Identification of a Novel Protein-Protein Interaction Motif Mediating Interaction of GPCR-Associated Sorting Proteins with G Protein-Coupled Receptors

GPCR desensitization and down-regulation are considered key molecular events underlying the development of tolerance in vivo. Among the many regulatory proteins that are involved in these complex processes, GASP-1 have been shown to participate to the sorting of several receptors toward the degradation pathway. This protein belongs to the recently identified GPCR-associated sorting proteins (GASPs) family that comprises ten members for which structural and functional details are poorly documented. We present here a detailed structure–function relationship analysis of the molecular interaction between GASPs and a panel of GPCRs. In a first step, GST-pull down experiments revealed that all the tested GASPs display significant interactions with a wide range of GPCRs. Importantly, the different GASP members exhibiting the strongest interaction properties were also characterized by the presence of a small, highly conserved and repeated “GASP motif” of 15 amino acids. We further showed using GST-pull down, surface plasmon resonance and co-immunoprecipitation experiments that the central domain of GASP-1, which contains 22 GASP motifs, is essential for the interaction with GPCRs. We then used site directed mutagenesis and competition experiments with synthetic peptides to demonstrate that the GASP motif, and particularly its highly conserved core sequence SWFW, is critically involved in the interaction with GPCRs. Overall, our data show that several members of the GASP family interact with GPCRs and highlight the presence within GASPs of a novel protein-protein interaction motif that might represent a new target to investigate the involvement of GASPs in the modulation of the activity of GPCRs.


Introduction
G protein-coupled receptors (GPCRs) represent one of the most diversified protein families in humans. They modulate a large panel of physiological processes making them unrivalled targets for development of new therapeutic agents. GPCRs translate extracellular stimuli into intracellular signals, and the intensity and duration of these are determined by complex regulation mechanisms. Internalization, whereby agonist-activated GPCRs are rapidly silenced by translocation from the cell surface to endocytic compartments, represents a central event for the modulation of receptor activity [1]. Upon internalization, GPCRs can be recycled back to the membrane or degraded. Although mechanisms that govern the postendocytic fate of GPCRs are not fully understood, several proteins have been shown to modulate this phenomenon via a direct interaction with their carboxyl-terminal intracellular tail (C-tail) [2][3][4][5]. One of these regulatory proteins is GPCR-associated sorting protein 1 (GASP-1), which was identified in a yeast two-hybrid screen using the d-opioid receptor (DOR) cytoplasmic C-tail as bait [6,7] and has been shown to be involved in the sorting of receptors that are quickly degraded following agonist-promoted internalization [7]. This phenomenon has been proposed to form the molecular basis for analgesic tolerance to cannabinoids [8,9]. GASP-1 was also found to interact with numerous other receptors both in vitro [6,7,10] and in vivo [11,12]. Although, its has been proposed that the binding epitope for GASP-1 to these receptors is large an covers major parts of the cytoplasmic C-tail of receptors (Heydorn et al., 2004), we have shown that helix 8, located near the seventh transmembrane helix, is critically involved in this interaction [6]. Concerning GASP-1, little is known about which region within this protein is required for its interaction with GPCRs.
GASP-1 is part of a novel protein family of ten members that were identified by sequence homology searches [6,13]. The carboxyl-terminal 250 amino acids (AA) region of GASP-1 displays remarkable sequence identity with the nine other members of the family. Furthermore, GASP genes, except GASP-8, contain a single coding-exon and are located within two 200kilobase clusters on two adjacent contigs on chromosome X, suggesting that they arose from a common ancestral gene [13,14]. Altogether, these data indicate that GASPs do indeed form a protein family and might display similar functions. However, except GASP-1 and GASP-2, interaction of the other members of the family with GPCRs has not been investigated so far.
In this study, we present a detailed structure-function relationship analysis of the molecular interaction between GASPs and GPCRs. GST-pull down experiments revealed that, besides GASP-1 and GASP-2, different members of the GASP family can interact with a wide range of GPCRs. Using surface plasmon resonance (SPR) analysis with two full-length GPCRs in solution and co-immunoprecipitation of GPCRs expressed in HEK293 cells, we further showed that the central domain of GASP-1 is critical for its interaction with GPCRs. Finally, we identified within this domain and in several other GASPs a conserved and repeated sequence of 15 amino acids that we called ''GASP motif'' and demonstrated that this motif plays a critical role in the interaction with GPCRs, both by site directed mutagenesis and competition experiments with synthetic peptides. Overall, our results demonstrate that GASPs indeed represent a novel family of GPCR-interacting proteins that can be divided into two subfamilies depending on the presence of the GASP motif. Our data clearly show that this sequence represents a novel protein-protein interaction motif that is critical for the interaction between GASPs and GPCRs.

[ 35 S]-labelled GASPs in vitro Production
cDNAs encoding GASP-1 and GASP-2 were cloned into the pcDNA3 eukaryotic expression vectors [6]. cDNAs encoding GASP-3, -6, -7, and -9 were subcloned into pcDNA3.1 vectors. All the clones were controlled by DNA sequencing. The in vitro production was performed by using the TNT Quick-coupled Transcription/Translation T7 kit (Promega) in presence of [ 35 S]-Methionine according to the instructions from the manufacturer.

GST-pull Down Assay
GST-fusion proteins were immobilized on glutathione-coupled Sepharose beads (GE Healthcare) for 1 h at 4uC and incubated with [ 35 S]-labelled in vitro translated GASP proteins or their truncated or mutated forms in ice-cold binding buffer containing: 20 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM DTT, 10% glycerol and 1% Triton X-100. The mixture was incubated for 1 h at 4uC with gentle rocking. The beads were washed five times with the same ice-cold buffer, resuspended in 30 ml SDS-PAGE sample buffer, incubated 10 min at 65uC, and pelleted for 60 s at 3000 g. The supernatants were then analyzed by SDS-PAGE, the gels were stained with Coomassie blue to visualize GST-fusion proteins, dried and analyzed using a Phosphor-imager (Personal Molecular Imager FX, Biorad) to visualize [ 35 S]-labelled GASPs. Quantification was performed with the Quantity One software (Biorad). GST-pull down quantification data were analyzed using GraphPad Prism 4.03 (GraphPad Software). Quantifications presented are means of at least three independent experiments.

GST-pull Down Competition Experiments
GASP peptide or its corresponding scrambled version was incubated with GST-fused proteins and [ 35 S]-labelled GASPs and the binding was analyzed as described for the GST-pull down assay. In a first step, dose-effects of GASP peptides were assessed in competition experiments between [ 35 S]-labelled GASP-2 and GASP peptides (concentrations ranging from 1 to 250 mM) for the interaction with GST-ADRB1. Subsequent competition experiments with other GASPs and GPCR C-tails were performed by using a single GASP peptide concentration of 150 mM.

Production and Purification of Full-length GPCRs
The human ADRB2 and CNR2 receptors were produced using the P. pastoris expression system as previously described [15,16].
After methanol-induced receptor expression, cells were washed with PBS pH 7.0 and resuspended in ice-cold buffer A (50 mM Tris-HCl pH 7.4, 500 mM NaCl, 10% glycerol, 5 mM EDTA, 1 mM PMSF, 2 mM DDT). Cells were then lysed with two cycles of 20 s shaking and 20 s cooling on ice using 0.5 mm glass beads in a FastPrep 24 device. Unbroken cells and cell debris were removed by centrifugation (3000 g, 5 min, 4uC) and the membrane fraction from the supernatant was pelleted by ultracentrifugation (100,000 g, 45 min, 4uC). Membranes were resuspended in buffer B (50 mM Tris-HCl pH 7.4, 500 mM NaCl, 10% glycerol, 1 mM PMSF, 2 mM DTT) using a Dounce homogenizer, and then successively washed with urea (buffer B with 4 M urea) and NaOH (buffer B with 10 mM NaOH), and ultracentrifuged (100,000 g, 45 min, 4uC). Finally membranes were resuspended in buffer B and quantified with the bicinchoninic acid (BCA) method (Pierce).

Purification of Free GST and GST-fused Central Part of GASP-1
Free GST or GST-fused central part of GASP-1 (AA 380-1073) were purified on an Ä KTApurifier (GE Healthcare) using a GSTrap 4B 1 ml column (GE Healthcare) according to the instructions from the manufacturer. Proteins were eluted with a buffer containing 50 mM Tris-HCl pH 8.0 and 10 mM reduced glutathione. The buffer was finally exchanged for 50 mM Tris-HCl pH 8.0 and glycerol 10% using a HiTrap 5 ml desalting column (GE Healthcare). Purified proteins were quantified using the BCA assay (Pierce), analyzed by SDS-PAGE and stored at 280uC.

SPR Measurements
A Biacore X100 SPR instrument (GE Healthcare) equilibrated to 25uC and equipped with a Sensor Chip CM5 (GE Healthcare) was used for all SPR measurements.
Affinity purified polyclonal anti-GST antibody (GE Healthcare) was covalently immobilized in two flow cells using Amine Coupling Kit (GE Healthcare) with a running buffer containing 10 mM HEPES pH 7.4, 150 mM NaCl, 0.05% Surfactant P-20 (GE Healthcare) and 3 mM EDTA and a flow rate of 5 ml/min. 50 mg/ml antibody in 10 mM sodium acetate pH 5.0 was injected for 5 min to a surface activated by a 7 min injection of a 1:1 mixture of 0.4 M 1-ethyl-3-(3-dimetrylaminopropyl) carbodiimide HCl and 0.1 M N-hydroxysuccinimide. The remaining active groups were deactivated by a 7 min injection of 1 M ethanolamine HCl pH 8.5. This procedure resulted in immobilization of more than 10000 resonance units (RU < pg/mm 2 [17]) of anti-GST antibody.
Capture and binding experiments were performed with a running buffer containing 50 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.1% DDM, and 0.01% CHS. Each binding cycle started with capture of GST in the first flow cell by a 0.5-1 min injection of 12.5 nM GST, followed by a 0.5-1 min injection of 94 nM of GST-tagged central domain of GASP-1 in the second flow cell and ended with regeneration by a 2 min injection of 10 mM glycine-HCl pH 2, all at a flow rate of 10 ml/min. Typical capture densities were 50-150 Resonance Units (RU) for GST and 100-200 RU for the central domain of GASP-1. No significant dissociation of the captured proteins was observed in the experimental time frame.
For saturation binding, a range of ADRB2 or CNR2 concentrations were injected after GST and GASP capture. For peptide competition, a fixed concentration of ADRB2 and CNR2 was mixed with a range of GASP peptide or control peptide concentrations and incubated for at least 1 h before injection. Several blank cycles were included for all samples. For the competition curves, an additional blank injection with the appropriate concentration of peptide was injected just before the injection of the peptide and receptor mixture. For all sample injections, the contact time was 3 min, the dissociation phase was 3 min, and the flow rate was 30 ml/min.
The binding curves were double referenced by (i) subtraction of the signal from the GST captured flow cell from the signal of the flow cell with captured central domain of GASP-1 and (ii) subtraction of the appropriate blank injection from the receptor injections with or without peptide [18]. The curves were adjusted for the slightly decreasing GASP capture density during a binding series (20-40 RU from the first to the last GASP capture) by normalization to the response from the first GASP capture. The high peptide concentrations resulted in large spikes at the beginning and end of each injection; these points were clearly artifacts and they were removed for presentation purposes. Endpoint responses were read 20 s after injection end.
The double-referenced dose-response curves were fitted with Biacore X100 Evaluation Software 2.0.1 (GE Healthcare). The dissociation phases were first fitted separately to a one-to-one binding model to obtain the dissociation rate constants. The association and dissociation phases were then fitted simultaneously to a model with two parallel independent one-to-one reactions (heterogeneous ligand), with the dissociation rate constants fixed to the single value found by fitting of the dissociation phases separately. The mass transport contribution was negligible in both data sets. The two association rate constants obtained from these fits were used to estimate the affinity ranges. All curves were fitted by global analysis [18].

Co-immunoprecipitation Experiments
HEK293 cells stably expressing N-terminal GFP-tagged ADRB1, ADRB2, CALCR and M 1 or MyrPalm-mYFP were transfected with pcDNA3.1 containing the sequence for a central domain of GASP-1 (AA 380 to 1073) using JetPEI reagent (polyplus transfection) according to the instructions from the manufacturer. After 48 h of expression, cells were washed twice with PBS and lysed in a buffer containing 10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 25 mM KCl, 0.3% Triton X-100 and cOmplete protease inhibitors (Roche) for 1 h at 4uC under agitation. 500 mg of cleared lysates were incubated with 1 mg of a mouse monoclonal anti-GFP antibody (Invitrogen) overnight at 4uC, followed by an incubation with 40 ml of protein A-Sepharose beads for an additional 2 h at 4uC. Beads were then washed five times in lysis buffer and precipitates were resolved on an 8% gel by SDS-PAGE and electrotransferred to immobilon-P membranes (Millipore) in 50 mM Tris-boric acid for 1 h at 300 mA. Membranes were blocked with 5% non-fat powder milk in TBS-Tween (50 mM Tris-HCl pH 8, 150 mM NaCl, 0.4% Tween-20) for 1 h at room temperature. Detection of GFP-tagged receptors was performed by a 2 h incubation of the blots with goat anti-GFP coupled to horseradish peroxidase-conjugated antibody (Abcam; 1:10,000). For GASP detection, blots were incubated for 1 h with anti-GASP polyclonal serum from rabbit (1:2,500) and for 1 h with HRP-conjugated anti-rabbit antibody (New England Biolabs; 1:5,000). For immunodetection of GASP or GFP-tagged receptors directly from cell lysates, 10 mg of proteins were loaded on SDS gels and analyzed following the procedure described above. Finally, membranes were washed in TBS-Tween and detected by chemiluminescence (ECL prime, GE Healthcare).

GASPs form a Family of G Protein-coupled Receptor Interacting Proteins
In our previous study we have shown that GASP-1 and GASP-2 interact with the carboxyl-terminal domain of several GPCRs. Based on sequence homology searches, we identified eight additional members of the GASP family ( [6,13]; Figure 1). All members display sequence similarities in their carboxyl-terminal domain (last 250 amino acids) and the first five members contain a repeated motif of 15 amino acids outside this domain, that we named the GASP motif ( Figure 1). Moreover, a crosswise comparison of the conserved carboxyl-terminal domain of the different members of the GASP family revealed very high sequence similarities between GASP-1 and GASP-2 as well as high similarities between GASP-6, -7, -8 and -9 ( Figure S1). Although GASP-1 and GASP-2 have been shown to interact with GPCRs [6,7,10], there is no experimental evidence showing that the other members of the GASP family also interact with these receptors. In order to examine this possibility we performed GSTpull down experiments with radiolabelled GASP-1, -2, -3, -6, -7, and -9 and the cytoplasmic C-tail of twelve GPCRs fused to GST ( Figure S2), all comprising the helix 8 that was shown to be critically involved in the interaction with GASP-1. As controls, we used free GST and GST-fused C-tails of the type III transforming growth factor b receptor (TGF b ) and the insulin growth factor I receptor (IGF 1 ), which are non-GPCR receptors. Equivalent amounts of radiolabelled GASPs were incubated with saturating concentrations of GST-fused receptor C-tails immobilized on glutathione-sepharose beads. Radiolabelled proteins that were retained by receptor C-tails were then analyzed by SDS-PAGE and quantified by Phosphor-imaging. As shown in Figure 2, the two non-GPCR receptor C-tails (TGF b and IGF 1 ) did not interact with any GASPs (Figure 2A-B). Although sometimes weak, we observed significant interactions of the tested GASPs with several GPCR C-tails.
Altogether, these results show that in addition to GASP-1 and -2, other members of the GASP family can interact with a wide range of GPCR C-tails. As describe in Figure 1, members of the GASP subfamily 1 contain 22 (GASP-1) or 2 repeated (GASP-2 to -5) GASP motif. Combined with the fact that subfamily 1 displays a higher level of interaction with GPCR C-tails compared to subfamily 2, this observation prompted us to investigate the significance of the GASP motif in the interaction with GPCRs.

Two Regions of GASP-1 are Involved in the Interaction with the C-tail of GPCRs
In a previous study, we have shown that the carboxyl-terminal region of GASP-1, corresponding to AA 924 to 1395, displays a strong interaction with the DOR C-tail [6]. Within this region, a 250 AA carboxyl-terminal domain displays high sequence similarities with the other GASPs ( Figure 1). We therefore hypothesized that this conserved carboxyl-terminal domain could be critical for the interaction of GASPs with GPCRs. To test this hypothesis, we assessed the interaction of three GPCR C-tails, DOR, ADRB1 and M 1 , that display medium to strong interaction with GASPs (Figure 2A), with truncated mutants of GASP-1 in GST-pull down experiments.
In a first set of experiments, we tested three truncated mutants of GASP-1: mutant 380-1395 that lacks the N-terminal part, mutant 1025-1395 corresponding to the conserved C-terminal domain, and mutant 380-1073 corresponding to a central portion of GASP-1 that contains 19 GASP motif. As shown in Figure 3A, DOR, ADRB1 and M 1 C-tails interacted similarly with full-length GASP-1 and mutant 380-1395, indicating that the N-terminal In order to delineate more precisely which regions of GASP-1 are important for the interaction with DOR, we evaluated the interaction of the DOR C-tail with four additional truncated mutants of GASP-1 ( Figure 3B). The first two were derived from mutant 1025-1395 by extension with 50 or 100 AA at the N-terminus to include one or two GASP motifs (mutants A and B) and the other two were derived from mutant 924-1395, which contains 2 GASP motifs, by deletion of 45 or 95 AA from the Cterminus (mutants C and D). When incubated with DOR C-tail, mutants A and B displayed increasing interaction compared to the C-terminal domain of GASP-1, which were similar to those obtained with full-length GASP-1. Mutant C displayed strong interaction as well, while it was completely lost with mutant D. Altogether, these results indicate that the GASP motifs from the central domain are important to warrant full interaction with DOR and that the integrity of the GASP-1 carboxyl-terminal portion is required for this interaction. From these results we identified two types of interactions between GPCRs and GASP-1: some GPCRs interact exclusively with the central part of GASP-1 Figure 2. GST-pull down experiments with radiolabelled GASP-1, -2, -3, -6, -7 and -9 and GST-fused receptor C-tails. A, GASP-1, -2 and -3 showed medium to strong interactions with some GPCR C-tails but no interaction was detected with the two one-transmembrane receptor C-tails (TGF b and IGF 1 ). B, GASP-7 showed weak to medium interactions with some GPCR C-tails. GASP-6 and -9 showed very weak interactions with all tested receptors. No interaction was detected with TGF b and IGF C-tails. Data were quantified by Phosphor-imaging. Results are shown as percent of the [ 35 S]-GASPs input retained by the GST-fused receptor C-tails and correspond to the mean 6 S.E.M of three independent experiments. Lower panels correspond to representative gel images. 5HT 7 , 5-hydroxytryptamine 7 receptor; ADRB1, b 1 adrenergic receptor; CALCR, calcitonin receptor; DOR, d-opioid receptor; FZ 4 , frizzled 4 receptor; H 2 , histamine 2 receptor; IGF 1 , insulin growth factor I receptor; KOR, k-opioid receptor; M 1 , muscarinic M 1 acetylcholine receptor; M 2 , muscarinic M 2 acetylcholine receptor; MOR, m-opioid receptor; ORL 1 , opioid receptor-like 1; TXA 2 , a isoform of the thromboxane A 2 receptor; TGF b , type III transforming growth factor b receptor. doi:10.1371/journal.pone.0056336.g002 The GASP Motif Mediates GASP-GPCR Interaction PLOS ONE | www.plosone.org

Purified Full-length Receptors Bind dose-dependently to the Central Domain of GASP-1
In order to further characterize the importance of the central part of GASP-1 for the interaction with GPCRs, we evaluated the interactions between a central domain of GASP-1 (AA 380 to 1073), which contains 19 GASP motifs, and two purified fulllength GPCRs: the b 2 adrenergic receptor (ADRB2) and the cannabinoid receptor type 2 (CNR2). We first showed that the C-  tails of these two receptors interact with GASP-1 in GST Pulldown experiments and that the central domain of GASP-1 is mandatory and sufficient for this interaction ( Figure 4A), which is in agreement with previous observation for ADRB1 and M1 ( Figure 3A). Both ADRB2 and CNR2 full-length receptors were produced in P. pastoris and purified as previously described [15,16], while GSTtagged central domain of GASP-1 was purified from BL21 E. coli strain ( Figure S3). Biacore SPR was used to monitor the binding of GPCRs to GASP-1 in real-time and thus to determine the kinetics and the affinity of the GASP-1-GPCR interactions. The GSTtagged central domain of GASP-1 was first captured with anti-GST antibodies immobilized on a CM5 sensor chip. As a control, free GST was captured in another flow cell of the sensor chip. A range of concentrations of the purified GPCRs was then injected in both flow cells and the interaction was monitored as a function of time. All curves were double referenced by subtraction of the signal from the GST and from the buffer of receptor injections. Using this set-up we observed dose-dependent binding of ADRB2 and CNR2 to the central domain of GASP-1 ( Figure 4B-C). The kinetics of the two interactions were very similar: both interactions were very stable and had a slow association rate. For both sets of binding curves, we found complex kinetics in the association phase and a unimodal dissociation phase ( Figure S4). Analysis of the dissociation phases showed a dissociation rate constant of k d = 4.560.1610 24 s 21 for ADRB2 and k d = 6.560.1610 24 s 21 for CNR2. The complex association phases did not allow accurate determination of the association rate constants, but we estimated the dissociation constant K d to be in the ranges , 14-140 nM for ADRB2 and , 5-50 nM for CNR2.

The Central Domain of GASP-1 Interacts with GPCRs in Cells
To confirm the relevance of our previous observations we sought to examine the interaction between different GPCRs and the central domain of GASP-1 in a cellular context. To this purpose we performed co-immunoprecipitation experiments from HEK cells stably expressing GFP-tagged ADRB1, ADRB2, CALCR or M 1 receptor, transfected with the central domain of GASP-1 (AA 380-1073). As expected, the four GFP-tagged GPCRs where mainly localized at the plasma membrane while the central domain of GASP-1 displayed a cytoplasmic distribution that was similar to that observed for the full length GASP-1 [7]. As shown in Figure 5, the central domain of GASP-1 co-immunoprecipitated with the four different GPCRs but not with myristoylated-palmitoylated mYFP (MyrPalm-mYFP), which is targeted to the plasma membrane and particularly enriched in lipid rafts where GPCRs are also preferentially targeted [19]. These data demonstrated that the central part of GASP-1 displays binding activity for full length GPCRs in a cellular context, thus confirming the relevance of our in vitro measurements with carboxyl-terminal tails or purified receptors.

The GASP Motif is Critical for the Interaction of GASPs with GPCRs
Results obtained with GASP-1 truncated mutants, SPR and coimmunoprecipation experiments pointed to the GASP motifs from the central domain of GASP-1 as critical elements for interaction with GPCRs. This conserved motif is also present twice in the corresponding central parts of GASP-2 to GASP-5, thus suggesting that it could also play an important role in the interaction between GPCRs and these GASP subfamily members. Thus, we evaluated the interaction of truncated GASP-2 mutants with GST-fused ADRB1, M 1 , and CALCR C-tails, for which we had previously shown a strong interaction with full-length GASP-2 (see Figure 2). As shown in Figure 6A, mutant 377-838, resulting from deletion of the N-terminal part of GASP-2 (upstream from the first repeated motif), displayed interaction with all three receptor Ctails ranging from 30% to 60% compared to the full-length GASP-2. Conversely, mutant 470-838, corresponding to the C-terminal domain of GASP-2, displayed almost no interaction with the three GPCR C-tails tested ( Figure 6A). As it was observed for GASP-1, these results suggested that the central portion and the two GASP motifs of GASP-2 are important for its interaction with GPCRs.
In a second step, we focused on the two GASP motifs of GASP-2. As shown in Figure 6B, we replaced the four most conserved residues from these motifs (SWFW) by alanines, either individually (GASP2-m1 and GASP2-m2) or in combination (GASP2-dm). The resulting mutants were probed for their interaction with ADRB1, M 1 , and CALCR C-tails: compared to wild-type GASP-2, both GASP2-m1 and -m2 displayed a strong decrease in their interaction with ADRB1 and M 1 C-tails, but they retained around 70% interaction with CALCR. For the double mutant, GASP2dm, almost no interaction was detected with the ADRB1 and M 1 C-tails and a weak interaction was measured with the CALCR Ctail. Altogether, these results indicated that the GASP motif plays a crucial role in the interaction between GASP-2 and GPCRs. Interestingly, site directed mutagenesis experiments pointed to the highly conserved SWFW sequence within the GASP motif as a key element for interaction with GPCR C-tails.

A Small Synthetic Peptide Derived from the GASP Motif is Capable of Disrupting Interactions between GASPs and GPCRs
We further investigated whether small synthetic peptides containing the GASP motif could compete with GASPs for interaction with GPCRs. In a first step, we performed competition experiments with a synthetic peptide of 24 amino acids containing the first GASP motif of GASP-2 (GASP peptide). Peptide concentrations ranging from 1 to 250 mM were tested for their capacity to disrupt the interaction between ADRB1 C-tail and GASP-2. As shown in Figure 7A, increasing amounts of peptide led to a decrease in the amount of GASP-2 that was retained by the ADRB1 C-tail, while addition of the highest dose of a scrambled peptide did not affect this interaction. Inhibitions of 6566% and 7666% were obtained with 100 mM and 250 mM GASP peptide, respectively.
In a second step, we used a single peptide concentration (150 mM) in competition experiments with four different receptor C-tails, ADRB1, M 1 , CALCR, and TXA 2 , and three different GASPs, GASP-1 and -2 that contain GASP motifs and GASP-7 that does not. Figure 7B displays a representative experiment for ADRB1 C-tail and the three GASPs. Addition of 150 mM GASP peptide almost completely prevented the interaction between GASP-1, -2, -7 and the ADRB1 C-tail, while scrambled peptide did not. Quantifications revealed that in the presence of 150 mM of GASP peptide, the four receptor C-tails retained between 6% and 35% of GASP-1, -2 and -7 compared to control values ( Figure 7C). Altogether, these results confirmed that the GASP motif is mandatory for the interaction of GASPs with GPCRs. Moreover, the interaction between GASP-7, which does not exhibit the GASP motif, and GPCRs was also inhibited by the GASP peptide suggesting that the two GASP subfamilies (with or without GASP motif) most likely interact with the same region of the GPCR carboxyl-terminal domain but with a distinct mode of binding.
Using SPR, we have shown that the central part of GASP-1 can efficiently interact with full-length ADRB2 and CNR2. In a second series of experiments, we examined whether the GASP peptide could also compete with binding of full-length receptors to the central domain of GASP-1. A single concentration of purified ADRB2 or CNR2 were preincubated with different concentrations of peptide before injecting the mixture to a surface with central domain of GASP-1 prepared the same way as for the doseresponse experiments (Figure 4). Signals from free GST and from buffer injections were subtracted from the curves. As expected from GST-pull down competition experiments (Figure 7), preincubation with 250 mM GASP peptide strongly decreased the binding of both receptors to the central domain of GASP-1 ( Figure 8A-B). This competition was dose-dependent with an IC 50 estimated to be around 100 mM ( Figure 8C). In contrast, a control peptide where the conserved SWFW motif was changed to AAAA only had a minor effect on the interaction (Figure 8). Overall, these results indicate that the interaction between GPCRs and the central part of GASP-1 is specific and that the GASP motifs in the central domain are critical for the interaction with GPCRs.

Conclusions
In our previous study we have identified a novel family of proteins from which the two first members, GASP-1 and GASP-2, interact with GPCRs [6]. We provide here the first evidence that other members of the GASP family can also interact with the Ctail of GPCRs. Although some members, including GASP-6 and GASP-9, displayed very weak level of interaction (about 1% to 2% of the GASPs input), it was significantly higher than non-specific interaction that was observed with GST alone or some receptor Ctails fused to GST, such as MOR or TGF b . As GASP-1 has been shown to modulate the postendocytic sorting of some GPCRs, our results raise the possibility that other GASPs could be implicated in similar functions, thus adding another level of complexity in agonist-induced intracellular trafficking of GPCRs. A key issue will then be to define how the selectivity of interaction between GASPs and GPCRs is achieved in a cellular context.
Among the receptor C-tails that we have tested, ADRB1 is the one that displays the highest level of interaction in vitro with GASP-1 and GASP-2. This result is in contrast with previous results showing a low interaction of GASP-1 with ADRB1 C-tail [10]. This discrepancy is most likely due to the fact that we used fulllength GASPs to perform our experiments, while Heydorn and collaborators used a truncated form of GASP-1, corresponding to amino-acids 898 to 1395. When we performed GST-pull down experiments with a similar fragment of GASP-1 (corresponding to amino acids 924 to 1395), only 5% of the input was retained by the ADRB1 C-tail (data not shown), which is in agreement with the results of Heydorn et al. [10]. Therefore, our results suggest that the repertoire of GPCRs that interacts with GASP-1 could prove to be even larger than previously anticipated [10].
While numerous GPCRs have been shown to interact with GASP-1 [6,7,10], little is known about the molecular mechanisms underlying the GASP-GPCR interaction. In the present study, we show that a small repeated motif of 15 AA present 22 times in GASP-1 and twice in GASP-2 to -5, is critical for the interaction of GASPs with GPCR. Previous studies have only focused on the conserved carboxyl-terminal region of the GASP family in the interaction with GPCRs [6,7,10]. Although our results do not exclude a role of this region (see Figure 3B), they clearly show that this motif, that we named ''GASP motif'', is mandatory for the interaction of GASPs from subfamily one with GPCRs and represents a new protein-protein interaction motif. These results, together with the fact that GASP-1 contains 22 GASP motifs, suggest that one GASP-1 molecule could interact with several receptors. Moreover, beside their interaction with GPCRs, recent studies have shown that different members of the GASP family can interact with several non-GPCR proteins, including growth factor receptors and ubiquitin ligases [13]. It is therefore tempting to speculate that, like arrestins or multi-PDZ proteins, GASP-1 could function as an adaptor protein assembling GPCRs and other proteins in order to promote receptor function, signaling or trafficking. Further studies are required to evaluate the functional relevance of these interactions. Concerning GASPs from subfamily two that do not contain GASP motifs, although weaker, we also observed interactions in vitro with different GPCRs (Figures 2 and  7). Moreover, the interaction of GASP-7 with these different GPCRs was also blocked by a synthetic peptide containing a GASP motif (Figure 7). These data suggest that the two GASP subfamilies (with or without GASP motif) most likely interact with the same region within carboxyl-terminal domain of GPCRs but with a distinct mode of binding. Although the region within GASPs from subfamily 2 that promote the interaction with GPCRs remains to identify, we propose that the conserved carboxyl-terminal domain of the GASP family is involve in this interaction as it is in GASP-1 and DOR interaction ( Figures 3A  and 3B). Figure 6. The GASP motif is critical for the interaction of GASP-2 with GPCRs. A, GST-pull down experiments with two truncated mutants of GASP-2 and ADRB1, M 1 and CALCR C-tails. Grey boxes represent the 15 AA GASP motifs. Deletion analysis revealed that the central domain of GASP-2, which contains the two GASP motifs, is critical for the interaction between GASP-2 and ADRB1, M 1 and CALCR C-tails. B, GST-pull down experiments with full-length GASP-2 where one (GASP2-m1 and GASP2-m2) or both GASP motifs (GASP2-dm) were mutated. Grey boxes represent the wild-type motifs and X represent the mutant motifs. Consensus sequences are given for wild-type and mutant motifs. Mutated amino acids are underlined. Site directed mutagenesis analysis of these two repeated motifs showed that they played a crucial role in the interaction of GASP-2 with the three receptor C-tails tested here. Results are shown as percent of the wild-type GASP-2 interaction and correspond to the mean 6 S.E.M of three independent experiments. doi:10.1371/journal.pone.0056336.g006 Figure 7. A small synthetic peptide derived from the GASP motif of GASP-2 blocks the interaction between GASPs and GPCR C-tails in GST-pull down experiments. A, GASP peptide competes for the interaction between GASP-2 and GST-fused ADRB1 Ctail. The scrambled peptide displayed no significant effect on the interaction between GASP-2 and ADRB1. B, A fixed concentration of GASP peptide (150 mM) inhibits the interaction between GASP-1, -2 or -7 with ADRB1 C-tail, but not the scrambled peptide. C, Phosphorimaging quantification of the competition experiments for the interaction between GASP-1, -2 and -7 and four different receptor Ctails with GASP peptide. A fixed concentration of GASP peptide (150 mM) strongly inhibited interactions of GASPs with ADRB1, M 1 , CALCR and TXA 2 C-tails. Results are represented as percent of the interaction between the corresponding GASPs and GPCRs in absence of peptide (mean 6 S.E.M of three independent experiments). doi:10.1371/journal.pone.0056336.g007 In this study we have set up an assay to examine the interaction between immobilized GASP protein and detergent solubilized fulllength GPCRs using SPR. This approach is well-suited for a quantitative study of the interactions between the GASP family of proteins and any GPCR, because (i) most soluble proteins can be expressed as GST fusion proteins and captured by anti-GST antibodies without disturbing their function, (ii) immobilization of GPCRs, which can easily impair their function [20], is avoided and (iii) a minimum of GPCR handling is required. Optical biosensors have previously been used to study the interaction between recombinant GPCRs and G proteins [21][22][23][24], but this is one of the first times that interactions between full-length GPCRs and another GPCR interacting protein are studied with such techniques. We have determined the dissociation rate constant for the interaction of the central domain of GASP-1 with full-length ADRB2 and CNR2 and estimated the affinity within an order or magnitude. The dissociation rate constants and the affinities of the two receptors are very close, indicating that the mode of interaction with the central domain of GASP-1 is similar. The slow dissociation constants (k d ,10 23 s 21 ) suggest a high stability of the GASP-GPCR complex. The relatively slow association rate constant found in both cases does however not necessarily mean that the association rate is slow under native conditions (i.e. in cells), since the association rate-in contrast to the dissociation rate-depends on the local concentration of the interacting proteins. In agreement with GST-pull down experiments, SPR competition experiments revealed that the GASP motif and the amino-acids SWFW within this motif are strongly involved in formation of GASP-GPCRs complexes. As GASP-1, arrestins can form stable complexes with GPCRs [25,26], and have been suggested to play a role in the postendocytic sorting of receptors [7,27]. Interestingly, cell studies have indicated that arrestins can also form transient complexes with GPCRs and that these receptors are rapidly recycled instead of being degraded or slowly recycled [26]. Whether GASPs can form transient complexes with some GPCRs remains to be shown.
In summary, we have shown here that the GASP family is divided into two sub-families based on their interaction with C-tail of GPCRs: subfamily 1 comprises GASP-1 to -5 that strongly interact with receptor C-tails and contain a small repeated motif, the GASP motif, while sub-family 2 includes GASP-6 to -10 that weakly interact with the receptor C-tails and does not contain the GASP motif. We also report here the first molecular characterization of the interaction between GASPs and GPCRs. Our data cleary demonstrate that the GASP motif mediates the interaction of GASPs with G protein-coupled receptors and that a small peptide containing this motif is capable of preventing the interaction of GASPs with receptor C-tails and also full-length GPCRs. This study clearly highlight that we have identified a novel protein-protein interacting motif that is implicated in GPCR interactions and might be a new target for investigation of the role played by GASP in the modulation of the activity of GPCRs in vivo. Figure S1 Crosswise comparison of the conserved carboxyl-terminal domain of GASPs. Red color corresponds to sequence identity between 90% and 100%, orange dark between 75% and 90%, orange light between 45% and 75%, blue dark between 25% and 44% and blue light less than 25%. In addition to figure 1, this table shows that all GASPs display sequence similarities in their carboxyl-terminal tail and reveals very high sequence similarities between GASP-1 and GASP-2 as well as high similarities between GASP-6, -7, -8 and -9.