Temperature-dependent variations of ligand-receptor contact points in hAT(1).

Photoaffinity labelling is regularly used to investigate proteins, including peptidergic G protein-coupled receptors (GPCR). To this purpose benzophenone photolabels have been widely used to identify many contact residues in ligand-binding pockets. The three-dimensional binding environment of the human angiotensin II type 1 receptor hAT(1) has been determined using an iterative methionine mutagenesis strategy based on the photochemical properties and preferential incorporation of benzophenone onto methionine. This has led to the construction of a ligand-bound receptor structure. The present study investigated the effect of temperature on the accessibility of some of these contact points. The hAT(1) receptor and two representative Met mutants (H256M-hAT(1) and F293M-hAT(1)) from the iterative mutagenesis study were photolabelled with the benzophenone-ligand (125)I-[Sar(1), Bpa(8)]AngII at temperatures ranging from - 15 degrees C to 37 degrees C. Labelled receptors were partially purified and digested with cyanogen bromide to identify the contact points or segments. There were no changes in receptor contacts or labelling in the 7th transmembrane domains (TMD) of hAT(1) and F293M-hAT(1) across the temperature range. However, a temperature-dependent change in the ligand-receptor contact of H256M-hAT(1) was observed. At - 15 degrees C, H256M labelling was identical to that of hAT(1), indicating that the interaction was specific to the 7th TMD. Significant labelling changes were observed at higher temperatures and at 37 degrees C labelling occurred almost exclusively at mutated residue H256M-hAT(1) in the 6th TMD. Simultaneous competitive labelling of different areas of this target protein indicated that the ligand-receptor structure became increasingly fluctual at physiological temperatures, while a more compact, low mobility, and low energy conformation prevailed at low temperatures.

Angiotensin II is bound to both receptors AT1 and AT2, parallel to the transmembrane domains and in an extended form.

UNLABELLED: We have applied photoaffinity labelling methods combined with site-directed mutagenesis towards the two principal angiotensin II (AnglI) receptors AT1 and AT2 in order to determine contact points between AngII and the two receptors. We have first identified the receptor contact points between an N- and a C-terminal residue of the AngII molecule and the AT1 receptor and constructed with this stereochemical restriction a molecular model of AT1. A similar approach with a modified procedure of photoaffinity labelling has allowed us now to determine contact points also in the AT2 receptor. Molecular modelling of AT2 on the rhodopsin scaffold and energy minimisation of AngII binding into this AT2 model produced a model strikingly similar to the AT11 structure. Superposition of the experimentally obtained contact points of AngII with AT2 upon this model revealed excellent congruence between the experimental and modelling results. CONCLUSIONS: (i) athough AT1 and AT2 have quite low sequence homology, they both bind AngII with similar affinity and in an almost identical fashion, as if the ligand dictates the way it has to be bound, and (ii) in its bound form, AngII adopts an extended conformation in both AT1 and AT2, contrary to all previous predictions.

Regulation of VEGF-induced endothelial cell PAF synthesis: role of p42/44 MAPK, p38 MAPK and PI3K pathways.

1. Vascular endothelial growth factor (VEGF) is a potent angiogenic and inflammatory mediator. We have recently shown that this latter effect requires the activation of Flk-1 receptor and subsequent endothelial cell (EC) PAF synthesis. However, the intracellular events that regulate EC PAF synthesis upon Flk-1 stimulation by VEGF remain to be elucidated. 2. Using specific inhibitors and Western blot analysis, we herein report that in bovine aortic endothelial cells (BAEC), VEGF induces the synthesis of PAF through the cascade activation of Flk-1 receptor, phospholipase Cgamma (PLCgamma), protein kinase C (PKC) and p42/44 mitogen-activated protein kinases (MAPK). 3. Moreover, we demonstrate that VEGF-mediated PAF synthesis requires the activation of p38 MAPK, likely by directing the conversion of lyso-PAF to PAF. 4. Interestingly, we observed that VEGF also promoted the activation of the phosphatidyl inositol-3-phosphate kinase (PI3K) pathway, and that its blockade potentiated PAF synthesis following a VEGF treatment. Consequently, it appears that the PI3K pathway acts as a negative regulator of EC PAF synthesis. 5. Taken together, these results allow a better understanding of the intracellular events activated upon EC stimulation by VEGF, and shed a new light on the mechanisms by which VEGF induces PAF synthesis.

Thiol-reactive agents biphasically regulate inositol 1,4,5-trisphosphate binding and Ca(2+) release activities in bovine adrenal cortex microsomes.

Within all endocrine cells, the inositol 1,4,5-trisphosphate (InsP(3)) receptor plays an important role in regulation of the intracellular Ca(2+) concentration. In the present study we showed that a single short-term treatment with either N-ethylmaleimide (known to decrease InsP(3) receptor activity) or thimerosal (known to increase InsP(3) receptor activity) caused time-dependent biphasic effects on the InsP(3) binding activity of bovine adrenal cortex microsomes. The early potentiating effect of thiol-reactive agents translated into a 2-fold increase in binding affinity and Ca(2+) release efficiency. The late dampening effect of thiol-reactive agents translated into a continuous reduction of the maximal binding capacity of the microsomes with a concomitant decrease in Ca(2+) release efficiency. Under these conditions, Western blot analyses demonstrated that the level of InsP(3) receptor protein was not modified. Sequential treatments with thimerosal and the reducing agent dithiothreitol showed that the InsP(3) receptor can readily oscillate between high and low affinity states that are related to its alkylation state. Our results suggest a common mode of action of thiol-reactive agents on the InsP(3) receptor. These results also support the contention that cellular mechanisms of thiol group modification could play important roles in regulation of the intracellular Ca(2+) concentration.

FK506 blocks intracellular Ca2+ oscillations in bovine adrenal glomerulosa cells.

The inositol 1,4,5-trisphosphate (InsP(3)) receptor is a ligand-gated Ca(2+) channel playing an important role in the control of intracellular Ca(2+). In the study presented here, we demonstrate that angiotensin (AngII), phorbol ester (PMA), and FK506 significantly increase the level of InsP(3) receptor phosphorylation in intact bovine adrenal glomerulosa cells. With a back-phosphorylation approach, we showed that the InsP(3) receptor is a good substrate for protein kinase C (PKC) and that FK506 increases the level of PKC-mediated InsP(3) receptor phosphorylation. With a microsomal preparation from bovine adrenal cortex, we showed that PKC enhances the release of Ca(2+) induced by a submaximal dose of InsP(3). We also showed that FK506 blocks intracellular Ca(2+) oscillations in isolated adrenal glomerulosa cells by progressively increasing the intracellular Ca(2+) concentration to a high plateau level. This effect is consistent with an inhibitory role of FK506 on calcineurin dephosphorylation of the InsP(3) receptor, thus keeping the receptor in a phosphorylated, high-conductance state. Our results provide further evidence for the crucial role of the InsP(3) receptor in the regulation of intracellular Ca(2+) oscillations and show that FK506, by maintaining the phosphorylated state of the InsP(3) receptor, causes important changes in the Ca(2+) oscillatory process.

Permeability of endothelial monolayers to albumin is increased by bradykinin and inhibited by prostaglandins.

Using monolayers of bovine aortic endothelial cells (BAEC) in modified Boyden chambers, we examined the role of prostaglandins (PGs) in the bradykinin (BK)-induced increase of albumin permeability. BK induced a concentration-dependent increase of the permeability of BAEC, which reached 49.9 +/- 1% at the concentration of 10(-8) M. Two inhibitors of the prostaglandin G/H synthase, indomethacin (2.88 microM) and ibuprofen (10 microM), potentiated BK-induced permeability 1.8- and 3.9-fold, respectively. Exogenously administered PGE2 and iloprost, a stable analog of prostacyclin, attenuated the effect of BK in a concentration-dependent manner. Butaprost equally reduced the effect of BK, suggesting the participation of the EP2 receptor in this phenomenon. However, the EP4-selective antagonist AH-23848 did not significantly inhibit the protective effect of PGE2. The inhibitory effect of PGE2 was reversed by the adenylate cyclase inhibitor MDL-12330A (10 microM). These results suggest that BK-induced increase of permeability of BAEC monolayer to (125)I-labeled albumin is negatively regulated by PGs. This postulated autocrine activity of PGs may involve an increase in the intracellular level of cAMP.

Prostaglandin E(2) increases cyclic AMP and inhibits endothelin-1 production/secretion by guinea-pig tracheal epithelial cells through EP(4) receptors.

Prostaglandin E(2) (PGE(2)) increased adenosine 3' : 5'-cyclic monophosphate (cyclic AMP) formation in tracheal epithelial cells and concomitantly decreased the production/secretion of immunoreactive endothelin (irET). Naturally occurring prostanoids and selective and non-selective EP receptor agonists showed the following rank order of potency in stimulating cyclic AMP generation by epithelial cells: PGE(2) (EP-selective)>16,16-dimethyl PGE(2) (EP-selective)>11-deoxy PGE(2) (EP-selective)>>>iloprost (IP/EP(1)/EP(3)-selective), butaprost (EP(2)-selective), PGD(2) (DP-selective), PGF(2alpha) (FP-selective). The lack of responsiveness of the latter prostanoids indicated that the prostanoid receptor present in these cells is not of the DP, FP, IP, EP(1), EP(2) or EP(3) subtype. Pre-incubating the cells with the selective TP/EP(4)-receptor antagonists AH23848B and AH22921X antagonized the PGE(2)-evoked cyclic AMP generation. This suggested that EP(4) receptors mediate PGE(2) effects. However, in addition to any antagonistic effects at EP(4)-receptors, both compounds, to a different extent, modified cyclic AMP metabolism. The selective EP(1), DP and EP(2) receptor antagonist (AH6809) failed to inhibit PGE(2)-evoked cyclic AMP generation which confirmed that the EP(2) receptor subtype did not contribute to the change in cyclic AMP formation in these cells. The PGE(2)-induced inhibition of irET production by guinea-pig tracheal epithelial cells was due to cyclic AMP generation and activation of the cyclic AMP-dependent protein kinase since this effect was reverted by the cyclic AMP antagonist Rp-cAMPS. These results provide the first evidence supporting the existence of a functional prostaglandin E(2) receptor that shares the pharmacological features of the EP(4)-receptor subtype in guinea-pig tracheal epithelial cells. These receptors modulate cyclic AMP formation as well as ET-1 production/secretion in these cells.

Assays for radiolabel-photoaffinity labeling of Angiotensin receptors.

Photoaffinity labeling is a useful method to covalently bind two interacting moieties whether they be substrate and enzyme or ligand and receptor. Irreversibly labeling any particular molecule is a practical way of detecting the latter throughout the course of a characterization or a purification procedure.

Photolabeling identifies position 172 of the human AT(1) receptor as a ligand contact point: receptor-bound angiotensin II adopts an extended structure.

An angiotensin II (AngII) peptidic analogue in which the third residue (valine) was substituted with the photoreactive p-benzoyl-L-phenylalanine (Bpa) was used to identify ligand-binding sites of the human AT(1) receptor. High-affinity binding of the analogue, (125)I-[Bpa(3)]AngII, to the AT(1) receptor heterologously expressed in COS-7 cells enabled us to efficiently photolabel the receptor. Chemical and enzymatic digestions of the (125)I-[Bpa(3)]AngII-AT(1) complex were performed, and receptor fragments were analyzed in order to define the region of the receptor with which the ligand interacts. Results show that CNBr hydrolysis of the photolabeled receptor gave a glycosylated fragment which, after PNGase-F digestion, migrated as a 11.4 kDa fragment, circumscribing the labeled domain between residues 143-243 of the AT(1) receptor. Digestion of the receptor-ligand complex with Endo Lys-C or trypsin followed by PNGase-F treatment yielded fragments of 7 and 4 kDa, defining the labeling site of (125)I-[Bpa(3)]AngII within residues 168-199 of the AT(1) receptor. Photolabeling of three mutant receptors in which selected residues adjacent to residue 168 were replaced by methionine within the 168-199 fragment (I172M, T175M, and I177M) followed by CNBr cleavage revealed that the bound photoligand (125)I-[Bpa(3)]AngII forms a covalent bond with the side chain of Met(172) of the second extracellular loop of the AT(1) receptor. These data coupled with previously obtained results enable us to propose a model whereby AngII adopts an extended beta-strand conformation when bound to the receptor and would orient itself within the binding domain by having its N-terminal portion interacting with the second extracellular loop and its C-terminus interacting with residues of the seventh transmembrane domain.

Molecular cloning of a ferret angiotensin II AT(1) receptor reveals the importance of position 163 for Losartan binding.

A complementary DNA for the angiotensin II (AngII) type 1 (AT(1)) receptor from Mustela putorius furo (ferret) was isolated from a ferret atria cDNA library. The cDNA encodes a protein (fAT(1)) of 359 amino acids having high homologies (93-99%) to other mammalian AT(1) receptor counterparts. When fAT(1) was expressed in COS-7 cells and photoaffinity labeled with the photoactive analogue (125)I-¿Sar(1), Bpa(8)AngII, a protein of 100 kDa was detected by autoradiography. The formation of this complex was specific since it was abolished in the presence of the AT(1) non-peptidic antagonist L-158,809. Functional analysis indicated that the fAT(1) receptor efficiently coupled to phospholipase C as demonstrated by an increase in inositol phosphate production following stimulation with AngII. Binding studies revealed that the fAT(1) receptor had a high affinity for the peptide antagonist ¿Sar(1), Ile(8)AngII (K(d) of 5. 8+/-1.4 nM) but a low affinity for the AT(1) selective non-peptidic antagonist DuP 753 (K(d) of 91+/-15.6 nM). Interestingly, when we substituted Thr(163) with an Ala residue, which occupies this position in many mammalian AT(1) receptors, we restored the high affinity of this receptor for Dup 753 (11.7+/-5.13 nM). These results suggest that position 163 of the AT(1) receptor does not contribute to the overall binding of peptidic ligands but that certain non-peptidic antagonists such as Dup 753 are clearly dependent on this position for efficient binding.

Different populations of inositol 1,4,5-trisphosphate receptors expressed in the bovine adrenal cortex.

The inositol 1,4,5-trisphosphate (InsP3) receptor forms a tetrameric channel responsible for the release of Ca2+ from intracellular stores. In the present study we showed that the experimental approach used to separate bound and free ligands may discriminate between two populations of InsP3 binding sites in bovine adrenal cortex microsomes. A large population of low affinity sites and a small population of high affinity sites were detected with centrifugation and filtration approaches, respectively. Both populations were found in the supernatant and the cytoskeleton fractions of Triton X-100 solubilized microsomes. After treatment of microsomes with thimerosal, an alkylating reagent known to increase InsP3 receptor affinity, the filtration and the centrifugation approaches yielded identical results. With selective anti-InsP3 receptor antibodies, we showed that types 1, 2 and 3 InsP3 receptors are present in intact microsomes and in the cytoskeleton fraction. Binding studies on immunoprecipitated receptors revealed that anti-type 1 antibody recognizes a large population of low affinity sites whereas anti-type 2 antibody recognizes a small population of high affinity sites. Our results indicate that the three types of InsP3 receptors are expressed at different levels in the bovine adrenal cortex. The presence of different types of InsP3 receptors with different ligand binding affinities and their association with the cytoskeleton offer a convenient way for the cell to simultaneously regulate its intracellular Ca2+ concentration and reorganize the spatial distribution of its Ca2+ stores.

Adenosine induces cyclic-AMP formation and inhibits endothelin-1 production/secretion in guinea-pig tracheal epithelial cells through A(2B) adenosine receptors.

1. The adenosine receptor subtype mediating adenosine 3' : 5'-cyclic monophosphate (cyclic AMP) formation and the effect of its activation on endothelin-1 (ET-1) secretion were studied in primary cultures of tracheal epithelial cells. 2. Adenosine analogues showed the following rank order of potency (pD(2) value) and intrinsic activity on the generation of cyclic AMP by tracheal epithelial cells: 5'-N-ethylcarboxyamidoadenosine (NECA, A(1)/A(2A)/A(2B), pD(2): 5.44+/-0.16)>adenosine (ADO, non selective, pD(2): 4.99+/-0. 09; 71+/-9% of NECA response) >/=2-Cl-adenosine (2CADO, non selective, pD(2): 4.72+/-0.14; 65+/-9% of NECA response)>>>CGS21680 (A(2A); inactive at up to 100 microM). 3. Cyclic AMP formation stimulated by NECA in guinea-pig tracheal epithelial cells was inhibited by adenosine receptor antagonist with the following order of apparent affinity (pA(2) value): Xanthine amine congeners (XAC, A(2A)/A(2B), 7.89+/-0.22)>CGS15943 (A(2A)/A(2B), 7.24+/-0. 26)>ZM241385 (A(2A), 6.69+/-0.14)>DPCPX (A(1), 6.51+/-0. 14)>3n-propylxanthine (weak A(2B), 4.30+/-0.10). This rank order of potency is typical for A(2B)-adenosine receptor. 4. Adenosine decreased basal and LPS-stimulated irET production in a concentration-dependent manner. Moreover, NECA but not CGS21680 inhibited LPS-induced irET production. 5. The inhibitory effect of NECA on LPS-induced irET production was reversed by XAC (pA(2)=8.84+/-0. 12) and DPCPX (pA(2)=8.10+/-0.22). 6. These results suggested that adenosine increased cyclic AMP formation and inhibited irET production/secretion by guinea-pig tracheal epithelial cells through the activation of a functional adenosine receptor that is most likely the A(2B) subtype. This adenosine receptor may be involved in the regulation of the level of ET-1 production/secretion by guinea-pig tracheal epithelial cells in physiological as well as in pathophysiological conditions.

Synthesis, acid-base behavior, and binding properties of 6-modified myo-inositol 1,4,5-tris(phosphate)s.

myo-Inositol 1,4,5-tris(phosphate) was modified at position 6. The analogues synthesized are reported in this publication are 6-deoxy-myo-inositol 1,4,5-tris(phosphate), 6-fluoro-6-deoxy-myo-inositol 1,4,5-tris(phosphate), epi-inositol 1, 4,5-tris(phosphate), and 6-amino-6-deoxy-myo-inositol 1,4, 5-tris(phosphate). These derivatives showed poor affinity for the Ins(1,4,5)P(3) receptors. The inframolecular acid-base behavior and the cooperative effects between the phosphate groups could help explain the loss of affinity of these 6-modified analogues.

Agonist-dependent AT(4) receptor internalization in bovine aortic endothelial cells.

Recent studies have characterized a specific binding site for the C-terminal 3-8 fragment of angiotensin II (Ang IV). In the present study we looked at the internalization process of this receptor on bovine aortic endothelial cells (BAEC). Under normal culture conditions, BAEC efficiently internalized (125)I-Ang IV as assessed by acid-resistant binding. Internalization of (125)I-Ang IV was considerably decreased after pretreatment of cells with hyperosmolar sucrose or after pretreatment of BAEC with inhibitors of endosomal acidification such as monensin or NH(4)Cl. About 50% of internalized (125)I-Ang IV recycled back to the extracellular medium during a 2 h incubation at 37 degrees C. (125)I-Ang IV remained mostly intact during the whole process of internalization and recycling as assessed by thin layer chromatography. As expected, internalization of (125)I-Ang IV was completely abolished by divalinal-Ang IV, a known AT(4) receptor antagonist. Interestingly, (125)I-divalinal-Ang IV did not internalize into BAEC. These results suggest that AT(4) receptor undergoes an agonist-dependent internalization and recycling process commonly observed upon activation of functional receptors.

Role of N-glycosylation in the expression and functional properties of human AT1 receptor.

The role of N-glycosylation in the pharmacological properties and cell surface expression of AT1 receptor was evaluated. Using site-directed mutagenesis, we substituted both separately and simultaneously the asparagine residues in all three putative N-linked glycosylation consensus sequences (N-X-S/T) of AT1 receptor (positions 4, 176, and 188) with aspartic acid. Expression of these mutant receptors in COS-7 cells followed by photolabeling with [125I]-[p-benzoyl-Phe8]AngII and SDS-PAGE revealed ligand-receptor complexes of four different molecular sizes, indicating that the three N-glycosylation sites are actually occupied by oligosaccharides. Binding studies showed that the affinity of each mutant receptor for [Sar1,Ile8]Ang II was not significantly different from that of wild-type AT1 receptor. Moreover, the functional properties of all mutant receptors were unaffected as evaluated by inositol phosphate production. However, the expression levels of the aglycosylated mutant were 5-fold lower than that of the wild-type AT1 receptor. Use of green fluorescent protein-AT1 receptor fusion proteins in studying the cellular location of the aglycosylated mutant demonstrated that it was distributed at a much higher density to the ER-Golgi complex than to the plasma membrane in HEK 293 cells. Together, these results suggest an important role of N-glycosylation in the proper trafficking of AT1 receptor to the plasma membrane.

Determination of peptide contact points in the human angiotensin II type I receptor (AT1) with photosensitive analogs of angiotensin II.

To identify ligand-binding domains of Angiotensin II (AngII) type 1 receptor (AT1), two different radiolabeled photoreactive AngII analogs were prepared by replacing either the first or the last amino acid of the octapeptide by p-benzoyl-L-phenylalanine (Bpa). High yield, specific labeling of the AT1 receptor was obtained with the 125I-[Sar1,Bpa8]AngII analog. Digestion of the covalent 125I-[Sar1,Bpa8]AngII-AT1 complex with V8 protease generated two major fragments of 15.8 kDa and 17.8 kDa, as determined by SDS-PAGE. Treatment of the [Sar1,Bpa8]AngII-AT1 complex with cyanogen bromide produced a major fragment of 7.5 kDa which, upon further digestion with endoproteinase Lys-C, generated a fragment of 3.6 kDa. Since the 7.5-kDa fragment was sensitive to hydrolysis by 2-nitro-5-thiocyanobenzoic acid, we circumscribed the labeling site of 125I-[Sar1,Bpa8]AngII within amino acids 285 and 295 of the AT1 receptor. When the AT1 receptor was photolabeled with 125I-[Bpa1]AngII, a poor incorporation yield was obtained. Cleavage of the labeled receptor with endoproteinase Lys-C produced a glycopeptide of 31 kDa, which upon deglycosylation showed an apparent molecular mass of 7.5 kDa, delimiting the labeling site of 125I-[Bpa1]AngII within amino acids 147 and 199 of the AT1 receptor. CNBr digestion of the hAT1 I165M mutant receptor narrowed down the labeling site to the fragment 166-199. Taken together, these results indicate that the seventh transmembrane domain of the AT1 receptor interacts strongly with the C-terminal amino acid of [Sar1, Bpa8]AngII interacts with the second extracellular loop of the AT1 receptor.

The bradykinin B2 receptor couples less efficiently than the angiotensin AT1 receptor to the G protein Gq in transiently transfected COS-7 cells.

The purpose of this study was to compare the efficiency of two different Gq protein-coupled receptors (AT1 receptor for angiotensin II and B2 receptor for bradykinin) to activate phospholipase C (PLC). When the receptors were expressed at a similar level of 0.5 pmol/mg of protein, inositol trisphosphate (IP) accumulation elicited by AT1 receptor was four times higher than that elicited by B2 receptor. Genistein and pertussis toxin did not modify AT1 receptor- or B2 receptor-induced IP accumulation. These results indicate that in COS-7 cells, the two receptors activate PLC beta through G proteins of the Gq family. AT1 or B2 receptors were co-expressed with the alpha subunit of either Gq or G11. Both alpha subunits potentiated to the same extent AT1 receptor-induced IP accumulation. alpha 11 was also as efficient as alpha q to potentiate B2 receptor-induced response. Interestingly, however, the potentiating effect of alpha q and alpha 11 was more important (by 5-fold) on AT1 receptor-mediated response than on B2 receptor-mediated response. These results demonstrate that the extent of activation of PLC beta by different Gq-coupled receptors depends on the level of expression of these receptors and on their coupling efficiency. These are important parameters that determine the relative contribution of specific hormones to different biological processes.

Effect of uncoupling agents on AT1 receptor affinity for antagonist analogs of angiotensin II.

AT1 receptor is responsible for most of the physiological effects of Angiotensin II (Ang II). AT1 receptor belongs to the G-protein-coupled receptor (GPCR) family, and it mediates its actions through the coupling of the Gq/11 protein with phospholipase C beta. Classical pharmacology has used the sensitivity of GPCR ligands to uncoupling agents as a criteria to discriminate agonists (which are sensitive) from antagonists (which are insensitive). In this study, the uncoupling agents GTP gamma S and pentosan sulfate (PS) (a low molecular weight polyanion) were used to further characterize the molecular interactions between Ang II analogs and the AT1 receptor. We show that some Ang II antagonists are sensitive to the conformational change of the AT1 receptor induced by uncoupling agents. These results demonstrate that there is no direct relationship between the intrinsic activity of a ligand and its affinity for different conformations of the AT1 receptor and that the sensitivity of GPCR ligands to uncoupling agents can not be used as a criteria to discriminate agonists from antagonists.