Tagging with green fluorescent protein reveals a distinct subcellular distribution of L-type and non-L-type Ca²⁺ channels expressed in dysgenic myotubes

Abstract

Expression of cardiac L-type Ca²⁺ channels in dysgenic myotubes results in large Ca²⁺ currents and electrically evoked contractions resulting from Ca²⁺-entry dependent release of Ca²⁺ from the sarcoplasmic reticulum. By contrast, expression of either P/Q-type or N-type Ca²⁺ channels in dysgenic myotubes does not result in electrically evoked contractions despite producing comparably large Ca²⁺ currents. In this work we examined the possibility that this discrepancy is caused by the preferential distribution of expressed L-type Ca²⁺ channels in close apposition to sarcoplasmic reticulum Ca²⁺ release channels. We tagged the N termini of different α₁ subunits (classes A, B, C, and S) with a modified green fluorescent protein (GFP) and expressed each of the fusion channels in dysgenic myotubes. Each GFP-tagged α₁ subunit exhibited Ca²⁺ channel activity that was indistinguishable from its wild-type counterpart. In addition, expression of GFP-α_1S and GFP-α_1C in dysgenic myotubes restored skeletal- and cardiac-type excitation-contraction (EC) coupling, respectively, whereas expression of GFP-α_1A and GFP-α_1B failed to restore EC coupling of any type. Laser-scanning confocal microscopy revealed a distinct expression pattern for L-type compared with non-L-type channels. After injection of cDNA into a single nucleus, GFP-α_1S and GFP-α_1C were present in the plasmalemma as small punctate foci along much of the longitudinal extent of the myotube. In contrast, GFP-α_1A and GFP-α_1B were not concentrated into punctate foci and primarily were found adjacent to the injected nucleus. Thus, L-type channels possess a targeting signal that directs their longitudinal transport and insertion into punctate regions of myotubes that presumably represent functional sites of EC coupling.

Excitation-contraction (EC) coupling of striated muscle (skeletal and cardiac) occurs as a close interplay between the L-type Ca²⁺ channels or dihydropyridine receptors (DHPRs) of the plasma membrane and the Ca²⁺ release channels or ryanodine receptors (RyRs) of the sarcoplasmic reticulum (SR) (for review see ref. 1). In cardiac muscle, rapid Ca²⁺ influx through L-type channels triggers RyRs to open and, thereby, causes a massive release of SR Ca²⁺ (2). In contrast, in skeletal muscle there is no need for Ca²⁺ influx (3). Instead, membrane depolarization is thought to induce a conformational change in the DHPR (or voltage sensor) of the plasma membrane (4, 5), which subsequently activates the SR RyR, possibly via a direct protein–protein interaction (6–8). Even though the two mechanisms of EC coupling differ, both require a close physical association between the DHPR and the RyR. Indeed, in both cultured mouse skeletal muscle (9) and chicken myocardial cells (10), DHPRs and RyRs colocalize in clusters that can be detected as punctate foci at the light microscopic level.

Experiments analyzing wild-type and chimeric skeletal and cardiac DHPRs expressed in dysgenic myotubes have helped to elucidate molecular mechanisms of EC coupling. Dysgenic myotubes are prepared from the skeletal muscle of mice homozygous for a frame-shift in the coding region of the α_1S subunit of the skeletal DHPR (11). Dysgenic myotubes lack skeletal-type EC coupling (12) and slowly activating L-type Ca²⁺ currents (13). However, injection of cDNA expression plasmids encoding the skeletal DHPR into the nuclei of dysgenic myotubes restores both slowly activating L-type Ca²⁺ current and skeletal-type EC coupling (i.e., electrically evoked contractions in the absence of extracellular Ca²⁺) (14). In contrast, expression of the cardiac DHPR in dysgenic myotubes results in both cardiac-like, rapidly activating Ca²⁺ currents and cardiac-type EC coupling (evoked contractions requiring extracellular Ca²⁺) (15).

Surprisingly, in dysgenic myotubes expressing the brain N-type (16) or P/Q-type (17) Ca²⁺ channels, Ca²⁺ influx-dependent contractile activity (i.e., cardiac-type EC coupling) is almost never observed even though the Ca²⁺ current density is comparable to that for the cardiac DHPR. The inability of large, non-L-type Ca²⁺ currents to trigger SR Ca²⁺ release might be because of a lack of targeting of the non-L-type Ca²⁺ channels into clusters at functional sites of EC coupling, namely, the junctions of the SR with either the surface membrane (peripheral couplings) or the transverse (T) tubules (18–20). Accordingly, few (or none) of the non-L-type Ca²⁺ channels would be positioned close enough to RyRs to trigger SR Ca²⁺ release.

Thus, the objective of the present work was to compare the subcellular distribution of four different kinds of voltage-gated Ca²⁺ channels expressed in the dysgenic cell system: the skeletal DHPR (α_1S; ref. 21), the cardiac DHPR (α_1C; ref. 22), the neuronal P/Q-type Ca²⁺ channel (α_1A; ref. 23), and the neuronal N-type Ca²⁺ channel (α_1B; ref. 16). To visualize the subcellular distribution of the expressed channels, the N terminus of each α₁ subunit was fused with the green fluorescent protein (GFP), previously cloned from the pacific jellyfish Aequorea victoria (24). Interestingly, the addition of this relatively large (27 kDa) protein to the Ca²⁺ channel α₁ subunits did not appear to interfere with the physiological properties of any of the four Ca²⁺ channels. Macroscopic channel properties (activation rate, voltage-dependence, and expression level) and restoration (or lack of restoration) of EC coupling were indistinguishable from their untagged, wild-type counterparts. By using confocal microscopy, we also were able to demonstrate differential subcellular targeting of L-type versus non-L-type Ca²⁺ channels. Our results indicate that only the L-type channels are efficiently targeted throughout the myotube and cluster in punctate foci, presumably reflecting localization to functional sites of EC coupling.

MATERIALS AND METHODS

Plasmid Construction.

The coding sequences of the α₁ subunits of the skeletal muscle DHPR (α_1S) (21), cardiac muscle DHPR (α_1C) (22), neuronal P/Q-type (α_1A) (23), and N-type (α_1B) (16) Ca²⁺ channels were inserted “in-frame” and downstream of the coding region of a modified GFP, cloned in a proprietary mammalian expression vector, kindly provided by P. Seeburg (Center for Molecular Biology, University of Heidelberg, Germany). The HindIII to EcoRI segment of the polylinker of pSP72 (Promega) was inserted directly after the last GFP codon (Lys-238). Transcription was under the control of a cytomegalovirus promoter. To avoid the potential loss of the GFP label because of the known, physiological truncation of the carboxyl terminus of various Ca²⁺ channel α₁ subunits (25–29) we fused the GFP to the N terminus of each α₁ subunit. To allow in-frame ligation of the 5′ terminus of the α₁ cDNAs, a SalI site was introduced by PCR in-frame directly in front of the initiating ATG. Thus, the first ATG was 24 bp downstream of the final GFP codon. Ligation of the 3′ terminus of the α₁ cDNAs into polylinker sites was accomplished either by native or by previously introduced restriction sites (14, 16) downstream of the termination codon: EcoRI (nucleotide 5676) for α_1S (14), HpaI (nucleotide 6609) into SmaI for α_1C, BamHI (nucleotide 7614) for α_1A, and EcoRI (nucleotide 7327) for α_1B (16). GFP cDNA was modified by introducing point mutations via PCR mutagenesis (30) to optimize for brightness and spectral properties (mutation S65T) (31), and to suppress thermosensitivity (mutations V163A, I167T, and S175G) (32). Finally, PCR-modified GFP and α₁ cDNA stretches were sequenced to confirm sequence integrity.

Expression of cDNA.

Primary cultures of myotubes isolated from newborn dysgenic mice were prepared as described previously (33). Approximately 1 week after plating, myotubes were microinjected (14) into a single nucleus with solutions of expression plasmids (300–400 ng/μl) carrying cDNAs for either GFP, GFP-α_1S, GFP-α_1C, GFP-α_1A, or GFP-α_1B. Injected myotubes subsequently were examined for the development of green fluorescence. Expressing cells were evaluated for restoration of contractility, the presence of macroscopic Ca²⁺ currents, and subcellular channel distribution.

Contractile Recordings.

Fluorescent myotubes were tested for the ability to contract in response to electrical stimulations (10 ms, 80 V) through an extracellular pipette placed near the cell (14). Contractions were recorded in a normal rodent Ringer solution (145 mM NaCl/5 mM KCl/2 mM CaCl₂/1 mM MgCl₂/10 mM Hepes, pH 7.4 with NaOH) or Ca²⁺-free rodent Ringer’s (equimolar substitution of Ca²⁺ by Mg²⁺). Contractions were videotaped, and the movement of a characteristic surface marker on the contracting cell was analyzed frame by frame (30 frames/s) by using an EVO-9700 Hi8 Desktop Video (SONY, Institute of Applied Video Technology, Hollywood, CA).

Electrophysiological Characterization.

Macroscopic Ca²⁺ currents were measured by using the whole-cell patch clamp technique (34). Whole-cell patch-clamp pipettes were pulled from borosilicate glass and had resistances of 1.5–1.9 MΩ when filled with an internal solution containing 140 mM cesium-aspartate, 10 mM cesium-EGTA, 5 mM MgCl₂, and 10 mM Hepes (pH 7.4 with CsOH). The composition of the external bath solution was 10 mM CaCl₂, 145 mM tetraethylammonium-Cl, 3 μM tetrodotoxin, and 10 mM Hepes (pH 7.4 with tetraethylammonium-OH). Test pulses were preceded by a 1-s prepulse to −30 mV to inactivate endogenous T-type Ca²⁺ currents (35). Test currents were corrected for linear components of leakage and capacitative currents by digitally scaling and subtracting a preceding control current elicited by a hyperpolarizing voltage step (20–40 mV amplitude) applied from the holding potential of −80 mV. Ca²⁺ current densities (expressed in pA/pF) were normalized by linear cell capacitance. Values are given as mean ± SD. All recordings were performed at room temperature.

Fluorescence Analysis.

Myotubes cultured on 35-mm culture dishes were superfused with a normal rodent Ringer solution (see above) mounted under a glass coverslip and examined by using a Zeiss Axiophot epifluorescence microscope equipped with a HBO 50 mercury lamp and a filter set suitable for the detection of fluorescein. Fluorescent myotubes were visualized by using a Zeiss 40× PlanNeofluar (numerical aperture 0.75) objective and digitized by using a charge-coupled device video camera and a DEI-470 processing unit (Optronics, Galeta, CA). Images were later processed by using the Adobe Photoshop software (ADOBE Systems, Mountain View, CA).

Laser-Scanning Confocal Microscopy.

Myotubes were superfused in a normal rodent Ringer solution and mounted under a glass coverslip as described above. The culture dish subsequently was fastened upside-down on the stage of a Nikon Inverted microscope. Fluorescing cells were analyzed by using a Sarastro 2000 confocal laser-scanning microscope (Molecular Dynamics) with a Nikon 60× PlanApo oil immersion objective (numerical aperture 1.40) and the ImageSpace software (Silicon Graphics, Mountain View, CA). GFP excitation/emission was achieved with a filter set (488 nm/510 nm) designed for fluorescein detection. Image were 1,024 × 1,024 pixels with a pixel size of 0.11 μm. Step size between confocal sections was 2 μm. To enable comparison between different myotubes, all images were recorded at the same adjustments of laser power (17 mW/30% transmission) and photomultiplier sensitivity (750 V) and were processed (Adobe Photoshop software) by using identical values for contrast and brightness.

RESULTS AND DISCUSSION

N-Terminal Tagging of α₁ Subunits with Modified GFP.

GFP has been used successfully as a protein tag on both the amino- and carboxyl termini of a wide range of cytosolic and membrane-bound proteins (for review see ref. 36) and in some cases physiological differences were reported depending on the terminus carrying the tag (e.g., ref. 37). We tagged the N terminus of the Ca²⁺ channel α₁ subunits (Fig. 1) because of the known physiological truncation of the C termini of several α₁ subunits (25–29). N-terminal, in-frame tagging of α₁ subunits (skeletal L-type: α_1S, cardiac L-type: α_1C, P/Q-type: α_1A, and N-type: α_1B) with a modified GFP and expression in dysgenic myotubes led to fully functionally Ca²⁺ channels as detailed below. Initial experiments showed that a GFP modified at Ser-65 (S65T, ref. 31) exhibited bright cytosolic fluorescence when expressed alone, but that S65T-GFP fused with the Ca²⁺ channel α₁ subunits exhibited only little fluorescence (data not shown). Lowering the incubation temperature for injected cells from 37°C to 30°C greatly enhanced fluorescence caused by the temperature sensitivity of the development of GFP fluorescence (31, 36, 38). However, the lower incubation temperature resulted in the complete loss of myotube contractile ability (from both normal myotubes and dysgenic myotubes expressing either GFP-α_1S or GFP-α_1C). Therefore, we introduced additional point mutations into GFP(S65T) to suppress its thermosensitivity (V163A, I167T, S175G; ref. 32) (subsequently referred to as “GFP” for simplicity). Ca²⁺ channel α₁ subunits tagged by the modified GFP generated sufficient fluorescence even when injected cells were incubated at 37°C.

Figure 1.

Schematic representation of GFP-α₁ subunit fusion proteins. The N terminus of L-type (α_1S and α_1C) and non-L-type (α_1A and α_1B) Ca²⁺ channel α₁ subunits was fused to the C terminus of a modified version of the GFP. in, intracellular; out, extracellular.

Distinct Cellular Distribution of L-Type and Non-L-Type Ca²⁺ Channels.

Two to four days after injection of expression plasmids into single nuclei, myotubes were screened with epifluorescence microscopy for the development of green fluorescence. Surprisingly, fluorescence of skeletal and cardiac L-type Ca²⁺ channels was not restricted to the regions surrounding the injected nucleus. Instead, the fluorescence was distributed throughout a large section of the cell (see GFP-α_1C in Fig. 2 A and B). However, the fluorescence associated with expression of the neuronal P/Q-type and N-type channels was strongly concentrated around the injected nucleus (see GFP-α_1A in Fig. 2 C and D). The observation of a different distribution of L-type and non-L-type channels in skeletal myotubes points to the possibility that different targeting mechanisms exist for the different Ca²⁺ channel subtypes. However, it is also possible that the tagging of the Ca²⁺ channel α₁ subunits may have artifactually caused an improper expression of the fusion channel. Thus, we characterized the functional properties (macroscopic currents and EC coupling) of the GFP-α₁ subunit fusion proteins.

Figure 2.

Different expression patterns of L-type and non-L-type Ca²⁺ channels in dysgenic skeletal myotubes. Phase contrast (A and C) and corresponding fluorescence images (B and D) of dysgenic skeletal myotubes 4 days after mononuclear injection of GFP-α_1C (A and B) or GFP-α_1A (C and D) cDNAs. GFP-α_1C (and also GFP-α_1S, data not shown) exhibited a widespread distribution of fluorescence (B), whereas GFP-α_1A (and also GFP-α_1B, data not shown) exhibited a predominantly perinuclear distribution of fluorescence (D). (Scale bar = 10 μm.)

GFP-Tagged Ca²⁺ Channels Are Functionally Unaltered.

Despite the obvious differences in the fluorescence distribution of the four classes of GFP-tagged Ca²⁺ channels, each fusion protein resulted in a functional voltage-gated Ca²⁺ channel on expression in dysgenic myotubes. For example, expression of GFP-α_1S in dysgenic myotubes restored slowly activating L-type Ca²⁺ current that was indistinguishable from its untagged (14) counterpart (Fig. 3A). Similarly, expression of GFP-α_1C, GFP-α_1A, and GFP-α_1B in dysgenic myotubes led to the appearance of rapidly activating Ca²⁺ currents (Fig. 3A) similar to their respective untagged, wild-type channels (15–17). Whole-cell peak currents (mean ± SD) were 4.8 ± 2.1 pA/pF (n = 14) for GFP-α_1S, 23.9 ± 12.5 pA/pF (n = 11) for GFP-α_1C, 30.7 ± 11.2 pA/pF (n = 10) for GFP-α_1A and 16.1 ± 10.5 pA/pF (n = 11) for GFP-α_1B (Fig. 3B). These peak currents are similar to values obtained for the wild-type channels on expression in the same system (15–17). Thus, tagging the N terminus of Ca²⁺ channel α₁ subunits with this relatively large (27 kDa) GFP protein does not interfere with the normal electrophysiological properties of the channel.

Figure 3.

GFP-α₁ subunits exhibit typical class-specific macroscopic Ca²⁺ current properties on expression in dysgenic skeletal myotubes. (A) Representative whole-cell Ca²⁺ currents attributable to the four GFP-α₁ subunit fusion channels. Macroscopic Ca²⁺ currents were elicited by 200-ms step depolarizations to the indicated potentials (after a prepulse to inactivate T-type currents) from a holding potential of −80 mV. Current amplitudes were normalized by linear cell capacitance. (B) Comparison of peak Ca²⁺ currents attributable to the four classes of GFP-α₁ subunit fusion channels (gray columns) with their corresponding unfused, wild-type counterparts (black columns). Bars represent the mean ± SD of 10–31 recordings, except for untagged α_1A (17) and α_1B (16), which are the mean ± SEM. Values for the untagged α₁ subunits were as published previously: α_1S and α_1C (15), α_1A (17), and α_1B (16).

We also investigated the ability (or inability) of the GFP-α₁ subunit fusion proteins to activate SR RyRs and, thus, support either skeletal- or cardiac-type EC coupling. As shown in Fig. 4 dysgenic myotubes expressing GFP-α_1S (Upper) exhibited strong contractions on extracellular stimulation that were indistinguishable from those of normal myotubes. Moreover, these contractions resulted from a skeletal-type EC coupling mechanism because contractions persisted on removal of extracellular Ca²⁺. Alternatively, a cardiac-type or Ca²⁺-induced-Ca²⁺-release mechanism was found on expression of GFP-α_1C in dysgenic myotubes because contraction was abolished on removal of extracellular Ca²⁺ (Fig. 4, Lower). Identical results have been obtained previously with the untagged, wild-type channels (14, 15). Thus, tagging of the N termini of muscle L-type Ca²⁺ channels apparently is without effect on the ability of these channels to participate efficiently in functional EC coupling.

Figure 4.

Restoration of electrically evoked contractions in dysgenic skeletal myotubes expressing GFP-α_1S and GFP-α_1C. Electrically evoked contractions obtained from myotubes expressing GFP-α_1S (Upper) and GFP-α_1C (Lower) in normal rodent Ringer (2 mM Ca²⁺, first column) and in Ca²⁺ free Ringer solution (equimolar substitution of Ca²⁺ by Mg²⁺, middle column). Removal of extracellular Ca²⁺ blocked contractions of GFP-α_1C but not of GFP-α_1S expressing cells. Recovery of electrically evoked contractions in the GFP-α_1C expressing myotube after a return to the normal rodent Ringer solution is shown (Lower, far right). Vertical scale bar: movement distance of 10 μm (GFP-α_1S expressing myotube) or 7 μm (GFP-α_1C expressing myotube). Arrows indicate the time of application of the electrical stimulus (10 ms, 80 V). Note: In the presence of extracellular Ca²⁺, contractions were never observed for either GFP-α_1A or GFP-α_1B expressing myotubes (n > 50 for each construct).

Electrical stimulation was applied to more than 50 myotubes exhibiting intense green fluorescence for each of the neuronal constructs (GFP-α_1A and GFP-α_1B). None of these myotubes displayed contractile activity, even in the presence of extracellular Ca²⁺ (data not shown). These results are in agreement with previous observations on the untagged counterparts (16, 17). The inability of the P/Q-type channel (GFP-α_1A) to support Ca²⁺-induced-Ca²⁺-release is especially striking because this channel produced peak Ca²⁺ current densities that were statistically indistinguishable from those of the cardiac L-type channel (see Fig. 3B). Comparable GFP-α_1A and GFP-α_1C current densities should have provided similar total Ca²⁺ influx, though only that attributable to the cardiac L-type channels was capable of triggering the release of SR Ca²⁺. Our observation that the P/Q-type and N-type channels were distributed differently than L-type channels (Fig. 2) raised the possibility that the neuronal channels are not directed to the functional sites of EC coupling. To compare the distribution of the different types of Ca²⁺ channels with higher resolution, we used laser-scanning confocal microscopy.

Only L-Type Ca²⁺ Channels Cluster in Punctate Foci.

Confocal imaging allows the visualization of fluorescent signals in a restricted plane of focus, thus limiting out-of-focus blur. Moreover, serial optical sections can be reconstructed by computer calculations to produce three-dimensional projections that reveal a much higher resolution compared with normal epifluorescence microscopy. As shown in the calculated projection in Fig. 5A, myotubes expressing GFP alone (Top) looked opaque and densely filled because of the cytosolic localization of free GFP. The topmost section of these control cells exhibits a relatively homogeneous fluorescence (Fig. 5B, Top). Similar results also were found for confocal sections through the midlevel of the cell (Fig. 5C). In contrast, myotubes expressing GFP-α_1S (Fig. 5, second row) or GFP-α_1C (Fig. 5, third row) exhibited a more inhomogeneous appearance. For both constructs, there were many clusters on or near the surface membrane, as indicated by the topmost optical section (Fig. 5B) and the midlevel optical section (Fig. 5C, note the clusters near the periphery). These clusters near the surface membrane may correspond to sites of peripheral couplings between the SR and surface plasmalemma, which are relatively plentiful early during myogenesis (18, 19). The midlevel section also revealed clusters that appeared to be far from the surface and that may represent nascent junctions between the SR and T-tubules (19). Clustered DHPRs previously have been demonstrated in mouse myotubes by using immunolabeling, which further revealed that the DHPR clusters were colocalized with RyR clusters (9). Thus, it seems likely that clusters of GFP-α_1S also were colocalized with RyRs. Because the confocal images reveal a subcellular distribution for GFP-α_1C similar to GFP-α_1S, it also appears likely that α_1C expressed in dysgenic myotubes colocalizes with the endogenous skeletal RyRs, which would seem a prerequisite for the demonstrated ability of α_1C to produce cardiac-type EC coupling in dysgenic myotubes.

Figure 5.

GFP-α_1S and GFP-α_1C channels, but not GFP-α_1A and GFP-α_1B channels, cluster in a punctate distribution on expression in dysgenic skeletal myotubes. Laser-scanning confocal images of dysgenic myotubes (in vivo) performed 4 days after mononuclear injection of cDNA expression plasmids containing GFP only (top row), GFP-α_1S (second row), GFP-α_1C (third row), GFP-α_1A (fourth row), and GFP-α_1B (bottom row). (A) Calculated three-dimensional projection (“maximum intensity”) of 8–10 confocal sections (2-μm distances) of myotubes expressing each of the indicated constructs. (B and C) Topmost confocal section of the cell (B) and of a more central slice (C), 6–8 μm deeper than in B of the same myotube as depicted in A. Identical laser and photomultiplier adjustments were used for all images. (Scale bar = 10 μm.)

Myotubes expressing GFP-α_1A and GFP-α_1B exhibited an entirely different spatial pattern (Fig. 5, fourth and bottom rows). The non-L-type channel fluorescence spread only a very limited distance away from the injected nucleus and was concentrated perinuclearly, which is especially evident in the midlevel sections (Fig. 5C, fourth and bottom rows). The topmost section (Fig. 5B, fourth and bottom rows) revealed only diffuse staining, which may indicate that the neuronal channels are incorporated diffusely into the plasma membrane only near the injected nucleus. Alternatively, it may be that the surface density of the brain channels was too low to be detected by fluorescence and that the signal from the topmost section represents out-of-focus fluorescence from deeper within the cell. However, some channels must have been present in the plasmalemma to account for the large Ca²⁺ currents supported by the GFP-α_1A and GFP-α_1B constructs (Fig. 3).

An argument that the surface density of the brain channels is indeed low compared with L-type channels is provided by the measurements of Adams et al. (17). They found that although current densities were similar in myotubes expressing α_1A and α_1C, the amount of maximal intramembrane charge movement (Q_max) was much lower for α_1A than α_1C. Because Q_max provides an indirect measurement of the number of channels in the plasmalemma, they postulated that the P/Q-type channels are present in the plasmalemma at a much lower density than the cardiac L-type channels, but that the P/Q-type channels have a much higher open probability. Our results support a much lower surface density for the P/Q-type channels than for L-type channels: even if all the apparently, surface-associated signal of α_1A (and α_1B) represented channels in the plasmalemma, this fluorescence was still clearly much less than the surface-associated fluorescence in myotubes expressing α_1S or α_1C.

If it is true that the GFP-tagged α_1A and α_1B subunits are present in the plasmalemma at only a very low density, they may have been undetectable by fluorescence measurements. Nonetheless, it is clear that the neuronal channels did not form high-density clusters like the muscle L-type channels. Thus, Ca²⁺ entry through α_1A and α_1B channels would be expected to couple very inefficiently to SR Ca²⁺ release because most or all of the SR RyRs would be far away from the nearest α_1A or α_1B channels. In contrast, comparable amounts of Ca²⁺ ions entering through cardiac L-type channels located in close proximity to SR Ca²⁺ release sites (i.e., in punctate clusters at peripheral couplings and SR/T-tubule junctions) throughout much of the myotubes’ extension would ensure a very efficient triggering of SR Ca²⁺ release.

Taken together, our data unequivocally demonstrate that N-terminal labeling of different classes of voltage-gated Ca²⁺ channels yields fully functional Ca²⁺ channels on expression in dysgenic myotubes. Thus, GFP tagging of Ca²⁺ channels should provide an extremely valuable tool in studying Ca²⁺ channel physiology and ultrastructure. Moreover, our results provide evidence for the importance of two essential factors required for efficient EC coupling in cultured myotubes: (i) an abundant, widespread distribution of Ca²⁺ channels along the myotube plasma membrane and (ii) a punctate distribution of plasmalemmal Ca²⁺ channels, presumably reflecting their close association with SR RyRs. The ability of transport and insertion of Ca²⁺ channels throughout the cell and the targeting of channels to functional sites of EC coupling appear to be features unique to the muscle L-type Ca²⁺ channels. Thus, it should be possible to identify regions of the L-type channels essential for their transport, targeting, and clustering through the use of GFP-tagged L-type/non-L-type chimeras.

ABBREVIATIONS

SR

sarcoplasmic reticulum

GFP

green fluorescent protein

EC

excitation-contraction

DHPR

dihydropyridine receptor

RyR

ryanodine receptor