GIP, a G-protein-coupled receptor interacting protein

GIP, a G-protein-coupled receptor interacting protein

Regulatory Peptides 109 (2002) 173 – 179 www.elsevier.com/locate/regpep GIP, a G-protein-coupled receptor interacting protein Giulio Innamorati a,*, ...

405KB Sizes 1 Downloads 11 Views

Regulatory Peptides 109 (2002) 173 – 179 www.elsevier.com/locate/regpep

GIP, a G-protein-coupled receptor interacting protein Giulio Innamorati a,*, Michael Insuk Whang b, Raffaella Molteni a, Christian Le Gouill c, Mariel Birnbaumer d a

Scientific Institute San Raffaele, Universtita’ Vita-Salute, via Olgettina 58, I-20132 Milan, Italy b Department of Anesthesiology, UCLA School of Medicine, Los Angeles, CA 90095, USA c De´partement d’immunologie, Universite´ de Sherbrooke, Sherbrooke, Que´bec, Canada J1H 5N4 d Environmental Biology Program, Division of Intramural Research, Research Triangle Park, NC 27709, USA

Abstract A novel protein was cloned while screening for partners interacting with the second intracellular loop of the V2 vasopressin receptor (V2R). The protein was named GIP as in G-protein-coupled receptor Interacting Protein; the corresponding gene was located on the 17th chromosome where three exons encode for a 379-amino-acid protein. GIP subcellular localization was studied by immunocytochemistry and also using a biotinylating agent. The protein was found to be localized, at least in part, on the plasma membrane, probably in the form of a trimer. The results indicated that GIP is a transmembrane protein and the most part of the molecule is intracellular. Sequence homology inferred that GIP cytosolic domain is folded as a collagen-like helix followed by a globular domain. The interaction of the globular domain with the V2R was confirmed by pull-down experiments indicating that this structural motif can also interact with cytosolic proteins. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Vasopressin; V2 vasopressin receptor; C1q; ACRP30; TNF-alpha

1. Introduction G-protein-coupled receptors (GPCRs) are capable of activating simultaneously different intracellular signalling pathways, in some cases even bypassing the G-protein itself [1]. To trigger these responses, these receptors interact with various proteins involved either in transducing their signal, or in regulating their function. One of the most representative examples is arrestin, which not only can promote receptor desensitization and endocytosis, but can also mediate signalling through c-Src and its homologs [2]. These ‘interactors’ are often assembled as a multisubunit machinery, defined as signalosome. Its composition is likely to be changing continuously while the activated receptor traffics inside the cell. The combination of the different components of the complex may determine the type of signalling transduced by the receptor at a given time.

*

Corresponding author. Tel.: +39-2-2643-4731; fax: +39-2-2643-4723. E-mail address: [email protected] (G. Innamorati).

The V2 vasopressin receptor (V2R) is a seven-transmembrane domain receptor, mainly coupled to Gs. It regulates water readsorption in the kidney in reference to the blood levels of the nine amino acids neuropeptide arginine– vasopressin (AVP). A ‘two-hybrid’ screening was set up to search for new proteins interacting with the second intracellular loop of the V2R. This stretch of the protein contains an arginine that represents the most conserved residue within a signature characteristic of this family of receptors. As a result of the screening, a novel protein was identified and was named Gprotein-coupled receptor Interacting Protein (GIP).

2. Materials and methods 2.1. Yeast two-hybrid screening The second intracellular loop of the V2R (DRHRAICRPMLAYRHGSGAHWNRP) was utilized as a bait to screen a Human Kidney Matchmaker cDNA library (Clontech, Palo Alto, CA). The screening was performed according

0167-0115/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 0 11 5 ( 0 2 ) 0 0 2 0 1 - X

174

G. Innamorati et al. / Regulatory Peptides 109 (2002) 173–179

to the manufacturer’s instructions using a matchmaker GAL4 two-hybrid system 2 and CG-1945 as a reporter host strain. 2.2. Cloning A library for rapid amplification of cDNA ends (RACE) was prepared from human heart mRNA (Ambion) or HEK293 mRNA using the Marathon cDNA Amplification

kit (Clontech) according to the manufacturer’s instructions. The PCR fragments were separated by electrophoresis in a 0.8% agarose gel, extracted using the QIAquick gel extraction kit (Qiagen), and subcloned into a TA cloning vector, pCRII or pCR 2.1 (Invitrogen). Several overlapping clones were obtained, including one clone from human genomic DNA from blood cells. The DNA was sequenced and compared to complete the sequence deposited in genebank (accession number AF232905).

Fig. 1. GIP nucleotides and amino acid sequence. (A) GIP cDNA sequence. Underlined is the portion derived from the Y2 clone. Bold letters identify the codon corresponding to the first methionine, preceded by a stop codon in italic. (B) Predicted peptide sequence of GIP. Underlined is the predicted transmembrane domain [7]. The glycines involved in the collagen repeat are in italic. Boxed are the h-strand regions as predicted according to the crystal structure of ACRP30. In bold are the consensus common to both C1q and TNF families.

G. Innamorati et al. / Regulatory Peptides 109 (2002) 173–179

175

2.3. Microscopy

2.5. Other

To detect epitope-tagged full-length GIP and V2R, cells were stained as previously described [3] utilizing monoclonal antibodies anti-HA or anti-myc epitopes directly coupled to Alexa probes (Alexa 488 and Alexa 495). Digital imaging of stained cells was obtained using a Biorad laser scanning confocal microscope MRC-1024 (Bio-Rad Laboratories, Hercules, CA) and processed by Adobe Photoshop.

GST fusion proteins were obtained in pGEX-4T (Amersham Pharmacia). Preparation of GST-fusion proteins in Escherichia coli, in vitro translation of cDNA fragments, and binding of GST-fusion proteins to in vitro-translated fragments have been described [6]. Transient transfection of GIP-myc or the V2R in pCDNA3 was obtained utilizing Fugene (Roche) according to the manufacturer’s instructions.

2.4. Cell lysate and protein biotinylation 3. Results Cells were scraped, resuspended in Dulbecco’s phosphate-buffered saline (PBS), pelleted and lysed by addition of 1 vol of 2  Laemmli’s sample buffer [4] at a ratio of 150 Al per 106 cells. Boiling the sample was avoided to prevent the disappearance of the higher molecular weight band. Biotinylation of surface proteins was achieved according to Fabbri et al. [5]: briefly, cells were plated at 80% confluency on 60-mm-diameter dishes, washed twice in PBS and incubated for 1 h on ice in the presence of 0.5 mg/ml NHS-SS-biotin in PBS. Labelled cells were then washed carefully in cold PBS, collected by scraping and lysed in 500 Al of a solution containing 10 mM Tris pH 7.2, 150 mM NaCl, 1 mM CaCl2, 1 mM MgCl2 and 1% Nonidet P-40 for 10 min on ice. After immunoprecipitation, proteins were eluted from the beads in non-reducing sample buffer, separated by gel electrophoresis and transferred onto nitrocellulose. The membrane was blocked in 5% BSA and incubated with HRP-conjugated streptavidin. The peroxidase reaction was developed using enhanced chemiluminescence (ECL; Amersham).

3.1. GIP cloning To identify new components involved in the intracellular signalling mediated by the V2R, its second intracellular loop was utilized as a bait in a ‘‘two-hybrid system’’ screening of a HEK cells cDNA library. Out of 15  106 transformants, 12 positive clones were finally selected; 4 encoded the same clone Y2 (Fig. 1a). The full-length cDNA containing the Y2 clone was completed by a RACE-PCR-based approach. The template consisted of retro-transcribed cDNA libraries derived from human heart and HEK cells in addition to genomic DNA from human blood cells. In the 5Vuntranslated region, a stop codon was detected 36 bases upstream of the first methionine thus identifying the beginning of the protein (GI12053808) (Fig. 1a). Using GIP nucleotide sequence in a blast search against the human genome, it was possible to assign the gene to chromosome 17 at locus 17q25, encoded by three exons. The full-length clone has a predicted sequence of 379 amino acids. When GIP sequence was run against all non-redundant

Fig. 2. Interaction of GIP with the second intracellular loop of the V2R. pACTII carrying the V2R second intracellular loop (V2R IC2) or its C-terminus (V2R C-terminus) was introduced into the yeast strain CG-1945 (MAT a), in parallel Y2 in pAS2-1 was transformed into Y187 (MAT a). Transformant colonies were selected on synthetic medium lacking tryptophan or leucine, mated and plated on the same medium. The resulting colonies were lifted on filters and a h-galactosidase assay was performed. Y2 and the V2R IC2 show a positive staining similar to the positive control indicating the occurrence of the interaction.

176

G. Innamorati et al. / Regulatory Peptides 109 (2002) 173–179

Fig. 3. Direct binding of GIP to the second intracellular loop of the V2R. In the left panel, aliquots of in vitro translated Y2 or a1c Ca2 + channel subunit C-terminal segment labeled with 35S Met/Cys. In pull-down experiments, 10 times. The input of the lanes in the left panel was used for each sample in pull-down experiments. The results are shown in the right panel, GST fusion proteins of different intracellular segments of the V2R were compared in their ability to form a complex with the Y2 domain of GIP or with a portion of the a1c Ca2 + channel subunit. Only the beads adsorbed with the second intracellular loop of the V2R efficiently interacted with GIP. None of the GST fusion proteins, or GST alone showed significant interaction with the fragment of the a1c Ca2 + channel subunit.

GenBank CDS translations + PDB + SwissProt + PIR + PRF sequences, the closest homolog found was ZSIG37, a mouse protein of unknown function (GI12053808). The homology between the two proteins is extremely high (77% identity), but it covers only the carboxyl terminal two thirds of GIP since ZSIG37 is short of 98 amino acids. The interaction between GIP and the second intracellular receptor loop was verified by a two-hybrid system. The bait and the prey (V2R second intracellular loop in pAS2-1 and Y2 in pACT2) were reintroduced in two different yeast strains, CG1945 or Y187 (MAT a and MAT a, respectively).

Mating the resulting transformants yielded colonies testing positive for h-galactosidase activity (Fig. 2). The direct interaction was further confirmed by ‘pulldown experiments’. Fusion proteins were prepared linking either the C-terminus or the second or the third intracellular loop of the V2R to the glutathione-S-transferase (GST). Y2 (the GIP fragment initially obtained from the screening) was in vitro translated and labeled with 35S-Met and 35S-Cys; the resulting mixture was incubated with GST fusion proteins previously adsorbed to glutathione-Sepharose. After 30 min at room temperature, the Sepharose beads were washed, the

Fig. 4. Northern blot analysis of GIP distribution. (A) GIP mRNA appears to be expressed in many different tissues. The relative level of expression is particularly high in the heart. (B) The comparable loading of the different lanes was assessed using h-actin.

G. Innamorati et al. / Regulatory Peptides 109 (2002) 173–179

177

relatively low signal corresponding to GIP mRNA in the kidney is not surprising, at least in the context of the interaction with the receptor. 3.3. GIP subcellular localization To define GIP subcellular distribution, a ‘myc-epitope’ was added to the C-terminal amino acid. The resulting construct was transiently transfected in HEK or COS cells together with an epitope (HA)-tagged V2R. GIP distribution was compared to the V2R by the confocal microscope after immunostaining. No staining over the background could be detected under non-permeabilizing conditions (data not shown). However, permeabilizing the cells beforehand revealed intracellular vesicular structures surrounding the nucleus (Fig. 5). In addition, a weak staining was present in close proximity to the cell surface. Promoting receptor internalization by treating the cells with 100 nM AVP for 60 min at 37 jC did not cause a major GIP redistribution. Nevertheless, under these conditions, both the V2R and GIP could be colocalized in certain endosomal vesicles. These findings are consistent with the fact that the C-terminal portion of GIP interacts with the intracellular loop of the V2R. Two different bands appeared by SDS-PAGE when the epitope-tagged form of GIP was transiently expressed in HeLa or in COS cells and revealed by western blot (Fig. 6). The smaller band migrated with an apparent mass of approximately 42 kDa, consistent with the predicted moleFig. 5. Immunofluorescence of V2R-HA and GIP-myc. Transfected cells were stained by antibodies raised against the two epitopes and coupled to fluorophores. The V2R-HA (in green) was decorated under non-permeabilizing condition in the absence (upper panel) or presence (lower panel) of AVP. GIP-myc (in red) was stained after fixation and permeabilization. The majority of the protein is localized intracellularly in a vesicular pattern, probably contributed also by neosynthesized GIP-myc. A smaller percentage of the protein can be seen in close proximity of the plasma membrane. In a discrete number of vesicles, some overlap is seen whenV2R-HA internalization is obtained in presence of 100 nM AVP.

adsorbed complexes were resolved by SDS-PAGE and finally autoradiographed (Fig. 3). Only the second intracellular loop efficiently pulled down the C-terminal portion of GIP, whereas none of the GST fusion proteins interacted with the C-terminal portion of the a1c subunit of the L-type Ca2 + channel. 3.2. Endogenous GIP distribution To describe GIP distribution in different human tissues, the Y2 fragment was utilized as a probe in a northern blot. As shown in Fig. 4, a predominant band was found in heart while fainter bands were present in many other tissues including kidney. The signal detected in the heart is surprisingly high and certainly it does not appear to be correlated to V2R expression. Vice versa, given that the V2R is only expressed in the principal cells of the collecting duct, the

Fig. 6. Western blot analysis of epitope-tagged GIP. GIP-myc was transiently expressed in COS cells. The whole cells lysate was analyzed by SDS-PAGE and western blot was performed using anti-myc antibody. The arrows point to two bands appearing upon transfection. In the first lane, a lysate from mock-transfected cells was loaded.

178

G. Innamorati et al. / Regulatory Peptides 109 (2002) 173–179

Fig. 7. Western blot analysis of biotinylated GIP. GIP-myc was transiently expressed in COS cells. SS-biotin, applied extracellularly and under nonpermeabilizing conditions, was utilized to label proteins exposed to the extracellular milieu. Epitope-tagged GIP was immunoprecipitated and, after electrophoresis and transferring to nitrocellulose membrane, the biotinylated proteins were visualized using horseradish peroxidase linked to streptavidin. As a positive control, the h1 subunit of endogenous integrins was immunoprecipitated instead of GIP. JAB was used as a negative control as its intracellular localization prevents it from being biotinylated.

cular weight; the second and slower band appeared predominant and three times larger in size (f 130 kDa). To further explore the topology of GIP, non-permeabilized transfected cells were exposed to NHS-SS-biotin in PBS. GIP was next immunoprecipitated with anti-myc antibody and, after electrophoresis, transferred to nitrocellulose to be probed with an HRP-conjugated streptavidin. A single band appeared in the gel; its size was consistent with the larger molecular mass form, assumed to be trimeric (see discussion). Only the latter form appears exposed to the extracellular environment (Fig. 7).

comes from sequence homology to ACRP30 and C1q. GIP is an additional member of the C1q family [9], which is characterized by a collagen-like sequence (Gxy repeats) followed by a globular domain. The collagen domain is made of 14 glycine repeats, likely to be involved in the formation of a helix stabilizing the trimeric complex. In C1q, the correspondent domain is involved in the heteromerization of 18 different chains composed by three groups of six chains each. The sequence homology to ACRP30, and in general to the C1q family, continues with the ‘globular domain’. This domain is structured in a 10-strand jelly-roll folding topology that in GIP is approximately 140 residues long (Fig. 1b). Based on C1q and ACRP30 crystal structures, the resulting trimer is bell shaped with a hydrophilic region at the apex and a hydrophobic region at the base (Fig. 8). Albeit the absence of sequence homology, the same structure is found in the TNFa family [9]. Each protomer’s globular domain contains a hydrophobic core that is properly packed thanks to four residues (Y, G, F, L) that are present in GIP as well as in all known members of the C1q/TNF family. The similarity between C1q and the intracellular portion of GIP is about 30%. In the case of ACRP30, the level of similarity is slightly higher (36%). In both, C1q and ACRP30, a considerably higher number of glycine repeats (26 or 23, respectively) is present. The reason is probably the need of stabilizing a structure with a higher degree of complexity that consists of four homotrimers in ACRP30 and of six heterotrimers in C1q.

4. Discussion The data presented describe GIP, a new integral membrane protein that according to HMMTOP [7], contains a predicted transmembrane spanning helix from amino acid 106 to 123 (similar results were obtained utilizing different software). The extracellular N-terminus of GIP contains a TNF/ NGFR family cysteine-rich region signature [8]. This domain is found in the homologous region of a number of proteins, some of which are known to be receptors for growth factors. Often the domain contains three or four modules of about 40 residues containing six conserved cysteines. In the case of GIP, only two overlapping modules are present. Given the probable trimeric nature of the protein, it is possible that some interchain interactions might occur at the amino-termini level. The hypothesis that GIP might be assembled in a ‘bouquet-like structure’ formed by three identical monomers

Fig. 8. GIP putative topology across the plasma membrane. The extracellular domain of each chain is folded by disulphide bonds; intracellularly, the domains are tightly joined by glycine repeats stabilizing the interactions among the C-termini organized in globular domains.

G. Innamorati et al. / Regulatory Peptides 109 (2002) 173–179

Both C1q and ACRP30 are circulating proteins. C1q is part of the complement cascade while ACRP30 regulates glucose metabolism by enhancing insulin action [10]. Given the intracellular localization of GIP globular domain, no obvious information about its function can be gleaned from these homologs. C1q is the only molecule for which, to date, a function has been ascribed to the globular module. The domain is in fact interacting with the Fc regions of IgG and IgM that are crucial for the formation of the immune complex. Despite ACRP30 and C1q being secreted proteins, receptors for C1q globular domain have been also found intracellularly. The first example is gC1q receptor, also known as p32. Although it was originally described as a cell surface receptor, its intracellular distribution is widespread and recent studies demonstrated that it is tightly associated with mitochondria and various vesicular compartments [11]. This protein of 279 amino acids (in the prepro-form) binds to multiple proteins including the nuclear pre-mRNA splicing factor SF2/ASF and numerous other nuclear and cell surface proteins. gC1q receptor was found to bind to the alpha 1B adrenergic receptor regulating its expression level and subcellular localization. The interaction occurs at the C-terminus of the GPCR within the arginine-rich region [12]. Furthermore, direct interaction of gC1q receptor with various isoforms of PKC regulates their kinase activity [13]. The second example of an intracellular receptor for C1q is calreticulin: similarly to gC1q receptor, this protein was found at many different locations including membranebound organelles, the cell surface and the extracellular environment. The latter’s multiple functions are better characterized and include the modulation of store-operated Ca2 + influx. These examples suggest that proteins containing a globular domain could play a role by organizing or regulating the signalosomes. Overexpressing GIP did not affect V2R expression, signalling to adenylyl cyclase or receptor internalization (data not shown). The function played by GIP remains unknown; however, its modular structure could serve the purpose of recruiting together different V2R molecules. This speculation appears particularly appealing given the number of GPCRs that have recently been found to oligomerize [14]. There are examples of coiled-coil structures playing a role in the nucleus, in the cytosol and extracellularly [15]. It is possible that a similar widespread employment might occur for the collagen helix and the globular domain. Currently, our efforts are directed in this direction. In synthesis we demonstrated the interaction between the second intracellular loop of the V2R and a GIP, a novel transmembrane protein. These findings and the initial char-

179

acterization of the protein suggest a role for the globular domain in cell signalling.

Acknowledgements This work was supported by National Institutes of Health Grant DK-41-244 (to M.B.). We thank Dr. Elisabetta Bianchi and Dr. Grazisa Rossetti for the valuable discussion.

References [1] Brzostowski JA, Kimmel AR. Signaling at zero G: G-protein-independent functions for 7-TM receptors. Trends Biochem Sci 2001; 26(5):291 – 7. [2] Pierce KL, Lefkowitz RJ. Classical and new roles of beta-arrestins in the regulation of G-protein-coupled receptors. Nat Rev Neurosci 2001;2(10):727 – 33. [3] Innamorati G, Le Gouill C, Balamotis M, Birnbaumer M. The long and the short cycle. Alternative intracellular routes for trafficking of G-protein-coupled receptors. J Biol Chem 2001;276(16):13096 – 103. [4] Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227(259):680 – 5. [5] Fabbri M, Fumagalli L, Bossi G, Bianchi E, Bender JR, Pardi R. A tyrosine-based sorting signal in the beta2 integrin cytoplasmic domain mediates its recycling to the plasma membrane and is required for ligand-supported migration. EMBO J 1999;18(18):4915 – 25. [6] Qin N, Olcese R, Bransby M, Lin T, Birnbaumer L. Ca2 + -induced inhibition of the cardiac Ca2 + channel depends on calmodulin. Proc Natl Acad Sci U S A 1999;96(5):2435 – 8. [7] Tusnady GE, Simon I. The HMMTOP transmembrane topology prediction server. Bioinformatics 2001;17(9):849 – 50. [8] Naismith JH, Sprang SR. Modularity in the TNF-receptor family. Trends Biochem Sci 1998;23(2):74 – 9. [9] Shapiro L, Scherer PE. The crystal structure of a complement-1q family protein suggests an evolutionary link to tumor necrosis factor. Curr Biol 1998;8(6):335 – 8. [10] Berg AH, Combs TP, Du X, Brownlee M, Scherer PE. The adipocytesecreted protein Acrp30 enhances hepatic insulin action. Nat Med 2001;7(8):947 – 53. [11] Soltys BJ, Kang D, Gupta RS. Localization of P32 protein (gC1q-R) in mitochondria and at specific extramitochondrial locations in normal tissues. Histochem Cell Biol 2000;114(3):245 – 55. [12] Xu Z, Hirasawa A, Shinoura H, Tsujimoto G. Interaction of the alpha(1B)-adrenergic receptor with gC1q-R, a multifunctional protein. J Biol Chem 1999;274(30):21149 – 54. [13] Robles-Flores M, Rendon-Huerta E, Gonzalez-Aguilar H, MendozaHernandez G, Islas S, Mendoza V, et al. p32 (gC1qBP) is a general protein kinase C (PKC)-binding protein; interaction and cellular localization of P32 – PKC complexes in ray hepatocytes. Biol Chem 2002;15(277(7)):5247 – 55. [14] Angers S, Salahpour A, Bouvier M. Biochemical and biophysical demonstration of GPCR oligomerization in mammalian cells. Life Sci 2001;68(19 – 20):2243 – 50. [15] Beck K, Brodsky B. Supercoiled protein motifs: the collagen triplehelix and the alpha-helical coiled coil. J Struct Biol 1998;122(1 – 2): 17 – 29.