CaBP1 Regulates Voltage-dependent Inactivation and Activation of Ca_V1.2 (L-type) Calcium Channels

Abstract

CaBP1 is a Ca²⁺-binding protein that regulates the gating of voltage-gated (Ca_V) Ca²⁺ channels. In the Ca_V1.2 channel α₁-subunit (α_1C), CaBP1 interacts with cytosolic N- and C-terminal domains and blunts Ca²⁺-dependent inactivation. To clarify the role of the α_1C N-terminal domain in CaBP1 regulation, we compared the effects of CaBP1 on two alternatively spliced variants of α_1C containing a long or short N-terminal domain. In both isoforms, CaBP1 inhibited Ca²⁺-dependent inactivation but also caused a depolarizing shift in voltage-dependent activation and enhanced voltage-dependent inactivation (VDI). In binding assays, CaBP1 interacted with the distal third of the N-terminal domain in a Ca²⁺-independent manner. This segment is distinct from the previously identified calmodulin-binding site in the N terminus. However, deletion of a segment in the proximal N-terminal domain of both α_1C isoforms, which spared the CaBP1-binding site, inhibited the effect of CaBP1 on VDI. This result suggests a modular organization of the α_1C N-terminal domain, with separate determinants for CaBP1 binding and transduction of the effect on VDI. Our findings expand the diversity and mechanisms of Ca_V channel regulation by CaBP1 and define a novel modulatory function for the initial segment of the N terminus of α_1C.

Introduction

EF-hand Ca²⁺-binding proteins (CaBPs)² regulate a wide variety of voltage- and ligand-gated ion channels (1–4). For voltage-dependent Ca²⁺ (Ca_V) channels, Ca²⁺-binding proteins can differentially regulate inactivation, activation, open probability (P_o), Ca²⁺-dependent facilitation, and surface expression. These diverse forms of regulation may expand the functional repertoire of Ca_V channels (5–13), tailoring Ca²⁺ signals to the specific needs of excitable cells.

Inactivation is a negative feedback process that prevents excessive Ca²⁺ entry via Ca_V channels. For Ca_V1 and Ca_V2 channels, inactivation occurs by two routes: voltage-dependent inactivation (VDI) and Ca²⁺-dependent inactivation (CDI) (14). Both CDI and VDI involve multiple determinants in the cytoplasmic domains of the pore-forming subunit (α₁) (15–17). CDI depends on calmodulin (CaM) interactions with the C-terminal tail of the Ca_V1 and Ca_V2 α₁-subunits (18–21). CaBPs can interact with CaM-binding sites in the Ca_V α₁-subunit (5, 6, 12, 22, 23) and may act by displacing or mimicking CaM (e.g. Refs. 6 and 11).

CaBP1 is a Ca²⁺-binding protein enriched in the brain and retina (2). It has emerged as a prominent regulator of Ca_V channel CDI (4). The molecular determinants underlying CaBP1 regulation have been most well characterized for Ca_V1.2 (α_1C) channels, which play crucial roles in Ca²⁺ signaling and excitability in the nervous and cardiovascular systems (14). In contrast to Ca²⁺/CaM, which stabilizes the inactivated state of Ca_V1.2 (24–26), CaBP1 fully eliminates CDI and supports Ca²⁺-dependent facilitation (6, 27). CaBP1 binds to the C-terminal CaM-binding α_1C sites but also interacts with the α_1C N-terminal domain, the deletion of which blunts the effect of CaBP1 on CDI (28). The N terminus of α_1C regulates several aspects of channel gating as well as modulation by protein kinase C (29–32). Within the N terminus, two modular segments are involved in CaM binding (33, 34) and regulation of P_o (35). A “long N-terminal” alternatively spliced α_1C isoform possesses a 20-amino acid (aa) initial segment (N-terminal inhibitory (NTI) module), which controls the channel's maximal P_o (30, 35). A second module in the central part of the N-terminal domain is a CaM-binding site, the N-terminal spatial Ca²⁺-transforming element (NSCaTE), which regulates CDI in Ca_V1.3 (33), but its role in Ca_V1.2 is less clear (34). The exact location of the CaBP1-binding site within the N terminus of α_1C has not been determined, and whether CaBP1 influences parameters of Ca_V1.2 function other than CDI is unknown.

Here, we report that CaBP1 expression in Xenopus oocytes regulates activation gating and accelerates VDI of Ca_V1.2. Although CaBP1 binds to a site in the distal N-terminal domain that is distinct from the CaM-binding site, CaBP1 regulation of VDI requires an intact proximal domain in the N terminus of α_1C. Deletion of the entire N-terminal domain inhibits but does not fully abolish effects of CaBP1 on VDI, which indicates the presence of additional relevant molecular determinants. These results suggest a new modulatory role for the α_1C N terminus, in which CaBP1 binding and transduction of the modulation of VDI are encoded by distinct N-terminal domains.

EXPERIMENTAL PROCEDURES

Expression Systems, DNAs, and RNAs

Maintenance of female frogs (Xenopus laevis), preparation of oocytes, and in vitro RNA synthesis were carried out as described previously (30). Oocytes were injected 3–5 days before the experiment with equal amounts (by weight, 0.8–5 ng of RNA) of cRNAs from the three channel subunits, α_1C, non-palmitoylated Ca_Vβ_2b (GenBank accession no. X64297), and skeletal muscle α₂δ-1 (P13806) (30), with or without 10 ng of cRNA from CaBP1-S (AF169148), unless mentioned otherwise. We used rabbit long N-terminal α_1C (LNTα_1C; X15539) or SNTα_1C, based on rabbit LNTα_1C, with the initial segment (46 aa long) replaced with that encoded by exon 1 (16 aa long) as in the human short isoform (Z34815) (35). RNAs of the N-terminal truncation mutants Δ46-, Δ20-, Δ20–46-, and Δ139-LNTα_1C (see Fig. 3A) were as described previously (30, 35). The cDNA of CaBP1 (6) was in pGEM-HJ, and the cDNA of CaM (M19312) was in the pBF vector (36).

FIGURE 3.

Mapping of the CaBP1-binding site in the N terminus of α_1C. A, upper, linear representation of the cytosolic N termini of LNTα_1C (aa 1–154) and SNTα_1C (aa 1–124). Most of the deletion constructs used in this study are indicated below the sequences. The locations of parts of the N terminus with known function (NTI module and CaM-binding site/NSCaTE) and of the newly identified CaBP1-binding segment are indicated by boxes. Lower, schematic representation of the GST-fused segments of LNTα_1C used in this study. B, in vitro synthesized, ³⁵S-labeled CaBP1 binds to the distal part of the N terminus in a Ca²⁺-independent manner. The two gels are from one experiment. The upper panels show the autoradiograms of ³⁵S-radiolabeled CaBP1 bound to the GST fusion proteins indicated on the bottom. The GST fusion proteins eluted from the beads at the end of experiment are shown in the lower panels, after staining the gel with Coomassie Blue. Interactions between CaBP1 and the GST fusion proteins were measured with the addition of 1 mm CaCl₂ or EGTA as indicated. input represents the signal from one-sixtieth of the total labeled protein used in the reaction.

Electrophysiological studies of HEK293T cells involved transient transfection of HEK293T cells with Ca_V1.2 with or without CaBP1. In some experiments, we utilized T-REx-293 cells (Invitrogen) stably transfected with GFP-tagged CaBP1 under tetracycline-inducible expression. Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum at 37 °C under 5% CO₂. Cells plated in 35-mm tissue culture dishes were grown to 65–80% confluency and transfected with FuGENE 6 transfection reagent (Roche Applied Science) according to the manufacturer's protocol. Cells were transfected with 2 μg of DNA coding for the rabbit LNTα_1C or rat rbcIIα_1C variant or their truncated forms, 1 μg of β_2a, 1 μg of α₂δ, and 0.05 μg of enhanced GFP with or without 0.05–0.1 μg of CaBP1, all in the pcDNA3 vector. For stably transfected T-REx-293 cells, cells were transfected only with Ca_V1.2 subunit cDNAs and 0.05 μg of mCherry plasmid to mark transfected cells. Tetracycline (1 μg/ml) was added to the culture medium 6–8 h before recording to induce CaBP1 expression. Patch-clamp recordings of all cells were made at least 24 h after transfection. There were no significant differences between channel properties or regulation by CaBP1 in the transiently transfected and stably transfected cells, so the data from both systems were combined for analysis.

Electrophysiology and Data Analysis

In Xenopus oocytes, the procedures were essentially as described (35). In brief, whole cell currents were recorded with the GeneClamp 500 amplifier Molecular Devices using a two-electrode voltage clamp. 25–30 nl of 50 mm BAPTA (Ca²⁺ chelator) was routinely injected into the oocytes 0.5–3 h before the measurement of currents. This procedure usually blocked the endogenous Ca²⁺-dependent Cl⁻ currents; cells with residual Cl⁻ currents (distinguished by long-lasting inward tails at −80 mV) were excluded from analysis. The extracellular solutions contained 40 mm Ba(OH)₂ or Ca(OH)₂, 50 mm NaOH, 2 mm KOH, and 5 mm HEPES, titrated to pH 7.5 with methanesulfonic acid. In each oocyte, the net current was obtained by subtraction of the residual currents recorded with the same protocols after applying 200 μm CdCl₂ to the same solution (supplemental Fig. S1A) (35). Stimulation, data acquisition, and analysis were performed using pCLAMP 10.2 software (Molecular Devices). All experiments were performed at 20–22 °C. Currents were measured by 400-ms, 1000-ms, or 10-s pulses from a resting potential of −80 mV to different potentials, with 60-s intervals for the 10-s pulses and with 10-s intervals for shorter pulses. A current-voltage (I-V) curve was fitted to the Boltzmann equation in the form I = G_max(V_m − V_rev)/(1 + exp(−(V_m− V_a)/K_a)), where K_a is the slope factor, V_a is the half-maximal activation voltage, G_max is the maximal macroscopic conductance, and V_rev is the reversal potential of the current. The parameters obtained for G_max and V_rev were then used to calculate fractional conductance at each V_m using the equation G/G_max = I/(G_max(V_m − V_rev)).

The waveform of decay of the Ba²⁺ current (I_Ba) was fitted, using a Levenberg-Marquardt algorithm, to a two-exponential equation in the form f(t) = A_faste^−t/τ_fast + A_slowe^−t/τ_slow + C, where A is the contribution of each kinetic component (τ_fast or τ_slow) of the decay of I_Ba, and C is the non-inactivating component (supplemental Fig. S1B and Table S2).

In transfected HEK293T cells, Ba²⁺ currents were recorded in whole cell patch-clamp configuration at room temperature. Extracellular recording solutions contained 150 mm Tris, 2 mm MgCl₂, and 10 mm BaCl₂. Intracellular recording solutions contained 130 mm N-methyl-d-glucamine, 60 mm HEPES, 1 mm MgCl₂, 2 mm MgATP, and 5 mm EGTA. The pH of the extracellular and intracellular recording solutions was adjusted to 7.3 with methanesulfonic acid. Reagents used for electrophysiological recordings were obtained from Sigma. Currents were recorded with an EPC 9 patch-clamp amplifier driven by PULSE software (HEKA Electronics, Lambrecht/Pfalz, Germany). Leak and capacitive transients were subtracted using a P4 protocol. To measure VDI, I_Ba was evoked by 1-s depolarizations to +10 mV from −80 mV.

r₄₀₀, r₁₀₀₀, and r₂₀₀₀ are defined as the residual current (I_res) left after 400, 1000, and 2000 ms, respectively, divided by the maximal current, r_x = I_res/I_peak. Data were analyzed using pCLAMP (Molecular Devices) or routines written in IGOR Pro software (WaveMetrics, Lake Oswego, OR). Averaged data are presented as means ± S.E. Statistical differences between two groups were determined by Student's t test, and multiple group comparisons were done by one-way analysis of variance followed by Dunnett's or Bonferroni tests. Graphs and statistical analysis were done with SigmaPlot (SPSS, Inc., Chicago, IL).

Pulldown Assays with GST-fused Proteins

Procedures were essentially as described (29). In brief, [³⁵S]Met/Cys-labeled CaBP1 or CaM was translated on the template of an in vitro synthesized RNA using a rabbit reticulocyte translation kit (Promega). DNAs of α_1C segments designed to create GST fusion proteins were cloned into the pGEX-4T-1 vector. GST-fused proteins were produced in Escherichia coli (strain BL21-RIL) transfected with the cDNAs (see Fig. 3A) and grown in standard medium at 37 or 18 °C after induction with isopropyl β-d-thiogalactopyranoside. The GST fusion proteins were extracted from E. coli using an Amersham Biosciences kit. The protein concentration was estimated using a Bio-Rad protein assay kit. Purified GST fusion proteins (5–10 μg) or purified GST (∼10 μg) was incubated with 5 μl of the lysate containing the ³⁵S-labeled proteins in PBS containing 0.05% Tween 20 with 1 mm CaCl₂ or EGTA. The final reaction volume was 300 μl. GST fusion protein was immobilized on glutathione-Sepharose beads (Amersham Biosciences) for 30 min at 4 °C and washed. Following washing, GST fusion proteins were eluted with 20 mm reduced glutathione in 120 mm NaCl and 100 mm Tris-HCl (pH 8) and analyzed by SDS-PAGE. Gels were stained to detect proportional protein concentration with Coomassie Brilliant Blue R-250 (Bio-Rad). The labeled products were identified by autoradiography using a PhosphorImager (Molecular Dynamics).

RESULTS

CaBP1 Accelerates VDI of Ca_V1.2

CaBP1 inhibition of CDI in Ca_V1.2 has been described for a SNTα_1C isoform from rat brain, rbcII (6, 28). However, a LNTα_1C isoform is also expressed in the brain (37). The LNTα_1C isoform includes a 46-aa initial segment encoded by exon 1a (38), which contains the NTI module (aa 1–20) (see Fig. 3A and supplemental Fig. S4 for sequence) that regulates the channel's P_o (35). The SNTα_1C isoform (39–41) has an initial N-terminal segment encoded by the alternative exon 1 (42, 43). This 16-aa segment is partly homologous to aa 6–20 of the long N terminus (see Fig. 3A and supplemental Fig. S4) but does not retain functionality of the NTI module in terms of modulation of P_o of Ca_V1.2 (35). Considering the importance of the rbcIIα_1C N terminus in CaBP1 regulation of CDI (28), we examined whether CaBP1 might differentially regulate the LNTα_1C and SNTα_1C isoforms. We employed two-electrode voltage-clamp recordings in Xenopus oocytes injected with RNAs encoding LNTα_1C or SNTα_1C together with the auxiliary α₂δ-1 and Ca_V β_2b-subunits. CDI and VDI were studied with Ca²⁺ and Ba²⁺ as the permeant ions, respectively. Ba²⁺ (I_Ba) or Ca²⁺ (I_Ca) currents were isolated by subtracting Cd²⁺-insensitive currents from the total current (supplemental Fig. S1A) (35). Ba²⁺ currents via Ca_V channels are widely considered as inactivating essentially by the VDI mechanism, and their decay provides a reliable measure of VDI kinetics (e.g. Ref. 25).

I_Ca demonstrated typical strong CDI, which was substantially slowed by coexpression of CaBP1 in both SNTα_1C and LNTα_1C isoforms (I_Ca) (Fig. 1A). Thus, Xenopus oocytes recapitulate the effect of CaBP1 on CDI described in HEK293T cells (6, 27). However, a novel observation was the acceleration of VDI seen with Ba²⁺ as the permeant ion (I_Ba) (Fig. 1A). The extent of inactivation was measured as r₄₀₀ (24), i.e. the residual current remaining at the end of a 400-ms test pulse normalized to the peak current amplitude. For both α_1C isoforms, CaBP1 significantly accelerated VDI (decreased r₄₀₀) at all positive voltages (Fig. 1, B and C).

FIGURE 1.

CaBP1 regulates both CDI and VDI. A, representative normalized currents elicited by 400-ms pulses from −80 to +20 mV in Xenopus oocytes expressing LNTα_1C (left panel) or SNTα_1C (right panel) without (control, black) or with CaBP1 (gray). Currents were recorded in 40 mm Ca²⁺ solution (I_Ca) or Ba²⁺ solution (I_Ba). α₂δ-1 and Ca_V β_2b-subunits were coexpressed in all experiments done in Xenopus oocytes. B, summary of changes in r₄₀₀ at different voltages in the presence (○) or absence (●) of CaBP1. Data are presented as means ± S.E. (n = 14–39). C, summary of the effects of CaBP1 on r₄₀₀ in the LNTα_1C and SNTα_1C isoforms in +20 mV. Bars represent means ± S.E.; the numbers within the bars indicate n. In B and C, the statistical significance of differences between the control and CaBP1 was determined by t test. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

CaBP1 Enhances VDI Mainly by Changing the Fraction of Fast and Slow Components of Inactivation

Given the relatively slow time course of Ca_V1.2 VDI, we next used 10-s depolarizing pulses to characterize the effects of CaBP1 on the kinetics of VDI (Fig. 2). For LNTα_1C and SNTα_1C alone, I_Ba showed faster decay kinetics at more positive membrane voltages, a hallmark of VDI (Fig. 2, A–C, and supplemental Fig. S3). The decay of I_Ba was best fitted with a double-exponential function (Fig. 2A and supplemental Fig. S1, B and C), giving time constants of τ_fast ∼ 0.6 s and τ_slow ∼ 3 s at +10 mV (Fig. 2C, Table 1, and supplemental Fig. S1B and Table S1), supporting the involvement of at least two processes in VDI of Ca_V1.2 (21). CaBP1 accelerated I_Ba decay kinetics at all voltages in both LNTα_1C and SNTα_1C isoforms (Fig. 2B and supplemental Fig. S3A). Acceleration of VDI by CaBP1 was dose-dependent and maximal with 5–10 ng of CaBP1 RNA/oocyte in both LNTα_1C and SNTα_1C isoforms (supplemental Fig. S2A). The kinetic analysis of current decay showed that, at less depolarized voltages, CaBP1 decreased both τ_fast and τ_slow. At more depolarized voltages, CaBP1 significantly increased the fast inactivating fraction (A_fast) with a concomitant decrease in the slowly inactivating fraction (A_slow) (Fig. 2C). Together, these changes account for the acceleration of VDI by CaBP1 at all voltages.

FIGURE 2.

Kinetic analysis of CaBP1-induced changes in VDI in LNTα_1C. A, records of I_Ba from a representative cell expressing LNTα_1C evoked by 10-s depolarizing voltage steps from −80 mV to different voltages. Net I_Ba, obtained by subtraction of Cd²⁺-insensitive currents, is shown in gray. The decay time course was fitted to a two-exponential function (see supplemental Table S1 for fitted parameters), and the fitted curve (black line) is shown superimposed on the original trace. B, CaBP1 accelerates the kinetics of decay of I_Ba of LNTα_1C. Currents were elicited by 10-s pulses from −80 mV to various voltages. Traces, averaged from 10–13 oocytes, show the decay phase of I_Ba, starting from the peak and to the end of the voltage step. For clarity, S.E. is shown only for every 20th acquisition point. Student's t test was used to determine statistical significance at the t = 2000 ms time point. C, summary of the two-exponential fit analysis of I_Ba decay in LNTα_1C. Averaged values of A_slow and A_fast (the contribution of each kinetic component), C (the non-inactivating current), and τ_slow and τ_fast (the time constants) at different voltages are shown with (○) or without (●) CaBP1. Statistical significance between values was calculated at each voltage using a non-paired t test. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

TABLE 1.

Parameters of VDI (r₄₀₀ and r₂₀₀₀) at +10 and +20 mV in the different channel constructs

NS, not significant.

RNA constructs injected	r400 ± S.E.	Δr400 (effect of CaBP1)	n	r2000 ± S.E.	Δr2000 (effect of CaBP1)	n
At 10 mV
LNTα1C	0.77 ± 0.02	0.20a	39	0.49 ± 0.05	0.22b	10
+CaBP1	0.57 ± 0.03		29	0.27 ± 0.04		13
SNTα1C	0.77 ± 0.03	0.23a	16	0.38 ± 0.05	0.19b	7
+CaBP1	0.55 ± 0.05		14	0.19 ± 0.04		7
Δ46-LNTα1C	0.73 ± 0.05	0.01 (NS)	10	0.40 ± 0.04	0.01 (NS)	10
+CaBP1	0.72 ± 0.04		14	0.39 ± 0.04		14
Δ139-LNTα1C	0.62 ± 0.04	0.12c	6	0.23 ± 0.05	0.07 (NS)	6
+CaBP1	0.50 ± 0.03		11	0.16 ± 0.03		11
At 20 mV
LNTα1C	0.73 ± 0.04	0.23a	39	0.41 ± 0.05	0.21b	10
+CaBP1	0.50 ± 0.04		29	0.20 ± 0.04		13
SNTα1C	0.73 ± 0.02	0.22a	16	0.36 ± 0.03	0.19a	7
+CaBP1	0.50 ± 0.03		14	0.17 ± 0.03		7
Δ46-LNTα1C	0.66 ± 0.05	0.03 (NS)	10	0.33 ± 0.04	0.03 (NS)	10
+CaBP1	0.63 ± 0.03		14	0.30 ± 0.03		14
Δ139-LNTα1C	0.58 ± 0.03	0.13b	6	0.18 ± 0.04	0.05 (NS)	6
+CaBP1	0.45 ± 0.02		11	0.13 ± 0.03		11

^a p < 0.001.

^b p < 0.01.

^c p < 0.05.

CaBP1-binding Site in the N Terminus, and No Overlap with the NTI or NSCaTE Module

The N-terminal domain of α_1C binds both CaM and CaBP1 (28, 29). The binding of CaM is Ca²⁺-dependent and has been mapped to aa 60–100, with crucial binding determinants between aa 80 and 92 (33, 34) conserved in all isoforms of α_1C (supplemental Fig. S4A), as well as in Ca_V1.3α₁ (α_1D). We tested if CaBP1 and CaM share similar molecular determinants for binding in the α_1C N-terminal domain in GST pulldown assays. GST fusion proteins that cover the complete long N terminus of α_1C, as shown schematically in Fig. 3A, were used to pull down in vitro synthesized radiolabeled CaBP1. CaBP1 bound to aa 95–154 and to a shorter segment (aa 95–140) but not to segments located N-terminally to aa 94 (Fig. 3B). Importantly, CaBP1 did not bind to the segment at aa 60–100, which has been shown to strongly bind CaM in the same assay (supplemental Fig. S4B) (34). These results locate the CaBP1-binding site to aa 95–140, a segment fully conserved in all α_1C isoforms; the crucial binding determinants are probably between aa 120 and 140. Thus, unlike the C terminus (6), in the N terminus, there is no overlap between CaM- and CaBP1-binding sites (Fig. 3A).

Consistent with previous findings (28), CaBP1 binding to the N-terminal fragments was Ca²⁺-independent in that binding was similar in the presence of either Ca²⁺ or EGTA (Fig. 3B). We estimate that the Ca²⁺-dependent binding of CaBP1 to the N terminus is weaker than that of CaM because the ratio of bound to loaded protein (“input,” i.e. initial amount of protein in the binding reaction) was always smaller for CaBP1 than for CaM (compare Fig. 3B and supplemental Fig. S4B). These results further underscore differences in the physical interactions of CaBP1 and CaM with Ca_V1.2 that could contribute to their distinct modulation of channel function.

CaBP1 Depolarizes Voltage-dependent Activation of Ca_V1.2

Because CaBP1 reduces the voltage sensitivity of activation in P/Q-type (Ca_V2.1) channels (5, 11), we also examined the possibility of such regulation in Ca_V1.2 channels expressed in Xenopus oocytes. CaBP1 caused a 10–12.5-mV depolarizing shift in the I-V curves and the corresponding conductance-voltage (G-V) curves of the LNTα_1C and SNTα_1C isoforms (Fig. 4; see Table 2 for Boltzmann fit parameters). CaBP1 inhibited Ca_V1.2 activation in a dose-dependent manner with the maximal effect achieved with 5–10 ng of CaBP1 RNA/oocyte (supplemental Fig. S2B). CaBP1 specifically altered the midpoint value of the G-V curve, the V_a. All other parameters of the Boltzmann equation did not change significantly (Table 2).

FIGURE 4.

CaBP1 depolarizes voltage dependence of activation. A, I_Ba was evoked by depolarizing pulses from a holding voltage of −80 mV to various voltages in oocytes expressing LNTα_1C, SNTα_1C, and the Δ46-LNTα_1C and Δ139-LNTα_1C constructs. Normalized current-voltage (I-V) curves averaged from 9–62 oocytes are shown. I-V curves were fitted to the Boltzmann equation in each cell. The values of V_a and k_a, averaged from the individual fit in all cells, were used to draw the solid lines of the I-V curves. B, conductance-voltage (G-V) curves. Each point shows the average G/G_max at each voltage calculated for the same cells as in A. The solid lines have been drawn using the Boltzmann equation for G/G_max utilizing the values of K_a and V_a obtained in the fits of the I-V curves. The Boltzmann fit parameters and number of cells analyzed are shown in Table 2.

TABLE 2.

Parameters of Boltzmann fits of I-V curves in the different α_1C constructs expressed in oocytes and the effect of CaBP1

Values are means ± S.E.

	IBa (40 mm Ba2+)		ΔVa (t test)
	Without CaBP1	With CaBP1
LNTα1α2β2b
Gmax	23.75 ± 2.54	19.39 ± 2.38
Vrev	68.67 ± 1.47	69.47 ± 0.78
Va	−0.16 ± 1.04	10.01 ± 0.80	10.17a
Ka	7.35 ± 0.14	7.18 ± 0.20
n	62	51
SNTα1α2β2b
Gmax	16.68 ± 4.3	26.37 ± 5.68
Vrev	69 ± 0.65	71.25 ± 0.79
Va	0.5 ± 1.41	12.98 ± 0.8	12.48a
Ka	6.7 ± 0.35	7.73 ± 0.17
n	9	16
Δ46-LNTα1α2β2b
Gmax	44.77 ± 8.08	31.06 ± 4.85
Vrev	62.58 ± 0.59	68.65 ± 1.21
Va	−2.34 ± 1.52	9.68 ± 0.92	12.02a
Ka	6.86 ± 0.21	7.02 ± 0.14
n	19	25
Δ139-LNTα1α2β2b
Gmax	38.16 ± 7.68	28.00 ± 5.64
Vrev	62.05 ± 1.08	64.74 ± 1.14
Va	1.57 ± 1.15	8.77 ± 1.29	7.2a
Ka	7.21 ± 0.24	7.15 ± 0.33
n	14	8
Δ139-LNTα1α2β2a (20 mm Ba2+)
Gmax	25.95 ± 6.72	25.82 ± 4.90
Vrev	61.27 ± 0.38	63.55 ± 0.73
Va	−1.84 ± 1.10	7.31 ± 1.27	9.15a
Ka	6.83 ± 0.47	7.34 ± 0.29
n	4	7
WIRα1α2β2b
Gmax	22.44 ± 2.47	22.21 ± 2.35
Vrev	63.99 ± 0.90	67.10 ± 0.89
Va	−2.44 ± 0.86	5.51 ± 0.96	7.94a
Ka	6.41 ± 0.26	6.86 ± 0.19
n	22	21
SNTα1α2β2b
Gmax	48.9 ± 8.72	48.29 ± 8.35
Vrev	71.48 ± 3.04	79.22 ± 8.15
Va	6.1 ± 1.19	14.34 ± 1.01	8.24a
Ka	6.52 ± 0.32	6.72 ± 0.16
n	7	9
Δ20-LNTα1α2β2b
Gmax	27.21 ± 5.33	36.57 ± 2.59
Vrev	76.18 ± 4.78	71.60 ± 1.32
Va	5.40 ± 1.12	11.66 ± 0.82	6.26a
Ka	6.71 ± 0.76	6.64 ± 0.26
n	4	7
Δ20–46-LNTα1α2β2b
Gmax	24.63 ± 3.93	21.69 ± 2.61
Vrev	72.36 ± 5.68	72.97 ± 4.26
Va	5.38 ± 2.20	15.48 ± 0.70	10.11b
Ka	7.60 ± 0.54	6.79 ± 0.10
n	4	5

^a p < 0.001.

^b p < 0.01.

To assess whether the N terminus of α_1C is involved in CaBP1-induced changes in activation, we used several mutants. Δ139-LNTα_1C lacks most of the cytosolic N terminus of α_1C, including both the NSCaTE module and the CaBP1-binding site. Δ46-LNTα_1C lacks the initial segment of the N terminus (46 aa encoded by exon 1a in LNTα_1C or 16 aa encoded by exon 1 in SNTα_1C) but retains the CaM-binding site (NSCaTE module) and the CaBP1-binding segment. The shift in the G-V curve was preserved in both Δ46-LNTα_1C and Δ139-LNTα_1C (Fig. 4). CaBP1 similarly affected the voltage dependence of channels with other N-terminal deletions and mutations: Δ20-LNTα_1C, Δ20-LNTα_1C, the WIR mutant (a triple mutation that abolishes CaM binding to the N terminus (33, 34)), and channels in which the β_2b-subunit was replaced by β_2a or β₃. In all cases, a 6–10-mV shift in V_a was observed upon coexpression of CaBP1 (Table 2). In all, the effect of CaBP1 on voltage-dependent activation appears to be independent of the N terminus of α_1C.

Effect of CaBP1 on VDI Is Regulated by the First 46 aa of the α_1C N Terminus

We next used the N-terminal mutants to examine whether the N terminus of α_1C is involved in CaBP1 regulation of VDI. To quantitate the extent of VDI, in addition to r₄₀₀, we also used r₂₀₀₀ (the fraction of residual current after 2000 ms), which might better reflect changes in slower components of VDI compared with r₄₀₀. We used Δ46-LNTα_1C, Δ139-LNTα_1C, and two additional mutants in which we deleted either the proximal 20 aa (Δ20-LNTα_1C) or the distal 26 aa (Δ20–46-LNTα_1C) from the 46-aa-long initial segment (Fig. 3A) (35). Except for Δ20–46-LNTα_1C, which contains the NTI module and has the low P_o characteristic of LNTα_1C (35), the biophysical properties of these four deletion constructs are generally similar to those of the SNTα_1C isoform.

Our previous work suggested that VDI was accelerated in the Δ139-LNTα_1C mutant (29). In agreement with this, the kinetics of VDI in Δ139-LNTα_1C and, to a smaller extent, in SNTα_1C and Δ46-LNTα_1C are faster than those in wild-type LNTα_1C (Fig. 5, Table 1, and supplemental Fig. S5A). Changes in VDI in Δ20-LNTα_1C and Δ20–46-LNTα_1C were less pronounced and not statistically significant (supplemental Fig. S5A). Together, these results implicate the N terminus of α_1C in the regulation of VDI.

FIGURE 5.

N terminus of α_1C is involved in the regulation of VDI by CaBP1. A, comparison of the time course of decay of I_Ba in various channel constructs tested. Currents were recorded at +10 (black) and +20 (dark gray) mV in the indicated constructs expressed in oocytes without CaBP1. In addition, currents recorded at +20 mV in the presence of CaBP1 are presented (light gray). Currents were normalized and averaged (n = 14). S.E. values are shown for each 20th point (n = 4–14 oocytes). B, the CaBP1-induced shift in activation does not account for its effect on VDI. Normalized I_Ba recorded at +10 mV without CaBP1 (black) is compared with I_Ba recorded at +20 mV in the presence of CaBP1 (gray) in representative cells. C, summary of the experiments shown in A. Values of r₂₀₀₀ measured at +10 mV (black bars) or +20 mV (dark gray bars) and +20 mV when coexpressing CaBP1 (light gray bars) are summarized. The statistical significance of the differences between the control at +10 or +20 mV and CaBP1 at +20 mV (t test) is shown above each pair of bars. The net effect of CaBP1, i.e. the difference in r₂₀₀₀ in CaBP1 (20 mV) versus the control (LNTα_1C at 10 mV) is shown by white bars. Negative values indicate acceleration of VDI by CaBP1. Compared with LNTα_1C, the effect of CaBP1 on Δ46-LNTα_1C and Δ139-LNTα_1C was significantly different, as determined by one-way analysis of variance followed by the Bonferroni t test. These statistical differences are indicated under the corresponding white bars. *, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant.

Unexpectedly, Δ46-LNTα_1C, which still contains the CaBP1-binding site (Fig. 3A), showed little effect of CaBP1 on VDI (Fig. 5A and supplemental Fig. S6A). CaBP1 caused no significant change in r₂₀₀₀ of Δ46-LNTα_1C, compared with an ∼45–50% reduction in r₂₀₀₀ in LNTα_1C or SNTα_1C (at +10 or +20 mV) (Table 1). Similarly, in Δ139-LNTα_1C, CaBP1 barely accelerated VDI, although a small residual effect of CaBP1 was apparent, especially in r₄₀₀ (Fig. 5A, Table 1, and supplemental Fig. S6C).

Considering that the approximately +10-mV shift in activation caused by CaBP1 may alter VDI kinetics, we compared VDI of the various channel constructs at +10 mV in the absence of CaBP1 and at +20 mV in the presence of CaBP1 (Fig. 5, A and B). This comparison also revealed a strong attenuation of the effect of CaBP1 on Δ46-LNTα_1C and Δ139-LNTα_1C channels. The small residual effect of CaBP1 on r₂₀₀₀ seen in both mutants (correlated with residual changes in A_fast and A_slow) (supplemental Fig. S6) was significantly smaller than that seen in the wild-type channels (p < 0.001 compared with LNTα_1C, one-way analysis of variance) (Fig. 5C). We conclude that the initial segments of both LNTα_1C and SNTα_1C isoforms play a major role in transducing the CaBP1-induced acceleration of VDI.

We confirmed these observations in HEK293T cells expressing either the rat SNTα_1C isoform (rbcII) or the rabbit LNTα_1C isoform and several of the N-terminal deletion mutants (supplemental Fig. S5B). Inactivation was measured as r₁₀₀₀ in HEK293T cells as described previously (6). In the wild-type channels, CaBP1 significantly accelerated VDI, decreasing r₁₀₀₀ by 31% in rbcII and by 27% in rabbit LNTα_1C. Deletion of the first 64 aa in rbcIIα_1C or of 46 aa in LNTα_1C completely abolished the CaBP1-induced change in r₁₀₀₀ of I_Ba.

To further define the determinants within the initial 46 aa required for CaBP1 modulation of VDI, we analyzed the Δ20-LNTα_1C and Δ20–46-LNTα_1C deletion mutants. However, both Δ20-LNTα_1C and Δ20–46-LNTα_1C manifested enhancement of VDI by CaBP1 similar to the wild-type channels (Fig. 5 and supplemental Fig. S7), suggesting that either the first or last half of this proximal N-terminal region is sufficient for modulation of VDI by CaBP1.

DISCUSSION

Here, we have demonstrated that the neuronal Ca²⁺-binding protein CaBP1 profoundly regulates multiple aspects of gating of the Ca_V1.2 channel. In addition to the previously described regulation of CDI, we have shown that CaBP1 also regulates activation gating and VDI. The N terminus of α_1C is important for the regulation of VDI by CaBP1. CaBP1 binds to the distal third of the N-terminal tail, adjacent to the first transmembrane domain of the channel, yet deletion mutagenesis showed that the functionally important part of N terminus is its initial segment, which does not bind CaBP1 in a direct biochemical assay. Importantly, the initial segments of both LNTα_1C and SNTα_1C isoforms, encoded by either one of the alternative first exons of the α_1C gene CACNA1C (exon 1a or 1, respectively), function as regulators of the effect of CaBP1 on VDI. This demonstrates, for the first time, a novel modulatory function for the initial segment of the N terminus of the short isoform of α_1C.

Role of the N Terminus of α_1C and Mechanism of Action

Our results indicate that the initial segment of α_1C plays a crucial role in CaBP1 regulation of VDI. The channel construct lacking the initial segment (Δ46-LNTα_1C), which starts essentially with the protein sequence encoded by the obligatory exon 2 (44, 45), retains only a residual small regulation of VDI by CaBP1. Importantly, the effect of CaBP1 is equally restored by the addition of either the short 16-aa initial segment encoded by exon 1 or the long 46-aa segment encoded by the alternative exon 1a. Thus, regulation by CaBP1 is observed in both isoforms of α_1C. (A third N-terminal isoform of α_1C starting with a 9-aa segment encoded by the alternative exon 1c is expressed in rat smooth muscle, but no such isoform was identified in humans (46).)

The ability of the 16 aa-initial segment of the SNTα_1C isoform to regulate CaBP1 modulation contrasts with the inability of this region to regulate the channel's maximal open probability, a function that is unique to the partially homologous 20 aa of the LNTα_1C isoform, the NTI module (35). This distinction indicates that the initial segment of the N terminus regulates maximal P_o and VDI by distinct molecular mechanisms. This concept is supported by the finding that another Ca²⁺-binding protein, KChiP (K_V channel-interacting protein), which interacts with the N terminus of α_1C, regulates only P_o but not inactivation of the cardiac LNTα_1C isoform (10).

Our finding that deletion of most of the N terminus does not influence the effect of CaBP1 on voltage dependence of activation highlights the specificity with which the N terminus regulates the VDI modulation by CaBP1. Specificity is also supported by the preservation of CaBP1-dependent regulation of VDI in mutants in which the first or second portion of the 46-aa segment was deleted, Δ20-LNTα_1C and Δ20–46-LNTα_1C. These two mutants differ by the presence of the NTI module (in Δ20–46-LNTα_1C) and the corresponding ∼7-fold difference in maximal P_o (35). Our finding that CaBP1 still modulates VDI in Δ20-LNTα_1C and Δ20–46-LNTα_1C raises the possibility that the length of the N terminus is the critical parameter. A similar proposal has been made for the N terminus of Ca_Vβ based on the ability of truncations of this region to prevent effects of Ca_Vβ on voltage-dependent inactivation (47, 48). However, we cannot rule out that there are two distinct regions that mediate the effect of CaBP1 on VDI, which are present in the first and second halves of the 46-aa initial segment and in the 16 aa of SNTα_1C.

Interestingly, CaBP1 did not interact directly with the initial N-terminal segment of LNTα_1C, although we cannot rule out that low affinity interactions may have evaded our pulldown assays. Because deletion of most of the N terminus including the CaBP1-binding site does not alter VDI regulation by CaBP1 any more than deletion of the initial segment alone, we propose that the initial segment plays a key role in transducing the effect of CaBP1. This effect may require CaBP1 binding to the distal N terminus and/or to the previously characterized site in the C-terminal domain (6). The initial segment of the N terminus may allosterically couple CaBP1 binding to VDI modulation. We proposed a similar mechanism to account for the ability of the Ca_V β-subunit to inhibit P_o via the NTI module despite lack of a direct Ca_Vβ-N terminus interaction (30).

Complex Regulation of VDI by the N Terminus of α_1C

The mechanism underlying inactivation of high voltage Ca_V channels is not completely understood but may involve “hinged lid” or pore block-type mechanisms (20, 49). The N-terminal tail of α_1C has previously been proposed to regulate the inactivation process (29, 31), in addition to other well established determinants (the C-terminal tail, cytosolic loop I, and the pore itself in α_1C; Ca_Vβ; and CaM) (18, 20, 21, 50). Our results further suggest a complex regulation of the inactivation process by multiple parts of the N-terminal tail of the Ca_V1.2 α_1C-subunit. Thus, despite the crucial role of the initial segment in regulation of VDI by CaBP1, it only moderately affects the VDI process itself: the overall kinetics of VDI are only slightly accelerated by the deletion of the initial segment (46 or 16 aa, respectively) compared with SNTα_1C. Although removal of a part of a protein can distort its conformation and affect functions that are normally not controlled by this part, we note that the Δ139-LNTα_1C mutant shows otherwise standard activation gating, P_o, amplitude, and regulation by Ca_Vβ similar to those of the SNTα_1C isoform (35).

CaBP1 Regulates VDI in Ca_V1.2, and Possible Physiological Implications

The depolarizing shift in the voltage dependence of activation and the acceleration of VDI are expected to reduce Ca²⁺ entry in excitable cells. On the other hand, CDI is reduced by CaBP1, which would enhance Ca²⁺ entry (6). Dual opposite regulation of VDI and CDI resembles the effect of sorcin (8) and may be physiologically relevant. For Ca_V1.2 channels in ventricular cardiomyocytes, VDI may primarily regulate channel function under basal conditions, whereas CDI is dominant upon β-adrenergic stimulation (51, 52). The importance of VDI as a Ca_V1.2 regulatory mechanism is further underscored by findings that the Timothy syndrome mutation G406R, which leads to autism, cardiac arrhythmia, and developmental abnormalities (53), inhibits VDI but spares CDI (54). CaBP1 is colocalized with L-type Ca²⁺ channels in somatodendritic areas of neurons in the brain (4) and is well posed to regulate their function. Therefore, CaBP1 regulation of VDI and CDI may contribute significantly toward activity-dependent regulation of neuronal Ca²⁺ signals.