INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Excitation-contraction (EC) coupling is perhaps the earliest recognized example of Ca²⁺ signaling in an excitable cell (Ebashi et al., 1969). This is a fast process in which a single action potential induces a transient elevation of cytosolic Ca²⁺, which in turn triggers a transient shortening of the cell. EC coupling takes place at specialized junctions between transverse tubules (t-tubules) and sarcoplasmic reticulum (SR) membranes. At this location, the gap between these two membranes is minimal and key proteins are highly concentrated. Two molecular complexes colocalized across the t-tubule/SR junction are the dihydropyridine receptor (DHPR) in the t-tubule membrane and the ryanodine receptor (RyR) in the SR membrane (Block et al., 1988). The well-established paradigm of EC coupling is that in response to depolarization, the L-type Ca²⁺ channel formed by the DHPR produces a signal that briefly opens RyR channels, leading to the release of SR-stored Ca²⁺.

DHPRs of skeletal muscle consist of ₁, ₂/, , and  subunits (Perez-Reyes and Schneider, 1994).  subunits are ~55-65-kDa proteins that bind strongly to the _1S subunit at the intracellular loop between repeats I and II. Biochemical studies showed that  subunits are present as a 1:1 complex with _1S and other subunits (Leung et al., 1987). The 18-amino acid motif in the I-II loop of ₁ that binds  (identified as the AID region; Pragnell et al., 1994) and the 30-amino acid motif in  that binds ₁ (identified as the BID region; De Waard et al., 1994) are highly conserved among ₁ and  subunits. The AID/BID interaction is highly specific, has an affinity in the nanomolar range, and survives membrane solubilization (Scott et al., 1997). Both the equimolar stoichiometry and the tight binding suggest ₁ and  are unlikely to separate from each other during the lifetime of the DHPR complex, although pools of free  subunits are known to be present in cells (Wicher et al., 1995). There are four  subunit genes and each produces multiple isoforms by use of alternate exons.  isoforms can be divided into five regions based on the amount of identity between them (see Fig. 1). Two highly conserved central regions (regions 2 and 4) are flanked by highly divergent N- and C-termini (regions 1 and 5) and linker regions (region 3). Region 4 contains the BID region essential for binding to the ₁ subunit (De Waard et al., 1994). Features of the primary structure (Ruth et al., 1989), biochemical data (Wicher et al., 1995; Scott et al., 1997), and recent predictions based on sequence homology searches (Hanlon et al., 1999), have suggested the  subunit is a peripheral membrane protein rich in secondary structure with homology domains typical of signaling proteins (SH3 domain) and receptor clustering proteins (MAGUK domain) (Craven and Bredt, 1998).

A knockout of the mouse 1 gene, encoding isoforms expressed in skeletal muscle (1a) and brain (1b, 1c) (Powers et al., 1992), was previously described (Gregg et al., 1996). 1-null mice die at birth due to the lack of EC coupling in the skeletal musculature. 1-null myotubes fail to contract in response to electrical stimulation despite the presence of normal action potentials, a normal Ca²⁺ storage capacity, and normal caffeine-sensitive Ca²⁺ release. 1-null cells have a low density of L-type Ca²⁺ current and charge movements and do not produce Ca²⁺ transients in response to depolarization (Strube et al., 1996, 1998). However, Ca²⁺ sparks due to the activity of ryanodine receptors are highly abundant in these cells (Conklin et al., 1999). These studies have suggested that 1-null cells fail to transduce depolarization into SR Ca²⁺ release due to one of two fundamental reasons. Either the density of DHPRs on the cell membrane of 1-null cells is too low to produce Ca²⁺ transients detectable by available techniques, or the absence of  renders membrane-located DHPRs unable to initiate EC coupling. In the former case, the EC coupling null phenotype would originate from the mistargeting of otherwise functional DHPRs. In the latter case, the null phenotype would primarily reflect an intrinsic dysfunction of the DHPR. In order to distinguish between these possibilities, we expressed different  isoforms in null cells and investigated the recovered phenotype. Expression of the skeletal muscle 1a isoform results in a quantitative recovery of the L-type Ca²⁺ current density, the intramembrane charge movement density, and the amplitude and voltage dependence of intracellular Ca²⁺ transients (Beurg et al., 1997). In contrast, expression of the nonskeletal muscle 2a isoform produced an entirely different result (Beurg et al., 1999b). 2a, like 1a, restored a Ca²⁺ current with a density, voltage-dependence, and kinetics identical to that of 1a-transfected cells. Yet 2a could not entirely restore skeletal-type EC coupling since Ca²⁺ transients evoked by voltage were significantly smaller at all potentials (Beurg et al., 1999b). These observations suggested that a unique region of 1a was required for normal EC coupling in the 1-null cell. To test this hypothesis further, we expressed several deletion mutants of 1a, 2a, and chimeric 2a-1a isoforms. Whole-cell voltage-clamp and confocal imaging analyses showed that a quantitative recovery of the EC coupling in 1-null myotubes required a 1a isoform with an intact C-terminus. Thus a region distinct from the BID is required for the normal EC coupling function of the DHPR. Part of these results appeared in abstract form (Beurg et al., 1999a).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Primary cultures of mouse myotubes

Primary cultures were prepared from hindlimbs of 18-day-old 1-null embryos as described elsewhere (Beurg et al., 1997). Dissected muscles were incubated for 9 min at 37°C in Ca²⁺/Mg²⁺-free Hanks' balanced salt solution (136.9 mM NaCl, 3 mM KCl, 0.44 mM KH₂PO₄, 0.34 mM NaHPO₄, 4.2 mM NaHCO₃, 5.5 mM glucose, pH 7.2) containing 0.25% (w/v) trypsin and 0.05% (w/v) pancreatin (Sigma, St. Louis, MO). Mononucleated cells were resuspended in plating medium containing 78% Dulbecco's modified Eagle's medium with low glucose (DMEM, Gibco BRL, Gaithersburg, MD), 10% horse serum (HS, Sigma), 10% fetal bovine serum (FBS, Sigma), 2% chicken embryo extract (CEE, Gibco), and plated on plastic culture dishes coated with gelatin at a density of ~1 × 10⁴ cells per dish. Cultures were grown at 37°C in 8% CO₂. After the fusion of myoblasts (~7 days), the medium was replaced with an FBS-free medium (88.75% DMEM, 10% horse serum, 1.25% CEE) and cells were incubated in 5% CO₂. All media contained 0.1% v/v penicillin and streptomycin (Sigma).

cDNA transfection

cDNAs were subcloned into a pSG5 expression plasmid (Stratagene, La Jolla, CA) containing the early simian virus-40 (SV40) promoter, an 1-globin intron to enhance RNA processing, and an SV40 polyadenylation signal. The original vector was modified so that 11 amino acids of the phage T7 gene 10 protein are fused to the N-terminus of the expressed  subunits. The T7 tagged-1a subunit had a functional expression indistinguishable from the untagged 1a subunit. We also added AgeI and NotI restriction enzyme sites downstream of the T7 tag to simplify cloning of the modified  subunits. Cotransfection of the pSG5 expression plasmid and a separate marker plasmid encoding the T-cell membrane antigen CD8 was performed with the polyamine LT-1 (Panvera, Madison, WI). Cotransfected cells were recognized by incubation with CD8 antibody beads (Dynal, Oslo, Norway).

cDNA constructs

Deletion and chimeric constructs of  subunits (Fig. 1) were made using PCR strategies. For deletion constructs, two oligonucleotide primers were designed to encompass the region of interest. Each primer had 20-25 bases identical to the original sequence and an additional 10-15 bases that resulted in an amplified product with an AgeI site at the 5' end, and a stop codon and NotI restriction site at the 3' end. The PCR products were subcloned into the pCR-Blunt vector (Invitrogen Inc., Carlsbad, CA), excised by digestion with AgeI and NotI and cloned into pSG5.

wt 1a

wt 2a

wt 1c

1-3't

1-5't

2-3't1

2-3't2

Chimeric 2-1

2-1-3't

Ca²⁺current and charge movements

Whole-cell recordings were performed as described previously (Strube et al., 1996) with an Axopatch 200B amplifier (Axon Instruments, Foster City, CA). Linear capacitance and leak currents were compensated with the circuit provided by the manufacturer. Effective series resistance was compensated up to the point of amplifier oscillation with the Axopatch circuit. All experiments were performed at room temperature. The external solution was (in mM) 130 TEA methanesulfonate, 10 CaCl₂, 1 MgCl₂, 10³ TTX, and 10 HEPES titrated with TEA(OH) to pH 7.4. The pipette solution consisted of (in mM) 140 cesium aspartate, 5 MgCl₂, 0.1 EGTA (when Ca²⁺ transients were recorded), or 5 EGTA (all other recordings), and 10 MOPS titrated with CsOH to pH 7.2. Patch pipettes had a resistance of 2-5 M when filled with the pipette solution. For recordings of charge movement, the external solution was supplemented with 0.5 mM CdCl₂ and 0.1 mM LaCl₃ to block the ionic Ca²⁺ currents. A prepulse protocol previously described (Beurg et al., 1999b) was used to measure the immobilization-resistant component of charge movement. Voltage was first stepped up from holding potential 80 mV to 20 mV for 1 s, then to 50 mV for 5 ms, then to test potential P for 25 ms, then to 50 mV for 30 ms and finally to the 80 mV holding potential. Subtraction of linear components was assisted by a P/4 procedure following the pulse paradigm listed above. P/4 pulses were in the negative direction, had a duration of 25 ms, and were separated by 500 ms.

Confocal fluorescence microscopy

Confocal line-scan measurements were performed as described elsewhere (Conklin et al., 1999). Cells were loaded with 4 µM fluo-3 acetoxymethyl (AM) ester (Molecular Probes, Eugene, OR) for 20 to 40 min at room temperature. A 1-mg sample of fluo-3 AM (Molecular Probes) was dissolved in 1 ml DMSO and kept frozen until use. All experiments were performed at room temperature. Cells were viewed with an inverted Olympus microscope with a 20× objective (N.A. = 0.4) and an Olympus Fluoview confocal attachment (Melville, NY). The 488-nm spectrum line necessary for fluo-3 excitation was provided by a 5 mW argon laser attenuated to 20% with neutral density filters. The fluorescence intensity, F, was calculated by densitometric scanning of line-scan images and was averaged over the entire width of the cell. The fluorescence intensity Fo was averaged in the same manner from areas of the same image before the voltage pulse.

Immunostaining

Cells were transfected with T7-tagged  constructs for four days and later fixed in 100% methanol and processed for immunostaining as previously described (Gregg et al., 1996). The primary antibody was a mouse monoclonal against the T7 epitope (Novagen, Madison, WI) and was used at a dilution of 1:1000. The secondary antibody was a fluorescein conjugated polyclonal goat anti-mouse IgG (Boehringer Mannheim, Indianapolis, IN) and was used at a dilution of 1:1000. Confocal images of 1024 × 1024 pixels (0.35 µm/pixel) were obtained in the Olympus Fluoview using the 488-nm spectral line for dye excitation and a 40× oil-immersion objective (N.A. 1.3) for capturing emission. Images were Kalman-averaged three times and the pixel intensity displayed as 16 levels of gray in reverse. All images were acquired using minimal laser power (6% of maximum 5 mW) and predetermined PMT settings to avoid pixel saturation and for accuracy in the comparison of images.

Curve-fitting

For each cell the voltage-dependence of charge movements (Q), Ca²⁺ conductance (G), and peak intracellular Ca²⁺ (F/F) was fitted according to a Boltzmann distribution

(1)

Where A_max was either Q_max, G_max, or F/F_max, V_1/2 is the potential at which A = A_max/2, and k is the slope factor. The time constant, ₁, describing activation of the Ca²⁺ current was obtained from a fit of the pulse current at each voltage according to

(2)

Where K is constant and ₂ describes inactivation. Parameters from a fit of separate cells are shown in Table 1. Parameters from a fit of averages of many cells (population averages) are shown in Figs. 4, 6, and 8.

View this table: [in this window] [in a new window]   TABLE 1   Parameters of the Ca2+ conductance, charge movement, and Ca2+ transient expressed by  constructs in 1-null myotubes

cDNA sequencing

All  constructs were sequenced before their use in experiments at the Biotechnology Center, University of Wisconsin.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The  subunits analyzed in this study are shown in Fig. 1. Sequence comparison between the full-length 1a and 2a isoforms revealed two highly conserved central regions (regions 2 and 4), a nonconserved linker region between the two conserved domains, and distinct N- and C-termini. We tested the participation of the nonconserved regions 1, 3, and 5 of 1a in the recovery of Ca²⁺ conductance, charge movements, and EC coupling in 1-null cells. These regions were respectively deleted in constructs 1-5't, the splice-variant 1c (Powers et al., 1992), and 1-3't. In addition, we tested region 5 of 2a (constructs 2-3't and 2-3t2) and region 5 of a chimera composed of an N-terminus half of 2a and a C-terminus half of 1a (constructs 2-1 and 2-1-3't). Measurements were made in whole-cell voltage-clamped myotubes at day 8 to 12 after cDNA transfection. We previously showed that within this period, the Ca²⁺ conductance of transfected 1-null cells remains relatively constant (Beurg et al., 1997). Therefore, a precise synchronization of cell cultures was not required for a quantitative comparison of the functional expression of DHPRs in different batches of transfected cells. All transfected  constructs carried an 11-amino acid T7 tag at the N-terminus, which was first tested in the wt 1a cDNA and was found not to interfere with function. This epitope was useful for determining whether a given construct was expressed in 1-null cells.

View larger version (23K): [in this window] [in a new window]   FIGURE 1   Alignment to scale of the amino acid sequence of the tested  constructs. The boxes indicate conserved domains described elsewhere (Perez-Reyes and Schneider, 1994). The amino acid coordinates of each construct are indicated in Materials and Methods.

Fig. 2 shows close-up confocal views of myotubes fixed and immunostained with T7 antibody and a fluorescein-conjugated secondary antibody. There was abundant expression of each of the tested constructs throughout the length of myotubes, and in many cases expression was heavily concentrated in the cell periphery. The latter is consistent with the known location of DHPRs in the periphery of cultured myotubes where couplings are established between the plasma membrane and the sarcoplasmic reticulum membrane (Takekura et al., 1994). The CD8 cDNA was used as a transfection marker and micron-size beads, with absorbed CD8 antibody, were used to identify transfected cells (see asterisks). Better than 95% of cells expressing CD8 also expressed  as determined from the coincidence of the T7 immunostain and CD8 beads in a given cell and the coincidence of CD8 beads and a high density of L-type Ca²⁺ current in a given cell.

View larger version (88K): [in this window] [in a new window]   FIGURE 2   Immunofluorescence of 1-null myotubes coexpressing a T7 tagged  construct. Cells transfected with T7 tagged constructs and CD8 cDNA were incubated with CD8 antibody beads, fixed, and stained with T7 primary/fluorescein-conjugated secondary antibodies. Confocal images of 1024 × 1024 pixels (360 × 360 µm) were converted to a 16-level inverted gray scale so that high-intensity pixels appear black and low-intensity pixels appear white. Asterisks indicate some CD8 antibody beads bound to cotransfected cells. NF indicates nontransfected myotubes in the same field of 1a transfected cells (A). Other panels show expression of T7 tagged 1c (B), 1-5't (C), 1-3't (D), 2a (E), 2-3't (F), 2-3't2 (G), 2-1 (H), and 2-1-3't (I).

Fig. 3 shows the L-type Ca²⁺ currents of cells transfected with each of the tested  constructs in response to the pulse potentials indicated in the top left set of traces. Each whole-cell clamped myotube was subjected to a total of 20 voltage pulses of increasing amplitude and a constant duration of 300 ms starting from a holding potential of 40 mV. For ease of comparison, only four traces of currents are shown in each case. Currents have been normalized according to the cell capacitance. The current scale is the same for all cells except for the two cells expressing the chimeric 2-1 constructs. The Ca²⁺ current recovered by the constructs had a threshold for activation more positive than 30 mV, had remarkably slow activation kinetics, and a fast deactivation. These features are entirely consistent with the known properties of skeletal L-type Ca²⁺ currents of control (wt) mouse myotubes in cell culture (Garcia et al., 1994a; Beurg et al., 1997). These features contrast with those of I-null, the Ca²⁺ current of nontransfected 1-null myotubes, which is much faster and has a significantly lower density (Beurg et al., 1997; Strube et al., 1998).

View larger version (22K): [in this window] [in a new window]   FIGURE 3   Traces of Ca2+ currents are from the same 1-null cell expressing the indicated  construct. The pulse duration was 300 ms and the holding potential was 40 mV. The vertical scale is 3pA/pF in all cases.

The voltage-dependence of the Ca²⁺ conductance recovered by each of the tested constructs is shown in Fig. 4. The Ca²⁺ conductance was estimated from the extrapolated reversal potential and the end-pulse current in each of the 20 300-ms pulses. The curves are a Boltzmann fit of the population mean with parameters indicated in the figure legend. The curve without symbols corresponds to the conductance of nontransfected cells obtained in a previous study (Beurg et al., 1997). In addition to these data, the Boltzmann parameters of the Ca²⁺ conductance averaged for 9 to 15 cells in each case are shown in Table 1. In Fig. 4 A we examined G-V curves of cells expressing deletion constructs of the 1a isoform. Deletion of region 3 between the two conserved domains (construct 1c) or deletion of almost the entire region 1 (construct 1-5't) had no effect on the recovered maximum Ca²⁺ conductance, which for these constructs and for wt 1a was ~160 pS/pF. However, the deletion of region 5 (construct 1-3't) resulted in a 1.8-fold (161/88) reduction in the recovered maximum Ca²⁺ conductance that was statistically significant (unpaired t-test with p = 0.0001). Scaled G-V curves of cells expressing 1-3't and wt 1a are shown in Fig. 4 A, top. The 3' truncation had no effect of the steepness of the G-V curve, although it produced an ~5 mV positive shift, which according to an unpaired t-test was not significant. In Fig. 4 B we examined whether a 3' truncation of 2a, which in full-length was shown to express skeletal L-type Ca²⁺ currents at a density similar to that of control (wt) myotubes (Beurg et al., 1999b), also curtailed the expressed Ca²⁺ current density. A partial deletion of region 5 of 2a (construct 2-3't) or a complete deletion of the 185 residues encompassing region 5 of 2a (construct 2-3't2) had no effect on the maximum Ca²⁺ conductance or the steepness of the G-V curve (unpaired t-tests with p > 0.8). Both parameters were similar to those of cells expressing 1a (Table 1). However, the more severe 2-3't2 truncation produced an ~10 mV positive shift of the G-V curve, shown in Fig. 4 B, top, that was statistically significant (unpaired t-test with p < 0.04). In Fig. 4 C we examined the voltage-dependence of the Ca²⁺ conductance produced by the 2-1 chimera and the 3' truncated chimera. The "full-length" chimera expressed a remarkably large specific Ca²⁺ conductance of ~320 pS/pF or twice that of wt 1a or wt 2a. Truncation of the 3' end of this chimera, encompassing the entire region 5 of 1a, resulted in a twofold (320/160) reduction in Ca²⁺ conductance that was highly significant (unpaired t-test with p = 0.001). The scaled G-V relationships shown in Fig. 4 C, top revealed that the steepness and midpoints of both curves did not change. In summary, the expression of wt 1a or wt 2a in 1-null cells restored a Ca²⁺ current with a density typical of normal (wt) myotubes (Beurg et al., 1997). The voltage-dependence of the Ca²⁺ currents strongly suggests these must have originated from complexes of , _1S, and the other subunits of the skeletal DHPR complex. A deletion of the nonconserved region 5 of 1a, but not regions 1 or 3, resulted in a significant reduction of the expressed Ca²⁺ current. This reduction was specific for region 5 of 1a as demonstrated by deletion approaches using wt 2a and the 2-1 chimera.

View larger version (27K): [in this window] [in a new window]   FIGURE 4   Voltage-dependence of the population average Ca2+ conductance in (A) 1a (n = 9), 1c (n = 12), 1-5't (n = 10), and 1-3't (n = 15); (B) in 2a (n = 15), 2-3't (n = 10), 2-3't2 (n = 12); and (C) 2-1 (n = 11) and 2-1-3't (n = 14) transfected myotubes. The curve without symbols represents the Boltzmann fit of G-V curve of nontransfected cells (Gmax = 16.2 pS/pF; V1/2 = 21.5 mV, and k = 9.4 mV). Curves correspond to a Boltzmann fit of the population mean G-V curve. Parameters of the fit were in (A) Gmax = 162.3, 168.5, 165.5, 87.6 pS/pF; V1/2 = 14.2, 9.9, 16.6, 20 mV, and k = 5.7, 4.6, 6.1, 5.8 mV for 1a, 1c, 1-5't, and 1-3't, respectively; (B) Gmax = 152.7, 149, 136.8 pS/pF; V1/2 = 10.4, 14.2, 19.7 mV and k = 5.4, 5.8, 5.8 mV for 2a, 2-3't, 2-3't2; and (C) Gmax = 320.8, 159.8 pS/pF; V1/2 = 18.8, 18.3 mV and k = 5.7, 5.5 mV for 2-1 and 2-1-3't transfected cells, respectively. The top panels show the conductance normalized according to the mean maximum (Gmax) of each group of cells.

The kinetics of activation of the Ca²⁺ current of skeletal muscle is the slowest among voltage-gated Ca²⁺ channels, and this characteristic is determined in part by repeat I of the _1S subunit (Tanabe et al., 1991). Therefore, the kinetics of the Ca²⁺ current recovered in 1-null cells should provide critical information on whether the expressed  subunits rescued a skeletal-type DHPR. In these experiments we used a 1.5-s depolarizing pulse from a holding potential of 40 mV to fit the activation and inactivation phases of the Ca²⁺ current. However, none of the constructs altered the inactivation rate to any great extent and in all cases the peak current inactivated <20% at the end of this relatively long pulse. The pulse current was fitted with Eq. 2, which conforms to a linear kinetic scheme with closed, open, and inactive states and assumes that for a sufficiently long pulse, inactivation is complete. Because pulses longer than 1.5 s invariably resulted in a loss of the pipette seal, this assumption could not be verified. In all cases we found an excellent agreement between the fit and the pulse current as previously shown for 1-null cells expressing wt 1a or wt 2a (Beurg et al., 1997, 1999b).

Fig. 5 shows the time constant of activation, ₁ of Eq. 2, fitted to the Ca²⁺ current recovered by each  construct at positive potentials. In this range of potentials, the activation rate of the recovered Ca²⁺ currents slowed for increasingly positive potentials by a factor of ~2, which is characteristic of the slow skeletal L-type Ca²⁺ channel (Strube et al., 1996; Dirksen and Beam, 1995). Data are shown for each of the constructs labeled with the same symbols as in the previous figures. With two exceptions, the activation time constant of any two  constructs within each panel (A-C) in Fig. 5 were not significantly different at any test potential. One exception was the activation of 1a-3't (black triangles), which at +30, +40, or +50 mV was slightly faster than that of wt 1a (black circles) or the other  constructs according to unpaired t-tests at each voltage (p < 0.05). The other was the activation of 2a-3't (white squares) which at +10, +20, +30, or +40 mV was slightly slower than that of wt 2a (white circles) (p < 0.04). The activation time constants for all constructs, except for 1a-3't and 2a-3't, averaged 50 to 70 ms at +30 mV. These values agreed with previous determinations in normal myotubes (Strube et al., 1996) and in myotubes expressing chimeras of _1S and _1C when repeat I was from _1S (Tanabe et al., 1991). The slow activation observed in cells transfected with  constructs suggested that all  constructs formed complexes with _1S, rather than with _1C-type isoforms that could potentially be expressed in the myotube (Chaudhari and Beam, 1993; Pereon et al., 1997). We also compared the inactivation time constant, ₂ of Eq. 2. The inactivation time constant, fitted with the limitation of the pulse duration discussed above, was 4.5 ± 1.3 s for cells expressing 1a, 4.5 ± 0.67 s for cells expressing 2a, and 4.8 ± 0.5 s for cells expressing the 2-1 chimera. Thus, notwithstanding the two exceptions described above that produced mild changes in activation kinetics, the main conclusion from these experiments was that the kinetics of activation and perhaps also that of inactivation of the L-type Ca²⁺ current rescued by  constructs in 1-null cells was either weakly modified or not modified at all by  constructs.

View larger version (17K): [in this window] [in a new window]   FIGURE 5   Voltage-dependence of the time constant of activation 1 (mean ± SE) obtained from the fit in Eq. 2 for the indicated number of cells. The symbols represent the same cell types shown in all figures (see Fig. 4).

The bulk of the immobilization-resistant nonlinear charge movements in myotubes in culture originates from DHPRs that include the _1S subunit (Adams et al., 1990). These DHPR-mediated charge movements are directly responsible for the restoration in dysgenic (_1S-null) myotubes of skeletal-type EC coupling (Tanabe et al., 1990). Both observations have strongly suggested that charge movements provide a critical index of the density of functional DHPRs that is independent of whether these DHPRs function as L-type Ca²⁺ channels. Here we measured the immobilization-resistant charge movements with a pulse of 20 ms and a protocol design to minimize contamination by the Na⁺ channel gating current and ionic current (Strube et al., 1996). Charge movements were calculated by integration of the ON component on the nonlinear capacitance after verification that ON and OFF components differed by 20% or less.

Fig. 6 shows population average Q-V relationships for the tested  constructs. Cells were rejected unless the ON and OFF components agreed within 20%. The curves correspond to a Boltzmann fit of the population average Q-V curves with parameters indicated in the figure legend. The curve without symbols corresponds to the mean charge movements of nontransfected 1-null cells obtained in a previous study using the same pulse protocol (Beurg et al., 1997). Averages of Boltzmann parameters fitted separately to each cell are shown in Table 1. The onset of charge movements occurred at ~10 mV and increased with voltage until a plateau was reached at potentials more positive than +40 mV. Fig. 6 A shows Q-V relationships of cells expressing deletions mutants of 1a. In all cases, the maximum charge movements, Q_max, were significantly larger than the Q_max of nontransfected cells, which averaged 2.5 ± 0.2 nC/µF (Beurg et al., 1997). This result indicated a robust recovery of membrane-associated DHPRs by the 1 constructs. The Q_max of cells expressing wt 1a was ~6.5 nC/µF, a value that agreed with determinations in normal (wt) myotubes and in _1S-transfected dysgenic myotubes (Garcia et al., 1994a; Strube et al., 1996; Beurg et al., 1997). In addition, the Q_max of cells expressing 1-3't or 1-5't was not significantly different from that of cells expressing wt 1a. Furthermore, neither the midpoint nor the steepness of these Q-V curves was significantly different (see Table 1). However, the Q_max of cells expressing 1c was the lowest for this group of constructs, yet the difference between 1c and wt 1a (5 ± 0.8 vs. 6.8 ± 0.9 nC/µF, respectively) was not significant (p = 0.08).

View larger version (16K): [in this window] [in a new window]   FIGURE 6   Voltage-dependence of immobilization-resistant charge movements (mean ± SE) in 1a (n = 6), 1c (n = 7), 1-5't (n = 5), and 1-3't (n = 5); (B) in 2a (n = 11), 2-3't (n = 7), 2-3't2 (n = 5); and (C) 2-1 (n = 7) and 2-1-3't (n = 5) transfected myotubes. Curves correspond to a Boltzmann fit of the population mean Q-V curve. Parameters of the fit are (A) Qmax = 6.5, 5.3, 5.9, 6.3 nC/µF; V1/2 = 16.9, 21.8, 20.6, 19.6 mV; k = 13.9, 11.5, 11.9, 10.7 mV for 1a, 1c, 1-5't and 1-3't, respectively; (B) Qmax = 2.6, 4.9, 3.9 nC/µF; V1/2 = 12, 18.6, 21.4 mV; k = 15.5, 14.9, 10.4 mV for 2a, 2-3't, 2-3't2; and (C) Qmax = 5.4, 3.8 nC/µF; V1/2 = 25.8, 19.4 mV; k = 14, 10.5 mV for 2-1 and 2-1-3't transfected cells, respectively. The curve without symbols is a fit of Q-V curve of nontransfected cells (Qmax = 2.5 nC/µF; V1/2 = 6 mV; k = 12 mV).

The results of Fig. 6 A clearly demonstrated that the lower Ca²⁺ conductance of cells expressing the 3'-truncated 1 subunit could not be explained by a lower density of membrane-associated DHPRs recovered by this construct. In Fig. 6 B we examined the charge movements produced by the truncated 2a constructs. The Q_max of cells expressing wt 2a was much lower than that of wt 1a-expressing cells and, in fact, indistinguishable from that of nontransfected cells. Although this result agreed with a previous determination (Beurg et al., 1999b), it is a puzzling one because the density of the Ca²⁺ currents recovered by wt 2a was similar to that recovered by wt 1a (see Fig. 4 and Table 1). The previous study showed that the same difference in Q_max between 1a and 2a expressing cells was measured from a more negative holding potential (120 mV), indicating that the low Q_max of 2a expressing cells was not due to a selective immobilization of charge produced by 2a (Beurg et al., 1999b). We thus surmise that the charge movements associated with the opening of 2a-recovered Ca²⁺ channels must have been masked by the background charge movements present in the 1-null myotube. Quite surprisingly, Fig. 6 B shows that the C-terminus deletion mutants 2-3't and 2-3't2 expressed charge movements significantly higher than those of the full-length construct. Table 1 shows that in the case of 2-3't there was a statistically significant (unpaired t-test with p = 0.00001) doubling of the Q_max from 2.6 ± 0.2 nC/µF produced by wt 2a to 4.9 ± 0.4 nC/µF produced by the 3' truncated 2a. This result is in contrast with the identical Ca²⁺ conductance produced by both constructs (Table 1), which were 152.7 ± 12 pS/pF and 149 ± 13.8 pS/pF, respectively. A complete deletion of region 5 (construct 2-3't2) produce a 1.5-fold increase in Q_max that also was highly significant (p = 0.0008). Both results clearly indicated that the C-terminus of 2a interfered with the expression of DHPR charge movements. Furthermore, the recovery of charge movements without a concomitant recovery of L-type Ca²⁺ current indicated that in the DHPR these two events, namely voltage-induced movements of electrical charges and opening the Ca²⁺ channel, are not uniquely associated. In Fig. 6 C we examined the charge movements produced by the 2-1 chimera and its 3' truncated form. The Q_max produced by this chimera was closer to that produced by wt 1a than to that produced by wt 2. The truncation slightly reduced the Q_max, although the difference was not statistically significant (unpaired t-test with p = 0.4). In summary, all  constructs except wt 2a recovered saturable movements of charge with a maximum density significantly higher than the background charge movements of nontransfected cells. Thus, except for wt 2a, all  constructs recovered electrically detectable amounts of membrane-associated DHPRs. A Boltzmann fit of the Q-V relationships (Table 1) showed that 1) the midpoints were not modified by the tested  isoforms; 2) the steepness factor was modestly affected; 3) the Q_max produced by deletion mutants of 1a were not affected; 4) the Q_max produced by the C-terminus truncated 2a constructs was increased; and 5) there was no unique correlation between the density of expressed charge movements and the density of expressed Ca²⁺ currents in any of the three groups (panels A-C) of  isoforms tested.

The contribution of the expressed  constructs to EC coupling was examined by measurements of intracellular Ca²⁺ using confocal line-scan imaging of fluo-3 fluorescence. Transfected cells were loaded with the cell-permeant form of fluo-3 and were whole-cell voltage-clamped. In some nontransfected cells slowly evolving Ca²⁺ "waves" could be evoked by depolarization from 80 mV that were presumably due to Ca²⁺-induced Ca²⁺ release produced by Ca²⁺ entry via T-type channels (not shown). To avoid the contribution of the T-type current, the holding potential was set at 40 mV (Strube et al., 1996). Extensive controls (15 of 15 cells) convinced us that no changes in cytosolic Ca²⁺ occurred in nontransfected 1-null cells from this holding potential.

Fig. 7 shows Ca²⁺ transients in cells expressing the indicated  constructs stimulated by a 50-ms depolarization to +70 mV from 40 mV. In the line-scan images time increases from left to right. The depolarizing pulse was delivered 100 ms after the start of the line scan as indicated in the bottom of the figure. The line-scan direction was in most instances across the myotube width rather than parallel to the length of the myotube. The magnification was the same in all cases and was adjusted so that the top and bottom borders of the line-scan image would roughly correspond with the edges of the cell. Also, the laser power, photomultiplier gain, and pixel size were kept constant to minimize errors when comparing the fluorescence of different cells. The traces under each image correspond to the fluorescence in F/F units averaged across the entire line-scan. Ca²⁺ release started at the onset of the depolarization and peaked at ~100 ms in all cases. The decay phase of the transient outlasted the depolarization by a significant amount of time, in agreement with studies in normal rat and mouse myotubes in culture (Garcia and Beam, 1994b; Beurg et al., 1997, 1999b). The peak fluorescence was in excess of four F/F for cells expressing the endogenous wt 1a construct and for cells expressing the 2-1 chimera. In both cases, the C-terminus truncation resulted in a dramatic decease in the peak fluorescence to <2 F/F units. However, cells expressing wt 2a produced a modest Ca²⁺ transient that increased when this isoform was truncated.

View larger version (43K): [in this window] [in a new window]   FIGURE 7   Confocal line-scan images of fluo-3 fluorescence in response to a 50-ms pulse to +70 mV from a holding potential of 40 mV. The curves show the time course of the normalized fluorescence intensity (F/F) obtained by integration of the image fluorescence. Images have a dimension of 2.05 s (horizontal) and 45, 42.5, 21.5, 43, 28.5, 73.7 µm for 1a, 1-3't, 2a, 2-3't2, 2-1, and 2-1-3't, respectively.

The peak fluorescence at different voltages is shown in F/F units in Fig. 8. The curves correspond to a Boltzmann fit of the population average F/F-V curves with parameters indicated in the figure legend. The curve without data corresponds to the fluo-3 fluorescence of nontransfected cells obtained in a previous study (Beurg et al., 1997). Averages of Boltzmann parameters fitted separately to each cell are shown in Table 1. All  constructs, without exception, recovered F/F-V curves that saturated at large positive potentials. This is expected of skeletal-type EC coupling but not of Ca²⁺-entry dependent (cardiac-type) EC coupling (Garcia and Beam, 1994b). A bell-shaped F/F-V curve, indicative of cardiac-type EC coupling, was restored by coexpression of the cardiac isoform _1C and wt 1a using the same pulse protocol and confocal imaging technique (Ahern et al., 1999). The recovery of skeletal EC coupling by the  constructs is entirely consistent with a recovery of DHPRs that include _1S, a conclusion reached earlier by analyses of the voltage-dependence and kinetics of the recovered Ca²⁺ current. Deletion analyses of 1a in Fig. 8 A indicated that removal of region 3 (construct 1c) did not alter the maximum amplitude of Ca²⁺ transients reached at positive potentials. However, a significant decrease in the maximum F/F was produced by removal of the C-terminus (construct 1a-3't) or by partial removal of the N-terminus (construct 1a-5't) of 1a.

View larger version (15K): [in this window] [in a new window]   FIGURE 8   Ca2+ transients were elicited by 50-ms voltage steps to the indicated potential from a holding potential of 40 mV. F/F at the peak of the transient (population mean ± SE) was obtained by integration of the confocal line-scan images. Curves correspond to Boltzmann fit of the population mean F/F-V curve. Parameters of the fit are (A) F/Fmax = 3.1, 3.4, 1.8, 0.8; V1/2 = 3, 2.4, 4.1, 0.8 mV; k = 8.2, 6.8, 8.3, 6.4 mV for 1a (n = 9), 1c (n = 11), 1-5't (n = 6), and 1-3't (n = 3), respectively; (B) F/Fmax = 1, 2, 0.9; V1/2 = 0.6, 1.5, 4.5 mV; k = 7, 5.8, 10.3 mV for 2a (n = 7), 2-3't (n = 4), 2-3't2 (n = 6); and (C) F/Fmax = 3.7, 1.4; V1/2 = 5.4, 6.3 mV; k = 8.4, 7.6 mV for 2-1 (n = 7) and 2-1-3't (n = 10) transfected cells, respectively.

As shown in Table 1, averages of several cells indicated a fivefold reduction in maximum F/F in the former case (3.3/0.65) and a 1.7-fold in the latter case (3.3/1.9) with high statistical significance in both cases (unpaired t-tests with p = 0.0002 and 0.05, respectively). Because the Q_max recovered by each of the two constructs was not significantly different from that recovered by wt 1a (unpaired t-test, p > 0.2), the reduction in voltage-evoked Ca²⁺ release could not be explained by a reduction in the amount of DHPR complexes present in the cell surface of these myotubes. More likely, the deleted regions were necessary for the EC coupling function of the DHPR. In Fig. 8 B we investigated whether the low EC coupling produced by wt 2a was related to the C-terminus of 2a, which has a unique composition and is much larger in mass than region 5 of 1a. The deletion of 115 amino acids from the C-terminus of 2a (construct 2a-3't) resulted in a 1.8-fold restoration (2.0/1.1) of F/F_max that was marginally significant compared to that produced by wt 2a (unpaired t-test with p = 0.06). This increase in F/F_max was consistent with the partial restoration of Q_max produced by the same construct, and suggested that the C-terminus of 2a could interfere with either the targeting of DHPRs to the cell surface, or EC coupling, or both. However, further C-truncation of 70 amino acids (construct 2a-3't2) did not increase F/F_max. In fact, the opposite was the case as F/F_max of 2a-3't2 was less than that of 2a-3't, although the difference was not significant (unpaired t-test with p = 0.18). The loss in activity produced by the deeper truncation could be caused by incorrect protein folding, because in this region there is a partially conserved  helix/ strand motif that was removed by the deletion (Hanlon et al., 1999).

Fig. 8 C shows that the 2-1 chimera fully restored the F/F_max present in cells expressing wt 1a. This indicated that the N-terminal half of 1a was interchangeable with that of 2a, and thus presumably the C-terminus half of 1a was specifically required for enhancing EC coupling. If this were the case, the C-terminus truncation of the chimera should produce the same result as the C-terminus truncation of wt 1a. This result is also shown in Fig. 8 C. The F/F_max of the 2-1-3't truncated chimera was reduced 2.7-fold (3.8/1.4) compared to that produced by the "full-length" chimera, and the difference was statistically significant (unpaired t-test with p = 0.005). In summary, the confocal imaging of Ca²⁺ transients in 1-null cells expressing  constructs demonstrated that 1) the linker region 3 of 1a is nonessential for skeletal-type EC coupling; 2) a domain of the 1a subunit, near the C-terminus (region 5), and the N-terminus (region 1) strengthened skeletal-type EC coupling but did not affect Q_max; 3) a domain of the 2a subunit, also near the C-terminus (region 5), weakened EC coupling; and 4) the N-terminus half of 1a could be interchanged with the N-terminus half of 2a without a loss in the strength of EC coupling as determined by the maximum amplitude of Ca²⁺ transients.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We previously showed that full-length 1a or 2a expressed in 1-null cells became integrated into functional skeletal-type DHPR complexes that include _1S, 1a or 2a, and presumably ₂- and  subunits. This conclusion was based on many similar characteristics of the expressed DHPRs such as the Ca²⁺ current density and voltage-dependence, the slow kinetics of activation of the Ca²⁺ current, and estimations of the single channel currents based on nonstationary variance analyses (Beurg et al., 1999b). Critical differences between 1a and 2a were observed in the characteristics of the recovered EC coupling, which suggested that 2a was incapable of fully substituting for 1a as a component of the voltage sensor that triggers the Ca²⁺ transient. In the present study we used deletion mutants and chimeras of 1a and 2a to determine the molecular basis for 1) the ability of 1a to recover EC coupling when expressed in 1-null cells; and 2) the inability of 2a to recover charge movements and EC coupling with normal characteristics in the same cells. We identified the participation of the N- and C-termini of 1a in skeletal-type EC coupling and the interference of the C-terminus of 2a in the same process.

Because the N-terminus of 1a and 2a have amino acid sequences that are entirely divergent, yet the N-terminus half of 2a supported EC coupling, the N-terminus of 1a was a far less critical determinant of EC coupling than the C-terminus of 1a. Since a domain of the pore-forming _1S subunit is also required for EC coupling in skeletal myotubes (Nakai et al., 1998b), the present observations suggest two distinct hypotheses. The identified C-terminal region of 1a could either assist _1S, or function in parallel with _1S, to bring about EC coupling with normal characteristics. For example, the identified domain could bring about a close colocalization of DHPRs and RyRs across from the junctional membrane that would enable _1S to signal opening of the RyR. Alternatively, the identified domain could be an element of the signal that opens the RyRs. These two possibilities, although seemingly dissimilar, may represent two manifestations of the same molecular interaction between  and the RyR or between  and other junctional proteins.

The deletion analyses of the present study indicated that the nonhomologous C-termini of 1a and