β-Arrestin 1 and Gαq/11 Coordinately Activate RhoA and Stress Fiber Formation following Receptor Stimulation*

β-Arrestins were initially shown, in conjunction with G protein-coupled receptor kinases, to be involved in the desensitization and internalization of activated seven-transmembrane receptors. Recently, β-arrestin 2 has been shown to act as a signal mediator in mitogen-activated protein kinase cascades and to play a positive regulatory role in chemotaxis. We now show that β-arrestin 1 is required to activate the small GTPase RhoA leading to the re-organization of stress fibers following the activation of the angiotensin II type 1A receptor. This angiotensin II type 1A receptor-directed RhoA activation and stress fiber formation also require the activation of the heterotrimeric G protein Gαq/11. Whereas neither β-arrestin 1 nor Gαq/11 activation alone is sufficient to robustly activate RhoA, the concurrent recruitment of β-arrestin 1 and activation of Gαq/11 leads to full activation of RhoA and to the subsequent formation of stress fibers.

Stress fiber formation and the polymerization of actin structures are fundamental processes required for cell motility, adhesion, and contraction (1)(2)(3). The Rho family of small GTPases (Rho, Rac, and Cdc42) is involved in such cytoskeletal rearrangement (4,5). RhoA in particular is important for stress fiber formation, whereas Rac and Cdc42 lead to lamellipodia and filopodia formation, respectively. Rho is activated by cytokine receptors, the epidermal growth factor receptor and several 7TMs (6,7) including AT 1 R, the angiotensin type II 1 receptor (8,9). Physiologically, the angiotensin type II 1 receptor plays important roles in the vasculature and renal systems and is implicated in several disease states including hypertension, heart failure, vascular thickening, and cardiac hypertrophy and remodeling (10,11). Stress fibers in cardiac myocytes, characteristic of cardiac hypertrophy, can be induced by angiotensin II (Ang II) 1 -activated AT 1A R (12). Furthermore, Rho and the downstream effector of Rho, Rho-associated kinase (ROCK), have been implicated in Ang II-induced effects in vivo (13,14) and in vitro (15)(16)(17)(18)(19).
7TM-mediated Rho activation has been shown to occur primarily through the heterotrimeric G proteins G ␣q/11 , G ␣12 , and G ␣13 (6). The AT 1A R signals primarily through the heterotrimeric proteins G ␣q/11 to downstream effectors such as phospholipase C and protein kinase C (10). Some controversy exists as to the role of G ␣q/11 in Rho activation, as some reports clearly show that G ␣q/11 is responsible for Rho activation (20 -26), whereas other reports demonstrate no involvement of G ␣q/11 in Rho activation (27)(28)(29)(30). Furthermore, in studies where G ␣q/11 activate Rho, this activation can occur in either a phospholipase C (PLC)-and protein kinase C (PKC)-dependent manner or PLC-and PKC-independent manner (26). Hence, a variety of heterotrimeric G proteins have been shown to link 7TMs to Rho through several distinct pathways that are both 7TM-and cell type-dependent.
Upon the activation of 7TMs, such as the AT 1A R, scaffolding proteins called ␤-arrestins are recruited to the cytoplasmic face of the receptor where they desensitize heterotrimeric G protein signaling (31)(32)(33) and facilitate receptor endocytosis by recruiting components of the endocytotic machinery such as clathrin and AP-2 (34,35). Furthermore, ␤-arrestin 2 is required for signaling from 7TMs through the mitogen-activated protein kinase (MAPK) cascade to activate ERK1/2, c-Jun N-terminal kinase, and p38 (36), and in some instances, ␤-arrestin 2 can activate ERK1/2 in the absence of heterotrimeric G protein activation (37). In these studies, the AT 1A R proved advantageous for distinguishing between heterotrimeric G protein and ␤-arrestin 2 signaling because of the availability of mutant receptors (38) and Ang II analogs (39) that fail to couple the AT 1A R to heterotrimeric G proteins while preserving ␤-arrestin 2 recruitment and signaling (37).
There is growing evidence for a role of ␤-arrestins in facilitating small GTPase-mediated events. For example, ␤-arrestin 2 has been implicated in chemotaxis (40 -42) and both ␤-arrestin 1 and 2 have been shown to directly interact with the small GTPase Arf6 and the guanine nucleotide exchange factors, ARNO (43) and Ral-GDS (44). In light of reports of ␤-arrestin 2-dependent signaling (37,45,46) and ␤-arrestin involvement in small GTPase pathways (40 -44), this study was undertaken to investigate the potential role of ␤-arrestins in RhoA activation and stress fiber formation. As a model, we utilized human embryonic kidney 293 (HEK293) cells stably expressing the AT 1A R (AT 1A R-HEK293). Here we show that in a concerted mechanism with G ␣q/11 , ␤-arrestin 1 is a critical component of RhoA activation and stress fiber formation following Ang II activation of the AT 1A R.

EXPERIMENTAL PROCEDURES
Materials-All of the tissue culture reagents, human Ang II, and G ␣i/o inhibitor pertussis toxin were purchased from Sigma. ROCK inhibitor Y-27632, PLC inhibitors ET-18-OCH 3 and U-73122, and PKC inhibitors Gö-6976, Gö-6983, and RO-31-8425 as well as the tyrosine kinase inhibitor genistein were all purchased from Calbiochem (Darmstadt, Germany). S 1 I 4 I 8 Ang II (SII) was purchased from the Cleveland Clinic Core Synthesis Facility (Cleveland, OH). cDNA constructs for G ␣q and G ␣q QL were generous gifts from the laboratory of Dr. Patrick Casey (Duke University, Durham, NC).
Cell Culture-HEK293 cells were cultured in minimum Eagle's medium supplemented with 10% fetal bovine serum and 1% penicillin/ streptomycin (Sigma). To make HEK293 cells stably expressing the AT 1A R, cells were transfected with rat AT 1A R cDNA that contained a zeocin-selectable marker using FuGENE 6 (Roche Diagnostics, Basel, Switzerland) according to the manufacturer's instructions. Stable clones were selected that contained 1.6 Ϯ 0.2 pmol/mg protein of AT 1A R expression as determined by radioligand binding and were maintained in zeocin (Invitrogen).
RNAi for ␤-Arrestin 1, ␤-Arrestin 2, and G ␣q/11 -AT 1A R-HEK293 cells were split to a density of 100,000 cells/well into 6-well dishes at least 24 h prior to the transfection of control siRNA or siRNAs targeted against ␤-arrestin 1, ␤-arrestin 2, or G ␣q/11 using the Gene Silencer transfection reagent (Gene Therapy Systems, San Diego, CA) as described previously (45). The siRNA sequence targeting overlapping regions of G ␣q and G ␣11 between nucleotides 931 and 951 relative to their start codons was 5Ј-AAGATGTTCGTGGACCTGAAC-3Ј. RhoApulldown and stress fiber formation assays were performed 3 days following the transfection of siRNA(s).
GST-Rhotekin Pulldown Assay-GST-Rhotekin beads were prepared as described previously (47) with the exception that glutathione beads (Amersham Biosciences) with bound GST-Rhotekin were stored at 4°C for not more than 7 days before being discarded. Following stimulation, cells were lysed in 1ϫ ice-cold Mg 2ϩ lysis buffer (125 mM HEPES, pH 7.5, 750 mM NaCl, 5% Igepal CA-630, 50 mM MgCl 2 , 5 mM EDTA, 10% glycerol, 10 g/ml aprotonin, and 10 g/ml leupeptin) (Upstate, Waltham, MA). Lysed cells were then incubated on ice for ϳ10 min and then scraped into prechilled 1.5-ml microcentrifuge tubes. Lysates were centrifuged at 15,800 ϫ g (13,000 rpm in International Equipment Company Micromax tabletop) for 5 min. A volume of lysate containing ϳ65-80 g of protein was then pipetted into a fresh prechilled 1.5-ml microcentrifuge tube and mixed with 30 g of GST-Rhotekin beads in a final volume of ϳ300 l. Lysates with beads were allowed to rotate for 1 h at 4°C before beads were washed 3ϫ with 0.5 ml of 1ϫ Mg 2ϩ lysis buffer. Following the last wash, the majority of supernatant was removed by pipette and the beads were aspirated to dryness with a flat gel-loading tip (Marsh Biomedical, Rochester, NY).
F-actin Cytoskeletal Staining and Quantification-AT 1A R-HEK293 cells were treated as indicated in the figure legends before cells were fixed with phosphate-buffered saline (PBS) containing 4% paraformaldehyde without methanol. Cells were then washed and permeabilized with 0.1% Triton X-100 before being stained with phalloidin as per the manufacturer's protocol (Molecular Probes, Eugene, OR). 25-40 individual cells were then visualized by three-dimensional confocal microscopy (LSM-510META, Carl Zeiss) per treatment group in each experiment. Acquisition parameters were kept constant for all of the experiments. Cell images were then coded and blindly placed into one of four bins: no stress fibers; low stress fibers; medium stress fibers; or high stress fibers. The medium and high stress fiber bins were then combined and shown as the percentage of total cells containing stress fibers.
Cross-linking Immunoprecipitation of ␤-Arrestins and HA-AT 1A R-Immunoprecipitation of HA-epitope-tagged AT 1A R was performed in 100-mm dishes using AT 1A R-HEK293 cells. Cells were starved for 2 h in 4 ml of PBS containing 10 mM HEPES and were subsequently stimulated as described in the figure legends. Incubations were terminated by the addition of 1 ml of PBS containing 2.5 mM dithiobis(succinimidylpropionate) (Sigma) in 50% (v/v) dimethyl sulfoxide and 10 mM HEPES. Monolayers were agitated gently for 30 min at room temperature and then neutralized for 15 min in 50 mM Tris-HCl, pH 7.0. Monolayers were then washed with ice-cold PBS, solubilized in 0.6 ml of radioimmune precipitation assay buffer (50 mM Tris-HCl, 150 mM NaCl, 5 mM EDTA, 1.0% (v/v) Nonidet P-40, 0.1% sodium deoxycholate, 100 M Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, 2 g/ml aprotinin, 1 g/ml pepstatin A, 15 g/ml benzamidin, and 10 g/ml soybean trypsin inhibitor), and clarified by centrifugation. 25-l aliquots of clarified cross-linked detergent lysates were mixed with an equal volume of 2ϫ Laemmli sample buffer. The remainder of each lysate was agitated overnight at 4°C with 20 l of 50% slurry of monoclonal anti-HA affinity-agarose (Sigma). After washing, immunoprecipitates were resolved by SDS-PAGE as described above.
Inositol Phosphate Determination-AT 1A R-HEK293 cells were transiently transfected in 100-mm dishes with either empty plasmids, plasmids containing cDNAs for G ␣q or G ␣q QL, or siRNA targeted against G ␣q/11 . One (plasmid) or two (siRNA) days post-transfection, cells were plated onto poly-D-lysine-coated 12-well plates (BD Biosciences Labware, San Jose, CA). To assay for inositol phosphate production, cells were incubated overnight at 37°C in labeling medium (2.5 Ci (1 Ci ϭ 37 GBq) of myo-[ 3 H]inositol in 0.5 ml of minimum Eagle's medium with 10% fetal bovine serum/well). Cells were washed with Hank's balanced salt solution containing 20 mM LiCl for 15 min at 37°C and then treated with agonist in Hank's balanced salt solution for 10 min. The reactions were terminated by the addition of perchloric acid, and total inositol phosphates were isolated by anion-exchange chromatography on Dowex AG1-X8 columns as described previously (48).
Statistical Analysis-All of the data were graphed and statistically analyzed using GraphPad Prism (GraphPad Software Inc., San Diego, CA).

RESULTS
In AT 1A R-HEK293 cells, RhoA activation by 10 nM Ang II was quite rapid, reaching a maximal response at the earliest time point measured (1 min) (Fig. 1, A-C). This level of RhoA activation was sustained for 10 min before diminishing, reaching 49% maximal response at 20 min. To determine the role of ␤-arrestin 1 and ␤-arrestin 2 in AT 1A R-mediated RhoA activation, RNAi was utilized, which reduced ␤-arrestin levels by ϳ70% in all of the experiments (Fig. 1, A and D). Depletion of ␤-arrestin 2 by RNAi led to a slight increase in the levels of RhoA activation by Ang II over time ( Fig. 1A and 1B, ANOVA p Ͻ 0.01) but did not significantly affect RhoA activation at any specific time point measured. In contrast, ␤-arrestin 1 depletion by RNAi dramatically reduced RhoA activation by 60% across the entire time course (Fig. 1, A and C, ANOVA p Ͻ 0.0001).
Ang II activated RhoA in a dose-dependent fashion with an EC 50 value of 10 nM and maximal RhoA-GTP levels at 100 nM Ang II (Fig. 1, D-E). Following the depletion of ␤-arrestin 1 by RNAi, maximal RhoA activation was significantly inhibited by 70% (Fig. 1, D and E next determined. In the absence of Ang II stimulation, control cells, ␤-arrestin 2-depleted cells, and ␤-arrestin 1-depleted cells showed little to no stress fiber formation (Fig. 2, A and B). Stimulation with 10 nM Ang II for 30 min provoked robust reorganization of F-actin and stress fiber formation in 73 Ϯ 6.1% control cells. The depletion of ␤-arrestin 2 had no significant effect on stress fiber formation as 61 Ϯ 10.2% cells showed actin re-organization upon Ang II stimulation. However, upon depletion of ␤-arrestin 1, only 23 Ϯ 5.2% cells showed stress fiber formation. This 70% reduction in stress fiber formation was consistent with the reduced RhoA activation observed under these conditions. To further elucidate the downstream effector of RhoA involved in the stress fiber formation, cells were pretreated for 15 min with 10 M ROCK inhibitor, Y-27632. Under these conditions, stress fiber formation was inhibited by 100%, further substantiating RhoA and ROCK as the upstream components of a pathway leading to stress fiber formation following AT 1A R activation by Ang II.
Recently, an analog of Ang II, SII, that binds the AT 1A R but does not activate heterotrimeric G proteins, has been shown to recruit ␤-arrestin 2 and activate ERK1/2 (37). However, it is not known whether SII is capable of recruiting ␤-arrestin 1 and, if it does, whether this recruitment in the absence of heterotrimeric G protein activation would be sufficient to activate RhoA. Therefore, the extent of ␤-arrestin 1 recruitment to the AT 1A R and subsequent RhoA activation was determined using SII. Because the affinities of SII and Ang II are markedly different with K d values of 310 and 1.6 nM, respectively (37), the AT 1A R-HEK293 cells were stimulated with 10 M SII or 1 or 100 nM Ang II for 1 min, the HA-AT 1A R was subsequently immunoprecipitated and subjected to SDS-PAGE, and the nitrocellulose was blotted for endogenous ␤-arrestins (Fig. 3, A  and B). Whereas 1 nM Ang II weakly recruited ␤-arrestin 1 and ␤-arrestin 2 after 1 min, 100 nM Ang II robustly recruited both ␤-arrestin 1 and ␤-arrestin 2 with a preference for ␤-arrestin 2. Upon stimulation with 10 M SII, ␤-arrestin 1 and ␤-arrestin 2 were clearly recruited to the AT 1A R in a similar pattern (␤arrestin 2 greater than ␤-arrestin 1) as that induced by Ang II.
To determine whether SII could induce the sequestration of ␤-arrestin 1 into endocytotic vesicles in a fashion similar to that induced by Ang II, GFP-␤-arrestin 1 was transiently transfected into AT 1A R-HEK293 cells. Following 20 min of 10 nM Ang II or 10 M SII stimulation, GFP-␤-arrestin 1 had moved from the cytoplasm into internalized vesicles (Fig.  3C), typical of an Ang II-induced pattern of ␤-arrestin 1 recruitment. Quantification of the percentage of cells showing GFP-␤-arrestin 1 or GFP-␤-arrestin 2 in vesicles following 100 nM Ang II or 10 M SII stimulation revealed that SII was less effective than Ang II at inducing ␤-arrestin endocytosis. Specifically, GFP-␤-arrestin 1 was sequestered into vesicles in 94 Ϯ 3.9 and 23.0 Ϯ 3.0% cells for Ang II and SII, respectively, whereas GFP-␤-arrestin 2 was sequestered into vesicles in 97.0 Ϯ 1.3 and 72.0 Ϯ 11.0% cells for Ang II and SII, respectively.
To determine whether the SII-mediated recruitment of ␤-arrestin 1 in the absence of heterotrimeric G protein coupling would be sufficient to activate RhoA, AT 1A R-HEK293 cells were stimulated with 10 M SII for 5 min and the levels of activated RhoA were determined. SII was able to reproducibly activate RhoA, but the levels of RhoA-GTP were only 5.7 Ϯ 4.5% compared with a maximal stimulation of RhoA-GTP following 100 nM Ang II (Fig. 4, A and B). Additionally, 10 M SII was incapable of inducing stress fiber formation after 30 min of stimulation (Fig. 4C). However, in the same experiment, SII stimulated the ␤-arrestin 2-mediated MAPK component ERK1/2 by 35 Ϯ 16% Ang II-stimulated ERK1/2, consistent with previous findings (37).
Given that ␤-arrestin 1 seemed to be necessary but not sufficient to significantly activate RhoA or to induce stress fiber formation in HEK293 cells, the role of the AT 1A R-coupled heterotrimeric G proteins G ␣q/11 and G ␣i was assessed. G ␣q/11 was depleted in AT 1A R-HEK293 cells by ϳ90% using RNAi (Fig.  5A). To show that this level of G ␣q/11 reduction led to a significant loss of AT 1A R coupling and signaling through this G protein, a PI hydrolysis assay was performed revealing almost complete inhibition of PI turnover when G ␣q/11 was silenced (Fig. 5B). When Rho activation was determined following the depletion of G ␣q/11 , it was found to be inhibited by 68% follow-ing a 5-min stimulation with 10 nM Ang II (Fig. 5C). However, neither PLC (ET-18-OCH 3 and U-73122), PKC (Gö-6976, Gö-6983, and RO-31-8425) nor tyrosine kinase (genistein) inhibitors were able to inhibit RhoA activation (Fig. 5D). Furthermore, the G ␣i/o inhibitor, pertussis toxin, had no significant effect on RhoA activation.
The dependence of stress fiber formation upon G ␣q/11 then was determined in AT 1A R-HEK293 cells with G ␣q/11 or simultaneous G ␣q/11 and ␤-arrestin 1 depletion. In the absence of Ang II stimulation, control cells and cells depleted of G ␣q/11 or G ␣q/11 and ␤-arrestin 1 showed little to no stress fiber formation (Fig.  6, A and B). Stimulation with Ang II showed stress fiber formation in 76 Ϯ 2.3% control cells. However, only 33 Ϯ 4.2% G ␣q/11 -depleted cells, 23 Ϯ 5.2% ␤-arrestin 1-depleted cells (shown again from Fig. 2B), or 21 Ϯ 3.5% G ␣q/11 and ␤-arrestin 1-depleted cells showed significant stress fiber formation upon Ang II stimulation. This 57-72% inhibition of stress fiber formation following G ␣q/11 and ␤-arrestin 1 depletion is again consistent with the reduced RhoA activation observed under similar conditions. In addition, because silencing both G ␣q/11 and ␤-arrestin 1 simultaneously did not lead to more RhoA inhibition than silencing either G ␣q/11 or ␤-arrestin 1 alone, FIG. 5. The role of heterotrimeric G proteins in Rho activation by Ang II. A, representative blots of endogenous G ␣q/11 , RhoA-GTP, or Total RhoA. B, PI hydrolysis was determined over a range of Ang II concentrations in control cells (f), and cells were depleted of G ␣q/11 (OE) by RNAi. Data represent the mean Ϯ S.E. for three independent experiments. C, similarly, RhoA activation by 10 nM Ang II was determined in control cells or cells depleted of G ␣q/11 . Data represent the mean Ϯ S.E. for four independent experiments. D, RhoA activation was determined following pretreatment with various inhibitors including PLC inhibitors ET-18-OCH 3 and U-73122, PKC inhibitors Gö-6976, Gö-6983, and RO-31-8425, tyrosine kinase inhibitor genistein, and G ␣i/o inhibitor pertussis toxin (PTX). Ang II-stimulated controls were normalized to 100%. Data represent the mean Ϯ S.E. for three independent experiments. AT 1A R-mediated RhoA Activation Is Dependent on ␤-Arrestin 1 and G ␣q/11 these data suggest that both molecules act concurrently in the same pathway.
The weak ability of SII to activate Rho, coupled with the apparent involvement of G ␣q/11 in Ang II stimulation, raised the possibility of a concerted mechanism requiring both ␤-arrestin 1 and G ␣q/11 . Such a hypothesis would explain the failure of SII to significantly activate Rho as a consequence of its inability to trigger G protein activation and would further predict that supplementation of SII with G ␣q activity would lead to Rho activation. To test this idea, we transfected cells with either wild type G ␣q or constitutively active G ␣q QL (Fig.  7A). As might be expected, overexpression of G ␣q or G ␣q QL led to modest but significant increases in basal PI turnover (Fig.  7B) but did not cause any significant alteration in sensitivity to Ang II stimulation. Importantly, SII failed to elevate PI turnover even in the presence of overexpressed G ␣q or G ␣q QL, further strengthening the evidence that this Ang II analog is incapable of activating heterotrimeric G proteins.
As shown in Fig. 7, C and D, transfection of G ␣q or G ␣q QL led to weak activation of RhoA similar to that produced by 10 M SII in the presence of endogenous levels of G ␣q . Strikingly, however, the combination of SII stimulation in the presence of transfected G ␣q or G ␣q QL led to synergistic activation of RhoA. In contrast, the activation of RhoA by Ang II was not significantly increased by transfection of G ␣q (p ϭ 0.24, n ϭ 8) or G ␣q QL (p ϭ 0.24, n ϭ 12) (Fig. 7A). Thus, these data are consistent with the notion of a concerted ␤-arrestin 1/G ␣q/11 mechanism for Rho activation in response to Ang II stimulation. DISCUSSION We show that the activation of the small GTPase RhoA as well as subsequent stress fiber formation following stimulation of the AT 1A R requires ␤-arrestin 1 as well as G ␣q/11 . Furthermore, we find that, in HEK293 cells, activated RhoA signals through ROCK to activate stress fiber reorganization, consistent with previous findings (49).
␤-Arrestins, which were initially shown to be involved in the desensitization of heterotrimeric G proteins and the internalization of 7TMs, have only recently been recognized as playing roles in cellular events ranging from activation of MAPK cascade components (37,46) to apoptosis (50,51) and chemotaxis (41,42). Most recently, following stimulation of the AT 1A R, ␤-arrestin 2 was shown to signal to the MAPK ERK1/2 in the absence of heterotrimeric G protein activation (37). Additional studies revealed that the ␤-arrestin 2-dependent activation of ERK1/2 was further enhanced when ␤-arrestin 1 was silenced, indicating that ␤-arrestin 1 reciprocally regulates ␤-arrestin 2-mediated ERK1/2 activation (46). This reciprocal regulation was hypothesized to be the result of competition between ␤-arrestin 2 and ␤-arrestin 1 for the same binding sites on the activated AT 1A R. Consequently, upon the depletion of ␤-arrestin 1, more ␤-arrestin 2 is able to bind the phosphorylated C terminus of the AT 1A R, leading to greater ERK1/2 activation.
We find that the recruitment of ␤-arrestin 1 to the AT 1A R not only regulates ␤-arrestin 2 activation of ERK1/2 as previously shown (46) but also leads to the activation of the small GTPase RhoA. Furthermore, ␤-arrestin 1 activation of RhoA leads to actin reorganization and stress fiber formation. Not surprisingly, it was also observed that ␤-arrestin 2 reciprocally regulates ␤-arrestin 1 activation of RhoA over time in a manner analogous to the reciprocal regulation of ␤-arrestin 2 by ␤-arrestin 1 with regard to ERK1/2 activation described above. Hence, it now appears that upon the activation of the AT 1A R, ␤-arrestin 1 and ␤-arrestin 2 are competitively recruited to the receptor leading to RhoA activation through a synergistic mechanism requiring ␤-arrestin 1 and G ␣q/11 and to ERK1/2 activation through independent ␤-arrestin 2 and G ␣q/11 mechanisms (Fig. 8). Furthermore, with respect to RhoA activation, ␤-arrestin 1 and G ␣q/11 appeared to be acting coordinately to activate RhoA as silencing ␤-arrestin 1, G ␣q/11 , or both together lead to nearly an equivalent (70 -80%) loss in RhoA activation and/or stress fiber formation. Importantly, the remaining 20 -30% RhoA activation and/or stress fiber formation following silencing of ␤-arrestin 1 and/or G ␣q/11 is probably reflective of incomplete silencing of ␤-arrestin 1 and G ␣q/11 obtained using RNA interference techniques.
Although the role of the heterotrimeric G proteins G ␣q/11 in mediating RhoA activation has been somewhat controversial (20), our data support the findings that G ␣q/11 is involved in the activation of RhoA and stress fiber formation. Furthermore, we have expanded upon previous findings to demonstrate that, in response to Ang II, ␤-arrestin 1 is a critical component of this G ␣q/11 -dependent RhoA pathway. Interestingly, the concurrent activation of both ␤-arrestin 1 and G ␣q/11 by Ang II leads to far FIG. 7. ␤-Arrestin 1 and G ␣q/11 work coordinately to activate RhoA. AT 1A R-HEK293 cells were transiently transfected with empty pcDNA3 vector, G ␣q , or G ␣q QL cDNAs. A, representative blots of G␣q/11 , RhoA-GTP, or Total RhoA. B, two days following transfection, average PI hydrolysis was measured in cells stimulated for 5 min with various concentrations of Ang II in pcDNA3 (f), G ␣q (OE), or G ␣q QL (•) transfected cells or various concentration of SII in pcDNA3 (Ⅺ), G ␣q (‚), or G ␣q QL (E) transfected cells. Data represent the mean Ϯ S.E. for three independent experiments. C and D, RhoA activation following 5 min of 10 nM Ang II or 10 M SII was determined in control cells or cells containing overexpressed G ␣q or G ␣q QL. Data represent the mean Ϯ S.E. of 5-7 experiments. Ang II-activated RhoA levels from pcDNA3-transfected cells were normalized to 100%. more robust RhoA activation than does the overexpression of wild type or constitutively active G ␣q alone or ␤-arrestin 1 recruitment alone.
Previously, it has been found that G ␣q/11 activates RhoA in either a PLC-and PKC-dependent manner and/or a tyrosine kinase-dependent manner (17,20). However, we find that PLC and PKC inhibitors as well as the tyrosine kinase inhibitor genistein do not inhibit AT 1A R-mediated RhoA activation. In contrast, PKC inhibitors Gö-6983 and RO-31-8425 and the tyrosine kinase inhibitor genistein actually increased AT 1A Rmediated RhoA activation, suggesting that PKC and tyrosine kinase might act to inhibit RhoA activation in this system.
The precise mechanism by which ␤-arrestin 1 and G ␣q/11 activate RhoA in a concerted manner remains unknown. How-ever, it is known that ␤-arrestins can act as scaffolding proteins for ARNO (43) and Ral-GDS (44), the guanine nucleotide exchange factors for Arf6 and Ral, respectively. Interestingly, the Rho-guanine nucleotide exchange factors Lbc (22) and LARG (52) have recently been shown to interact with G ␣q/11 , which leads to the hypothesis that ␤-arrestin 1 might recruit Lbc or LARG to the activated G ␣q/11 to facilitate RhoA activation and the subsequent stress fiber formation. However, whether or not Lbc, LARG, or other Rho-guanine nucleotide exchange factors play a role in AT 1A R-mediated RhoA activation in HEK293 cells remains to be determined.
Our findings that ␤-arrestin 1 works in concert with G ␣q/11 to activate RhoA-dependent stress fibers following the activation of the AT 1A R advances the idea that ␤-arrestins are not only required for receptor desensitization and internalization but are also involved in cellular signaling. As ␤-arrestins are recruited to the vast majority of 7TMs, these findings are relevant not only to small GTPase signaling pathways but also to 7TM-stimulated pathways, in general. Furthermore, these findings stress the potential importance of novel ligands such as the Ang II analog SII that are capable of activating subsets of 7TM-mediated pathways. Such selective activation of signaling pathways may lead to the development of drugs that are capable of conferring therapeutic effects with less capability for stimulating additional signaling pathways that may inflict harmful side effects.
FIG. 8. Model of ␤-arrestin-dependent signaling to RhoA or ERK1/2. Present data support the hypothesis that ␤-arrestin 1 and G ␣q/11 work coordinately to activate RhoA. Activated RhoA then leads to stress fiber formation via the downstream effector of RhoA, ROCK. Neither ␤-arrestin 1 recruitment alone by SII nor overexpression of G ␣q or G ␣q QL is sufficient to robustly activate RhoA in a manner analogous to that induced by Ang II, thereby suggesting that both arrestin and heterotrimeric G protein activation is required. Conversely, recent data has been published in which ␤-arrestin 2 signals to the MAPK component ERK1/2 independently from the heterotrimeric G protein pathway to ERK1/2 via PKC. Therefore, it seems that ␤-arrestin 1 and ␤-arrestin 2 are endowed with the capacity to signal in a heterotrimeric G proteinand ␤-arrestin 1-dependent manner for RhoA activation and stress fiber formation and a heterotrimeric G protein-independent and ␤-arrestin 2-dependent manner for ERK1/2 activation.