Abstract
Human dopamine D2 (hD2) and D3 (hD3) receptors were expressed at similar, high expression levels in Chinese hamster ovary (CHO) cells, and their coupling to G proteins and further signal transduction pathways were compared. In competition radioligand-binding experiments, guanosine-5′-O-(3-thio)triphosphate (GTPγS) treatment of hD2S- or hD3-CHO cell membranes induced a rightward shift and steeping of the dopamine inhibition curve. This effect was pronounced for hD2 receptors and small for hD3 receptors. Activation of G proteins was investigated in [35S]GTPγS-binding assays. Dopamine stimulated [35S]GTPγS binding 330 and 70% over basal levels on hD2-CHO and hD3-CHO cell membranes, respectively. (+)-7-(Dipropylamino)-5,6,7,8-tetrahydro-2-naphthalenol and PD128907 were partial agonists for both receptors. Haloperidol, risperidone, raclopride, and nemonapride inhibited dopamine-stimulated [35S]GTPγS binding with potencies comparable to their binding affinities for hD2 and hD3 receptors in CHO cell membranes; inverse agonism could not be detected with this assay. Receptor stimulation by dopamine inhibited forskolin-induced cyclic AMP formation in hD2-CHO and hD3-CHO cells by 70%. Furthermore, the extracellular acidification rate increased when hD2-CHO and hD3-CHO cells were stimulated by dopamine; this effect was abolished by pertussis toxin pretreatment. In this study, we could demonstrate clear functional effects at different levels of the signaling cascade of hD2and hD3 receptors in CHO cells when expressed at high levels. High-affinity agonist binding to hD2 and hD3 receptors was still present, but effects of receptor-G protein uncoupling at hD3 receptors were small, indicating that hD3 receptors maintain relatively high-affinity agonist binding in the absence of G proteins.
Dopaminergic transmission is involved in the control of locomotor activity and cognitive and neuroendocrine functions (Dolan et al., 1995; Williams and Goldman-Rakic, 1995). Dysfunction of the dopaminergic system, due to degeneration of the nigrostriatal dopaminergic pathway, leads to Parkinson’s disease, whereas increased mesolimbic dopaminergic activity is believed to be involved in schizophrenia (Heimer, 1994).
The action of dopamine is mediated by receptors that belong to the superfamily of G protein-coupled receptors. Originally, dopamine receptors were classified into two families (D1and D2) based on pharmacology and signal transduction (Kebabian and Calne, 1979). The cloning of various dopamine receptor genes led to the description of the D1-like family (D1 and D5 receptors) and the D2-like family (D2, D3, and D4 receptors) (Missale et al., 1998).
Human dopamine D2 (hD2) and D3 (hD3) receptors display considerable amino acid sequence similarity. All known antipsychotics bind to D2 receptors, and most also bind to D3 receptors (Leysen et al., 1998). In general, dopamine and several dopaminergic agonists have a higher affinity for hD3 than for hD2 receptors, whereas the affinity of antagonists is usually slightly higher for hD2 receptors (Sokoloff et al., 1992). The distribution of hD3 receptors in the brain seems to be confined to mesolimbic areas, whereas hD2receptors are found in all dopaminergic brain areas (De Keyser, 1993;Landwehrmeyer et al., 1993). Both hD2 and hD3 receptors possess a large third intracytoplasmic loop and a short carboxyl-terminal tail, a characteristic of receptors that couple to the Gαi/o subfamily of G proteins (Dohlman et al., 1991). Modulation of agonist binding at hD2 receptors by guanine nucleotides and of hD2 receptor-activated signaling pathways has been clearly demonstrated. In contrast, conflicting results have been reported regarding the G protein-coupling and -signaling properties of hD3 receptors. First, several groups failed to find modulation of agonist binding at hD3receptors by guanine nucleotides (Freedman et al., 1994; Tang et al., 1994a; Akunne et al., 1995; McAllister et al., 1995). However, other groups have reported a small rightward shift of the dopamine inhibition binding curve by guanine nucleotides (Sokoloff et al., 1992; MacKenzie et al., 1994). Second, there are several reports of negative findings on hD3-mediated inhibition of adenylyl cyclase activity, stimulation of arachidonic acid release, or phospholipase C activity (Freedman et al., 1994; MacKenzie et al., 1994; Tang et al., 1994a). Also, stimulation of adenylyl cyclase subtype II by hD3 receptors could not be demonstrated (Watts and Neve, 1997). In contrast, a slight inhibition of adenylyl cyclase activity was shown by McAllister et al. (1995) and Griffon et al. (1996). Third, a slight activation of G proteins measured as [35S]guanosine-5′-O-(3-thio)triphosphate (GTPγS) binding by hD3 receptors in transfected mammalian cell lines has been reported (Gardner et al., 1996; Malmberg et al., 1998). However, further downstream effects of hD3 receptor signaling, such as inhibition of dopamine synthesis and release or stimulation of neurite outgrowth, have been clearly demonstrated in neuronal cells, indicating a functional role for hD3 receptors (Tang et al., 1994b; O’Hara et al., 1996; Swarzenski et al., 1996). Most surprisingly, some reports have described coupling to signal transduction pathways but no effect of guanine nucleotides on agonist binding, and vice versa (MacKenzie et al., 1994; Tang et al., 1994a,b;McAllister et al., 1995; O’Hara et al., 1996).
From these conflicting findings, it is clear that the hD3 receptor-signaling mechanism has not yet been fully established. We report an extensive study of the signaling of hD3 receptors expressed in Chinese hamster ovary (CHO) cells, investigated in parallel with hD2Lreceptors expressed in CHO cells. We studied the ligand-binding and -signaling properties of the same type of parent CHO cells with the same high expression levels of hD2L and hD3 receptors. In addition, we investigated a hD3-CHO cell clone with an 8-fold lower expression level. The effect of receptor-G protein uncoupling was investigated by examination of the effect of GTPγS on the dopamine-binding curve. Using the [35S]GTPγS binding assay, we compared the activation of G proteins by hD2L and hD3 receptors on binding of dopamine, (+)-7-(dipropylamino)-5,6,7,8-tetrahydro-2-naphthalenol (7-OH-DPAT), and PD128907. Haloperidol, risperidone, raclopride, and nemonapride were tested for their ability to inhibit dopamine-stimulated [35S]GTPγS binding. We investigated the inhibition by dopamine of forskolin-stimulated cyclic AMP (cAMP) formation in CHO cells stably expressing cloned human D2L or D3 receptors (hD2L-CHO; D3-CHO). In addition, we compared extracellular acidification rates on dopamine stimulation of hD2L-CHO and hD3-CHO cells and the effect of pertussis toxin pretreatment thereupon. The various investigations revealed clear signal transduction phenomena for both dopamine receptor subtypes, indicating that both receptors couple to Gαi/o proteins.
Experimental Procedures
Cell Culture and Membrane Preparation.
The cDNA clones of hD2L and hD3 receptors were purchased, corrected by polymerase chain reaction techniques, and inserted into the pKCRE expression vector (Stam et al., 1992). The sequence was verified by DNA sequencing. The expression constructs were stably transfected into CHO cells by the calcium phosphate method, and clonal cell colonies were isolated in medium containing 800 μg/ml G418. Clones expressing high levels of receptor were selected (hD2L-CHO and hD3-CHO-high;Schotte et al., 1996); for hD3-CHO, we also selected a clone with a lower expression level. CHO cells expressing hD2L or hD3 receptors were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 2 mM l-glutamine, 1 mM sodium pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% fetal calf serum (FCS), in a humidified atmosphere of 5% CO2 at 37°C. Cells were subcultured at 80 to 90% confluence.
For the preparation of membranes, cells were subcultured from 175-cm2 tissue culture flasks to 145-cm2 Petri dishes. At 90% confluence, 5 mM sodium butyrate was added to increase the receptor expression level (Palermo et al., 1991), and the cells were incubated for an additional 24 h. The medium was removed, and the Petri dishes were washed once with 5 ml of ice-cold PBS and stored at −70°C. Petri dishes were thawed, and 5 ml of 10 mM Tris-HCl, pH 7.4, containing 1 mM EDTA and 1 mM 4-(2-aminoethyl)-benzenesulfonylfluoride hydrochloride (buffer A) was added to each dish. The cells were harvested and homogenized by 10 strokes with a Dual homogenizer (motor-driven Teflon pestle and conical glass tube). The homogenate was centrifuged (10 min at 1000g at 4°C), and the resulting pellet was resuspended in buffer A and centrifuged again (10 min at 1300g at 4°C). The two supernatants were pooled and centrifuged at 50,000gfor 1 h at 4°C. The resulting pellet was resuspended in 50 mM Tris-HCl, pH 7.4, containing 10% glycerol and stored in aliquots at −70°C. The protein concentration in membrane preparations was measured with the Bradford protein assay, with BSA as a calibration standard.
Radioligand-Binding Assays.
hD2L-CHO and hD3-CHO cell membranes were thawed on ice and suspended in 50 mM Tris-HCl buffer, pH 7.4, with 120 mM NaCl. For the radioligand binding assay, 5 to 10 μg membrane protein/assay was used. For [3H]spiperone binding, the incubation volume was 0.5 ml and incubation was performed for 30 min at 37°C. For [125I]iodosulpride binding, the incubation mixture contained 0.1% BSA, the assay volume was 0.25 ml, and incubation was performed for 30 min at 25°C. Nonspecific binding was estimated in the presence of 10 μM haloperidol for both hD2L and hD3 receptors. The reaction was terminated by filtration through Whatman GF/B filters presoaked in 0.1% polyethyleneimine. Filters were rinsed twice with 5 ml of ice-cold incubation buffer. The filter-bound radioactivity was measured in a liquid scintillation spectrometer (Tricarb; Packard, Meriden, CT) with 3 ml of scintillation fluid. Specific binding was calculated as the difference between total binding and nonspecific binding. For ligand concentration binding isotherms, [3H]spiperone was used at 10 to 12 concentrations in the range 0.01 to 1 nM, and [125I]iodosulpride was used at 10 concentrations in the range 0.1 to 3 nM. Ligand concentration binding isotherms were fitted to a rectangular hyperbola by nonlinear regression analysis in which the Kdand Bmax values were free parameters.
In competition binding experiments, serial dilutions of unlabeled compounds were incubated with [3H]spiperone (0.5 nM) or [125I]iodosulpride (0.4 nM) for the hD2 and hD3 receptor, respectively. For inhibition of [3H]spiperone binding by dopamine, the incubation buffer was 50 mM Tris-HCl, pH 7.4, containing 10 mM MgCl2 and 1 mM EGTA. Competition curves were fitted to a sigmoid by nonlinear regression analysis, in which the pIC50 value (pIC50 = −log IC50, concentration of the compound producing 50% inhibition of the specific binding of the radioactive ligand) and the Hill coefficient were free parameters.Ki values were calculated according toCheng and Prusoff (1973). Nonlinear regression analysis was performed according to algorithms described by Oestreicher and Pinto (1987). The Prism program (GraphPad Software, San Diego, CA) was used to fit curves to one- and two-site models, and the F test was used to evaluate the statistical significance of the difference in goodness-of-fit. Assays were run in duplicate in ligand concentration binding isotherms and in singlets in competition binding experiments and repeated in independent experiments (n). Curves were calculated for individual experiments, and the mean of the derived parameters was calculated.
[35S]GTPγS-Binding Assays.
The [35S]GTPγS-binding assay was performed essentially as described by Gardner et al. (1996). Briefly, the [35S]GTPγS-binding assay was carried out in a final volume of 0.5 ml containing 50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 5 mM MgCl2, 1 mM EGTA, 0.1 mM dithiothreitol, 1 μM guanosine diphosphate, and 0.2 nM [35S]GTPγS. Membranes (5–10 μg of membrane protein) and ligands were preincubated without [35S]GTPγS for 30 min at 30°C to obtain steady-state receptor occupation. After the addition of [35S]GTPγS, membranes were further incubated for 30 min. Basal [35S]GTPγS binding was measured in the absence of ligands. Reactions were terminated by rapid filtration (see above) through Whatman GF/B filters soaked in incubation buffer. The amount of radioactivity collected on the filter was determined by liquid scintillation counting. The maximum amount of [35S]GTPγS binding was always less than 10% of [35S]GTPγS added. In preliminary experiments, nonspecific binding was measured in the presence of 100 μM GTPγS; this never exceeded 10% of basal binding. Nonspecific binding was not subtracted; all values represent the total levels of [35S]GTPγS binding.
Percent stimulation was calculated as 100 times the difference between the amount of agonist-stimulated and basal binding divided by the amount of basal binding. GraphPad Prism was used to fit dose-response curves to a sigmoid in which the pEC50value (pEC50 = −log EC50, EC50, concentration of the compound producing 50% effect) and the Hill coefficient were free parameters. Antagonists were tested for inhibition of dopamine-stimulated (10 μM) [35S]GTPγS binding. The IC50 values obtained from the inhibition curves were corrected as follows:
Measurement of cAMP Content.
Cells were grown overnight in a Falcon multiwell 96-plate (Becton Dickinson Labware Ekembodegem, Belgium) at 10,000 cells/well. Sodium butyrate (5 mM) was added 24 h before the experiment. On the day of the experiment, cells were washed with controlled salt solution (CSS; 120 mM NaCl, 5 mM KCl, 0.8 mM MgCl2, 1.8 mM CaCl2, 15 mM glucose, 0.04 mM phenol red in 25 mM Tris-HCl, pH 7.4) at 37°C. After removal of CSS, the cells were preincubated for 20 min at 37°C in 60 μl of CSS containing 0.1% BSA. Dopamine was not included during the preincubation. Then, the cells were further incubated for 20 min at 37°C by the addition of 60 μl of CSS supplemented with 0.1% BSA, 2 mM 3-isobutyl-1-methylxanthine, 2 μM pargyline (a monoamine oxidase inhibitor), 2 μM cocaine (a dopamine uptake inhibitor), 100 μM forskolin, and the appropriate concentration of dopamine. After the addition of 20 μl of 1 M HClO4 (ice-cold) to terminate the reaction, the plates were frozen and thawed, and 20 μl of ice-cold KOH/K3PO4(0.5 M, pH 13.5) was added to neutralize the samples (final pH 7.4). After formation of the KClO4 precipitate (30 min at 4°C), the plates were centrifuged (10 min at 650g, 4°C). The amount of cAMP in each well was determined with a commercial 125I-labeled cAMP radioimmunoassay kit according to the procedure recommended by the manufacturer. Results are calculated as percentages of forskolin levels. GraphPad Prism was used to fit dose-response curves to a sigmoid in which the pEC50 value and the Hill coefficient were free parameters.
Measurement of Extracellular Acidification Rate.
Microphysiometry was performed as described elsewhere (Owicki et al., 1990). Briefly, cells in DMEM containing 10% FCS were seeded onto Transwell capsules at 24 h before the experiment. The capsules were briefly rinsed in running medium (DMEM without serum and without NaHCO3) and set up in the Cytosensor (Molecular Devices, Munchen, Germany). The superfusion speed was 100 μl/min, and a cycle was 120 s, composed of 90 s of pumping, 3 s of rest, 25 s of measurement, and 2 s of rest. Results are expressed as percentages of basal value; basal acidification rates were measured between 50 and 70 μV/s. Cells were stimulated each half hour for 4 min with increasing concentrations of dopamine. In one series of experiments, hD2L-CHO and hD3-CHO cells were pretreated in the Transwell capsules with pertussis toxin for 6 h at 100 ng/ml. Subsequently, the cells were washed in toxin-free medium and used for microphysiometry.
Preliminary experiments comparing serum-free and serum-containing medium, the presence and absence of butyric acid, and the different cell densities (100,000–250,000 cells/capsule) demonstrated that the results were independent of these variables (data not shown).
Materials.
[3H]Spiperone (3.5 TBq/mmol), [125I]iodosulpride (74.1 TBq/mmol), and [35S]GTPγS (±40.7 TBq/mmol) were purchased from Amersham Pharmacia Biotech (Little Chalfont, UK). Pargyline and forskolin were obtained from Sigma-Aldrich (St. Louis, MO). 3-Isobutyl-1-methylxanthine was obtained from Fluka (Buchs, Switzerland). DMEM and FCS were purchased from Life Technologies (Gaithersburg, MD). The protein assay kit was obtained from Bio-Rad (Hercules, CA). The 125I-labeled cAMP radioimmunoassay kit (SMP001J) was obtained from Dupont-NEN (Boston, MA). PD128907, lisuride, and dopamine were purchased from Research Biochemicals, Inc. (Natick, MA). Raclopride and nemonapride were purchased from Astra Arcus (Stockholm, Sweden) and Yamanouchi (Tokyo, Japan), respectively. TL99 was obtained from ICN Pharmaceuticals (Costa Mesa, CA). Haloperidol, domperidone, risperidone, and spiperone are original products of Janssen Pharmaceutica (Beerse, Belgium). 7-OH-DPAT was synthesized in-house. Guanosine diphosphate, GTPγS, and 4-(2-aminoethyl)-benzenesulfonylfluoride hydrochloride were purchased from Boehringer Mannheim (Mannheim, Germany). All other reagents were of analytical grade and obtained from Merck (Haasrode, Belgium) or Sigma Chemical Co. (St. Louis, MO). GF/B glass-fiber filters were purchased from Whatman (Kent, UK). The scintillation fluid (Ultima Gold MV) was purchased from Packard (Meriden, CT). Transwell capsules (Costar) were obtained from Elscolab (Kruibeke, Belgium). Dopamine, 7-OH-DPAT, and PD128907 were dissolved and diluted in assay buffer. Lisuride and TL99 were dissolved and diluted in ethanol. Haloperidol, spiperone, domperidone, risperidone, raclopride, and nemonapride were dissolved and diluted in dimethyl sulfoxide (DMSO). For compounds that were dissolved and diluted in DMSO or ethanol, the final, 20-fold dilution step in assay buffer was performed just before addition to the assay mixture, in which the dilution was 10-fold. In control assays, ethanol or DMSO was added to a final concentration of 0.5%.
Results
Characterization of hD2L and hD3 Receptor Binding in Transfected CHO Cell Membranes.
The hD2L and hD3 receptor binding was investigated by use of [3H]spiperone and [125I]iodosulpride;Kd andBmax values are listed in Table1. No binding was detectable in wild-type CHO cell membranes (results not shown).
Several compounds were tested for inhibition of [3H]spiperone or [125I]iodosulpride binding to hD2L-CHO or hD3-CHO-high cell membranes, respectively. Inhibition curves are shown in Fig.1; pIC50 andKi values are listed in Table2. All agonists displayed higher affinities for hD3 than for hD2L receptors; PD128907 was the most selective compound for hD3 receptors. The pKi values for antagonists were in general slightly higher for hD2L than for hD3. In this series, haloperidol and domperidone showed the biggest differences (20–30-fold) in affinities for hD2L versus hD3 receptors.
Effect of GTPγS on Dopamine Binding.
The effect of G protein uncoupling on dopamine binding to hD2L and hD3 receptors was investigated by measurement of the inhibition of [3H]spiperone binding by dopamine in the absence and presence of GTPγS (100 μM; Table3; curves shown in Fig.2). For both hD2Land hD3 receptors, GTPγS induced a steepening (hD2L, p < .01; hD3, p < .05) and a rightward shift (hD2L, 0.73 log unit, p < .005; hD3, 0.33 log unit, p < .005) of the dopamine inhibition curve when fitted to a single binding site (paired two-tailed Student’s t test).
However, in the absence of GTPγS, the data for hD2L-CHO were significantly better fitted to a two-site competition curve with a high- and a low-affinity site (F test, p < .05). In contrast, no significant improvement was observed if the dopamine inhibition curve for hD3-CHO-high (see Table 1) was fitted to a two-site model (F test, p > .05). pIC50 values and Hill coefficients are summarized in Table 3. In the presence of GTPγS, a one-site binding better described the data for both hD2L-CHO and hD3-CHO-high.
The effect of GTPγS on the dopamine inhibition curve for hD3-CHO-low (see Table 1) was similar; we found the same slight rightward shift and steepening of the inhibition curve as for hD3-CHO-high (results not shown).
hD2L and hD3 Receptor-Mediated Modulation of [35S]GTPγS Binding.
The effects of dopamine, 7-OH-DPAT, and PD128907 on [35S]GTPγS binding to hD2L-CHO, hD3-CHO-high, and hD3-CHO-low cell membranes are shown in Fig.3; pEC50 values and the levels of stimulation are listed in Table4. We used dopamine as a reference agonist. 7-OH-DPAT and PD128907 did not produce the maximum stimulation compared with dopamine and hence seem to be partial agonists (see Table4). 7-OH-DPAT inhibited dopamine-stimulated [35S]GTPγS binding to the level of its partial agonistic effect in hD2L-CHO and in both hD3-CHO clones, but PD128907 did not show any antagonistic effect at either receptor (results not shown).
In addition, dopamine antagonists were used to inhibit dopamine-stimulated (10 μM) [35S]GTPγS binding in hD2L-CHO and hD3-CHO-high cell membranes. All compounds inhibited dopamine-stimulated [35S]GTPγS binding in a dose-dependent manner (Fig.4). Corrected IC50(cIC50) values were calculated as described in Experimental Procedures (Table5). The potencies, as indicated by the cIC50 values, of the compounds to antagonize dopamine-stimulated [35S]GTPγS binding corresponded well with the affinities (Ki values) measured by inhibition of [3H]spiperone binding to the hD2L receptor or [125I]iodosulpride binding to the hD3 receptor. No inverse agonist activity (i.e., no leveling off of curves below basal [35S]GTPγS binding) was detected for any of these antagonists under the experimental conditions applied.
hD2L and hD3 Receptor-Mediated Inhibition of cAMP.
Forskolin-stimulated (50 μM) cAMP levels were 4.5 ± 0.7 pmol/well for hD2L-CHO (n= 5), 14.0 ± 3.4 pmol/well for hD3-CHO-high (n = 7), and 10.0 ± 4.2 pmol/well for untransfected CHO cells (n = 2). Basal cAMP levels never exceeded 5% of the forskolin-stimulated level. Wild-type CHO cells did not show any response on the addition of dopamine.
The dopamine inhibition of forskolin-induced cAMP formation in hD2L-CHO and hD3-CHO-high cells is shown in Fig. 5. The pEC50 values were 7.28 ± 0.08 for hD2L-CHO (n = 4) and 7.38 ± 0.35 for hD3-CHO (n = 7). Dopamine inhibited forskolin-stimulated cAMP formation by 73 ± 2% in hD2L-CHO cells and by 69 ± 8% in hD3-CHO cells.
hD2L and hD3 Receptor-Mediated Acidification of Cell Culture Medium.
The dose-response curves for the acidification rate of the cell culture medium after the application of dopamine to hD2L-CHO cells and hD3-CHO-high cells, and the effect of pertussis toxin pretreatment thereupon, are shown in Fig.6. Dopamine stimulated the extracellular acidification rates up to 15 and 45% over basal levels in hD3-CHO and hD2L-CHO cells, respectively. In the absence of pertussis toxin, derived pEC50 values were 7.87 ± 0.16 (n = 4) and 8.40 ± 0.22 (n = 6) for hD2L-CHO and hD3-CHO cells, respectively. Pertussis toxin treatment almost completely abolished hD2L and hD3receptor-stimulated extracellular acidification.
Discussion
In this study, we compared the binding and signaling properties of hD2L and hD3 receptors in the same background of CHO cells. Discrepancies between previous reports regarding hD3 receptor coupling to G proteins and/or the generation of second messengers necessitated clear demonstration of these mechanisms in parallel with hD2L receptors.
Therefore, we performed experiments with CHO cells expressing approximately equal, high levels of hD2L and hD3 receptors. Both receptors bound [3H]spiperone and [125I]iodosulpride with high affinity. The pharmacological properties of the receptors corresponded well with previously reported data (Sokoloff et al., 1990). Like dopamine, the agonists tested had an apparent preference for hD3 over hD2L receptors in ligand-binding assays with CHO cell membranes. PD128907 was the most selective ligand for hD3 receptors, and lisuride was the least selective. The antagonists investigated showed nanomolar binding affinities for both hD2L and hD3 receptors. We observed a 20- to 30-fold higher binding affinity of haloperidol and domperidone for hD2L over hD3 receptors, which is in agreement with data reported by Sokoloff et al. (1990).
GTPγS was used to modulate the dopamine inhibition of [3H]spiperone binding. In the absence of GTPγS, the data for hD2L-CHO were best explained by a two-site competition curve with a high- and a low-affinity site. In contrast, the hD3-CHO-high dopamine inhibition curve was best fitted to a one-site model. GTPγS induced a significant rightward shift and steepening of the dopamine inhibition curve for both hD2L-CHO and hD3-CHO-high, although the effect was less pronounced for hD3-CHO-high. However, the latter was observed consistently and was statistically significant. Similar experiments with another hD3 clone (hD3-CHO-low), expressing about 8-fold fewer receptors, also yielded a slight rightward shift and steepening of the dopamine inhibition curve. Our observations suggest that the hD3 receptor exists in different conformational states, but dopamine does not seem to distinguish much between the G protein-coupled conformation and the uncoupled receptor conformation, as suggested by the small shift in the dopamine inhibition curve in the presence of GTPγS. Observations on dopamine binding to hD3 receptors, expressed in Escherichia coli providing the receptor in the complete absence of G proteins, support this hypothesis (unpublished observation). In contrast, the hD2L receptor displays a low- and high-affinity conformation for agonist binding, presumably representing the uncoupled and G protein-coupled receptor conformation, respectively.
In the present study, the [35S]GTPγS binding assay was used to investigate G protein activation by agonist stimulation of hD2L and hD3receptors in CHO cell membranes. The effects of GDP concentration and incubation time on agonist-stimulated [35S]GTPγS binding at hD2L-CHO have already been described (Gardner et al., 1996). In preliminary experiments, these results were confirmed (results not shown). We have demonstrated that dopamine, PD128907, and 7-OH-DPAT concentration-dependently stimulated [35S]GTPγS binding in hD2L-CHO and hD3-CHO membranes. Dopamine stimulated [35S]GTPγS binding 5-fold more in hD2L-CHO cell membranes than in hD3-CHO-high cell membranes, but with a 40-fold lower potency. Maximum stimulation of [35S]GTPγS binding was 2-fold higher in hD3-CHO-high than in hD3-CHO-low, suggesting an increase in G protein coupling due to the higher number of receptors in the membranes (i.e., higher [receptor]/[G protein] ratio). This could mean that the receptor density drives this reaction. The amount of stimulation at hD3-CHO-high in this study was more than 2-fold higher than that recently reported (Malmberg et al., 1998). Based on the partial agonism of 7-OH-DPAT and PD128907, we tried to inhibit dopamine-stimulated [35S]GTPγS binding by using these compounds. 7-OH-DPAT inhibited dopamine stimulation whereas PD128907 did not, indicating that 7-OH-DPAT is a partial agonist with antagonist properties and that PD128907 is a partial agonist that lacks antagonist properties. In addition, we used a number of structurally different antagonists to inhibit dopamine-stimulated [35S]GTPγS binding. Haloperidol is a butyrophenone, risperidone is a benzisoxazole, and raclopride and nemonapride are benzamides. All compounds inhibited dopamine-stimulated [35S]GTPγS binding with potencies similar to those found in radioligand-binding experiments (see Tables 2 and 5). We could not observe inverse agonism under the conditions used (i.e., curves did not level out below basal [35S]GTPγS binding), although haloperidol and raclopride have been reported to be inverse agonists (Hall and Strange, 1997; Malmberg et al., 1998). The concentration of GDP that we used (1 μM) probably decreased basal levels of [35S]GTPγS binding to such an extent that inverse agonism became hard to detect. So far, inverse agonism at hD2 receptors has been clearly shown only on second messenger formation in whole cells, which represents a more physiological condition than [35S]GTPγS binding in membrane preparations.
In this study, we were able to demonstrate that dopamine inhibited forskolin-stimulated cAMP formation in a dose-dependent way in both hD2L-CHO and hD3-CHO-high cells. Surprisingly, a similar, strong dopamine inhibition of cAMP formation (up to 70%) was apparent in both types of cell, whereas earlier reports on hD3 receptors have shown only 30 to 40% inhibition of forskolin levels by dopamine (McAllister et al., 1995; Griffon et al., 1996) or even no inhibition at all (Freedman et al., 1994; MacKenzie et al., 1994; Tang et al., 1994). However, it should be noted that forskolin-stimulated cAMP levels were 2- to 3-fold lower in hD2L-CHO cells than in hD3-CHO-high or in wild-type cells. This suggests a tonic inhibition of cAMP formation by hD2Lreceptors. Interestingly, we found nearly the same potency of dopamine for stimulating hD2L-CHO and hD3-CHO-high cells in cAMP assays, which contrasts with our results from [35S]GTPγS-binding experiments involving cell membrane preparations, where dopamine was at least 10-fold less potent at hD2L receptors. Because the slopes of the dopamine curves are near unity in the [35S]GTPγS binding and cAMP experiments, the higher shift in potency at hD2L-CHO compared with hD3-CHO cannot be attributed to promiscuous coupling to a wide variety of G proteins with different affinities. The shift in potency between these assays could be attributed to the different assay conditions (e.g., temperature, buffer composition). The most striking difference in assay conditions is the use of membranes instead of intact cells. Indeed, comparison of the potencies of dopamine in membranes (radioligand- and [35S]GTPγS-binding experiments) and in intact CHO cells (cAMP content measurement and microphysiometry experiments) indicates that dopamine is more potent in intact CHO cells. This suggests that during the membrane preparation, receptor-G protein interaction may be disturbed, leading to lower potencies of dopamine; the hD2L receptor would be more susceptible to this perturbation than the hD3 receptor (see Fig.2).
In microphysiometry experiments, dopamine stimulated extracellular acidification rates for hD2L-CHO and hD3-CHO-high cells. As in [35S]GTPγS-binding experiments, we found a higher efficacy of dopamine stimulation of hD2Lreceptors than of hD3 receptors but with a higher potency at the latter. Also, dopamine was more potent at hD2L and hD3 receptors in microphysiometry experiments than in cAMP assays, which might indicate a different level of signal amplification. The effect of dopamine could be blocked almost completely by pertussis toxin pretreatment of the cells, indicating involvement of G proteins from the Gαi/o family in the signaling of hD2L and hD3receptors.
We demonstrated clear activation of hD2L and hD3 receptors in CHO cells by using various signal transduction assays. The high levels of stimulation, compared with previous reports (which have 2- to 40-fold lower expression levels), could be ascribed to the high expression level. Hence, it can be supposed that for G protein coupling of the D3receptor in vivo, high expression levels in specific cells may be required.
In conclusion, we compared the signal transductions of hD2L-CHO and hD3-CHO using [35S]GTPγS-binding experiments, dopamine inhibition of forskolin-stimulated cAMP formation, and microphysiometry experiments. In all cases, we found stimulation of hD2L and hD3 receptors. In [35S]GTPγS-binding and microphysiometry experiments, stimulation of hD3-CHO was much less pronounced than stimulation of hD2L-CHO. Interestingly, we found dopamine inhibition of forskolin-stimulated adenylyl cyclase activity to be almost equally high in hD3-CHO and hD2L-CHO cells. These data suggest that functional effects of hD3receptors can be measured reliably in CHO cells when expressed at high levels.
Acknowledgments
We thank Dr. Katty Josson for scientific advice; Ilse Van den Wyngaert, Inez Van de Weyer, and Geert Nobels for preparing the hD2L and hD3 receptor cDNA clones; and Dr. Anne Lesage and Paul van Gompel for expression of the receptors in CHO cells. Martine Ercken’s practical tips were greatly appreciated. In particular, we would like to thank Walter Gommeren, who died in Nepal in the Fall of 1998. We could always rely on his invaluable experience, gained from 25 years of working in receptor binding technology. His ironic, yet enthusiastic, approach to life was stimulating to all of us. We miss him deeply.
Footnotes
-
Send reprint requests to: Dr. Josée E. Leysen, Department of Biochemical Pharmacology, Janssen Research Foundation, Turnhoutseweg 30, B-2340 Beerse, Belgium. E-mail:jleysen2{at}janbe.jnj.com
-
1 This work was supported by a grant from the IWT (Vlaams Instituut voor de Bevordering van het Wetenschappelijk-Technologisch Onderzoek in de Industrie); project IWT 940232.
- Abbreviations:
- human dopamine D2
- cAMP, cyclic AMP
- CHO
- Chinese hamster ovary
- CSS
- controlled salt solution
- DMEM
- Dulbecco’s modified Eagle’s medium
- DMSO
- dimethyl sulfoxide
- hD2L-CHO
- CHO cells stably expressing cloned human D2L receptors
- hD3-CHO
- CHO cells stably expressing cloned human D3 receptors
- cIC50
- corrected IC50
- [35S]GTPγS
- [35S]guanosine-5′-O-(3-thio)triphosphate
- 7-OH-DPAT
- (+)-7-(dipropylamino)-5,6,7,8-tetrahydro-2-naphthalenol
- hD2
- hD3, human dopamine D3
- FCS
- fetal calf serum
- pIC50 = −log IC50
- IC50, concentration of the compound producing 50% inhibition of the specific binding of the radioactive ligand)
- pEC50 = −log EC50
- EC50, concentration of the compound producing 50% effect)
- Received February 1, 1999.
- Accepted March 26, 1999.
- The American Society for Pharmacology and Experimental Therapeutics