Regulation of Maximal Open Probability Is a Separable Function of Ca_vβ Subunit in L-type Ca²⁺ Channel, Dependent on NH₂ Terminus of α_1C (Ca_v1.2α)

Abstract

β subunits (Ca_vβ) increase macroscopic currents of voltage-dependent Ca²⁺ channels (VDCC) by increasing surface expression and modulating their gating, causing a leftward shift in conductance–voltage (G-V) curve and increasing the maximal open probability, P_o,max. In L-type Ca_v1.2 channels, the Ca_vβ-induced increase in macroscopic current crucially depends on the initial segment of the cytosolic NH₂ terminus (NT) of the Ca_v1.2α (α_1C) subunit. This segment, which we term the “NT inhibitory (NTI) module,” potently inhibits long-NT (cardiac) isoform of α_1C that features an initial segment of 46 amino acid residues (aa); removal of NTI module greatly increases macroscopic currents. It is not known whether an NTI module exists in the short-NT (smooth muscle/brain type) α_1C isoform with a 16-aa initial segment. We addressed this question, and the molecular mechanism of NTI module action, by expressing subunits of Ca_v1.2 in Xenopus oocytes. NT deletions and chimeras identified aa 1–20 of the long-NT as necessary and sufficient to perform NTI module functions. Coexpression of β_2b subunit reproducibly modulated function and surface expression of α_1C, despite the presence of measurable amounts of an endogenous Ca_vβ in Xenopus oocytes. Coexpressed β_2b increased surface expression of α_1C approximately twofold (as demonstrated by two independent immunohistochemical methods), shifted the G-V curve by ∼14 mV, and increased P_o,max 2.8–3.8-fold. Neither the surface expression of the channel without Ca_vβ nor β_2b-induced increase in surface expression or the shift in G-V curve depended on the presence of the NTI module. In contrast, the increase in P_o,max was completely absent in the short-NT isoform and in mutants of long-NT α_1C lacking the NTI module. We conclude that regulation of P_o,max is a discrete, separable function of Ca_vβ. In Ca_v1.2, this action of Ca_vβ depends on NT of α_1C and is α_1C isoform specific.

INTRODUCTION

Voltage-dependent Ca²⁺ channels are grouped into three families, Ca_v1–Ca_v3 (Ertel et al., 2000). The main structural component of all Ca_v channels is the α₁ subunit that bears the archetypal features of a voltage-dependent channel, with four membrane-spanning domains and a large cytosolic domain comprising the NH₂- and COOH-terminal parts of the protein (NT and CT, respectively), and three large intracellular loops, L1–L3, connecting the membrane-spanning domains (Fig. 1 A). In addition, members of Ca_v1 and Ca_v2 families also contain at least two auxiliary subunits, β (Ca_vβ1–Ca_vβ4) and α₂δ (Isom et al., 1994; Varadi et al., 1995; De Waard et al., 1996; Birnbaumer et al., 1998; Walker and De Waard, 1998; Striessnig, 1999; Catterall, 2000). The α₂δ subunit regulates channel expression and trafficking to the plasma membrane (PM) (Shistik et al., 1995; Yasuda et al., 2004; Canti et al., 2005), increases the open probability (P_o) of α_1C (Shistik et al., 1995), and regulates some pharmacological properties of the channel (De Waard et al., 1996). Ca_vβ subunits are modular MAGUK-type proteins with an SH₃-like and a guanylate kinase (GK)-like domain (Chen et al., 2004; McGee et al., 2004; Opatowsky et al., 2004; Van Petegem et al., 2004). The latter binds with high affinity to a conserved AID (α-interaction domain) motif within the first intracellular loop (L1) of α₁ (Pragnell et al., 1994). The β subunits profoundly modulate the properties of voltage-dependent Ca²⁺ channels. The most prominent effect is a great increase in the magnitude of macroscopic Ca²⁺ currents, caused by the expression of Ca_vβ on top of α₁ or α₁+α₂δ in most heterologous expression systems (Mori et al., 1991; Singer et al., 1991; Varadi et al., 1991; Williams et al., 1992; Castellano et al., 1993; Lory et al., 1993), or by the expression of Ca_vβ in cardiac cells (Wei et al., 2000; Colecraft et al., 2002). Accordingly, depletion or elimination of endogenous Ca_vβ subunits by knockdown/knockout strategies greatly reduces voltage-dependent Ca²⁺ currents in various excitable cells (Gregg et al., 1996; Strube et al., 1996; Leuranguer et al., 1998; Namkung et al., 1998; Chu et al., 2004).

Figure 1.

The structure of the α_1C subunit and its NH₂ terminus. (A) Schematic presentation of the α_1C subunit of the L-type Ca²⁺ channel, Ca_v1.2. Numbering is shown according to rabbit long-NT α_1C-wt (Mikami et al., 1989). The principal NH₂-terminal deletion mutants used in this work are indicated by short lines. (B) Comparison of protein sequences of NH₂ termini of rabbit and human long-NT and short-NT isoforms. Dots stand for full sequence homology, dashes show gaps, and shaded areas show the conserved motif TxxYxP. Arrows indicate the location of the inserted methionine at the beginning of each NH₂-terminal deletion mutant. The initial segments of long-NT and short-NT isoforms are encoded by exons 1a and exon 1, accordingly; the beginning of the transmembrane segment IS1 corresponds to the beginning of exon 3.

Extensive studies in heterologous expression systems revealed a multitude of effects of β subunits on biosynthesis and gating of VDCCs. First, Ca_vβ increases the surface expression of α₁, by improving its trafficking from the ER to the PM (Chien et al., 1995; Brice et al., 1997; Tareilus et al., 1997; Gao et al., 1999), probably by relieving an ER retention signal (Bichet et al., 2000). Second, Ca_vβ regulates several gating properties of VDCCs. The extent of regulation varies among subtypes and isoforms of Ca_vβ and Ca_vα₁. The most prominent changes in macroscopic currents include hyperpolarizing (leftward) shifts in current–voltage (I-V) (or conductance–voltage, G-V) and steady-state inactivation curves, an increase in the rate of activation, and changes in the kinetics of voltage-dependent inactivation (for reviews see Birnbaumer et al., 1998; Stotz and Zamponi, 2001; Dolphin, 2003). The shift in G-V curve reflects an improved coupling between gating charge movement and channel opening (Neely et al., 1993, 2004). On the single channel level, coexpression of Ca_vβ increases the open probability, P_o, without changing the single channel conductance (Wakamori et al., 1993; Shistik et al., 1995; Costantin et al., 1998).

The overall increase in macroscopic Ca²⁺ channel currents caused by heterologous expression of Ca_vβ results both from increased surface expression, and from changes in gating that lead to increased P_o. Regulation of trafficking by Ca_vβ is clearly separable from modulation of gating; changes in trafficking and gating occur on distinct time and Ca_vβ concentration scales, and mutations that disrupt the high-affinity interaction between AID and Ca_vβ disrupt colocalization of α₁ and β in the PM and membrane targeting of α₁, but spare some or all of the β-induced changes in gating parameters. Thus, the high-affinity binding of Ca_vβ to AID is obligatory for the regulation of trafficking but not gating (Yamaguchi et al., 1998; Canti et al., 1999, 2001; Gerster et al., 1999; Hullin et al., 2003; McGee et al., 2004; Leroy et al., 2005; Maltez et al., 2005). The SH₃–GK domain interaction may also play an important role in regulating trafficking (Takahashi et al., 2005). Among gating effects of Ca_vβ, at least one, the regulation of kinetics of voltage-dependent inactivation (VDI), is separable from the others. A palmitoylated isoform of β_2a (usually designated simply as β_2a) decelerates the VDI of several Ca_vα, whereas all other β subunits, including a nonpalmitoylated isoform of β_2a (np-β_2a of rabbit, and its human orthologue β_2b), accelerate VDI. At the same time, the enhanced trafficking of α₁ and the hyperpolarizing shift in G-V curve are independent of palmitoylation of Ca_vβ (Olcese et al., 1994; Chien et al., 1996; Qin et al., 1996, 1998; Gao et al., 1999). The separability of different effects of Ca_vβ implies that they are determined by distinct molecular interactions between different parts of β and/or α₁. Some actions of β may rely upon low-affinity interactions of β (the SH₃ domain in particular) with regions outside the AID in α₁, either in L1 or in the CT in some types of α₁ (Tareilus et al., 1997; Walker et al., 1999; Takahashi et al., 2004; Maltez et al., 2005). At present, it is unclear whether gating effects of Ca_vβ other than change in kinetics are also separable, and what is the molecular basis of separate effects of Ca_vβ on different gating parameters.

It is notable that cells most widely used for heterologous expression of Ca²⁺ channels, Xenopus oocytes and human embryonic kidney (HEK) cells, contain small but measurable amounts of an endogenous Ca_vβ protein that undoubtedly aids the “β-less” channels to reach the PM (Tareilus et al., 1997; Canti et al., 2001; Leroy et al., 2005). Arguably, the presence of a minimal amount of endogenous Ca_vβ may be obligatory for surface expression of at least some subtypes of Ca_v1 and Ca_v2 channels (Tareilus et al., 1997; Leroy et al., 2005). An absence of such endogenous Ca_vβ may explain the reports that in some cell lines transfected with α_1C alone or even with α_1C+ α₂δ, no functional Ca²⁺ channel expression is observed (Gao et al., 1999; Harry et al., 2004; Kobrinsky et al., 2004). The presence of the endogenous Ca_vβ, which is permissive for trafficking of α₁ to PM, does not impair the ability of coexpressed or exogenously added Ca_vβ protein to modulate the biophysical properties of the channel (Tareilus et al., 1997; Yamaguchi et al., 1998; Garcia et al., 2002). Therefore, it has been proposed that the endogenous β only “chaperones” α₁, helping it to leave the ER without staying with it in the PM; the added exogenous β then binds to α₁ and modulates the gating (the single Ca_vβ-binding model). An alternative multiple Ca_vβ-binding model contends that the endogenous β remains irreversibly bound to AID, and additional β subunit(s) modulate channel gating by interacting with other parts of α₁ (discussed by Birnbaumer et al., 1998; Jones, 2002; Dolphin, 2003). A recent study that used a β_2b subunit tethered to the end of α_1C strongly supports a functional 1:1 α₁-β stoichiometry (Dalton et al., 2005); unfortunately, not all functions of β were fully recovered by the tethered β, leaving this fundamental issue open for argument.

Unfortunately, the exact extent of regulation of different Ca_v channels by various Ca_vβ is still debated (discussed in Yasuda et al., 2004), and the contribution of different mechanisms to the increase in whole-cell current has not yet been precisely assessed. The variations are aggravated by apparent inconsistencies between results obtained in different cells and the use of different isoforms of Ca_vα₁ and Ca_vβ. To properly understand the molecular principles underlying the different actions of Ca_vβ, one needs to reliably monitor both the surface expression of α₁ and a defined set of gating parameters, in a well characterized system. In this report, we implemented such an approach to analyze the mechanism of regulation of Ca_v1.2 (α_1C), the L-type channel present in most excitable tissues, by β_2b. The study focused on the parameters of channel activity that lead to changes in the magnitude of the macroscopic Ba²⁺ current (I_Ba), leaving the regulation of inactivation out of scope. We have previously found that in the cardiac (long-NT) isoform of α_1C, the first 46 aa of the cytosolic NT constitute an NH₂-terminal inhibitory (NTI) module whose removal greatly increases the macroscopic current and the P_o. The presence of the NTI module is also crucial for β_2b-induced increase in I_Ba via the long-NT isoform α_1C in Xenopus oocytes. We therefore proposed that the β subunit acts, in part, by functionally counteracting the inhibitory effect of this module (Shistik et al., 1998; Ivanina et al., 2000). Here we demonstrate that the long-NT initial segment selectively regulates a single action of β_2b: the elevation of P_o,max. This modulation is absent in deletion mutants lacking the initial NT segment, and in a short-NT isoform of α_1C (smooth muscle/brain subtype), in which the initial 46 aa encoded by exon 1a are replaced by a partially homologous stretch of 16 aa encoded by exon 1. Other effects of β (trafficking, G-V curve shift) are preserved. This is the first report on a discrete regulation by Ca_vβ of P_o,max in Ca_v1.2. These findings bear upon the isoform-specific physiological properties of the L-type Ca²⁺ channel and upon the physiological role of Ca_vβ in different tissues.

MATERIALS AND METHODS

DNA Constructs and mRNA

cDNAs of rabbit heart α_1C (X15539), rabbit heart np-β_2a (L06110) (here termed β_2b), skeletal muscle α₂δ-1 (P13806) subunits, and the α_1C NH₂-terminal truncation mutants α_1CΔ5, α_1CΔ20, α_1CΔ46, and α_1CΔ139 were prepared and used as described previously (Shistik et al., 1998). The mutants Δ6-20, Δ21-46, T₁₀A/Y₁₃F/P₁₅A, NT_SL, and NT_LS chimeras were constructed by standard PCR methods. To create α_1C-HA, the hemagglutinin (HA) tag (SRYPYDVPDYA; the first two amino acids SR were added in order to create an XbaI restriction site in the corresponding cDNA sequence) has been inserted into the extracellular loop after the S5 transmembrane segment of α_1C-wt, between amino acids Q713 and T714, by two consecutive PCRs. The Δ21-46-HA and NT_LS-HA were then produced by standard subcloning procedures. All mutations and PCR products were verified by nucleotide sequencing at the Tel Aviv University Sequencing Facility.

All cDNA constructs of α_1C and mutants were inserted into the same vector, pGEM-HE-GSB (Shistik et al., 1998), which is a derivative of pGEM-HE (Liman et al., 1992). This vector provides the necessary 5′ and 3′ untranslated regions (UTR) from Xenopus α-globin. Therefore, only the coding sequences of all α_1C derivatives, without any residual original UTRs, were inserted into the vector. To minimize any variability that may be caused by variations in quality and quantity of RNAs, in each series of experiments all tested RNAs (all mutants of α_1C under study and the wt α_1C) were synthesized anew on the same day.

Oocyte Culture and Electrophysiology

All the experiments were performed in accordance with the Tel Aviv University Institutional Animal Care and Use Committee (permits no. 11–99-47 and 11–05-064). Xenopus laevis frogs were maintained and operated, and oocytes were collected, defolliculated, and injected with RNA as previously described (Dascal and Lotan, 1992). In brief, female frogs, maintained at 20 ± 2°C on an 11-h light/13-h dark cycle, were anaesthetized in a 0.15% solution of procaine methanesulfonate (MS222), and portions of ovary were removed through a small incision on the abdomen. The incision was sutured, and the animal was returned to a separate tank until it had fully recovered from the anesthesia, and afterwards was returned to a large tank where, together with the other postoperational animals, it was allowed to recover for at least 4 wk until the next surgery. The animals did not show any signs of postoperational distress. The oocytes were injected with the mRNAs of α_1C or its mutants, α₂δ, and β_2b, according to the design of experiment (0.3–5 ng for electrophysiology and 2.5–5 ng for imaging). Unless indicated otherwise, equal amounts (by weight) of different RNAs were injected. The oocytes were incubated for 3–5 d at 20–22°C in ND96 solution (96 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 5 mM HEPES, pH 7.5) supplemented with 2.5 mM Na-pyruvate and 50 μg/ml gentamycin. In some batches with high endogenous chloride currents, oocytes were injected with 25–30 nl/oocyte of the Ca²⁺ chelator EGTA (50–100 mM), giving a final concentration of 2–5 mM within the oocyte (assuming oocyte's free water volume of 0.5 μl).

Whole cell currents were recorded using the Gene Clamp 500 amplifier (Axon Instruments) using the two-electrode voltage clamp technique, in a solution containing 40 mM Ba(OH)₂, 50 mM NaOH, 2 mM KOH, and 5 mM HEPES, titrated to pH 7.5 with methanesulfonic acid (Shistik et al., 1995). Stimulation, data acquisition, and analysis were performed using pCLAMP software (Axon Instruments). Current–voltage (I-V) relation of Ba²⁺ currents was measured by 60-ms, 10-mV steps given every 10 s from a holding potential of −80 mV. In each cell, the net I_Ba was obtained by subtraction of the residual currents recorded with the same protocols after applying 200 μM Cd²⁺.

Immunocytochemistry and Confocal Imaging

Immunocytochemistry in giant PM patches was done essentially as previously described (Singer-Lahat et al., 2000; Peleg et al., 2002), as illustrated in Fig. 3. Oocytes were devitellinized by peeling off the vitelline membrane in ND96, and placed on plastic coverslips (Thermanox plastic coverslip; Nunc). After sticking to the coverslip, the oocyte was removed mechanically and/or by washing with a strong jet of solution. Pieces of membranes strongly attached to the coverslip were continuously washed until the membrane patch became transparent without any visible cytosolic content and pigment granules. After fixation for 10 min in 1% formaldehyde, the membranes were washed three times with 1% BSA dissolved in TBS solution (135 mM NaCl, 10 mM Tris-HCl, pH 7.4). Blocking of nonspecific binding sites was done with donkey immunoglobulin G (IgG, whole molecule, 1/200, Jackson ImmunoResearch Laboratories) for 30 min. Each coverslip was incubated for 1 h with CT4 antibody against α_1C COOH terminus (1:500; provided by M. Hosey, Northwestern University, Chicago, IL; see Gao et al., 2001) or against L2 (1:500; Alomone Labs.). Residual antibody was washed out with 1% BSA three times, 5 min each. This was followed by a 30-min incubation with secondary antibody (Cy3 donkey anti–rabbit IgG, 1:400; Jackson ImmunoResearch Laboratories). Free secondary antibody was then washed out with 1% BSA three times, 5 min each in darkness and the coverslips were mounted on a glass slide. The fluorescent labeling was examined by a confocal laser scanning microscope (LSM 410 or LSM 510, Zeiss, Germany). 40× NA/1.2 C-apochromat water-immersion lens (Axiovert 135 M, Zeiss) was used for imaging. The Cy3-conjugated secondary antibody was excited at 488 nm and the emitted light at >568 nm was collected. The fluorescent signals were analyzed by measuring total luminosity (optical density) of the whole image using the Tina 2.1 (Raytest Isotopenmelgerilte GmbH) or Carl Zeiss MicroImaging, Inc. LSM5 software. In all confocal imaging procedures, care was taken to completely avoid saturation of the signal. The gain of the photomultiplier was kept <75% of maximum. In each experiment, all oocytes from the different groups were studied using a constant set of imaging parameters. The normalized intensities were always calculated relative to the control group of the same experiment. Net fluorescence intensity per unit area was obtained by subtracting an averaged background signal measured in the same way in membranes of native (uninjected) oocytes from the same batch.

Figure 3.

NT mutations and deletions do not alter the surface expression of α_1C. (A) A simplified presentation of the method used to measure the surface expression of α_1C in giant membrane patches (see Materials and Methods for details). A devitellinized intact oocyte is placed for 10–20 min on coverslip with its animal (dark) hemisphere facing the coverslip (a). After attachment, the oocyte is swept away and patches of membrane (usually >100 μm in diameter) remain stuck to the coverslip (b). The cytosolic surface of the PM, facing the external solution, is thoroughly washed until the membrane appears transparent (c) and then stained with an antibody directed against a cytosolic part of the channel. (B) Examples of confocal images of α_1C-wt, α_1CΔ46, α_1CΔ20, and α_1CTYP in giant patches of oocytes of the same batch. (C) NT mutations do not alter the surface expression of α_1C (a), despite the typical differences in I_Ba (measured at +20 mV) in oocytes of one of these batches (b). The amount of injected RNA was 5 ng/oocyte.

Immunocytochemistry and imaging of whole oocytes has been done as follows. 3–4 d after the injection of RNA, the oocytes were fixated in 4% formaldehyde (37%) in Ca-free ND96 solution for 15 min. Blocking of nonspecific binding sites was done by 5% skim milk for 1 h. Then the oocytes were incubated for 1 h with the mouse monoclonal IgG_2a antibody against HA (Santa Cruz Biotechnology), diluted 1:400 in 2.5% skim milk. Residual antibody was washed out with 2.5% skim milk three times, 5 min each. This was followed by 1 h incubation with the secondary antibody (Alexa-conjugated anti–mouse IgG, 1:400; Jackson ImmunoResearch Laboratories) in dark. Free secondary antibody was then washed out with Ca-free ND96. Oocytes were placed in a chamber with a transparent bottom, and fluorescence imaging of optical slices was performed with LSM 510 (×20 objective, zoom = 1, pinhole 3 Airy units). Alexa was excited at 594 nm and the emitted light was collected using long-pass (LP) 615-nm filter. The fluorescent signals were usually analyzed as described in Fig. 4 D. Alternatively, the intensity of fluorescence in the PM was measured by averaging the signal obtained from six standard circular regions of interest. Net fluorescence intensity per unit area was obtained by subtracting the background signal measured in native oocytes.

Figure 4.

Coexpression of β_2b increases the surface expression of α_1C-wt and of NT mutants. (A) Examples of confocal images of α_1C-wt, α_1CΔ46, α_1CΔ20, and α_1CTYP in giant PM patches, in absence and presence of β_2b subunit. (B) Surface expression of the mutant proteins in α₁α₂δ subunit composition, without β_2b, normalized to α_1C-wt. This series of experiments included two oocyte batches and was separate from that shown in Fig. 3. Injected RNA was 5 ng/oocyte. (C) Summary of the effects of β_2b on the surface expression of α_1C mutants. For each construct, the surface expression in the presence of β_2b was normalized to the average expression measured without β_2b. (D) The effect of β_2b on surface expression of α_1C-HA; panel a shows representative confocal images of whole oocytes, either native (uninjected) or expressing α_1C-HA + α₂/δ (“α_1C-HA”) or α_1C-HA + α₂/δ + β_2b (“α_1C-HA+β_2b”). Injected RNA was 5 ng/oocyte. The dimensions of the image are 0.45 × 0.45 mm. The rightmost image exemplifies the method used to analyze the intensity of fluorescent labeling. A rectangle of a fixed width was superimposed on the image, and a profile of optical density (shown in b) along the axis approximately perpendicular to the PM (arrow) was obtained using the TINA software. (Optical density and distance were measured in arbitrary units inherent to the software). (b) Quantitative analysis of optical density profiles. The intensity was defined as the integral of the shaded area. The width of this area was defined as constant for all images taken in an experiment. The net intensity in each α_1C-HA–expressing oocyte was calculated by subtracting the average intensity measured in uninjected oocytes of the same experiment, and then normalized to the average net intensity of [α_1C-HA+α₂/δ]-expressing oocytes of this experiment. The normalized values, shown in c, were summarized across all experiments (N = 3).

Data Analysis

Current–voltage (I-V) curve was fitted to the Boltzmann equation in the form

(1)

where K_a is the slope factor, V_1/2 is the voltage that causes half maximal activation, G_max is the maximal macroscopic conductance, V_m is membrane voltage, I_Ba is the current measured at the same voltage, and V_rev is the reversal potential of I_Ba. The obtained parameters of G_max and V_rev were then used to calculate fractional conductance at each V_m, G/G_max, using the equation

(2)

where G is the total macroscopic conductance at V_m. The conductance–voltage (G-V) curves were plotted with the values of V_1/2 and K_a obtained from the fit of the I-V curves, using the following form of the Boltzmann equation:

(3)

The results were summarized from many groups of oocytes from different donors (“batches”). To avoid series resistance artifacts associated with very large currents (Schreibmayer et al., 1994), or inaccuracies resulting from a contribution from the small endogenous oocyte's Ca²⁺, Cl⁻, or K⁺ channel currents when the macroscopic I_Ba was low (mainly in oocytes expressing α_1C+ α₂δ without β_2b; Dascal, 1987; Dascal et al., 1992), these quantitative analyses were performed only in experiments in which, at the peak of the I-V curve, 0.1 μA ≤ I_Ba ≤ 6 μA.

The calculation of changes in P_o,max was based on the following considerations. The total macroscopic conductance, estimated by measuring peak currents at different voltages, is a linear function of P_o:

(4)

where γ is the single channel conductance, and N is the total number of functional channels in the PM (Hille, 2002). In Ca_v channels, P_o is voltage dependent whereas γ and N are not. P_o reaches a maximal value, P_o,max, at positive voltages where a maximal macroscopic conductance, G_max, is attained. Both G_max and P_o,max are empirical parameters that are considered voltage independent and are interrelated as follows:

(5)

The fold change in N caused by a treatment, R_N, is defined as N_treatment/N_control. From here, if γ is constant, the change in P_o,max caused by a treatment is given by

(6)

In this study, R_N was monitored by imaging measurements in giant PM patches or in whole oocytes. Eq. 6 is similar to those used previously (Wei et al., 1994; Takahashi et al., 2004) to assess changes in P_o; in these studies, R_N was estimated from the measurement of Q_max (the maximal gating charge).

To compare parameters (I₄₀, surface α_1C labeling, or G_max) observed or calculated in oocytes of different batches, data from individual cells were averaged across batches using the following normalization procedure (Sharon et al., 1997). The value of the parameter in each cell was normalized to the mean value of this parameter in the control group of the same batch. This procedure was also applied to the cells of the control group, thus providing a useful measure of variability in this group and enabling an accurate statistical analysis. Normalized values were then averaged from all cells across all batches. The results are presented as means ± SEM Multiple group comparisons were done by one-way analysis of variance (ANOVA) test followed by Tukey's test. Two-group comparisons were done using Student's t test.

Online Supplemental Material

Table S1 (available at http://www.jgp.org/cgi/content/full/jgp.200609485/DC1) shows the mean values of G_max and V_rev obtained in Boltzmann equation fits for some of the α_1C constructs used in this study.

RESULTS

Mapping the Part of NT Crucial for the Regulation of G_max

Deletions of initial segment of long-NT α_1C increase the amplitude of I_Ba but attenuate the current enhancement caused by Ca_vβ. It is not known if the same structural elements of NT alter both macroscopic currents and modulation by Ca_vβ. To address this problem, we sought to delineate the regions that are important for these modulations. Fig. 1 B shows the protein sequences of the NH₂-terminal part of long-NT and short-NT isoforms of rabbit and human α_1C (the corresponding sequences in rat and mouse are also highly homologous; see Blumenstein et al., 2002). The cytosolic NT consists of 124 (short-NT) or 154 aa (long-NT), which are the products of transcription of alternative initial exons 1 or 1a, respectively, and an invariable exon 2 encoding 108 aa. Amino acids 6–20 of the long-NT show partial homology with the first 16 aa of the short-NT (Shistik et al., 1998); in particular, the amino acids T₁₀, Y₁₃, and P₁₅ (which we call TxxYxP motif; numbering by long-NT) are conserved in both isoforms (Fig. 1 B, gray boxes). In the previous studies, we have created several NT deletion mutants of rabbit cardiac long-NT α_1C. The logic of these deletions was dictated by the comparison of nucleotide sequences of long- and short-NT isoforms. Initially, we deleted amino acid stretches from the first methionine up to the positions corresponding to the boundaries of the region of partial homology (mutants α_1CΔ5 and α_1CΔ20), and then up to the beginning of the high homology region (α_1CΔ46). We also deleted almost all of the cytosolic part of the NT (α_1CΔ139). In the previous studies, only α_1CΔ46 and α_1CΔ139 have been characterized in detail; voltage-dependent characteristics, protein expression levels, and the effect of β subunit have not been compared systematically in most mutants. Such a comparative study has been done in the experiments described below. In addition, we have deleted the region of partial homology with short-NT α_1C, aa 6–20, from long-NT α_1C (mutant α_1CΔ6-20). We also mutated the TxxYxP motif, creating the T₁₀A/Y₁₃F/P₁₅A (α_1CTYP) mutant.

To delineate the region of long-NT that regulates the magnitude of I_Ba, either long-NT wild-type α_1C (α_1C-wt) or the various NT deletion mutants were expressed without Ca_vβ. The expression of α_1C-wt alone in Xenopus oocytes results in very small (a few nA) Ba²⁺ currents that are difficult to reliably resolve and analyze, and often difficult to separate from currents via oocyte's endogenous Ca_v channels, which are non–L type (Singer et al., 1991; Dascal et al., 1992; Singer-Lahat et al., 1994). Therefore, α₂δ was coexpressed in all experiments. The latter greatly increases I_Ba compared with α_1C-wt alone, but does not increase the current via the endogenous channels. α₂δ aids trafficking α_1C to the PM (Shistik et al., 1995; Yasuda et al., 2004; Canti et al., 2005) and produces certain synergistic (more than additive) effects with some β subunits (Singer et al., 1991; Yamaguchi et al., 2000), but most of the actions of β_2b are similar in the presence or absence of α₂δ (see Table IV). In oocytes expressing α_1C+α₂δ or α_1C+α₂δ+β_2b, the contribution of endogenous channels to I_Ba was negligible; I_Ba was reduced by >95% by 10 μM of the dihydropyridine L-type Ca²⁺ channel blockers nitrendipine and nifedipine (Singer-Lahat et al., 1994; unpublished data).

Table IV.

A Summary of Published Quantitative Analyses of the Effects of Coexpression of Ca_vβ Subunits on Ca_v1.2 Channels in Heterologous Expression Systems

Surface expression, I-V shift, and P_o are usually presented for α_1c+α₂δ+β subunit combination vs. α_1c+α₂δ, unless indicated otherwise. Mammalian cells were derivatives of HEK, COS, or similar cell lines lacking significant amounts of an endogenous α_1C. In one study (Colecraft et al., 2002), β subunits were overexpressed in cardiac myocytes.

^aChanges in N were measured from Q_max. We have calculated P_o,max using Eq. 6 and the presented data.

^bNo α₂δ present.

^cEstimated from current increase at peak, by comparing early (1 h) and late times after the injection of the β₃ protein; the shift in I-V is already complete by 1 h, and the difference in peak current is only due to surface expression (see Fig. 4 in Yamaguchi et al., 1998).

^dPurified β protein was injected into oocytes.

^eBy imaging an intracellular HA tag added to α_1C at the NT. Oocytes were permeabilized with Triton-X100 and stained with an anti-HA antibody.

^fIn the presence of a dihydropyridine agonist.

^gFrom total membrane [³H]PN200-110 binding. This method does not exclude the labeling of α_1C located in ER or Golgi.

^hBoth by immunoprecipitation of α_1C from separated PM and by counting channels in cell-attached patches.

ⁱBy counting channels in patches. Measurement by surface biotinylation confirmed the increase in N, but there is no quantitative estimate.

^jβ_2a was overexpressed in cardiomyocytes on the background of existing cardiac channels, probably the long-NT α_1C type (Blumenstein et al., 2002; Pang et al., 2003). P_o,max was measured in single channel records at +40 mV.

^kA deletion mutant of the rabbit long-NT α_1C missing the first 60 aa.

^lThe types of α_1C are defined as follows: Mikami (Mikami et al., 1989) is the first cloned long-NT. This is the rabbit cardiac isoform, used also in the present study (termed α_1C-wt). Identical or nearly identical cDNAs were later cloned by Wei et al. (1991), Biel et al. (1991) and Lory et al. (1993).^m Schultz (Schultz et al., 1993) is a human α_1C bearing 93% similarity to Mikami's clone but missing the first 59 aa at the NT (the protein starts with Met₆₀ encoded by the first in-frame ATG of exon 2). Biel (Biel et al., 1990) is a rabbit lung short-NT α_1C with the 16-aa initial segment, and variations in a few downstream exons compared to Mikami's clone. Snutch (Snutch et al., 1991) is a short-NT (16 aa initial segment) rat cDNA; the most widely used is the rbII clone.

^mThe cDNA is identical to that of Mikami et al. (1989) except for an alternative exon encoding for a part of IVS3.

Fig. 2 A shows the routine experimental protocol used to analyze the I-V characteristic of the expressed channels. Ca²⁺ channel currents were elicited by 60-ms depolarizing pulses from a holding potential of −80 mV, in 10-mV steps, in a solution containing 40 mM Ba²⁺ using the two-electrode voltage clamp technique. Net I_Ba was obtained by subtracting currents elicited by the same voltage protocol in the presence of 200 μM CdCl₂ (Fig. 2 A, left). The absolute amplitude of I_Ba depended on the amount of RNA injected, period of incubation, and also varied among oocyte batches. The amount of the injected RNA (of both α_1C-wt and α₂δ) varied in our experiments between 1 and 5 ng/oocyte, depending on the experimental design. In this range, greater amounts of RNA always resulted in larger I_Ba. The right panel of Fig. 2 A shows a typical I-V curve of net I_Ba, averaged from six oocytes of the same batch (donor frog). Without Ca_vβ, I_Ba was maximal at ∼30 mV. In the following, to compare I_Ba across different groups and treatments, we chose to use the value of I_Ba measured at +40 mV (I₄₀) rather than at the peak of the I-V curve, because the latter varied between different treatments. Also, at +40 mV, the whole-cell Ba²⁺ conductance is close to maximum under most conditions (Figs. 2 and 4), therefore changes in I₄₀ should approximately reflect changes in G_max.

Figure 2.

Effects of NH₂-terminal deletions and T₁₀A/Y₁₃F/P₁₅A mutation on I_Ba in the absence of Ca_vβ (the α_1Cα₂δ subunit composition). (A) The standard procedure used to monitor I_Ba, and an averaged I-V curve. See explanations in the text. (B) Comparison of I_Ba in α_1C-wt and α_1CΔ5 (2.5 ng RNA/oocyte of each subunit). Panel a shows representative current traces recorded at +40 mV in oocytes of one batch (donor). A current trace obtained in an oocyte expressing the α₁Δ46 mutant is shown for comparison. Panel b shows averaged I-V curves from α_1C-wt and α₁Δ5 groups (n = 9 oocytes, N = 2 batches). The normalized values of I₄₀ are shown in panel c. (C) Panel a shows the summary of relative I₄₀ in the various mutants. I₄₀ in each oocyte was normalized to the average I₄₀ of α_1C-wt of the same batch, as explained in Materials and Methods. Panels b and c show normalized I-V and G-V curves, respectively, of the indicated channel constructs, averaged from oocytes of two representative batches (n = 8–14, N = 2). Amount of injected RNA was varied from 0.3 ng/oocyte in NT mutants to 5 ng/oocyte in α₁-wt to reach comparable I_Ba amplitudes in order to minimize the fitting artifacts. Note that Boltzmann fits have been performed separately in each oocyte. For illustration purposes, in averaged I-V and G-V curves shown in C and in the following figures, the solid lines through averaged experimental points were drawn using Boltzmann equation with values of V_1/2 and K_a from Table I. In this and the following figures, the numbers above bars indicate the number of cells tested, and asterisks indicate statistically significant differences, as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

As reported previously (Shistik et al., 1999), in the α_1CΔ5 mutant, I₄₀ increased about twofold compared with α_1C-wt, without any significant change in the voltage dependence of activation (Fig. 2 B). The calculated increase in G_max in the α_1CΔ5 mutant was 66% (Table II). In comparison, the deletion of 20 or more aa of the long-NT initial segment caused a robust seven to ninefold increase in I₄₀ and in the calculated G_max (Fig. 2, Ba and Ca, and Table II). For a summary of the absolute values of G_max in the different groups, injected with the same RNA amount of 1 ng/oocyte, see Table S1 (available at http://www.jgp.org/cgi/content/full/jgp.200609485/DC1).

TABLE II.

Relative Changes in Surface Expression, G_max, I₄₀, and P_o in NT Mutants in α_1C+α₂/δ Subunit Composition in the Absence of Ca_vβ

	Fold increase in PM α1C labeling			Fold increase in Po
Construct		Fold increase in Gmax	Fold increase in I40	from Gmax	from I40
α1C-wt	1 ± 0.05 (40; 7)	1 ± 0.04 (56; 11)	1 ± 0.07 (56; 11)
α1CΔ5	1.21 ± 0.18 (6; 2)	1.66 ± 0.08c (9; 2)a	1.76 ± 0.24b (9; 2)a	1.37	1.45
α1CΔ20	1.22 ± 0.12 (24; 3)	7.64 ± 0.98d (30; 6)	7.68 ± 0.34d (30; 6)	6.26	6.3
α1CΔ46	0.99 ± 0.12 (16; 3)	7.63 ± 0.81d (30; 6)	7.75 ± 0.95d (30; 6)	7.7	7.82
α1CΔ139	Not examined	6.81 ± 0.93d (19; 5)	6.88 ± 0.94d (19; 5)
α1CTYP	0.86 ± 0.1 (17; 3)	5.71 ± 0.93d (23; 3)	5.95 ± 0.95d (23; 3)	6.63	6.91
α1CΔ6-20	Not examined	8.22 ± 1.01d (14; 3)	8.6 ± 1.23d (14; 3)
α1C-short	1.16 ± 0.14 (9; 2)	8.56 ± 0.98d (12; 2)	8.71 ± 1.71d (12; 2)	7.38	7.5
NTSL	1.13 ± 0.19 (11; 2)	5.77 ± 0.56d (16; 3)	5.98 ± 0.74d (16; 3)	5.1	5.3
HA-labeled constructs
α1C-HA	1 ± 0.12 (10; 1)	1 ± 0.11 (9; 3)	1 ± 0.31 (9; 3)
NTLS-HA	1.15 ± 0.09 (11; 1)	13.6 ± 1.85d (15; 3)	14.08 ± 2.0d (15; 3)	11.8	12.2
α1CΔ21-46-HA	0.88 ± 0.09 (11; 1)	1.41 ± 0.29d (9; 3)	1.47 ± 0.29d (9; 3)	1.6	1.67

Surface expression was measured by confocal imaging of immunostained α_1C in giant membrane patches or, for HA-labeled constructs, in intact oocytes. Numbers in parentheses indicate n; N (the total number of oocytes and the number of oocyte batches tested). To calculate fold change in a parameter, normalization procedure was performed as described in Materials and Methods. Statistical comparisons were performed using one way ANOVA (multiple comparisons versus control group, α_1C-wt or α_1C-HA, in oocytes injected with the same amount of RNA, followed by Tukey's test).

^aIn two batches where α_1C-wt and α_1CΔ5 were compared, the statistical analysis has been done separately from other batches, since the amount of injected RNA was 2.5 ng per oocyte rather than 1 ng. Averaged normalized fold increase in G_max and I₄₀ for α_1C-wt in these two batches was 1 ± 0.16 (n = 9, N = 2) and 1 ± 0.2 (n = 9, N = 2), respectively. The statistical analysis was done using t test.

^bP < 0.05.

^cP < 0.01.

^dP < 0.001.

As reported earlier (Wei et al., 1996; Shistik et al., 1998), the I-V curves in α_1CΔ46 and α_1CΔ139 appeared similar to α_1C-wt. However, an exhaustive comparison of a large amount of records revealed a small hyperpolarizing shift caused by the NT deletions, as illustrated in normalized averaged I-V curves (Fig. 2 Cb). To assure that the shift was not an artifact resulting form a larger current amplitude in the NT deletion mutants, in two batches (14–16 oocytes) oocytes were injected with amounts of RNA adjusted to produce I_Ba of similar amplitude in α_1C-wt, α_1CΔ46, and α_1CΔ20. The I-V curve shift was observed in all cases. Fitting I-V curves to Boltzmann equation (solid lines in Fig. 2 Cb) indicated a statistically significant 5–7 mV shift in the half activation voltage, V_1/2, in all NT deletion mutants studied except α_1CΔ5 (Table I). The leftward shift in the G-V curves is more clearly seen in conductance–voltage (G-V) curves that were drawn through the data points using the Boltzmann equation parameters obtained in I-V curve fits (Fig. 2 Cc). In addition, the slope appears to be increased by the NT deletions. Indeed, the slope factor K_a was slightly but statistically significantly reduced, from 9.6 mV in α_1C-wt to ∼7.9 mV in most mutants (Table I). None of the mutations caused any changes in the reversal potential of I_Ba (see Table S1).

TABLE I.

Effects of NT Deletions and Mutations of α_1C on the Gating Parameters of Ba²⁺ Current Activation

	V1/2, mV		Ka, mV
Construct	No β2b	With β2b	No β2b	With β2b
α1C-wt	15.66 ± 1.23 (56)	1.49 ± 0.79 (56)d	9.62 ± 0.31 (56)	7.2 ± 0.13 (56)d
α1Δ5	15.83 ± 3.17 (6)	0.91 ± 0.37 (9)d	9.48 ± 0.92 (6)	5.59 ± 0.08 (9) b d
α1Δ20	9.28 ± 0.98 (19)b	−1.14 ± 1.14 (19)d	7.87 ± 0.18 (19)b	6.41 ± 0.23 (19)b d
α1Δ46	9.49 ± 0.99 (30)b	−1.40 ± 0.80 (20)d	7.88 ± 0.24 (30)b	5.87 ± 0.18 (20)b d
α1Δ139	11.02 ± 1.00 (16)a	−1.83 ± 3.58 (13)d	7.93 ± 0.15 (16)b	5.38 ± 0.35 (13)b d
α1Δ6-20	8.61 ± 1.56 (8)	−1.28 ± 2.21 (14)c	7.83 ± 0.23 (8)	6.29 ± 0.23 (14)b d
α1TYP	10.28 ± 0.94 (23)b	−2.36 ± 0.57 (12)d	8.88 ± 0.24 (23)	6.13 ± 0.27 (12)b d
α1C-short	9.83 ± 1.28 (22)b	−2.19 ± 1.20 (26)d	7.15 ± 0.28 (22)b	6.06 ± 0.19 (26)b c
α1Δ21-46	20.75 ± 1.73 (9)	5.68 ± 1.93 (12)d	9.15 ± 0.33 (9)	6.94 ± 0.11 (12)d
NTSL	9.77 ± 1.40 (13)a	−1.89 ± 2.59 (18)d	8.03 ± 0.17 (13)b	5.95 ± 0.3 (18)b d
NTLS	13.86 ± 2.18 (13)	−3.59 ± 2.9 (9)d	6.58 ± 0.97 (13)b	4.45 ± 0.62 (9)b c
HA-labeled constructs
α1C-HA	17.88 ± 0.59 (6)	1.43 ± 1.38 (6)d	8.53 ± 0.21 (6)	6.55 ± 0.27 (6)d
α1CΔ21-46-HA	18.37 ± 1.08 (7)	3.6 ± 1.16 (6)d	8.34 ± 0.39 (7)	6.8 ± 0.17 (6)d
NTSL-HA	7.12 ± 1.67 (6)b	−3.06 ± 1.54 (6)d	6.76 ± 0.4 (6)b	4.89 ± 0.49 (6)a d

The values of V_1/2 and K_a were obtained in each cell by fitting the I-V curve to the Boltzmann equation (Eq. 1 in Materials and Methods). Mean ± SEM are shown, with numbers of cells in parentheses.

^aStatistically significant difference in V_1/2 or K_a as compared with α_1C-wt (or α_1C-HA) by one-way ANOVA followed by the Dunnett's test. P < 0.05.

^bStatistically significant difference in V_1/2 or K_a as compared with α_1C-wt (or α_1C-HA) by one-way ANOVA followed by the Dunnett's test. P < 0.01.

^cStatistically significant difference compared to no β_2b group by t test. P < 0.01.

^dStatistically significant difference compared to no β_2b group by t test. P < 0.001.

The robust increase in I_Ba and the shift in G-V curve were also observed in the α_1CΔ6-20 mutant lacking the 16 aa of the partial homology region. Mutation of the conserved TxxYxP motif produced similar but milder changes (Fig. 2 and Table I). The changes in G_max (calculated from the Boltzmann fits) were similar to changes in the measured I₄₀ (Table II). Taken together, these data imply that amino acids 6–20 of the long-NT isoform constitute a crucial part of the NTI module. The increase in I_Ba and in G_max is accompanied by a moderate hyperpolarizing shift in the G-V curve; this effect is also fully dependent on the presence of the same 16 aa. The T₁₀A/Y₁₃F/P₁₅A mutation interferes with the function of the NTI module by a mechanism yet to be determined.

Previously, controversial results have been reported regarding the surface expression of α_1C with NT deletions of 40 aa or more from a long-NT α_1C, expressed in Xenopus oocytes. Wei et al. (1996) reported an approximately sixfold increase in surface expression on the basis of measurements of Q_max in cut-open oocytes. In contrast, we did not observe any changes in surface expression in α_1CΔ46 either by counting channels in cell-attached patches, or by immunoprecipitating α_1C from manually separated plasma membranes (Shistik et al., 1998). Here we confirm these results by an independent imaging method (Singer-Lahat et al., 2000; Peleg et al., 2002). Proteins expressed in the PM were visualized by immunostaining in giant membrane patches in which the cytosolic surface of the PM is exposed to external solution (Fig. 3 A). The membrane protein in the patch is labeled with an antibody against a cytosolic segment, then with a fluorescently labeled secondary antibody, and visualized using a confocal microscope (see Materials and Methods for details). The image captures a randomly selected 105 × 105 μm area within a (larger) patch. The large size of the imaged area ensures a fair averaging of channel density even if the channels are clustered. Patches from many tens of oocytes can be screened during a 1-d experiment, providing statistically reliable data. This is a big advantage over the previously employed method of immunoprecipitation of α_1C from manually separated PM of metabolically labeled oocytes (Shistik et al., 1995, 1998), which produced one experimental point for 20–30 whole-oocyte membranes.

Fig. 3 B shows images of representative membrane patches from oocytes of one batch expressing α_1C-wt or various NT mutants of α_1C without the β subunit. The antibody labeling was clearly detected in all channel constructs, and no significant labeling was observed in uninjected oocytes at these imaging parameters. The data from two oocyte batches are summarized in Fig. 3 Ca, showing that there were no significant differences in the expression of any of the channel variants tested. This result was reproduced later in a separate series of experiments (see Fig. 4, A and B) and also using an alternative imaging method (see Fig. 7 C). At the same time, I_Ba measured in oocytes of one of these two batches was enhanced by the NT mutations, as usual (Fig. 3 Cb). We conclude that physical or functional removal of the NTI module of the long NH₂ terminus does not alter the surface expression of α_1C. Since the single channel conductance is not changed either (Shistik et al., 1998), all of the increase in G_max caused by these mutations must result from an increase in P_o,max. Indeed, calculations using Eq. 6 show that all these mutations cause a greater than sixfold increase in P_o,max (Table II), except α_1CΔ5, which shows only a 37% increase. The increase in I₄₀ was essentially identical (Table II).

Figure 7.

Design and surface expression of α_1C-short, short/long NT chimeras, and α_1CΔ21-46. (A) The design of initial NT segments of new constructs. Dashes show gaps. The homologous 15-aa sequence in long- and short-NT is highlighted in yellow and magenta, respectively. (B and C) Surface expression of the NT chimeric and deletion constructs in the absence and presence of β_2b monitored in giant membrane patches (B) or in intact oocyte by imaging the external HA tag (C). The amount of injected RNA was 2.5 ng/oocyte in B and 5 ng in C. Examples of confocal images are shown in panel a, and the summary of the surface expression in the absence of β subunit (relative to α_1C-wt) is shown in b. The effect of coexpression of β_2b for each one of the constructs tested in these experiments is summarized in c. Numbers above bars indicate the number of patches (B) or oocytes (C) imaged.

Regulation by β_2b Subunit of P_o,max, But Not of Other Parameters, Depends on NT of α_1C

We used the nonpalmitoylated np-β_2a (Hullin et al., 1992), which is the rabbit orthologue of human β_2b (96% identity), the most abundant cardiac β subunit in rat and humans (Colecraft et al., 2002; Hullin et al., 2003). To avoid confusion between palmitoylated and nonpalmitoylated forms of β_2a, we refer to np-β_2a as β_2b. Unlike the palmitoylated splice variant isoform (β_2a) of the same gene, β_2b does not slow down the VDI and does not reside in PM in the absence of α_1C (Olcese et al., 1994; Chien et al., 1998). When expressed in Xenopus oocytes, β_2b accelerates the VDI of Ca_v1.2, and in addition causes a robust increase in the whole-cell current and affects all gating parameters (Hullin et al., 1992; Shistik et al., 1995). However, it is not clear whether coexpression of either β_2a or β_2b improves the trafficking of α_1C to the PM in Xenopus oocytes. No change (Neely et al., 1993, 2004) or a mild ∼50% increase (Shistik et al., 1995) have been reported (see Table IV and Discussion).

Using the imaging method shown in Fig. 3, we have examined the effect of coexpression of β_2b on the amount of α_1C-wt, α_1CΔ20, α_1CΔ46, and α_1CTYP in two to four batches of oocytes; α₂δ was always coexpressed (Fig. 4). As before, in the absence of Ca_vβ, the surface expression of NT mutants was similar to that of α_1C-wt (Fig. 4, A and B). In contrast, β_2b induced a mild but reproducible and statistically significant ∼70% increase (P < 0.001) in the surface expression in all α_1C constructs (Fig. 4, A and C).

The effect of β_2b on surface expression of α_1C was additionally verified using an independent method, using a modified α_1C (termed α_1C-HA) with an extracellular HA tag inserted in the extracellular loop following the S5 segment of domain I. A similar construct was previously used to study the trafficking of α_1C in mammalian cells (Altier et al., 2002). Ba²⁺ currents via channels formed by α_1C-HA, coexpressed with α₂/δ with or without β_2b, did not differ in amplitude or voltage dependency from α_1C-wt (unpublished data; Table I). Surface expression of the HA label was measured using a confocal microscope in whole intact oocytes fixated with formaldehyde and treated with an anti-HA antibody and a secondary antibody conjugated to a fluorescent dye (Fig. 4 D). As shown in Fig. 4 Da, the expression of β_2b caused a clear increase in surface expression of α_1C-HA. The results were quantified as explained in Fig. 4 legend, and showed a 2.09 ± 0.34-fold increase in the intensity of fluorescence caused by β_2b (Fig. 4 Dc). This estimate, though somewhat higher than by measuring the effect of β_2b in giant PM patches, was not significantly different when compared in the same oocyte batch (not depicted). These results unequivocally demonstrate that, despite the theoretical possibility that some of the fluorescent signal in the giant membrane patches arises from α_1C found in submembrane ER, the latter method provides a realistic estimate of α_1C levels in the PM. We conclude that, despite the presence of an endogenous Ca_vβ in the oocytes, the expressed β_2b further increases the surface expression of Ca_v1.2 channels about twofold (in the presence of α₂δ). The ability of β_2b to do so is not affected by the elimination of the NTI module.

Coexpression of β_2b also increased the macroscopic currents in wt and all NT mutants. Examples are shown in Fig. 5 A for α_1C-wt and α_1CΔ46. Note that β_2b accelerated the inactivation kinetics in both cases, despite the fact that without Ca_vβ, they were already faster in α_1CΔ46 than in α_1C-wt. This was a recurrent result in most mutants (unpublished data). Thus, although VDI is out of the scope of this paper, and although it is clear that the NH₂ terminus itself plays a role in VDI (see also Shistik et al., 1998; Kobrinsky et al., 2004), we construe that the elimination of the NTI module does not impair the ability of β_2b to speed up the VDI.

Figure 5.

Effects of mutations in NT, and of the β_2b subunit, on voltage dependence of activation of I_Ba. (A) Representative net I_Ba of α_1C-wt and α_1CΔ46 constructs in the absence and presence of coexpressed β_2b subunit, recorded at +40 mV. (B) I-V curves of channels in α₁α₂δ and α₁α₂δβ_2b subunit compositions (n = 6–8, N = 1). Representative averaged I-V curves are shown for α_1C-wt, α_1CΔ5, α_1CΔ6-20, and α_1CΔ46. (C and D) Normalized averaged I-V curves (C) and G-V curves (D) from all experiments and all mutants tested, in absence (gray symbols) and presence (pink symbols) of the β_2b subunit. The solid lines were drawn for illustration using the Boltzmann equation with the averaged parameters of V_1/2 and K_a from Table I and V_rev from Table S1. Data are from 2–11 experiments, 14–56 cells.

The I-V curves of all α_1C mutants, and of the α_1C-wt, were shifted to the left by coexpression of β_2b. Examples of averaged I-V curves of oocytes from representative batches are shown in Fig. 5 B, and a full summary of normalized I-V and G-V curves averaged from all oocytes is presented in Fig. 5 (C and D). The hyperpolarizing shift produced by β_2b is observed both in α_1C-wt and in all mutants lacking the NTI module. A closer examination of the results of Boltzmann fits in Table I reveals that the net change in V_1/2 and K_a parameters caused by β_2b in α_1C-wt (14 and 2.4 mV, respectively) was somewhat greater than in some of the NT mutants (11–13 and 1.5–2 mV). However, these differences are small and may be within the experimental or fitting error. In all, we conclude that elimination of NT inhibitory module does not impair the ability of β_2b to cause a hyperpolarizing shift in the voltage dependency of channel activation.

In contrast to parameters considered so far, the extent of increase in I_Ba and in G_max caused by coexpression of β_2b was altered dramatically by NT mutations (Figs. 5 B, Fig. 6, and Table III). In agreement with previous reports regarding α_1CΔ20, α_1CΔ46, and α_1CΔ139 (Shistik et al., 1998, 1999), all mutations that impaired the function of the NTI module also greatly reduced the ability of β_2b to increase I_Ba (Fig. 6, A and B). The differences were more pronounced at less positive potentials, because at least part of the increase in I_Ba at these potentials is due to the leftward shift in the I-V curve, as illustrated in Fig. 6 A for some of the mutants. However, even at +40 mV, the difference between α_1C-wt and the various mutants is still very substantial (Fig. 6 B). The β_2b-induced increase in the calculated G_max, which in α_1C-wt was 4.65 ± 0.39-fold, was diminished to only 1.92 ± 0.23-fold in α_1CΔ46 (P < 0.01; Table III). A similar, though slightly milder, reduction was observed in α_1CΔ20 and α_1CTYP (Table III). The only outlier is α_1CΔ5. The β_2b-induced change in I₄₀ was even greater in this mutant than in α_1C-wt, as measured in two batches of oocytes (Fig. 6 C). This is at odds with a previous observation (Fig. 2 F in Shistik et al., 1999), but a comprehensive investigation into the reasons of controversy is not possible, because in the experiment reported in 1999, the effect of β_2b was studied only marginally (Shistik et al., 1999); I_Ba was measured only at +10 mV, and no analysis of voltage dependency has been performed. Thus, dose-dependent or batch-dependent effects of β_2b on α_1CΔ5 cannot be excluded at present.

Figure 6.

The NTI module regulates the effect of β_2b subunit. (A) The increase in I_Ba caused by coexpression of β_2b (relative to currents observed in the same α_1C constructs without β_2b) is much smaller in NT deletion mutants than in α_1C-wt, at all voltages. Shown are results of two representative experiments (n = 10–18). (B) A general summary of the effect of coexpression of β_2b. Shown is the fold increase in relative I₄₀ in the various constructs; numbers of assayed cells are indicated above the bars. N = 2–7 experiments. (C) The increase in I_Ba caused by coexpression of β_2b is potentiated in the α_1CΔ5 mutant. N = 2.

Table III.

Relative Changes in Surface Expression, G_max, I₄₀, and P_o Caused by Coexpression of β_2b

				Fold increase in Po
Group	Fold increase in PM α1C labeling	Fold increase in Gmax	Fold increase in I40	from Gmax	from I40
α1C-wt	1.69 ± 0.19 (20; 4)	4.65 ± 0.39 (45; 7)	4.9 ± 0.37 (45; 7)	2.75	2.9
α1CΔ20	1.65 ± 0.11 (11; 2)	2.48 ± 0.58a (19; 4)	2.36 ± 0.26a (19; 4)	1.5	1.43
α1CΔ46	1.63 ± 0.08 (20; 3)	2.01 ± 0.22a (28; 5)	1.92 ± 0.23a (28; 5)	1.23	1.18
α1CTYP	1.79 ± 0.04 (9; 2)	2.49 ± 0.21a (20; 3)	2.24 ± 0.23a (20; 3)	1.39	1.25
α1C-short	1.64 ± 0.13 (12; 3)	1.25 ± 0.14a (11; 2)	1.31 ± 0.15a (11; 2)	0.76	0.8
NTSL	1.58 ± 0.17 (9; 2)	4.12 ± 0.41 (16; 3)	4.07 ± 0.49 (16; 3)	2.6	2.57
HA-labeled constructs
α1C-HA	2.22 ± 0.21 (13; 1)	8.04 ± 0.93 (13; 3)	8.49 ± 1 (13; 3)	3.6	3.82
NTLS-HA	1.99 ± 0.22 (7; 1)	2.36 ± 0.31a (9; 3)	2.29 ± 0.28a (9; 3)	1.18	1.15
α1CΔ21-46-HA	2.86 ± 0.38 (11; 1)	10.9 ± 1.38 (12; 3)	11.03 ± 1.54 (12; 3)	3.8	3.85

All comparisons and normalization procedures are as in Table II.

^aP < 0.01, compared with α_1C-wt.

To summarize, the main conclusion from these experiments is that the ability of β_2b to increase G_max is greatly reduced by the removal of a crucial part of the NT inhibitory module, aa 6–20. The unique aa 2–5 of the long-NT channel are a part of the NTI module, since their removal already causes a mild increase in P_o,max; however, it also potentiates the enhancing effect of β_2b on P_o,max, instead of reducing it. This peculiar effect deserves further attention in the future.

The β_2b-induced increase in G_max still remaining after the elimination of the NTI module is comparable to the increase in the surface expression caused by β_2b. This indicates that in these mutants, the increase of the macroscopic Ca²⁺ channel conductance caused by β_2b may be the result of an increase in the amount of the functional channels, N, whereas P_o,max is not substantially changed. This is illustrated by comparing the values of β-induced change in P_o,max calculated using Eq. 6 (Table III).

The Initial Segment of the Short-NT Isoform of α_1C Is Not an NTI Module

Are aa 6–20 of long-NT sufficient to act as an NTI module? Does the initial segment of the short-NT isoform (aa 2–16) act as an inhibitory module, similarly to the partially homologous aa 6–20 segment in the long-NT? To answer these questions, we created a cDNA encoding for a short-NT isoform, α_1C-short, in which exon 1a of the long-NT was replaced by the exon 1 of short-NT α_1C (Fig. 7 A). The rest of the molecule was left as in the cardiac, long-NT α_1C. This design was intended to single out the effects of NT and to avoid additional changes in channel's properties that can arise if one is using one of the several cloned short-NT isoforms, with additional variations in the other parts of the protein (Abernethy and Soldatov, 2002). We also created cDNAs encoding two chimeric proteins in which parts of short- and long-NT initial segments were switched (Fig. 7 A): (1) NT_SL, which is a short-NT α_1C, in which aa 2–16 of the short-NT α_1C are replaced by aa 6–20 of the long-NT α_1C; (2) NT_LS, which is a long-NT α_1C in which aa 6–20 are replaced by aa 2–16 of the short-NT. Two additional groups tested in these experiments included the standard α_1CΔ46 mutant (which misses the entire initial segment), and a new internal deletion of long-NT α_1C, α_1CΔ21-46. External HA-tagged constructs were also made for all new mutants except short-NT α_1C; they exhibited G-V curves and macroscopic I_Ba very similar to the untagged counterparts (Table I; unpublished data). Immunohistochemical measurements in giant membrane patches or in intact oocytes (for HA-tagged constructs) showed that all new constructs expressed similarly to α_1C-wt in the PM, and coexpression of β_2b increased the amount of all constructs in the PM similarly to α_1C-wt (Fig. 7, B and C; Tables II and III).

Fig. 8 (A–C) reports on the results of two series of experiments (a and b) in which the macroscopic properties of I_Ba in the new constructs were tested, in different combinations and at two RNA doses. Without the β subunit, I_Ba in α_1C-short, NT_SL, and NT_LS was substantially greater than in α_1C-wt, and a small but significant leftward shift in V_1/2 and a decrease in K_a were also observed. All these parameters were comparable to those of α_1CΔ46 (Fig. 8; Tables I and II). I₄₀ tended to be somewhat smaller in NT_SL than in α_1C-short or α_1CΔ46, but the difference did not reach statistical significance. In contrast, both macroscopic I_Ba and all G-V curve parameters of α_1CΔ21-46 were very similar to α_1C-wt.

Figure 8.

Chimeras of long- and short-NT help demarcate the NTI module. Two separate series of experiments, that used different doses of RNA, are shown in A–C. In Aa, B, and Ca, all RNAs were injected at 1 ng/oocyte; in Ab and Cb, at 2.5 ng/oocyte. (A) Comparison of normalized I₄₀ in the absence of coexpressed β subunit. N = 2–3. (B) Effect of β_2b on the voltage dependency of I_Ba in the same experiments as in Aa. Normalized averaged G-V curves are shown, in the absence (black symbols) and presence (red symbols) of the β_2b subunit. (C) Summary of the effect of coexpression of β_2b on I₄₀. (D) The voltage dependency of the relative effects of β_2b on I_Ba in selected NT mutants. The results were summarized from experiments shown in Figs. 6 B and 8. In each experiment, the increase in I_Ba caused by the coexpression of β_2b was normalized to the average increase observed in α_1C-wt group of oocytes at the same voltage, and the results were averaged across all experiments. The linear regression (dotted lines) is shown for illustrative purposes.

Coexpression of β_2b shifted the G-V curve to the left by 12–17 mV in all constructs (Fig. 8 B; Table I), whereas the effects on I_Ba amplitude were again diverse. The increase in I₄₀ caused by β_2b was poor (less than twofold) and comparable to α_1CΔ46 in α_1C-short and NT_LS, which lack the crucial 16 aa of the long-NT. In NT_SL, I₄₀ increased much more than in α_1CΔ46 but less than in α_1C-wt, whereas α_1CΔ21-46 again behaved almost exactly as α_1C-wt (Fig. 8 C). The increase in P_o,max caused by β_2b almost disappeared in α_1CΔ46 and NT_LS mutants; in the α_1C-short, the calculation even showed a small decrease in P_o,max by β_2b (Table III). In contrast, in NT_SL, the calculated P_o,max was increased by β_2b almost like in the long-NT α_1C-wt. α_1CΔ21-46 behaved exactly like α_1C-wt (Fig. 8 Cb). Taken together, these results strongly suggest that aa 1–20 of long-NT fully restore the function of the NTI module, whereas aa 6–20 of long-NT alone are insufficient. aa 2–16 of short-NT cannot replace the homologous 15 aa of the long-NT as a crucial component of NTI module.

The difference in effects of β_2b on α_1C-wt and the various NT deletion mutants is further explored in Fig. 8 D, where the β_2b-induced increase in I_Ba at each voltage was normalized to that observed in α_1C-wt. The rationale for this analysis is as follows. If the NTI module regulates the effect of β_2b on P_o,max (a nominally voltage-independent parameter) and not on the voltage dependence of activation, then the removal of the NTI module should alter the effect of β_2b to a similar extent at all voltages. In general, this was indeed the case (Fig. 8 D), although the relative effect of β_2b showed a slight voltage sensitivity; the difference between α_1C-wt and NT mutants was somewhat greater at more negative voltages. This, however, is a predictable consequence of the mild hyperpolarizing shift in the G-V curve caused by the removal of the NTI module (see Fig. 2 C and Fig. 8 B). The latter leads to a slightly stronger voltage dependency of β_2b-induced increase in I_Ba in α_1C-wt than in the mutants. Fig. 8 D also illustrates the fact that, of all constructs tested, the increase in I_Ba by β_2b was the greatest in α_1C-wt and the smallest in α_1C-short (even compared with α_1CΔ46). The substantial recovery of the effect of β_2b in NT_SL and the full recovery in α_1CΔ21-46 are also clearly visualized.

DISCUSSION

Overview of Findings and Conclusions

Despite the importance of regulation of Ca²⁺ channels by β subunits and the wide interest in structure, function, and interactions of Ca_vα and Ca_vβ, controversies still abound regarding the mechanisms and even the phenomenology of the effects of Ca_vβ. In this study we attempt to sort out some of the discrepancies, and to better understand the mechanism by which Ca_vβ increases Ca²⁺ channel currents, focusing on a single type of Ca_vα₁ and Ca_vβ.

(a) We show that Ca_vβ-induced increase in maximal open probability (P_o,max), at least in Ca_v1.2, is separable from the other mechanisms (shift in G-V curve and change in PM expression) that underlie the increase in macroscopic Ca²⁺ currents by Ca_vβ. This effect is crucially dependent on the presence of an NTI module. This separability, and the separability of trafficking from all gating effects, implies that Ca_vβ acts on α_1C to increase P_o,max via molecular determinants (in β and/or in α₁) different from those used to change trafficking or to shift the G-V curve.

(b) Using deletion mutagenesis and chimeric constructs, we map the necessary and sufficient component of the NTI module to aa 1–20 of the long-NT and demonstrate that the NTI reduces P_o,max, thus restraining the activation of the channel at all voltages. Loss of NTI module enhances I_Ba by increasing P_o,max ∼6–10-fold and, in parallel, specifically eliminates the effect of Ca_vβ on P_o,max; addition of the module restores the small amplitude of I_Ba and the enhancement of P_o by Ca_vβ. This correlation supports the hypothesis (Shistik et al., 1998) that part of the effect of β on α_1C is due to a suppression of the inhibitory action of the NTI module. In the absence of the NTI module, Ca_vβ cannot further increase P_o,max, which is already high.

(c) We establish that the NTI module is completely missing from a short-NT Ca_v1.2, which represents a class of Ca_v1.2α isoforms abundant in the smooth muscle and brain. This is a striking example of a physiologically meaningful disparate regulation of two very similar isoforms of an ion channel by an auxiliary subunit.

(d) Using two independent methods, we unequivocally demonstrate that coexpression of Ca_vβ increases the surface expression of α_1C in Xenopus oocytes approximately twofold, independently of the presence of the NTI module. By a critical examination of literature and separation of studies that used different NT isoforms, we resolve apparent controversies and single out remaining problems, regarding the effects of Ca_vβ on α_1C.

A Summary of the Effects of Ca_vβ on α_1C

It is known that the effects of β subunits differ among the various α₁ subunits (Dolphin, 2003). However, even for a single species of Ca_vα₁, the details and magnitude of effects of coexpressed Ca_vβ often remain controversial. For instance, estimates of the increase in macroscopic maximal I_Ba of Ca_v1.2 (α_1C) in mammalian cells range between 15 and 20-fold (Lory et al., 1993; Takahashi et al., 2004) to almost no effect (Yasuda et al., 2004). The comparison between the different reports is thwarted by the unavoidable difference in recording conditions and often by an absence of quantitative estimates of some of the electrophysiological parameters. To understand the origins of the controversies, we summarized the data on Ca_vβ effects on Ca_v1.2 from reports that contain an explicit quantitative estimation of parameters that affect the macroscopic Ca²⁺ channel currents (Table IV). Data collected from short-NT (or mutants of α_1C missing the initial NT segment) and the long-NT isoforms of α_1C are separated. Macroscopic inactivation is unimportant in determining the whole-cell current when Ba²⁺ is used as the charge carrier (see below), and is left out of scope. Reports in which effects of different Ca_vβ have been compared but there are no data on channels expressed without Ca_vβ are not included. If several β subunits have been compared in the same study, we usually show β_2a or np-β_2a (β_2b), which were the most widely used. While the summary in Table IV may not be exhaustive, it is quite revealing.

A Distinction between P_o and P_o,max

Whole-cell Ca²⁺ channel currents and/or conductance (G) depend both on the amount of functional channels (N) and on the open probability, P_o, such that G = γ × N × P_o, and G_max = γ × N × P_o,max (see Materials and Methods, Eqs. 4 and 5). Treatments used in our experiments (NT deletions or coexpression of β_2b) do not affect γ (Shistik et al., 1995, 1998; Hullin et al., 2003), thus only N and P_o were changing. N is voltage independent, and assessing the impact of changes in N (once monitored) on G is straightforward.

In comparison, the analysis of P_o and the assessment of its effect on G are not trivial. First, in voltage-dependent channels, P_o is voltage dependent, being described by the same Boltzmann equation as G/G_max (Hille, 2002) (Eq. 3). P_o ranges from zero at negative potentials to P_o,max at “saturating” membrane voltages where the G-V curve reaches a plateau. Note that in this classical treatment G_max and P_o,max are voltage independent. An agent that improves the coupling between gating charge movement and pore opening, such as Ca_vβ (Neely et al., 1993), and shifts the G-V curve to the left will increase P_o (and macroscopic currents) in a range of “nonsaturating” voltages, whether P_o,max is changed or not. In contrast, an increase in P_o,max leads to an increase in Ca²⁺ conductance at all voltages except those at which the channel is fully shut. Therefore, to understand the molecular mechanism of β's actions, it is important to distinguish between changes in P_o caused by a G-V curve shift vs. an increase in P_o,max.

In single-channel recordings it is possible to directly determine P_o, which is the fraction of time that a channel spends in open state(s), out of total observation time (Colquhoun and Sigworth, 1995). Although estimation of P_o can be distorted by the presence of inactivated states (Colquhoun and Sigworth, 1995), in Ca_v1.2 with Ba²⁺ as charge carrier, the inactivation is slow, and such distortion can be avoided by using short analyzed time segments (a few tens of milliseconds). Unfortunately, the estimation of changes in P_o,max caused by Ca_vβ is almost impossible in single-channel recording of Ca_v channels. Most publications report single channel currents recorded between −10 and +10 mV in high-Ba²⁺ solutions, to ensure high signal-to-noise ratio; a full I-V curve is usually not constructed. (An exception is a study in cardiomyocytes [Colecraft et al., 2002], where a full I-V relation was explored; see Table IV.) Under these conditions, G (and thus P_o) is very far from reaching a maximum, as evident from typical G-V curves shown in Figs. 2, 5, and 8, and the measured P_o does not provide any estimate of P_o,max.

In comparison, measurements of changes in macroscopic G_max in conjunction with a reliable estimation of N should provide a good measure of relative changes in P_o,max (Wei et al., 1994; Takahashi et al., 2004). Here, G/G_max is calculated using the values of peak macroscopic current that are independent of the duration of voltage step, therefore the latter does not affect the calculated relative change in P_o,max. Our estimates of P_o,max were not affected by changes in the steady-state inactivation curve in the NT deletion mutants, because the availability of the channels for opening at the holding potential of −80 mV remained 100% (Shistik et al., 1998). There are two possible sources of inaccuracy. First, acceleration of the inactivation process by β_2b may decrease peak I_Ba, if the rate constants of activation and inactivation are comparable (Aldrich et al., 1983; Hille, 2002). However, this artifact is probably negligible in Ca_v1.2. Here, the time constant of the faster of the two components of macroscopic VDI of I_Ba is ∼75 ms for a short-NT isoform (Shi and Soldatov, 2002) and >150 ms with the long-NT α_1C (unpublished data). This is at least twofold slower than first latency time constant observed in single channel recordings (∼25–35 ms, see Hullin et al., 2003). Another artifact may arise when G_max is estimated by fitting experimental I-V curves to Boltzmann equation. In this case, G usually appears to reach its maximum at +40 to +50 mV. G-V curves based on the measurement of tail currents, which are devoid of certain inaccuracies inherent to I-V curves (such as very small I_Ba at V_m close to the reversal potential), often imply that the G-V curve has two components and does not reach a plateau at +40 to +50 mV, but close to 100 mV (e.g., Takahashi et al., 2004). Assuming that this is the case and using the data from Takahashi et al. (2004), we calculated that we might have overestimated the changes in G_max, but by no more than 30% (unpublished data). Since in α_1C-wt the changes in G_max caused by NT deletions or by β_2b are between 260 and 770% (Tables II and III), such an error would not affect our main conclusions.

Ca_vβ-induced Changes in Voltage Dependency (G-V curve) and in P_o,max Are Separable

(a) Only the Ca_vβ-induced increase in P_o,max, but not the shift in G-V curve, depends on the presence of the NTI module. In all mutants lacking the crucial part of the NTI module (aa 6–20 of long NT), G-V curve shifts are like in α_1C-wt, but P_o,max is not changed by β_2b (Table III). The mild approximately twofold increase in G_max (or I₄₀) caused by β_2b in these mutants can be almost fully accounted for by the increase in N, which is 1.6–2.2-fold. Strikingly, in the short-NT isoform, the coexpression of β_2b even causes an ∼20% decrease in calculated P_o, despite a well-pronounced threefold increase in I_Ba at 0 mV (a V_m relevant for comparison with single-channel records). This increase is undoubtedly governed by the increase in N (see below) and the leftward shift in the G-V curve.

Published results (Table IV) fully support our data; separation of long-NT and short-NT isoforms eliminates controversies and allows solid conclusions. Coexpression of β subunits increases the P_o measured in single-channel patches at −10 to +10 mV in high-Ba²⁺, in oocytes and in mammalian cells. However, the increase is consistently greater in the long-NT isoform (range: six to eightfold) than in the short-NT α_1C forms (range: two to fivefold). Data necessary for the calculation of the relative change in P_o,max are available from fewer works, but also fit our results; coexpression of Ca_vβ increases P_o,max of the long-NT α_1C two to fourfold, whereas a mild ∼20% decrease is observed in a short-NT α_1C isoform.

(b) Changes in P_o,max caused by coexpression of Ca_vβ are similar in Xenopus oocytes and mammalian cells (Table IV). In contrast, there is a consistent and striking difference in the reported shifts in the G-V curve: changes in ΔV_1/2 range from 12 to 22 mV in oocytes but are always <10 mV in mammalian cells. This intriguing difference calls for study and may provide new clues to understanding VDCC gating; but it also supports the disparity of mechanisms of the effects of Ca_vβ on P_o,max and on voltage dependence of activation.

On the basis of these arguments, we conclude that, most intriguingly, the two gating actions of Ca_vβ that are most important in determining the amplitude of whole-cell Ca²⁺ currents are separable, and therefore may rely on distinct molecular mechanisms. It is remarkable that one of the most prominent actions of Ca_vβ can be selectively eliminated by removing a short segment of α_1C that does not even bind β, whereas other functions of Ca_vβ remain unchanged. This finding provides a hint at how a single Ca_vβ may regulate several different functions of α₁, without the need to assume the involvement of multiple β subunits.

Changes in Surface Expression Caused by β_2b

Surface expression of voltage-dependent ion channels can be gauged by quantitative immunohistochemistry of various kinds, or by measuring the maximal nonlinear gating charge movement, Q_max. The immunohistochemical methods that use monitoring of external surface-exposed parts of channel with or without genetically introduced epitopes reliably monitor the total number of channels in PM (e.g., Zerangue et al., 1999; Altier et al., 2002; Canti et al., 2005; Leroy et al., 2005) but do not guarantee that all detected channels are conducting (functional). Similarly, Q_max reflects the displacement of the voltage sensors in all channels in the PM, whether conducting or not. Relative changes in number of channel (N) caused by a regulatory factor (e.g., Ca_vβ) are faithfully reflected in changes of either surface labeling or Q_max; however, the latter is also altered by factors that alter the charge or movement of the voltage sensors (Aggarwal and MacKinnon, 1996; Barrett et al., 2005).

The prevailing view is that the Ca_vβ-induced increase in the amount of α_1C in the PM, reproducibly observed in mammalian cells including even cardiomyocytes (Table IV), does not occur in Xenopus oocytes. This notion is based on the reported absence of a significant Ca_vβ-induced increase in Q_max measured in the oocytes using the cut-open voltage clamp, in which gating and ionic currents are measured from a large part of an internally perfused cell (Neely et al., 1993, 2004). In contrast, Yamaguchi et al. (1998, 2000) reported a threefold increase in α_1C caused in Xenopus oocytes by either injecting a purified β₃ subunit or by coexpressing β_2a. Although imaging of an intracellular HA tag in permeabilized oocytes, used in these works to monitor surface expression of α_1C, does not exclude the labeling of α_1C located in submembrane ER, the approximately threefold increase is also supported by observing the enhancement of I_Ba at different times after the injection of the β₃ protein (Yamaguchi et al., 1998; see Table IV legend for details). We have previously detected a β_2b-induced, 20–55% increase in N measured by counting channels in cell-attached patches and by immunoprecipitating α_1C from manually separated PM (Shistik et al., 1995), but it did not reach statistical significance. Interestingly, in the absence of α₂δ, even this small increase was not observed (Table IV).

To address the controversy and to measure changes in surface expression of α_1C in oocytes as accurately as possible, in this work we employed two additional, independent methods: immunolabeling of α_1C, tagged by an extracellular HA epitope, in whole oocytes; and imaging of immunolabeled α_1C in giant membrane patches of oocytes. Both methods clearly demonstrated a reproducible, highly statistically significant, isoform-independent increase in PM expression of α_1C induced by coexpression of β_2b. The estimates obtained by the widely accepted external tag measurement were slightly but usually insignificantly greater than by imaging in giant patches, when measured with the same construct of the channel (2.2-fold vs. 1.7-fold increase in α_1C-wt, respectively). We conclude that β_2b causes an approximately twofold increase in the PM content of α_1C, in the presence of α₂δ. The reason for the discrepancy with Neely et al. (1993, 2004) remains unclear; it would be interesting to see whether a change in Q_max occurs when α₂δ is present. The increase in surface expression caused by coexpression Ca_vβ is absolutely independent of NTI module, in sharp contrast with P_o,max. These results corroborate the previously demonstrated separability of the effects of Ca_vβ on surface expression and gating (Canti et al., 1999, 2001; Gerster et al., 1999).

The NTI Module and the Effect of Ca_vβ

In this report we demonstrated that α_1C variants lacking the NTI module (the various mutants and the short NT isoform) express in the PM at the same level as the long-NT isoform, though they give much larger macroscopic currents. These and our previous results (Shistik et al., 1998) firmly establish the role of the initial segment of long-NT as a regulator of gating, rather than expression, of α_1C. In support, Agler et al. (2005) recently found that the NT also constitutes an inhibitory module in the N-type (Ca_v2.2) channel; in contrast to Ca_v1.2, in Ca_v2.2 the inhibitory effect of the NT is G protein gated.

Since the main effect of NTI is to regulate the voltage-independent parameter P_o,max, it is possible that the removal of the NTI module causes a modal shift in gating, like Ca_vβ (Costantin et al., 1998; Colecraft et al., 2002). If NTI module “tonically” restrains the opening of the channel's gate at all voltages (Shistik et al., 1998), its removal would cause a voltage-independent transition from a low-P_o mode to a high-P_o mode, in which the voltage independent transition from the last closed to the open state is more favorable than in the low-P_o mode. The slight change in the voltage dependence of activation in NT mutants lacking the NTI module may be due to a reequilibration among closed states, the transitions between which are voltage dependent. Further studies including single channel recordings will be required to test this conjecture.

We find that aa 6–20 constitute a crucial and necessary part of the inhibitory module; its deletion causes a maximal enhancement of P_o,max and an almost complete disappearance of the enhancing effect of β_2b on P_o,max. Nevertheless, aa 6–20 alone only partially recover the function of the NTI when added at the NH₂ terminus of a α_1CΔ46 mutant, which lacks all 46 initial amino acid residues (NT_SL chimera). Full recovery of both functions of NTI (reduction in P_o,max and enhancement of β_2b-induced increase in P_o,max) is achieved by the addition of aa 1–20 of the long-NT, which is thus established as a necessary and sufficient structural determinant of the NTI module. The function of aa 21–46, which link between the NTI module and the beginning of the highly conserved part of NT encoded by exon 2, is unknown at present. Our data conclusively demonstrate that the initial 16-aa NT segment of the short α_1C isoform is not an inhibitory module. These 16 aa were unable to restore the NTI functions even in the context of a full 46-aa initial segment, i.e., in the presence of aa 2–5 and 21–46 of the long NT (NT_LS chimera).

The interacting partners of NT in Ca_v1.2 are unknown. The regulation of inactivation kinetics of α_1C (Ca_v1.2α) by β_2a involves voltage-dependent rearrangements of NT regions of α_1C and β_2a, as deduced from fluorescent resonance energy transfer (FRET) experiments (Kobrinsky et al., 2004). However, no physical interaction between NT of α_1C and Ca_vβ could be detected by direct biochemical measurements (Pragnell et al., 1994; Shistik et al., 1998), and the exact molecular interaction underlying the proximity between the NT of α_1C and β_2a indicated by FRET is yet to be determined. No direct interaction of NT with L1 (which would explain the intensive interplay with Ca_vβ function) could be detected in α_1C (Agler et al., 2005). Another possible interactor is the CT of α_1C. Similarly to NT, CT acts to prevent channel opening (Wei et al., 1994; Klockner et al., 1995; Gao et al., 2001; Mikala et al., 2003). Joint removal of NT and distal CT causes a more-than-additive increase in I_Ba (Ivanina et al., 2000), suggesting that the relaxed state of the channel is a high-P_o one, and the termini are vital to prevent uncontrolled entry of Ca²⁺.

Using Xenopus Oocytes vs. Mammalian Cells to Study the Effects of Ca_vβ on Ca²⁺ Channels

The validity of conclusions regarding Ca_vβ effects, especially changes in surface expression, obtained in Xenopus oocytes has often been met with reservations, because of the presence of an endogenous β subunit. However, HEK cells also contain endogenous β (Leroy et al., 2005). In this work we demonstrate that all major effects of Ca_vβ can be reproducibly observed and reliably studied in Xenopus oocytes. The Ca_vβ-induced increase in surface α_1C in oocytes is in qualitative agreement with that in mammalian systems, although quantitatively the increase reported in mammalian cells is usually greater, more than fivefold (Table IV; but see Yasuda et al., 2004). The variability among different cell types may reflect variable levels of endogenous Ca_vβ. The changes in gating parameters are also at least qualitatively (shift in G-V curve; see above) and quantitatively (change in P_o,max) similar.

Physiological Significance

Ca_v1.2 channels are crucial for contraction of cardiac and smooth muscles, hormone secretion, and regulation of gene expression in the brain (Reuter, 1983; Finkbeiner and Greenberg, 1998; Catterall, 2000; West et al., 2001; Lipscombe et al., 2004). In humans, the Ca_v1.2 gene, CACNA1C (Soldatov, 1994) contains 55 exons, of which 19 are subject to alternative splicing, giving rise to a very large number of isoforms (Tang et al., 2004). The exact exon combinations underlying the isoforms of α_1C in different tissues are yet to be catalogued (Abernethy and Soldatov, 2002; Jurkat-Rott and Lehmann-Horn, 2004; Tang et al., 2004). Long-NT isoform (isoforms?) is predominant in the heart but is also present in the brain, short-NT isoforms prevail in smooth muscle and brain (Slish et al., 1989; Koch et al., 1990; Snutch et al., 1991; Shistik et al., 1999; Blumenstein et al., 2002; Dai et al., 2002; Saada et al., 2003). At present we cannot exclude the possibility that in some of the short-NT isoforms of α_1C the function of the NT initial segment is different than in the isoform that we used here, due to an interaction with some alternative distant parts of α_1C; this remains to be investigated.

The isoforms may substantially differ in their physiological and pharmacological properties. For instance, splice variations in the COOH-terminal region regulate the inactivation of the channel and its sensitivity to certain pharmacological agents (Abernethy and Soldatov, 2002). The present work reveals a strikingly disparate regulation of two very similar isoforms of a Ca_v channel by a Ca_vβ subunit. It is notable that, in addition to the two isoforms of α_1C corresponding to splice variants with the alternative first exons 1 and 1a, another short-NT isoform lacking either of these exons and starting from exon 2 may exist in human tissues (Schultz et al., 1993; Saada et al., 2003). This would correspond to the truncation mutant α_1CΔ46 or, if the protein starts with the first methionine of exon 2, to α_1CΔ59 (see Fig. 1). The differences in function of NH₂ termini of these isoforms, discovered in this paper, are expected to have significant physiological implications. They will result in substantial differences in amplitudes of macroscopic Ca²⁺ currents and in the effects of Ca_vβ (whose association with α₁ may change in development and growth; see Spafford et al., 2004) or regulators that act in a Ca_vβ-dependent manner, such as protein kinase A (Bunemann et al., 1999).