Introduction

The skeletal muscle L-type Ca2+ channel (Cav1.1) is localized in regions of the T-tubular membrane that are closely apposed to the sarcoplasmic reticulum (i.e., the T-SR junction), and the primary role of Cav1.1 is to serve as the voltage sensor for skeletal muscle-type excitation–contraction (EC) coupling [20, 23, 45]. While Ca2+ currents through other L-type channels (Cav1.2, Cav1.3, and Cav1.4) have fairly well-understood physiological roles, the functional significance of Ca2+ current through Cav1.1 has remained uncertain [29, 33, 34, 44, 58]. This uncertainty stems from the fact that during single twitches of skeletal muscle fibers, fewer than 5% of Cav1.1 channels are thought to function as calcium channels, and twitches proceed normally in the absence of external Ca2+ [1, 2, 53]. However, some evidence indicates that Ca2+ entry through Cav1.1 may have significance for the long-term function of skeletal muscle. For example, it has been suggested that during sustained contractile activity, the presence of external Ca2+ and its flux through Cav1.1 contributes to maintaining contractile force [30, 46]. It has also been suggested that, in aged mammalian skeletal muscle, Ca2+ current through Cav1.1 may play a role in the mechanism of EC coupling and in sustaining contractile force [12, 36].

Dynamic changes in intracellular Ca2+ concentration ([Ca2+]i) are often the direct result of electrical signals in cells. It is now clear that many ion channels are subject to feedback modulation by changes in [Ca2+]i. For example, Ca2+-dependent modulation of channel activity has been demonstrated for cyclic nucleotide-gated channels, N-methyl-d-aspartate receptors, ryanodine receptors (RyRs), inositol 1,4,5-triphosphate receptors, and high-voltage-activated Ca2+ channels [42, 32]. Many of these channels display Ca2+-dependent modulation, either inactivation or facilitation (or both), that is mediated by calmodulin (CaM) [42].

The Cav1 family of channels contains four isoforms that exhibit distinct tissue-specific expression: Cav1.1 (skeletal muscle), Cav1.2 (cardiac myocytes and neurons), Cav1.3 (neurons and sino-atrial node cells), and Cav1.4 (retinal neurons). The Ca2+-dependent modulation of Cav1.2 has been extensively studied; this channel displays Ca2+-dependent inactivation (CDI) and facilitation (CDF), in addition to voltage-dependent inactivation [32, 37]. Both Ca2+-dependent processes are mediated by CaM [37, 39, 47, 60]. The structural determinants of CDI have been assigned to the proximal region of the C terminus of Cav1.2 [4, 29, 61]. Three domains have been identified within this region: a Ca2+-binding EF-hand motif, a CaM-tethering site, and a CaM-binding IQ motif (Fig. 1). The EF-hand motif, located ~16 residues beyond the end of the last transmembrane segment (IVS6), is absolutely necessary for CDI [37]. The CaM-tethering site, which consists of both preIQ3 [16] and IQ motifs [35, 38], resides 50 amino acids downstream from the EF-hand motif and binds Ca2+-free CaM (apo-CaM) at resting [Ca2+]i. The IQ motif resides downstream from the EF-hand motif and the pre-IQ3 domain, and it binds Ca2+-CaM [35, 43]. When interaction of CaM with either of these domains is compromised, CDI is reduced or eliminated.

Fig. 1
figure 1

Alignment of the proximal carboxyl termini of Cav1 family channels. The proximal 197 residues of each C terminus, beginning immediately downstream from the end of transmembrane segment IVS6, are compared. Sequences are from mouse Cav1.2 (Genbank accession number NM_009781), mouse Cav1.1 (XM_983862), rat Cav1.3 (NM_017298), and rat Cav1.4 (NM_053701)

Full size image

The initial 200 amino acids of the C terminus of all members of the Cav1 family are highly conserved and contain the above-described domains (Fig. 1). However, there are striking differences in Ca2+ modulation between these different channel isoforms. For example, Cav1.2 mediates a fast-activating Ca2+ current that exhibits both fast CDI and CDF, Cav1.3 mediates a fast-activating Ca2+ current that displays fast CDI and slow CDI, and Cav1.4 mediates a fast-activating Ca2+ current that completely lacks CDI [33, 37, 59]. By contrast, Cav1.1 mediates a slowly activating Ca2+ current that does not display significant inactivation [57]. Whether Cav1.1 exhibits any form of Ca2+-dependent modulation has been controversial [3, 22, 27]. In earlier studies, adult skeletal muscle fibers were used to investigate the underlying mechanism of inactivation of the skeletal L-type Ca2+ current [3, 22, 25]. Results from these studies led to two controversial explanations: (1) L-type Ca2+ current displays only slow voltage-dependent inactivation and no Ca2+-dependent inactivation and (2) L-type Ca2+ current exhibits Ca2+-dependent inactivation in addition to very slow voltage-dependent inactivation. However, the question of Ca2+-dependent modulation of Cav1.1 has been difficult to address using adult skeletal muscle fibers because these cells possess an extensive transverse tubule system (TTS) and numerous ionic conductances.

In the present work, cultured mouse myotubes were used to explore the relationship between Cav1.1 and CaM. These cells are advantageous for this purpose because although they lack extensively invaginated TTS, the organization and position of Cav1.1 and RyR1 are the same as in adult fibers where Cav1.1 is organized in groups of four (tetrads), and each tetrad in T-tubules apposes every other RyR1 in junctional SR [20]. Importantly, the outward ionic currents in myotubes can be effectively blocked by appropriate internal and external solutions. Furthermore, cultured myotubes can be made to express exogenous proteins through direct nuclear microinjection.

The results presented in this work demonstrate that Cav1.1 displays slow CDI that is mediated by CaM. Additionally, fluorescent resonance energy transfer (FRET) measurements show, for the first time, that CaM associates with Cav1.1 channel in vivo.

Experimental procedures

Molecular biology

The coding sequences for rat wild-type CaM (CaMwt) and rat mutant CaM incapable of binding Ca2+ (CaM1234) were excised from their original vector (pcDNA3) [37] using XbaI and KpnI and ligated into the multiple cloning site of XbaI- and KpnI-digested EYFP-C1 (BD Biosciences Clontech, CA). All constructs were verified by restriction digest analysis and sequencing.

Electrophysiology

Primary myotubes were cultured from normal newborn mouse skeletal muscle as previously described [52]. Approximately 1 week after plating, macroscopic Ca2+ currents were measured with the standard ruptured-patch whole-cell technique. In experiments assessing the effects of CaM on Ca2+ currents, normal myotubes (~1 week in culture) were injected with expression plasmids encoding CaMwt or CaM1234 (gift of Dr. Yue [37]) and green fluorescent protein (pEGFP-C1, BD Biosciences Clontech) at concentrations of 0.7 and 0.02 μg/μl, respectively. Successfully transfected myotubes were identified 36–48 h after injection by their green fluorescence under UV illumination. Patch pipettes were constructed of borosilicate glass and had resistances of 1.8–2.5 MΩ when filled with the standard internal solution, which contained (in mM) 145 Cs-aspartate, 10 Cs2-EGTA, 5 MgCl2, and 10 HEPES (pH 7.4 with CsOH), or with BAPTA internal solution, in which 10 mM Cs2-BAPTA replaced Cs2-EGTA. The external solution contained (in mM) 145 tetraethylammonium chloride (TEA-Cl), 10 CaCl2 or 10 BaCl2, 0.03 tetrodotoxin, and 10 HEPES (pH 7.4 with TEA-OH). The holding potential was −80 mV, and test pulses were preceded by a 1-s prepulse to −30 mV to inactivate endogenous T-type Ca2+ currents. Recorded membrane currents were corrected off-line for linear components of leakage and capacitance by digitally scaling and subtracting the average of ten preceding control currents, elicited by hyperpolarizing voltage steps (30-mV amplitude) from the potential of −50 mV. Ca2+ currents were normalized by linear cell capacitance (expressed in pA/pF). Values for G max, the maximal Ca2+ conductance, were obtained by fitting the measured currents according to the equation:

$$I_{{{\text{peak}}}} = G_{{{\text{max}}}} {{\left( {V - V_{{\text{R}}} } \right)}} \mathord{\left/ {\vphantom {{{\left( {V - V_{{\text{R}}} } \right)}} {{\left\{ {1 + \exp {\left[ { - {{\left( {V - V_{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}} } \right)}} \mathord{\left/ {\vphantom {{{\left( {V - V_{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}} } \right)}} k}} \right. \kern-\nulldelimiterspace} k} \right]}} \right\}}}}} \right. \kern-\nulldelimiterspace} {{\left\{ {1 + \exp {\left[ { - {{\left( {V - V_{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}} } \right)}} \mathord{\left/ {\vphantom {{{\left( {V - V_{{1 \mathord{\left/ {\vphantom {1 2}} \right. \kern-\nulldelimiterspace} 2}} } \right)}} k}} \right. \kern-\nulldelimiterspace} k} \right]}} \right\}}}$$
(1)

where I peak is the peak current activated at the test potential V, V R is the extrapolated reversal potential, V 1/2 is the potential for half-maximal activation of the Ca2+ conductance, and k is a slope factor. The fraction of current remaining at the end of an 800-ms test pulse was divided by the peak current (r 800 = I end/I peak) and used to quantify the level of inactivation according to the equation:

$${\text{Inactivation }}{\left[ \% \right]} = {\left( {1 - r_{{800}} } \right)} \times 100\% $$
(2)

All recordings were performed at room temperature (~20°C), and data are reported as mean ± SEM; n indicates the number of myotubes tested. Data sets were statistically compared by an unpaired, two-sample Student’s t test, with a confidence interval of at least 95%.

Confocal microscopy and FRET measurements

Primary cultures of myotubes isolated from newborn dysgenic mice were plated onto 35-mm culture dishes with integral No. 0 glass coverslip bottoms (MatTek). Approximately 1 week after plating, myotubes were injected with expression plasmids encoding fluorescently tagged Cav1.1 (pECFP-Cav1.1; gift of Dr. K. Beam) and CaM (YFP-CaM). The rabbit skeletal muscle Cav1.1 was in CFP-C1 mammalian expression vector (BD Biosciences Clontech). Successfully transfected myotubes were identified 36–48 h after injection by their cyan or yellow fluorescence under UV illumination. Cells were bathed in rodent ringer (in mM: 146 NaCl, 5 KCL, 2 CaCl2, 1 MgCl2, 11 glucose, 10 HEPES; pH 7.4 adjusted with NaOH) and examined with an LSM 510 META laser scanning microscope (Zeiss, Thornwood, New York) with 40× oil objective. Two laser lines (458 and 514 nm) of the argon laser (30 mW maximum output, operated at 50% or 6.3 A), respectively, were used to excite CFP and YFP fluorophores. Emissions of CFP and YFP were recorded in multitrack configuration with a bandwidth filter of 465–495 nm (Chroma, Rockingham, VT, USA) and long-pass filter of 530 nm, respectively. Under these conditions, there is no cross-talk between the CFP and YFP fluorescence because CFP is not excited at 514 nm and YFP does not emit in the 465–495 nm range. To determine the magnitude of FRET, YFP-tagged CaM was photobleached by repeated (40–100) scans with the 514-nm line set at maximum laser intensity, which is 20- to 40-fold higher than the intensity used to elicit yellow fluorescence from YFP. Upon completion of YFP photobleaching, fluorescence intensities (I CFPpost and I YFPpost) were recorded under the same conditions as before photobleaching. The increased emission of CFP-tagged Cav1.1 indicated that FRET had occurred between donor (CFP) and acceptor (YFP) fluorophores. Fluorescence intensities were analyzed by the 510 LSM Image Examiner and 3D Zeiss software packages (Zeiss), and FRET efficiencies were calculated as:

$$E = {\left( {\frac{{I_{{{\text{CFPpost}}}} - I_{{{\text{CFPpre}}}} }}{{I_{{{\text{CFPpost}}}} }}} \right)} \cdot 100\% $$
(3)

where I CFPpre and I CFPpost are the background-corrected CFP fluorescence intensities before and after photobleaching YFP, respectively.

Results

The native skeletal muscle L-type channel (Cav1.1) displays CDI

Normal myotubes were used to test the hypothesis that endogenously expressed Cav1.1 exhibits CDI (Fig. 2). Whole-cell currents were evoked by 800-ms depolarizations, which was the longest duration test pulse that was experimentally compatible with stable recordings and myotube survival. During a depolarization of this length, Ca2+ currents mediated by Cav1.1 underwent significant inactivation (Fig. 2a). The current remaining at the end of the pulse (r 800) displayed a U-shaped voltage-dependence (Fig. 2e, filled circles), consistent with a current-dependent inactivation process [7]. In such a process, the extent of inactivation varies in proportion to the amplitude of the inward calcium current, which in turn depends on the number of conducting channels and the electrochemical driving force on calcium. Test pulses to +10 mV and above elicited currents of sufficient amplitude to trigger current-dependent inactivation, as previously described for adult frog and mouse muscle fibers [3, 25] and heterologously expressed Cav1.2 and Cav1.3 [37, 59]. Inactivation was maximal at a test potential of +20 mV, as reflected by a minimum r 800 value of 0.68 ± 0.03 (n = 53; Fig. 2e). Correspondingly, the Ca2+ current attained its maximum conductance at +20 mV (Fig. 2d and Table 1).

Fig. 2
figure 2

Ba2+ virtually eliminates inactivation of CaV1.1 current. Representative whole-cell currents were recorded from normal myotubes using first Ca2+ (a) and then Ba2+ (b) as charge carrier. The currents shown in a and b were recorded from the same myotube [cell 4 (10–30–01); linear capacitance C = 267 pF], initially using Ca2+ as charge carrier and then again after exchange of the external solution. Currents were elicited by 800-ms depolarizations to the indicated test potentials (+10 to +40 mV). Note that the scales are different for Ca2+ and Ba2+ currents. c The Ca2+ (black trace) and Ba2+ (dark grey trace) currents evoked by a depolarization to +20 mV in a and b were superimposed following normalization. d The average peak current density (I peak) is plotted as a function of membrane potential (V test). Data were obtained from the indicated number of myotubes for each group. The smooth lines through the data were generated by using the Eq. 1 (“Experimental procedures”) and the average values. e The fraction of current remaining at the end of 800-ms depolarizations (r 800) is plotted as a function of test potential (V test) for Ca2+ and Ba2+ currents recorded from normal myotubes. Symbols and error bars represent mean ± SEM, with the number of myotubes in parentheses

Full size image
Table 1 Ca2+ and Ba2+ conductances and their voltage-dependent characteristics in normal myotubes
Full size table

To investigate whether inactivation of Cav1.1 is ion specific as was previously shown for other Ca2+ channels [7, 8], currents were first recorded with 10-mM external Ca2+; this solution was then replaced with one containing equimolar Ba2+ in place of Ca2+, and whole-cell currents were again recorded in the same myotubes. Previous studies have shown that Ba2+ is ~100 times less effective than Ca2+ in triggering current-dependent inactivation of Cav1.2 [8, 19]. Exchange of Ba2+ for Ca2+ resulted in a significant increase in maximal ion conductance (from 239 to 383 nS/nF; p < 0.02), a significant shift of about −11 mV (p < 0.001) in the current/voltage (I–V) relationship, and the virtual elimination of inactivation. Altogether, inactivation was reduced from 30 to 2% at test potentials eliciting maximum currents (Table 1; Fig. 2b,d and e, open circles). These effects, which are similar to those observed for other L-type channels [8, 13, 54], indicate that inactivation of Cav1.1 is strongly dependent on Ca2+ permeation through the channel. Furthermore, the U-shaped voltage dependence of r 800 when Ca2+ is the charge carrier (Fig. 2e) and the absence of inactivation when Ba2+ is the charge carrier indicate that Cav1.1 undergoes CDI.

In some of the myotubes, a small, slowly activating outward current was observed during 800-ms depolarizations, when Ca2+ currents were recorded with 10-mM external Ca2+ and 10-mM internal EGTA (Fig. 2a). This small outward current was eliminated when Ba2+ replaced external Ca2+ or when glutamate replaced external Cl−, and it was also eliminated when BAPTA was substituted for EGTA in the pipette solution. Furthermore, addition of 0.2 mM niflumic acid (NiF), a known blocker of Ca2+-activated Cl− channels, to 10-mM external Ca2+ solution largely eliminated this small outward current, consistent with the possibility that this small outward current represents Cl− influx through Ca2+-activated Cl− channels. Importantly, however, neither NiF nor replacement of external Cl− with glutamate significantly affected the inactivation phenotype, i.e., the currents still displayed CDI and the U-shaped voltage dependence of r 800 was preserved (data not shown).

CDI of Cav1.1 is sensitive to fast dynamic changes in [Ca2+]i within the T-SR junction

CDI of other members of the Cav1 family, as well as members of the Cav2 family, has been shown to be mediated by CaM. A model of Ca2+-dependent modulation via CaM has been proposed for Cav1 and Cav2 families in which Ca2+ influx through the channel pore selectively activates the C-lobe of CaM, whereas more global increases in [Ca], for example those resulting from Ca2+ release from internal stores, preferentially activate the N-lobe of CaM [34]. Ca2+ binding to the N-lobe of CaM is prevented by the presence of Ca2+ buffers (either EGTA or BAPTA, at concentrations ranging from 1–10 mM), whereas Ca2+ binding to the C-lobe is not [32].

To assess whether CDI of Cav1.1 is affected by Ca2+ buffering, and thus indirectly investigate which lobe of CaM is responsible, Ca2+ currents were recorded from normal myotubes with either EGTA or BAPTA in the pipette solution. EGTA is a slow, high-affinity Ca2+ buffer that effectively buffers Ca2+ in the bulk myoplasm, but not in subcellular regions such as the T-SR junction (where Cav1.1 is localized) which experience fast dynamic changes in [Ca2+]. In the presence of 10 mM EGTA, Ca2+ entering the T-SR junction can freely diffuse up to 140 nm before being bound by EGTA, which means that Ca2+ concentration can change dynamically by an order of magnitude within the T-SR junction. By contrast, 10 mM BAPTA is able to bind Ca2+ within 30 nm from the point of entry, thus BAPTA can effectively buffer Ca2+ concentration within the T-SR junction as well as within the bulk myoplasm [48, 50].

Figure 3 shows Ca2+ currents recorded from normal myotubes with either 10 mM EGTA or BAPTA in the pipette. Currents were recorded >5 min after achieving the whole-cell configuration to allow for adequate intracellular dialysis with the pipette solution. In the presence of EGTA, Ca2+ currents through Cav1.1 displayed a characteristic CDI (a, c, and e), indicating that EGTA does not prevent permeating Ca2+ from triggering CDI. By contrast, intracellular BAPTA virtually eliminated CDI and additionally produced significant changes in the I–V relationship. At test potentials that elicited maximal currents, r 800 values were increased from 0.68 ± 0.03 to 0.94 ± 0.02 (p < 0.001) in BAPTA-dialysed cells (Fig. 3b and e). The elimination of CDI by BAPTA suggests that intracellular accumulation of Ca2+ within the T-SR junction, or near the cytoplasmic mouth of the channel, is the trigger for CDI of Cav1.1. BAPTA also decreased the peak current density from −9.4 to −5.5 pA/pF (p < 0.004). However, the maximal Ca2+ conductance was not significantly reduced (249 ± 24 nS/nF with BAPTA, n = 8 compared to 272 ± 21 nS/nF with EGTA, n = 53; Fig. 3d and Table 2), owing to a negative shift in the reversal potential (see below). Finally, BAPTA shifted the midpoint (V 1/2) of current activation by 6 mV (p < 0.01), shifted the reversal potential (V R) by −10.4 mV (p < 0.02), and decreased the slope factor (k) by 1.4 mV (p < 0.02; Table 2).

Fig. 3
figure 3

Intracellular BAPTA eliminates CDI of CaV1.1. Representative whole-cell currents were recorded from normal myotubes with either 10 mM EGTA (a) or 10 mM BAPTA (b) in the pipette solution. Currents were elicited by 800-ms depolarizations to the indicated test potentials (+20 to +40 mV). a Cell 3 (11–5–01), C = 250 pF; b cell 4 (9–11–03), C = 254 pF. c Currents evoked at +20 mV with either EGTA (black trace) or BAPTA (dark grey trace) in the pipette solution were superimposed following normalization. These are the same +20-mV currents as illustrated in a and b. d The average peak current density (I peak) is plotted as a function of membrane potential (V test). Data were obtained from the indicated number of myotubes for each group. The smooth lines through the data were generated by using the Eq. 1 (“Experimental procedures”) and the average values. e r 800 values are plotted as a function of test potential (V test) for Ca2+ currents recorded from normal myotubes using either EGTA or BAPTA in the pipette solution. Symbols and error bars represent mean ± SEM, with the number of myotubes in parentheses

Full size image
Table 2 Maximal Ca2+ conductance and its voltage-dependent characteristics recorded with either 10 mM EGTA or BAPTA as intracellular Ca2+ buffer
Full size table

CaM mediates Ca2+-dependent inactivation of skeletal L-type Ca2+ current

The finding that Cav1.1 exhibits CDI in conjunction with the presence of a highly conserved CaM-interacting region within the C terminus of Cav1.1 prompted me to investigate whether CDI of Cav1.1 is mediated by CaM through a mechanism similar to that previously found for Cav1.2 and Cav1.3 [37, 44, 60]. One way to test for CaM involvement is to overexpress mutant calmodulin (CaM1234) that does not bind Ca2+. CaM1234 has aspartate (D)-to-alanine (A) substitutions in all four EF-hand motifs to abolish Ca2+ binding [37]. Overexpression of CaM1234 produces a dominant negative effect that eliminates CDI of Cav1.2 and Cav1.3 [15, 16, 44, 59].

To test this approach, normal myotubes were injected with cDNAs encoding wild-type CaM (CaMwt) or the mutant CaM1234, and Ca2+ currents were recorded 48 h later. Skeletal muscle L-type Ca2+ currents recorded from myotubes overexpressing CaMwt and CaM1234, respectively, are shown in Fig. 4a and b. Overexpression of CaMwt in normal myotubes did not significantly affect the I–V relationship (Fig. 4d and Table 3) or the extent of CDI compared to uninjected myotubes (Fig. 4e, half-filled squares), suggesting that the concentration of endogenous CaM is sufficient to fully populate native Cav1.1. However, in normal myotubes overexpressing CaM1234, the extent of CDI was greatly suppressed (Fig. 4e, open diamonds), although the I–V relationship was unaffected (Fig. 4d and Table 3). On average, overexpression of CaM1234 significantly increased r 800 from 0.68 ± 0.03 to 0.92 ± 0.02 (at a test potential of +20 mV; p < 0.01), reflecting a large decrease in the extent of CDI. This result is similar to previous data obtained for Cav1.2 coexpressed with CaM1234 in HEK cells [37], and also for Cav1.2 and CaM1234 coexpressed in dysgenic myotubes, where the r 800 increased from 0.7 to 0.9 in comparison with myotubes expressing Cav1.2 alone [51]. Altogether, these results indicate that CaM mediates CDI of Cav1.1.

Fig. 4
figure 4

CaM mediates CDI of Cav1.1. Representative Ca2+ currents recorded from normal myotubes that overexpressed either CaMwt (a) or CaM1234 (b). Currents were elicited by 800-ms depolarizations from −80 mV to the indicated test potentials (+20 to +40 mV). a Cell 3 (6–26–02), C = 813 pF; b cell 1 (2–5–02), C = 343 pF. c Currents recorded at +20 mV from myotubes overexpressing either CaMwt (black trace; from a) or CaM1234 (grey trace; from b) were superimposed following normalization. d The average peak current density (I peak) is plotted as a function of membrane potential (V test). Data were obtained from the indicated number of myotubes for each group. The smooth lines through the data were generated by using the Eq. 1 (“Experimental procedures”) and the average values. e r 800 values are plotted as a function of test potential (V test) for Ca2+ currents recorded from uninjected myotubes and myotubes overexpressing CaMwt or CaM1234. Symbols and error bars represent mean ± SEM, with the number of myotubes in parentheses

Full size image
Table 3 Ca2+ conductance and its voltage-dependent characteristics in normal myotubes overexpressing CaMwt or CaM1234
Full size table

CaM associates with Cav1.1 in vivo

It has been shown that CaM is tethered to Cav1.2 in the Ca2+-free form (apo-CaM) and functions as the Ca2+ sensor for CDI. The CaM-tethering site binds apo-CaM and consists of two short peptide sequences (preIQ3 and IQ motifs [14]), and these sequences are conserved in Cav1.1 (Fig. 1). To date, only in vitro approaches have been used to investigate potential interactions between Cav1.1 and CaM [35, 43]. No previous studies have tested for in vivo association of CaM with Cav1.1.

To examine this issue, confocal microscopy and FRET were used to test whether Cav1.1 and CaM interact in vivo. The FRET method is an appropriate tool for analyzing the relative position of donor and acceptor fluorophores that are separated by distances of 10 nm or less; such distances are relevant to protein–protein interactions in vivo [14], and FRET has been used for determining the in vivo association of Cav1.2 and CaM [15, 16]. The efficiency of FRET between CFP-Cav1.1 and YFP-CaMwt was determined by means of acceptor photobleaching (“Experimental procedures”). Figure 5 shows confocal images of cyan and yellow fluorescence from a dysgenic myotube expressing both CFP-Cav1.1 and YFP-CaMwt. Expression of CFP-Cav1.1 resulted in the appearance of small cyan fluorescence foci located near the cell surface. The small foci correspond to groups of Cav1.1 localized to T-SR junctions; these foci are similar in size and distribution to those of Cav1.1 foci revealed by immunohistochemistry [21]. Expression of YFP-CaM resulted in the appearance of homogeneous yellow fluorescence, as CaM is a cytoplasmic protein. Panel a shows emission of cyan and yellow fluorescence before photobleaching. Panel b shows the same cell after acceptor (YFP) photobleaching. After photobleaching YFP-CaMwt, there was a notable increase in the cyan signal from foci. FRET efficiency (E) was determined according to Eq. 3 (see “Experimental procedures”) and was found to have a value of E = 12%. Overall, seven cells were examined, and the average FRET efficiency was estimated as E = 4.06 ± 1.11% (n = 7). For comparison, the FRET efficiency between Cav1.2-CFP and CaMwt-YFP coexpressed in HEK293 cells was estimated as E = 10 ± 1.1% (n = 8) [16]. It should be noted that the FRET signal between Cav1.2 and CaMwt is stronger due to the position of the donor fluorophore at the C terminus of Cav1.2, whereas in the present experiment, the donor fluorophore was positioned at the N terminus of Cav1.1, and the fluorophores were thus farther apart. Despite the N-terminal location of CFP in the present experiments, the measured FRET signal between CFP-Cav1.1 and YFP-CaMwt in dysgenic myotubes is highly relevant, and it strongly suggests that CaM associates with Cav1.1 in vivo.

Fig. 5
figure 5

FRET measurements indicate that Cav1.1 interacts with CaMwt in vivo. CFP-Cav1.1 and YFP-CaMwt were coexpressed in dysgenic myotubes. Confocal microscopy was used to measure the intensities of cyan and yellow fluorescence both before (a) and after (b) photobleaching the acceptor fluorophore (EYFP-CaM). Before and after measurements were obtained from the same myotube. The cyan fluorescence signal was significantly increased after photobleaching the acceptor, indicating the presence of significant FRET. Note that cyan fluorescence appears as small foci (arrows). The efficiency of FRET (E) was determined according to Eq. 3 (see “Experimental procedures”) and had a value of E = 12% for the myotube shown in this figure. The scale bar corresponds to 5 μm

Full size image

Discussion

The present study provides new information about the skeletal muscle L-type Ca2+ channel (Cav1.1). Specifically, the data demonstrate, for the first time, that Cav1.1 undergoes Ca2+-dependent inactivation that requires CaM. Additionally, the results show that CaM associates with Cav1.1 in vivo.

The skeletal muscle L-type Ca2+ current has been extensively studied in adult amphibian and mammalian muscle fibers [3, 9–13, 17, 18, 22, 24, 25, 27], and several of these studies have investigated the mechanisms underlying inactivation. In general, these studies have concluded either that Cav1.1 displays only voltage-dependent inactivation [9, 27] or, alternatively, that Cav1.1 exhibits both Ca2+-dependent and voltage-dependent inactivation [3, 24]. However, whether Cav1.1 exhibits CDI has remained unresolved.

The present experiments demonstrate that natively expressed Cav1.1 displays classic features of CDI, including U-shaped voltage dependence and elimination of inactivation when Ba2+ replaces Ca2+ as charge carrier or when the fast Ca2+ buffer BAPTA is present intracellularly. The U-shaped voltage dependence indicates that inactivation is current dependent and is greatest at potentials eliciting the greatest inward Ca2+ current. Ba2+ is widely used as a charge carrier in studies of Ca2+ channel function, and it is generally accepted that Ba2+ cannot effectively activate the Ca2+-binding sites of Ca2+-activated proteins such as CaM. Thus, it is generally assumed that voltage-gated Ca2+ channels exhibit only voltage-dependent inactivation when Ba2+ is used as the charge carrier [37, 44]. However, this is not strictly true, as Ba2+ currents mediated by Cav1.2 display both current-dependent and voltage-dependent inactivation, although it is important to note that the affinity of Ba2+ for the “inactivation site” on the Cav1.2 channel is 100-fold lower than that of Ca2+ [19]. The ability of BAPTA to eliminate CDI of Cav1.1 suggests that CDI is triggered by accumulation of Ca2+ within the T-SR junction.

This study was performed on primary cultures of skeletal muscle myotubes, which provide a natural cellular environment for Cav1.1. It is important to recognize that myotubes differ morphologically from adult skeletal muscle fibers. Myotubes grown in culture for 6–7 days, like those employed here, possess fully developed sarcomeric organization and distinct SR and T tubular systems (TTS) [20, 23], and the mechanism of skeletal muscle-type EC coupling appears to be functionally mature in such cells [49]. Importantly, myotubes differ from adult fibers by the fact that in myotubes, most (~88%) of the T-tubules are oriented longitudinally and do not penetrate very far into the cell’s interior [20, 23]. Thus, T-SR junctions are located close to the myotube’s surface, and the diffusion of extracellular ions to the junction is relatively unrestricted, in contrast to the situation which exists in adult fibers. Consequently, Ca2+ depletion within the TTS, which has been suggested to contribute to inactivation of the L-type current in adult fibers [3], is unlikely to play a significant role in CDI of Cav1.1 in myotubes.

One interesting finding of the present study is that CDI of Cav1.1 is considerably slower (τ i ~ 500 ms) than the previously reported “fast” CDI of Cav1.2 and Cav1.3 (τ i ~ 20 ms), which was shown to be triggered by Ca2+ binding to the C-lobe of CaM [37, 44]. The mechanism underlying the relative slowness of CDI by Cav1.1 remains to be determined. However, it may be a consequence of the slow activation rate of Cav1.1, as channel activation and inactivation are interdependent processes [28]. Cav1.1 activates at least tenfold slower than other members of the Cav1 family. For instance, at room temperature, Cav1.1 activates with a time constant (τ a) of ~20–200 ms, whereas Cav1.2 activates with a τ a of ~2 ms [56]. Alternatively, the relatively slow CDI of Cav1.1 may reflect a different underlying mechanism than that which underlies fast CDI of Cav1.2. The speed at which CaV1.1 undergoes CDI is very similar to the “slow” CDI of Cav1.3 (τ i ~ 400 ms; estimated from Fig. 2 of [58], which was previously shown to be triggered by Ca2+ binding to N-lobe of CaM [37, 44]. This similarity suggests that CDI of Cav1.1 may also be triggered by Ca2+ binding to N-lobe of CaM, according to a model proposed for Ca2+/CaM-dependent regulation of Cav1 and Cav2 family channels [32]. Further experiments (currently in progress) will be needed to test this possibility.

The results obtained with CaM1234 (Fig. 4) demonstrate that CaM is required for CDI of Cav1.1. This demonstration suggests that the general mechanism of CDI may be similar for all channels of the Cav1 and Cav2 families. Consistent with this possibility, the structural determinants of CaM-dependent CDI have been identified within the proximal C terminus of the Cav subunit, and these determinants are conserved among all members of the Cav1 and Cav2 families [4, 29, 60]. Previously, information concerning the interaction between Cav1.1 and CaM has been limited to in vitro experiments. For example, it has been reported that the preIQ3 and IQ motifs of Cav1.1 can bind apo-CaM and Ca-CaM, respectively, in a fashion similar to the corresponding domains of Cav1.2 [35, 43]. The present results with overexpression of CaM1234 in normal myotubes (Fig. 4) are the first to indicate that CaM modulates the function of Cav1.1 in vivo. Furthermore, the significant FRET signal detected between CFP-Cav1.1 and YFP-CaMwt in dysgenic myotubes (Fig. 5) provides the first indication that CaM associates with Cav1.1 in vivo.

Intracellular BAPTA produced significant effects on some of the biophysical properties of macroscopic Cav1.1 currents (Fig. 3). For example, BAPTA decreased the peak current density, shifted the I–V relationship to more positive potentials, shifted the reversal potential to more negative potentials, and decreased the slope factor (k) of the conductance–voltage relationship. These results are consistent with a previous study in which the presence of 5 mM BAPTA plus 0.5 mM CaCl2 shifted activation of the I–V relationship by approximately +6 mV [31]. Additionally, in dialyzed adult frog skeletal muscle fibers, the presence of 20 mM BAPTA resulted in a significant decrease in the maximum quantity of intramembrane charge movement (Q max), which reflects the activity of the voltage-sensors for EC coupling [49]. Charge movement is an intramembrane current caused by the voltage-dependent movement of charged S4 segments within each transmembrane domain of Cav1.1, and the “ON” charge movement in response to depolarization precedes the onset of inward L-type Ca2+ current [28]. Furthermore, L-type Ca2+ currents have been reported to “run down” more rapidly in the presence of intracellular BAPTA [18, 26]. These effects of BAPTA may indicate that dynamic changes in Ca2+ concentration within the T-SR junction play a role in voltage-sensing by Cav1.1. It is conceivable that, when BAPTA is present in the intracellular environment of myotube or adult muscle fiber, the voltage sensed by Cav1.1 is altered. In this regard, it has been demonstrated that Ca2+-free BAPTA can penetrate into the lipid bilayer and remove Ca2+ from either extracellular or intracellular face, thereby reducing the negative charge screening effect of Ca2+ and changing the electric field across the membrane [41]. Alternatively, BAPTA may produce additional effects on channel function that are independent of its Ca2+-buffering properties. This possibility is suggested by previous studies of slow muscarinic inhibition of Cav channels showing that BAPTA disrupts channel modulation via a mechanism independent of Ca2+ buffering [5, 6].

Despite decades of study, the physiological significance of the skeletal muscle L-type Ca2+ current has remained unclear. Single contractions of skeletal muscle fibers do not require Ca2+ current through Cav1.1 because skeletal muscle-type EC coupling is mediated by a mechanical interaction between the voltage sensor for EC coupling (i.e., Cav1.1) and the Ca2+ release channels of the SR [45, 40]. Thus, skeletal muscle fibers produce normal twitches in the temporary absence of external Ca2+. However, during sustained or repetitive contractions, Ca2+ current through Cav1.1 may be necessary for maintaining force [30, 46]. The mechanism underlying regulation of contractile force by external Ca2+ is unknown. It is conceivable that Ca2+ entry can enhance contraction directly, or entering Ca2+ may be sequestered into the SR and thereby increase the amount of Ca2+ available for release in response to subsequent action potentials, which would lead to an increase in contractile force. Furthermore, during repetitive EC coupling in rat and frog skeletal muscle fibers, the resultant increase in intracellular [Ca2+] triggers a significant increase in Q max, suggesting a potential physiological role for Ca2+ influx in regulating the voltage-sensor function of Cav1.1 [11, 17, 49, 50]. Finally, in skeletal muscle fibers of aged mice and humans, the L-type Ca2+ current through Cav1.1 seems to play an important role both in EC coupling and in maintaining contractile force [12, 36]. Because normal daily activities require sustained activation of skeletal muscles, Ca2+ influx through Cav1.1 should play an important role. Ca2+-dependent inactivation of Cav1.1 may thus represent an important mechanism for fine-tuning regulation of intracellular [Ca2+] and consequently of skeletal muscle function.