Differential Interactions of the C terminus and the Cytoplasmic I-II Loop of Neuronal Ca2+ Channels with G-protein α and βγ Subunits

Interactions of G-protein α (Gα) and βγ subunits (Gβγ) with N- (α1B) and P/Q-type (α1A) Ca2+ channels were investigated using the Xenopus oocyte expression system. Gi3α was found to inhibit both N- and P/Q-type channels by receptor agonists, whereas Gβ1γ2 was responsible for prepulse facilitation of N-type channels. L-type channels (α1C) were not regulated by Gα or Gβγ. For N-type, prepulse facilitation mediated via Gβγ was impaired when the cytoplasmic I-II loop (loop 1) was deleted or replaced with the α1C loop 1. Gα-mediated inhibitions were also impaired by substitution of the α1C loop 1, but only when the C terminus was deleted. For P/Q-type, by contrast, deletion of the C terminus alone diminished Gα-mediated inhibition. Moreover, a chimera of L-type with the α1B loop 1 gained Gβγ-dependent facilitation, whereas an L-type chimera with the N- or P/Q-type C terminus gained Gα-mediated inhibition. These findings provide evidence that loop 1 of N-type channels is a regulatory site for Gβγ and the C termini of P/Q- and N-types for Gα.

A family of membrane-associated guanine nucleotide-binding regulatory proteins (G-proteins) 1 is essential for mediating signal transduction between cell-surface receptors and intracellular effectors such as adenylate cyclase, phospholipase C, phospholipase A 2 , and ion channels (1)(2)(3)(4). G-proteins are composed of three subunits termed ␣, ␤, and ␥. The ␣ subunit (G␣) contains a binding site for guanine nucleotides and possesses GTPase activity. Upon receptor stimulation, heterotrimeric Gproteins disassociate into an ␣-GTP complex and ␤/␥ dimer. In most systems, a GTP-bound G␣ activates or inhibits an effector system, and the functional half-life is determined by the intrinsic GTPase activity of G␣. Recently, it has been shown that the ␤␥ dimer (G␤␥) is significantly important in signal transduction as well (3).
High voltage-activated (HVA) Ca 2ϩ channels are negatively regulated by G-proteins in a membrane-delimited manner (2,4). This response is primarily mediated by pertussis toxinsensitive G-proteins (G o /G i ), in which G o ␣ has been shown to inhibit current from HVA Ca 2ϩ channels (5)(6)(7). Additionally, it has been shown that G␤␥ also transduces an inhibitory signal to HVA Ca 2ϩ channels (8,9). It remains to be determined, however, which subunit arm of the G-protein complex preferentially interacts with N-and P/Q-types of HVA Ca 2ϩ channels. Recently, it has been determined that the intracellular loop joining motif I and II (referred to as "loop 1" in the present study) is an interaction site on neuronal HVA Ca 2ϩ channels for G␤␥ (10 -13). Nevertheless, mapping of region(s) on HVA Ca 2ϩ channels responsible for interactions with G␣ and/or G␤␥ is still very incomplete (14).
To address these issues at the molecular level, we have functionally expressed ␣ 1A , ␣ 1B , and ␣ 1C of HVA Ca 2ϩ channels in Xenopus oocytes. These subunits were derived from rabbit brain N-type, P/Q-type, and cardiac L-type Ca 2ϩ channels, respectively. In addition, we have co-expressed ␦-opioid receptor (DOR) together with G␣ or G␤␥ as we did in determining a region of the muscarinic-gated K ϩ channel critical for activation by G␤␥ with the presumption that co-expression with G␣ or G␤␥ determines which kind of modulation takes place (15). In this paper, interactions of G␣ and G␤␥ with Ca 2ϩ channels were characterized using mutant and chimeric N-(␣ 1B ) and P/Q-type (␣ 1A ) Ca 2ϩ channels. The results, together with evidence for a direct binding provided by the companion paper (16), define the interaction sites of Ca 2ϩ channels for G␣ and G␤␥.

In Vitro Transcription
The 1.4-kb ApaI/ApaI and 6.8-kb HindIII/HindIII fragment containing the entire coding regions of DOR (17) and the Ca 2ϩ channel ␣ 1C subunit (18) were inserted into the HindIII site of the pSPA2 vector (19), to yield pSPDOR and pSPCDR, respectively. The plasmid pSPCDR was digested with XbaI, blunted with T4 DNA polymerase, and ligated with a SalI linker to yield pSPCDRS. The ␣ 1C subunit cDNA was kindly provided by Drs. Atsushi Mikami and Tsutomu Tanabe. The pSPA1, pSPA2, pSP72, pSP65, and pSP64 recombinant plasmids carrying the entire protein-coding sequences of G i3 ␣, G␤ 1 , G␥ 2 , and Ca 2ϩ channel * This investigation was supported in part by Ministry of Education, Science and Culture of Japan Research Grants 08770519 (to T. F.), 02557013, 06264101, 08680855 (to T. N.), and 04807013 (to M. Y.) and by National Institutes of Health Grant P01 HL22619-20 (to Y. M. and M. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Construction of Mutant and Chimeric Ca 2ϩ Channels
B3T⌬1-The plasmid pSPB3 carrying the entire protein-coding sequences of ␣ 1B (21) was digested with BamHI and circularized with T 4 DNA ligase to yield pSPB3BH. The 5.5-kb NotI/SrfI fragment excised from pSPB3BH was ligated with the 55-bp NotI/BglI fragment from pSPB3 and the annealed oligodeoxyribonucleotides, GGGCTGGCG-GTCC and GGACCGCCAGCCCCGC. The 1.7-kb NotI/PflMI fragment from the resulting plasmid was ligated with the 8.6-kb NotI/PflMI fragment from pSPB3 to obtain pSPB3S. The plasmid pSPB3S was digested with SrfI and SpeI, blunted with T 4 DNA polymerase, and ligated with the annealed oligonucleotides, GGGCTTAGCTGCG-GAGAAGAGTTCTGAGACGTGCACCGGTT and AACCGGTGCACGT-CTCAGAACTCTTCTCCGCAGCTAAGCCC, to yield pSPB3T⌬1. In this plasmid, the codon TTC for Tyr-1913 was replaced with the codon TAG for termination.
B3TCD-The 8.7-kb SrfI/SalI fragment excised from pSPB3S was ligated with the 2.2-kb ScaI/SalI fragment from pSPCDRS to obtain pSPB3TCD. In this plasmid, the codon CGG for Arg-1911 of the ␣ 1B subunit was replaced with the codon CAC for (His), and the segment encoding amino acid residues 1658 -2171 of the ␣ 1C subunit was substituted for amino acid residues 1912-2339 of the ␣ 1B subunit.
CDT⌬1-The plasmid pSPCDR was partially digested with AvrII, blunted with T 4 DNA polymerase, and circularized with T 4 DNA ligase to obtain pSPCDTD1. In this plasmid, the codon AGG and CCC for Arg-1980 and Pro-1981 of the ␣ 1C subunit was replaced with the codon AGC (Ser) and TAG for termination.
CDTB3-The 5.2-kb HindIII/ScaI and the 3.0-kb HindIII/SalI fragments excised from pSPCDRS were ligated with the 1.6-kb SrfI/SalI fragment from pSPB3 to obtain pSPCDTB3. In this plasmid, the segment encoding amino acid residues 1912-2339 of the ␣ 1B subunit was substituted for amino acid residues of 1658 -2171 of the ␣ 1C subunit, and the codon TAC for Tyr-1657 of the ␣ 1C subunit was replaced with the codon TGG (Trp).
CDLB3 and CDLB3TB3-To delete an internal SacI site, the plasmids pSPCDRS and pSPCDTB3 were partially digested with SacI, blunted, and circularized to produce pSPCDRSS and pSPCDTB3S. Another SacI site on pSPCDRSS was deleted by the same procedure. The resulting plasmid and the plasmid pSPCDTB3S were digested with SacI and StuI and blunted. The 9.5-kb SacI/StuI fragment from the former or the 8.9-kb SacI/StuI fragment from the latter was ligated with the 890-bp XhoI/ApaI fragment that was excised from pSPB3 and blunted with T 4 DNA polymerase, in order to yield pSPCDLB3 or pSPCDLB3TB3. In these plasmids, the segment encoding amino acid residues 242-537 of the ␣ 1B subunit were substituted for amino acid residues 318 -610 of the ␣ 1C subunit, and the codon CTC for Leu-242 of the ␣ 1B subunit and CAG for Gln-611 of the ␣ 1C subunit were replaced with the codon GTC for (Val) and GAG for (Glu), respectively. B3L⌬1-The 9.6-kb PmlI/PflMI fragment from pSPB3S was ligated with the 99-bp PmlI/HhaI and 610-bp KpnI/PflMI fragments excised from pSPB3 and the annealed oligonucleotides, CGAGAGAGAGCT-CAACGGGTAC and CCGTTGAGCTCTCTCTCGCG, to yield pSPB3L⌬1. In this plasmid, the segment encoding amino acid residues 366 -383 of the ␣ 1B subunit were deleted.
Subcloning and mutagenesis procedures were verified by restriction enzyme analysis and DNA sequencing.
In order to unmask the effect of endogenous G␣ (16), a deoxyoligonucleotide 20-mer (AGO) of the following sequence was used in antisense experiments, CATGACTGCTCGGGGGGGGA. The AGO antisense oligonucleotide is complementary to nucleotides (Ϫ17 to 3) of the Xenopus G o ␣ mRNA (23). The endogenous Xenopus G o ␣ nucleotide sequence shows 40% identity with the corresponding nucleotide sequence of G o ␣ cRNA injected. This antisense oligonucleotide (0.1 g/l, 50 nl) was injected 12-16 h prior to electrophysiological measurements.
The oocytes were positioned in a recording chamber (1.0 ml in volume) and were perfused with a Ba 2ϩ solution containing 40 mM Ba 2ϩ , 50 mM Na ϩ , 2 mM K ϩ , and 5 mM HEPES (pH 7.5 with methanesulfonic acid). Membrane currents through the expressed Ca 2ϩ channels were measured with the two-microelectrode voltage-clamp method as described previously (15). Also, the membrane potential recorded by the potential electrode was monitored. The membrane was held at Ϫ80 or Ϫ100 mV, and step depolarizations were applied to activate the Ca 2ϩ channels. Microelectrodes were filled with 3 M KCl, and those showing resistances of 0.5-1.5 megohms were used.
We noticed slow tail currents upon repolarization as shown in Fig. 1. In these cases, the time resolution of clamping was within 4 ms and the potential error was within 3% of the command pulse, indicating no serious space-clamping problems in characterizing Ca 2ϩ channel currents.
Unless otherwise stated, statistical data were represented by the mean and S.E.

Functional Expression of the N-, P/Q-, and L-type Ca 2ϩ
Channels in Xenopus Oocytes-To establish a recombinant expression system, where current inhibition mediated by G-proteins can be reconstituted individually, HVA N-, P/Q-, and L-type Ca 2ϩ channels were co-expressed in Xenopus oocytes by injection of cRNAs for three (␣ 1 , ␣ 2 , and ␤ 1 ) Ca 2ϩ channel subunits and the ␦-opioid receptor (DOR). Their responses to the opioid peptide, Leu-enkephalin (Leu-EK), were examined by the two-microelectrode voltage-clamp technique. Fig. 1 illustrates inward membrane currents recorded from Xenopus oocytes that were injected with the N-type ␣ 1B (Fig. 1, A and B), P/Q-type ␣ 1A (Fig. 1, C and D), and L-type ␣ 1C (Fig.  1, E and F) cRNA in combination with ␣ 2 and ␤ 1a subunits and DOR. As shown by the current-voltage (I-V) relationships in Fig. 1, step depolarizations from a holding potential of Ϫ80 mV produced long lasting inward currents at potentials more positive than Ϫ30 mV for oocytes injected with ␣ 1B (Fig. 1B) and ␣ 1A subunits (Fig. 1D) and at potentials positive to Ϫ50 mV with ␣ 1C (Fig. 1F).
As shown in Fig. 1A, inward currents recorded from oocytes implanted with ␣ 1B , ␣ 2 , ␤ 1a and DOR showed a time-dependent inactivation and a sensitivity to 0.1 M -conotoxin GVIA (-CTx), an N-type Ca 2ϩ channel blocker. This current was not blocked by 0.3 M -agatoxin IVA (-Aga), a P/Q-type Ca 2ϩ channel blocker (n ϭ 3), nor 10 M nifedipine, a dihydropyridine (DHP)-derivative L-type Ca 2ϩ channel blocker (n ϭ 12). Application of Leu-EK (1 M) to the bathing solution inhibited inward current from N-type channels within seconds (Fig. 1B). The membranes were held at Ϫ80 mV and depolarized by a 250 (or 300)-ms test pulse from Ϫ80 mV to ϩ50 mV with 10 mV steps. Peak currents before (open circles) and during (filled circles) exposure to 1 M Leu-EK and after removal of Leu-EK (filled triangles) are plotted against the membrane potential of test pulses. Note that the peak current is inhibited prominently by Leu-EK either in B or D. Therefore, in the following experiments, the amplitude of peak currents was used as a measure of the response to Leu-EK. In practice, peak currents were measured before and after application of a receptor agonist, and the change was expressed as their ratio.
The inhibited current displayed "kinetic slowing" of the current activation as well as an overall reduction in peak current (24,25).
As shown in Fig. 1, C and D, oocytes expressing ␣ 1A , ␣ 2 , ␤ 1a , and DOR exhibited inward currents which were blocked by 0.3 M -Aga (Fig. 1C) but not by 0.3 M -CTx (n ϭ 3) nor 10 M nifedipine (n ϭ 9), consistent with previous findings. Application of Leu-EK to oocytes expressing ␣ 1A also displayed Ca 2ϩ channel modulation, similar to that observed with ␣ 1B . Since DOR translates a signal to downstream effectors through activation of G-proteins (26), it is conceivable that the Leu-EKinduced inhibition of ␣ 1B and ␣ 1A currents is mediated by endogenous oocyte G-proteins.
Following the injection of L-type ␣ 1C cRNA in combination with ␣ 2 , ␤ 1a , and DOR cRNAs, inward currents were observed (Fig. 1, E and F), which were sensitive to 10 M nifedipine ( . By contrast to the -CTx-sensitive N-type or -Aga-sensitive P/Q-type currents, these currents were not inhibited by Leu-EK (Fig. 1F).
Effects of G␣ and G␤␥ on the N-and P/Q-type Ca 2ϩ Channels-To determine which arm of the G-protein complex contributes to regulation of N-and P/Q-type Ca 2ϩ channels, either G i3 ␣ cRNA or G␤ 1 plus G␥ 2 cRNAs were injected into oocytes in combination with Ca 2ϩ channel ␣ 1 (␣ 1B or ␣ 1A ), ␣ 2 , and ␤ 1a subunits cRNAs and DOR cRNA.
As detailed in the companion paper (16), agonist-induced inhibition of N-type ␣ 1B currents (Figs. 2A and 3A, B3) and P/Q-type ␣ 1A currents (Fig. 4, A and C, B1) was further pronounced in oocytes injected with G i3 ␣ cRNA. By contrast, inhibition of ␣ 1B and ␣ 1A channels was not potentiated in oocytes co-expressed with G␤ 1 ␥ 2 . However, Ba 2ϩ currents recorded from oocytes expressed with Ca 2ϩ channel ␣ 1B , ␣ 2 , and ␤ 1a subunits, DOR and G␤ 1 ␥ 2 , were increased by a large conditioning depolarization to ϩ80 mV without receptor stimulation (Fig. 3A, B3, open bar; also see Fig. 2 in the companion paper). This may indicate that the exogenous G␤␥ can inhibit the N-type Ca 2ϩ channel by itself, therefore not requiring receptormediated activation of G-proteins. Prepulse facilitations were not prominent, but still significant, for the ␣ 1A channel when injected with G␤␥ (Fig. 4C, B1). Moreover, L-type ␣ 1C currents were never inhibited by the application of agonist nor facilitated by administration of a prepulse ( Figs. 2A and 3A, CD).
Combination of ␣ 1B , ␣ 2 , and ␤ 1a Subunits Is Required for the Inhibitory Regulations by G i3 ␣ and G␤␥-The ␣ 1 subunit of the Ca 2ϩ channel forms the channel pore (4). As a result of this, N-type Ca 2ϩ channel currents were not detectable without the injection of ␣ 1B subunit cRNA (n ϭ 13). However, when the ␣ 1B subunit was expressed without the ␣ 2 and ␤ 1a subunits, Leu-EK still produced channel inhibition and slowing of the ␣ 1B currents via G i3 ␣ (n ϭ 8). Moreover, the opioid-induced inhibition of ␣ 1B currents was larger in the absence of ␤ 1a subunit (n ϭ 15) and did not change without the ␣ 2 subunit (n ϭ 8) (27,28). In addition, the prepulse facilitation of ␣ 1B currents mediated via G␤ 1 ␥ 2 (see Fig. 3) was also present without the ␣ 2 and ␤ 1a subunits (n ϭ 5). These results suggest that both G␣ and G␤␥ can interact with the ␣ 1 subunit regardless of subunit composition and are able to produce channel modulation.
Effects of Prepulse on the Inhibitions of Mutant and Chimeric N-type Ca 2ϩ Channels via G i3 ␣ or G␤ 1 ␥ 2 -The experiments described above, in which the wild-type ␣ 1B and ␣ 1A channels were used, demonstrated that G␣ plays a significant role in G-protein-mediated inhibition of neuronal Ca 2ϩ channels. However, there is a possibility that G␣ exerts its effect indirectly upon Ca 2ϩ channels through G␤␥. To exclude this possibility, it was necessary to investigate further molecularly and structurally the dependence of G␣ on G␤␥ when interacting with Ca 2ϩ channel ␣ 1 subunits.
In order to clarify further the modulation sites on the ␣ 1B channel by G i3 ␣ and G␤ 1 ␥ 2 , responses to Leu-EK and a large prepulse were studied in oocytes implanted with mutant and chimeric Ca 2ϩ channel ␣ 1 , ␣ 2 , and ␤ 1a subunits, DOR, and G i3 ␣ (or G␤ 1 ␥ 2 ) (Fig. 3). For clarity, changes induced by application of a prepulse without Leu-EK, in oocytes expressing G␤ 1 ␥ 2 , and differences in the response to Leu-EK between oocytes with and without expression of G i3 ␣, are summarized in Fig. 3 (⌬ Response). In the case of wild-type ␣ 1B channels, co-expressed with G␤ 1 ␥ 2 , a remarkable prepulse facilitation was observed in the absence of Leu-EK (Fig. 3A, B3, open bar), whereas the ␣ 1B chimera, B3LCD (having a loop 1 region derived from ␣ 1C ), when co-expressed with G␤ 1 ␥ 2 , displayed a complete loss of prepulse facilitation (Fig. 3A, B3LCD, open bar). By contrast, . Deletion or replacement of the C terminus and/or substitution of the intracellular loop between segment I and II (loop 1) were carried out in the two types of ␣ 1 subunit of Ca 2ϩ channels. Nomenclature is as follows: B3, wild-type ␣ 1B ; CD, wild-type ␣ 1C ; T or Tail, C terminus; L or L1, loop 1; ⌬, deletion. Functional expression of Ca 2ϩ channels is also indicated (ϩ and Ϫ). A, right, responsiveness of mutant and chimeric ␣ 1 channels to 1 M Leu-EK in oocytes implanted with DOR, ␣ 1 , ␣ 2 , and ␤ 1a in combination with G i3 ␣ (filled boxes), G␤ 1 ␥ 2 (hatched boxes), or no exogenous G-protein (open boxes). Positive and negative responses represent inhibition and facilitation of channels, respectively. In oocytes from which endogenous Ca 2ϩ currents were recorded, ␣ 1 subunit was not co-expressed (No exogenous Ca 2ϩ channel). In other oocytes, wild-type, mutant, and chimeric ␣ 1 subunits as indicated for each on the left side were co-expressed. The antisense oligonucleotide, AGO, was used. The responses to Leu-EK were measured (see Fig. 1 legend) and expressed as ratios of inhibition. The number of oocytes examined for each data are 4 -68. B, representative current traces for the mutant ␣ 1B (B3T⌬1) and ␣ 1C (CDT⌬1) channels and the chimeric ␣ 1B /␣ 1C (B3LCD, B3LCDT⌬1, CDTB3, CDLB3, and CDLB3TB3) channels in oocytes coexpressed with DOR, G i3 ␣, and Ca 2ϩ channel ␣ 2 and ␤ 1a subunits. The pulse protocol was identical to that in Fig. 1 for ␣ 1B channels. Concentrations of Leu-EK, -CTx, and nifedipine used were 1, 0.3, and 10 M, respectively. The antisense oligonucleotide, AGO, was used. the ␣ 1C chimera, CDLB3 (having a loop 1 region derived from ␣ 1B ), restored the prepulse facilitation when G␤ 1 ␥ 2 was coexpressed ( Fig. 3A, CDLB3, open bar). Moreover, deletion of the C terminus of ␣ 1B enhanced the prepulse facilitation in oocytes co-expressed with G␤ 1 ␥ 2 (Fig. 3A, B3T⌬1, open bar).
When G i3 ␣, instead of G␤ 1 ␥ 2 , was co-expressed, the agonistinduced inhibition of Ca 2ϩ currents was strengthened in wildtype ␣ 1B channels as compared with control oocytes, in which no exogenous G-proteins were co-expressed (Fig. 3A, B3, filled  bar). This large inhibition was abolished by applying a large conditioning prepulse (filled circle). In chimeric ␣ 1B channels, B3LCD, such a potentiation of current inhibition by G i3 ␣ was still detectable (Fig. 3A, B3LCD, filled bar) and almost entirely relieved by applying a prepulse (filled circle). Furthermore, deletion of the C terminus of B3LCD abolished the sensitivity to the agonist-induced inhibition with G i3 ␣ (Fig. 3A,  B3LCDT⌬1, filled bar). By contrast, the ␣ 1C chimera, CDTB3, having a C terminus derived from ␣ 1B , acquired sensitivity to the agonist-induced current inhibition with G i3 ␣ (Fig. 3A,  CDTB3, filled bar), but the prepulse procedure failed to influence this inhibition (filled circle). The deletion alone of the C terminus of ␣ 1B channel did not affect the channel responsiveness to G i3 ␣ (Fig. 3A, B3T⌬1, filled bar).
To gain a clearer understanding of contributions of loop 1 in more detail, Ba 2ϩ currents through mutant ␣ 1B channels with four kinds of loop 1 deletions were studied (Fig. 3B). In the mutant channel, B3L⌬2, with a deletion of amino acid residues 384 -420 of ␣ 1B (Fig. 5A, L⌬2), the prepulse facilitation in oocytes co-expressed with G␤ 1 ␥ 2 was diminished (Fig. 3B, B3L⌬2, open bar). However, currents through the mutant channel, B3L⌬3, with a deletion of amino acid residues 421-470 of ␣ 1B (Fig. 5A, L⌬3), were facilitated by a prepulse when G␤ 1 ␥ 2 was co-expressed (Fig. 3B, B3L⌬3, open bar). In both mutant channels, the potentiation by G i3 ␣ of Leu-EK-induced inhibition of currents was observed (filled bar). These characteristics of B3L⌬2 indicate that deletion of loop 1, which nearly abolished interaction of the ␣ 1B subunit with G␤␥, did not impair interactions with G␣. On the other hand, we could not detect expression (n ϭ 6) of the mutants, B3L⌬1 and B3L⌬4, in which either a part of the loop 1 of ␣ 1B (amino acid residues 366 -383, see Fig. 5A, L⌬1) or a part of the loop 1 of ␣ 1B that combines the regions covered by L⌬2 and L⌬3 (amino acid residues 384 -470, see Fig. 5A) were deleted. Moreover, currents through the mutant channel, B3L⌬2, were not detectable in the absence of Ca 2ϩ channel ␤ subunit expression (n ϭ 5). Because B3L⌬2 was devoid of the segment corresponding to the major binding site for the ␤ subunit (31), this indicates that there may be another interaction site on the ␣ 1B channel for ␤ subunits (32).
Interaction Site on the P/Q-type Ca 2ϩ Channel for G-protein-In order to determine the interaction site on the P/Q-type Ca 2ϩ channel for G␣ and G␤␥, procedures similar to those for ␣ 1B channels (Figs. 2 and 3) were applied to ␣ 1A channels (Fig.  4). In oocytes co-expressed with G i3 ␣ or G␤ 1 ␥ 2 together with FIG. 3. The C terminus and the loop 1 of ␣ 1B channels determining the interactions with G i3 ␣ and G␤ 1 ␥ 2 . A and B, upper, comparisons of wild-type, mutant, and chimeric ␣ 1B and ␣ 1C channels with respect to the Leu-EK-induced inhibition via G i3 ␣ and to the prepulse facilitation via G␤ 1 ␥ 2 . The responses of 7 different channel types, as indicated with schemes, to 1 M Leu-EK (horizontal bars), prepulse (open circles), or both (filled circles) were measured in oocytes co-expressing with DOR, ␣ 2 , and ␤ 1a in combination with G-protein subunit as indicated. The pulse protocols were as follows: a 200-ms test pulse was applied to ϩ10 mV from a holding potential of Ϫ100 mV, which was preceded, if necessary, by a depolarizing prepulse (30 ms in duration) to ϩ80 mV and then by a 20-ms repolarization to Ϫ100 mV. The antisense oligonucleotide, AGO, was used. The number of oocytes examined for each data are 4 -23 in A and 4 -10 in B. The deletion sites for these mutant ␣ 1B channels in B are represented schematically in Fig. 5A. A and B, lower, Leu-EK-induced inhibition as mediated by G i3 ␣ (filled bars) and prepulse-induced facilitation as mediated by G␤ 1 ␥ 2 (open bars). Differences in the response to Leu-EK between oocytes with and without expression of G i3 ␣ and changes induced by prepulse without Leu-EK in oocytes expressing G␤ 1 ␥ 2 , as shown in upper, are represented as ⌬ Response.
Next, the mutant (B1T⌬2) and chimeric (CDTB1) ␣ 1A channels were further characterized using a double-pulse protocol. Fig. 4C illustrates responses to application of a prepulse and Leu-EK by these channels and also demonstrates changes induced by prepulse without Leu-EK in oocytes expressing G␤ 1 ␥ 2 as well as differences in the response to Leu-EK between oocytes with and without expression of G i3 ␣. The potentiation by G i3 ␣ of Leu-EK-induced inhibition observed in the wild-type  Fig. 2A, except that B1 denotes wild-type ␣ 1A . Functional expression of Ca 2ϩ channels is also indicated (ϩ and Ϫ). A, right, responsiveness of mutant ␣ 1A and chimeric ␣ 1A /␣ 1C channels to 1 M Leu-EK in oocytes implanted with DOR, ␣ 1 , ␣ 2 , and ␤ 1a in combination with G i3 ␣ (filled boxes), G␤ 1 ␥ 2 (hatched boxes), or no exogenous G-protein (open boxes). In oocytes from which endogenous Ca 2ϩ currents were recorded, ␣ 1 subunit was not co-expressed (No exogenous Ca 2ϩ channel). In other oocytes, wild-type, mutant, and chimeric ␣ 1 subunits as indicated for each on the left side were co-expressed. The antisense oligonucleotide, AGO, was used. The responses to Leu-EK were measured (see Fig. 2A) and expressed as ratios. The number of oocytes examined for each data are 4 -18. B, representative current traces for the mutant (B1T⌬2) and chimeric (CDTB1) ␣ 1A channels in oocytes implanted with DOR, G i3 ␣, ␣ 2 , and ␤ 1a in combination with each mutant or chimeric ␣ 1 . The pulse protocol was identical to that as in Fig. 1 for the ␣ 1A channel. Concentrations of Leu-EK and nifedipine used were 1 and 10 M, respectively. The antisense oligonucleotide, AGO, was used. C, comparisons of wildtype, mutant, and chimeric ␣ 1A channels with respect to the Leu-EK-induced inhibition via G i3 ␣ and to the prepulse facilitation via G␤ 1 ␥ 2 . The responses of three different channel types, as indicated, to 1 M Leu-EK (horizontal bars), prepulse (open circles), or both (filled circles) were measured (see Fig. 3A) and expressed as ratios. The antisense oligonucleotide, AGO, was used. The number of oocytes examined for each data are 4 -18. ␣ 1A (Fig. 4C, B1, filled bar) almost disappeared in the mutant ␣ 1A , B1T⌬2, having the deletion of C terminus (Fig. 4C, B1T⌬2, filled bar). By contrast, the ␣ 1C /␣ 1A chimera, CDTB1 (having the C terminus derived from ␣ 1A ), conferred Leu-EK sensitivity via G i3 ␣ (Fig. 4C, CDTB1, filled bar). In both the wild-type ␣ 1A and the chimera CDTB1, the prepulse did not abolish the potentiation of Leu-EK-induced inhibition via G i3 ␣ (filled circles). In addition, when G␤ 1 ␥ 2 was co-expressed instead of G i3 ␣, a small facilitation by prepulse was observed in the wild-type ␣ 1A (Fig. 4C). In the mutant (B1T⌬2) and chimera (CDTB1) channels, prepulse facilitation was not detected. DISCUSSION In the present study, the -CTx-sensitive N-type (␣ 1B ) and -Aga-sensitive P/Q-type (␣ 1A ) Ca 2ϩ channels were functionally expressed in Xenopus oocytes, an in vivo expression system. As described, we found that G i3 ␣ co-expressed in oocytes mediated receptor agonist-induced inhibition of N-type ␣ 1B and P/Q-type ␣ 1A channels. On the other hand, a depolarizing prepulse relieved current inhibition caused by the G␤ 1 ␥ 2 complex, and the facilitatory effects were more pronounced in ␣ 1B than in ␣ 1A . Because responsiveness of the ␣ 1B and ␣ 1A channels to the inhibition mediated by G i3 ␣ and G␤ 1 ␥ 2 was maintained even in the absence of the Ca 2ϩ channel auxiliary subunits ␣ 2 and ␤ 1 , the ␣ 1 subunit should bear the interaction sites for both the G␣ subunit and the G␤␥ dimer. Finally, we defined loop 1 of ␣ 1B as an interaction site for G␤␥ and the C termini of ␣ 1B and ␣ 1A for G␣, based on the responses of mutant and chimeric channels to G␣ and G␤␥.
The Native Type ␣ 1B , ␣ 1A , and ␣ 1C Channels Expressed in Xenopus Oocytes-The electrophysiological and pharmacological properties of the ␣ 1B , ␣ 1A , and ␣ 1C channels determined were identical to those of the N-, P/Q-, and L-type Ca 2ϩ channels described previously (18,20,33). This indicates that ␣ 1B (N-type), ␣ 1A (P/Q-type), and ␣ 1C (L-type) Ca 2ϩ channels were functionally expressed with the Ca 2ϩ channel ␣ 2 and ␤ 1 subunits in Xenopus oocytes. When DOR was further co-expressed, ␣ 1B and ␣ 1A channel currents, but not ␣ 1C channel currents, were inhibited within seconds when stimulated by Leu-EK. It is likely that agonist-induced inhibitions of ␣ 1B and ␣ 1A channels are mediated by endogenous oocyte G-proteins that are coupled to the receptor, because the inhibitions of ␣ 1B and ␣ 1A channels were reduced when the antisense oligonucleotide AGO against Xenopus G o ␣ was injected (16).
The Loop 1 of ␣ 1B Channel as an Interaction Site for G␤␥-When G␤ 1 ␥ 2 was co-expressed, the Leu-EK-induced inhibition was not potentiated in either ␣ 1B or ␣ 1A channels. In the case of N-type ␣ 1B , however, a depolarizing prepulse to ϩ80 mV facilitated the currents in the absence of the receptor agonist, suggesting that the exogenous G␤␥ inhibits the ␣ 1B channel by itself (8,9). Thus, the difference between the current traces before and after the prepulse should correspond to an ␣ 1B current component that is mainly inhibited by exogenous G␤ 1 ␥ 2 .
A diminished response to prepulse application was found for an ␣ 1B channel (B3LCD) chimerized with the ␣ 1C loop 1 and a mutant ␣ 1B (B3L⌬2) with a partial deletion of the loop 1 (including the binding site for Ca 2ϩ channel ␤ subunit) (see Fig.  5A), when G␤ 1 ␥ 2 was co-expressed. On the other hand, an ␣ 1C channel (CDLB3) chimerized with the ␣ 1B loop 1 conferred properties of facilitation by prepulse on this channel when expressed with G␤␥. Taken together, these findings indicate that the loop 1 plays an essential role for the interaction of the ␣ 1B channel with G␤␥. This is consistent with recent evidence that has shown the direct binding of G␤␥ to the loop 1 of ␣ 1B (10, 11, 16). The prepulse facilitation could not be abolished completely by deleting the loop 1 of ␣ 1B (B3L⌬2), whereas it FIG. 5. Schematic representation of the sites on ␣ 1B subunit for making the deletion mutants. A, positions of the deletion (L⌬1, L⌬2, L⌬3, and T⌬1), the loop 1, and the C terminus are indicated by the number of the amino acid residues for ␣ 1B subunit (33) and ␣ 1A (BI-1 ␣ 1 ) subunit (20) in parentheses. The deletion sites are indicated by the crossing bars, and the cytoplasmic side below the horizontal lines. The asterisk denotes the binding site for Ca 2ϩ channel ␤ subunit (31), and the filled circle denotes the phosphorylation sites for protein kinase C (10). B, mutations introduced into the ␣ 1B and ␣ 1A subunits. The numbers of amino acid residues deleted are indicated. could be totally abolished by replacing the whole loop 1 of ␣ 1B with that of ␣ 1C (B3LCD). As described in the companion paper (16), a loop 1 peptide (PL1) corresponding to amino acid residues 366 -384 of ␣ 1B almost abolished the prepulse facilitation. This means that the interaction site on the ␣ 1B loop 1 for G␤␥ might cover both the deletion site in B3L⌬2 and the region determining PL1 (10,11). The flanking region of the ␣ 1B loop 1 (amino acid residues 421-470) including the phosphorylation sites for protein kinase C (10) was not critical for the interaction between the channel and G␤␥ as examined by B3L⌬3, a mutant with a partial deletion of the ␣ 1B loop 1 (see Fig. 5A).
On the other hand, a mutant ␣ 1B channel devoid of the normal C terminus (B3T⌬1) never failed to induce prepulse facilitation via G␤ 1 ␥ 2 . Inversely, the ␣ 1C channel chimerized with the ␣ 1B C terminus (CDTB3) did not facilitate the current by prepulse in the presence of G␤ 1 ␥ 2 . These findings indicate that the C terminus of ␣ 1B channel is not involved in the prepulse facilitation via G␤ 1 ␥ 2 .
In the case of the P/Q-type ␣ 1A channel, prepulse facilitation was not as prominent as observed in the ␣ 1B channel when G␤ 1 ␥ 2 was co-expressed (16). Nonetheless, the prepulse facilitation via G␤ 1 ␥ 2 appeared to be abolished in a mutant ␣ 1A channel (B1T⌬2) that contained a deletion of the C terminus. This is in contrast to a similar ␣ 1B mutant (B3T⌬1), in which the prepulse facilitation via G␤ 1 ␥ 2 was rather enhanced. However, an ␣ 1C channel (CDTB1) chimerized with the ␣ 1A C terminus did not confer facilitation by prepulse. These results suggest that the contribution of the C terminus to channel inhibition via G␤ 1 ␥ 2 is small, although the possibility of interaction of the ␣ 1A C terminus with G␤␥ (34) has not been excluded.
The C Termini of ␣ 1B and ␣ 1A Channels as an Interaction Site for G␣-Receptor stimulation by agonist is known to catalyze activation of G␣ and lead to dissociation of the G␣␤␥ heterotrimer (1). In fact, application of a prepulse did not facilitate N-type ␣ 1B channels when co-expressed with G i3 ␣ and DOR, unless DOR was stimulated by Leu-EK. The potentiating action of G i3 ␣ on the agonist-induced inhibition of ␣ 1B channels was abolished by application of a large conditioning prepulse. This suggests that exogenous G␣, unlike G␤␥, does not influence the ␣ 1B channel by itself and stays in its inactive form. It appears, therefore, that potentiation of agonist-induced inhibition via exogenous G␣ results from the interaction of the channel with activated exogenous G␣ and, probably, with an endogenous G␤␥ dissociated from the G␣. This idea is further evidenced by the observation for mutant (B3T⌬1) and chimera (B3TCD) ␣ 1B channels, in which loss of the C terminus (a possible interaction site for G i3 ␣, see below) did not affect the potentiation of inhibition via G i3 ␣. However, a further loss of the loop 1, an interaction site for G␤ 1 ␥ 2 , of the mutant B3T⌬1 ␣ 1B channel (B3LCDT⌬1) eliminated the potentiation of inhibition via G i3 ␣ as well as the prepulse facilitation via G␤ 1 ␥ 2 . Moreover, an ␣ 1B channel chimerized with the ␣ 1C loop 1 (B3LCD), in which the interaction site exclusively for G␤ 1 ␥ 2 was lost, retained the potentiation of inhibition via G i3 ␣ but no longer the prepulse facilitation via G␤ 1 ␥ 2 . These results indicate that there are two distinct interaction sites, namely loop 1 and the C terminus, for G␣ and G␤␥ on N-type Ca 2ϩ channels and that the two sites regulate the channel activity independently when they receive inhibitory signals from G␣ and G␤␥. This independence of loop 1 and the C terminus in the ␣ 1B channel modulation is supported by the observation that the G i3 ␣-dependent potentiation of inhibition was not affected by single application of the loop 1 peptide (PL1) or a C-terminal peptide (PB3T4) inside the oocyte but abolished by simultaneous application of both of them (16).
As in the case of B3LCD described above, mutant ␣ 1B channels devoid of the normal loop 1 (B3L⌬2 and B3L⌬3) never failed to potentiate the current inhibition in response to Leu-EK via G i3 ␣. Inversely, the ␣ 1C channel chimerized with the ␣ 1B loop 1 (CDLB3) did not potentiate the inhibition via G i3 ␣. These findings indicate that there is an interaction site for G␣ outside the loop 1 of ␣ 1B channel (14).
The ␣ 1B channel chimerized with the ␣ 1C loop 1 (B3LCD), which was devoid of the interaction site for G␤␥, did not impair G i3 ␣-dependent potentiation in the inhibitory response to Leu-EK unless its C terminus was deleted (B3LCDT⌬1). Contrary, an ␣ 1C channel chimerized with the ␣ 1B C terminus (CDTB3) potentiated the inhibition via G i3 ␣. Together, the results indicate that the C-terminal segment of ␣ 1B is essential for the interaction with G␣.
In the case of P/Q-type ␣ 1A channel, Leu-EK-induced inhibition was markedly potentiated when co-expressed with G i3 ␣, similar to the ␣ 1B channel. In contrast to the ␣ 1B channel, prepulse failed to abolish potentiation of inhibition via G i3 ␣. Furthermore, G i3 ␣-dependent potentiation of inhibition of the ␣ 1A channel was impaired only when the C terminus was deleted (B1T⌬2). This is probably due to a minor contribution of G␤␥ to the potentiation of inhibition by agonist, despite the direct binding of G␤␥ to the loop 1 of ␣ 1A (11,16). In fact, the potentiation of inhibition via G␣ was abolished by intracellular application of a C-terminal peptide (PPQT1) alone (16). Moreover, an ␣ 1C channel chimerized with the ␣ 1A C terminus (CDTB1) potentiated the Leu-EK-induced current inhibition via G i3 ␣. These findings indicate that, as in the case of ␣ 1B , the C terminus of ␣ 1A is also essential for the interaction with G␣ and that, in contrast to ␣ 1B , the loop 1 of ␣ 1A appears not to be essential in the channel regulation by G-proteins.
In conclusion, N-and P/Q-type Ca 2ϩ channels are differentially regulated by G␣ and G␤␥ in a way that prepulse preferentially facilitates N-type and that the C terminus and the loop 1 of N-type are equally involved in agonist-induced inhibition, whereas the C terminus of P/Q-type is mainly involved. Further studies using an in vitro binding assay will be necessary to determine the direct interaction of G␣ and G␤␥ with N-type ␣ 1B and P/Q-type ␣ 1A Ca 2ϩ channels.