INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

THE MANY BIOLOGICAL ACTIONS of kinins in mammalian species are mediated by the activation of two subtypes of kinin receptors, B₁ and B₂, belonging to the large family of G protein-coupled receptors (32). Typically, kinin receptors couple to G proteins of the G_q subtype, which trigger phospholipase C-mediated pathways signaling through inositol 1,4,5-trisphosphate [Ins(1,4,5)P₃] and Ca²⁺ (29, 30) before they are silenced by sequestration and phosphorylation (8, 21). B₁ receptors (B₁R) are stimulated by carboxyterminally truncated kinins such as [des-Arg⁹]bradykinin and [des-Arg¹⁰]kallidin and inhibited by [des-Arg⁹-Leu⁸]bradykinin. In contrast, the B₂ receptors (B₂R) are activated by the full-length peptides bradykinin and Lys-bradykinin (kallidin) and inhibited by kinin analogs having a D-aromatic residue in position 7, as exemplified by HOE 140 (4). Birds have yet another subtype of kinin receptor, namely the ornithokinin receptor (B_oR; also termed OKR), which couples to the G_q subtype (34). B_oR is stimulated by the avian kinin homologue, i.e., [Thr⁶,Leu⁸]bradykinin ("ornithokinin") but is unresponsive to bradykinin or [des-Arg⁹]bradykinin. Unexpectedly, HOE 140, a widely used B₂R antagonist, was found to be a full agonist for B_oR (34). Depending on the species origin of the kinin receptors, HOE 140 has been characterized as an inverse agonist (23, 28), neutral antagonist (14, 37), partial agonist (10, 11, 17), or full agonist (8, 34). Furthermore, HOE 140 was reported to act as an inverse agonist or partial agonist on constitutively active kinin receptors (10, 37). Thus a single ligand HOE 140 appears to display a spectrum of activities ranging from inverse and partial to full agonism, depending on the cell system expressing the kinin receptor. At present, the molecular mechanisms causing differential efficacy of HOE 140 for the various kinin receptor (sub)types are still being debated.

Drug efficacy denotes the capacity of a ligand L to activate or to inactivate a receptor response (19, 24, 36). The allosteric ternary complex model for G protein-coupled receptors (33) explains differences in ligand efficacy by the spontaneous isomerization of a given receptor between a basal, inactive conformational state (R) and an active conformational state (R*) that associates with a G protein to form R* G, which triggers an intracellular response (18). Any perturbation in the equilibrium between R and R* will modulate receptor activity (25). In this model, ligand efficacy critically depends on the capacity of the ligand to stabilize R* within the ternary complex of L · R* · G thereby shifting the equilibrium between R and R* toward the active, G protein-coupled receptor conformation (6, 25, 33).

Here we have addressed the molecular mechanisms underlying differential efficacy of the peptidic ligand HOE 140 for kinin receptors. Using wild-type and chimeric kinin receptors expressed in Chinese hamster ovary (CHO) or COS-7 cells, we demonstrate that the structural variability in the amino-terminal domains translates into differential ligand affinity for the corresponding inactive receptor conformers and thus into grossly varying ligand efficacy for the different kinin receptor types.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Construction of chimeric receptors. The cDNAs of avian B_oR and human B₂R (B₂R) were cloned into the EcoRI site of pBluescript (pBS, from Stratagene). A single HpaI restriction site was introduced into both receptor cDNAs using the QuickChange Site-directed Mutagenesis Kit (Stratagene) with the following primers: B_oR-HpaI-CT 5'-CTATGTAAAGCTGTTAACACAATCAA-CTAC-3'; B_oR-HpaI-NT 5'-GTAGTTGATTGTGTTAACAGCTTTACATAG-3'; B₂R-HpaI-CT 5'-CTCTGCCGCGTGGTTAACGCCATTATCTCC-3'; and B₂R-HpaI-NT 5'-GGAGATAATGGCGTTAACCACGCGGCCAGAG-3'.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Structure of chimeric kinin receptors. Predicted protein structures of chimeric kinin receptors CKR-1 and -2 are depicted. Columns correspond to the transmembrane regions. A: extracellular domains. B: intracellular domains. HpaI and AvrII cleavage sites and the corresponding flanking residues of the protein structures are shown. Numbers denote the relative positions of the residues in the receptor sequences. Domains derived from chick B0R (human B2R) are shown in black (gray).

Cell culture. The pED4 vectors harboring the relevant cDNAs were transfected into dihydrofolate reductase-deficient (DHFR) CHO cells using lipofectamine (Life Technologies). After 24 h, the transfected cells were split 1:10 into medium containing 250 nM methotrexate. Two weeks later transfectants were isolated by single cell cloning. Receptor expression was probed by specific binding using [3-(4- hydroxyphenyl-propyl)¹²⁵I]HOE 140 (kindly provided by Hoechst, Frankfurt, Germany) or [2,3-prolyl-2,4-³H(N)]bradykinin (DuPont); alternatively ligand-promoted total inositol phosphate (InsP_n) accumulation was measured to monitor receptor expression. COS-7 cells were grown in DMEM supplemented with 10% fetal calf serum and transfected using the diethylaminoethyl-dextran method (34).

Determination of inositol phosphates. Inosital phosphate hydrolysis was determined by the original method (2) as subsequently modified (15). Clonal CHO cells were grown to confluence in 24-well plates and labeled with myo[2-³H]inositol (1 µCi/ml) for 12 h in serum-free medium. After incubation with 10 mM LiCl for 15 min, the cells were stimulated for 10 min at 37°C with various ligands at the indicated concentrations. The released inositol phosphates were purified by anion exchange chromatography (Dowex 1×8) and quantified by liquid scintillation counting. CHO clones were stimulated with aluminum tetrafluoride AlF as described (31).

Ligand binding studies. Radioligand binding assays were done as described (26) with minor modifications. Briefly, cells expressing kinin receptors were grown on culture plates and rinsed three times with 0.15 M NaCl and 0.1 M phosphate, pH 7.4 (PBS). Washed cells were scraped off and collected by centrifugation. Membrane fractions were prepared and incubated at 4°C overnight in 500 µl DMEM to reach equilibrium conditions; DMEM was supplemented with 0.5% bovine serum albumin containing 1-100 nM [³H]bradykinin (>50 Ci/mmol) or 50 pM [3-(4-hydroxyphenyl-propyl)¹²⁵I]HOE 140 (2,100 Ci/mmol), subsequently designated [¹²⁵I]HOE 140, and competing ligands at the indicated concentrations. Bound ligand was separated from free ligand by filtration through a Whatman GF/C filter followed by three washes with PBS, and the activity of the samples was determined by scintillation counting. For binding studies in the presence of guanylyl-5'-imidodiphosphate [Gpp(NH)p], cell membranes were prepared in 10 mM KH₂PO₄, pH 6.8, 1 mM EDTA, 100 µM phenylmethanesulfonyl fluoride, 1 µM leupeptin, and 0.014 mg/ml bacitracin (22). Freshly prepared membranes were incubated with 10 µM Gpp(NH)p for 15 min, and the binding assays were performed as above. Data were analyzed by the Origin 4.0 program (MicroCal). IC₅₀ values reflecting apparent affinities are presented as means ± SD for five independent experiments, each performed in triplicate. The efficacy of HOE 140 (1 µM) was calculated as a fraction of the maximum response by the cognate ligand (set to 100%), i.e., 1 µM bradykinin for B₂R, CKR-2, and CKR-1, and 1 µM ornithokinin for B_oR. Receptor density was calculated from Scatchard transformation of saturation binding analysis data using Origin 4.1. To measure bradykinin-induced contraction of guinea pig ileum, sections of the terminal ileum (2 cm long) were dissected, washed free of fecal contents, hung in Tyrode solution, and bubbled with 98% O₂-2% CO₂ (5). After 1 h of relaxation, with frequent changes of buffer, a dose-response curve to bradykinin was determined, allowing 5 min between challenges. New compounds to be tested were added to the bath, followed by an ED₅₀ dose of bradykinin 30 s later. If the tissue contracted following the addition of the new compound alone, agonist potency was determined relative to the ED₅₀ of bradykinin. If the compound was an antagonist, its potency was measured as the pA₂ value, defined as the negative logarithm of the dose of antagonist that reduces the effect of a 2× dose agonist to that of an x dose in the absence of antagonist (1).

Peptide synthesis. Peptides were synthesized by the standard BOC-benzyl solid phase synthesis (35), purified and characterized by high performance liquid chromatography, amino acid analysis, and laser desorption mass spectrometry (Table 1).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

B₂R antagonists act as B_oR agonists. HOE 140, the "classic" antagonist of mammalian kinin B₂ receptors, is a potent agonist of the chick ornithokinin receptor (34). Initially, we sought to determine whether this effect is specific for HOE 140, or whether the phenomenon holds for other established B₂R antagonists. To this end we selected five prototypic B₂R antagonists (Table 1) and correlated their antagonistic potency for B₂R in the guinea pig ileum contraction assay with their agonistic potency for B_oR recombinantly expressed in CHO cells. We found a linear correlation between the EC₅₀ values for B_oR agonism (measured as G_q-phospholipase C-mediated InsP_n accumulation) and the pA₂ values for B₂R antagonism (r = 0.88 ± 0.047; Fig. 2). Using 15 distinct HOE analogs (data not shown), we found the same correlation coefficient (0.88 ± 0.001). Similar results were obtained with COS-7 cells transiently expressing B_oR (data not shown). The linear correlation between agonist and antagonist capacity indicated that the structurally divergent ligands recognize similar, if not identical, binding sites on the two kinin receptor subtypes but have opposite effects. The same ligands showed partial agonism for human B₂R stably expressed in CHO cells or transiently expressed in COS-7 cells (data not shown). Thus HOE 140 and its derivatives recognize similar or even identical binding motifs at the various kinin receptors but exhibit markedly different efficacies.

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Correlation of agonistic and antagonistic potency of kinin ligands. Agonistic potency is given as EC50 calculated from the dose-response curves of ligand-induced inositol trisphosphates (InsPn) accumulation of BoR (A), CKR-1 (B), and CKR-2 (C). Data are presented as means ± SD from three independent experiments done in duplicate each. Antagonistic potency is given as pA2 of the dose-dependent inhibition of the bradykinin-induced contraction of guinea pig ileum. Numbers denote the HOE 140 derivatives; their structures are presented in Table 1. Assays were done with Chinese Hamster ovary (CHO) cells (BoR, CKR-1) or COS-7 cells (CKR-2).

Chimeric receptors show enhanced efficacy of HOE 140. Because differential ligand efficacy may reflect the structural context of the various receptors, we created chimeric receptors by combining complementary domains of human B₂R and chick B_oR. CKR-1 consisted of the amino- and carboxy-terminal portions of B₂R and a short segment of B_oR, including TM2 and ED2 (Fig. 1). B₂R and CKR-1 were recombinantly expressed in COS-7 cells, and dose-response curves of InsP_n accumulation were analyzed after stimulation with 1 µM bradykinin or HOE 140. Bradykinin showed similar EC₅₀ values (Fig. 3A) and almost identical efficacies (data not shown) for B₂R and CKR-1. Also the maximal effect (E_max) were almost identical for both kinin receptors (Table 3) suggesting that similar copy numbers of functional B₂R and CKR-1 receptors are expressed on the cell surface. By contrast HOE 140 showed a 2.3-fold decrease in EC₅₀ values (Fig. 3B), a 1.9-fold increase in efficacy (Table 2), and a 1.8-fold increase in the E_max values (Table 3) for CKR-1 compared with B₂R, indicating an enhanced potency and efficacy of HOE 140 at the chimeric receptor. Similar results were obtained for HOE 140 and CKR-1 stably expressed in CHO cells where a 2.7-fold decrease in EC₅₀ values and a 2.3-fold increase in efficacy was found (Table 2). We also constructed chimeric receptor CKR-2 with an extended amino-terminal portion of B_oR fused to the B₂R sequence at the junction of ED2 and TM3 (Fig. 1) and expressed it in COS-7 cells. Dose-response curves demonstrated that EC₅₀ (Fig. 3A), efficacy (data not shown), and E_max values (Table 3) of bradykinin were similar for CKR-2 and wild-type B₂R. Thus efficacy, potency, and E_max for bradykinin were almost uniform at CKR-1, CKR-2, and B₂R. For HOE 140 and CKR-2, EC₅₀ values dropped 17-fold, efficacy increased 3.1-fold, and E_max values increased 3-fold compared with wild-type B₂R matching the corresponding parameters for HOE 140 and wild-type B_oR (Fig. 3B; Tables 2 and 3). Hence, replacement of the entire amino-terminal portion of B₂R (amino acid positions 1-127) by the corresponding segments of B_oR "transfers" full agonistic activity of HOE 140 from chicken B_oR to human B₂R without apparent loss of potency of the authentic B₂R ligand bradykinin.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Dose-response curves for bradykinin (A) and HOE 140 (B) at wild-type and CKR-1 and -2. Phosphoinositide hydrolysis was measured as InsPn accumulation after stimulation of COS-7 cells expressing the corresponding receptor with 10 pM to 500 nM of bradykinin or 10 pM to 1 µM of HOE 140. Data are means ± SE from three independent experiments done in triplicate each.

Increased HOE 140 efficacy does not reflect increased receptor density. We asked whether different expression levels are responsible for the change in efficacy of HOE 140 and whether the increased efficacy of HOE 140 at CKR receptors also holds for cells other than COS-7. Studying three distinct CHO clones expressing varying copy numbers of B₂ receptors, we found that the variation of E_max for bradykinin (7- to 18-fold InsP_n accumulation) was not correlated to receptor density (Fig. 4). For instance, clone 2 had a 2.1-fold higher receptor density than clone 3, yet the latter clone had a 1.8-fold higher E_max than the former (Fig. 4, left). By contrast E_max correlated well with InsP_n accumulation induced by the nonspecific G protein activator aluminum tetrafluoride (AlF), suggesting that the observed clonal variability in E_max may well be due to different levels and activities of G_q and/or phospholipase C. We obtained similar results when we studied clones 4 and 5 overexpressing CKR-1, i.e., E_max was not correlated with receptor density, whereas it paralleled ligand-independent AlF-induced InsP_n accumulation (Fig. 4, right). Thus ligand efficacy of HOE 140 is not correlated with receptor density. These results also demonstrate that increased HOE 140 efficacy at CKR-1 is not a COS cell-specific phenomenon.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Relative efficacies of bradykinin and HOE 140 in single CHO clones. Phosphoinositide hydrolysis was measured as InsPn accumulation after stimulation of single clones (arbitrarily numbered 1 through 5) with 100 nM bradykinin (black bars) or 1 µM HOE 140 (white) and is given as the x-fold increase in transfected cells over nontransfected cells. For comparison, clones were stimulated with AlF (gray bars), an unspecific activator of G proteins mimicking phospholipase C-coupled receptor stimulation. Receptor density (fmol/mg of solubilized protein) was calculated from radioligand saturation binding experiments. Efficacy of 1 µM HOE 140 was calculated as the fraction of the maximum response (set to 100%) by 1 µM bradykinin. Results are the means ± SE of at least three independent experiments. Background, i.e., activity in the absence of ligand, was 1,200 counts/min.

Increased HOE 140 efficacy is not due to change in ligand binding site. The observation that HOE 140 efficacy at chimeric receptors is markedly increased prompted us to ask whether this effect reflects structural changes in the corresponding ligand binding sites and/or introduction of a new HOE 140 binding site into the chimeras. To address the first question, we measured the agonistic potencies of five selected HOE 140 derivatives for CKR-1 and CKR-2 and compared them with their antagonistic potencies in the B₂R system (Fig. 2 and Table 1). A linear correlation was found, consistent with the classic pharmacological criteria for identical HOE 140 binding sites at the various receptors (r = 0.98 ± 0.003 and 0.99 ± 0.001 for CKR-1 and CKR-2, respectively). We also determined the rank orders of the potency of the various ligands with the wild-type receptors B₂R and B_oR, as well as for the chimeric receptors CKR-1 and CKR-2: they were invariably HOE 140 > B10038  S911486 > B8852 > S900765 (Table 1), suggesting that the binding sites were essentially identical for the various kinin receptors. Thus increased HOE 140 efficacy is not simply due to an altered ligand binding site.

Efficacy correlates inversely with apparent ligand affinity for uncoupled receptor state. Given that the ligand binding sites among the various kinin receptor types are largely identical, we asked whether differential ligand efficacy may reflect differences in binding affinity to the various receptor states. We measured the apparent binding affinities of HOE 140 and bradykinin for coupled receptors in membranes isolated from untreated cells and for uncoupled receptors in membranes preincubated with 10 µM Gpp(NH)p. We employed homologous displacement assays with [¹²⁵I]HOE 140 ([³H]bradykinin) as the radioligands and unlabeled HOE 140 (bradykinin) as the competitors. In the absence of Gpp(NH)p, the apparent IC₅₀ values for HOE 140 were 1.8 ± 0.2 (B₂R), 1.7 ± 0.4 (CKR-1), 1.2 ± 0.2 (CKR-2), and 1.9 ± 0.3 nM (B_oR), and for bradykinin were 2.2 ± 0.2 (B₂R), 2.4 ± 0.3 (CKR-1), 2.3 ± 0.4 (CKR-2), indicating indiscriminate high affinity binding of HOE 140 and bradykinin to the coupled states of all receptor subtypes and chimeras (Fig. 5, A and C). Uncoupling of the receptors from their G proteins in the presence of 10 µM Gpp(NH)p drastically changed this uniform picture, whereas the apparent IC₅₀ of bradykinin for B₂R, CKR-1, and CKR-2 did not significantly change under these conditions (Fig. 5D), the apparent IC₅₀ for HOE 140 decreased as much as 500-fold to 35.0 ± 6.0 nM (B₂R), 78.5 ± 7.2 nM (CKR-1), 611 ± 63 nM (CKR-2), and 692 ± 78 nM (B_oR) (Fig. 5B). For HOE 140, the rank order of apparent IC₅₀ for the uncoupled receptors, i.e., B₂R > CKR-1 >> CKR-2  B_oR, exactly mirrored the order of efficacy, i.e., B_oR  CKR-2 > CKR-1 > B₂R, indicating an inverse relationship between efficacy and apparent affinity of HOE 140 for the uncoupled receptor state.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Homologous displacement of [125I]HOE 140 and [3H]bradykinin. Membranes were prepared from CHO (B2R, BoR, and CKR-1) or COS-7 cells (CKR-2) and incubated in the absence (left) or presence of 10 µM Gpp(NH)p (right). Binding of 50 pM [125I]HOE 140 (A and B) or 1 nM of [3H]bradykinin (C and D) was determined in the presence of increasing concentrations of unlabeled ligand. Specific binding of the radioligand in the absence of a competitor was set to 100%. Data are presented as means ± SD from three independent experiments performed in duplicate.

Change in receptor conformation modulates ligand efficacy. Differential ligand affinity for uncoupled receptors may reflect differences in the conformations of the various receptors. To address this hypothesis we applied heterologous competition binding assays using [¹²⁵I]HOE 140 as the radioligand and bradykinin as the competitor or vice versa. The IC₅₀ values for wild-type B₂R were almost the same for the two competition modes, i.e., 1.8 ± 0.2 nM for [¹²⁵I]HOE 140/HOE 140 in the homologous displacement assay (Fig. 5A) and 2.0 ± 0.3 nM for [³H]bradykinin/HOE 140 in the heterologous assay (data not shown). The apparent IC₅₀ values for the chimeric receptors differed markedly: 1.7 ± 0.4 nM (homologous) vs. 25 ± 5 µM (heterologous) for CKR-1, and 1.2 ± 0.2 nM vs. >100 µM for CKR-2 (Fig. 5A and data not shown). Similar results were obtained when [³H]bradykinin and [¹²⁵I]HOE 140 were competed by bradykinin: 1.4 ± 0.2 nM (homologous) vs. 6.4 ± 1.8 nM (heterologous) for B₂R, 1.8 ± 0.4 nM vs. 85 ± 25 µM for CKR-1, and 1.2 ± 0.2 nM vs. >100 µM for CKR-2 (Figs. 5C and 6). We also noted that the rank order of HOE 140 efficacy for the various kinin receptors, i.e., CKR-2 > CKR-1 > B₂R, is inversely correlated with its heterologous binding affinity, i.e., B₂R > CKR-1 > CKR-2. Thus the differential efficacy of HOE 140 for the various kinin receptors likely reflects structural differences in the uncoupled receptor conformers, which translate into different ligand affinities for the inactive, but not for the active receptor states.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Heterologous displacement of [125I]HOE 140. Membranes were prepared from CHO (B2R, BoR, and CKR-1) or COS-7 cells (CKR-2) incubated with 50 pM [125I]HOE 140, and radioligand binding was determined in the presence of increasing concentrations of unlabeled bradykinin. The data are presented as means ± SD from three independent experiments done in duplicate.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The fundamental nature of the molecular mechanisms by which agonist binding to G protein-coupled receptors promotes biological activity is a central issue in molecular pharmacology (33). The ternary complex model for G protein-coupled receptors predicts that the agonistic activity of a given ligand is governed by the ratio of ligand affinities for the G protein-coupled, active receptor state R* and for the uncoupled, inactive receptor state R, respectively (19). Thus neutral antagonists, partial agonists, and full agonists are thought to differ by progressively increasing their binding affinity for R* . The major accomplishment of the work presented herein is the experimental demonstration that the same effect, i.e., an increase in efficacy, can be achieved by selectively changing the structure of kinin receptors, thereby reducing the ligand affinity for the uncoupled state R while maintaining the high ligand affinity for R* , and thus shifting the equilibrium toward the coupled, active receptor state. We have exemplified this effect for wild-type and CKR and their major synthetic ligand HOE 140 and conclude that the ratio of the active complex L · R* · G and the inactive complex L · R increases among the various kinin receptor types because the apparent affinity of HOE 140 for the corresponding R states decreases by a factor of >500, whereas the apparent affinity for R* remains constant. The net effect of this equilibrium shift is an increased efficacy of HOE 140 in the order B_oR > CKR-2 > CKR-1 > B₂R.

The progressively lowered affinity of HOE 140 for the uncoupled receptor state, i.e., B₂R > CKR-1 > CKR-2 > B_oR, is not simply due to structural differences in the corresponding ligand binding sites because both HOE 140 and bradykinin show high affinity binding for their corresponding receptors in homologous competition assays. Furthermore, the linear correlation between agonist potency for B_oR and antagonist potency for B₂R of a panel of structurally divergent ligands lends further credit to the notion that the various ligands recognize similar, if not identical, binding sites on the various kinin receptors. Alternatively, each receptor might exist in two distinct conformers, one binding bradykinin and the other HOE 140 and its derivatives. Indeed, mutagenesis studies have revealed that HOE 140 and bradykinin exhibit overlapping though distinct binding sites on B₂R, and that mutated kinin receptors may selectively bind HOE 140 but not bradykinin (16). The most striking effect was found for the mutation F261A in TM6, which lowered the affinity for bradykinin by a factor of >2,000 but left the affinity of HOE 140 almost unchanged (16). We took advantage of this finding and extended it to B_oR: mutation of the conserved phenylalanine residue (F261A) in B_oR drastically reduced the apparent affinity of the authentic ligand ornithokinin, but had little, if any, effect on HOE 140 binding (C. Schroeder, unpublished experiments). These findings reemphasize the idea 1) that B₂R and B_oR expose similar HOE 140 binding sites, and 2) that the two receptors have overlapping but not identical binding sites for the various bradykinin-like ligands.

The observed reduction in apparent affinity in heterologous competition binding studies for B_oR and CKR could be explained by a decreased allosteric transformation rate among receptor conformers, which reduces the apparent affinity of the competing ligand. Isomerization between different receptor conformers has been suggested for the neurokinin NK₁ receptor where the demonstration of two distinct NK₁ receptor conformers blunted an unsuccessful quest for new NK₁ subtypes (27). Mutations affecting allosteric transitions between receptor conformers without concomitant change in ligand affinity have also been reported for the thrombin receptor (13) and for the nicotinic acetylcholine receptor (12). At present we can only speculate that similar conformational changes slowing allosteric transitions of kinin receptors may directly or indirectly reduce the apparent affinity of HOE 140 for the uncoupled receptor state. For instance, the slow interconversion rate among receptor conformers may have allowed us to "capture" a reduced affinity in the pseudosteady state of our radioligand binding assay.

With the advent of constitutively active receptors (33) and overexpression systems for wild-type receptors (6, 20), it became apparent that ligands preferentially binding to the basal receptor state may reduce the spontaneous receptor activity and therefore act as inverse agonists. Indeed HOE 140 has been shown to be an inverse agonist at rat myometrial B₂R (23) or at human B₂R in COS-7 cells (28). Furthermore, HOE 140 is a partial agonist for B₂R from species such as trout (17), sheep (11), guinea pig (9), and humans (10). Finally, conversion from the inverse to partial agonism or vice versa (7) has been reported for HOE 140 with human B₂R and constitutively active mutants thereof (10, 28). One possible explanation for these seemingly discrepant findings is that the specific action of HOE 140 and related compounds may be regulated by the levels of spontaneous activity and basal desensitization of B₂R, which vary with cellular context (22, 28). Because such a possibility is in direct conflict with the two-state receptor model, a three-state model has been proposed that hypothesizes an intermediate, partially active B₂R conformer that is not fully functionally coupled to the G protein (28).

A more straightforward interpretation holds that the ratio of binding affinities of HOE 140 toward the R and R* state(s) dictates the efficacy of the ligand, and that apparent binding affinity may vary among the conformers of various kinin receptor subtypes. At one extreme a very low binding affinity of HOE 140 for R drives almost all of the receptors toward the active LR* G state, and thus HOE is a full agonist. At the other extreme a very high binding affinity of HOE 140 for R shifts the equilibrium between R and R* further toward R thus making HOE 140 an inverse agonist. In this scenario the efficacy or "intrinsic" activity is not an endogenous feature of a given ligand; rather it depends on receptor subtype and/or cellular context, and therefore may vary widely. Thus our findings may help explain species- and tissue-specific differences in the efficacy of HOE 140 and reconcile apparently conflicting results published for HOE 140 and related compounds such as NPC17731 (23, 28). Clearly, the pharmacological classification of a prototypical kinin receptor ligand such as HOE 140 is only valid in a given cellular setting.