Abstract
Studies of Ca channels expressed in oocytes have identified kurtoxin as a promising tool for functional and structural studies of low-threshold T-type Ca channels. This peptide, isolated from the venomous scorpion Parabuthus transvaalicus, inhibits low-threshold α1G and α1H Ca channels expressed in oocytes with relatively high potency and high selectivity. Here we report its effects on Ca channel currents, carried by 5 mmBa2+ ions, in rat central and peripheral neurons. In thalamic neurons 500 nm kurtoxin inhibited T-type Ca channel currents almost completely (90.2 ± 2.5% at −85 mV;n = 6). Its selectivity, however, was less than expected because it also reduced the composite high-threshold Ca channel current recorded in these cells (46.1 ± 6.9% at −30 mV;n = 6). In sympathetic and thalamic neurons, 250–500 nm kurtoxin partially inhibited N-type and L-type Ca channel currents, respectively. It similarly reduced the high-threshold Ca channel current that remains after a blockade of P-type, N-type, and L-type Ca channels in thalamic neurons. In contrast, kurtoxin facilitated steady-state P-type Ba currents in Purkinje neurons (by 34.9 ± 3.7%; n = 10). In all cases the kurtoxin effect was voltage-dependent and entailed a modification of channel gating. Exposure to kurtoxin slowed current activation kinetics, although its effects on deactivation varied with the channel types. Kurtoxin thus appears as a unique gating-modifier that interacts with different Ca channel types with high affinity. This unusual property and the complex gating modifications it induces may facilitate future studies of gating in voltage-dependent ion channels.
Peptide toxins derived from animal venoms have become valuable tools for studies of voltage-gated ion channels. Although the origin of these proteins may be diverse, their modes of action fall within two major categories. Pore-blocking toxins bind to the external vestibule of the channel pore and physically obstruct the movement of ions (MacKinnon and Miller, 1988). Gating modifiers bind close to the channel voltage sensor and alter the energetics of voltage-dependent gating (Cahalan, 1975). In the case of voltage-gated Ca channels, toxins that display high selectivity have been particularly useful. The pore blocker ω-conotoxin-GVIA (ω-CgTX) (Ellinor et al., 1994; Stocker et al., 1997), which inhibits N-type Ca channels (McCleskey et al., 1987) selectively (Aosaki and Kasai, 1989; Jones and Marks, 1989; Plummer et al., 1989), and the gating modifier ω-Aga-IVA (Mintz et al., 1992a), which blocks P-type Ca channels with high affinity (Mintz et al., 1992b), have helped to characterize the gating behavior of these channels (Plummer et al., 1989; Usowicz et al., 1992; Tottene et al., 1996; Stocker et al., 1997), identify their structure (Williams et al., 1992; Sather et al., 1993; Berrow et al., 1997), and study their roles in controlling neuronal excitability (Llinás et al., 1989; Gorelova and Reiner, 1996; Magee and Carruth, 1999) and transmitter release (Pfrieger et al., 1992; Takahashi and Momiyama, 1993). These two toxins and others with less restricted selectivity (Hillyard et al., 1992; Lampe et al., 1993; McDonough et al., 1996) also have been instrumental in defining important motifs in the voltage-sensing domain of high-threshold Ca channels (Li-Smerin and Swartz, 1998) or in their pore region (Ellinor et al., 1994). As a result, a wealth of structural, biophysical, and functional data is available for high-threshold Ca channels.
In contrast, information on low-threshold Ca channels is still limited. T-type Ca channels play important roles that can be inferred from their voltage dependence and their characteristic activation, deactivation, and inactivation properties (Carbone and Lux, 1984). They may boost synaptic potentials (Hirsch et al., 1985), mediate Ca influx during action potentials (McCobb and Beam, 1991; Scroggs and Fox, 1992), and generate low-threshold Ca spikes (LTS) (Jahnsen and Llinás, 1984;Crunelli et al., 1989) and pacemaker activities (Destexhe et al., 1998), yet the lack of selective antagonist has limited the study of these roles (Huguenard, 1996).
After the recent cloning of T-type Ca channels (Lambert et al., 1998;Perez-Reyes et al., 1998; Lee et al., 1999), the screening of new inhibitors has led to the identification of a promising antagonist, kurtoxin (Chuang et al., 1998). This peptide acts with high selectivity on T-type Ca channels expressed in Xenopus oocytes. It inhibits low-threshold α1G and α1H Ca channels potently but does not affect high-threshold α1A, α1B, α1C, and α1E Ca channels. Its selectivity, however, is not absolute, because kurtoxin also affects the inactivation of voltage-gated Na channels in oocytes (Chuang et al., 1998). Its effects on native ion channels have not yet been studied.
To characterize kurtoxin selectivity on neuronal Ca channels, we studied its effects on a variety of identified Ca channel currents in rat central and peripheral neurons. We found that these effects differed remarkably from those reported in oocytes. In neurons the kurtoxin interacted with high affinity with T-type, L-type, N-type, and P-type Ca channels, producing complex gating modifications that were specific to each channel type. This unique property will make it an asset for structural studies of gating in voltage-gated ion channels.
MATERIALS AND METHODS
Enzymatic dissociation of thalamic, sympathetic, and Purkinje neurons. Thalamic neurons from the ventral posteromedial (VPM) nucleus (Paxinos and Watson, 1986) were freshly dissociated with enzyme (Mintz et al., 1992b). Briefly, 400-μm-thick coronal slices were cut in ice-cold Ringer's solution from brains of 10- to 13-d-old rats. The VPM nucleus was dissected out under 400× magnification. Then it was incubated for 8 min in a solution and maintained at 37°C under constant stirring, which contained (in mm): 81.4 Na2SO4, 30 K2SO2, 5.8 MgCl2, 10 Na-HEPES, 20.4 glucose, 0.5% phenol red, 3 mg/ml protease type XXIII (Sigma, St. Louis, MO), pH 7.4, with NaOH. After the incubation with enzyme, the brain tissue was rinsed three times in a MEM solution (reference 11090-073, Life Technologies, Grand Island, NY) supplemented with 20 mmglucose and 10 mm Na-HEPES, pH 7.4, with NaOH (at 37°C). Then cells were released by gentle trituration through a fire-polished Pasteur pipette into the MEM solution containing 20 mmglucose, 10 mm Na-HEPES, 1 mg/ml trypsin inhibitor, and 1 mg/ml bovine serum albumin (Fraction V, Sigma), pH 7.4, with NaOH. Cells were kept at 14–16°C and remained viable for 5–6 hr after preparation.
Purkinje neurons were dissociated from the cerebellar vermis of 9- to 11-d-old rats (Mintz et al., 1992b). The cerebellum was dissected out in ice-cold Ringer's solution and cut into three to four pieces, which were incubated with enzyme by following the protocol described above. After dissociation the Purkinje neurons were identified morphologically by their large cell bodies (15–25 μm in diameter) with single dendritic remnants.
Sympathetic neurons were prepared from 9- to 14-d-old rats (Boland et al., 1994; McDonough et al., 1997a). The superior cervical ganglia were dissected out in ice-cold oxygenated Leibovitz's L-15 medium (reference 11415-064, Life Technologies). Each ganglion was cut into two pieces before being incubated for 20 min at 37°C in a calcium-free Tyrode's solution that contained (in mm): 150 NaCl, 4 KCl, 2 MgCl2, 10 glucose, 10 Na-HEPES, 0.5 EDTA, and 2 l-cysteine, pH 7.4, with NaOH, plus 25 U/ml papain (Worthington Biochemicals, Lakewood, NJ). After this incubation the ganglia were transferred into calcium-free Tyrode's solution containing 1 mg/ml collagenase (type I) and 8 mg/ml dispase (Boehringer Mannheim, Indianapolis, IN). This incubation was performed at 37°C for 40 min. The ganglia were rinsed three times, and the cells were released by gentle trituration into the MEM solution supplemented with 20 mm glucose, 10 mm Na-HEPES, 1 mg/ml trypsin inhibitor, and 1 mg/ml bovine serum albumin, pH 7.4, with NaOH (at 37°C).
Voltage-clamp recording. Patch-clamp recordings of Ca channel currents were performed in the whole-cell configuration (Hamill et al., 1981) by using 5 mmBa2+ ions as charge carrier. Patch pipettes were pulled from borosilicate glass capillaries (Fisher Scientific, Pittsburgh, PA), coated with Sylgard (Dow Corning, Midland, MI), and fire-polished to achieve a resistance of 1.5–2 MΩ for recordings of Purkinje neurons and 3–3.5 MΩ for recordings of thalamic and sympathetic neurons. In all of the recordings used in this report, the capacitance and access series resistance were compensated to minimize voltage errors to <5 mV. From 3 to 10 GΩ seals were obtained routinely. We waited at least 5 min to allow the recording currents to stabilize. Ba currents usually ran down by ∼5–10% within the first 2 min of their recording. They stabilized within 5 min, after which time rundown was negligible and the experiment was started. Current recordings are presented without leak correction. They were obtained with an Axopatch 200A patch-clamp amplifier (Axon Instruments, Foster City, CA). Voltage-step commands and data acquisition were controlled by using the external operation (XOP) Pulse Control (Herrington and Bookman, 1994) in IGOR Pro (WaveMetrics, Lake Oswego, OR) and the ITC-16 analog-to-digital converter from Instrutech (Port Washington, NY). In most experiments the currents were digitized at 100 μsec and filtered at 2 kHz. For experiments focusing on tail currents, recordings were digitized at 20 μsec and filtered at 10 kHz.
For data analysis and illustrations, we used IGOR Pro.
Statistics are given as mean ± SEM.
Solutions. The solution in the recording pipette contained (in mm): 108 cesium methanesulfonate, 4 MgCl2, 9 EGTA, 10 HEPES, 4 Mg-ATP, 0.3 GTP-Tris, 14 creatine phosphate (Tris salt), pH 7.3, with CsOH. The extracellular solution contained (in mm): 160 tetraethylammonium (TEA)-chloride, 5 BaCl2, 10 HEPES, 0.1 EGTA plus 1 μm tetrodotoxin and 1 mg/ml cytochrome c, pH 7.4, with TEA-OH.
Stock solutions of 5 μm kurtoxin (generously provided by Dr. K. J. Swartz, National Institute of Neurological Diseases and Stroke, National Institutes of Health, Bethesda, MD), 100 μm ω-Aga-IVA (Peptide Institute, Osaka, Japan), 0.5 mm ω-conotoxin-GVIA (Peninsula Laboratories, San Carlos, CA), 100 μm nimodipine, and 100 μm Bay K 8644 (Research Biochemicals, Natick, MA) were prepared in the extracellular recording solution, aliquoted, and stored at −80°C. All other chemicals were purchased from Sigma.
The pipette tips, the recording chamber, and the vials that contain toxin stocks were all siliconized to minimize toxin loss via nonspecific binding.
Drug application and determination of kurtoxin potency. For most experiments the cells were recorded in a minichamber (100 μl) to which small volumes of concentrated (10×) toxin solution could be applied directly with a micropipette (Mintz et al., 1992b). With this protocol the recorded cells can be exposed to high concentrations of kurtoxin without using large amounts of toxin. Experiments investigating the kurtoxin inhibition of Bay K 8644-enhanced L-type Ba currents (see Fig. 7) and the Ba current that was resistant to blockers of P-type, L-type, and N-type Ca channels (see Fig. 8) were all performed in the minichamber (including studies of reversibility) to minimize the use of ω-Aga-IVA and ω-CgTX.
Although kurtoxin application to the minichamber allowed for reliable estimates of steady-state effects, it did not describe the kinetics of kurtoxin association accurately, because the toxin application was neither instantaneous nor homogenous. In most instances, current inhibition could not be fit with a monoexponential function. A number of experiments thus were performed by using a conventional array of gravity-fed glass microcapillaries (100 μm in diameter) that were connected with Teflon tubing to plastic syringes. Complete exchange of solution occurred within 1–2 sec, after the lateral displacement of the recorded cell from the opening of one capillary to the next. This application system was used to study the reversibility of kurtoxin effect on T-type (see Fig. 3A), P-type (see Figs.4A2, 5), and N-type (see Fig.6A1,A2) Ca channels. This application system also has limitations. It is difficult to ensure complete saturation of kurtoxin nonspecific binding sites to the glass capillaries. As for ω-Aga-IVA, a prolonged perfusion (of ∼10 min) of the glass capillary with the toxin-containing solution is necessary to saturate these binding sites. Because purified kurtoxin is not available in large quantities, most experiments were performed after a 2–4 min “loading” perfusion of the kurtoxin-containing solution. As a result, the time constants for current inhibition were overestimated, and it is possible that the current reduction measured in some experiments (see Figs.4A2, 6A2) had not reached steady state.
These limitations account for the difference in kurtoxin potency when determined from the steady-state effects of kurtoxin (measured in the minichamber) and from the on-rate and off-rate time constants (measured by using microcapillaries).
All experiments were done at room temperature (20–22°C).
RESULTS
Kurtoxin inhibition of T-type and high-threshold Ba currents
We investigated the effects of kurtoxin on identified T-type Ca channel currents recorded in thalamic neurons that were freshly isolated from the VPM nucleus. These bursting neurons display robust low-voltage-activated T-type calcium currents in addition to a composite high-threshold Ca channel current (Huguenard, 1996). In these cells test depolarizations elicited maximum T-type Ca channel currents when applied from a −120 mV holding potential (Kuo and Yang, 2001). At less negative holding potentials, measurable T-type current inactivation occurred. Nevertheless, because most cells constantly held at −120 mV became unstable over the 1–2 hr length of our recordings, we used a −80 mV holding potential and conditioned all test depolarizations with a 2-sec-long prepulse to −120 mV. The amplitude and duration of this prepulse were sufficient to promote the complete recovery of inactivated T-type Ca channels. Longer or more negative prepulses had no additional effects on the amplitude of T-type Ca channel currents in these cells.
Consistent with its inhibition of low-threshold Ca channels expressed in oocytes (Chuang et al., 1998), 500 nm kurtoxin nearly abolished T-type Ba currents in thalamic neurons. Figure1A illustrates the low-threshold (left) and high-threshold (right) Ca channel currents carried by 5 mm Ba in such a cell. The transient Ba current elicited by a 200 msec depolarization to −75 mV was reduced by ∼90% after exposure to 500 nm kurtoxin. Current reduction was maximal because an increase in kurtoxin concentration to 1 μm had no additional effect (Fig.1A, left). The dose–response relationship (Fig. 1C) suggests a dissociation constant (KD) for kurtoxin inhibition of thalamic T-type channels of ∼50 nm.
Surprisingly, the composite high-threshold Ca channel current evoked in thalamic neurons by stronger depolarizations (Kammermeier and Jones, 1997) also was reduced by kurtoxin. Current inhibition was incomplete, however (Fig. 1A, right), and appeared to be less potent than that seen for T-type Ca channel currents. Although toxin concentrations of 500 nm and 1 μm were equally effective on low-threshold Ba currents (as seen in the I–V relationship for test pulses below −40 mV), 1 μm kurtoxin produced greater inhibition of the high-threshold Ba currents triggered by test pulses above −30 mV (Fig. 1B).
Selectivity
The prominent effect of kurtoxin on the composite high-threshold Ca channel current of thalamic neurons suggests that this toxin may target multiple Ca channel types. We investigated this issue in rat Purkinje, sympathetic, and thalamic neurons by using pharmacological conditions that isolate large P-type, N-type, and L-type Ca channel currents, respectively.
Figure 2A(left) illustrates the effects of kurtoxin on a Purkinje neuron P-type Ba current, which was recorded after a blockade of N-type and L-type Ca channels with 2.5 μmω-conotoxin-GVIA (ω-CgTX) and 2.5 μmnimodipine (Mintz et al., 1992b). Kurtoxin had little effect on the peak current elicited by a 100 msec test pulse to −20 mV, but it slowed the kinetics of current activation remarkably. In a total of nine Purkinje neurons similarly exposed to 500 nmkurtoxin, the current that was measured 10 msec after the onset of the test depolarization (applied from −80 to −20 mV) amounted to 49.2 ± 3.3% of the control current, whereas 90 msec later the current amplitude had returned to its control level (94.8 ± 1.9%; n = 9). Because kurtoxin produced minimal changes in steady-state current, we estimated the dose dependence of its effect by measuring the percentage of increase in the activation time constant of the Ba current elicited at the peak of theI–V relationship (% Δτ/τ). The resulting dose–response curve (Fig. 2A, right) indicates a high-affinity interaction (estimatedKD of ∼15 nm) between kurtoxin and Purkinje neuron P-type Ca channels.
In Figure 2B (left), 500 nm kurtoxin partially inhibited a sympathetic neuron N-type Ca channel current, which was recorded in isolation after a block of L-type Ca channels by 2.5 μmnimodipine (Mintz et al., 1992b; Boland et al., 1994). Like Purkinje neuron P-type Ba currents, N-type Ba currents displayed slower activation kinetics in the presence of kurtoxin. Still, at steady state, current reduction remained significant, and it amounted to 46.9 ± 1.2% of the control current (at −10 mV;n = 10). Measurements of the peak current reduction for different concentrations of kurtoxin are represented in the dose–response curve of Figure 2B (right). Kurtoxin inhibition of N-type Ca channels appeared to be notably weaker (estimated KD of ≈450 nm) than its effects on T-type and P-type Ca channels.
We assessed the sensitivity of L-type Ca channels to 500 nmkurtoxin in thalamic neurons (Fig. 2C). To maximize the contribution of L-type Ca channels to the recorded currents, we performed experiments after a block of N-type and P-type Ba currents with 2.5 μm ω-CgTX and 200 nm ω-Aga-IVA and after enhancement of L-type Ca channel currents with the dihydropyridine agonist Bay K 8644 (3 μm) (Hess et al., 1984; Nowycky et al., 1985). In the continuous presence of ω-CgTX and ω-Aga-IVA the application of Bay K 8644 resulted in a 10- to 15-fold facilitation of the small Ba current elicited by depolarizations to −50 mV, indicating that this current was carried primarily through L-type Ca channels with little contamination from other high-threshold Ca channels (Fig.2C, left). Bay K 8644 also induced a pure L-type Ca channel tail current, which deactivated with slow kinetics, consistent with a prolongation of the open time of these channels (Nowycky et al., 1985). On average, 500 nmkurtoxin inhibited Bay K 8644-enhanced L-type Ca channel currents by 74.1 ± 3.7% (current reduction measured at −50 mV;n = 6). The dose–response curve illustrated in Figure2C (right), which was obtained in the same recording conditions, shows that kurtoxin inhibition of thalamic L-type Ba currents is potent, with an estimatedKD of ∼70 nm.
These results demonstrate that kurtoxin affected each Ca channel type (T-type, L-type, N-type, and P-type) that can be identified pharmacologically in mammalian neurons. Surprisingly, P-type and N-type channels, which are most alike structurally, showed dramatically different responses to kurtoxin. These differences motivated a detailed investigation of kurtoxin actions on T-type, P-type, N-type, and L-type Ca channel currents.
Kurtoxin inhibition of T-type Ba currents
Reversibility
In thalamic neurons the kurtoxin inhibition of T-type Ca channel currents was fully reversible (Fig.3A, left;n = 6). In the experiment of Figure 3A, a saturating concentration of 500 nm kurtoxin inhibited T-type Ca channel current recorded at −60 mV by ∼77% within 60 sec of its application. With toxin removal the current fully recovered to its control value within 100 sec. Both current inhibition and current recovery followed monoexponential time courses (τon = 6.7 and τoff = 13 sec for 500 nm kurtoxin).
Gating modifications
Within the range of membrane potentials (−85 to −60 mV), in which T-type Ca channels activated with little contamination from high-threshold Ca channels, the magnitude of T-type current inhibition by kurtoxin appeared to be voltage-dependent. Although kurtoxin abolished the Ba current elicited at −85 mV, the magnitude of this effect decreased when currents were triggered by larger test depolarizations (Fig. 3B). For example, T-type Ba currents measured at −60 mV were inhibited by only 55.3 ± 2.6% (n = 11).
At all test voltages kurtoxin also slowed the current activation and inactivation. This is illustrated in Figure 3C by superimposed, normalized T-type Ba currents recorded before and after the addition of 500 nm kurtoxin. Kurtoxin increased the current activation and inactivation time constants (measured at −60 mV) by similar factors (of 2.6 and 2.1, respectively;n = 10). Kurtoxin also increased the rate of current deactivation (Fig. 3D). This was seen best when tail currents were recorded at −110 mV to minimize current inactivation (Kuo and Yang, 2001). Normalized tail currents recorded before and after the addition of kurtoxin showed a clear increase in their rate of deactivation (n = 4). Because they were poorly fit with single exponentials, we used the later phase of deactivation to quantify this effect (τcont = 3.6 and τkurt = 2.8 msec; Fig. 3D).
Kurtoxin facilitation of P-type Ba currents
Figure 4 illustrates kurtoxin effects on P-type Ba currents in Purkinje neurons. To suppress the small current fractions that flow through N-type and L-type Ca channels in these cells, we recorded currents in the continuous presence of 2.5 μm each ω-CgTX and nimodipine. As already seen in Figure 2A, 500 nm kurtoxin altered the activation kinetics of P-type Ca channel significantly. In Purkinje neurons that were recorded in these conditions (n = 9), current activation (measured at −15 mV) was well described by a single exponential in control (τ = 1.2 ± 0.1 msec) and after exposure to kurtoxin (τ = 4.8 ± 0.5 msec).
Reversibility
To monitor the time course of kurtoxin action (Fig.4A2), we chose to measure current amplitude 5 msec after the onset of the test pulse to −15 mV, because kurtoxin had little effect on the steady-state current amplitude (Fig.4A1). The effect of 500 nmkurtoxin on P-type Ca channel currents was relatively fast and readily reversible (n = 6). In the experiment of Figure4A2 the time course of current reduction was well described by a single exponential function (τonof ≈12 sec). Removal of the toxin led to rapid and complete current recovery, which followed a monoexponential time course (τoff of ≈51 sec).
In many cells the current recovered after toxin removal was bigger than in control conditions (Fig. 4A1, trace 3). Such current run-up is not specific to kurtoxin. Often it is seen when P-type Ba currents in Purkinje neurons recover from a block by ω-Aga-IVA (Mintz et al., 1992b) or Cd ions (I. M. Mintz, unpublished results).
Current facilitation
Kurtoxin facilitated the P-type Ba currents measured at steady state during large test depolarizations (Fig.4B,C1–C3) while it reduced the currents elicited by smaller test pulses. These effects are illustrated by theI–V relationships presented in Figure 4C2 and the corresponding activation curves in Figure 4C3. The latter were well described by single Boltzmann distributions. They showed that kurtoxin shifted the voltage for half-maximum activation (V1/2) toward positive potentials by 2.4 ± 0.54 mV (n = 9) and did not affect the curve slope factor. This shift in V1/2may reflect an inhibition of P-type Ca channel current during small test pulses (because toxin-bound channels require larger depolarization to open) but also, to some extent, an artifact. In the presence of kurtoxin the slowly activating currents recorded at the foot of theI–V relationship may not have reached steady state within the duration of the test pulses (20 msec).
The facilitation of P-type Ba currents during large test depolarizations was accompanied with a parallel enhancement of the tail currents seen on repolarization. In Purkinje neurons that were exposed to 500 nm kurtoxin, the tail currents were consistently larger (by 34.9 ± 3.7%; n = 10) and slower (see the normalized tail currents in the inset, Fig.4B) than in control conditions. The effect on tail current deactivation was fully reversible (Fig. 4A1). These findings suggest that kurtoxin facilitation of P-type Ca channels is mediated, at least partially, by an increase in the channel open time, confirming that kurtoxin is a gating modifier of P-type Ca channels.
It has been proposed that gating modifiers recognize a conserved binding motif in voltage-gated ion channels (Li-Smerin and Swartz, 1998), suggesting that kurtoxin and ω-Aga-IVA, two gating modifiers of P-type Ca channels, might interact at the same binding site. To test this hypothesis, we investigated how kurtoxin affects ω-Aga-IVA inhibition of P-Type Ba currents in Purkinje neurons.
Kurtoxin and ω-Aga-IVA binding sites
Because the effects of kurtoxin and ω-Aga-IVA on P-type Ca channels are fully reversible, we were able to assess, in the same cell, the inhibition of P-type Ba currents by ω-Aga-IVA when it was applied alone or in the presence of kurtoxin. In the experiment of Figure 5, the application of 500 nm kurtoxin did not affect the amplitude of the peak current that was elicited by a 20 msec test depolarization to −20 mV (Fig. 5A). The progression of its effect was signaled, however, by changes in current activation (Fig. 5B). After kurtoxin effect had reached steady state, the coapplication of 200 nm ω-Aga-IVA led to a slow (τon of ≈60 sec) and partial inhibition of P-type Ba currents (Fig. 5C). This result is very different from the complete and rapid reduction of P-type Ba currents normally produced by 200 nm ω-Aga-IVA in Purkinje neurons. Indeed, on washout of both toxins and complete current recovery (which was facilitated by the administration of 39 pulses to +140 mV), the application of 200 nmω-Aga-IVA rapidly abolished the recovered P-type Ba current (τon of ≈12 sec; Fig. 5C).
These findings are consistent with overlapping binding sites for kurtoxin and ω-Aga-IVA in P-type Ca channels.
Kurtoxin inhibition of N-type Ba currents
Figure 6 illustrates the characteristics of N-type Ca channel current inhibition by kurtoxin. For these studies the Ba currents carried through N-type Ca channels were isolated in rat sympathetic neurons after a block of L-type Ca channels with 2.5 μm nimodipine (Boland et al., 1994).
Current block and recovery exhibit complex kinetics
Kurtoxin (500 nm) inhibition of N-type Ba currents in sympathetic neurons followed complex kinetics that were well described by a dual exponential time course, with fast (τfast = 13 sec) and slow (τslow = 247 sec) components (Fig.6A2). The fast component was predominant, because ∼80% of the total current inhibition occurred within 60 sec of the toxin application.
With the removal of kurtoxin the current recovery was incomplete (n = 12). In the experiment of Figure6A2, the current elicited at −5 mV recovered to ∼80% of its control value, after a simple monoexponential time course, with a time constant τoff of ∼20 sec. Because we selected cells with very stable control currents for these experiments, the current fraction that did not recover after wash of the toxin likely reflected irreversible current inhibition by kurtoxin rather than current rundown.
Interestingly, current recovery was more complete when currents were recorded at potentials below −20 mV (n = 6). In Figure6B, the N-type Ba current elicited by a small test depolarization returned to its control level after washout of the toxin (Fig. 6B, left). In the same cell, after the same duration of wash, the currents elicited by larger depolarizations recovered only partially, by ∼70% at −15 mV (Fig.6B, middle) and by ∼60% at −10 mV (Fig. 6B, right).
Gating modifications
Kurtoxin reduction of N-type Ca channel currents in sympathetic neurons also was accompanied by a slowing of activation (Fig.6A1,B,C). On average, the activation time constant of currents recorded at −10 mV was 2.34 ± 0.07 msec in control conditions and 3.42 ± 0.07 msec in the presence of 500 nm kurtoxin (n = 10). Interestingly, there was no apparent change in deactivation. In Figure6C, tail currents were measured after a test pulse to −5 mV at different repolarizing potentials (−80, −70, and −50 mV). After normalization the tail currents recorded before and after kurtoxin application were, in each case, identical.
The magnitude of kurtoxin effect on N-type Ba currents was voltage-dependent. In the I–V relationships of Figure6D1, larger current reduction occurred at more negative test potentials. For example, kurtoxin inhibited 39.3 ± 1.8% of the peak current measured at −10 mV and only 25.4 ± 1.7% of that evoked at +10 mV (n = 10). Although activation curves recorded in control conditions were poorly fit by single Boltzmann distributions (Fig. 6D2), they consistently showed a slight shift toward positive potentials (on average by 4.0 ± 0.5 mV at V1/2;n = 10) after the addition of kurtoxin.
Kurtoxin inhibition of L-type Ba currents
We investigated kurtoxin inhibition of L-type Ca channel currents in thalamic neurons, using the same experimental conditions described in Figure 2C1 to enhance L-type Ca channel currents. The recorded cells were held at −80 mV to inactivate T-type Ca channels. They were exposed continuously to ω-CgTX (2.5 μm) and ω-Aga-IVA (100–200 nm) to block N-type and P-type Ca channels and to the dihydropyridine agonist Bay K 8644 (3 μm) to augment L-type Ca channel open probability.
In the experiment of Figure7A1, the application of 3 μm Bay K 8644 increased the Ba currents elicited by a small depolarization to −50 mV 10-fold, confirming that, at this potential, >90% of the recorded Ba current was carried through L-type Ca channels. Consistent with an increase in L-type Ca channel open time (Nowycky et al., 1985), the agonist also slowed the deactivating tail currents seen on repolarization (Fig. 7A1,inset). Kurtoxin (200–250 nm) inhibited the Bay K 8644-enhanced L-type Ba currents on average by 74.1 ± 3.7% (at −45 mV; n = 6) as well as the Bay K 8644-enhanced tail currents recorded at −70 mV.
This inhibition resembled that seen for T-type and N-type Ca channel currents in being voltage-dependent and being accompanied by a slowing of activation (Fig. 7A1, right). At −45 mV the kurtoxin almost doubled the time constant for current activation (τcontrol = 7.9 ± 1.1 and τkurtoxin = 14.0 ± 2.3 msec;n = 12). The magnitude of current inhibition was more pronounced at negative test potentials (Fig. 7A1,A2). Kurtoxin inhibited ∼80% of the Bay K 8644-enhanced Ba current recorded at −45 mV, ∼25% of those recorded at −30 mV, and only 5% of the currents recorded at −10 mV (Fig. 7A2). Like T-type Ca channels, L-type Ca channels exposed to kurtoxin displayed faster deactivation (Fig. 7A1, inset). On average, the deactivation rate measured at −70 mV increased by 49 ± 6% (n = 12).
Kurtoxin inhibition of Bay K 8644 was also fully reversible (Fig.7B).
Although our data clearly demonstrate that kurtoxin inhibition of L-type Ba current is voltage-dependent, they could not be used for precise quantification. Kurtoxin facilitated the voltage-dependent relief of P-type Ba current block by ω-Aga-IVA (data not shown), which was manifest in the presence of kurtoxin at potentials as low as 0 mV. Unblocked P-type Ca channels thus may contribute to the currents recorded during large test depolarizations. Because of the prolongation of their open time (as in Fig. 4B), unblocked kurtoxin-modified P-type Ca channels also may contaminate the late phase of the Bay K 8644-enhanced tail currents. The classical approach, which uses the late component of the Bay K 8644-enhanced tail currents as a measure of L-type Ca channel activation, was therefore not applicable.
High-threshold Ba currents resistant to antagonists of L-type, N-type, and P-type Ca channels
In thalamic neurons as in other neurons (Eliot and Johnston, 1994;Penington and Fox, 1995; Randall and Tsien, 1995; Turner et al., 1995;Hilaire et al., 1997), a significant proportion of the total Ba current remained unaffected after blockade of P-type, N-type, and L-type Ca channels. In the experiment of Figure 8, we used saturating concentrations of 2.5 μm nimodipine, 2.5 μm ω-CgTX, and 1 μm ω-Aga-IVA to maximize inhibition of L-type, N-type, and P-type Ba currents. Within 30 min this combination of antagonists reduced the peak control current elicited at −35 mV by ∼90%. The remaining Ba current showed a relatively low threshold for activation, because measurable activation already had occurred at −50 mV (Fig. 8B) plus significant inactivation during a 500 msec test pulse (Fig. 8C) and fast deactivation kinetics (Fig. 8A, inset). It was, however, clearly distinct from T-type Ca channel currents (Fig.8C, inset), which completely inactivated at holding potentials as negative as −90 mV, showed a much lower threshold for activation of approximately −80 mV (Fig.1B), and deactivated with slower kinetics (by a factor of ∼10 at −70 mV).
This remaining Ba current, isolated after a block of T-type, P-type, N-type, and L-type Ca channels, was inhibited partially by 500 nm kurtoxin (Fig. 8A). The magnitude of current inhibition showed clear voltage dependence, although less dramatic than that seen for other Ca channel types (Fig. 8B). For example, kurtoxin decreased the currents measured at −40 mV by 56.2 ± 2.4% and those measured at +10 mV by 35.0 ± 2.8% (n = 3). Like T-type, L-type, N-type, and P-type Ba currents, this current activated with slower kinetics in the presence of kurtoxin; current deactivation, however, appeared to be unaffected (Fig. 8A, inset).
DISCUSSION
This study demonstrates that kurtoxin affects a large variety of Ca channels in mammalian neurons. Consistent with its potent blockade of α1G and α1H T-type Ca channels studied in Xenopusoocytes (Chuang et al., 1998), we find that kurtoxin inhibited low-threshold T-type Ca channel currents in thalamic neurons. However, it also targeted neuronal high-threshold Ca channels, including P-type, N-type, and L-type Ca channels, and others that still are unidentified pharmacologically. To our knowledge this is the first example of a gating modifier that acts with high potency on such a variety of neuronal low-threshold and high-threshold Ca channels.
Unlike other gating modifiers of Ca channels, such as grammatoxin (McDonough et al., 1997a) and ω-Aga-IVA (Mintz et al., 1992b; Sather et al., 1993; McDonough et al., 1997b), kurtoxin produced very different sets of gating modifications in different Ca channel types. Kurtoxin inhibited T-type, L-type, and N-type Ca channels and facilitated P-type Ca channels. It accelerated deactivation in T-type and L-type Ca channels, slowed it in P-type Ca channels, and left it unaffected in N-type Ca channels.
These unexpected findings rule out the use of kurtoxin for functional studies of T-type Ca channels. Still, the variety of gating alterations it promotes is unique and may lead to new insights on how gating modifiers interact with voltage-gated Ca channels.
Determination of kurtoxin potency
We measured steady-state current reduction (for T-type, N-type, and L-type Ba currents) and increases in activation time constants (for P-type Ba currents) to evaluate the affinity of kurtoxin for each channel type (see Materials and Methods). The resulting dose–response curves were described well by hyperbolic functions, suggesting a bimolecular binding. Their fits yielded values of 50 nm(T-type), 15 nm (P-type), 460 nm (N-type), and 70 nm (L-type Ba currents) as respective dissociation constants (KD values) for kurtoxin.
Although imprecise, these numbers are informative. The potency for kurtoxin inhibition of native T-type Ca channels is comparable with that estimated for α1G Ca channels (15 nm) (Chuang et al., 1998). The kinetics, however, are quite different, because the off rate for kurtoxin binding to thalamic T-type Ca channels was ∼20 times faster than that seen in α1G Ca channels. Kurtoxin modulation of P-type, L-type, and even N-type Ca channels was surprisingly potent, considering that concentrations as high as 350 nm have no effect on α1A, α1B, α1C, and α1E Ca channels (Chuang et al., 1998). We do not know whether this discrepancy reflects differences in the primary structure of native neuronal Ca channels and Ca channel clones expressed in Xenopus oocytes, distinct post-translational modifications, or their association with different auxiliary subunits.
Association and dissociation kinetics
As described in Materials and Methods, accurate measurements of kurtoxin on rate constants were impaired in our recording conditions (because of toxin nonspecific binding to glass capillaries or its nonhomogeneous application into the minichamber). Nevertheless, the time courses depicted in the figures illustrate the reversibility of kurtoxin effects and provide reasonable estimates of the toxin off rate. Except for N-type Ca channel current, the effects of kurtoxin were fully reversible, and current recovery followed simple monoexponential time courses. These results yielded off rate time constants of ∼13 sec (T-type), 51 sec (P-type), 15 sec (L-type), and 20 sec (for N-type Ba currents).
Kurtoxin inhibition of N-type Ba currents showed complex kinetics that were well described by the sum of two exponential functions. Current recovery followed a monoexponential time course, but its magnitude was voltage-dependent. Although currents elicited by small test potentials returned to their control values after a wash of kurtoxin, currents measured at more positive potentials recovered only partially. These findings may reflect the structural heterogeneity of N-type Ca channels in sympathetic neurons. A subtype (possibly α1BΔET variants) may carry the small current component that was inhibited slowly and irreversibly by kurtoxin (Lin et al., 1997). Another (possibly α1B+ETvariants), with a lower threshold for activation, may carry the larger component that was inhibited rapidly and reversibly by kurtoxin (Lin et al., 1997). Alternatively, two kurtoxin molecules may bind to a single N-type Ca channel with a first reversible step and a second irreversible one (Hess et al., 1975).
Gating modifications
Activation and deactivation
Kurtoxin modified T-type and L-type Ca channel currents by promoting slower activation and faster deactivation and by reducing steady-state currents. These effects were voltage-dependent. They resemble ω-Aga-IVA inhibition of P-type Ca channels in Purkinje neurons and might be explained, similarly, by the stabilization of closed states (McDonough et al., 1997b). Toxin-bound channels then require larger depolarizations to gate into the open state. An expected corollary to these observations is the destabilization of kurtoxin binding to open channels. It will be interesting to see whether trains of large depolarizations increase the rate of kurtoxin unbinding as they do to ω-Aga-IVA binding to P-type Ca channels. During the brief depolarizations that have been tested so far, there was no indication of kurtoxin unbinding from T- and L-type Ca channels.
In Purkinje neurons the kurtoxin also decreased the rate of P-type current activation and reduced currents recorded at negative test potentials, shifting the activation voltage dependence toward more positive potentials. All of these effects are consistent with a stabilization of closed states. Kurtoxin, however, produced other gating modifications in P-type Ca channels, because it clearly facilitated their steady-state currents and slowed their deactivation. The prolongation of P-type Ba tail currents suggests an increase in the open time of toxin-bound channels. We do not know whether this increase fully accounts for the steady-state current facilitation observed during large test pulses because it was contaminated with current run-up. Run-up was clearly different from the gating modifications produced by kurtoxin. It occurred by simple and irreversible scaling of the current amplitude, with no change in activation and deactivation kinetics. Its presence, however, impeded the true measure of steady-state current facilitation by kurtoxin.
Kurtoxin inhibition of N-type Ca channel currents was qualitatively different. As in other channel types, kurtoxin slowed activation, but at the resolution of our recordings it had no apparent effect on channel deactivation. Current reduction at steady state was also voltage-dependent, but to a much smaller extent. Current inhibition was also much less potent. Such differences are striking considering that N-type and P-type Ca channels are likely to be highly homologous (Williams et al., 1992; Berrow et al., 1997; Bourinet et al., 1999). These results imply that a limited number of amino acid substitutions (Fig. 9) may alter completely the nature of kurtoxin interaction with these channels, leading to opposite effects on gating.
Inactivation
Kurtoxin belongs to the family of α-scorpion toxins, which slow inactivation of voltage-gated Na channels (Catterall, 1980) by inhibiting their transition from open to inactivated state without changing their activation (Wang and Strichartz, 1985). Kurtoxin, however, consistently slowed activation in parallel to its effect on inactivation. Both effects are seen for voltage-gated Na currents (Chuang et al., 1998), T-type Ba currents (Fig. 3), and the currents that remained after a blockade of P-type, L-type, and N-type Ca channels (Fig. 8). In the case of T-type Ba currents the similar magnitudes of kurtoxin effects on inactivation and activation suggest that kurtoxin may act simply by reducing the number of open channels available for inactivation (Kuo and Yang, 2001). We have not examined the effects of kurtoxin on the inactivation of high-threshold Ca channels because this process was not fully reversible, especially for P-type Ca channels (Regan, 1991), in our recording conditions.
Kurtoxin binding sites
Because kurtoxin affects both voltage-gated Na and T-type Ca channels, it has been proposed that kurtoxin recognizes a motif common to these channel types at a site located in the S3–S4 linker of domain IV close to the voltage sensor (Chuang et al., 1998). This hypothesis is particularly attractive because the same (or equivalent) channel region appears to be critical for high-affinity binding of other gating modifiers to voltage-gated K and Ca channels (Fig. 9). For example, the substitution of a particular glutamate (in position 1613) with neutral amino acids weakens binding of the α-scorpion toxin V fromLeiurus quinquestriatus (LqTX) to Na channels (Rogers et al., 1996), whereas its equivalent, in Shab (drk1) K channels, appears to be critical for the binding of hanatoxin and grammotoxin (Li-Smerin and Swartz, 1998, 2000) and, in Ca channels, for the high-affinity binding of ω-Aga-IVA (Winterfield and Swartz, 2000).
In this context it is interesting to see that kurtoxin interfered noticeably with ω-Aga-IVA inhibition of P-type Ca channels in Purkinje neurons. Kurtoxin diminished the magnitude of P-type Ba current block by ω-Aga-IVA and slowed the time course of this inhibition. These effects suggest significant overlap between the binding sites for each toxin, and, because ω-Aga-IVA binding has been mapped to the S3–S4 linker of domain IV (Winterfield and Swartz, 2000), they identify this region as one likely target for kurtoxin binding to P-type Ca channels.
We do not know whether kurtoxin binds to domain IV S3–S4 linker in T-type, L-type, and N-type Ca channels. This question is of particular interest in the case of N-type Ca channels for which the response to kurtoxin differs dramatically from that of P-type Ca channels despite their structural similarity (Fig. 9).
Our results suggest that kurtoxin is suited ideally for structural studies of voltage-dependent gating. Once conditions are identified for the expression of Ca channel clones that mimic the diversity of native Ca channels in their response to kurtoxin, this gating modifier may provide a powerful approach to the identification of key mechanisms of activation and deactivation in voltage-gated Ca channels.
Footnotes
This work was supported by National Institutes of Health Grant R01 NS36794. We thank Dr. Kenton J. Swartz for his generous gift of kurtoxin and helpful suggestions and Drs. Enrico Nasi and Abdul Traish for their comments on data analysis.
Correspondence should be addressed to Isabelle M. Mintz, Department of Pharmacology and Experimental Therapeutics, Boston University Medical Center, 80 East Concord Street, Boston, MA 02118. E-mail:imintz{at}bu.edu.