Multiple Modulation Pathways of Calcium Channel Activity by a β Subunit

In order to study the precise mechanisms of α1 subunit modulation by an auxiliary β subunit of voltage-dependent calcium channels, a recombinant β3 subunit fusion protein was produced and introduced into oocytes that express the human α1C subunit. Injection of the β3 subunit protein rapidly modulated the current kinetics and voltage dependence of activation, whereas massive augmentation of peak current amplitude occurred over a longer time scale. Consistent with the latter, a severalfold increase in the amount of the α1C subunit in the plasma membrane was detected by quantitative confocal laser-scanning microscopy after β3subunit injection. Pretreatment of oocytes with bafilomycin A1, a vacuolar type H+-ATPase inhibitor, abolished the increase of the α1C subunit in the plasma membrane, attenuated current increase, but did not affect the modulation of current kinetics and voltage dependence by the β3 subunit. These results provide clear evidence that the β subunit modifies the calcium channel complex in a binary fashion; one is an allosteric modulation of the α1 subunit function and the other is a chaperoning of the α1 subunit to the plasma membrane.

complex have been extensively studied using coexpression of cDNAs in heterologous systems. It is well established that coexpression of the ␣ 1 subunit with a ␤ subunit results in an increase of peak current density (5), acceleration of activation and inactivation kinetics, a leftward shift of the current-voltage relationship, and increased dihydropyridine (DHP) 1 binding activity (6 -12). However, there have been inconsistencies in the reported mechanism(s) by which these effects occur. Varadi et al. (6) reported a 10-fold increase in the number of DHPbinding sites by coexpression of the ␣ 1 subunit with the ␤ subunit, suggesting an increase in available channels within the plasma membrane. In contrast, an increase in current amplitude without affecting charge movement by ␤ subunit coexpression was shown by Neely et al. (8). These same authors later reported two modes of activation of the ␣ 1C subunit (13). Coexpression of a ␤ subunit potentiated current by an increase of the fast-activating component, an acceleration of the slow component, and a larger proportion of long openings. An increase in DHP binding and current density without a change in the amount of ␣ 1 subunit protein in the plasma membrane was reported by Nishimura et al. (10). These reports suggest that the ␤ subunit modulates channel properties by "assisting" the ␣ 1 subunit in establishing a proper conformation suitable for a functional Ca 2ϩ channel, rather than affecting expression, trafficking, or stability of the ␣ 1 subunit (reviewed by Catterall (14)).
By using immunocytochemical methods, Chien et al. (15) showed that the ␤ 2a subunit acts as a chaperone-like molecule to facilitate membrane targeting of the ␣ 1C subunit, without affecting the total amount of expressed ␣ 1C subunit in human embryonic kidney cells. Coexpression of the ␣ 1C subunit with ␤ 2a resulted in a marked increase in localization of the ␣ 1C subunit to the plasma membrane. These observations were confirmed by Brice et al. (16) who coexpressed the ␣ 1A with ␣ 2 /␦ and several different ␤ subunits. By using Xenopus oocytes as an expression system, Shistik et al. (17) reported opposite results in that no change in the amount of ␣ 1 subunit protein was found in the plasma membrane when coexpressed with a ␤ subunit. Interestingly, an increase in charge movement has been reported when ␣ 1C and ␤ subunits were coexpressed in a mammalian cell system (18,19). These discrepancies have been attributed to differences in expression systems employed (20). Taken together, the mechanisms by which the ␤ subunit modulates calcium channel activity remain unclear. The limitation of previous studies are that they were all done using coexpression of the ␣ 1 and ␤ subunits, which makes it difficult to answer the question as to which level of the biosynthesis of the channel complex is the ␤ subunit working, i.e. translation, trafficking, and/or direct binding to the membrane-incorporated ␣ 1 subunit.
We address these questions by injecting a recombinant ␤ subunit protein into Xenopus oocytes expressing the ␣ 1C subunit and observing the time course of modulation. We found that changes in voltage dependence of activation and kinetics occurred at an earlier stage in the time course, compared with increases in current amplitude which required a substantially longer time to reach plateau. This suggests that the effects are functionally uncoupled and occur by distinct mechanisms. By using confocal laser-scanning microscopy (CLSM), we found that the amount of the ␣ 1 subunit in the plasma membrane was substantially increased by the ␤ subunit. Furthermore, pretreatment of oocytes with bafilomycin A 1 , a vacuolar type H ϩ -ATPase (V-ATPase) inhibitor (21), inhibited the ␤ subunit effect on current amplitude, abolished the increase in channel amount in the plasma membrane, but did not influence the ␤ effects on voltage dependence and current kinetics.

Bacterial Production and Purification of ␤ 3 Subunit Fusion Protein-
The human calcium channel ␤ 3 cDNA clone (22) was subcloned into pBluescript SK(ϩ) between the HindIII and BamHI sites. This plasmid was cleaved with HindIII, and the protruding end was filled in with T4 DNA polymerase and then cut with BamHI to liberate the ␤ 3 -coding fragment. The pET15(b) vector (Novagen) was cleaved with NdeI, filled in with T4 DNA polymerase, and then cut with BamHI. Finally, the ␤ 3 fragment was ligated into the blunt end/BamHI sites. This strategy has generated an in-frame cloning of the ␤ 3 protein with the His 6 -tag sequence and resulted in a fusion product that had the following peptide fused to the N-terminal sequence of the human ␤ 3 protein, MGSSHH-HHHHSSGLVPRGSHKLDP. The sequence of the construct was verified by sequencing through the junction regions.
Escherichia coli BL21(DE3) (Novagen) cells were transformed with the above construct and used for mass production of the His 6 -tagged ␤ 3 fusion protein. The cells were cultured in 500 ml of LB medium in the presence of 50 g/ml ampicillin and grown at 37°C until an A 600 of 0.8 was reached. The production of fusion protein was induced by 0.4 mM isopropyl ␤-D-thiogalactopyranoside (IPTG) for various times at temperatures between 23 and 37°C (we found induction being optimal at 23°C for 8 h). The cells were pelleted by low speed centrifugation, washed with buffer A (50 mM Tris-HCl, 2 mM EDTA, pH 8.0), resuspended in 10 ml of binding buffer (20 mM Tris-HCl, 1 mM imidazole, 500 mM NaCl, pH 7.9), and disrupted by sonication. The lysate was further fractionated by differential centrifugation. The resulting pellet (1,000 ϫ g, P 12 ) contained nondisrupted cells and inclusion bodies. The supernatant was further centrifuged at 10,000 ϫ g, resulting in a pellet (P 3 ) and a clear supernatant fraction (S), which was purified on a His-bound column (Ni 2ϩ -agarose, Novagen). The column was washed with 60 mM imidazole, pH 7.9, and finally the specifically bound protein was eluted with 1 M imidazole and 500 mM NaCl, pH 7.9. We observed that the eluted protein product aggregated when stored frozen in the elution buffer. To prevent aggregation we dialyzed the eluate gradually against 100 mM EDTA⅐Tris, pH 7.6, and stored at Ϫ80°C in small aliquots. The protein fractions and purification products were analyzed by SDS-PAGE and stained with Coomassie Brilliant Blue. Protein concentrations were determined with the Pierce protein assay kit.
In Vitro Translation of the h␤ 3 Subunit and Binding of GST Fusion Proteins-The I-II, II-III, and III-IV intracellular loops were extracted from the human heart calcium channel cDNA (23) via PCR. By using PCR we performed site-directed mutagenesis, placing unique BamHI and XhoI restriction sites at both the 5Ј and 3Ј ends of these fragments. The PCR products containing the I-II loop (nucleotides 1089 -1499), the II-III loop (nucleotides 2133-2633), and the III-IV loop (nucleotides 3402-3574) were subcloned into the pGEX-4T-1 vector (Amersham Pharmacia Biotech), which is under the control of the tac promoter and contains the glutathione S-transferase (GST) tag. The resulting subclones, pGEXI-II, pGEXII-II, and pGEXIII-IV, were sequenced, and analysis confirmed the presence of the intracellular loops and the absence of additional mutations. The pGEX-4T-1, pGEXI-II, pGEXII-III, and pGEXIII-IV plasmids were transformed into Novablue (DE3) cells (Novagen). Production and purification of GST fusion proteins were done according to the manufacturer's protocol.
The pET15(b) plasmid carrying the h␤ 3 subunit cDNA was used to synthesize h␤ 3 using the coupled in vitro transcription/translation reticulocyte lysate system (Promega) in the presence of 35 S-labeled methionine (Amersham Pharmacia Biotech), and a protease inhibitor mixture (1 g/ml chymostatin, 1 g/ml aprotinin, 1 g/ml leupeptin, 1 g/ml antipain, 1 g/ml pepstatin A, 1 g/ml phenylmethylsulfonyl fluoride) was used to minimize proteolysis.
Screening of fusion proteins with the radioactively labeled h␤ 3 subunit was performed three times, as described previously (24,25). Bacterial colonies containing the plasmids were grown overnight at 37°C. A 1:10 dilution was made in fresh media and grown at 37°C for ϳ2-4 h, until an A 600 of 1.0 was reached. The bacterial cultures were then induced with 1 mM IPTG for 4 h at 23°C. Cultures were sedimented by centrifugation and resuspended in ice-cold 1ϫ PBS, 1% Triton X-100 buffer and disrupted by sonication for ϳ20 s. After sedimentation of the lysate by centrifugation at 4°C, the soluble fractions containing the GST fusion proteins were adsorbed to glutathione-bound Sepharose (1 volume lysate, 0.1 volume of 50% (v/v) slurry of glutathione-Sepharose 4B (Amersham Pharmacia Biotech)) and incubated for 5 min at room temperature. The slurry was centrifuged at 4°C and washed 3 times with 1ϫ PBS, 1% Triton X-100 and resuspended in an equal volume of ice-cold 1ϫ PBS, 1% Triton X-100, and 35 S-labeled h␤ 3 subunit was added in vitro. This mixture was incubated for 30 min at room temperature. The slurry was centrifuged and washed 4 times with 1ϫ PBS, 1% Triton X-100 and eluted with 10 mM glutathione, 50 mM Tris-HCl, pH 8.0 (Amersham Pharmacia Biotech). After elution, the slurry was spun for 5 min at 4°C, and the supernatant was removed and added to an equal volume of 2ϫ SDS sample buffer. One-half of the eluate was run on a 5-14% gradient SDS-PAGE gel (Bio-Rad) and stained with Coomassie Blue for protein detection, and the remainder of the eluate was run on another 5-14% gradient SDS-PAGE gel. This gel was fixed, treated with Amplify (Amersham Pharmacia Biotech) for fluorography for 30 min, dried, and exposed to Hyperfilm MP (Amersham Pharmacia Biotech).
Electrophysiology-Xenopus laevis oocytes (stage V-VI) were prepared as described previously (26). Capped cRNA was synthesized from XbaI-linearized human heart L-type calcium channel ␣ 1 subunit DNA template (23) and injected into the oocytes (50 nl, 0.2 g/l). After 4 -5 days of incubation at 19°C in the solution containing (in mM) 96 NaCl, 2 KCl, 1 MgCl 2 , 1.8 CaCl 2 , 5 HEPES, 2.5 sodium pyruvate, 0.5 theophylline, pH 7.5, supplemented with 100 units/ml penicillin and 100 g/ml streptomycin, oocytes were injected with 50 nl of either ␤ 3 subunit fusion protein solution (90 mM 1,2-bis(2-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid, 10 mM EDTA, 340 ng/l protein, pH 7.4, with Tris) or vehicle (same composition except protein). The final concentrations in the oocytes were 300 nM ␤ 3 subunit fusion protein, 4.5 mM 1,2-bis(2-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid, 0.5 mM EDTA, assuming the volume of oocytes was 1 l. We found this was a saturating amount of protein since in our preliminary experiments higher concentrations did not result in greater effects. In the experiments indicated, oocytes were preincubated in a solution containing 500 nM bafilomycin A 1 for 2-3 h prior to the injection of the protein. After injection, they were further incubated in the presence of bafilomycin A 1 until currents were measured. Bafilomycin A 1 was dissolved in dimethyl sulfoxide as a 1 mM stock solution. The final concentration of dimethyl sulfoxide was 0.05%, which had no effect on the currents or the ␤ effect.
The currents were recorded 1-4 h after injection with the ␤ protein or vehicle using the standard two-electrode voltage-clamp technique at room temperature (20 -21°C). One oocyte was used for only one recording. The recording medium contained (in mM): 40 Ba(OH) 2 , 50 Nmethyl-D-glucamine, 2 KOH, 5 HEPES, 0.5 niflumic acid, pH 7.4, with methanesulfonic acid. The voltage recording electrode and the current injection electrode were filled with 3 M KCl and had resistances of 0.5-1 M⍀. Currents were recorded using an Axoclamp-2A (Axon Instruments) amplifier. Pulses were applied from a holding potential of Ϫ80 mV every 15 s. Whole cell leakage and capacitative currents were digitally subtracted using the P/4 protocol. Data were filtered at 1 kHz, sampled at 1-10 kHz, and stored on a hard disk. Batches of oocytes that showed significant (Ͼ20 nA) endogenous Ca 2ϩ channel current were excluded from the analysis. The unpaired t test and analysis of variance were performed for the statistical analyses.
Immunofluorescence Studies-DNA sequence complementary to the YPYDVPDYA epitope sequence of influenza virus hemagglutinin (HA) (27), which is recognized by the 12CA5 mouse monoclonal antibody, was introduced to the N terminus immediately after the initiator methionine of human heart L-type calcium channel ␣ 1C subunit DNA by PCR-based mutagenesis. The cRNA of the epitope-tagged ␣ 1 subunit was transcribed and injected into X. laevis oocytes to express tagged channels. The functional characteristics of HA epitope-tagged ␣ 1 were tested in a separate set of experiments, and we found that the expressed currents and modulation of them by the ␤ 3 subunit were indistinguishable from those of wild type (data not shown). After 4 -5 days of incubation, oocytes were injected with the ␤ 3 subunit fusion protein as described above. Four hours after injection, oocytes were fixed with 3.7% formaldehyde, 0.25% glutaraldehyde, and permeabilized with 0.2% Triton X-100 at room temperature. Oocytes were then post-fixed in 100% methanol at Ϫ20°C overnight, incubated with 2% bovine serum albumin for 1 h at room temperature, followed by incubation with 10 g/ml fluorescein-conjugated anti-HA monoclonal antibody (Boehringer Mannheim) at 4°C overnight. Oocytes were then washed extensively in PBS and examined by CLSM.

Production and Purification of a Recombinant Human ␤ 3
Subunit Fusion Protein-The human ␤ 3 subunit fusion protein was purified from the soluble fraction of a bacterial culture induced with IPTG at 23°C for 8 h. The soluble fraction was then applied to a Ni 2ϩ -agarose column. Approximately 88.5% of total protein was in the flow-through. A smaller amount of total protein (ϳ8%) weakly bound to the resin and was washed out by 60 mM imidazole. Approximately 3.5% of the total protein bound firmly to the Ni 2ϩ -agarose and was eluted with 1 M imidazole and 0.5 M NaCl (Fig. 1A).
Analysis of the P 12 fraction showed large quantities of the 57-kDa ␤ 3 protein, which was insoluble and not suitable for purification. The P 3 and S fractions contained sizable amounts of the ␤ 3 protein. Ni 2ϩ -agarose affinity purification of the S fraction resulted in highly pure ␤ 3 protein (estimated purity is higher than 99.5%) with the predicted molecular weight (Fig.  1B, lane 7).
In Vitro Binding of the ␤ 3 Subunit to the I-II Intracellular Loop-To determine whether the recombinant His 6 -tagged ␤ 3 subunit was able to interact with the calcium channel intracellular I-II loop, as identified by Pragnell et al. (28), we created GST fusion proteins of the I-II, II-III, and III-IV loops of the ␣ 1C subunit and screened with an in vitro transcribed and translated, 35 S-labeled His 6 -tagged ␤ 3 subunit. The data from these experiments clearly show that the His 6 -tagged ␤ 3 subunit is able to interact with the I-II loop in a highly specific manner ( Fig. 2A). In addition, these results indicate that the ␤ 3 subunit did not interact with the control GST, II-III, and the III-IV loops. A sample of the eluate was run on an 4 -15% SDS-PAGE gel and stained with Coomassie Blue to determine whether the purified fusion protein could also be eluted from the glutathione-Sepharose. All four GST fusion proteins were purified and present in the binding assay eliminating the possibility of nonspecific interaction with other protein components (Fig.  2B). These experiments confirm that the His 6 -tagged ␤ 3 subunit maintains its ability to interact with the I-II loop of the ␣ 1C subunit in vitro. Electrophysiological Properties of Expressed Human ␣ 1C Subunit in Xenopus Oocytes and Time-dependent Effects of Injected Human ␤ 3 Subunit Fusion Protein-Oocytes were injected with human ␣ 1C subunit cRNA and incubated for 4 -5 days. The control (␣ 1C subunit alone) Ca 2ϩ channels expressed in oocytes showed peak barium currents of 149 Ϯ 14 nA (n ϭ 14) when depolarized from a holding potential of Ϫ80 mV to test potentials between Ϫ30 and ϩ60 mV. The control current exhibited a slow activation and very little inactivation (Fig.  3A), in agreement with previous studies (2,9). The threshold potential for current activation was found between Ϫ20 and Ϫ10 mV, and the current peaked at ϩ40 mV. After injection of the ␤ 3 subunit fusion protein, an increase in current amplitude, a change in the voltage dependence of activation, and changes in activation and inactivation kinetics occurred in a time-dependent manner (Fig. 3A). After injection of the ␤ 3 subunit fusion protein the activation threshold shifted to hyperpolarizing potentials between Ϫ30 and Ϫ20 mV, and the current peaked between ϩ20 and ϩ30 mV. tude and the Influence of Bafilomycin A 1 Treatment- Fig. 4 depicts the average time course of peak Ba 2ϩ current amplitude enhancement after injection of the ␤ 3 subunit fusion protein. We observed an increase in peak current amplitude, 2.3fold at 1 h and 2.9-fold at 2 h after injection of the ␤ 3 subunit. However, more than half of the increase occurred after 2 h, reaching a plateau at 3 h (6.5-and 6.6-fold for 3 and 4 h after injection, respectively). The effect of the ␤ subunit on the current amplitude is clearly slower than the effect on the voltage dependence of activation (cf. Fig. 5). The time of the halfmaximal effect for the increase in current amplitude was between 2 and 3 h. Injection of the vehicle had no significant effect on peak current amplitude, although after 4 h the current amplitude appeared to be smaller in some cases. We attribute this to a nonspecific mechanistic disruption by the injection itself.
It has been established that the V-ATPase is responsible for the maintenance of the luminal acidic environment within cell organelles including the Golgi complex, lysosomes, and endosomes (29). Inhibition of the V-ATPase by bafilomycin A 1 impairs the intracellular transportation of glycoproteins via alkalinization of the organelles (30). Thus, if the ␤ subunit assists the translocation of the ␣ 1 subunit to the plasma membrane, inhibition of this pathway may be imposed by consequences on one or all of the auxiliary ␤ subunit modulatory function(s). Therefore, we pretreated oocytes with bafilomycin A 1 to determine whether the ␤ subunit participates in the intracellular translocation of the ␣ 1C subunit. Pretreatment of oocytes expressing the ␣ 1C subunit with bafilomycin A 1 for 2-3 h showed a slight decrease in the control (␣ 1C subunit alone) current (114 Ϯ 14 nA, n ϭ 7) compared with the untreated control. However, these data were not statistically different (p ϭ 0.13 versus untreated control). The effect of the ␤ 3 subunit protein on peak current amplitude was effectively blocked by preincubation with bafilomycin A 1 prior to subunit injection. Only 1.5-, 1.8-, 1.8-, and 1.7-fold current increase was observed at 1-4 h after injection of the ␤ 3 subunit protein, respectively.
Effects of ␤ 3 Subunit Protein on the Voltage Dependence of Activation-Steady-state activation curves were derived from I-V relationships, and the ␤ 3 subunit effect on the voltage dependence of activation was further analyzed (Fig. 5A). Control currents showed a shallow activation curve with a halfactivation potential of 29.9 Ϯ 1.5 mV (n ϭ 14). Injection of the ␤ 3 subunit protein caused a leftward shift of the curve and an increase in the slope. Most dramatic changes occurred within 1 h, and no significant changes were observed between 2, 3, and 4 h after injection. As shown in Fig. 5B, ϳ70% of the negative shift in the half-activation potential occurred within 1 h after injection, reaching a plateau at 2 h. The time of the halfmaximal effect was less than 1 h. The vehicle had no effect on the voltage dependence of activation. Bafilomycin A 1 treatment did not influence the ability of the ␤ 3 subunit to shift the half-activation potential in the negative direction (Fig. 5B).
Effects of the ␤ 3 Subunit Protein on Current Kinetics-We measured the time to half-peak current as a parameter of the macroscopic activation kinetics. As shown in Fig. 3A, the control (␣ 1 subunit alone) current showed slow activation. When depolarized to ϩ30 mV from a holding potential of Ϫ80 mV, the time to half-peak was 19.2 Ϯ 1.0 ms (n ϭ 9) (Fig. 6A). The injection of oocytes with the ␤ subunit protein rapidly facilitated activation kinetics, reaching a plateau within 2 h, with the time of the half-maximal effect occurring within 1 h. The vehicle did not change activation kinetics. Pretreatment of oocytes with bafilomycin A 1 did not change activation kinetics of control (␣ 1 subunit alone) currents and did not modify acceleration of activation kinetics by the ␤ subunit protein (Fig. 6A).
We also investigated the kinetics of inactivation for macroscopic Ba 2ϩ currents. The rate of inactivation was quantitated as the inactivated fraction of Ba 2ϩ current 1 s after the application of a depolarizing test pulse. In control oocytes, the current elicited by the ␣ 1 subunit alone showed very little or no inactivation (see Fig. 3A). After injection with the ␤ subunit protein the current exhibited a time-dependent faster inactivation (Fig. 6B), similar to the time-dependent changes of voltage dependence and activation kinetics. Again, changes occurred maximally within 2 h, with the time of the half-maximal effect being less than 1 h. The vehicle did not significantly affect the inactivation of Ba 2ϩ current. In addition, pretreatment of oocytes with bafilomycin A 1 did not influence the control current nor the effect of the ␤ subunit protein on inactivation kinetics (Fig. 6B), suggesting this mechanism of modulation is independent of protein translocation.
␤ Subunit Protein Increases the Amount of ␣ 1 Subunit in the Plasma Membrane-We tested whether the injection of the ␤ 3 subunit protein into oocytes promoted ␣ 1C protein delivery to the plasma membrane. The HA epitope-tagged ␣ 1C subunit was expressed in Xenopus oocytes that were then injected with the ␤ 3 subunit protein. We found no difference in either the characteristics of Ba 2ϩ current or effects of the ␤ 3 subunit protein between the wild-type channels and epitope-tagged channels. The epitope-tagged ␣ 1C subunits were detected in the plasma membrane immunocytochemically using CLSM and quantitated by measuring the pixel intensity. When the epitopetagged ␣ 1C was expressed alone in Xenopus oocytes, we observed very little fluorescence in the plasma membrane. In the presence of the ␤ 3 subunit protein, the amount of the ␣ 1C subunit in the plasma membrane increased severalfold (Fig. 7). Furthermore, pretreatment of oocytes with bafilomycin A 1 completely abolished this increase, strongly suggesting that this process is dependent on protein translocation. We obtained similar results in 10 -15 oocytes for each group over two different batches of oocytes.

DISCUSSION
Modulation of calcium channel activity by auxiliary subunits has been extensively studied using many different recombinant expression systems. These studies have clearly demonstrated that coexpression of the ␤ subunit alters calcium channel characteristics by increasing current density, shifting the voltage dependence of activation and accelerating channel kinetics. Taken together, these modulatory alterations are believed to impart the inherent current characteristics observed in native preparations. An issue that remains unresolved is whether the ␤ subunit increases the amount of the ␣ 1 subunit in the plasma membrane or modulates calcium channel functions solely through an allosteric pathway. Experimental observations to date present conflicting results. When the ␣ 1 subunit was coexpressed with the ␤ subunit in Xenopus oocytes, no change in channel expression was observed, when measured as a function of gating charge movement (8). Moreover, in the same experimental system, the amount of 35 S-labeled ␣ 1 subunit in the plasma membrane did not change (17). Furthermore, Nishimura et al. (10) have also shown no change in the ␣ 1 subunit content of membrane fractions by immunoblotting analysis of cells transfected with the ␤ subunit. In contrast, several studies using mammalian cells have shown an increase in charge movement when transfected with ␤ (18,19) and the involvement of the ␤ subunit in translocation of the ␣ 1 subunit to the membrane (15,16). However, the biochemical mechanism(s) responsible for this modulation have remained unclear due to insufficient experimental methods to monitor time-dependent biological events occurring intracellularly. In an effort to overcome this limitation, and to resolve apparent conflicting data, we expressed the L-type calcium channel ␣ 1 subunit in oocytes, and after injecting a highly purified recombinant ␤ 3 subunit protein, we examined the time-dependent changes to peak current, voltage dependence, and kinetics. Our results demonstrate for the first time a time-dependent uncoupling of ␤ 3 subunit modulation of voltage dependence and kinetics from enhancement of peak current density. Moreover, these results suggest that ␤ subunit modulation occurs via an allosteric mechanism and through facilitation of protein translocation. ␤ 3 Subunit Binding to the I-II Intracellular Loop-In order to determine whether the His 6 -tagged ␤ 3 subunit interacts with the ␣ 1 interaction domain (AID) of the intracellular I-II loop as described by Pragnell et al. (28), we screened GST fusion proteins containing each of the three intracellular loops I-II, II-III, and III-IV with an 35 S-labeled His 6 -tagged ␤ 3 subunit. We found that the ␤ 3 subunit was able to interact with the I-II loop, suggesting that the presence of six histidine residues at N terminus does not interfere with the binding of this subunit to its intracellular binding site. Since Walker and co-workers (31) have shown that the modulatory functions of the ␤ subunit are largely dependent on this interaction, and we have shown that the His 6 -tagged ␤ 3 subunit can interact with the I-II loop, the ␤ 3 subunit should modulate calcium channel function when injected into oocytes as a purified fusion protein.
It was also clear from the results of these experiments that the ␤ 3 subunit did not interact with either the intracellular II-III or the III-IV loops. Although our in vitro experiment showed highly specific interaction between the ␤ 3 subunit and the ␣ 1C I-II loop, we cannot exclude possible weak or transient interactions with other intracellular regions.
Injection of Oocytes Expressing the Ca 2ϩ Channel ␣ 1 Subunit with a Recombinant ␤ Subunit Fusion Protein Induces Effects Comparable to ␣ 1 -␤ Coexpression-Effects of ␤ subunit coexpression on the ␣ 1C subunit have been studied using Xenopus oocytes (7,8,17,32,33) as well as in mammalian cells (6,9,10,15,18). It is now generally accepted that coexpression of the ␤ subunit results in a 2-fold to more than 100-fold increase in peak current amplitude, a hyperpolarizing shift of the voltage dependence of activation, and acceleration of kinetics of the current (to a different extent, depending on the different types of ␤ subunits). Since our results showed comparable effects of the ␤ subunit on these parameters (e.g. ϳ6.5-fold increase in current amplitude, ϳ16 mV hyperpolarizing shift of half-activation potential, and acceleration of activation and inactivation kinetics) within 3 h of ␤ subunit protein injection, it is unlikely that the ␤ subunit is enhancing protein synthesis of ␣ 1 subunits in the endoplasmic reticulum. Employing our experimental conditions, it takes at least 3 days from the time of coinjection of the ␣ 1C and ␤ subunit cRNAs and at least 4 days for the ␣ 1C subunit cRNA alone to get measurable current through these expressed channels. Therefore, even if we assume that the presence of the ␤ subunit somehow facilitates protein synthesis of the ␣ 1 subunit, its contribution to the observed effects should be minimal, since our recording time scale is shorter. Thus, we believe that the effects of the ␤ subunit are exerted mainly on mechanism(s) downstream of protein synthesis.
The Time Course of the ␤ Subunit Effects Can Be Categorized in Two Distinct Patterns-In our present study, we analyzed the time course of four parameters after injection of the ␤ subunit protein as follows: 1) peak Ba 2ϩ current amplitude; 2) voltage dependence of activation; 3) activation kinetics; and 4) inactivation kinetics. Among these, only the peak current amplitude increased within a slow time framework; it took 3 h to reach plateau and the time of the half-maximal effect was between 2 and 3 h. The other three parameters behaved very similarly to each other and changed with a faster time course, reaching a steady level in 2 h, and the time of the half-maximal effect was less than 1 h. According to the in vitro binding study by De Waard et al. (24), AID and the ␤ 1b subunit bound at a rate constant of 0.1 min Ϫ1 ⅐M Ϫ1 , and when using 500 nM AID, the rate constant corresponded to a half-time of about 20 min. Since our "faster" time course falls within the same time range as their results, we believe that the faster time course indicates direct binding of the injected ␤ 3 subunit protein to the ␣ 1C subunits that already exist in the plasma membrane. We do not know the exact time it takes for the injected ␤ 3 subunit protein to diffuse, reach the plasma membrane, and build to a saturated concentration. However, since an excess amount of ␤ 3 subunit protein was injected (final concentration in the oocyte cytoplasm was 300 nM, which is about 5.4 times higher than the reported K d of AID and ␤ 3 subunit by De Waard et al. (24), 55.1 nM), we assume it takes less than 1 h. Taken together, it seems that the binding of the injected ␤ 3 subunit protein to the ␣ 1C subunits in the plasma membrane reaches equilibrium within 2 h. Bafilomycin A 1 treatment did not influence the ␤ effects that are categorized with a faster time course but did abolish the increase of ␣ 1 subunit by the ␤ subunit. This finding also supports the concept that allosteric modulation of ␣ 1 subunits in the plasma membrane by the ␤ subunit is responsible for the faster effects, whereas the "slower" component implies the existence of a distinct mechanism for modulation. In order to clarify this, we addressed the question whether the ␤ subunit modulation of current amplitude occurs during protein translocation.
Influence of the Pretreatment with Bafilomycin A 1 -Bafilomycin A 1 , a V-ATPase inhibitor, inhibits intracellular glycoprotein transport by impairing the acidification of organelles (30).
In the presence of this compound the ␤ subunit failed to increase the amount of ␣ 1 subunit in the plasma membrane. Bafilomycin A 1 also significantly blocked the effect of the ␤ subunit on current amplitude. The results strongly suggest that the effect of the ␤ subunit protein on current amplitude is largely dependent on intracellular translocation of nascent ␣ 1 subunits. The ␤ subunit elicited a ϳ1.8-fold increase of current when oocytes were treated with bafilomycin A 1 , despite a complete loss of increase of the ␣ 1 subunit in the plasma membrane. An attractive explanation for the current increase is that the injected ␤ subunit allosterically modulates the population of ␣ 1 subunits that are already inserted in the membrane. This is in agreement with single channel analyses done by Wakamori et al. (33), in which the coexpression of the ␣ 1 subunit with the ␤ subunit resulted in ϳ2-fold increase in the channel open probability.
In the present study, pretreatment with bafilomycin A 1 slightly decreased the control (␣ 1 subunit alone) current amplitude, without affecting other characteristics of the current. Assuming that there is turnover of channels in the plasma membrane (15), some breakdown of functional channels will occur and will be replaced by the translocation of new channels from the cytosolic region. Therefore, it seems reasonable that inhibition of protein translocation results in a decreased number of functional channels and, consequently, decreased current amplitude. The contribution of other mechanism(s), such as destabilization of membrane-incorporated channels, cannot be excluded.
Possible Role of Xenopus Oocyte Endogenous ␤ Subunit-Recently Tareilus et al. (25) cloned an endogenous ␤ subunit from Xenopus oocytes (␤ XO ) which is highly homologous to the mammalian ␤ 3 subunit. Injection of Xenopus ␤ antisense oligonucleotides significantly reduced the current through the expressed ␣ 1C and ␣ 1E subunit of the Ca 2ϩ channel. Based on this, the authors proposed that the "␣ 1 alone" channels are in fact forming an ␣ 1 subunit-endogenous ␤ subunit complex (␣ 1 ␤ XO ). However, it is not possible at this time to determine whether the endogenous ␤ XO was present in high enough concentration to complex with the ␣ 1 subunit in the absence of exogenous ␤. Thus, it seems likely that our observations incor-porate the effects of exogenous ␤ 3 subunit on the ␣ 1 ␤ XO complex. However, the basic characteristics of the control (␣ 1 alone) current in the present study, viz. slow activation, slow inactivation, smaller current amplitude, and a shifted voltage dependence to the depolarized direction, and the modulation of these parameters by the application of exogenous ␤ subunit are in good agreement with previous studies using a variety of mammalian cells (6,9,10,15,19) in which no endogenous ␤ subunits have been reported.
The strategy and molecular reagents used throughout these studies are thought to be required for efficient interaction among all calcium channel ␣ 1 and ␤ subunits (25,28,34,35), viz. the I-II intracellular connecting loop and C-terminal tail of ␣ 1 and a full-length ␤ subunit. Therefore, it is reasonable to conclude that the mechanisms described in the present study should apply for all calcium channel ␣ 1 and ␤ subunit interactions.
In summary, we have provided compelling evidence that the ␤ subunit modulates the function of the ␣ 1 subunit of the voltage-dependent Ca 2ϩ channel in two distinct modes, i.e. allosteric modulation and chaperoning of channels to the plasma membrane.