The Rat D4 Dopamine Receptor Couples to Cone Transducin (Gαt2) to Inhibit Forskolin-stimulated cAMP Accumulation*

Based on its expression pattern and pharmacology, the D4 dopamine receptor may play a role in schizophrenia. Thus it is of interest to know what signaling pathways are utilized by this receptor. Previously, we showed that activation of D4 receptors in a mouse mesencephalic neuronal cell line (MN9D) inhibited forskolin-stimulated cAMP accumulation in a pertussis toxin-sensitive (Ptx-sensitive) fashion. Of the known Ptx-sensitive G-protein α subunits, MN9D-expressed Gαi2, GαoA, and GαoB; however, none of these coupled to the D4 receptor. Using a low stringency polymerase chain reaction cloning method, we found an additional Ptx-sensitive G-protein cone transducin (Gαt2) expressed in the MN9D cells. We also found that Gαt2 mRNA is highly expressed in rat mesencephalic tissue. To test the hypothesis that the D4 receptor couples to Gαt2, we cotransfected MN9D cells with the D4 receptor and a mutagenized Ptx-resistant Gαt2 subunit (mGαt2). Application of the dopaminergic agonist quinpirole to cotransfected cells inhibited forskolin-stimulated cAMP accumulation in the presence or absence of Ptx. To our knowledge, this is the first report demonstrating that the D4 dopamine receptor functionally couples to a specific G-protein and that a non-opsin-like receptor can couple with a transducin subunit.

Dopamine is a major neurotransmitter in a wide variety of organisms. In mammalian brain, dopamine modulates motor, affective, cognitive, and neuroendocrine functions. There are five cloned dopamine receptors, D 1-5, that can be distinguished by pharmacological and physiological criteria (1)(2)(3)(4)(5)(6)(7)(8). Of these, the D 4 receptor has engendered much interest due to its possible involvement in schizophrenia. The D 4 receptor exhibits a high affinity for clozapine, the prototype for a new class of antipsychotic drugs with reduced extrapyramidal side effects (8). This is in keeping with its distribution in limbic and cortical brain regions as well as its relative absence in striatal regions (6,8,9). Studies have also shown that D 4 receptors may be increased 2-fold in schizophrenic brain (10). In rat striatum, treatment for 1 month with the antipsychotic haloperidol elevates the density of D 4 receptors by about 2-fold, whereas D 2 and D 3 are only modestly affected (11).
Characterization of the signal transduction pathways of D 4 and other dopamine receptors in vivo has been difficult because of the heterogeneity of tissue and cell types as well as the lack of truly selective agonists and antagonists. Transiently or stably expressed D 4 receptors in several cell lines show similar binding characteristics and effects of guanine nucleotides on agonist binding (8,(12)(13)(14)(15). For example, in HEK and CHO cells transfected with the D 4 receptor, forskolin-stimulated cAMP accumulation is reduced by the D 2 , D 3 , and D 4 receptor agonist, quinpirole (8,14,15). Previous studies in our laboratory found that D 4 receptors are present in the photoreceptor layer of mouse retina and that dopaminergic agonists reduce light-sensitive cAMP levels with a D 4 -like pharmacological profile (16). We also demonstrated that D 4 receptors inhibited forskolin-stimulated cAMP accumulation in a dopaminergic cell line, MN9D (17,18). To identify the G-protein(s) that couples to the D 4 receptor, MN9D cells were cotransfected with D 4 and a pertussis toxin-resistant (Ptx-resistant) 1 mutant of G␣ i2 , G␣ i3 , G␣ oA (17), or G␣ oB . 2 However, none of the G-protein ␣ subunits tested could rescue receptor-mediated inhibition of forskolin-stimulated cAMP accumulation from Ptx treatment. Based on this result, we attempted to clone additional G-protein ␣ subunits from these cells using PCR-based strategies. Here we show that cone transducin (G␣ t2 ) is expressed in MN9D cells and demonstrate that it couples to the D 4 receptor.  mRNA Analysis-Total RNA was isolated from various tissues and MN9D cells as described (20). G␣ t2 -specific messages (21) were detected using primers o-955 (5Ј-TTCTCTAGAGCTGGAGAAGAAGCTG-3Ј, identical to nucleotides 58 -74) and o-978 (5Ј-CTAGAGTGGACATG-GCTC-3Ј, complementary to nucleotides 262-278) at 94°C for 30 s, 50°C for 30 s, and 72°C for 30 s for 30 cycles. In other experiments, primers o-955 and o-956 (5Ј-TATTCTAGATGGCCAGGATGGACTG-3Ј, complementary to nucleotides 241-258) were used. Products were electrophoresed, transferred to a nylon membrane, and hybridized with a radiolabeled internal primer o-1053 (5Ј-ATTGTCAAACAGATGAA-GAT-3Ј). Image analysis was performed using a PhosphorImager system (Molecular Dynamics, Sunnyvale, CA). The mRNA levels were normalized for equal amounts of the 18 S fragment of ribosomal RNA (22) as described (6).

Materials-All
Site-directed Mutagenesis-An expression vector containing the bovine G␣ t2 cDNA was kindly provided by N. Gautam (Washington University School of Medicine, St. Louis, MO). A Ptx-resistant G␣ t2 clone was constructed by introducing a cysteine-to-serine change four residues from the carboxyl terminus (23) using standard techniques for site-directed mutagenesis (24) and the primers o-983 (5Ј-ACCTCAA(A/ G)GACAGCGG(A/G)CTCTTCT-3Ј, nucleotides 1045-1066) and its complement, o-984 (25). The resulting clone (mG␣ t2 ) was confirmed by sequence analysis.
Tissue Culture and Transfection-MN9D cells (26,27) were maintained and transfected as described previously (17,18). Each stable subclone was screened for the rat D 4 receptor and mG␣ t2 mRNA expression by RT-PCR assays. Several rounds of further subcloning by limiting dilution were performed to establish homogeneous cell populations.
Membrane Preparation, Binding Assays, and cAMP Measurement-Membrane preparation, receptor binding assays, and cellular cAMP measurements were performed exactly as described (18). Saturation binding data were analyzed using programs from Lundon Software (Chagrin Falls, OH). For Ptx treatment, cells were incubated overnight in 50 ng/ml Ptx, which was previously shown to be a maximally effective dose (18).
Bead Capture Experiments-Experiments using the Capture Tec pHook-2 kit (Invitrogen) were performed according to manufacturer protocol. pHook-2 encodes a fusion protein consisting of a signal peptide, a single chain antibody against the hapten, phOx, and a transmembrane domain to anchor the antibody in the plasma membrane (28 -30). pHook-2-expressing cells can be isolated from whole cultures magnetically using beads coated with phOx.
Statistical Analysis-All data were normally distributed. For multiple comparisons within a given cell line, one-way analysis of variance was used to estimate significance followed by post hoc t tests corrected for multiple comparisons by the method of Bonferroni. For the single comparisons using a single cell line, paired t tests were used. All analyses were adjusted for inequality of variances when appropriate. All analyses were completed using the SAS suite of programs (SAS Institute Inc., Cary, NC).

Identification of G-protein ␣ Subunits in MN9D-Previously,
we showed that the G-protein ␣ subunits G␣ i2 and G␣ oA that are expressed in MN9D cells (18) do not couple to D 4 receptors in order to decrease cAMP levels (17). Similar studies have also ruled out coupling to G␣ i3 (17), G␣ i1 , and G␣ oB . 2 To determine whether other known or novel G␣ subunits were present in this cell line, we used a low stringency nested PCR approach. Sets of degenerate primers derived from conserved domains among the Ptx-sensitive G␣ subunits were synthesized and used to amplify cDNA derived from MN9D cells. PCR products were subcloned and transformed then randomly selected. Clones were further identified by sequence analysis. As shown in Table I, G␣ i2 and G␣ oA sequences were recovered (18) as well as a single G␣ i3 clone. No G␣ i1 or G␣ t1 clones were detected in this screen (not shown). Unexpectedly, 40% of the clones sequenced corresponded to mouse G␣ t2 . To verify that G␣ t2 was expressed in MN9D cells, we used mouse-specific primer pairs to amplify G␣ t2 mRNA. The predicted G␣ t2 band was seen at a level comparable to that present in mouse retina (Fig. 1A).
Tissue Distribution of G␣ t2 mRNA-Because MN9D cells were derived from dopaminergic mesencephalic cells (26,27), we examined mesencephalic tissue as well as other central nervous system and peripheral tissues for the presence and relative abundance of G␣ t2 transcripts. As suggested by its presence in MN9D cells, G␣ t2 mRNA was abundant in the mesencephalon as well as in the retina (Fig. 1B). Moreover, detectable but relatively low levels of G␣ t2 transcripts were seen in other central nervous system and peripheral tissues, with the exception of the kidney, where G␣ t2 mRNA was more abundantly expressed. Taken together, these data suggest that G␣ t2 is expressed outside of the retina, where it may couple to non-opsin-like receptors.
Establishment of a D 4 Receptor/mG␣ t2 Cell Line-To test the hypothesis that D 4 receptors can couple to G␣ t2 , we cotransfected MN9D cells with the rat D 4 receptor and a Ptx-resistant mG␣ t2 . In this paradigm, G␣ t2 has been mutated such that the expressed protein is incapable of being ADP-ribosylated, thus rendering it insensitive to Ptx (23). If D 4 receptors couple to G␣ t2 to inhibit cAMP accumulation, then cotransfection with the mutagenized G␣ t2 should rescue the receptor-stimulated inhibition of cAMP accumulation from Ptx treatment. Cotransfected, stable cell lines (17) were characterized by RT-PCR to confirm D 4 receptor expression (not shown) and G␣ t2 (Fig. 2A).  Because the primers chosen were from the bovine G␣ t2 3Јflanking region, the endogenous mouse cone message would not have been detected. Next, we used an indirect approach to look for mG␣ t2 protein expression because we were unable to obtain antibodies directed against bovine G␣ t2 . Since G␣ t2 cDNA was expressed together with the fusion protein in the pHook vector system, we measured expression of the pHook cell surface antibody. As shown in Fig. 2B nM (mean Ϯ S.D., n ϭ 3) (Fig. 2C). No specific [ 3 H]spiperone binding was detected in untransfected MN9D cells (18). D 4 Receptors Can Couple to G␣ t2 to Inhibit cAMP Accumulation-Previously, we showed that agonist stimulation of D 4 receptors in MN9D cells inhibited forskolin-stimulated cAMP accumulation by 30 -40% (17,18). Fig. 3 shows comparable and significant inhibition of cAMP accumulation in cells with or without G␣ t2 . D 4 -mediated inhibition of cAMP accumulation could be blocked by overnight pretreatment with 50 ng/ml Ptx (Fig. 3A and Ref. 18). In contrast, in the cotransfected MN9D-D 4 /mG␣ t2 cells, Ptx could no longer block the effects of agonist treatment (Fig. 3B). Thus the D 4 receptor can couple to G␣ t2, making it, to our knowledge, the first non-opsin-like receptor to couple to this retinal-enriched G-protein.
Transducins are known to couple to cyclic nucleotide phosphodiesterases (PDEs) leading to decreased cyclic nucleotide levels and closure of cation-specific channels in the plasma membrane (31). To test the hypothesis that D 4 -mediated inhibition of cAMP levels is achieved via coupling to G␣ t2 followed by activation of a cAMP PDE, we performed cAMP accumulation assays in the presence of several PDE inhibitors. We first tested 1 mM 3-isobutyl-1-methylxanthine (IBMX), a non-selective PDE inhibitor. However, IBMX did not block the inhibition of forskolin-stimulated cAMP accumulation by quinpirole (n ϭ 4) nor did 8-methoxymethyl-IBMX (25 M), a specific inhibitor of PDE type I, trequinsin (10 nM), a PDE type III specific inhibitor, rolipram (10 M), a type IV inhibitor, or a type V inhibitor (4-{[3Ј,4Ј-(methylenedioxy)benzyl]amino}-6-methoxyquinazoline, 1 M). We have also looked for but not found any changes in cGMP levels associated with receptor stimulation. Thus the downstream target of D 4 receptor-activated G␣ t2 (or its associated ␤␥ dimer) in MN9D cells is unclear at present. DISCUSSION PCR-based cloning techniques allowed us to survey the repertoire of Ptx-sensitive G-protein ␣ subunits in the mesencephalic-derived MN9D cell line (26,27). G␣ t2 was isolated with about the same frequency as G␣ i2 using several different primer sets. G␣ t2 was also expressed in nontransfected mesencephalic tissue as well as other central nervous system and peripheral tissue. Cellular expression of a Ptx-resistant mutant of G␣ t2 restored agonist-stimulated inhibition of cAMP accumulation in D 4 -expressing cells in which the endogenous G-proteins were uncoupled from the receptor by pretreatment with Ptx. This is the first report of a non-opsin-like receptor coupling to a transducin as well as the first identification of a specific G-protein coupled to the D 4 dopamine receptor.
Previously, we have shown that the D 4 receptor is most abundantly expressed in the inner segment layer of mouse photoreceptors (16). Moreover, in this tissue, the light-sensitive pool of cAMP can be eliminated in the dark by application of ligands with the rank order of affinities characteristic of the D 4 receptor. Thus, in vivo the D 4 receptor is physiologically coupled to the modulation of cAMP levels. Our finding that the D 4 receptor can couple to G␣ t2 in transfected cells raises the possibility that this also occurs within the retina.
Numerous studies have suggested that G␣ t2 is specifically expressed in the retina (25). Recently, however, Zigman et al. (32) found G␣ t2 mRNA in pancreatic islet cells, adrenal gland, and pituitary. The presence of G␣ t2 in immortalized cells from the central nervous system as well as in various central nervous system and peripheral tissues confirms and extends these observations. Interestingly, the D 4 receptor is expressed in some but not all of these regions. For example in rats, D 4 receptors are present in kidney (33) and retina (16) but not in mesencephalic tissues (6). Given the high level of G␣ t2 in the latter region (Fig. 1B), this suggests that G␣ t2 can couple to other receptors as well.
What is the function of G␣ t2 outside of the retina? Within photoreceptors, G␣ t1 and G␣ t2 activate cGMP PDE, leading to decreased cGMP levels and closure of cation-specific channels in the plasma membrane (31). However, recent studies by Margolskee and co-workers (34,35) have shown that rod transducin is also present in vertebrate taste cells, where it appears to activate a cAMP-specific PDE. In this cell type, phosphodiesterase-mediated degradation of cAMP appears to activate a cyclic nucleotide-suppressive channel, leading to depolarization and Ca 2ϩ influx (35). Thus, rod transducin also appears to be functionally important outside of the retina, although the receptor to which it is coupling has yet to be identified.
We tested the hypothesis that D 4 -mediated inhibition of cAMP accumulation is achieved via coupling to G␣ t2 followed by activation of a cAMP PDE by performing cAMP accumulation assays in the presence of several PDE inhibitors. To date, however, none of the PDE inhibitors tested appeared to have any effect. We have also looked for but not found any changes in cGMP levels associated with receptor stimulation. Conceivably, the ␤/␥ subunits associated with G␣ t2 might modulate adenylyl cyclase activity, as observed with ␤/␥ subunits associated with G␣ i or G␣ o (36). Alternatively, IBMX-insensitive PDEs might be involved. Further investigations are needed to determine what the downstream signaling molecule is in MN9D cells.
Although D 1 and D 2 dopamine receptors appear to mediate most of the known biological effects of dopamine receptors in the central nervous system, the presence and expression patterns of the less abundantly expressed receptors such as D 4 make their function all the more intriguing. The mechanisms mediating the effects of D 4 are of particular interest. Since the D 4 receptor may be involved in schizophrenia, knowledge of its signal transduction pathway might be of therapeutic importance as well.