Inactivation determinants in segment IIIS6 of Ca_v3.1

Abstract

1. Low threshold, T-type, Ca²⁺ channels of the Ca_v3 family display the fastest inactivation kinetics among all voltage-gated Ca²⁺ channels. The molecular inactivation determinants of this channel family are largely unknown. Here we investigate whether segment IIIS6 plays a role in Ca_v3.1 inactivation as observed previously in high voltage-activated Ca²⁺ channels.

2. Amino acids that are identical in IIIS6 segments of all Ca²⁺ channel subtypes were mutated to alanine (F1505A, F1506A, N1509A, F1511A, V1512A, F1519A, FV1511/1512AA). Additionally M1510 was mutated to isoleucine and alanine.

3. The kinetic properties of the mutants were analysed with the two-microelectrode voltage-clamp technique after expression in Xenopus oocytes. The time constant for the barium current (I_Ba) inactivation, τ_inact, of wild-type channels at −20 mV was 9.5 ± 0.4 ms; the corresponding time constants of the mutants ranged from 9.2 ± 0.4 ms in V1512A to 45.7 ± 5.2 ms (4.8-fold slowing) in M1510I. Recovery at −80 mV was most significantly slowed by V1512A and accelerated by F1511A.

4. We conclude that amino acids M1510, F1511 and V1512 corresponding to previously identified inactivation determinants in IIIS6 of Ca_v2.1 (Hering et al. 1998) have a significant role in Ca_v3.1 inactivation. These data suggest common elements in the molecular architecture of the inactivation mechanism in high and low threshold Ca²⁺ channels.

Low voltage-activated Ca²⁺ channels of the Ca_v3 family (also known as α_1G/H/I or T-type channels; Ertel et al. 2000) mediate Ca²⁺ influx into neurons, endocrine cells and muscle cells. Ca²⁺ entry through voltage-gated ion channels is affected by the rate of channel inactivation.

High voltage-activated Ca²⁺ channels such as Ca_v1.2 and Ca_v2.1 inactivate by at least two voltage-dependent mechanisms (fast and slow inactivation) and by an additional Ca²⁺-dependent inactivation (see Hering et al. 2000). Low voltage-activated Ca²⁺ channels display the fastest inactivation kinetics among all Ca²⁺ channel subtypes and inactivate in a voltage-dependent manner (Perez-Reyes et al. 1998). As observed with voltage-gated Na⁺ and high voltage-activated Ca²⁺ channels, the α₁ subunit of T-type channels contains four internal repeats, with each repeat containing six transmembrane segments and a pore loop.

Although there is no direct information on the structure of these α₁ subunits, data from the X-ray crystal structure of a potassium channel has been used to glean structural insights (Doyle et al. 1998). Thus, the S6 segments of α₁ subunits are thought to line the inner pore as observed with the second transmembrane segments (TM2) of KcsA. These segments are likely to move during channel gating (see Perozo et al. 1998; Yellen, 1999), making them interesting targets for mutagenesis studies. Such studies on high threshold Ca²⁺ channels have identified a number of important residues in binding of calcium channel blockers, as well as important motifs involved in channel inactivation of Ca_v2.1 (Hering et al. 1997, 1998; Sokolov et al. 2000; Berjukow et al. 2001). Crucial residues have been identified in segments S5, S6, pore loops (Cens et al. 1999; Sokolov et al. 1999), intracellular domain linkers (Herlitze et al. 1997; Spaetgens & Zamponi, 1999; Stotz et al. 2000), and the carboxy (Soldatov et al. 1998) and amino (Stephens et al. 2000) terminal ends.

The current hypotheses on the mechanisms of high threshold Ca²⁺ channel inactivation comprise hinged lid mechanisms (I-II loop), ball and chain mechanisms (carboxy terminus), and non-covalent interaction between pore-forming S6 segments (Hering et al. 2000).

In contrast, little is known about the inactivation determinants of the Ca_v3 family. The data of Staes et al. (2001) suggest a crucial role of the carboxy terminus. Based on our previous studies showing a role of IIIS6 segments in inactivation, we decided to evaluate the role of the IIIS6 segment in Ca_v3.1 inactivation gating by mutating six residues to alanine (F1505A, F1506A, N1509A, F1511A, V1512A, F1519A). In addition we mutated M1510 to either isoleucine or alanine. These amino acids (except M1510) are identical in the IIIS6 segments of all voltage-gated Ca²⁺ channels. M1510, F1511 and V1512 correspond to the IFV-inactivation motif that was previously identified in segment IIIS6 of Ca_v2.1 (Hering et al. 1997, 1998; Sokolov et al. 2000). In order to compare the impact of the corresponding sequence stretches in Ca_v3.1 and Ca_v2.1 we analysed entry and recovery from inactivation of the corresponding Ca_v3.1 mutants at different potentials with the two-microelectrode voltage-clamp technique after expression in Xenopus oocytes. Here we demonstrate that residues in low and high threshold Ca²⁺ channels have a qualitatively similar impact on channel inactivation.

METHODS

Generation of Ca_v3.1 mutants

The residues of segment IIIS6, identical in Ca_V3.1 and high threshold Ca²⁺ channels, were replaced by alanine resulting in mutants F1505A, F1506A, N1509A, F1511A, V1512A and F1519A. Additionally, residue M1510 was replaced by alanine and isoleucine (M1510A; M1510I) and two double-point mutants were created: MF1510/1511AA and FV1511/1512AA.

The Ca_v3.1 mutants were constructed by introducing point mutations into the corresponding rat α₁ subunit cDNA (accession number AF027984, Perez-Reyes et al. 1998) by the ‘gene SOEing’ technique (Horton et al. 1989). All constructs were inserted into the transcription plasmid pGEM-HEA and verified by sequence analysis (Chuang et al. 1998).

Electrophysiology

All animal experiments were approved by the Austrian Federal Ministry of Education, Science and Culture. Preparation of stage V-VI oocytes from Xenopus laevis, synthesis of capped run-off cRNA transcripts from linearized cDNA templates, and injection of cRNA were performed as previously described in detail by Grabner et al. (1996). Stage V and VI oocytes were harvested from female Xenopus laevis frogs under anaesthesia with 3-aminobenzoic acid ethyl ester (Sigma, 1 mg ml⁻¹ in a water bath for 20 min). After oocyte removal, the abdominal incision was re-sutured and the frog was allowed to recover for 3–5 h in 1000 ml of water. Following the final oocyte harvest, the frogs were humanely killed. Follicle membranes from isolated oocytes were enzymatically digested with 2 mg ml⁻¹ collagenase (Type 1A, Sigma).

Barium currents (I_Ba) were studied 2–14 days after microinjection of wild-type Ca_v3.1 and IIIS6-mutants cRNAs using the two-microelectrode voltage-clamp technique. The bath solution contained (mm): 40 Ba(OH)₂, 50 mm NaOH, 5 Hepes, 2 CsOH, adjusted to pH 7.4 with methanesulfonic acid as previously described (Berjukow et al. 2000). Voltage-recording and current-injecting microelectrodes were filled with 2.8 m CsCl, 0.2 m CsOH, 10 mm EGTA and 10 mm Hepes (pH 7.4), and had resistances of 0.3–2 MΩ. If not otherwise stated, the holding potential was set to −80 mV.

The voltage of half-maximal inactivation (V_0.5,inact) under quasi-steady-state conditions was measured using a multi-step protocol: a control test pulse (30 ms to the peak potential of the I–V curve) was followed by a 1 s step to −100 mV, a 3 s conditioning step, a 1 ms step to −100 mV and a subsequent test pulse to the peak potential of the I–V curve. Inactivation during the 3 s conditioning pulse was calculated as:

The pulse sequence was applied every 1 min from a holding potential of −100 mV and the estimated inactivation curves fitted to a Boltzmann equation:

where V is the membrane potential, V_0.5,inact is the midpoint voltage of the inactivation curve, k is a slope factor, and I_ss represents the fraction of non-inactivated current (I_ss was close to zero).

The voltage of half-maximal activation (V_0.5,act) was estimated from g_peak/g_peak,max = 0.5, where g_peak = I_peak/(V - E_rev), g_peak,max is the maximum value of g_peak measured at the descending part of the I–V curve and E_rev is the reversal potential.

Recovery from inactivation was studied using a conventional double pulse protocol: after depolarising the channels from a holding potential of −100 mV for 1 s to the peak current potential of the I–V curve, 30 ms test pulses were applied at various time intervals to the same voltage. Peak I_Ba values were normalised to the peak current measured during the prepulse and the time course of I_Ba recovery from inactivation was fitted to a biexponential function:

The onset of inactivation at −50 and −40 mV was studied as the effect of variable prepulse durations (from −80 mV to −50 and −40 mV) on the peak current at −20 mV. A 1 ms gap at −80 mV was introduced between the pre and test pulses. Subsequently the time courses of channel inactivation were fitted to the monoexponential function:

At −30, −20 and −10 mV the time course of I_Ba inactivation was also fitted to a monoexponential function.

The pCLAMP software package version 8.0 (Axon Instruments, Inc.) was used for data acquisition and preliminary analysis. Microcal Origin 5.0 was employed for analysis and curve fitting. Data are given as means ± s.e.m. Statistical significance was calculated according to Student's unpaired t test (P < 0.05).

RESULTS

Amino acids in segment IIIS6 of Ca_V3.1 affect the time constant of channel inactivation

Channels of the Ca_V3 family display the fastest current inactivation kinetics among all voltage-gated calcium channels (Perez-Reyes et al. 1998, 1999). To elucidate the role of IIIS6 amino acids in inactivation of Ca_v3.1 we mutated those residues that are identical in all voltage-gated Ca²⁺ channels to alanine (Fig. 1). Some of these residues have previously been shown to modulate the inactivation properties of Ca_v2.1 (Hering et al. 1997; Kraus et al. 1998; Sokolov et al. 1999). Six out of the seven mutations (F1505A, F1506A, F1511A, V1512A, FV1511/1512AA and F1519A, but not N1509A) resulted in functional channels when the cRNAs were injected into Xenopus oocytes. In a previous study on Ca_v2.1 and Ca_v2.1/1.1 chimeric channel constructs, alanine substitutions of the adjacent I1612 in the Ca_v2.1 sequence (or Ca_v1.1, see Fig. 1) significantly slowed current inactivation (Hering et al. 1997). We have, therefore, mutated the corresponding M1510 in Ca_v3.1 to either alanine or isoleucine (Fig. 1).

Figure 1. Segment IIIS6 of voltage-gated Ca²⁺ channel α₁ subunits.

A, putative transmembrane topology of a voltage-gated Ca²⁺ channel α₁ subunit. Each of the four domains consists of six transmembrane segments (segment IIIS6 is shown in black). B, sequence alignment of IIIS6 segments of various voltage-gated Ca²⁺ channel subtypes. Conserved amino acids are framed. The numbers are given for Ca_v3.1 sequence according to Perez-Reyes et al. (1998); accession number AF027984.

The inactivation properties of I_Ba through the mutant channels were investigated using the two-microelectrode voltage clamp (see Methods). Figure 2 illustrates the effects of IIIS6 mutations on the Ca_v3.1 kinetics. Representative families of I_Ba are shown on the left panels (A–D). The most prominent slowing in I_Ba inactivation was observed in mutants M1510I and F1511A; almost no change in current inactivation was observed for V1512A. The time constants of I_Ba decay of all mutants at −20 mV are summarised in Fig. 2E.

Figure 2. Voltage dependence of I_Ba through wild-type Ca_v3.1 and mutant channels.

A–D, left panels, families of the representative I_Ba through wild-type Ca_v3.1 (A) and selected Ca_v3.1 mutant channels (B–D). I_Ba were evoked by depolarising pulses from a holding potential of −80 mV to the indicated test potentials. Middle panels, normalised current-voltage relationships of the peak I_Ba through wild-type (A, •) and the mutant channels (B–D, ○). Right panels, voltage dependence of I_Ba steady-state inactivation of wild-type (A) and mutant channels (B–D). Data points were fitted to a Boltzmann function (see Methods). Dashed lines in B–D illustrate the position of the fitted activation and inactivation curves of wild-type Ca_v3.1. E, mean time constants of channel inactivation at −20 mV. F, the mean values of the corresponding half-maximal activation and inactivation potentials (V_0.5,act and V_0.5,inact). *P < 0.05 vs. wild-type.

The voltages of half-maximal inactivation ranged between −54.7 ± 1.4 mV in wild-type and −49.5 ± 1.5 mV in mutant F1511A (Fig. 2A and C, right panels). Thus, only F1511A displayed a statistically significant shift of the steady-state inactivation curve to more depolarised voltages (P < 0.05). The voltage dependence of channel activation was significantly shifted in mutants F1511A (-37.2 ± 1.2 mV, Fig. 2C, middle panel), V1512A (-26.1 ± 1.5 mV, Fig. 2D) and F1519A (-25.9 ± 1.1 mV compared to wild-type 32.1 ± 0.9 mV, Fig. 2A). The V_0.5,act and V_0.5,inact values are summarised in Fig. 2F.

The IIIS6 mutations produced different effects on the voltage dependence of inactivation. The time courses of channel inactivation at −50 mV and the corresponding voltage dependencies of inactivation are illustrated on the left and right panels of Fig. 3A–E. Prominent slowing of the I_Ba decay at −20 mV was observed for mutants M1510I (4.8-fold) and F1511A (3.5-fold, see Fig. 2). Mutants V1512A did not significantly affect the inactivation time course at −20 mV (Fig. 2D) but significantly slowed inactivation at −50 mV (2.4-fold, Fig. 3E and F). The different levels of steady-state inactivation at −50 mV produced by the various mutations are illustrated in Fig. 3G.

Figure 3. Mutations of amino acids in segment IIIS6 alter kinetics of channel inactivation.

A–E, left panels, mean time courses of channel inactivation at −50 mV. Smooth curves depict monoexponential functions fitted to the mean inactivation time courses. The steady-state levels of inactivation are shown as dotted lines. Dashed lines in B–E illustrate the monoexponential function fitted to the mean time course of inactivation in the wild-type Ca_v3.1. Inset in A illustrates the voltage protocol. A–E, right panels, voltage dependencies of the inactivation time constants. The dashed lines in B–E correspond to wild-type. Insets in right panels, scaled superimposed families of I_Ba (at −20 mV) through wild-type and selected Ca_v3.1 mutant channels. F, mean time constants of channel inactivation at −50 mV. G, steady-state levels of inactivation at −50 mV. The dotted line corresponds to the steady-state inactivation of wild-type. *P < 0.05 vs. wild-type.

In analogy to the Ca_v2.1 double mutant IF1612/1613AA (see Sokolov et al. 2000) we created a corresponding Ca_v3.1 construct, MF1510/1511AA. At −20 mV MF1510/ 1511AA inactivated significantly more slowly than wild-type (3.2-fold, Fig. 2E). The double mutation FV1511/1512AA also caused a significantly slower time course of inactivation at −20 mV (2.7-fold slowing). At −50 mV the inactivation of MF1510/1511AA was accelerated, whereas inactivation of FV1511/1512AA was not different from wild-type (Fig. 3F). In other words, dominant effects of a single point mutation on the inactivation kinetics of a double mutant occurred only in a certain voltage range (Fig. 3A–E, right panels).

IIIS6 mutations modulate the stability of channel inactivation

In order to elucidate the impact of the IIIS6 mutations on the stability of the inactivated channel conformations, we analysed the I_Ba recovery at −80 mV. The time course of channel recovery from inactivation was described by a double exponential function (Fig. 4A). Recovery of wild-type channels occurred with a fast time constant (τ_fast) of 0.038 ± 0.003 s and a slow time constant (τ_slow) of 0.27 ± 0.03 s. The most pronounced acceleration of repriming was observed for mutant F1511A (τ_fast = 0.018 ± 0.002 s, τ_slow = 0.097 ± 0.005 s). As evident from the decelerated recovery, mutation V1512A stabilised the inactivated channel states (τ_fast = 0.140 ± 0.011 s, τ_slow = 0.60 ± 0.05 s, Fig. 4).

Figure 4. Mutations in the segment IIIS6 of Ca_v3.1 affect recovery from inactivation.

A, mean time courses of I_Ba recovery from inactivation at −80 mV after a 1 s conditioning prepulse to −20 mV in wild-type (•), mutant F1511A (♦) and V1512A (▪) channels. B and C, mean time constants of I_Ba recovery (τ_fast and τ_slow) at −80 mV. *P < 0.05 vs. wild-type.

The double mutant FV1511/1512AA recovered from inactivation with comparable kinetics to V1512A (see Fig. 4B and C).

DISCUSSION

Structural determinants of high threshold Ca²⁺ channel inactivation are localised in transmembrane S4, S5 and S6 segments, pore loops, intracellular loops (particularly in intracellular domain linkers), and the carboxyl terminus (Zhang et al. 1994; Herlitze et al. 1997; Soldatov et al. 1998; Berjukow et al. 1999; Bourinet et al. 1999; Spaetgens & Zamponi, 1999; Sokolov et al. 2000; Stotz et al. 2000; see Hering et al. 1998, 2000).

The structural determinants of low threshold Ca²⁺ channels are much less characterised. Staes et al. (2001) identified a negatively charged region of 23 amino acids at the carboxy terminus of the Ca_v3.1 as a crucial determinant of fast inactivation and suggested a model where this sequence stretch serves as an intracellular acceptor for a yet to be identified ball segment hypothetically located at an intracellular channel part.

In the present study we focused on the role of segment IIIS6 in Ca_v3.1 inactivation. By substituting six amino acids that are identical in the IIIS6 segments of all Ca²⁺ channel types and an additional M1510, we identified a hot spot of inactivation determinants that has previously been shown to play an essential role in inactivation of the high threshold Ca_v2.1 (Fig. 1).

Hot spot of inactivation determinants in segment IIIS6

Alanine substitutions of IIIS6 residues revealed that three consecutive amino acids play an important role in Ca_v3.1 inactivation gating. At −20 mV significant slowing of the I_Ba decay was observed for four single amino acid substitutions (M1510I > F1511A > M1510A > F1519A). M1510I and F1511A induced the most pronounced slowing at −20 mV, suggesting that these residues play a critical role in Ca_v3.1 inactivation. At −50 mV the most prominent slowing of channel inactivation was caused by mutation V1512A, while F1511A induced an acceleration (Fig. 2 and Fig. 3).

Recovery of wild-type channels from inactivation was better fitted by two than by one exponential(s) suggesting the existence of at least two (fast and slow) inactivated channel conformations. Point mutations in segment IIIS6 affected both time constants. Recovery from the fast inactivated state at −80 mV was most significantly slowed by V1512A and accelerated by F1511A (Fig. 4). In general, a slowing or acceleration of the fast component in recovery was accompanied by a slowing or acceleration of the slow component (Fig. 4B and C). These data suggest that amino acid substitutions in segment IIIS6 affect the stability of the fast inactivated state as well as the recovery from a second (presumably slow) inactivated channel conformation. Further studies will characterise the mutational effects on the two inactivated states in more detail.

In previous studies on Ca_v2.1 and chimeric Ca_V2.1/1.1 channel constructs (Hering et al. 1997; Sokolov et al. 2000) we observed a pronounced slowing of channel inactivation if I1612, F1613 and V1614 in segment IIIS6 were substituted by alanine (IFV-motif, Hering et al. 1998). I1612, F1613 and V1614 in Ca_v2.1 correspond to M1510, F1511 and V1512 in Ca_V3.1 (Fig. 1B). In order to analyse if a common structural motif in segment IIIS6 determines inactivation in high and low threshold Ca²⁺ channels, we have substituted for M1510 in Ca_v3.1 the Ca_V2.1 isoleucine and additionally alanine. Both mutations affected the time course of Ca_v3.1 inactivation (Fig. 2 and Fig. 3). These data suggest that, in analogy to the IFV-motif in Ca_v2.1, M1510, F1511 and V1512 form a ‘hot spot’ of inactivation determinants in segment IIIS6 of Ca_V3.1. Interestingly, in sodium channels F1468A (corresponding to F1511 in IIIS6 of Ca_V3.1) also modulates inactivation (Yarov-Yarovoy et al. 2001) suggesting common inactivation determinants in sodium and Ca²⁺ channels.

Implications for the mechanism of Ca_v3.1 inactivation

The double mutation MF1510/1511AA slowed Ca_v3.1 inactivation like the corresponding IF1612/1613AA in Ca_v2.1 (Sokolov et al. 2000). Interestingly, mutations of neighbouring amino acids on segment IIIS6 (F1511, V1512) lead - at a given membrane voltage - to different and in some cases even opposite effects on the channel kinetics. At −20 mV mutant F1511A inactivated slower than wild-type while V1512A did not significantly change state transitions towards inactivation (Fig. 2C and D). Recovery was substantially slowed by V1512A and accelerated by F1511A (Fig. 4).

At −20 mV the kinetic phenotype of the single point mutant F1511A was dominant for FV1511/1512AA. At −50 mV inactivation occurred at an intermediate rate, suggesting an impact of both amino acid substitutions. Recovery of the double mutant FV1511/1512AA occurred with similar kinetics to wild-type while single point mutants either significantly accelerated or slowed channel repriming. Comparable observations were made for M1510A, F1511A and the corresponding double mutation MF1510/1511AA; at −20 mV, inactivation of MF1510/1511AA occurred with a similar time course to F1511A, while channel recovery at −80 mV was apparently determined by M1510A. Thus, our data suggest that point mutations in segment IIIS6 affect inactivation gating predominantly in a certain voltage range. Some of our mutations also affected channel activation. Thus, mutants F1511A and FV1511/1512AA activate apparently with slower kinetics than wild-type (Fig. 1, time to peak of FV1511/1512AA is 15.5 ± 0.7 ms and of F1511A is 16.0 ± 0.4 ms vs. 9.5 ± 0.3 ms in wild-type Ca_v3.1 at −20 mV, P < 0.05, see also changes in the voltage dependence of steady-state activation in Fig. 2F).

It is tempting to speculate that spontaneous mutations in S6 and adjacent segments played an important role in fine tuning of Ca²⁺ entry through different Ca²⁺ channel subtypes during evolution. The recent identification of point mutations in this region causing a number of channelopathies that are associated with disturbances in inactivation gating support this hypothesis (see Jen, 1999). Future studies of the impact of Ca²⁺ channel subtype-specific amino acids in other S6 segments including also residues that are specific for Ca_v3 (see Fig. 1) will help to understand the molecular basis of the different inactivation phenotypes.

In summary, we report for the first time a significant role of a S6 segment in Ca_v3.1 inactivation. We have identified a ‘hot spot’ of inactivation determinants (M1510, F1511, V1512) in segment IIIS6. Our data suggest common elements in the architecture of the inactivation machineries in high and low threshold Ca²⁺ channels. We hypothesize that these residues play a critical role in helix packing within this putative bundle crossing region of the corresponding α₁ subunits (Doyle et al. 1998) thereby modulating the closure of the channel pore during depolarisation (Perozo et al. 1998; Yellen, 1999). Alternatively, structural changes in S6 segments may directly or indirectly affect a pore receptor for a gate formed by the carboxy terminus (Staes et al. 2001).