Presynaptic Ca²⁺ channels of the Ca_v2 family conduct N-, P/Q-, and R-type Ca²⁺ currents (Ertel et al., 2000) and are inhibited by neurotransmitters acting through G protein-coupled receptors (Hille, 1994). This negative feedback mechanism has an important effect on regulation of neurotransmitter release and synaptic transmission (Lipscombe et al., 1989; Brody and Yue, 2000). G protein-mediated inhibition is caused by a positive shift in the voltage dependence of activation (Bean, 1989) and is reversed by strong positive prepulses or by trains of action potential-like stimuli, producing prepulse facilitation of Ca²⁺ channel activity (Marchetti et al., 1986; Tsunoo et al., 1986; Brody and Yue, 2000). These effects are mediated by binding of G protein βγ subunits (Gβγ) to Ca_v2 channels (Herlitze et al., 1996; Ikeda, 1996). Gβγ binds to at least three sites on Ca_v2 channels. The primary site, located in the intracellular loop connecting domains I and II (L_I-II), was identified by mutagenesis studies and G protein binding studies (De Waard et al., 1997; Herlitze et al., 1997; Page et al., 1997; Zamponi et al., 1997). A second site of interaction is located in the C-terminal domain of Ca_v2.3 subunits (Zhang et al., 1996; Qin et al., 1997; Furukawa et al., 1998; Simen et al., 2001). The third site of interaction is located at the N terminus and in domain IofCa_v2.2 and Ca_v2.3 channels (Page et al., 1998; Simen and Miller, 1998, 2000; Canti et al., 1999). All these binding sites appear to be involved in G protein binding and subsequent modulation of channel gating, but their relative functional roles have not been clearly resolved.

N-type Ca²⁺ currents in nerve terminals play a central role in regulating the release of neurotransmitters (Hirning at al., 1988: Turner et al., 1993; Dunlap et al., 1995). N-type Ca²⁺ currents are conducted by Ca_v2.2 channels (Dubel et al., 1992; Williams et al., 1992; Ertel et al., 2000), which are highly concentrated in nerve terminals (Westenbroek et al., 1992, 1995). Like the other members of the Ca_v2 family, Ca_v2.2 channels are inhibited by activation of G protein-coupled receptors (Hille 1994; Dunlap et al., 1995) through binding of Gβγ subunits (Herlitze et al., 1996; Ikeda, 1996). Localization of G protein interaction sites and determination of their functional effects on Ca_v2.2 channels is important for understanding the molecular mechanisms of modulation of the activity of the Ca²⁺ channel and subsequent regulation of synaptic transmission. The initial evidence for a functional role of the C-terminal domain in G protein modulation came from studies of a family of Ca_v2.2/Ca_v1.2 chimeras (Zhang et al., 1996), which led to the conclusion that domain I, L_I-II, and the C-terminal domain are required for establishment of G protein modulation in an unresponsive Ca_v1.2 channel. Consistent with these results, Qin et al. (1997) defined a C-terminal G protein binding site that is required for modulation in Ca_v2.3, which has weak G protein modulation compared with Ca_v2.1 and Ca_v2.2 channels (Toth et al., 1996); however, Furukawa et al. (1998) found that neither a deletion of the C-terminal region nor a replacement of the C-terminal domain of Ca_v2.2 by that of Ca_v1.2 affected Gβγ-dependent facilitation of the channel. In contrast, they found that a Ca_v2.2 chimera containing a substitution of L_I-II from Ca_v1.2 displayed a complete loss of Gβγ-dependent prepulse facilitation. From these reports, the functional roles of the G protein binding sites in the C-terminal domains of Ca_v2.2 and related Ca²⁺ channels remain unresolved, and the location of the Gβγ binding site in the C-terminal domain of Ca_v2.2 is unclear. In this study, we have identified a Gβγ binding site near the C terminus of Ca_v2.2b channels through in vitro binding of fusion proteins and in vivo binding in yeast cells, and we have tested the functional significance of the C-terminal Gβγ-binding domain by whole-cell voltage clamp recording. Our results show that Gβγ subunits bind to a segment of 80 amino acid residues near the C terminus of Ca_v2.2b, which contributes to prepulse facilitation of Ca²⁺ channel activity primarily at the basal level of G protein activation.

Materials and Methods

Construction of Ca²⁺ Channel Plasmids. A recombinant cDNA encoding Ca_v2.2b from rat brain (originally called rbB-II; Dubel et al., 1992) was subcloned into pMRC-CMV (Invitrogen; Zhong et al., 2001). The β1b subunit was in the vector pMT2XS, α2δ in pZEM228, and CD8 in EBO-pcD (Zhong et al., 2001). To construct the C-terminal deletion mutant of Ca_v2.2b (Ca_v2.2bΔCT5), C at nt 6768 was mutated to G by a two-step PCR mutagenesis such that a TAG stop codon was introduced at residue 2256. A PCR fragment (nt 6526–7011) containing the mutation was subcloned into the Ca_v2.2b expression plasmid by replacing its wild-type counterpart. To identify the G protein binding sites on the Ca²⁺ channel, NT (nt 1–285, aa 1–95), L_I-II (nt 1069–1449, aa 357–483), L_II-III (nt 2131–3429, aa 711-1143), L_III-IV (nt 4252–4422, aa 1418–1474), CT1 (nt 5128–6130, aa 1710–2043), CT2 (nt 6130–7011, aa 2044–2336), CT3 (nt 6628–7011, aa 2211–2336), CT4 (nt 6130–6747, aa 2044–2249), and CT5 (nt 6768–7022, aa 2257–2336) were subcloned into the pAS2–1 vector (BD Biosciences Clontech, Palo Alto, CA) for expression of GAL4 DNA-binding domain (GBD) hybrids in yeast. Plasmids expressing the GAL4 activation domain (GAD)-Gβ hybrids and the yeast strain for the yeast two-hybrid assay have been described by Garcia et al. (1998). Plasmids expressing hexahistidine-tagged channel fragments were described earlier (Sheng et al., 1994). L_I-II was also cloned into the pGEX4T-1 (Pfizer, Inc., New York, NY) as a EcoRI-SalI fragment and CT5 into pGEX3X (Pfizer, Inc.) as a BamHI-PstI fragment to express GST fusion proteins.

Protein Purification. Ca_v2.2b fragments were subcloned in both pTrcHisC and the GST-fusion protein expression vectors pGEX3X or pGEX4T-1 by PCR amplification and subcloning. The constructs including the PCR-amplified Ca²⁺ channel inserts were confirmed by DNA sequencing. These plasmids were transformed into B21DE3 (Novagen, Madison, WI) for protein expression. To purify the hexahistidine-tagged fusion protein, 500 ml of LB medium containing 75 μg/ml ampicillin was seeded with a 10-ml overnight culture containing the expression plasmid. Isopropyl β-d-thiogalactoside (0.1 mM) was added to the culture when the cell density generated an optical density at 600 nm of 0.7. After a 3-h induction, the culture was harvested, and the cell pellet was resuspended in the phosphate buffer (QIAGEN, Valencia, CA) containing protease inhibitors (200 μm phenylmethylsulfonyl fluoride, 2 μg/ml leupeptin, 1 μg/ml pepstatin, 1 μg/ml aprotinin). Following sonication, 1% Triton X-100 was added to the lysate. The lysate was incubated on ice for 30 min and subjected to centrifugation for 50 min at 15,000 rpm at 4°C. The supernatant was saved and incubated for 2 h with Ni-NTA beads (QIAGEN) that had been equilibrated with phosphate buffer. After washing by phosphate buffer and the same buffer containing 20 mM imidazole, the hexahistidine-tagged protein was eluted by a series of concentrations of imidazole from 50 to 300 mM. The fraction containing concentrated pure protein was used for in vitro binding assays. The GST fusion proteins were expressed similarly in bacteria and harvested. Cell pellets were resuspended in phosphate-buffered saline containing 5 mM EDTA and protease inhibitors mentioned above. Cell lysates were obtained by sonication and centrifugation as described above. The lysates were incubated for 1 h with glutathione-Sepharose (Pfizer, Inc.) equilibrated with PBS. Following incubation, the beads were washed with PBS containing 0.1% Triton X-100, 5% glycerol, 0.5 mM EDTA, and 0.2 mM phenylmethylsulfonyl fluoride, and the GST-fusion protein was eluted with 20 mM reduced glutathione in 50 mM Tris-HCl, pH 8.0. The untagged wild-type Gβ₁γ₂ protein, hexahistidine-tagged wild-type Gβ₁γ₂, and the five hexahistidine-tagged Gβ₁ mutants in complex with the γ₂ subunit were purified as described (Kozasa and Gilman, 1995).

In Vitro Binding Assay. Affinity-purified wild-type Gβ₁γ₂ complex (50 nM) was incubated with 100 nM hexahistidine-tagged Ca²⁺ channel fragments in a volume of 400 μl of buffer A (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Tween 20, 20 mM imidazole) for 2 h at 4°C. Ni-NTA beads (20 μl) that had been equilibrated with buffer A were then added to the preincubated protein complexes followed by tumbling for another 2 h at 4°C. The beads were then centrifuged briefly at top speed in a microcentrifuge, and washed with buffer A four times. The hexahistidine-tagged Ca²⁺ channel fragments and their associated proteins were eluted from the Ni-NTA beads with 600 mM imidazole in Tris-buffered saline (10 mM Tris-HCl, pH 8.0, 150 mM NaCl) at room temperature for 30 min on a vortex mixer. The eluates were then subjected to Western blot analysis with anti-Gβ₁ antibodies.

In experiments that tested the interactions of L_I-II and CT5 with Gβ₁, binding assays were performed in a similar fashion. In brief, approximately 100 nM GST-fusion proteins were incubated with increasing amounts of wild-type or mutant hexahistidine-tagged Gβ₁ in complex with γ₂. After 2 h, Ni-NTA beads were added to the protein mix and incubated for another 2 h. Beads were pelleted and washed with buffer A. Bound proteins were eluted with 600 mM imidazole in Tris-buffered saline. Eluates were subjected to Western blotting with anti-GST antibodies (Zymed Laboratories, South San Francisco, CA).

Yeast Two-Hybrid Assay. The assay was performed as described (Garcia et al., 1998). In brief, plasmids expressing GBD and GAD hybrids were cotransformed into the yeast reporter strain, PJ69-4A. Transformants carrying these two plasmids were first isolated and subsequently transferred to the yeast medium to select those that express the reporter genes as a result of two-hybrid interactions.

Transfection and Electrophysiology. tsA-201 cells were maintained in Dulbecco's modified Eagle's medium/Ham's F-12 medium (1:1) supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA) at 37°C under 10% CO₂. Cells plated in 35-mm tissue culture dishes were grown to 60 to 80% confluence and transfected by the Ca²⁺ phosphate method with a total of 4 μg of DNA including a 1:1 M ratio of cDNAs encoding Ca²⁺ channel subunits and 0.3 μg of a CD8 expression plasmid for identification of transfected cells. The cells were subcultured at 24 h after the transfection. At least 48 h after transfection, positive transfected cells were visually identified by labeling with anti-CD8 antibody. Barium currents were recorded by whole-cell patch clamp technique using a List EPC-7 amplifier and filtered at 5 kHz with an eight-pole Bessel filter. Leak and capacitance currents were measured using hyperpolarizing pulses and subtracted using the p/-4 method. The extracellular recording solution contained 120 mM Tris, 4 mM MgCl₂, and 10 mM BaCl₂ with pH adjusted to 7.3 by methanesulfonic acid. The internal pipette solution consisted of 120 mM aspartic acid, 5 mM CaCl₂, 2 mM MgCl₂, 10 mM Hepes, 10 mM EGTA, and 2 mM Mg-ATP with pH adjusted to 7.3 by CsOH. In some experiments, guanosine 5′-3-O-(thio)triphosphate (GTPγS) was added to the internal solution at a concentration of 0.2 mM. Somatostatin (SST) was dissolved in water to give a stock solution of 1 mM and added to the extracellular solution at a final concentration of 1 μM. N-Ethylmaleimide (NEM) was prepared in distilled water at 50 mM and added to the extracellular solution to give a final concentration of 50 μM. All agents were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise mentioned. All averaged values represent mean ± S.E.M.

Results

Interaction of the Intracellular Loops and C Terminus of Ca_v2.2 with Gβγ. To understand the mechanism by which Gβγ subunit complexes modulate the N-type Ca²⁺ channels, we sought to identify the regions on Ca²⁺ channel that interact with the Gβγ subunits. We used an in vitro biochemical binding assay and the in vivo yeast two-hybrid assay in intact yeast cells (Fields and Song, 1989). In the biochemical assay, intracellular loops were expressed in bacteria as hexahistidine-tagged fusion proteins and were purified by nickel affinity columns. The interactions between the fusion proteins and affinity-purified Gβ₁γ₂ protein were assessed by examining whether Gβ₁γ₂ could be coprecipitated by the Ni-NTA agarose beads that bind to the hexahistidine moiety. In these experiments, the intracellular loop connecting domains II and III (L_II-III) of the skeletal muscle Ca²⁺ channel Ca_v1.1 was used as a negative control. Consistent with previous results, our experiments showed that L_I-II of Ca_v2.2b interacted strongly with the Gβ₁γ₂ subunits, whereas the L_II-III and L_III-IV did not interact (Fig. 1, A and B). We next focused on the C-terminal tail of Ca_v2.2b, which contains the last 666 amino acid residues. To facilitate the purification of hexahistidine-tagged fusion proteins of this region, the C-terminal domain was divided in half and expressed and purified as two separate proteins for assay. We observed that Gβ₁γ₂ interacted effectively with the C-terminal half (CT2, aa 2044–2336), but not the N-terminal half of the C-terminal domain (CT1, aa 1710–2043; Fig. 1B), suggesting that this may be a region involved in mediating G protein modulation.

Binding of L_I-II and CT2 to Gβγ was also confirmed by the yeast two-hybrid assay, which detects protein-protein interactions in intact cells. In our assays, the Ca²⁺ channel fragments were expressed as GBD hybrids and the Gβ₁ subunit was expressed as a GAD hybrid. Interactions between the GBD and GAD hybrids in the yeast reporter strain can reconstitute GAL4 transcriptional activator function to activate the expression of two reporter genes, ADE2 and HIS3, that are involved in adenine and histidine synthesis, respectively. Thus, yeast cells expressing interacting hybrid proteins are able to grow on medium lacking these two amino acids. When we tested the intracellular loops and the C-terminal tail, we observed that only yeast cells coexpressing Gβ₁ and L_I-II or CT2 grew on the selection medium, indicating specific interactions (Fig. 2). In these experiments, no mammalian γ subunit was introduced into the yeast cells. Therefore, the Gβ₁ subunit may complex with the endogenous yeast γ subunit, or the overexpressed GAD-Gβ₁ fusion protein may be stable enough to interact separately with Ca²⁺ channel fragments.

To identify the minimal domain in the C terminus of the channel that is required for the interaction with Gβγ, three deletion constructs were made in the CT2 region (aa 2044–2366) (Fig. 3A) and introduced into the yeast reporter strain with the GAD-Gβ₁-expressing plasmid. The two-hybrid analyses showed that the last 80 residues (CT5, aa 2257–2336) of Ca_v2.2b are necessary and sufficient for interaction with the Gβ₁ subunit (Fig. 3A). This region was also shown to bind to the G protein in the in vitro biochemical assay in which the last 80 residues were expressed as a GST-fusion protein and the Gβ₁ subunit was tagged with hexahistidine. GST-CT5, but not GST was coprecipitated with hexahistidine-Gβγ by the Ni-NTA beads (Fig. 3B). These results confirmed that the C-terminal Gβ binding site is located within the last 80 amino acid residues at the C terminus of the polypeptide. The location of this site is different from the previously reported Gβ binding site in Ca_v2.3, which is located in the middle of the C-terminal region (Qin et al., 1997); however, these sites contain a conserved sequence motif (see Discussion).

Specificity of Interaction of the C-Terminal Gβγ-Binding Site with Different Gβ Subunits. It has been reported that effectors modulated by the βγ subunits of the G proteins have differential affinities toward different subtypes of G proteins. For example, a G protein receptor kinase (GRK), GRK2, binds to Gβ₁ and Gβ₂, but not Gβ₃ (Daaka et al., 1997). We examined the interaction of the last 80 residues of Ca_v2.2b with the known five subtypes of Gβ subunit. In contrast to L_I-II, which showed differential affinities toward these Gβ subunits (Garcia et al., 1998), yeast cells cotransformed with GBD-CT5 and the five GAD-Gβ-expressing plasmids showed similar growth rates on the selection medium (Fig. 4), indicating similar interaction strength between the five paired hybrids. Since the five Gβ subunit hybrids were expressed at a similar level (Garcia et al., 1998), it is probable that CT5 interacts with the five Gβ subtypes with similar affinities. These data indicate that the C-terminal Gβ binding site is unable to distinguish among the different subtypes of G protein β subunits. It may function together with L_I-II to mediate specific modulation by these G proteins.

Kinetics and Voltage Dependence of Activation of Ca_v2.2b and Ca_v2.2bΔCT5 Channels. Ca²⁺ channels consisting of the Ca_v2.2b or Ca_v2.2bΔCT5 α1 subunits, α2δ subunits, and β1b subunits were expressed in the tsA-201 subclone of human embryonic kidney 293 cells as described under Materials and Methods. These two Ca²⁺ channels exhibit similar fast activation kinetics (Fig. 5A). Although the midpoint of the conductance-voltage curve of Ca_v2.2b is more positive (V_a = 18.6 ± 2.2 mV, n = 24; Fig. 5B) and less steep in comparison to Ca_v2.2bΔCT5 (V_a = 15.4 ± 1.6 mV, n = 28; Fig. 5B), the difference is not statistically significant (p > 0.05). These results show that deletion of CT5 has no effect on basic Ca²⁺ channel function.

Intrinsic and Somatostatin-induced G Protein Activation and Prepulse Facilitation of Ca_v2.2b and Ca_v2.2bΔCT5 Channels. To examine G protein modulation of these Ca²⁺ channels, we induced prepulse facilitation without and with SST in the bath solution to activate endogenous SST receptors and G proteins (see Fig. 6A, inset). Facilitation during the prepulse was assessed by comparing the tail current following Test Pulse 2 with the tail current following Test Pulse 1. In the absence of SST, Ca_v2.2b channels activate rapidly at relatively negative membane potentials (Fig. 6, A and B). A depolarizing prepulse has no obvious effect on the rate of activation and causes only a small shift in the voltage dependence activation (Fig. 6, A and B). Depolarizing prepulses cause a small degree of facilitation, which is maximum at +30 mV (9.7 ± 2.3%; Fig. 6B) and is significant compared with the baseline variation in barium current (Fig. 10). As this facilitation is observed without exogenous activation of G proteins, we have designated it intrinsic facilitation. Intrinsic facilitation of Ca_v2.2 channels has been observed previously in neuroblastoma/glioma cells and transfected tsA-201 cells (Kasai, 1991; Zhong et al., 2001). The intrinsic facilitation indicates that, in the absence of exogenous activators, the cells have constitutively active G proteins that cause tonic inhibition of Ca²⁺ currents and allow prepulse facilitation. Activation of G proteins with SST (1 μM), which reacts with receptors constitutively expressed in the tsA 201 cells, slows activation of Ca_v2.2b and shifts its voltage dependence of activation to more positive membrane potentials (Fig. 6, C and D). Depolarizing prepulses cause an obvious facilitation (22.7 ± 1.8% at +30 mV; Fig. 6, C and D), resulting from voltage-dependent reversal of G protein inhibition.

In the absence of SST, Ca_v2.2bΔCT5 channels activate rapidly at relatively negative membrane potentials, similar to Ca_v2.2b (Fig. 7, A and B). A depolarizing prepulse has no obvious effect on the rate or voltage dependence of activation (Fig. 7, A and B). Using a single test pulse to +30 mV, depolarizing prepulses cause a 2.6 ± 2.5% facilitation of Ca_v2.2bΔCT5, signifcantly less than wild-type intrinsic facilitation (p < 0.02) and not significantly different from zero (Fig. 10). These results revealed that the C-terminal Gβγ binding site contributes to intrinsic facilitation of Ca_v2.2b. Activation of G proteins with SST (1 μM) slows activation of Ca_v2.2bΔCT5, shifts its voltage dependence of activation to more positive membrane potentials, and increases prepulse facilitation (Fig. 7, C and D) as in wild type. At +30 mV test pulse, the level of facilitation (19.3 ± 1.6%; Fig. 7) is slightly, but not significantly, less than that of wild-type Ca_v2.2b (22.7 ± 1.8%; p = 0.11; Fig. 10).

GTPγS-Induced G Protein Activation and Prepulse Facilitation of Ca_v2.2b and Ca_v2.2bΔCT5 Channels. Activation of G proteins by SST may only activate part of the intrinsic G protein pool that is available for modulation of Ca²⁺ channels. To examine modulation by the full complement of G proteins, we used GTPγS in the recording pipette to fully activate endogenous G proteins (Fig. 8). In Ca_v2.2b channels, activation of G proteins with GTPγS slows activation and shifts its voltage dependence to more positive membrane potentials (Fig. 8, A and B). Depolarizing prepulses cause an obvious facilitation, with a maximum at +30 mV (36.0 ± 2.9%; Fig. 8, A and B). Similar to the results with Ca_v2.2b, Ca_v2.2bΔCT5 channels are slowly activated and have strong prepulse facilitation with G protein activation by GTPγS (Fig. 8, C and D). Prepulse facilitation +30 mV is 32.2 ± 1.2%, slightly but not significantly less than wild-type (p > 0.05; Fig. 10). Thus, as for activation by SST, there is a small but not statistically significant reduction in modulation by GTPγS with deletion of CT5.

To test whether the prepulse facilitation of Ca_v2.2b is truly dependent on G protein action, we used a G protein inhibitor. The sulfhydryl reagent NEM is a potent inhibitor of G protein action on Ca_v2.2 channels (Shapiro et al., 1994). Addition of 50 μM NEM to the extracellular solution blocked both SST-induced and intrinsic facilitation of Ca_v2.2b in our transfected cells (Fig. 9). Previous reports suggested that NEM rapidly eliminates agonist-induced inhibition whereas sparing the tonic inhibition of N-type Ca²⁺ currents in neuroblastoma/glioma cells (Kasai, 1991). Our results show that NEM blocks both somatostatin-induced and intrinsic facilitation of these channels in tsA-201 cells and that both these forms of modulation are dependent on G protein activation.

Primary Role of the C-Terminal Gβγ Binding Site in Intrinsic Facilitation. Comparison of the effects of deletion of the C-terminal Gβγ binding site in Ca_v2.2 on intrinsic, SST-induced, and GTPγS-induced facilitation revealed that the primary effects of the C-terminal Gβγ binding site is on intrinsic facilitation (Fig. 10). Intrinsic facilitation was reduced from 9.7% to essentially zero. In contrast, there was a trend toward reduction of SST-induced and GTPγS-induced facilitation but it was not statistically significant at the p < 0.05 confidence level. Possible mechanisms underlying these effects are considered under Discussion.

Discussion

Gβγ-Binding Sites on Ca²⁺ Channels. Ca²⁺ channels are modulated by G proteins through direct binding of Gβγ subunits to the α1 subunits. In this article, we studied two regions of the N-type Ca²⁺ channel Ca_v2.2b subunit that can interact with the βγ subunits of the G protein. Our results confirmed Gβγ binding to L_I-II, consistent with previous reports (De Waard et al., 1997; Herlitze et al., 1997; Zamponi et al., 1997; Furukawa et al., 1998). In addition, we have identified a novel site of Gβγ interaction located near the C terminus. Four separate lines of evidence support our identification of this Gβγ binding site. First, we observed specific binding interactions between fusion proteins containing segments of the C-terminal domain and Gβγ. Second, we measured this specific binding interaction in vivo in yeast cells with the yeast two-hybrid assay. Third, we detected a specific functional effect, loss of basal G protein modulation in transfected cells, upon deletion of this C-terminal Gβγ binding site. Fourth, we found that this Gβγ binding site has substantial amino acid sequence similarity to G protein regulatory sites in Ca_v2.3 channels and in the family of PAK kinases (see below). Together, these four lines of evidence constitute strong support for the conclusion that this segment of the distal C terminus of Ca_v2.2 channels is a genuine binding site for Gβγ.

We tested the specificity of G protein binding in vivo in yeast cells using the yeast two-hybrid assay. Previous work showed that L_I-II binds Gβγ subunits with the same specificity as modulation of channel function (Garcia et al., 1998). Gβ₁ and Gβ₂ interacted strongly with L_I-II, Gβ₅ interacted weakly, and Gβ₃ and Gβ₄ failed to interact in a yeast two-hybrid assay. In contrast to L_I-II (Garcia et al., 1998), our present results show that the C-terminal binding site of Ca_v2.2b has similar interaction with all five Gβ subunits. These data suggest that the C terminus does not contribute to the differential effects of the five subtypes of Gβ subunits on voltage-dependent facilitation, in contrast to the site in L_I-II (Garcia et al., 1998). The C-terminal site in Ca_v2.2 channels may interact with a conserved site on the Gβ subunit.

Comparison of the Gβγ-Binding Site in the C-Terminal Domains of Ca_v2.2b and Ca_v2.3. Comparing the Gβγ binding site we discovered near the C terminus of the Ca_v2.2b with the previously identified Gβγ binding site in C-terminal domain of the Ca_v2.3 channel (Qin et al., 1997) reveals 40% amino acid sequence identity (Scheme 1). Both of these sequences contain an SSL motif, which is also found in the Gβ binding of PAK kinase family members (Leeuw et al., 1998), strongly suggesting the site in Ca_v2.2 identified in this study is an authentic one. Interestingly, the C-terminal Gβγ binding site in the Ca_v2.3 channel is located near the middle of the C-terminal domain (Qin et al., 1997), rather than near the end of the C terminus as in Ca_v2.2b. R-Type currents mediated by Ca_v2.3 are not as strongly modulated by G proteins as N- and P/Q-type currents mediated by Ca_v2.1 and Ca_v2.2 (Toth et al., 1996; Simen and Miller, 1998).

The low level of modulation of Ca_v2.3 is entirely dependent on the C-terminal Gβγ binding site (Qin et al., 1997), in contrast to Ca_v2.2 where the Gβγ site in L_I-II is more important when G proteins are strongly activated and the C-terminal Gβγ site is significant primarily for basal modulation due to the low level of intrinsic G protein activity. It is possible that the comparatively weak modulation of the Ca_v2.3 channel reflects the weak regulatory capacity of the C-terminal Gβγ sites compared with the site in L_I-II, which is not functional in Ca_v2.3.

Function of the C-Terminal Gβγ-Binding Site. To understand the functional importance of the C-terminal Gβγ binding site in G protein modulation, we analyzed the deletion mutant Ca_v2.2bΔCT5. The current amplitude, voltage-dependent activation, and inactivation of the mutant channel were similar to wild type, suggesting that the deletion does not produce global structural changes; however, prepulse facilitation of this mutant channel is significantly smaller than that of the wild-type channel under basal conditions where intrinsic G protein activity modulates the channel (Fig. 10). These data can be interpreted in terms of a reversible equilibrium between “willing” and “reluctant” channel states with differing kinetics and voltage dependence of activation (Bean, 1989). Many factors alter G protein modulation by affecting the equilibrium between these two states, including G proteins, membrane depolarization, Ca_vβ subunits, and protein kinase C (Herlitze et al., 2001). G protein activation shifts the equilibrium to the reluctant state, whereas depolarizing prepulses shift the channel back to the willing state and relieve the G protein-dependent inhibition resulting in prepulse facilitation. Based on this model, our experimental data suggest that the C-terminal Gβγ binding site regulates voltage-dependent activation primarily by enhancing the G protein-dependent shift of the channel state when G protein activation is at the basal level, perhaps by enhancing the affinity for Gβγ binding. Prepulse facilitation is a sensitive method to test the changes of channel activity due to G protein modulation, even at the basal level. Although the intrinsic facilitation of Ca_v2.2b is less than 10%, the prepulse facilitation protocol clearly detected the difference between the wild-type and Ca_v2.2bΔCT mutant. Moreover, our results showed that both SST-induced and intrinsic facilitation are blocked by NEM, providing strong support for the conclusion that intrinsic facilitation is dependent on endogenous G protein activation. The simplest explanation of the data is that when the C-terminal site is present, a fraction of Ca_v2.2 channels are reluctant to open due to the inhibition by Gβγ bound at least in part to the C-terminal site. When this region is deleted, Gβγ binds with lower affinity; the Ca_v2.2 channel is less inhibited by endogenous activated Gβγ and is less available for prepulse facilitation because fewer of the channels are in the reluctant state.

Although experiments probing G protein regulation are usually designed to activate a single dominant G protein-coupled signaling pathway, the situation is often more complex in vivo since numerous signaling pathways are likely to be activated at a low level simultaneously. For example, central neurons receive simultaneous input from low level activation of G protein signaling pathways at synapses from many presynaptic neurons using different neurotransmitters. Enhanced regulation of Ca_v2.2 channels at low levels of G protein activation is likely to be physiologically significant in this in vivo situation. The function of the C-terminal Gβγ-binding domain in basal modulation of Ca_v2.2 channels in our experiments is consistent with a physiologically significant role at low levels of G protein activation in central neurons and other cell types in vivo.

Molecular Mechanisms of G Protein Binding and Function. Our data support the proposal that the functional target of voltage-dependent Gβγ interaction includes the loop L_I-II (De Waard et al., 1997; Herlitze et al., 1997; Page et al., 1997; Zamponi et al., 1997). However, our results do not support the proposal that C-terminal binding site is the primary one that mediates the action of Gβγ for Ca_v2.2, as suggested for Ca_v2.3 by Qin et al. (1997), since we find that deletion of this site in Ca_v2.2 reduces intrinsic facilitation but does not have a statistically significant effect on modulation by SST and GTPγS. Whether the C-terminal region comprises an independent Gβγ binding site or is part of a single multifaceted site is not clear. Our working hypothesis is that the C terminus of Ca_v2.2b may function as a secondary anchoring site for the Gβγ, which enhances its affinity for Ca_v2.2 channels and allows regulation by Gβγ interaction with L_I-II at lower concentrations of activated Gβγ. Consistent with this idea, Zhao et al. (2003) have recently reported that amino acid residues in two of the β-propellor domains of Gβγ are involved in regulation of inwardly rectifying potassium channels. By analogy, L_I-II and the C-terminal domain of Ca_v2.2 channels may bind to two different β-propellors of Gβγ subunits to give optimal regulation at low levels of G protein activation. Alternatively, more than one Gβγ complex may bind to the channel with additive or synergistic effects. The C-terminal region may have a higher affinity for Gβγ and contribute primarily to intrinsic facilitation at low level activation of G proteins. When G proteins are activated more strongly by G protein-coupled receptors or by GTPγS, L_I-II, and the N-terminal Gβγ binding site may dominate channel regulation and exceed the effects of Gβγ binding to the C-terminal region. These multiple functional sites may allow response to a wider range of activation of G protein signaling pathways. Further studies on the binding affinity of the N terminus, L_I-II and the C-terminal site to Gβγ in intact mammalian cells using fluorescence resonance energy transfer or bioluminescence resonance energy transfer methods with Ca_v2.2 channels labeled in these three Gβγ-binding domains may shed more light on the mechanism of G protein modulation. It is likely that structural studies will be required to finally resolve the number of distinct G protein binding sites and the segments of the Ca_v2.2 channels that form them.