Folding of Active Calcium Channel β1b -Subunit by Size-exclusion Chromatography and Its Role on Channel Function*

Voltage-gated calcium channels mediate the influx of Ca2+ ions into eukaryotic cells in response to membrane depolarization. They are hetero-multimer membrane proteins formed by at least three subunits, the poreforming α1-subunit and the auxiliary β- and α2δ-subunits. The β-subunit is essential for channel performance because it regulates two distinct features of voltage-gated calcium channels, the surface expression and the channel activity. Four β-subunit genes have been cloned, β1–4, with molecular masses ranging from 52 to 78 kDa, and several splice variants have been identified. The β1b-subunit, expressed at high levels in mammalian brain, has been used extensively to study the interaction between the pore forming α1- and the regulatory β-subunit. However, structural characterization has been impaired for its tendency to form aggregates when expressed in bacteria. We applied an on-column refolding procedure based on size exclusion chromatography to fold the β1b-subunit of the voltage gated-calcium channels from Escherichia coli inclusion bodies. The β1b-subunit refolds into monomers, as shown by sucrose gradient analysis, and binds to a glutathione S-transferase protein fused to the known target in the α1-subunit (the α-interaction domain). Using the cutopen oocyte voltage clamp technique, we measured gating and ionic currents in Xenopus oocytes expressing cardiac α1-subunit (α1C) co-injected with folded-β1b-protein or β1b-cRNA. We demonstrate that the co-expression of the α1C-subunit with either folded-β1b-protein or β1b-cRNA increases ionic currents to a similar extent and with no changes in charge movement, indicating that the β1b-subunit primarily modulates channel activity, rather than expression.

Changes in the intracellular calcium concentration regulate a variety of cellular functions such as neurotransmission, muscle contraction, hormone secretion, and gene expression. High threshold voltage-activated calcium channels are the main route for calcium entry in electrically excitable cells. They are membrane protein complexes composed of at least three nonhomologous subunits, the ␣ 1 -, ␤-, and the ␣ 2 /␦-subunit. Through molecular cloning, at least 10 genes encoding mammalian ␣ 1 -subunits (␣ 1A-I and ␣ 1S ) have been identified in different cell types (1). Although the ␣ 1 -subunit encompasses all the structural elements of a functional voltage-activated calcium channel, such as the ion-conduction pathway, the voltage sensor, and drug-binding sites, the ␤-subunit seems to be essential for channel performance (2) and to be acting at two levels: (i) channel expression by interfering with the ␣ 1 -subunit endoplasmic reticulum (ER) 1 -retention signal to facilitate intracellular trafficking (3,4), and (ii) channel activity by modifying the electrophysiological properties of the channel (5-7).
Two highly conserved sequences have been identified as the primary interaction site between the ␣ 1 -and the ␤-subunit, the ␣ 1 -subunit interaction domain (AID) that lies within the cytoplasmic I-II loop of the ␣ 1 -subunit (8), and the ␤-subunit interaction domain that lies within the second conserved sequence domain of the ␤-subunit (9). Four ␤-subunit isoforms (␤ [1][2][3][4] with molecular masses between 52-78 kDa from four nonallelic genes have been cloned, each encoding multiple splice variants. Sequence analysis of all ␤-subunits reveals five homology domains, two of them highly conserved among all ␤-subunits and known to be important for channel function (10). Structural modeling based on homology searches has been carried out for one of the ␤-subunit isoforms, the ␤ 1b . This study proposes a modular structure of the protein containing a PDZ-like, an Src homology 3, and a guanylate kinase domain (11). However, the functional relevance of these structural domains is yet to be established. Direct biochemical and structural characterization has been hampered by the lack of appropriate expression and purification protocols to produce soluble and stable protein in sufficient amounts. The ␤ 1b -subunit has a length of 597 amino acids and a calculated molecular mass of 65.679 Da, and although it expresses at high levels in Escherichia coli, the majority of the recombinant protein accumulates in inclusion bodies. We describe here a procedure to purify and refold the ␤ 1b -subunit from inclusion bodies using sizeexclusion chromatography (SEC). Sedimentation profile in a sucrose density gradient is consistent with a monomeric state of the folded protein. This protein is stable in solution up to 5-6 mg/ml for about 2 weeks and displays binding affinity for the AID site of the cardiac isoform of the ␣ 1 -subunit (␣ 1C ). We investigated the biological function of the ␤ 1b -subunit, either * This work was supported by Fondo Nacional de Desarrollo Científico y Tecnológico Grants 1020899, ICM P99-037-F (to A. N.), and Deutsche Forschungsgemeinschaft Grants FOR 450 and TP 1 (to P. H.). 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.
¶ Student of Programa de Doctorado, Facultad de Ciencias de la Universidad de Chile and Universidad de Valparaíso.
** To whom correspondence should be addressed: Institut fü r Physiologie, RWTH Aachen, Pauwelsstrasse 30, 52074 Aachen, Germany. Tel.: 49-24180-88810 or -88873; Fax: 49-24180-82434; E-mail: patty@ physiology.rwth-aachen.de. originated from cRNA or directly from the folded protein, by testing its ability to modify the activity of calcium channels expressed in Xenopus laevis oocytes. To distinguish between the effect of the ␤-subunit on channel expression and activity, we used the cut-open oocyte voltage clamp technique (12) to record simultaneously ionic currents and gating currents. The latter, being proportional to the number of channels in the membrane, reports changes in expression levels (6). The injection of either ␤ 1b -cRNA or folded-␤ 1b -protein into ␣ 1C -subunitexpressing oocytes increases ionic currents with little or no changes in charge movement, indicating that the ␤ 1b -subunit modulates primarily channel activity.

EXPERIMENTAL PROCEDURES
Protein Preparation-The cDNA encoding the ␤ 1b -subunit (Gen-Bank™ accession number X61394) residues 24 -597 was subcloned between the BamHI and EcoRI sites by conventional PCR methods into pRSETB vector containing a polyhistidine (His 6 ) at the N-terminal end (Invitrogen). The entire cDNA was sequenced after being subcloned in pRSETB vector, and seven amino acid substitutions were detected (F158L, S348F, R417S, V435A, V449A, W492R, and V511A) when compared with the sequence reported by Pragnell et al. (13) (Swiss-Prot accession number P54283). The His 6 -␤ 1b -tagged protein was used for Western blotting analysis with anti-His antibodies according to the manufacturer's instructions (Qiagen). E. coli BL-21 transformed with the ␤ 1b -pRSET vector were grown at 25°C to an optical density of 0.7 and induced 90 min with 0.5 mM isopropyl-␤-D-thiogalactopyranoside.
Cells were harvested by centrifugation and stored at Ϫ80°C until use. After lysis by sonication and centrifugation, the pellet fraction was recovered, repeatedly washed using mild detergent (B-PER™, Pierce) according to the manufacturer's instructions, dissolved in denaturing buffer containing 6 M guanidinium chloride (GdmCl), 100 mM NaH 2 PO 4 , and 10 mM Tris base, pH 8.0, and the resulting suspension was centrifuged for 5 mins at 10,000 ϫ g to remove particulate. Dithiothreitol (DTT) was added to a final concentration of 10 mM, and the sample was incubated at 95°C for 10 mins. The DTT-treated sample was allowed to cool down, and 5 ml were loaded onto a Superdex™ 200 column 26/60 (340 ml bed volume) attached to a fast performance liquid chromatography system (Amersham Biosciences) and pre-equilibrated at 12°C with refolding buffer A (50 mM Tris, 300 mM NaCl, 1 mM EDTA, pH 8.0). The elution profile was monitored at 280 nm with the use of a UV detector in all chromatography steps. The fractions containing the protein were verified by SDS-PAGE, pooled, concentrated to 3-6 mg/ml, and stored at 4°C until use. The protein concentration was measured by absorbance at 280 nm using the absorbance coefficient of A 280 mg/ml ϭ 1.3 (1-cm cell path length). The cDNA encoding the I-II loop residues 422-534 of the ␣ 1C calcium channel subunit (GenBank™ accession number X15539) that encompass the AID site, was subcloned by PCR methods into the glutathione S-transferase (GST) gene fusion pGEX vector (Amersham Biosciences). E. coli BL-21 pLysS transformed with the pGEX GST-AID ␣1C vector were grown at 23°C to an optical density of 0.7 and induced with 0.5 mM isopropyl-␤-D-thiogalactopyranoside for 2 h. Cells were harvested by centrifugation and stored at Ϫ80°C until use.
The GST-AID ␣1C fusion protein was purified from the cleared cell lysate by glutathione-affinity chromatography (glutathione-Sepharose 4 Fast Flow, Amersham Biosciences), according to the manufacturer's instructions, except that 10 mM DTT and 1% reduced Triton X-100 (Sigma) was added to the buffers throughout the whole purification procedure, and the Tris-HCl elution buffer was replaced by phosphatebuffered saline buffer. The eluted fractions containing GST-AID ␣1C and verified by SDS-PAGE, were pooled, concentrated 10 -15 times, loaded onto a Superdex™ 200 column 10/30 (24 ml bed volume) pre-equilibrated at 12°C with buffer A supplemented with 0.1% Triton and eluted at 0.5 ml/min. The fractions that contained the protein and eluted within the included volume of the column were collected and stored at 4°C until use. The concentration of the GST-AID ␣1C was estimated by the intensity of the band in a reducing SDS-PAGE and confirmed by bicinchoninic acid assay for protein quantitation (Uptima, Montluçon, France). The GST protein fused to 126 amino acid residues derived from the carboxyl-terminal end of the chloride channel from human skeletal muscle ClC-1 (Swiss-Prot accession number P35523, GST-ClC 126 ) was kindly provided by Christoph Fahlke (RWTH Aachen, Germany) Binding Assay-4 g of purified GST-AID ␣1C fusion protein were diluted into 500 l of binding buffer containing 50 mM Tris-HCl, 150 mM NaCl, and 0.1% Triton, pH 8.0, and added to ϳ20 l of glutathione-Sepharose resin previously washed and resuspended in binding buffer. This mixture was incubated for 1 h at room temperature and washed twice with binding buffer. The binding reaction was performed by adding 8 g of purified ␤ 1b -subunit diluted in 500 l of binding buffer, incubated 60 min at room temperature, and washed three times with binding buffer. The bound proteins were released from the resin with 20 l of reducing SDS-loading buffer, incubated 5 min at 95°C, spun down, and the supernatant was analyzed by reducing SDS-PAGE. Control binding assays were carried out with 4-fold excess of either GST alone or of GST-ClC 126 .
Sucrose Density Gradient Centrifugation-Sucrose gradients 0 -15% in buffer A were generated by a gradient mixer to a final volume of 4 ml. 200 g of purified ␤ 1b -subunit together with protein markers or on separate gradients were layered on top of the gradient and centrifuged in a Beckman SW-56 rotor at 100,000 ϫ g for 16 h at 4°C. After sedimentation, individual gradients were fractionated bottom-to-top by drop-wise collection into 16 tubes (250 l each). Aliquots of each fraction were analyzed by SDS-PAGE, and UV-absorbance was measured at 280 nm (Ultrospec 2000, Pharmacia Biothec).
Xenopus laevis Oocytes Injection and Electrophysiological Recordings-Xenopus laevis females were anesthetized with 0.5% ethyl 3-amino-benzoato methanesulfonic acid salt and the oocytes were removed, defolliculated by 1-to 2-h treatment with 3 mg/ml collagenase (Worthington Biochemical Corporation, Lakewood, NJ) at room temperature, washed, and kept at 18°C in oocyte Ringer's solution (ND-96) containing 96 mM NaCl, 2 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES, pH 7.4, until used. Capped cRNA was synthesized from HindIII-linearized ␤ 1b and ␣ 1C -DN60 cDNA (14) template using the MESSAGE machine (Ambion, Austin, TX), resuspended in 10 l of water, and stored in 2-l aliquots at Ϫ80°C until use. The ␣ 1C -DN60 bears a deletion of 60 amino acids at the amino-terminal end that increase expression. The cRNA was injected using a nanoliter injector (nanoliter 2000, World Precision Instruments, Sarasota, FL), and oocytes were kept at 18°C in ND-96 solution supplemented with 2.5 mM sodium pyruvate and 100 g/ml gentamycin sulfate until recording. To assay the functional properties of the folded-␤ 1b -protein, 4 -6 days after cRNA injection, oocytes were injected with 50 nl of ␤ 1b (2-4 mg/ml in buffer A). Electrophysiological recordings were carried out as described (15) between 2-7 h after protein injection. Briefly, macroscopic currents were recorded using the cut-open oocyte voltage-clamp technique (12) with a CA-1B amplifier (Dagan Corp., Minneapolis, MN). The membrane of the oocyte exposed to the bottom chamber was permeabilized by a brief treatment with 0.1% saponin. The external solution contained 10 mM Ba 2ϩ , 96 mM n-methylglucamine, and 10 mM HEPES and was adjusted to pH 7.0 with methanesulfonic acid. The internal solution contained 120 mM n-methylglucamine, 10 mM EGTA, and 10 mM HEPES, pH 7.0 (adjusted with methanesulfonic acid). Data acquisition and analysis were performed using the pCLAMP system and software (Axon Instruments Inc., Foster City, CA). Currents were filtered at 2 kHz and digitized at 10 kHz. Linear components arising from charging the membrane and voltage-independent leakage conductance were eliminated by P/Ϫ4 prepulse protocol, as described previously (6). For the average currentvoltage plot (see Fig. 5D), currents were evoked by 70-ms pulses from Ϫ40 to ϩ 70 mV in 10 mV increments, and the amplitude (I m ) was measured at the end of the pulse. The normalized conductance-voltage plot (see Fig. 5E) was obtained by using Eqs. 1 and 2 as follows: Eq. 1 was fitted to the I-V plot, and the values of G max and E rev were obtained from the fit used to calculated the normalized membrane conductances (G m ) according to Eq. 2.
where G max is the maximum conductance, V m is the membrane potential, E rev is the reversal potential, z is the effective valence, and V1 ⁄2 is the voltage at which the conductance is 50% of G max .

RESULTS
Refolding of ␤ 1b -Subunit-The ␤ 1b -subunit expressed in E. coli accumulates intracellularly in large insoluble aggregates or inclusion bodies (Fig. 1A) that are recovered in the pellet fraction following cell lysis and centrifugation (Fig. 1A,  lane 6). Successive washes with B-PER™ reagent were done to remove contaminants (Fig. 1A, lane 7), and the purified inclusion bodies were solubilized in denaturing buffer containing 6 M GdmCl. After the refolding procedure was completed (see below), the ␤ 1b purified from the inclusion bodies was analyzed by Western blotting using anti-His antibody (Fig. 1B).
Our attempts to refold the protein using conventional methods for denaturant removal such as dialysis or fast dilution failed, and therefore, we investigated a new strategy to refold the ␤ 1b -subunit from the inclusion bodies (16,17). The method uses SEC to exchange the buffer, remove aggregates, and promote folding in a single step during elution with refolding buffer. The solubilized inclusion bodies were incubated for 10 mins at 95°C under reducing conditions (10 mM DTT), loaded onto a Superdex-200 26/60 column equilibrated with refolding buffer (buffer A), and eluted at a flow rate of 2.5 ml/min at 12°C. The eluted fractions were monitored at 280 nm and further analyzed by reducing SDS-PAGE. The elution profile shows three peaks ( Fig. 2A) containing pure ␤ 1b , as judged by SDS-PAGE (data not shown). One peak eluted within the column void volume (V 0 ), indicating the presence of higher aggregates of ␤ 1b . Two distinct peaks (peaks I and II) eluted within the column included volume, suggesting that the ␤ 1b -subunit may have been refolded into at least two relatively discrete states. When the higher molecular weight peak (peak I) was recovered from the column, concentrated to 1.5 mg/ml, and reloaded onto the column, a significant portion eluted in the void volume and became the predominant peak within a couple of days (data not shown), indicating that the protein aggregates over time. In contrast, when peak II was recovered and concentrated to 3.7 mg/ml, it eluted predominantly as a single peak with a minor shoulder (Fig. 2B). This profile remained essentially unchanged upon storage for about 2 weeks, even at higher protein concentrations, indicating that the protein was stable and likely in a fully folded conformation. Protein isolated from peak II was, therefore, referred to as folded-␤ 1b and used for further studies. The integrity of the folded-␤ 1b was confirmed by amino acid analysis.
To optimize the buffer conditions for the folded-␤ 1b , a 1 mg/ml protein stock solution was diluted six times in each different test buffer and loaded onto the Superdex 200 10/30 equilibrated column with the same dilution buffer. The SEC elution profiles for the folded-␤ 1b were superimposable at pH 8.0 and above and at salt concentrations higher than 200 mM NaCl (data not shown). However, at lower salt concentrations or at pH below 6.2 no peak was detected. Removal of EDTA from buffer A did not alter the elution curve.
A SEC calibration curve prepared with a set of globular protein standards yielded an apparent molecular mass of 188 kDa for the folded-␤ 1b (Fig. 2C), more than twice the molecular mass predicted by the amino acid sequence (66 kDa). This finding may indicate either formation of oligomers (dimers or trimers) or alternatively, an anomalous SEC behavior due to molecule asymmetry, because the molecular mass of native proteins estimated by SEC depends highly upon the shape of the molecule (18).
Sucrose Gradient Analysis-We used sucrose gradient analysis to resolve independently from SEC whether or not the folded-␤ 1b forms multimers. The folded protein was subjected to 0 -15% sucrose gradient centrifugation either in a separate gradient or along with two protein markers, albumin that has a comparable molecular mass (67 kDa) and catalase that has a higher molecular mass (232 kDa) (Fig. 3). Analysis of the different sucrose gradient fractions by reducing SDS-PAGE and UV-absorbance at 280 nm showed that the folded-␤ 1b sediments at one position along the gradient (Fig. 3A), indicating the presence of a homogenous protein population in a defined conformation. This analysis also shows that folded-␤ 1b co-sediments with albumin while it separates from catalase (Fig. 3B). Albumin and catalase sediment at the same position of the gradient fraction when run alone (data not shown). It is readily apparent from Figs. 2C and 3B that, although catalase behaves as expected from the calculated where V e is the elution volume of the protein, V 0 is the void volume of the column calibrated with blue dextran, and V t is the total bed volume. molecular mass in SEC and sucrose gradient analysis, it migrates faster in SDS-PAGE. The same fast migrating band in SDS-PAGE was obtained when the catalase fraction was recovered from either the SEC column or the sucrose gradient and ran in parallel with fresh protein (data not shown). The recovery of the folded-␤ 1b in the same fraction as the albumin indicates that the folded protein behaves as a monomer in sucrose gradient analysis and, accordingly, its larger apparent molecular mass, deduced by SEC, suggests an anomalous behavior common to asymmetric molecules.
Binding in Vitro of the Folded-␤ 1b to the AID Site-Because refolded proteins may not necessarily adopt their native conformation with retention of biological activity, we investigated the ability of the folded-␤ 1b to bind to its natural target on the ␣ 1 pore-forming subunit, the highly conserved AID site. The I-II loop of the ␣ 1C -subunit (residues Gly 422 -Arg 534 ) bearing the AID site was fused to GST (GST-AID ␣1C ) and used for in vitro binding assays, as described under "Experimental Procedures." The folded-␤ 1b bound to the GST-AID ␣1C fusion protein (Fig.  4A) but not to GST protein alone, even at a 4-fold excess over GST-AID ␣1C (Fig. 4B) or to GST fused to an unrelated sequence, such as the carboxyl-terminal end of the hClC1 channel protein (Fig. 4C, GST-ClC 126 ). These results show that the folded-␤ 1b binds specifically to the AID ␣1C site.
Effect of the Folded-␤ 1b on ␣ 1C -Calcium Channels Expressed in Xenopus laevis Oocytes-To prove further that the ␤ 1b -subunit folds to an active form and to investigate the functional effects of this ␤-subunit isoform on calcium channel activity, we examined the effect of the ␤ 1b -subunit originated from cRNA and the folded protein on ␣ 1C expressing Xenopus oocytes. We used the cut-open oocyte voltage-clamp technique to simultaneously record ionic and gating currents. The ability to measure gating and ionic currents in the same oocyte allow us to discriminate between the effect of the ␤-subunit on channel expression from the one on channel activity as follows. The movement of charged residues in the channel protein, as it undergoes conformational changes, gives rise to gating currents reflected as small outward transients preceding ionic currents at the onset of depolarizing pulses (19). The integral of these gating currents yields the charge movement during channel activation and is a function of the number of channels present in the membrane and the number of charged residues per channel that move within the electric field during activation (20). On the other hand, the amplitude of ionic currents, carried by Ba 2ϩ (I Ba ) in our case, is a function of the number of channels, the single-channel conductance, and the probability of channel opening. Thus, changes in charge movement report alteration in channel expression, whereas Ba 2ϩ currents, when normalized by charge movement, reveal changes in channel activity. Using this approach, it has been shown that in Xenopus oocytes injected with cRNA for the pore-forming ␣ 1C -subunit alone or together with cRNA encoding the ␤ 2a -subunit, a similar number of ␣ 1C -subunits reach the membrane, but coexpression of ␤ 2a -subunit yields larger ionic currents (6). This increase was also observed when co-injecting cRNA for the ␤ 1a -subunit (15).
Here, we injected Xenopus oocytes with ␣ 1C -DN60-cRNA alone or together with ␤ 1b -cRNA or folded-␤ 1b -protein and measured gating and ionic currents (Fig. 5A). Currents recorded in the absence of exogenous ␤ 1b -subunit showed a rapid outward transient corresponding to gating currents, followed by a small non-inactivating ionic current (Fig. 5A, upper panel). In contrast, oocytes co-injected with ␤ 1b -cRNA displayed inward currents that are larger in amplitude than gating currents (Fig. 5A, middle panel). Likewise, injection of the folded-␤ 1b protein to ␣ 1C -DN60-expressing oocytes also leads to an increase in ionic current amplitude relative to the gating currents (Fig. 5A, lower panel). Another property shared by calcium channel ␤-subunits is their ability to shift the voltagedependence of the activation curve toward more hyperpolarized potentials. Fig. 5D shows the voltage dependence of averaged current amplitudes, and Fig. 5E shows the voltage-dependence of the normalized conductance from oocytes expressing ␣ 1C -DN60 alone (OE), co-injected with ␤ 1b -cRNA (q), or with folded-␤ 1b -protein (E). From these measurements, G max for ␣ 1C -DN60 (2.2 Ϯ 0.3 s; n ϭ 9) compares to 2.4 Ϯ 0.3 s (n ϭ 15) and 2.1 Ϯ 0.3 s (n ϭ 15) measured in the presence of ␤ 1b -cRNA or folded-␤ 1b -protein , respectively. On the other hand, the mid-  4. Binding of the folded-␤ 1b to a GST-AID ␣1C fusion protein. A, binding of ␤ 1b to GST-AID ␣1C . The folded-␤ 1b was added to the GST fusion protein bound to glutathione agarose resin. After the binding reaction was completed, the resin was washed, resuspended in SDS loading buffer, and the supernatant was analyzed by SDS-PAGE (10% acrylamide gel). Binding activity is only observed with GST-AID ␣1C . point of the activation, V1 ⁄2 , shifts from 26 Ϯ 1 mV to 4 Ϯ 1 mV after the injection of folded-␤ 1b , which compares to the V1 ⁄2 measured in oocytes co-expressing ␤ 1b -cRNA (Ϫ1 Ϯ 1 mV). These results show that the folded-␤ 1b protein is nearly as effective as the ␤ 1b -subunit derived from cRNA in modulating channel activity, and that the expressed ␣ 1C -calcium channels are fully available to interact with injected ␤ 1b -protein. Injection of folded-␤ 1b previously heated at 100°C for 15 min produces no apparent changes in the relationship between ionic and gating currents (data not shown).
To resolve the role of the ␤ 1b -subunit in channel expression and/or activity, we estimated the number of channels contributing to the recorded current by integrating the first 2 ms of gating currents evoked by a pulse from Ϫ80 to Ϫ30 mV (Q on ) for the three subunit combinations (␣ 1C , ␣ 1C ϩ␤ 1b RNA, and ␣ 1C ϩ␤ 1b protein, Fig. 5B). At this membrane potential, the ionic current component is minimal and is delayed by about 4 ms with respect to the gating currents. Because the voltage-and time-dependence of the charge movement during the initial 2 ms is not affected by the presence of the ␤-subunit (6), Q on is expected to be proportional to the number of ␣ 1C channels present in the recorded membrane (21). From these measurements, we found that there is no significant difference in expression levels among the different subunit combinations (Fig. 5B: 17.1 Ϯ 3.0 pC for ␣ 1C , 17.7 Ϯ 5.7 pC for ␣ 1C /␤ 1b RNA, and 19.5 Ϯ 3.9 pC for ␣ 1C /␤ 1b protein) as revealed by pair-wise comparison with standard t test. In contrast, ionic currents at the end of a 60-ms pulse to 0 mV normalized by Q on (Fig. 5C) were about 20-fold larger when the ␤ 1b -cRNA was co-injected (0.26 Ϯ 0.06 nA/pC for ␣ 1C -cRNA alone versus 5.25 Ϯ 0.28 nA/pC for ␣ 1C /␤ 1b -cRNA). Likewise, when folded-␤ 1b protein was injected to oocytes expressing ␣ 1C alone, the ionic current versus charge movement ratio increased more than 12-fold (Fig. 5C, 3.28 Ϯ 1.09 nA/pC). The difference between ␣ 1C /␤ 1b -cRNA and ␣ 1C /␤ 1b -protein was not significant and, therefore, we conclude that the ␤ 1b -cRNA and the folded-␤ 1b protein increase the ratio between ionic current versus charge movement to similar levels and without significant effect on the charge movement magnitude. Fig. 6 shows the time course of the charge movement alone and the ionic current versus charge movement ratio after injection of FIG. 5. Gating and ionic currents from oocytes injected with ␣ 1C -cRNA alone or combined with ␤ 1b -cRNA or folded-␤ 1b -protein. A, representative gating and ionic current traces from oocytes expressing three different subunit combination during 60-ms voltage pulses to Ϫ30, 0, and ϩ30 mV from a holding potential of Ϫ80 mV (upper traces). Horizontal bars indicate the 30 ms displayed on an expanded scale as inset for the pulse to Ϫ30 mV. The initial 2 ms under the outward transient corresponding to gating currents was integrated to obtain Q on (highlighted as the gray shaded area). Vertical bars in the inset correspond to 10 nA. B, average Q on from oocytes expressing the different subunit combination. Similar values were obtained: 17.1 Ϯ 3.0 pC (n ϭ 16) for oocytes expressing ␣ 1C -cRNA alone; 17.7 Ϯ 5.7 pC (n ϭ 8) for ␣ 1C -/␤ 1b -cRNA, and 19.5 Ϯ 3.9 pC (n ϭ 14) for ␣ 1C -cRNA/␤ 1bprotein. C, ratios of ionic currents at 0 mV versus Q on . Oocytes expressing ␣ 1C alone transport 0.26 Ϯ 0.06 nA/pC, which is significantly smaller (p Ͻ 0.01) than when combined with ␤ 1b injected as cRNA (5.25 Ϯ 0.28 nA/PC) or as folded protein (3.28 Ϯ 1.09 nA/pC). D, average current-voltage plot for the different subunit combinations, oocytes expressing ␣ 1C -cRNA alone (n ϭ 9), ␣ 1C -/␤ 1b -cRNA (n ϭ 15), and ␣ 1C -cRNA/␤ 1b -protein (n ϭ 15), measured at the end of 70-ms pulses in 10 mV increments. E, voltage-dependence of the normalized conductances obtained from the current-voltage measurements (discussed under "Experimental Procedures"). At least three batches of Xenopus oocytes were analyzed for each subunit combination.  Fig. 5. B, Q on measured as in Fig. 5 from several batches were averaged for different time windows. The first time window includes all measurements from 10 to 30 min after protein injection; the second bin includes values from 31 to 60 min; the third bin is from 61 to 120 min. Time intervals of 120 min were used thereafter. C, ratios of ionic currents versus Q on from several batches were averaged for different time windows as in B.
folded-␤ 1b . During the initial 30 min after injection of folded-␤ 1b , we did not observe changes with respect to the recordings in the absence ␤ 1b (Fig. 6A). After this time, ionic current versus charge movement ratio increased gradually until it reached a plateau at 2 h (Fig. 6C). We did not observe any secondary increase for up to 7 h after protein injection, and during this time period, there was no sign of an increase in the number of functional channels, as assessed by the Q on values at Ϫ30 mV (Fig. 6B), which indicate that most of the changes arise from the modulation of channels pre-existing in the membrane. DISCUSSION We used size-exclusion chromatography to fold the ␤ 1b -subunit from voltage-activated calcium channels into active monomers. The ␤ 1b -subunit with 65 kDa is the largest protein so far refolded by this method that was initially applied to fold smaller proteins with molecular masses between 14 -20 kDa (17,22,23). The successful use of this method for the ␤ 1bsubunit might be associated to its multi-domain structure (11) and, whether the SEC-based refolding procedure can be generally applied to larger proteins lacking such a feature, can not be anticipated from this work. The aggregation reaction during folding was not totally inhibited, because a peak eluting within the void volume of the column was consistently observed. During the on-column folding reaction, two main conformational states, represented by peaks I and II, were detected. Peak II corresponds to the native-like state of the protein as judged by the stability and functional activity, whereas peak I includes a mixture of more expanded states that aggregate fast over time. The average recovery of active ␤ 1b -subunit from the Superdex-200 column was 50%. The folded protein is stable at pH 8 -10, with salt concentrations above 200 mM NaCl, and is independent of the presence of divalent cations. The folded-␤ 1b interacts specifically with the highly conserved AID site because it binds to GST-AID␣ 1C fusion protein but not to GST alone or GST-ClC 126 . It also interacts with the ␣ 1C -subunit of the voltage gated-calcium channels expressed in Xenopus oocytes, showing that folded-␤ 1b retains biological activity.
When injected into ␣ 1C -subunit-expressing oocytes, the folded-␤ 1b shifts the activation curve toward more negative potentials and increases significantly the ratio of ionic current versus charge movement, whereas charge movement itself remains more or less invariant. Oocytes co-injected with ␣ 1C -and ␤ 1b -cRNA behave similarly. These results show that ␤ 1b modulates function through an interaction site that is equally accessible to ␤ 1b synthesized from cRNA or to folded-␤ 1b protein, and that channel density, as assessed by the gating current measurements, is not affected by the presence of exogenous ␤ 1b -subunit. The increase in ionic current versus charge movement ratio becomes apparent after 30 min elapsed from the injection of the folded-␤ 1b and it develops slowly to a single plateau 2 h later. The charge movement remains constant over the whole period of time after protein injection. Consequently, the effect of the ␤ 1b -subunit on channel activity must be a separate action from the effect on channel expression. This conclusion is also consistent with the recent finding that in spherical vesicles derived from frog and mouse skeletal muscle plasma membranes, ionic currents through calcium channels are increased without changes in charge movement by ␤ 1a purified from COS-transfected cells (24). On the other hand, the lack of effect of the ␤ 1b -subunit on channel expression observed here might be explained by sufficient amount of endogenous ␤-subunit (␤ 3xo ) that was available to release the ␣ 1C -subunit from the ER, bypassing the need of exogenous ␤ 1b -subunit. An alternative scenario to explain the increase in ionic current with no increase in channel density is that the folded-␤ 1b subunit binds to the ␣ 1C retained in the ER and gives rise to functional channels with increased ionic current versus charge movement. In such a case, and to maintain Q on , the ␣ 1C /␤ 1b -channels should replace virtually all ␣ 1C -or ␣ 1C /␤ 3xo -channels within the first 2 or 3 h after protein injection. Although it seems unlikely, specific experiments need to be carried out to fully address this possibility.
Two models have been proposed for the double action of the ␤-subunit (25): (i) the ␤-subunit interacts sequentially with the same ␣ 1 -subunit binding site at different stages of channel cycle to modify channel expression and later function while maintaining a one-to-one stoichiometry (one-site model); and (ii) the different functions result from binding at two different sites (two-site model). The second hypothesis implies that binding of the first ␤-subunit in the ER facilitates expression of the ␣ 1 -subunit and that binding of a second ␤-subunit modifies channel activity, as proposed by Tareilus et al. (26). Although our experiments cannot discard a two-to-one-stoichiometry, they clearly show that the binding surface responsible for the functional changes is available to the folded-␤ 1b protein despite the presence of an endogenous ␤, and thus, binding to this surface occurs downstream to the release from the ER, which favors the idea of a sequential binding.
In summary, we have successfully refolded the ␤ 1b -subunit of the voltage-gated calcium channel from E. coli inclusion bodies using SEC. The folded-␤ 1b subunit is monomer in solution and binds in vitro to the AID␣ 1C domain. Our data are consistent with the ␤ 1b -subunit acting primarily on calcium channels preexisting in the plasma membrane to regulate the ion conduction properties of the channel.