Stable interaction between beta-arrestin 2 and angiotensin type 1A receptor is required for beta-arrestin 2-mediated activation of extracellular signal-regulated kinases 1 and 2.

Binding of beta-arrestins to seven-membrane-spanning receptors (7MSRs) not only leads to receptor desensitization and endocytosis but also elicits additional signaling processes. We recently proposed that stimulation of the angiotensin type 1A (AT(1A)) receptor results in independent beta-arrestin 2- and G protein-mediated extracellular signal-regulated kinases 1 and 2 (ERK1/2) activation. Here we utilize two AT(1A) mutant receptors to study these independent pathways, one truncated at residue 324, thus removing all potential carboxyl-terminal phosphorylation sites, and the other bearing four mutations in the serine/threonine-rich clusters in the carboxyl terminus. As assessed by confocal microscopy, the two mutant receptors interacted with beta-arrestin 2-green fluorescent protein with much lower affinity than did the wild-type receptor. In addition, the mutant receptors more robustly stimulated G protein-mediated inositol phosphate production. Approximately one-half of the wild-type AT(1A) receptor-stimulated ERK1/2 activation was via a beta-arrestin 2-dependent pathway (suppressed by beta-arrestin 2 small interfering RNA), whereas the rest was mediated by a G protein-dependent pathway (suppressed by protein kinase C inhibitor). ERK1/2 activation by the mutant receptors was insensitive to beta-arrestin 2 small interfering RNA but was reduced more than 80% by a protein kinase C inhibitor. The biochemical consequences of ERK activation by the G protein and beta-arrestin 2-dependent pathways were also distinct. G-protein-mediated ERK activation enhanced the transcription of early growth response 1, whereas beta-arrestin 2-dependent ERK activation did not. In addition, stimulation of the truncated AT(1A) mutant receptor caused significantly greater early growth response 1 transcription than did the wild-type receptor. These findings demonstrate how the ability of receptors to interact with beta-arrestins determines both the mechanism of ERK activation as well as the physiological consequences of this activation.

Binding of ␤-arrestins to seven-membrane-spanning receptors (7MSRs) not only leads to receptor desensitization and endocytosis but also elicits additional signaling processes. We recently proposed that stimulation of the angiotensin type 1A (AT 1A ) receptor results in independent ␤-arrestin 2-and G protein-mediated extracellular signal-regulated kinases 1 and 2 (ERK1/2) activation. Here we utilize two AT 1A mutant receptors to study these independent pathways, one truncated at residue 324, thus removing all potential carboxyl-terminal phosphorylation sites, and the other bearing four mutations in the serine/threonine-rich clusters in the carboxyl terminus. As assessed by confocal microscopy, the two mutant receptors interacted with ␤-arrestin 2-green fluorescent protein with much lower affinity than did the wild-type receptor. In addition, the mutant receptors more robustly stimulated G protein-mediated inositol phosphate production. Approximately one-half of the wild-type AT 1A receptor-stimulated ERK1/2 activation was via a ␤-arrestin 2-dependent pathway (suppressed by ␤-arrestin 2 small interfering RNA), whereas the rest was mediated by a G protein-dependent pathway (suppressed by protein kinase C inhibitor). ERK1/2 activation by the mutant receptors was insensitive to ␤-arrestin 2 small interfering RNA but was reduced more than 80% by a protein kinase C inhibitor. The biochemical consequences of ERK activation by the G protein and ␤-arrestin 2-dependent pathways were also distinct. Gprotein-mediated ERK activation enhanced the transcription of early growth response 1, whereas ␤-arrestin 2-dependent ERK activation did not. In addition, stimulation of the truncated AT 1A mutant receptor caused significantly greater early growth response 1 transcription than did the wild-type receptor. These findings demonstrate how the ability of receptors to interact with ␤-arrestins determines both the mechanism of ERK activation as well as the physiological consequences of this activation.

Stimulation
of seven-membrane-spanning receptors (7MSRs) 1 leads to G protein coupling as well as receptor phos-phorylation by G protein-coupled receptor kinases (GRKs) and the recruitment of arrestins. Phosphorylation of 7MSRs has been shown to occur primarily in the carboxyl-terminal tail (C-tail) and to be critical for stable and high affinity arrestin binding (1)(2)(3). Binding of arrestins not only mediates 7MSR desensitization by physically preventing the interaction between G protein and receptor but also initiates receptor internalization. Some receptors, such as the AT 1A receptor and V2 vasopressin receptor, bind ␤-arrestins tightly and internalize with them into endosomal vesicles ("class B") (4). Others, such as the ␤2-adrenergic receptor, bind ␤-arrestins with relatively low affinity, dissociate from them in coated pits, and internalize without ␤-arrestins ("class A") (4). These patterns are easily distinguishable by confocal microscopy.
Upon agonist stimulation, the AT 1A receptor is phosphorylated on its carboxyl terminus by GRKs and protein kinase C (PKC) (5,6). The primary PKC phosphorylation sites have been mapped to Ser 331 , Ser 338 , and Ser 348 (5), which do not appear to be involved in ␤-arrestin binding (3). In addition, there are also three distinct serine/threonine-rich regions in the COOH terminus. The most carboxyl-terminal of these regions (Ser 346 , Ser 347 , Ser 348 ) appears to play a minimal role in ␤-arrestin binding (2, 3), whereas mutations in either the most aminoterminal region (Ser 328 , Ser 329 , Ser 331 , Thr 332 ) or the middle region (Ser 335 , Thr 336 , Ser 338 ) change the pattern of ␤-arrestin 2-GFP recruitment from class B to class A, suggesting that these two regions are important for ␤-arrestin 2 binding (2). Consistent with this, alanine substitutions for Thr 332 , Ser 335 , Thr 336 , and Ser 338 of the AT 1A receptor result in a marked decrease in agonist-induced ␤-arrestin 1 association (3). Furthermore, an AT 1A mutant receptor truncated at residue 325, which lacks all of the serine and threonine residues of the carboxyl terminus and shows no agonist-dependent phosphorylation (5), lacks the ability to interact with ␤-arrestin 1 in an agonist dependent manner as assessed by co-immunoprecipitation (3).
Accumulating evidence strongly suggests that ␤-arrestins also scaffold signaling pathways such as those leading to mitogen-activated protein kinase activation. In particular, ␤-arrestins have been shown to form complexes with the AT 1A , neurokinin receptor 1, and V2 vasopressin receptors, which scaffold and facilitate the activation of the ERK kinase cascade (RAF, MEK, ERK) while targeting the activated ERK to endo-cytic vesicles in the cytoplasm (7)(8)(9). We recently showed that stimulation of a mutant AT 1A receptor (DRY/AAY) with angiotensin II (Ang II) or a wild-type receptor with an Ang II analog ([Sar 1 ,Ile 4 ,Ile 8 ]Ang II) fails to activate classical heterotrimeric G protein signaling but does lead to ␤-arrestin 2 recruitment and ERK1/2 activation (10). In addition, depletion of cellular ␤-arrestin 2 via siRNA completely abolishes such G proteinindependent activation of ERK1/2. Thus the DRY/AAY mutant AT 1A receptor and [Sar 1 ,Ile 4 ,Ile 8 ]Ang II mutant ligand appear to be selective for the ␤-arrestin 2-dependent G protein-independent pathway leading to ERK1/2 activation. Here we set out to develop corresponding AT 1A receptor mutants capable of activating only G protein and independent of ␤-arrestin 2-mediated signaling; this was accomplished by mutating the carboxyl-terminal GRK phosphorylation sites in the receptor that are thought to be necessary for high affinity ␤-arrestin binding. This approach also permitted us to explore the correlation between ␤-arrestin/receptor interactions and ␤-arrestin-mediated ERK1/2 signaling.

EXPERIMENTAL PROCEDURES
Materials-The radiolabeled compounds, 125 I-Tyr-4-Ang II and myo-[ 3 H]inositol, were obtained from PerkinElmer Life Sciences. Human Ang II was purchased from Peninsula Laboratories, Inc.
[Sar 1 ,Ile 4 ,Ile 8 ]Ang II was synthesized in the Cleveland Clinic core synthesis facility (Cleveland, OH). Ro-31-8425 was purchased from Calbiochem. Chemically synthesized double-stranded siRNAs corresponding to human ␤-arrestin 2 and a non-silencing control were described previously (11). GeneSilencer transfection reagents were from Gene Therapy Systems (San Diego, CA). All other reagents were purchased from Sigma. The pcDNA3.1 expression plasmid encoding hemagglutinin (HA) epitope-tagged AT 1A receptors was provided by M. G. Caron (Duke University). ␤-Arrestin 2-GFP was provided by Sudha Shenoy.
Construction of AT 1A Receptor Mutants-Mutant AT 1A receptors were generated by mutagenesis PCR using pcDNA3.1-HA-AT 1A receptor as a template and a QuikChange multisite-directed mutagenesis kit (Stratagene) according to the manufacturer's instructions. The primers for TSTS/A and 324⌬ mutant receptors were 5Ј p-CTCAAGC-CTGTCTGCGAAAATGGCCGCTTGCTTACCGGCCTTC 3Ј and 5Ј p-C-CCCCAAAGGCCTAGTCCCACTCAAGCC 3Ј, respectively. Mutations were confirmed by DNA sequencing.
Cell Culture and DNA Transfection-HEK-293 cells were grown in Eagle's minimal essential medium with Earle's salts supplemented with 10% (v/v) fetal bovine serum and a 1% penicillin/streptomycin mixture (Sigma). Cells were transiently transfected using FuGENE 6 reagent (Roche Applied Science) according to the manufacturer's instructions.
Inositol Phosphate Determination-Transiently transfected HEK-293 cells in 10-cm dishes were plated onto poly-D-lysine-coated 12-well plates (BD Biosciences). To assay for inositol phosphate (IP) production, cells were incubated overnight at 37°C in labeling medium (1 Ci of myo-[ 3 H]inositol in 0.5 ml of Eagle's minimal essential medium with 10% fetal bovine serum/well). Cells were washed with 20 mM HEPES containing 20 mM LiCl for 20 min at 37°C and then treated with different concentrations of Ang II for 20 min. Total inositol phosphates were extracted and separated as described previously (12).
siRNA Transfection-HEK-293 cells were simultaneously transfected in 10-cm dishes with 1 g of pcDNA3.1-HA-AT 1A receptor or 3 g of TSTS/A or 324⌬ mutant AT 1A receptor plasmids and 20 g of ␤-arrestin 2 siRNA or control siRNA as described previously (13). Similar receptor expression levels (150 -200 fmol of receptors/mg of protein) were obtained under these experimental conditions as measured by radioligand binding. Forty-eight hours after transfection, cells were divided into 6-well plates, and cellular extracts were prepared three days after transfection.
Confocal Microscopy-HEK-293 cells were transiently transfected in 10-cm dishes with 1 g of pcDNA3.1-HA-AT 1A receptor or 3 g of TSTS/A or 324⌬ mutant AT 1A receptor plasmids and 0.3 g of the ␤-arrestin 2-GFP plasmid. One day after transfection, cells were split onto collagen-coated 35-mm plastic dishes with glass bottoms (MatTek, Ashland, MA) and cultured overnight at 37°C. Confocal microscopy was performed at ϫ100 magnification with a Zeiss laser-scanning microscope (LSM-510). Images were collected using a 488-nm excitation and 515-540-nm emission filter.
Quantitative Real Time Reverse Transcription-PCR-Total RNA was purified from cells using an RNeasy mini kit (Qiagen) according to the manufacturer's instructions. Two micrograms of total RNA was then used to generate cDNA using an Omniscript RT kit (Qiagen). For each PCR reaction, cDNAs corresponding to 300 ng of total RNA were used. Real time PCR was performed using a SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer's instructions on a Mx3000P™ real time PCR system (Stratagene). PCR reactions were carried out by incubating reactions at 95°C for 10 min to inactive the reverse transcriptases and activate the DNA polymerase. Forty-five standard PCR cycles were carried out at 95°C for 30 s for denaturing, 58°C for 30 s for annealing, and 72°C for 20 s for extension. Primers for human EGR-1 were CAGCACCTTCAACCCTCAG (sense) and CA-CAAGGTGTTGCCACTGTT (antisense). Additionally, real time reverse transcription-PCR for human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was conducted as a reference for normalization. Primers for GAPDH were ACCACAGTCCATGCCATCAC (sense) and TCCAC-CACCCTGTTGCTGTA (antisense) In each experiment, reverse transcription-PCR reactions for both EGR-1 and GAPDH were performed in triplicate for each RNA sample. The average threshold cycle values for EGR-1 were normalized to the threshold cycle value of GAPDH and converted to a linear scale.

RESULTS
To generate mutant AT 1A receptors that are defective in their ability to associate with ␤-arrestin 2, we mutated the carboxyl-terminal sites in the receptor thought to be necessary for high affinity ␤-arrestin binding. The two mutant AT 1A receptors created were TSTS/A, which has alanine substitutions for Thr 332 , Ser 335 , Thr 336 , and Ser 338 , and 324⌬, which truncates the AT 1A receptor from residue 325 and lacks all of the serine and threonine residues at the carboxyl terminus (Fig. 1A).
To test whether TSTS/A and 324⌬ mutant receptors were defective in their ability to associate with ␤-arrestin 2, we examined the ability of wild-type, TSTS/A, and 324⌬ mutant AT 1A receptors to interact with ␤-arrestin 2-GFP upon agonist stimulation using confocal microscopy. Ang II induced translocation of ␤-arrestin 2-GFP to endocytic vesicles (class B pattern) in cells expressing wild-type AT 1A receptors (Fig. 1B), demonstrating a stable interaction between ␤-arrestin 2-GFP and internalized AT 1A receptors (2). Conversely, in cells expressing TSTS/A mutant receptor, Ang II induced translocation of ␤-arrestin 2-GFP only to the plasma membrane (class A pattern) (Fig. 1B). This is consistent with the previous report that the mutated serine/threonine sites are important for the high affinity binding of ␤-arrestin (2, 3). Ang II also induced translocation of ␤-arrestin 2-GFP to the plasma membrane in cells expressing the 324⌬ mutant receptor but in a pattern that seemed weaker than the classic class A pattern of ␤-arrestin 2 recruitment (Fig. 1B) (4).
The carboxyl terminus of AT 1A receptors is not thought to be critical for G protein coupling (14). Consistent with this, both TSTS/A and 324⌬ mutant receptors still coupled efficiently to G proteins as indicated by inositol phosphate production turnover assay (Fig. 2). In fact, both mutant receptors mediated somewhat greater constitutive and agonist-stimulated IP production than the wild-type receptor. Both the affinity and maximum activity of the mutant receptors were also increased (Fig.  2). This presumably reflects impaired ␤-arrestin-mediated desensitization. Thus, both the TSTS/A and 324⌬ receptors are defective in their ability to associate with ␤-arrestin 2 while retaining the ability to interact with G protein.
We have previously proposed that the AT 1A receptor can stimulate ERK1/2 activation by two pathways mediated, respectively, by G protein and ␤-arrestin 2 (10). Because both TSTS/A and 324⌬ can still couple to G protein, activation of G protein-mediated ERK1/2 should not be compromised for either of these two mutant receptors. Consistent with this idea, stimulation of wild-type AT 1A receptor, TSTS/A, and 324⌬ mutant receptors with Ang II all led to robust ERK1/2 activation (Fig.  3, A and B), which, if anything, was increased with the mutant receptors. To determine the role of ␤-arrestin 2 in the activation of ERK1/2 by these mutant receptors, we compared the effect of ␤-arrestin 2 siRNA on the ERK1/2 activation induced by the wild-type AT 1A receptor, TSTS/A, or 324⌬ mutant receptors. Down-regulation of ␤-arrestin 2 significantly decreased the ERK1/2 activation induced by wild-type AT 1A receptor by ϳ50 -60% (Fig. 3, A and C) as reported previously (10). This suggests that about one-half of the ERK1/2 activation by the wild-type AT 1A receptor is mediated by ␤-arrestin 2. In contrast, no significant inhibition in the ERK1/2 activation was observed when either of the mutant receptors was stimulated with Ang II in ␤-arrestin 2 siRNA-transfected cells (Fig. 3, A  and C). These results indicate that ERK1/2 activated by TSTS/A or 324⌬ mutant receptors is largely independent of ␤-arrestin 2 and mediated primarily by G proteins. In addition, these results demonstrate a good correlation between the ability of ␤-arrestin 2 to interact with the receptors (as observed by confocal microscopy) and the ability of the receptors to mediate ␤-arrestin 2-dependent ERK1/2 activation.
We have shown previously that the G protein-mediated ERK1/2 activation by the wild-type AT 1A receptor in HEK-293 cells requires activation of PKC, whereas the ␤-arrestin 2-dependent ERK1/2 activation does not (10). If the ERK1/2 activation induced by TSTS/A or 324⌬ mutant receptor is mostly dependent on the G protein-mediated pathway, then ERK1/2 activation induced by these mutant receptors should be more sensitive to PKC inhibition than activation induced by the wild-type AT 1A receptor. As predicted, pretreatment with the PKC inhibitor Ro-31-8425 inhibited the wild-type AT 1A receptor-stimulated ERK1/2 activation by ϳ55%, whereas it inhibited TSTS/A or 324⌬ mutant receptor-stimulated ERK1/2 activation by Ͼ80% (Fig. 4, A and B). A combination of ␤-arrestin 2 siRNA and Ro-31-8425 almost completely abolished the activation of ERK1/2 induced by the wild-type AT 1A receptor as well as TSTS/A and 324⌬ mutant receptors (Fig. 4, C and D).
Ang II-induced EGR-1 transcription is mediated by ERK1/2 in vascular smooth muscle cells and Chinese hamster ovary cells (15,16). To determine whether Ang II-induced transcription of EGR-1 is ERK1/2-dependent in HEK-293 cells, we tested the effect of U0126, a specific inhibitor of MEK, just upstream of ERK1/2, on Ang II-induced EGR-1 transcription. Pretreatment with U0126 completely abolished Ang II-induced ERK1/2 activation by the wild-type AT 1A receptor (data not shown) and inhibited Ͼ97.3 Ϯ 12.3% (n ϭ 3) of Ang II-induced EGR-1 transcription, indicating that Ang II-induced transcription of EGR-1 is dependent on ERK1/2 activation in HEK-293 cells. Because mitogen-induced gene expression requires nu- It has been demonstrated previously that ␤-arrestins facilitate 7MSR-mediated ERK activation but retain activated phospho-ERK1/2 in the cytosol, which is incapable of inducing ERKdependent transcription (7,8,18,19). Thus, we would expect that G protein-dependent phospho-ERK1/2 translocates to the nucleus, leading to ERK-dependent transcription, whereas the ␤-arrestin 2-dependent phospho-ERK1/2 remains in the cytosol and cannot activate ERK-dependent transcription. To determine whether the induction of EGR-1 transcription by Ang II is mediated by the G protein-dependent but not by ␤-arrestin 2-dependent phospho-ERK1/2, we tested the effect of the PKC inhibitor Ro-31-8425 on the induction of EGR-1 transcription by Ang II. Stimulation of wild-type AT 1A receptors with Ang II for 1 h led to an almost 22-fold increase in EGR-1 transcripts (Fig. 5). Pretreatment with the PKC inhibitor all but abolished this induction, suggesting that stimulation of EGR-1 transcription by Ang II is mediated by the G protein-dependent phospho-ERK. To further confirm that the ␤-arrestin 2-dependent phospho-ERK1/2 is unable to enhance transcription of EGR-1, we tested the ability of [Sar 1 ,Ile 4 ,Ile 8 ]Ang II, which we have previously shown to activate ERK1/2 only via the ␤-arrestin 2-dependent pathway and not by the G protein-dependent pathway (10), to induce EGR-1 transcription. Stimulation of wild-type AT 1A receptors with [Sar 1 ,Ile 4 ,Ile 8 ]Ang II did not lead to a significant increase in EGR-1 mRNA levels (Fig. 5A).
The 324⌬ and TSTS/A mutant AT 1A receptors are defective in their ability to tightly bind ␤-arrestins (Fig. 1B) and are more capable of stimulating G protein signaling than the wildtype receptor (Fig. 2). Thus, they might be expected to induce EGR-1 transcription even more strongly than the wild-type AT 1A receptor. Consistent with our IP production data, a significant increase in the basal EGR-1 transcripts was observed in cells expressing TSTS/A or 324⌬ mutant AT 1A receptor (Fig.  5C). In addition, stimulation of the 324⌬ mutant AT 1A receptor with Ang II for 1 h led to a significant increase in the amount of EGR-1 transcripts as compared with the amount of EGR-1 transcripts induced by stimulation of the wild-type AT 1A receptor (Fig. 5C), whereas TSTS/A induced an intermediate level of EGR-1 transcripts. DISCUSSION We have previously proposed that the AT 1A receptor can induce ERK1/2 activation through two independent pathways mediated by G protein and ␤-arrestin 2, respectively (10). Here we provide support for this idea by showing that the nature of the interaction between ␤-arrestin 2 and the AT 1A receptor is determinative for ␤-arrestin 2-mediated ERK1/2 activation. Using EGR-1 mRNA as an indicator for nuclear ERK1/2 activation, we further show that G protein-mediated ERK1/2 activation induces EGR-1 transcription, whereas ␤-arrestin 2-mediated ERK1/2 activation does not. Not only is this consistent with previous reports that ␤-arrestins sequester phospho-ERK1/2 in the cytosol (7,18,19), but it is also indicative of a clear functional divergence in the consequence of ERK1/2 activation by the two pathways. Although several cytosolic proteins such as GRK2, Bcl2, and tau are regulated by ERK1/2 phosphorylation (20 -22), the physiological substrates of the cytosolic phospho-ERK1/2 activated by ␤-arrestin 2 remain unknown. Nevertheless, it is clear that the ␤-arrestin 2-and G protein-mediated signaling pathways have different physiological consequences.
The 324⌬ mutant receptor lacks all the phosphorylation sites in its C-tail and cannot be phosphorylated in response to Ang II stimulation (5). However, our confocal data clearly indicate that this mutant receptor can recruit ␤-arrestin 2 to the plasma membrane in an agonist-dependent manner. These results show that ␤-arrestin 2 can bind to AT 1A receptors in a phosphorylation-independent manner, albeit with lower affinity. A, HEK-293 cells were transfected with expression vectors encoding wild-type (WT), TSTS/A, or 324⌬ mutant AT 1A receptors. Cells were pretreated with or without Ro-31-8425 for 20 min followed by 5 min of stimulation by Ang II. The activation of ERK1/2 was determined by immunoblotting with a phospho-ERK1/2-specific antibody. B, the effects of Ro-31-8425 on Ang II-induced ERK1/2 activation were compared by normalizing each phospho-ERK1/2 signal to the response induced by Ang II in non-inhibitor-treated cells (n ϭ 4). **, p Ͻ 0.01, compared with the wild-type AT 1A receptor-stimulated ERK1/2 activation in the presence of Ro-31-8425. C, HEK-293 cells were transfected with control (CTL) siRNA or ␤-arrestin 2 siRNA (␤-arr2) and expression vectors encoding wild-type, TSTS/A, or 324⌬ mutant AT 1A receptors. Three days after transfection, cells were pretreated with or without Ro-31-8425 as indicated for 20 min followed by stimulation by 100 nM Ang II for 5 min. The amount of p-ERK1/2, total ERK1/2, and ␤-arrestin 2 were determined by immunoblotting. D, the effects of ␤-arrestin 2 siRNA and Ro-31-8425 on Ang II-induced ERK1/2 activation by wild-type, TSTS/A, or 324⌬ mutant AT 1A receptors were compared by normalizing each phospho-ERK1/2 signal to the response induced by Ang II in non-inhibitor-treated cells (n ϭ 4).
However, in the case of AT 1A receptors, this phosphorylationindependent binding of ␤-arrestin 2 does not apparently trigger ERK1/2 activation. Although phosphorylation of the C-tail ap-pears to be required for high affinity binding of ␤-arrestins (2, 3), several receptors have been reported to interact with ␤-arrestins through the third intracellular loop. For example, both the third intracellular loop and the C-tail are important for ␤-arrestin binding to the neurokinin receptor 1 (23), delta opioid receptor (24,25), and lutropin/choriogonadotropin receptor (26). In each case, Ser/Thr residues in the third intracellular loop also appear to be critical for ␤-arrestin binding. However, we found that an AT 1A mutant receptor with alanine substitutions for all three Ser/Thr residues in the third intracellular loop showed no difference in recruiting ␤-arrestin 2-GFP compared with the wild-type AT 1A receptor (data not shown). In addition, an AT 1A mutant receptor-combining truncation of the C-tail and alanine substitutions for all three Ser/Thr residues in the third intracellular loop can still recruit ␤-arrestin 2-GFP to the plasma membrane in an agonist-dependent manner (data not shown). Together with previous reports that a C-tail truncation mutant of AT 1A receptor at residue 325 cannot be phosphorylated upon Ang II stimulation (3), these results suggest that the phosphorylation-independent association of ␤-arrestin 2 with the AT 1A receptor is driven by conformational changes in the receptor, which occur upon agonist binding. Such a conformational change-dependent recruitment of ␤-arrestins is not unique. For example, the interaction between ␤-arrestin 2 and the lutropin/choriogonadotropin receptor depends mostly on receptor activation rather than on receptor phosphorylation (27), and an Asp residue in the third intracellular loop is important for ␤-arrestin 2 binding to this receptor (28). At first glance, the pattern of agonistdependent recruitment of ␤-arrestin 2-GFP to the plasma membrane by the 324⌬ mutant receptor is similar to the pattern of ␤-arrestin 2-GFP recruitment by class A receptors (4). However, the binding of ␤-arrestin 2 to 324⌬ mutant receptors and the binding of ␤-arrestin 2 to a class A receptor might be fundamentally different because ␤-arrestin 2 recruitment by a class A receptor, such as the ␤2-adrenergic receptor, is primarily dependent on receptor phosphorylation (29).
The inability of TSTS/A and 324⌬ mutant receptors to stably interact with ␤-arrestins resulted in dramatic inhibition of receptor internalization and increased IP production (3). As a consequence, G protein-mediated ERK1/2 activation was augmented, further reducing the role of the ␤-arrestin 2-dependent pathway. This may explain why the proportion of ␤-arrestin 2-dependent ERK1/2 activation induced by stimulation of the TSTS/A or 324⌬ mutant receptors is so low, making it hard to detect by the ␤-arrestin 2 siRNA method. Our data further suggest that the low affinity association of ␤-arrestin 2 with the AT 1A receptor, which is not dependent on receptor phosphorylation, is less able to initiate ␤-arrestin 2-dependent signaling and that a high affinity state of ␤-arrestin 2 binding, which requires receptor phosphorylation, is necessary for ␤-arrestin 2-dependent signaling.