting that, despite the absence of sequence homology, conserved structural determinants could exist in the proximal carboxyl-terminal sequence of GPCRs from families A and B. Alignment of the sequences of GPCR COOH-termini tested in GST pull-down experiments with the rhodopsin sequence (Fig. 2b) revealed the presence of two conserved residues in all of the sequences, including CALCR, suggesting that these amino acids could play a critical role in the interaction with GASP-1. In rhodopsin, these residues (F313 and R314) are located in a region shown to form a cytoplasmic helix (named helix 8) perpendicular to the membrane (see Fig. 2b and Palczewski et al. 2000). Mutation of these two residues in alanine in DOR and β1-AR COOH-termini resulted in a decrease of 77 ± 3 and 90 ± 3% (respectively) of GASP-1 binding compared with wild-type receptor tails (Fig. 2c). Altogether, these results point to helix 8 as a critical structural determinant of GPCR carboxyl-terminal tail for the interaction with GASP-1.
GASP-1 mRNA is predominantly and broadly expressed in the central nervous system
To further examine the relevance of GASP-1 interaction with a large number of GPCRs, we investigated the distribution of GASP-1 mRNA in human tissues as well as in mouse brain (Fig. 3). Northern blot from human tissues hybridized with a human GASP-1 probe revealed a single transcript of approximately 6 kb expressed predominantly in central nervous system, while other tissues expressed lower or undetectable levels of this transcript (Fig. 3a). Following normalization with actin, expression levels of GASP-1 transcript were found to be homogeneous within the different regions of the central nervous system, except for medulla, spinal cord, corpus callosum and substancia nigra, where lower signals were observed. In situ hybridization with a mouse GASP-1 probe on mouse brain sections confirmed that GASP-1 transcript is broadly expressed in brain with some regions showing higher expression levels like olfactory bulb, hippocampus, hypothalamus or some thalamic nuclei (Fig. 3b). Finally, Blast searches conducted with the GenBank revealed that the large majority of ESTs encoding GASP-1 originate from nervous tissue, thus confirming our previous results. Collectively, these results are consistent with an interaction of GASP-1 with a large number of GPCRs, the higher levels of transcript being observed in the central nervous system where a large range of GPCRs is expressed.
GASP-1 is a member of a novel gene family
Sequence homology searches performed with the SPTrEMBL database indicated that GASP-1 contains a carboxyl-terminal 250 amino acids region with similarities with several mammalian proteins that we named GASP-2 to GASP-10 (see Fig. 4). This region displays 20 to 77% identity in pair-wise comparisons between GASP-1 and these other proteins and extends across the entire 250 amino acid domain, except for GASP-10. In addition, outside of this domain, a 15-amino acid motif is repeated 22 times in GASP-1, and is also present at least two times in four over the nine GASP-1 homologous sequences (see Fig. 4), thus defining a subfamily of five members. An alignment of the conserved 250 amino acids carboxyl-terminal domains showing that conserved positions are present throughout this core domain is shown in Fig. 5. Additional homology searches with the whole human and mouse genomes revealed that these proteins, except GASP-8, are encoded by a single exon and that their genes are grouped within two 200-kb clusters on two adjacent contigs in chromosome X, indicating that they derived from the same ancestral gene. In human, the gene encoding GASP-8 is located on chromosome seven, contains seven coding exons and gives rise to two splice isoforms. An intronless identical coding sequence, except for two missing amino acid (aa) stretches encoded by exon 2 (34 aa) and exon 6 (25 aa), is also present on chromosome three, suggesting that this gene also has evolved from the same common ancestor. We observed a similar organization for GASP-8 in mouse genome. Altogether, results from our sequence analysis therefore suggest that these proteins constitute a novel family.
Tissue distribution of ESTs from the GenBank for GASP-2, -4, -5, -6, -7, -8 and -10, together with RT-PCR distribution of GASP-3 and 9 mRNAs (as found in the HUGE database, http://www.kazusa.orjp/huge/), revealed that, except GASP-4 that seems homogeneously distributed and GASP-9 that is strongly expressed in ovary and kidney, the other GASP transcripts are predominantly expressed in the central nervous system (see Table 1). In addition, high levels of GASP transcripts are also found in immune cells, lung and liver for GASP-6, immune cells, pancreas and skin for GASP-8, heart, brain, placenta, lung and prostate for GASP-9 and placenta and skin for GASP-10. Finally, low expression levels for all GASP transcripts were detected in numerous organs. As for GASP-1, these results are consistent with an interaction of GASP proteins with a large number of GPCRs.
Table 1. Tissue distribution of GASP-1–10 transcripts
Predominantly expressed in CNS
Northern blot/in situ hybridization
Predominantly expressed in CNS
ESTs (18 CNS/21 other)
Predominantly expressed in CNS
ESTs (22 from various tissues)
Enriched in CNS
ESTs (12 CNS/24 other)
Enriched in CNS and highly expressed in immune cells, lung and liver
ESTs (24 CNS/64 other)
Predominantly expressed in CNS
ESTs (27 CNS/32 other)
Enriched in CNS, immune cells pancreas and skin
ESTs (20 CNS/55 other)
Enriched in kidney and ovary and highly expressed in heart, brain, placenta, lung and prostate
Enriched in CNS, placenta and skin
ESTs (13 CNS/30 other)
The closest GASP-1 paralogue also interacts with COOH-terminus of various GPCRs
Within this family, several members share sequence similarities outside of the carboxyl-terminal 250-amino acid domain. In particular, GASP-2 displays sequence homology with GASP-1 throughout its entire sequence with the NH2-terminal region (aa 1–367) showing 50% identity with the NH2-terminal part of GASP-1, and the COOH-terminal region (aa 368–838) being 68% identical to the COOH-terminal sequence of GASP-1. We therefore investigated whether GASP-2 also binds in vitro to the carboxyl-terminal tail of GPCRs previously tested with GASP-1 (see above). As shown in Fig. 6, in vitro translated GASP-2 strongly bound to GST-fusions from M1 and CALCR and not to GST alone. Lower but significant levels of binding were observed for GST-fusions from 5-HT7, β1-AR, M2 and H2, while no binding was observed for opioid and ORL1 receptors. These results indicate that at least another member of this protein family interacts with the carboxyl-terminal region of various GPCRs with a different selectivity profile from GASP-1 protein.
Using the two-hybrid system, we have cloned a cDNA encoding a protein of 1395 residues that interacts with GPCRs. We found that GASP-1 is a member of an as yet undescribed family of proteins containing a conserved 250-residue COOH-terminal domain. Consistent with the notion of a family, genes encoding these proteins display the same organization and are all located, but one, in a very small portion of chromosome X, both in human and mouse genomes. Interestingly, in human, this region is associated with five different mental retardation diseases including Allan–Hernson and Partington syndromes. We did not find any protein showing significant sequence homology with GASP-1 within the complete genome of eukaryotes other than mammals, suggesting that this protein family appeared very recently during evolution.
Our experiments indicate that the core domain of GASP-1 is involved in the binding to the cytoplasmic tail of GPCRs from subfamilies A and B, and it is therefore likely that the conserved domain of the other members of the family also binds to GPCRs. In line with this, we have shown that the closest paralogue of GASP-1 also binds to GPCR carboxyl-terminal tails, although with a different selectivity profile. Our data therefore suggest that the core domain of this protein family represents a novel protein–protein interaction domain. We could not find any homology with other domains previously described to be involved in protein–protein interactions. We propose that each member of this family could associate with a different subset of GPCRs to fulfil a specific function that remains to be elucidated. In this respect, it will be interesting to thoroughly examine the binding capacity of each of these proteins to each subfamily of GPCRs.
On the receptor side, we have shown that a small conserved region within the COOH-terminal domain of opioid receptors is critical for binding to GASP-1. This region contains the putative helix 8 described in rhodopsin structure and there is evidence suggesting the presence of an α-helix in the corresponding region of other GPCRs (Wakamatsu et al. 1992; Jung et al. 1996) from subfamily A. This helix could therefore represent a critical structural determinant for the interaction with GASP-1. Consistently, GASP-1 interacts with the carboxyl-terminal tail of all GPCRs from subfamily A that we tested. In addition, sequence comparisons revealed the presence of two conserved residues within helix 8 of these receptors and their mutation in DOR and β1-AR strongly reduced their interaction with GASP-1. Although one cannot exclude that additional motifs could account for the differences in GASP-1 binding we observed between the different GPCR carboxyl-terminal tails, our results suggest a critical role of helix 8 in the interaction with GASP-1. We also observed a strong interaction of GASP-1 with one GPCR from subfamily B. Although there is no direct experimental evidence for the presence of an alpha helix in the proximal part of COOH-terminal tail of GPCRs from subfamily B, it is of note that this region displays a high degree of similarity among the different receptors within this subfamily, indicating that it could play an important role. The fact that the two conserved residues between GPCRs from subfamilies A and B are located within the sequence that forms helix 8 in rhodopsin further suggests that in subfamily B the proximal COOH-terminal region could also adopt a helical structure critical for the interaction with GASP-1. Helix 8 has previously been implicated in the interaction with G-proteins (Ernst et al. 2000; Marin et al. 2000) suggesting that GASP-1 could compete with G-proteins binding at this site.
Recently, Whistler et al. (Whistler et al. 2002) have also identified GASP-1 as an interacting partner of DOR from a two-hybrid screen. These authors isolated a shorter cDNA encoding the last 273 carboxyl-terminal residues of this protein, which confirms that the core carboxyl-terminal domain of GASP-1 is critical for its association with GPCRs. In their report, GASP-1 was proposed to participate to post-endocytic sorting of DOR, and was named GASP for candidate G protein-coupled receptor associated sorting protein. This proposal was based on several lines of evidence, including the fact that GASP-1 strongly interacts with DOR, which is preferentially targeted to lysosomes upon sustained stimulation, but poorly interacts with MOR that rather undergoes recycling. Although we confirm that MOR weakly interacts with GASP-1 in vitro, we also found that this protein interacts strongly with a broad range of GPCRs, including fast recycling receptors like M1 (Vogler et al. 1998). This suggests that, independently from endocytic sorting, GASP-1 protein might play another role in response to GPCR activation. In line with this, Matsuki and co-workers (Matsuki et al. 2001) recently identified GASP-1 as an interacting partner for the transcription factor Per1 and named it PIPS for Per1 interacting protein of the suprachiasmatic nucleus. These authors showed that co-expression of both proteins in COS cells affects their distribution between cytoplasmic and nuclear compartments. Per1 is one of the essential elements involved in the transcription/translation-based autoregulatory loop of the endogenous master clock (see Cermakian and Sassone-Corsi 2000). In mammals, Per1 is strongly expressed in the master-clock centre, but is also expressed in many other brain regions, suggesting additional roles for this protein (Yamamoto et al. 2001). In addition, the activation of several GPCRs, including dopamine, serotonin and neuropeptide Y receptors, has been shown to modulate the response to light (see Cermakian and Sassone-Corsi 2000). From these observations, it is conceivable that upon GPCRs activation GASP-1/PIPS could recruit Per1 or related transcription factors, thus providing a direct coupling between GPCRs activation and regulation of transcription. Such direct signalling pathways have recently been observed for the transcription factor tubby, which is released from its association with plasma membrane phosphatidylinositol upon activation of phospholipase C-β by GPCRs (Santagata et al. 2001), and for CREB2 which dissociates from the COOH-terminal tail of GABAB receptor upon activation by an agonist (White et al. 2000). In the last few years, a growing body of evidence has emerged suggesting that endocytic transport is important in mediating the formation of specialized signalling complexes (see Sorkin and Von Zastrow 2002). It is therefore reasonable to consider that GASP-1 could be involved both in sorting of GPCRs and regulation of transcription following activation of these receptors. This dual function is reminiscent of other interacting partners of GPCRs, β-arrestins, which are responsible for GPCR desensitization and function as adaptor proteins that facilitate the activation and subcellular localization of signalling cascades, particularly mitogen-activated protein kinase cascades (see Pierce and Lefkowitz 2001).
Further studies are now required to evaluate the functional consequences of this interaction.
We thank Pascal Dollé and Valérie Fraulob for in situ hybridization; Olivier Poch and Jean-Cristophe Amé for their help in sequence comparisons; IGBMC core facilities and particularly Gilles Duval for antibody generation. We are grateful to Renaud Wagner for the cDNAs encoding β1-AR, CALCR, H2, 5-HT7, TXA2α; Brigitte Ilien for M1 and M2; Lakshmi Devi for FlagDOR. We thank Claire Gavériaux-Ruff for critical reading of the manuscript. This work was supported by le Centre National de la Recherche Scientifique, l'Hôpital Universitaire de Strasbourg, l'Institut Nationale de la Santé et de la Recherche Médicale, L'Université Louis Pasteur and by a grant from l'Association pour la Recherche sur le Cancer No.5770.