The present invention relates to nucleic acid molecules encoding modified adrenergic receptors being constitutively activating, cells including modified adrenergic receptors and methods of screening compounds using such cells.

Background of the Invention

The adrenergic receptors (AR) mediate the effects of the sympathetic hormones epinephrine and norepinephrine by transducing G-proteins which activate effector molecules intracellularly. αi-Adrenergic receptors (αi-ARs) are members of the G-protein-coupled receptor (GPCR) superfamily. These receptors share many features of conserved structural motifs such as the seven membrane- spanning domains joined by extracellular and intracellular loops. Functional features are also conserved in that the coupling to specific G-protein appears to involve the second and third intracellular loops. αi-ARs can activate multiple effectors, such as phospholipase C(PLC) and phospholipase A ₂ (P A ₂ ) through their respective pathways, via coupling to distinct G-proteins (1). The phenomena of multiple effector coupling is now well documented for many members of the CPCR superfamily. Although signalling by these receptors most likely involves an agonist-induced conformational change in the receptor-protein that translates down to the coupled G-protein, the mechanism(s) whereby a single activated receptor subtype can couple to several distinct G-proteins has not been addressed. Adrenergic ligands bind to residues in the extracellular half of the putative transmembrane α-helical domains.

There have been previous reports of the production of mutated adrenergic receptors being constitutively activated (14, 22). These receptors, however, are activated through both phospholipase pathways (P C and P A ₂ ) and exhibit higher affinity for agonists. The

previously reported mutated adrenergic receptors were produced by substituting amino acids at position Ala 293, located near the carboxyl terminus of the third intracellular loop of otiR- R. There is however no suggestion in these previous reports of the possibility of producing a constitutively activating adrenergic receptor in which the activation is mediated through only one pathway.

Constitutively activating adrenergic receptors in which activation is mediated through only one pathway and cells possessing these receptors are required to more clearly understand the mechanisms of adrenergic receptor binding and coupling. Such cells will also be useful in studying drug interactions with adrenergic receptors and in the development of new drugs.

In order to produce constitutively activating ARs, the present inventors have mutated the α^g-AR in the third transmembrane region and observed a dramatic increase in affinity for a particular class of agonists, phenethylamines, but not for imidazolines. This increased affinity of phenethylamines correlated with a corresponding increase in its potency to activate polyphosphoinositide (PI) hydrolysis but, interestingly, not of arachidonic acid release. In addition, this mutant had elevated basal levels of inositol phosphates (PLC pathway) but not of arachidonic acid release (P A2 pathway) indicative of a constitutively activating receptor. These data suggest the ability of a receptor to form multiple conformations for G-protein interactions (high affinity states) that are specific for a particular G-protein and that there exists multiple sites for agonist binding that promote these specific interactions. Summary of the Invention In a first aspect, the present invention consists in a nucleic acid molecule encoding a modified αig-adrenergic receptor, the modification being such that the receptor is

constitutively activating and is activated through only one pathway.

In a preferred embodiment of the first aspect of the present invention the one pathway is the phospholipase C pathway.

In a further preferred embodiment of the first aspect of the present invention, the nucleic acid molecule includes a codon for an amino acid other than cysteine at position 128 of the receptor. More preferably, the nucleic acid molecule includes a codon for the amino acid phenylalanine at position 128 of the receptor.

In a second aspect, the present invention consists in a cell including a modified 0.1 ₃ -adrenergic receptor, the modification being such that the receptor is constitutively activating and is activated through only one pathway. Preferably, the cell is transformed or transfected with the nucleic acid molecule of the first aspect of the present invention such that the cell expresses the constitutively activating adrenergic receptor.

In a preferred embodiment of the second aspect of the present invention, the one pathway is the phospholipase C pathway.

In a further preferred embodiment of the second aspect of the present invention, the adrenergic receptor includes an amino acid other than cysteine at position 128 of the receptor. More preferably, the amino acid at position 128 of the receptor is phenylalanine.

In a third aspect, the present invention consists in a method of determining whether a compound exerts an effect on an adrenergic receptor, the method comprising exposing the cell of the second aspect of the present invention to the compound and measuring change in phospholipase activity in the cell. In order that the nature of the present invention may be more clearly understood, preferred forms thereof

will be described with reference to the following drawings and examples.

Brief Description of Drawings

Figure 1 shows the chemical structures of phenethylkamines and imidazolines with respective sites of substitutional "R" groups.

Figure 2 shows agonist binding and effect or activation of wild-type and mutated αi-ARs . Panels A, Epinephrine and cirazoline binding by membranes of COS-1 cells expressing the wild-type αiR-AR, the Cys Phe mutation and the Ala 293Glu mutation. The K of epinephrine binding was 2223nM for the wild type, 133nM

128 293 for Cys Phe and 91nM for Ala Glu. The K of cirazoline binding was 1337nM for the wild type, 1066nM for Cys 128Phe and 1160nM for Ala 293Glu. Panels B, Epinephrine and cirazoline potencies for PI hydrolysis in COS-1 cells

128 expressing wild-type αχB~ R, the Cys Phe mutation and the

Ala 293Glu mutation. The EC50 of epinephrine was 40 nM for the wild type, 5 nM for Cys Phe and 3nM for the Ala Glu receptors. The EC ₅₀ of cirazoline was 21 nM for the wild type and 12nM for Cys ¹²⁸ Phe and 25nM for the Ala ²⁹³ Glu receptors. Panels C, Epinephrine and cirazoline potencies for arachidonic acid release in COS-1 cells expressing wild-type O-IB-AR, the Cys 128Phe mutation and the Ala293Glu mutation. The EC ₅₀ of epinephrine was 42 nM for the wild- type, 27 nM for Cys ¹²⁸ Phe and 31ρM for the Ala ²⁹³ Glu receptor. The EC ₅₀ of cirazoline was 22 nM for the wild type and 35 nM for both of the mutated receptors. Each value shown is the mean ± standard error of the mean (vertical lines), for at least three individual experiments . K values for other agonists are shown in Table 1. Receptor expression in panel B and C was 3, 2.2 and 2.2 pmol/mg membrane protein for the wild type,

Cys128Phe and Ala293Glu mutations, respectively. Figure 3. (A)IP3 production or (B) arachidonic acid

128 release in the absence of agonist by the Cys Phe or

Ala 293Glu c-i-AR mutations. Basal levels of IP ₃ and arachidonic acid release in COS-1 cells expressing the mutated αiβ- Rs are represented relative to levels in cells transfected with the vector alone (mock) . Expression levels of receptors were 1.9, 2.2 and 1.4 pmol/mg membrane protein for the wild-type, Cys 128Phe and

Ala 293Glu, respectively. Results are normalized to receptor densities (determined by Scatchard analysis) in the transfected COS-1 cells by dividing the total receptor-specific pmoles of IP ₃ or arachidonic acid release per 60mm dish by the total number of receptors per dish. PTX (lμg/ml) was added directly to the medium 24 hours prior to assay. Prazosin concentration (PRZ) was at 1 μg/ml) was added directly to the medium 24 hours prior to assay. Prazosin concentration (PRZ) was at 1 μM final and was added 16-24 hour prior to assay. *indicates statistical significance as evaluated by the Student's t test (p<0.001). Values are the mean ± standard error of the mean (vertical lines) of three independent experiments determined in triplicate.

Description of Invention

The inventors mutated a single residue, Cys 128, in the third transmembrane domain of the αiβ-AR to a phenylalanine. This residue is situated approximately one helical turn below Asp 125 in the third transmembrane segment. Since Asp 125 is the putative counterion that binds the protonated amine of adrenergic ligands, one can postulate that an agonist-induced conformational change of the third transmembrane segment may be involved in receptor signalling. Pertubation, therefore, of this critical agonist-binding interaction might induce changes in the conformation of this helix. The present inventors have found that a point mutation (Cys 128Phe) in the third transmembrane segment of the O-IB- R results in a conformation that partially mimics the activated-state of the receptor for only a single effector pathway. This

contrasts with another constitutively-activating mutation (Ala 293Glu) that promiscuously activates more than one effector pathway. In addition, the Cys 128Phe mutation seems to favour signalling through a single effector pathway that is promoted by a particular class of agonists called phenethylamines, suggesting that the structure of an agonist may be able to induce the receptor into a particular signalling mode. The ability to direct receptor-mediated second messenger pathways from outside the cell has major therapeutic potential. MATERIALS AND METHODS

Site-directed Mutagenesis. The construct used was the hamster αiβ-AR (2) which included a manufactured EcoRl restriction site at the 5' end and a region encoding an octapeptide tag (ID4) at the end of the coding region which was used to evaluate membrane expression using a monoclonal antibody (anti-ID4). The attachment of this epitope after the coding region does not affect protein expression or the functional nature of the receptor. The construct ends with a stop codon and a Not 1 restriction site. The cDNA was divided into two 800bp fragments by restriction with BamHl endonuclease and each fragment was inserted individually into M13mpl9. This was done to reduce the potential incidence of spurious modifications in the DNA due to the M13 system as reported with large constructs (3). Site-directed mutagenesis was performed as previously described utilizing the oligonucleotide-mediated double primer method (3). This utilizes a 20 base mutagenic primer encoding the codon mismatches to achieve the point mutation and the Universal primer for extension on single-stranded M13 templates. After transformation of the extended products into DH5αF cells, plaques were screened for the mutation on nitrocellulose lifts and probed with the 32P-end labelled mutagenic primer 5 C below the calculated T _m . The efficiency of mutagenic incorporation was 5% of total

plaques. Positive plaques were purified and the DNA isolated and sequenced by the dideoxy method (sequenase, Amersham) to verify the mutation. The replicative form (RF) and the DNA was isolated from the Ml3 and the insert removed and subcloned into the expression vector, pMT2' . The full length plasmid DNA was again sequenced to verify the mutation.

Cell Culture and Transfection. COS-1 cells (ATCC) were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Cells (1 x 10 ) were plated in 60 mm dishes for transfection. cDNAs encoding the wild type αi _B -AR and various mutants were subcloned into the mammalian expression vector pMT2' , as previously described (4). Plasmid DNA, purified by Csl gradient centrifugation and Biogel A-50m (Bio-Rad) column chromatography, was used to transfect cells. Transient expression in COS-1 cells was accomplished by the DEAE-dextram method (3). DNA levels (2-4μg per 1 x 10 cells) were varied to achieve similar receptor expression among the constructs for the dose-response studies. Cells were harvested or assayed 60 hours post-transfeetion.

Membrane Preparation. COs-1 membranes were prepared as previously described (4). Briefly, membranes were prepared by washing the culture plates twice with warm HBSS. One millilter of HBSS was added, and the plates were scraped and transferred to a 50 ml centrifuge tube. The intact cells were centrifuges at 1000 x g in a Sorvall RT6000B rotor for 5 min, and the pellet was resuspended in 5 ml of 0.25M sucrose. The cell suspension was centrifuged again at 1000 x g for 5 min, and the pellet resuspended in 10ml of 0.25M sucrose containing the following protease inhibitors: 20 μg/ l aprotinin, 20μg/ml leupeptin, 20μg/ml bacitracin, 20μg/ml benza idine, and 17μg/ml phenylmethylsulfonylfluoride. The cells were disrupted by N ₂ cavitation and then homogenised in a Dounce homogenizer by 10 strokes with a

loose fitting (B) pestile. The mixture was then centrifuged at 1260 x g for 5 min. Buffer, 50mM Tris, pH 7.4, 12.5 mM MgC12, and 5 mM EGTA was added to the supernatant fraction, which was then centrifuged at 30,000 x g for 15 min. The resulting pellet was resuspended in 50ml of buffer and recentrifuged for 15 min. The resulting pellet was resuspended in 1ml of buffer containing 10% glycerol and stored in aliquots at -70°C. The protein concentration was measured using the method of Bradford (5) .

Radioligand Binding. The ligand-binding characteristics of the expressed receptors were determined in a series of radioligand binding studies using the αi-antagonist radioligand [I 125] HEAT as previously described(4) . Competition reactions (total volume, 0.25ml) contained

1 ?"i

20mM HEPES, pH 7.5, 1.4mM EGTA, 12.5mM MgCl ₂ , 200pM [ I] HEAT, COS-1 membranes, and increasing amounts of unlabelled ligands known to interact with ARs..

Nonspecific binding was determined in the presence of 10 ^" M phentolamine. Reactions were stopped by the addition of cold HEPES buffer and were filtered onto Whatman GF/C glass fiber filters with a Brandel cell harvester. Filters were washed fives times with HEPES buffer, and bound radioactivity was determined using a Packard Auto-gamma 500 counter. Binding data were analysed by the iterative curve-fitting program LIGAND.

Hill coefficients were determined from a Hill plot. For saturation binding studies, [I 125] HEAT concentrations ranging from 25 to 800pM were used. Saturation curves were obtained by incubating cell membranes with increasing amounts of [ 125IJHEAT in the same buffer system used for the competition studies. To reduce interassay variation, all binding assays were performed simultaneously with all three constructs. PI Hydrolysis. Inositol phosphate determination were performed as described (6). Cells expressing the receptor

constructs, grown in 60 mm dishes (3 x 10 cells/dish)

3 were labelled for 16-24 hours with [ H] inositol (Du Pont

- New England Nuclear) at luCi/ml in DMEM supplemented with 5% fetal bovine serum. After 24 hours, cells were washed three times with DMEM (no serum) and incubated in the serum-free media for 30 min, followed by a 30 min incubation in DMEM containing lOmM LiCl. Agonists were then added for 30 min., the media was removed and the cells lysed with 2ml of ice-cold 0.4M perchloric acid. One half volume of 0.72N KOH/0.6M KHC0 ₃ was added and the sample was centrifuged to settle the precipitate. The supernatant was applied to 1ml packed AG1-X8 (BioRad) columns (100-200 mesh, formate form) and total inositol phosphates were eluted with 1 M ammonium formate/0.1 M formic acid, after the column was washed with 16ml of 0.1M formic acid. For basal measurements performed in separate

₃ studies, IP ₃ production was determined using a H-IP ₃ - radioreceptor assay kit (Du Pont) according to the manufacturer's specifications. This kit contains known IP ₃ standards to generate curves for quantitation.

Scatchard analysis of assay plates were performed in parallel with functional studies to quantitate receptor expression. Receptor-specific IP ₃ generation (pmole/plate) was determined by subtracting out the IP ₃ generated by cells transfected with vector alone (mock) . Scatchard analyses of equilibrium binding studies performed on parallel plates were used to determine receptor number/ g of membrane protein. Bradford assays were used to determine mg of membrane protein per 60mm dish. This allowed the pmole IP ₃ /pmole receptor/dish to be calculated. To reduce interassay variability, functional studies were always done for all three constructs in a single assay with a particular drug. To block receptor signalling- prazosin (1 μM) was added to the media 16-24 hours before the assay.

Arachidonic acid assay. For arachidonic acid release, transfected cells in 60 mm dishes (3 x 10 cells/dish)

3 were incubated overnight with [ H]-arachidonic acid (Du Pont-New England Nuclear) (luCi/ml media or 40nmole total) . The cells were washed three times with DMEM (no serum) . After the final wash, the cells were incubated with DMEM for 30 min. followed by the addition of the dual lipoxygenase and cyclooxygenase pathway- inhibitor, phenidone (100 μM final) for 30 min to allow free achidonic acid to accumulate. Agonists were then added in the media and incubated for 30 min. The media was removed and the cells lysed with 0.4M perchloric acid and extracted with a methanol/chloroform/HCl mixture (40:40:0.5). The upper phase is removed to vials, 20ml of scintillation cocktail (Ecoscint A; National Diagnostic) added and the samples counted after dark-adaptation. Here and in previous studies (6), the inventors verified the identity of the radioactive extracts as released arachidonic acid by thin layer chromatography. The extracts were evaporated to dryness, resuspended in 50μl of chloroform, and applied to silica gel thin layer chromatography plates (LK5D; Whatman) . The plates were developed in a heptane/diethyl ether/acetic acid (75:25:4, v/v/v) solvent system. Nonradioactive standards (2ug) were run in each lane as carriers. Carriers were visualised with iodine vapour and radioactivity was quantified by scraping the plates and adding 10ml scintillation cocktail to the resin. Samples were dark- adapted overnight before being counted. To reduce interassay variability, functional studies were always done for all three constructs in a single assay with a particular drug. Pmole of arachidonic acid release was calculated based on the specific activity (100 Ci/mmole) . Pertussis toxin (PTX) (lμg/ml) was added directly to the media 24 hours prior to assay. In previous studies (6), the inventors determined that 1 μg/ml of PTX is sufficient

32 to totally block [ P]-NAD incorporation into a 41-kDa protein in the COS-l system.

Molecular Modelling. The coordinates of the alpha-carbon positions were determined by an overlay of the putative cti-AR transmembrane residues with the transmembrane coordinates of bacteriorhodopsin (7) using data files generated using the Insight II molecular modelling software from Biosym Technologies. The boundaries of the putative transmembrane domains were determined by an algorithm based upon the weighted pair-wise comparisons of adjacent residues. The αi-AR model was then minimised and conflicts adjusted as previously described (8).

Assumptions of key amino acids involved in ligand binding, such as the Asp 125 _t are based upon previous mutagenesis work and proposed models for the β-AR (9).

Materials. Drugs were obtained from the following manufacturers: WB4101, 5-methylurapidil, Research Biochemicals Inc. (Natick, MA); (-) epinephrine, (-) norepinephrine, prazosin, oxymetazoline, clonidine, methoxamine, phenylephrine, phentolamine, Sigma; rauwolscine, Roth (Germany); Cirazoline was a gift from

195 ^ 3

Pfizer (CT. [ I]HEAT, [ H] inositol [ H] arachidonic acid and the H-IP ₃ radioreceptor kit were from Du Pont - New England Nuclear (Boston, MA) . AG1-X8 and Biogel A-50m resin were from Bio-Rad.

BESHLIS

Many structure-function relationships have been explored for the adrenergic receptor family. Most of this work has been confined to the β-AR and some to the 0- ₂ -AR systems but little work has been done on oti-ARs per se. Since αi-AR subtypes bind very different types of antagonists with high affinity, the ligand binding pocket must be distinctive between β, α^ and G- ₂ -ARs despite the similar binding of the natural agonists. The inventors sought to explore the effect the Cys 128Phe mutation had the antagonist binding pocket of αχ-ARs. The ligand

binding pocket of αi-ARs is very long and, indeed, good αi-antagonists such as 5-methylurapidil and WB4101 tend to be very long molecules. However, in the 01 ₂ -AR, antagonists are shorter molecules and at the equivalent amino acid position of 128, the α ₂ -AR contains a Phe residue. Therefore, a mutation at position 128 in the a^- AR that mimics the 0. ₂ -AR might shorten the ligand binding pocket to long antagonists. This phenotype was not observed, however, as this mutation was constitutively activated and further study indicated a selective activation of a single effector pathway.

Choice of Expression Systems — To measure the effect of a single type of receptor, the inventors chose a cell line for transfections that does not endogenously express αi- ARs or other adrenergic receptors. The criticism of this system is that this cell line might not contain the correct isoforms of the G-proteins and effect or molecules that normally couple to the αχ-ARs. However, the inventors have previously shown in the overexpressed transiently-transfected COS-1 cell system (a green monkey kidney cell line) that a single αχ-AR can couple to multiple G-proteins (6). Thus in COS-1 cells, the O-IB- R can activate a phospholipase A ₂ and generate arachidonic acid. This activation proceeds through a pertussis toxin- sensitive G-protein. The GCIB-AR can also couple to inositol phosphate generation through a pertussin toxin- insensitive G-protein. Both of these pathways have been shown to be independent from one another. However, αχ~ AR/G-protein coupling is not promiscuous in the COS-1 system since similar results were obtained in a stably- transfected CHO cell line which expresses receptor at levels observed in vivo. Moreover, similar signalling pathways that utilise PTX-sensitive and -insensitive G- proteins were observed for 0- ₁ -ARs in cell lines endogenously expressing the receptor (6). However, the COS-1 system because of its high receptor expression and

low signalling background is ideally suited for signalling studies and, thus, was the cell line of choice.

128

Characterisation of Cys Phe Pharmacology — As shown in Table 1, the binding of the αχ-AR antagonists, prazosin, WB4101, 5-methylurapidil and [ ¹²⁵ I]HEAT by t he Cys ¹²⁸ Phe mutant, was not significantly different from that of the wild-type receptor. Thus, the original premise that the

Cys 128Phe mutation would shorten the antagonist binding pocket and, thus, prevent or impair antagonist binding was incorrect. Further analysis, however, revealed that this mutant recognised the catecholamines, norepinephrine and epinephrine, as well as other phenethylamine agonists (Table 1, Figure 1 and 2A) with higher affinity (5-16 fold) . Scatachard analysis also indicated no change in the affinity for the radiolabel, [ 125I]HEAT, used in the binding studies as compared to the wild type receptor

(Table 1). The αi-antagonist [ 125I]HEAT apparently labels a homogeneous population of binding sites in the membranes prepared from these mutations. However, the binding of a different structural class of agonists, the imidazoline agonists (Table 1; Figure 1 and 2A) by the Cys 128Phe mutant was unaltered. Thus, this mutant distinguishes between different classes of adrenergic agonists, indicating that the determinants for the binding of phenethylamines differ from those of imidazolines. This difference cannot be attributable to the low intrinsic activity commonly observed for imidazolins, since the binding of the partial agonist, methoxamine, a phenethylamine, was also of a higher affinity and this increase in affinity was comparable to that observed with the full agonist, norepinephrine. Nevertheless, it is likely that the determinants of phenethylamine and imidazoline binding although distinct, are overlapping. Previous studies of the β- and 0- ₂ -ARs, for example, suggest conservation of key amino acids (e.g. Asp 113, 65 208

Ser , Ser ) for agonist binding (9), and both classes

of agonists share common pharmacophores, such as a phenyl ring and a protonated amine, located three bond lengths from the aromatic ring. The present findings of differential agonist binding are important, since they provide direct evidence for an earlier hypothesis based on pharmacological data; thus, it was proposed that phenethylamines and imidazolines interact differentially with 0- ₂ - and αχ-ARs (10, 11). Moreover, differential binding sites might suggest different molecular mechanisms of activation as has been recently indicated for the C. ₂ -AR where imidazolines and phenethylamines differentially couple to distinct G-proteins (12).

COS-1 cell membranes transfected with the pMT2 ' expression vector containing either the wild type hamster oti _B cDNA or the mutated Cys 128Phe or Ala293Glu cDNAs were incubated with the αχ-AR antagonist [ 125I]HEAT, in the absence or presence of increasing concentrations of various agonists or antagonists *. Each point represents the mean of at least two to five individual experiments, in duplicate. Ten concentrations of each ligand were treated, and the points were chosen to be on the linear portion of the displacement curve. Ki values were generated using the iterative curve-fitting program

LIGAND. Values in parentheses are the ratio of wild type to mutant Ki's.

TABLE 1 Pharmacological characterisation of wild-type and mutant αi _B -ARs

LIGAND T Cyε ¹²⁸ Phe Ala ²⁹³ Glu

Agonist? Ki(nM)

Phenethylamines (-)epinephrine 2,223 133 (16.7) 91(24) (-)norepinephrine 3,700 373 (9.9) 194(19) (+)norepinephrine 247,666 16,066 (15) methoxamine 450,000 47,333 (9.5) 66,666 (6.8) phenylephrine 10,366 1,913 (5.4) 933 (11.1)

Imidazolines oxymetazoline 596 560 (1.1) 1270 (0.5) cirazoline 1,337 1,066 (1.3) 1160 (1.2) clonidine 1,506 1,113 (1.4) -

Antagonists

Prazosin 0.35 0.28 (1.3) 0.43 (0.8)

5-methylurapidil 99 60 (1.7) 50 (1.9) WB4101 15 12 (1.3) 7.7 (1.9)

Rauwolseine 2,993 1,660 (1.8) —

Yohimbine 1,230 1,323 (0.9 —

*In saturation bindinα. studies the K _D and Bm _ax values for the binding αf [ I]HEAT by the wild type, Cys Phe and Ala Glu receptors was 96.2 and 1.9, 157.2 and 2.8, and 142.9pM and 1.4 pmol/mg, respectively

Characterisation of Constitutive Activation and Negative

Antagonism — Previous studies on constitutive activation and the ternary complex model (13) predict two phenotypic criteria. Constitutive activity where signalling occurs without the addition of agonists and an increased affinity and potency for agonists (H) are the two primary characteristics of the "activated" (i.e. G-protein precoupled) form of the receptor (R*) that is seen with the formation of the high affinity "ternary" complex, HR*G (13). Because of the higher affinity of the Cys 128Phe mutant for phenethylamine agonists and not for antagonists, the inventors investigated if this receptor might be constitutively active. As shown in Figure 2B,

the Cys128Phe mutant increased basal levels of inositol phosphates by approximately 40% and increased its potency for epinephrine by 10-fold as compared to wild type. However, no changes in potency could be seen when using an imidazoline full agonist, cirazoline. This finding is consistent with the observation that only the affinity of phenethylamines and not imidazolines was increased with the Cys 128Phe mutation (Figure 2A and Table 1). In separate experiments (Figure 3A, the inventors also measured the amount of IP ₃ production using a highly sensitive radioreceptor assay in which most of the IP ₃ released is measured and not just a fraction of a labelled phospholipid pool. This method of measurement showed increases in IP ₃ production of about 200% and could be blocked by the extended exposure (16-24 hours) to the αχ~ specific antagonist, phentolamine (lOOμM). Blockade of this constitutive response with prazosin indicates that this compound is a negative antagonist that can shift the equilibrium from R* by stabilising the "inactive" (R) or ground state conformation of the receptor. This is in

₂₉₃ agreement with previous studies on the Ala Glu mutation in which blockade of the constitutive inositol phosphate signal was achieved with 5 μM of prazosin added with the ₃ H-inositol 18-24 hours prior to assay (14). One may predict from computer simulations (15) that the binding affinity of a negative antagonist in this system to be lower. This would be due to a slower on-rate since negative antagonists bind to the inactive (ground)-state (R) of the receptor and with constitutively-active receptors the resident time in the R state compared to the R* state can be predicted to be shorter than with wild- type receptor. The present inventors found that prolonged incubation with antagonist is required to achieve blockade of constitutively active receptors. However, as shown in Table 1, no significant changes in affinity for any antagonists for either mutant were noted. This phenomena

can be explained in that these mutations are not intrinsically fully activated, therefore, the proportion of HR*G might not be large enough to observe this effect.

128

Effect of Cys Phe on the Activation of the PLA _∑ Pathway — Since the Cys 128Phe mutation appears to constitutively activate inositol phosphate release, the present inventors explored the effect this mutation had on another o-i-AR- coupled signalling pathway in the COS-1 system, arachidonic acid release via PLA2. However, basal levels of arachidonic acid release and changes in potencies

(EC ₅₀ ) could not be observed for either phenethylkamines or imidazolines when compared to the wild type receptor

(Figures 2C and 3B) . Thus, it appears that the Cys 128Phe mutation is selective in its constitutive activity, activating only the PLC pathway and not PLA ₂ .

Comparison of Cys 128 Phe with Ala 293 Glu — Since selective constitutive activity might be caused by differential efficacies of αχ-AR coupling to the PLC and PLA ₂ pathways or to different relative amounts of Gαq to Gαi in the COS- 1 cell line, the inventors sought to compare this mutant in the same cell line with another constitutively active mutation. The only other published constitutively active dχ-AR mutations are the Ala 293 mutations (14). To compare the phenotype of such mutants with the Cys 128Phe mutation, a hamster αis Ala 293Glu mutant was constructed which was shown previously to have the highest degree of constitutive activity of all amino acid substitutions in the 293 position (14). This mutant also recognised phenethylamines but not imidazolines with higher affinity (Table 1), and also increased basal IP ₃ levels (Figure

3A) . However, an important difference between the

Cys 128Phe and the Ala293Glu mutations was evident from dose-response studies. As shown in Figure 2B, by comparison with the wild-type receptor, the EC ₅₀ for activation of PI hydrolysis by the phenethylamine, (-) epinephrine, but not by the imidazoline, cirazoline.

decreased with the Cys128Phe mutant. However, with the

Cys 128Phe mutation the EC ₅₀ ^ror activation of arachidonic acid release was unchanged for both types of agonist (Figure 2C) . By contrast, with the Ala 293Glu mutant the EC ₅₀ for (-) epinephrine was decreased for both the PLC and P A ₂ pathways (Figure 2B and 2C). In separate experiments, basal measurements of arachidonic acid release also showed minimum activation with the Cys 128Phe mutant but increased levels with the Ala 293Glu mutant as compared with the wild-type receptor (Figure 3B) . The arachidonic acid release response could be blocked with prazosin or pertussis toxin, indicating that the increased arachidonic acid release by the Ala 293Glu mutation is due to selective precoupling to a pertussis toxin-sensitive G- protein and not to promiscuous coupling to additional pertussis-toxin insensitive G-proteins. This same argument holds for IP ₃ production in which constitutive activity is fully pertussis toxin-insensitive for both mutations (Figure 3A) . Thus, comparisons in the same cell system and with the same receptor subtype with different constitutive active mutations reveals selective constitutive activity for Cys 128Phe which is not due to an artefact of the system.

Induced-Conformational Pleiotropy — Since αi-ARs in the COS-1 cell system can activate both PLA ₂ and PLC by coupling to two distinct G-proteins, one pertussis toxin (PTX) - sensitive and one insensitive (6, 16, 17), respectively, these findings indicate that the Cys 128Phe mutation mimics a partially-activated conformation that precouples to a PTX-insensitive G-protein, and allows preferential activation of PLC through a phenethylamine.

The Ala 293Glu mutant, on the other hand, partially mimics a high-affinity complex that can promiscuously couple to both pathways, perhaps even favouring an interaction with P A ₂ , since the degree of potency shift for (-) epinephrine and the amount of basal activity is greater

for the PLA ₂ pathway. It is likely that the high agonist- affinity observed for phenethylamines with both mutant receptors is an intrinsic property of the mutants and does not require G-protein coupling. To test for possible modulation by G-proteins, receptor binding studies were also performed in the presence of guanine nucleotides. Gpp(NH)p (O.lmM) did not cause a change in the ligand binding affinity (Ki) or in the Hill coefficients. Hill coefficients were 0.8-1.0 for all binding curves without guanine nucleotides and indicated a best-fit to a single site model. These observations of guanine nucleotide- independent binding were also previously described with the Ala 293Glu mutation (14). Receptor binding profiles also could not be altered when membranes were prepared from transfected cells that were treated for 24 hours with 1 μg/ml of pertussis-toxin, which uncouples the receptor/G-protein interaction. In addition, in these studies agonist-binding was determined following transient transfection. This results in high level receptor expression and an excess of free receptor uncoupled from G-protein. Moreover, even with lower levels of receptor- expression as occurs in vivo, it is notoriously difficult with αi-adrenergic receptors to show the increase in agonist affinity with receptor/G-protein coupling that is readily apparent with other receptors, such as β- adrenergic receptors. Since the alanine residue involved in the Ala 293Glu mutation is located in the highly conserved carboxyl-end of the third intracellular loop and is contained in all adrenergic receptors, this could be an explanation for its precoupling to both pathways. This region may not be able to adopt a "fixed" conformation and, in fact, mutation of the Ala 293 residue in this region with any other amino acid results in constitutive activity (14). This suggests that the native receptor is "constrained", and that any mutation at that site releases this constraint. By

contrast, preliminary results from site-saturation studies of the Cys128 residue indicates that only selected residues are able to confer constitutive activity. Recent data has also suggested that the third intracellular loop may contain separate and distinct molecular determinants that divergently stimulate separate signalling pathways. For the β-AR, activation of Na+-H+ exchange is discrete from that for adenylate cycase activation, and involves coupling via distinct regions of the third intracellular loop to Gχ ₃ and Gs, respectively (18,19). The second intracellular loop has also been implicated in determining signalling specificity (20,21). Therefore, it is plausible that specific conformations of the receptor are needed for each distinct signalling pathways that utilize the second intracellular loop in conjunction with different regions of the third intracellular loop.

The inventors are the first to find that a G-protein coupled receptor can actually be induced into a conformation that constitutively signals only through a particular pathway. Thus, the Cys 128Phe mutation has induced a conformational change in the receptor structure that partially mimics the activated state for the PLC pathway, but not for the PLA ₂ pathway. This conformational change is associated with an increase in the affinity of the receptor for phenethylamine agonists. By contrast, the affinity of this receptor mutant for imidazolines is unaltered. This indicates that at least some of the residues binding imidazolines are distinct from those binding phenethylamines. In keeping with this latter conclusion, the binding affinity of imidazolines is also unaltered in the Ala 293Glu mutation, even though this mutant is constitutively activated for both PLC and PLA ₂ pathways. This implies that by contrast with phenethylamines, imidazolines bind to both the activated and ground state of the receptor with similar affinity. Importantly, however, although both mutations result in

receptor/G protein precoupling, the increase in receptor affinity for phenethylamine agonists is due to a direct effect on the effect on the ligand binding residues and not to an indirect effect on the ligand binding pocket that results from receptor/G-protein coupling. This is evident, as discussed above, from the failure of either pertussis-toxin or Gpp(NH)p, which uncouple receptor/G- protein interactions, to alter agonist affinity.

A possible mechanism of constitutive activation for the Cys 128Phe mutation could be due to the bulky Phe residue pushing the third transmembrane domain away from the second, thus, allowing a closer and tighter contact of the ASP 125 counterion to the protonated amine on the phenethylamine ligand. This transmembrane movement could mimic a partial conformation needed in full agonist induced activation. Imidazolines, on the other hand, might be binding further back in the binding pocket and their protonated amine might be located further from the counterion thus accounting for why many imidazolines are partial agonists and for the differential binding affinities for phenethylamines and imidazolines seen in these mutations.

Based on these findings with the Cys 128Phe mutation, one can predict that it should also be possible to develop a mutant that induces the receptor into a conformation that mimics the activated state for PLA ₂ and its coupled pertussis-toxin-sensitive G-protein, rather than for PLC and the pertussis-toxin-insensitive G-protein. Together with the Cys 128Phe mutation, such a mutant should be uniquely useful for ultimately designing signalling- specific drugs (agonists or negative agonists) that can induce the receptor to adopt a single activated conformation, thus, achieving pharmacological diversity through a single receptor subtype. Such compounds would not only allow the contribution of each distinct receptor-

activated pathway to be evaluated, but may also be an important new class of therapeutic agents. References

1. Hwa J., DeYoung M, Perez DM, and Graham RM. In: The Autonomic Nervous System. G Burnstock, Ed, Volume

VIII: The nervous control of the heart, Volume editors: J Shepherd and SF Vatner. Harvard Academic Press, 1994.

2. Cotecchia S, Schwinn DA, Randall RR, Lefkowitz RJ, Caron MG and Kobilka BK. PNAS 85: 7159-7163, 1988.

3. Sambrook J, Fritsch EF, and Maniaris T Molecular Cloning: A Laboratory Manual . Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 1989.

4. Perez DM, Piascik MT and Graham RM. Mol . Pharmacol. 40:876-883, 1991.

5. Bradford MM. Anal. Biochem.72:248-254, 1976.

6. Perez DM, DeYoung MB and Graham RM. Mol. Pharmacol 44:784-795, 1993.

7. Henderson R, Baldwin JM, Ceska TA, Zemlin F, Beckmann E and Downing KH. J. Mol. Biol. 213:899-

929, 1990.

8. Sung S-S, Riek P, Handschumacher M, Novotony J and Graham RM. FASEB 5:A804, 1991.

9. Strader CD, Sigal IS and Dixon RA, FASEB J. 3: 1832, 1989.

10. Ruffolo RR, Turowski BS and Patil PN, J. Pharm. Pharmacol. 29: 378-380,1977.

11. Ruffolo RR, Rice PJ, Patil PN, Hamada A and Miller DD. Eur. J. Pharmacol. 86:471-475, 1983. 12 Eason MG, Jacinto MT and Liggett SD. Mol. Pharmacol. 45:696-702, 1994.

13. DeLean A, Stadel JM and Lefkowitz RJ. Biol. Chem. 255:7108-7117, 1980.

14. Kjelsberg MA, Cotecchia S, Ostrowski J, Caron MG and Lefkowitz RJ. J. Biol. Chem. 267 (3) : 1430-1433,

1992.

15. Costa T, Ogino Y, Munson PJ, Onaran HO and Rodbard D. Mol. Pharmacol. 41:549-560,1992.

16. Lomasney JW, Hirakata A, Allen LF, Capel D, Proia AD, Caron MG and Lefkowitz RJ. Society for Neurosciences Abstract, Volume 18:255.1, 1992.

17. Burch RM, Luine A and Axelrod J. PNAS 83:7201-7205, 1986.

18. Barber DL and Ganz MB. J. Biol. Chem. 269:20607- 20612, 1992. 19. Voyno-Yasenetskaya T, Conklin BR, Gilbert RL,

Hooley R, Bourne HR and Barber DL. J Biol. Chem. 269:4721-4724, 1994. 20. Franke RR, Konig B, Sakmar TP, Khorana HG and Hoffmann KP. Science 250: 123-125, 1990. 21. Wong SK and Ross EM. J. Biol. Chem. 269:18968- 18976, 1994. 22. Allen LF, Lefkowitz RJ, Caron MG and Cotecchia S. PNAS. 88:11354-11358, 1991.

It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.