Role of Transmembrane Helix IV in G-protein Specificity of the Angiotensin II Type 1 Receptor*

G-protein activation by G-protein coupled receptors (GPCRs) is accomplished through proper interaction with the cytoplasmic loops rather than through sequence-specific interactions. However, the mechanism by which a specific G-protein is selected by a GPCR is not known. In the current model of GPCR activation, agonist binding modulates helix-helix interactions, which is necessary for fully determining G-protein specificity and stimulation of GDP/GTP exchange. In this study, we report that a single-residue deletion in transmembrane helix IV leads the angiotensin II type 1 (AT1) receptor chimera CR17 to retain GTP-sensitive high affinity for the agonist angiotensin II but results in complete inactivation of intracellular inositol phosphate production. The agonist dissociation profile of CR17 in the presence of guanosine 5′-3-O-(thio)triphosphate suggests that the activation-induced conformational changes of the chimeric receptor itself remain intact. Insertion of an alanine at position 149 (CR17▿149A) in this chimera rescued the inactive phenotype, restoring intracellular inositol phosphate production by the chimera. This finding suggests that in the wild-type AT1 receptor the orientation of transmembrane helix IV-residues following Cys149 is a key determinant for effectively distinguishing among various structurally similar G-proteins. The results emphasize that the contacts within the membrane-embedded portion of transmembrane helix IV in the AT1 receptor is important for specific G-protein selection.

The octapeptide hormone angiotensin II (Ang II) 1 (DRVYI-HPF) mediates diverse biological functions through multiple signals (1)(2)(3)(4). Two distinct integral membrane receptors, type 1 (AT 1 ) and type 2 (AT 2 ), belonging to the G-protein-coupled receptor (GPCR) superfamily, mediate the intracellular response to Ang II. The predicted secondary structures of AT 1 and AT 2 are characterized by seven transmembrane ␣-helical segments (TM-I through TM-VII), three connecting loops on either side of the membrane, an extracellular N terminus, and a cytoplasmic C terminus. The extracellular loops and the transmembrane domain are involved in binding Ang II, whereas the cytoplasmic loops form the site for binding the G-protein (guanine nucleotide-binding regulatory protein) upon activation. The AT 1 receptor is a prototypical GPCR, mediating most known functions of Ang II, and the mechanism for the functions of the AT 1 receptor are relatively well studied (3)(4)(5)(7)(8)(9)(10). The AT 2 receptor is believed to transduce cell growth inhibitory signals and may regulate apoptosis during embryonic development and tissue remodeling in the adult (11)(12)(13)(14)(15). The molecular basis for the AT 2 receptor function is largely unknown. Notably, the binding of the agonist Ang II to the AT 2 receptor is insensitive to analogues of GTP, and second-messenger assays for detection of AT 2 receptor activation have been unsuccessful in various recombinantly expressed surrogate cell systems (10 -14).
Analysis of the primary structures reveals a 32% amino acid sequence homology between AT 1 and AT 2 receptors (11,12), but several chimeras of AT 1 and AT 2 receptors (constructed to elucidate the molecular basis of the subtype-specific functions of the Ang II receptor) were fully active. However, a chimera in which AT 1 receptor residues 99 -359 have been replaced with a topologically identical segment from the AT 2 receptor was inactive, whereas a chimera with residues 165-359 replaced with a segment from the AT 2 receptor was fully active (Fig. 1). Members of the GPCR family are believed to have the same basic molecular architecture. Therefore, one can construct chimeras that retain distinct subtype-specific functions or exhibit functions of both parent subtypes simultaneously or that display unique defects. The study of chimeric receptors helps us understand the signal transduction mechanism. To gain potentially important insight into the mechanism of G-protein recruitment in Ang II signaling, we analyzed several chimeras containing portions of AT 2 receptor segments. The results indicate that the integrity of TM-IV of the AT 1 receptor is critical for specific G-protein coupling. Cloning, Mutagenesis, and Expression of Ang II Receptors-The synthetic rat AT 1 receptor gene, cloned in the shuttle expression vector pMT3, was used for expression as described earlier (5,(7)(8)(9)(10)16). The cDNA of the AT 2 receptor was cloned from mRNA isolated from adrenal medulla of spontaneously hypertensive rat by reverse transcriptasecoupled polymerase chain reaction. The cloned cDNA was fully sequenced and modified for expression in COS1 cells (1) contain a consensus Marilyn-Kozak sequence and a unique EcoRI site at the 5Ј end and NotI site at the 3Ј end of the gene (2) and to encode an octapeptide (ETSQVAPA) epitope tag for a monoclonal antibody 1D4 at the 3Ј end before the stop codon. The epitope-tagged cDNA was subcloned into a shuttle vector pMT3 in which the cDNA would be transcribed from a polyoma major late promoter. Mutant Ang II receptors were prepared by the restriction fragment replacement method and the polymerase chain reaction method. DNA sequence analysis was performed to confirm each mutant or receptor chimera. For expression of receptor proteins, 10 g of CsCl purified plasmid DNA per 10 7 cells was used in transfection. COS1 cells (American Type Culture Collection), cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, were transfected by the DEAE-dextran method. Transfected COS1 cells cultured for 72 h were harvested. Cell membranes were prepared by the nitrogen PARR bomb disruption method and suspended in hypotonic buffer (HME) (50 mM HEPES, pH 7.2, 12.5 mM MgCl 2 , 1.5 mM EGTA) containing 10% glycerol as described earlier (5,(7)(8)(9)(10). The receptor expression was assessed in each case by immunoblot analysis (data not included) and by 125 I-[Sar 1 ,Ile 8 ]Ang II saturation binding analysis.

Materials-The
Ligand Binding Study-The ligand binding experiments were carried out under equilibrium conditions as described before (5,(7)(8)(9)(10) GTP␥S-dependent Affinity Analysis-To measure the competitive dissociation of 125 I-Ang II from the receptors in the presence and absence of G-protein activation receptor bound radioligand was challenged with cold Ang II or a mixture of cold Ang II and GTP␥S. To ensure that the 125 I-Ang II binding reaction reached equilibrium, the radioligand and membrane were mixed for 30 min at 22°C. After this time, either cold Ang II or a mixture of cold Ang II and GTP␥S (100 M) was added, and the incubation was allowed to continue for another 60 min at 22°C. At the end of the incubation period, the membranes were collected by filtration and washed, and the bound radioactivity was determined. In a typical experiment, the total 125 I-Ang II per tube was 330,000 cpm, of which specific binding without any competitor and GTP␥S added was 28,000 cpm and the nonspecific binding in the presence of 10 Ϫ4 M Ang II was 960 cpm. To measure the effect of GTP␥S on Ang II binding, membrane samples (10 -15 g of protein) were incubated with 300 pM 125 I-Ang II in the presence and absence of varying concentrations of GTP␥S. Dissociation of the radioligand was initiated by adding 10 Ϫ6 M cold 127 I-Ang II. The dissociation constants were calculated by nonlinear regression using single or double exponential decay equations.
Production of Inositol Phosphate (IP)-Transfected COS1 cells were cultured in 60-mm Petri dishes for 24 h after transfection, then labeled for 24 h with [ 3 H]myoinositol (1.5 Ci/Petri dish), specific activity 22 Ci/mmol (Amersham Pharmacia Biotech), at 37°C in Dulbecco's modi-fied Eagle's medium containing 10% bovine calf serum. At 48 h after transfection, the labeled cells were washed with serum-free medium three times and incubated with Dulbecco's modified Eagle's medium containing 10 mM LiCl for 20 min; agonists were added, and incubation was continued for another 45 min at 37°C. At the end of incubation, the medium was removed, and total soluble IP was extracted from the cells by the perchloric acid extraction method, as described previously (8 -11). The amount of [ 3 H]IP eluted from the column was counted, and a concentration response curve generated using iterative nonlinear regression analysis (see Refs. 8 -10 and 15 for additional details). The results of IP production were examined by unpaired Student's t test, and p values less than 0.05 were considered significant.

Ang II Receptors in COS1 Cell
Model System-Recombinant expression in transiently transfected COS1 cells was employed for analysis of the structure-function relationship of the wildtype and chimeric AT 1 or AT 2 receptors as described earlier (5,(7)(8)(9)(10). Expression, in each case, was measured by immunoblotting with the C-terminal epitope-directed monoclonal antibody 1D4 (data not shown), followed by 125 I-[Sar 1 ,Ile 8 ]Ang II binding and competition binding to subtype-specific antagonists (Table  I) and then by second-messenger production to measure function. The affinity of the expressed AT 1 and AT 2 receptors were 0.3 Ϯ 0.09 and 0.21 Ϯ 0.05 nM, respectively, for the nonselective peptide antagonist 125 I-[Sar 1 ,Ile 8 ]Ang II, and 11.8 Ϯ 1.7 and 17.2 Ϯ 6.4 nM, respectively, for losartan and PD123319. The K d values of the AT 1 receptor for the agonists [Sar 1 ]Ang II and Ang II were 0.33 Ϯ 0.07 and 3.1 Ϯ 0.33 nM, respectively. The B max values estimated for the two receptors were very similar (ϳ5 pmol/mg). Scatchard plot analysis indicated a single affinity class for both receptors. Competition binding studies employing Ang II, [Sar 1 ]Ang II, losartan, and PD123319 demonstrated that the receptors expressed in COS1 cells preserve the selectivity and affinity profiles previously described for native tissue receptors as well as recombinantly expressed receptors. The protein expression of all the mutants described in this report was within 25% of the level of the expression of the wild-type receptor ( Table I). The variation in cell surface receptor number, estimated by acid labile binding of 125 I-[Sar 1 ,Ile 8 ]Ang II in intact cells, was 1.3-1.7 ϫ 10 5 sites per cell in this study.
The ability of AT 1 receptor to activate IP production in COS1 cells has been described before by our laboratory (5,(7)(8)(9)(10). The basal IP production in transfected COS1 cells without [Sar 1 ]Ang II treatment is 5 Ϯ 0.5% (ϳ4000 cpm) when compared with the maximal IP response elicited by [Sar 1 ]Ang II concentration Ͼ10 7 M (taken as 100%). This value is significantly higher than that measured in the mock-transfected cells (Fig. 2). Alterations in the functional activity of different mutant receptors could be accurately measured in the COS1 cells, because the maximal IP response elicited by the AT 1 receptor is  1 and AT 2 receptors to elucidate the molecular basis of the Ang II receptor subtypespecific functions. A chimera containing the AT 1 receptor in which residues 99 -359 had been replaced with a topologically identical segment from the AT 2 receptor was inactive, whereas a chimera with residues 165 through 359 replaced with that of AT 2 receptor was fully active. Both of these chimeras express well and bind Ang II with high affinity. 2 Analysis of chimeras CR5, CR7, CR8, and CR17 ( Fig. 1) suggests that the dominant loss of function is linked to the TM-IV segment derived from the AT 2 receptor. These chimeras were constructed to individually evaluate the role of TM-III, TM-IV, and the interhelical cytoplasmic loop connecting the two helices that include residues 99 -165 of the AT 1 receptor. The expression levels of the chimeras were comparable to that of the wild-type AT 1 receptor (Table I). The affinity of [Sar 1 Ile 8 ]Ang II was slightly reduced, and the affinity of losartan was reduced more than 100,000-fold in all of the chimeras. The affinity of PD123319 varied between chimeras.
Chimeras CR5 and CR7 were inactive (Fig. 2), suggesting that neither the second cytoplasmic segment (CR5) nor the TM-III and the second cytoplasmic segment (CR7) from the AT 1 receptor could rescue the function. The chimera CR8, which contains TM-III from the AT 2 receptor but the second cytoplasmic loop and TM-IV segment of the AT 1 receptor, demonstrated increased (12%) basal activity and was fully activated by [Sar 1 ]Ang II. The putative mechanism of constitutive activity of CR8 has been fully described previously (9). Restoration of function in CR8 (which contains TM-IV and the adjacent loop segment of the AT 1 receptor) implies that the defect in CR5 and CR7 function is related to the TM-IV segment derived from the AT 2 receptor.
This conclusion is further supported by results from the CR17 chimera, in which the TM-IV region in the AT 1 receptor (Lys 146 -Ile 166 shown within the box in Fig. 1A) was replaced by a topologically identical 19-residue segment (Tyr 162 -Tyr 180 ) of the AT 2 receptor. Although chimera CR17 expressed well and its ligand affinity profile is comparable to that of the wild-type receptor (Table I), it did not activate IP production upon Ang II stimulation (Fig. 2). Because a larger segment exchange (i.e. residues 166 -359) between the two receptors yielded functional chimeras, the lack of function in CR17 indicates incompatibility of TM-IV helices in the two Ang II receptors.
The inability of CR17 to stimulate IP production could be caused by the lack of Ang II-induced isomerization to the active state or by the inability of the activated form of the receptor to interact with G-proteins. To distinguish between these two possibilities, G-protein interaction with receptor was evaluated by measuring the dissociation rate of 125 I-Ang II from wild-type and CR17 chimeric receptors in the presence and absence of the nonhydrolyzable GTP analogue GTP␥S (Fig. 3A). In the absence of added GTP␥S, the higher affinity of the 125 I-Ang II in the ternary agonist-receptor-G-protein complex is indicated from 6 -10-fold higher molar concentration of the cold Ang II at which the radioligand binding is inhibited. The magnitude of the GTP-induced affinity shift is more pronounced with the CR17 chimera than with the wild-type receptor, perhaps suggesting that different G-proteins are responsible for the shift observed in each. This suggestion needs further confirmation.
In the presence of the added GTP␥S, the dissociation of 125 I-Ang II is faster from both wild-type and CR17 chimeric recep- tors, suggesting the existence of a low affinity binary agonistreceptor complex. This implies that CR17 chimera is capable of stimulating GDP/GTP exchange in G-protein bound to the receptor. The GTP␥S-dependent destabilization of the ternary complex implies that the binding and dissociation of G-protein from the agonist-activated receptor is intact. Fig. 3B shows that the dissociation kinetics exhibited by the CR17 chimera is comparable to that exhibited by the wild-type receptor, suggesting that the CR17 receptor interacted productively with a G-protein, even though the IP formation was completely defective. These observations suggest that the ligand activated CR17 chimeric receptor couples to a G-protein from a family other than the phospholipase C (PLC) activating G q family of G-proteins. Pertussis toxin treatment did not unmask PLC activation in CR17-transfected cells. 3 Exposure to pertussis toxin, which uncouples receptors from G i /G o , the abundant pool of G-proteins in the cell, is expected to unmask nonspecifically coupled receptors, as has been shown previously (16 -19). The lack of an increase in IP formation from this treatment indicates that G i /G o did not block coupling to G q and that the G-protein specificity of CR17 is not relaxed but altered. However, there was no alteration in the intracellular concentrations of cAMP, cGMP, calcium, and arachidonic acid in CR17transfected cells following [Sar 1 ]Ang II stimulation. 2 Thus, the identity of the G-protein coupled to CR17 and of the downstream signal activated is currently unknown.
Because the CR17 chimera contains all the intracellular loops of the AT 1 receptor presumed essential for G-protein interaction, it is able to couple to a G-protein other than G q . We speculate that this defect must be the result of disruption of the G-protein selection mechanism in the CR17 chimera. The homology between the two TM-IV segments of the AT 1 and AT 2 receptors is 40% (8 out of 20 residues conserved), which is greater than the overall homology (32%) between the two receptors (Fig. 4A). Among the eight conserved residues, Trp 153 , Ala 156 , and Pro 162 are present in Ͼ90% of the members of the GPCR superfamily (see Fig. 4 for conservation of residues) (20). Anchoring these residues, sequence alignment reveals, perhaps most strikingly, a single-residue deletion at position Cys 149 upon segment replacement (Fig. 4A). Rather interestingly, this is the only deletion within all seven TM domains when the primary sequence of two receptors is aligned. The remaining 3 S. Chang and S. S. Karnik, unpublished observations. changes resulting from replacement include substantial differences in residue size due to the Ile 150 3 Pro, Leu 154 3 Cys, Gly 157 3 Cys, Ala 159 3 Ser, and Ala 163 3 Thr alterations; in charge due to the Lys 146 3 Tyr alteration; and in aromaticity due to the Lys 146 3 Tyr, Val 164 3 Phe, and Ile 165 3 Tyr alterations. These differences are expected to perturb helical packing. Because the peptide agonist specificity of CR17 is not substantially different from that of the wild-type AT 1 receptor, these residue changes are unlikely to be responsible for inducing the functional defect through perturbation in the ligandexposed phase (Fig. 4B). Finally, it is clear that the changes observed in CR17 do not represent a global unfolding of the protein compared with the wild-type.
Insertion of an Ala at Position 149 of CR17 Fully Rescued the Function-To examine whether the predicted deletion of a single amino acid residue was responsible for the inactivation of CR17 in IP production, the CR17ƒ149A mutant was prepared by insertion of an Ala at position 149 of CR17. The CR17ƒ149A mutant receptor activated IP production as much as did the wild-type AT 1 receptor (Fig. 2). The ligand affinity profiles for the mutant were identical to that of CR17 and were distinguishable from the wild-type AT 1 receptor. As shown in Fig.  3B, the GTP␥S-induced dissociation of 125 I-Ang II from the CR17ƒ149A chimera closely resembled the profile of wild-type AT 1 receptors. Insertion of an Ala at the identical position in the CR5 (CR5ƒ149A) and CR7 (CR7ƒ149A) chimeras fully rescued Ang II-dependent IP production (Fig. 2). In each in-stance, the ligand affinity profiles resembled that of the inactive chimera (Table I). Strikingly, the CR5ƒ149A chimeric receptor exhibited partial constitutive IP production, comparable to that exhibited in the CR8 chimera, consistent with our previous findings (9).
Thus, the defective phenotype resulting from the singleresidue deletion in TM-IV is dominant over the constitutive activation induced by substitution of the TM-III of the AT 1 receptor with that from the AT 2 receptor. Rescue of function in CR5, CR7, and CR17 with Ala insertion indicates that these chimeric receptors all share the same molecular defect. Furthermore, the results demonstrate that Cys 149 does not play a specific role in G-protein selection because an Ala substitution also leads to restoration of function. Moreover, Cys is not a highly conserved residue at this position.

DISCUSSION
Activation of the AT 1 receptor, as in other GPCRs, must involve conformational changes that result from agonist binding within the TM domain of the receptor. The nature of the conformational changes associated with AT 1 receptor activation is not known at the present time, but several lines of evidence suggest the occurrence of rather well characterized events similar to those in prototypical GPCRs such as visual rhodopsin and ␤-adrenergic receptor. For instance, the high affinity state for agonists is believed to result from the formation of a ternary ligand-receptor/G-protein complex, which is FIG. 4. A, secondary structure cartoon of the CR17 chimera and the primary sequence of the substituted AT 2 receptor segment aligned with the sequence of the AT 1 receptor (within the box). The residues conserved in Ͼ90% of GPCRs are indicated by vertical lines (also see Ref. 20). The gap in the AT 2 receptor-derived sequence, corresponding to Cys 149 of the AT 1 receptor (*), is proposed to be a single-residue deletion leading to the defect in IP formation in the CR17 chimera. B, linearized projection map of TM-IV and its cytoplasmic extension (shaded light gray) for the wild-type AT 1 , CR17ƒ149A, and CR17 receptors. The primary sequence was transformed to a helical projection map, assuming 3.6 residues per turn (20). Linearization was done for a 33-residue segment, Pro 133 -Ile 165 , of the wild-type AT 1 receptor. The proposed alignment for the CR17ƒ149A and CR17 chimeric receptors is based on their phenotype and assumes that the ␣-helical structure is maintained, forcing the position of the highly conserved residues (indicated in boldface circles) to match that of the wild-type receptor. The net effect is displacement of residue positions (indicated by boldface letters) beyond Val 149 in CR17 relative to CR17ƒ149A. The G-protein-interacting phase of TM-IV cytoplasmic helical extension is not defined at the present time for any receptor. The putative ligand-exposed phase of the TM-IV helix is shown in white, and phases exposed to the lipid/transmembrane portion of the protein are shown in dark gray.
abolished by guanine nucleotides that promote G-protein activation and dissociation of G-protein subunits from the receptorligand binary complex. Uncoupling of the receptor from Gprotein prevents development of both the ternary complex and the GTP␥S-sensitive high affinity state (1)(2)(3)(4)21).
In this study, the functionally inactive AT 1 /AT 2 receptor chimeras formed the ternary receptor-ligand/G-protein complex that demonstrated GTP␥S-sensitive dissociation, suggesting that defective IP formation in these chimeras is not because of their inability to adopt an activated conformation or to trigger G-protein activation. Instead, the ability to selectively couple to the G q -PLC pathway upon Ang II-activation is lost. A different G-protein that coupled to the chimeras selectively dissociated upon GTP␥S treatment. This observation per se is not novel because several mutagenesis studies of GPCRs have provided examples of relaxed specificity-simultaneous coupling to different G-proteins-by chimeric and mutant receptors, whereas the wild-type receptors exhibit considerable specificity (1-4, 16 -19, 21). However, the G-proteins coupled to the AT 1 /AT 2 receptor chimeras were not the pertussis toxin-inactivated G i /G o proteins, which by their abundance potentially could couple through mass action and block coupling to the G q -PLC pathway. Furthermore, all three defective chimeras harbored the same defect, which could be overcome by the insertion of a single amino acid in the transmembrane region of TM-IV outside the putative G-protein contact region in all three defective chimeras. These novel and unique characteristics imply that TM-IV of the AT 1 receptor has a role in controlling G-protein specificity.
The third cytoplasmic loop and the C-terminal tail of the AT 1 receptor have been shown to be essential for coupling to the G q -PLC pathway (3,22,23). But the phenotype of CR5ƒ149A, a chimera that binds the AT 2 receptor-selective antagonist PD123319 and activates AT 1 receptor-specific function, indicates that G-protein specificity is not controlled by these two cytoplasmic regions of the receptor. Instead, the observation suggests that the third cytoplasmic loop and the C-terminal tail of the AT 2 receptor could also support G q -PLC coupling quite efficiently. All chimeras activated by insertion of Ala 149 contain the second loop and the cytoplasmic extension of TM-IV of the AT 1 receptor beyond the predicted site of deletion, indicating that this region is crucial for G-protein selection. The G-protein-binding specificity of GPCRs is not controlled by discrete consensus sequences in the cytoplasmic loop. Rather, concerted interaction of multiple cytoplasmic segments is required, within a specific and stringent context, for a GPCR to effectively distinguish between various structurally similar G-proteins and productively couple to a specific one. Taken together, the results presented here indicate that the cytoplasmic region of TM-IV could be responsible for controlling G-protein specificity of the wild-type AT 1 receptor. This is likely a general mechanism in the GPCR family in view of the characteristics of TM-IV discussed below.
How did the G q -PLC uncoupling occur when the particular TM-IV segment (residues 146 -166) of the AT 1 receptor was replaced by the topologically equivalent TM-IV segment (residues 162-180) of the AT 2 receptor? The mechanism is not clear in light of the plasticity displayed by TM segments in GPCRs that allow their exchange, sometimes even between distant members, without significantly compromising the function. The mechanism suggested below is based on recent observations in bovine rhodopsin that suggest that rigid body movements of individual TM ␣-helices upon light activation (or agonist binding) result in distal changes in structures at the cytoplasmic surface (24, 25). Altenbach et al. (24) determined the topography of the intracellular loop-2 region for bovine rhodopsin. The membrane/aqueous boundary is 3 residues from the conserved Arg in TM-III and 10 residues from the conserved Trp in TM-IV. Baldwin (20) has proposed a model for packing of ␣-helices in rhodopsin, which is consistent with the 9 Å electron diffraction density map (6) and the site-directed electron paramagnetic resonance mapping studies. This model is based on helical packing, taking into account the location of highly conserved residues and residue polarity in each of the TM-helices leading to identification of putative surfaces of tertiary interactions. Extending Baldwin's modeling approach to the AT 1 receptor, 2 the assumption that length and packing arrangements of helices are the same as in rhodopsin allows us to predict that the TM-IV helical region has a cytoplasmic extension of nearly six residues involved in tertiary interaction with neighboring helices (Fig. 4B). The model predicts that Lys 146 faces TM-V and perhaps is directly involved in tertiary interaction with TM-V. Electron paramagnetic resonance analysis in rhodopsin indicates that the movement in the TM-IV region in the activated state is limited to Ͻ5 Å (there is no vertical displacement with respect to other helices) and that the scope for disengagement of the cytoplasmic extension from the G-protein-interacting surface or reorganization of secondary structure is restricted. This is because TM-IV is unique in that it has no polar-accommodating site in the middle of the hydrophobic core; it has a large lipid-exposed surface with no polar residues that must limit its plasticity. Perhaps due to the lack of flexibility and accommodation of structure, all residues subsequent to the site of deletion (Cys 149 ) in TM-IV would be back-rotated by 100 degrees (Fig. 4B). This anomaly will result in the presence of a new secondary arrangement of residues in the putative G-protein-exposed TM-IV helix/intracellular loop 2 region, which could explain coupling to a different G-protein in the chimeras. This unique defect, which is overcome by a single-residue insertion, implies that the secondary arrangement of residue in the cytoplasmic extension of TM-IV provides a unique interaction surface for G-proteins, which must be preserved for the purpose of specific G-protein selection. The propagation of structural changes toward the ligand pocket and extracellular region may be far less because the agonist affinity profile is only minimally affected, although perturbation in antagonist specificity clearly suggests at least some structural change (Table I). Thus, the rigidity of TM-IV appears to be a key regulatory factor in receptor/G-protein selection.
In summary, substitution in the TM-IV helical segment of the AT 1 receptor with a topologically equivalent segment from the AT 2 receptor results in defective G q -PLC coupling. The defect is consistent with deletion of a single residue, Cys 149 , in the TM-IV of the AT 1 receptor. The functional defect caused by deletion of a single residue in TM-IV does not affect agonist interaction or receptor activation but does affect specificity of coupled G-protein. The phenotypes of defective and the cognate reverent chimeras obtained through predetermined insertion of a single residue can be rationalized in terms of displacement of specific residues in an ␣-helical structure. The observations reported here suggest that helical movements are remarkably conserved among subfamilies of GPCRs.