Abstract
The stable interaction of a G-protein coupled receptor and a particular partner G-protein was made possible by creating tandems between the α2A adrenergic receptor (α2A-R) and pertussis toxin-resistant mutants of different Gα subunits of heterotrimeric G-proteins. Both α2A-R-Gαoand α2A-R-Gαi proved able to reconstitute agonist-induced voltage-dependent inhibition of N-type calcium channels (CaV2.2) similar to the wild-type α2A-R when expressed in COS-7 cells. The interaction of Gq with the Gi/o signaling pathways was studied by expressing either Gαq or a chimeric construct based on Gαqcontaining the last five amino acids of Gαz, which is activated by α2A-R. It was found that Gαqz5activated by the wild-type α2A-R inhibited CaV2.2 currents in a voltage-independent fashion. Furthermore, Gαqz5 counteracted the voltage-dependent inhibition resulting from α2A-R-Gαoactivation. We subsequently investigated the basis for the behavior of Gαqz5. Our evidence suggests that this occurs as a result of a downstream effect of activation of Gαqz5 because it was blocked by C-terminal construct of phospholipase Cβ1. Furthermore it is likely to occur in part via protein kinase C (PKC) activation, because the PKC activator phorbol dibutyrate mimicked the effects of Gαqz5 in α2A-R-Gαo-transfected cells. Conversely, cells expressing both α2A-R-Gαo and Gαqz5 exhibited a partial restoration of voltage-dependent inhibition in the presence of the PKC inhibitor bisindolylmaleimide I (GF 109203X). The potential sites of phosphorylation are discussed.
Calcium influx in any cell requires fine tuning to guarantee the correct balance between activation of calcium-dependent processes, such as muscle contraction and neurotransmitter release, and calcium-induced cell damage. G-protein-coupled receptors (GPCRs) play a role in negative feedback of the activity of voltage-dependent calcium channels (Dolphin, 1995). Establishing the basis for the specificity of the relationships between membrane receptors, G-proteins, and effectors has proven elusive, in part because of the promiscuity of the partners involved when expressed in heterologous systems. When different G-protein subunits are over-expressed together with GPCRs and calcium channels, the degree of specificity is rather low. For example, the α2A-adrenergic receptor (α2A-R) couples to all members of the Gi/o family, including the pertussis toxin (PTX)-sensitive Go and Gi, and the PTX-insensitive Gz (for review, seeHille, 1994).
In native systems, however, receptors display a more selective activation of endogenous G-proteins subtypes, with Go being more important than Gi in the inhibition of calcium currents in sensory neurons (Campbell et al., 1993). Furthermore, in sympathetic neurons, muscarinic activation of G-protein-activated inward-rectifier (GIRK) channels is mediated by Gi, whereas muscarinic inhibition of N-type calcium channels is mediated by GoA (Fernández-Fernández et al., 2001). These results point to the importance of the cellular localization of each receptor and G-protein subtype.
For GPCRs that associate with PTX-sensitive G-proteins, production of Gβγ dimers seems to be responsible for the direct voltage-dependent inhibition of N- and P/Q-type channels (Herlitze et al., 1996; Stephens et al., 1998), although it has also been proposed that in chick sensory neurons, Gβγ results in activation of PKC, to mediate the voltage-dependent inhibition caused by norepinephrine (Diversé-Pierluissi et al., 1995). Furthermore, Gα subunits have also been implicated in mediating G-protein modulation (Diversé-Pierluissi et al., 1995).
One way to identify the direct effects of a specific G-protein on calcium channel activity is to link the G-protein α subunit to the receptor of choice to form a tandem construct. One of the advantages of this approach is the elimination of one of the signal amplification steps, occurring at the receptor/G-protein interaction level, because the two components are constrained to work with a 1:1 stoichiometry. Furthermore, there is increasing evidence against the established model which sees G-proteins shuttling between receptor and effector, and toward a view that there is a close localization of signal transduction elements in distinct membrane domains (Seifert et al., 1999). We used fusion proteins between the α2A-R and either Gαi1 or Gαo1, both of which were rendered PTX-insensitive by means of a point mutation at residue 351 (Bahia et al., 1998). The Ile351 Gα mutants were chosen over other possible PTX-resistant mutants because they resulted in the strongest activation by α2A-R (Bahia et al., 1998). Activation of these tandems by the α2A-R agonist clonidine was studied in COS-7 cells coexpressing N-type channels (CaV2.2) and comparing the response to that produced by the activation of the wild-type α2A-R. These tandems have been found able to interact with endogenous G-proteins to a certain extent (Burt et al., 1998). In the present study, treatment of cells with PTX before recording allowed the receptor/G-protein tandems to be studied in isolation, effectively removing the contribution of endogenous Gi/o proteins.
The carboxyl terminus of the Gα subunit is not only a determinant of its sensitivity to PTX-dependent ADP-ribosylation but is also essential to confer specificity of coupling to GPCRs (Conklin et al., 1993). To examine whether Gβγ dimers liberated from Gq could also inhibit N-type Ca2+ channels, we exploited a chimeric Gαq-protein. This construct was formed by a Gαq subunit in which the last 5 amino acids were substituted for the corresponding amino acids from Gαz. The resulting Gαqz5, unlike Gq itself, is able both to couple to the α2A-R and to activate effectors specific to the Gq family, such as phospholipase C and the downstream protein kinase C (PKC) (Conklin et al., 1996). We report the effects of such a construct in isolation and when coexpressed with the α2A-R-Gαo fusion protein and compare these effects with those of the wild type Gq subunit. The involvement of downstream effectors of Gαqz5 is also examined.
Materials and Methods
Constructs.
COS-7 cells were transiently transfected with the following cDNAs: rabbit CaV2.2 (GenBank accession no. D14157); rat β1b (GenBank accession no. X11394); and mut-3 green fluorescent protein (GFP).
The PTX-resistant α2A-R-G-protein fusion proteins used throughout this study were prepared as described previously (Cavalli et al., 2000). In brief, Cys351 of rat Gαi1 and Gαo1 was mutated to Ile by site-directed mutagenesis and then used to create the α2A-R-Gα fusion proteins using porcine α2A-R in pcDNA3. The Ile19Ala, Glu20Ala (IE) mutant of Gαo1 was constructed, based on studies of an equivalent mutation (Ile25Ala, Glu26Ala) of Gαq (Evanko et al., 2000), and this was then incorporated into the PTX-resistant α2A-R-Gαo fusion protein. The wild-type Gαq subunit (Gαq w.t.) and the Gαqz5 subunits described previously (Conklin et al., 1993) were subcloned into pMT2. Gα-transducin (Gαt) was in pcDNA3. The pEGFP-PLC-β1ct fusion construct of the C terminus of phospholipase Cβ (PLC-β1ct) was described previously (Kammermeier and Ikeda, 1999).
Cell Culture and Transfections.
Cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% newborn calf serum, penicillin (100 IU/ml) and streptomycin (100 μg/ml) (all from Invitrogen, Paisley, UK) at 37°C, 5% CO2, and passaged every 3 to 4 days. For transient transfections of the different constructs, a cDNA mixture was made containing the voltage-dependent calcium channel CaV2.2 subunit cDNA in a ratio of 3:1 with all the other constructs, β1b, α2A-R, α2A-R-G-protein tandems, and/or the Gα subunits. Mut-3 GFP cDNA was also included at a ratio of 0.2. For transfection, 10 μl of GenePORTER reagent (Qbiogene, Harefield, UK) and 2 μl of cDNA mixture were preincubated in 1 ml of Dulbecco's modified Eagle's medium at 20°C for 1 h before addition to 35-mm Petri dishes containing approximately 2 × 106 cells. Cells were cultured at 37°C for 72 h, replated using a nonenzymatic cell dissociation medium (Sigma, Poole, UK), and maintained at 27°C for 1 to 8 h, before recording. PTX (Sigma) was used to inactivate the endogenous Gαi/o subunits by adding it to the culture medium at a concentration of 40 to 100 ng/ml for 16 h before replating the cells.
[3H]RS-79948-197 Binding.
To determine the levels of expression of the various α2A-R-G-protein fusion proteins, the specific binding of [3H]RS-79948-197 was measured as described previously (Ward and Milligan, 2002).
[35S]GTPγS Binding.
[35S]GTPγS binding experiments were performed essentially as described for receptor-G-protein tandems incorporating Gα11 (Carrillo et al., 2002). These were initiated by the addition of membranes containing 50 fmol of the fusion constructs to an assay buffer [20 mM HEPES, pH 7.4, 3 mM MgCl2, 100 mM NaCl, 1 μM guanosine 5′-diphosphate, 0.2 mM ascorbic acid, and 50 nCi of [35S]GTPγS] in the absence or presence of clonidine (10 μM). Nonspecific binding was determined in the same conditions but in the presence of 100 μM GTPγS. Reactions were incubated for 15 min at 30°C and were terminated by the addition of 0.5 ml of ice-cold buffer containing 20 mM HEPES, pH 7.4, 3 mM MgCl2, and 100 mM NaCl. The samples were centrifuged at 16,000g for 15 min at 4°C, and the resulting pellets were resuspended in solubilization buffer (100 mM Tris, 200 mM NaCl, 1 mM EDTA, and 1.25% Nonidet P-40) plus 0.2% SDS. Because all the α2A-R-G-protein tandems used in these studies incorporated a hemagglutinin (HA) epitope tag at the N terminus of the receptor, samples were precleared with Pansorbin (Calbiochem, Nottingham, UK), followed by immunoprecipitation with the anti-HA antiserum 12CA5 (Roche Diagnostics, Lewes, UK). Finally, the immunocomplexes were washed twice with solubilization buffer, and bound [35S]GTPγS was measured by liquid scintillation counting.
Immunoprecipitation and Immunodetection Studies.
To analyze the interaction of α2A-R-Gαo with Gβγ dimers, cells were transfected with α2A-R-Gαo or α2A-R-Ile19Ala, Glu20Ala Gαo in the absence or presence of plasmids encoding G-protein β1 and γ2 subunits. Cells were washed once with ice-cold phosphate-buffered saline and immediately homogenized in a lysis medium containing 50 mM HEPES, pH 7.4, 10 mM Na4P2O7, 100 mM NaF, 10 mM EDTA, 0.1 mM Na3VO4, 1% Triton X-100, and a protease inhibitor cocktail (Complete; Roche). Cell lysates were centrifuged (15 min, 13,000 rpm) and the supernatants precleared for 1 h with nonspecific serum and protein A. Next, samples were incubated overnight with a polyclonal antiserum directed against the C-terminal decapeptide of Gαo1 (Mullaney and Milligan, 1990). The immunocomplexes were then captured with protein A-agarose.
For immmunoblotting, cell lysates or immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis (PAGE). Proteins were transferred to polyvinylidene fluoride (PVDF) membranes and blocked for 2 h with 5% nonfat dried milk in 0.05% Tween 20/Tris-buffered saline (TTBS). Then, the PVDF membranes were probed overnight at 4°C with an antiserum (BN) directed against the N-terminal decapeptide of the G-protein β1 subunit (Green et al., 1990) and washed with TTBS. The PVDF membranes were incubated for 20 min with horseradish peroxidase conjugated to anti-rabbit IgG (1:20,000) (Amersham Biosciences). Finally, they were washed with TTBS and developed by enhanced chemiluminescence.
Electrophysiology.
Fluorescent COS-7 cells expressing GFP were chosen for whole-cell, patch-clamp recording. Borosilicate glass electrodes were used with a resistance of 2 to 5 MΩ when filled with a solution containing 140 mM cesium aspartate, 5 mM EGTA, 2 mM MgCl2, 0.1 mM CaCl2, 2 mM K2ATP, and 20 mM HEPES, pH adjusted to 7.2 with CsOH, 310 mOsM with sucrose. Cells were perfused with an extracellular solution containing 160 mM tetraethylammonium-Br, 2 mM KCl, 1.0 NaHCO3, 1.0 MgCl2, 10 mM HEPES, 4 mM glucose, and 10 mM BaCl2, pH 7.4, 320 mOsM with sucrose. Barium currents were recorded using an Axopatch-1D amplifier (Axon Instruments, Union City, CA). Data were filtered at 2 kHz, digitized at 5 to 10 kHz, and analyzed using pCLAMP 6 (Axon Instruments) and Origin 5.0 (Microcal, Northampton, MA). Cell capacitance compensation and series resistance compensation between 65 and 80% were applied electronically. Records are shown after leak subtraction (P/4 or P/8 protocol).
Facilitation was assessed by using a double-pulse protocol (see Fig.1a, top). A first 30-ms step (P1) usually to 0 mV was followed by a 300-ms period of repolarization to −100 mV. A strongly depolarizing prepulse PP of 30 to +100 mV was then delivered before a second pulse (P2) to the same voltage as the first test pulse, to assess the voltage-dependence of current inhibition. The PP and the second pulse were separated by a 10-ms repolarization time to −100 mV. Pulses were delivered every 15 s. Currents were measured 10 ms after the onset of both P1 and P2 and the average over a 2-ms period was calculated and used for subsequent analysis. The 300-ms interval between P1 and PP was sufficient to minimize the voltage-dependent calcium channel inactivation caused by P1. The duration and amplitude of the PP were chosen to produce maximal facilitation in the conditions used (data not shown). Experiments were performed at room temperature (20–24°C). Drugs were applied by the use of a gravity-fed, electronically controlled, multibarrelled perfusion system. Current density-voltage (I-V) relationships were fitted with a modified Boltzmann equation as follows: I =Gmax (V −Vrev)/(1 + exp(−(V −V50,act)/k)), whereI is the current density (picoamperes per picofarad),Gmax is the maximal conductance (nanosiemens per picofarad), Vrev is the reversal potential, V50,act is the mid-point voltage for current activation, and k is the slope factor.
The time constant of activation (τact) was calculated by fitting a single exponential to the current traces:I = A × exp(−t/τact) + C, whereA is the amplitude of the component with time constant τ, and C is a constant. Data are expressed as mean ± S.E.M., and statistical significance between conditions was examined using Student's t test or paired t test, as appropriate.
Materials.
[3H]RS-79948-197 (90 Ci/mmol) was from Amersham Biosciences (Little Chalfont, Buckinghamshire, UK), [35S]GTPγS (1250 Ci/mmol) was from PerkinElmer Biosciences (Warrington, UK). Clonidine hydrochloride (Calbiochem) was prepared as a 10−2 M stock in H2O. The protein kinase C activator phorbol-12,13-dibutyrate (PDBu; Calbiochem) and the PKC inhibitor bisindolylmaleimide I (GF 109203X, Calbiochem) were prepared as 10−2 M stock in DMSO. All drugs were diluted in the experimental solutions to the final concentrations indicated.
Results
Effect of the α2A Adrenergic Receptor-Gαi and -Gαo Tandems.
We first expressed either α2A-R w.t. or the PTX-insensitive receptor-Gα tandems α2A-R-Gαo1C351I (α2A-R-Gαo) or α2A-R-GαiC351I (α2A-R-Gαi) together with the CaV2.2 calcium channel. The inhibition of the expressed Ba2+ currents (IBa) by activation of the α2A-R w.t. was compared with the effect of the receptor G-protein tandems (Fig. 1). Overall, the α2A-R agonist clonidine (10 μM) inhibited N-type IBa via activation of both the free α2A-R and the tandem α2A-R-Gα constructs, as exemplified by the current traces in Fig. 1, a−c. The inhibition was rapid (< 15 s) and reversible upon washing (data not shown). The extent of IBa inhibition at 0 mV is given in Fig. 1e (■). In the absence of PTX, IBa was similarly reduced by both the wild-type α2A-R (64.2 ± 6.6%, n = 9, Fig. 1a) and the tandems α2A-R-Gαo (77.6 ± 6.6%, n = 5, Fig. 1b) and α2A-R-Gαi (64.1 ± 4.0%, n = 8, Fig. 1c). Thus, removal of the amplification step between receptor and G-protein did not affect the ability of Gi/o to produce inhibition of CaV2.2 IBa.
It has been observed previously that chimeric receptor-Gα constructs are able to activate not only the tethered Gα subunit but also endogenous subunits of the Gi/o family (Burt et al., 1998). The use of PTX therefore allows isolation of the effects of exogenous Gα subunits mutated to be PTX-resistant by rendering the endogenous Gi/o subunits unable to couple to the receptor. Preincubation of the cells with PTX greatly reduced the inhibition produced by the α2A-R w.t. (see traces in Fig. 1d and mean results in Fig. 1e). Conversely, PTX did not significantly affect the functioning of the two PTX-insensitive receptor G-protein tandems. The calcium channel currents at 0 mV were still reduced by 74.1 ± 6.5% (n = 18, Fig. 1b) and 62.9 ± 9.1% (n = 8, Fig. 1c) with the Gαo and the Gαi fusion proteins, respectively, after pretreatment with the toxin (Fig. 1e). Experiments repeated with a lower concentration of clonidine (100 nM) gave comparable results in terms of degree of inhibition, demonstrating that maximal receptor activation was achieved at the concentration of agonist used (data not shown).
Inhibition of N-type currents by the receptor-Gαi/o tandems was largely voltage-dependent, as seen by using a double pulse voltage-clamp protocol (Fig. 1, a—-d). The PP was able to reverse the agonist-induced inhibition induced by either α2A-R w.t. (Fig. 1a) or the α2A-R-Gαo (Fig. 1b) and α2A-R-Gαi (Fig. 1c) tandems, whereas incubation with PTX eliminated the voltage-dependent effects of the α2A-R w.t. (Fig. 1d). The amount of inhibition by clonidine before and after the PP is summarized in Fig. 1e. The resultant “facilitation” (determined as the P2 current amplitude divided by P1 current amplitude) was substantial for all three receptor constructs. In all cases, however, removal of inhibition during P2 by the PP to +100 mV was never complete, indicating a voltage-independent inhibitory component.
As a corollary of the voltage-dependence of the inhibition of IBa by clonidine, it should also be abolished at large step potentials. The voltage-clamp protocol used to examine this was similar to that shown in Fig. 1 with the exception that both test pulses (P1 and P2) were varied from −40 to +70 mV in 10-mV increments. Example traces are shown in Fig. 2a, whereas the mean I-V plots for values measured in P1, before and during application of clonidine, for cells expressing the α2A-R-Gαo(n = 8) are shown in Fig. 2b. With all receptor constructs, the agonist caused both a reduction in IBa and a depolarizing shift in the I-V relationship. The V50,act during P1 was significantly depolarized for cells expressing α2A-R-Gαo, from −6.4 ± 3.1 to +9.0 ± 4.0 mV (p < 0.05,n = 6, Fig. 2b) and, for cells expressing α2A-R-Gαi, from +2.5 ± 2.4 to +9.7 ± 0.7 mV (p < 0.05,n = 6). No significant differences in theVrev or in theGmax were detected (Fig. 2b; data not shown). The P2/P1 facilitation ratios for the different test potentials are reported in Fig. 2, c–d. The PP revealed some tonic facilitation in the absence of the agonist (■), which was more marked when expressing the α2A-R w.t., where P2/P1 was 2.3 ± 0.3 at 0 mV (Fig. 2c). Clonidine enhanced the voltage-dependent facilitation, although the effects were much greater for the α2A-R-Gαotandem than for the α2A-R w.t. (Fig. 2d). Maximal facilitation was obtained at −10 or 0 mV and it was absent above +20 mV.
Not only did activation of the α2A-R-Gα tandems cause a reduction in current amplitude but the activation phase of the current was typically slowed during P1; this effect was reversed by the PP (e.g., Fig. 1, a–c). For example, for those cells transfected with the α2A-R-Gαo tandem, the τact at 0 mV during P1 was 3.7 ± 0.5 ms in control and 6.1 ± 1.1 ms during clonidine application (n = 10, p < 0.05, see Fig. 1b). This slowed activation was reversed by Gαt, which acts as a Gβγ sink to sequester free Gβγ subunits but does not couple to the α2A-R. Example traces are shown in Fig. 3a (top). After cotransfection of Gαt with α2A-R-Gαo, there was no longer a difference in the τact values measured in control and clonidine during P1 (2.9 ± 0.6 ms and 3.4 ± 0.5 ms, respectively, n = 9). Along with this effect, Gαt was able significantly to reduce inhibition by clonidine at 0 mV from 74.1 ± 6.5 to 43.0 ± 6.7% (p < 0.001; Fig. 3b) and to reduce the P2/P1 facilitation ratio in the presence of clonidine to 1.54 ± 0.24 at 0 mV, although this was still significantly greater than the P2/P1 ratio under control conditions (Fig. 3c).
Given that these data were obtained in the presence of PTX, to prevent promiscuous coupling of the tandems to additional endogenous Gi/o proteins, these findings indicate that the α2A-R tandems are able to reconstitute inhibitory effects on CaV2.2 calcium channel currents by means of the tethered Gαi/o that are almost identical to the wild-type receptor coupling to endogenous G-proteins and that such effects are very likely to be mediated purely by Gβγ dimers. It has been found previously that mutation of both Ile25 and Glu26 of Gqα to Ala severely limits interaction with the Gβγ complex (Evanko et al., 2000). These residues are highly conserved in other G-protein α subunits. We thus constructed a form of the PTX-resistant α2A-R-Gαo tandem (IE) that also incorporated the equivalent mutations of Ile19Ala and Glu20Ala in Gαo. Application of clonidine to cells expressing the IE form of the α2A-R-Gαo tandem produced no inhibition of IBa, and no effect on facilitation (Fig. 3, a, bottom, and b–c). It is also evident that these CaV2.2 currents show some tonic modulation, being slowly activating and facilitated by a prepulse, although this is no greater than for the free α2A-R (Fig. 2c).
To examine the binding of Gβγ to the IE mutant of α2A-R-Gαo, either α2A-R-Gαo or the IE form of this construct was cotransfected together with plasmids encoding the Gβ1 and Gγ2 subunits. Cell lysates were subsequently immunoprecipitated with an antiserum (OC) that identifies the C-terminal decapeptide of Gαo1. Such samples were then resolved by SDS-PAGE, transferred to PVDF membranes, and immunoblotted with an antiserum (BN) that identifies the N-terminal decapeptide of Gβ1. Although the α2A-R-Gαo tandem allowed coimmunoprecipitation of β1 subunit (Fig. 3d, lane 2), this was not observed for the IE form of the tandem receptor (Fig. 3d, lane 3).
Investigation of α2A-R-Gαq and α2A-R-Gαqz5 Chimeras.
Because the expression of the receptor/G-protein tandems indicated that the release of activated Gα subunits, Gαi and Gαo, does not play any direct role in G-protein-effector coupling for calcium channel inhibition, we were interested in studying whether Gβγ released from another class of G-protein, Gq, could also participate in the inhibitory process. However Gq is known not to couple efficiently to the α2A-R (Dorn et al., 1997). To use the same receptor for activation of both Gi/o and Gq pathways, we therefore employed the chimeric construct Gαqz5. This subunit conserved the main structure of Gαq but the last five amino acids were substituted for those of Gαz, a PTX-resistant member of the Gi/o family that does couple to the α2A-R (Conklin et al., 1993). Tandem α2A-R-Gαq and α2A-R-Gαqz5 constructs were assembled, and their functionality was assessed biochemically.
Evidence of the activation of the PTX-resistant G-proteins within the α2A-R-Gα tandems by clonidine was obtained by monitoring agonist-induced binding of [35S]GTPγS. Expression levels of the α2A-R-containing fusion proteins in membranes of PTX-treated cells were quantified by saturation ligand binding studies employing the high-affinity α2-adrenoceptor antagonist [3H]RS-79948-197. [35S]GTPγS binding studies were performed in the presence and absence of clonidine (10 μM) on membrane fractions expressing equal amounts of the various fusion proteins. After this, the anti-HA antibody 12CA5 was used to immunoprecipitate the samples, because all of these constructs contained an N-terminal HA epitope tag. Significant levels of [35S]GTPγS binding were observed for both the Gαo- and Gαi-containing fusion proteins; this was stimulated markedly by the presence of clonidine (Fig.4). In contrast, little binding of [35S]GTPγS was observed to the α2A-R-Gαq and α2A-R-Gαqz5 constructs, even in the presence of clonidine, consistent with a lack of activation of these G-proteins by the associated α2A-R. The inability of clonidine to promote binding of [35S]GTPγS to the fusion proteins containing Gαq does not reflect the well appreciated difficulty in monitoring nucleotide exchange for such G-proteins in standard [35S]GTPγS binding assays. We have recently shown that combination of use of receptor-G-protein tandems and selective immunoprecipitation allows a 30-fold stimulation of binding in the presence of agonist when such G-proteins are linked in tandem with appropriate receptors (Carrillo et al., 2002). A preliminary investigation also failed to show any clonidine-mediated inhibition of CaV2.2 via the α2A-R-Gαqz5 tandem, but because these receptor-Gαq tandems were nonfunctional biochemically, their coupling to CaV2.2 was not further examined.
We therefore employed free Gαq and Gαqz5 to examine whether Gβγ released from Gq or Gqz5 can signal to N-type calcium channels (Fig. 5a). We confirmed, by coexpressing the α2A-R w.t. with Gαq w.t. in cells treated with PTX, that Gαq did not couple directly to the α2A-R. Perfusion of clonidine induced only 7.1 ± 1.1% reduction in the current (n = 5, Fig.5b). In contrast, expression of Gαqz5 with the α2A-R w.t. resulted in significantly greater inhibition of CaV2.2 currents by clonidine (35.8 ± 8.6%, n = 9, Fig. 5, a and b). Surprisingly however, this was not removed by a PP to +100 mV, the inhibition in P2 being 34.5 ± 5.4% (n = 9, Fig.5b). Thus, the inhibition elicited by Gαqz5 was much greater than that elicited by Gαq w.t. (p < 0.001) but was not voltage-dependent. The P2/P1 facilitation ratio in the presence of Gαqz5 was around unity and was unaffected by the presence of agonist (0.98 ± 0.07 in control, 1.20 ± 0.23 in clonidine, p> 0.05, Fig. 5c). Current traces in the presence of Gαqz5 showed no evidence of slowing of the kinetics of activation in response to clonidine (e.g., traces in Fig.5a and data not shown). To determine whether voltage-dependent inhibition was completely absent for Gαqz5, we also examined the voltage-dependence of inhibition over a range of potentials. However, no obvious facilitation was evident at any test potential (data not shown). These results demonstrate that the C-terminal modification of Gq allowed Gαqz5 to couple to the α2A-R, causing a reduction in IBa, although the inhibition was voltage-independent and smaller than that elicited by the tandems α2A-R-Gαo and α2A-R-Gαi or the wild type α2A-R coupling to endogenous G-proteins.
Interaction between α2A-R-Gαo and Gαqz5.
It has been observed previously that chimeric receptor-Gα constructs are able to activate not only the tethered Gα subunit but also endogenous subunits of the Gi/o family (Burt et al., 1998). Accordingly, Gαqz5 might be expected also to interact with, and to be activated by, the α2A-R-Gαo tandem used in this part of the study. We investigated this potential interaction by coexpressing the tandem α2A-R-Gαo with the Gαqz5 subunit, and treating all cells with PTX.
The first observation was that the inhibition of IBa obtained when coexpressing α2A-R-Gαo with Gαqz5 was significantly smaller than in cells expressing α2A-R-Gαoalone (Fig. 6a). Inhibition was 21.0 ± 12.4% in P1 (n = 16, p < 0.01, Fig. 6b). Interestingly, the presence of Gαqz5also almost abolished facilitation by the PP at all potentials examined (Fig. 6c). For example the P2/P1 facilitation ratio in clonidine was 1.24 ± 0.38 at 0 mV and 0.87 ± 0.11 at +10 mV (n = 11, both p < 0.01 compared with the much greater facilitation shown by the α2A-R-Gαo alone). As a corollary of this, no agonist-induced depolarizing shift of the I-V relationship for IBa was detected (data not shown). Furthermore, no slowing of the activation kinetics was evident during P1 (e.g., Fig. 6a). In summary, coexpression of Gαqz5 with α2A-R-Gαo reduced the inhibition and reversed the P2/P1 facilitation observed upon activation of the α2A-R-Gαo alone.
Mechanism of Action of Gαqz5.
We addressed the possibility that the effects produced by Gαqz5on CaV2.2 channel modulation might be mediated by a signaling pathway downstream from Gq rather than directly by the Gαqz5 subunit. It has been proposed that overexpression of any Gα subunit could abolish calcium channel inhibition by sequestering Gβγ subunits, which would therefore become unavailable for receptor activation (Jeong and Ikeda, 1999). However, this will depend on the balance between Gα activation to form free Gα-GTP and Gβγ interaction with the Gα-GDP species. In such a scenario, coexpression of Gαqz5 could buffer the effect of the Gβγ released upon activation of α2A-R-Gαo, in a similar way to transducin; as we have shown, however, Gαqz5 is able to be activated. Once activated, it would then lead to stimulation of phospholipase C (Conklin et al., 1993), causing breakdown of phosphatidylinositol 4,5-bisphosphate (PIP2) into inositol 1,4,5-trisphosphate and diacylglycerol, the latter stimulating PKC. Activation of PKC has been reported to counter G-protein modulation of rat CaV2.2 (Zamponi et al., 1997; Hamid et al., 1999). However, elevation of PIP2 has also been shown to modulate CaV2.1, mimicking that by Gβγ (Wu et al., 2002). To investigate whether the reduction in inhibition and loss of facilitation in our coexpression studies with Gαqz5 were caused by a Gβγ buffering effect or by a specific downstream effect of activated Gαqz5 protein, we first chose to block the downstream action of activated Gαqz5 by coexpressing the C-terminal peptide of phospholipase C-β1 (PLC-β1ct), which binds activated Gαq and acts as a GTPase-activating protein (Kammermeier and Ikeda, 1999). Inhibition by clonidine in cells coexpressing α2A-R-Gαo and Gαqz5 together with PLC-β1ct returned to levels comparable with when α2A-R-Gαo was expressed alone (Fig. 6b). Furthermore the P2/P1 facilitation ratio in the presence of clonidine was increased relative to that in the presence of α2A-R-Gαo and Gαqz5 at all potentials between 0 and +20 mV, being 4.88 ± 2.48 at 0 mV and 2.25 ± 0.66 at +10 mV (Fig.6d, n = 6, p < 0.05 relative to facilitation in clonidine for α2A-R-Gαo and Gαqz5 alone at 0 and +10 mV).
Because PLC-β1ct only binds activated Gqspecies and was able to reverse the effect of Gαqz5, this must occur via its GTP-bound form. We therefore examined the role of downstream effectors of Gq. We investigated the effect of activating PKC to mimic the presence of Gαq as a signal transduction component, and simultaneously removed its presence as a potential Gβγ buffering agent. We used PDBu, an activator of PKC, on cells expressing the α2A-R-Gαo fusion protein. After assessing the inhibition of CaV2.2 currents and the voltage-dependent facilitation elicited by clonidine alone, cells were perfused with PDBu (500 nM) in the presence of clonidine (Fig. 7, a and b). Within 5 min after the start of PDBu application, IBapartially recovered from inhibition by clonidine. During P1, inhibition by clonidine was reduced from 77.8 ± 6.1 to 56.1 ± 9.4% in the additional presence of PDBu (Fig. 7c, n = 7,p < 0.001). Application of PDBu also resulted in reduced current facilitation (Fig. 7d, n = 7). After application of PDBu and clonidine, the current during P1 showed a rapid activation phase, further evidence for the loss of voltage-dependent inhibition (e.g., Fig. 7a, traces). Both the loss of inhibition and the reduction of facilitation are similar to the effect of Gαqz5. Application of PDBu (500 nM) in the absence of receptor activation did not cause any increase of CaV2.2 IBa, rather reducing it by 37 ± 9% after application for 3 min, with a loss of control facilitation (n = 6, data not shown).
In a second approach to examine the involvement of PKC in the effects of Gαqz5, we observed that the PKC inhibitor GF109203X partially restored the voltage-dependence of G-protein modulation in the presence of Gαqz5. After a 30-min preincubation with 1 μM GF 109203X, application of clonidine to cells expressing α2A-R-Gαo and Gαqz5 produced a 54 ± 15% inhibition of IBa at 0 mV (n = 9), and the P2/P1 facilitation ratio approached that in the absence of Gαqz5 [2.4 ± 0.5 (n = 9)]. These two pieces of data indicate that PKC activation is at least in part responsible for the effects of Gαqz5.
Discussion
The Advantage of Using GPCR-G Protein Tandems.
We sought to recreate proximity between a GPCR, the α2AR, and a specific G-protein by using tandem constructs. Both the chimeric receptors α2A-R-Gαi and α2A-R-Gαo reconstituted N-type current inhibition, comparable with the α2AR w.t. Similarly, it has been found that a tandem between the muscarinic m2 receptor and Gαz was able to modulate GIRK channels by release of Gβγ (Vorobiov et al., 2000). This is in contrast to their inability to activate downstream effectors via the Gα moiety (Sautel and Milligan, 1998; Burt et al., 1998), presumably because the Gα-subunits are not amplified and also because they are constrained. The conclusion of these results is that the release of Gβγ from both the activated GPCR tandems is completely sufficient to produce typical voltage-dependent inhibition of N-type calcium channels. This is confirmed by the inability of the IE mutant of α2A-R-Gαo, which does not bind Gβγ, to mediate inhibition of CaV2.2 by clonidine. Although tonic facilitation was seen with this mutant in the absence of agonist (Fig. 3c), this was no greater than for the nontandem α2A-R (Fig. 2c), where inhibition by clonidine was observed (Fig. 1e).
It has been proposed that members of the Gosubfamily are responsible for the voltage-dependent inhibition of calcium channels in sympathetic neurons, whereas Gi produced only a voltage-independent effect (Delmas et al., 1999). However, we did not find a clear correlation between the Gα-subunit in the tandem and the voltage-dependence of the inhibition, although there was a slightly greater voltage-dependent effect with the α2A-R-Gαo fusion protein. This may relate to the endogenous Gβγ dimers with which the Gα subunits preferentially associate. Indeed, the kinetics and voltage-dependence of Gβγ dissociation and reassociation are dependent on the nature of the Gβγ dimers (Stephens et al., 1998).
Effects of Gαq on G-Protein Modulation of Calcium Channels.
The role of Gq in G-protein modulation of calcium currents remains unclear. It has been shown that Gq is not involved in modulation by the α2A-R of the (largely N type) calcium currents in mouse sympathetic neurons (Haley et al., 2000). In the present study, expression of Gαq produced negligible inhibition of N-type channels, consistent with its very low ability to couple to the α2A-R (Chabre et al., 1994). In contrast, the chimeric counterpart, Gαqz5, allowed significant inhibition of CaV2.2, indicating that substitution of the C terminus of Gαz enhanced the coupling to the α2A-R (Conklin et al., 1993). However, Gαqz5 showed a reduced ability to inhibit IBa compared with Gi/o. The inhibition also showed a lack of voltage-dependence; together, these results suggested that Gqz5 acts via a different or modified signaling mechanism compared with Gi/o. A similar voltage-independent inhibition of Ca2+ channels by the Gq-coupled muscarinic m1 receptor was shown to involve both the Gαq and Gβγ subunits (Kammermeier et al., 2000). Furthermore, the voltage-independent inhibition was converted into voltage-dependent inhibition by sequestering activated Gαq (Kammermeier and Ikeda, 1999).
In the present study, coexpression of Gαqz5with α2A-R-Gαo caused first a reduction of clonidine-induced inhibition of CaV2.2 and second a loss of voltage-dependent facilitation. This action of Gαqz5 could result from a number of mechanisms: 1) Gβγ buffering, as suggested for Gαq (Jeong and Ikeda, 1999), or 2) Gαqz5 might interact with, and be activated by, the α2A-R-Gαo tandem. It has been observed previously that chimeric receptor-Gα constructs are able to activate not only the tethered Gα subunit but also endogenous subunits of the Gi/o family (Burt et al., 1998). In the case of Gαqz5 this would result in downstream activation of phospholipase C, resulting in elevation of inositol 1,4,5-trisphosphate and diacylglycerol and concomitant reduction of PIP2. One potential downstream pathway would be PKC activation and subsequent phosphorylation of either the calcium channel or the α2A-R to suppress G-protein modulation. Another potential downstream pathway would be via reduction of PIP2, because elevation of PIP2 mimics and may play an essential role in G-protein modulation (Wu et al., 2002).
We have addressed these possibilities in turn. If the mechanism were Gβγ sequestration, Gαqz5 should act identically to Gαt. However Gαt reduced inhibition of CaV2.2 via α2A-R-Gαo from 75 to 43% but did not abolish facilitation in the presence of clonidine (Fig. 3a, traces). In contrast, Gαqz5 reduced inhibition by clonidine to 36% but completely abolished facilitation (Fig. 5a, traces). Furthermore, in cells coexpressing α2A-R-Gαo and Gαqz5, it was possible to restore typical Go-mediated facilitation by enhancing the GTPase activity of activated Gαqz5 with PLC-β1ct. This suggests that the effect of Gαqz5 is downstream from its activation.
Two pieces of evidence indicate that PKC activation is involved in the response to Gαqz5, although it may not represent the entire story. Firstly, application of a PKC agonist to cells expressing α2A-R-Gαo mimicked the effects of Gαqz5, resulting in reduced inhibition by clonidine and loss of PP facilitation. Secondly, in cells coexpressing α2A-R-Gαoand Gαqz5, Go-mediated inhibition and facilitation were restored with a PKC inhibitor. Taken together, these results suggest that activation of the PLC-β signaling pathway by the α2A-R coupling to Gqz5 can oppose G-protein-mediated inhibition of CaV2.2.
Potential Targets for PKC Phosphorylation.
PKC-mediated phosphorylation might occur at several sites, either separately or in combination. PKC activation has been shown to cause phosphorylation-dependent desensitization of the α2A-R in COS-7 and Chinese hamster ovary cells (Liang et al., 1998). It is possible that this process may play a role in the effect of Gαqz5, although it is unclear how this could result in a selective loss of facilitation while substantial clonidine-mediated inhibition remains. Alternatively, PKC could phosphorylate one or more calcium channel subunits, thus rendering the channel less responsive to G-protein mediated inhibition and abolishing facilitation. There are a number of possible mechanisms by which this might occur. Phosphorylation of CaV2.2 or an accessory β subunit may result in the loss of its ability to be modulated by Gβγ. Residues in the I-II linker of rat CaV2.2 have been proposed to be a target of phosphorylation by PKC and to be responsible for PKC antagonism of G-protein modulation (Zamponi et al., 1997). Evidence has been presented recently that this process involves binding of Gαq and PKC to the C terminus of CaV2.2 (Simen et al., 2001). Direct activation of PKC has been shown to counteract inhibition of N-type calcium channels by norepinephrine (Shapiro et al., 1996). Subsequently, the importance was examined of phosphorylation sites on the I-II linker of rat CaV2.2, including Thr422and Ser425, in the modulation by PKC (Zamponi et al., 1997). An increase in calcium current was observed when mimicking channel phosphorylation on either of these residues by mutation to Glu, and a reduction of G-protein modulation by somatostatin when Thr422 was mutated to Glu (Hamid et al., 1999). However, this effect was subsequently observed only with Gβ1 and not with other Gβ subunits, calling into question its general relevance (Cooper et al., 2000). Indeed, we have shown that G-protein modulation of CaV2.2 is not dependent on the presence of the I-II linker of a modulatable calcium channel, whereas the N terminus is essential (Canti et al., 1999).
Although the rabbit CaV2.2 used in the present study shows very high overall sequence conservation in the I-II linker, and retains Ser425, there is Ala at position 422, thus ruling out the role of phosphorylation of this residue in removing G-protein modulation. In addition, Ser425 is not an optimal consensus PKC phosphorylation motif (AAAKKSRSD). Furthermore, we did not observe any increase of N-type IBa upon application of PDBu in the absence of G-protein modulation. This would agree with results in superior cervical ganglion neurons, where the only effect of PKC activation was antagonism of G-protein inhibition (Barrett and Rittenhouse, 2000).
The mechanism of the reduction of G-protein modulation and loss of P2/P1 facilitation that is characteristic of coexpression of Gαqz5 thus seems to involve its activation, and at least in part involves downstream activation of PKC, but the main target site(s) for phosphorylation may not be the calcium channel α1 subunit. Facilitation involves the unbinding of Gβγ subunits from the channel during the depolarizing PP (Stephens et al., 1998); this requires the functional interaction of CaVβ subunits (Canti et al., 2000, 2001; Meir et al., 2000). In the absence of coexpressed CaVβ subunit, we observed previously that activation of the D2 dopamine receptor produced only a small voltage-independent inhibition of CaV2.2 calcium channels, whereas in the presence of CaVβ subunits, the inhibition was much larger and voltage-dependent (Meir et al., 2000). It is therefore possible that phosphorylation of the CaVβ subunit might mediate the loss of facilitation resulting from Gαqz5 coexpression. Indeed, another CaVβ subunit, β2a, is phosphorylated stoichiometrically by PKC (Puri et al., 1997). We will examine in a future study whether phosphorylation of CaVβ subunits by PKC is responsible for the effects of Gαqz5.
Acknowledgments
We are grateful to N. Balaguero, W. S. Pratt, M. Nieto-Rostro, and K. Chaggar for technical assistance, to Drs. K. M. Page and N. S. Berrow for molecular biology expertise, and to A. Meir for helpful discussion. We thank the following for gifts of cDNA: Dr. L. E. Limbird (Vanderbilt University, Nashville, TN) (α2A-R); Dr/B. Conklin (J. David Gladstone Institutes, San Francisco, CA) (Gαq and Gαqz5); and Dr. S. Ikeda (National Institutes of Health, Bethesda, MD) (PLC-β1ct).
Footnotes
- Abbreviations:
- GPCR
- G-protein-coupled receptor
- α2A-R
- α2A-adrenergic receptor
- PTX
- pertussis toxin
- GIRK
- G-protein-coupled inward rectifier K channel
- GFP
- green fluorescent protein
- w.t.
- wild type
- RS-79948-197
- [8aR,12aS,13aS]5,6,8a,9,10,11,12,12a,13,13a-decahydro-12-ethanesulfonyl-3-methoxy-6H-isoquino[2,1-g]-[1,6]naphthyridine hydrochloride)
- GTPγS
- guanosine 5′-O-(3-thio)triphosphate
- HA
- hemagglutinin
- PAGE
- polyacrylamide gel electrophoresis
- PVDF
- polyvinylidene fluoride
- TTBS
- Tween 20/Tris buffered saline
- PDBu
- phorbol dibutyrate
- GF 109203X
- bisindolylmaleimide I
- PKC
- protein kinase C
- PLC-β1ct
- phospholipase C-β1 C terminus
- PP
- prepulse
- Gαt
- Gα-transducin
- PIP2
- phosphatidylinositol 4,5-bisphosphate
- IE
- Ile19Ala, Glu20Ala
- Received September 11, 2002.
- Accepted January 14, 2003.
- The American Society for Pharmacology and Experimental Therapeutics