Constitutively active homo-oligomeric angiotensin II type 2 receptor induces cell signaling independent of receptor conformation and ligand stimulation.

Members of the G-protein-coupled receptor superfamily (GPCRs) undergo homo- and/or hetero-oligomerization to induce cell signaling. Although some of these show constitutive activation, it is not clear how such GPCRs undergo homo-oligomerization with transmembrane helix movement. We previously reported that angiotensin II (Ang II) type 2 (AT(2)) receptor, a GPCR, showed constitutive activation and induced apoptosis independent of its ligand, Ang II. In the present study, we analyzed the translocation and oligomerization of the AT(2) receptor with transmembrane movement when the receptor induces cell signaling. Constitutively active homo-oligomerization, which was due to disulfide bonding between Cys(35) in one AT(2) receptor and Cys(290) in another AT(2) receptor, was localized in the cell membrane without Ang II stimulation and induced apoptosis without changes in receptor conformation. These results provide the direct evidence that the constitutively active homo-oligomeric GPCRs by intermolecular interaction in two extracellular loops is translocated to the cell membrane and induces cell signaling independent of receptor conformation and ligand stimulation.

Members of the G-protein-coupled receptor superfamily (GPCRs) undergo homo-and/or hetero-oligomerization to induce cell signaling. Although some of these show constitutive activation, it is not clear how such GPCRs undergo homo-oligomerization with transmembrane helix movement. We previously reported that angiotensin II (Ang II) type 2 (AT 2 ) receptor, a GPCR, showed constitutive activation and induced apoptosis independent of its ligand, Ang II. In the present study, we analyzed the translocation and oligomerization of the AT 2 receptor with transmembrane movement when the receptor induces cell signaling. Constitutively active homo-oligomerization, which was due to disulfide bonding between Cys 35 in one AT 2 receptor and Cys 290 in another AT 2 receptor, was localized in the cell membrane without Ang II stimulation and induced apoptosis without changes in receptor conformation. These results provide the direct evidence that the constitutively active homo-oligomeric GPCRs by intermolecular interaction in two extracellular loops is translocated to the cell membrane and induces cell signaling independent of receptor conformation and ligand stimulation.
Recently, there have been many new insights regarding Gprotein-coupled receptor superfamily (GPCRs), 1 such as on homo-and hetero-oligomerization (1,2), constitutive activation (3,4), and receptor-associated protein. GPCR oligomers may be "contact" oligomers (5), where two binding pockets are formed from regions donated by both monomers instead of one binding site formed by one "subunit" receptor. Moreover, the dissociation of receptor homo-oligomers by reducing agents has been observed for several rhodopsin-like GPCRs (6 -9). ␦-Opioid re-ceptor is unable to form oligomers when the terminal 15 amino acids are truncated (6), indicating that the C terminus might be a site of intermolecular interaction. Although these observations indicate that a disulfide linkage is important in the formation of receptor oligomers, it is not known whether the bond is inter-or intramolecular. Here we demonstrate that constitutively active homo-oligomeric (by intermolecular contact) angiotensin II type 2 receptor induces cell signaling independent of receptor conformation and ligand stimulation.
Cell Cultures and Treatments-Mouse fibroblast R3T3 cells and Chinese hamster ovary (CHO)-K1 cells (ATCC CCL-61) were maintained in 10% fetal bovine serum with penicillin-and streptomycin-supplemented Dulbecco's modified Eagle's essential medium (Invitrogen) in 5% CO 2 at 37°C. Apoptosis was measured in cells maintained for 48 h in Dulbecco's modified Eagle's essential medium without serum. Cell viability was Ͼ95% by trypan blue exclusion analysis in control experiments. The level of AT 2 receptor stimulation was measured in the presence of 0.1 M [Sar 1 ]Ang II (K d ϭ 0.3 nM), and inhibition was measured in the presence of 10 M AT 2 selective non-peptide antagonist PD123319 (K d ϭ 5 nM). [Sar 1 ]Ang II was added every 12 h to compensate for potential degradation. Inhibition experiments were performed in the presence of the p38 mitogen-activated protein kinase (p38 MAPK)-specific inhibitor SB203580 (10 M) and the caspase-3 inhibitor (1 M) for 48 h after serum starvation.
CHO Cell Transfection and Establishment of CHO Cell Lines-The rat AT 2 receptor gene described earlier was used (10). The wild-type (WT), N127G, inactive mutant AT 2 receptor (4), and C35A, C290A, and C35A/C290A genes (10) cloned in the pEGFP (enhanced green fluorescent protein), pEYFP (enhanced yellow fluorescent protein)), pECFP (enhanced cyan fluorescent protein), and pDsRed (discosoma red fluorescent protein) vectors that contain the neomycin (G418) resistance gene were transfected into CHO cells using the Lipofectamine 2000 liposomal reagent according to the manufacturer's instructions (Roche Applied Science). Clonal cell lines that permanently expressed AT 2 receptor were selected by 800 g/ml G418 and evaluated by fluorescence-activated cell sorter.
PAGE, Western Blotting, and Immunoprecipitation-Cell membrane and cytoplasmic fraction were prepared as described previously (11,12). For Western blotting, equal amounts of samples on a protein basis as determined using the Bradford reagent (Bio-Rad) were resolved on non-reducing 12% SDS-PAGE. Cell membrane was incubated with or without 1 mM dithiothreitol (DTT) at 30°C for 1 h and with loading buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris-HCl (pH 7.4), 1 mM EDTA, 0.2 mM sodium orthovanadate, 0.2 mM phenylmethylsulfonyl fluoride, 0.5% Nonidet P-40) at room temperature for 30 min. Western blot analysis was performed with primary antibodies as specified in each case. Horseradish peroxidase-conjugated secondary antibody and the enhanced chemiluminescent substrate system (Amersham Biosciences) were used as a detection system. The signal was independently quantified by a digital image analysis system.
Analysis of Apoptosis-Nuclear fragmentation and condensation analysis with 4Ј, 6-diamidino-2-phenylindole (DAPI) staining was used to identify apoptotic cells at the cellular level with a DMIRB digital fluorescent microscope (Leica) and a LSM410 laser scanning confocal microscope (Carl Zeiss Co. Ltd.). For imaging, the cells were plated on lysine-coated microscope slides and then treated for up to 48 h. The cells were washed and fixed with 1% paraformaldehyde at 4°C for 15 min and then stained with DAPI. Five images of separate locations were taken for each sample, and the percentage of apoptotic cells was quantified. The percentage of apoptotic cells was also determined by the in situ cell death detection kit, TMR red TM , according to the manufacturer's protocol (Roche Diagnostics). Briefly, cells were fixed in 2% paraformaldehyde, and terminal deoxynucleotide transferase-mediated deoxyuridine 5Ј-triphosphate nicked end labeling (TUNEL) reaction mix was added. Cells were analyzed by flow cytometry.
Caspase-3-like Activity Assay-Caspase-3-like protease activity was determined by colorimetric assay using the CaspACE TM assay system (Promega) according to the manufacturer's instructions. Transfected and control cells (10 6 ) were lysed in lysis buffer by freezing and thawing followed by centrifugation (15,000 ϫ g for 20 min at 4°C). Caspase-3like activity was measured in supernatant following proteolytic cleavage of the colorimetric substrate Ac-DEVD-pNA. DEVD-pNA was used as a standard in the assay buffer ( max 405 nm).
Radioligand Binding Studies-K d and B max of the receptor were estimated by 125 I-[Sar 1 ,Ile 8 ]Ang II binding experiments were carried out under equilibrium binding and Scatchard plot analysis, as described previously (11,12). Membranes expressing the WT or the mutant receptor were incubated with 300 pM 125 I-[Sar 1 ,Ile 8 ]Ang II for 1 h at 22°C in a 125-l volume. The binding experiments were stopped by filtering the binding mixture through Whatman GF/C glass fiber filters, which were extensively washed further with binding buffer. The bound ligand fraction was determined from the cpm remaining on the membrane. Binding kinetics values were determined using the computer program Ligand.
Reactions with MTS Reagents by Reporter Cysteine Accessibility Mapping-Aliquots of cell membranes (20 l) were incubated with or without MTS reagents at the stated concentrations (0.1-12.5 mM) at 22°C for the indicated times (2-10 min) in 20 mM HEPES buffer (pH 7.4). The reaction mix was then diluted 75-fold with cold buffer to stop the reaction, centrifuged for 10 min at 16,000 ϫ g at 4°C, and resuspended in 200 l of buffer. A 150-l aliquot was used for 125 I-[Sar 1 ,Ile 8 ]Ang II binding analysis. The percent inhibition of 125 I-[Sar 1 ,Ile 8 ]Ang II binding was calculated as {1-[(specific binding after MTS reagent)/(specific binding without reagent)]} ϫ 100%. The experiment was performed 4 -10 times.
Statistical Analysis-The results are expressed as the mean Ϯ S.E. of three or more independent determinations. Significant differences in measured values were evaluated with an analysis of variance using Bonferroni's test and the unpaired Student's t test. Pearson product moment correlation coefficients were calculated. Statistical significance was set at Ͻ0.05.

RESULTS AND DISCUSSION
The PC12W pheocromocytoma cell line has been used as a model for study because apoptosis in these cells (13) resembles the in vivo situation described in ovarian granulosa cells and other remodeling tissues (14 -16). Serum starvation leads to a ϳ5-fold increase in AT 2 receptor expression (Fig. 1a). Immunoblotting showed that the AT 2 receptor underwent homo-oligomerization in the cell membrane (Fig. 1a). AT 2 receptorexpressing native cell lines are not a suitable model for concluding the localization and homo-oligomerization of constitutively active AT 2 receptor for apoptosis and are even less appropriate for concluding a role for the induced overexpression of AT 2 receptor protein. Since CHO cells are an established epithelial lineage of non-transformed cells that are capable of growing under serum starvation with appropriate supplements and since we previously reported that exogenous overexpression of the AT 2 receptor induced apoptosis in CHO cells (4), we used these cells as surrogate models to link the de novo expression of the AT 2 receptor to apoptosis.
Using molecular imaging techniques, which show the translocation of the AT 2 receptor to the cell membrane in living cells, we analyzed the translocation of AT 2 -WT-EGFP receptor that had been permanently transfected in CHO cells (Fig. 1b). Although the expression level of the AT 2 receptor in normal adult tissues is ϳ10 -20 fmol/mg of protein (B max ), it is re-expressed ϳ5-25-fold over basal levels during the remodeling of tissues, where the AT 2 receptor is thought to play a role in the apoptosis of smooth muscle cells, fibroblasts, and endothelial cells (14,15). Therefore, a CHO cell line, which showed an expression level of 341 Ϯ 13 fmol/mg of protein in the cell membrane (Table I), was used for pharmacological examinations. As shown in Fig. 1b, AT 2 -WT-EGFP receptor was localized in the cell membrane after 24 h under serum-free conditions, whereas serum starvation by itself did not induce EGFP translocation. The expression level of AT 2 -WT-EGFP receptor in the cell membrane after 24 h under serum-free conditions was about 2.4-fold higher than the expression under serum conditions (Table I). In addition, AT 2 -WT-EGFP receptor was localized in the cell membrane after 24 h under serum-free conditions ( Fig.  1b), suggesting that AT 2 -WT-EGFP receptor translocation in the cell membrane may induce cell signaling.
Morphological features associated with apoptotic cells, such as nuclear DNA condensation and fragmentation, began ϳ24 h (the time at which growth arrest was established) after shifting to serum starvation and was complete after 48 h in a CHO cell line expressing AT 2 -WT-EGPF receptor (Fig. 1c). The TUNEL method was used to analyze the percentage of apoptotic cells under various treatments (Fig. 1d); the CHO cell line expressing AT 2 -WT-EGPF receptor showed 64 Ϯ 5% apoptosis. These cells also exhibited membrane blebbing, cytoplasmic shrinkage, and inhibition of protein synthesis (data not shown). Although we established five different AT 2 -WT-EGPF receptorexpressing CHO cell lines, all of them showed similar apoptotic effects under serum-free conditions. In contrast, DAPI staining of cells that expressed AT 2 -N127G-EGFP receptor, which is an inactive AT 2 receptor (4), revealed cell nuclei with sharp round edges and diffuse chromatin (Fig. 1c). In addition, the expression level of AT 2 -N127G receptor in the cell membrane after 24 h under serum-free conditions was about 2.5-fold higher than the expression under serum conditions, and this increase in the expression level of AT 2 -N127G receptor is comparable with that of AT 2 -WT receptor ( Table I). The translocation of the AT 2 receptor to the cell membrane may be necessary but not sufficient to induce apoptosis. Since we previously reported that AT 2 receptor-induced apoptosis in A7r5 cells was blocked by the pharmacological inhibition of p38 MAPK and caspase-3 (4), the effects of p38 MAPK inhibitor and caspase-3 inhibitor were analyzed in this cell system. The p38 MAPK-specific inhibitor SB203580 blocked AT 2 receptor-induced apoptosis by about 60%. The caspase-3 inhibitor Ac-DEVD-cmk (1 mM) inhibited AT 2 receptor-induced apoptosis by about 40% (Fig. 1d). Combined treatment with SB203580 and Ac-DEVD-cmk almost completely blocked AT 2 -WT-receptor-induced apoptosis.
To determine whether the homo-oligomerization of the AT 2 receptor in the cell membrane induces apoptotic signaling, immunoblotting of the AT 2 -WT and AT 2 -N127G receptors was performed. AT 2 -WT receptor homo-oligomerized in the cell membrane under serum-free conditions for 24 h (Fig. 1e). Although AT 2 -N127G receptor mainly homo-oligomerized in the cell membrane, a monomer band was also observed. After being pretreated with DTT under non-reducing conditions, AT 2 -WT  immunoblots (a, d, and f) and pictures (b and c) are shown. Three independent determinations were performed, and similar results were observed (a-d and f).
receptor showed only a monomer band, suggesting that disulfide bonding may be important for homo-oligomerization (Fig.  1e). In addition, whereas both receptors mainly homo-oligomerized in the cell membrane under serum-free conditions for 48 h (Fig. 1f), only homo-oligomerized AT 2 -WT receptor induced apoptosis. Although the AT 2 receptor contains consensus sites for phosphorylation, it is not phosphorylated in response to Ang II and does not internalize (17). These results suggest that a constitutively active homo-oligomerized AT 2 -WT receptor but not an inactive homo-oligomerized AT 2 receptor (N127G)-induced apoptosis.
Next, we co-transfected CHO cells with AT 2 -WT-EGFP and AT 2 -WT-DsRed receptors to confirm the homo-oligomerization of AT 2 -WT receptor in the cell membrane. AT 2 -WT-EGFP and AT 2 -WT-DsRed receptors were co-localized in the cell membrane after 24 h under serum-free conditions (Fig. 2a). Many studies on GPCR oligomerization have used the co-immunoprecipitation of differentially epitope tagged forms of GPCRs. Coexpression of both the EGFP-and the DsRed-fusion forms of the AT 2 receptors in this study followed by immunoprecipitation with anti-EGFP antibody and immunoblotting of the precipitated samples with an anti-DsRed antibody detected the presence of DsRed-fused AT 2 receptors. In addition, AT 2 receptors appear to be localized in a perinuclear compartment, and this may be an artifact of oligomerization. These results indicate that AT 2 -WT-EGFP and AT 2 -WT-DsRed receptors homooligomerized (Fig. 2b). As shown in Fig. 2c, CHO cells, which were co-transfected with AT 2 -WT-EGFP and AT 2 -WT-DsRed receptors, also showed morphological features of apoptotic cells, such as irregular nuclei and membrane blebbing under serum-free conditions. Although early studies using radioligand binding and cross-linking had predicted homo-oligomeric receptors, the significance of these findings was not clear at that time. Although the functional importance of homooligomerization is much better defined for other receptors, such as tyrosine-kinase and steroid-hormone receptors (18), in light of several reports on GPCR oligomers, it is not yet clear whether or not these receptors exist functionally as homooligomers. Although constitutive activity has been observed for more than 60 wild-type GPCRs and different species including humans (19), this is the first report in which a constitutively active GPCR, the AT 2 receptor, has been shown to be functionally homo-oligomerized in the cell membrane.
The agonist [Sar 1 ]Ang II and the antagonist PD123319 do not modulate apoptosis in AT 2 receptor-transfected A7r5 cells (4). Serum starvation by itself induced the translocation of AT 2 -WT-EGFP receptor from the cytoplasm to the cell membrane, whereas treatment with [Sar 1 ]Ang II or PD123319 did not induce the delocalization of AT 2 -WT-EGFP from the cell membrane to the cytoplasm (Fig. 2d). Immunoblotting indicated that the homo-oligomerization of AT 2 -WT-EGFP receptors was not affected by [Sar 1 ]Ang II or PD123319 for up to 24 h under serum-free conditions (Fig. 2e). In addition, the expression level of AT 2 -WT-EGFP receptor in the cell membrane also did not show any changes in the presence of [Sar 1 ]Ang II or PD123319 after 24 h under serum-free conditions. AT 1 receptor is stabilized in the R state (inactive state), and agonist binding causes transition to the R* state (fully active state). In the AT 1 receptor, this transition may proceed through a relaxed intermediate activated RЈ state (partially active state) (20). In the case of the AT 2 receptor, constitutively active AT 2 receptor may be in the RЈ state but not in the R state. Since PD123319 is a neutral antagonist but not an inverse agonist for constitutively active AT 2 receptor-induced apoptosis (4), the antagonist may not affect AT 2 receptor translocation, and the AT 2 receptor may remain in the RЈ state in the cell membrane. Taken together, ligand-independent AT 2 receptor-induced signaling may cause apoptosis associated with AT 2 receptor translocation and homo-oligomerization.
Fluorescence resonance energy transfer (21) and bioluminescence resonance energy transfer (22) are valuable approaches for analyzing the oligomerization of receptor after ligand stimulation. Although fluorescence resonance energy transfer analysis was performed using CHO cells, which co-expressed both the AT 2 -EYFP and the AT 2 -ECFP receptors (data not shown), there was no energy transfer after Ang II stimulation because the AT 2 receptor undergoes constitutive homo-oligomerization independent of Ang II. Since fluorescence resonance energy transfer and bioluminescence resonance energy transfer were not useful in this study, several AT 2 mutant receptors were used to analyze the specificity of the constitutive homo-oligomerization of the AT 2 receptor. Four Cys residues, Cys 35 , Cys 117 , Cys 195 and Cys 290 , in the AT 2 receptor are located in the extracellular domain, where the redox environment may facilitate two disulfide bonds. A disulfide bond linking transmembrane (TM) 3 and extracellular loop 2 occurs in Ͼ91% of GPCRs (23). In the case of the AT 2 receptor, a disulfide bond between Cys 117 and Cys 195 in the extracellular domain is critical for the polypeptide to gain a ligand binding conformation, and in its absence, the polypeptide is irreversibly misfolded (15). Cys 35 in the N terminus and Cys 290 in extracellular loop 3 in the AT 2 receptor may play a central role in homo-oligomerization of the receptor through intermolecular or intramolecular interaction by disulfide linkage. AT 2 receptor translocation was analyzed using C35A-EGFP, C290A-EGFP, and C35A/C290A-EGFP receptors (Fig. 3a). C35A-EGFP and C290A-EGFP receptors were translocated to the cell membrane after 24 h under serum-free conditions, whereas C35A/C290A was found in the cytoplasm (Fig. 3a). C35A-EGFP and C290A-EGFP receptors but not C35A/C290A-EGFP receptor were homo-oligomerized by immunoblotting (Fig. 3b) and induced apoptosis after 48 h under serum-free conditions (Fig. 3c), suggesting that the interaction of Cys 35 and Cys 290 in the homo-oligomerization of the AT 2 receptor may be critical for inducing apoptosis. Since AT 2 receptor expression-linked apoptosis has been shown to activate caspase-3 (4), we examined caspase-3-like activity as an AT 2 receptor-induced signal. Basal caspase-3like activity in cell lines that expressed WT-EGFP, C35A-EGFP, and C290A-EGFP receptors was much higher than that in cell lines that expressed EGFP as a control and N127G receptor, which is an inactive receptor, under serum-free conditions (Fig. 4a). Basal caspase-3-like activity in C35A/C290A-EGFP was comparable with that in cell lines that expressed EGFP or N127G (Fig. 4a). However, [Sar 1 ]Ang II caused a significant increase in caspase-3-like activity in cell lines that expressed WT, C35A, C290A, and N127G receptors but not C35A/C290A receptors. The increased activity was not suppressed when cells were pretreated with PD123319. As shown in Table I, the expression level of C35A receptor in the cell membrane after 24 h under serum-free conditions was 2.1-fold higher than that in serum conditions, which is comparable with the results in WT, whereas there was no significant difference in the expression level of C290A receptor between serum and serum-free conditions because C290A receptor is already located in the cell membrane under serum conditions. Homooligomeric AT 2 receptor in the cell membrane is not sufficient but may be necessary for caspase-3-like activation. The activation of caspase-3 is thought to be important for specific cleavage of cell cycle regulatory proteins, such as p21 waf1 and p27 kip1 , in apoptotic cells (24).
Since we previously reported that constitutive activation of AT 1 receptor induces movement of the TM2-TM7 helix by reporter cysteine accessibility mapping using sulfhydryl-specific reaction with MTSEA (11,12), we also analyzed AT 2 receptor TM movement by reporter cysteine accessibility mapping (Fig. 4, b and c). Exposure of native Cys residues to a water-accessible ligand pocket was measured. The percentage of inhibition of 125 I-[Sar 1 ,Ile 8 ]Ang II-specific binding treated with 2.5 mM MTSEA ϩ for 2-10 min in the inactive N127G receptor was significantly lower than that in the constitutively active WT receptor (Fig. 4b). Exposure of the constitutively active WT receptor to MTSEA for 2 min reduced specific binding of the peptide antagonist 125 I-[Sar 1 ,Ile 8 ]Ang II by nearly 47%, whereas exposure of the inactive N127G receptor reduced specific binding by 18% (Fig. 4c). In the case of N127G receptor, the Cys residues are moved to an area inaccessible by water, and this may reduce the specific binding of the peptide antagonist 125 I-[Sar 1 ,Ile 8 ]Ang II. There are three Cys residues in the upper site of the plasma membrane that face the extracellular space: Cys 169 and Cys 172 in TM IV and Cys 269 in TM VI. Since one or more of these Cys residues may expose the ligand pocked in WT receptor, we replaced these Cys residues by Ala (C169A, C172A, and C269A). Cys accessibility for C169A was only 10%, suggesting that Cys 169 in TM IV was exposed to the ligand pocket in WT receptor. Interestingly, Cys accessibilities for C35A, C290A, and C35A/C290A receptors are similar to that of WT receptor, indicating that these mutations did not induce conformational changes in the AT 2 receptor. In addition, Cys accessibilities for C35A/N127G, N127G/C290A and C35A/ N127G/C290A receptors are similar to that for N127G receptor. C35A/C290A receptor but not N127G receptor did not express and form a homo-oligomer in the cell membrane, suggesting that both Cys 35 and Cys 290 are important for making a disulfide bond and translocating the receptor to the cell membrane but not for creating a constitutively active form (Fig. 3b). EGFP and DsRed in C35A/C290A receptors did not mainly overlap in the cytoplasm (Fig. 4d). Most C35A/C290A receptors did not form a homo-oligomer in the cytoplasm and mainly formed a monomer by immunoblotting (Fig. 4e). Several recent studies have identified novel intercellular proteins that interact with the C-terminal tails of GPCRs (25). AT 2 receptor-interacting protein has a novel coiled-coil domain that contains protein that interacts with the AT 2 receptor (26). This kind of protein may still form a homo-oligomer even if interaction between Cys 35 and Cys 290 is disrupted.
To reconcile the observations presented here with those reported earlier, we propose a model of constitutively active homo-oligomeric AT 2 receptor in the cell membrane (Fig. 4f). Inactive N127G receptor forms a homo-oligomer in the cell membrane but does not induce apoptosis. If Cys 35 in the AT 2 receptor binds to Cys 290 in another AT 2 receptor, C35A/C290A receptors do not form a homo-oligomer. Although the conformation of C35A/C290A receptor is similar to that of WT receptor, C35A/C290A mutant did not form a homo-oligomer in cytoplasm. Therefore, disulfide bonding in the extracellular loop in the WT receptor was disrupted by DTT treatment, and Cys 35 in the AT 2 receptor may establish some intermolecular contact such as a disulfide bond with Cys 290 in another AT 2 receptor. Other regions of intermolecular contact that have been proposed in other GPCRs include a region or regions within TM domains 1-3 from one receptor interacting with TM domains 1-3 from another for the V2 vasopressin receptor (27) and docking of helix 8 with the loop between helices 5 and 6 for rhodopsin (28). The relative arrangement of receptors and the FIG. 3. a, AT 2 -C35A-EGFP and AT 2 -C290A-EGFP but not AT 2 -C35A/C290A-EGFP receptors were localized in the cell membrane after 24 h under serum-free conditions. CHO cell lines that expressed AT 2 -C35A-EGFP and AT 2 -C290A-EGFP but not AT 2 -C35A/C290A-EGFP receptors were grown under serum or 24 h under serum-free conditions and stained with DAPI to visualize nuclear morphology by a laser scanning confocal microscope. b, an immunoblot analysis of CHO cell lines expressing AT 2 -C35A-EGFP or AT 2 -C290A-EGFP receptors showed the homo-oligomerization of AT 2 -WT-EGFP receptor. CHO cells expressing AT 2 -C35A-EGFP and AT 2 -C290A-EGFP receptors were subjected to SDS-gel electrophoresis under non-reducing conditions and immunoblotted with anti-EGFP antibody as described under "Experimental Procedures." 30 mg of protein was used in each lane. Arrows indicate the monomer and dimer species of the receptor. c, AT 2 -C35A-EGFP and AT 2 -C290A-EGFP receptors but not AT 2 -C35A/C290A-EGFP receptor induced apoptosis after 48 h under serum-free conditions. CHO cell lines were grown under serum conditions or for up to 48 h under serum-free conditions and then stained with DAPI and imaged by a digital fluorescent microscope. Data are shown as the percentage of apoptotic cells in three independent experiments as assessed by TUNEL as described under "Experimental Procedures." *, p Ͻ 0.05 versus serum conditions. Representative pictures (a and c) and immunoblots (b) are shown. Four independent determinations were performed, and similar results were observed (a-c).
orientation of helix 8 are also based on the crystalline array of squid rhodopsin and on the hypothesis that there is both a symmetrical dimer interface that involves TM domain 4 and an interaction that involves helix 8 with the third intracellular loop. Site-directed mutagenesis studies have indicated that two conserved extracellular Cys residues (Cys 140 and Cys 220 ), which are present in almost all GPCRs of the rhodopsin family, play key roles in the formation of disulfide-linked m3Ј receptor dimmers (8). Since the two conserved Cys residues are likely to form an intramolecular disulfide bond, their findings indicate that the two conserved extracellular Cys residues can also participate in the formation of intermolecular disulfide bonds. In the case of constitutively active AT 2 receptor, two conserved Cys residues form an intramolecular disulfide bond. Two other Cys residues (Cys 35 and Cys 290 ) may form intermolecular disulfide bonds independent of ligand stimulation and receptor conformation changes, and the homo-oligomer induces cell signaling. Although oligomerization affects GPCR signaling, other  (n ϭ 6). b, inhibition of 125 I-[Sar 1 ,Ile 8 ]Ang II-specific binding after treatment with 2.5 mM MTSEA ϩ . After the reaction of MTSEA ϩ with cell membranes of CHO cells that expressed AT 2 -WT-EGFP, AT 2 -C35A-EGFP, AT 2 -C290A-EGFP, and AT 2 -C35A/C290A-EGFP receptors for 1-10 min, the reaction mixture was diluted 120-fold, and membranes were centrifuged and suspended to carry out saturation binding analysis as described under "Experimental Procedures." Maximal specific binding to each control sample without MTSEA ϩ reagent is taken to be 100%. roles still need to be identified. This discovery should help us to elucidate what is likely to be a common underlying mechanism for oligomerization of constitutively active GPCRs and to obtain a better understanding of the structure and function of these receptors in cells.