Regulation of Arrestin-3 Phosphorylation by Casein Kinase II*

Arrestins play an important role in regulating the function of G protein-coupled receptors including receptor desensitization, internalization, down-regulation, and signaling via nonreceptor tyrosine kinases and mitogen-activated protein kinases. Previous studies have revealed that arrestins themselves are also subject to regulation. In the present study, we focused on identifying potential mechanisms involved in regulating the function of arrestin-3. Using metabolic labeling, phosphoamino acid analysis, and mutagenesis studies, we found that arrestin-3 is constitutively phosphorylated at Thr-382 and becomes dephosphorylated upon β2-adrenergic receptor activation in COS-1 cells. Casein kinase II (CKII) appears to be the major kinase mediating arrestin-3 phosphorylation, since 1) Thr-382 is contained within a canonical consensus sequence for CKII phosphorylation and 2) wild type arrestin-3 but not a T382A mutant is phosphorylated by CKII in vitro. Functional analysis reveals that mutants mimicking the phosphorylated (T382E) and dephosphorylated (T382A or T382V) states of arrestin-3 promote β2-adrenergic receptor internalization and bind clathrin, β-adaptin, and Src to comparable levels as wild type arrestin-3. This suggests that the phosphorylation of arrestin-3 does not directly regulate interaction with endocytic (clathrin, β-adaptin) or signaling (Src) components and is in contrast to arrestin-2, where phosphorylation appears to regulate interaction with clathrin and Src. However, additional analysis reveals that arrestin-3 phosphorylation may regulate formation of a large arrestin-3-containing protein complex. Differences between the regulatory roles of arrestin-2 and -3 phosphorylation may contribute to the different cellular functions of these proteins in G protein-coupled receptor signaling and regulation.

Arrestins play an important role in regulating the function of G protein-coupled receptors including receptor desensitization, internalization, down-regulation, and signaling via nonreceptor tyrosine kinases and mitogen-activated protein kinases. Previous studies have revealed that arrestins themselves are also subject to regulation. In the present study, we focused on identifying potential mechanisms involved in regulating the function of arrestin-3. Using metabolic labeling, phosphoamino acid analysis, and mutagenesis studies, we found that arrestin-3 is constitutively phosphorylated at Thr-382 and becomes dephosphorylated upon ␤ 2 -adrenergic receptor activation in COS-1 cells. Casein kinase II (CKII) appears to be the major kinase mediating arrestin-3 phosphorylation, since 1) Thr-382 is contained within a canonical consensus sequence for CKII phosphorylation and 2) wild type arrestin-3 but not a T382A mutant is phosphorylated by CKII in vitro. Functional analysis reveals that mutants mimicking the phosphorylated (T382E) and dephosphorylated (T382A or T382V) states of arrestin-3 promote ␤ 2 -adrenergic receptor internalization and bind clathrin, ␤-adaptin, and Src to comparable levels as wild type arrestin-3. This suggests that the phosphorylation of arrestin-3 does not directly regulate interaction with endocytic (clathrin, ␤-adaptin) or signaling (Src) components and is in contrast to arrestin-2, where phosphorylation appears to regulate interaction with clathrin and Src. However, additional analysis reveals that arrestin-3 phosphorylation may regulate formation of a large arrestin-3-containing protein complex. Differences between the regulatory roles of arrestin-2 and -3 phosphorylation may contribute to the different cellular functions of these proteins in G protein-coupled receptor signaling and regulation.
Arrestins are cytosolic proteins primarily involved in the regulation of GPCR 1 function (1,2). Arrestins were initially described to attenuate or arrest intracellular signaling of GPCRs such as rhodopsin and ␤ 2 AR. In the continuous presence of a stimulus, many cellular receptors lose their responsiveness to stimuli through a process known as desensitization. Desensitization of GPCRs is largely mediated by two protein families, G protein-coupled receptor kinases (GRKs) and arrestins. GRKs specifically phosphorylate agonist-activated receptors. Subsequently, arrestins translocate from the cytosol and bind to plasma membrane-localized phosphorylated receptors. Binding of arrestin to the intracellular surface of receptors physically blocks interaction of the receptor with its cognate heterotrimeric G protein, thus terminating signal transduction.
Recent studies reveal that the role of arrestins in GPCR regulation is not limited to receptor desensitization. Nonvisual arrestins have also been demonstrated to mediate internalization (endocytosis) of various GPCRs. Both arrestin-2 and-3 directly bind clathrin and AP2, the two major structural components of the clathrin-based endocytic machinery (7,8). By doing so, they target receptors to clathrin-coated pits and promote receptor internalization. Nonvisual arrestins additionally mediate interaction with the GDP-bound form of ADP-ribosylation factor-6, a plasma membrane-specific regulator of endocytic membrane trafficking, and with ADP-ribosylation factor nucleotide binding site opener, an ADP-ribosylation factor guanine nucleotide exchange factor (9). Arrestin-2 also interacts with N-ethylmaleimidesensitive fusion protein (NSF), an ATPase that binds to SNARE complexes and plays an essential role in membrane trafficking (10). The interaction of nonvisual arrestins with various proteins involved in protein trafficking supports an important functional role for arrestins in receptor internalization.
Recently, arrestins have been shown to play a role not only in "negative" regulation of GPCRs but also in transmitting "positive" signals from receptors via interaction of arrestins with Src family tyrosine kinases and mitogen-activated protein kinases such as extracellular signal-regulated kinases 1 and 2 and c-Jun N-terminal kinase 3 (11)(12)(13)(14)(15)(16)(17). In the case of extracellular signal-regulated kinases 1 and 2 and c-Jun N-terminal kinase 3 interaction, arrestin-3 also directly binds the mitogenactivated protein kinase kinase kinases Raf-1 and ASK1, respectively, thereby serving as a scaffolding molecule and mediating GPCR-stimulated activation of the specific mitogenactivated protein kinase cascades (15,16).
Despite numerous studies suggesting an essential role of arrestins in GPCR regulation, the detailed mechanism under-lying the function of arrestins and receptor-mediated modulation of arrestin activity is unclear. However, several reports on arrestin phosphorylation provide some insight into potential regulation of arrestin function. Studies with Drosophila visual arrestin-2 demonstrate that it undergoes light-dependent phosphorylation by calcium/calmodulin-dependent kinase II and that phosphorylation is necessary for dissociation of arrestin from rhodopsin (18,19). In comparison, mammalian arrestin-2 was reported to be constitutively phosphorylated at Ser-412 by extracellular signal-regulated kinases 1 and 2 and become dephosphorylated following ␤ 2 AR stimulation (20,21). Arrestin-2 phosphorylation was proposed to regulate the interaction of arrestin with clathrin and Src, since a S412D mutant mimicking phosphorylated arrestin-2 had reduced binding to these proteins (11,20). Notably, arrestin-3 shares ϳ80% amino acid identity with arrestin-2 but does not contain the corresponding serine residue. Thus, arrestin-3 might be phosphorylated at other sites or perhaps be regulated by different mechanisms. Here, we report that arrestin-3 is phosphorylated by casein kinase II at Thr-382 and that arrestin-3 phosphorylation may regulate its interaction with novel binding partners.
Cell Culture and Transient Transfection-COS-1 and HEK293 cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum, 100 units/ml penicillin G, and 100 g/ml streptomycin sulfate in a humidified atmosphere of 95% air and 5% CO 2 at 37°C. Transient transfection was done with Fugene-6 (Roche Molecular Biochemicals) according to the manufacturer's recommendations.
Metabolic Labeling-COS-1 cells transfected with pcDNA3-FLAG-␤ 2 AR, pcDNA3-arr2, or pcDNA3-arr3 constructs were incubated in phosphatefree DMEM (Invitrogen) for 30 min and labeled for 2 h in the same medium containing 0.1 mCi/ml [ 32 P]orthophosphate (PerkinElmer Life Sciences). For ␤ 2 AR stimulation, 10 M (Ϫ)-isoproterenol was added to the labeling medium 15 min before the end of labeling. Cells were then washed with ice-cold Tris-buffered saline (TBS) and harvested for immunoprecipitation of arrestins.
Subcellular Fractionation-COS-1 cells grown on 60-mm dishes, labeled with [ 32 P]orthophosphate, and incubated with or without isoproterenol were washed with TBS twice, incubated with 250 g/ml concanavalin A in TBS on ice for 20 min, and scraped into 3 ml of homogenization buffer (20 mM Tris, pH 7.4, 0.25 M sucrose, 2 mM EDTA, 50 mM NaF, 10 mM sodium pyrophosphate, and a mixture of protease inhibitors (10 g/ml aprotinin, 10 g/ml leupeptin, 100 g/ml benzamidine, 10 g/ml pepstatin, and 1 mM phenylmethylsulfonyl fluoride). Cells were then lysed using a Dounce homogenizer (15 strokes), and unbroken cells and nuclei were removed by centrifugation at 1000 ϫ g for 10 min. A crude plasma membrane fraction was prepared by centrifugation of the cell homogenate at 3000 ϫ g for 15 min, and the pellet was then suspended in 1 ml of immunoprecipitation buffer (IP buffer) (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100, 50 mM NaF, 10 mM sodium pyrophosphate, and protease inhibitors). The 3000 ϫ g supernatant was then centrifuged at 300,000 ϫ g for 30 min. The resulting pellet (vesicle membrane fraction) was dissolved in 1 ml of IP buffer. The remaining supernatant (cytosolic fraction) was diluted with an equal volume of IP buffer containing 2% Triton X-100. The protein concentration of the three subcellular fractions was determined using BCA protein assay kit (Pierce).
Immunoprecipitation of Radiolabeled Arrestins and Autoradiography-COS-1 cells labeled with [ 32 P]orthophosphate were washed with ice-cold TBS twice and lysed in 1 ml of IP buffer. Insoluble proteins were removed by centrifugation at 55,000 ϫ g for 20 min. The cell lysate or various subcellular fractions were precleared by incubation with protein A-agarose for 1 h at 4°C. Arrestin-2 or -3 proteins were immunoprecipitated by incubating with purified polyclonal antibodies, KEE and 182, respectively, for 2 h and subsequently with protein A-agarose for an additional 2 h. KEE is a highly specific anti-arrestin-2 antibody generated against the C-terminal 16 amino acids of arrestin-2 (KEEEEDGTGSPRLNDR) (24). 182 was generated against a glutathi-FIG. 1. Constitutive phosphorylation of arrestin-3 in COS-1 cells. a, arrestin-2 and -3 were transiently expressed in COS-1 cells, and 48 h after transfection, the cells were incubated in phosphate-free medium for 30 min and then in media containing 0.1 mCi/ml [ 32 P]orthophosphate for 2 h. The cells were then washed twice with ice-cold TBS and lysed in 1 ml of IP buffer, and arrestin-2 and -3 were immunoprecipitated using purified polyclonal antibodies. The resulting pellets were washed four times with IP buffer, solubilized in SDS sample buffer, subjected to SDS-PAGE, transferred to a PVDF membrane, and exposed to x-ray film (upper panel). Phosphorylated arrestin-2 and -3 are marked with arrows, and an additional phosphoprotein of ϳ150 kDa that appears to specifically immunoprecipitate with arrestins is denoted with an arrowhead. Arrestins were also visualized by immunoblotting the PVDF membrane using an anti-arrestin monoclonal antibody and chemiluminescence (lower panel). b, phosphoamino acid analysis of phosphorylated arrestins. The arrestin bands and the nonspecific phosphoprotein band (Ϫ lane) in a were excised and washed successively with methanol and water, and the proteins were eluted and digested with 6 N HCl for 1 h at 110°C. The samples were then lyophilized, dissolved in 3 l of 5 mM Tris-HCl, pH 8, separated on a silica gel plate, and detected by autoradiography. The positions of phosphoserine (p-Ser) and phosphothreonine (p-Thr) standards are marked with arrows. WB, Western blot. one S-transferase (GST) fusion protein containing amino acid residues 350 -409 of arrestin-3 and has less than 5% cross-reactivity with arrestin-2 in Western blot analysis. After washing with IP buffer four times, samples were subjected to SDS-PAGE and transferred to a PVDF membrane (DuPont). The membrane was then exposed to x-ray film at Ϫ80°C, and phosphorylation of arrestins was determined by autoradiography. The total amount of immunoprecipitated arrestins was detected by probing a comparable membrane with monoclonal antibody F4C1, an antibody that detects an epitope common to all mammalian arrestins (25), horseradish peroxidase-conjugated goat secondary antibodies, and enhanced chemiluminescence (Pierce).
Phosphoamino Acid Analysis-Metabolically radiolabeled and immunoprecipitated arrestins were resolved by SDS-PAGE and transferred to a PVDF membrane. The arrestin bands were excised, washed successively with methanol and water, eluted, and digested with 6 N HCl for 1 h at 110°C. The samples were lyophilized, dissolved in 3 l of 5 mM Tris-HCl, pH 8.0, and then spotted on a 20 ϫ 20-cm silica gel plate (Eastman Kodak Co.) along with phosphoamino acid standards. The amino acids were separated by thin layer chromatography (isopropyl alcohol/HCl/water ϭ 7:1.5:1.5) for 5 h. Phosphoamino acid standards were developed using 0.2% ninhydrin solution spray (Sigma), and [ 32 P]phosphoamino acids were detected by autoradiography.
In Vitro Kinase Assay-One microgram of purified arrestin-3 was incubated with 0.1 M casein kinase I (CKI), casein kinase II (CKII), GRK2, or GRK5 in 20 l of 50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 5 mM dithiothreitol, 100 M ATP, and 2 Ci of [␥-32 P]ATP for 30 min at 30°C. The reaction was stopped by adding 5 l of 6ϫ SDS-sample buffer, and the samples were run on a 10% SDS-polyacrylamide gel. The gel was stained with Coomassie Blue, dried, and exposed to x-ray film. CKI and CKII were purchased from New England Biolabs, and GRK2 and GRK5 were purified from Sf9 insect cells infected with recombinant baculoviruses.
For in vitro phosphorylation of arrestins in COS-1 extracts, cells overexpressing wild type or T382A mutant arrestin-3 proteins were harvested in buffer (20 mM Tris-HCl, pH 7.4, 100 mM NaCl, 2 mM EDTA, 50 mM NaF, 10 mM sodium pyrophosphate, and protease inhibitors) and ruptured by freezing and thawing. Insoluble materials were removed by centrifugation at 55,000 ϫ g for 20 min. Cell extracts (60 g of protein) were then incubated with or without 0.05 M CKII in the presence of 10 mM MgCl 2 , 50 M ATP, and 10 Ci of [␥-32 P]ATP for 1 h at 30°C. Reactions were stopped on ice, and Triton X-100 was added to a final concentration of 1%. Arrestins were immunoprecipitated as described above, and the phosphorylation was assessed by autoradiography.
Internalization Assay-Internalization of FLAG-␤ 2 AR was assessed by enzyme-linked immunosorbent assay as described (28). Briefly, COS-1 cells transfected with pcDNA3-FLAG-␤ 2 AR and various arrestin constructs were split into poly-L-lysine-coated 24-well tissue culture plates after 24 h. The next day (48 h post-transfection), cells were treated with 10 M (Ϫ)-isoproterenol and 0.1 mM ascorbate at 37°C for 0, 2, 5, 10, 20, or 30 min, fixed with 3.7% formaldehyde in TBS for 5 min at ambient temperature, and washed three times with TBS. Cells were then blocked with TBS containing 1% BSA (TBS/BSA) for 45 min, incubated with a primary antibody (M1 (Sigma), 1:1000 dilution in TBS/BSA) for 1 h, washed three times with TBS, blocked again with TBS/BSA for 15 min, incubated with a secondary antibody (goat antimouse IgG conjugated with alkaline phosphatase (Bio-Rad), 1:1000 dilution in TBS/BSA) for 1 h at room temperature, and washed three times with TBS. Colorimetric visualization of antibody binding was performed using an alkaline phosphatase substrate kit (Bio-Rad), and samples were read at 405 nm in a microplate reader using Microplate Manager software (Bio-Rad). The reading from cells that did not express flag-␤ 2 AR was used as blank and subtracted. The percentage of surface receptor loss was determined by calculating the change of antibody-accessible FLAG-␤ 2 AR.
In Vitro Translation of Arrestin-3 and Clathrin Binding Assay-Arrestin-3 constructs were transcribed and translated in vitro using the TNT-coupled reticulocyte lysate system (Promega) according to the manufacturer's recommendations. Briefly, pcDNA3-arr3 (wild type, T382A, or T382E) were added to a reaction mixture containing TNT buffer, 1 l of T7 RNA polymerase, 1 mM amino acid mixture (without leucine), 1 l of RNasin, 25 l of rabbit reticulocyte lysate, and 5 l of [ 3 H]leucine (1 mCi/ml, 179 Ci/mmol; PerkinElmer Life Sciences) in a total volume of 50 l. The reaction was incubated for 90 min at 30°C and stopped by snap-freezing. To calculate incorporated radioactivity, 1 l of reaction mix was spotted onto Whatman filter paper and then successively incubated in ice-cold 10% (w/v) trichloroacetic acid for 10 min, boiling 5% (w/v) trichloroacetic acid for 10 min, and 100% ethanol for 1 min. The precipitated materials were dissolved in 1% SDS and 50 mM NaOH, and radioactivity was counted in a liquid scintillation counter with scintillant. The in vitro translated proteins were also analyzed by SDS-PAGE.
The clathrin cage binding assay has been described previously (7). Briefly, 0.5 nM of in vitro synthesized 3 H-labeled arrestins were incubated with or without 200 nM clathrin cages for 10 min at ambient temperature in a total volume of 50 l in 100 mM Na-MES, pH 6.8, 1 mM dithiothreitol, 1 g/ml aprotinin, 1 g/ml leupeptin, 1 g/ml pepstatin, and 100 g/ml benzamidine. Samples were then cooled on ice, loaded onto a 75-l cushion of 0.2 M sucrose in 100 mM Na-MES, pH 6.8, and centrifuged at 100,000 rpm in a TLA100.1 rotor for 5 min at 4°C. Pellets were solubilized in 25 l of SDS-sample buffer supplemented with 100 mM Tris-HCl, pH 8, heated at 100°C for 5 min, and run on 10% SDS-polyacrylamide gels. The gels were stained with Coomassie Blue, destained, incubated with ENTENSIFY (PerkinElmer Life Sciences), dried, and exposed to x-ray film. After development, films were scanned with a densitometer (Molecular Dynamics, Inc., Sunnyvale, CA) to quantitate the amount of radiolabeled arrestins that were co-sedimented with clathrin cages. The percentage of bound arrestin was calculated by comparing the signal from the co-sedimented arrestins to a set of arrestin standards. Binding of in vitro translated arrestins to clathrin-coated vesicles (CVs) was done in the same manner as the clathrin cage binding assay using 50 nM CVs (50 nM clathrin trimers). Clathrin cages and CVs were generously provided by Dr. James H. Keen (Thomas Jefferson University).
In Vitro Binding of Arrestins to GST-␤ 2 -Adaptin-GST or GST-␤ 2adaptin (residues 700 -937) fusion proteins were expressed in BL21(De3)pLysS cells and purified on glutathione-agarose beads. Five micrograms of purified GST fusion proteins were incubated with 10 g of COS-1 cell lysates containing overexpressed arrestin-3, arr3(T382V), or arr3(T382E) in binding buffer (20 mM HEPES, pH 7.2, 120 mM  , and cytosol fractions were then prepared by differential centrifugation. Arrestin-3 phosphorylation in the total cell lysate and each fraction was determined as described under "Experimental Procedures." b, arrestin-3 and FLAG-␤ 2 AR were transiently expressed in COS-1 cells. At 48 h post-transfection, cells were incubated with (ϩ) or without (Ϫ) 10 M isoproterenol for 2 min at 37°C. The cells were then lysed, and arrestin was immunoprecipitated using an arrestin-3-specific polyclonal antibody as described under "Experimental Procedures." Coimmunoprecipitated PP2A was detected using a monoclonal antibody against the catalytic subunit of PP2A (PP2A-C) (upper panel), whereas immunoprecipitated arrestin-3 was detected using the monoclonal antibody F4C1 (lower panel). Endogenous PP2A-C in the lysate is shown in the upper panel.
potassium acetate, 0.1 mM dithiothreitol, 0.2% Triton X-100) for 1 h at 4°C (in a total volume of 100 l). The beads were then washed twice with 500 l of binding buffer, and bound arrestins were eluted by boiling the beads in SDS sample buffer for 5 min. The samples were electrophoresed on a 10% polyacrylamide gel, transferred to a nitrocellulose membrane, and detected by immunoblotting using purified rabbit polyclonal anti-arrestin-3 antibody 182.
Kinetics of Arrestin-3 Degradation-pcDNA3-arr3, arr3(T382A), or arr3(T382E) was transfected into COS-1 cells on 100-mm culture dishes, and 6 h later cells were split into four 60-mm culture dishes. At 24 h post-transfection, cells were rinsed twice with Met/Cys-free DMEM and incubated for 1 h to deplete the cellular content of methionine and cysteine. Cells were then metabolically labeled with [ 35 S]methionine and [ 35 S]cysteine by incubating in Met/Cys-free DMEM containing 0.2 mCi/ml of Trans-35 S-LABEL (ICN) for 2 h at 37°C (pulse). Cells were then washed twice with DMEM containing 10% fetal bovine serum and harvested either immediately (0 h chase) or after the indicated chase period (8,16, and 32 h). Arrestins were immunoprecipitated and electrophoresed on a 10% SDS-polyacrylamide gel, and radioactive arrestins were detected by fluorography. Levels of radioactive arrestin were quantified by determining band densities using a densitometer.
Search for Phosphorylation-dependent Arrestin-3-binding Proteins-Wild type arrestin-3, arr3(T382V), and arr3(T382E) were expressed in HEK293 cells (100-mm dishes), the cells were lysed in 1 ml of coimmunoprecipitation buffer after 48 h, and arrestin-3 was immunoprecipitated as described above. Immunocomplexes were incubated with 2ϫ SDS-sample buffer for 2 h at 27°C, resolved on a 10% SDS-polyacryl-amide gel, and transferred to a nitrocellulose membrane. Proteins were stained with Coomassie Blue, and arrestins were detected by probing the membrane with monoclonal antibody F4C1.

Constitutive Phosphorylation of Arrestin-3 in COS-1 Cells-
Previous studies have shown that arrestin-2 is constitutively phosphorylated in HEK293 cells and that the dephosphorylation of arrestin-2 that occurs upon stimulation of ␤ 2 AR is important for receptor internalization (20). To test whether arrestin-3 is also a phosphoprotein, COS-1 cells transiently expressing arrestin-2 or -3 were metabolically labeled with [ 32 P]orthophosphate, and the arrestins were immunoprecipitated using specific polyclonal antibodies (KEE and 182). The immunoprecipitated arrestins were run on an SDS-polyacrylamide gel, transferred to a PVDF membrane, and detected using an anti-arrestin monoclonal antibody. As shown in Fig.  1a (lower panel), KEE specifically precipitated arrestin-2, whereas 182 precipitated arrestin-3 efficiently and arrestin-2 to a lesser extent. Subsequent autoradiographic analysis of the immunoprecipitated arrestins revealed that arrestin-3 is indeed a phosphoprotein (Fig. 1a, upper panel). It also confirmed that arrestin-2 is phosphorylated, albeit more weakly than arrestin-3. The phosphorylation of arrestin-3 seems to be constitutive, since the cells were not stimulated during the metabolic labeling.
There was an additional major phosphoprotein of ϳ150 kDa on the autoradiogram (Fig. 1a, upper panel, marked with an  arrowhead). We explored the possibility that this protein might be the clathrin heavy chain, since nonvisual arrestins specifically interact with clathrin. However, coimmunoprecipitated clathrin heavy chain did not co-migrate with the 150-kDa protein (data not shown). Therefore, the 150-kDa phosphoprotein may represent a novel arrestin binding protein.
Next, we performed phosphoamino acid analysis of the radiolabeled arrestins. The arrestins were excised from the PVDF membrane, and the proteins were subjected to acid hydrolysis. A phosphoprotein that migrated to a position similar to arrestin-3 (from the control cells without arrestin overexpression) was also included in the analysis. As shown in Fig. 1b, arrestin-2 contained only phosphoserine as previously reported, whereas arrestin-3 had a major phosphothreonine band and a The transfected cells were plated on 35-mm glass-bottomed culture dishes 1 day before the experiment. Cells expressing equivalent amounts of arr3-GFP proteins were selected, and the distribution of the fusion proteins was visualized with a confocal microscope before (0) and after (100-, 200-, and 300-s) treatment with 10 M isoproterenol. All three forms of arr3-GFP fusion proteins translocated to the plasma membrane with similar kinetics. minor phosphoserine band. A phosphoserine band was also present in the control sample (without arrestin overexpression). From subsequent mutagenesis studies and phosphoamino acid analyses (see below), we conclude that arrestin-3 is mainly phosphorylated on threonine.
Mapping of Phosphorylation Sites in Arrestin-3-To identify the specific residues phosphorylated in arrestin-3, we first examined the phosphorylation of various truncated arrestin-3 proteins. Amino-or carboxyl-terminally truncated proteins were expressed in COS-1 cells, and their phosphorylation was compared with wild type arrestin-3 by metabolic labeling and immunoprecipitation. Amino-terminally truncated proteins, arr3-(201-409) and arr3-(284 -409), were phosphorylated similar to wild type arrestin-3, suggesting that the phosphorylation sites reside after residue 284 (Fig. 2a). Analysis of the carboxyl-terminally truncated proteins arr3-(1-360) and arr3-(1-370) suggested that the phosphorylation occurs from residue 370 to 409, since neither of these proteins were phosphorylated (Fig. 2b). All of the truncated proteins were expressed and immunoprecipitated comparably with wild type protein except for arr3-(1-360) that was immunoprecipitated less well (data not shown).
We next made several point mutations where threonine and serine residues in the carboxyl terminus of arrestin-3 were changed to alanine. As shown in Fig. 2c, mutation of Thr-382 to alanine (T382A) completely abolished arrestin-3 phosphorylation, whereas mutation of Ser-337 (S337A) or Ser-360 and Thr-365 (S360A/T365A) were still phosphorylated, albeit to a reduced level. Therefore, we believe that the constitutive phosphorylation of arrestin-3 occurs primarily at Thr-382. It is still possible, however, that there is additional phosphorylation on serine or threonine residues of arrestin-3 that might occur after the primary phosphorylation of Thr-382.
Arrestin-3 Is Phosphorylated by CKII-To identify the protein kinase responsible for arrestin-3 phosphorylation, we initially performed in vitro kinase assays using purified recombinant arrestin-3 and various purified protein kinases. Since Thr-382 is in an acid-rich region of arrestin-3 ( 382 TDDD; Fig.  2d), we tested whether arrestin-3 was a substrate for various acidotropic kinases. These included CKI and GRK2, both of which prefer substrates that contain acidic residues aminoterminal to the phosphorylation site (29,30), and CKII, which has a consensus sequence of (S/T)XX(D/E) (31). We also tested GRK5 since one would predict that GRKs and arrestins are in close proximity when associated with GPCRs. These studies revealed that arrestin-3 is effectively phosphorylated by CKII but is not a substrate for CKI, GRK2, or GRK5 (Fig. 3a). Additional analysis showed that GRK2 and GRK5 were also unable to phosphorylate arrestin-3 in the presence of phospholipids or light-activated rhodopsin, two GRK activators (data not shown) (32,33). CKII seems to phosphorylate a single residue, because arrestin-3 phosphorylation by CKII was saturable over time with a stoichiometry of ϳ0.5 mol of phosphate/ mol of arrestin-3 (Fig. 3b). In contrast, arrestin-2 was not significantly phosphorylated by CKII.
To test whether CKII phosphorylates Thr-382, we prepared lysates from COS-1 cells overexpressing wild type or T382A arrestin-3 and performed in vitro kinase assays in the presence or absence of purified CKII. Arrestins were then immunoprecipitated, and the phosphorylation was analyzed by SDS-PAGE and autoradiography. Wild type arrestin-3 was effectively phosphorylated by exogenous CKII, although there was some phosphorylation in the absence of additional kinase (possibly by endogenous CKII) (Fig. 3c). In contrast, the T382A mutant was not phosphorylated by either the endogenous kinases or by recombinant CKII. Phosphorylation of Thr-382 by CKII was also confirmed using purified glutathione S-transferase fusion proteins of arrestin-3 as substrates. While CKII efficiently phosphorylated GST-arr3-(330 -409), it failed to phosphorylate a T382A mutant of GST-arr3-(330 -409) (data not shown). Taken together, these results demonstrate that CKII phosphorylates Thr-382 and is probably the kinase responsible for arrestin-3 phosphorylation in intact cells.
Arrestin-3 Is Dephosphorylated upon ␤ 2 -Adrenergic Receptor Stimulation-To gain further insight into arrestin-3 phosphorylation, we examined the phosphorylation status of arrestin-3 during GPCR stimulation. COS-1 cells expressing arrestin-3 and ␤ 2 AR were metabolically labeled with [ 32 P]orthophosphate and treated with or without the ␤-agonist isoproterenol for 15 min, and the phosphorylation of arrestin-3 was determined by immunoprecipitation, SDS-PAGE, and autoradiography. Initial analysis revealed that arrestin-3 phosphorylation decreased ϳ35% after ␤-agonist treatment, suggesting that arrestin-3 is dephosphorylated upon activation of ␤ 2 AR (Fig. 4a,  upper panel).
Arrestins are cytosolic proteins that translocate to the plasma membrane upon GPCR activation and subsequent phosphorylation. Dependent on the receptor subtype, nonvisual arrestins can then traffic with the receptor initially to clathrin-coated pits and in some cases to endocytic vesicles (26). Because of the dynamics of arrestin localization, we next examined the phosphorylation status of arrestin-3 in various intracellular compartments. After metabolic labeling and stimulation of ␤ 2 AR, COS-1 cells were lysed and fractionated by differential centrifugation. Agonist-dependent dephosphorylation of arrestin-3 was significant in the plasma membrane fraction, whereas the extent of arrestin-3 phosphorylation in the vesicle membrane and cytosolic fractions remained largely unchanged (Fig. 4a). This suggests that dephosphorylation of arrestin-3 mainly occurs at the plasma membrane or that dephosphorylated arrestin-3 is preferentially recruited to the plasma membrane. It is noteworthy that significant arrestin-3 dephosphorylation also occurred in cells with endogenous ␤ 2 AR, suggesting that this process is highly efficient. Although the detailed mechanism of arrestin-3 dephosphorylation is unknown, it is interesting to note that the catalytic subunit of endogenous PP2A coimmunoprecipitates with expressed arrestin-3 in COS-1 cells (Fig. 4b). Previous studies implicated PP2A in ␤ 2 AR dephosphorylation, and it was present in a multicomponent complex containing ␤ 2 AR, protein kinase A, PP2B, and gravin (A kinase anchoring protein 250) (34,35).
Phosphorylation of Arrestin-3 Does Not Regulate Its Translocation to the Plasma Membrane after ␤ 2 AR Stimulation-To investigate the potential functional role of arrestin-3 phosphorylation, we generated the mutant constructs arr3(T382A) and arr3(T382V) to mimic the dephosphorylated form of arrestin-3, and we generated arr3(T382E) to mimic phosphorylated arrestin-3. To study arrestin localization, we generated fusion proteins in which enhanced green fluorescent protein was fused to the carboxyl terminus of arr3, arr3(T382V), and arr3(T382E). The various arr3-GFP fusion proteins were then expressed in HEK293 cells that had been stably transfected with ␤ 2 AR and visualized using time lapse confocal microscopy. Before receptor activation, wild type and mutant arrestins were found exclusively in the cytosol (Fig. 5, time 0), suggesting that phosphorylation does not regulate the basal intracellular localization of arrestin-3. We next studied the translocation of arrestin-3 after ␤ 2 AR stimulation, an approach proven useful for investigation of real time arrestin translocation and modeling of arrestin/receptor interaction (36). As shown in Fig. 5, ␤ 2 AR FIG. 7. Arrestin-3 interaction with clathrin and ␤-adaptin. a, binding of in vitro translated arrestins to clathrin cages. In vitro synthesized tritiated arrestin-3, arr3(T382A), or arr3(T382E) was incubated with (ϩ) or without (Ϫ) 200 nM clathrin cages for 10 min at ambient temperature. Clathrin-bound arrestins were pelleted by centrifugation, run on SDS-PAGE, and detected by fluorography. Radiolabeled arrestins were quantified using a densitometer. Std, 20% of arrestin-3 input. The autoradiogram (upper panel) is representative of three independent experiments, and the bar graph is the mean Ϯ S.E. from the three experiments. b, co-immunoprecipitation of arrestin and ␤-adaptin. FLAG-␤ 2 AR and arrestin-3, arr3(T382A), or arr3(T382E) were expressed in COS-1 cells, and at 48 h post-transfection, cells were treated with (ϩ) or without (Ϫ) 10 M isoproterenol for 2 min at 37°C. Arrestins were immunoprecipitated from cell lysates, and co-immunoprecipitated endogenous ␤-adaptin was detected with a monoclonal antibody as described under "Experimental Procedures." c, in vitro binding of arrestins to GST-␤ 2 -adaptin. Five micrograms of purified GST or a GST-␤ 2 -adaptin appendage domain (residues 700 -937) fusion protein bound to glutathione-agarose beads were incubated with 10 g of COS-1 cell lysate expressing wild type or mutant arrestin-3 for 1 h at 4°C. After two washes, bound arrestin was eluted from the beads, electrophoresed on a 10% SDS-polyacrylamide gel, transferred to nitro- cellulose, and detected with purified polyclonal antibody against arrestin-3. d, binding of in vitro translated arrestins to clathrin-coated vesicles. In vitro synthesized tritiated arrestin-3, arr3(T382A) or arr3(T382E) were incubated with (ϩ) or without (Ϫ) 50 nM clathrincoated vesicles for 10 min at ambient temperature. The samples were then processed and analyzed as described for a. stimulation results in rapid translocation of arr3, arr3(T382V), and arr3(T382E) from the cytosol to the plasma membrane. Quantification reveals that the initial rates of translocation were similar for arr3, arr3(T382V), and arr3(T382E) and suggests that the rate of receptor association is not regulated by arrestin-3 phosphorylation. These results are similar to arrestin-2, where the phosphorylation of Ser-412 by extracellular signal-regulated kinases 1 and 2 appears to have no effect on receptor interaction (36). Our results also suggest that the decrease in phosphorylated arrestin-3 that occurs upon ␤ 2 AR stimulation (Fig. 4) is probably due to dephosphorylation at the plasma membrane rather than the preferential translocation of the dephosphorylated protein to the plasma membrane.
Internalization and Recycling of ␤ 2 AR Is Not Regulated by Arrestin-3 Phosphorylation-Because ␤ 2 AR stimulation promotes arrestin-3 dephosphorylation, we hypothesized that the phosphorylation state of arrestin-3 might influence its ability to regulate ␤ 2 AR function. Nonvisual arrestins mediate internalization of many GPCRs by concomitant binding to receptors and components of the endocytic machinery including clathrin and ␤-adaptin (7,8). Therefore, we tested whether agonistpromoted ␤ 2 AR internalization was regulated by the phosphorylation state of arrestin-3. In COS-1 cells, the rate and extent of FLAG-␤ 2 AR internalization was significantly increased by overexpression of wild type arrestin-3 (Fig. 6a). Similar results were observed with arr3(T382A) and arr3(T382E) (Fig. 6a). Additional studies that varied the expression level of arr3, arr3(T382A), and arr3(T382E) also did not detect differences in their ability to promote ␤ 2 AR internalization (not shown). Similarly, arr3, arr3(T382A), arr3(T382V), and arr3(T382E) promoted internalization of ␤ 2 AR(Y326A), a mutant receptor partially impaired in agonist-promoted phosphorylation and internalization (37,38), to a similar extent (not shown). Moreover, arr3, arr3(T382A), and arr3(T382E) all promoted a similar level of FLAG-␤ 2 AR internalization as assessed by immunofluorescence microscopy (not shown). Taken together, these results suggest that the phosphorylation state of arrestin-3 does not significantly alter its ability to promote ␤ 2 AR internalization. These findings are in contrast to arrestin-2, where previous studies found that arr2(S412D), which mimics the phosphorylated form of arrestin-2, functioned as a dominant negative mutant and inhibited ␤ 2 AR internalization. This defect of arr2(S412D) has been ascribed to an inability to interact with clathrin (20).
Via its ability to promote ␤ 2 AR internalization, arrestin-3 also promotes subsequent resensitization and recycling of ␤ 2 AR (39). Thus, we next tested whether the phosphorylation state of arrestin-3 influences ␤ 2 AR recycling. COS-1 cells expressing arr3, arr3(T382A), or arr3(T382E) along with FLAG-␤ 2 AR were stimulated with isoproterenol for 30 min, washed extensively with media, and treated with the ␤-antagonist alprenolol for 90 min to allow internalized receptors to recycle back to the plasma membrane. The profiles of FLAG-␤ 2 AR internalization and recycling in the presence of the various forms of arrestin-3 were indistinguishable (Fig. 6b). Therefore, internalization and recycling of the ␤ 2 AR do not appear to be regulated by arrestin-3 phosphorylation.
Phosphorylation of Arrestin-3 Does Not Regulate Interaction with Clathrin or ␤-Adaptin-Although we did not see a defect in the ability of arr3(T382E) to promote ␤ 2 AR internalization, we were interested in determining whether arrestin-3 phosphorylation directly affects clathrin interaction, since Thr-382 is close to the primary clathrin binding site in arrestin-3 (Fig. 2d). This was tested using an established in vitro clathrin binding assay (7). Radiolabeled arrestins were prepared by in vitro translation and incubated with purified clathrin cages, and clathrin-bound arrestins were pelleted by centrifugation, resolved by SDS-PAGE, and detected by fluorography. As shown in Fig. 7a, arr3, arr3(T382A), and arr3(T382E) bound to clathrin cages to a similar extent. We also analyzed the ability of arr3, arr3(T382A), and arr3(T382E) expressed in COS-1 cells to bind to purified GST-clathrin terminal domain and found a comparable level of binding (data not shown). Taken together, these results suggest that phosphorylation of arrestin-3 does not directly influence its interaction with clathrin.
Since nonvisual arrestins also interact with the ␤-adaptin subunit of the adaptor protein AP2 (8), we next investigated whether phosphorylation of arrestin-3 affects its interaction with ␤-adaptin. This was assessed by co-expressing arrestins and ␤ 2 ARs in COS-1 cells, treating the cells with or without ␤-agonist for 2 min, and analyzing arrestin/␤-adaptin interaction by immunoprecipitating the arrestin and blotting for endogenous ␤-adaptin. As shown in Fig. 7b, arrestin-3 binding to ␤-adaptin is detected in unstimulated cells but is increased ϳ2-fold by receptor stimulation. Arr3(T382A) and arr3(T382E) bind ␤-adaptin comparably with wild type arrestin-3, suggesting that the interaction of these proteins is not regulated by arrestin-3 phosphorylation. Similarly, we found that arr3, arr3(T382A), and arr3(T382E) expressed in COS-1 cells bound comparably with purified GST-␤ 2 -adaptin appendage domain (Fig. 7c). To confirm these results, we also analyzed interaction of arrestin-3 with CVs. CVs contain clathrin and AP2 and thus represent a more physiological model for clathrin-coated pits. Similar to the clathrin cage binding assay, arr3, arr3(T382A), and arr3(T382E) bound equally well to CVs (Fig. 7d). Overall, these results suggest that arrestin-3 phosphorylation at Thr-382 does not directly regulate interaction with either clathrin or ␤ 2 -adaptin.
Interaction between Arrestin-3 and Src Is Independent of Arrestin-3 Phosphorylation-Recently, a role for arrestins as positive transducers of GPCR signaling has been appreciated (40). Arrestins can interact with Src family members (11,14,17) and function as scaffolds in GPCR-mediated activation of mitogen-activated protein kinases (12,13,15,16). Interestingly, the interaction between Src and arrestin-2 was predicted to be regulated by the phosphorylation state of arrestin, since arr2(S412D) did not bind Src (11). Accordingly, we tested whether the phosphorylation state of arrestin-3 would regulate Src interaction. Arrestins were co-expressed with Src and ␤ 2 ARs in COS-1 cells, ␤ 2 ARs were stimulated with agonist for 5 min to promote arrestin/Src interaction, and the interaction was then assessed by immunoprecipitating the arrestin and blotting for Src. Src was found to coprecipitate with arr3, arr3(T382A), arr3(T382V), and arr3(T382E) to a comparable extent, suggesting that the phosphorylation of arrestin-3 does not affect Src interaction (Fig. 8).
Cellular Degradation of Arrestin-3 Is Not Regulated by Its Phosphorylation-Recently, arrestin-3 has been reported to be ubiquitinated by the E3 ubiquitin ligase Mdm2 (41). Because ubiquitin-mediated degradation is often regulated by the phosphorylation state of the target protein (42,43), we next evaluated the half-life of the various arrestin-3 mutants by pulsechase labeling. Arrestins were expressed in COS-1 cells, labeled with [ 35 S]methionine and [ 35 S]cysteine for 2 h, and then immunoprecipitated following a chase with nonradioactive media. The half-lives of arrestin-3, arr3(T382A), and arr3(T382E) were all ϳ13.5 h (Fig. 9). Therefore, degradation of arrestin-3 is not regulated by phosphorylation at Thr-382.
The Phosphorylation of Arrestin-3 Modulates Formation of a Protein Complex-Since our studies suggest that arrestin-3 phosphorylation does not regulate its interaction with several known binding partners including clathrin, ␤-adaptin, and Src, we decided to take a more general approach to identify a target interaction. This involved expressing arr3, arr3(T382V), and arr3(T382E) in HEK293 cells and immunoprecipitating the arrestin using a purified anti-arrestin-3 polyclonal antibody bound to protein A-conjugated agarose beads. The beads were washed and incubated in SDS sample buffer for 2 h at 27°C, and the bound proteins were electrophoresed on a 10% SDSpolyacrylamide gel, transferred to nitrocellulose, and stained with Coomassie Blue. As seen in the upper panel of Fig. 10, protein staining detected the ϳ50-kDa arrestin-3 proteins as well as the ϳ100-kDa IgG heavy chain dimers. In addition, we noted some protein staining near the top of the gel (marked with a question mark) that was consistently most prominent in the arr3(T382V) lane, less in the wild type arrestin-3 lane, and least prominent in the arr3(T382E) lane. We hypothesize that this band represents a protein or proteins that selectively interact with dephosphorylated arrestin-3. Using a lower percentage gel, the size of the band(s) was estimated to be Ͼ300 kDa.
Interestingly, the high molecular weight band(s) were also detected by immunoblotting with an anti-arrestin monoclonal antibody (Fig. 10, lower panel). Similar to the protein staining, arrestin-3 was most prominent in the arr3(T382V) lane and was not evident in the arr3(T382E) sample. Therefore, we believe that the high molecular weight band represents a multiprotein complex consisting of dephosphorylated arrestin-3 and unidentified arrestin-3 binding protein(s). The reason why the individual protein components of the complex were not dissembled with treatment of SDS and ␤-mercaptoethanol in the SDS-sample buffer is not clear. However, there are examples where protein complexes such as oligomeric GPCRs are not resolved by denaturing SDS-PAGE (44 -46).
In summary, we found that arrestin-3 is constitutively phosphorylated at Thr-382 by casein kinase II in COS-1 cells and becomes dephosphorylated after stimulation of ␤ 2 AR. The phosphorylation state of arrestin-3 does not appear to regulate interaction with ␤ 2 AR, clathrin, ␤-adaptin, or Src. Rather, it seems to regulate interaction with an unidentified arrestin-3binding protein. Differences between the regulatory role of arrestin-2 and -3 phosphorylation may contribute to the different cellular functions of these highly homologous proteins in GPCR signaling and regulation.