β-Arrestin Is Crucial for Ubiquitination and Down-regulation of the Insulin-like Growth Factor-1 Receptor by Acting as Adaptor for the MDM2 E3 Ligase*

  • Leonard Girnita
    Affiliations
    Department of Oncology and Pathology, Division of Cellular and Molecular Tumor Pathology, Cancer Center Karolinska, R8:04, Karolinska Hospital, SE-171 76 Stockholm, Sweden and the
    Search for articles by this author
  • Sudha K. Shenoy
    Affiliations
    Department of Medicine, Duke University, Medical Center, Durham, North Carolina 27710
    Search for articles by this author
  • Bita Sehat
    Affiliations
    Department of Oncology and Pathology, Division of Cellular and Molecular Tumor Pathology, Cancer Center Karolinska, R8:04, Karolinska Hospital, SE-171 76 Stockholm, Sweden and the
    Search for articles by this author
  • Radu Vasilcanu
    Affiliations
    Department of Oncology and Pathology, Division of Cellular and Molecular Tumor Pathology, Cancer Center Karolinska, R8:04, Karolinska Hospital, SE-171 76 Stockholm, Sweden and the
    Search for articles by this author
  • Ada Girnita
    Affiliations
    Department of Oncology and Pathology, Division of Cellular and Molecular Tumor Pathology, Cancer Center Karolinska, R8:04, Karolinska Hospital, SE-171 76 Stockholm, Sweden and the
    Search for articles by this author
  • Robert J. Lefkowitz
    Footnotes
    Affiliations
    Department of Medicine, Duke University, Medical Center, Durham, North Carolina 27710

    Howard Hughes Medical Institute and
    Search for articles by this author
  • Olle Larsson
    Correspondence
    To whom correspondence should be addressed. Tel: 46-8-51770366; Fax: 46-8-321047;
    Affiliations
    Department of Oncology and Pathology, Division of Cellular and Molecular Tumor Pathology, Cancer Center Karolinska, R8:04, Karolinska Hospital, SE-171 76 Stockholm, Sweden and the
    Search for articles by this author
  • Author Footnotes
    ∥ An investigator with the Howard Hughes Medical Institute.
    * This study was supported by grants from the Swedish Cancer Society, the Cancer Society in Stockholm, the Jubilee Fund of King Gustaf V, the Swedish Children Cancer Society, Alex and Eva Wallström's foundation, the Karolinska Institute, and National Institutes of Health Grants HL16037 and HL 70631 (to R. J. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
      The insulin-like growth factor-1 receptor (IGF-1R) plays important roles in physiological growth and aging as well as promoting several crucial functions in cancer cells. However, the molecular mechanisms involved in expression and down-regulation of IGF-1R are still poorly understood. Here we provide evidence that β-arrestin, otherwise known to be involved in the regulation of G protein-coupled receptors, serves as an adaptor to bring the oncoprotein E3 ubiquitin ligase MDM2 to the IGF-1R. In this way, β-arrestin acts as a crucial component in the ubiquitination and down-regulation of the receptor. Both MDM2 and β-arrestin co-immunoprecipitated with the IGF-1R. The β-arrestin isoform 1 appeared to be more strongly associated with the receptor than isoform 2, and in a molecular context it was 4-fold more efficient in inducing polyubiquitination of IGF-1R, a reaction that required the presence of β-arrestin and MDM2. Ligand stimulation accelerated IGF-1R ubiquitination. In mouse P6 cells (overexpressing human IGF-1R) absence of β-arrestin 1, but not of β-arrestin 2, blocked ubiquitination of IGF-1R. Conversely, in the two studied human melanoma cell lines both β-arrestin isoforms seemed to be involved in IGF-1R ubiquitination. However, because depletion of β-arrestin 1 almost completely eliminated degradation, and IGF-1 induced down-regulation of the receptor in these cells, whereas β-arrestin 2 only had a partial effect, β-arrestin 1 seems to the more important isoform in affecting the expression of IGF-1R. To our knowledge this is the first study demonstrating a defined molecular role of β-arrestin with direct relevance to cell growth and cancer.
      Insulin-like growth factor-1 receptor (IGF-1R),
      The abbreviations used are: IGF-1R, insulin-like growth factor-1 receptor; E3, ubiquitin-protein isopeptide ligase; GPCR, G protein-coupled receptor; siRNA, small interfering RNA; GST, glutathione S-transferase; E1, ubiquitin-activating enzyme; E2, ubiquitin carrier protein.
      1The abbreviations used are: IGF-1R, insulin-like growth factor-1 receptor; E3, ubiquitin-protein isopeptide ligase; GPCR, G protein-coupled receptor; siRNA, small interfering RNA; GST, glutathione S-transferase; E1, ubiquitin-activating enzyme; E2, ubiquitin carrier protein.
      together with its ligands IGF-1 and IGF-2, plays several crucial physiologic roles. IGF-1R is necessary for embryonic development and differentiation and is critical for normal postnatal growth. Tissue growth is centrally mediated by growth hormone causing local release of IGF-1 in the peripheral tissues, which in turn activates the IGF-1R (
      • Adams T.E.
      • Epa V.C.
      • Garrett T.P.
      • Ward C.W.
      ,
      • Dupont J.
      • Holzenberger M.
      ). The IGF-1 pathway has recently been demonstrated to be involved in regulation of aging (
      • Holzenberger M.
      • Dupont J.
      • Ducos B.
      • Leneuve P.
      • Geloen A.
      • Even P.C.
      • Cervera P.
      • Le Bouc Y.
      ). In addition, the IGF-1R is crucial for development and growth of malignant cells (
      • Baserga R.
      ,
      • Baserga R.
      • Hongo A.
      • Rubini M.
      • Prisco M.
      • Valentinis B.
      ,
      • Girnita L.
      • Girnita A.
      • Brodin B.
      • Xie Y.
      • Nilsson G.
      • Dricu A.
      • Lundeberg J.
      • Wejde J.
      • Bartolazzi A.
      • Wiman K.G.
      • Larsson O.
      ,
      • Girnita A.
      • Girnita L.
      • Prete F.d.
      • Bartolazzi A.
      • Larsson O.
      • Axelson M.
      ). The IGF-1R is a tetramer consisting of two ligand-binding extracellular α-subunits and two β-subunits involving a transmembrane domain, an intracellular tyrosine kinase domain and a C-terminal domain (
      • LeRoith D.
      • Werner H.
      • Beitner-Johnson D.
      • Roberts Jr., C.T.
      ). The mechanisms behind the regulation of IGF-1R in tumor cells are still poorly understood. Recent studies have reported an intriguing linkage between p53 and the IGF-1R pathway (
      • Girnita L.
      • Girnita A.
      • Brodin B.
      • Xie Y.
      • Nilsson G.
      • Dricu A.
      • Lundeberg J.
      • Wejde J.
      • Bartolazzi A.
      • Wiman K.G.
      • Larsson O.
      ,
      • Prisco M.
      • Hongo A.
      • Rizzo M.G.
      • Sacchi A.
      • Baserga R.
      ,
      • Buckbinder L.
      • Talbott R.
      • Velasco-Miguel S.
      • Takenaka I.
      • Faha B.
      • Seizinger B.R.
      • Kley N.
      ,
      • Werner H.
      • Karnieli E.
      • Rauscher F.J.
      • LeRoith D.
      ). This connection is interesting because it builds a bridge between a nuclear suppressor oncogene and a highly tumor-relevant tyrosine kinase at the cell surface. Recently we demonstrated that MDM2, an E3 ubiquitin ligase, whose most established role is to target and degrade p53, is associated with and ubiquitinates the IGF-1R (
      • Girnita L.
      • Girnita A.
      • Larsson O.
      ). Under conditions where expression of p53 was inhibited, MDM2 was redistributed and caused ubiquitination and proteasome-dependent degradation of the IGF-1R (
      • Girnita L.
      • Girnita A.
      • Larsson O.
      ).
      Although MDM2 is mostly known for its role in interaction with p53, it also binds to and interacts with other molecules (
      • Shenoy S.K.
      • McDonald P.H.
      • Kohout T.A.
      • Lefkowitz R.J.
      ,
      • Perry M.E.
      • Piette J.
      • Zawadzki J.A.
      • Harvey D.
      • Levine A.J.
      ,
      • Daujat S.
      • Neel H.
      • Piette J.
      ,
      • Li L.
      • Liao J.
      • Ruland J.
      • Mak T.W.
      • Cohen S.N.
      ,
      • Wang P.
      • Gao H.
      • Ni Y.
      • Wang B.
      • Wu Y.
      • Ji L.
      • Qin L.
      • Ma L.
      • Pei G.
      ). An interesting example is its interaction with the β-arrestins, which function as adaptor proteins for agonist-occupied G protein-coupled receptors (GPCR) that are phosphorylated by G protein-coupled receptor kinases (
      • Shenoy S.K.
      • McDonald P.H.
      • Kohout T.A.
      • Lefkowitz R.J.
      ). Activation of the prototypic GPCR, the β2-adrenergic receptor, results in ubiquitination of both β-arrestin and the receptor. Ubiquitination of β-arrestin is required for endocytosis of β2-adrenergic receptor by clathrin-coated pits. In cells lacking MDM2, β-arrestin ubiquitination does not occur, and receptor internalization is impaired, although receptor degradation is not (
      • Shenoy S.K.
      • McDonald P.H.
      • Kohout T.A.
      • Lefkowitz R.J.
      ). The nature of the ligase(s) ubiquitinating the β2-adrenergic receptor is still unknown.
      Unexpectedly, β-arrestin has been demonstrated to bind to ligand-stimulated IGF-1R and promote IGF-1R internalization (
      • Lin F.T.
      • Daaka Y.
      • Lefkowitz R.J.
      ). The role of β-arrestin in this context is unknown. Accordingly, in the present study we sought to investigate the potential role of β-arrestins in the MDM2-dependent ubiquitination and regulation of the IGF-1R.

      EXPERIMENTAL PROCEDURES

      Reagents—Polyclonal IGF-1R antibodies (N-20, C-20, and H-60) and a monoclonal antibody to phosphotyrosine (PY99) were from Santa Cruz Biotechnology Inc., Santa Cruz, CA. Rabbit polyclonal antibodies against β-arrestin 1/2 were generated in the Lefkowitz laboratory. A mouse monoclonal antibody against the human IGF-1R, a mouse monoclonal antibody to MDM2, and the proteasome inhibitor MG 132 were from Calbiochem. All other reagents unless stated otherwise were from Sigma.
      Cell Cultures—The human melanoma cell lines BE and DFB have been described elsewhere (
      • Girnita L.
      • Girnita A.
      • Brodin B.
      • Xie Y.
      • Nilsson G.
      • Dricu A.
      • Lundeberg J.
      • Wejde J.
      • Bartolazzi A.
      • Wiman K.G.
      • Larsson O.
      ). The R-(IGF-1R knock-out) and P6 (IGF-1R overexpressing) mouse cell lines (
      • Rubini M.
      • Hongo A.
      • D'Ambrosio C.
      • Baserga R.
      ) were gifts from Dr. Renato Baserga (Thomas Jefferson University, Philadelphia, PA). Mouse embryonic fibroblast were cultured as described elsewhere (
      • Shenoy S.K.
      • McDonald P.H.
      • Kohout T.A.
      • Lefkowitz R.J.
      ).
      Small Interfering RNAs (siRNAs)—Chemically synthesized, double-strand siRNAs, with 19-nucleotide duplex RNA and 2-nucleotide 3′-dTdT overhangs were purchased from Xeragon (Germantown, MD) in deprotected and desalted form. The siRNA sequences targeting human β-arrestin 1 and β-arrestin 2 have been reported previously (
      • Ahn S.
      • Nelson C.D.
      • Garrison T.R.
      • Miller W.E.
      • Lefkowitz R.J.
      ) and were used to deplete endogenous β-arrestin levels in BE and DFB cell lines. The siRNA sequences targeting mouse β-arrestin 1 and 2 are 5′-AAAGCCUUCUGUGCUGAGAAC-3′ and 5′-AAGGACCGGAAAGUGUUCGUG-3′. A non-silencing RNA duplex (5′-AAUUCUCCGAACGUGUCACGU-3′), as the manufacturer indicated, was used as a control for both human and mouse cells.
      Transfections—40–50% confluent cells in 25-cm2 flasks, split 24 h before transfection, were transfected with siRNA (see above) using the Lipofectamine 2000 (Invitrogen) according to the modified manufacturer's instructions. Briefly, 10 μl of transfection reagent was added to 300 μl of serum-free medium, while RNA mixtures containing 12 μlof20 μm (3.5 μg) RNA, and 188 μl of medium were prepared. Both solutions were allowed to stand 5–10 min at room temperature and mixed by inversion. After a 10–20-min incubation at room temperature, the entire transfection mixture was added to cells in a flask containing 3–4 ml of fresh, serum-free medium. After cells were incubated for 24 h at 37 °C, the medium was replaced with normal (serum-containing) growth medium. After additional incubation for 24 h, cells were divided into two flasks or 6-well plates for further experiments. All assays were performed at least 2 days after siRNA transfection. Transfections with pcDNA3 β-arrestin 1 and 2 and MDM mutants were performed as described elsewhere (
      • Shenoy S.K.
      • McDonald P.H.
      • Kohout T.A.
      • Lefkowitz R.J.
      ).
      Immunoprecipitation—The isolated cells were lysed as described elsewhere (
      • Girnita L.
      • Wang M.
      • Xie Y.
      • Nilsson G.
      • Dricu A.
      • Wejde J.
      • Larsson O.
      ). To detect IGF-1-stimulated ubiquitination in the absence of MG132 incubation, 10 mm N-ethylmaleimide was added to the lysis buffer. Fifteen μl of protein G plus-A/G-agarose and 1 μg of antibody were added to 1 mg of protein material. After overnight incubation at 4 °C on a rocker platform, the immunoprecipitates were collected by centrifugation in a microcentrifuge at 2,500 rpm for 2 min. The supernatant was discarded, whereupon the pellet was washed and then dissolved in a sample buffer for SDS-PAGE.
      SDS-PAGE and Western Blotting—Protein samples were dissolved in a sample buffer containing 0.0625 m Tris-HCl (pH 6.8), 20% glycerol, 2% SDS, bromphenol blue, and dithiothreitol. Samples corresponding to 50–100 μg of cell protein were analyzed by SDS-PAGE with a 7.5 or 10% separation gel. Molecular weight markers (Bio-Rad) were run simultaneously. Following SDS-PAGE the proteins were transferred overnight to nitrocellulose membranes (Amersham Biosciences) and then blocked for 1 h at room temperature in a solution of 5% (w/v) skimmed milk powder and 0.02% (w/v) Tween 20 in phosphate-buffered saline (pH 7.5). Incubation with the appropriate primary antibodies was performed for 1 h at room temperature. This was followed by washes with phosphate-buffered saline and incubation with either a horseradish peroxidase-labeled or a biotinylated secondary antibody (Amersham Biosciences) for 1 h. Following the biotinylated secondary antibody, incubation with streptavidin-labeled horse peroxidase was performed. The detection was made with either ECL (Amersham Biosciences) or by Supersignal West Pico reagents (Pierce). The films were scanned by Fluor-S (Bio-Rad).
      In Vitro Ubiquitination—In vitro ubiquitination of IGF-1R was performed essentially as described (
      • Fang S.
      • Jensen J.P.
      • Ludwig R.L.
      • Vousden K.H.
      • Weissman A.M.
      ). Recombinant glutathione S-transferase (GST)-MDM2 and -β-arrestin 1 (pGEX-4T1) constructs and His6-β-arrestin 2 (
      • Shenoy S.K.
      • McDonald P.H.
      • Kohout T.A.
      • Lefkowitz R.J.
      ) were expressed in Escherichia coli and purified using glutathione-Sepharose (Pierce). IGF-1R was isolated from P6 cells by immunoprecipitation with a polyclonal rabbit antibody directed against β-subunit (H60) and protein G-Sepharose (Amersham Biosciences). IGF-1R-Sepharose beads were mixed with or without GST-MDM2, rabbit E1 (Calbiochem), E2 bacterial recombinant UbcH5B (Calbiochem), His6-ubiquitin (Calbiochem), and with or without β-arrestins, in a 30-μl reaction. After a 1-h incubation at 37 °C the reaction was stopped by the addition of SDS sample buffer. Reaction products were loaded on a 7.5% polyacrylamide gel, transferred to nitrocellulose membrane, and detected using either antibody against IGF-1R (C20) or an anti-ubiquitin antibody (Santa Cruz).
      Analysis of IGF-1R Degradation—After the indicated experimental procedures, cells were transferred to methionine-free medium supplemented with 10% fetal bovine serum and 100 μCi/ml l-[35S]methionine (specific activity > 1,000 Ci/mm, Amersham Biosciences) for 24 h. The cells were carefully washed and transferred to radioactive-free methionine-containing medium supplemented with 10% fetal bovine serum for the indicated time periods. Cells were then quickly washed twice with ice-cold phosphate-buffered saline and lysed in radioimmune precipitation assay buffer (1× phosphate-buffered saline, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% SDS) containing dissolved protease inhibitor tablets (Roche Diagnostics). An equal amount of protein from each sample was immunoprecipitated with antibodies for the IGF-1R β-subunit (H-60) collected by protein A-Sepharose, resolved by SDS-PAGE, and visualized by autoradiography.

      RESULTS

      Identification of Domains in MDM2 Required for IGF-1R Ubiquitination—We have previously identified MDM2 as an E3 ubiquitin ligase that can ubiquitinate the IGF-1R (
      • Girnita L.
      • Girnita A.
      • Larsson O.
      ). To identify the regions involved in IGF-1R ubiquitination, we tested different deletion mutants of MDM2 for their ability to ubiquitinate IGF-1R. We transfected P6 cells (mouse cells overexpressing human IGF-1R) and BE cells (human melanoma cells) with wild type MDM2 (MDM21–491) and the C-terminal truncated variants MDM21–400 and MDM21–161. The former mutant contains the β-arrestin binding site but lacks the ligase domain, whereas MDM21–161 lacks both of these domains. The expression of the three different MDM2 variants, along with endogenous MDM2, was confirmed in Fig. 1A.
      Figure thumbnail gr1
      Fig. 1Full-length MDM2 is critical for ubiquitination of IGF-1R. P6 cells and BE cells were transfected with MDM21–491 (wild type), MDM21–400 (including the β-arrestin binding site but not the ligase domain), and MDM21–161 (excluding both the β-arrestin binding site and the ligase domain) for 48 h. Expression of the MDM2 variants was assayed by Western blot with the Ab-1 antibody (A). B, analysis of the effects of three different MDM2 variants on ubiquitination of IGF-1R, in the absence or presence of the proteasome inhibitor MG 132 (50 μm) (here denoted PI) during the last 3 h before harvesting. Equal amounts of immunoprecipitated (IP) IGF-1R were used for detection with ubiquitin antibodies. C, association of MDM2 variants with IGF-1R was assessed by immunoprecipitation with anti-MDM2 followed by Western blot detection of IGF-1R β-subunit. The bar graphs depict the quantification of MDM2-IGF-1R association. The cell lysates from the same samples were analyzed for expression of IGF-1R. Actin was used as loading control (C). All experiments were repeated at least three times with similar results. IB, immunoblot.
      We then investigated whether any of the MDM2 constructs could induce IGF-1R ubiquitination. Transfected P6 and BE cells either remained untreated or were treated with the proteasome inhibitor MG132 before harvesting. The cell lysates were immunoprecipitated for IGF-1R, and the same amounts of the immunoprecipitates (Fig. 1B, bottom panels) were fractionated and immunoblotted with an ubiquitin antibody. As seen in Fig. 1B, there is a basal level of ubiquitination of IGF-1R in the control cells. The expression of MDM21–161 did not affect this ubiquitination level, whereas MDM21–491 clearly increased it (especially in P6 cells). In the presence of proteasome inhibitor MG132 the level of ubiquitinated IGF-1R clearly increased suggesting that ubiquitinated IGF-1R is targeted for degradation by the proteasome system. The high molecular (90–190 kDa) ubiquitin smears are typical for polyubiquitinated IGF-1R (
      • Girnita L.
      • Girnita A.
      • Larsson O.
      ,
      • Vecchione A.
      • Marchese A.
      • Henry P.
      • Rotin D.
      • Morrione A.
      ,
      • Shan Y.X.
      • Yang T.L.
      • Mestril R.
      • Wang P.H.
      ). Interestingly, cells transfected with MDM21–400 did not show any IGF-1R ubiquitination at all (Fig. 1B).
      We also investigated the effects of the three MDM2 constructs in their ability to bind the IGF-1R and on IGF-1R expression. Immunoprecipitation of MDM2 (using an antibody recognizing the N-terminal part of MDM2) followed by detection of the IGF-1R β-subunit was carried out. Fig. 1C confirms that MDM2 is associated with IGF-1R in the untransfected P6 and BE control cells. The MDM2/IGF-1R association in cells expressing MDM21–161 was unaffected. This association was significantly increased in MDM21–400-expressing cells (p < 0.005 in both cell lines) (Fig. 1C, graph panels), whereas it was unaffected or abolished in cells overexpressing wild type MDM2. These observations, taken together with the findings that expression of IGF-1R was reduced or deleted (Fig. 1C, bottom panels) and IGF-1R ubiquitination increased (Fig. 1B) in cells transfected with MDM21–491, suggests that full-length MDM2 associates with, ubiquitinates, and degrades the IGF-1R. Because cells transfected with MDM21–400, in contrast to MDM21–161, did not show any IGF-1R ubiquitination at all (Fig. 1B) and “stabilized” IGF-1R (Fig. 1C), it appears to produce a dominant negative effect in this context.
      Involvement of β-Arrestin in MDM2-dependent IGF-1R Ubiquitination and Down-regulation—The impact of the β-arrestin binding site on MDM2-dependent ubiquitination suggests a role of β-arrestins in this process. Therefore we now investigated whether β-arrestins can associate with IGF-1R and MDM2. Lysates of cultured BE cells were subjected to immunoprecipitation of β-arrestin 1 or 2 or MDM2. The obtained immunoprecipitates were analyzed for the β-subunit of IGF-1R (Fig. 2, upper panel). Both β-arrestin immunoprecipitates contained clearly detectable levels of IGF-1R. However, the IGF-1R β-subunit appears to be more associated with β-arrestin 1 than with β-arrestin 2 (Fig. 2, upper panel). The IGF-1R·β-arrestin 1 complex was expressed at higher levels in the presence of the proteasome inhibitor. It is also confirmed in this panel that MDM2 is associated with IGF-1R but only in the presence of the proteasome inhibitor. The middle panel of Fig. 2 shows that β-arrestin 1 and 2 as well as IGF-1R immunoprecipitates contain MDM2. Finally, the IGF-1R and MDM2 immunoprecipitates contain both β-arrestin isoforms (Fig. 2, bottom panel). Together, these data suggest that β-arrestins bind to both IGF-1R and MDM2 at the same time.
      Figure thumbnail gr2
      Fig. 2β-Arrestins are associated with IGF-1R and MDM2. BE cells were harvested with or without a prior 6-h treatment with proteasome inhibitor (PI) MG 132 (50 μm). Cell lysates were subjected to immunoprecipitation (IP) with β-arrestin 1 or 2 antibodies (here denoted β1 and β2) or antibodies against MDM2 and then detected for β-subunit of IGF-1R (upper panel), or subjected to immunoprecipitation with β-arrestin 1 or 2 antibodies or antibodies against IGF-1R β-subunit and detected for MDM2 (middle panel), or immunoprecipitated with an IGF-1R or MDM2 antibody and detected for β-arrestin 1 and 2 (bottom panel).
      As mentioned above the region 161–400 of MDM2, which appears essential for IGF-1R ubiquitination, contains the earlier reported β-arrestin binding domain. This and the fact that β-arrestin can be recruited to IGF-1R (Fig. 2) raises the possibility of the involvement of β-arrestin in mediating IGF-1R ubiquitination. We investigated this possibility using an in vitro assay. We utilized recombinant MDM2, β-arrestin 1 and 2, and IGF-1R immunoprecipitated from P6 cells. All samples, which were also supplied with ligase E1 and E2, were analyzed by Western blotting for ubiquitin. The addition of MDM2 to an IGF-1R containing sample induced a clearly detectable polyubiquitination (Fig. 3A). When the β-arrestin isoforms 1 or 2 were added, the ubiquitin signals increased substantially. β-Arrestin 1 was almost 4-fold more efficient than β-arrestin 2 in this respect (Fig. 3A, lower panel). As also seen in Fig. 3A, MDM2 could, in the absence of exogenous β-arrestin, still mediate some ubiquitination of the IGF-1R. This could be explained by its interaction with endogenous β-arrestin. Consistently, we found that β-arrestin immunodepleted IGF-1R, which should be devoid of co-purifying endogenous β-arrestin, did not exhibit any ubiquitin signals when MDM2 was added (Fig. 3A, first lane from the right). In the absence of IGF-1R but in the presence of MDM2, β-arrestins, and all other reagents, no ubiquitination occurred (Fig. 3A, first lane from left). Taken together, these data strongly suggest involvement of β-arrestin in MDM2-dependent ubiquitination of IGF-1R.
      Figure thumbnail gr3
      Fig. 3Involvement of β-arrestins in IGF-1R ubiquitination. A, in vitro ubiquitination of IGF-1R. Sepharose beads of IGF-1R, isolated from P6 were mixed with ubiquitin (Ubiq) reagents with or without GST-MDM2 or with or without GST-β-arrestin 1 or His6-β-arrestin 2 as described under “Experimental Procedures.” Similar procedures were undertaken for R-(IGF-1R knock-out) as negative controls (first lane). To delete any co-purifying β-arrestins in the IGF-1R immunoprecipitate, immunodepletion using β-arrestin antibodies was performed. This β-arrestin-immunodepleted IGF-1R immunoprecipitate is shown in the last lane (on the right, marked –/– for β-arrestin). Equal amounts of the IGF-1R immunoprecipitates (IP, verified in the middle panel) were resolved by SDS-PAGE and finally detected by Western blotting using an antibody to ubiquitin. The bottom panel shows quantifications of ubiquitin signals (mean values and S.D.). All experiments were repeated four times with similar results. B, BE and DFB cells were treated with β-arrestin 1 or 2 siRNA (here denoted -β1 and -β2) for 72 h. The upper panel shows that siRNA had reduced β-arrestin 1 or 2 specifically at this time point. Separate cultures were treated with the proteasome inhibitor for additional 3 h. Equal amounts of immunoprecipitated IGF-1R were subjected to Western blot. Detection was made by ubiquitin antibodies. The bottom panel depicts quantification of the ubiquitin signals. IB, immunoblot. C, P6, BE and DFB cells were transfected with β-arrestin 1 and 2 (here denoted +β1 and +β2). Expression of each β-arrestin type (indicated as β-arr 1 and β-arr 2) is compared with wild type expression (upper panels) and effects on expression of IGF-1R β-subunit (lower panels). The untreated control cells were incubated with transfection reagents. The quantified IGF-1R signals are shown in the lower panel. All experiments were repeated at least three times with similar results.
      Next, we investigated whether β-arrestin was also required for IGF-1R ubiquitination in cultured cells. For this purpose we employed the siRNA technique to down-regulate the expression of endogenous β-arrestin 1 and 2 in BE and DFB cells. Analysis of the expression of β-arrestin 1 and 2 in the cell lysate confirmed that it was specifically decreased by siRNA (Fig. 3B, upper panel). The densitometry of three different determinations revealed that the decrease was 66 ± 6 and 72 ± 4% for β-arrestin 1 and 72 ± 6 and 68 ± 5% for β-arrestin 2 in BE and DFB cells, respectively. The lower panels of Fig. 3B show the effects of β-arrestin 1 and 2 siRNA on the ubiquitination of IGF-1R. The inhibition of β-arrestin 1 or 2 decreased IGF-1R ubiquitination. The signals were dramatically stronger when cells were incubated in the presence of the proteasome inhibitor before assay, suggesting that the ubiquitinated proteins are rapidly degraded in a proteasome-dependent manner. In DFB cells treated with the proteasome inhibitor, β-arrestin 1 siRNA had a stronger inhibitory effect (p < 0.01) on the ubiquitination of IGF-1R compared with siRNA against β-arrestin 2 (Fig. 3B, bottom panel). In BE cells there was no significant difference in this respect.
      In Fig. 3C the effect of overexpression of β-arrestin 1 or 2 on IGF-1R expression in P6, BE, and DFB cells is shown. In Fig. 3C, upper panel, the increased expression of the respective β-arrestin isoform is confirmed. The highest overexpression of the β-arrestins was seen in P6 cells. In all the three cell lines β-arrestin 1 and 2 transfections decreased the IGF-1R expression, suggesting that overexpression of β-arrestin accelerates ubiquitination and degradation of the receptor.
      Taken together the results presented in Fig. 3 suggest that β-arrestin functions as a cofactor or adaptor for MDM2 in ubiquitination of the IGF-1R, and in this way it can affect the expression levels of the receptor. To further elucidate the effect of β-arrestin on IGF-1R expression, we analyzed its possible influence on degradation of the receptor using pulse-chase labeling with [35S]methionine. As seen in Fig. 4,[35S]methionine-labeled IGF-1R of control BE and DFB cells (incubated in the presence of transfection reagents and control siRNA) exhibited a gradual decrease, and after 24 h ∼60% of the receptor had degraded in both cell lines. After reduction of β-arrestin 1, or both isoforms, there was almost no degradation of IGF-1R (Fig. 4). This stabilizing effect on IGF-1R was confirmed to be strongly significant in both cell lines (p < 0.005). The reduction of β-arrestin 2 in DFB cells did not affect IGF-1R degradation, which thus was comparable with the control. However, in β-arrestin 2 siRNA-treated BE cells there was a strongly accelerated decrease that significantly exceeded the control (Fig. 4B). In this experiment it was confirmed that siRNA decreased the levels of endogenous β-arrestin 1 and 2 adequately, substantially, and equally (65–76%) (data not shown).
      Figure thumbnail gr4
      Fig. 4Role of β-arrestins in degradation of IGF-1R under basal conditions. A, BE and DFB, were treated with β-arrestin 1 or β-arrestin 2 siRNA (here denoted -β-arr 1 and -β-arr 2) or transfection reagents (control, C) for 72 h. During the last 24 h of this period, the cells were labeled with [35S]methionine (100 μCi/ml). The cells were then washed and transferred to radioactive-free methionine-supplemented medium containing β-arrestin 1 or 2 siRNA or control siRNA for 3, 6, 12, and 24 h. IGF-1R β-subunit was immunoprecipitated and resolved by SDS-PAGE, and finally visualized by autoradiography. B, shows the densitometry data. The results are representative of three independent experiments.
      Ligand-induced IGF-1R Ubiquitination and Down-regulation—The data presented so far demonstrate the requirement of β-arrestin in MDM2-mediated ubiquitination and down-regulation of IGF-1R under basal conditions, i.e. cells were maintained in serum-supplemented medium during the entire experiment without addition of exogenous IGF-1. The usage of basal culture conditions, as opposed to ligand-stimulated ones, has the advantage that it better emulates physiological conditions. On the other hand, IGF-1-stimulated experimental conditions enable studies on molecular events that are specifically related to activation of the receptor. Therefore, IGF-1-stimulated conditions were now undertaken to determine effects on IGF-1R ubiquitination.
      We first investigated whether a short stimulation with IGF-1 could induce ubiquitination of IGF-1R and in such case whether β-arrestins are needed. To obtain strong signals this experiment was performed on the IGF-1R overexpressing cell line (P6). Fig. 5A depicts a representative result where the IGF-1R was immunoprecipitated and probed with an ubiquitin antibody. The ubiquitin signals were clearly increased in IGF-1-stimulated cells. The size and pattern of the bands differ considerably from those observed in BE and DFB cells under basal conditions. Here the size is in the 90-kDa range instead of 90–190 kDa. In addition, the signal is more distinct (Fig. 5A). This suggests that IGF-1R had undergone monoubiquitination following the 15-min IGF-1 stimulation. No polyubiquitination was seen. To ensure that the band actually represented monoubiquitinated IGF-1R, P6, and R-cells (IGF-1R knock-out) were treated with IGF-1 for 15 min. Ubiquitin immunoprecipitates from both P6 and R-cells were probed with an antibody specific to the IGF-1R β-subunit. The IGF-1R β-subunit protein bands were detected in the P6 cells but not in the R-cells (Fig. 5B). The IGF-1R monoubiquitination was increased with IGF-1 stimulation. No polyubiquitination signals were observed.
      Figure thumbnail gr5
      Fig. 5Role of β-arrestins in ligand dependent IGF-1R ubiquitination. A, P6 cells, stably expressing IGF-1R, were serum-starved for 2 h and then treated with IGF-1 for 15 min. Receptor protein was immunoprecipitated (IP) with IGF-1R β-subunit antibody and protein A and G-Sepharose beads. N-Ethylmaleimide (10 mm) was used in the lysis buffer. The top panel displays the signal obtained with an ubiquitin (Ubiq) antibody in a Western blot. The same blot was stripped and reprobed with a horseradish peroxidase-linked phosphotyrosine (pTyr) antibody (Calbiochem), lower panel. B, serum-depleted P6 and R-cells were kept either untreated or treated with IGF-1 (100 ng/ml) for 15 min. Cells were harvested, immunoprecipitated with ubiquitin antibody (P4D1, Santa Cruz), and blotted with an antibody specific to IGF-1R. Identical results were obtained in two independent experiments. C, P6 cells, transfected with control, β-arrestin 1, or β-arrestin 2 siRNA (mouse sequence-specific) for 72 h were serum-depleted and then treated with IGF-1 (100 ng/ml) for 15 min. IGF-1R was immunoprecipitated and probed for either ubiquitin or phosphotyrosine. Equal amounts of cell lysates were used to detect the expression levels of β-arrestins (bottom panel). D, the ubiquitinated IGF-1R bands in C were quantified by densitometry and graphically represented to summarize four independent experiments ± S.E. #, p < 0.02 less than control siRNA + IGF-1 (CTL); **, p < 0.002 less than β-arrestin 2 siRNA + IGF-1 as determined by paired t test. E, P6 cells, growing under basal conditions, were treated with β-arrestin 1 or β-arrestin 2 siRNA (here denoted -β1 and -β2) for 72 h, in the absence or presence of MG132. Equal amounts of immunoprecipitated IGF-1R were subjected to Western blot. Detection was made by ubiquitin antibodies. IB, immunoblot.
      To determine the role of β-arrestin isoforms in IGF-1-stimulated ubiquitination of IGF-1R, we employed siRNA to down-regulate β-arrestin 1 and 2. The specificity and effects of β-arrestins are seen in the lysate blots shown in Fig. 5C, left bottom panel. β-Arrestin 1 and 2 levels were decreased by 85 and 80% after quantification with densitometry, respectively (Fig. 5C). The quantification of IGF-1R ubiquitination from four independent experiments is shown in Fig. 5D. β-Arrestin 1 depletion nearly abolished the agonist-induced IGF-1R ubiquitination, whereas β-arrestin 2 depletion had minimal effect. There were no corresponding changes in the amount of ligand-induced phosphotyrosine signals in the receptor immunoprecipitates suggesting that the loss of IGF-1-induced ubiquitination was not because of a lack of receptor activation (Fig. 5C, middle panel). Rather the reduction of receptor ubiquitination upon β-arrestin 1 down-regulation indicates that β-arrestin 1 is necessary to mediate receptor monoubiquitination following receptor stimulation.
      To elucidate whether this requirement for β-arrestin 1 was specific for the experimental condition (IGF-1 stimulation) or was specific for the cell type, a complementary experiment under basal conditions with and without proteasomal inhibition was carried out on P6 cells. As demonstrated in Fig. 5E, P6 cells exhibited polyubiquitination of IGF-1R under this condition. This polyubiquitination was almost completely deleted by β-arrestin 1 siRNA, whereas β-arrestin 2 siRNA had no detectable effect (Fig. 5E). Thus, the P6 mouse cell line differs from the human melanoma cells in this respect (cf. Fig. 3B).
      The next step was to investigate whether β-arrestin may affect IGF-1R down-regulation in IGF-1-stimulated cells. BE and DFB cells, transfected with β-arrestin 1 or 2 siRNA for 72 h, were treated with IGF-1 for 3–24 h, after which the expression of IGF-1R was analyzed. It was confirmed that the two endogenous β-arrestin isoforms were adequately reduced (70–75%). In the controls of both cell lines (cells incubated in the presence of transfection reagents and control siRNA), there was a rapid decrease in IGF-1R expression, and by 12 h the level was reduced by as much as 60–70% (Fig. 6). After 12 h, on the other hand, the rate of IGF-1R decrease was considerably reduced. In both cell lines the absence of β-arrestin 1 completely stabilized the IGF-1R levels (p < 0.0005 at time points 12 and 24 h). In contrast, the absence of β-arrestin 2 actually enhanced the initial rate of decrease in IGF-1R levels such that as early as 3 h the receptor levels had dropped by as much as 40%, and there was no difference between the two cell lines in this respect (Fig. 6B). However, after this time point (3 h) there was hardly any further decrease in IGF-1R upon β-arrestin 2 reduction.
      Figure thumbnail gr6
      Fig. 6Role of β-arrestins in IGF-1 induced IGF-1R down-regulation. A, BE and DFB, which had been treated with β-arrestin 1 or β-arrestin 2 siRNA (here denoted -β-arr 1 and -β-arr 2) or transfection reagents (control, C) for 72 h, cells were treated with IGF-1 (100 ng/ml) for 3, 6, 12, and 24 h, in the absence or presence of β-arrestin 1 or 2 siRNA. The control cells were incubated with transfection reagents without siRNA. IGF-1R expression was determined by Western blotting and, B, quantifications by densitometry (after normalization to actin) (lower panel).
      From these experiments, we can conclude that β-arrestins are also important components for ligand-induced IGF-1R ubiquitination and down-regulation. Overall, the β-arrestin isoform 1 was substantially more efficient than isoform 2.

      DISCUSSION

      The present study provides strong evidence that β-arrestins, otherwise known for their role in regulation of seven-membrane-spanning receptors, serve as adaptors bringing the MDM2 ligase to the IGF-1R for ubiquitination and down-regulation of the receptor. Considered together with previous data demonstrating that β-arrestins increased IGF-1-induced extracellular signal-related kinase activation (
      • Lin F.T.
      • Daaka Y.
      • Lefkowitz R.J.
      ), a dual function for IGF-1R ubiquitination mediated by MDM2/β-arrestin is suggested; down-regulation of the receptor in conjunction with activation of IGF-1R signaling.
      The β-arrestins, β-arrestin 1 (
      • Lohse M.J.
      • Benovic J.L.
      • Codina J.
      • Caron M.G.
      • Lefkowitz R.J.
      ) and β-arrestin 2 (
      • Attramadal H.
      • Arriza J.L.
      • Aoki C.
      • Dawson T.M.
      • Codina J.
      • Kwatra M.M.
      • Snyder S.H.
      • Caron M.G.
      • Lefkowitz R.J.
      ), are ubiquitously expressed proteins (
      • Luttrell L.M.
      • Lefkowitz R.J.
      ). They play a well established role in the termination of receptor/G protein coupling. G-coupled protein receptor kinase phosphorylation and β-arrestin binding are central events for desensitization, sequestration, recycling, and down-regulation of most mammalian GPCRs. Although the HECT-type E3 ubiquitin ligase, AIP4, has been reported to lead to lysosomal trafficking of the GPCR, CXCR4, the role of β-arrestin in the process remains to be elucidated. The nature of the E3 ubiquitin ligases involved in β-arrestin-dependent degradation of GPCRs is largely unknown.
      Studies of mouse embryonic fibroblasts lacking either or both β-arrestins or of mammalian cells treated with siRNA have demonstrated that β-arrestin 1 and 2 exhibit functional specialization (
      • Luttrell L.M.
      • Lefkowitz R.J.
      ,
      • Lefkowitz R.J.
      • Whalen E.J.
      ,
      • Shenoy S.K.
      • Lefkowitz R.J.
      ,
      • Ahn S.
      • Wei H.
      • Garrison T.R.
      • Lefkowitz R.J.
      ). For example, reconstitution of β-arrestin expression in double knock-out mouse embryonic fibroblasts revealed that β-arrestin 2 is 100-fold more potent than β-arrestin 1 in the mediation of internalization of the β2-adrenergic receptor. Additionally for both the β2-adrenergic receptor and the V2R, β-arrestin 2 and not β-arrestin 1 was shown to be required for mediating receptor ubiquitination.
      The predominating role of β-arrestin 2 in β2-adrenergic receptor and vasopressin 2 receptor ubiquitination is diametrically opposite to the case in MDM2-dependent ubiquitination of IGF-1R shown in the present study. We observed that β-arrestin 1 is severalfold more efficient than β-arrestin 2 in inducing IGF-1R ubiquitination in a molecular context. Furthermore, β-arrestin 1 was more potent in regulating down-regulation of the IGF-1R, both under basal and IGF-stimulated conditions. Reducing β-arrestin 1 expression (using siRNA technology) effectively stabilized the cellular expression of IGF-1R, whereas β-arrestin 2 depletion had weaker effects. Only upon overexpression could β-arrestin 2 induce a strong down-regulation of the receptor. Indeed, inhibition of β-arrestin 2 abrogated or decreased the ubiquitin signals in the two human melanoma cell lines, BE and DFB, respectively, but this might at least partially explained by the fact that this treatment caused a strong decrease in IGF-1R expression. Overall, in contrast to other (7TM receptor) systems in which β-arrestin 2 is involved in receptor ubiquitination and degradation, it is evident β-arrestin 1 is the superior adaptor for MDM2 ligase E3 in ubiquitination of the IGF-1R. Actually, this is the only system thus far described in which a β-arrestin is shown to function as an adaptor to bring a specifically identified ligase to a receptor.
      The MDM2/β-arrestin-mediated IGF-1R ubiquitination and its potentially functional consequences are also supported by previous findings. For example, microinjection of a β-arrestin 1 antibody was observed to specifically inhibit IGF-1 mitogenic actions but had no effect on epidermal growth factor or insulin action (
      • Dalle S.
      • Ricketts W.
      • Imamura T.
      • Vollenweider P.
      • Olefsky J.M.
      ). In contrast, neither metabolic events, such as GLUT4 translocation, nor mitogenic signaling by insulin were dependent on these protein components (
      • Dalle S.
      • Ricketts W.
      • Imamura T.
      • Vollenweider P.
      • Olefsky J.M.
      ).
      Although it is often thought that receptor tyrosine kinases and GPCRs represent functionally different classes of cell surface receptors, involved in completely different physiological and pathophysiological processes, increasing evidence indicates that this is not strictly the case (
      • Dhanasekaran N.
      • Heasley L.E.
      • Johnson G.L.
      ,
      • Daub H.
      • Wallasch C.
      • Lankenau A.
      • Herrlich A.
      • Ullrich A.
      ,
      • Pace A.M.
      • Faure M.
      • Bourne H.R.
      ,
      • Maudsley S.
      • Pierce K.L.
      • Zamah A.M.
      • Miller W.E.
      • Ahn S.
      • Daaka Y.
      • Lefkowitz R.J.
      • Luttrell L.M.
      ,
      • Chen L.
      • Agrawal S.
      • Zhou W.
      • Zhang R.
      • Chen J.
      ,
      • Cao W.
      • Luttrell L.M.
      • Medvedev A.V.
      • Pierce K.L.
      • Daniel K.W.
      • Dixon T.M.
      • Lefkowitz R.J.
      • Collins S.
      ,
      • Luttrell L.M.
      • van Biesen T.
      • Hawes B.E.
      • Koch W.J.
      • Touhara K.
      • Lefkowitz R.J.
      ,
      • Povsic T.J.
      • Kohout T.A.
      • Lefkowitz R.J.
      ). Our findings provide further support for the idea of β-arrestins serving as nodes for molecular cross-talk between these two different receptors classes. For example, the accessibility of β-arrestin for a receptor belonging to one class may be affected by the activation status of receptors in the other class.
      Currently, almost nothing is known about the role of β-arrestins in normal and cancer cell growth. However, our present study, demonstrating a strict requirement for β-arrestins for the interaction between the oncoprotein MDM2 and the IGF-1 receptor, which is a highly relevant to cell growth and cancer, suggests a role of this adaptor protein in both physiological and malignant cell growth. Further studies are needed to elucidate the function of β-arrestin in relation to p53, MDM2, and IGF-1R in malignant cells. Perhaps β-arrestin works as a director in this network by regulating the traffic of MDM2 E3 ligase between p53 and IGF-1R. It would also be interesting to evaluate the role of β-arrestin 1 as a potential target to regulate the IGF-1 signaling in the context of neoplasia. Such an approach seems attractive, because blocking of β-arrestin 1 does not result in inhibition of the insulin receptor, which is highly homologous to the IGF-1R, and its associated metabolic events (
      • Dalle S.
      • Ricketts W.
      • Imamura T.
      • Vollenweider P.
      • Olefsky J.M.
      ).

      References

        • Adams T.E.
        • Epa V.C.
        • Garrett T.P.
        • Ward C.W.
        Cell Mol. Life Sci. 2000; 57: 1050-1093
        • Dupont J.
        • Holzenberger M.
        Birth Defects Res. C Embryo Today. 2003; 69: 257-271
        • Holzenberger M.
        • Dupont J.
        • Ducos B.
        • Leneuve P.
        • Geloen A.
        • Even P.C.
        • Cervera P.
        • Le Bouc Y.
        Nature. 2003; 421: 182-187
        • Baserga R.
        Cancer Res. 1995; 55: 249-252
        • Baserga R.
        • Hongo A.
        • Rubini M.
        • Prisco M.
        • Valentinis B.
        Biochim. Biophys. Acta. 1997; 1332: 105-126
        • Girnita L.
        • Girnita A.
        • Brodin B.
        • Xie Y.
        • Nilsson G.
        • Dricu A.
        • Lundeberg J.
        • Wejde J.
        • Bartolazzi A.
        • Wiman K.G.
        • Larsson O.
        Cancer Res. 2000; 60: 5278-5283
        • Girnita A.
        • Girnita L.
        • Prete F.d.
        • Bartolazzi A.
        • Larsson O.
        • Axelson M.
        Cancer Res. 2004; 64: 236-242
        • LeRoith D.
        • Werner H.
        • Beitner-Johnson D.
        • Roberts Jr., C.T.
        Endocr. Rev. 1995; 16: 143-163
        • Prisco M.
        • Hongo A.
        • Rizzo M.G.
        • Sacchi A.
        • Baserga R.
        Mol. Cell. Biol. 1997; 17: 1084-1092
        • Buckbinder L.
        • Talbott R.
        • Velasco-Miguel S.
        • Takenaka I.
        • Faha B.
        • Seizinger B.R.
        • Kley N.
        Nature. 1995; 377: 646-649
        • Werner H.
        • Karnieli E.
        • Rauscher F.J.
        • LeRoith D.
        Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 8318-8323
        • Girnita L.
        • Girnita A.
        • Larsson O.
        Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 8247-8252
        • Shenoy S.K.
        • McDonald P.H.
        • Kohout T.A.
        • Lefkowitz R.J.
        Science. 2001; 294: 1307-1313
        • Perry M.E.
        • Piette J.
        • Zawadzki J.A.
        • Harvey D.
        • Levine A.J.
        Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 11623-11627
        • Daujat S.
        • Neel H.
        • Piette J.
        Trends Genet. 2001; 17: 459-464
        • Li L.
        • Liao J.
        • Ruland J.
        • Mak T.W.
        • Cohen S.N.
        Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 1619-1624
        • Wang P.
        • Gao H.
        • Ni Y.
        • Wang B.
        • Wu Y.
        • Ji L.
        • Qin L.
        • Ma L.
        • Pei G.
        J. Biol. Chem. 2003; 278: 6363-6370
        • Lin F.T.
        • Daaka Y.
        • Lefkowitz R.J.
        J. Biol. Chem. 1998; 273: 31640-31643
        • Rubini M.
        • Hongo A.
        • D'Ambrosio C.
        • Baserga R.
        Exp. Cell Res. 1997; 230: 284-292
        • Ahn S.
        • Nelson C.D.
        • Garrison T.R.
        • Miller W.E.
        • Lefkowitz R.J.
        Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 1740-1744
        • Girnita L.
        • Wang M.
        • Xie Y.
        • Nilsson G.
        • Dricu A.
        • Wejde J.
        • Larsson O.
        Anti-Cancer Drug Des. 2000; 15: 67-72
        • Fang S.
        • Jensen J.P.
        • Ludwig R.L.
        • Vousden K.H.
        • Weissman A.M.
        J. Biol. Chem. 2000; 275: 8945-8951
        • Vecchione A.
        • Marchese A.
        • Henry P.
        • Rotin D.
        • Morrione A.
        Mol. Cell. Biol. 2003; 23: 3363-3372
        • Shan Y.X.
        • Yang T.L.
        • Mestril R.
        • Wang P.H.
        J. Biol. Chem. 2003; 278: 45492-45498
        • Lohse M.J.
        • Benovic J.L.
        • Codina J.
        • Caron M.G.
        • Lefkowitz R.J.
        Science. 1990; 248: 1547-1550
        • Attramadal H.
        • Arriza J.L.
        • Aoki C.
        • Dawson T.M.
        • Codina J.
        • Kwatra M.M.
        • Snyder S.H.
        • Caron M.G.
        • Lefkowitz R.J.
        J. Biol. Chem. 1992; 267: 17882-17890
        • Luttrell L.M.
        • Lefkowitz R.J.
        J. Cell Sci. 2002; 115: 455-465
        • Lefkowitz R.J.
        • Whalen E.J.
        Curr. Opin. Cell Biol. 2004; 16: 162-168
        • Shenoy S.K.
        • Lefkowitz R.J.
        Biochem. J. 2003; 375: 503-515
        • Ahn S.
        • Wei H.
        • Garrison T.R.
        • Lefkowitz R.J.
        J. Biol. Chem. 2004; 279: 7807-7811
        • Dalle S.
        • Ricketts W.
        • Imamura T.
        • Vollenweider P.
        • Olefsky J.M.
        J. Biol. Chem. 2001; 276: 15688-15695
        • Dhanasekaran N.
        • Heasley L.E.
        • Johnson G.L.
        Endocr. Rev. 1995; 16: 259-270
        • Daub H.
        • Wallasch C.
        • Lankenau A.
        • Herrlich A.
        • Ullrich A.
        EMBO J. 1997; 16: 7032-7044
        • Pace A.M.
        • Faure M.
        • Bourne H.R.
        Mol. Biol. Cell. 1995; 6: 1685-1695
        • Maudsley S.
        • Pierce K.L.
        • Zamah A.M.
        • Miller W.E.
        • Ahn S.
        • Daaka Y.
        • Lefkowitz R.J.
        • Luttrell L.M.
        J. Biol. Chem. 2000; 275: 9572-9580
        • Chen L.
        • Agrawal S.
        • Zhou W.
        • Zhang R.
        • Chen J.
        Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 195-200
        • Cao W.
        • Luttrell L.M.
        • Medvedev A.V.
        • Pierce K.L.
        • Daniel K.W.
        • Dixon T.M.
        • Lefkowitz R.J.
        • Collins S.
        J. Biol. Chem. 2000; 275: 38131-38134
        • Luttrell L.M.
        • van Biesen T.
        • Hawes B.E.
        • Koch W.J.
        • Touhara K.
        • Lefkowitz R.J.
        J. Biol. Chem. 1995; 270: 16495-16498
        • Povsic T.J.
        • Kohout T.A.
        • Lefkowitz R.J.
        J. Biol. Chem. 2003; 278: 51334-51339