Evidence that a deviation in the kynurenine pathway aggravates atherosclerotic disease in humans.

BACKGROUND: The metabolism of tryptophan (Trp) along the kynurenine pathway has been shown to carry strong immunoregulatory properties. Several experimental studies indicate that this pathway is a major regulator of vascular inflammation and influences atherogenesis. Knowledge of the role of this pathway in human atherosclerosis remains incomplete. OBJECTIVES: In this study, we performed a multiplatform analysis of tissue samples, in vitro and in vivo functional assays to elucidate the potential role of the kynurenine pathway in human atherosclerosis. METHODS AND RESULTS: Comparison of transcriptomic data from carotid plaques and control arteries revealed an upregulation of enzymes within the quinolinic branch of the kynurenine pathway in the disease state, whilst the branch leading to the formation of kynurenic acid (KynA) was downregulated. Further analyses indicated that local inflammatory responses are closely tied to the deviation of the kynurenine pathway in the vascular wall. Analysis of cerebrovascular symptomatic and asymptomatic carotid stenosis data showed that the downregulation of KynA branch enzymes and reduced KynA production were associated with an increased probability of patients to undergo surgery due to an unstable disease. In vitro, we showed that KynA-mediated signalling through aryl hydrocarbon receptor (AhR) is a major regulator of human macrophage activation. Using a mouse model of peritoneal inflammation, we showed that KynA inhibits leukocyte recruitment. CONCLUSIONS: We have found that a deviation in the kynurenine pathway is associated with an increased probability of developing symptomatic unstable atherosclerotic disease. Our study suggests that KynA-mediated signalling through AhR is an important mechanism involved in the regulation of vascular inflammation.

Constitutive activation of tethered-peptide/corticotropin-releasing factor receptor chimeras.

Constitutive activity, or ligand-independent activity, of mutant G protein-coupled receptors (GPCRs) has been described extensively and implicated in the pathology of many diseases. Using the corticotropin-releasing factor (CRF) receptor and the thrombin receptor as a model, we present a ligand-dependent constitutive activation of a GPCR. A chimera in which the N-terminal domain of the CRF receptor is replaced by the amino-terminal 16 residues of CRF displays significant levels of constitutive activation. The activity, as measured by intracellular levels of cAMP, is blocked in a dose-dependent manner by the nonpeptide antagonist antalarmin. These results support a propinquity effect in CRF receptor activation, in which the amino-terminal portion of the CRF peptide is presented to the body of the receptor in the proper proximity for activation. This form of ligand-dependent constitutive activation may be of general applicability for the creation of constitutively activated GPCRs that are regulated by peptide ligands such as CRF. These chimeras may prove useful in analyzing mechanisms of receptor regulation and in the structural analysis of ligand activated receptors.

Mutational analysis of the interaction of the N- and C-terminal ends of angiotensin II with the rat AT(1A) receptor.

1. The role of different residues of the rat AT(1A) receptor in the interaction with the N- and C-terminal ends of angiotensin II (AngII) was studied by determining ligand binding and production of inositol phosphates (IP) in COS-7 cells transiently expressing the following AT(1A) mutants: T88H, Y92H, G196I, G196W and D278E. 2. G196W and G196I retained significant binding and IP-production properties, indicating that bulky substituents in position 196 did not affect the interaction of AngII's C-terminal carboxyl with Lys(199) located three residues below. 3. Although the T88A mutation did not affect binding, the T88H mutant had greatly decreased affinity for AngII, suggesting that substitution of Thr(88) by His might hinder binding through an indirect effect. 4. The Y92H mutation caused loss of affinity for AngII that was much less pronounced than that reported for Y92A, indicating that His in that position can fulfil part of the requirements for binding. 5. Replacing Asp(278) by Glu caused a much smaller reduction in affinity than replacing it by Ala, indicating the importance of Asp's beta-carboxyl group for AngII binding. 6. Mutations in residues Thr(88), Tyr(92) and Asp(278) greatly reduced affinity for AngII but not for Sar(1) Leu(8)-AngII, suggesting unfavourable interactions between these residues and AngII's aspartic acid side-chain or N-terminal amino group, which might account for the proposed role of the N-terminal amino group of AngII in the agonist-induced desensitization (tachyphylaxis) of smooth muscles.

Binding of norbinaltorphimine (norBNI) congeners to wild-type and mutant mu and kappa opioid receptors: molecular recognition loci for the pharmacophore and address components of kappa antagonists.

Molecular modifications of both the kappa opioid antagonist norbinaltorphimine (norBNI, 1) and the kappa receptor have provided evidence that the selectivity of this ligand is conferred through ionic interaction if its N17' protonated amine group (an "address") with a nonconserved acidic residue (Glu297) on the kappa receptor. In the present study, we have examined the effect of structural modifications on the affinity of norBNI analogues for wild-type and mutant kappa and mu opioid receptors expressed in COS-7 cells. Compounds 2, 3, and 7, which have an antagonist pharmacophore and basic N17' group in common with norBNI, retained high affinity for the wild-type kappa but exhibited greatly reduced affinity for mutant kappa receptors (E297K and E297A). Modification of the phenolic or N-substituent groups of the antagonist pharmacophore (4 and 5) or removal of basicity at the address N17' center (6) led to greatly reduced affinity for the wild-type and mutant receptors. The reduced affinity upon modification of the kappa receptor is consistent with the ionic interaction of the protonated N17' group of kappa antagonists (1-3, 7) with the carboxylate group of E297 at the top of TM6. This was supported by the greatly enhanced affinity of compounds 1-3 for the mutant mu receptor (K303E), as compared to the wild-type mu receptor, given that residue K303 occupies a position equivalent to that of E297 in the kappa receptor. In view of the high degree of homology of the seven TM domains of the kappa and mu opioid receptors, it is suggested that the antagonist pharmacophore is bound within this highly conserved region of the kappa or mutant mu receptor and that an anionic residue at the top of TM6 (E297 or K303E, respectively) provides additional binding affinity.

Steric hindrance mutagenesis versus alanine scan in mapping of ligand binding sites in the tachykinin NK1 receptor.

Residues in transmembrane domain (TM)-III, TM-V, TM-VI, and TM-VII believed to be facing the deep part of the presumed main ligand-binding pocket of the NK1 receptor were probed by alanine substitution and introduction of residues with larger and/or chemically distinct side chains. Unaltered or even improved binding affinity for four peptide agonists, substance P, substance P-O-methyl ester, eledoisin, and neurokinin A, as well as normal EC50 values for substance P in stimulating phosphatidylinositol turnover indicated that these mutations did not alter the overall functional integrity of the receptor. The alanine substitutions in general had only minor effects on nonpeptide antagonist binding. However, the introduction of the larger and polar aspartic acid and histidine residues at positions corresponding to the monoamine binding aspartic acid in TM-III of the beta 2-adrenoceptor (ProIII:08, Pro112 in the NK1 receptor) and to the presumed monoamine interacting "two serines" in TM-V (ThrV:09, Thr201; and IleV:12, Ile204) impaired by > 100-fold the binding of a group of nonpeptide antagonists, including CP96,345, CP99,994, RP67,580, RPR100,893, and CAM4092. In contrast, another group of nonpeptide antagonists, LY303,870, FK888, and SR140,333, were little or not at all affected by the space-filling substitutions. Two of these compounds, FK888 and LY303,870, were those most seriously affected (75-89-fold) by alanine substitution of PheVI:20 located in the upper part of the main ligand-binding crevice. Surprisingly, substitution of AlaIII:11 (Ala115), which is located in the middle of TM-III, conceivably pointing toward TM-VII, with a larger valine residue increased the affinity for all 13 ligands tested, presumably by creating a closer interhelical packing. It is concluded that the introduction of larger side chains at positions at which molecular models indicate that this is structurally allowed can be a powerful method of locating ligand-binding sites due to the considerable difference between positive and negative results. Such steric hindrance mutagenesis strongly indicates that one population of nonpeptide antagonists bind in the deep pocket of the main ligand-binding crevice of the NK1 receptor, whereas another group of nonpeptide antagonists, especially SR140,333, was surprisingly resistant to mutational mapping in this pocket.

Constitutive activity of glucagon receptor mutants.

Increased constitutive activity has been observed in the PTH receptor in association with naturally occurring mutations of two residues that are conserved between members of the glucagon/vasoactive intestinal peptide/calcitonin 7TM receptor family. Here, the corresponding residues of the glucagon receptor, His178 and Thr352, were probed by mutagenesis. An elevated level of basal cAMP production was observed after the exchange of His178 into Arg, but not for the exchange into Lys, Ala, or Glu. However, for all of these His178 substitutions, an increased binding affinity for glucagon was observed [dissociation constant (Kd) ranging from 1.1-6.4 nM, wild type: Kd = 12.0 nM]. A further increase in cAMP production was observed for the [H178R] construct upon stimulation with glucagon, albeit the EC50 surprisingly was increased approximately 10-fold relative to the wild-type receptor. Substitution of Thr352, located at the intracellular end of transmembrane segment VI, with Ala led to a slightly elevated basal cAMP level, while the introduction of Pro or Ser at this position affected rather the binding affinity of glucagon or the EC50 for stimulation of cAMP production. The large extracellular segment, which is essential for glucagon binding, was not required for constitutive activation of the glucagon receptor as the introduction of the [H178R] mutation into an N-terminally truncated construct exhibited an elevated basal level of cAMP production. The analog des-His1-[Glu9]glucagon amide, which in vivo is a glucagon antagonist, was an agonist on both the wild-type and the [H178R] receptor and did not display any activity as an inverse agonist. It is concluded that the various phenotypes displayed by the constitutively active glucagon receptor mutants reflect the existence of multiple agonist-preferring receptor conformers, which include functionally active as well as inactive states. This view agrees with a recent multi-state extension of the ternary complex model for 7TM receptor activation.

Ligand binding pocket of the human somatostatin receptor 5: mutational analysis of the extracellular domains.

The ligand binding domain of G protein-coupled receptors for peptide ligands consists of a pocket formed by extracellular and transmembrane domain (TM) residues. In the case of somatostatin (SRIF), however, previous studies have suggested that the binding cavity of the octapeptide analog SMS201-995 (SMS) is lined by residues in TMs III-VII. The additional involvement of the extracellular domains for binding SMS or the natural SRIF ligands (SRIF-14, SRIF-28) has not been clarified. Using a cassette construct cDNA for the human somatostatin 5 receptor (sst5R), we systematically examined the role of exofacial structures in ligand binding by creating a series of mutants in which the extracellular portions have been altered by conservative segment exchange (CSE) mutagenesis for the extracellular loops (ECLs) and by deletion (for the NH2-terminal segment) or truncation analysis (ECL3). CHO-K1 cells were stably transfected with wild type or mutant human sst5R constructs, and agonist binding was assessed using membrane binding assays with 125I-LTT SRIF-28 ligand. Deletion of the NH2 terminus or CSE mutagenesis of ECL1 and ECL3 produced minor 2-8-fold decreases in affinity for SRIF-14, SRIF-28, and SMS ligands. Truncation of ECL3 to mimic the size of this loop in sst1R and sst4R (the two subtypes that do not bind SMS) did not interfere with the binding of SMS, SRIF-14, or SRIF-28. In contrast, both ECL2 mutants failed to bind 125I-LTT SRIF-28. Immunocytochemical analysis of nonpermeabilized cells with a human sst5R antibody revealed that the mutant receptors were targeted to the plasma membrane. Labeled SMS (125I-Tyr3 SMS) also failed to bind to the mutant ECL2 receptors. These results suggest a potential contribution of ECL2 (in addition to the previously identified residues in TMs III-VII) to the SRIF ligand binding pocket.

Dual agonistic and antagonistic property of nonpeptide angiotensin AT1 ligands: susceptibility to receptor mutations.

Two nonpeptide ligands that differ chemically by only a single methyl group but have agonistic (L-162,782) and antagonistic (L-162,389) properties in vivo were characterized on the cloned angiotensin AT1 receptor. Both compounds bound with high affinity (K(I) = 8 and 28 nM, respectively) to the AT1 receptor expressed transiently in COS-7 cells as determined in radioligand competition assays. L-162,782 acted as a powerful partial agonist, stimulating phosphatidylinositol turnover with a bell-shaped dose-response curve to 64% of the maximal level reached in response to angiotensin II. Surprisingly, L-162,389 also stimulated phosphatidylinositol turnover, albeit only to a small percentage of the angiotensin response. The prototype nonpeptide AT1 agonist L-162,313 gave a response of approximately 50%. The apparent EC50 values for all three compounds in stimulating phosphatidylinositol turnover were similar, approximately 30 nM, corresponding to their binding affinity. Each of the three compounds also acted as angiotensin antagonists, yet in this capacity the compounds differed markedly, with IC50 values ranging from 1.05 x 10(-7) M for L-162,389 to 6.5 x 10(-6) for L-162,782. A series of point mutations in the transmembrane segments (TMs) of the AT1 receptor had only minor effect on the binding affinity of the nonpeptide compounds, with the exception of A104V at the top of TM III, which selectively impaired the binding of L-162,782 and L-162,389. Substitutions in the middle of TM III, VI, or VII, which did not affect the binding affinity of the compounds, impaired or eliminated the agonistic efficacy of the nonpeptides but with only minor or no effect on the angiotensin potency or efficacy. Thus, in the N295D rat AT1 construct, L-162,782, L-162,313, and L-162,389 all antagonized the angiotensin-induced phosphatidylinositol turnover with surprisingly similar IC50 values (90-180 nM), and they all bound with unaltered, high affinity (22-36 nM). However, L-162,313 and L-162,782 could stimulate phosphatidylinositol turnover to only 20% of that of angiotensin. It is concluded that minor chemical modifications of either the compound or the receptor can dramatically alter the agonistic efficacy of biphenyl imidazole compounds on the AT1 receptor without affecting their affinity, as determined in binding assays, and that a number of substitutions in the middle of the TM segments affect the efficacy of nonpeptide agonists as opposed to angiotensin.

Engineering of metal-ion sites as distance constraints in structural and functional analysis of 7TM receptors.

G-protein-coupled receptors with their seven transmembrane (7TM) segments constitute the largest superfamily of proteins known. Unfortunately, still only relatively low resolution structures derived from electron cryo-microscopy analysis of 2D crystals are available for these proteins. We have used artificially designed Zn(II) metal-ion binding sites to probe 7TM receptors structurally and functionally and to define some basic distance constraints for molecular modeling. In this way, the relative helical rotation and vertical translocation of transmembrane helices TM-II, TM-III, TM-V, and TM-VI of the tachykinin NK-1 receptor have been restricted. Collectively, these zinc sites constitute a basic network of distance constraints that limit the degrees of freedom of the interhelical contact faces in molecular models of 7TM receptors. The construction of artificially designed metal-ion sites is discussed also in the context of probes for conformational changes occurring during receptor activation.

Radioligand-dependent discrepancy in agonist affinities enhanced by mutations in the kappa-opioid receptor.

A series of kappa/mu receptor chimeras and a number of kappa receptors substituted in the second transmembrane segment (TM-II) were investigated using as radioligands, respectively, the kappa-selective agonist [3H]C1977 and the nonselective opioid antagonist [3H]diprenorphine (DIP). All of the receptor constructs bound [3H]DIP with similar and high affinity, whereas the apparent affinity of the nonpeptide agonist C1977, when estimated in competition binding with the antagonist [3H]DIP, was impaired between 42- and > 500-fold in the kappa/mu chimeras and between 64- and 153-fold in three of the kappa receptor mutants that had been substituted in the TM-II segment. However, homologous competition binding experiments, using [3H]C1977 as radioligand, showed that the high affinity binding of this nonpeptide agonist was in fact not impaired in four of the kappa/mu chimeras and in three TM-II substituted kappa receptors compared with the wild-type kappa receptor. In all cases in which mutations decreased the apparent affinity of C1977 without affecting its actual affinity, as determined in homologous assays using [3H]C1977, the calculated number of receptor sites (Bmax) was decreased. In three of the kappa/mu constructs, binding of [3H]C1977 was undetectable, indicating that in these chimeras the affinity of the nonpeptide agonist had actually been affected. Also, for the kappa-selective peptide agonist dynorphin A(1-8), the measured affinity for the receptor mutants was strongly dependent on whether it was determined using the antagonist [3H]DIP or the agonist [3H]C1977 in that < or = 800-fold higher Ki values were determined in competition with the antagonist. It is concluded that mutations in the kappa-opioid receptor can cause large discrepancies between the affinity determined for agonists in homologous versus heterologous competition binding assays and that this pattern, which is compatible with a partial uncoupling of receptors, is observed in surprisingly many types of receptor mutations.

Construction of a high affinity zinc switch in the kappa-opioid receptor.

Very limited structural information is available concerning the superfamily of G-protein-coupled receptors with their seven-transmembrane segments. Recently a non-peptide antagonist site was structurally and functionally replaced by a metal ion site in the tachykinin NK-1 receptor. Here, this Zn(II) site is transferred to the kappa-opioid receptor by substituting two residues at the outer portion of transmembrane V (TM-V), Asp223 and Lys227, and one residue at the top of TM-VI, Ala298, with histidyl residues. The histidyl residues had no direct effect on the binding of either the non-peptide antagonist [3H]diprenorphine or the non-peptide agonist, [3H]CI977, just as these mutations/substitutions did not affect the apparent affinity of a series of other peptide and non-peptide ligands when tested in competition binding experiments. However, zinc ions in a dose-dependent manner prevented binding of both agonist and antagonist ligands with an apparent affinity for the metal ion, which gradually was built up to 10(-6) M. This represents an increase in affinity for the metal ion of about 1000-fold as compared with the wild-type kappa receptor and is specific for Zn(II) as the affinity for e.g. Cu(II) was almost unaffected. The direct transfer of this high affinity metal ion switch between two only distantly related receptors indicates a common overall arrangement of the seven-helix bundle among receptors of the rhodopsin family.

Analysis of selective binding epitopes for the kappa-opioid receptor antagonist nor-binaltorphimine.

The structural determinants for the selective binding of the nonpeptide opioid receptor antagonist nor-binaltorphimine (nor-BNI) to the kappa-opioid receptor were characterized using a systematic series of chimeras between the kappa receptor and the homologous mu-opioid receptor. All 10 chimeric constructs bound the nonselective antagonists (-)-naloxone and diprenorphine with similar affinities, as did the two wild-type receptors. Introduction of amino-terminal segments of increasing length, extending to and including transmembrane segment VI, from the mu receptor into the kappa receptor did not impair the high affinity binding of nor-BNI, and neither did introduction of the intracellular carboxyl-terminal extension of the mu receptor. In contrast, nor-BNI binding was impaired > or = 600-fold in constructs in which extracellular loop 3 and transmembrane segment VII originated from the mu receptor. The exchange of a single residue within this region, Glu297, for lysine, the corresponding residue from the mu receptor, reduced the binding affinity of nor-BNI 142-fold, without affecting the binding the nonselective compounds (-)-naloxone and diprenorphine. It is concluded that the selective binding of nor-BNI to the kappa-opioid receptor is determined by nonconserved residues located in extracellular loop 3 and transmembrane segment VII and that Glu297, located just outside transmembrane segment VI, plays a major role in the kappa-selective binding characteristics of nor-BNI.

Interaction between the nonpeptide angiotensin antagonist SKF-108,566 and histidine 256 (HisVI:16) of the angiotensin type 1 receptor.

His256 (HisVI:16) of transmembrane segment (TM)-VI of the rat angiotensin type 1 (AT1) receptor was targeted for mutagenesis to investigate its potential involvement in ligand binding. Substitution of His256 with alanine, phenylalanine, glutamine, or isoleucine did not affect the binding of either angiotensin II or nine different biphenylimidazole AT1 antagonists. In contrast, the binding affinity of the prototype imidazoleacrylic acid antagonist SKF-108,566 was reduced 15-fold by the exchange of His256 with alanine. Substitution of His256 with either isoleucine or phenylalanine yielded similar results, whereas a glutamine residue was able to substitute for His256, suggesting that the epsilon-nitrogen of His256 could be involved in the interaction with the imidazoleacrylic acid. To identify the chemical groups on SKF-108,566 that interact with His256 and with Asn295, a previously identified interaction point for nonpeptide antagonists located in TM-VII, we tested the binding of 15 analogs of SKF-108,566 in which different chemical moieties were systematically exchanged. The results indicated that the carboxyphenyl group of SKF-108,566 interacts with the imidazole side chain of His256. The data did not point to any particular contact group on the antagonist for Asn295. It is concluded that the imidazoleacrylic acid antagonists share some interactions in TM-VII of the AT1 receptor with the biphenylimidazole antagonists, but the binding of the imidazoleacrylic acid compounds is uniquely dependent on His256 in TM-VI, possibly through the carboxyphenyl moiety.

Cloning and expression in insect cells of two pancreatic lipases and a procolipase from Myocastor coypus.

The physiological role of pancreatic lipases has traditionally been assigned solely to triacylglyceride metabolism, while the digestion of phospholipids requires the presence of the pancreatic phospholipase A2, a 14-kDa enzyme unrelated to pancreatic lipases. However, in the guinea pig, it was observed that the pancreatic phospholipase A2 was absent and that a guinea pig pancreatic-lipase-related protein 2 (GPL-RP2) was responsible for phospholipase activity, in contrast to the situation observed in other mammalian species. As the guinea pig is a member of the hystricomorph rodents, it was of interest to investigate if other species within this evolutionary suborder display similar characteristics. The coypu (Myocastor coypus) also a member of the hystricomorph rodents, was chosen for further investigations. The cDNAs encoding two pancreatic lipases and a procolipase from the coypu were cloned, expressed and characterized. One lipase, CoPL-RP2, was identified as belonging to the RP2 subfamily, while the second, CoPL, was found to belong to the classical pancreatic lipase subfamily. Enzymic characterization and sequence data suggest a role for coypu colipase as a specific cofactor for CoPL, while this coypu colipase cannot be an important cofactor for CoPL-RP2 in vivo. Also, the new lipase cDNA sequences were used in a phylogentic analysis to reinvestigate the taxonomical position of the hystricomorph rodents (e.g. coypu and guinea pig) with respect to the myomorph rodents (e.g. rat and mouse).

Non-peptide angiotensin agonist. Functional and molecular interaction with the AT1 receptor.

Non-peptide ligands for peptide receptors for the G-protein-coupled type are generally antagonists, except in the opiate system. Recently, it was observed that a subset of biphenylimidazole derivatives surprisingly possessed angiotensin-like activity in vivo. In COS-7 cells transfected with the rat AT1 receptor a prototype of these compounds, L-162,313 stimulated phosphoinositide hydrolysis with an EC50 of 33 +/- 11 nM. The maximal response to the compound was 50% of that of angiotensin II in COS-7 cells but only 3% in stably transfected Chinese hamster ovary cells. The agonistic effect of L-162,313 was blocked by the AT1-specific antagonist L-158,809 and was not observed in untransfected cells. In Chinese hamster ovary cells, L-162,313 also acted as an insurmountable antagonist of the angiotensin stimulated phosphoinositide hydrolysis. In contrast to previously tested non-peptide ligands, L-162,313 bound with reasonably high affinity to the Xenopus laevis AT1 receptor. In the human receptor, the binding of L-162,313 was found to be unaffected by point mutations in transmembrane segments III and VII, which impaired the binding of biphenylimidazole antagonists. Substitutions in the extracellular domains of the human and rat receptor, which impaired the binding of angiotensin II, did not affect the binding of L-162,313. It is concluded that a subset of biphenylimidazole compounds can act as high affinity partial agonists on the AT1 receptor. These compounds have molecular interactions with the receptor which appear to differ both from that of the structurally similar non-peptide antagonists and from that of their functional counterpart, the peptide agonist.

Identification of peptide binding residues in the extracellular domains of the AT1 receptor.

To locate essential determinants for angiotensin II binding, we have performed a systematic mutational analysis of the exterior domain of the AT1 receptor. Receptor mutants, deficient in peptide binding, were analyzed using radiolabeled nonpeptide ligand as an important tool. Two independent strategies for mutagenesis were employed: conservative segment exchange and point mutagenesis of evolutionarily conserved residues. Results from the conservative segment exchange in which 6-17 residues were replaced with chemically similar, yet different, amino acid sequences of the same length suggested that important peptide ligand binding epitopes are located in the N-terminal extension of the AT1 receptor, in particular adjacent to the top of transmembrane segment I (TM-I), and in the third extracellular loop, close to the top of TM-VII. The substitution of residues from either of these regions resulted in a 5,000-20,000-fold decrease in affinity for the peptide agonist angiotensin II (AII) and the peptide antagonist [Sar1,Leu8]AII without affecting the binding of nonpeptide antagonists. Alanine substitution of evolutionarily conserved residues demonstrated that peptide binding was dependent on several residues in the N-terminal extension, near the top of TM-I, a tyrosine residue located in extracellular loop 1, close to TM-II, and 2 aspartate residues positioned in extracellular loop 3 on the same face of an alpha-helical extension of TM-VII. In all cases the binding of nonpeptide antagonist was unaffected by these substitutions. It is concluded that important epitopes involved in angiotensin II binding are located around the top of transmembrane segments I, II, and VII which conceivably are in close spatial proximity in the folded receptor structure.