Ca²⁺-sensitive inactivation of L-type Ca²⁺ channels depends on multiple cytoplasmic amino acid sequences of the α_1C subunit

Abstract

Ca²⁺-dependent inactivation of Ca²⁺ currents is a physiological phenomenon widely associated with L-type Ca²⁺ channels. Although the pore-forming α_1C subunit of the channel is the target for Ca²⁺ binding, the amino acid sequences involved in the binding and/or in the coordination of Ca²⁺-dependent inactivation are still unclear. Based on previous experiments, we have prepared truncation mutants of a human α_1C subunit by systematically deleting an EF-hand motif and sequences in a segment of 80 amino acids in the carboxyl-terminal tail. We found that the rate as well as the Ca²⁺ dependence of inactivation of currents through these mutated channels were very different. We have identified three amino acid sequences, the presence of which is important for Ca²⁺-dependent inactivation: (i) a putative Ca²⁺-binding EF-hand motif, (ii) two hydrophilic residues (asparagine and glutamic acid) 77–78 amino acids downstream of the EF-hand motif, and (iii) a putative IQ calmodulin binding motif. We suggest that Ca²⁺-dependent inactivation is a cooperative process involving several amino acid sequences in cytoplasmic segments of the α_1C subunit.

The voltage-gated L-type Ca²⁺ channel is an essential part of signal transduction systems in many cell types. The channel is a heteromeric protein complex that occurs in several subunit combinations (1, 2). The cardiac type Ca²⁺ channel consists of the pore-forming α_1C subunit and auxiliary β and α₂δ subunits. Several splice variants of the gene for the α_1C subunit of human L-type Ca²⁺ channels have been identified (3–5), and some of them have been expressed in Xenopus oocytes (4, 6, 7). This work has led to the identification of α_1C subunits through which currents with very different inactivation properties could be activated. Ba²⁺ currents through a subunit named α_1C,77 exhibited slow inactivation kinetics, whereas those of currents through a chimeric α_1C,86 channel were fast (7). Furthermore, with the physiological Ca²⁺ ions as charge carriers, inactivation of the current through α_1C,77 was greatly accelerated, in contrast to α_1C,86, where it was very similar in both Ba²⁺ and Ca²⁺ solutions. The two α_1C subunits differed in a stretch of 80 or 81 amino acids in the carboxyl-terminal tail of the protein (7).

Ca²⁺-dependent inactivation is a typical feature of many L-type Ca²⁺ channels (8). It may be a physiological safety mechanism against a harmful Ca²⁺ overload in the cell, notably during large Ca²⁺ currents through the channels. It has been shown to be caused by binding of Ca²⁺ to the pore-forming α_1C subunit (9–12), although the location of the Ca²⁺ binding sites are highly controversial. There is general agreement that the long cytoplasmic, carboxyl-terminal tail of α_1C subunits is involved in Ca²⁺-dependent inactivation. However, although one group (13) found the distal half of the C-terminal tail important, others (7, 14, 15) identified amino acid sequences in the proximal part as being essential for Ca²⁺-dependent inactivation. An EF-hand motif in this region has been suggested as a crucial Ca²⁺-binding site for Ca²⁺-dependent inactivation (14, 16). However, this was not confirmed by Zhou et al. (15). Based on our previous experiments with the α_1C,77 and α_1C,86 subunits (7), we now have approached the problem by systematically deleting the EF-hand motif and segments within the stretch of 80 amino acids that we had identified as being critical for Ca²⁺-dependent inactivation (7). By this method we have found that the putative Ca²⁺-binding EF-hand motif, two hydrophilic amino acid residues (asparagine and glutamic acid) downstream of the EF-hand motif, and a consensus IQ calmodulin binding motif help to support this type of inactivation. However, other sites may well be of additional importance in the coordination of this process (13).

METHODS

Construction and in Vitro RNA Transcription of α_1C,77 Deletion Mutants.

Details of the preparation of the plasmid pHLCC77 harboring the cDNA that encodes the subunit α_1C,77 have been described elsewhere (6). The NsiI–BamHI fragment of pHLCC77 has been subcloned into the PsiI–BamHI polylinker sites of pBluescript SK(−). This construct (p77NB) was used for all subsequent manipulations of C-terminal sequences of α_1C,77. The deletion mutant 77d80 (Fig. 1) has been constructed by using the Seamless Cloning Kit (Stratagene). All other deletions (Fig. 1) have been introduced into p77NB by PCR, taking advantage of the unique feature of Pfu DNA Polymerase (Stratagene) to create blunt-ended amplification products. For each deletion two pairs of phosphorylated primers have been designed to amplify two fragments that flank the sequences to be deleted. After isolation and proper restriction of the fragments, they have been cloned into p77NB by triple ligations and were further subcloned into the full-length pHLCC77. All mutations have been confirmed by sequencing. Plasmids encoding mutated α_1C subunits as well as the auxiliary subunits β₁ (17, 18) and α₂δ (19), kindly provided by V. Flockerzi (University of Saarland, Homburg/Saar) and F. Hofmann (Technical University, Munich), were appropriately linearized before in vitro transcription by using Ambion’s T7 mMessage mMachine Kit.

Figure 1.

Schematic presentation of the L-type Ca²⁺ channel α_1C,77 subunit (Top) and its deletion constructs used in this study. The pore-forming α_1C subunit consists of four homologous, membrane-spanning domains (I–IV). For simplicity, our full-length reference channel α_1C,77 is called 77, and deletions derived from 77 are defined as 77dXX, where XX stands for the number of amino acid residues deleted in the C terminus of α_1C,77, except for 77dEF, where EF stands for the deleted EF-hand motif. The deletion mutants constructed are (positions of deleted amino acid residues): 77d80 (1572–1651), 77d42 (1572–1613), 77d31 (1572–1602), 77d21 (1572–1592), 77d12 (1572–1583), 77d68 (1584–1651), 77d38 (1614–1651), 77d48 (1584–1631), 77d18 (1614–1631), 77d08 (1624–1631), 77d20 (1632–1651), 77d30 (1584–1613), 77d10 (1572–1581), 77d50 (1584–1613, 1632–1651), and 77dEF (1496–1524). Bars underneath the constructs highlight amino acids sequences important for Ca²⁺-dependent inactivation.

Functional Expression and Electrophysiological Recordings of Constructs in Xenopus Oocyte.

Xenopus oocytes were isolated according to standard procedures (20). cRNAs for mutated α_1C subunits were mixed with those for α₂δ and β₁ subunits in equimolar ratios at concentrations of 0.1–1.0 ng/nl. About 50 nl of these mixtures were injected into defolliculated Xenopus laevis oocytes. The oocytes were stored at 18°C for 5–8 days before the electrophysiological recordings began. Whole-cell Ba²⁺ currents (I_Ba) or Ca²⁺ currents (I_Ca) were measured by a conventional two-electrode voltage-clamp method. The glass electrodes were filled with 3 M CsCl and had resistances between 0.2 and 1 MΩ. During I_Ba recordings, the oocyte was constantly superfused with a solution containing 40 mM Ba(OH)₂, 50 mM NaOH, 1 mM KOH, 10 mM Hepes (adjusted to pH 7.4 with methanesulfonic acid). When I_Ca was measured, Ba(OH)₂ was replaced by Ca(NO₃)₂, and the oocytes were injected beforehand with 50 nl of a 40 mM BAPTA (1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid Na₄ salt) solution (pH 7.4) to inhibit Ca²⁺-activated Cl⁻ currents.

In some instances an endogenous I_Ba or I_Ca could be detected in injected oocytes (21). This current is dihydropyridine (DHP)-insensitive and therefore pharmacologically distinct from currents through overexpressed α_1C channels. The deletion mutants described in this study expressed peak I_Ba averaging more than 1 μA, except for 77d48 (0.59 ± 0.17 μA, n = 5), 77d12 (0.44 ± 0.04 μA, n = 7), and 77dEF (0.80 ± 0.32 μA, n = 6). The contaminating fraction of endogenous current for constructs with low expression was determined by bath application of 5 μM (+)-isradipine at the end of an experiment. Experiments showing substantial DHP-insensitive residual current (>10% of peak current) have been discarded. In the case of construct 77d12, I_Ca could be measured reliably in only one experiment because of contaminating endogenous I_Ca.

Voltage-clamp protocols, current recordings, and leak subtractions were performed by using the EPC software (Cambridge Electronic Design, Cambridge, UK). Currents were filtered at 1 Hz and sampled at 2 Hz. For the construction of inactivation curves, 2-s prepulses (PPs) to different voltages, from a holding potential (V_h) −90 mV, were followed by a test pulse (TP) to +20 mV. The 2-s PPs were not quite long enough to achieve steady-state inactivation (7). We refer therefore to our inactivation curves with 2-s PPs as “isochronic” instead of steady-state.

RESULTS

Analysis of Currents Through α_1C,77 and α_1C,77d80.

In a previous paper (7) we have compared electrophysiological properties of the α_1C subunit of a human fibroblast L-type Ca²⁺-channel isoform (α_1C,77) with a chimeric α_1C,86 homologue. The latter was prepared by incorporation of a partial cDNA clone, isolated from human hippocampus (5), into the nucleotide sequence of the recombinant plasmid for α_1C,77. This resulted in the replacement of 80 amino acid residues (1572–1651) in α_1C,77 by 81 nonidentical amino acids of α_1C,86. This replacement produced a marked acceleration of inactivation and a loss of Ca²⁺ sensitivity of inactivation of Ca²⁺ currents (7). Like in the previous study, α_1C,77 is our reference channel in this paper. Sequence alignment revealed more than 93% identity with α_1C from rabbit cardiac muscle.

We began this study by deleting the 80 amino acids in the α_1C,77 subunit (Fig. 1) that previously had been replaced by 81 amino acids in α_1C,86 (7). For simplicity, from here on we will use the nomenclature 77 instead of α_1C,77, and for the deletion mutants 77d. Thus, 77d80 refers to the α_1C,77 subunit where the 80 amino acids 1572–1651 have been deleted (Fig. 1). When either 77 or 77d80 were coinjected with the β₁ and α₂δ subunits into Xenopus oocytes at a 1:1:1 molar ratio, the resulting currents were very different (Fig. 2A). Whereas Ba²⁺ currents (I_Ba) through the 77 channel showed the slow inactivation behavior typical for α_1C, inactivation of I_Ba through 77d80 was as fast as in the previously described chimeric channel α_1C,86 (7). Furthermore, Ca²⁺ currents (I_Ca) through the 77 channel showed a pronounced acceleration of inactivation (7), whereas the rate of inactivation of currents through 77d80 was very similar with both Ba²⁺ and Ca²⁺ as charge carriers (Fig. 2A). These findings suggest that the presence of the 80 amino acids in 77 is important for (i) the slow rate of inactivation of I_Ba, and (ii) the Ca²⁺ sensitivity of inactivation in this α_1C subunit.

Figure 2.

Properties of I_Ba and I_Ca through constructs 77 and 77d80 expressed in Xenopus oocytes. (A) C-terminal amino acid sequences between positions 1572 and 1651 (Fig. 1) are shown schematically above the current traces. Ionic currents have been elicited by applying in 30-s intervals a series of 400-ms TPs with 10 mV increments from V_h = −90 mV. Representative families of current traces elicited by TPs between 0 and +50 mV are shown. Bath solutions contained either 40 mM Ba²⁺ (Upper, filled symbols) or 40 mM Ca²⁺ (Lower, open symbols) as charge carriers. Normalized I–V curves (B) and isochronic inactivation curves (C) for I_Ba (filled symbols) and I_Ca (open symbols) through 77 (squares) and 77d80 (circles) were obtained and plotted according to the procedures described in Table 1. Shown are means ± SEM.

Fig. 2 B and C shows some other more subtle differences between 77 and 77d80. The potential ranges for activation of I_Ba and I_Ca through 77 were more negative than for 77d80, whereas those for inactivation were more positive in 77 than in 77d80 (Table 1). These differences are again very similar to those of α_1C,77 and α_1C,86 in our previous work (7).

Table 1.

Parameters of nonlinear fittings of I-V curves and isochronic (2 s) inactivation curves

α1C subunit	I-V curve*			Isochronic inactivation curve†
	V0.5 (mV)			V0.5 (mV)
	IBa	ICa	n(IBa/ICa)	IBa	ICa	n(IBa/ICa)
77	4.45  ±  0.77	10.30  ±  0.90	13/15	−6.14  ±  0.86	−15.17  ±  1.66	10/6
77d80	18.64  ±  1.08	25.46  ±  1.42	10/4	−17.30  ±  1.28	−20.44  ±  2.02	8/4
77d42	20.41  ±  0.42	27.25  ±  1.23	5/3	−5.21  ±  0.69	−10.22  ±  1.31	5/3
77d31	10.83  ±  3.05	19.26  ±  3.03	6/3	−22.64  ±  1.12	−25.67  ±  2.23	6/3
77d21	8.64  ±  2.51	16.40  ±  2.84	5/3	−18.14  ±  0.68	−22.45  ±  1.01	4/3
77d12	18.95  ±  1.15	17.03	7/1	−8.23  ±  1.20	−16.95	7/1
77d68	18.22  ±  2.32	19.84  ±  2.22	6/6	−16.54  ±  1.68	−23.21  ±  2.49	6/5
77d38	17.73  ±  1.53	22.02  ±  1.44	7/5	−10.82  ±  1.30	−16.03  ±  1.59	6/5
77d48	20.15  ±  2.51	21.23  ±  1.44	5/4	−13.54  ±  2.39	−21.39  ±  2.45	5/4
77d18	19.42  ±  2.36	22.71  ±  1.95	6/4	−9.81  ±  1.93	−14.98  ±  1.53	5/4
77d08	13.57  ±  1.34	18.02  ±  1.92	10/6	−14.96  ±  1.65	−16.36  ±  2.33	5/5
77d20	−0.72  ±  0.87	6.00  ±  1.94	7/5	−17.07  ±  1.22	−27.52  ±  2.22	8/4
77d30	3.41  ±  2.52	9.92  ±  1.66	5/7	−14.55  ±  1.33	−21.64  ±  2.85	6/4
77d10	4.69  ±  2.19	13.16  ±  2.12	9/5	−20.24  ±  1.28	−23.93  ±  2.68	5/3
77d50	10.34  ±  3.38	18.83  ±  1.01	4/3	−15.57  ±  0.07	−22.58  ±  1.13	3/3
77dEF	13.35  ±  1.75	21.60  ±  1.55	6/3	−10.31  ±  3.43	−13.61  ±  2.00	4/3

^*TPs in the range of −40 to +80 mV were applied in 10 mV increments from V_h = −90 mV. I-V curves were fitted by I = G_max(V − E_rev)/{1 + exp[(V − V_0.5)/k]}, where G_max = maximum conductance; E_rev = reversal potential; V_0.5 = voltage at 50% of G_max; k = negative slope factor. 

^†From V_h = −90 mV 2-s conditioning PP were applied in 10 mV increments from −60 up to +20 mV, each followed by a 1-s TP to +20 mV. The interval between TPs was 30 s. Recorded peak current amplitudes were normalized to I_Ba at −60 mV. Isochronic inactivation curves were fitted by a Boltzmann function: I = 1/{1 + exp[(V − V_0.5)/k]}, where V is the conditioning PP voltage, V_0.5 is the voltage at half-maximal inactivation, and k is a slope factor. Values are means ± SEM. n = number of tested oocytes. 

We also have studied the kinetics of inactivation and of the recovery from inactivation of currents through 77 and 77d80 channels, both in Ba²⁺ and in Ca²⁺ solutions. Fig. 3A shows time constants (τ) of the rates of inactivation of I_Ba and I_Ca through 77 and 77d80. Whereas the time courses of I_Ba inactivation in 77 were fitted best by single exponentials, those of I_Ca required a bi-exponential fit, as did those for I_Ba and I_Ca through 77d80 (Table 2). In Fig. 3 A and B only the time constants of the initial exponential decays have been plotted over the potential range from 0 or +10 mV to +50 mV. Fig. 3A shows the slow inactivation kinetics of I_Ba through 77 (▪), and the large decrease in inactivation time constants with Ca²⁺ as charge carrier (□). Note also the increase in τ of I_Ca through 77 at potentials more positive than +20 mV, which is an indication of Ca²⁺-dependent inactivation (8). The currents through 77d80 showed a very different behavior. Fast-inactivation time constants were strongly voltage-dependent, and those of I_Ca (○) were slightly larger than those of I_Ba (•). This result is shown more clearly in Fig. 3B, where all inactivation τs were normalized to those measured for currents at +50 mV.

Figure 3.

Time constants of inactivation and of recovery from inactivation in constructs 77 and 77d80. (A) I_Ba (filled symbols) and I_Ca (open symbols) have been evoked by applying 400-ms pulses to various potentials (abscissa) from V_h = −90 mV. Inactivation time constants (τ) were estimated by exponential fits of I_Ba and I_Ca (Table 2). Only τ of the fast component of the bi-exponential fit has been plotted, except for I_Ba through 77, which was best-fitted, even with 1-s pulses, by a single exponential. The numbers of experiments were: 77 (▪, n = 8; □, n = 10–11), 77d80 (•, n = 8; ○, n = 4). (B) Data from A are normalized to τ at +50 mV. (C) The fractional recovery from inactivation was determined by a two-pulse protocol. A conditioning PP applied from V_h = −90 mV to +20 mV was followed by a TP to +20 mV after variable pulse intervals at V_h. The peak current amplitudes at TPs were plotted against the pulse interval. The PP duration was 3 s for 77 I_Ba, and 400 ms for all other experiments. Full recovery (>98%) was reached for 77 I_Ba after 16 s (▪, n = 4) and for I_Ca after 2 s (□, n = 4); for 77d80 I_Ba after 1 s (•, n = 6) and for I_Ca after 250 ms (○, n = 4). Smooth lines are bi-exponential fits of the averaged data between 5 and 250 ms. All data points are means ± SEM (error bars are in most cases smaller than symbols).

Table 2.

Kinetics of I_Ba and I_Ca inactivation at +20 mV or +30 mV (according to maximal I_Ba)

α1C subunit	Potential at max IBa	IBa				ICa
		τ1 (ms)	τ2 (ms)	A1 (%)	n	τ1 (ms)	τ2 (ms)	A1 (%)	n
77	+20 mV	321.7  ±  19.0	–	–	8	27.8  ±  1.6	202.0  ±  9.0	68.2  ±  1.9	10
77d80	+30 mV	28.8  ±  2.2	105.4  ±  7.6	82.0  ±  2.0	8	46.6  ±  8.7	127.0  ±  12.1	72.8  ±  2.4	4
77d42	+30 mV	37.5  ±  1.6	140.5  ±  7.1	67.1  ±  2.1	5	45.8  ±  2.8	152.8  ±  10.0	59.6  ±  1.1	3
77d31	+20 mV	44.1  ±  4.3	173.3  ±  14.3	46.7  ±  2.3	6	69.9  ±  14.1	261.4  ±  63.3	60.7  ±  0.8	3
77d21	+20 mV	39.2  ±  1.9	189.1  ±  7.1	45.3  ±  2.2	5	49.9  ±  7.8	171.8  ±  24.3	51.1  ±  2.1	3
77d12	+30 mV	41.2  ±  2.2	177.0  ±  13.5	61.2  ±  1.5	7	38.7	139.9	41.3	1
77d68	+30 mV	32.4  ±  3.1	107.6  ±  8.7	73.2  ±  2.7	5	36.6  ±  4.2	112.9  ±  12.4	67.0  ±  3.7	6
77d38	+30 mV	80.1  ±  6.7	301.6  ±  50.9	61.9  ±  2.7	7	84.9  ±  4.9	286.8  ±  24.2	43.5  ±  8.3	5
77d48	+30 mV	48.7  ±  9.8	199.0  ±  47.7	76.3  ±  2.4	4	65.6  ±  8.3	224.9  ±  32.2	69.3  ±  1.7	4
77d18	+30 mV	96.0  ±  17.7	368.4  ±  88.3	62.8  ±  2.8	6	104.3  ±  11.5	337.2  ±  56.8	53.9  ±  2.9	4
77d08	+30 mV	80.3  ±  4.6	332.3  ±  46.9	60.0  ±  2.5	8	101.5  ±  7.3	318.3  ±  26.5	51.2  ±  5.5	6
77d20	+20 mV	359.5  ±  21.3	–	–	6	47.9  ±  6.5	199.2  ±  8.4	52.6  ±  4.5	4
77d30	+20 mV	212.7  ±  7.8	–	–	5	67.7  ±  6.8	239.5  ±  27.8	50.2  ±  4.0	6
77d10	+20 mV	67.5  ±  4.3	362.7  ±  29.9	36.1  ±  1.4	7	71.0  ±  7.4	237.1  ±  23.6	45.5  ±  3.5	5
77d50	+20 mV	52.7  ±  3.3	260.3  ±  9.9	45.5  ±  3.4	4	77.5  ±  13.7	228.2  ±  41.7	35.8  ±  6.0	3
77dEF	+30 mV	31.0  ±  1.7	134.0  ±  10.0	68.7  ±  3.0	6	37.8  ±  0.1	139.8  ±  1.0	55.4  ±  3.8	3

During a 400-ms TP to the indicated potential, inactivation time constants were estimated by fitting the inactivating component of the current trace to the following equation: I =I_o + I₁ exp(−t/τ₁) + I₂ exp(−t/τ₂); I_o is the residual current amplitude at equilibrium, I₁ and I₂ are the amplitudes of the current components. A₁ has been calculated as I₁/(I₁ + I₂). I_Ba of 77, 77d20, and 77d30 were best-fitted by a single-exponential equation. Values are means ± SEM. n = number of tested oocytes. 

The time course of recovery from inactivation was measured by a two-pulse protocol. For 77, the first voltage pulse (PP) from a holding potential of −90 mV to +20 mV lasted 3 s, and the second pulse (TP) to +20 mV lasted 400 ms. This pulse was applied after variable intervals at −90 mV. During the PP the currents became inactivated by 80–90%. The recovery of I_Ba from inactivation during the intervals at −90 mV initially was fast (τ₁ = 17.4 ± 0.9 ms; n = 4) and later on very slow (τ₂ = 779.4 ± 101 ms; n = 4). The fast and the slow components amounted to about 46% and 54% of the total recovery process, respectively (Fig. 3C). The recovery from inactivation of I_Ca through 77 was slightly slower than that of I_Ba in its first component (≈40%; τ₁ = 30.9 ± 2.3 ms; n = 4) and considerably faster in its second component (≈60%; τ₂ = 157.6 ± 11.3 ms; n = 4). In 77d80 I_Ba and I_Ca recovered almost completely during a 250-ms time interval, although the initial time constant of recovery was slightly slower for I_Ca than for I_Ba (27.9 ± 3.0 ms vs. 19.5 ± 0.9 ms; n = 4). The slow components of recovery (τ₂ = 179.2 ± 24.4 ms and 121.8 ± 33.8 ms, respectively) contributed 17% and 25% to the total time course of I_Ca and I_Ba.

The increase in the rate of inactivation when Ca²⁺ instead of Ba²⁺ was the charge carrier through the 77 channel (Fig. 3 A and B) points to a Ca²⁺-dependent inactivation process (8). A crucial test for Ca²⁺-dependent inactivation is the demonstration that it does depend on the size of I_Ca through the open channel. This was tested by a standard two-pulse protocol (Fig. 4A). The potentials during the first pulses (400 ms) were changed in 10-mV steps between −40 and +80 mV from a holding potential of −90 mV. During these PPs, current-voltage relationships were established for I_Ba and subsequently for I_Ca in the same oocytes. After PPs, the potential was held for 50 ms at −90 mV before a constant TP to +20 mV was applied. The extent of recovery of I_Ba or I_Ca after PP during the 50-ms time interval at −90 mV was measured by reactivating the currents at TP. If the current at TP is smallest when the current during PP is largest, a current-dependent inactivation mechanism is suggested. The results of such experiments are shown in Fig. 4B. The peak currents activated by PP (current-voltage relationships) under each conditions were normalized for both channels to the respective maximal current amplitudes. As described above, activation of the currents during PP occurred about 14 mV more positive for 77d80 than for 77 (Fig. 4B, filled symbols). Normalization of the current amplitudes was necessary because of the variability of expression of the channels between different oocytes, and because of the smaller current amplitude of I_Ca as compared with I_Ba. The latter is because of the lower permeability of the channels for Ca²⁺ than for Ba²⁺ (22). In Ba²⁺ solution the peak currents through 77 and 77d80 at TP, which were normalized to those measured after a PP to −40 mV, show a slight tendency to become smaller with increasing PP potentials (Fig. 4B, Upper, open symbols). In contrast to Ba²⁺, with Ca²⁺ as charge carriers (Fig. 4B, Lower) there is an important difference between the 77 and 77d80 channels in the relative current amplitudes at TP. For 77d80, peak I_Ca and I_Ba at TP show a similar tendency to become smaller with increasing PPs (Fig. 4B, ○), whereas I_Ca through 77 displays a bell-shaped relationship between normalized current at TP and PP potentials (Fig. 4B, Lower, □). Peak I_Ca at TP is smallest when I_Ca activated by PPs to +20 or +30 mV is largest (7). Subtraction of I_Ba from I_Ca at TP yields the Ca²⁺-sensitive component of inactivation during PP for both channel types. This result is shown more clearly in Fig. 4C. The open circles show the subtracted currents [1 − (I_Ca − I_Ba)] for 77d80, and the open squares show those for 77. The difference between the two curves in Fig. 4C is an indication for the Ca²⁺ sensitivity of inactivation of I_Ca in the 77 channel. The larger the Ca²⁺ influx through the channels during PP, the larger is the extent of Ca²⁺-dependent inactivation measured at TP.

Figure 4.

Ca²⁺-dependent inactivation depends on Ca²⁺ influx. (A) Two-pulse voltage protocol. Conditioning PPs applied in 10 mV increments from −90 mV to voltages ranging from −40 mV to +80 mV are each followed by a TP to +20 mV after a 50-ms pulse interval at −90 mV. (B) For 77 and 77d80, peak currents at PPs were normalized to the maximal current amplitudes (filled symbols), and those at TPs to the current amplitude obtained after a PP to −40 mV (open symbols). Normalized current amplitudes for I_Ba (Upper) and I_Ca (Lower) are plotted against the PP potential. For I_Ca through 77 (□) the bell-shaped relationship at TP indicates Ca²⁺-dependent inactivation. I_Ba and I_Ca were always recorded from the same Xenopus oocyte. Numbers of experiments: 77 (▪, □, n = 5), 77d80 (•, ○, n = 4). (C) Subtraction of I_Ba from I_Ca at TP for both constructs yields a U-shaped relationship for 77 (□) and a more steady decrease for 77d80 (○). The difference between the two curves is a measure of Ca²⁺-dependent inactivation. Data represent means ± SEM.

Analysis of Currents Through Other Deletion Mutants of α_1C,77.

The main task of our study was the attempt to determine sequences within the segment of 80 amino acids in the carboxyl-terminal region of α_1C,77 that are involved in Ca²⁺-dependent inactivation. Therefore, a similar analysis, as described extensively above for the α_1C channels 77 and 77d80, has been made for all the constructs shown in Fig. 1. The results for half-maximal voltages (V_0.5) of activation and inactivation of I_Ba and I_Ca through these channels, as well as time constants of inactivation, and the respective fractional components of inactivation governed by these time constants, are summarized in Tables 1 and 2. The most interesting differences between the various deletion mutants, however, concern their sensitivities to Ca²⁺-dependent inactivation.

For the analysis of data where I_Ba and I_Ca are directly compared, only those experiments have been selected where both currents were measured in the same oocyte. However, the channel 77d12 expressed poorly, and therefore I_Ba and I_Ca both could be measured in only one oocyte where no obvious differences in the time courses could be seen (Table 2). In Fig. 5, I_Ca and I_Ba through eight α_1C constructs are depicted. All current amplitudes have been scaled to the respective peak I_Ba. The reference channel 77 showed prominent Ca²⁺-dependent inactivation, as described above. The deletion mutants 77d68, 77d42, and 77d08 showed fast or moderately fast (77d08) inactivation in Ba²⁺ and Ca²⁺ solutions and no clear increase in the rates of inactivation by Ca²⁺. The slight slowing of the initial rate of inactivation in Ca²⁺ solution also was observed with other deletion mutants (Table 2). Constructs 77d68 and 77d42 include deletions of either the 18 amino acids 1614–1631 or the 12 amino acids 1572–1583, respectively. In construct 77d08 only the eight amino acids 1624–1631 have been deleted. By contrast, the deletion mutants 77d50, 77d30, and 77d20 contain the full amino acid sequences 1572–1583 and 1614–1631, but 50, 30, or 20 amino acids are absent in other parts of the sequences. In 77d10 only the segment 1572–1583 was truncated by the first 10 amino acids. The currents through these four channels all exhibited Ca²⁺-sensitive inactivation (Fig. 5, lower four traces), though to different extents. The inactivation rates of currents in Ca²⁺ solution were similar in all four deletion mutants. However, the contribution of Ca²⁺-dependent inactivation to the total inactivation process was different. The current through channel 77d20 had slow inactivation rates in Ba²⁺ solution and showed the most prominent Ca²⁺-dependent inactivation. I_Ba through 77d50 showed the fastest inactivation rate, whereas that of I_Ca was only moderately accelerated.

Figure 5.

I_Ba and I_Ca traces of selected α_1C constructs. A 400-ms pulse was applied from V_h = −90 mV to potentials that evoked approximately maximal I_Ba at +20 mV for 77, 77d20, 77d30, 77d10, 77d50, and at +30 mV for 77d68, 77d42, and 77d08. After switching from a 40 mM Ba²⁺- to 40 mM Ca²⁺-containing solution, the same TPs were applied. I_Ca traces were scaled to the amplitude of the corresponding I_Ba trace to compare time courses of I_Ba and I_Ca. Deletions of C-terminal amino acid residues between positions 1572 and 1651 for each construct are shown above each pair of current traces. Stretches of amino acids in the sequences 1572–1584 and 1614–1631 that were found to be important for Ca²⁺-dependent inactivation in each construct (Fig. 1) are underlined.

To establish the dependence of the Ca²⁺ sensitivity of inactivation on Ca²⁺ influx through all these channel mutants, we have applied the same double-pulse protocols and procedures as described for 77 and 77d80 (Fig. 4 A–C). For constructs 77d10, 77d20, 77d30, and 77d50 the currents at TP became smallest when the preceding I_Ca at PP was largest. Such a relationship was not found for the other deletion mutants. Normalized peak I_Ba at TP was subtracted from each normalized I_Ca at the same potential for every oocyte. The subtracted currents [1 − (I_Ca − I_Ba)] are plotted (Fig. 6 A and B) for all constructs shown in Fig. 5. Although clear current minima at TP +20 mV were obtained for the reference channel 77 and for the deletion mutants 77d10, 77d20, 77d30, and 77d50 (Fig. 6A), no such minima could be seen for currents through 77d68, 77d42, and 77d08 (Fig. 6B) and for all other constructs in Tables 1 and 2. Notably the channel 77d08 lacks only the eight amino acids 1624–1631 (Fig. 1). Thus, we conclude that the two amino acids 1582–1583 (channel 77d10) plus the eight amino acids 1624–1631 (channel 77d08) are important sequences involved in Ca²⁺-dependent inactivation. Moreover, deletion of an EF-hand motif (14–16), 47 amino acids upstream of the stretch of 80 amino acids studied here, also led to a loss of Ca²⁺-dependent inactivation and to a pronounced acceleration of I_Ba inactivation (Table 2).

Figure 6.

Ca²⁺ sensitivity of inactivation in selected deletion mutants. By using a double-pulse protocol and following the same procedure as described in Fig. 4, normalized I_Ba at TP were subtracted from normalized I_Ca and plotted against the PP potential. (A) Reference channel 77 (▪, n = 5) and its deletion mutants 77d50 (◊, n = 5), 77d30 (○, n = 6), 77d20 (▴, n = 5), and 77d10 (•, n = 5) show U-shaped curves, with minima at +20 to +30 mV. (B) For 77d68 (▾, n = 6), 77d42 (⧫, n = 3), and 77d08 (□, n = 6) no clear minima were seen. (C) A calcium sensitivity index has been calculated and plotted for 77, 77d10, 77d20, 77d30, and 77d50, on the basis of the data presented in A. Normalized currents at TP elicited after a PP to −40 mV (I₋₄₀) and +80 mV (I₊₈₀) were averaged and subtracted from the normalized current minima at TP (I_min) for each construct: [(1 − I_min) − (I₋₄₀ − I₊₈₀)/2]. For all other deletion mutants the calcium sensitivity index was smaller than 0.1. Shown are means ± SEM.

To provide a more quantitative picture of Ca²⁺-dependent inactivation of currents through the α_1C constructs 77, 77d10, 77d20, 77d30, and 77d50, we have calculated a “calcium sensitivity index.” The index is based on the data shown in Fig. 6A. The average of the normalized currents at TP elicited after PPs to −40 mV and +80 mV (I₋₄₀ and I₊₈₀) have been subtracted from the normalized minimal currents (I_min) at TP for each construct [(1 − I_min) − (I₋₄₀ − I₊₈₀)/2]. The difference is a measure of the Ca²⁺-dependent fraction of inactivation. This calcium sensitivity index has been plotted in Fig. 6C and shows that Ca²⁺-dependent inactivation is strongest in 77 and weakest in 77d50. All other constructs had an index of less than 0.1, with I_min after a PP to +30 mV. Other sequences in the carboxyl-terminal cytoplasmic stretch of 80 amino acids, notably within the segment 1584–1623, are significant in determining the rates of voltage-dependent inactivation (Table 2; see also ref. 23).

A summary of our conclusions is provided in Fig. 7. The stretch of 80 amino acids in the carboxyl-terminal end of α_1C,77 is shown together with the deletions. The bars above the amino acids correspond to the underlined sequences in the upper scheme of mutations in Fig. 7. Elimination of either one of these sequences abolished Ca²⁺-dependent inactivation. The dashes above the bars indicate amino acid sequences that were essential for the conservation of Ca²⁺-dependent inactivation in our experiments with channels 77d10 and 77d08. Inactivation properties of the currents through all channel mutants are indicated on the right. The list provides a qualitative survey of inactivation rates of the currents in Ba²⁺ solutions and of their Ca²⁺ sensitivity in Ca²⁺ solutions.

Figure 7.

Summary of inactivation characteristics for all deletion mutants studied. The schematic presentation of α_1C and its deletion mutants (Top and Left) is as in Fig. 1. The two columns to the right qualitatively assess the inactivation rates of I_Ba (Table 2) and the calcium sensitivity of each construct. Calcium sensitivities are based on calcium sensitivity indices as shown in Fig. 6C. (Bottom) The amino acid sequence between positions 1572 and 1651 of α_1C,77 is shown. The heavy bars above the sequence correspond to the underlined sequences in the upper schemes of mutants. The dashed lines above the bars indicate amino acid residues that were found to be essential for the conservation of Ca²⁺-dependent inactivation.

DISCUSSION

Recently, several articles have been published indicating an important role for Ca²⁺-dependent inactivation of the carboxyl-terminal cytoplasmic tail in the pore-forming α_1C subunit of L-type Ca²⁺ channels (7, 13–15). However, the exact location of this regulation is still very controversial. A putative Ca²⁺-binding EF-hand motif near the transmembrane region IVS6 has been shown by one group (14) to be essential for Ca²⁺-dependent inactivation. But others (15) could not confirm this finding. Disruption of possible Ca²⁺-coordination sites, by replacement of charged amino acids with alanine in the EF-hand motif, had little effect on Ca²⁺-dependent inactivation. Deletion of the EF-hand sequence prevented functional expression of the α_1C channel in the surface membrane of the oocytes. Those authors (15), however, found that replacement of 134 amino acids in α_1E by the homologous 142 amino acid sequence of α_1C, downstream of the EF-hand motif, conferred Ca²⁺-dependent inhibition to the normally Ca²⁺-insensitive α_1E subunit. In our 77 channel, deletions have been made within the same sequence. Although our previous (7) and the present results agree with those by Zhou et al. (15) in showing the importance of sequences downstream of the EF-hand motif, in our hands deletion of this motif did not impede expression of the protein, and it eliminated Ca²⁺-dependent inactivation. These findings indicate that multiple amino acid sequences in the first half of the putative cytoplasmic carboxyl-terminal region of α_1C subunits can participate in the coordination of Ca²⁺ binding that is required for Ca²⁺-dependent inactivation. Further sequences that have been described to be important for Ca²⁺-dependent inactivation are the putative cytoplasmic loops between transmembrane domains I-II and II-III (13).

Recently, Soldatov et al. (23) have shown that replacement of amino acid segments 1572–1598 or 1595–1652 in the carboxyl terminus of α_1C,77 by the corresponding sequences of α_1C,86 converts the slowly inactivating α_1C,77 into the fast-inactivating α_1C,86 channel. This finding agrees with our results where deletion of segments 1572–1583 or 1614–1631 in α_1C,77 produced fast Ca²⁺-independent inactivation. They also showed that replacement of the amino acids 1572–1576 in α_1C,77 produced a moderately fast, but still Ca²⁺-sensitive inactivation. This sequence is part of our 77d10 deletion mutant that showed a similar behavior. However, Soldatov et al. (23) could further accelerate inactivation and remove its Ca²⁺ dependence by the additional replacement of amino acids 1600–1604. Although we have not tried a corresponding deletion in our experiments, we found that deletion of the more distal sequence 1624–1631 converts the slow Ca²⁺-sensitive inactivation of channel 77 into the faster Ca²⁺-insensitive inactivation of channel 77d08.

As pointed out to us by R. W. Tsien and K. Deisseroth (personal communication), the sequence 1624–1631 (see Fig. 7) is the initial part of a stretch of 12 amino acids that appears to be a consensus IQ motif for calmodulin binding. This is of special interest not only because of our finding that this motif is involved in Ca²⁺-dependent inactivation, but also in view of its possible importance for intracellular signaling mechanisms (24). The conserved positions in the consensus sequence of several Ca²⁺-independent calmodulin-binding motifs are IQXXXRG (or K)XXXRX (25). If Ca²⁺ binds to calmodulin attached to this motif, this reaction could become part of the possible cooperativity between multiple sites in the α_1C subunits that are involved in Ca²⁺-dependent inactivation. Such a hypothesis appears to be incompatible with the finding that the calmodulin inhibitor calmidazolium has no effect on Ca²⁺-dependent inactivation (9, 26). We have confirmed this result (unpublished observation). However, it is possible that calmidazolium has a very low affinity to calmodulin when it is bound in a Ca²⁺-independent way to the IQ motif. The IQ consensus motif also is conserved in α_1D- and α_1S subunits of L-type Ca²⁺ channels (27). Although Ca²⁺-dependent inactivation of I_Ca seems to be present in α_1D of neuroendocrine cells (26), this is not the case for α_1S (13, 28). Because, however, other molecular structures that are involved in Ca²⁺-dependent inactivation in α_1C are absent in α_1S (13), this may explain the difference between the two channels. The IQ motif is absent and the EF-hand motif scores low in the Tufty–Kretsinger test (14) in α_1E, which does not exhibit Ca²⁺-dependent inactivation (14, 15). Another puzzling result concerns the Hill coefficient near 1 for Ca²⁺-induced inactivation (12). This result could be explained, however, by the sum of several steps with low cooperativity involved in Ca²⁺-dependent inactivation.

In conclusion, we favor a hypothesis for Ca²⁺-dependent inactivation that requires multiple amino acid sequences in the cytoplasmic regions of the α_1C subunit of L-type Ca²⁺ channels. Two possible Ca²⁺-binding sites that may be directly involved in the cooperative interaction between several sequences in α_1C could be the putative Ca²⁺-binding EF-hand and the calmodulin-binding IQ motifs. We imagine coiling and interaction of the flexible cytoplasmic sequences and, thereby, plugging of the pore. Further studies involving extensive site-directed mutagenesis are required to substantiate this hypothesis.

ABBREVIATIONS

I_Ba

whole-cell Ba²⁺ currents

I_Ca

whole-cell Ca²⁺ currents

PP

prepulse

TP

test pulse