Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Determination of the intrinsic activity of receptor ligands remains a chief issue in the pharmacological characterization of the activity profile of a compound. The resolving capacity by which compounds with various degrees of intrinsic activity can be differentiated is codetermined by the receptor's G protein coupling state. Fine-tuning of the assay system is often necessary to measure the entire window of pharmacological activities, ranging from efficacious agonism via neutral, silent antagonism to inverse agonism. Although monitoring of different degrees of ligand-mediated positive efficacy has become accessible in some instances (Strange, 2002), the detection of inverse agonism implies that a receptor is constitutively active. This prerequisite can be observed under native conditions [i.e., for wild-type (wt) benzodiazepine, muscarinic, and -opioid receptors; see Schutz and Freissmuth, 1992]; however, for most G protein-coupled receptors either a high receptor expression level and/or introduction of activating mutation is necessary to yield detectable constitutive receptor activity (Kenakin, 1997; Pauwels and Wurch, 1998). Several activating mutations have been described, such as Ala²⁹³Glu in the 3ICL of the _1B-adrenoceptor (_1B-AR; Kjelsberg et al., 1992).

The dopamine D₂ receptor is of particular interest because it constitutes the principally targeted receptor by neuroleptic drugs (Strange, 2001). These compounds comprise a wide variety of chemical structures that act as clinically effective agents in the treatment of schizophrenia and are considered antagonists at D_2-like receptors (Strange, 2001). Former studies suggested that haloperidol acts as an inverse agonist by increasing prolactin release from GH₄C₁ pituitary cells expressing a D_2short receptor (Nilsson and Eriksson, 1993; Nilsson et al., 1996) or by decreasing basal arachidonic acid release from a Chinese hamster ovary (CHO) ^pro-5 cell line stably expressing a D_2long receptor isoform (Nilsson et al., 1998). More recent reports evaluating larger series of neuroleptic drugs at either wt or mutant D₂ receptors also suggested that most of these antagonists may act as inverse agonists (Hall and Strange, 1997; Bullock et al., 2001; Wilson et al., 2001). The amplitude of inverse agonism by these compounds seems to be comparable for almost each of the tested neuroleptic drugs. The putative absence of a difference in the ligand's amplitude of inverse agonism may perhaps be due to the low resolving capacity of the investigated assay systems. Thus, it is of interest to develop more sensitive assay systems to measure the magnitude of inverse agonism at D₂ receptors.

Analysis of chimeric receptors has proven to be useful to study the structural basis of the function of G protein-coupled receptors. In pioneering work, Kobilka et al. (1988) described chimera of ₂-AR and ₂-AR to demonstrate that transmembrane domains (TM) V and VI and the 3ICL are determinants of coupling to adenylyl cyclase and that the subtype selectivity of ₂- and ₂-AR ligands is strongly determined by TM VII. A similar approach has been used for the _2A-AR by exchanging its 2ICL and 3ICL for the corresponding portions of either a ₂-AR or serotonin 5-HT_1A receptor to investigate receptor desensitization (Liggett, 1991; Eason and Liggett, 1996; Jewell-Motz et al., 1998). Analyses of _2A/_2C-AR chimera in their 3ICL indicated that agonist-mediated phosphorylation of ₂-ARs is highly dependent on the conformation of the 3ICL (Jewell-Motz et al., 2000).

In the present study, whole-cell [³⁵S]GTPS binding experiments were performed in a first attempt on digitonin-permeabilized CHO-K1 cells, stably expressing a Thr³⁴³Ser D_2short receptor to maximally preserve the integrity of the receptor/G protein coupling. Inverse agonism was detected, for instance with nemonapride, but its amplitude was weak (about 20%). To enhance the magnitude of the inverse agonist response, chimeric D₂/_1B receptor constructs with different portions of the _1B-AR were prepared. The wt _1B-AR and mutants in its 3ICL alanine²⁹³ position have previously been reported to display various degrees of constitutive activity (Kjelsberg et al., 1992). One particular chimeric receptor construct (D₂/_1B Ala²⁷⁹Glu 3ICL), obtained by exchange of the 3ICL of the D₂ receptor for that of an _1B-AR and containing an activating mutation in its 3ICL (Ala²⁷⁹Glu equivalent to the Ala²⁹³ position in _1B-AR), displayed a large amplitude (about 60%) of inverse agonism in the presence of a G₁₁ protein. This chimeric receptor construct was further characterized with different agonists and putative antagonists using both kinetic Ca²⁺ and inositol phosphates (IP) responses and with respect to the determination of the ligands' intrinsic responses.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of Chimeric Dopamine D₂ Receptor Constructs. The short splice form of the human wt dopamine D₂ receptor (R.C. 2.1.DA.02; GenBank accession number: S69899) was modified by exchanging its 2ICL, 3ICL, and/or its C-terminal intracellular portions by those of the human _1B-AR (R.C. 2.1.ADR.A1B; GenBank accession number: U03865; see Fig. 1). A modified overlap extension polymerase chain reaction approach was applied by using overlapping complementary oligonucleotide primers allowing the junction of the D₂ receptor and _1B-AR portions without addition of restriction sites (Wurch et al., 1998). Incorporation of the Ala²⁷⁹Glu mutation was performed using a QuickChange site-directed mutagenesis kit, as described by the supplier (Stratagene, La Jolla, CA). The chimeric D₂/_1B receptor constructs were inserted into a pCR3.1 mammalian expression vector and fully sequenced on an ABI Prism 310 genetic analyzer using a Big Dye Terminator cycle sequencing ready reaction kit confirming the respective chimeric sequences (Applied Biosystems, Foster City, CA).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Schematic representation of various chimeric D2/1B receptor constructs. , transmembrane domain; , extra-membrane portion.

Cell Culture and Transfection Procedures. Cos-7 and CHO-K1 cell lines were cultured, respectively, in Dulbecco's modified Eagle's medium and Ham's F-12 medium, each one supplemented with 10% heat-inactivated fetal calf serum. Transfection was performed by electroporation [Bio-Rad Gene pulser apparatus (250 mV, 250 µF); Bio-Rad, Hercules, CA] using 10 µg of indicated receptor plasmid in either the presence or absence of 10 µg of recombinant G protein plasmid. The culture was prolonged in complete culture medium containing 1% dimethyl sulfoxide under conditions dependent on the monitored response. The CHO-K1 cell line stably expressing a mutant Thr³⁴³Ser D_2short receptor was generated upon dilution of transfected cells (10- to 1000-fold) and selection in complete Ham's F-12 medium containing 1.25 mg of geneticin/ml. The selected cell line showed 5.01 ± 0.04 pmol/mg of protein of specific [³H]nemonapride binding sites on cellular membranes.

Membrane Preparation and Radioligand Binding Experiments. Cells were scraped mechanically in 10 mM Tris-HCl and 0.1 M EDTA (pH 7.5) and centrifuged for 10 min at 45,000g. The pellet was homogenized in the same buffer and centrifuged under similar conditions. Membrane preparations were diluted in 50 mM Tris-HCl, 120 mM NaCl, and 5 mM KCl (pH 7.4). Incubation mixtures consisted of 0.4 ml of membrane preparations (50-150 µg of protein), 0.05 ml of [³H]nemonapride, and 0.05 ml of compound for inhibition or 10 µM of (+)-butaclamol to determine nonspecific binding. The reactions were stopped after a 1-h incubation at 25°C by addition of 3.0 ml of ice-cold 50 mM Tris-HCl (pH 7.7) and rapid filtration over Whatman GF/B glass fiber filters (Whatman, Clifton, NJ) using a Brandel harvester (Brandel, Inc., Gaithersburg, MD), washed, and radioactivity was counted (Pauwels et al., 2001a). Scatchard analysis was performed as described (Pauwels et al., 2001a) using concentrations of radioligand ranging from 3 pM to 3 nM. The protein level was estimated with a dye-binding assay using a Bio-Rad kit; bovine serum albumin was used as a standard (Bradford, 1976).

Measurement of [³⁵S]GTPS Binding Responses on Digitonin-Permeabilized Cells. CHO-K1 cells stably expressing a human Thr³⁴³Ser D_2short receptor were grown in 24-well plates at a density of 75,000 cells/well in complete Ham's F-12 medium containing 1.25 mg/ml geneticin. The [³⁵S]GTPS binding procedure on permeabilized cells was adapted from Wieland et al. (1995). Briefly, cells were kept in permeation buffer (50 mM triethanolamine-HCl, 5 mM MgCl₂, 1 mM EDTA, 150 mM NaCl, and 10 µM digitonin, pH 7.4) for 20 min at 37°C. Upon removal of the permeation buffer, GTPS binding buffer was applied (50 mM triethanolamine-HCl, 1 mM MgCl₂, 150 mM KCl, 30 µM GDP, and 0.2 nM [³⁵S]GTPS, pH 7.4) in either the absence or presence of compound, and cells were incubated for 30 min at 37°C. Cells were washed twice with ice-cold Hank's balanced salt solution. Five hundred microliters of scintillation liquid was added to extract radioactivity, and scintillation counting was performed on a TopCount microplate reader (PerkinElmer Life Sciences, Boston, MA).

Measurement of Inositol Phosphates Formation. Cos-7 cells expressing the indicated receptor construct and G protein combination (for cloning details, see Pauwels et al., 2001b) were loaded with [³H]myoinositol (4 µCi/well of a 24-well plate) for 48 h in Dulbecco's modified Eagle's medium supplemented with 2% dialyzed fetal calf serum. Cells were washed with 1.0 ml of controlled salt solution and incubated for 1.5 h at 37°C in 1.0 ml controlled salt solution containing 10 mM LiCl either in the presence or absence of compound. The reaction was stopped by the addition of 0.25 ml of sample buffer (30 mM Na₂B₄O₇ and 3 mM EDTA), and the fraction of total [³H]IP was separated from the other ³H derivatives by chromatography on an anion exchange AG1-X8 resin as described (Wurch et al., 1999). pIC₅₀ and pEC₅₀ values were defined as the concentration of compound at which 50% of its own, respectively maximal inhibitory and stimulatory effect was obtained.

Measurement of Intracellular Ca²⁺ Responses. CHO-K1 cells expressing the indicated receptor construct and G protein combination were assayed for Ca²⁺ responses at 24 to 48 h post-transfection upon a 1-h loading with the Ca²⁺ indicator dye Fluo-3 (2 µM). Dopaminergic ligands were assayed for their Ca²⁺ response, as previously described (Pauwels et al., 2001a). Data were expressed in arbitrary fluorescence units (AFU) and were not converted into Ca²⁺ concentrations. Fluorescent readings were made every 2 s for a 3-min time period using a fluorometric image plate reader (Molecular Devices Corp., Sunnyvale, CA). E_max values were defined as the ligand's maximal high-magnitude response versus that obtained with 10 µM DA. pEC₅₀ values correspond to a ligand concentration at which 50% of its own E_max value was measured. Antagonists were preincubated for 10 min before DA (10 µM) to prevent the high-magnitude Ca²⁺ phase in the antagonist-bound state (Pauwels et al., 2001a). Antagonist capacity (%) of DA-mediated high-magnitude Ca²⁺ response was defined as the property of the ligand (1 µM, added 10 min before 10 µM DA) to antagonize the high-magnitude response. This was calculated as the surface area between the DA and ligand condition for a period of 4 min upon addition of DA.

Statistical Analysis. Statistical significance was determined by comparisons performed for the [³H]nemonapride binding data (expressed in pK_i for the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct versus wt D_2short receptor), [³⁵S]GTPS binding data (ligand-mediated versus basal [³⁵S]GTPS binding level), and IP formation data (ligand-mediated versus basal IP level). All statistical comparisons were based on a two-tailed Student's t test. Relationships between antagonist capacities were calculated using a Pearson correlation test.

Materials. The ABI Prism 310 genetic analyzer and Big Dye Terminator cycle sequencing ready reaction kit were from Applied Biosystems. The pCR3.1 expression was purchased from Invitrogen (Carlsbad, CA). The QuickChange site-directed mutagenesis kit was obtained from Stratagene. Cos-7 and CHO-K1 cells were obtained from the American Type Culture Collection (Manassas, VA). [³⁵S]GTPS was from Amersham Biosciences (Les Ulis, France). [³H]Nemonapride (85 Ci/mmol) and 2-[³H](N)-myoinositol (20 Ci/mol) were obtained from PerkinElmer Life Sciences (Les Ulis, France). DA, raclopride, haloperidol, fluphenazine, ()-sulpiride, chlorpromazine, (10,11)-dihydroxy-N-n-propylnorapomorphine (NPA) enantiomers, risperidone, (+)-butaclamol, olanzapine, ()-epinephrine, prazosin, clonidine, bromocriptine, and domperidone were from Sigma/RBI (Natick MA). (+)-UH 232 was from Tocris Cookson, Inc., Ballwin, MO). Bromerguride was from Schering (Berlin, Germany). Nemonapride, tropapride, olanzapine, ziprasidone, and S 14066 were synthesized at the Centre de Recherche Pierre Fabre (Castres Cédex, France).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Inverse Agonism at the Mutant Thr³⁴³Ser Dopamine D_2short Receptor. [³⁵S]GTPS binding experiments were performed on intact CHO-K1 cells, stably expressing a D_2short receptor carrying a Thr³⁴³Ser mutation in the distal BBXXB motif (Lys³⁴⁰-Lys-Ala-Thr-Gln) of its 3ICL, upon digitonin (10 µM) permeation. DA (10 µM) strongly stimulated (424 ± 46% versus basal) the binding of [³⁵S]GTPS. The putative antagonists nemonapride, haloperidol, and (+)-butaclamol significantly reduced at a 1 µM concentration basal [³⁵S]GTPS binding by 15 to 19% (Fig. 2). The inactive enantiomer ()-butaclamol (1 µM) did not affect the basal [³⁵S]GTPS binding response. In contrast, (+)-UH 232 (1 µM) significantly stimulated the [³⁵S]GTPS binding response by 39% (Fig. 2). Expression of recombinant G_i1, G_i3, or G_o proteins in the CHO-Thr³⁴³Ser D_2short cell line did not enhance the magnitude of decreased basal [³⁵S]GTPS binding response mediated by the putative dopamine antagonists investigated in this study. The compounds behaved under similar experimental conditions as silent ligands in CHO-K1 cells stably expressing a wt D_2short receptor (not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   [35S]GTPS binding responses on digitonin-permeabilized CHO-K1 cells, stably expressing a mutant Thr343Arg D2short receptor. Cell permeation and labeling was performed as described under Materials and Methods. Maximal stimulation of [35S]GTPS binding by 10 µM DA corresponded to 194 ± 9 fmol/mg of protein. Bar graphs are calculated as a percentage of ligand-mediated (1 µM) stimulation or attenuation of basal [35S]GTPS binding response (36 ± 2 fmol/mg of protein). Results are expressed as the mean values ± S.E.M. of nine independent experiments, each experimental data point performed in triplicate. (, p < 0.001 versus basal [35S]GTPS binding response; Student's t test)

Coupling of Various Chimeric D₂/_1B Receptor Constructs to the Phospholipase C Pathway. Although no measurable stimulation of IP formation could be observed for the wt D_2short receptor with 10 µM DA in the absence of an exogenous G protein, a weak Ca²⁺ response (896 ± 194 AFU) was apparent (Table 1). This Ca²⁺ response could be enhanced by coexpression with a G₁₁ protein and, in particular, with a G_q/o protein. This latter response was similar to the strong Ca²⁺ response (11703 ± 573 AFU) of the ()-epinephrine-stimulated (10 µM) wt _1B-AR, independently of the coexpressed G protein subunit. To identify portions of the _1B-AR necessary for its coupling to cognate G_q/11 proteins and to enhance the amplitude of inverse agonism, the intracellular portions (2ICL, 3ICL, and C-terminal end) of the D₂ receptor were exchanged for the corresponding domains of the _1B-AR (Fig. 1). Only the introduction of the 3ICL of the _1B-AR yielded a 17-fold enhanced DA-mediated Ca²⁺ response compared with the wt D_2short receptor. The addition of an Ala²⁹³ to Glu mutation in the distal portion of the _1B-AR's 3ICL (respectively, Ala²⁷⁹Glu position in the chimeric D₂/_1B 3ICL receptor construct) affected the Ca²⁺ response by 34 ± 6% (Table 1). The 2ICL and the C-terminal intracellular portion or a combination of the three _1B-AR portions were unable to restore a DA-mediated Ca²⁺ response (Table 1). Coexpression with a G₁₁ protein slightly increased (Table 1) the DA-mediated Ca²⁺ responses of the chimeric D₂/_1B 2ICL and D₂/_1B 2ICL + Ala²⁷⁹Glu3ICL + C-term receptor constructs, although less than observed with the wt D_2short receptor. Further experimental characterization was performed with the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct.

View this table: [in this window] [in a new window]   TABLE 1 Ca2+ responses of wt dopamine D2short receptor, 1B-AR and various chimeric D2/1B receptor constructs in either the absence or presence of G11 and Gq/o proteins DA- (10 mM) or ()-epinephrine- (10 mM, for the wt 1B-AR) mediated Ca2+ responses were measured in transfected CHO-K1 cells, as described under Materials and Methods. Data correspond to mean ± S.E.M. values for a minimum of three independent transfection experiments.

Binding Properties of the Chimeric D₂/_1B Ala²⁷⁹Glu 3ICL Receptor Construct. Saturation binding analysis to a membrane preparation containing the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct, using [³H]nemonapride as a radioligand, indicated the presence of a single population of high-affinity binding sites with an equilibrium dissociation constant and maximal binding capacity of 45.6 ± 14 pM and 3.69 ± 0.73 pmol/mg of protein, respectively (Table 2). Transient expression of the wt D_2short receptor yielded a similar dissociation constant for [³H]nemonapride (K_D, 47.1 ± 0.9 pM) but a 2.5-fold lower maximal binding capacity (B_max, 1.48 ± 0.07 pmol/mg of protein; Table 2). The binding profile for a series of putative dopamine antagonists resembled that of the wt D_2short receptor (Table 2). Slight but significant differences (2.6- to 3.6-fold) were observed for bromerguride, (+)-butaclamol, and ()-sulpiride. An 11- and 22-fold increased binding affinity was observed, respectively, with the agonists DA and ()-NPA at the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct compared with the wt D_2short receptor. The higher agonist binding affinities were not longer observed with the chimeric D₂/_1B3ICL receptor construct without the Ala²⁷⁹Glu mutation (Table 2). The ₁-AR ligands ()-epinephrine, clonidine, and prazosin (Schwinn et al., 1995), at a concentration of 10 µM, did not demonstrate specific binding to the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct, as was also the case for the wt D_2short receptor (Table 2).

View this table: [in this window] [in a new window]   TABLE 2 Binding affinities for a series of dopaminergic and adrenergic ligands at the chimeric D2/1B Ala279Glu 3ICL receptor construct compared to the wt dopamine D2-short receptor pKi values were determined on Cos-7 cellular membranes transiently expressing either the chimeric D2/1B Ala279Glu 3ICL receptor construct or the wt D2 short receptor using [3H]nemonapride (0.14 nM), as described under Materials and Methods. Values between brackets correspond to the chimeric D2/1B3ICL receptor construct. Competition curves yielded Hill coefficients close to 1.00 (0.93  nH  1.27). Monophasic Scatchard analysis yielded dissociation constants (46.0 ± 1.42 and 47.1 ± 0.9 pM) and maximal [3H]nemonapride binding capacity (3.69 ± 0.73 and 1.48 ± 0.07 pmol/mg of protein) for respectively chimeric D2/1B Ala279Glu 3ICL and wt D2short receptor constructs. Data are the mean ± S.E.M. of three independent experiments, each experimental data point performed in duplicate. Statistical analysis was performed for comparison of the ligands; pKi values between the D2/1B Ala279Glu 3ICL and wt D2short receptor constructs using a Student's t test.

Ca²⁺ Responses by the Chimeric D₂/_1B Ala²⁷⁹Glu 3ICL Receptor Construct. Both DA and ()-epinephrine were able to induce Ca²⁺ responses at the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct as well as at the wt parental D_2short receptor and _1B-AR (Table 3). The Ca²⁺ kinetic data indicated a typical profile for both the wt D_2short receptor and _1B-AR. A high-magnitude Ca²⁺ response was observed with ()-epinephrine, as previously found (Pauwels et al., 2001a) for DA at the wt D_2short receptor with a G_q/o protein; the onset time to obtain the maximal Ca²⁺ pulse was 16 to 20 s (shown for DA in Fig. 3A). Thereafter, the signal decreased to reach 53 ± 2% of its maximal amplitude at 3 min. DA at the wt _1B-AR, regardless the G protein subtype, initially showed a sharp Ca²⁺ peak at 6 to 8 s, followed by a shoulder of a lower Ca²⁺ magnitude at 20 to 30 s (Fig. 3C). Similar results were obtained with ()-epinephrine (not shown). The Ca²⁺ kinetic profile of the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct in both the absence and presence of a G₁₁ or G_q/o protein was qualitatively similar to that of the wt D_2short receptor; the onset time of maximal response by both DA and ()-epinephrine was delayed to 35 ± 2 s (shown for DA in Fig. 3B).

View this table: [in this window] [in a new window]   TABLE 3 Emax and/or pEC50 values for Ca2+ responses by a series of dopaminergic ligands and ()-epinephrine as obtained with the wt D2short receptor, wt 1B-AR, and chimeric D2/1B Ala279Glu 3ICL receptor constructs Concentration-Ca2+ response curves were performed for DA and ()-epinephrine in CHO-K1 cells cotransfected with the indicated receptor and G protein combination, as described under Materials and Methods. The magnitude of maximal stimulation in AFU is indicated in Table 1 for each experimental condition. Maximal Ca2+ responses were measured for bromocriptine, (+)-NPA, and (+)-UH 232 and were calculated as a percentage versus that obtained with 10 µM DA. pEC50 and Emax values correspond to the mean or mean ± S.E.M. values for a minimum of three independent transfection experiments.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Time-dependent DA-mediated Ca2+ responses at wt D2short, chimeric D2/1B Ala279Glu 3ICL receptor construct, and wt 1B-AR in either the absence or presence of a G11 and Gq/o protein in CHO-K1 cells. Cotransfection of 10 µg of either wt D2short receptor (A), chimeric D2/1B Ala279Glu 3ICL receptor construct (B), or wt 1B-AR (C) plasmid, and 10 µg of empty plasmid or plasmid containing a G11 or Gq/o protein was performed as described under Materials and Methods. DA (10 µM) was applied at minute 0 and Ca2+ responses were measured every 2 s for 3 min, as described under Materials and Methods. Curves illustrate a representative experiment of a minimum of 12 independent experiments.

11

q/o

11

11

q/o

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Correlation between the ligands' magnitude of preventing effect of the high-magnitude Ca2+ response by putative dopaminergic antagonists at the wt D2short receptor coexpressed with a Gq/o protein compared with either chimeric D2/1B Ala279Glu 3ICL or D2/1B 3ICL receptor constructs. Antagonists (1 µM) were added 10 min before DA (10 µM). High-magnitude Ca2+ responses were measured as described under Materials and Methods. Ca2+ responses were quantified as a percentage remaining of the surface area of the Ca2+ response obtained with DA (10 µM) alone for a period of 4 min, as previously described (Pauwels et al., 2001a). Surface areas are expressed as mean values ± S.E.M. of three independent transfection experiments for the D2/1B Ala279Glu 3ICL and D2/1B 3ICL receptor constructs. Mean values for the wt D2short receptor were taken from Table 1 in Pauwels et al. (2001a).

Inositol Phosphates Responses at the Chimeric D₂/_1B Ala²⁷⁹Glu 3ICL Receptor Construct. In the absence of recombinant G proteins, DA (10 µM) stimulated (836 ± 65%) the formation of IP at the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct, whereas tropapride (1 µM) did not significantly affect the basal signaling. Coexpression with various recombinant G proteins (Fig. 5) demonstrated a significant enhancement (+86 to + 226%) of the basal IP response with G₁₁, G_q, and G₁₅ proteins, a maximal effect being observed with the G₁₁ protein. Tropapride dose dependently (pIC₅₀, 9.10 ± 0.01) attenuated basal IP production by 59 ± 9% (Fig. 6A). Tropapride (0.1 µM) also antagonized the DA-mediated IP response in an insurmountable manner (Fig. 6B). A stereoselective inverse agonist response was observed for butaclamol; the (+)-enantiomer was as efficacious as tropapride, whereas ()-butaclamol (10 µM) did not affect the basal IP response (Table 4). Clozapine, olanzapine, and raclopride (1 µM) behaved as less efficacious inverse agonists with a maximal attenuation of the IP response by, respectively, 31, 67, and 71% (versus maximal inhibition by tropapride 1 µM; Table 4). The potencies (pIC₅₀) of clozapine, olanzapine, and raclopride to decrease basal IP levels were, respectively, 6.81 ± 0.25, 8.21 ± 0.27, and 7.81 ± 0.08 (Fig. 6C). Other putative dopaminergic antagonists attenuated basal IP formation with a similar magnitude as tropapride in contrast to bromerguride and (+)-UH 232, which yielded efficacious positive agonism (respectively, +63 and + 88% versus 10 µM DA; Table 4). The bicyclic derivative ziprasidone did not affect the basal IP formation (+5 ± 3%; p > 0.05; Student's t test). It potently and competitively antagonized both tropapride-mediated (pK_B, 8.52 ± 0.27; Fig. 6A) and DA-mediated (pK_B, 7.61 ± 0.27; Fig. 6B) stimulation and inhibition of basal IP formation. Inverse agonist responses were not longer observed at the chimeric D₂/_1B 3ICL receptor construct in which the Ala²⁷⁹Glu mutation was suppressed (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Influence of recombinant G protein expression on the IP response mediated by a chimeric D2/1B Ala279Glu 3ICL receptor construct. Cos-7 cells were transfected with 10 µg of chimeric D2/1B Ala279Glu 3ICL receptor plasmid and 10 µg of either empty plasmid or the indicated G protein plasmid and assayed for IP formation, as described under Materials and Methods. , basal IP level; , DA-mediated (10 µM) IP level; (), tropapride-mediated (1 µM) IP level. Data are expressed as the mean disintegrations per minute ± S.E.M. per 105 cells of three to five independent experiments, each experimental data point performed in triplicate. Statistical analysis was performed using a Student's t test for comparison of basal IP level in the presence of each G protein versus the absence of G protein (p < 0.01) (a) and tropapride-mediated versus basal IP level (p < 0.01) (b).

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Dose-dependent attenuation of basal and DA-mediated IP formation by a series of inverse agonists and antagonism by ziprasidone using a chimeric D2/1B Ala279Glu 3ICL receptor construct in the copresence of a G11 protein. Cos-7 cells were cotransfected with 10 µg of chimeric D2/1B Ala279Glu 3ICL receptor construct and 10 µg of G11 protein plasmid and assayed for IP formation, as described under Materials and Methods. A, attenuation of basal IP formation by tropapride in either the absence () or presence of ziprasidone (0.1 µM; ). B, stimulation of IP formation by DA in either the absence () or presence of tropapride (0.1 µM; ) or ziprasidone (0.1 µM; ). C, attenuation of basal IP formation by raclopride (), olanzapine (), or clozapine (). Data are calculated as a percentage of basal (A and B) or of tropapride-mediated (1 µM) attenuation of basal (C) IP formation and expressed as the mean ± S.E.M. of three independent experiments, each experimental data point performed in triplicate.

View this table: [in this window] [in a new window]   TABLE 4 Inositol phosphates response for a series of putative dopaminergic antagonists mediated by a chimeric D2/1B Ala279Glu 3ICL receptor construct in the copresence of a G11 protein in Cos-7 cells Cos-7 cells were transfected with 10 µg of chimeric D2/1B Ala279Glu 3ICL receptor construct and 10 µg of G 11 protein plasmids and assayed for IP formation, as described under Materials and Methods. Basal IP level corresponded to 9476 ± 598 dpm/105 CHO-K1 cells (n = 12), mean maximal attenuation of IP formation by tropapride (1 µM) was 3287 ± 310 dpm/105 cells (n = 12), and mean maximal stimulation by DA (10 µM) was 42677 ± 3162 dpm/105 cells (n = 12). Data are calculated as a percentage of the tropapride-induced (1 µM) inhibition of basal IP formation except for bromerguride and (+)-UH 232 (underlined), which are expressed versus the maximal stimulation by DA (10 µM). Ligands were investigated at a concentration of 1 µM, a concentration effective for both antagonism of the DA-mediated high-magnitude and reversal of the low-magnitude Ca2+ responses (Figs. 5 and 6; Pauwels et al., 2001a). Data are expressed as the mean ± S.E.M. of three to eight independent experiments, each experimental data point performed in triplicate. Statistical analysis was performed on the ligand-induced versus tropapride-induced (1 µM) IP level (based on data expressed in disintegrations per minute per well) by using a Student's t test.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Chimeric D₂/_1B receptor constructs were prepared to enhance constitutive receptor activation and to differentiate between the amplitude of inverse agonism produced by different dopamine receptor ligands. The construct D₂/_1B Ala²⁷⁹Glu 3ICL displayed the largest constitutive activation and was highly sensitive to inverse agonism for most of the tested putative dopamine antagonists, which behaved as efficacious inverse agonists. Two atypical neuroleptic drugs, olanzapine and clozapine, as well as the benzamide-derived compound raclopride displayed partial inverse agonist activities, whereas ziprasidone behaved as a silent antagonist. The results based on the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct extend the resolution capacity of model systems with a lower level of constitutive D₂ receptor activation (Fig. 2; Hall and Strange, 1997; Wilson et al., 2001) and strongly suggest that the inverse agonist feature is common to most neuroleptic drugs. Two compounds, (+)-UH 232 and bromerguride, displayed positive agonism and have to be considered as partial agonists rather than antagonists.

The D₂ receptor state was switched to a more active conformation to increase the basal receptor activation R to R* by exchanging the 3ICL of the _1B-AR in a D₂ receptor backbone and containing the activating mutation Ala²⁷⁹Glu in the distal BBXXB motif. Simulations of molecular dynamics revealed that the distal portion of the _1B-AR's 3ICL folds into an -helix and faces the -helical proximal portion of the same loop (Scheer et al., 1996). The mutant Ala²⁹³Glu _1B-AR structure indicated that the Glu²⁹³ residue is involved in hydrogen bonds with both Tyr²²⁷ and Lys²³¹ at the N-terminal portion of the 3ICL (Scheer et al., 1996) in contrast to the wt Ala²⁹³ residue. It is likely that this intramolecular interaction is also present in the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct as the entire _1B-AR's 3ICL was exchanged. One can further postulate a conformational link between the Glu²⁷⁹ residue and the Asp-Arg-Tyr (DRY) motif in the D₂ receptor-derived 2ICL, as previously shown for the wt _1B-AR (Scheer et al., 1997). The presence of the _1B-AR's 3ICL combined with the Ala²⁷⁹Glu mutation in a D₂ receptor backbone is apparently sufficient to modify the overall conformation of the protein to generate constitutive activity. The ligand binding characteristics at the chimeric construct, although being not fully identical, largely resembled that of a wt D_2short receptor, suggesting that the investigated ligands are predominantly interacting with the receptor's TMs rather than with the _1B-AR-derived 3ICL. Furthermore, the rank order of dopamine antagonists to prevent the DA-mediated high-magnitude Ca²⁺ response in the antagonist-bound receptor state fitted that observed for the wt D_2short receptor. This indicates that the recognition pattern of the chimeric D₂/_1B receptor construct by the dopamine antagonists is functionally conserved and, therefore, appropriate to investigate their pharmacological responses.

The binding affinities for the agonists DA and ()-NPA were increased at the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL construct compared with the wt D_2short receptor. The increased agonist binding affinity is probably due to the formation of a high-affinity chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor state (R*) generated by the Ala to Glu mutation since the chimeric D₂/_1B 3ICL receptor did not demonstrate modified binding affinities. A comparable increased binding affinity was also observed for DA at the mutant Thr³⁴³Ser D_2short receptor (Wilson et al., 2001), suggesting that this mutant receptor behaves as a constitutively active receptor as confirmed by the [³⁵S]GTPS binding responses reported in this study. The chimeric D₂/_1B receptor constructs containing the 3ICL of the _1B-AR displayed functional characteristics of phospholipase C activation, such as induction of Ca²⁺ responses and stimulation of IP formation. Coexpression with three G proteins of the G_q family indicated the following rank order of enhanced basal IP formation: G₁₁ > G₁₅ > G_q. This may suggest a preferential coupling of the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct to a G₁₁ protein. It cannot be excluded that a difference in the expression level between these G proteins may also explain this G protein effect. The unavailability of an antibody that recognizes the three G proteins makes this investigation difficult.

Increasing the concentration of G proteins and consequently the amount of R*G complex has previously been reported to enhance constitutive receptor activity [i.e., muscarinic M₁ and M₃ receptors (Burstein et al., 1997) and Thr³⁷³Lys _2A-AR (Pauwels et al., 2000)]. Nevertheless, over expression of either a G_oCys³⁵¹Ile or a chimeric G_q/o protein in combination with a mutant Thr³⁴³Arg D_2short or Thr³⁷²Arg D_2long receptor did not yield constitutive receptor activation (this study; Pauwels et al., 2001b,c). The enhanced constitutive activity of the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct was also associated with an increased potency for the agonists DA and ()-epinephrine and an increased efficacy for the partial agonists (+)-NPA and bromocriptine and the putative antagonists bromerguride and (+)-UH 232. By monitoring forskolin-stimulated cAMP formation, Hall and Strange (1997) suggested (+)-UH 232 to be a weak partial inverse agonist at the stably transfected wt D_2short receptor rather than a truly neutral antagonist. Both enantiomers of UH 232 have been characterized as partial agonists by measuring the extracellular acidification rate at the D_2long receptor stably transfected in CHO-K1 cells (Coldwell et al., 1999). We found (+)-UH 232 as a partial agonist by measuring [³⁵S]GTPS binding responses in digitonin-treated CHO-K1 cells transfected with the Thr³⁴³Ser D_2short receptor. (+)-UH 232 also behaved as a partial agonist in CHO-K1 cells transiently coexpressing a related mutant Thr³⁴³Arg D_2short receptor and a G_oCys³⁵¹Ile protein (Pauwels et al., 2001b). It cannot be excluded that these minor differences in intrinsic activities for UH 232 reflect effector-dependent features. (+)-UH 232 could achieve with the D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct a maximal positive response of nearly 90%, close to that of DA. Therefore, this compound together with bromerguride are different from the dopamine antagonists investigated here, most of them acted as efficacious inverse agonists irrespective to their chemical structure. Clozapine displayed a partial inverse agonist response reaching about 30% to that of tropapride. This compound acted as an inverse agonist at the mutant Thr³⁴³Arg D_2short receptor with a tendency to be less efficacious than haloperidol (Wilson et al., 2001). In other experimental systems, such as sensitization of adenylate cyclase by mutant rat Thr³⁴⁴Arg D_2short receptor (Bullock et al., 2001) and potentiation of forskolin-stimulated cAMP accumulation by wt D_2long receptor (Kozell and Neve, 1997), clozapine also tend to behave as a submaximal inverse agonist. In contrast, clozapine exhibited efficacious inverse agonist responses at a chimeric D₁/D₂[1-4,7] receptor construct (containing the [TMV-3ICL-TMVI] portion of a D₁ receptor in a D₂ receptor backbone) like (+)-butaclamol and haloperidol, although the ligand binding profile of this chimeric D₁/D₂ receptor construct was modified compared with the wt D₂ receptor (Kozell and Neve, 1997). The benzamide derivative raclopride yielded as well partial inverse agonism (~70% versus that of 1 µM tropapride) at the chimeric D₂/_1B Ala²⁷⁹Glu 3ICL receptor construct. This compound has previously been reported as a silent antagonist by measuring the prolactin release response in GH₄C₁ cells transfected with a D_2short receptor, whereas haloperidol displayed inverse agonism (Nilsson et al., 1996). The magnitude of the inverse agonist response is mainly determined by the amount of the basal, ligand-independent activation of the receptor (Kenakin, 2001). Hence, compounds that display weak inverse agonism may be better observable in an expression system in which the receptor displays a high level of constitutive activity (i.e., D₂/_1B Ala²⁷⁹Glu 3ICL construct). It is possible that the detection of weaker inverse agonists by measuring basal prolactin release in GH₄C₁ cells is limited.

We recently reported that dopamine antagonists differ in terms of their ability to prevent the high-magnitude Ca²⁺ response in the antagonist-bound receptor state and to reverse the low-magnitude Ca²⁺ response in the DA-bound state (Pauwels et al., 2001a,c). Among the investigated dopamine antagonists, tropapride, nemonapride, and haloperidol were most efficacious to reverse the DA-mediated low-magnitude Ca²⁺ response at the wt D_2short receptor in the DA-bound state. The antagonist properties of haloperidol for the agonist-mediated low-magnitude Ca²⁺ response by the Thr³⁷²Arg D_2long receptor were attenuated, whereas those of (+)-butaclamol were almost absent (Pauwels et al., 2001c). Hence, putative dopamine antagonists, which behaved as efficacious inverse agonists in our study, can nevertheless be differentiated on the basis of their antagonist properties. The partial inverse agonists clozapine and olanzapine were only capable to reverse the DA-mediated low-magnitude Ca²⁺ response at the wt D_2short receptor under submaximal DA (0.1 µM) stimulation and were virtually inactive upon maximal DA (10 µM) stimulation (Pauwels et al., 2001a), suggesting a weaker antagonist efficacy compared with tropapride, and that is in line with their partial inverse agonist responses.

The inverse agonist activity by putative dopamine antagonists seems to be a common feature to most of the neuroleptic drugs investigated in this study. These compounds differ in terms of chemical structure and therapeutic class. Nevertheless, there is apparently no link between the intrinsic activity of the neuroleptic drugs compared with their clinical efficacy and full and partial inverse agonists, as well as the silent antagonist ziprasidone being effective antischizophrenic agents (Strange, 2001). Interestingly, the partial agonist (+)-UH 232 identified here did not exhibit antipsychotic properties (Lahti et al., 1998). Atypical neuroleptic drugs (i.e., clozapine, ziprasidone, and olanzapine) have weak potency to elicit catalepsy in a rat model for extrapyramidal side effects in schizophrenic patients (Hoffman and Donovan, 1995). Raclopride produced catalepsy, but with a wide separation between the effective doses inducing catalepsy or activity (Hoffman and Donovan, 1995); this feature is also found back for clozapine, olanzapine, and ziprasidone but not for the typical neuroleptic drugs such as haloperidol and chlorpromazine (Hoffman and Donovan, 1995). Thus, a link between the magnitude of inverse agonism and induction of catalepsy in rats may exist.

In conclusion, increased isomerization in an active receptor state was achieved by the coupling of a dopamine D₂ receptor to the phospholipase C pathway via the 3ICL of an _1B-AR, the incorporation of an activating mutation (Ala²⁷⁹Glu) in the distal BBXXB motif of its 3ICL and the coexpression with a G₁₁ protein. Under these experimental conditions, the chimeric D₂/_1B receptor construct displayed strongly enhanced basal IP formation. This could be reversed by a large series of dopamine antagonists acting as inverse agonists with the exception of ziprasidone, which behaved as a silent antagonist. The model described here makes it possible to differentiate between efficacious to partial dopamine inverse agonists as well as silent dopamine antagonists. The precise determination of the intrinsic activity of putative dopamine antagonists may help to explore either therapeutic efficacy and/or adverse effects of neuroleptic drugs.