Gating of skeletal and cardiac muscle sodium channels in mammalian cells

Abstract

1. Sodium channel ionic current (I_Na) and gating current (I_g) were compared for rat skeletal (rSkM1) and human heart Na⁺ channels (hH1a) heterologously expressed in cultured mammalian cells at ∼13 °C before and after modification by site-3 toxins (Anthopleurin A and Anthopleurin B).

2. For hH1a Na⁺ channels there was a concordance between the half-points (V_½) of the peak conductance-voltage (G–V) relationship and the gating charge-voltage (Q–V) relationship with no significant difference in half-points. In contrast, the half-point of the Q–V relationship for rSkM1 Na⁺ channels was shifted to more negative potentials compared with its G–V relationship with a significant difference in the half-points of −8 mV.

3. Site-3 toxins slowed the decay of I_Na in response to step depolarizations for both rSkM1 and hH1a Na⁺ channels. The half-point of the G–V relationship in rSkM1 Na⁺ channels was shifted by −8.0 mV while toxin modification of hH1a Na⁺ channels produced a smaller hyperpolarizing shift of the V_½ by −3.7 mV.

4. Site-3 toxins reduced maximal gating charge (Q_max) by 33% in rSkM1 and by 31% in hH1a, but produced only minor changes in the half-points and slope factors of their Q–V relationships. In contrast to measurements in control solutions, after modification by site-3 toxin the half-points of the G–V and the Q–V relationships for rSkM1 Na⁺ channels demonstrated a concordance similar to that for hH1a.

5. Q_maxvs. G_max for rSkM1 and hH1a Na⁺ channels exhibited linear relationships with almost identical slopes, as would be expected if the number of electronic charges (e⁻) per channel was comparable.

6. We conclude that the faster kinetics in rSkM1 channels compared with hH1a channels may arise from inherently faster rate transitions in skeletal muscle Na⁺ channels, and not from major differences in the voltage dependence of the channel transitions.

Voltage-gated sodium (Na⁺) channels form one of the main families of ion channels, and are present in many different species of animals as well as in many different tissues within the same animal. Although Na⁺ channel isoforms give rise to rapid and large inward ion currents, they are products of different genes that allow for specialized characteristics matched to their specific functions. Cardiac Na⁺ channels, which are responsible for the rapid conduction of the action potential throughout the heart muscle, are also found in immature and denervated skeletal muscle (White et al. 1991). They have different functional properties compared with the dominant Na⁺ channel isoform expressed in innervated skeletal muscle. For example, channel kinetics are more rapid in skeletal muscle compared with heart while the heart Na⁺ channel has a higher affinity for lidocaine (lignocaine), an important antiarrhythmic medication (Nuss et al. 1995; Wang et al. 1996a, b).

To help understand the structural basis of the functional differences between skeletal and cardiac Na⁺ channels, we recorded ionic and gating currents under similar experimental conditions from each of the two Na⁺ channel isoforms expressed in mammalian cells. In addition, we studied the effects of site-3 toxins (Anthopleurin A and Anthopleurin B) on the two Na⁺ channel isoforms. Both single channel studies (Kirsch et al. 1989; El-Sherif et al. 1992) and whole cell studies (Hanck & Sheets, 1995) with site-3 toxins have shown that site-3 toxins slow decay of I_Na by inhibition of the open-to-inactivated (O ↔ I)transition with little effect on channel activation. Accompanying the modification of the (O ↔ I)kinetic transition, site-3 toxins caused a reduction in the maximal gating charge (Q_max) by 33 % which was associated with inhibition of gating charge movement specifically associated with, or tightly coupled to the (O ↔ I)kinetic transition (Sheets & Hanck, 1995). To investigate whether rSkM1 Na⁺ channels had a similar response to modification by site-3 toxins, studies were undertaken by expression of the rat skeletal muscle (rSkM1) Na⁺ channel in fused tsA201 cells (a variant of HEK 293 cells that had undergone cell fusion with polyethylene glycol). Parallel experiments were done on fused cells expressing hH1a in order to (1) allow for direct comparison with rSkM1 Na⁺ channels, and (2) determine whether heterologously expressed cardiac Na⁺ channels had similar responses to site-3 toxin modification to native Na⁺ channels in cardiac muscle.

METHODS

Cell preparation

Human embryonic kidney cells HEK 293, or SV40-transformed HEK 293 (tsA201) cells, were fused together forming large mammalian cells using polyethylene glycol as previously described (Sheets et al. 1996). After fusion, the cells were placed in cell culture for several days in order to remodel before they were transiently transfected using standard transfection methods of calcium phosphate (Gibco, Grand Island, NY, USA) or lipofectamine (Gibco). The hH1a channel was kindly provided by H. Hartmann and A. Brown (Hartmann et al. 1994), and the rSkM1 channel (Trimmer et al. 1989) was provided by G. Mandel. Both Na⁺ channel isoforms had been subcloned directionally into the mammalian expression vector pRcCMV (Invitrogen, Carlsbad, CA, USA). The rat β1 subunit (Satin et al. 1994) was also subcloned directionally into pRcCMV. For all rSkM1 and hH1a studies, both α and β subunits were cotransfected. Three to six days after transfection fused cells were trypsinized and studied electrophysiologically.

Recording technique, solutions, and experimental protocols

Recordings were made using a large bore, double-barrelled glass suction pipette for both voltage clamp and internal perfusion as previously described (Sheets et al. 1996). I_Na was measured with a virtual ground amplifier (Burr-Brown OPA-101) with a 2.5 MΩ or 5 MΩ feedback resistor. Voltage protocols were imposed from a 12-bit (or 16-bit) DA converter (Masscomp 5450, Concurrent Computer, Tinton Falls, NJ, USA) over a 30/1 voltage divider. Data were filtered by the inherent response of the voltage-clamp circuit (corner frequency near 125 kHz) and recorded with a 16-bit AD converter on a Masscomp 5450 at 300 kHz. A fraction of the current was fed back to compensate for series resistance.

A cell was placed in the aperture of the pipette and transferred to one of four experimental chambers. After a high resistance seal had formed, the cell membrane inside the pipette was disrupted with a manipulator-controlled platinum wire. Voltage control was assessed by evaluating the time course of the capacitive current, and the steepness of the negative slope region of the peak current-voltage relationship based upon previously published criteria for cardiac I_Na (Hanck & Sheets, 1992). The holding membrane potential (V_h) was -150 mV which provided full Na⁺ channel availability for both isoforms. To maximize the signal-to-noise ratio, gating current (I_g) protocols contained up to four repetitions at each test voltage that were one-quarter of a 60 Hz cycle out of phase.

The control extracellular solution for I_Na measurements contained (mM): 15 Na⁺, 185 tetramethylammonium (TMA⁺), 2 Ca²⁺, 200 Mes⁻ and 10 Hepes (pH 7.2). Intracellular solution contained (mM): 200 TMA⁺, 75 F⁻, 125 Mes⁻, 10 EGTA and 10 Hepes (pH 7.2). For measurements of gating currents Na⁺ was replaced with TMA⁺ and saxitoxin (Stx, Calbiochem Corp., San Diego, CA, USA) was added to the extracellular solution. For hH1a Na⁺ studies saxitoxin was used at a concentration of 2.5 μm. To conserve toxin by taking advantage of the higher affinity of rSkM1 Na⁺ channels for Stx, the extracellular solution contained 200 nM Stx in rSkM1 studies. The site-3 toxin used to modify hH1a channels was Anthopleurin A (Ap-A) toxin (Sigma Chemical Co, St. Louis, MO, USA) at a concentration of 1 μm, which is at least three orders of magnitude greater than the K_D (Hanck & Sheets, 1995). rSkM1 Na⁺ channels have an order of magnitude lower affinity for Ap-A toxin compared with Anthopleurin B (Ap-B) toxin (Benzinger et al. 1997), consequently the high affinity toxin isoform Ap-B (kindly supplied by K. Blumenthal, University of Cincinnati, OH, USA) was used for rSkM1 studies at a concentration of 200 nM which is more than two orders of magnitude greater than its K_D.

Changes in bath solution were achieved by placing the pipette with the cell adjacent to the inlet of one of four parallel experimental chambers containing the experimental solution. Modification of I_Na by site-3 toxins was achieved by waiting at least 3 min in the presence of toxin while maintaining a V_h of -150 mV. To wash site-3 toxins from Na⁺ channels, the membrane potential was depolarized to -10 mV for ∼8 min in control solution. Temperature was controlled using a Sensortek (Physiotemp Instruments, Inc., Clifton, NJ, USA) TS-4 thermoelectric stage mounted beneath the bath chambers which typically varied less than 0.5°C during an experiment. Cells were typically studied between 12 and 13°C.

Data analysis

Data were capacity corrected using 4 to 16 scaled current responses to steps between -150 and -190 mV. Leak resistance (R_L) was calculated as the reciprocal of the linear conductance between -190 and -110 mV, and the mean was 72 MΩ (n= 27 cells, range 18-296 MΩ). When the large circumference of the pipette is taken into account, these seal resistances are equivalent to gigaohm seals with conventional whole cell pipettes. Cell capacitance was measured from the integral of the current responses to voltage steps between -150 and -190 mV, and the mean ±s.d. was 589 ± 278 pF (n= 27).

Peak I_Na was taken as the mean of four data samples clustered around the maximal value of data digitally filtered at 5 kHz and leak corrected by the amount of the time-independent linear leak extrapolated from current measurements elicited between -110 and -190 mV. For gating charge measurements data were leak corrected by subtracting the mean of the current recordings typically taken between 8 and 10 ms for test potentials (V_t) < 0 mV and between 6 and 8 ms for V_t≥ 0 mV. Running integrals exhibited a stable plateau except occasionally at the most positive potentials when a small outward ionic current, that developed after a delay of several milliseconds, was present. To determine time constants of I_g relaxations, the decay of gating current tracings was fitted by a sum of exponentials with DISCRETE (Provencher, 1976), a program that provided a modified F statistic in order to evaluate the number of exponential components that best described the data.

Data were analysed and plotted on a SUN Sparcstation using SAS (Statistical Analysis System, Cary, NC, USA). Unless otherwise specified all summary statistics are expressed as means ± one standard deviation (s.d.). Regression parameters are reported as the estimate and the standard error of the estimate (s.e.e.).

RESULTS

Ionic currents for hH1a and rSkM1 in fused cells

It has previously been recognized that cardiac Na⁺ currents have slower channel kinetics than the dominant Na⁺ currents in skeletal muscle and neurons. Figure 1A illustrates this phenotypic difference for rat skeletal muscle and human heart Na⁺ channels expressed heterologously in fused cells studied at ∼12°C. Inspection shows that the onset and decay of I_Na was faster for hSkM1 than for hH1a channels, and as expected, the time to peak I_Na for rSkM1 I_Na was shorter than that for hH1a I_Na (Fig. 1B). Cardiac I_Na activated at more negative potentials and reached a maximal inward value at a more negative potential (-40 mV) compared with rSkM1 I_Na (-25 mV) as shown in Fig. 1C. Boltzmann equation fits to normalized peak conductance- voltage (G-V) relationships (Fig. 1D) demonstrate that the half-point (V_½) for cardiac I_Na occurred at a significantly more negative potential (by -16 mV) compared with the V_½ of rSkM1 I_Na, although the slope factors were similar for the two isoforms.

Figure 1. rSkM1 (•) and hH1a (○) I_Na expressed in fused cells under control conditions.

A shows a family of I_Na responses to step depolarizations between -70 and +30 mV for rSkM1 Na⁺ channels (top) and hH1a Na⁺ channels (bottom). The I_Na responses were capacity corrected but not leak corrected, and digitally filtered at 5 kHz. (Cells T4.01 and P2.01.) B shows the time to peak I_Na (mean ±s.e.m.) for rSkM1 (14 cells) and hH1a (10 cells) isoforms. Values were significantly different at P < 0.05 except at test potentials of -65, -55 and -45 mV. C shows the mean peak I-V relationships for 6 fused cells transfected with rSkM1 and for 5 cells transfected with hH1a. The lines connect the mean values. D shows the peak G-V relationships for the same cells as in C with the values for each cell normalized to the G_max value of each cell. The lines represent the best fits to the grouped data with parameters in the table below to a transform of a Boltzmann distribution:

(1)

where I_Na is the peak current in the depolarizing step and V_t is the test potential. The parameters estimated by the fit were V_rev, the reversal potential; G_max, the maximum peak conductance; V_½, the half-point of the relationship; and s, the slope factor (mV).

Na+ channel isoform	Gmax (μS)	V1/2 (mV)	Slope factor (mV)
rSkM1 (n= 6 cells)	0.61 ± 0.23	−40 ± 4	−6.5 ± 0.6
hH1a (n= 5 cells)	0.56 ± 0.30	−56 ± 5*	−6.6 ± 0.7

^*

Significantly different at P < 0.05.

We have previously shown that kinetic indices of cardiac I_Na in cardiac Purkinje cells (e.g. the half-point of peak conductance-voltage relationship) shift to more negative potentials as a linear function of time (Hanck & Sheets, 1992) typically at a rate of ∼0.5 mV min⁻¹. I_Na kinetics also shift in fused cell preparations, although at a much reduced rate. Figure 2 shows the shift of the half-point (V_½) of the peak conductance-voltage (G-V) relationship for both Na⁺ channel isoforms. Under our experimental conditions, the shift of V_½ for the hH1a isoform was -0.13 ± 0.05 mV min⁻¹, n= 9 cells, and for the rSkM1 isoform it was -0.15 ± 0.07 mV min⁻¹, n= 10 cells. Similar shifts in channel kinetics occurred for the V_½ of voltage-dependent Na⁺ channel availability (e.g. steady-state inactivation relationship); it was -0.15 ± 0.06 mV min⁻¹ (n= 9 cells) for hH1a and -0.12 ± 0.05 mV min⁻¹ (n= 10 cells) for rSkM1 (data not shown). Since the rate of shift of I_Na kinetics was much lower in fused cell preparations compared with cardiac cells, correction for shifts in kinetics was not performed except where specifically stated.

Figure 2. Time dependent shift of I_Na kinetics.

Time dependent shift of the half-point (V_½) of the G-V relationship for 10 cells transfected with rSkM1 (A) and 9 cells transfected with hH1a (B). The line in each panel represents the mean of the individual linear fits to the V_½ of each cell. The mean slope for rSkM1 was -0.15 ± 0.07 mV min⁻¹ and for hH1a it was -0.13 ± 0.05 mV min⁻¹.

Gating currents for hH1a and rSkM1

We would expect the more rapid kinetics that are evident in I_Na traces of rSkM1 Na⁺ channels to be reflected in the time course of gating currents (I_g) such that the decay time constants of I_g relaxations would be faster for rSkM1 Na⁺ channels compared with hH1a. Gating currents were recorded at ∼12°C after addition of Stx to the extracellular solution and all Na⁺_o replaced by TMA⁺_o. Figure 3 shows capacity and leak corrected I_g traces and their corresponding integrals for rSkM1 channels (Fig. 3A) and for hH1a channels (Fig. 3B). To compare the time course of I_g relaxations in response to step depolarizations, I_g relaxations were trimmed until the decay phase was clearly identified, and then fitted with up to two exponential terms. For rSkM1 I_g relaxations, a single exponential fitted best for 91 % of the time, while for I_g relaxations from hH1a, a single exponential fitted better for 88 % of the time. Figure 3C shows the single exponential fits for both isoforms over a voltage range from -90 to +60 mV. As anticipated, the decay time constants (τ) for I_g relaxation were faster for rSkM1 than for hH1a, although only at test potentials positive to -40 mV where the time to peak I_Na was less than 2 ms. At more negative test potentials where the times to peak I_Na become longer (almost 10 ms), the decay time of I_g relaxations was slightly faster for hH1a than for rSkM1. One explanation for such a finding would be if hH1a contained a component of gating charge that is so slow that it is not captured in I_g recordings near threshold potentials (see Discussion).

Figure 3. Family of Na⁺ channel gating currents and their integrals.

Family of gating currents (top) and their integrals (bottom) from fused cells expressing rSkM1 Na⁺ channels (A) and hH1a Na⁺ channels (B). Step depolarizations ranged from -120 to +40 mV from a holding potential of -150 mV. Data are capacity and leak corrected, digitally filtered at 15 kHz, and with every sixth point plotted. (Cells T4.10 and Q7.03.) C shows the single exponential fits to I_g relaxations for rSkM1 (•) and hH1a (○) Na⁺ channels. The data are the means ±s.e.m. for 10 cells expressing rSkM1 and 6 cells expressing hH1a. Time values (τ) were significantly different at P < 0.05 when V_t≥ 0 mV.

We have previously reported a concordance between the half-points of the peak G-V and Q-V relationships in native Na⁺ channels in canine cardiac cells (Hanck et al. 1990). A similar concordance was found to exist for hH1a Na⁺ channels expressed in fused cells (Fig. 4B) with no significant difference in half-points (0 ± 1 mV, n= 5 cells). In contrast, the half-point of the Q-V relationship for rSkM1 Na⁺ channels is shifted to a more negative voltage compared with its G-V relationship (Fig. 4A) with a significant difference in their half-points of -8 mV (± 3 mV, n= 6 cells). These data confirm that the difference between the half-points of the two relationships for rSkM1 compared with hH1a is a result of the Na⁺ channel isoforms themselves, and not from different experimental conditions. Previous studies of Na⁺ channels in squid giant axon (e.g. Stimers et al. 1985) and rat brain IIA Na⁺ channels expressed in Xenopus oocytes (Conti & Stühmer, 1989) have also shown a leftward shift of the Q-V relationship compared with the G-V relationship similar to that for rSkM1 Na⁺ channels.

Figure 4. Superimposed peak G-V and Q-V relationships in control conditions.

Superimposed peak Q-V (•) and G-V (dotted lines, from Fig. 1D) relationships for rSkM1 Na⁺ channels (A) and for hH1a Na⁺ channels (B). The Q-V relationships are from the same cells as shown in the G-V relationships from Fig. 1D. The data plotted in each panel are the means ±s.e.m. and the charge was normalized to the Q_max of each cell. The continuous lines represent the means of the best fits to each cell by a Boltzmann distribution:

(2)

where fractional Q_max is the normalized charge during depolarizing step; V_t is the test potential; V_½ is the half-point of the relationship; and s is the slope factor (mV). The parameters from the best fits are:

Na+ channel isoform	Qmax (pC)	V1/2 (mV)	Slope factor (mV)
rSkM1 (n= 6 cells)	3.4 ± 1.3 −	48 ± 5*	−10.6 ± 1.4
hH1a (n= 5 cells)	2.9 ± 1.4	−56 ± 5	−10.6 ± 1.8

^*

Significantly different at P < 0.05.

Effects of site-3 toxins on rSkM1 and hH1a Na⁺ channels

The faster time constants of I_g relaxations in rSkM1 could result from inherently faster rates of activation and/or inactivation. The work of Aldrich et al. (1983) in neuronal Na⁺ channels proposed that inactivation is rapid compared with activation, and has been interpreted to cause the large negative shift of the G-V relationship after removal of Na⁺ channel inactivation in neuroblastoma (Gonoi & Hille, 1987) and GH3 cells (Cota & Armstrong, 1989). However, site-3 toxin modification of whole cell I_Na in cardiac Purkinje cells resulted in only a very small shift of the G-V relationship, consistent with little overlap between channel opening and inactivation from the open state (Hanck & Sheets, 1995).

The effect of site-3 toxins on I_Na for fused cells expressing rSkM1 or hH1a Na⁺ channels is shown in Fig. 5. As expected after extracellular application of site-3 toxins, the decay of I_Na in response to step depolarizations was dramatically slowed for both isoforms. The effects of site-3 toxins on peak I_Na-voltage (I-V) relationships and normalized G-V relationships are shown for fused cells transfected with rSkM1 (Fig. 6A and B) and fused cells transfected with hH1a (Fig. 6C and D). Similar to native cardiac Na⁺ channels, site-3 toxin modification of hH1a Na⁺ channels caused only a minor hyperpolarizing shift of the V_½ of -3.7 ± 2.4 mV (n= 5) while the shift in the G-V relationship for rSkM1 Na⁺ channels was greater, -8.0 ± 1.1 mV (n= 6) (see Table 1). For both channel isoforms there was a slight steepening of the slope factor of the G-V relationship by 0.9 ± 0.1 mV in rSkM1, and by 0.8 ± 0.1 mV in hH1a. Maximal conductance (G_max) increased in both isoforms; although in these solutions it increased by a larger amount for hH1a (73 %) than for rSkM1 (17 %). This difference is considered in Discussion.

Figure 5. Effects of site-3 toxins on I_Na.

Effects of site-3 toxins on a family of I_Na responses to step depolarizations between -70 and +30 mV for fused cells expressing rSkM1 (A) and hH1a Na⁺ channels (B). These are the same cells as shown in Fig. 1A. The data are shown capacity corrected but not leak corrected, and digitally filtered at 5 kHz. (Cells T4.01 and P2.01.)

Figure 6. Effect of site-3 toxins on peak I_Na voltage and G-V relationships.

Effect of site-3 toxins on peak I_Na-V relationships (A and C) and peak G-V relationships (B and D) for 6 fused cells transfected with rSkM1 (A and B), and for 5 cells transfected with hH1a (C and D). All control (○) data are the mean values of each cell value measured before toxin modification and after wash of toxin except for one cell transfected with hH1a which had no wash. Data obtained after modification by site-3 toxins are shown (•). For B and D control data were normalized to G_max for each cell in control, and toxin data to G_max for each cell in toxin. The lines in A and C connect the data points. The lines in C and D represent the means of the best fits of each cell to eqn (1) with the parameters listed in Table 1.

Table 1.

Effect of site-3 toxins on the Boltzmann parameters of fits to peak G–V relationships for rSkM1 and hH1a Na⁺ channels

Na+ channel isoform	External solution	Gmax (μS)	V1/2 (mV)	Slope factor (mV)
rSkM1 (n= 6 cells)	Control	0.64 ± 0.24	−42 ± 3	−6.4 ± 0.5
	Site-3 toxin	0.75 ± 0.22*	−50 ± 3*	−5.1 ± 0.3*
hH1a (n= 5 cells)	Control	0.57 ± 0.31	−58 ± 5	−6.6 ± 0.6
	Site-3 toxin	0.96 ± 0.46*	−61 ± 6*	−5.8 ± 0.5*

^*Toxin values compared with control are significant at P < 0.05. G_max is the maximum peak conductance, V_1/2 is the half-point of the relationship (see eqn (1)).

The effect of site-3 toxin modification on the gating currents of both Na⁺ channel isoforms is shown in Fig. 7. The top panels show capacity and leak corrected I_g traces and their corresponding integrals are shown below for the same two cells shown in Fig. 3. Compared with control, the magnitude of the gating currents and their integrals were reduced in the presence of site-3 toxins. The magnitude of the reduction in gating currents by site-3 toxins is readily apparent by comparison of the Q-V relationships both before and after toxin modification (Fig. 7). For rSkM1 Q_max was reduced by 33 ± 7 % with a small, non-significant change in the half-point and slope factor of the Q-V relationship (Table 2). After washing out the toxin, the Q-V relationship returned to control (shown as the open squares). Similar changes in the Q-V relationships after toxin modification were noted in cells expressing hH1a Na⁺ channels (Fig. 7D). For hH1a channels the reduction in Q_max was 31 ± 4 % associated with a small, non-significant change in slope factor, and a small shift of the V_½ of the Q-V relationship by -5 mV (Table 2). The effects of site-3 toxins on the Q-V relationship of hH1a Na⁺ channels were similar to those previously reported for native Na⁺ channels in cardiac cells (Sheets & Hanck, 1995).

Figure 7. Effect of site-3 toxins on Q-V relationships of transfected cells.

Effect of site-3 toxins on Q-V relationships of 6 cells transfected with rSkM1 (A and C) and 5 cells transfected with hH1a (B and D). A and B show families of gating currents (top) and their integrals (bottom) after leak correction and digitally filtering at 15 kHz. Data plotted are means ±s.e.m. for cells in control (○), toxin (•) and after wash (□), except one cell from each isoform had no wash. Gating charge in toxin and wash were normalized to the Q_max determined for each cell in control solution. The continuous lines represent the best fits (see Table 2) to a Boltzmann distribution (eqn (2)).

Table 2.

Effect of site-3 toxins on the Boltzmann parameters of fits to Q–V relationships for rSkM1 and hH1a Na⁺ channels

Na+ channel isoform	External solution	Qmax (normalized)	V1/2 (mV)	Slope factor (mV)
rSkM1 (n= 6 cells)	Control	1.0	−48 ± 5	−10.6 ± 1.4
	Site-3 toxin	0.67 ± 0.07*	−50 ± 6	−12.3 ± 1.7
	Wash	0.99 ± 0.08	−52 ± 3	−9.8 ± 1.2
hH1a (n= 5 cells)	Control	1.0	−55 ± 5	−10.6 ± 1.8
	Site-3 toxin	0.69 ± 0.04*	−62 ± 9*	−11.5 ± 1.8
	Wash	0.93 ± 0.04	−57 ± 5	−12.0 ± 1.4

Differences in toxin values compared with control are significant at P < 0.05. Q_max is the maximal gating charge normalized to that in control, and V_1/2 is the half-point of the relationship (see eqn (2)).

It is important to note that an obvious reduction in charge was not apparent over the entire voltage range for either isoform, but the reduction in charge became obvious at potentials greater than -55 mV for both channel isoforms. This was the case despite the fact that V_½ of the Q-V relationship for rSkM1 was 8-10 mV more positive than that for hH1a, which means that relative to their mid-points the reduction of gating charge by site-3 toxins for rSkM1 became apparent at potentials more negative than those for hH1a Na⁺ channels.

Previously (see Fig. 4), we have shown that the relationship of the G-V curve to the Q-V curve is different between the two Na⁺ channel isoforms in control conditions with the half-point of the G-V curve shifted to more negative potentials compared with the Q-V curve for rSkM1 while the two half-points for hH1a were similar. If Na⁺ channel inactivation contributed to the difference in half-points between isoforms, then the modification of inactivation by site-3 toxins may help mitigate this difference. Figure 8 shows the G-V and Q-V relationships superimposed for both rSkM1 and hH1a Na⁺ channels after modification of Na⁺ channels by site-3 toxins. In contrast to the relationships in control, the striking feature is the similarity of the two relationships for both rSkM1 and hH1a. Although the relationships for hH1a continue to be about 10 mV more negative than sSkM1, both isoforms demonstrated a concordance between the G-V and Q-V relationships. The similarity between the data occurred because the half-point of the G-V relationship for rSkM1 was shifted to more negative potentials by ∼8 mV after modification by toxin, while the V_½ of the Q-V curve was little changed (Table 3). The shift in the G-V relationship was similar in magnitude to that reported for site-3 toxins in neuroblastoma cells (Gonoi & Hille, 1987), and supports the view that inactivation plays a greater role in determination of the G-V relationship for skeletal muscle Na⁺ channels than for heart Na⁺ channels.

Figure 8. Superimposed peak G-V and Q-V relationships for modified Na⁺ channels.

Superimposed peak G-V (•) and Q-V (○) relationships for site-3 toxin modified rSkM1 Na⁺ channels (A) and for site-3 toxin modified hH1a Na⁺ channels (B). The results plotted in each panel are the means ±s.e.m. for 6 cells expressing rSkM1 Na⁺ channels and 5 cells expressing hH1a Na⁺ channels. Data were normalized either to G_max for G-V relationships or to Q_max for Q-V relationships. The lines represent the means of the best fits (seeTable 3) of each cell to a Boltzmann distribution (eqns (1) and (2)).

Table 3.

Comparison of Boltzmann parameters of fits to peak G–V and Q–V relationships for rSkM1 and hH1a Na⁺ channels after modification by site-3 toxins

Na+ channel isoform	Ionic or gating current relationship	V1/2 (mV)	Slope factor (mV)
rSkM1 (n= 6 cells)	G–V	−50 ± 3	−5.1 ± 0.3
	Q–V	−50 ± 6	−12.3 ± 1.7*
hH1a (n= 5 cells)	G–V	−61 ± 6	–5.8 ± 0.5
	Q–V	−62 ± 8	−11.5 ± 1.8*

^*Q–V values compared with G–V values for each isoform are significantly different at P < 0.05. V_1/2 is the half-point of the relationship (see eqns (1) and (2)).

Q_max compared with G_max for rSkM1 and hH1a Na⁺ channels

Although the maximal number of electronic charges (e⁻) that can move during a voltage step depolarization are not precisely known for Na⁺ channels, there is evidence to suggest that skeletal muscle Na⁺ channels (Hirschberg et al. 1995) may move more e⁻ than do hH1a (Sheets & Hanck, 1995). To help investigate this question we compared Q_max with G_max for the two different isoforms of Na⁺ channels under the same experimental conditions. If Q_max were much greater for rSkM1 channels than for hH1a, then it might be expected that the slope of the Q_maxvs. G_max relationship would be steeper for rSkM1 Na⁺ channels than for hH1a. Figure 9 shows the results for 14 cells expressing rSkM1 and 13 cells expressing hH1a. Surprisingly, the slopes of the regression are not different. However, there are important limitations to the comparison of Q_max with G_max between isoforms, and a full discussion of these limitations is given below.

Figure 9. Relationship of Q_maxvs.G_max.

Relationship of Q_maxvs.G_max for rSkM1 (•) and hH1a (○) under control conditions. Q_max and G_max were obtained from the best fit of Boltzmann distributions to Q-V and G-V relationships for 14 fused cells expressing rSkM1 Na⁺ channels, and 13 cells expressing hH1a Na⁺ channels. The continuous line is the best fit by least squares regression with an intercept set to zero for rSkM1 (slope = 5.3 pC nS⁻¹, r²= 0.99), and the dotted line is the least squares regression for hH1a (slope = 5.4 pC nS⁻¹, r²= 0.97).

DISCUSSION

Sodium currents and gating currents from rSkM1 or hH1a Na⁺ channels expressed in fused cultured mammalian cells (either HEK 293 or tsA201 cells) allowed for comparison of the gating characteristics of the two Na⁺ channel isoforms. Use of a mammalian expression system facilitated the comparison of Na⁺ channel isoforms because it avoided the problem of abnormal I_Na kinetics of skeletal muscle Na⁺ channels that can occur when they are expressed in Xenopus oocytes (Krafte et al. 1988; Zhou et al. 1991). Secondly, the fused cell mammalian preparation minimized the shift of kinetic parameters of I_Na during voltage clamp of internally perfused cells compared with native Na⁺ channels in cardiac cells. The mean shift of the half-points of the G-V relationship and Na⁺ channel availability curves (i.e. steady-state inactivation curve) varied from -0.12 to -0.15 mV min⁻¹, and are much less than those reported for cardiac Na⁺ channels in native heart cells which typically approximate -0.5 mV min⁻¹ (Hanck & Sheets, 1992).

Comparison of the Q-V to G-V relationships for rSkM1 and hH1a Na⁺ channels

Similar to previous studies of human skeletal muscle Na⁺ channels (Wang et al. 1996a), the half-point of the G-V relationship for rSkM1 occurred at more positive potentials compared with that for hH1a Na⁺ channels, and channel kinetics, demonstrated by the shorter time to peak I_Na, were faster for rSkM1. In this study we show additional differences between the two Na⁺ channel isoforms by comparison of the Q-V relationship with the G-V relationship under control conditions. For hH1a Na⁺ channels the two relationships were concordant with almost identical half-points (see Fig. 4B), equivalent to what we have shown for cardiac Na⁺ channels in native heart cells (Hanck et al. 1990). However, comparison of the two relationships for rSkM1 showed the half-point of the Q-V relationship to be leftward (i.e. more negative) of the half-point for the G-V relationship, and with the appearance of a ‘foot’ to the Q-V relationship (see Fig. 4A). Similar ‘feet’ in the Q-V relationships have also been shown for neuronal Na⁺ channels including those in squid giant axon (e.g. Stimers et al. 1985) and rat brain IIA Na⁺ channels expressed in Xenopus ooctyes (Conti & Stühmer, 1989). Since the ionic and gating current recordings reported here were made under almost identical experimental conditions, the differences between the G-V and Q-V relationships result from the isoforms themselves. The appearance of a ‘foot’ to the Q-V relationship for rSkM1 may, in part, result from the more rapid kinetic transitions of rSkM1a compared with hH1a. During small step depolarizations where the probability that a channel will open is small and where the delays to channel openings are long, gating currents are expected to be slow and relatively small. As a consequence, the small gating current signal may be too small to raise it completely out of the baseline noise, resulting in an underestimation of the total gating charge. In the case of rSkM1 Na⁺ channels with their faster channel kinetics, measurement of gating charge near threshold potentials would be expected to be more complete compared with that for hH1a, and thus a ‘foot’ in the Q-V relationship may appear for rSkM1 and may be absent for hH1a. A similar explanation may explain the faster time course of I_g relaxations for hH1a compared with those of rSkM1 at test potentials near -60 mV (see Fig. 3C). Due to the slower kinetics of hH1a Na⁺ channels, a component of gating charge near threshold potentials may be so slow that it is almost completely absent in the I_g recordings.

Effects of site-3 toxins on Na⁺ channel ionic currents

Site-3 toxins resulted in the slowing of the decay of I_Na in response to step depolarizations for both rSkM1 and hH1a Na⁺ channels consistent with inhibition of channel inactivation from the open state(s). We have previously shown that the major effect of site-3 toxins is inhibition of the open-to-inactivated state transition for native Na⁺ channels in cardiac Purkinje cells (Hanck & Sheets, 1995). In this study, we found that hH1a Na⁺ channels heterologously expressed in fused mammalian cells are modified by site-3 toxins in an almost identical manner to those for native cardiac Na⁺ channels. Modified hH1a Na⁺ channels were characterized by a small negative shift of the V_½ of G-V curve (-2.5 mV in native Na⁺ channels vs. -3.7 mV in hH1a), a small change in the slope factor (1.0 mV in native Na⁺ channels vs. 0.8 mV in hH1a) and a relatively large increase in G_max (77 % in native Na⁺ channels vs. 73 % in hH1a) when the solutions contained TMA⁺ as the major cation. However, the large increase in G_max of cardiac I_Na by site-3 toxins may be, in part, due to relief of the previously reported voltage-dependent block of I_Na by intracellular TMA⁺ (O'Leary & Horn, 1994). In our previous studies with Cs⁺ as the major cation, G_max was found to increase by only 26 % for I_Na in heart cells (Hanck & Sheets, 1995). This later value was similar to the increase of 17 % for the G_max of rSkM1 channels. Even though the increase in G_max was less for rSkM1 Na⁺ channels, the half-point of the G-V relationship for rSkM1 Na⁺ channels was shifted more by site-3 toxins (-8 mV) compared with hH1a. The larger shift in the V_½ after toxin modification suggests that channel inactivation has a stronger influence on peak I_Na in normal, unmodified rSkM1 Na⁺ channels than on the peak I_Na of normal hH1a channels. Similar conclusions on the large shift of G-V relationships after inhibition of I_Na inactivation have been reported for I_Na in neuroblastoma cells (Gonoi & Hille, 1987) and for GH3 cells (Cota & Armstrong, 1989).

Effects of site-3 toxins on Na⁺ channel gating currents

The major effect of site-3 toxins on the Q-V relationships in both Na⁺ channel isoforms was a reduction in Q_max by 33 % in rSkM1 and by 31 % in hH1a channels. Only minor changes in the half-points and slope factors of the Q-V relationships occurred after toxin modification for both channel isoforms. These changes were nearly identical to the findings of site-3 toxins on cardiac Na⁺ channels in native heart cells (Sheets & Hanck, 1995). In those studies, it was demonstrated that site-3 toxins inhibit gating charge movement that is associated with, or tightly coupled to, the open-to-inactivated state transition. It was surprising to find that site-3 toxins reduced Q_max by almost the same magnitude in rSkM1 Na⁺ channels compared with cardiac Na⁺ channels. As the channel kinetics of skeletal muscle Na⁺ channels are more rapid than those of heart Na⁺ channels (Wang et al. 1996a; Chahine et al. 1996), and the maximal number of electronic charges (e⁻) that are associated with skeletal muscle Na⁺ channels is thought to be as high as 12 e⁻ (Hirschberg et al. 1995) compared with estimates of 5-6 e⁻ for heart Na⁺ channels (Sheets & Hanck, 1995), it is reasonable to expect that more gating charge may be associated with channel activation in rSkM1 Na⁺ channels compared with cardiac Na⁺ channels. If this were correct, it might be anticipated that the reduction of Q_max in rSkM1 Na⁺ channels by site-3 toxins would be less than that for cardiac Na⁺ channels. However, the reduction in Q_max was nearly identical for the two isoforms suggesting that the total amount of gating charge associated with channel activation may be 5-6 e⁻ for both isoforms. Because gating charge is a measure of the voltage dependence of kinetic state transitions, these results suggest that the voltage dependence of kinetic transitions are similar between rSkM1 and hH1a Na⁺ channels. If this were the case, then the primary difference between the channel kinetics of rSkM1 and hH1a channels would result from inherently faster time constants of channel activation and inactivation in rSkM1 compared with hH1a.

In contrast to recordings made in control solutions, modification of rSkM1 channels by site-3 toxins resulted in the concordance of the Q-V and G-V relationships similar to that for hH1a (see Fig. 8). The concordance in toxin-modified rSkM1 Na⁺ channels resulted, in large part, from a negative shift of the G-V relationship by 8 mV. This suggests the difference in the two relationships in control conditions for rSkM1 may result from relatively smaller peak I_Na magnitudes near threshold potentials caused by more rapid channel inactivation. As inactivation becomes slowed by site-3 toxin modification of rSkM1, peak I_Na would be expected to increase as the probability of channels occupying the open state increased, thus resulting in a leftward shift of the G-V relationship.

Q_maxvs. G_max relationships for rSkM1 and hH1a Na⁺ channels

Although the maximum gating charge associated with rSkM1 or hH1a remains controversial, it has been estimated to vary between 5 and 12 e⁻ (Hille, 1992; Hirschberg et al. 1995; Sheets & Hanck, 1995). To measure the gating charge associated with a single channel accurately, it is necessary to determine both total gating charge and the number of channels in the same preparation. So far, this has not been possible for Na⁺ channels. However, a relative comparison of gating charge associated with Na⁺ channels can be made between isoforms especially if the measurements are performed under similar experimental conditions. Figure 9 showed the relationship of Q_max to G_max for both rSkM1 and hH1a isoforms measured under similar conditions. Surprisingly, the slopes of the two relationships were almost identical suggesting that the total e⁻ per channel were equivalent. It is important to note that relative comparisons of G_max in different Na⁺ channel isoforms, even obtained under identical experimental conditions, do not necessarily allow for an accurate comparison of the number of Na⁺ channels expressed in each cell. This results from the fact that G_max of each cell is dependent not only upon the total number of channels, but also upon single channel conductance and the probability of single channels being open at peak I_Na. Single channel conductance of skeletal and heart Na⁺ channels has been shown to be similar (see Fozzard & Hanck, 1991), and is unlikely to contribute to a large error in the comparison. However, the channel kinetics clearly differ between the two isoforms, and thus G_max, a measurement dependent upon the number of simultaneous open channels, may not be directly comparable. Despite these limitations, the nearly identical slopes of Q_max to G_max suggest the maximum number of e⁻ per channel may be comparable between rSkM1 and hH1a Na⁺ channels. A similar conclusion results from comparison of the amino acid sequences between the two Na⁺ channel isoforms. Of the 554 amino acids assigned to the membrane spanning region based upon hydropathy analysis 86 % of the amino acids are identical and 94 % are highly homologous. In particular, all the charged residues within the membrane spanning region (which includes the putative voltage sensors) are identical with only a single exception, an asparagine to lysine near the extracellular end of domain IV segment 2. Due to such conservation between the two Na⁺ channel isoforms it may not be surprising if the total number of e⁻ was found to be similar.