Introduction

Summary Introduction Procedures Results Discussion References

Dopamine is an important neurotransmitter that is involved in the control of locomotor activity, emotion and affect, and neuroendocrine secretion. The actions of dopamine are mediated by a family of receptors that are part of the large superfamily of G protein-coupled receptors. On the basis of their structure, pharmacology, and signaling properties (for reviews, see Refs. 1 and 2), the dopamine receptors are divided into two subgroups: D₁-and D₂-like receptors. The dopamine D₃ receptor has been of particular interest because of its high affinity for the neuroleptic drugs used to treat schizophrenia and its localization within limbic areas of the brain, which are presumed to be involved in the control of emotion and affect.

The structure of the D₃ receptor is typical of G protein-coupled receptors that are coupled to G_i and mediate the inhibition of AC. This receptor has a large third intracellular loop and a short carboxyl-terminal tail, ending at the putatively palmitoylated cysteine residue. In addition, the D₃ receptor is highly homologous to the D₂ receptor, which inhibits AC (1, 2). Thus, after its cloning, it came as a surprise that the D₃ receptor did not inhibit AC in transfected cells (3). Since that time, the D₃ receptor has been shown to inhibit cAMP accumulation in some cases (4-7) as well as modulate other second messenger pathways that are affected by the D₂ receptor, such as ion channel activity and stimulation of Na⁺/H⁺ exchange (4, 8). However, the modulation of these signaling pathways by the D₃ receptor is quite weak, especially in comparison with the effects of the D₂ receptor on the same pathways. In addition, the modulation of agonist binding to the D₃ receptor by guanine nucleotides, an indicator of receptor/G protein coupling, often is not observed (3, 5, 7, 9, 10). Thus, to date it has seemed that the D₃ receptor does not modulate any second messenger system to a substantial extent.

The D₃ receptor is expressed only in the central nervous system, with a very restricted distribution and generally low levels of expression in the brain. This, coupled with a lack of agonists and antagonists that are highly selective between the D₂ and D₃ receptors, has made it difficult to examine the biochemical and physiological functions of the D₃ receptor. As a result, D₃ receptor signaling has been studied almost exclusively by its expression in heterologous cell lines that usually are not neuronal. This raises the possibility that the lack of signaling by the D₃ receptor could be the result of the absence of essential components of a signal transduction pathway in the cultured cells that have been used to study this receptor.

ACV was cloned independently by three groups, from canine heart (11), rat liver (12), and rat striatum (13). This AC is stimulated by G_s, inhibited by G_i and Ca²⁺, and unaffected by subunits (for reviews, see Refs. 14 and 15). In addition, ACV is stimulated by protein kinase C phosphorylation and inhibited by protein kinase A phosphorylation (16, 17). This AC isoform seems to be fairly widely expressed in the periphery (11, 12); however, Northern blotting of brain structures indicated that ACV is only expressed in the striatum. Furthermore, in situ hybridization in the rat brain demonstrated that it is expressed in the nucleus accumbens and olfactory tubercle, in addition to the striatum (13). More detailed in situ hybridization studies revealed that ACV is expressed in other brain regions at lower levels, including the islands of Calleja, and most strikingly, that its distribution is limited to dopaminergically innervated regions (18).

The correlation of ACV localization with regions of expression of dopamine receptors led us to hypothesize that this effector could be involved in the signal transduction of the dopamine receptors. In particular, its expression in brain areas in which the D₃ receptor is present, such as the nucleus accumbens, olfactory tubercle, and islands of Calleja, suggested that the D₃ receptor might regulate ACV activity. We coexpressed the D₃ receptor with various AC isoforms in human embryonic kidney 293 cells to examine the ability of this receptor to inhibit specific AC subtypes.

	Experimental Procedures

Summary Introduction Procedures Results Discussion References

Materials. ACV (canine) cDNA was the gift of Dr. Y. Ishikawa (Columbia University, New York, NY). ACVI (rat) cDNA was the gift of Dr. R. Premont (Duke University Medical Center, Durham, NC), and D₃ receptor (human) cDNA was from Dr. B. Sahagan (DuPont-Merck Pharmaceutical, Wilmington, DE). G_i subunit cDNAs were the gift of Dr. P. Casey (Duke University Medical Center). Antiserum 116 was the gift of Dr. D. Manning (University of Pennsylvania, Philadelphia, PA), and antiserum EC/2 was from DuPont-New England Nuclear (Boston, MA). Cell culture reagents were from Cellgro (Herndon, VA) except for FBS, which was from Atlanta Biologicals (Norcross, GA). Quinpirole was obtained from Research Biochemicals (Natick, MA), and other chemicals were from Sigma Chemical (St. Louis, MO). [³H]Adenine, [¹⁴C]cAMP, and [³⁵S]GTPS (1250 Ci/mmol) were from DuPont-New England Nuclear. [³H]Spiperone (95-105 Ci/mmol) was from Amersham (Arlington Heights, IL), and PTX was from Calbiochem (La Jolla, CA).

Cell culture. Human embryonic kidney 293 cells (CRL 1573; American Type Culture Collection; Rockville, MD) were grown in minimum essential medium containing 10% FBS and 50 µg/ml gentamicin in a humidified atmosphere of 5% CO₂/95% air at 37°. Cells were transfected by the calcium phosphate method (19). The day before transfection, cells were plated at a density of 2.5 × 10⁶ cells/100-mm dish. On the morning after transfection, the cells were shocked with 15% glycerol in PBS and then supplied with fresh medium. In the afternoon of the same day, cells were plated onto 12-well dishes or 150-mm dishes. Cells were used for assays ~60 hr after transfection. Receptor expression was monitored by saturation binding with [³H]spiperone as previously described (20).

Whole-cell cAMP assay. Transfected cells plated onto 12-well dishes were labeled with media containing 5% FBS, gentamicin, and 1 µCi/ml [³H]adenine overnight before the assay. On the day of the assay, the labeling media was aspirated, and the cells were gently washed with PBS. The assay was carried out as previously described (20). Conversion of [³H]ATP to [³H]cAMP was assessed by sequential chromatography over Dowex and alumina columns as described by Salomon (21). For experiments including treatment with PTX, 100 ng/ml PTX was added with the labeling media.

GTPS binding assay. Transiently transfected 293 cells were plated onto 150-mm dishes 24 hr after transfection. On the day of the assay, the cells were washed twice with cold PBS and detached from the plate. Cells were collected with a brief spin and resuspended in lysis buffer (10 mM Tris·HCl, pH 7.4, 5 mM EDTA). The [³⁵S]GTPS binding assay was carried out essentially as described by Barr et al. (22). Briefly, membranes were homogenized with a Teflon pestle, centrifuged at 40,000 × g, and resuspended in Tris/MgCl₂/EDTA buffer (50 mM Tris·HCl, pH 7.4, 4.8 mM MgCl₂, 2 mM EDTA, 100 mM NaCl). Assays contained 20 µg of membranes in Tris/MgCl₂/EDTA buffer plus 1 µM GDP, 30 nM [³⁵S]GTPS, and the indicated drugs in a volume of 62 µl. The reaction was carried out at 30° for 5 min and terminated by the addition of IP buffer (50 mM Tris·HCl, pH 7.4, 20 mM MgCl₂, 150 mM NaCl, 0.5% Nonidet P-40, 1 µg/ml aprotinin, 100 µM GDP, 100 µM GTP). Membranes were solubilized for 30 min at 4°. Samples were precleared for 20 min at 4° with 20% protein A-Sepharose (Pharmacia, Piscataway, NJ) in IP buffer containing 2% bovine serum albumin. Supernatants were then transferred to tubes containing antibody, and 100 µl of protein A-Sepharose was added. G_i1 and G_i2 were immunoprecipitated with antiserum 116, and G_i3 was immunoprecipitated with antiserum EC/2. After 1 hr of incubation at 4°, the beads were washed three times with IP buffer and once with wash buffer (50 mM Tris·HCl, pH 7.4, 20 mM MgCl₂, 150 mM NaCl, 100 µM GDP, 100 µM GTP). The immune complexes were eluted from the beads with 0.5% sodium dodecyl sulfate at 95° for 5 min, and the eluate, containing immunoprecipitated [³⁵S]GTPS, was counted in a liquid scintillation counter.

Data analysis. Data are presented as mean ± standard error unless indicated otherwise. For cAMP assays, cAMP accumulation in the presence of 1 µM forskolin was normalized to 100%, and cAMP levels in the presence of forskolin plus quinpirole were expressed as a percentage of the cAMP accumulation stimulated by forskolin. For [³⁵S]GTPS binding experiments, binding of [³⁵S]GTPS in the absence of agonist was normalized to 1, and binding in the presence of agonist was expressed as fold-increase of this value. Dose-response curves were analyzed with the nonlinear curve-fitting program Prism (GraphPAD Software, San Diego, CA). Statistical analyses were performed using Student's t test; p < 0.05 was considered statistically significant.

	Results

Summary Introduction Procedures Results Discussion References

We tested the ability of the D₃ and D₂ receptors to inhibit cAMP accumulation in 293 cells. In cells transiently transfected with the human D₃ receptor, the agonist quinpirole caused no inhibition of forskolin-stimulated cAMP levels. In contrast, when the D₃ receptor was cotransfected with canine ACV, quinpirole caused a 71 ± 7% inhibition of forskolin-stimulated cAMP (Fig. 1). In comparison, when the human D₂ receptor was transfected alone, agonist evoked a 44 ± 9% inhibition of forskolin-stimulated cAMP accumulation, whereas when the D₂ receptor and ACV were cotransfected, this inhibition increased to 87 ± 4% (Fig. 1). Thus, the extent of inhibition of ACV by D₂ and D₃ receptors was not significantly different (p > 0.05), and inhibition of ACV by the D₃ receptor was more effective than inhibition of endogenous 293 cell ACs by the D₂ receptor. Inhibition of ACV by both the D₃ and D₂ receptors was also observed in COS-7 cells. The D₃ receptor caused 53 ± 9% inhibition of ACV and the D₂ receptor caused 70 ± 12% inhibition of ACV in COS-7 cells.

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Inhibition of ACV by D3 and D2 receptors. 293 cells were transfected with receptor alone (Mock) or receptor and ACV. cAMP accumulation in the presence of 1 µM forskolin plus 10 µM quinpirole was normalized as described in Experimental Procedures. Forskolin-stimulated cAMP percent conversion was as follows. D3 receptor, 0.73 ± 0.29; D3 receptor plus ACV, 5.4 ± 1.6; D2 receptor, 0.64 ± 0.17; and D2 receptor plus ACV, 4.9 ± 1.5. Data are mean ± standard error for three independent experiments. Significant decrease compared with Mock, *, p < 0.01 and **, p < 0.0005. Significant decrease compared with forskolin-stimulated cAMP, ++, p < 0.05.

In all experiments, we attempted to match the levels of receptor expression between the conditions being compared. For example, in the experiments presented in Fig. 1, receptor expression was as follows: D₂ receptor, 5.2 ± 0.9 pmol/mg of protein; D₂ receptor plus ACV, 6.1 ± 0.9 pmol/mg of protein; D₃ receptor, 6.6 ± 1.9 pmol/mg of protein; and D₃ receptor plus ACV, 6.6 ± 1.9 pmol/mg of protein. Receptor expression among the different conditions was not significantly different as determined by analysis of variance. Receptor expression levels varied between 1 and 4 pmol/mg of protein in the remaining experiments and were never significantly different between the conditions being compared.

To determine the specificity of the inhibition of ACV by the D₃ receptor, both the D₃ and D₂ receptors were tested for their ability to affect the activity of other AC isoforms. ACs are divided into three major subgroups. Members of each of the subfamilies were tested for modulation of their activity by the D₃ and D₂ receptors. ACI, which is a member of the Ca²⁺/calmodulin-stimulated subgroup, and ACII, which is the prototypical -stimulated isoform, were tested. ACVI, which belongs to the same subfamily as ACV, was also tested.

The D₃ receptor was unable to inhibit forskolin-stimulated ACI (Fig. 2A). Interestingly, the D₂ receptor caused only a 13 ± 2% inhibition of cAMP in the presence of ACI, whereas when the D₂ receptor was expressed alone, it inhibited cAMP accumulation by 37 ± 11% (Fig. 2B). This difference is significant (p < 0.05), suggesting that the D₂ receptor does not inhibit ACI activity as effectively as it inhibits other AC isoforms present in 293 cells. Similar results were obtained when PGE₁, an endogenous receptor that activates G_s, was used to stimulate cAMP formation (data not shown). In these experiments, the forskolin-stimulated cAMP accumulation was not greater in cells transfected with ACI than in 293 cells not expressing this isoform (see legend to Fig. 2). ACI has been shown to be inhibited by G protein subunits (23). In these experiments, expression of ACI was confirmed by a decrease in forskolin-stimulated cAMP when 12 subunits were cotransfected with ACI, an effect that was not observed when 293 cells were transfected with 12 alone (Fig. 2C).

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Effect of D3 and D2 receptors on cAMP accumulation in 293 cells. A, The D3 receptor was transfected in 293 cells with or without the indicated AC isoforms. B, Transfection of the D2 receptor in the presence or absence of the indicated AC isoforms. Data are the mean ± standard error of three to five experiments. cAMP accumulation in the presence of 1 µM forskolin plus 10 µM quinpirole was normalized as described in Experimental Procedures. Forskolin-stimulated cAMP (percent conversion) for each condition was as follows: D3 receptor, 0.74 ± 0.36; D3 receptor plus ACI, 0.95 ± 0.31; D3 receptor plus ACII, 0.58 ± 0.25; and D3 receptor plus ACVI, 0.89 ± 0.28 (A); and D2 receptor, 0.69 ± 0.09; D2 receptor plus ACI, 0.79 ± 0.28, D2 receptor plus ACII, 0.58 ± 0.19; and D2 receptor plus ACVI, 0.82 ± 0.22 (B). *, p < 0.05; **, p < 0.01, significantly different from cAMP accumulation in the presence of receptor alone. +, p < 0.05, significant decrease compared with forskolin-stimulated cAMP. C, Confirmation of expression of transfected ACs. Left, effect of the absence (Mock) or presence of transfected 12 subunits on 1 µM forskolin-stimulated cAMP accumulation. *, p < 0.05, significantly different from mock transfected cells. Right, stimulation of cAMP formation by the simultaneous addition of 1 µM forskolin and 1 µM PGE1 in mock transfected 293 cells or cells expressing ACV or ACVI. *, p < 0.05, significantly different from mock transfected cells.

Stimulation of the ACII isoform by subunits has been shown to depend on simultaneous receptor-mediated activation of G_s (23, 24). We observed that this stimulation by subunits is also evident when forskolin is used to activate ACII in place of activation of G_s by hormone. When the D₃ receptor and ACII were coexpressed, forskolin plus the agonist quinpirole significantly stimulated cAMP accumulation to 125 ± 15% of forskolin-stimulated cAMP levels (Fig. 2A). This suggests that the D₃ receptor is capable of providing subunits required for stimulation of ACII. In comparison, in the presence of forskolin and quinpirole, the D₂ receptor stimulated ACII to a greater extent (155 ± 11%) than the D₃ receptor (Fig. 2B). Similar results were obtained when ACII was activated with PGE₁ (data not shown). Similar to the case for ACI, we did not observe increases in forskolin-stimulated cAMP when ACII was transfected in 293 cells (see legend to Fig. 2). The expression of this isoform was confirmed by an increase in forskolin-stimulated cAMP levels in cells transfected with ACII and 12, whereas this was not observed in cells transfected with 12 only (Fig. 2C).

The D₃ receptor was unable to significantly inhibit ACVI (Fig. 2A), despite its similarity to ACV. Likewise, the D₂ receptor caused no greater inhibition of cAMP accumulation in the presence of ACVI than when the receptor was expressed alone (Fig. 2B). As was the case for ACI and ACII, ACVI does not seem to be substantially stimulated by forskolin in 293 cells (see legend to Fig. 2). However, ACV and ACVI have been observed to be synergistically activated by forskolin and G_s (25). We found that in cells transfected with ACVI, cAMP levels generated by the simultaneous addition of forskolin and PGE₁ were equivalent to those generated in the presence of ACV, whereas a significantly lower amount of cAMP is generated by forskolin and PGE₁ in untransfected 293 cells (Fig. 2C). Thus, the expression levels of ACV and ACVI are presumed to be similar.

The coexpression of the D₃ receptor with ACV was also found to affect the potency of quinpirole to inhibit forskolin-stimulated cAMP levels (Fig. 3). In many experiments, no inhibition of cAMP accumulation by the D₃ receptor was observed (e.g., Fig. 1). However, in some cases, a slight inhibition of forskolin-stimulated cAMP levels occurred. In these experiments, the IC₅₀ value of quinpirole was 0.56 ± 0.18 nM. When the D₃ receptor and ACV were coexpressed, the IC₅₀ value of quinpirole decreased significantly to 0.26 ± 0.04 nM (p < 0.05). Fig. 3 shows one experiment selected from the occasional experiments in which the D₃ receptor inhibited cAMP accumulation in the absence of ACV. The IC₅₀ value of quinpirole at the D₂ receptor was not changed by the coexpression of ACV. The increased potency for inhibition of ACV by the D₃ receptor could reflect a situation in which ACV is inhibited more strongly by G proteins activated by this receptor than are other AC subtypes.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Dose-response curves for the inhibition of cAMP accumulation by the D3 receptor in 293 cells. The D3 receptor was expressed in the absence or presence of ACV (representative of four independent experiments). The experiment was performed in triplicate. Error bars, standard deviation. Forskolin-stimulated cAMP (percent conversion): D3 receptor, 1.5 ± 0.5; D3 receptor plus ACV, 3.9 ± 2.3. Data were normalized as described in Experimental Procedures.

The involvement of G proteins in the inhibition of ACV by the D₃ receptor was examined. To test for the involvement of members of the G_i class of G proteins, cells were treated overnight before cAMP assay with 100 ng/ml PTX. Treatment with PTX prevented the inhibition of forskolin-stimulated ACV by the D₃ receptor (Table 1). Inhibition of ACV by the D₂ receptor was also sensitive to PTX (data not shown). This suggests that inhibition of ACV by the receptors is mediated through G_i or G_o. Interestingly, PTX treatment decreased the maximal forskolin-stimulated cAMP level that could be obtained by approximately half (Table 1). This effect was observed only when ACV was expressed in 293 cells, not when ACI or ACII were overexpressed or in the absence of transfected ACs (data not shown).

Possible explanations for the apparent selectivity of the inhibition of ACV by the D₃ receptor are that this receptor activates a G protein that specifically inhibits ACV or activates a G protein that inhibits ACV to a greater extent than other ACs. To test this hypothesis, we evaluated the ability of the D₃ and D₂ receptors to activate G_i subunits by using an assay consisting of [³⁵S]GTPS binding, followed by immunoprecipitation of a G_i subunit. We chose to test G_i subunits because of the PTX sensitivity of the inhibition of ACV by the D₃ receptor. To produce a detectable signal, individual G_i subunits were coexpressed with each receptor and ACV. No receptor-dependent [³⁵S]GTPS binding was observed in the absence of cotransfected G_i subunits (data not shown).

These experiments indicated that the D₃ receptor weakly activated all three G_i subunits (Fig. 4). Incubation with quinpirole caused a 1.2-1.7-fold increase in [³⁵S]GTPS binding over that observed in the absence of agonist. In contrast, the D₂ receptor stimulated the activation of G_i1 and G_i2 subunits to a greater extent than the D₃ receptor. [³⁵S]GTPS binding was increased ~3-fold by the D₂ receptor when quinpirole was included in the incubation. The activation of G_i3 by the D₂ receptor was more variable and was not significant but seemed to be greater than the activation of G_i3 by the D₃ receptor. The overall profile of activation of G_i subunits by the D₃ and D₂ receptors is quite similar and differs primarily in the extent of G protein activation, with activation by the D₃ receptor much weaker than that by the D₂ receptor.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Binding of GTPS to overexpressed Gi subunits in 293 cells. The D2 or D3 receptor and the indicated Gi subtype were transiently transfected in 293 cells. Data are expressed as fold-increase in [35S]GTPS immunoprecipitated in the presence of 50 µM quinpirole compared with binding in the absence of agonist (cpm immunoprecipitated: D2/Gi1: quin, 934 ± 68; +quin, 2881 ± 963; D3/Gi1: quin, 4561 ± 1075; +quin, 5034 ± 1158; D2/Gi2: quin, 804 ± 163; +quin, 2313 ± 656; D3/Gi2: quin, 2097 ± 913; +quin, 3301 ± 921; D2/Gi3: quin, 944 ± 283; +quin, 1441 ± 452; D3/Gi3: quin, 8603 ± 1284; +quin, 9949 ± 1837). Values are mean ± standard error for three to five experiments. *, p < 0.05, significant increase over [35S]GTPS binding in the absence of agonist.

	Discussion

Summary Introduction Procedures Results Discussion References

Despite the expression of the D₃ receptor in numerous cell culture lines, no second messenger pathways for this receptor have been clearly identified. When second messenger coupling has been demonstrated, the magnitude of the response is considerably weaker than that generated by the D₂ receptor. This raised the possibility that these cell lines lack a necessary component of the D₃ receptor signaling pathway. The identification of ACV as being preferentially expressed in dopaminergically innervated brain regions, and particularly its expression in the islands of Calleja and nucleus accumbens (18), in which the D₃ receptor is expressed, suggested that it might be a specific downstream target of the D₃ receptor. Strikingly, this AC subtype does not seem to be expressed in any cultured cell line, although the related ACVI is expressed in several lines (26). In this study, we used the human embryonic kidney 293 cell line. The expression of AC isoforms in these cells has been characterized by several groups using reverse transcription-polymerase chain reaction cloning. It seems that 293 cells express ACVI and ACVII (26-28) and probably also express ACI, ACII, and ACIII (27, 28).

Coexpression of the D₃ receptor and ACV in both 293 and COS-7 cells conferred on the receptor the ability to effectively and consistently inhibit forskolin-stimulated cAMP accumulation, which did not occur in the absence of ACV or the presence of other AC subtypes. Similar results were observed with the D₂ receptor, although a critical difference is that this receptor is able to strongly and consistently inhibit cAMP accumulation in the absence of transfected ACs in 293 cells.

Treatment of the cells with PTX prevented the inhibition of forskolin-stimulated ACV by the agonist quinpirole. This suggested that the D₃ receptor might be coupled to a member of the G_i family. Surprisingly, PTX treatment substantially reduced the extent to which forskolin could stimulate cAMP accumulation in cells expressing ACV but not other ACs. One property of ACV is that it is inhibited by micromolar concentrations of Ca²⁺ (14). A possible explanation for the reduction of forskolin-stimulated activity is that PTX might raise intracellular Ca²⁺ levels sufficiently to inhibit ACV activity but have little or no effect on other ACs. Testing of the effects of PTX on ACVI activity, which is also inhibited by Ca²⁺, may shed further light on this possibility. Despite this decrease in forskolin-stimulated cAMP in PTX-treated cells, the D₃ receptor did not cause any further inhibition of cAMP accumulation in the presence of PTX.

The coexpression of ACV with the D₃ receptor increased the potency of agonist to inhibit cAMP accumulation compared with the D₃ receptor alone. This suggested that one or more G proteins activated by the D₃ receptor might inhibit ACV more effectively than other AC isoforms. When the ability of the D₃ receptor to activate specific G_i subunits was examined more directly, we found that this receptor weakly activates all three G_i subtypes. The D₂ receptor exhibited a very similar specificity of activation of G_i subunits, although the activation was to a much greater extent (Fig. 4). The activation of G_i1 and G_i2 by the D₂ receptor was ~3-fold over the basal level, an amount that is consistent with that observed for other receptors by this method (22).

The expression levels of the G_i subunits and the efficiency of the antibodies to immunoprecipitate these subunits were not examined in this study. Therefore, the extent of activation of the G_i subunits by the D₃ receptor cannot be compared, and we cannot make a strong case that G_i2 is more activated more robustly by the D₃ receptor than G_i1 or G_i3. However, the conditions for GTPS binding for each G_i subtype were identical when either the D₂ or D₃ receptor was expressed. Thus, we can, for example, compare the extent of activation of G_i1 by the D₂ and D₃ receptors. In this case, it is clear that the D₂ receptor promotes greater activation of G_i subunits than the D₃ receptor.

These results support our hypothesis that ACV may be more efficiently inhibited by G proteins activated by the D₃ receptor than are other ACs; apparently only a small amount of activated G_i is required to robustly inhibit this AC. Studies using recombinant G_i subunits have shown that the three subtypes are equivalent in their ability to inhibit ACV (29). However, recombinant G_i subunits are clearly able to inhibit ACV to a greater extent than they inhibit forskolin- or G_s-activated ACI (29, 30). The small stimulation of forskolin-activated ACII by the D₃ receptor may also be a reflection of the weak G protein activation by this receptor, which would supply only a small amount of free subunits in the cell. Furthermore, the D₂ receptor activates all three G_i subtypes much more robustly than the D₃ receptor, and this receptor can efficiently inhibit the endogenous 293 cell ACs as well as stimulate transfected ACII to a 2-fold-greater extent than does the D₃ receptor (Fig. 2).

However, the explanation that ACV is inhibited more effectively by G_i subunits than other ACs does not fully account for the specific inhibition of ACV by the D₃ receptor. For example, ACVI is inhibited by recombinant G_i subunits to a similar extent and with a similar potency as ACV (29), but ACVI, which is endogenously expressed in 293 cells, does not seem to be inhibited by the D₃ or D₂ receptors (Fig. 2). This suggests that there could be differences in the actions of G proteins on ACV and ACVI or that other signaling molecules are involved in this process. One possibility is that the D₃ receptor, G proteins, and ACV are compartmentalized within the cell in such a way that preferential coupling between these components occurs, whereas ACVI, and perhaps other ACs, is excluded from this compartment. Another formal possibility is that a complex of the D₃ receptor and G_i subunits exists that exhibits increased potency for inhibition of ACV but not other ACs. Additional studies are necessary to elucidate the mechanism by which the D₃ receptor inhibits ACV in greater detail, by, for example, testing the ability of the D₃ receptor to activate other G protein subtypes.

The D₃ receptor seems to demonstrate remarkable specificity of signaling for ACV rather than other ACs. This selectivity at the level of an effector may be a general mechanism for tailoring second messenger signals in a particular cell. In cardiac myocytes, the EGF receptor has been shown to stimulate cAMP accumulation. This effect seems to require the expression of ACV and does not occur when ACI, ACII, or ACVI is expressed (31). The mechanism of stimulation of cAMP accumulation by the EGF receptor is via activation of G_s (32), and an antibody against G_s blocks the stimulation of ACV by the EGF receptor in 293 cells (31). This suggests that the mechanisms determining specificity of receptor/AC coupling are not yet fully understood.

In summary, in heterologous systems, the D₃ receptor does not activate G proteins robustly, as reflected by our measurements of [³⁵S]GTPS binding, as well as by its lack of a shift in agonist affinity in the presence of GTP and its weak modulation of second messengers in most cell lines (see above). However, when it is coexpressed with ACV, as it may be in vivo, D₃ receptor activation results in a significant signal (i.e., inhibition of cAMP accumulation). This suggests that the D₃ receptor may generate meaningful second messenger signals in vivo. This also raises the possibility that specific isoforms of other dopamine receptor effectors, such as K⁺ or Ca²⁺ channels, may exist that are more sensitive to the weak G protein activation generated by the D₃ receptor than those present in the cultured cells examined to date.

The physiological significance of results obtained in transfected cell lines is often difficult to interpret. In these experiments, the D₃ receptor and ACV were expressed at higher levels than those that presumably exist in vivo; therefore, demonstration of selective coupling of the D₃ receptor to ACV is subject to the usual caveat of results obtained with overexpressed proteins in heterologous systems. However, regardless of whether inhibition of ACV plays an important role in the function of the D₃ receptor in vivo, our discovery of the generation of a robust second messenger signal by this receptor in a defined environment is highly significant. First, it indicates that this receptor is capable of substantially modulating second messenger levels under some conditions, which had previously been unclear. Second, this system provides a tool for further study of the D₃ receptor. Through coexpression of the D₃ receptor and ACV, new studies on the biology of this receptor can be undertaken, such as experiments designed to examine the regulation of the activity of this receptor. In addition, this discovery provides a system for the positive identification of D₃ receptor agonists and antagonists. The development of ligands that are selective for the D₃ receptor and clearly demonstrated to be agonists or antagonists is an important and necessary step in elucidating the physiological role of this poorly understood receptor.