Abstract
A range of ligands displayed agonism at the long isoform of the human dopamine D2 receptor, whether using receptor-G protein fusions or membranes of cells in which pertussis toxin-resistant mutants of individual Gαi-family G proteins could be expressed in an inducible fashion. Varying degrees of efficacy were observed for individual ligands as monitored by their capacity to load [35S]GTPγS onto each of Gαi1,Gαi2,Gαi3, and Gαo1. By contrast, (S)-(–)-3-(3-hydroxyphenyl)-N-propylpiperidine was a partial agonist when Gαo1 was the target G protein but an antagonist/inverse agonist at Gαi1,Gαi2, and Gαi3. In ligand binding assays, dopamine identified both high- and low-affinity states at each of the dopamine D2 receptor-G protein fusion proteins, and the high-affinity state was eliminated by guanine nucleotide. (S)-(–)-3-(3-Hydroxyphenyl)-N-propylpiperidine bound to an apparent single state of the constructs in which the D2 receptor was fused to Gαi1,Gαi2, or Gαi3. However, it bound to distinct high- and low-affinity states of the D2 receptor-Gαo1 fusion, with the high-affinity state being eliminated by guanine nucleotide. Likewise, although dopamine identified guanine nucleotide-sensitive high-affinity states of the D2 receptor when expression of pertussis toxin-resistant forms of each of Gαi1, Gαi2, Gαi3, and Gαo1 was induced, (S)-(–)-3-(3-hydroxyphenyl)-N-propylpiperidine identified a high-affinity site only in the presence of Gαo1. p-Tyramine displayed a protean ligand profile similar to that of (S)-(–)-3-(3-hydroxyphenyl)-N-propylpiperidine but with lower potency. These results demonstrate (S)-(–)-3-(3-hydroxyphenyl)-N-propylpiperidine to be a protean agonist at the D2 receptor and may explain in vivo actions of this ligand.
A large number of G protein-coupled receptors (GPCRs) are able to generate a variety of intracellular signals, and for those with a rich pharmacology of synthetic small-molecule ligands, it has often been possible to observe differential pharmacology for individual end points (Perez and Karnik, 2005). This has resulted in an appreciation that different ligands may stabilize distinct conformational states of GPCRs (Kenakin, 2001; Perez and Karnik, 2005) and in an expansion of the simple “active” or “inactive” “two-state” model (Leff, 1996) of GPCR function into “three-state” models (Leff et al., 1997) and subsequent chemical and physical considerations of GPCRs that allow the potential for an essentially unlimited number of states (Milligan and IJzerman, 2000; Vauquelin and Van Liefde, 2005). Although GPCRs are defined by their capacity to activate heterotrimeric G proteins, a number of ligand-induced signals seem not to require G protein interactions (Wei et al., 2003; Gesty-Palmer et al., 2006). In the case of the β2-adrenoceptor, for example, such separation of signal transduction has resulted in the identification of ligands that can be defined as inverse agonists for their effects on adenylyl cyclase activity but as agonists for their capacity to stimulate phosphorylation of the extracellular signal-regulated kinases 1 and 2 mitogen-activated protein kinases (Azzi et al., 2003; Galandrin and Bouvier, 2006). Ligands that display either positive or negative efficacy when assessed in different assays or in different experimental conditions have been described as “protean” ligands (Kenakin, 2001) and have been of particular value in defining the ability of GPCRs to adopt different conformational states. Many GPCRs are also able to couple to a number of different G proteins, and differences in agonist pharmacology to regulate signals via different G proteins are described frequently as “agonist-directed trafficking” (Kenakin, 1995). In many of these studies, observations have concentrated predominantly on measuring varying efficacy of ligands to regulate the production of two separate second messengers rather than directly measuring differential activation of two individual G proteins (Berg et al., 1998).
The dopamine D2 receptor has been one of the most studied monoaminergic GPCRs, not least because of the affinity of a wide range of antipsychotic agents for this receptor (Akam and Strange, 2004). As with a series of GPCRs that interact with members of the pertussis toxin-sensitive subgroup of G proteins, this receptor is able to initiate signals via each of Gαi1, Gαi2, Gαi3, and Gαo1 (Gazi et al., 2003). However, ligand pharmacology can be influenced greatly by the ratio of GPCR to G protein expression (Milligan, 2000), and this can be difficult to define in cells, particularly in studies designed to compare activation and function of different G proteins. One means to overcome this issue is to use GPCR-G protein fusions that ensure a fixed 1:1 stoichiometry of GPCR and G protein (Milligan et al., 2004). Because pertussis toxin-sensitive G proteins are endogenously expressed by all cells, we have also used previously variants of each of Gαi1,Gαi2,Gαi3, and Gαo1 that have been rendered insensitive to the ADP-ribosyltransferase activity of pertussis toxin by mutation of the cysteine, four amino acids from the C terminus that is the site of modification, to isoleucine (Bahia et al., 1998; Wise et al., 1999).
By using both fusions of the long isoform of the human dopamine D2 (D2L) receptor with pertussis toxin-resistant cysteine-isoleucine variants of each of Gαi1,Gαi2,Gαi3, and Gαo1 and cell lines stably expressing the D2L receptor in which varying amounts of each G protein can be expressed in an entirely tetracycline-dependent manner, we now demonstrate that both (S)-(–)-3-(3-hydroxyphenyl)-N-propylpiperidine [S-(–)-3-PPP] and p-tyramine are protean ligands at the D2L receptor, being agonists for activation of Gαo1 but antagonists/inverse agonists at Gαi1,Gαi2, and Gαi3. Such observations provide further evidence for the concept that the dopamine D2 receptor can exist in multiple conformational states and indicate that it is possible to selectively control the nature of signals generated by D2 receptor “agonists.” Because individual pertussis toxin-sensitive G proteins are expressed differentially pre- and postsynaptically (Aoki et al., 1992), these observations may be relevant to the reported in vivo actions of S-(–)-3-PPP (Arnt et al., 1983; Hjorth et al., 1983).
Materials and Methods
Materials. [3H]Spiperone (65–140 Ci/mmol) was from GE Healthcare (Chalfont St. Giles, Buckinghamshire, UK), and guanosine 5′-O-(3-[35S]thio)triphosphate ([35S]GTPγS; 1250 Ci/mmol) was from PerkinElmer Life and Analytical Sciences (Boston, MA). (+)-Butaclamol, dopamine, (–)-quinpirole, m-tyramine, p-tyramine, S-(–)-PPP, (R)-(+)-3-(3-hydroxyphenyl)-N-propylpiperidine [R-(+)-3-PPP], (R)-(–)-10,11-dihydroxy-N-n-propylnorapomorphine (NPA), (R)-(+)-7-hydroxy-DPAT hydrobromide (7-OH-DPAT), and GTPγS were purchased from Sigma (Gillingham, Dorset, UK). Spiperone hydrochloride was from Tocris (Bristol, UK). Oligonucleotides were from ThermoElectron (Ulm, Germany), and all materials for tissue culture were from Invitrogen (Paisley, UK). All other reagents were obtained as indicated.
D2 Dopamine Receptor Subcloning into pcDNA3. D2L was initially in the vector pDEST12.2. D2L cDNA was amplified by PCR using the following primers: sense, AAA AGA ATC CGC CAC CAT GGA TCC ACT GAA TCT GTC C; antisense, AAA ACT CGA GTC AGC AGT GGA GGA TCT TCA GGA AGG. Underlined bases indicate the restriction sites EcoRI (sense) and XhoI (antisense). The resulting PCR fragment was digested with EcoR I and XhoI and inserted into pcDNA3.
Construction of the Myc-D2L-G-Protein α Subunit Fusion Proteins. Pertussis toxin-resistant α2A-adrenoceptor-G-protein fusion proteins had been prepared as described previously (Wise and Milligan, 1997; Cavalli et al., 2000). In brief, Cys351 of rat Gαi1,Gαi3, and Gαo1 (Cys352 in Gαi2) was mutated to isoleucine by site-directed mutagenesis and then used to create the α2A-adrenoceptor-Gα fusion proteins using the porcine α2A-adrenoceptor in pcDNA3. These constructs were cloned into pcDNA3 using a created 5′ KpnI site and 3′ EcoRI site with a NcoI site between receptor and G-protein α subunit cDNAs. To create D2L:G protein α subunit proteins, the first step was to remove the NcoI site from within the D2L cDNA by site-directed mutagenesis using a QuikChange Mutagenesis kit (Stratagene, La Jolla, CA) and the following primers: sense, 5′-CC GAC CCG TCC CAT CAT GGT CTC CAC AG-3′; antisense, 5′-CT GTG GAG ACC ATG ATG GGA CGG GTC GG-3′. Boldface letters indicate altered bases. The PCR product was then digested with DpnI and transformed into bacteria. In a similar manner, NcoI sites were removed from both the Gαi1 and Gαo1 cDNAs in the respective α2A-adrenoceptor-Gα fusion protein cDNAs using the following primers: Gαi1 sense, 5′-TT GCC ATC ATT AGA GCG ATG GGG AGA TTG AAA ATC G-3′; antisense, 5′-C GAT TTT CAA TCT CCC CAT CGC TCT AAT GAT GGC AA-3′; and Gαo1 sense, 5′-CC ATT GTG CGG GCG ATG GAT ACT CTG GG-3′; antisense, 5′-CC CAG AGT ATC CAT CGC CCG CAC AAT GG-3′.
Myc-D2L (NcoI-) was amplified by PCR using the following primers: sense, 5′-AGA ACG GGG TAC CTT ATG GAA CAA CAA AAA CTT ATT TCT GAA GAA GAT CTG GAT CCA CTG AAT CTG TCC TGG TAT GAT G-3′; antisense, 5′-AAAAAAAACCAT GGAGTGGAGGATCTTCAGGAAGGC-3′. Underlined bases indicate introduced restriction sites (sense, KpnI; antisense, NcoI), bases in bold-face type indicate introduced N-terminal Myc tag. The PCR fragment was digested using KpnI and NcoI.
The α2A-adrenoceptor-Gα fusion proteins (NcoI-) were excised from pcDNA3 using KpnI and EcoRI, digested with NcoI, and the Gα subunit cDNA was purified. The Gαi subunit cDNAs were then cloned into pcDNA3 with the Flag-D2L PCR fragment to create the four D2L:G-protein α subunit fusion proteins.
Flp-In Constructs. Previously, pertussis toxin-resistant mutants of rat Gαi1,Gαi3, and Gαo1 were created by mutation of Cys351 to isoleucine (Cys352 for Gαi2) by site-directed mutagenesis. These were cloned into pcDNA3. cDNAs were excised using KpnI and ApaI (Gαi1–3) or ApaI (Gαo1) and subcloned into the pcDNA5/FRT/TO vector (Invitrogen).
Cell Culture and Transfection. HEK293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 0.292 g/l l-glutamine and 10% (v/v) newborn calf serum at 37°C in a 5% CO2 humidified atmosphere. Cells were grown to 60 to 80% confluence before transient transfection. Transfection was performed using Lipofectamine transfection reagent (Invitrogen) according to the manufacturer's instructions.
Generation of Stable Flp-In T-REx HEK293 Cells. To generate Flp-In T-REx HEK293 cells able to inducibly express the G protein α subunit of interest, the cells were transfected with a mixture containing the desired G protein α subunit cDNA in the pcDNA5/FRT/TO vector and pOG44 vectors (1:9) using Lipofectamine according to the manufacturer's instructions. Cell maintenance and selection were as described previously (Milasta et al., 2006). Clones were screened for G protein expression by Western blotting. To constitutively stably coexpress the D2L receptor in inducible cell lines, the appropriate cells were further transfected with the D2L receptor cDNA in pcDNA3 as described above, and resistant cells were selected in the presence of 1 mg/ml G418. Resistant clones were screened for receptor expression using specific [3H]spiperone binding. Cells were treated with 1 μg/ml doxycycline 24 to 48 h before assay to induce the expression of G protein α subunits cloned into the Flp-In locus.
Membrane Preparation. Cells were collected by centrifugation (1700g, 5 min, 4°C) frozen at –80°C for at least 1 h and resuspended in 15 ml of buffer (10 mM Tris and 0.1 mM EDTA, pH 7.4). Cell suspensions were then homogenized using an Ultra Turrax for 3 × 20 s. The homogenate was centrifuged at 1700g for 10 min, and the supernatant was collected and centrifuged at 48,000g for 45 min at 4°C. The resulting pellet was resuspended in buffer and stored at –80°C in aliquots of 1 ml.
Saturation Binding Assays Using [3H]Spiperone. Cell membranes (10 μg of protein) were incubated in triplicate with [3H]spiperone (0.001–2 nM) in a total volume of 1 ml of buffer (20 mM HEPES, 6 mM MgCl2, 1 mM EDTA, and 1 mM EGTA, pH 7.4). Nonspecific binding was determined by the inclusion of 10 μM(+)-butaclamol. The reaction was initiated by the addition of membranes, and the tubes were incubated at 25°C for 3 h. The reaction was terminated by rapid filtration using a Brandel cell harvester with three 5-ml washes of ice-cold phosphate-buffered saline (140 mM NaCl, 10 mM KCl, 1.5 mM KH2PO4, and 8 mM Na2HPO4). The filters were soaked in 3 ml of scintillation fluid, and radioactivity present was determined by liquid scintillation spectrometry.
Agonist Competition versus [3H]Spiperone Binding. Cell membranes (10 μg of protein) were incubated with 0.05 nM [3H]spiperone and various concentrations of agonists, in triplicate, in a final volume of 1 ml of buffer (20 mM HEPES, 6 mM MgCl2, 1 mM EDTA, 1 mM EGTA, and 40 μM ascorbic acid, pH 7.4). Nonspecific binding was determined by the inclusion of 10 μM (+)-butaclamol. The reactions were initiated, incubated, and terminated as described above. The effect of guanine nucleotides on dopamine binding was assessed by the addition of 100 mM NaCl and 100 μM GTP to the buffer.
[35S]GTPγS Binding Assays. Cell membranes (10 μg) were incubated in 900 μl of buffer (20 mM HEPES, 100 mM NaCl, 6 mM MgCl2, and 40 μM ascorbic acid, pH 7.4) containing 10 μM GDP and various concentrations of ligands. All experiments were performed in triplicate. The reaction was initiated by the addition of cell membranes and incubated at 30°C for 30 min. A 100-μl volume of [35S]GTPγS (0.1 nM final concentration) was then added, and the incubation continued for a further 30 min. The reaction was terminated by rapid filtration with a Brandel cell harvester and three 4-ml washes with ice-cold phosphate-buffered saline. Radioactivity was determined as described for saturation analysis. For antagonist dose-response assays, an EC50 concentration of dopamine was added along with various concentrations of antagonist.
[35S]GTPγS Binding Assay: Agonist Stimulation of [35S]GTPγS Binding by Fusion Proteins. [35S]GTPγS binding assays were performed at room temperature in 384-well format. Membranes (10 μg/point) were diluted to 0.4 mg/ml in assay buffer (20 mM HEPES, 100 mM NaCl, and 10 mM MgCl2, pH 7.4) supplemented with saponin (10 mg/l) and preincubated with 10 μM GDP and wheat germ agglutinin SPA beads (GE Healthcare) (0.5 mg) and incubated at room temperature for 45 min with agitation. Various concentrations of D2 dopamine receptor agonists were added, followed by [35S]GTPγS (1170 Ci/mmol; GE Healthcare) at 0.3 nM (total volume of 46 μl), and binding was allowed to proceed at room temperature for 4 hours. Bound [35S]GTPγS was determined by scintillation counting on a ViewLux ultraHTS Microplate Imager (PerkinElmer).
Data Analysis. Data were analyzed using Prism software (GraphPad Software Inc., San Diego, CA).
Results
Studies with Dopamine D2 Receptor-G Protein Fusions. It has been established previously that the dopamine D2 receptor is able to interact with and activate each of the pertussis toxin-sensitive G proteins Gαi1, Gαi2, Gαi3, and Gαo1 (Gazi et al., 2003). Because particularly Gαi2 and Gαi3 are expressed endogenously by virtually all cells and we wished to examine potential variations in the ability of ligands at the D2L receptor to activate the different G proteins, cysteine to isoleucine, pertussis toxin-insensitive mutants of the α subunit of each of Gi1,Gi2,Gi3, and Go1 (Wise et al., 1999) were linked in-frame with the C-terminal tail of the human D2L (Fig. 1A). This ensured that the stoichiometry of receptor to G protein would be identical for each G protein to be studied and that the relative cellular distribution and orientation of receptor and G protein would be uniform. Each fusion protein was expressed transiently in HEK293 cells. After pertussis toxin treatment (25 ng/ml, 24 h) of these cells to cause ADP-ribosylation of the endogenously expressed forms of “Gαi” and membrane preparation, saturation binding assays using [3H]spiperone indicated each fusion protein to be expressed to similar levels and to bind this ligand with similar and high affinity (Table 1). Membranes of HEK293 cells mock-transfected or transfected to express D2L-Cys351Ile Gαo1 and treated or not with pertussis toxin were used in [35S]GTPγS binding studies. In the absence of D2L-Cys351Ile Gαo1, binding of the nucleotide was low, essentially unaffected by pertussis toxin treatment, and not modulated by the addition of dopamine (Fig. 1B). With expression of D2L-Cys351Ile Gαo1, [35S]GTPγS binding in the absence of dopamine was increased, and this level was increased substantially further in the presence of dopamine (Fig. 1B). Pertussis toxin treatment produced a small decline in dopamine-stimulated [35S]GTPγS binding, consistent with the D2L receptor within the fusion being able to access endogenously expressed G proteins, but after pertussis toxin treatment, the elevation of [35S]GTPγS binding by dopamine remained robust (Fig. 1B), indicating direct activation of the fused G protein by the D2L receptor. Equivalent [35S]GTPγS binding assays demonstrated all of the fusion proteins to be activated in a pertussis toxin-insensitive manner by dopamine, NPA, quinpirole, m-tyramine, and R-(+)-3-PPP (Fig. 1C), although compared with dopamine, only NPA was a full agonist at each construct, and potency of the individual ligands varied significantly at the various fusion constructs (Table 2). With the exception of NPA and 7-OH DPAT, potency of the ligands was greatest for D2L-Cys351Ile Gαo1 (Table 2), whereas the potency of quinpirole was particularly low at D2L-Cys352Ile Gαi2 (Table 2). Unlike the ligands mentioned above, although both p-tyramine and S-(–)-3-PPP displayed agonism at the D2L-Cys351Ile Gαo1 fusion (Table 2), they did not enhance [35S]GTPγS binding to any of the other fusions. Although p-tyramine displayed greater agonist efficacy than S-(–)-3-PPP at the D2L-Cys351Ile Gαo1 fusion, subsequent detailed studies used S-(–)-3-PPP because its potency as an agonist at D2L-Cys351Ile Gαo1 was 300-fold greater than p-tyramine (Table 2).
The ability of dopamine to compete with [3H]spiperone to bind to the various D2L-G protein fusions (Fig. 2) was best fit by a two-site model in which between 30 and 50% of the sites displayed higher affinity (pKh = 7.1–7.7) and the remainder lower affinity (pKl = 5.6–5.8) for dopamine (Table 3). The presence of 100 μM GTP in such assays resulted in this competition becoming monophasic (pK = 5.5–5.9) in each case (Table 3). By contrast, the ability of S-(–)-3-PPP to compete with [3H]spiperone (Fig. 2) was monophasic in the absence of GTP and essentially unaffected by the presence of GTP (pK = 6.2–6.5) for each of the fusions except for D2L-Cys351Ile Gαo1 in which a biphasic competition curve (pKh = 8.4, pKl = 6.2) was converted to a monophasic curve (pK = 6.3) in the presence of GTP (Table 3).
To explore the details of this apparent protean (Kenakin, 2001) characteristic of S-(–)-3-PPP at the D2L, we compared effects at D2L-Cys351Ile Gαo1 and D2L-Cys352Ile Gαi2 because Gαi2 and Gαo1 have the lowest sequence identity among the four G proteins studied, and the observed variation in ligand potency for activation of Gαo1 and the other G proteins was most consistent for Gαi2. Increasing concentrations of S-(–)-3-PPP inhibited the capacity of an EC50 concentration of dopamine to enhance binding of [35S]GTPγSto both fusion constructs with pIC50 = 4.68 ± 0.07 (D2L-Cys352Ile Gαi2) and 5.09 ± 0.12 (D2L-Cys351Ile Gαo1) (Fig. 3A). However, the maximal effect of S-(–)-3-PPP at D2L-Cys351Ile Gαo1 in this assay confirmed its partial agonist action at Gαo1 because it failed to reduce binding of [35S]GTPγS to the level observed in the absence of dopamine. By contrast, at D2L-Cys352Ile Gαi2, S-(–)-3-PPP completely blocked dopamine stimulation of [35S]GTPγS and, indeed, acted as an efficacious inverse agonist (Fig. 3A). Spiperone is frequently described as an inverse agonist at the dopamine D2 receptor, and accordingly, spiperone also acted as an effective inverse agonist at D2L-Cys352Ile Gαi2 (Fig. 3B). Furthermore, this ligand also completely reversed the effect of dopamine at D2L-Cys351Ile Gαo1 (Fig. 3B). Further studies demonstrated that both spiperone and S-(–)-3-PPP also acted as antagonists/inverse agonists at D2L-Cys351Ile Gαi1 and D2L-Cys351Ile Gαi3 (Fig. 3, C and D).
Cell Lines that Express Gαi1,Gαi2,Gαi3, or Gαo1 in anInducible Manner. Although the fusion proteins described above have a major advantage in defining and ensuring the receptor-to-G protein ratio for each G protein, they are an inherently artificial system. To examine whether S-(–)-3-PPP would also behave as a protean agonist at the D2L receptor when regulating different G proteins in a system expressing separated receptor and G protein, we generated a series of HEK293 cell lines based on the Flp-In T-REx system (Ellis et al., 2006; Milasta et al., 2006). In these cell lines, the D2L receptor was expressed stably and constitutively, whereas the pertussis toxin-resistant cysteine to isoleucine forms of Gαi1, Gαi2, Gαi3, and Gαo1 were cloned into the Flp-In locus, allowing their expression in an entirely inducible manner from the same single defined chromosomal locus by the addition of tetracycline. In concert with pertussis toxin treatment, to cause ADP-ribosylation of endogenously expressed forms of Gαi, we anticipated that this would produce a second alternative system in which D2L receptor-mediated stimulation of [35S]GTPγS binding must reflect only activation of a single defined G protein. Initial studies confirmed expression of the G protein of interest in all cases in a “tetracycline-on” fashion (Fig. 4). At a 24-h time point, maximal expression of each G protein was achieved by treatment of the cells with between 0.5 and 1.0 μg/ml tetracycline (Fig. 4). Levels of the D2L receptor constitutively expressed by each of the cell lines were not affected (p > 0.05) by tetracycline-induced turn-on of the G proteins (Table 4). As anticipated, in membranes of pertussis toxin-treated cells that were not treated with tetracycline, there was no capacity of dopamine to stimulate binding of [35S]GTPγS, whereas in equivalent membranes of cells treated with tetracycline to cause the expression of Cys352IleGαi2 dopamine produced a robust stimulation of [35S]GTPγS binding (Fig. 5A). Although this effect of dopamine was both substantial and concentration-dependent (Fig. 5B), S-(–)-3-PPP was unable to enhance [35S]GTPγS binding at all and, indeed, tended to reduce basal [35S]GTPγS binding (Fig. 5B). In membranes of pertussis toxin-treated cells expressing the D2L receptor and induced to express Cys351IleGαo1, dopamine was also able to stimulate binding of [35S]GTPγS (Fig. 6A) in a concentration-dependent fashion (Fig. 6B), and now S-(–)-3-PPP also enhanced [35S]GTPγS binding but functioned as a partial agonist (Fig. 6B).
The ability of dopamine to compete with [3H]spiperone to bind the D2L receptor was monophasic, of low-affinity, and insensitive to GTP in membranes of pertussis toxin-treated cells not induced to express Cys352Ile Gαi2 (Fig. 7A). However, tetracycline induction of Cys352Ile Gαi2 expression resulted in the appearance of a high-affinity site for dopamine that was eliminated in the presence of GTP (Fig. 7A). By contrast, both with and without tetracycline induction of Cys352Ile Gαi2 expression, the capacity of S-(–)-3-PPP to compete with [3H]spiperone was monophasic and unaffected by the presence of GTP (Fig. 7B). In equivalent membranes of pertussis toxin-treated cells that allowed inducible expression of Cys351IleGαo1, dopamine again identified both high- and low-affinity sites in [3H]spiperone competition binding studies only after treatment with tetracycline (Fig. 8A). As anticipated, the high-affinity site was absent in the presence of GTP (Fig. 8A). It is interesting that although not as pronounced as with dopamine, S-(–)-3-PPP also identified both high- and low-affinity states in membranes of cells induced to express Cys351IleGαo1 (Fig. 8B).
Finally, we explored the pharmacology of both spiperone and S-(–)-3-PPP in membranes of cells expressing the D2L receptor and induced to express either Cys352Ile Gαi2 (Fig. 9A) or Cys351IleGαo1 (Fig. 9B). Dopamine-mediated stimulation of [35S]GTPγS binding was completely reversed by both spiperone and S-(–)-3-PPP when Cys352Ile Gαi2 was the target (Fig. 9A), but whereas spiperone also fully reversed dopamine stimulation of [35S]GTPγS binding to Cys351IleGαo1 (Fig. 9A), S-(–)-3-PPP produced only partial inhibition (Fig. 9B).
Discussion
A number of recent studies have provided evidence that different agonist ligands at a single GPCR can selectively identify and stabilize distinct confirmations or sets of conformations of the receptor (Ghanouni et al., 2001; Krueger et al., 2005; Yao et al., 2006). This can result in differences in the ability of individual agonists to control pairs of signal transduction pathways that are modulated by that GPCR (Perez and Karnik, 2005). Such observations are important in that they provide suggestions of mechanisms that may explain differential functional properties of individual agonist ligands that are believed to be selective for a single GPCR and give insights into the flexibility of GPCR structures. Furthermore, a number of studies have identified differences in the potency or efficacy of agonist ligands to stimulate different G proteins if a GPCR has the potential to interact with more than one G protein (Cussac et al., 2002). Examples of this have been reported for GPCRs that interact selectively with members of the Gi group of pertussis toxin-sensitive G proteins, including the dopamine D2 receptor (Cordeaux et al., 2001). Because the members of this G protein subfamily that are widely expressed are similar in sequence and, hence, presumably in structure, it is not surprising that many GPCRs can interact with and activate more than one member of the family.
Studies of potential selectivity are hampered by the coexpression of a number of these G proteins in essentially all mammalian cells and tissues. One means to overcome this has been to use insect cell systems (Clawges et al., 1997; Cordeaux et al., 2001) in which levels of expression, or the sequence conservation, of such G proteins is low. When mammalian G proteins are introduced into such cells along with a GPCR of interest, activation is largely restricted to the exogenous G protein. An alternative that allows the use of mammalian cell lines has been to use pertussis toxin-resistant variants (Ghahremani et al., 1999; Wise et al., 1999) in which the cysteine residue that is the target for toxin-mediated ADP-ribosylation is altered to another amino acid but retains the capacity to interact with GPCRs. When such mutants are expressed in mammalian cells, treatment of the cells with pertussis toxin results in ADP-ribosylation of the endogenously expressed forms of Gi and an inability of GPCRs to cause their activation. Studies detailed previously on Gαi1 replaced the relevant cysteine with every other natural amino acid and assessed the impact on GPCR-mediated activation (Bahia et al., 1998), and similar studies were subsequently performed with Gαi3 (Dupuis et al., 2001).
However, these widely used approaches are still not ideal for detailed studies on differential agonist actions at different G proteins if expression levels are not carefully controlled. Most importantly, although it is well-established that alterations in GPCR-to-G protein ratios can alter ligand function and receptor pharmacology (Milligan, 2000), this can be a challenge to control. For example, in previous studies examining interactions between the D2L and different G protein α subunits in insect Sf9 cells, receptor-to-G protein ratios ranging from 1:3 to 1:14 were reported for the different G proteins (Gazi et al., 2003). In the current studies, we therefore combined the use of pertussis toxin-resistant forms of the various Gi-like G proteins with both GPCR-G protein fusion technologies (Milligan et al., 2004) and with the inducible expression of individual G proteins from a single defined site of chromosomal integration in Flp-In T-REx HEK293 cells (Ellis et al., 2006; Milasta et al., 2006).
Although an artificial system, GPCR-G protein fusions can greatly improve signal-to-background in [35S]GTPγS binding studies (Milligan, 2003; Milligan et al., 2006), and they ensure the same receptor to G protein stoichiometry for each construct. Although the Flp-In T-REx HEK293 cells cannot ensure exactly the same level of expression of each G protein, each G protein is produced from the same single chromosomal location, and this overcomes the potential for different clones to have more than a single site of integration of the cDNA of interest and that different sites of integration might alter the effectiveness of expression. Furthermore, the inducible nature of expression from this locus, combined with the use of pertussis toxin-insensitive mutants and pertussis toxin pretreatment, provided a null background for expression of each G protein.
As anticipated, the majority of dopamine D2 receptor agonists stimulated [35S]GTPγS binding to all of the four G proteins, although, as reported by others (Gazi et al., 2003), significant variation in potency and efficacy could be observed. However, although both p-tyramine and S-(–)-3PPP were agonists, at Cys351IleGαo1, they both failed to act in this manner for the other three Gα subunits. The potency of p-tyramine was sufficiently low to limit its usefulness for detailed studies. The higher potency of S-(–)-3PPP, however, allowed concentration-response curves to demonstrate that it was able to fully inhibit dopamine-stimulated binding of [35S]GTPγS to Cys352Ile Gαi2 and, indeed, acted as an inverse agonist. By contrast, even at maximally effective concentrations, S-(–)-3PPP was unable to fully reverse dopamine-stimulated binding of [35S]GTPγS to Cys351Ile Gαo1, consistent with the direct measures of its partial agonist activity. In competition studies using two agonist ligands of varying efficacy, full receptor occupancy by the ligand with lower efficacy is expected to result in a direct measure of the efficacy of that ligand.
In further support of the protean effect of S-(–)-3PPP at different G proteins, ligand binding studies identified both high- and low-affinity states of the D2L receptor for dopamine with each of the four G proteins but distinct high- and low-affinity states of the receptor for S-(–)-3PPP only for Cys351Ile Gαo1. Similar results were obtained using both the receptor-G protein fusions and cells able to produce the G protein of choice on demand. Indeed, the Flp-In T-REx cells were particularly useful in this regard because, with pertussis toxin treatment but without tetracycline-induction of expression of an appropriate G protein, all of the [3H]spiperone binding sites displayed monophasic and low-affinity interactions with both dopamine and S-(–)-3PPP, whereas induction of expression resulted in the development of a high-affinity state for dopamine, no matter which of the G proteins was expressed. By contrast, only expression of Cys351Ile Gαo1 resulted in the appearance of a high-affinity state for S-(–)-3PPP.
In general, the potency of the ligands used was higher for activation of Gαo than for Gαi2. There were not statistically valid differences between the values obtained for Gαi1,Gαi2, and Gαi3 for enough compounds to allow us to convincingly state rank-order potency differences (which might reflect selective stabilization of distinct states of the receptor) for interactions with Gαi1 versus Gαi2, for example. This may reflect that the amino acid sequence identities for Gαi1,Gαi2, and Gαi3 are all between 86 and 94%, whereas for each of these against Gαo, sequence identity lies between 70 and 73%. It might, therefore, be postulated that greater variation in ligand conformational states could be observed for interactions with Gαo versus the others rather than between Gαi1, Gαi2, and Gαi3.
These studies may have implications for the action of S-(–)-3PPP. This ligand has been described to have agonist and antagonist properties in physiologically relevant end-points (Arnt et al., 1983; Hjorth et al., 1983). Of course, one explanation of such observations might relate to its partial agonist function at Gαo1 and antagonist/inverse agonist function at Gαi1,Gαi2, and Gαi3. In two studies with S-(–)-3-PPP, patients with schizophrenia showed improvements in both positive and negative symptoms but a limited duration of effectiveness (Lahti et al., 1998). It has been postulated that this is due to the action of S-(–)-3-PPP as a D2-like dopamine receptor partial agonist and that this would have a “dopamine system stabilization effect,” that is, normalization of both dopamine hypo- and hyperactivity in the pathologically affected dopaminergic tracts observed in patients with schizophrenia (Lieberman, 2004). Likewise, the atypical antipsychotic aripiprazole, which now has Food and Drug Administration approval for the treatment of schizophrenia, has also been characterized as a partial agonist at D2 receptors and again to have a dopamine stabilization effect (Cosi et al., 2006). However, the intrinsic activity and potency of aripiprazole at the D2 receptor is both cell line- and assay-dependent. For example aripiprazole has been shown to be a partial agonist for inhibition of cAMP accumulation in a Chinese hamster ovary cell line but an antagonist for [35S]GTPγS binding. Most significant, however, is the observation that, like S-(–)-3-PPP, this drug is reported to antagonize postsynaptic D2 receptors but partially activate presynaptic autoreceptors (Kikuchi et al., 1995). In agreement with these findings, a recent study has demonstrated the differential signaling of aripiprazole for several D2L-mediated pathways (Urban et al., 2007). These observations are consistent with aripiprazole, like S-(–)-3-PPP, having differential pharmacology at different signaling pathways, and it will be interesting to ascertain whether aripiprazole has similar protean characteristics in terms of G protein coupling. It will now be interesting to explore the more general contribution of protean effects of clinically relevant ligands that seem to target the same receptor.
Acknowledgments
J. Robert Lane thanks the Biotechnology and Biosciences Research Council for a CASE studentship.
Footnotes
-
Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org.
-
doi:10.1124/mol.106.032722.
-
ABBREVIATIONS: GPCR, G protein-coupled receptor; 7-OH-DPAT, R-(+)-7-hydroxy-2-dipropylaminotetralin hydrobromide; D2L, long isoform of the human dopamine D2 receptor; [35S]GTPγS, guanosine 5′-O-(3-[35S]thio)triphosphate; NPA, R-(–)-10,11-dihydroxy-N-n-propylnorapomorphine; R-(+)-3-PPP, (R)-(+)-3-(3-hydroxyphenyl)-N-propylpiperidine; S-(–)-3-PPP, (S)-(–)-3-(3-hydroxyphenyl)-N-propylpiperidine; PCR, polymerase chain reaction; HEK, human embryonic kidney.
- Received November 15, 2006.
- Accepted February 7, 2007.
- The American Society for Pharmacology and Experimental Therapeutics