Journal of Neurophysiology

Ef-Current Contributes to Whole-Cell Calcium Current in Low Calcium in Frog Sympathetic Neurons

Haoya Liang, Keith S. Elmslie

Abstract

Because Ca2+ plays diverse roles in intracellular signaling in neurons, several types of calcium channels are employed to control Ca2+ influx in these cells. Our experiments focus on resolving the paradox of why whole-cell current has not been observed under typical recording conditions for one type of calcium channel that is highly expressed in frog sympathetic neurons. These channels, referred to as Ef-channels, are present in the membrane at a density greater than the channels that carry ∼90% of whole-cell current in low Ba2+; but, Ef-current has not been detected in low Ba2+. Using Ca2+ instead of Ba2+ as the charge carrier, we recorded a possible E-type current in frog sympathetic neurons. The current was resistant to specific blockers of N-, L-, and P/Q-type calcium channels but was more sensitive to Ni2+ block than was N- or L-current. Current amplitude in Ca2+ is slightly greater than that in Ba2+. In 3 mM Ca2+, the current contributed ∼12% of total current at peak voltage and increased at voltages more hyperpolarized to the peak, reaching ∼40% at −30 mV, where whole-cell current starts to activate. The presence of Ef-current in 3 mM Ca2+ suggests a potential role for Ef-channels in regulating calcium influx into sympathetic neurons.

INTRODUCTION

Calcium current resistant to blockers of N-, L-, and P/Q-type channels are widely distributed in the nervous system, where they have been implicated in dentritic integration (Delmas et al. 2000) and synaptic release of neurotransmitters and hormones (Wang et al. 1999;Wu et al. 1998). These currents have been receiving increased attention since the cloning of α1Echannels (CaV 2.3). Resistant current (R-current) has been detected in frog sympathetic neurons in 2 mM Ba2+(Elmslie et al. 1992) as well as in 100 mM Ba2+ (Elmslie et al. 1994). In 2 mM Ba2+, it has been shown that 5% of whole-cell current is resistant to both ω-conotoxin GVIA (ωCGVIA) and dihydropyridines (DHPs). Except for its insensitivity to ωCGVIA, the resistant current in low Ba2+ shares several biophysical properties with N-current, including a similar current–voltage relationship (I-V), voltage-dependent inhibition by norepinephrine (NE), and increased inactivation by phosphorylation (Elmslie et al. 1992;Werz et al. 1993). We therefore refer to it as “N-like” current.

Surprisingly, the properties of the current resistant to ωCGVIA and DHPs change in high Ba2+. In 100 mM Ba2+, the current contributes 30–50% of the whole-cell current and activates and inactivates at voltages more hyperpolarized than those of N-current (Elmslie et al. 1994). The current is not inhibited by NE but can be preferentially blocked by Ni2+. Pharmacological characteristics of the resistant current resemble those of current through channels expressed from E-class mRNA (α1E channels) (Bourinet et al. 1996). Therefore, the resistant current in high Ba2+ was termed Ef-current (Elmslie 1997), which is the abbreviation for E-type current in frog neurons.

Single-channel studies in isotonic Ba2+ have identified Ef-channels and possible N-like channels (Elmslie 1997). Both channel types are insensitive to ωCGVIA and DHPs. Voltage-dependent properties of Ef-channels match those of whole-cell Ef-current in 100 mM Ba2+. In contrast, the potential N-like channels were indistinguishable from N-channels in their activation voltage range (>0 mV), single-channel conductance (∼20 pS), or unitary current amplitude (∼1.4 pA at 0 mV). Single-channel evidence also shows that the density of putative Ef-channels is equal to or greater than that of ωCGVIA-sensitive N-channels (Elmslie 1997). N-channels are responsible for the majority of calcium current in frog sympathetic neurons. The objective of our experiments is to search for a current that may be attributed to channels coded by E-type genes under more physiological recording conditions.

If channels expressed from E-class mRNA are present in frog sympathetic neurons, they should pass larger current in Ca2+than they do in Ba2+, according to the relative permeability of the two divalent cations in α1Echannels (Bourinet et al. 1996). L- and N-channels, which have been identified in sympathetic neurons (Elmslie et al. 1992), carry smaller current in Ca2+than they do in Ba2+ (Bourinet et al. 1996; Hess et al. 1986; Wakamori et al. 1998). These channel properties suggest that using Ca2+ as the charge carrier will facilitate detection of potential E-type current. The criteria used to identify E-type current were based on three prominent characteristics of currents through α1E recombinant channels, i.e., resistance to specific blockers of N-, L-, and P/Q-type calcium channels; sensitivity to being blocked by Ni2+; and current amplitude in Ca2+ that is the same as or larger than that in Ba2+.

Methods

Cells

Adult bullfrogs (Rana catesbeiana) were chilled to 4°C, brain pithed, decapitated, and spine pithed before their paravertebral sympathetic ganglia were removed. The method of sacrifice was approved by the Institution Animal Care and Usage Committee. Neurons were dissociated with collagenase/dispase digestion and trituration (Elmslie et al. 1992; Jones 1987; Kuffler and Sejnowski 1983). Cells were maintained for 1–14 days at 4°C in L-15 medium, which was supplemented with 10% fetal bovine serum and penicillin/streptomycin.

Electrophysiology

Neurons were voltage-clamped in the whole-cell configuration. Series resistances ranging from 0.3 to 1.5 MΩ were compensated at 80%. Currents were recorded using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA). Experiments were controlled with a Macintosh II computer (Apple Computer, Cupertino, CA) running S3 data acquisition software written by Dr. Stephen Ikeda (Guthrie Research Institute, Sayre, PA). Currents were digitized with a MacAdios II analog–digital converter (GW Instruments, Somerville, MA) and stored on a hard disk. Leak current was subtracted using a P/4 protocol. Voltage steps were 10 ms in length unless otherwise noted. Step and tail currents were sampled at 50 kHz and were typically filtered at 5 kHz. In the ramp voltage protocol, the membrane was depolarized from −80 to +80 mV during 160 ms. Ramp currents were sampled at 5 kHz and filtered at 1 kHz. Ramp I-Vs were produced by plotting ramp current against ramp voltage after subtracting leak current, which was estimated by fitting the current between −80 and −60 mV to linear function. All recordings were carried out at 25°C.

Solutions

To isolate calcium currents, Na+ and K+ were replaced in the internal and external solutions with the impermeant cationN-methyl-d-glucamine (NMG+). The internal solution contained (in mM) 51.6 NMG-Cl, 6.0 MgCl2, 14 creatine-PO4, 2.5 NMG-HEPES, 5 Tris2-ATP, 20 NMG–bis-(o-aminophenoxy)-N,N,N,N′-tetraacetic acid (BAPTA), and 0.3 Li2-guanosine 5′-triphosphate (GTP). External solutions contained either Ba2+ or Ca2+ at 3 or 100 mM as the charge carrier. Other components in external solutions included (in mM) 10 NMG-HEPES and 10 tetraethylammonium chloride (TEA-Cl). Finally, NMG-Cl was added to maintain iso-osmolarity between internal and external solutions. The osmolarity of the internal solution was within the range of 260 to 290 mOs and that of the extenal solutions was from 260 to 320 mOs. All solutions were titrated to pH 7.2 with NMG base.

Data analysis

Data were analyzed using Igor Pro (WaveMetrics, Lake Oswego, OR) running on a Quadra 630 Macintosh computer. Step current was measured as the average of 10 points on the current trace at the end of the voltage step. Fractional block = 1 − (I Ca during block/I Ca control). For the Ni2+ block experiments, rundown was corrected by averaging control before and after each Ni2+concentration was applied, i.e., I Cacontrol = (I Ca before +I Ca after)/2. Half-maximum block (IC50) was estimated from a least-square regression fit of the data points according to the mass action equationB/B max = 1/(1 +K d/[Ni2+]o). Group data were calculated as mean ± SD throughout the study.

Chemicals

ωCGVIA and ω-conotoxin MVIIC (ωCMVIIC) were obtained from Bachem Bioscience (King of Prussia, PA). (±) Bay K 8644 and nimodipine were obtained from Research Biochemicals (Woburn, MA). Li2-GTP was obtained from Boehringer Mannheim (Indianapolis, IN). All other chemicals were obtained from Sigma (St. Louis, MO).

RESULTS

Whole-cell current in Ba2+ versus Ca2+

Previous reports suggest that, in frog sympathetic neurons, current resistant to ωCGVIA and DHPs results from at least two types of current, Ef and N-like. The former was observed in isotonic Ba2+ but not in 2 mM Ba2+ whereas the latter was the only detectable ωCGVIA- and DHPs-resistant current in 2 mM Ba2+(Elmslie et al. 1992, 1994). If E-type current contributes to resistant current in low Ca2+ and Ba2+, we postulated that the current would be more prominent with Ca2+ as the charge carrier. Because Ef-current was detected in high but not in low Ba2+, we first compared ramp currents in 100 mM Ba2+ with those in 100 mM Ca2+. Ef-current was observed as a prominent “shoulder” on I-V in 100 mM Ba2+ when compared with 3 mM Ba2+ (Fig.1 A). The difference in the shape of the ramp I-V was highlighted after the peak currents were normalized to unity and the I-V was shifted along the voltage axis to superimpose at the peaks (Fig.1 B). These shifts were required to compensate for the surface charge effects of different concentrations and/or types of divalent cations (Elmslie et al. 1994; Zhou and Jones 1995). In 100 mM Ca2+, a shoulder on the I-V was also observed that was even more pronounced than that in Ba2+. At 30 mV negative to the peak, the relative amplitude of the shoulder was 25 ± 10% of peak current in Ca2+, which was significantly larger than the 14 ± 7% of peak in Ba2+ (n = 5, P < 0.01 from one-tail paired t-test). This is consistent with the idea that the relative permeation ratio of Ca2+ to Ba2+ is higher in Ef-channels than in N- and L-channels. Interestingly, when the current in 3 mM Ba2+ was compared with that in 3 mM Ca2+, there was also a larger shoulder component in Ca2+ (Fig. 1,C and D). Subsequent experiments were designed to determine the identity of the current generating the shoulder in low Ca2+.

Fig. 1.

Ramp current–voltage relationship (I-V) reveals a similar current component in high Ba2+ and low Ca2+. A: ramp I-Vs recorded in 3 mM Ba2+, 100 mM Ba2+, and 100 mM Ca2+ from a single cell. B:I-Vs in different external solutions were superimposed after being normalized to their corresponding peaks. Prominent shoulder (arrow) was observed on the rising phase of the rampI-Vs both in 100 mM Ba2+ and 100 mM Ca2+. C: ramp I-Vs in 3 mM Ba2+ and 3 mM Ca2+.D: normalized ramp currents in 3 mM Ba2+ and 3 mM Ca2+ were superimposed at peak. There is a larger shoulder in Ca2+ than in Ba2+ that is similar in its relative activation voltage to the shoulder in high Ba2+ and high Ca2+.

Fraction of resistant current in 3 mM Ba2+ versus in 3 mM Ca2+

If Ef-current underlies the larger shoulder in 3 mM Ca2+, a larger fraction of the whole-cell current should be resistant to ωCGVIA and DHPs in Ca2+ than in Ba2+. Contribution from each current component was evaluated pharmacologically with 3 μM each of ωCGVIA, nimodipine, and ωCMVIIC sequentially applied to the external solutions containing either 3 mM Ba2+ or 3 mM Ca2+ (Fig. 2). ωCMVIIC was used to test if the N-like current was caused by P/Q-current. The amplitude of peak current decreased with time after blockers were applied in Ba2+ (Fig.2 A) and in Ca2+ (Fig. 2 D). In 3 mM Ba2+, 85 ± 3, 10 ± 2, and 2 ± 1% of total current at peak voltage (0 mV) was blocked by ωCGVIA, nimodipine, and ωCMVIIC, respectively; the remaining 3 ± 1% was resistant to all three blockers (n = 7). In 3 mM Ca2+, the percentages of peak current (10 mV) sensitive to ωCGVIA, nimodipine, and ωCMVIIC were 70 ± 6, 14 ± 4, and 5 ± 2%, respectively. On average, 12 ± 5% persisted in the presence of all blockers (n= 7). Hereafter, we refer to the current that is resistant to N-, L- and P/Q-channel blockers as resistant current.

Fig. 2.

Current components in 3 mM Ba2+ and 3 mM Ca2+. AC: in Ba2+. A: time course of peak current with sequential addition of 3 μM each of ω-conotoxin GVIA (ωCGVIA), nimodipine, and ω-conotoxin MVIIC (ωCMVIIC). Thick lines mark the period of time when the blockers were applied.B: ramp currents were recorded in control and after blockers were added. Each ramp current corresponds to the average of 3 traces at the time marked (a, b, c, d) in A. C: fractional current vs. voltage for ωCGVIA-sensitive current (N-current, ▪), nimodipine-sensitive current (L-current, ♦), and resistant current (●). The fraction of each component was calculated as the ratio to control current in 3 mM Ba2+. DF: in Ca2+. D: time course of peak current with sequential application of blockers. E: ramp currents in control and after blockers were added. Each ramp current is the average of 3 traces at time noted in D. F: fractional current vs. voltage for ωCGVIA-sensitive current (N-current, ▪), DHP-sensitive current (L-current, ♦), and resistant current (●). The fraction of each component was normalized to control in 3 mM Ca2+.

The increased effect of ωCMVIIC in Ca2+relative to Ba2+ was surprising. However, we noticed that the ωCGVIA block in Ca2+ was slower than that in Ba2+ (Fig. 2). Thus it seemed likely that the increased ωCMVIIC block resulted from residual N-current. To test this idea, we examined the effect of ωCMVIIC after ∼9 min application of 3 μM ωCGVIA (range 8–10 min). Under this condition, the ωCGVIA block was complete and ωCMVIIC had no additional effect (1 ± 1%, n = 7, not shown). Thus the block by ωCMVIIC in Ca2+ (Fig.3) appears to result from an incomplete block of N-current by ωCGVIA. Therefore our estimation of each component in 3 mM Ca2+ is 75% N-current (ωCGVIA + ωCMVIIC sensitive components), l4% L-current, and 12% resistant current.

Fig. 3.

Comparison of the fractions of current components at the peak in Ba2+ (0 mV) vs. in Ca2+ (10 mV). The fractions of different components were estimated as the ratio to control current in 3 mM Ba2+ or 3 mM Ca2+. Data from 7 cells in Ba2+ and 7 different cells in Ca2+ were used to estimate the mean. Error bars, standard deviation from the mean. The fraction of each component in Ba2+ was significantly different from that in Ca2+ (*P < 0.05; **P < 0.01).

Comparing each component in Ba2+ with that in Ca2+ (Fig. 3) showed that the largest changes were in the reduced fraction of ωCGVIA-sensitive current (N-current) and the increased fraction of resistant current. The fraction of resistant current increased at voltages more hyperpolarized to the peak both in Ca2+ and Ba2+, but the percentage of resistant current in Ca2+ was significantly higher (P < 0.05) than that in Ba2+. Although the percentage of L-current also increased with hyperpolarization both in Ca2+ and Ba2+, there was no significant difference (P > 0.05) in its percentage at 30 mV hyperpolarized to the peak when external solution was switched from Ca2+ to Ba2+. These findings provide quantitative support for the hypothesis that the more prominent shoulder on the I-V in 3 mM Ca2+ results from resistant current. The hyperpolarized activation voltage of resistant current is consistent with a substantial contribution from Ef-current.

Amplitude of resistant current in 3 mM Ba2+ versus 3 mM Ca2+

Because of the slow washout of peptide toxins, the comparison of component currents (Figs. 2 and 3) was carried out on data collected from two groups of cells, one group in Ba2+ and the other group in Ca2+. To determine the relative permeability of Ca2+ versus Ba2+ through channels underlying resistant current, we further examined the effect of switching between Ba2+ and Ca2+ on the amplitude of resistant current in the same cell. With 3 μM each of ωCGVIA, nimodipine, and ωCMVIIC in the external solutions,I-V of resistant current in 3 mM Ba2+was compared with that in 3 mM Ca2+ (Fig.4). Peak current occurred at more hyperpolarized voltages in 3 mM Ba2+ than it did in 3 mM Ca2+, as expected from more effective screening of surface charge by Ca2+. The amplitudes of the peak current in Ca2+ were consistently larger than those in Ba2+ regardless of the order in which Ca2+ and Ba2+ were applied. On average, the ratio of peak current in 3 mM Ca2+ to that in 3 mM Ba2+ was 1.14 ± 0.07 (n = 5). Because α1E recombinant channels carry larger Ca2+ current than Ba2+ current (Bourinet et al. 1996), the results support the idea that the activity of Ef-channels contributes to resistant current.

Fig. 4.

I-Vs of resistant current in 3 mM Ba2+ and 3 mM Ca2+ from the same cell. In the presence of 3 μM each of ωCGVIA, nimodipine, and ωCMVIIC,I-V shifted to depolarized voltages when the switch was made from Ba2+ to Ca2+. PeakI-V current in Ca2+ was larger than that in Ba2+ within the same cell regardless of which divalent external was applied first. Similar results were obtained in all cells examined (n = 5).

Ni2+ sensitivity of N-, L-, and resistant current in 3 mM Ca2+

The data presented so far are consistent with the hypothesis that the larger shoulder in 3 mM Ca2+ results from Ef-current. Another characteristic of Ef-current in high Ba2+ was its higher sensitivity to block by Ni2+ (Elmslie et al. 1994). Therefore the Ni2+ sensitivity of resistant current in 3 mM Ca2+ was examined and compared with that of N- and L-current. Estimation of the IC50 of Ni2+ for blocking N-current was carried out in the presence of 3 μM nimodipine to block L-current. By eliminating L-current, the Ni2+sensitivity of the remaining current would be weighed toward that of N-current. In the example cell (Fig.5 A), step current to +10 mV decreased with increasing [Ni2+]o and the fractional block was best fit with a single-site mass action equation. The IC50 of Ni2+ was calculated from the fit to be 245 μM in the cell illustrated (Fig.5 C). On average, Ni2+ blocked N-current with IC50 of 322 ± 73 μM (n = 7). Our experiment probably underestimated the value of IC50 of Ni2+ for blocking N-current because 1) in the presence of 3 μM nimodipine, whole-cell current is the mixture of N-current and resistant current and the latter will be shown to be more sensitive to Ni2+ block and 2) we had evidence that, at concentrations >1 mM, Ni2+right-shifted the N-channel I-V because the brief outward current on stepping to +10 mV was decreased. The outward current is thought to result from the movement of gating particles (Jones and Marks 1989) and its reduction suggests that Ni2+ is screening surface charge, leading to a shift in channel activation to more depolarized voltages. Therefore the fractional block may be overestimated when [Ni2+]o >1 mM because the decrease in current size was the cumulative effect of the block and a shift in the I-V caused by Ni2+.

Fig. 5.

Estimation of Ni2+ sensitivity of N-current in 3 mM Ca2+. Ni2+ block of calcium current was recorded in the presence of 3 μM nimodipine. A: step currents in control and in the presence of 30, 300, and 3,000 μM Ni2+. B: time course of Ni2+block for the currents in A. Application of Ni2+ is indicated by thick bars. C: steady-state fractional block by Ni2+ from the example cell was fitted with a single-site mass action equation that yielded IC50 of 245 μM.

The Ni2+ sensitivity of L-current was determined by measuring Ni2+ block of the slow tail current induced by 1 μM Bay K 8644 (Fig. 6). Under these conditions, tail current deactivation at −40 mV could be fitted with the sum of two exponential functions, yielding τ fast (τf = 0.3 ms) and τ slow (τs = 2.3 ms). The fast deactivating current resulted from non-L-current because τf was similar to the deactivation time constant at the same tail voltage in the presence of nimodipine (Fig. 6, inset). The slow deactivating component has been shown to result exclusively from L-current (Elmslie et al. 1992; Plummer et al. 1989). L-current amplitude was measured at 4 ms after repolarization, when the deactivation of the fast component was complete, and plotted against [Ni2+]o (Fig.6 C). Ni2+ block was best fit with a single-site mass action equation yielding an IC50of 261 μM in the example cell and 244 ± 39 μM in all cells examined (n = 6).

Fig. 6.

Estimation of Ni2+ sensitivity of L-current in 3 mM Ca2+. 1 μM Bay K 8644 was added in the external solutions. A: membrane was stepped to 10 mV and repolarized to −40 mV in the presence of 30, 300, and 3,000 μM Ni2+. Inset: currents in the presence of 3 μM nimodipine and 1 μM Bay K 8644 were normalized to peak tail current and superimposed. With 1 μM Bay K 8644, tail current deactivation could be fit with a double exponential function. The slow component was caused by deactivation of L-current, which was measured as the average between 4.0 and 4.2 ms after repolarization to −40 mV (arrow). B: time course of L-current block by increasing [Ni2+]o. Application of Ni2+ is indicated by thick bars. C: steady-state fractional block vs. [Ni2+]o was fitted with single-site mass action equation from which IC50 was calculated to be 261 μM in the example cell.

Ni2+ sensitivity of resistant current was determined after other currents were blocked with 3 μM each of ωCGVIA, nimodipine, and ωCMVIIC (Fig.7). IC50 of Ni2+ block was 76 μM according to a single-site mass action equation in the example cell and 83 ± 21 μM for the group (n = 10). Given the possibility that Ef- and N-like current both contribute to resistant current, we tried to further dissect resistant current based on Ni2+ sensitivity. To this end, fractional block of resistant current was also fitted with the sum of two mass action equations, which gave a slightly closer fit than assuming a single site. From two-site fit, 79% of the current was blocked by Ni2+ with IC50 of 49 μM and 21% of the current was blocked with IC50 of 491 μM for the example cell. In all cells tested, the IC50 of Ni2+ at the high affinity site was 39 ± 9 μM and that at the low affinity site was 436 ± 109 μM (n = 10). In Fig.8, fractional block versus [Ni2+]o relationship was compared among N-, L-, and resistant current. Resistant current was more sensitive to Ni2+ block than either N- or L-current. The IC50 of Ni2+for blocking resistant current in Ca2+ is close to that required for blocking current through expressed α1E channels (IC50 = 25 μM) (Jouvenceau et al. 2000). The results support the idea that resistant current in low Ca2+predominantly derives from Ef-current. The Ni2+-insensitive component of resistant current, which was ∼20% of resistant current according to the two-site fit, could arise from the activity of the N-like channels.

Fig. 7.

Estimation of Ni2+ sensitivity of resistant current in 3 mM Ca2+. Resistant current was isolated by applying 3 μM each of ωCGVIA, nimodipine, and ωCMVIIC. A: step current to 0 mV in control and in the presence of 30, 300, and 3,000 μM Ni2+. Points within the first 0.7 ms after step to 0 mV and within the first 0.3 ms after repolarization to −40 mV were blanked to eliminate the transients. B: time course of block with increasing [Ni2+]o. Application of Ni2+ is indicated by thick bars. C: steady-state fractional block vs. [Ni2+]o was fitted with single-site (thin line) as well as with two-site (thick line) mass action equations. IC50 from single-site fit was 76 μM. Using a two-site equation gives a slightly closer fit to the data and yields IC50 of 49 and 491 μM for 79 and 21% of the overall resistant current, respectively.

Fig. 8.

Ni2+ sensitivity of current components in 3 mM Ca2+. Mean fractional block of N-current (▪), L-current (♦), and resistant current (●) were plotted against [Ni2+]o. Error bars, standard deviation. Data were best fitted with a single-site mass action equation for ▪ and ♦, with both single-site (thin line) and two-site (thick line) mass action equations for ●. IC50 calculated from the fit were 308 μM (n = 4 to 6) for N-current and 254 μM for L-current (n = 3 to 4). IC50 for blocking resistant current were 70 μM with single-site fit and 46 μM (79%) and 447 μM (21%) with two-site fit (n = 3 for 3 μM Ni2+;n = 9 to 12 for other concentrations).

Ni2+-sensitive component of whole-cell current in 3 mM Ca2+

Finally, we examined the effect of 30 μM Ni2+ on whole-cell calcium current to test our initial hypothesis that the larger shoulder on the I-V in Ca2+ (Fig. 1 D) resulted from Ef-current. Based on the IC50 for Ni2+ blocking different current components, 30 μM Ni2+ should block 40% of the Ni2+-sensitive component in resistant current and 10% of N- and L-current. Thus this concentration of Ni2+ was expected to suppress the shoulder but have less effect on the peak current. In the presence of 30 μM Ni2+ (Fig. 9), the shoulder in 3 mM Ca2+ was substantially decreased. On average, 30 μM Ni2+ blocked a significantly larger fraction of the total current at the shoulder (measured at 30 mV hyperpolarized to the peak) than at the peak (28 ± 9% vs. 10 ± 3%, n = 11,P < 0.01, paired t-test). The results support the hypothesis that the larger shoulder on the I-Vthat is observed in low Ca2+ results from Ef-current.

Fig. 9.

Effect of 30 μM Ni2+ on whole-cell current in 3 mM Ca2+. Ramp currents in 3 mM Ca2+ plus 30 μM Ni2+ (thin solid line) and 3 mM Ba2+ (thick broken line) were normalized to control current in 3 mM Ca2+ (thick solid line). Normalized currents in 3 mM Ba2+ and 3 mM Ca2+ control were superimposed at peak. There was a larger shoulder component in 3 mM Ca2+than in 3 mM Ba2+, which was significantly diminished after the addition of 30 μM Ni2+.

DISCUSSION

Our data show that, in frog sympathetic neurons, a notable fraction (∼10%) of the whole-cell current in physiological Ca2+ exhibits characteristics distinct from those of N- and L-current. Several lines of evidence support the identification of this current as Ef-current. First, this current shares a pharmacological profile with α1E recombinant channels, including resistance to specific blockers of N-, L-, and P/Q-channels, and sensitivity to Ni2+ block. Second, the current was larger with Ca2+ as the charge carrier than it was with Ba2+, which has been shown to be a salient property of α1Erecombinant channels. Finally, the current activates at voltages negative to N-current, as has been shown for Ef-current in high Ba2+(Elmslie et al. 1994). As a result of the hyperpolarized activation voltage, the current forms a shoulder on the I-Vrelationship that is highlighted when the I-V in Ca2+ is superimposed on that in Ba2+.

Composition of whole-cell current in Ba2+ versus Ca2+

N-, L-, Ef-, and N-like current have been identified at the whole-cell and single-channel levels in frog sympathetic neurons (Elmslie 1997; Elmslie et al. 1992; Jones and Marks 1989). In low Ba2+, the percentage of N-current from our data (85%) is consistent with earlier findings. The development of N-current block by ωCGVIA followed a single exponential function with a time constant of 20.9 ± 2.8 s (n = 11), which agrees well with reported data (Boland et al. 1994). In low Ca2+, the percentage of N-current was ∼75%, which was significantly lower than that in Ba2+. ωCGVIA blocked N-current with a time constant of 67.3 ± 8.9 s (n = 7). Previous reports demonstrated the effect of divalent cations on ωCGVIA block (Boland et al. 1994). The effect has been interpreted either as a charge screening effect or as divalent competition for binding (Boland et al. 1994; McDonough et al. 1996). Regardless of the mechanism, Ca2+more strongly slows ωCGVIA binding than does Ba2+.

The fraction of L-current increased slightly when the switch is made from Ba2+ (10%) to Ca2+(14%). This increase was unexpected because it has been shown that Ba2+ current is greater than Ca2+ current in L- and N-channels (Bourinet et al. 1996; Hess et al. 1986;Wakamori et al. 1998). One reason may be that L-channels pass Ca2+ better than do N-channels. Within-cell comparisons showed that peak current (primarily N-current) in 3 mM Ba2+ is 2.0 ± 0.2 (n = 16) times greater than that in 3 mM Ca2+. Single-channel data from cardiac myocytes provide an estimate for relative unitary L-current amplitude in Ba2+versus Ca2+ (Hess et al. 1986). From the current–concentration relationship, unitary current in 3 mM Ba2+ is 1.4 to 1.6 times greater than that in 3 mM Ca2+. This suggests that switching from Ba2+ to Ca2+ may have less effect on the current size through L-channels than that through N-channels. Therefore the percentage of L-current in whole-cell current increased after 3 mM Ba2+ was replaced with 3 mM Ca2+ whereas that of N-current decreased. The relative amount of L-current in low Ba2+ was higher than reported previously (∼5%, Elmslie et al. 1992). The reason for the larger fraction of L-current in our experiments is not clear. One possible explanation is that the creatine phosphate added to the internal solution in our experiments slowed the rate of L-current rundown.

Our experiments suggest that there is no detectable P/Q-current in frog sympathetic neurons. In 3 mM Ba2+, application of ωCMVIIC had a negligible effect on peak current size (2 ± 1%,n = 7), which could not be differentiated from current rundown. In 3 mM Ca2+, the reduction in peak current size in response to ωCMVIIC was caused by block of residual N-current because prolonged ωCGVIA application completely eliminated ωCMVIIC effect.

Although Ef-current is more sensitive to Ni2+ block than are the other current components in frog sympathetic neurons, there is no Ni2+concentration that would selectively block only Ef-current in low Ca2+. 30 μM Ni2+ was used to examine the potential contribution of Ef-current to whole-cell current in 3 mM Ca2+ because, at this concentration, Ni2+ should produce a substantial block of the Ef-current (∼40%) with minimum block (∼10%) of N- and L-current. The observation that 30 μM Ni2+ reduced the shoulder in low Ca2+ (Fig. 9) supports the idea that Ef-current is a major component of whole-cell current at voltages hyperpolarized to the peak.

Molecular basis of resistant current

In rat cerebellar granule cells, resistant current could also be dissected into Ni2+-sensitive and -insensitive components (Tottene et al. 1996, 2000). Single-channel conductance of the Ni2+-sensitive component matches that of Ef-channels in frog neurons whereas that of the Ni2+-insensitive component is larger than that of Ef-channels. Interestingly, both Ni2+-sensitive and -insensitive components were diminished after the neurons were injected with α1E antisense mRNA, which suggests that they are coded by the same gene (Tottene et al. 2000).

Other results support the idea that multiple genes can give rise to resistant currents. In one case, expressed α1Dchannels in HEK 293 cells were shown to be incompletely blocked by DHPs but sensitive to Ni2+ block (Xu and Lipscombe 2000). In a second example, substantial resistant current was observed in neurons isolated from α1E knockout mice. The identity of this channel was speculated to be α1A (Wilson et al. 2000). These results demonstrated that the type of channel generating resistant currents cannot be determined by pharmacology alone. We used pharmacology to detect and isolate the resistant current in frog sympathetic neurons, but it was the permeation properties of the current that helped us to further identify the likely gene.

The molecular basis of resistant current in frog sympathetic neurons has not been examined in our experiments, but our data are consistent with the notion that α1E gene product gives rise to the Ni2+-sensitive component of resistant current, which we have tentatively identified as Ef-current. The Ni2+-insensitive component of resistant current could arise from a distinct gene. Alternatively, it is possible that, despite their difference in Ni2+ sensitivity, both components of resistant current derive from the activity of channels coded by the E-class gene, mirroring findings in rat cerebellar granule cells.

Functional significance of Ef-channels

Single-channel experiments in 100 mM Ba2+indicated that the density of Ef-channels is as high as that of N-channels in frog sympathetic neurons but, in 3 mM Ca2+, Ef-current makes up only ∼10% of the total current at peak. This inconsistency may be attributed to three major differences in Ef- and N-channels. First, Ef-channels tend to gate in a low Po mode with brief open times (Elmslie 1997) whereas N-channels normally gate in a mode characterized by high Po (0.8–0.9) and long open times (Lee and Elmslie 1999). Second, single-channel conductance as well as unitary current are smaller for Ef-channels than they are for N-channels (Elmslie 1997). Last, steady-state inactivation of Ef-channels at resting membrane potential (−80 mV) may result in fewer channels being available for opening on depolarization, as compared with N-channels (Elmslie 1997; Elmslie et al. 1994).

Although Ef-current accounts for only ∼10% of whole-cell current at peak, the fraction of the current increases at hyperpolarized voltages. This negative activation voltage suggests that Ef-channels can open with moderate membrane depolarization, such as that induced by large excitatory postsynaptic potentials. Given the density of Ef-channels, influx of Ca2+ through these channels is likely to have widespread effects by increasing local [Ca2+]i and initiating cellular processes that are highly sensitive to Ca2+, such as the opening of Ca2+-activated channels and the assembly of machinery that leads to neurotransmitter release. Therefore the unique physiological role of Ef-channels may lie in the fine-tuning of neuronal firing patterns and neurotransmitter release.

Acknowledgments

We thank Drs. Geoffrey G. Schofield and Norman R. Kreisman for valuable comments on the manuscript.

This study was supported by National Institute of Neurological Disorders and Stroke Grant NS-33671.

Footnotes

  • Address for reprint requests: K. S. Elmslie, Dept. of Physiology SL-39, Tulane University Health Science Center, 1430 Tulane Ave., New Orleans, LA 70112 (E-mail: kelmslie{at}tulane.edu).

REFERENCES

View Abstract