.
`0028-895X/93/081094-11$03.00/0 (cid:9)
`Copyright 0 by The American Society for Pharmacology arid Experimental Therapeutics •
`All rights of reproduction in any form reserved.
`MOLECULAR PHARMACOLOGY, 44:10E4-1104
`
`Structural arid Conformational Features Determining Selective
`Signal Transduction in the 03-Adrenergic Receptor
`
`NATHALIE BUN, LUC CAMOIN, BERNARD MAIdRET, and A. DONNY STROSBERG
`Institut Cochin de Géneitique Moléculaire, CNRS-UPR 0415, and Universite Paris VII, 75014 Paris, France (N.B., L.C., A.D.S.), and Laboratolre de
`Chimie Théorique, Université de Nancy I, 54506 Vandoeuvre Les Nancy, France (B.M.)
`Received April 23, 1993; Accepted September 11, 1993
`
`SUMMARY •
`With respect to the (31- and ß2-adrenergic receptors (ARs), the
`83-AR induces specific physiological effects in a few target
`tissues and exhibits atypical pharmacological properties that
`distinguish It unambiguously from its counterparts. Therefore,
`the 03-AR represents a suitable model to study the molecular
`mechanism responsible for receptor subtype selectivity and
`icecificity. Potent /33-AR ligands newly characterized in Chinese
`hamster ovary cells expressing the /33-AR were also evaluated
`in Chinese hamster ovary cells expressing 01- and 02-ARs and
`were classified into three groups according to their pharmaco-
`logical properties. Among the /31/02/133 agonists BRL 37344 and
`LY 79771 exhibit 03 selectivity in stimulating adenylyl cydase;
`among the /31/82 antagonists displaying 03 agonistic effects ICI
`201651 exhibits 03-AR binding selectivity, whereas among the
`01/82/03 antagonist clast bupranolol is this most efficient (but
`not selective) 83-AR antagonist. The structures of these ligands
`
`were simulated and compared using computer-generated molec-
`ular modeling. Structure-activity relationship analysis indicates
`that potent or selective 03-AR compounds, in addition to pos-
`sessing a pharmacophore common to all 13-AR ligands, contain
`a long and bulky alkylamine substituent moiety, which is able to
`adopt and exchange extended and stacked Conformations. Com-
`puterized three-dimensional models of the p1-, 02-, and 03-AR
`binding sites show that more bulky amino acid side chains point
`inside the groove of the 81 and 02 sites, compared with the fi3
`site, in a region implicated in signal processing. The long alkyls-
`mine chain of compounds behaving as /31//32 antagonists and
`133 agonists may thus adopt either a stacked conformation, in the
`encumbered ßl- and 82-AR sites, leading to antagonistic effects,
`or an extended conformation in the less encumbered 03 site,
`thus interacting with specific residues implicated in signal trans-
`duction.
`
`Sympathetic stimulation via humoral (adrenergic) and neu-
`ronal (noradrenergic) pathways induces a number of physiolog-
`ical effects, such as modulation of heart rate, vascular tonus,
`bronchospasm, and glucose and lipid metabolism. Lands et ca.
`(1) first subdivided the ft.-AR-mediated effects into 01 and /32,
`on the basis of the rank order of potency of epinephrine and
`norepinephrine in different tissues. Since this classification,
`
`This work was supported by grants from the Centre National de la Recherche
`Scientifique, the Institut National de Is Sante et de is Recherche Medicate, the
`Ministkre de la Recherche et de l'Espace, the Université Paris V. Bristol-Myers-
`Squibb Company (Princeton, NJ), the Fondation pour Is Recherche Medicate,
`the Association pour le Développement de Is Recherche cur le Cancer, and the
`Ligue Nationale Française contra le Cancer.
`
`many clinically active drugs, mimicking or blocking the effects
`of natural hormones, have been synthesized and shown to
`discriminate between 81- and ß2-AR-mediated effects.
`In the, following years, however, a number of novel com-
`pounds revealed atypical /3-AR properties in various tissues.
`BRL 37344 was thus characterized as a potent thermogenic
`and lipolytic /3-AR agonist in rat adipose tissue (2) and SR
`68611A as an atypical /3-AR agonist mediating relaxation in
`precontracted guinea pig ileum (3). Several 81//32 antagonists
`displayed atypically low binding affinities in these tissues as
`well as low potencies in inhibiting responses mediated by these
`novel compounds, thus suggesting the existence of a novel 8-
`AR pharmacological profile. However, partly because of its low
`
`ABBREVIATIONS: ft-AR, ft-adrenergic receptor; CHO, Chinese hamster ovary; CH0-8, Chinese hamster ovary cells expressing the (3-adrenergic
`receptor; IA, intrinsic activity; ICYP, iodocyanopindolol; MD, molecular dynamics; RMS, root mean square index; TM, transmembrane domain; BRL
`37344, (RR,SS)-(±)-4-(2'42-hydroxy-2-(3-chloropheny0ethylamlno]propy0phenoxyacetate sodium salt sesquihydrate; bucindolol, 242-hydroxy-3-([2-
`(3-indoly1)-1,1-dimethylethyl]amino)propoxylbenzonitrile hydrochloride; bupranolol, 1-(2-chloro-5-methylphenoxy)-3-[(1,1-dimethylethyl)amino1-2-pro-
`panol; CGP 12177A, (±)-4-(34-butylamino-2-hydroxypropoxy)benzimidazol-2-one; CGP 20712A, (±)-(2-(3-carbamoy1-4-hydroxyphenoxy)ethylamino]
`-344-(1-methy1-4-trifluoromethy1-2-imidazolyl)phenoxy]-2-propanolisopropylamino-2-propanol hydrochloride; cimaterol, 2-amino-5-(1-hydroxy-2-[(1-
`methylethy0aminOjethyljbenZOnitrile; clenbuterol, 4-amino-3,5-dichloro-a-[(1,14imethylethyl)amino]benzenemethanol; ICI 118551, o-(±)-1-(7-meth-
`ylindan-4-yloxy)-3-isopropylaminobutan-2-ol; ICI 201651, (R)-4-(2-hydroxy-3-phenoxypropylaminoethoxy)-N-(2-methoxyethyl)phenoxyacetic acid; LY
`79771, (RS)-(±)-4-(2'-[(2-hydroxy-3-phenylethy0amino]buty0benzyl alcohol; SM 11044, L-3-(3,4-dihydroxypheny1)-N43-(4-flqoropheny0propyl]serine
`pyrrolidine amide hYdrobromide; SR 58611A, (RS)-N-U2S)-7-ethoxycarbonylmethoxy-1,2,3,4-tetrahydronaphth-2-y1)-(2R)-2-(3-chloropheny0-2-hy-
`droxyethanamine hydrochloride; PBS, phosphate-buffered saline; HEPES, 4-(2-hydroxyethy0-1-piperazineethanesultonic acid.
`1094
`
`Sawai Ex. 1006
`Page 1 of 11
`
`

`

`;
`
`1
`
`affinity for available fl-AR radioligands and primarily because
`of the lack of suitable tools to study its expression among a
`population of conventional fl-ARs, this atypical 0-AR remained
`difficult to characterize unambiguously by a classical pharma-
`cological apprbach, and some inconsistencies were described
`between drug affinities identified in binding studies and those
`measured in functional assays e(4).
`After the initial cloning of the 02- (5) and 01-ARs (6), a third
`gene, coding for a novel 0-AR subtype (the (33-AR) sharing 51%
`and 46% identity with the human 01- and 132-AR amino acid
`sequences, respectively, was cloned from a human genomic
`library (7). The presence of human 03-AR mRNA transcripts '
`has been demonstrated in human fat tissues as well as in gall
`bladder and colon biopsies (8), and evidence for a functional
`03-AR in human fat cells has been recently shown by lipolysis
`stimulation studies (9). Functional /33-ARs, cloned from either
`human (7), mouse (10), or rat (11, 12) tissues, were character-
`ized in transfected CHO cells, and their pharmacological pat-
`tern indicated that.the 03-AR is closely related to the atypical
`0-AR in adipose tissues (13, 14). However, minor differences
`between the human and rodent 03-ARS as well as between
`atypical 0-Alts from different tissues have led some authors to
`question whether these are actually the same pharmacological
`subtypes (11, 12, 15). (cid:9)
`•
`To settle this point, we performed a systematic pharmaco-
`logical analysis in CHO-ß3 (human) and CHO-133 (mouse) using
`a large panel of 0-AR ligands, and we showed (i) that both the
`human and the rodént 03-:ARs display well defined ribarniaco-
`logical properties that distinguish them unambiguously from
`the in- and /32-ARs, (ii) that the 03-AR is the prototype of the
`atypical ß sites described, in a few, target tissues (adipose, gut,
`and cardiac tissues) where it., induces specific physiological
`effects, and (iii) that some compounds (BRL 37344; bucindolnli
`bupranolol, CGP 12177A, cimaterol, ICI 201651, LY 79771, SR
`58613.A, and SM 11044) exhibit potent affinities or activities
`in CHO-03.1 These atypical and specific properties make the
`/33-AR a model receptor to study the molecular basis of subtype
`selectivity, using these new pharmacological tools.
`In this study, we analyzed the selectivity of the subtype by
`evaluating pharmacological receptor binding and adenylyl cy-
`clase activation properties of P-AR ligands in CHO cells ex-
`pressing human 01-, 02-, or 03-ARs. Results led us to classify
`compounds into pharmacological classes, and the structure-
`activity relationship of these ligands was analyzed using MD
`. simulations. Structural features of 03-efficient agonists and
`antagonists were examined to define a putative pharmacophore,
`as well as to provide new insights into the molecular mechanism
`responsible for the 03-AR potency and selectivity.
`•
`Materials and Methods •
`
`Chemicals. Bucindolol and nadolol were provided by Bristol-Myers
`Squibb (Princeton, NJ). CGP 12177A, CGP 20712A, alprenolol, and
`oxprenolol were gifts from Ciba-Geigy Corporation (Basel, Switzer-
`land). ICI 118551 and ICI 201651 were obtained from Imperial Chem-
`ical Industries (Macclesfield, England). Cimaterol and LY 79771 were
`donated by American Cyanamid (Pearl River, NY) and Lilly Research
`Labs (Indianapolis, IN), respectively. Clenbuterol was obtained from
`Roussel Uclaf (Romainville, France). Pindolol and cyanopindolol were
`
`N. Blin, C. Nahmias, M. F. Drumare, and A. D. Strosherg. The /13-adrenergic
`receptor a single subtype responsible for atypical ft-adrenergiereceptor-mediated
`effects. Submitted for publication.
`
`Structure-Activity Relationship of /13-AR Uganda (cid:9)
`
`1095
`
`provided by Sandoz (Basel, Switzerland). (±)- and (-)-Bupranolol were
`gifts from Schwarz Pharma (Monheim, Germany), BRL 37344 was
`obtained from SmithKline Beecham Pharmaceuticals (Epsom, Eng-
`land). SM 11044 and SR 58611A were given by Sumitomo Pharmaceu-
`ticals (Osaka, Japan) and Sanofi-Midy (Milano, Italy), respectively.
`(-)-Isoproterened and propranolol were purchased from Sigma Chem-
`ical Co. (St. Louis, MO).
`Cell culture. Subclones of CHO cells stably transfected with human
`$1-, /32-, or /33-ARs were grown as described previously (7, 16).
`Receptor binding assays. Preconfluent cells were harvested by
`treatment with Versen-EDTA (Seromed) and were washed with Hanks'
`balanced salt solution supplemented with 1 mm ascorbic acid and
`buffered with 20 rum HEPES to achieve a pH of 7.4. Aliquots of 10°
`cells were incubated with (-)-(3-12°I1ICYP (2000 Ci/mmol; Amersham,
`England), in the absence or presence of competitor, in a buffered 500-
`p1 final volume with 0.1% (w/v) bovine serum albumin (Sigma) and 4
`AIM desipramine (Sigma). The reaction was performed for 45 mm at
`37', with shaking, in the dark. After dilution with ice-cold PBS, pH
`7.4, cells were immediately filtered and extensively washed over glass
`fiber disks (VVhatman GF/C) that had been presoaked with 0.3%
`polyethyleneimine (Sigma). Radioactivity was measured in a LKB 1282
`7-radiation counter..
`Saturation experiments were performed with ICYP concentration's
`ranging from 5 to 500 pm for the /31- and #2-ARs and from 50 to 5000
`pm for the /33-AR. Nonspecific binding was determined in the Presence
`of 2 Am (±)-propranolol for CHO-ß1 and CHO-02 or 100 ph4 (-)-
`isoproterenol for CHO-ß3. Competition experiments were perfonned.
`with ICYP concentrations of 50 pm for the 451 and 452 subtypes and 1
`nm for the /33 subtype and various concentrations of competitor ranging
`from. 1 pm to 100 Am. .Ligand lipophilicity indexes ,(log #F) were
`calculated using the TSAR software (Oxford Molecular, Oxford, Eng-
`land). .
`Aderiyly1 cyclase binding assays. Because forskolin directly stim-
`ulates the catalytic subunit of adenylyl cyclase and displays greater
`• efficacy and potency whèn its. catalirtic domain. interacts with the. a,
`subunit of the G protein (17), forskolin binding experiments, were
`performed with adherent transfected CHO-/3 in the absence or presence
`of ft-AR ligands. (cid:9)
`•
`Preconfluent cells in six-well dishes (g..-1.2 x 10° cells/well) were
`washed twice with 2 ml of ice-cold PBS, added to 1 ml of ice-cold
`Ham's F12 medium buffered with 20 MIN HEPES, pH 7.4, and kept on
`ice for 30 min before the binding study. Cella were incubated at 4° for
`1 hr, with slow shaking, in 500 prof buffered (12-81.11forskolin (20-35
`Ci/mmol; New England Nuclear), in the absence or presence of non-
`tritiatedTorskolin or ft-AR ligands. Cells were then washed three times
`with 2 ml of PBS and dissolved in 1 ml of 1 N NaOH for 30 min at 37'
`before the homogenate was counted in a LKB-Wallac 1410 scintillation
`counter.
`Cholera toxin ADP-ribosylates G., irreversibly blocking it GTPase
`• activity and maintaining the stability of the aecyclase complex in (
`way that is independent of receptor occupancy. Cells were treated with
`cholera toxin (2 ag/m1 in culture medium; Sigma) for 5 hr at 37' before
`measurement of forskolin binding at 4, a temperature that allows
`stabilization of the transient complex but probably leads to underesti-
`mation of the maximal complex association ût 37°.
`Adenylyl cyclase stimulation assays. CHO-fil, CHO-a, and
`CHO-453 were grown to preconfluence in six-well dishes (m..1.2 x 106
`cells/well). After washing with 1 ml of Ham's F12 medium buffered
`with 20 mm HEPES, 01'7.4, and supplemented with 1 ram ascorbic
`acid and 1 mM 3-isobutylmethylxanthine (Sigma), cell monolayera were
`incubated for 30 mm at 37° in ml of buffer, in the absence (basal
`level, 5-25 pmo1/105 cells) or in the presence of 10 Am •(-)-isoproterenol
`(maximal stimulation mediated by /3-AR, 170-400 pmo1/108 cells), 25
`pM forskolin (direct adenylyl cyclase stimulation, 420-850 pmo1/105
`cells), or 1 pm to 100 Am ligand. The reaction was stopped by one wash
`with 1 ml of PBS and immediate addition of 500 prof 1 N NaOH. After
`a period of 20 mm at 37', dissolved cells were collected, buffered with
`
`Sawai Ex. 1006
`Page 2 of 11
`
`

`

`1096 (cid:9)
`
`sun at at.
`
`1 N acetic acid, and centrifuged at 3000 x g for 10 mm at 4'. The total
`cAMP amount contained in an aliquot of supernatant was determined
`using the Amershani PliicAMP assay or r23I]-cAMP scintillation
`proximity assay.
`For inhibition studies of adenylyl cyclase stimulation, cella were
`preincubated with the antagonist at 37 for 10 min before addition of
`a reference agonist [i.e., (--1-isoproterenoll at its K.... concentration (5
`rim) and incubation for a subsequent 20-min period. • •
`Data analyses. The data were expressed as the means t standard
`errors of at least three independent experiments performed in duplicate,
`except for forskolin binding data, which resulted from two experiments
`only. Saturation experiments were computer analyzed with the EBDA
`program (Biosoft-Elsevier, Cambridge, UK) using the Scatehard plot
`representation. 1C60 'and EC60 parameters obtained from binding com-
`petition experiments or adenylyl cyclase activation or inhibition exper-
`iments were determined using a computerized, iterative, nonlinear,
`least squares curve-fitting program (Inplot 4.0, written by H. J. Motul-.
`sky, GraphPad Software, San Diego, pm. IC6.6 values measured in
`binding competition or cyclase antagonism experiments were corrected
`(Ki value) according to the method of Cheng and Pnisoff. The IA of a
`compound was measured relative to the.reaxinial cYclase Stimulation
`obtained for (—)-isoproterenol.. Ligands ithat 'possessed IA v'alima of
`<0.90 were defined ae Partial agonists.
`Molecular modeling. The conformations of the arylethanolamine- • •
`related comPounds' that were incorporated into the analysis werePb-
`taMed using the BIOSYM molecular Modeling software (BIOSYM
`Technologies, Inc.,. San Diego, CA) on a Silicon .Graphics workstation.
`.1 Initial structures were built using the,. Insight, U Builder module,
`which directly produced :coarse three-dimensional starting structures.
`To mimic ionization it neutral pH, an sp3+ hybridization was assigned
`' to the' aniine of thieraiiii alkyl chain; increasing the meileCular electro-
`static total charge by +1.
`Energy minimization and MD simulations were performed with the
`Insight II Discover Module, using the consistent Valence force field. All
`calculations were , performed for in uacuo conditions, using, in the
`description ofthe poulombic interaction a distance-dependent dielectric
`constant fixed to 3.5 to avoid. formation of intramolecular Salt bridges.
`The first step of medeling.-densisted• of Miniiniiing the stnicture
`previously cimatructed, to find a local energy minimum on the Potential
`energy hypersurface of the molecule. Calculations were performed
`according to several algorithms commonly used in molecular mechanics
`minimization for choosing descent directions, namely steepest descent,
`conjugate gradient, and Newton-Raphson methods.
`The second step of the conformational sampling procedure consisted
`of recording MD trajectories. By solving the equations of motion for a
`system of atoms,, MD has an advantage in that it is not restricted to
`harmonic motion about a single minima but allows molecules to cross
`energy barriera and explore Other stable conformations. 'Molecular
`conformers were sampled during a 1-nsec MD 'trajectory at 300°K. A
`time step 'of 5 fsec Wit8 used, and the system was equilibrated for.1 psec.
`conformation was stored each 5 psec, so that 200 conformations were
`--fecorded by the end of the MD simulation.
`All molecular conformations were compared using the Analysis
`module of InSight IL Conformational similarities were evaluated by
`calculating the RMS of deviation between heavy atoms for each possible
`pair of these 200 structures and by plotting the associated cluster graph.
`A thieshold value of'4 A was selected to plot the RMS evolution, so
`that numerous boxlike areas appeared along the diagonal, representing
`group of structures whose small RMS deviations (<1 A) and closeness
`in time suggested that they may belong to the same conformational
`family. Conformational representatives extracted from each family
`were compared for each compound, as well as between different ligands,
`using a superimposition procedure.
`
`Results and Discussion
`
`• (cid:9)
`
`•
`
`Selectivity of fl-AR Uganda in CHO-f11, CHO-a, and CH0-113
`Although fi-AR overexpression has been reported to affect
`adenylyl cyclase sensitivity (18-20), it offers the opportunity
`
`to thoroughly characterize receptors such as the /33-AR, for
`which high affinity radiolabeled antagonists have not been
`developed thus far. The human 131-, /32-, and 113-ARs overex-
`pressed in CHO cells displayed selectivity profiles for catechol-
`amines and reference fi-AR ligands that were consistent with
`those described in tissues characterized by Prevailing /31-, /32-,
`and ß3-AR populations (16). The presence of six. additional
`carboxyl-terminal residues in the sequence of the human ß3-
`AR; resulting from splicing of an intron in the. corresponding
`gene, has been reported (21, 22), but a recent pharmacological
`comparison failed to. detect any difference between the 408-
`.
`and 602-residue forms of this receptor (23). (cid:9)
`Because the apparent affinity of agonists at ICYP binding
`sites may be influenced by varying degrees of internalization,
`we verified that the lipophilicity indices (log P) of the ßl- and
`P2-AR agonists tested in CHO-fi were higher than that of the
`ICYP radioligand. For the. /33-AR, no bias in measurement ,qf
`IC1 values is expected, because this receptor subtype does not
`become sequestered (23). (cid:9)
`. (cid:9)
`• . . (cid:9)
`• •
`(cid:9) in each
`d
`Because differences in, the level of receptor exPrisse
`CH0-0. subchine [190,271. ± 16,796 sites/cell in CH0431 (hu-
`man), 74,885 t 22,461 ,.sites/cell ,in di40.432 ,(human), and
`108,785 ± 5,988 sites/cell in CHO-ß3 (hunian)]. and differences
`in receptor subtype coupling Might interfere with the measurer
`ment of c yclase,stiinulation potency, the ,,stoiehipmetrY pf re-
`ceptor-G1-adenyly1 cyclase interactions was assessed in CHO-
`Qi10-/32,. and PHO-ß3..Becattie isoproterenot-stiiatilated
`forsfcolie binding measurements revealed approximately the
`
`same number offorsitolin binding sites:in, the three types of
` toxin 'stimulatiein (Fig. 1), it
`cells as well as after chelers'
`
`appeared that all of,thé cholera toxin sensitive G protein cou
`pled adenylyl _Cyplase, existing n. CHOFA: Was. Stint-040 o'by.
`isoproterenol. Moreover, it apPeered that couPlingerfficiency of
`the three 13-AR subtypes shi5Uld not Wei the 'adenylyl erelgiàe
`stimulation potency measurements, thus allowing'comptirison
`of the /3 selectivity of ligands based on K,„.1 values.
`The selectivity of li-ARfigands exhibiting iiiteiesting 'phar-
`macological ProPertiees idthé /33 sitewas evaluated in cHO.:
`#1, CHO-‘32, and CH0-#3 'and led to the classification of the
`compounds into three group's, is., agónists at the three fi sites,'
`
`,T
`
`75000 -
`
`50000 -
`
`25000 -
`
`cno-02
`Fig. 1. Measurement of the rate of coupling in CH0-131 , CH0432, and
`CH0433. Forikolin , binding was evaluated in intact CH0431, CHO-02,
`and CH0-133 preincubated A or hot (0) with 2 pg/rni cholera tokin for 5
`hr at' 370 and Incubated with 100 OA isoproterenol for 1 hr at 40. Values
`are. the mean ± standard error of two separate experiments performed
`in duplicate.
`
`Sawai Ex. 1006
`Page 3 of 11
`
`

`

`#1/$2 antagonists displaying $3 agonistic properties, and an-
`tagonists for the #1-, /32-, and 03-ARs (Table 1).
`$1/#2/$3 agonists: The $3-AR was characterized by its
`potency for a class of arylethanolamine agonists that were
`initially found to be potent and selective activators of lipolysis
`and thermogenesis at the atypical #-ARs described in white
`and brown adipose tissues (Table 1). BRL 37344, the most
`representative compound of this class (2, 24), was a full agonist
`in CH0-#1 and CH0-#3, with partial agonistic effects (IA =
`0.8) in CH0-#2, and exhibited a 10-fold f13-AR selectivity,
`relative to the #1- and #2-ARs.. LY 79771; an activator of the
`metabolic rate in dogs (25), stimulated adenylyl cyclase with 5-
`and 17-fold greater potency in CH0-#3 than in CH0-#1 and
`CH0-#2, respectively. Thus, titypical 0-AR compounds, which
`. are potent in inducing thermogenesis in brown adipose tissue
`and in increasing the rates of cellular metabolism such as
`lipolysis in white adipose tissue, appeared to be $3-selective
`
`SR 58611A and SM 11044, which were relaxant agents in the
`precontracted rat colon (3) and guinea pig ileum (26); respec-
`tively, were "rather #2/#3-selective" agonists in cHo-g. SR
`58611A, the potent' and most selective compound of the phen-
`yfethanolaminotetralirie class, , induced rat colon relaxation
`with an EC E,0 of 3.5 nm (3), comPaied with a 1<„,,' of 25 nm in
`stimulating CH0-133 adenylyl cyclase. The SM 11044 func-
`tional selectivity order in guinea pig tissues, i.e., ileum relaxa-
`tion (atypical #-AR) > trachea or lung' relaxation ($2.:AR).>
`atrium rate increase (31-AR), was consistent with the selectiv-
`ity of this drug in CH0-#1; CF10-/32, and CHO-M. Although
`possessing rather low affinities at the $3 site (Ki range of 1-6
`Am), these compounds were efficient enough (I.C.q values be-
`tween 10 and 100'nm) to induce. ß3.:AR'-atediated functional
`relaxation in smooth muscle tissues.
`Cimaterol and plenbuterol, reported to induce protein accre-
`tion and to increase skeletal muscle mass in vivo (25), were
`"rather ßl/$2-selective" agonists in CHO-A In addition, cim-
`aterol exhibited high efficiency in stimulating the cyclase in
`CH0-#3, in agreement with its ability to potently activate
`lipolysis in rat white adipose tissue (25). ,
`#1//12 antagonists/$3 agonists. Among the 01/#2 antago-
`nists displaying #3 agonistic properties, some exhibited high
`binding affinities and agonistic potencies in CH0-#3 (Table 1).
`Bucindolol, described as a high affinity nonselective (3-AR
`antagonist (27), displayed the same binding affinities for P-
`end 02-AFts expressed in CHO cells (.1(1 values of 0.2 nm and
`0.1 nm, respectively) and possessed full and potent (K.e. = 7
`nm) $3 agonistic effects. ICI 201651, the in vivo metabolized
`form of ICI D7114 that is able to selectively stimulate brown
`adipose tissue actiVitSn (28), was .ti_wesk antagonist at the gl-
`and $2-AR sites (k; values of 0.55 .Am and 2.86 Am, respectively)
`but 'a potent full agonist in CHO-113 (K.ct.
`20 nm). ICI 201651
`was the most important compound of this class, exhibiting a
`#3 selectivity in binding affinities.
`CGP 12177A, oxprinolol, pindolol; and alprenolol were 10-
`100-fold less potent in stimulating the $3-AR than were the
`previously mentioned full agonists and, except for alprenolol,
`demonstrated partial agonistic effects. Pindolol maintained its
`cyclase stimulation potency when a cyano group was added to
`the indol function of the molecule (K„c, value of 153 nm,
`compared with 174 nM) but displayed an IA that increased
`from 0.55 to 0.82. These compounds bound to the $1- and #2-
`
`Structure-Activity Relationship of 03-AR Uganda (cid:9)
`
`1097.
`
`ARs with 10-100-fold higher affinities than those measured in
`CHO-M.
`Nadolol and propranolol were #1/#2 antagonists exhibiting
`weak (K„c, values in the micromolar range) and partial agonistic
`effects in C}{0-p3. In agreement with these results, Bond and
`Clarke (29) reported a biphasic effect for nadolol and propran-
`olol in antagonizing the isoproterenol-induced relaxation of
`precontracted guinea pig ileum strips.
`#1/#2/t33 antagonists. The third category of ligands in-
`cluded antagonists such as the in-selective CGP 20712A, the
`02-selective ICI 118551, and bupranolol (Table 1).
`Kaumann (30) earlier reported that heart atypical 13 agonistic
`effects were antagonized by 1 Am bupranolol but not propran-
`olol. Although (—)-bupranolol appeared to be the best antago-
`nist available to characterize the #3-AR (IC, value of 50 nm),
`its receptor binding order of selectivity in CH0-# was $2-AR
`> #1;AR > #3-AR.
`The selectivity profiles for these antagonists were CGP
`20712A = bupranolol > ICI 118551 in CHO-fil, bupranolol
`ICI 118551 > CGP 20712A in CH0-#2, and bupranolol > ICI
`118551 > CGP 20712A in CH0-$3. (cid:9)
`• _
`Taken together, our data show that #3-selective agonists
`(BRL 37344 and LY 79771), /13-selective (ICI 201651) and $3-
`potent (bucindolol and CGP 12177A) agonists that exhibit #1/
`$2 antagonistic properties, a /33-potent antagonist (bupranolol),
`and $1- and $2-selective antagonists (CGP 20712A and ICI
`118551, respectively) are useful tools that can help' to distin-
`guish $3-AR-mediated physiological effects from those me-
`diated by conventional #1- and fl2,ARs. To date, only 1uI]
`ICYP and [3HICGP 12177A have allowed direct characteriza-
`tion of tissue #3-ARs (13). In addition, radiolabeling of ICI
`201651, which' exhibitectbinding- selectivity towardsthe $3 site,
`should provide a new pharmacological tool for the characteriza-
`tion ,of the 03-AR in tissues. More selective compounds for the
`$3-AR, however, remain to be found, and analysis of the struc-
`ture-activity relationships for this large variety of compounds
`should help in determining the structural features responsible
`for the #3 potency and selectivity of ligancLs.
`Structural Features of 03-AR Uganda
`Fine specificity of the ligand recognition mechanism
`for G protein-coupled receptor's. Norepinephrine stimu-
`lated adenylyl cyclase in CHO-M with a 1600-fold higher
`potency, relative to dopamine, which is its metabolic precursor
`and is specific for dopaminergic receptors (7). Although thee
`compounds are structurally related, #-hydroxylation of thb
`alkylamine chain appears to be important fór ligand-receptor
`recognition. Indeed, this modification creates an asymmetrical
`center, leading to isornerization of the molecule, and this polar
`0-hydroxyl group may interact with an electrophilic center and
`form a hydrogen bond with an amino acid side chain inside the
`receptor groove.
`Similarly, a- and #-Alts were distinguished upon the basis of
`the potency order of isoproterenol, relative to norepinephrine
`and epinephrine, three catecholaininergic structures that
`are closely related. Indeed, isoproterenol differs from (nor)-
`epinephrine by a (di)methyl substitution, which increases steric
`bulk and lipophilicity at the end of the alkylamine chain, and
`the substitution of a methyl group on the protonated amine
`moiety of norepinephrine corresponds, in thermodynamic cal-
`culations, to a loss of 6-7 kcal (31). These modifications seem
`
`Sawai Ex. 1006
`Page 4 of 11
`
`

`

`• '1
`TABLE 1 (cid:9)
`,.
`.. . (cid:9)
`.
`. (cid:9)
`. (cid:9)
`.. (cid:9)
`Comparison of the pharmacological properties of human di-, 62-,-and 63-AR expressed In CHO cells (cid:9)
`•
`Binding competition assays were carried out with intact cells for 45 min at 37° in the presence if r°51]1CY12, as.chiscribed in Materials Ind Methods. Adenylylcyclase stimulation assays were pertained with-Intact cells
`preincubated or not with 5 flu isoproterenol for 10 min and incubated with drugs for .30 min at 37°, Conc,entration-response curves were fitted using least squares regression analysis. and binding comoetiton (K,) and
`adenylyi cyclase stimulation (K.,:) constants were deduced. IA was calculated for each drub relative to isdprotirenol4nduced1118)(11TO cAIR accumulation. Values are means t standard errors of at least three independent
`experiments performed in duplicate. Ligands were classified as /31/(12/03 igonists (more ß3-selective, more'02/03-ielective. or more P1432-seledlive agontsts). tlyil2 antagonists/03 agonists. or ßl/ß2/ß3 antagonists. All
`• (cid:9)
`' (cid:9)
`• (cid:9)
`:. (cid:9)
`data were obtained using similar experimental conditions. (cid:9)
`.
`*- . (cid:9) * . (cid:9)
`. (cid:9)
`. : (cid:9)
`f - (cid:9)
`.- (cid:9)
`. (cid:9)
`:Human /12-AR (cid:9)
`• Human /31-Alt; (cid:9)
`: (cid:9)
`Aden* cydase sting:Won (cid:9)
`Aden* cyclase stindation (cid:9)
`' (cid:9)
`IA
`IA (cid:9)
`
`,
`
`f (cid:9)
`
`Nixing lc,
`
`Human #3-All
`Adenyfyl cyclase stlnidatico
`
`Bincirg
`
`Baba lc
`
`NJ (cid:9)
`
`nu (cid:9)
`
`' au. (cid:9)
`
`MI . (cid:9)
`
`nm
`
`Kra (cid:9)
`. DM (cid:9)
`
`1,750 t 310 (cid:9)
`
`112 t 28 (cid:9)
`86± 8 (cid:9)
`• 36,500 t 13,400 12,000 t soo. (cid:9)
`190 t 20 (cid:9)
`18,100 t 1,700 (cid:9)
`. 0.64 t 0.15 (cid:9)
`. (cid:9)
`
`.
`
`.t 0.11 1 -120 t 380 (cid:9)_ , (cid:9)
`1.30 (cid:9)
`1.42 ± 0.30 (cid:9)
`0.96 t 0.07 ' 187 t 26 (cid:9)
`1.50±0.21 • 4,100 t 200 (cid:9)
`1.20 t 0.06 (cid:9)
`
`60. t 9 (cid:9)
`
`. (cid:9)
`
`e
`
`. (cid:9)
`287 t 92
`15 t 3 (cid:9)
`0.80 t 0.04 (cid:9)
`- 177 t 47 (cid:9)
`18 ± 3 (cid:9)
`555 t 71 (cid:9)
`325 t 121 - 0.22 t 0.03 (cid:9)
`• 36 t 19 - 0.87 t 0.07 6,640 t 960 . (cid:9)
`25 t 5 (cid:9)
`84±10 (cid:9)
`1.03±0:08. 1,300 t 200 (cid:9)
`62 t 6 . (cid:9)
`17 t 3 (cid:9)
`,0•57 t 0.002 0.98±0.03 4,700 t 1,710 - (cid:9)
`1.0 t 0.2 • 0.91 t 0:02' 1,100 t 200 (cid:9)
`1,050 t 130_ (cid:9)
`
`61/621/33 agonis:ts (cid:9)
`BRL 37344 (cid:9)
`LY 79771 (cid:9)
`SR 58611A (cid:9)
`SM 11044 (cid:9)
`Cimaterol (cid:9)
`Clenbuterol (cid:9)
`61/62 antagonists/63 agonists
`Bucindolol (cid:9)
`ICI 201651 (cid:9)
`CGP 12177A (cid:9)
`Oxprenoloi (cid:9)
`Pindoloi (cid:9)
`. (cid:9)
`Cyanopindolol (cid:9)
`8.8 t 0.2 (cid:9)
`Alprenotol (cid:9)
`40 ±6 - (cid:9)
`Nadoloi (cid:9)
`6.3 t 1.0 (cid:9)
`Prairanolol (cid:9)
`-
`61/62/63 antagonists (cid:9)
`Antagonist (cid:9)
`1.7 t 0.3 (cid:9)
`(-)-Bupranolol (cid:9)
`Antagatist: (cid:9)
`2.4± 0.5 . (cid:9)
`(t)-Bupranolol (cid:9)
`Antagonist- (cid:9)
`•120 t 3° (cid:9)
`ICI 118551 (cid:9)
`Antagonist (cid:9)
`. 1.5 t 0.2 - (cid:9)
`CGP 20712A. (cid:9)
`'Results reported by Nahmlas et al. (10), with IA expressed relative to isoproterenol.
`b Data reported by Emodne at at: (7), with IA expressed relative to.norepinephrine maximal cyciaser stimulation. (cid:9)
`'Data reported by Tate at al, (16).
`
`'190 t 30 (cid:9)
`
`L0.20 t 0.04 (cid:9)
`549 t 200 (cid:9)
`0.9 t 0.1 (cid:9)
`5.4 t 1.3 • (cid:9)
`3.4 t 0.7 (cid:9)
`
`- (cid:9)
`
`• (cid:9)
`
`Antagonist . (cid:9)
`Antagonist • (cid:9)
`Antagonist ' (cid:9)
`Antagonist ' (cid:9)
`Antagonist (cid:9)
`Angatonist. (cid:9)
`Antagonist • (cid:9)
`Antagonist: (cid:9)
`Antagonist (cid:9)
`
`0.10 t 0.03 ' ... Antagonist (cid:9)
`2,860 t 750 ' ._ Antagonist (cid:9)
`4 t 2' (cid:9)
`.- Antagonist (cid:9)
`, Antagonist (cid:9)
`1.5 t 0.4 (cid:9)
`••
`.2.3±0.9 ' •• 1 Antagonist (cid:9)
`..
`. (cid:9)
`Antagonist (cid:9)
`- Antagonist (cid:9)
`-1.5 t 0.3 (cid:9)
`- 14 t 5 (cid:9)
`Antagonist (cid:9)
`0.7 t 0.3 (cid:9)
`Antagonist (cid:9)
`
`0.4 ± 0.1--- Antagonist (cid:9)
`-
`0.5 t 0.1 I. ' Antagoni