The C-terminal Third Intracellular Loop of the Rat AT1A Angiotensin Receptor Plays a Key Role in G Protein Coupling Specificity and Transduction of the Mitogenic Signal*

To identify the role(s) of the third intracellular loop of the angiotensin II (AngII) type 1A (AT1A) receptor in G protein coupling specificity and receptor activation, several chimerae were constructed and characterized. The cDNA sequence encoding the C-terminal segment of the third intracellular loop of the AT1A receptor (residues 234–240) was replaced with the homologous regions of the α1B adrenergic (α1B-AR), the β2 adrenergic (β2-AR), and the AngII type 2 (AT2) receptors. These chimeric receptors were stably expressed in Chinese hamster ovary cells, and their pharmacological and functional properties were characterized, including AngII-induced inositol phosphate and cyclic AMP (cAMP) productions, [3H]thymidine incorporation into DNA, and internalization. The affinities of these chimeric receptors for [Sar1]AngII, [Sar1,Ile8]AngII, and losartan were essentially normal; however, the affinity of these mutants was increased by a factor of 10–40 for the AT2-specific ligand CGP42112A. The functional properties of the α1B-AR chimera were essentially identical to those of the wild type AT1A receptor. On the other hand, replacement with the β2-AR segment produced a partial reduction of the inositol phosphate production, a measurable AngII-induced cAMP accumulation, a reduced internalization, and a total impairment to transduce the mitogenic effect of AngII. The AT2 chimera presented a normal internalization, but was inactive in all the other functional tests. In conclusion, the distal segment of the third intracellular loop of the rat AT1A receptor plays a pivotal role in coupling selectivity and receptor signaling via G protein(s) as well as in the activation of the specific signaling pathways involved in the mitogenic actions of AngII.

To identify the role(s) of the third intracellular loop of the angiotensin II (AngII) type 1A (AT 1A ) receptor in G protein coupling specificity and receptor activation, several chimerae were constructed and characterized. The cDNA sequence encoding the C-terminal segment of the third intracellular loop of the AT 1A receptor (residues 234 -240) was replaced with the homologous regions of the ␣ 1B adrenergic (␣ 1B -AR), the ␤ 2 adrenergic (␤ 2 -AR), and the AngII type 2 (AT 2 ) receptors. These chimeric receptors were stably expressed in Chinese hamster ovary cells, and their pharmacological and functional properties were characterized, including

AngII-induced inositol phosphate and cyclic AMP (cAMP) productions, [ 3 H]thymidine incorporation into DNA, and internalization. The affinities of these chimeric receptors for [Sar 1 ]AngII, [Sar 1 ,Ile 8 ]AngII
, and losartan were essentially normal; however, the affinity of these mutants was increased by a factor of 10 -40 for the AT 2 -specific ligand CGP42112A. The functional properties of the ␣ 1B -AR chimera were essentially identical to those of the wild type AT 1A receptor. On the other hand, replacement with the ␤ 2 -AR segment produced a partial reduction of the inositol phosphate production, a measurable AngII-induced cAMP accumulation, a reduced internalization, and a total impairment to transduce the mitogenic effect of AngII. The AT 2 chimera presented a normal internalization, but was inactive in all the other functional tests. In conclusion, the distal segment of the third intracellular loop of the rat AT 1A receptor plays a pivotal role in coupling selectivity and receptor signaling via G protein(s) as well as in the activation of the specific signaling pathways involved in the mitogenic actions of AngII.
The vasoactive peptide angiotensin II (AngII) 1 acts on its target tissues via membrane-bound receptors of the G protein-coupled receptor family. On the basis of pharmacological, functional, and structural features, these receptors have been divided into two classes: AT 1 and AT 2 (for review, see Refs. 1 and 2). Both types of AngII receptors comprise a heptahelical structure and bind AngII with a high affinity (nanomolar range). The AT 1 receptors are characterized by a high affinity for nonpeptidic imidazolic compounds (Dup753), by a coupling to several G proteins, but mainly to a G protein of the G q/11 family that activates phospholipase C-␤ (PLC-␤) and, finally, by its ability to transduce most if not all of the physiological actions of AngII. In contrast, the AT 2 receptors display a high affinity for derivatives of spinacine (PD123177) or pseudopeptidic AngII analogs (CGP42112A), and their coupling to a G protein or a functional pathway remains questionable (3,4).
The AT 1 receptor cDNA has been cloned from several species (5)(6)(7)(8) and reveals a strong conservation of the amino acid sequence among mammals and the existence of two subtypes in rodents, called AT 1A and AT 1B . Mutagenesis of the rat and human AT 1 cDNAs or synthetic peptides competing with intracellular loops of the AT 1 receptors was used to study the involvement of various intracellular sequences or amino acid(s) in the receptor coupling to G i and/or G q/11 proteins. These reports all indicate the crucial role of the third intracellular (i3) loop in G q coupling (9 -12).
The role of the i3 loop in G protein coupling specificity has been investigated extensively for many heptahelical receptors, including adrenergic, muscarinic, and dopaminergic receptors (see, e.g., Refs. [13][14][15][16]. All these reports have identified amino acids and sequences of the proximal (Ni3) and distal (Ci3) parts of this third intracellular loop involved in G protein interactions and specificities. Moreover, modifications in the Ci3 segment are responsible for constitutive activation of several adrenergic receptors (AR) (17)(18)(19). This concept of constitutively activated receptor has been extended (i) to other classes of G protein-coupled receptors, such as thyroid-stimulating hormone or luteinizing hormone receptors, which contain activating mutations of the i3 loop (20,21); (ii) to other regions of the G protein-coupled receptor (21,22); and finally (iii) to the pathology, because several acquired or genetic diseases correspond to a constitutive activation of receptors, due to somatic or germinal mutations of the corresponding genes (23). However, there are very few activating mutations reported for heptahelical receptors binding to peptidic ligands, such as the AT 1 receptor (24).
Finally, accumulating and converging data seem to indicate that the G protein coupling differences between AT 1 and AT 2 receptors are the consequence of sequence differences in the i3 loop. Initially, the absence of activation or inhibition of either PLC or adenylate cyclase and the absence of modification of AngII affinity in the presence of GTP analogs suggested that the AT 2 receptor was not coupled to G proteins (3). Since then, it has been demonstrated that the AT 2 receptor could be coupled to G proteins at least in some tissues; Zhang and Pratt (4) have recently demonstrated that the AT 2 receptor is coimmunoprecipitated with G i antibodies in rat fetal tissues. Moreover, Buisson et al. (25) have demonstrated that AngII is able to inhibit T type calcium currents via a G protein coupled to AT 2 in NG108-15 cells. Moreover, the production of chimerae between the AT 1 and AT 2 receptors seems to indicate that the functional differences between these two receptors lie in the proximal and distal parts of the i3 loop (11).
Taken together, these data point out the major role of the i3 loop and especially the distal part of this loop in the signal transduction process. To understand more clearly the role of this segment, we have constructed several chimeric mutants of the Ci3 region of the rat AT 1A receptor. The functional comparison of these mutants and the wild type rat AT 1A receptor was performed to answer several questions. (i) Is this sequence involved in the selective binding properties of AngII receptors? (ii) Are these mutations able to produce constitutively activated AT 1 receptors? (iii) Is this segment responsible for the difference of G protein coupling and signaling between AT 1 and AT 2 AngII receptors? (iv) Is this segment involved in the mitogenic effect of AngII (DNA synthesis and mitogen-activated protein (MAP) kinase activation)? (v) Is this segment involved in internalization of the AT 1 receptor?

EXPERIMENTAL PROCEDURES
Construction of Mutated Receptor cDNAs-Expression plasmids coding for the mutated receptors were obtained using the previously described synthetic rat AT 1A cDNA containing multiple unique restriction sites (26), subcloned into the eucaryotic expression vector pECE (27). For the three chimeric mutants, residues 234 -240 of AT 1A receptor were substituted for residues 288 -294 of the hamster ␣ 1B -AR, residues 267-273 of the human ␤ 2 -AR, or residues 250 -256 of the rat AT 2 receptor (Fig. 1). This was achieved by replacing the KspI-KpnI fragment that corresponds to nucleotides 706 -768 of the synthetic AT 1A cDNA with synthetic oligonucleotide adapters. The identity of each mutant was confirmed by dideoxy sequencing with Sequenase version 2 (U. S. Biochemical Corp.).
Cell Culture and Transfections-CHO K1 cells were obtained from ATCC (catalog no. CCL61) and were cultured in Ham's F-12 medium supplemented with 10% fetal calf serum, 0.5 mM glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin (all from Boehringer Mannheim). To establish pure cell lines expressing the different mutants, CHO cells were co-transfected with 10 g of the plasmids and 2 g of the selection marker pSV2neo, using the calcium phosphate precipitation method (28). Transfected cells were selected for their resistance to 750 g/ml G418 (Life Technologies, Inc.) and cloned by limiting dilution. COS-7 cells were grown in Dulbecco's modified Eagle's medium (Boehringer Mannheim) supplemented with 10% fetal calf serum, 0.5 mM glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. Cells (3 ϫ 10 6 ) were plated in 75-cm 2 flasks and transfected by the DEAE-Dextran method using 15 g of DNA/dish. They were immediately transferred to 24- Ci/mmol; 1 Ci ϭ 37 GBq), was purified by high performance liquid chromatography. Saturation and competitive binding assays were performed as described (29). Cells were subcultured into 24-well culture trays and incubated for 45 min at 22°C with various concentrations of 125 I-[Sar 1 ]AngII or competitive ligands in 50 mM Tris-HCl, 6.5 mM MgCl 2 , 125 mM NaCl, 1 mM EDTA, 1 mg/ml bovine serum albumin, pH 7.6. Each experiment was carried out in duplicate. Binding data were analyzed with a nonlinear least-squares curve fitting procedure, using Ebda-Ligand software (Elsevier-Biosoft, Cambridge, United Kingdom).
Inositol Phosphate Determination-Stimulation by AngII of the inositol phosphate (IP) production was performed as described previously (30). Cells were subcultured in 12-well culture dishes and labeled for 20 -24 h with 2 Ci/ml myo-[ 3 H]inositol (Amersham). After 30 min of stimulation with increasing concentrations of AngII in the presence of 10 mM LiCl, IP was extracted and separated on Dowex AG1-X8 (Bio-Rad) columns. Total IP was eluted with 1 M ammonium formate, 0.1 M formic acid.
Cytosolic Calcium Measurement-Variations of intracellular calcium levels were measured by dual emission microfluorometry using the fluorescent dye Indo1-AM (Sigma) as described previously (31). The variations of the fluorescent ratio (measured at 405 and 480 nm) reflect the variations of the intracellular calcium concentrations in response to 10 Ϫ8 M AngII. Experiments were carried out at room temperature in a solution containing 5.5 mM glucose, 145 mM NaCl, 5.5 mM KCl, 0.9 mM MgCl 2 , 1.1 mM CaCl 2 , and 20 mM HEPES, pH 7.4.
cAMP Measurement-Cells were cultured in 24-well culture dishes until confluence. After two washes, cells were preincubated for 10 min at 37°C in standard phosphate-buffered saline solution supplemented with 1 mM CaCl 2 , 1 mM MgCl 2 , 1 mM Hepes, 0.1% bovine serum albumin, 0.01% bacitracin, 10 Ϫ6 M bestatin, 5 ϫ 10 Ϫ6 M indomethacin, pH 7.4. Cells were then incubated with increasing concentrations of AngII in the same buffer, for 4 min at 37°C. The incubation solution was removed, and 0.5 ml of 5% formic acid in ethanol was added in each well for 10 min. This fraction was kept and allowed to evaporate overnight. The cAMP content of each sample was determined by enzymo-immunoassay and the results expressed in picomoles/million cells/4 min. Each determination was performed in duplicate.
MAP Kinase Assay-CHO cells were grown to confluence in 6-well culture dishes and incubated overnight in serum free medium. AngII (0 and 100 nM) was added for 10 or 180 min at 37°C and the cells were washed three times with ice-cold phosphate-buffered saline. Solubilization, immunoprecipitation, and myelin basic protein phosphorylation were performed as described previously (32).

AngII-induced [ 3 H]Thymidine
Incorporation into DNA-Cells, in 12well culture dishes, were incubated for 48 h in starvation medium (50% Dulbecco's modified Eagle's medium, 50% Ham's F-12, 0.5 mM glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, nonessential amino acids (Life Technologies, Inc.), and transferrin (Sigma). Increasing concentrations of AngII were added to this medium for 16 h. Cells were then labeled with 1 Ci/ml [ 3 H]thymidine for 45 min at 37°C. The DNA was then precipitated with 10% trichloracetic acid, and the radioactive material was dissolved in 1 M NaOH and counted. Each determination was performed in triplicate.
Internalization Assay-Internalization of wild type and mutant AT 1A was measured using the biochemical acid washing procedure. Transfected cells were incubated with 0.4 nM 125 I-[Sar 1 ]AngII in binding buffer with or without 1 M [Sar 1 ]AngII for 180 min at 4°C, washed twice, and placed in binding buffer alone at 37°C for various times. Finally, cells were placed at 4°C and processed as described previously (33).
Statistics-Results are expressed as mean Ϯ S.E. Statistical significance was assessed by Student's t-test.

RESULTS
The third cytoplasmic loop of many G protein-coupled receptors has been shown to be implicated in the efficiency and the specificity of receptor coupling to G proteins. To assess whether or not this region was important for angiotensin II receptor transduction, different chimeric or insertion mutants were constructed (Fig. 1). In the AT 1A receptor, which is mainly coupled to phospholipase C via a G q protein, seven residues (234 -240) of the Ci3 segment were substituted with the homologous regions of either the ␣ 1B -AR (residues 288 -294, mutant C-␣ 1 ), sharing the same major coupling to G q protein, or the ␤ 2 -AR (residues 267-273, mutant C-␤ 2 ), which interacts with G s protein, or the AT 2 receptor (residues 250 -256, mutant C-AT 2 ).
Binding Properties of the Different Chimeric Receptors-The mutants were stably expressed in CHO cells and binding characteristics of the different receptors were determined on pure clonal cell lines. As shown in Table I, binding characteristics of the different compounds for all the mutants, except the CGP42112A were in the range of affinities classically described for the AT 1A receptor. The substitution of the residues 234 -240 of AT 1A by any homologous sequence allowed the receptor to bind the AT 2 -specific antagonist CGP42112A with a significant and at least a 10-fold higher affinity compared with the wild type AT 1A receptor.
Absence of Constitutive Activation of the Chimeric Receptors-Mutations in the Ci3 segment of bioamine and thyroid-stimulating hormone receptors have been shown to cause their constitutive activation. Therefore, to look for a possible effect of the substitutions of this region of the AT 1A receptor, the different constructions were transiently transfected in COS cells and the IP accumulation was measured with or without a 100 nM AngII stimulation (see Fig. 2). Relative AngII-induced IP production via the different mutants were comparable to that obtained with the pure CHO cell lines, as discussed in the next section. Basal levels of IP production (in absence of agonist) were compared with the IP content of cells transfected with the expression vector pECE alone. To verify the validity of the procedure, COS cells were transfected with cDNAs encoding either the wild type or a mutated ␣ 1B -AR containing a A293K substitution responsible for the constitutive activation of this receptor (18). Under basal conditions, this oncogenic ␣ 1B -AR induced a 327% stimulation of IP accumulation whereas all other receptors had no detectable effect. Therefore, under these conditions, none of the AT 1A mutants have acquired the ability  of coupling to this transduction pathway, in the absence of AngII.

Role of the i3 Loop Distal Segment in G Protein
Coupling-To assess the role of this intracellular region in the AngII receptor coupling specificity and efficiency, different functional tests were assayed on the stable CHO clones. The ability of these receptors to activate phospholipase C was investigated first. It has been reported that the amplitude of the IP production is linearly correlated with the number of binding sites (34). Therefore, for a quantitative analysis, the results were expressed as the ratio IP production/B max (Fig. 3). The doseresponse curve obtained for C-␣ 1 was comparable to that of the wild type receptor, with 250% and 211% maximal stimulation over basal, respectively. The AngII concentrations required for half-maximal response (EC 50 ) were 1.20 nM for the wild type receptor and 3.53 nM for C-␣ 1 . The C-␤ 2 chimera displayed an intermediary profile, inasmuch as, via this receptor, AngII induced a maximal stimulation of IP production of 103% over basal, with an EC 50 of 0.65 nM. Finally, the substitution with the AT 2 residues abolished this effect.
To analyze another signaling pathway, the cAMP accumulation in response to increasing doses of AngII was measured in the different pure CHO cell lines (Fig. 4). Neither the wild type AT 1A nor the C-␣ 1 and C-AT 2 mutants could mediate an increase of the intracellular cAMP concentration. The substitution of residues 234 -240 of AT 1A by the homologous region of the ␤ 2 -AR resulted in the mutant C-␤ 2 to undergo an AngII dose-dependent accumulation of cAMP (E max ϭ 7.82 pmol/million cells/4 min, EC 50 ϭ 5.64 nM).
Role of the i3 Loop Distal Segment on DNA Synthesis-The effect of AngII on DNA synthesis in the CHO cell lines expressing the different mutant receptors was examined. Fig. 5 shows the AngII-induced [ 3 H]thymidine incorporation into DNA mediated by the mutants compared with the wild type receptor. The C-␣ 1 mutant was as potent as the wild type receptor in transducing the mitogenic effect of AngII, with a maximal stimulation of 177% over basal versus 214% for AT 1A , and an EC 50 of 0.83 nM versus 0.94 nM. The initiation of the DNA synthesis signal in the C-AT 2 mutant was abolished completely. Surprisingly, the C-␤ 2 mutant also displayed a total absence of stimulation of [ 3 H]thymidine incorporation into DNA. This result has been confirmed with another independent CHO C-␤ 2 cell line, to exclude the possibility of a clonal effect (data not shown). This impairment is not likely to be due to the decreased ability of C-␤ 2 to stimulate phospholipase C because, under the same conditions, other mutants, which do not stimulate IP production more than the C-␤ 2 mutant, transduce the mitogenic action of AngII (data not shown). However, due to its reduced ability to stimulate IP production, the C-␤ 2 mutant could have been unable to elicit a transient increase of intracellular calcium concentration, in response to AngII. Therefore, the AngII-induced intracellular calcium mobilization observed in cells expressing the mutant was compared with that measured in a CHO cell line expressing the wild type receptor with a similar density of sites, using the fluorescent probe Indo1. As shown in Fig. 6b, the C-␤ 2 mutant mediates a measurable increase of [Ca 2ϩ ] i in response to a short (15 s) application of 10 Ϫ8 M AngII. The amplitude of this mobilization is, however, slightly reduced compared with that monitored for the wild type AT 1A receptor (Fig. 6a), in agreement with the AngII-induced IP production observed for these receptors. Therefore, the absence of mitogenic response to AngII in the cells expressing the C-␤ 2 is not the result of an incapacity of this receptor to mobilize intracellular calcium. Because the C-␤ 2 mutant increases cAMP production, the effect of the cAMP analog 8-bromo-cAMP at various doses (10 Ϫ12 to 10 Ϫ8 M) was tested on the AngII-induced thymidine incorporation. The AngII-induced mitogenic response in CHO AT 1A cells was comparable with or without the cAMP analog (data not shown). This demonstrates that the presence of low doses of cAMP in the cells does not seem to cause an inhibition of the AngII mitogenic effect. The absence of transduction of the mitogenic signal by the C-␤ 2 mutant justifies the analysis of the MAP kinase activation by AngII in these cells.
In CHO AT 1A cells, AngII induces a biphasic activation of the MAP kinases with a rapid early peak (5 min), followed by a sustained activation (3 h). This delayed MAP kinase activity was shown to be associated with the AngII-induced mitogenic response (35). The MAP kinase activity induced by 10-or 180min treatment with 100 nM AngII of the CHO C-␤ 2 and CHO AT 1A cells was analyzed. As shown in Fig. 7, the C-␤ 2 mutant is at least as potent as the wild type receptor in mediating MAP kinase activation. After 180 min of stimulation by AngII, the activity was still 2.84-fold higher than in unstimulated cells. Therefore, the absence of mitogenic actions of AngII via the C-␤ 2 mutant is not due to an impairment in coupling to this signaling pathway. In conclusion, the distal segment of the i3 loop is important for the signaling of the mitogenic effect of AngII. A signaling pathway independent of the MAP kinase activation is potentially involved in this effect, but its nature remains unknown.
Role of the i3 Loop Distal Segment in Internalization-Finally, the time course of ligand-induced internalization of the different mutated receptors was studied. The patterns of AngII-and [Sar 1 ]AngII-induced internalization of the AT 1A receptor are similar (data not shown). As shown in Fig. 8, [Sar 1 ]AngII binding to AT 1A , C-␣ 1 , and C-AT 2 triggered a maximal internalization of the ligand-receptor complexes that reached 79%, 70%, and 73%, respectively, of total specific binding within 15 min (t max ). The times for half-maximal internalization were also comparable (t1 ⁄2 : 4.17, 3.21, and 5.18 min, respectively). Under the same conditions, the ligand-induced internalization of the last mutant, C-␤ 2 , was reduced to a maximum of 33.91% of total specific binding, but with normal kinetic characteristics (t max ϭ 15 min, t1 ⁄2 ϭ 5.30 min). DISCUSSION In the present report, the role of the distal part of the third intracellular loop on AT 1A receptor activation was analyzed by producing several mutant receptors and characterizing their signaling pathways and functional features.
The first observation is that all the modifications of the Ci3 loop reported in the present paper are associated with a better affinity for the AT 2 -specific agonist CGP42112A. This rather surprising observation is not isolated, inasmuch as several previous reports indicate that mutations altering the binding site for nonpeptidic AT 1 analogs or the receptor activation improve the binding affinity for agonist CGP42112A (29,34,36). The fact that various mutations in very different parts of the AT 1 receptor are able to result in similar modifications of the affinity for this compound seems to indicate that nonspecific structural constraints of the wild type receptor are responsible for the low affinity of the agonist CGP42112A for the AT 1 receptor and that these constraints are relaxed by the mutations. However, no clear molecular explanation can be proposed for this observation.
In addition, none of the modifications of the Ci3 segment of the AT 1 receptor result in its constitutive activation, in contrast to observations for the adrenergic receptors (17,18,37). Recently, a mutation in the third transmembrane domain (N111A) of the AT 1 receptor was reported to be a constitutively activated mutation (24). This observation suggests that the active and inactive states depend on different structural motifs from one G protein-coupled receptor to another.
The Ci3 Segment Is Crucial for G Protein Coupling-The AT 1A receptor is coupled to G q/11 protein(s) that activate(s) a PLC-␤, which produces both inositol 1,4,5-trisphosphate, which mobilizes the intracellular calcium stores, and 1,2-diacylglycerol, which activates a set of protein kinases C. Several investigators have tried to delineate the sequences involved in the coupling of the AT 1 receptors to G q/11 proteins. Using sitedirected mutagenesis, AT 1 /AT 2 chimerae, or other approaches, it has been possible to unambiguously localize the sequences involved in G q/11 protein coupling in the i3 loop and more precisely in its proximal and distal segments (10,11). The possibility that the i2 loop of AT 1 and the proximal segment of the C-terminal tail contain sequences involved in the coupling to G q/11 protein or, more probably, G i protein is more controversial (9,38). However, like for many other G protein-coupled receptors, it seems that the acquisition of the active conformation of the AT 1 receptor results from the general arrangement of numerous discrete sequences. The m1 and m3 muscarinic receptors and the ␣ 1B -AR were the principal models to study the coupling to G q proteins. For these receptors, the proximal and distal parts of i3 contain either sequences rich in basic residues or hydrophobic sequences forming amphipathic ␣ helices, which both appear to be the major determinants of G protein coupling specificity. All the data agree with a model in which the recognition site for the G protein is a discontinuous structure composed of several segments of the receptor, some of them being masked in the basal state and being unmasked by ligand binding (39).
In this report, the modifications performed in the Ci3 segment of the AT 1A receptor can differentially modify the ability of the receptor to activate various signaling pathways. The first evidence of this fact was given by the study of the second messengers IP and cAMP. The replacement of the Ci3 segment of AT 1A by the corresponding region of the ␣ 1B -AR, which also activates a PLC via the coupling to a G q/11 protein, does not affect the signaling properties of the AT 1 receptor, which still transduces several effects of AngII. However, the replacement of the Ci3 segment of AT 1A by the homologous sequence of the AT 2 receptor results in the abolition of the coupling to a G q/11 protein and PLC. The i3 loop of the AT 2 receptor was demonstrated recently as important for the coupling to a G i protein and the transduction of an antiproliferative signal (40). The ability of the C-AT 2 chimera to couple to G i protein and transduce an antiproliferative signal remains to be tested. Finally, the sequence of the ␤ 2 -AR homologous to the Ci3 segment of AT 1A has been shown to participate in the coupling of this receptor to the G s protein (13,41). Its presence in the context of the AT 1A receptor results in a loss of selectivity of the active conformation of the receptor. Indeed, this receptor is still activated in response to AngII binding, and the active conformation of this receptor is less efficient in PLC stimulation but gains the ability to stimulate a cAMP accumulation in response to AngII. This observation is concordant with results obtained by Wong et al. (42), who have shown that the substitution of limited sequences of either the m1 or the m2 muscarinic receptor with the homologous sequence of the ␤ 2 -AR resulted in receptors that could not discriminate between various G proteins.
The Ci3 Segment and the Mitogenic Action of AngII-This study demonstrates that the conformational modifications produced by the substitution of the Ci3 segment of AT 1A with the ␤ 2 -AR sequence results in the selective abolition of the transmission of the mitogenic signal of AngII. This impairment does not seem to be related to the loss of selectivity of this mutant or to an absence of stimulation of the MAP kinases. Another hypothesis can be postulated; the signaling pathway activating the Janus kinases and the signaling transducers and activators of transcription (JAK/STAT pathway) is involved in the proliferative response to various growth factors, and it has been shown that in vascular smooth muscle cells and in CHO cells transfected with the recombinant AT 1A receptor, AngII can stimulate this pathway (43,44). Moreover, in vascular smooth muscle cells, the tyrosine kinase JAK2 seems to physically interact with the AT 1A receptor (43). Therefore, it might be postulated that the C-␤ 2 mutant is unable to interact with JAK2. Anyhow, these data demonstrate that the activation of the signaling pathway leading to the proliferative response to AngII depends on a conformational state of the AT 1A receptor, distinct from that required for the activation of other second messenger systems, which is selectively impaired by the substitution with the ␤ 2 -AR residues in the Ci3 segment of the receptor.
Recently, mutations in the third exoplasmic loop of the luteinizing hormone/choriogonadotropin receptor have been shown to abolish preferentially cAMP induction, being more permissive for the IP signaling (45). Moreover, a mutation in the third transmembrane domain of the ␣ 1 -AR produced a constitutive activation of the IP signaling, without affecting the phospholipase A 2 induction (46). Therefore, for these two receptors, it appears that the mechanisms underlying the transduction of their different signals could be highly independent. Most of the studies identifying residues involved in the activation mechanism of the AT 1A receptor have been performed almost exclusively in terms of PLC stimulation. Therefore, it would be interesting to delineate the extent of the divergence of the conformational requirements for various signaling pathways of the receptor. The Ci3 Segment and Internalization-The pivotal role of the Ci3 segment in the distinct conformational changes induced by the agonist binding to the AT 1A receptor is strengthened by the analysis of the internalization pattern of the different mutants. The independence of the G q coupling and internalization of the AT 1 receptor has been demonstrated (10,33). Here, we demonstrate that depending on the mutation performed in the Ci3 segment of the AT 1A receptor, it is possible either to inactivate the receptor without modifying its ligand-induced internalization (mutant C-AT 2 ) or to abolish its internalization without causing a drastic impairment of its G q coupling ability (mutant C-␤ 2 ). Thus, this demonstrates that part of the Ci3 segment of the AT 1A receptor can be independently involved in both coupling and internalization.
Coupling Differences between AT 1 and AT 2 Receptors-Finally, the G protein coupling differences between the AT 1 and AT 2 receptors are the consequence of sequences differing in the i3 loop. The present results clearly confirm those published recently using AT 1 /AT 2 chimerae showing that the replacement of the i3 loop of AT 1 by that of AT 2 suppresses the coupling to G q protein, whereas the replacement of the i3 loop of AT 2 by that of AT 1 restores the coupling to G q protein (11). More refined analysis of the sequences of the i3 loop involved in this coupling has identified two sequences: an Ni3 basic sequence ( 219 WKALKKA 225 ), which reduces the binding to G q protein by 70%, and a Ci3 sequence ( 232 KPRN 235 ), which overlaps the sequence replaced in this work and reduces the G q coupling by 50% (11). The replacement of both the Ni3 and Ci3 sequences by those of the AT 2 receptor results in an uncoupled receptor, indicating that the effects of these two sequences are additive. Interestingly, the present work shows that, if the replacement of the Ci3 segment is extended further in the C-terminal region of the loop, the coupling to G q is totally abolished in the absence of any modification of the Ni3 segment.
In conclusion, this study demonstrates that the activation of different signaling pathways by the AT 1A receptor have distinct conformational requirements. The C-terminal segment of the third intracellular loop of this receptor plays a crucial role in the acquisition of these conformations, inasmuch as, depending on the mutation performed in this region, it is possible either to totally inactivate the receptor or to modify the selectivity of the coupling and to selectively impair the ability of the receptor to transduce a precise effect of AngII. Finally, and in contrast to bioamine receptors, these different modifications of the Ci3 segment do not produce a constitutively activated receptor, suggesting that perhaps the transition state between the nonactivated and the activated forms of the receptor requires more energy than for the adrenergic receptors.