Ca2+-dependent inactivation (CDI) has recently been shown in heterologously expressed N-type calcium channels (CaV2.2), but CDI has been inconsistently observed in native N-current. We examined the effect of Ca2+ on N-channel inactivation in rat sympathetic neurons to determine the role of CDI on mammalian N-channels. N-current inactivated with fast (τ ∼ 150 ms) and slow (τ ∼ 3 s) components in Ba2+. Ca2+ differentially affected these components by accelerating the slow component (slow inactivation) and enhancing the amplitude of the fast component (fast inactivation). Lowering intracellular BAPTA concentration from 20 to 0.1 mM accelerated slow inactivation, but only in Ca2+ as expected from CDI. However, low BAPTA accelerated fast inactivation in either Ca2+ or Ba2+, which was unexpected. Fast inactivation was abolished with monovalent cations as the charge carrier, but slow inactivation was similar to that in Ba2+. Increased Ca2+, but not Ba2+, concentration (5–30 mM) enhanced the amplitude of fast inactivation and accelerated slow inactivation. However, the enhancement of fast inactivation was independent of Ca2+ influx, which indicates the relevant site is exposed to the extracellular solution and is inconsistent with CDI. Fast inactivation showed U-shaped voltage dependence in both Ba2+ and Ca2+, which appears to result from preferential inactivation from intermediate closed states (U-type inactivation). Taken together, the data support a role for extracellular divalent cations in modulating U-type inactivation. CDI appears to play a role in N-channel inactivation, but on a slower (sec) time scale.
Influx of Ca2+ through N-type calcium channels triggers neurotransmitter release at a variety of central and peripheral synapses (Meir et al. 1999). Thus modulation of N-channel activity can produce profound effects on neuronal communication. One regulatory mechanism is neurotransmitter-induced inhibition of N-type channels (Elmslie et al. 1990; Koh and Hille 1997). Inactivation is another mechanism by which N-channel activity can be reduced. Several inactivation pathways have been found for calcium channels. In L-type channels (CaV1), the two primary mechanisms are voltage-dependent inactivation (VDI) and Ca2+-dependent inactivation (CDI) (Giannattasio et al. 1991; Gutnick et al. 1989; Yue et al. 1990). VDI increases monotonically with increasing voltage and saturates at voltages where channels are fully activated. CDI shows U-shaped voltage dependence with a strong correlation with Ca2+ influx. A different form of VDI identified in N-channels has U-shaped voltage dependence (Jones and Marks 1989b; Patil et al. 1998) and has been termed U-type inactivation (Klemic et al. 2001). U-type inactivation is observed with either Ca2+ or Ba2+ as the charge carrier, which distinguishes it from CDI (Jones and Marks 1989b; Patil et al. 1998).
CDI has been observed in L-type and P/Q-type calcium channels (Giannattasio et al. 1991; Lee et al. 1999, 2000), but had been controversial for N-type channels (Cox and Dunlap 1994; Jones and Marks 1989b; Patil et al. 1998). Studies that examined N-current inactivation in bullfrog sympathetic neurons (Jones and Marks 1989a) and expressed mammalian N-channels in HEK293 cells failed to find evidence of CDI (Patil et al. 1998), whereas Cox and Dunlap (1994) provided solid evidence for CDI of chick N-current. It was recently shown that expressed N-channels showed CDI (Liang et al. 2003), but it was only obvious when N-channels were expressed with the CaVβ2a subunit as opposed to CaVβ1b or CaVβ3 subunits (Patil et al. 1998). The CaVβ2a subunit seemed to slow VDI sufficiently to reveal CDI (Liang et al. 2003). In addition, the concentration of intracellular Ca2+ buffer was low (0.5 mM EGTA), which was a condition also required for CDI of chick N-current (Cox and Dunlap 1994). However, N-current inactivation was not affected by intracellular Ca2+ chelator concentration in the other studies (Jones and Marks 1989b; Patil et al. 1998). To determine the effect of Ca2+ on inactivation of native mammalian N-current, we compared Ca2+ versus Ba2+inactivation of N-current recorded from rat sympathetic neurons. We found that Ca2+ could affect inactivation, but some of those effects were inconsistent with CDI. We conclude that the Ca2+ enhancement of fast U-type inactivation (τ = ∼50 ms) results from Ca2+ binding to an external site on the channel. CDI seems to play a role in N-channel inactivation, but on a slower (s) time scale.
Cell isolation procedure
Superior cervical ganglion (SCG) neurons were acutely isolated from adult Sprague-Dawley rats (150–350 g) as described previously (Ehrlich and Elmslie 1995). Briefly, rats were anesthetized with ether and decapitated, and the heads were placed in iced Hank's balanced salt solution. Neurons were dissociated from the isolated ganglia by enzymatic digestion at 37°C followed by vigorous shaking. The enzymatic digestion was stopped by the addition of 10% fetal calf serum to the media. The dissociated cells were plated in 35-mm culture dishes and stored in a humidified atmosphere at 4°C (for ≤30 h) until use.
The neurons were voltage-clamped using the whole cell configuration of the patch-clamp technique. Electrodes were fabricated from Corning 7740 glass (ID 0.90 mm; OD 1.5 mm; Garner Glass, Claremont, CA) using a Flaming/Brown P-97 pipette puller (Sutter Instrument, San Rafael, CA) and had resistances of 1–2 MΩ, which produced a mean series resistance (Rs) of 3.52 ± 1.13 (SD) MΩ (n = 70). Series resistance was compensated by ≥80%. Membrane currents were recorded using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA.) and digitized with a 12-bit A/D converter (GW Instruments, Cambridge, MA) after analog filtering with the amplifier's four-pole low-pass Bessel filter. The digitization rate was ≥5 times the filter frequency. All experiments were conducted at room temperature.
The standard internal solution contained (in mM) 120 N-methyl-d-glucamine (NMG)-Cl, 10 tetraethylammonium (TEA)-Cl, 11 NMG2-EGTA, 10 NMG-HEPES, 1 CaCl2, 6 MgCl2, 2 Li3GDP-β-S, 14 creatine phosphate, and 5 Tris2ATP. GDP-β-S was used to block endogenous G protein activation (Ikeda 1991). The standard external solution contained (mM) 140 TEA-Cl, 10 NMG-HEPES, 5 BaCl2 or 5 CaCl2, 1 MgCl2, and 15 glucose. The pH of both solutions was adjusted to 7.4 using NMG base.
For the high BAPTA internal solution, 11 mM EGTA was replaced with 20 mM BAPTA, and the concentration of NMG-Cl was reduced from 120 to 90 mM to maintain osmolarity. Experiments with low internal BAPTA (0.1 mM) and some 11 mM EGTA experiments used low Cl− internal and external solutions (Cl− substituted with methanesulfonate) to minimize possible artifact resulting from activation of Ca2+-activated chloride channels (Sanchez-Vives and Gallego 1994). We used 0.1 mM BAPTA as opposed to 0.5 mM EGTA used in other studies (Cox and Dunlap 1994; Liang et al. 2003), because it has been estimated to closely mimic the natural Ca2+ buffering capacity of cytoplasm (Beech et al. 1991). In addition, little or no difference in inactivation was observed between EGTA and BAPTA in this or previous studies (Jones and Marks 1989b; Patil et al. 1998). Experiments to examine the effect of monovalent cation permeation on N-current inactivation used an external solution containing (mM) 150 methylamine chloride (MA-Cl), 10 NMG-HEPES, 15 glucose, and 10 NMG2-EGTA. MA+ was chosen over Cs+ or Na+ because it selectively permeates calcium channels to provide excellent current isolation (Liang and Elmslie 2001) (see also Fig. 3). For experiments using higher concentrations of external Ca2+ or Ba2+, the concentration of TEA-Cl was lowered to maintain osmolarity. The osmolarity of all external solutions was set to 300 mOsm and that of all internal solutions was 290 mOsm.
Data acquisition and analysis
Voltage steps were generated and data taken using S3 (software developed by Dr. Stephen Ikeda, National Institutes of Health, NIAAA, Bethesda, MD) on a Macintosh Quadra 800 computer (Apple Computer, Cupertino, CA). IgorPro software (WaveMetrics, Lake Oswego, OR) was used to measure current amplitudes and to fit (Marquardt-Levenberg algorithm) currents and data plots. Graphs were exported from IgorPro to Canvas (Deneba Software, Miami, FL) for final polishing. One measure used to quantify inactivation was to divide current amplitude at the end of the inactivating step (5 s) by peak current (typically measured at 20 ms into the step). This was indicated in figures as 1 − (IEnd/IPk). In addition, we used double exponential fitting to separate and quantify the fast and slow components of inactivation during 5-s voltage steps (see Supplemental Fig. 11 ). The amplitude of each inactivating component was divided by the total amplitude, which was the sum of the two inactivating and one noninactivating components. This was indicated in figures as fractional inactivation, and the sum of the two components was always less than unity because we included the noninactivating component in the calculation of total amplitude. Data are represented as mean ± SD, and statistical significance was assessed using the Wilcoxon rank sum test or Kruskal-Wallis test.
Test solutions were applied from a gravity-fed perfusion system with five inputs and a single output. NMG, TEA, MA, creatine phosphate, ATP, EGTA, and HEPES were obtained from Sigma Chemical (St. Louis, MO). GDP-β-S was obtained from Roche Molecular Biochemicals (Indianapolis, IN). All other chemicals were reagent grade.
Calcium enhances inactivation
Calcium currents were recorded in single cells where the external solution was switched from 5 mM Ba2+ to 5 mM Ca2+ (Fig. 1A). This switch induced a 10-mV right shift in the current-voltage (I-V) relationship (Liang and Elmslie 2001) and decreased peak current by an average of 58 ± 9% (n = 43). We made several adjustments to compensate for these effects, because we were interested in the effect of Ca2+ on channel inactivation. Test voltages were increased 10 mV in Ca2+ compared with those in Ba2+, and currents in Ca2+ and Ba2+ were normalized to peak (Fig. 1A). Our initial observations were in neurons dialyzed with 20 mM BAPTA to tightly control intracellular Ca2+. During a 5-s depolarization to voltages that generate peak current (0 mV in Ba2+, +10 mV in Ca2+), we observed an average increase in inactivation of 13% in Ca2+ versus Ba2+ (Fig. 1B). The Ca2+ enhancement of inactivation was similar (10 ± 9%, P < 0.001, n = 36) in cells dialyzed with 10 mM EGTA. This enhancement did not result from a contaminating Cl− current such as Ca2+-activated Cl− current (Sanchez-Vives and Gallego 1994), because the Ca2+ effect on inactivation was similar (9 ± 6%, P < 0.01, n = 10) when Cl− in the internal and external solutions was replaced by methanesulfonate.
Because N-type channels comprises 60–90% of the total current in rat sympathetic neurons (Plummer et al. 1989; Regan et al. 1991; Zhu and Ikeda 1993), it is likely that N-channel inactivation is affected by Ca2+. This was confirmed using the N-channel specific blocker ω-conotoxin GVIA (ωCGVIA; Fig. 2). Inactivation was measured (5-s steps) in both Ca2+ and Ba2+ in neurons before and after exposure to 1 μM ωCGVIA. In four cells that survived the entire procedure, Ca2+ enhanced the inactivation of the total current (before ωCGVIA) by 12 ± 12%, but the enhancement was only 6 ± 12% after toxin application. The Ca2+-enhanced inactivation of ωCGVIA-sensitive N-current was 16 ± 13%, which shows that Ca2+ enhances N-current inactivation (Fig. 2).
Two components of inactivation in both Ca2+ and Ba2+
It was clear from examining the current during 5-s voltage steps that N-current inactivated with fast and slow components. We used double exponential fitting to determine the effect of Ca2+ on each component. N-current was well described by double exponential equations plus an offset (noninactivating component), which provided amplitude and τ for the fast and slow components (Supplemental Fig. 1). From this point, the fast and slow components of inactivation will be termed fast and slow inactivation, respectively. The speed of inactivation was derived from the τ for each component. We derived the fractional amplitudes of fast and slow components relative to the total current (see methods). Ca2+ enhanced the amplitude of fast inactivation (P < 0.001) and the speed of slow inactivation (P < 0.001; Fig. 3), which together generated the increased inactivation observed in Fig. 1.
Fast and slow inactivation resulted from the activity of N-type calcium channels because both components of inactivation were identified in ωCGVIA-sensitive currents. The fast inactivation τ was 167 ± 69 ms in Ba2+ and 191 ± 79 ms in Ca2+, and the slow inactivation τ was 3253 ± 994 ms and 2278 ± 867 ms in Ba2+ and Ca2+, respectively (n = 4; cf. with Fig. 3). Ca2+ enhanced the amplitude of the fast component by 18 ± 27%. None of the Ca2+–Ba2+ differences were significant for the ωCGVIA-sensitive current, which likely resulted from the small number of observations. Indeed, Ca2+ failed to enhance inactivation of control current (before ωCGVIA application) in one of the four cells. However, the trend of Ca2+-induced increase in fast inactivation amplitude and speeding of slow inactivation were preserved.
One obvious possibility is that the Ca2+-induced enhancement results from CDI. The main characteristics of CDI are 1) the increase in the magnitude and 2) speed of inactivation on switching from Ba2+ to Ca2+, 3) a U-shaped voltage dependence where the magnitude of inactivation is correlated with Ca2+ influx, and 4) reduced inactivation by high concentrations of intracellular Ca2+ chelators (Cox and Dunlap 1994; Giannattasio et al. 1991; Haack and Rosenberg 1994; Kalman et al. 1988; Liang et al. 2003; Peterson et al. 1999). Interestingly, Ca2+ had no effect on the speed of fast inactivation (cf. open bars of Fig. 3, Bb and Eb), which is inconsistent with CDI. However, Ca2+ accelerated the slow component as expected for CDI (cf. open bars of Fig. 3, Cb and Fb). To gain a better understanding the mechanism by which Ca2+ affects each component, we examined each of these characteristics starting with reduced the levels of intracellular Ca2+ buffering, which has been particularly effective at revealing N-channel CDI (Cox and Dunlap 1994; Liang et al. 2003).
Lowering internal BAPTA differentially affected each inactivation component. The speed of fast inactivation was increased (relative to high BAPTA) to the same degree in both Ca2+ and Ba2+ (Fig. 3, Bb and Eb). In addition, the amplitude of the fast component was reduced. CDI predicts a Ca2+ specific acceleration of inactivation, which was observed for slow inactivation in low BAPTA versus high BAPTA. This suggests that CDI could be affecting native N-channel availability on a time scale of seconds as opposed to 100 ms observed for N-channels expressed with CaVβ2a subunits in HEK293 cells (Liang et al. 2003).
Effect of altered divalent cation concentration on the components of N-channel inactivation
N-CHANNEL INACTIVATION IN THE ABSENCE OF EXTERNAL DIVALENT CATIONS.
Cox and Dunlap (1994) observed a dramatic decrease of N-channel inactivation when Na+ was used as the charge carrier in chick sensory neurons, which was used as support for CDI. We were interested to determine if mammalian N-channels responded in a similar fashion. We used methylamine (MA+) as the monovalent charge carriers, which has been shown to produce significantly better calcium current isolation than with inorganic monovalent cations (Jones and Marks 1989a; Liang and Elmslie 2002). To ensure that N-current was well isolated, we measured the block of MA+ current by ωCGVIA (Fig. 4, A and B). On average, 79 ± 18% of MA+ current was blocked by 1 μM ωCGVIA (n = 5), which shows that the majority of current is attributable to N-type channels. Peak MA+ current was observed at a voltage 10 mV hyperpolarized to that in 5 mM Ba2+, so when comparing inactivation at peak current, the stimulus voltage was hyperpolarized by 10 mV for MA+ compared with Ba2+. When compared in the same cell, inactivation was significantly decreased during permeation by MA+ versus Ba2+ (Fig. 4, C and D). It seemed that this decrease resulted from the absence of fast inactivation, because the current inactivated slowly during our 5-s voltage step. Inactivation in MA+ was best fit with a single exponential equation with τ = 4.0 ± 0.2 s (at −20 mV, n = 3). Inactivation in the same three cells recorded in 5 mM Ba2+ was best fit by two exponential equations with τ = 201.7 ± 85.5 ms and 4.1 ± 1.1 s (at −10 mV) for fast and slow inactivation, respectively. There was no significant difference in the slow inactivation τ measured in MA+ versus Ba2+ (P = 0.34).
The 10-mV left shift used to compare MA+ and Ba2+ data were derived from the I-V. To further test the validity of this shift, slow inactivation τ was compared over four voltages in cells exposed to both Ba2+ and MA+. Slow inactivation τ corresponded very well, provided the MA+ data were shifted 10 mV hyperpolarized (Fig. 4E). Thus the 10-mV left shift seems to be valid for comparison of inactivation between Ba2+ and MA+. In addition, this analysis shows that slow inactivation τ decreases with depolarization in both MA+ and Ba2+, which is consistent with a classical VDI mechanism.
Other effects of zero divalent cations were slow activation and deactivation of N-current (Fig. 4). Slow deactivation is consistent with previous observations from chick N-channels (Cox and Dunlap 1994) and may be explained by the left-shifted current-voltage relationship in zero divalent cations. However, such a shift cannot explain slowed activation. The effect on activation is similar to that observed during voltage-dependent inhibition of N-current by neurotransmitters (Elmslie et al. 1990), but we included 2 mM GDP-β-S in the internal solution to prevent the interference of tonically activated G proteins (Ikeda 1991). It is possible that divalent cations enhance gating kinetics of N-type calcium channels.
INCREASING EXTRACELLULAR [CA2+] ENHANCED INACTIVATION.
Because removing external divalent cation concentration prevented fast inactivation, we were interested to determine the effect of increased external Ca2+ and Ba2+ concentration on inactivation. These experiments were done using 11 mM EGTA because altering the type and concentration of internal Ca2+ chelator had no effect on the magnitude of inactivation measured over 5-s steps. An additional benefit of high EGTA was longer recording times, which permitted the measurement of inactivation in multiple Ca2+ or Ba2+ concentrations. To compensate for the increased surface charge screening with higher divalent cation concentrations (Zhou and Jones 1995), inactivation was always measured at the voltage generating peak inward current. Comparison of the I-V and inactivation versus voltage relationships showed that altered external divalent cation concentration induced a similar shift (data not shown). Increasing Ca2+ concentration from 5 to 30 mM significantly enhanced inactivation (Fig. 5). The average peak current for each Ca2+ concentration was 4.5 ± 4.0 (5 mM), 5.5 ± 4.2 (10 mM), and 6.2 ± 4.0 nA (30 mM; n = 5). On the other hand, there was no significant change in inactivation when Ba2+ concentration was increased (Fig. 5). The average peak current for each Ba2+ concentration was 5.3 ± 3.7 (5 mM), 9.0 ± 6.3 (10 mM), and 10.5 ± 6.6 nA (30 mM; n = 5). Perhaps the larger calcium currents resulting from higher external Ca2+ increased inactivation by inducing CDI. This idea was initially examined by determining the effect of changes in Ca2+ concentration on fast and slow inactivation, which showed that increased Ca2+ specifically increased the amplitude of fast inactivation (Fig. 6A) and the speed of slow inactivation (Fig. 6D). On the other hand, increasing Ba2+ concentration failed to significantly alter either fast or slow inactivation (data not shown). The Ca2+ specific acceleration of slow inactivation is consistent with CDI (Giannattasio et al. 1991).
Voltage dependence of fast inactivation
Up to this point, the effects of Ca2+ on fast inactivation are largely inconsistent with CDI. One additional test is current dependence, which tests the requirement for Ca2+ influx needed to activate calmodulin, the intracellular mediator of CDI. We looked for current dependence by examining inactivation over a wide range of voltages using a two-pulse protocol with 500-ms prepulses given to different voltages followed by a postpulse to the voltage producing peak inward current (Fig. 7). As a result of the shorter voltage steps (500 ms), this protocol specifically examines the voltage dependence of fast inactivation. A plot of postpulse current amplitude versus prepulse voltage revealed a U-shaped relationship in Ba2+, with maximal inactivation observed at voltages near those yielding maximal inward current and less inactivation at more positive and negative voltages (Fig. 7, A and B). This results from U-type inactivation, which is characterized by U-shaped voltage dependence of inactivation in Ba2+ and an absence of current dependence to inactivation (e.g., substantial inactivation at voltages yielding little or no current) (Jones and Marks 1989b; Patil et al. 1998). Ca2+ did not alter the voltage dependence of inactivation, but, as expected, the magnitude of inactivation was increased. The average peak inactivation was 62.1 ± 0.8% in Ca2+ (n = 8) and 44.4 ± 0.8% in Ba2+ (n = 10).
One possibility is that Ca2+ enhances U-type inactivation. Alternatively, it is possible that the enhancement results from CDI that is superimposed on U-type inactivation. CDI requires Ca2+ influx, which was tested by plotting the percentage enhancement of inactivation by Ca2+ (over Ba2+, 5 mM each) versus the normalized current amplitude in Ca2+ (Fig. 7D). Ca2+ enhanced inactivation in three of four cells examined for this analysis, and in each cell, there was no correlation between the enhancement of inactivation and current amplitude. In fact, the enhancement at voltages generating little or no current was just as large as those generating peak current (Fig. 7D). Thus the enhancement of fast inactivation does not seem to require Ca2+ influx, which is inconsistent with CDI.
Even though the above analysis showed no current dependence, it is possible that a dependence on Ca2+ influx could be revealed with higher Ca2+ concentrations (larger currents). As expected from our data using 5-s steps, inactivation observed from prepulse-postpulse protocols increased with Ca2+ concentration (Fig. 7E). Current dependence was tested by plotting the percentage increase of inactivation versus the percentage increase in current amplitude (5 vs. 30 mM Ca2+, a within-cell comparison, n = 5). Again there was no relationship between inactivation and current amplitude (Fig. 7F). From this, we conclude that Ca2+ influx is not required for enhancement of fast inactivation.
We examined the effect of Ca2+ on inactivation to determine if there is a role for CDI in the inactivation of native N-type calcium channels. In accordance with previous observations of CDI, we hypothesized a Ca2+-specific enhancement of the speed and magnitude of inactivation that was dependent on the internal Ca2+ concentration (CDI). This hypothesis was tested by lowering internal Ca2+ buffer concentration, increasing external Ca2+ concentration, and correlating inactivation with inward current amplitude. While these conditions did affect inactivation, most of the effects on fast inactivation (τ = 150 ms) were inconsistent with our expectations for CDI. Lowering the intracellular BAPTA concentration significantly accelerated fast inactivation, but that effect was not specific for Ca2+ as expected for CDI. Fast inactivation increased with external Ca2+ concentration, but that increase was not dependent on Ca2+ influx. Interestingly, fast inactivation was dependent on divalent cations (Cox and Dunlap 1994), because it was lost in our zero divalent cation solution (MA+ permeation). We conclude that the presence of divalent cations is required for fast inactivation, but Ca2+ influx is not. CDI may play a role in N-channel inactivation, but on a time scale of 1–2 s.
Slow N-channel inactivation
Early work on calcium currents in rat sensory neurons showed an inactivation pathway could be blocked when external Ca2+ was replaced by Mg2+ (Schroeder et al. 1990), which led to the conclusion that CDI was involved. These experiments used 3-s voltage steps to induce inactivation, which indicates that CDI could be slow in these neurons. We have also observed effects of Ca2+ on slow inactivation that are consistent with CDI. Ca2+ accelerated slow inactivation in a concentration-dependent manner, but increased Ba2+ concentration was without effect. In addition, lowering internal BAPTA significantly accelerated slow inactivation in Ca2+, but not Ba2+. Thus the data support CDI as a slow regulator of N-channel availability in rat sympathetic neurons.
The speed of our putative CDI (τ ∼ 1,500 ms) is slow relative to that observed in expressed N-channels (τ ∼ 100 ms) (Liang et al. 2003). N-channel CDI is mediated by Ca2+ binding to the N-lobe of CaM, which makes N-channel CDI more sensitive to changes in cytoplasmic Ca2+ than that of L-type channels (Liang et al. 2003). This implies that N-channel CDI would be sensitive to both intracellular Ca2+ buffering and cell size. HEK293 cells tend to be smaller (∼15 pF) than rat sympathetic neurons (∼30 pF), but current amplitude (1–5 nA) can be similar in these preparations. Thus the lower current density of sympathetic neurons could delay increased intracellular Ca2+ relative to HEK293 cells, which could contribute to slower CDI. However, CDI was not observed in previous studies of N-channel inactivation using HEK293 cells (Patil et al. 1998) that co-expressed CaVβ3 with the N-channel as opposed to CaVβ2a used by Liang et al. (2003). CaVβ3 has been shown to be the dominant isoform in rat sympathetic neurons (Lin et al. 1996), which could explain the apparent slow CDI in our recordings (Patil et al. 1998). Perhaps differences in CaVβ subunit expression can explain the disparate N-current inactivation observations from different neuronal preparations (Cox and Dunlap 1994; Jones and Marks 1989b; Schroeder et al. 1990).
In Ba2+, slow inactivation seems to involve a classic VDI mechanism, where inactivation increases monotonically with voltage (Fig. 3E). In Ca2+, this component should also be present as has been shown for L-type channels (Giannattasio et al. 1991). We believe that CDI is an additional inactivation component with a 1- to 2-s τ. Theoretically, this should appear as a third component to inactivation. However, this τ is not sufficiently different from that of VDI (2–4 s) to permit separation by exponential fitting. Thus CDI is superimposed on VDI, and we observe the effect of CDI as a Ca2+-induced increase in the speed of slow inactivation.
Physiological effect of slow CDI
The slow time-course of CDI suggests that its most prominent impact would be on long periods of action potential activity (>500 ms). Such activity would supply the Ca2+ needed over a sufficiently long period to activate CDI in a manor that is mimicked by our long (5 s) voltage steps. In addition, slow VDI would also be increasing during the long activity period. The resulting N-channel inactivation would reduce Ca2+ influx/action potential, which could lower excitability by reducing neurotransmitter release from presynaptic terminals. On the other hand, decreased Ca2+ influx in the soma could increase excitability as a result of reduced activation of Ca2+-activated potassium channels. These predictions are difficult to test because conditions that reduce CDI will also impair the other physiological effects of increased intracellular Ca2+.
REDUCED INTRACELLULAR CA2+ CHELATION.
Previous publications showed no effect on the speed of fast inactivation in Ba2+ when lowering the concentration of intracellular Ca2+ chelator (Liang et al. 2003). Thus we were surprised to find that fast inactivation in both Ca2+ and Ba2+ was significantly accelerated by lowering internal BAPTA concentration from 20 to 0.1 mM. The increased fast inactivation τ in low BAPTA seems to be inconsistent with CDI. However, two studies have concluded that Ba2+ influx could trigger L-channel CDI (Ferreira et al. 1997; Jouvenceau et al. 2000). This seems highly unlikely because calmodulin has been shown to be the Ca2+ detector for CDI in both CaV1 and CaV2 channels (Lee et al. 1999; Liang et al. 2003; Peterson et al. 1999; Zuhlke et al. 1999), and calmodulin and other E-F hand proteins have a very low affinity for Ba2+ (>1,000 μM) versus Ca2+ (∼2.5 μM) (Gu and Cooper 2000). In addition, Haack and Rosenberg (1994) showed that CDI could not be induced in L-channels with intracellular Ba2+ concentrations ≤1 mM, but CDI was induced with low concentrations of internal Ca2+ (10 μM). Thus it is unlikely that Ba2+ could activate the calmodulin-mediated CDI observed for N-type calcium channels (Liang et al. 2003).
It is also unlikely that accelerated fast inactivation results from the activity of contaminating currents, such as Ca2+-activated Cl− current, because Cl− in the internal and external solution was replace by methanesulfonate and the only permeant cations were Ba2+ or Ca2+. Alternatively low intracellular BAPTA could invite a build-up in the intracellular concentration of Ca2+ or Ba2+ that would decrease driving force during the 5-s voltage steps. This should affect both inactivation components, but slow inactivation was not affected when Ba2+ was the charge carrier. In addition, Goldman-Hodgkin-Katz rectification predicts a substantial increase of intracellular divalent cation concentration is needed to significantly affect calcium current amplitude. For example, raising internal Ca2+ from 1 nM to 100 μM decreases calcium current at 0 mV by 2% (5 mM external Ca2+), which is much too small an effect to explain the changes in inactivation we observe using low internal BAPTA. It seems more likely that BAPTA and EGTA are affecting inactivation in a manner independent of their Ca2+ chelating abilities. Such an effect has been observed as a BAPTA-mediated block of an intracellular second-messenger pathway linking muscarinic receptors to N-channel inhibition (Beech et al. 1991).
ALTERED EXTERNAL DIVALENT CATION CONCENTRATION.
Cox and Dunlap (1994), using chick sensory neurons, showed the complete loss of fast N-channel inactivation when monovalent cations carried the current. We confirmed this observation in rat sympathetic neurons using methylamine (MA+) as the charge carrier, which provided excellent N-current isolation (Liang and Elmslie 2002). Although reduced inactivation during monovalent permeation is expected for CDI, it is also consistent with the hypothesis that external Ca2+ is required for fast inactivation.
When extracellular Ca2+ concentration was increased from 5 to 30 mM, we observed a significant increase in the magnitude of fast inactivation, but inactivation was unaltered by a similar change in Ba2+ concentration. Such a differential effect is expected from CDI. Another expectation from CDI is that the effect of Ca2+ on inactivation is current dependent, because Ca2+ must enter the cell to bind calmodulin. Current dependence of Ca2+-enhanced fast inactivation was examined by comparing the effect of Ca2+ on inactivation over a wide range of voltages. This showed that Ca2+ enhanced inactivation regardless of current amplitude. Thus a mechanism that does not require Ca2+ influx appears to enhance fast inactivation.
HYPOTHESIZED MECHANISM FOR CA2+ ENHANCEMENT OF FAST INACTIVATION.
It is clear that Ca2+ affects fast inactivation, but the evidence fails to support CDI as the mechanism. Fast inactivation of N-current in rat sympathetic neurons is best explained by a U-type inactivation model where the channels preferentially inactivate from intermediate closed states (Patil et al. 1998). To explain our data, we hypothesize that Ca2+ binds to at least two sites to enhance fast N-channel inactivation. Because Ca2+ influx does not seem to be required, these sites should be exposed to the extracellular solution, which distinguishes this from CDI. One site (site 1) needs to be occupied by divalent cations for N-channels to undergo U-type inactivation. Site 1 can bind both Ca2+ and Ba2+ because fast inactivation can be observed in either ion, but not in their absence. One candidate for site 1 is the selectivity filter because it binds both Ca2+ and Ba2+ (Almers and McCleskey 1984; Carbone et al. 1997; Hess and Tsien 1984). The second site (site 2) seems to be selective for Ca2+, but its affinity for Ca2+ is low (mM). Inactivation selectively increases with Ca2+ concentrations >5 mM. EF-hand containing proteins are highly Ca2+ selective (Gu and Cooper 2000), but these protein are intracellularly located and Ca2+ influx does not seem to be required for enhanced fast inactivation. A putative EF-hand motif has recently been identified on the extracellular face of the N-channel (domain III, S5-H5) (Feng et al. 2001), which could provide the specific Ca2+ binding to site 2. However, it should be noted that site 2 seems to have very low Ca2+ affinity (millimolar) compared with the micromolar affinity of typical EF-hand binding proteins. On the other hand, ion sensitivity may be a general feature of U-type inactivation, because increased extracellular K+ concentration enhances U-type inactivation of potassium channels (Klemic et al. 1998, 2001). The mechanism by which increased ion concentration enhances U-type inactivation remains to be determined.
Extracellular ions seem to affect the function of many voltage-dependent ion channels. Ca2+ has been shown to be an important co-factor in sodium channel closing (Armstrong 1999; Armstrong and Cota 1991, 1999). K+ occupancy of the Shaker potassium channel pore seems to be required for the channel to maintain a functional permeation pathway and gating charge movement (Loboda et al. 2001; Melishchuk et al. 1998). In addition, K+ occupancy of the pore seems to retard C-type inactivation (Baukrowitz and Yellen 1995; Kiss and Korn 1998). Ca2+ has recently been shown to modulate dihydropyridine binding to L-type calcium channels (Peterson and Catterall 2006). Thus regulation by physiological ions seems to be a common theme among voltage-dependent ion channels. Extracellular divalent cations have recently been shown to regulate the binding of ωCGVIA to N-channels (Liang and Elmslie 2002). ωCGVIA is a very slowly reversible N-channel blocker in the presence of Ca2+ or Ba2+, but rapidly dissociates from the channel in divalent cation-free solutions. This effect was interpreted to result from a divalent cation induced conformational change in N-channel structure. Perhaps fast inactivation also requires this divalent cation-induced N-channel conformation.
This work was supported by Korea Science and Engineering Foundation Grant R05-2001-000-00620-0 to Y. S. Goo and by grants from the American Heart Association, the Louisiana Educational Quality Support Fund (LEQSF), and National Institute of Neurological Disorders and Stroke to K. S. Elmslie.
↵* Y. S. Goo and W. Lim contributed equally to this work.
↵1 The online version of this article contains supplemental data.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 by the American Physiological Society