INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The dihydropyridine receptor (DHPR) of skeletal muscle comprises _1S, _1a, ₂/, and  subunits. This complex serves as a voltage sensor for excitation-contraction (EC) coupling and is responsible for the L-type Ca²⁺ current present in these cells. Functional expression of cDNAs in dysgenic myotubes supports the view that the ₁ subunit of the DHPR determines, to a large extent, the properties of the Ca²⁺ current and the type of EC coupling expressed in the muscle cell (Tanabe et al., 1988, 1990a,b, 1991; Adams et al., 1990; Garcia-Martinez et al., 1994; Garcia-Martinez, 1994). The dysgenic mutation consists of a single base deletion in the murine gene encoding for the _1S subunit of the DHPR (Chaudhari, 1992; Varadi et al., 1995). Dysgenic myotubes do not have a functional _1S subunit, yet display a low-density L-type Ca²⁺ current that has been named I_dys (Adams and Beam, 1989; Shimahara and Bournaud, 1991). Presumably, I_dys is encoded by a cardiac-type _1C subunit, although this has not been entirely demonstrated (Chaudhari and Beam, 1993). I_dys activates much faster than, and inactivates much more slowly than, the normal L-type Ca²⁺ current (Adams and Beam, 1989). Under some conditions, I_dys mediates contractions that are dependent on external Ca²⁺, suggesting that this current may play a functional role in the fetal stages of muscle development (Adams and Beam, 1991).

The function of the  subunit of the DHPR in skeletal muscle has been investigated using gene targeting to inactivate the murine gene encoding _1a, the most abundant  subunit expressed in skeletal muscle (Gregg et al., 1996). Quite surprisingly, ₁-null myotubes displayed a phenotype that is similar to that of dysgenic myotubes consisting of EC uncoupling, a low density of charge movement, and a low-density L-type Ca²⁺ current named I_null (Strube et al., 1996). Ca²⁺ currents, charge movements, and intracellular Ca²⁺ transients are restored in ₁-null myotubes after transfection with _1a cDNA (Beurg et al., 1997). These results suggest that ₁ has a critical function in modulating both the functional expression of the DHPR voltage sensor and the Ca²⁺ current.

The low density of I_null, the L-type Ca²⁺ current of ₁-null myotubes, may indicate that functional _1S subunits are absent in these cells. If this is so, I_null could be similar to I_dys. Alternatively, I_null may represent a down-regulated L-type Ca²⁺ channel due to the specific absence of _1a from the skeletal DHPR complex. In this case, I_null should differ qualitatively from I_dys. To clarify the molecular origin of I_null, in the present study we analyzed conductive and gating properties of the L-type Ca²⁺ current of dysgenic and ₁-null myotubes under identical conditions in primary cell cultures. Part of this work has appeared previously in abstract form (Strube et al., 1997).

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell cultures

Mouse myotubes homozygous for the mdg allele (mdg/mdg) or the ₁ mutation cchb1^/ are called dysgenic and ₁-null, respectively. Collectively they are called mutant cells. Mouse myotubes with a normal phenotype were heterozygous for either mutation (mdg/+ or cchb1^/+) or wild type. All experiments were performed on primary cultures as previously described (Beurg et al., 1997). The hind limbs of 18-day-old fetuses were dissected free of skin and bones and washed in Ca²⁺-Mg²⁺-free Hanks' buffer. The tissues were incubated for ~10 min at 37°C in Ca²⁺-Mg²⁺-free Hanks' buffer containing 0.125% trypsin and 0.05% pancreatin (from porcine pancreas; Sigma Chemical Co., St. Louis, MO). After mechanical dispersion, the cell suspension was filtered through sterile gauze. After centrifugation of the filtrate and resuspension of the pellet in plating medium, the cells were preplated into 100-mm Falcon plastic dishes for 1 h to enrich the myoblasts. Final plating was done in 35-mm Falcon plastic petri dishes covered with 1% gelatin at 1-4 × 10⁴ cells/plate in 2 ml plating medium. Cells were grown in 8% CO₂ for 5-7 days and later in 5% CO₂ in fetal bovine serum-free medium. The plating medium was composed of 78% Dulbecco's modified Eagle medium, 10% horse serum, 10% fetal bovine serum, 2% chick embryo extract, 10 UI/ml penicillin, and 0.01 mg/ml streptomycin. The fetal bovine serum-free medium was composed of 88.75% Dulbecco's modified Eagle medium, 10% horse serum, 1.25% chick embryo extract, 10 UI/ml penicillin, and 0.01 mg/ml streptomycin.

Immunostaining

Cells were fixed and processed for immunostaining as described (Flucher et al., 1991). The DHPR _1S monoclonal antibody (Upstate Biotechnology, Lake Placid, NY) was used at a dilution of 1:50. The secondary antibody was a fluorescein conjugated polyclonal goat anti-mouse IgG (Cappel, IGN Pharmaceuticals, Irvine, CA) and was used at a dilution of 1:100. Fluorescence confocal images of 512 by 512 pixels (0.1-0.3 µm/pixel) were obtained on a BioRad 1000 confocal microscope (BioRad Instruments, Hercules, CA), using the 488-nm spectral line from an argon laser. Images were converted to a 16-level gray scale with National Institutes of Health-Image software.

Ca²⁺ currents

The standard patch-clamp technique was used in the whole-cell recording configuration. Ca²⁺ current was recorded as previously described (Strube et al., 1996). The external solution for current recording was (in mM) 130 tetraethylammonium methanesulfonate, 10 CaCl₂ or 10 BaCl₂, 1 MgCl₂, 10³ tetrodotoxin, and 10 HEPES-tetraethylammonium(OH), pH 7.4. The pipette solution consisted of (in mM) 140 Cs aspartate, 5 MgCl₂, 5 EGTA, 10 3-[N-morpholino]propane sulfonic acid-CsOH, pH 7.2. Standard patch electrodes had tip resistances between 2 M and 5 M when filled with the pipette solution. Recordings were made with an Axopatch 1D and a headstage with a 50-M feedback resistor (Axon Instruments, Foster City, CA). The effective series resistance was compensated up to the point of amplifier oscillation with the analog circuit provided by Axopatch. Three linear capacitive components and a leak component were canceled with a tunable analog circuit. Data acquisition was performed with a TL1 DMA interface controlled by pCLAMP software (Axon Instruments). The data were digitized at 50-400 µs/point and filtered at 1-3 kHz with an analog 8-pole Bessel filter. All experiments were performed at room temperature.

Variance analysis

The ensemble variance of I_dys and I_null was estimated from the ensemble average of the squared difference between consecutive current records (Heinemann and Conti, 1992). A set of 50 pulses to +20 mV was delivered to the same cell at a rate of one pulse every 5 s. The pulse cycle was delivered from a holding potential of 80 mV and consisted of a step to 30 mV for 750 ms, followed by the test pulse to +20 mV, followed by a step to 30 mV, followed by a step to the holding potential. Test pulse duration and sampling frequency were 25 ms and 40 kHz or 50 ms and 20 kHz for dysgenic cells. Test pulse duration and sampling frequency were 50 ms and 20 kHz or 100 ms and 10 kHz for ₁-null and normal cells. All records were low-pass-filtered at 4 or 2 kHz at the moment of acquisition with an 8-pole analog Bessel filter. Amplifier gain was set at 5 mV/pA, and the A/D resolution was 0.5 pA per bit. The set of 50 pulses was repeated several times, and one or two sets per cell were selected for analysis. Pairs of consecutive records {X_i1(t), X_i(t)} within a selected set were subtracted in an overlapped manner to generate 49 difference records, {X_i(t)  X_i1(t)}. A maximum of 10 difference records were discarded. The ensemble variance, ² (+), for the remaining records, n, was calculated according to Eq. 1 (Noceti et al., 1996): where µ(t) was the mean value of Y(t)_i and Y(t)_i = 1/2 {X_i(t)  X_i1(t)}. The variance at the holding potential was estimated in the same manner and was time-averaged for 10 ms before the voltage pulse. The resting variance was subtracted from ²(t), and the latter plotted against the ensemble mean current, I(t), of the same set of current records. The mean-variance relationship was fit by a nonlinear least-squares method according to Eq. 2 (Neher and Stevens, 1975; Sigworth, 1980), where i is the single-channel current and N_F is the number of functional channels activated by the voltage pulse. Some difference records were discarded from the analysis because of deterioration of the pipette seal resistance or excess of current run-down during the acquisition of one or both of the original current records. To discard these records objectively, i.e., without assumptions concerning the time course ²(t), we ranked difference records as suggested by Heinemann and Conti (1992). We calculated the variance of the time-averaged current in three segments of each difference record, a few milliseconds before the pulse, immediately after the onset of the pulse, and immediately before the ending of the pulse. Traces with the highest numerical score in each of these time segments, up to a total of 10 traces, were discarded. In these time segments, the traces with the highest scores were the most likely to be contaminated by, respectively, an increase in resting leak, a mismatch in the cancellation of residual capacitance transients, and an increase in leak or rundown during the pulse. Variance calculations were verified using pseudomacroscopic ensemble currents generated by a single-channel simulation program (CSIM, Axon Instruments).

Data and curve fitting

The density of the Ca²⁺ current of normal and mutant myotubes is approximately constant from days 8 to 16 of cell culture (Beurg et al., 1997). In the present study, we pooled and averaged data from cells between days 8 and 17 of culture. Curve fitting was done with Marquardt-Levenberg algorithms provided by Sigmaplot (Jandel, San Rafael, CA) and pClamp (Axon Instruments). The time constant ₁, describing activation of the Ca²⁺ current, was obtained from a fit of the pulse current at each voltage according to I(t) = K[1  exp(t/₁)]exp(t/₂) (Eq. 3), where K is a constant and ₂ describes inactivation. All averages are presented as mean ± SEM.

Chemicals

Deionized glass-distilled water was used in all solutions. All salts were reagent grade. Bay K 8644 was made as 5 mM stocks in absolute ethanol and stored in light-resistant containers. TTX was from Sigma Chemical Co. Bay K 8644 was from Calbiochem (La Jolla, CA).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Fig. 1 shows Ca²⁺ currents in normal, ₁-null, and dysgenic cells in culture. The transient component was in response to a pulse to 20 mV from a holding of 80 mV and was identified as T-type current (Adams and Beam, 1989; Strube et al., 1996). The sustained component shown in the bottom trace was in response to a pulse to +40 mV and was identified as L-type current in all cases. The sustained currents of the two mutant cells, here identified as I_null and I_dys, were much smaller than the L-type current of normal cells. This result is in agreement with previous studies performed separately in dysgenic and ₁-null myotubes in culture (Bournaud et al., 1989; Beurg et al., 1997). L-type and T-type components were not always present in each cell. This is shown in the right panels of Fig. 1, in which the density of L-type current is plotted in the abscissa, and the density of T-type current in the same cell is plotted in the ordinate. The open symbols correspond to cells without a detectable T-type current, whereas the filled symbols correspond to cells in which T-type currents were obvious. The population average density of the L-type current was computed separately for cells with or without T-type currents and is indicated in each graph. The presence or absence of T-type currents did not influence the L-type Ca²⁺ current density, except in the case of ₁-null cells, where the L-type current was lower in the subpopulation of cells without T-type current. This correlation was weak, and it did not extend to other properties of the L-type current of ₁-null cells, such as the kinetics of activation and sensitivity to Bay K 8644 described below. L-type Ca²⁺ currents were present in all normal cells (61 cells), about two-thirds of ₁-null cells (48 of 67 cells), and about two-thirds of dysgenic cells (52 of 67 cells). L-type and T-type Ca²⁺ currents were present in about one-third of all normal cells and about five-sixths of all mutant cells. In the remainder of this study, recordings were made either from a holding potential of 50 or 40 mV to inactivate the T-type current, or from 80 mV in cells without T-type currents.

View larger version (23K): [in this window] [in a new window]   FIGURE 1   Density of L-type and T-type currents in normal and mutant myotubes in culture. (Left) Traces of Ca2+ current for pulses to 20 mV and +40 mV from a holding potential of 80 mV. The pulse duration was 300 ms, and the cell capacitance was 314, 592, and 510 pF for normal, 1-null, and dysgenic, respectively. (Right) Density of L-type current and T-type current for each cell. Open symbols are for cells without measurable T-type currents (<50 pA). All other cells are shown with filled symbols. Mean ± SEM is shown separately for the two groups of cells. Current densities were computed at the peak of the I-V curve of each cell (+10 mV to +30 mV for L-type current and 20 mV for T-type current). The L-type current densities of cells with and without T-type currents were 7.46 ± 0.44 pA/pF (44 cells), 0.62 ± 0.05 pA/pF (31 cells), and 0.69 ± 0.1 pA/pF (28 cells) for normal, 1-null, and dysgenic cells, respectively. The L-type current densities of cells with T-type currents were 7.02 ± 0.73 pA/pF (14 cells), 0.63 ± 0.06 pA/pF (25 cells), and 0.56 ± 0.11 pA/pF (20 cells) for normal, 1-null, and dysgenic cells, respectively. The L-type current densities of cells without T-type currents were 7.66 ± 0.55 pA/pF (30 cells), 0.55 ± 0.07 pA/pF (6 cells), and 1.01 ± 0.18 pA/pF (8 cells) for normal, 1-null, and dysgenic cells, respectively.

Fig. 2 shows L-type Ca²⁺ currents in response to 1-s pulses from a holding potential of 50 mV in a normal, a ₁-null, and a dysgenic myotube. Currents were normalized to the cell capacitance, and it should be noticed that scales for normal and mutant cells are different. At all pulse potentials, the normal Ca²⁺ current was much larger than either I_null or I_dys. In addition, the normal current had a slower time to peak and displayed significant inactivation. The mutant currents had a similar density and showed little inactivation during the pulse. Current-voltage relationships are shown in the top panel of Fig. 3. The three Ca²⁺ currents activated at potentials more positive than 10 mV, but the peak densities of I_null and I_dys were 11- to 12-fold lower than the peak density of the normal current. Furthermore, I_dys activated at slightly more negative potentials than I_null. The bottom panel of Fig. 3 shows conductance-voltage relationships for the two mutant currents. A Boltzmann fit of the G_Ca-V curve of 31 ₁-null cells and 25 dysgenic cells showed that G_max (30.2 ± 2.5 pS/pF for ₁-null versus 30.5 ± 3.5 pS/pF for dysgenic) and the slope factor k (8.6 ± 0.4 mV versus 8.7 ± 0.4 mV) were identical for the two currents. However, there was a ~5 mV difference in V_1/2 (16.3 ± 1.1 mV versus 11.7 ± 1.0 mV) that was significant (p < 0.005, unpaired t-test). This result suggested that the voltage dependences of the DHPR complexes underlying I_null and I_dys were nonidentical.

View larger version (17K): [in this window] [in a new window]   FIGURE 2   Voltage dependence of normal and mutant Ca2+ currents. Whole-cell L-type Ca2+ currents are shown for a normal, 1-null, and dysgenic myotube at the indicated pulse potentials. The holding potential was 50 mV, and the pulse duration was 1 s. The cell capacitance was 267, 522, and 718 pF for the normal, 1-null, and dysgenic cells, respectively.

View larger version (16K): [in this window] [in a new window]   FIGURE 3   Current-voltage and conductance-voltage relationships of normal and mutant Ca2+ currents. (Top) Voltage dependence of the L-type Ca2+ current measured 300 ms after the onset of the pulse in normal (, 45 cells), 1-null (, 31 cells), and dysgenic (, 28 cells) myotubes. (Bottom) Voltage dependence of the L-type Ca2+ conductance of 1-null (, 31 cells) and dysgenic (, 25 cells) myotubes. The lines are a Boltzmann fit to the population average GCa-V curves. Parameters of the fit are Gmax = 29.2 and 29.8 pS/pF; V1/2 = 14.8 and 10.7 mV; k = 7.9 and 8.4 mV for 1-null and dysgenic myotubes, respectively.

To more fully understand the molecular nature of the DHPR complexes underlying I_null and I_dys, we investigated permeation, activation kinetics, and pharmacological properties of both currents. Fig. 4 shows current-voltage curves of ₁-null and dysgenic cells in which the external solution containing 10 mM Ca²⁺ was replaced by 10 mM Ba²⁺ and then returned to 10 mM Ca²⁺. In some experiments, this sequence was reversed so that Ba²⁺ was replaced with Ca²⁺, and then the external solution was returned to Ba²⁺. In all cases, the Ba²⁺ current was larger than the Ca²⁺ current by an average proportion of ~1.7-fold for I_null and ~1.9-fold for I_dys. This result was consistent with the identification of I_null and I_dys as L-type currents. However, there was a much larger separation of the I_Ba-V and I_Ca-V curves in ₁-null cells than in dysgenic cells. This is clearly seen in the normalized curves shown in the bottom panels of Fig. 4. The asterisks indicate Ca²⁺ and Ba²⁺ currents that at the same potential were significantly different (p < 0.02, unpaired t-test). Evidently, Ca²⁺ biased the voltage dependence of I_null more strongly than that of I_dys.

View larger version (25K): [in this window] [in a new window]   FIGURE 4   Ca2+ and Ba2+ current-voltage relationships for Inull and Idys. (Top left and right) L-type I-V curves for the same 1-null (five cells) or dysgenic (five cells) cells in external solution containing 10 mM Ca2+ (filled symbols) and 10 mM Ba2+ (open symbols). (Bottom left and right) The same data normalized to the peak of the I-V curve. The asterisks indicate Ca2+ and Ba2+ currents that were significantly different at the same potential (unpaired t-test, p < 0.02).

Fig. 5 shows scaled traces of the time course of normal, I_null, and I_dys currents for a depolarization to +20 mV. I_dys activated faster than I_null, and both activated faster than the normal current. Furthermore, I_null and I_dys displayed much less inactivation than the normal current. In fact, the inactivation of I_dys or I_null was barely detectable in most cases. The lines correspond to a fit of the pulse current according to Eq. 3. In most cells, this equation was sufficient to describe the time course of the Ca²⁺ current in the range of positive potentials. From the fit we extracted the time constant of current activation, which is shown at each voltage and for each cell type in the bottom panel of Fig. 5. I_dys activated two- to threefold faster than I_null and fivefold faster than the normal current. Furthermore, in this range of test potentials the activation rate of I_dys was essentially voltage-independent. In contrast, the activation rate of the normal and I_null currents slowed with increasingly positive potentials. The slowing of the normal current at positive potentials was similar to that described previously (Dirksen and Beam, 1995).

View larger version (16K): [in this window] [in a new window]   FIGURE 5   Kinetics of activation of Inull and Idys. (Top) Scaled traces at +20 mV of normal L-type Ca2+ current (), Inull (), and Idys () in response to a 1-s pulse from a holding potential of 50 mV. The solid lines correspond to a fit of the pulse current according to Eq. 3, with K = 1.16 pA, 1 = 68.6 ms, 2 = 1824 ms for normal, K = 1.00 pA, 1 = 16.3 ms, 2 = 6349 ms for 1-null, and K = 1.00 pA, 1 = 7.6 ms, 2 = 4251 ms for dysgenic myotubes. (Bottom) Tau activation, 1 of Eq. 3, obtained from a fit of the pulse current at each potential. Entries indicate the number of cells. Asterisks indicate 1 for Inull and Idys significantly different at the same potential (unpaired t-test, p < 0.005).

The stimulatory effects of the DHP Bay K 8644 are shown in Fig. 6. When cells were exposed to 5 µM Bay K 8644, the peak Ca²⁺ current increased by ~1.3-fold in normal, ~2.1-fold in ₁-null, and ~2.7-fold in dysgenic myotubes. Thus I_dys was stimulated more strongly than I_null. However, the difference was not significant (t-test, p = 0.2). Because Bay K 8644 is known to reduce the time to peak of the L-type current, we also investigated whether the kinetics of activation of each current were affected differently. The bottom panel of Fig. 6 shows time constants of activation fitted to the pulse current in the range of 10 to +40 mV in cells stimulated by 5 µM Bay K 8644. A comparison of time constants in Figs. 5 and 6 indicated that the DHP accelerated the kinetics of activation in all cases. However, the activation time constant was reduced more strongly in normal and ₁-null cells than in dysgenic cells. The ranking order for stimulation of the Ca²⁺ current was I_dys > I_null > normal, whereas that for stimulation of the activation rate was normal > I_null > I_dys. This result suggested that the DHPR complexes responsible for I_null and I_dys displayed different mechanisms of modulation by DHPs.

View larger version (28K): [in this window] [in a new window]   FIGURE 6   Stimulation of Inull and Idys by Bay K 8644. (Top) L-type I-V curves of normal, 1-null, and dysgenic myotubes during a control period (open symbols) and 10 min after exposure of cells to 5 mM Bay K 8644 (filled symbols). Current was normalized to the peak of the I-V curve during the control period. The fold stimulation of the peak Ca2+ current produced by Bay K 8644 was 1.3 ± 0.1 (11 cells) for normal, 2.1 ± 0.3 (11 cells) for 1-null, and 2.7 ± 0.3 (11 cells) for dysgenic myotubes. (Bottom) Activation time constants of the normal L-type current (squares), Inull (triangles), and Idys (circles) in myotubes exposed to 5 µM Bay K 8644. The pulse current at each potential was fitted by Eq. 3. Tau activation corresponds to 1 of Eq. 3. Entries indicate the number of cells. Asterisks indicate 1 for Inull and Idys significantly different at the same potential (unpaired t-test, p < 0.005). Tau activation at +30 mV in the absence and presence of Bay K 8644 was 49.2 ± 2.6 and 11.2 ± 1.5 ms (9 cells) for normal, 26 ± 3.3 and 7.6 ± 0.7 ms (10 cells) for 1-null, and 7.3 ± 1.1 and 3.4 ± 0.3 ms (4 cells) for dysgenic myotubes. The fold decrease in tau activation at +30 mV produced by Bay K 8644 was 4.4 ± 0.2 for normal, 3.4 ± 0.4 for 1-null, and 2.1 ± 0.3 for dysgenic myotubes.

Fig. 3 showed that the densities of I_null and I_dys were the same. However, the properties of the Ca²⁺ channels underlying these currents, namely, the unitary single-channel current, i, and the number of functional channels per cell, N_F, may not necessarily be identical. Thus we estimated i and N_F by mean-variance analysis. Fig. 7, A-C, shows the time course of the mean Ca²⁺ current (smooth trace) and its variance (noisy trace) estimated from 40 pulses to +20 mV in a normal, a ₁-null, and a dysgenic cell. In the normal cell, the variance increased in proportion to the mean current throughout the pulse. In contrast, the variance of I_null and I_dys saturated earlier than the mean current. To increase the accuracy of the variance determination, we selected mutant cells with relatively large current densities. I_max, the whole-cell Ca²⁺ current measured at the end of the pulse, was 0.5-2 pA/pF for the mutant cells and 4-7 pA/pF for normal cells. In cells with I_max < 0.2 pA/pF, the variance of the pulse current was barely distinguishable from the variance of the rest current immediately before the pulse. The latter component was subtracted from the pulse variance in all cases. After normalization for cell capacitance, the variance of I_null at times longer than the activation time constant amounted to a doubling of the rest variance, whereas those of I_dys or the normal current were at least three times as large. Hence I_dys was intrinsically noisier than I_null. This is clearly noticeable in a comparison of variance traces in Fig. 7, B and C, during the second half of the pulse. Fig. 7, D-F, shows mean-variance curves of the same data. The smooth line is a fit of the data according to Eq. 2. In agreement with the shape of each curve, N_F was the largest and p_max (I_max/iN_F) the lowest for normal cells. Because of the uncertainty in the fit of a quasilinear mean-variance curve with Eq. 2, for normal cells we could only estimate N_F as >300 channels/pF and p_max < 0.3 (six cells). In mutant cells, the mean-variance curves reached a maximum in all cases, and the parabolic fit resulted in unique parameters. For five ₁-null cells N_F was 45 ± 12 channels/pF and p_max was 0.68 ± 0.1, and for five dysgenic cells N_F was 24 ± 5 channels/pF and p_max was 0.62 ± 0.1. Evidently, a similar p_max was reached at the end of the pulse by the Ca²⁺ channels of ₁-null and dysgenic cells. Furthermore, I_max for these two groups of cells was the same (1 ± 0.2 pA/pF versus 1.4 ± 0.5 pA/pF, respectively), and the difference in channel density was not significant (t-test, p = 0.13). The initial slope of the mean-variance curve was much larger for dysgenic cells than for the other two cell types and resulted in a fitted i that was larger for the dysgenic cells. This is best shown by plots of the ratio variance/mean versus mean (Fig. 7, G-I). Because Eq. 2 can be rearranged as ²(t)/I(t) = i  (1/N_F)I(t), the single-channel current i may be conveniently obtained by linear extrapolation of the variance/mean ratio to zero mean current. The data showed that the extrapolated y intercept (i) was higher for I_dys than for either I_null or the normal current. The single-channel currents obtained by extrapolation of the ratio-mean plot and those obtained from the parabolic fit of the mean-variance plot were in close agreement and were (in fA) 22 ± 4 (six cells), 43 ± 7 (five cells), and 84 ± 9 (five cells) for normal, ₁-null, and dysgenic cells. In all combinations, these differences were significant (t-test, p < 0.02). It should be noted that the accuracy of the fit of N_F is low unless the mean-variance relationship is highly curved. This was not the case for all cells. On the other hand, i is fit with the same accuracy in mean-variance plots with high or low curvature, as long as the time course of the variance at low I(t) is accurate. In summary, the mean-variance analysis demonstrated that the Ca²⁺ permeation characteristics of dysgenic and ₁-null Ca²⁺ channels were significantly different.

View larger version (19K): [in this window] [in a new window]   FIGURE 7   Mean-variance relationship of the normal L-type Ca2+ current, Inull and Idys. (A-C) Superimposed time courses of the mean whole-cell Ca2+ current (smooth trace) and the ensemble variance (noisy trace) for a step potential to +20 mV after a 750-ms prepulse to 35 mV from a holding potential of 80 mV. The pulse duration was 100 ms for the normal cell and 50 ms for the 1-null and dysgenic cells. The ensemble variance was computed according to Eq. 1 from 40 pulses (39 difference traces). Variance traces were low-pass filtered at 1 kHz with a digital Gaussian filter. The vertical calibration bar corresponds to I = 1 pA/pF and 2 = 0.02 pA2/pF for normal, I = 0.2 pA/pF and 2 = 0.005 pA2/pF for 1-null, and I = 0.4 pA/pF and 2 = 0.02 pA2/pF for dysgenic cells. The horizontal calibration bar corresponds to 20 ms for normal and 10 ms for 1-null and dysgenic cells. (D-F) The ensemble variance plotted as a function of the mean Ca2+ current for the same cells. The origin of each graph is given by the intersection of the horizontal calibration bar (mean) and vertical calibration bar (variance). Lines are a fit of the data according to Eq. 1. Parameters of the fit are i (pA) = 0.02, 0.05, 0.08 and NF (channels/pF) = 575, 24, 39 for normal, 1-null, and dysgenic, respectively. (G-I) A linear fit of the variance/mean ratio plotted as a function of the mean current.

Many properties of I_null shown here and previously are consistent with the bulk of this current originating from a DHPR complex that includes the _1S subunit (Strube et al., 1996; Beurg et al., 1997). However, conventional epifluorescence of ₁-null cells labeled with a _1S monoclonal antibody failed to detect significant levels of expression of _1S (Gregg et al., 1996). We reexamined this question by immunolabeling cells at a higher antibody concentration (1:50 dilution instead of 1:200 used previously) and with a different _1S monoclonal antibody (Ohlendieck et al., 1991). Fig. 8 shows confocal gray scale images of a dysgenic (A) and ₁-null (B) myotubes labeled with a _1S-specific primary antibody and a fluorescein-congugated secondary antibody. We consistently observed a faint staining of the dysgenic myotubes. The latter result agreed with a previous report (Flucher et al., 1991) and presumably was due to nonspecific secondary antibody binding to the myotube. Background-level staining of cells of the kind observed in Fig. 8 A was also seen in many ₁-null cells (not shown). However, ~50% of the examined ₁-null myotubes displayed a fluorescence intensity significantly above the background intensity. The image in Fig. 8 B was representative of ₁-null myotubes displaying a high immunofluorescence. Evidently, _1S was expressed in some but not all ₁-null myotubes in culture. The fact that ~50% of ₁-null cells did not express _1S was consistent with the finding that ~30% of the ₁-null myotubes did not display L-type Ca²⁺ current, although no efforts were made to explore this correlation further.

View larger version (58K): [in this window] [in a new window]   FIGURE 8   Immunofluorescence of 1-null and dysgenic myotubes labeled with DHPR 1S antibody. Confocal images of 512 × 512 pixels were converted to a 16-level gray scale, and the color table was inverted so that high-intensity pixels appear black and low-intensity pixels appear white. (A) A dysgenic myotube. (B) A 1-null myotube at the same magnification.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

₁-null myotubes display an L-type Ca²⁺ current that differs from the normal L-type Ca²⁺ current in density, voltage dependence, and kinetics of activation (Strube et al., 1996; Beurg et al., 1997). I_null has many characteristics of the Ca²⁺ current produced by DHPR ₁ subunits expressed in the absence of DHPR  subunits (summarized by Strube et al., 1996). Thus I_null could originate from skeletal-type DHPR complexes that include _1S but are deficient in _1a. To address this issue, we compared I_null with I_dys, the L-type Ca²⁺ current of dysgenic myotubes that do not express a functional DHPR _1S (Knudson et al., 1989; Chaudhari, 1992; Varadi et al., 1995). Statistically significant differences were found in the midpoint of the G-V curves, the kinetics of activation, and the single-channel currents estimated by ensemble variance analysis.

Because i estimated for I_null is between the values estimated for the normal currrent and for I_dys, I_null could represent a mixed population of L-type channels. Some channels in this mixture could include the DHPR _1S subunit, whereas others include the ₁-dysgenic subunit. We based this explanation on the fact that a significant fraction of ₁-null cells expressed DHPR _1S and that there is no reason to assume that ₁-dysgenic may not be expressed in these cells as well. To test this hypothesis, we performed mean-variance analysis of ensemble currents generated by computer simulation of two independent channels of different single-channel currents. The estimated i varied in direct proportion to the number of trials of each channel included in the ensemble (not shown). Assuming that i estimated for ₁-null represents a weighted sum of the single-channel currents of x dysgenic channels and (1  x) _1S channels, then i_null = i_dysx + i_1S(1  x) and x  0.34. According to this two-channel population model for I_null, the fraction of dysgenic and _1S channels in ₁-null cells would be roughly one-third and two-thirds, respectively. Thus the bulk of the Ca²⁺ channels of I_null would have _1S as their pore subunit. We also considered two alternative explanations, that I_null was produced by a novel DHPR ₁ subunit with a distinct unitary conductance, and that a single type of DHPR complex without was responsible for the estimated single-channel current of ₁-null cells. Both explanations were considered unlikely. First, splice variants of _1S that could account for this putative subunit have not been reported. Second, mutagenesis experiments have shown that the single-channel conductance of Ca²⁺ channels is determined by domains of the ₁ subunit (Yang et al., 1993; Dirksen et al., 1997) and is the same when ₁ is expresssed alone or coexpressed with  subunits (Bourinet et al., 1996).

The maximum open channel probability, p_max, estimated by variance analysis in dysgenic, ₁-null, and some normal cells was significantly higher than previously determined by cell-attached patch recordings. In normal myotubes, p_max measured in many-channel patches in the presence of Bay K 8644 was 0.19 (Dirksen and Beam, 1995). In the absence of Bay K 8644, as in our case, this value is expected to be even lower. On the other hand, p_max is ~0.3 in cardiac L-type Ca²⁺ channels. Similar values were estimated by single-channel recordings and by whole-cell variance analysis in the absence of Bay K 8644 (Tsien et al., 1986). It is entirely possible that a few Ca²⁺ channels in normal, dysgenic, and ₁-null cells may have a high p_max, but the majority of them would have a low p_max and contribute little to the mean Ca²⁺ current or variance. Furthermore, dysgenic channels are intrinsically noisier than normal channels, and if their p_max is similar to that of cardiac channels, they could increase the p_max estimated in the ₁-null cell. In any case, the discrepancy in p_max does not compromise the estimation of single-channel currents. When we separated normal cells with high and low p_max, the single-channel currents estimated in the two groups were similar (not shown). Finally, the mean-variance data were consistent with the noise spectra of I_null and I_dys when measured near steady-state (not shown). The limiting noise power at low frequencies, S(0), and the cutoff frequency, f_1/2, were lower for I_null than for I_dys. Both results were expected, because S(0) increases with the square of the single-channel current and I_dys is a faster current (Conti et al., 1975).

To identify the subunit composition of I_null, we first considered that of I_dys. The latter Ca²⁺ current could conceivably originate from DHPR complexes that include _1C. Chaudhari and Beam (1993) showed that mRNA for the cardiac _1C subunit is abundant in dysgenic and normal muscle. Furthermore, mRNA for _1C has been reported in a dysgenic cell line (Varadi et al., 1995), as well as in primary cultures of normal myotubes (Bulteau et al., 1977) and in normal adult rodent skeletal muscle (Pereon et al., 1997). Cardiac-type Ca²⁺ currents are also present in the dysgenic cell line (Varadi et al., 1995). On the other hand, there appears to be a kinetic mismatch between the appearance of cardiac mRNA, which is higher in young fetal myotubes and declines thereafter (Chaudhari and Beam, 1993), and the appearance of I_dys, which has roughly the same density throughout fetal development (Shimahara and Bournaud, 1991). Because targeting of ₁ subunits to the transverse tubules may require other DHPR subunits (Flucher et al., 1991; Chien et al., 1995), the appearance of the Ca²⁺ current may be controlled by the expression levels of ₁ as well as by other factors. Additional support for a cardiac origin of I_dys comes from its functional profile, which is entirely consistent with that of a cardiac L-type Ca²⁺ current. In agreement with Adams and Beam (1989), our data showed that I_dys activated much faster and was stimulated more strongly by Bay K 8644 than the normal L-type current. Although I_dys did not inactivate during prolonged depolarization, it should be noticed that _1C does not either when expressed in dysgenic myotubes (Tanabe et al., 1990b; Adams et al., 1990). Finally, the mean-variance analysis provides a compelling reason to suspect that I_dys is a cardiac-type current. The estimated single-channel current of ~84 fA in 10 mM Ca²⁺ at +20 mV is consistent with estimations made by the same technique in ventricular myocytes, which were 130 fA in 10 mM external Ba²⁺ at +10 mV (Bean et al., 1984). In summary, our results are entirely consistent with I_dys originating from DHPR complexes that include _1C or an unidentified embryonic homolog of _1C.

Based on the conclusion above, we considered the possibility that I_dys and I_null originated from DHPR complexes of the same "_1C"-like subunit, but that complexes underlying I_null lacked _1a, the isoform found in skeletal muscle and absent in the ₁-null myotube. In heterologous systems, _1a produces a negative shift of the I-V curve of _1C Ca²⁺ channels, decreases the sensitivity of the Ca²⁺ current to Bay K 8644, and increases the Ca²⁺ current density (Singer et al., 1991; Wei et al., 1991; Lory et al., 1993; Nishimura et al., 1993; Perez-Garcia et al., 1995; Kamp et al., 1996). The effect of _1a on the kinetics of activation of _1C Ca²⁺ channels is more complex. Some investigators found an increase in the activation rate (Singer et al., 1991; Wei et al., 1991; Lory et al., 1993); others showed no effect (Itagaki et al., 1992) or even a decrease in the activation rate (Perez-Garcia et al., 1995), none of which seemed to correlate with the expression system. We reasoned that if I_null originated from the same DHPR complex as I_dys but without _1a, differences between I_null and I_dys would be analogous to those observed when _1C is expressed alone or _1C is coexpressed with _1a, respectively. The activation of I_dys at more negative potentials than I_null (Fig. 3) and the faster activation rate of I_dys (Figs. 5 and 6) fit this scenario. With respect to the first observation (Fig. 3), it must be pointed out that not only  subunits but also ₁ subunits produce strong effects on the midpoint of the G-V relationship. In dysgenic myotubes the G-V curve of expressed _1C Ca²⁺ channels is ~20 mV more negative than that of expressed _1S Ca²⁺ channels (Garcia-Martinez et al., 1994). Hence the fact that I_dys activates at more negative potentials than I_null may not be explained solely by the presence of _1a in the Ca²⁺ channel complex of I_dys and its absence from the I_null complex. Against the hypothesis that I_dys and I_null differ by the presence or absence of the _1a subunit, respectively, is the observation that Bay K 8644 stimulated I_dys more strongly than I_null. In heterologous expression systems, Ca²⁺ currents from _1S and _1a subunits or _1C and _1a subunits are always less sensitive to the agonist than those from -deficient complexes (Varadi et al., 1991; Singer et al., 1991; Lory et al., 1992; Itagaki et al., 1992; Hullin et al., 1992). A similar observation has been made in the ₁-null myotube, where expression of _1a results in a reduction in the sensitivity of the rescued Ca²⁺ current to Bay K 8644 (Beurg et al., 1997). Exemptions to this rule are found in the expression of _1C2a (Perez-Reyes et al., 1992; but see Hullin et al., 1992, for a different result), _1C4, and _1S4 (Castellano et al., 1993). In these cases, the sensitivity of the Ca²⁺ current to Bay K 8644 was unchanged by coexpression of ₁ and  subunits. Clearly, however, none of these cases involved _1a. In summary, the higher sensitivity of I_dys to Bay K 8644 and the fact that I_null and I_dys had the same current density are not in keeping with the hypothesis that I_dys and I_null differ only by the presence or absence of DHPR _1a.

The immunodetection of DHPR _1S in ₁-null and its absence in dysgenic myotubes agreed with the hypothesis that _1S is a component of I_null. DHPR _1S was previously thought to be present at extremely low levels in ₁-null myotubes (Gregg et al., 1996). In the present study we used a different and, evidently, a higher-affinity antibody to show significant expression of this subunit in ₁-null myotubes in culture. The distribution of DHPR _1S in ₁-null cells differed from that reported in normal cells in several respects. Unlike in normal myotubes, a significant fraction of ₁-null myotubes did not express DHPR _1S. This observation agreed with the fact that ~30% of ₁-null myotubes had undetectable levels of L-type Ca²⁺ current. Furthermore, the _1S immunolabeling of ₁-null myotubes lacked the punctuate appearance seen in normal cells (Gregg et al., 1996) and was dimmer than the immunofluorescence of normal myotubes, although well above background levels. The nonclustered distribution of _1S in ₁-null cells is consistent with the low density of tetrads observed in freeze fractures (Protasi and Franzini-Armstrong, unpublished observations) and suggests that _1S is not found in triadic junctions. In this respect, the distribution of DHPR _1S in ₁-null myotubes may be similar to the distribution of the DHPR ₂ subunit in dysgenic myotubes (Flucher et al., 1991). Finally, it is important to mention that the present data cannot explain the low density of I_null and the absence of a significant amount of _1S subunits in the transverse tubules previously inferred from charge movements (Strube et al., 1996). Recent studies indicate that DHPR  subunits play a role in targeting ₁ subunits to membrane sites (Chien et al., 1995). Expression of DHPR _1S and  subunits in double-mutant mdg/mdg cchb1^/ myotubes may represent a useful system for exploring this dual role of DHPR  subunits in skeletal muscle.