Partial Agonist Clonidine Mediates α₂-AR Subtypes Specific Regulation of cAMP Accumulation in Adenylyl Cyclase II Transfected DDT1-MF2 Cells

Materials and Methods

[³H]cAMP (30 Ci/mmol), [³²P]ATP (30 Ci/mmol), [³H]rauwolscine (81 Ci/mmol), [α-³²P]GTP (3000 Ci/mmol), and [¹²⁵I]cAMP radioimmunoassay kit were purchased from NEN Life Sciences Products (Les Ulis, France). ARC-239 bichloride (2[2-[4(O-methoxy-phen) piperazine-1-y]4,4dimethyl-1,3,2H-4H] isoquinolinedione) was a gift from Karl Thomae (Biberach, Germany). 4-Azidoanilide was obtained from Fluka Biochemicals (Saint Quentin Fallavier, France). 1-Ethyl-3-[3-(dimethylamino)propyl] carbodiimide HCl and Dowex 50W-X4 (100–200 mesh, hydrogen form) were from Bio-Rad (Ivry sur Seine, France). PEI-cellulose plates and aluminum oxide (90 active neutral) were purchased from Merck (Nogent sur Marne, France). Cell culture supplies were obtained from Life Technologies (Cergy-Pontoise, France). Pansorbin cells were supplied from Calbiochem (Meudon, France). GF/C glass-fiber filters were from Millipore (Saint Quentin en Yvelines, France). All remaining drugs were from Sigma-Aldrich (Saint Quentin Fallavier, France).

Cell Culture and Transfection.

DDT1-MF2 cells were grown and stably transfected with human α_2B(α₂C2) or rat α_2A/D(RG20) cDNA as described previously (Duzic and Lanier, 1992). Resistant clones were tested for their α₂-AR binding capacity using the selective tritiated antagonist [³H]rauwolscine as described under binding studies. Clones expressing a receptor density between 1 and 1.5 pmol/mg of membrane proteins, were further cotransfected with the ACII cDNA expression vector and the drug resistance cassette pHyg according to the transfection strategy described by Gorman (1986). Transfected cells were selected by their resistance to hygromycin B (750 mg/ml). Each clone was then analyzed for AC II expression by Northern blot using a³²P-labeled cDNA AC II probe [rat full-length (4 kilobases)] as described previously (Marjamaki et al., 1997) and tested for enzyme activity.

Partially Purified Membrane Preparation.

Membranes were prepared by hypotonic lysis in ice-cold lysis buffer (5 mM Tris-HCl, pH 7.5, 5 mM EDTA, 5 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride, 10 mg/ml aprotinin, and 10 mg/ml pepstatin A) and collected by centrifugation (17,000g for 15 min at 4°C). Membrane pellet was resuspended in 50 mM HEPES, pH 8.0, for adenylyl cyclase assay or 50 mM Tris, pH 7.5, 5 mM MgCl₂, 0.6 mM EDTA for binding studies. Protein concentration was determined according to the method of Schacterle and Pollack (1973) using bovine serum albumin as standard.

Binding Studies.

For saturation experiments, membranes (30–40 μg) were incubated with the required concentrations of [³H]rauwolscine [1–50 nM] for 20 min at 25°C in a final volume of 100 μl. Nonspecific binding was determined in the presence of 10 μM phentolamine. Competition studies were performed in presence of increasing concentrations (10 pM–50 μM) of various competitors and 8 nM [³H]rauwolscine (a concentration near theK_D value). Bound radioligand was separated from the free by vacuum filtration over GF/C glass-fiber filters as described previously (Bouet-Alard et al., 1997). Radioactivity was counted by liquid scintillation in a 1214 Rack-β spectrometer (LKB, Rockville, MD) with a counting efficiency of approximately 30%.

Data for saturation and competition studies were analyzed by a nonlinear least-squares, curve-fitting GraphPad program (Graph Pad Software, San Diego, CA). Iterative curve fitting to experimental data from one site model provided IC₅₀. IC₅₀ were converted toK_i values using the equation of Cheng and Prussof (1973).

Adenylyl Cyclase Assay and Determination of Intracellular cAMP Accumulation.

Adenylyl cyclase activity was measured as described previously (E. Duzic and S.M. Lanier, 1992) using 50 μg of crude membrane. For intracellular cAMP accumulation, cells were plated at a concentration of 5 × 10⁵ cells/well in six-well plates and incubated at 37°C for 24 h. One hour before starting the experiment, the medium was removed and replaced with 4 ml of serum-free DMEM containing 20 mM HEPES, pH 7.5, and 250 mM isobutyl-l-methylxanthine. Then, cells were incubated with drugs to be tested for 10 min at 37°C. The reaction was stopped by aspiration of the medium and cells were disrupted by the addition of 1 ml of 10% ice-cold trichloracetic acid per well. After recovering the cellular lysate by scrapping the wells, samples were centrifuged (10,000g, 15 min at 4°C). The supernatants were then extracted 3 times with diethyl ether (1v/4v) and cAMP contents were determined by a cAMP radioimmunoassay system obtained from NEN Life Sciences Products.

Immunoblot and Immunoprecipitation.

For immunoblotting, 50 μg of membranes obtained from DDT1-MF2 cells or tissues were resolved on 10% SDS-PAGE and transfer to polyvinylidene difluoride-transfer membrane POLYSCREEN (NEN Research Products, DuPont de Nemours, France). After transfer, the blots were probed with anti-Gα protein subtype specific antibodies as described previously (Cohen-Tannoudji et al., 1995). Briefly, after blocking, the polyvinylidene difluoride blots were incubated with the primary antibody [AS/7 (anti-Giα₂/Giα₁) or EC/2 (anti-Giα₃/Gαο) or GC/2 (anti-Gαο)] for 1 h in a high-detergent 5% nonfat dry milk/Tris-buffered saline at room temperature at a dilution of 1:1000. After four successive washes, the blots were incubated for 1 h with HRP-conjugated secondary antibody in high-detergent 5% nonfat dry milk/Tris-buffered saline. After washing, bound antibodies were visualized using enhanced chemiluminescence reagents (Amersham Pharmacia Biotech, les Ulis, France). For immunoprecipitation, antisera were raised against the C-terminal decapeptide (amino acids 345–354) of Giα₃ and against the C-terminal decapeptide (amino acids 345–354) that is shared by both Giα₁ and Giα₂ as described previously (Gettys et al., 1994). The antisera were characterized with respect to titer, specificity, and cross-reactivity using lysates from bacteria transformed with cDNA for each of the G proteins. Each antiserum was desalted and purified as described previously (Raymond et al., 1993; Gettys et al., 1994).

Photolabeling of Membrane G Proteins.

The 4-azidoanilido-[α-³²P]GTP ([α-³²P]AA-GTP) was synthesized according to the method described by Offermanns et al., 1990 and 1991), except that [α-³²P]AA-GTP was purified using a thin-layer chromatography and finally resuspended in water at a concentration of 4 × 10⁹ cpm/ml. Photoaffinity labeling of G proteins was performed as described by Offermanns et al. (1991). Briefly, cell membranes (50 μg) were preincubated with α₂-AR agonists and/or antagonists for 10 min at 30°C in 50 μl of assay buffer containing 30 mM HEPES, pH 7.5, 100 mM NaCl, 100 mM EDTA, 1 mM benzamidine, 5 mM MgCl_2, 50 mM leupeptin, and 3 μM GDP. Then, 10 μl of [α-³²P]AA-GTP (0.4 × 10⁹ cpm/ml) diluted in distilled water was added to each sample and the labeling reaction was allowed to proceed for 10 min. The reaction was stopped by the addition of 100 μl of ice-cold assay buffer and immediate 4°C centrifugation at 12,000gfor 10 min. All subsequent procedures were performed at 4°C. Supernatants containing free [α-³²P]AA-GTP were removed. Membrane pellets were rapidly resuspended in 55 μl of assay buffer supplemented with 2 mM dithiothreitol and exposed, on ice, to UV light (254 nm, 15 W) in the dark for 4 min. Before immunoprecipitation, photolabeled membranes (50 μl) were then solubilized by incubation in presence of 0.25% SDS at 60°C for 5 min and addition of 50 μl of an immunoprecipitation buffer (0.5% SDS, 2% Nonidet P40, 1% cholate, 150 mM NaCl). Solubilized samples were first incubated 30 min at 4°C with prewashed Pansorbin cells (25 μl of a 10% solution in 50 mM sodium phosphate, pH 7.4) to further minimize the nonspecific antibody binding. After removing Pansorbin cells by centrifugation (700 g), each sample was divided in two aliquots and incubated in presence of anti-Giα₂ or -Giα₃ IgG (1:50) over night at 4°C under constant rotation. The immunocomplexes were collected by the addition of 25 μl of Pansorbin cell suspension and centrifugation at 700g. Then, pellets were washed two times in PBS and resuspended in 30 μl of 1.5% SDS and 30 μl of Laemmli buffer (Laemmli, 1970). Samples were boiled for 5 min before 10% SDS-PAGE analysis. After drying the gel, photolabeled proteins were visualized by autoradiography on Kodak X-Omat AR-5 films (Sigma-Aldrich). Incorporation of [α-³²P]AA-GTP into immunoprecipitated G proteins α subunits was quantified by densitometric analysis of autoradiograms with an Imstar computer-assisted image analyzer. Results are expressed as -fold incorporation of [α-³²P]AA-GTP into immunoprecipitated G protein α subunits compared with unstimulated control subunits.

Results

Establishment of the Experimental System.

DDT1-MF2 cells express useful common and distinct signaling entities in comparison with pregnant myometrium. Indeed, they display a similar density of β₂ adrenoceptors and Gs proteins (Hadcock et al., 1991; Vivat et al., 1992). They also express the same isoforms of pertussis toxin (PTX)-sensitive G proteins (Gi₂and Gi₃) that exert a similar tonic inhibition of adenylyl cyclase activity in an agonist-independent manner (Tanfin et al., 1991; Cohen-Tannoudji et al., 1995). However, none of α₂-AR subtypes (Philippe et al., 1989; Duzic and Lanier, 1992) nor AC isoforms type II and IV could be detected (Marjamaki et al., 1997). Thus we established, in this cell line, an experimental system expressing α_2A/D- or α_2B-AR subtype in presence of AC II using stable gene transfection method to further assess their functional characterization.

DDT1-MF2 cells were transfected with the cDNA encoding the human α_2B-AR or the rat α_2A/D-AR. After Scatchard analysis of saturation binding studies using [³H]rauwolscine, cell lines expressing ∼ 1.5 pmol of receptor/mg of membrane proteins were isolated. No specific binding of [³H]rauwolscine was seen in control DDT1-MF2 cell membranes. As shown Fig.1A, competition studies revealed that [³H]rauwolscine-specific binding was inhibited by subtype-selective compounds such as oxymetazoline (α_2A-specific), chlorpromazine and ARC 239 (α_2B- specific) with pK_i values characteristic of human α_2B-AR (Bylund et al., 1988), thus indicating that the transfected α_2B-AR receptors displayed the expected ligand recognition properties. The α₂-AR selective agonist clonidine inhibited the β₂-AR-stimulated-cAMP production in a dose-dependent manner (significant at concentrations as low as 1 nM, *p < 0.05) (Fig. 1B). This inhibitory effect was mediated through α_2B-AR, because it was prevented by the α₂-AR antagonist yohimbine and was not observed in cells transfected with vector alone. Incubation of the cells with PTX completely abolished the inhibition of stimulated cAMP accumulation elicited by the activation of the expressed α_2B-AR (data not shown). Maximal reduction of the isoproterenol response was 59% ± 6 (EC₅₀ value of ∼ 30 nM). These additional data indicated that, besides retaining its binding features, α_2B-AR expressed in the plasma membranes of DDT1-MF2 cells was functional and implicated in a negative cross talk with the β₂-AR/Gs cascade through PTX-sensitive G proteins.

DDT1-MF2 cells expressing α_2B-AR and α_2A/D-AR were further stably cotransfected with the cDNA encoding adenylyl cyclase II isoform. RNA screening using AC II cDNA-specific probe indicated that transcripts of expected 4.2-kilobase size were detected in selected hygromycin resistant clones but not in control cells transfected with vector alone (data not shown). These clonal cell lines coexpressing the α_2B-AR and adenylyl cyclase II were also assessed for functional evaluation of enzyme activity in response to a saturating dose (10 μM) of GTPγS in comparison with DDT1-MF2 cells expressing α_2B-AR only. With regard to DDT1-MF2-α_2B and -α_2A/Dtransfectants, stimulation with GTPγS increased adenylyl cyclase activity by ∼6 fold in both DDT1-MF2-α_2B-ACII and -α_2A/D-ACII cotransfectants [from 650 ± 27 to 3650 ± 460 and 570 ± 70 to 4100 ± 330 pmol cAMP/10 min/mg of protein, respectively (Fig.2)]. Similar observations have been made in previous experiments in DDT1-MF2 cells stably transfected with adenylyl cyclase II cDNA alone (Marjamaki et al., 1997). These results clearly indicated that the adenylyl cyclase II transcript expressed in DDT1-MF2-α_2B-ACII or -α_2A/D-ACII cotransfectants encoded for an enzyme exhibiting the expected functional properties.

Effect of α2-AR Activation on Cellular cAMP in DDT1-MF2 Cotransfectants.

Whereas epinephrine stimulation potentiated the β₂-activated cAMP production in DDT1-MF2-α_2B-ACII cotransfectants, clonidine decreased it (Fig. 3A). Indeed, epinephine enhanced β₂-stimulated cAMP production up to 52% ± 2 at 10 μM with an ED₅₀ value of 114 ± 33 nM. Conversely, clonidine produced a dose-dependent attenuation of cAMP accumulation over a concentration range of 1 nM to 0.1 mM. Maximal inhibition (−43%) was obtained at 1 μM (*p < 0.05). Half-maximal inhibition (ED₅₀) occurred at 10 nM clonidine. With higher concentrations of clonidine, negative input persisted, although it was reduced. The inhibitory influence of the clonidine-activated α_2B-AR did not seem to be caused by low receptor expression in α_2B-ACII cells. Indeed, in these cells, B_max was 1.3 ± 1 pmol of receptor/mg of membrane protein, a receptor density equivalent to the one measured in α_2A/D-ACII cells (1.2 ± 0.045 pmol/mg). Furthermore, we tested six α_2B-ACII clones ranging in receptor density from 1 to 3 pmol/mg of protein; none produced significant potentiation of isoproterenol-stimulated cAMP accumulation using clonidine (data not shown). Conversely, they all reduced cAMP-β₂-AR dependent generation. These data demonstrated that in DDT₁-MF₂ cells expressing type II AC isoform, the α_2B-AR could translate into opposite response depending on the type of agonist used (epinephrine or clonidine).

In contrast to the divergent agonist effects observed in α_2B-AC II cotransfectants, both clonidine and epinephrine increased isoproterenol-stimulated cAMP production in α_2A/D-AC II cotransfectants (Fig. 3B). The maximal stimulation of cAMP accumulation was produced at 0.1 μM clonidine (60 ± 1%) and 1 μM epinephrine (115 ± 41%). ED₅₀ values were 9 ± 1.3 and 180 nM ± 48 nM, respectively.

In both cotransfectants (α_2B-ACII and α_2A/D-ACII), clonidine as well as epinephrine effects were blocked by yohimbine (Fig. 3, A and B) and prior treatment of cells with PTX (Fig. 4). These data were consistent with the fact that these two agonists potentiated or inhibited cAMP production acting on α₂-ARs through Gi/o family members endogenously expressed in DDT1-MF2. However, epinephrine also produced a small PTX-insensitive potentiation of cAMP levels when acting through α_2B-AR. One possible interpretation of this result is that α_2B-AR, when present in high density in the membrane, may also cross-react, to a low extent, with endogenous Gs proteins, as reported previously in other cell lines (Eason et al., 1992, 1994; Pepperl and Regan, 1993).

Altogether, these results indicated that, in the presence of transfected AC II, α_2B-AR was able to mediate opposite regulatory effects (positive input versus negative input) on Gs-stimulated cAMP production via PTX-sensitive G proteins. To gain insight on how α_2B-AR could switch from a positive to a negative regulation, we compared Gi protein coupling of epinephrine- and clonidine-activated α₂-AR subtypes.

Selective Recruitment of Gi Proteins by α_2B- or α_2A/D -AR in Response to Clonidine or Epinephrine.

Within the family of PTX-sensitive G proteins, DDT₁-MF₂ cells and myometrium express Gi₂ and Gi₃ (Fig. 5) (Hadcock et al., 1991; Tanfin et al., 1991, Cohen-Tannoudji et al., 1995). Thus, agonist-specific adenylyl cyclase II response could be the consequence of a differential α₂-AR subtype specific recruitment of Gi₂ and/or Gi₃. So, we questioned whether the differences observed in the receptor coupling to AC for clonidine and epinephrine in the cell models reflected specific coupling to Gi₂ or Gi₃. This issue was addressed by incubation of membranes with the photoreactive GTP analog, 4-azido-anilido-[α−³²P]GTP ([α-³²P]AA-GTP) in the presence of ligand, followed by cross-linking, solubilization, and selective immunoprecipitation of Gi₂ or Gi₃.

As shown in Fig. 6A, clonidine induced a dose-dependent labeling of Giα₂ protein exclusively. No significant incorporation of [α-³²P]AA-GTP was detected in Giα₃ protein. At 1 μM clonidine, maximal labeling of Gαi₂ with [α-P³²]−AA-GTP (∼ 2.5-fold compared with control fraction) was completely inhibited with yohimbine, thus indicating that recruitment of Giα₂ protein was strictly dependent upon α_2B-AR activation. It should be noted that 1 μM clonidine elicited maximal inhibition of Gs-stimulated cAMP production (Fig. 3A). In marked contrast, when experiments of similar design were conducted with epinephrine, both Gαi proteins (Giα₂ and Giα₃) were photolabeled (Fig. 6B). Maximal incorporation of [α-³²P]AA-GTP was obtained at 1 μM epinephrine for each endogenous Gi protein (∼2.6-fold compared with control fraction). This epinephrine-dependent [α-³²P]GTP azidoanilide labeling resulted from α_2B-AR activation, because it could be completely blocked by yohimbine. On membranes obtained from DDT1-MF2-α_2A/D-ACII cotransfectants where AC II potentiation also occurred, clonidine induced activation of both types of Gi proteins [∼2.6- and 2-fold, respectively, for Gαi₂ and Gαi₃ proteins compared with unstimulated fraction at 1 μM (Fig.7)].

Altogether, these data indicated that Gi₃activation is required for potentiation of β₂-AR stimulation of AC II by α_2A/D- or α_2B-AR in DDT1-MF2 cells. Furthermore, they suggested that heterotrimeric Gi2 and Gi3 proteins may have specific roles in modulating stimulated AC II activity in a given cell type.

Discussion

Activation of α₂-AR subtypes induces multiple cellular effects, including inhibition of adenylyl cyclase or, in some physiological models or cell systems, an increase of cAMP levels. The mechanisms responsible for inhibitory or stimulatory input on adenylyl cyclase activity/cAMP production greatly depend on their interaction with PTX-sensitive G inhibitory proteins and, additionally, also reflect the type of adenylyl cyclases expressed in various cells.

The present work was motivated by the observation that, in the midpregnancy and term myometrium, the cross talk between activated α₂- and β₂-ARs differently affect the degree of intracellular cAMP generation (Mhaouty et al., 1995) and, consequently, the relaxed or contractile state of the uterus (Do Khac et al., 1986). Molecular events that underlie these subtle changes in sensitivity of the smooth muscle to catecholamines may result from 1) the alteration of the α_2A/D-/α_2B-subtypes expression pattern (Bouet-Alard et al., 1997); 2) the drastic changes of Gi₂/Gi₃ protein ratio (Tanfin et al., 1991; Cohen-Tannoudji et al., 1995); and/or 3) the functional properties of the pregnant myometrium adenylyl cyclase population (Mhaouty-Kodja et al., 1997; Suzuki et al., 1997). As an initial approach, we investigated whether α_2A/D- and α_2B-AR subtypes could exert different regulatory roles on β₂-AR catalyzed cAMP production. To address this issue, DDT1-MF2 cells provided an interesting context, because they endogenously express some of the molecular entities (β₂-AR, Gi₂, Gi₃, and G_s proteins) involved in myometrium α₂-/β₂-AR cross talk but lack α₂-AR subtypes and AC type II isoform. Thus, we separately expressed α_2B- or α_2A/D-AR subtype in this cell line and, further on, cotransfected each clone with the adenylyl cyclase II isoform that potentiated cAMP production in response to α₂-AR agonists in pregnant myometrium. Selected clonal cell lines with comparable functional pools of ARs and adenylyl cyclase have provided useful test models for a careful examination on how α₂-AR differ in their ability to modulate adenylyl cyclase and to couple to endogenous Gi protein.

In this context, we found that, without any ACII expression, clonidine induced an inhibition of β₂-dependent cAMP production in both α_2A- and α_2B-AR transfectants (at 1 μM clonidine α_2A/D- and α_2B-ARtransfectants: 34 ± 3% for and 62 ± 2% rspectively). This result is consistent with those reported by Duzic and Lanier (1992) in the same cell line, demonstrating that α_2B- and α_2A/D-AR activation similarly inhibits forskolin-induced increase in intracellular cAMP. When AC II was coexpressed in DDT1-MF2-α_2A/D transfectants, epinephrine as well as clonidine were able to switch the inhibitory signal into a stimulatory input through PTX-sensitive G proteins. In DDT1-MF2-α_2B-AC II cotransfectants, despite a similar functional pool of AC II and an equivalent density of receptor, clonidine was unable to trigger such a switch, in contrast with epinephrine. Direct measurement of G protein activation by photoaffinity labeling with [α-³²P]AA-GTP followed by selective separation of individual G protein α subunits revealed that clonidine, acting on α_2B-AR, mediated an exclusive coupling to Gi₂, whereas the full agonist, epinephrine, led to the recruitment of both PTX-sensitive G proteins, Gi₂ and Gi₃. Thus, using this photoaffinity probe, it seems clear that, in DDT1-MF2 cells overexpressing AC II, the ability of α_2B-AR to switch from a negative to a positive input to Gs-stimulated cAMP production greatly depends on the recruitment of Gi_3.

Previous works have established that the potentiation of Gs-stimulated cAMP production is caused by the input of Gβγ released from Gi to AC II that synergizes with Gs to further elevate cAMP levels (Tang and Gilman, 1991; Federman et al., 1992). Nevertheless, this phenomenon can only occur from a threshold concentration of released Giβγ (Tang and Gilman, 1991). Thus, the persistent inhibitory effect observed when Gi₂ is activated alone could be explained by the following hypothesis: the total amount of Gi₂βγ released upon receptor activation would be insufficient to overcome inhibitory influences exerted by Giα₂ on endogenous AC. On the contrary, when both Gi proteins (Gi₂ and Gi₃) were recruited, the threshold concentration would be reached, thus allowing AC II potentiation. Here, we should note that the possibility of overcoming Gαi₂inhibition was probably reinforced by the low ability of Giα₃ to induce AC inhibition (Raymond et al., 1993; Gettys et al., 1994). This might also reflect the fact that Gi₂βγ dimers released upon α_2B-AR activation poorly interact or activate AC II compared with Gi₃ βγ. Although there is no direct evidence that Gi₂βγ and Gi₃βγ differ in their capacity to potentiate Gs-stimulated AC, such a hypothesis must be taken into account. Indeed, several studies report that the regulation of βγ-sensitive effectors depends on the composition of the Gβγ dimers with which they are interacting (Müller et al., 1997; Bayewitch et al., 1998; Maier et al., 2000). Finally, because post-translational modifications are considered important criteria in determining the potency by which Gβγ complexes modulate effectors (Ford et al., 1998), differential modifications of Gi₂β and/or -γ versus Gi₃β and/or -γ might also contribute to the clonidine-specific effect.

The present work also brings some evidence as to whether changes in the ratio of Gi₂ to Gi₃proteins in late pregnant myometrium may play a crucial role for the switch in the stimulatory versus inhibitory input to AC population from the α₂-AR/Gi protein signaling. The down-regulation of Gi₃ protein, together with Gi₂-increased expression (Cohen-Tannoudji et al., 1995), could prevent ACII potentiation, thus allowing the decrease of β₂-AR stimulated cAMP production at term. The inhibition of the synthesis of smooth muscle relaxation factor (cAMP) together with the increase of intracellular Ca²⁺would promote myometrial contractions at term.

In summary, these data provide the molecular basis of clonidine partial agonist effect when acting through α_2B-AR, because they reveal that this compound selectively uncouples the receptor from one of the normally targeted G proteins: Gi₃. From this finding, it can be predicted that, in the situation where clonidine behaves as the full agonist, the second messenger pathway would be exclusively regulated by Gi₂ in Gi₂/Gi₃ expressing cells. On the other hand, in systems in which clonidine acts as partial agonist or even as an antagonist, Gi₃ would play a determinant or exclusive role. Furthermore, they suggested that Gi₂ and Gi₃ have specific roles in modulating AC II effector through α and/or βγ subunits. Finally, they shed light upon a possible molecular mechanism that might allow the versatility of signal routed through myometrial α₂-AR during pregnancy involved changes in Gi₂/Gi₃ ratio.