β-Arrestins Regulate Mitogenic Signaling and Clathrin-mediated Endocytosis of the Insulin-like Growth Factor I Receptor*

β-Arrestins mediate agonist-dependent desensitization of G protein-coupled receptors and target the receptors to clathrin-coated pits for internalization. Here we report an expanded role of β-arrestins in promoting clathrin-mediated endocytosis of a tyrosine kinase growth factor receptor, i.e. the insulin-like growth factor I (IGF-1) receptor. β-Arrestins bind to the ligand-occupied IGF-1 receptors, promote their endocytosis, and enhance IGF-1-dependent mitogen-activated protein kinase phosphorylation and DNA synthesis. Our results suggest a role for β-arrestins in regulating mitogenic signaling and clathrin-mediated endocytosis of receptors not classically coupled to G proteins.

cytosis plays a dual regulatory role in receptor-mediated signaling pathways of both tyrosine kinase growth factor receptors and GPCRs. On the one hand, it attenuates some forms of extracellular signaling by reducing the number of cell surface receptors; on the other hand, it is critical for transducing some agonist-dependent intracellular responses. For example, it has been shown that receptor endocytosis is required for agonistinduced mitogenic signaling of various tyrosine kinase growth factor receptors, such as the receptors for epidermal growth factor (11), nerve growth factor (12), and IGF-1 (13), as well as GPCRs, such as the ␤ 2 -adrenergic receptor and receptors for lysophosphatidic acid, thrombin, and bombesin (14,15). Thus, inhibition of internalization of these receptors reduces agonistdependent MAP kinase activation.
The IGF-1 receptor is composed of a heterotetrameric ␣ 2 ␤ 2 structure and possesses ligand-induced tyrosine kinase activity (16). The activated receptor catalyzes both receptor autophosphorylation and tyrosine phosphorylation of exogenous proteins, including insulin receptor substrate-1 (IRS-1) (17,18) and -2 (IRS-2) (19) and the Shc proteins (20). Previous studies have suggested that a number of the biochemical signaling properties of the IGF-1 receptor are mediated through classical GPCR signaling pathways. For example, activation of MAP kinase by IGF-1 appears to utilize a G␤␥-mediated pathway essentially identical to that utilized by G i -coupled lysophosphatidic acid receptors (21). Thus, IGF-1 stimulation of MAP kinase phosphorylation in Rat-1 cells is blocked by pertussis toxin or by a G␤␥ sequestering reagent, the carboxyl terminus of G protein-coupled receptor kinase 2.
Based on the extensive structural homology to the insulin receptor, the ligand-occupied IGF-1 receptor is believed to utilize clathrin-coated vesicles for internalization (22,23). Because ␤-arrestins act as clathrin adaptors in GPCR endocytosis, we considered the possibility that ␤-arrestins might play a broader role in clathrin-mediated internalization of receptors other than those classically coupled to G proteins (heptahelical receptors).

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-The stable cell lines were generated as described (24). Only low passaged cells (less than 15 passages) were used in this study.
Immunoprecipitation and Western Blot Analysis-HEK 293 cells overexpressing the IGF-1 receptor and either form of ␤-arrestin were treated or not with 10 nM IGF-1 (Sigma), cross-linked or not with dithiobis(succinimidyl propionate) (DSP, Pierce) as described (25) and harvested in lysis buffer (24) containing 1 mM sodium vanadate, 10 mM sodium pyrophosphate, 50 mM sodium fluoride, and protease inhibitor mixtures. The IGF-1 receptor was immunoprecipitated using a monoclonal antibody specific to its ␣ subunit (Calbiochem), and the proteins were resolved by SDS-PAGE. Western blot analysis was performed using antibodies against both ␤-arrestin1 and ␤-arrestin2. The phosphorylated IGF-1 receptor was detected by an anti-phosphotyrosine antibody (RC20H, Transduction Labs). These immunoblots were visualized by ECL (Amersham Pharmacia Biotech). The nitrocellulose membranes were stripped and reprobed using a polyclonal antibody specific to the ␤ subunit of the IGF-1 receptor (Santa Cruz Biotech) to confirm equal expression of the IGF-1 receptor in each sample. The expression of K44A dynamin was detected by probing its immunoblot with a dynamin antibody (Transduction Labs). To detect tyrosine phosphorylation of Gab1, HEK 293 cells were co-transfected with pcDNA3/Gab1 and either a control vector or a ␤-arrestin expression plasmid. Cells were treated with IGF-1 or not for 5 min and were lysed in RIPA buffer (15) with 1 mM sodium vanadate and protease inhibitors. Comparably expressed Gab1 was immunoprecipitated with an antibody * This work was supported by the Howard Hughes Medical Institute and Grant HL16037 from the National Institutes of Health. 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.
‡ To whom correspondence should be addressed: Howard Hughes Medical Inst., Depts. of Medicine and Biochemistry, Duke University Medical Center, Box 3821, Durham, NC 27710. 1 The abbreviations used are: GPCR, G protein-coupled receptor; IGF, insulin-like growth factor; MAP, mitogen-activated protein; IRS, insulin receptor substrate; DSP, dithiobis(succinimidyl propionate); PAGE, polyacrylamide gel electrophoresis. raised against a glutathione S-transferase fusion protein comprising amino acids 23-189 of Gab1 and resolved by SDS-PAGE. The immunoblot was developed with an anti-phosphotyrosine antibody as described.
IGF-1 Receptor Internalization Determined by Flow Cytometry-Agonist-induced IGF-1 receptor internalization was determined by immunofluorescence flow cytometry analysis similar to that described for the ␤ 2 -adrenergic receptor (26). 2 days after transfection, cells were starved for 90 min and then treated with or without 10 nM IGF-1 at 37°C for 20 min. Next, cells were incubated on ice for 1 h with a monoclonal antibody specific to the ␣ subunit of the IGF-1 receptor (1:100). HEK 293 cells not exposed to the anti-IGF-1 receptor antibody were used as a control. Cells were washed and incubated with the fluorescein isothiocyanate-conjugated anti-mouse IgG (1:500, Jackson Labs) for an additional hour, then washed, and fixed in 8% formaldehyde in phosphate-buffered saline, pH 8. Flow cytometry analysis was performed to determine cell surface receptors. Receptor internalization was defined as the fraction of total cell surface IGF-1 receptors, which are removed from the plasma membrane following IGF-1 treatment.
Measurement of MAP Kinase Phosphorylation-The phospho-MAP kinase levels were measured as described (14,15).
[ 3 H]Thymidine Incorporation-Equal number of cells were starved for 30 h and then exposed to 10 nM IGF-1 for an additional 16 h. 1 Ci/ml of [methyl-3 H]thymidine (NEN Life Science Products) was added for 1 h, and then cells were rinsed twice with ice-cold phosphatebuffered saline, 5% trichloroacetic acid followed by ethanol. Cells were lysed in 0.1 N NaOH, and [ 3 H]thymidine incorporation was determined by counting in a liquid scintillation counter.

RESULTS AND DISCUSSION
To investigate the potential functional role of ␤-arrestins in clathrin-mediated endocytosis of the IGF-1 receptor, we first examined the ability of ␤-arrestins to interact with the IGF-1 receptor in HEK 293 cells. As shown in Fig. 1, IGF-1 treatment for 2 min followed by covalent cross-linking with DSP strikingly increased co-immunoprecipitation of the receptor with either ␤-arrestin1 (Fig. 1a) or ␤-arrestin2 (Fig. 1b) in cells overexpressing IGF-1 receptors and either form of ␤-arrestin. Agonist-promoted association of the IGF-1 receptor with endogenous ␤-arrestin1 and ␤-arrestin2 could also be detected in cells overexpressing IGF-1 receptors alone (Fig. 1c).
Previously we have demonstrated that phosphorylation/dephosphorylation of ␤-arrestin1 regulates clathrin-mediated endocytosis of the ␤ 2 -adrenergic receptor (24). Some of the biological functions of ␤-arrestin1 appear to be controlled by phosphorylation at a single serine, Ser-412, located close to the carboxyl terminus. Cytosolic ␤-arrestin1 is phosphorylated, and although competent to bind receptors, it is deficient in mediating downstream functions such as clathrin binding. Dephosphorylation of the ␤-arrestin1 occurs at the plasma membrane and is required for its activity to promote receptor internalization. The S412A and S412D mutants, which, respectively, mimic the dephosphorylated and phosphorylated forms of ␤-arrestin1, have been proven to be useful probes for exploring the roles of ␤-arrestin1 in GPCR function. The S412A mutant is generally more active in promoting GPCR internalization, whereas the S412D mutant serves as a dominant-negative.
As shown in Fig. 2a, transfection of either wild-type or mutant ␤-arrestins in HEK 293 cells stably overexpressing the IGF-1 receptor did not affect agonist-dependent tyrosine phosphorylation of the receptors. However, IGF-1 treatment for 1 min resulted in increased association of the IGF-1 receptor with wild-type ␤-arrestins and Ser-412 mutants (S412A and S412D) of ␤-arrestin1.
Next, we assessed the effect of ␤-arrestins on IGF-1 receptor internalization in HEK 293 cells. Incubation with IGF-1 for 20 min caused ϳ25% loss of cell surface IGF-1 receptors in control cells (Fig. 2b). Overexpression of either wild-type ␤-arrestin1 or ␤-arrestin2 increased the agonist-dependent loss of cell surface
Previously it has been shown that the process of receptor internalization plays an essential role in receptor-mediated mitogenic signaling (11)(12)(13)(14)(15). Similarly, inhibition of IGF-1 receptor internalization by the K44A mutant of dynamin significantly impaired IGF-1-stimulated MAP kinase phosphorylation (Fig. 3b). Considering the role of ␤-arrestins in IGF-1 receptor internalization, we next investigated IGF-1 receptormediated mitogenic signaling. First, we examined MAP kinase phosphorylation in HEK 293 cells stably overexpressing ␤-arrestins. At endogenous receptor levels, IGF-1-stimulated MAP kinase phosphorylation was moderately increased by overexpression of wild-type ␤-arrestin1 or ␤-arrestin2 and dramatically enhanced by overexpression of S412A ␤-arrestin1 (Fig. 4,  a and b). S412D ␤-arrestin1 acted as a dominant-negative mutant and significantly attenuated IGF-1-induced MAP kinase activation (Fig. 4, a and b). These results support the hypothesis that ␤-arrestin-mediated receptor endocytosis is required for MAP kinase activation via the IGF-1 receptor. Similar to the classical G i -coupled receptors, the IGF-1 receptor employs a mechanism involving both tyrosine phosphorylation and G ␤␥ subunits derived from pertussis toxin-sensitive G proteins in its MAP kinase signaling cascade (21). The enhancing effect of ␤-arrestins on IGF-1-stimulated MAP kinase activation was abolished by pertussis toxin (data not shown).
One of the earliest steps of the mitogenic signaling cascade triggered by the IGF-1 receptor is the receptor-catalyzed tyro- FIG. 4. Effect of ␤-arrestins on IGF-1 receptor-mediated MAP kinase phosphorylation. a, subconfluent HEK 293 cells stably overexpressing a control vector (mock), ␤-arrestin2, or ␤-arrestin1 (wild type, S412A, and S412D) were starved overnight. Cells were treated with or without IGF-1 for 3 min and harvested. Levels of phosphorylated ERK1 and ERK2 and total cellular ERK2 was determined as described above. The bottom panel is the same immunoblot stripped and reprobed with both anti-␤arr1 and anti-␤arr2 antibodies. b, IGF-1promoted MAP kinase phosphorylation was determined. Data shown represent the means Ϯ S.E. from four independent experiments. c, HEK 293 cells were co-transfected with pcDNA3/Gab1 plasmid and either an empty vector or a ␤-arrestin expression plasmid. Cells were starved overnight and incubated with or without IGF-1 for 5 min before lysis in RIPA buffer. Equally expressed Gab1 protein was immunoprecipitated, resolved by SDS-PAGE, and detected by an anti-phosphotyrosine antibody. Data shown are the means Ϯ S.E. from three independent experiments. The bottom panel is an immunoblot of Gab1 tyrosine phosphorylation. sine phosphorylation of IRS-1 and Shc proteins (17)(18)(19)(20). Recently it has been shown that the activated insulin receptor phosphorylates the Grb2-associated binder 1 (Gab1), which binds to and activates phosphatidylinositol 3-kinase (27). Phosphatidylinositol 3-kinase activity has been implicated in the G i -mediated MAP kinase signaling cascade upstream of Ras (28 -30). To determine the step in the MAP kinase signaling pathway at which ␤-arrestin-mediated receptor internalization is required, we examined IGF-1-stimulated tyrosine phosphorylation of IRS-1, Shc, and Gab1 proteins in HEK 293 cells overexpressing ␤-arrestins. We observed that the extent of IGF-1-induced tyrosine phosphorylation of IRS-1 and Shc proteins was not significantly altered by overexpression of wildtype or mutant ␤-arrestins (data not shown). Similar results have been reported for the tyrosine phosphorylation of Shc proteins mediated by the ␤ 2 -adrenergic receptor (15). In contrast, as shown in Fig. 4c, IGF-1-induced Gab1 tyrosine phosphorylation was enhanced by overexpressing S412A ␤-arres-tin1 but was impaired by overexpressing S412D ␤-arrestin1. This result suggests that ␤-arrestin-dependent internalization of the IGF-1 receptor is necessary for tyrosine phosphorylation of Gab1 and that Gab1 may serve as one of the important signaling intermediates in the IGF-1-mediated MAP kinase activation pathway.
In agreement with our results, Chow et al. recently reported that inhibition of IGF-1 receptor internalization in CHO cells by a chemical reagent, i.e. dansylcadaverine, impaired MAP kinase activation but not tyrosine phosphorylation of the receptor and IRS-1 (13). In contrast, however, they found that IGF-1-induced Shc tyrosine phosphorylation was reduced in the presence of dansylcadaverine. The reason for this apparent discrepancy is not yet clear, although it might relate to distinct mechanisms by which different reagents inhibit receptor endocytosis. The S412D ␤-arrestin1 mutant inhibits receptor endocytosis by interfering with receptor targeting to clathrin-coated pits (26). The mechanism of dansylcadaverine is not fully understood, although some evidence indicates that it inhibits receptor trafficking at a step proximal to the formation of endocytotic vesicles (31,32). It is possible that agonist-dependent tyrosine phosphorylation of Shc proteins proceeds at an earlier step that is inhibited by dansylcadaverine, but this is prior to ␤-arrestin-mediated receptor targeting to clathrincoated pits, whereas Gab1 phosphorylation following IGF-1 stimulation occurs thereafter.
To further investigate the regulatory role of ␤-arrestins in IGF-1 receptor-mediated mitogenic signaling, we assessed their effects on IGF-1-sitmulated DNA synthesis in these HEK 293 cell lines overexpressing ␤-arrestins. As shown in Fig. 5, IGF-1-dependent [ 3 H]thymidine incorporation was significantly increased by overexpression of S412A ␤-arrestin1, followed by ␤-arrestin2 and ␤-arrestin1 but was inhibited by overexpression of S412D ␤-arrestin1. Consistent with the effects of ␤-arrestins on the activation of MAP kinases, this result suggests that IGF-1 receptor endocytosis is also required for IGF-1-stimulated DNA synthesis.
Taken together, our findings suggest that ␤-arrestins may play a much wider role in receptor biology than had been previously imagined. In addition to their effects on internalization and signaling of GPCRs, they may be involved in regulating the function of a variety of receptor tyrosine kinases and other surface membrane receptors in response to ligand stimulation.