Cell Physiology

Molecular identification of a TTX-sensitive Ca2+current

Silvia Guatimosim, Eric A. Sobie, Jader dos Santos Cruz, Laura A. Martin, W. J. Lederer


The TTX-sensitive Ca2+ current [I Ca(TTX)] observed in cardiac myocytes under Na+-free conditions was investigated using patch-clamp and Ca2+-imaging methods. Cs+ and Ca2+were found to contribute to I Ca(TTX), but TEA+ and N-methyl-d-glucamine (NMDG+) did not. HEK-293 cells transfected with cardiac Na+ channels exhibited a current that resembledI Ca(TTX) in cardiac myocytes with regard to voltage dependence, inactivation kinetics, and ion selectivity, suggesting that the cardiac Na+ channel itself gives rise to I Ca(TTX). Furthermore, repeated activation ofI Ca(TTX) led to a 60% increase in intracellular Ca2+ concentration, confirming Ca2+ entry through this current. Ba2+ permeation ofI Ca(TTX), reported by others, did not occur in rat myocytes or in HEK-293 cells expressing cardiac Na+channels under our experimental conditions. The report of block ofI Ca(TTX) in guinea pig heart by mibefradil (10 μM) was supported in transfected HEK-293 cells, but Na+current was also blocked (half-block at 0.45 μM). We conclude thatI Ca(TTX) reflects current through cardiac Na+ channels in Na+-free (or “null”) conditions. We suggest that the current be renamedI Na(null) to more accurately reflect the molecular identity of the channel and the conditions needed for its activation. The relationship between I Na(null)and Ca2+ flux through slip-mode conductance of cardiac Na+ channels is discussed in the context of ion channel biophysics and “permeation plasticity.”

  • tetrodotoxin
  • excitation-contraction coupling
  • sodium channel
  • calcium channel
  • ventricular myocyte

imperfect selectivity is a property of ion channels, including cardiac Na+channels (13). Alterations in selectivity characteristics have been observed in diverse ion channels in many cell types and have been recognized as a means by which ion channels can be modulated (15), a feature we call “permeation plasticity.” In heart cells, the tetrodotoxin (TTX)-sensitive Na+ channel has been shown to become permeable to Ca2+ after protein kinase A (PKA)-dependent phosphorylation by a process called “slip-mode conductance” (28). A second TTX-sensitive Ca2+ current also has been identified in heart cells (6, 19) and attributed to a novel protein and not to the cardiac Na+ channel (1, 3, 12, 24). The novel conductance pathway was named I Ca(TTX)to recognize its distinctive identity (1).

A number of features of I Ca(TTX) support the hypothesis that this unusual current is due to a protein that is distinct from the TTX-sensitive cardiac Na+ channel. Ca2+ permeation is an important distinguishing feature because Ca2+ has been reported to block Na+channels (25). In addition, the inactivation kinetics ofI Ca(TTX) are slower than those of Na+ current (I Na) (1,3), and, like Ca2+ current (I Ca), it is permeable to both Ca2+and Ba2+ (19, 24). Recently,I Ca(TTX) has been linked to the T-type Ca2+ channel because both I Ca(TTX)and T-type Ca2+ current are blocked by 10 μM mibefradil. (12, 24). Together, this array of findings has been interpreted to suggest that I Ca(TTX) reflects current through a novel TTX-sensitive Ca2+ channel protein.

Here we investigate the molecular identity ofI Ca(TTX). While the data noted above support the notion that a novel protein may underlieI Ca(TTX), a single feature ofI Ca(TTX) raises doubt. The remarkable specificity of TTX for the Na+ channel (13) suggested to us that I Ca(TTX) may reflect the permeation properties of the cardiac Na+ channel itself in Na+-free conditions. This TTX-blockable Ca2+permeation of Na+ channels is reminiscent of slip-mode conductance because it, too, is blocked by TTX (7, 28). To rigorously test the hypothesis that I Ca(TTX) in heart cells is due to Ca2+ flux through Na+channels, we examined HEK-293 cells expressing cardiac Na+channels.

Preliminary reports have been presented to the Biophysical Society (10, 29).


Rat Cardiac Myocytes: Cell Isolation and Preparation

Adult rat heart cells were prepared by standard methods (28). Briefly, rats of either sex weighing between 200 and 300 g were killed by lethal intraperitoneal injection of pentobarbital sodium (100 mg/kg). The hearts were rapidly removed and perfused via the Langendorff method with Ca2+-free modified Tyrode solution (see Solutions) until the blood was washed out. Hearts were then perfused with Tyrode solution containing 50 μM CaCl2 along with 1.4 mg/ml collagenase (type 2; Worthington, Lakewood, NJ) and 0.04 mg/ml protease (type XIV; Sigma, St. Louis, MO) until they were soft (∼5 min). The hearts were removed from the perfusion apparatus, minced into ∼1-mm chunks, and stirred for 4 min in Tyrode solution containing 50 μM CaCl2, 0.7 mg/ml collagenase, and 0.02 mg/ml protease. Cells were filtered through a 200-μm mesh to remove tissue chunks, and extracellular Ca2+ concentration was raised to 0.5 mM over 10 min through three centrifuge cycles. Cells were stored in DMEM until they were used (within 8 h). Experiments were performed at room temperature (22–24°C).

HEK-293 Cells: Culture and Transfection

HEK-293 cells obtained from American Type Culture Collection (Manassas, VA) were cultured by standard methods. The Lipofectamine 2000 reagent (Life Technologies, Rockville, MD) was used to transfect cells at 60–95% confluence with the cardiac α-subunit of the cardiac Na+ channel (also called SNC5A, GenBank accession no. M77235; Ref. 9) and both β1- and β2-subunits (kindly provided by Lori Isom, University of Michigan). We also examined HEK-293 cells that were permanently transfected with only the α-subunit (kindly provided by Hali Hartmann, University of Maryland Biotechnology Institute). In all cases, the cells were replated on no. 1 glass coverslips and used within 72 h.


Solution compositions are noted below. Isolated myocytes and HEK-293 cells were initially superfused with modified Tyrode solution containing (in mM) 140 NaCl, 5 KCl, 5 HEPES, 1 NaH2PO4, 1 MgCl2, 2 CaCl2, and 10 glucose (pH 7.4). After the whole cell recording configuration was established, Ca2+-free modified Tyrode was washed in for 2 min, and then cells were superfused with either “Cs+ Tyrode” solution containing (in mM) 150 CsCl, 0 NaCl, 10 HEPES, 1 MgCl2, 2 CaCl2, and 10 glucose (pH 7.4 with CsOH) or “TEA+(tetraethylammonium) Tyrode,” which was identical to Cs+Tyrode except that TEA-Cl replaced CsCl and TEA-OH was used to adjust the pH to 7.4. For experiments with transfected HEK-293 cells, extracellular solutions contained 8 mM CaCl2 but were otherwise identical to those noted above. At least 2 min elapsed after each solution change before recordings were made. Two basic pipette filling solutions were used: the first contained (in mM) 150 CsCl, 5 EGTA, 10 HEPES, and 4 Mg-ATP (pH 7.2 with CsOH); the second was identical except that tetramethylammonium-Cl was used as a CsCl replacement and pH was adjusted to 7.2 with TEA-OH. In experiments with myocytes a slightly modified Cs+ filling solution was used, in which 20 mM TEA+ was added to the pipette solution, replacing 20 mM Cs+. Citrate-free TTX (Calbiochem) was used in these experiments. Other modifications to these solutions that were made during some experiments are noted in the text and figure legends.


Cells were voltage clamped in the whole cell mode with 200A and 200B patch-clamp amplifiers from Axon Instruments (Foster City, CA). Data acquisition was performed with pCLAMP (versions 6.01 and 7; Axon Instruments), and pCLAMP, Origin (version 6.0; Microcal, Northampton, MA), and IDL (Research Systems, Boulder, CO) were used for data analysis. Patch-clamp pipettes were pulled to an initial resistance of 1.5–2.0 MΩ, and series resistance was corrected by 60–70%. The holding potential was −100 mV in all experiments. In some experiments capacity compensation was partly achieved with a P/4 voltage-clamp protocol; reduced total capacitance was achieved in some experiments by coating the pipette tips with a thin layer of Sylgard 184 (Dow Corning, Lansing, MI).

For intracellular Ca2+ concentration ([Ca2+]i) imaging experiments in HEK-293 cells, a perforated-patch method was used, following the method previously described (7). Amphotericin B (250 μg/ml) was added to the Cs+ pipette solution. Electrical access that enabled adequate voltage control (less than ∼5 MΩ) was obtained 20–30 min after the gigaseal was formed.

Ca2+ Imaging in HEK-293 Cells

In those HEK-293 cells used for [Ca2+]i imaging, fluo 3-AM was used to load the cells with fluo 3, following the method previously described (7). [Ca2+]i imaging was done using an MRC600 confocal microscope (Bio-Rad). Images were analyzed and processed using COMOS (Bio-Rad), IDL5.2 (Research Systems), and CorelDraw (Corel) software.


Two groups (1, 3, 19) observed a TTX-blockable, voltage-gated Ca2+ current in cardiac myocytes and named itI Ca(TTX) to suggest that this current was due to a novel Ca2+ channel protein. The initial discovery ofI Ca(TTX) in 1995 was almost as surprising as the more recent discovery of a second TTX-blockable Ca2+ flux in heart in 1998 (7, 28). The second conductance, however, was shown to reflect Ca2+ permeation through cardiac Na+ channels after PKA activation in normal extracellular Na+ in a process called slip-mode conductance. The similar TTX sensitivity of I Ca(TTX) (1, 3,6, 19, 32) and slip-mode conductance of the cardiac Na+ channel (28) suggested to us that the two processes might be related. However, important differences that favor the novel protein hypothesis have been reported, including differences in Ba2+ permeation and “activation” requirements. Because no group has yet made a molecular identification ofI Ca(TTX), we sought to determine the identity of the protein responsible for I Ca(TTX).

Characteristics of ICa(TTX)

Rat heart cells.

Using Na+-free extracellular solutions, we examinedI Ca(TTX) as shown in Fig.1. With Cs+ as the Na+ replacement, Fig. 1 A shows sample recordings in 2 mM Ca2+ (control) and after addition of the specific Na+ channel blocker TTX (10 μM ). Virtually all of the voltage-gated membrane currents seen on step-depolarizations from −100 mV are blocked by TTX. A typical current-voltage (I-V) plot of the TTX-sensitive component of the currentI Ca(TTX) is shown in Fig. 1 C (filled circles). This membrane current is similar to that reported earlier (1, 3, 6, 12, 19, 24). Because Cs+ is not thought to carry inward current through Ca2+ channels in the presence of Ca2+ but has been shown to permeate Na+ channels, albeit poorly (7, 18), the Cs+ permeability of I Ca(TTX) may provide a clue to its molecular identity. We therefore repeated the experiment shown in Fig. 1 A with a Na+replacement cation that is known not to permeate cardiac Na+ channels (TEA+). Figure 1 B shows sample recordings made under control conditions when TEA+rather than Cs+ was used as the Na+ replacement and after the addition of 10 μM TTX, as well as the TTX-sensitive current. The time course of individual I Ca(TTX)recordings in TEA+ is similar to that seen in Cs+; however, the magnitude ofI Ca(TTX) is much less. Figure 1 Cpresents the voltage dependence of I Ca(TTX) in Cs+ (filled circles) and TEA+ (open circles) and shows that the peak I Ca(TTX) is more than eight times larger when measured in the presence of Cs+than in the presence of TEA+.

Fig. 1.

TTX-sensitive Ca2+ current [I Ca(TTX)] in isolated rat cardiac myocytes.A: sample membrane currents resulting from depolarizations from −100 mV are shown for control conditions (i), after addition of 10 μM TTX (ii), and the difference current (iii). Test potentials illustrated are −75, −70, −65, and −55 mV. The extracellular solution contained Cs+ (as the Na+ replacement cation) and 2 mM Ca2+.B: sample recordings in tetraethylammonium (TEA+; as the Na+ replacement cation) and 2 mM Ca2+ for control conditions (i), in the presence of 10 μM TTX (ii), and the difference current (iii) for depolarizations from −100 mV to −66, −62, −58, and −54 mV. C: current density vs. voltage (I-V) plots corresponding to the TTX-sensitive currents recorded inA and B in 2 mM Ca2+ with either Cs+ (●) or TEA+(○) as the Na+ replacement cation, respectively.

The effect of the Na+ replacement cation onI Ca(TTX) is explored further in Fig.2. The time course of the changes inI Ca(TTX) magnitude when TEA+ is replaced by Cs+ is shown in Fig. 2 A. Sample recordings (left) are shown to illustrate the plot of peakI Ca(TTX) on repeated depolarizations to −50 mV over several minutes (right). This voltage was chosen because I Ca(TTX) is large and the potentially interfering I Ca is near zero. Replacing TEA+ with Cs+ increasesI Ca(TTX) by a factor of 10 in this example, and TTX completely blocks I Ca(TTX), as shown at the 7-min point. Figure 2 B shows composite I-Vrelationships obtained with either Cs+ (open circles) or TEA+ (filled circles) as the Na+ replacement cation. These I-V plots display the characteristic “two-hump” shape revealing I Ca(TTX) at potentials negative to approximately −30 mV andI Ca at more positive potentials. The addition of TTX largely abolished the more negative hump, confirming the identity of this current as I Ca(TTX). The magnitude ofI Ca(TTX) is clearly larger with Cs+rather than with TEA+ in the external solution. To further investigate whether Cs+ permeation contributed toI Ca(TTX), we removed extracellular Ca2+ and replaced it with the impermeant divalent cation Mg2+. The results of these experiments are shown in Fig.2 C. In the presence of Cs+ and 3 mM Mg2+, I Ca(TTX) is present and large. From this finding, we conclude that Cs+ genuinely carries charge through the I Ca(TTX) channel. Furthermore, since Mg2+ does not carry charge throughI Ca, the overlapping current through the L-type Ca2+ channel is removed, as shown in Fig. 2 C. This finding suggests that I Ca(TTX) is not a “Ca2+ channel” in the usual sense because Cs+ competes well with Ca2+ for permeation through the channel responsible for I Ca(TTX). Additional investigation of this issue is addressed below.

Fig. 2.

Both Cs+ and Ca2+ contribute toI Ca(TTX) in isolated myocytes. A, left: sample current traces obtained sequentially from a cell repetitively depolarized from −100 to −50 mV in the presence of 2 mM Ca2+ and either TEA+ (i) or Cs+ (ii) as the Na+ replacement cation. The addition of TTX (10 μM ) blockedI Ca(TTX) (iii). Right: changes in the peak level of I Ca(TTX) are shown as a function of time. Sample recordings were taken as indicated.B: average I-V relationships forI Ca(TTX) current density in Na+-free conditions with TEA+ as the Na+ substitute (●, n = 12), with Cs+ as the Na+ substitute (○, n = 20), and with TEA+ as the Na+ substitute but with TTX (10 μM) present (×, n = 3). Ca2+ (2 mM) was present in all test solutions. Values are means ± SE.C: average I-V plot forI Ca(TTX) current density in Na+-free conditions with Cs+ as the Na+ substitute, with either 2 mM Ca2+ and 1 mM Mg2+(○) or 0 mM Ca2+ and 3 mM Mg2+(●) as divalent cations. In these experiments, currents were recorded under both conditions in the same cells (n = 10). Values are means ± SE.

Heterologous Expression of Cardiac Na+ Channels in HEK-293 Cells

To assess whether I Ca(TTX) is due to Ca2+ flux through a novel TTX-sensitive cardiac Ca2+ channel or through the well-known cardiac Na+ channel, we examined Na+ channels expressed in HEK-293 cells. If I Ca(TTX) arises from the former, then we should have observed no I Ca(TTX)in these cells. Figure 3 Ashows sample recordings of I Na (left) and the resulting I-V plot (right) obtained in HEK-293 cells expressing cardiac Na+ channels and recorded in external solution containing 140 mM Na+ and 8 mM Ca2+. Figure 3, B–D, showsI Ca(TTX) in these cells. In the absence of Na+ channel expression in the HEK-293 cells, neitherI Na nor I Ca(TTX) is observed (data not shown). To simplify the interpretation of experiments on I Ca(TTX) performed in Na+-free conditions, we eliminated Cs+ from the internal and external solutions because it appears to permeateI Ca(TTX), as noted above. Figure 3 Bshows sample recordings of I Ca(TTX)(left) and the I-V plot ofI Ca(TTX) (right) recorded with 140 mM TEA+ and 8 mM Ca2+ in the bath solution. The blockade of I Ca(TTX) by TTX (10 μM) is also shown. Figure 3, C and D, which displays pooled data, shows that I Ca(TTX) is present whether TEA+ (Fig. 3 C) orN-methyl-d-glucamine (NMDG+) (Fig.3 D) is used to replace Na+ and thatI Ca(TTX) is abolished when Mg2+replaces extracellular Ca2+. We thus note that the following four major features of I Ca(TTX)recorded in cardiac myocytes are also seen in HEK-293 cells that express cardiac Na+ channels: 1)I Ca(TTX) is activated at negative potentials,2) I Ca(TTX) is blocked by 10 μM TTX, 3) I Ca(TTX) disappears when Ca2+ is replaced by Mg2+, and 4)I Ca(TTX) magnitude is greater when Cs+ (rather than TEA+ or NMDG+) is the Na+ replacement cation (data not shown; see also Fig.6). From these experiments, we deduce thatI Ca(TTX) can be seen in HEK-293 cells expressing cardiac Na+ channels and that both Ca2+ and Cs+ can carry charge throughI Ca(TTX) but that TEA+, NMDG+, and Mg2+ cannot. As expected forI Ca(TTX), TTX can block the inward membrane current.

Fig. 3.

I Ca(TTX) in transfected HEK-293 cells. Patch-clamp recordings were made using HEK-293 cells either transiently transfected with the α-, β1-, and β2-subunits of the cardiac Na+ channel or stably transfected with the α-subunit. A: sample membrane currents (left) and resulting I-V plot (right) obtained with 140 mM Na+ and 8 mM Ca2+ in the extracellular solution. Sample current recordings for depolarizations were to −50, −45, −40, −35, −30, and −25 mV from a holding potential of −100 mV to revealI Na. B, left: sample currents as described in A recorded from the same cell with TEA+ fully replacing Na+. A current is present with 8 mM Ca2+ (top) that is blocked by 10 μM TTX (bottom). Right: I-V plots.C: composite I-V plot obtained in 140 TEA+, 0 Na+ , 8 mM Ca2+, and 1 mM Mg2+ (●, n = 7); in 140 TEA+, 0 Na+, 0 mM Ca2+, and 9 mM Mg2+ (×, n = 5); and in 140 TEA+, 0 Na+, 8 mM Ca2+, 1 mM Mg2+, and 10 μM TTX (○, n= 4). D: experiments identical to those in Cexcept that N-methyl-d-glucamine (NMDG+) replaced TEA+ in 8 mM Ca2+(●, n = 13); NMDG+ and 9 Mg2+ (×, n = 3); and NMDG+, 8 mM Ca2+, and TTX (○, n = 4). Curves in C and D were normalized to peak currents recorded at −30 mV in the presence of 8 mM Ca2+. See methods for other components of extracellular solutions. Values are means ± SE. *P < 0.05, by one-way ANOVA, followed by the Student-Newman-Keuls test of multiple comparisons.

We next sought to examine other characteristics ofI Ca(TTX) observed in previous studies. Because the measured Ba2+ current throughI Ca(TTX) was an important element in the suggestion that I Ca(TTX) represented membrane current through a novel Ca2+ channel, we attempted to determine whether Ba2+ could carry charge throughI Ca(TTX) in cardiac ventricular myocytes or in HEK-293 cells that express cardiac Na+ channels. Figure4 A shows sample recordings of membrane current when either Ca2+ (top left) or Ba2+ (bottom left) is present and the averageI-V plot for such data from heart cells (right). Figure 4 B shows sample recordings of membrane current and average I-V plots from HEK-293 cells expressing cardiac Na+ channels from experiments similar to those shown in Fig. 4 A. These data show thatI Ca(TTX) is reduced to zero when Ba2+ replaces Ca2+. We conclude from these experiments that Ba2+ cannot permeateI Ca(TTX). These data therefore cast doubt on the novel channel hypothesis.

Fig. 4.

Ba2+ does not permeate I Ca(TTX).A, left: sample current traces in myocytes elicited by depolarizations from −100 to −50 mV in 2 mM Ca2+(top) or 2 mM Ba2+ (bottom).Right: corresponding I-V plots in the presence of 2 mM Ca2+ (●; n = 5) or 2 mM Ba2+ (○; n = 5). B, left: sample current traces elicited in HEK-293 cells stably transfected with the α-subunit of the cardiac Na+channel. Depolarizations from −100 to −25 mV activated an inward current in 140 TEA+ and 8 mM Ca2+(top) but not when extracellular Ca2+ was replaced by 8 mM Ba2+ (bottom).Right: average I-V plots obtained in 8 mM Ca2+ (●, n = 8) or 8 mM Ba2+ (○, n = 8). Each curve was normalized to peak current recorded at −25 mV in TEA+-Tyrode solution plus 8 mM Ca2+. Cs+-free internal and external solutions were present in all cases. Values are means ± SE. *P < 0.05.

A surprising result reported by others, however, at first glance supports the hypothesis that I Ca(TTX) arises from a novel Ca2+ channel protein. This is the observation that I Ca(TTX) is blocked by mibefradil, a known Ca2+ channel blocker (12). However, for the novel protein hypothesis to be supported by this finding, mibefradil must be shown to be a relatively pure blocker of Ca2+channels, and I Na should not be affected significantly by mibefradil. We therefore examined the actions of mibefradil in our heterologous expression system. Figure5 A shows samples ofI Na measured on depolarization from −100 to −30 mV at different concentrations of mibefradil. The I-Vrelationships at different mibefradil concentrations are shown in Fig.5 B, and the dose-response curve is shown in Fig.5 C. Half-block was achieved at 0.45 μM mibefradil. Figure5 D shows that, similar to I Na,I Ca(TTX) observed in HEK-293 cells expressing cardiac Na+ channels is fully blocked by 10 μM mibefradil. Importantly, this is the concentration of mibefradil used to block I Ca(TTX) in the experiments of Heubach et al. (12). Thus we conclude that all of the effects of mibefradil on I Ca(TTX) can be explained by the actions of mibefradil on cardiac Na+ channels.

Fig. 5.

Mibefradil blocks I Na andI Ca(TTX). Whole cell recordings were performed in HEK-293 cells stably transfected with the α-subunit of the cardiac Na+ channel. A: sample recordings ofI Na elicited by step depolarizations from −100 to −30 mV in solutions containing 0 (control), 0.1, 1, and 10 μM mibefradil. Progressive block of I Na is observed. B: I-V plots corresponding to the sample records shown in A. C: pooled data showing the concentration dependence of I Na block by mibefradil (n = 7–8 at each concentration). The curve was fit by eye and suggests 50% block at ∼0.5 μM.D: sample traces elicited upon depolarizations from −100 to −20 mV in Na+-free extracellular solution containing Cs+ and 8 mM Ca2+ show that 10 μM mibefradil also blocks I Ca(TTX). On average, mibefradil (10 μM ) reduced the amplitude of I Ca(TTX) to 8 ± 5.3% of control (n = 8). All recordings were obtained with Cs+-containing pipette solution.

The absence of outward membrane current whenI Ca(TTX) is measured in heart cells could be explained by many mechanisms. Clearly, if a novel Ca2+channel were responsible for I Ca(TTX), then the absence of significant amounts of charge carrier (i.e., Ca2+) in the intracellular compartment could account for this observation (1, 3, 19), as is the case forI Ca measured in heart cells. If, however, the cardiac Na+ channel was responsible forI Ca(TTX) as suggested by our results, then outward current should be readily observed when a permeant charge carrier is present. To test this hypothesis, we measuredI Ca(TTX) with equal concentrations of Cs+ (150 mM) inside and outside the cell and with 8 mM extracellular Ca2+. Intracellular TEA+ was removed because it can block outward current through the Na+ channel. Figure 6 shows the inward current carried by both Cs+ and Ca2+and the outward current carried by Cs+. The presence of an outward current is important, as is the value of the reversal potential of I Ca(TTX) . The reversal potential for Cs+ is 0 mV, and under these conditions the reversal potential (E rev) ofI Ca(TTX) is 8 mV. If only Cs+ were to permeate the Na+ channel, thenE rev for I Ca(TTX) would be 0 mV. The observed E rev of +8 mV suggests that the ratio of Ca2+ to Cs+ permeability (P Ca/P Cs) is significant and has a value of 4.1 (5, 20).

Fig. 6.

I Ca(TTX) can be both inward and outward.I Ca(TTX) was recorded in HEK-293 cells stably transfected with the α-subunit of the cardiac Na+ channel in Na+-free extracellular solutions. Both internal and external solutions contained 150 mM Cs+, and the external solution additionally contained 8 mM Ca2+. A: sample current traces recorded upon depolarizations from −100 mV to potentials between −80 and +70 mV in 10-mV steps. Depending on the test potential, either inward or outward current was produced.B: average I-V plot produced in these experiments (n = 13). The reversal potential is approximately +8 mV, indicating that both Cs+ and Ca2+ permeate the channel. This reversal potential suggests that in Na+-free conditions, the ratio of Ca2+ to Cs+ permeability (P Ca/P Cs) = 4.1 according to a solution of the Goldman-Hodgkin-Katz equation (see text). The curve was normalized to the peak current. Values are means ± SE.

Ca2+ Influx via ICa(TTX)

The measured permeability of Ca2+ throughI Ca(TTX) (deduced from Fig. 6 and supported by Figs. 1-4) should be confirmed by measuring the accumulation of intracellular Ca2+ under conditions in which I Ca(TTX) is observed. Using HEK-293 cells to express the cardiac Na+ channel proteins, we examined how changes in [Ca2+]i were related to I Ca(TTX). Figure7 A shows sample images of two cells. The top pair of images (Fig. 7 A, i) shows a control cell that was transfected but did not expressI Na or I Ca(TTX) to any significant level; the bottom pair of images (Fig. 7 A, ii) shows a cell expressing 4.5 nA I Na and significant levels of I Ca(TTX) (26.2 pA). The images in Fig. 7 A, left, show that [Ca2+]i is at control levels in the absence of stimulation; the images in Fig. 7A, right, show how [Ca2+]i changed in response to a series of 200 depolarizations from −100 to −30 mV at 20 Hz. No significant change in [Ca2+]i was observed in the control cell, but a 57 nM increase in [Ca2+]i was observed in the cell with I Ca(TTX). Figure7 B shows how the increase in [Ca2+]i varies with the number of depolarizations. Consistent with the overall finding that the Ca2+ flux depends on Ca2+ movement through Na+ channels exhibiting I Ca(TTX), the magnitude of the I Ca(TTX) that has been linked to increases in [Ca2+]i is identified with I Na in the presence of Na+ in the same cells (Fig. 7C). Of the 28 cells in which Ca2+accumulation was examined, 17 had very small I Na(<3 nA/cell) in Na+-containing extracellular solution and had no measurable I Ca(TTX). Of the 11 cells withI Na > 3 nA, 4 had measurableI Ca(TTX), and all of these showed increases in [Ca2+]i when examined with the use of our imaging methods (Fig. 7D). We conclude from these findings thatI Ca(TTX) is associated with genuine Ca2+ influx. The measured I Ca(TTX)and related increases in [Ca2+]i are dependent on the level of expression of the cardiac Na+channels in the HEK-293 cells.

Fig. 7.

Measured Ca2+ flux throughI Ca(TTX). HEK-293 cells transfected with the α-, β1-, β2-subunits of the cardiac Na+ channel were voltage clamped in whole cell mode using a perforated-patch method (see methods). Cells were loaded with the Ca2+ indicator fluo 3-AM. A: confocal images before (left) and after (right) 200 depolarizations from −100 to −30 mV. An insignificant increase in fluorescence (F0 = 1 to F/F0 = 1.1) was observed in a cell that expressed a very smallI Na (i), whereas a substantial increase in fluorescence (and hence [Ca2+]i) was seen in a cell that had significant I Na and a clear I Ca(TTX) (ii). B: dependence of F/F0 on the number of depolarizations in HEK-293 cells that expressed I Ca(TTX). *P < 0.05. C: relationship betweenI Ca(TTX) in Na+-free solutions andI Na in Na+-containing solutions in cells that showed an increase in [Ca2+]i. The black line was obtained by linear regression analysis, and the red lines mark the 95% confidence interval. D: of 28 cells examined, 17 had small I Na, and none of these cells had a measurable I Ca(TTX). Of the 28 cells, 11 had large I Na (>3 nA), but of these 11, only 4 had measurable I Ca(TTX). Of those 4 cells with measurable I Ca(TTX), all 4 showed an increase in [Ca2+]i with repeated depolarizations.

Inactivation of ICa(TTX)

The kinetics of I Ca(TTX) inactivation were used by Aggarwal et al. (1) and Balke et al. (3) to distinguish I Ca(TTX) fromI Na. For that reason we compared the inactivation kinetics of I Ca(TTX) to those ofI Na in both rat ventricular myocytes and HEK-293 cells expressing the cardiac Na+ channel. Figure8 shows sample inward currents obtained in myocytes on step depolarizations from −100 to −60 and −50 mV either in the presence of 10 mM Na+ and 2 mM Ca2+ (Fig. 8 A) or in Na+-free conditions with TEA+ and 2 mM Ca2+ in the bath (Fig. 8 B). Currents inactivated more slowly when Ca2+ alone carried the charge throughI Ca(TTX) than when Na+ carried the charge through I Na, as shown by the scaled and superimposed sample recordings shown in Fig. 8 C (recordings at −55 mV). The exponential inactivation time constants (τ) forI Na and I Ca(TTX) are plotted in Fig. 8 D. Significant slowing in inactivation ofI Ca(TTX) is seen at all voltages measured over the range from −60 to −45 mV. Similar experiments were carried out in HEK-293 cells expressing cardiac Na+ channels. The measurement of I Na was carried out in 2 mM Na+ and 0 mM Ca2+ or in 8 mM Na+and 0 mM Ca2+, while I Ca(TTX) was carried out in 0 mM Na+ and 8 mM Ca2+. In both cases, TEA+ was used as the Na+ replacement. Sample recordings of I Na andI Ca(TTX) obtained on depolarization from −100 to −40 mV (top) and −30 mV (bottom) are displayed in Fig. 8, E and F, respectively. Scaled and superimposed recordings obtained at −35 mV are shown in Fig. 8 G. τ for I Na andI Ca(TTX) are plotted in Fig. 8 H. Compared with I Na, significant slowing in inactivation of I Ca(TTX) is seen at all potentials negative to −20 mV. Although there are small quantitative differences between rat heart cells and HEK-293 cells expressing Na+ channels, similar kinetic differences betweenI Na and I Ca(TTX) are observed in the two preparations. Additionally, the inactivation rates for I Na range from ∼5 to ∼1 ms from −60 to −10 mV, findings in agreement with those reported earlier forI Na (23). While these differences in inactivation kinetics could be indicative of two distinct channel proteins (1), they could reflect instead differences in Na+ channel behavior due to the permeant cation(s) (18, 33-35) (see discussion).

Fig. 8.

Inactivation kinetics of I Ca(TTX).A: sample current traces recorded in myocytes on depolarization from −100 to −60 (top) and −50 mV (bottom) in the presence of 10 mM Na+ and 2 mM Ca2+ with 130 mM TEA+ as the monovalent cation.B: sample current traces recorded in myocytes in Na+-free conditions but in the presence of 140 mM TEA+ and 2 mM Ca2+. Currents recorded in these conditions inactivated more slowly than those in the presence of Na+. C: declining phases of currents recorded at −55 mV from the same cell shown in A and Bplotted on a normalized scale to facilitate comparison of the inactivation kinetics. D: pooled time constants of inactivation (τ) for currents recorded in myocytes in TEA with 10 mM Na+ (○, n = 3) or with 2 mM Ca2+ (●, n = 4). Currents recorded in 140 mM TEA+ and 2 mM Ca2+ were fitted with two time constants; the faster time constant is plotted.E: sample current traces recorded in the presence of 2 mM Na+ in HEK-293 cells transfected with the cardiac Na+ channel α-subunit. F: current traces recorded in the same cell in TEA+ with 8 mM Ca2+ inactivated more slowly than those in the presence of Na+. G: declining phases of currents recorded at −35 mV from the cell shown in E and F plotted on a normalized scale. H: pooled time constants of inactivation for I Na (○, n = 3; recorded in 2 or 8 mM Na+) andI Ca(TTX) (●, n = 3) in transfected HEK-293 cells. In HEK-293 cells, as in myocytes, TTX-blockable current inactivated more slowly in TEA+ and Ca2+ than in the presence of Na+.


By examining TTX-sensitive current in Na+-free conditions in rat heart cells, we have been able to confirm the existence of a small but significant Ca2+-dependent current. Furthermore, by showing that an identical inward current appears in HEK-293 cells only when the cardiac Na+ channel is expressed, we have been able to identify the Na+ channel as the protein responsible for this Ca2+ flux. While Ca2+ contributes to this current, calledI Ca(TTX) up to now, Cs+ is also able to permeate. Together, our findings suggest four reasons that the membrane current called I Ca(TTX) should not be called a Ca2+ current. First, it is the Na+channel that is responsible for the current. Second, at least one other ion (i.e., Cs+) readily permeates the channel under diverse experimental conditions. Third, Ba2+ does not contribute toI Ca(TTX), while it permeates established Ca2+ channels as well as or better than Ca2+itself. Fourth, the current as such arises under special conditions, namely, Na+-free (“null”) conditions. We therefore prefer the term I Na(null) because it properly reflects these four important factors. While the need for Na+-free conditions suggests thatI Na(null) may not be physiologically important, the existence of I Na(null) and its properties help to broaden our understanding of the cardiac Na+channel and its biophysical properties.

Implications of INa(null)

Ca2+ influx via INa(null)and Ca2+ influx via slip-mode conductance.

The measurement of Ca2+ influx through cardiac Na+ channels is unexpected under any condition because Ca2+ has been reported to block Na+ channels (8, 25, 27). Nevertheless, a strong case has been made for such Ca2+ influx in two very distinct modes of behavior of the cardiac Na+ channel: 1) in the absence of extracellular Na+, Ca2+ permeates the channel via I Na(null), and 2) in the presence of extracellular Na+ and after activation of PKA, Ca2+ permeates the channel via slip-mode conductance. The finding that Ca2+ can permeate Na+ channels in these two very different circumstances does suggest that Ca2+ permeability is a property of the cardiac Na+ channel, even if it takes special circumstances to recruit that property. It is hard to directly compare the amount of Ca2+ that enters under two very different conditions, but it is clear that much more Ca2+ can permeate via slip-mode conductance than via I Na(null), as assessed by the measured increases in [Ca2+]i. When slip-mode conductance was activated, 100 depolarizations were necessary to produce an increase in [Ca2+]i in HEK-293 cells of 250 nM (7), whereas here [Ca2+]i only increased by 60 nM after 200 depolarizations to activate I Na(null). Work to date indicates that there is little, if any, Ca2+permeation via I Na (i.e., when extracellular Na+ is present and slip-mode conductance has not been activated). For the “Na+ channel null current”I Na(null) to exist, Na+-free conditions appear to be needed. Interestingly, slip-mode conductance occurs in normal Na+ (7, 28) but requires activation (e.g., PKA) and is blocked by very low Na+(i.e., 0.5 mM). Thus the Na+-free mode of Ca2+permeation appears to be I Na(null), while Ca2+ permeation in normal extracellular Na+depends on slip-mode conductance. In the absence of those two conditions, there is essentially no Ca2+ permeation via the cardiac Na+ channel.

Mechanisms of Ca2+ permeation through the Na+ channel to produce INa(null).

While slip-mode conductance requires the heterotrimeric cardiac Na+ channel (α, β1 and β2) or the α-subunit and at least one of the two β-subunits (see Refs.7 and 28), I Na(null) can be observed whenever the α-subunit is expressed and Na+-free conditions are established. In contrast, slip-mode conductance appears to need Na+ in the extracellular solution along with PKA activation. I Na(null) is readily observed in the absence of explicitly activated PKA, as shown in Figs. 1-8. Given the earlier reports (21, 22, 26) on changes inI Na after PKA activation, however, clarification of the actions of PKA activation on I Na(null)seemed warranted.

Role of PKA activation in INa(null).

Experiments examining PKA activation andI Na(null) in HEK-293 cells expressing the cardiac Na+ channel were carried out, and the results are shown in Fig. 9. With PKA inhibitory peptide (PKI) in the pipette (Fig. 9 A), a robustI Na(null) was observed, and forskolin did not alter the magnitude of I Na(null). We conclude from this experiment that PKA activation is not necessary forI Na(null) to be seen and that no independent action of forskolin (an activator of PKA) materially affectsI Na(null). In addition, activation of PKA by forskolin in the absence of PKI in the pipette failed to alter the magnitude of I Na(null) (Fig. 9 B). This finding is surprising in light of the results of Balke et al. (3), who showed that isoproterenol increased the magnitude of I Na(null) in heart cells. Experiments in myocytes demonstrated the integrity of our reagents, because forskolin altered the magnitude and voltage dependence of the L-type Ca2+ current only when PKI was absent from the pipette (data not shown). We then hypothesized that PKA activation affectsI Na(null) in a manner that depends strongly on the holding potential, as suggested for I Na by Ono et al. (26). Representative results displayed in Fig.9 C show that forskolin increases the magnitude ofI Na(null) when the holding potential is −120 mV. Additionally, in cardiac myocytes, isoproterenol can affectI Na via G protein signaling that is independent of PKA activation (21), and this may account for some of the observations of Balke et al. (3) in heart cells.

Fig. 9.

Effects of protein kinase A (PKA)-dependent phosphorylation on I Na(null). A: composite I-V plots obtained with PKA inhibitory peptide (PKI; 100 μM) in the pipette show thatI Na(null) is observed in Na+-free conditions with TEA+ and 8 mM Ca2+ in the bath in HEK-293 cells expressing cardiac Na+ channels (●, n = 4) and is not altered by the addition of forskolin (○, n = 4).B: plots identical to those in A without PKI in the pipette show that, with a holding potential of −100 mV, forskolin (50 μM) does not affect the magnitude ofI Na(null) (n = 8). The curves were normalized to the peak current recorded in TEA+ with 8 mM Ca2+. Values are means ± SE. C: repetitive depolarizations to −30 mV were applied to determine the time course of the magnitude of I Na(null) after forskolin was added. An increase in the amplitude ofI Na(null) was observed when depolarizations were applied from a holding potential (V H) of −120 mV but not when VH was −100 mV.

An additional difference between I Na(null) and slip-mode conductance is that Ca2+ flux through a slip-mode mechanism requires the presence of extracellular Na+(7), whereas the addition of a small amount (e.g., 50 μM) of extracellular Na+ appears to blockI Na(null) (6), suggesting that Ca2+ permeation via I Na(null) is in competition with Na+ permeation. In summary, then, the data indicate that the mechanism(s) that allows observation ofI Na(null) is different from those responsible for activating Ca2+ permeation via slip-mode conductance. In both cases, however, Ca2+ is permeating the cardiac Na+ channel and Ba2+ cannot substitute for Ca2+. We infer that when Na+ occupies a position in the channel pore that is required for permeation, Ca2+ cannot permeate via I Na(null). How well does Ca2+ interact with this critical binding site within the channel? While it is hard to directly compare Ca2+ flux through the Na+ channel inI Na(null) with Na+ flux through the Na+ channel via I Na, any measure of relative flux suggests that permeation is quite small. According to data from all sources, the maximum current carried by Ca2+via I Na(null) in heart cells with 2 mM Ca2+ when TEA+ replaces Na+ is ∼100 pA, while peak I Na is ∼100 nA. This finding suggests that Na+ can carry ∼1,000 times more charge through I Na than Ca2+ can carry through I Na(null). So far, no compelling results on P Ca/P Na during very low [Na+] (i.e., 0 < [Na+] < 0.5 mM) have been obtained. However, during moderately low [Na+] (i.e., 10 ≤ [Na+] ≤ 20 mM) and no activation of slip-mode conductance,P Ca/P Na was indistinguishable from 0 in control experiments in HEK-293 cells expressing cardiac Na+ channels (7). Thus essentially no Ca2+ would appear to permeate Na+ channels under normal conditions (in the presence of extracellular Na+ and in the absence of PKA activation), when the order of selectivity is Na+ = Li+ > K+ > Cs+ > TEA+ = NMDG+ = Ca2+= Ba2+ (7, 13, 18). The known ability of Ca2+ to block Na+ channels may simply reflect the ability of Ca2+ to interact with a putative permeation binding site and compete with Na+ for this site under these conditions.

Permeation Plasticity

Role of permeant ions in channel behavior and permeation plasticity.

Permeant ions have been known for some time to have significant effects on channel permeation (13). For example, the L-type Ca2+ channel can readily conduct Na+ and other monovalent cations in the absence of Ca2+ (2), and the relative permeabilities of this channel to Ca2+ and Ba2+ depend on the ionic concentrations (anomalous mole fraction effect) (11). Similar behaviors are observed in certain K+ channels (17) . That permeation depends on the concentration and identity of the permeant ion is thus a feature broadly seen. What has only been appreciated more recently is that the ionic milieu can also have effects on channel gating, such as the decrease in Na+ channel open probability observed upon removal of extracellular Na+ (34, 35) and the slowing of C-type inactivation by K+ occupancy of the pore in Shaker K+ channels (4, 16).I Na(null) displays aspects of both phenomena: the removal of intracellular and extracellular Na+ changes the permeability properties as well as the inactivation kinetics of the Na+ channel.

In summary, the primary conclusion from this study and other recent reports is that channel selectivity and kinetics are not rigid features of the assembled Na+ channel proteins but are remarkably dynamic, flexible, and subtle. There are now many examples of channels that change selectivity, including the cardiac Na+ channel (current study and Refs. 7 and 28), the ShakerK+ channel (30, 31, 36), and ligand-gated channels (14, 15). Channel properties can be influenced by phosphorylation or by interaction with permeant ions, even at sites apparently distant from the “selectivity” filter or channel gates. These features may depend critically on protein-protein interactions (e.g., among channel subunits or cytoskeleton proteins or among extracellular matrix proteins or other proteins). Selectivity can be dynamically modified by both physiological interventions (e.g., PKA-dependent phosphorylation in slip-mode conductance) and more severe interventions [e.g., Na+ removal forI Na(null) investigations].


We thank C. A. Frederick and N. Agarwal for laboratory support and E. Moczydlowski, A. Goldin, L. Isom, A. George, M. Cahalan, L. F. Santana, and H. Hartmann for reagents, clones, and advice.


  • * S. Guatimosim, E. A. Sobie, and J. dos Santos Cruz contributed equally to this work.

  • This work was supported by a grant from the National Heart, Lung, and Blood Institute, by DRIF and Medical Biotechnology Center special accounts funding from the University of Maryland, Baltimore, by the University of Maryland Biotechnology Institute, and by the National Institutes of Health Muscle Training Program at the University of Maryland, Baltimore.

  • Address for reprint requests and other correspondence: W. J. Lederer, Medical Biotechnology Center, Univ. of Maryland Biotechnology Institute, 725 W. Lombard St., Baltimore, MD 21201 (E-mail:lederer{at}umbi.umd.edu).

  • 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.


View Abstract