Major contribution of sarcoplasmic reticulum Ca²⁺ depletion during long-lasting activation of skeletal muscle

Depolarization of skeletal muscle fiber induces the release of Ca²⁺ ions from the SR through RyR Ca²⁺ release channels, which in turn activate contraction. Activity of the RyR is a tightly voltage-controlled process that involves the dihydropyridine receptors (DHPRs) anchored in the t-tubules, which act as voltage sensors and transduce membrane potential change into RyR opening (Ríos and Pizarro, 1991; Schneider, 1994; Melzer et al., 1995). The increase in cytosolic Ca²⁺ and development of force that ensue are transient in twitch muscle fibers while the membrane depolarization is maintained, and the progressive decay in Ca²⁺ release during continuous stimulation has been proposed as one of the mechanisms leading to muscle fatigue (Allen et al., 2008). Different mechanisms have been put forward to explain the cytosolic Ca²⁺ decline during maintained depolarization: (a) RyR activity may be inhibited by high cytosolic Ca²⁺ concentrations that are likely to be reached in the proximity of the outer cytoplasmic mouth of the channel during voltage-evoked SR Ca²⁺ release; (b) RyRs may enter a voltage-induced closed inactivated state driven by the DHPR; (c) SR Ca²⁺ efflux may lead to a progressive depletion of SR Ca²⁺. In frog skeletal muscle, it was demonstrated that Ca²⁺-dependent inactivation of the flux rate of Ca²⁺ release occurs within tens of milliseconds (Schneider and Simon, 1988; Simon et al., 1991), whereas voltage-dependent inactivation of Ca²⁺ release and SR Ca²⁺ depletion take place later on a slower time scale (Schneider et al., 1987; Brum et al., 1988). In adult mammalian skeletal muscle, early experiments have shown that fibers enter a voltage-dependent inactivated state of contractile refractoriness a few seconds after a K⁺ contracture has been elicited (Chua and Dulhunty, 1988). The voltage dependence of inactivation of Ca²⁺ release was later assessed by using conditioning depolarizing pulses of increasing amplitude in voltage-clamped mouse muscle fibers (Collet et al., 2004; Ursu et al., 2004). SR Ca²⁺ depletion is thought to precede the voltage-induced inactivation of RyRs, and a few hundreds of milliseconds appear to be sufficient to cause a reduction of SR Ca²⁺ to a stabilized level (Ursu et al., 2005; Royer et al., 2008). More recently, simultaneous measurements of Ca²⁺ changes inside the SR and in the cytosol under voltage clamp indicated that prolonged Ca²⁺ release results in major SR depletion, which is accompanied by a reduction in SR permeability during Ca²⁺ release (Sztretye et al., 2011).

The contributions of voltage-dependent inactivation of Ca²⁺ release and SR Ca²⁺ depletion during long-lasting, voltage-evoked cytosolic Ca²⁺ transients have been mainly investigated by applying conditioning depolarizations of variable amplitudes and durations. However, the measurement of Ca²⁺ changes in the cytosol did not allow to determine whether the alteration of Ca²⁺ transients resulted from SR depletion or voltage-dependent inactivation of Ca²⁺ release channels provoked by the conditioning depolarization. Yet, in a previous study, by measuring Ca²⁺ changes inside the SR in voltage-controlled mouse muscle fibers, we showed that depolarizing steps of very small amplitude from a holding potential of −80 mV induced a slow but substantial SR Ca²⁺ depletion resulting from voltage-induced opening of Ca²⁺ release channels that may not be detected by measuring cytosolic Ca²⁺ changes (Robin and Allard, 2012). Consequently, conditioning depolarizations that apparently did not give rise to a detectable change in cytosolic Ca²⁺ may, however, be responsible for SR depletion and subsequent inhibition of Ca²⁺ transient, which could be misinterpreted as voltage-dependent inactivation.

In this study, we took advantage of the use of the SR Ca²⁺ indicator Fluo-5N in voltage-clamped mouse muscle fibers to investigate the role of SR Ca²⁺ depletion and inactivation of Ca²⁺ release channels in the decline of long-lasting, voltage-evoked Ca²⁺ signals. We showed that conditioning depolarizations at voltages more negative than voltages for which voltage-dependent inactivation occurs produce SR depletion that reduces the pool of releasable Ca²⁺. Additionally, SR Ca²⁺ signals evoked by trains of action potentials gave evidence of a pronounced SR depletion with no sign of voltage-dependent inactivation, whereas cytosolic Ca²⁺ signals markedly declined during stimulation. These results suggest that SR Ca²⁺ depletion predominantly contributes to the decline of SR Ca²⁺ release during long-lasting depolarizations in skeletal muscle fibers.

All experiments were performed in accordance with the guidelines of the French Ministry of Agriculture (87/848) and of the European Community (86/609/EEC). Male OF1 mice were killed by cervical dislocation before removal of flexor digitorum brevis muscles. Single fibers were isolated by a 50-min enzymatic treatment at 37°C using Tyrode’s solution containing 2 mg/ml collagenase type I (Sigma-Aldrich).

The major part of a single fiber was electrically insulated with silicone grease, and a micropipette was inserted into the fiber through the silicone layer to voltage clamp or current clamp the portion of the fiber free of grease (50–100-µm length) using a patch-clamp amplifier (RK-400; Bio-Logic) in whole-cell configuration (Collet et al., 2004; Robin et al., 2012). Command voltage or current generation and data acquisition were done using the pClamp10 software (Molecular Devices) driving an A/D converter (Digidata 1400A; Molecular Devices). Analogue compensation was systematically used to decrease the effective series resistance. The tip of the micropipette was then crushed into the dish bottom to allow intracellular dialysis of the fiber with the intra-pipette solution (Robin et al., 2012).

Cells were dialyzed through the micropipette for 20 min using an intracellular solution containing 200 µM Fura-2. Fluorescence was measured on an inverted epifluorescence microscope (Diaphot; Nikon), allowing the detection of fluorescence above 510 nm using a 40× oil-immersion objective. Fluorescence was produced from a 75-W xenon lamp by using a monochromator (Polychrome II; Till Photonics), allowing a 50-Hz alternation of the 340- and 380-nm excitation wavelengths conducted with an optical fiber. Images from a circular region of interest of 450 µm² corresponding to 12,000 pixels centered on the silicone-free cell portion under study (Fig. S1) were captured with a charge-coupled device camera (CoolSNAP_EZ; Roper Scientific), and image acquisition and processing were performed using an imaging workbench (MetaFluor; Molecular Devices). Time of exposure ranged between 15 and 50 ms depending on the level of fluorescence obtained after dye loading in each fiber. Background fluorescence at both excitation wavelengths was measured from an equivalent area next to each fiber tested and was subtracted, and the ratio F340/F380 was calculated. The frequency of images captured is indicated in the figure legends.

Fibers were incubated in Tyrode’s solution containing 10 µM Fluo-5N-AM at room temperature for 2 h to allow entry of the dye into the SR and subsequent deesterification. Cells were then dialyzed for 20 min through the micropipette used for clamping, with an internal solution containing 50 mM EGTA to prevent possible contraction during imaging and to avoid contribution of cytosolic Ca²⁺ changes to SR Ca²⁺ signals. Fluo-5N fluorescence was measured using the same setup and experimental conditions that were used for Fura-2 experiments, except that a single excitation wavelength of 488 nm was selected and 4 × 4–pixel binning was used to increase emission light intensity. Time of exposure ranged between 40 and 70 ms depending on the level of fluorescence obtained after dye loading in each fiber. For each voltage pulse, changes in SR Ca²⁺ content were expressed as F/F₀ after background subtraction, where F₀ is the measured fluorescence at the steady state before applying voltage pulses. The frequency of images captured is indicated in the figure legends.

The external solution contained, except for Fig. 6 (A and B) (mM), 140 TEA-methanesulfonate, 2.5 CaCl₂, 1 MgCl₂, and 10 HEPES, adjusted to pH 7.2 with TEA-OH. The Tyrode’s solution (Fig. 6, A and B) contained (mM) 140 NaCl, 5 KCl, 2.5 CaCl₂, 1 MgCl₂, and 10 HEPES, adjusted to pH 7.2 with NaOH. The intra-pipette solution contained (mM) 50 EGTA-KOH, 50 K-glutamate, 5 Na₂-ATP, 5 Na₂-phosphocreatine, 5.5 MgCl₂, 5 glucose, and 5 HEPES, adjusted to pH 7.2 with KOH. Stock solutions of Fluo-5N AM and Fura-2 were dissolved at 1 mM in dimethylsulfoxide and 10 mM in water, respectively. Experiments were performed at room temperature.

Fits were performed with Microcal Origin (Microcal Software, Inc.). Data are given as means ± SEM.

Fig. S1 shows a fluorescence image of a voltage-clamped fiber loaded with Fura-2 and the regions of interest selected for measuring the changes in fluorescence. It is available.

Fig. 1 A shows the cytosolic Fura-2 Ca²⁺ signals produced by long-lasting depolarizing pulses of 50-s duration and of increasing amplitudes. Although the low frequency of Ca²⁺ image capture led to an important underestimation of the early peak of Ca²⁺ signal, we observed that the maximal value of the Ca²⁺ signal increased with the amplitude of the depolarization, giving rise to the classical S-shape voltage-dependent activation curve of SR Ca²⁺ release, relating the peak Ca²⁺ values to voltages. Fig. 1 A also shows that Ca²⁺ signals remained fairly stable for low depolarizing pulses but declined faster and faster during the voltage pulse as depolarization amplitude increased. In all the fibers tested and for each depolarizing pulse, the level of the Ca²⁺ signal at the pulse end was measured and normalized to the peak values, and these ratios were plotted as a function of voltage (Fig. 1 B). Decline in the Ca²⁺ signal was variable at −40 mV; among the eight cells tested, four did not display any decline, as illustrated by the data presented in Fig. 1 A, whereas a decline ranging from 25 to 42% was observed in the four other fibers. Fitting the global relationship with a Boltzmann equation in each cell indicated a mean voltage value inducing half-maximal decline of −13 ± 1.4 mV and a mean steepness factor of 12 ± 1.6 mV. However, meticulous examination of the relationship revealed that for voltages more negative than −20 mV, the Ca²⁺ signal decline tends to display a more negative voltage dependence than the one observed for higher depolarizations, suggesting that two distinct processes with different voltage dependences may contribute to the Ca²⁺ decline. We hypothesized that the increase of the rate of Ca²⁺ signal decline with depolarization could result from voltage-dependent inactivation of Ca²⁺ release channels and/or from SR Ca²⁺ depletion, two processes that may display distinct voltage dependence. To further examine the involvement of these two phenomena in these long-lasting Ca²⁺ signals, we next challenged fibers with very long conditioning pulses of increasing amplitudes.

Fig. 2 A presents the cytosolic Fura-2 Ca²⁺ signals elicited by depolarizing test pulses of 1-s duration to +20 mV, preceded by 2-min conditioning pulses of increasing amplitudes. Here again, the low sampling frequency impeded the measurement of the peak value of the Ca²⁺ signal during the test pulse. We thus took the first data point 1 s after the onset of the test pulse as an index of the maximal Ca²⁺ level reached during the test pulse. Conditioning pulses to voltages as low as −60 mV, which did not evoke any detectable change in cytosolic Ca²⁺, induced a reduction of the Ca²⁺ transient elicited during the test pulse. As expected, further increases in the amplitude of the conditioning depolarization led to the development of a Ca²⁺ signal of increasing amplitude during the prepulse, which was followed by a more and more reduced Ca²⁺ transient during the test pulse. The peak value of the Ca²⁺ signal during the test pulse was measured in all fibers tested, normalized to the value obtained from a holding potential of −80 mV, and mean values were plotted as a function of the conditioning pulse voltage (Fig. 2 B). Fitting the relationship with a Boltzmann equation in each cell indicated a mean conditioning voltage value inducing half-maximal reduction of −53 ± 2 mV and a mean steepness factor of 6 ± 1.7 mV. These experiments thus confirm that long-lasting depolarizations turn on a voltage-dependent process leading to reduction of Ca²⁺ release, which at this step could not be identified as voltage-dependent inactivation of Ca²⁺ release channels or as SR Ca²⁺ depletion, but exhibiting a very negative voltage dependence that may explain the apparent shift toward negative voltages of the left part of the relationship in Fig. 1 B.

To determine whether voltage-dependent inactivation of Ca²⁺ release channels or SR Ca²⁺ depletion predominates at low voltages, we performed a series of experiments using the SR Ca²⁺ indicator Fluo-5N. We previously showed that long-lasting depolarization led to a decrease in SR Ca²⁺ content caused by Ca²⁺ release followed by a spontaneous recovery phase while depolarization was maintained (Robin et al., 2012). This recovery phase was attributed to voltage-dependent inactivation of Ca²⁺ release channels together with SR Ca²⁺ refilling supported by Ca²⁺-ATPases. It can yet be considered that sarcolemmal Ca²⁺ influxes should be activated in response to depolarization and in this way might contribute to the time course of recovery of Fluo-5N signals. However, we demonstrated in a previous study that increasing the Ca²⁺ permeability of the sarcolemma using the Ca²⁺ ionophore A23187 in the same experimental conditions did not induce any increase in Fluo-5N fluorescence, indicating that a potentiated sarcolemmal Ca²⁺ influx cannot alter Fluo-5N signals (Robin and Allard, 2012). Fluo-5N is thus very relevant to specifically investigate voltage-dependent inactivation of Ca²⁺ release channels during long-lasting depolarizations. Indeed, in contrast to cytosolic measurement of Ca²⁺ signals in which voltage-dependent inactivation of Ca²⁺ release channels and SR Ca²⁺ depletion both lead to a decline in fluorescence and cannot be distinguished, in Fluo-5N experiments, inactivation of Ca²⁺ release channels gives rise to an increase in Fluo-5N fluorescence, whereas SR depletion gives rise to the opposite. Fig. 3 A shows the Fluo-5N signals induced by long-lasting 50-s depolarizing pulses and of increasing amplitudes. In response to a depolarizing pulse to −30 mV, SR Ca²⁺ content decreased as a result of SR Ca²⁺ release and tended to stabilize at the end of the depolarizing step, strongly suggesting that Ca²⁺ release channels did not inactivate. The absence of a distinct recovery phase cannot however exclude that inactivation occurred but, if it were the case, our data indicate that depletion predominated over inactivation. Higher depolarizations induced faster SR Ca²⁺ content decrease but overall led, before the pulse end, to a Fluo-5N fluorescence recovery phase, the rate of which increased as depolarization went higher, indicating that Ca²⁺ release channels entered faster and faster the closed inactivated state. Comparable results were obtained when the highest voltages were applied first. As a consequence of inactivation of Ca²⁺ release induced by high voltage pulses, the lower level of cytosolic Ca²⁺ that ensued led to a slowing down of SR reloading upon repolarization. We also observed that for depolarizations to +20 mV, which induced maximal Ca²⁺ release and depletion in less than 1 s (Robin et al., 2012), inactivation occurred in an average of 2.1 ± 0.2 s (n = 6) after the onset of the pulse (not depicted), confirming that depletion precedes inactivation at this voltage. The peak to pulse end recovery of the Fluo-5N signal was measured in all fibers for each depolarizing step, normalized to the peak value of the SR Ca²⁺ drop, and mean values were plotted as a function of voltage (Fig. 3 B). Fitting the relationship with a Boltzmann equation in each cell indicated a mean voltage value inducing half-maximal inactivation of −13 ± 2.1 mV and a mean steepness factor of 4.7 ± 0.4 mV. These experiments likely suggest that during long-lasting depolarizations, voltage-dependent inactivation of Ca²⁺ release channels occurs for high depolarizations with a V_1/2 of −13 mV, whereas SR Ca²⁺ depletion takes place in response to much lower depolarizations with a V_1/2 of −53 mV. Along this line, superimposing the relationship between decline of cytosolic Ca²⁺ signals and voltage (Fig. 1 B) on the relationship between rate of recovery of SR Ca²⁺ signals and voltage corroborates that two voltage-dependent processes with distinct voltage sensitivities are responsible for the decline of cytosolic Ca²⁺ signals during long-lasting depolarizations.

To confirm this hypothesis, we next explored the effect of long-lasting conditioning depolarizations of increasing amplitudes on voltage-evoked SR Ca²⁺ signals. We recently demonstrated that SR Ca²⁺ release channels are under the active control of DHPRs at resting membrane potentials so that depolarizing pulses to voltages as low as −60 mV elicit SR Ca²⁺ efflux (Robin and Allard, 2012). Indeed, Fig. 4 A shows that a conditioning depolarizing pulse to −60 mV evoked a slow SR Ca²⁺ efflux that led within the 2-min conditioning pulse to a substantial SR depletion. Consequently, the subsequent test depolarization to +20 mV induced a SR Ca²⁺ decrease caused by active Ca²⁺ release of lower amplitude than the one evoked from a holding potential of −80 mV. Higher conditioning depolarizations to −50 and −40 mV induced more pronounced depletions that led to further reduction of the SR Ca²⁺ decrease produced by the test pulse. The amplitude of the peak SR Ca²⁺ decrease evoked by the test pulse was normalized relative to the total SR Ca²⁺ decrease induced by the conditioning pulse plus the test pulse in each cell for all conditioning pulses, and mean values were plotted as a function of conditioning voltages. Fitting the relationship with a Boltzmann equation in each cell indicated a mean voltage value inducing half-maximal depletion of −51 ± 2 mV and a mean steepness factor of 3.3 ± 0.5 mV (Fig. 4 B). The fact that the voltage dependence of SR depletion induced by conditioning pulses is very close to the voltage dependence of inhibition of cytosolic Ca²⁺ signals induced by long conditioning pulses (−53 mV) likely implies that SR depletion is the predominant process that leads to Ca²⁺ signal decline during long-lasting depolarizations. As expected, when the duration of the conditioning pulse to a given voltage was augmented, the magnitude of SR depletion increased, thereby reducing the SR Ca²⁺ decrease produced by the test pulse. Fig. 5 A illustrates the effects of increasing the duration of a conditioning pulse given to −60 mV. The best fit with an exponential function to the relationship between mean values of relative SR Ca²⁺ decrease evoked by the test pulse and duration of the −60-mV conditioning pulse in all cells tested indicated an average time constant of inhibition of 95 s (Fig. 5 B).

In physiological conditions, activation of skeletal muscle Ca²⁺ release is triggered by trains of action potentials. To determine whether the decline of Ca²⁺ signals provoked by SR depletion together with voltage-dependent inactivation of Ca²⁺ release channels in response to long-lasting depolarizations also occurs in response to a train of action potentials, we measured cytosolic and SR Ca²⁺ signals in response to 1-min-long trains of action potentials under current-clamp conditions in the presence of Tyrode’s solution. Fig. 6 A shows that cytosolic Fura-2 Ca²⁺ signals declined in average of 55 ± 1% after 1-min trains of action potentials. This decay of Ca²⁺ signals was much less pronounced than the decay displayed by Ca²⁺ signals evoked by a single voltage step to +20 mV because 50 s after the onset of the train of action potentials and 50 s after the onset of the voltage pulse, Ca²⁺ signals had decayed by 52 ± 2% and of 84 ± 2%, respectively. In contrast, Fluo-5N–monitored SR depletion induced by trains of action potentials reached a plateau value without ever recovering during the 1-min-long stimulation in all the cells tested, indicating that voltage-dependent inactivation of Ca²⁺ release does not occur under these conditions of stimulation (Fig. 6 B). These experiments strongly suggest that the decline of the cytosolic Ca²⁺ signal during long-lasting stimulations with trains of action potentials exclusively results from SR depletion. Yet, Fluo-5N experiments under voltage-clamp conditions indicated that inactivation of Ca²⁺ release started to develop in response to voltage pulses around −20 mV (see Fig. 3 B), a potential value far below the one reached by action potentials. We thus hypothesized that the spike depolarization, even reaching high voltage values close to +40 mV, is too short during an action potential for inactivation of Ca²⁺ release to be turned on. To test this hypothesis, we measured SR Ca²⁺ signals in response to a 30-s train of voltage pulses delivered at 50 Hz to +20 mV and lasting 2, 5, 10, and 15 ms. Fig. 6 C shows that in response to trains composed of pulses of 2- and 5-ms duration, the SR signals remained stable until the end of the train, whereas inactivation of Ca²⁺ release started to develop for 10-ms pulses and reinforced for 15-ms pulses and the single 30-s depolarizing step. The lengthening of pulse duration also gave rise to an increase in the rate of SR Ca²⁺ decrease, suggesting an enhancement of the kinetic of Ca²⁺ release. Comparable results were obtained when the longest voltage pulses were applied first. Plotting the peak to train end recovery of the Fluo-5N signal measured in all fibers tested and normalized to the peak value of the SR Ca²⁺ drop as a function of pulse duration indicated that Ca²⁺ release inactivated to an extent equal to half of the inactivation obtained with a single 30-s depolarizing pulse for pulses of 16.2 ± 0.5–ms mean duration (Fig. 6 D).

SR Ca²⁺ depletion, Ca²⁺-dependent inactivation of Ca²⁺ release channels, and voltage-dependent inactivation of voltage sensors are the three mechanisms that have been put forward to produce the decline of Ca²⁺ signals and of force, as a consequence, during continuous stimulation, whereas store-operated, voltage-gated, or excitation-coupled Ca²⁺ influxes, which are possibly activated during stimulation of long duration, might contribute to favor force recovery (Melzer et al., 1995; Berbey and Allard, 2009; Launikonis et al., 2010). In this study, our experimental conditions allowed us to preclude the possible involvement of sarcolemmal Ca²⁺ influx, and our recording conditions with a time scale of over tens of seconds did not offer the required resolution to investigate Ca²⁺-dependent inactivation that is known to fully develop in a few tens of milliseconds after the onset of depolarization (Schneider and Simon, 1988). These conditions allowed us to focus on SR Ca²⁺ depletion and voltage-dependent inactivation of Ca²⁺ release and to determine to what extent these two processes contribute to the decline of Ca²⁺ signals during long-lasting depolarization of skeletal muscle fibers. By measuring cytosolic Ca²⁺ signals and SR Ca²⁺ changes, we demonstrated that SR Ca²⁺ depletion together with inactivation of Ca²⁺ release channels are involved in the decline of long-lasting, depolarization-evoked Ca²⁺ signals but display different kinetics and voltage dependences. Since the work of Hodgkin and Horowicz (1960), it has been known that prolonged depolarization of muscle fibers causes spontaneous relaxation after contracture, leading to contractile refractoriness. Numerous studies have further demonstrated that this spontaneous relaxation results from an inactivation process that promotes closure of SR Ca²⁺ release channels in a voltage- and time-dependent manner. As we explored in the present study, time and voltage dependence of the inactivation process have been investigated by examining the effects of conditioning pulses on voltage-induced Ca²⁺ or force transients, or indirectly of high K⁺ solutions on K⁺-induced contractures (Hodgkin and Horowicz, 1960; Lüttgau and Spiecker, 1979; Caputo et al., 1984; Brum et al., 1988; Schnier et al., 1993). These steady-state or double-pulse inactivation protocols revealed that inactivation of Ca²⁺ release requires depolarizations lasting seconds or tens of seconds to develop and is associated with immobilization of intramembrane charge movements, suggesting that inactivation of Ca²⁺ release channels arises from transition of the voltage sensor that controls its opening into an inactivated state (Chandler et al., 1976; Brum and Rios, 1987). In the present study, monitoring SR Ca²⁺ changes under voltage control allowed us to focus on the inactivation process and to demonstrate that inactivation does occur in response to long-lasting depolarizing pulses, but for depolarizations of unexpected high intensity with a V_1/2 of −13 mV. Early K⁺ contracture experiments performed in mammalian skeletal muscle provided a much lower voltage value, producing half-inactivation (V_1/2) of −52 mV with 3-min conditioning pulses (Chua and Dulhunty, 1988). More recent voltage-clamp studies on mouse muscle fibers indicated a similar V_1/2 for Ca²⁺ release flux and cytosolic Ca²⁺ transients of −50 and −42 mV using 30 s and stabilized conditioning potential, respectively (Collet et al., 2004; Ursu et al., 2004). These values are very close to the values we obtained in the present study for cytosolic Ca²⁺ signals (−53 mV) as well as for voltage-dependent inhibition of SR Ca²⁺ signals (−51 mV) using 2-min conditioning pulses. However, our SR Ca²⁺ measurements clearly indicated that such low depolarizing conditioning pulses only produced SR Ca²⁺ depletion but were too small to induce inactivation of Ca²⁺ release channels, the voltage dependence of which was found to be 40 mV more positive. This suggests that the voltage-dependent inhibition of K⁺ contractures, Ca²⁺ release flux, and transients induced by conditioning pulses in the cited studies did not result from an inactivation process but mainly likely from SR Ca²⁺ depletion at least for low voltages.

The involvement of SR Ca²⁺ depletion in the decline of voltage-evoked Ca²⁺ signals was first established in frog muscle (Schneider et al., 1987). The Ca²⁺ release flux was found to recover from conditioning depolarizing pulses of a few hundred milliseconds at a rate that was dependent on the amount of Ca²⁺ released during the conditioning pulses, demonstrating that SR Ca²⁺ depletion is responsible for the decline of Ca²⁺ signals before Ca²⁺ release channels inactivate. In voltage-clamped mouse muscle fibers, it was further estimated that SR Ca²⁺ content decreased by 80% in response to 100-ms depolarizing pulses above +10 mV (Ursu et al., 2005). More recently, analysis of the evolution of release flux during pulses of longer duration (1–2 s) and simultaneous measurement of cytosolic and luminal SR Ca²⁺ revealed that +30-mV evoked Ca²⁺ release provokes a 65% reduction of the SR Ca²⁺ content and is associated with a decrease in the ability of the SR to empty its Ca²⁺ content (Royer et al., 2008; Sztretye et al., 2011). Our study confirms that depletion precedes inactivation during long-lasting depolarization because inactivation occurred 2.1 s after the onset of the pulse. Overall, our work is the first to provide the voltage dependence of SR Ca²⁺ depletion in parallel with the one of inactivation of Ca²⁺ release in skeletal muscle fibers.

Our findings that SR depletion occurs in response to depolarizations of unforeseen small amplitude that do not provoke inactivation have physiological relevance, especially in the context of sustained muscle activity. Intense muscle exercise leading to fatigue is known to induce considerable electrolyte changes, particularly loss of intracellular K⁺ and accumulation of interstitial and tubular K⁺ (Sejersted and Sjøgaard, 2000). These K⁺ shifts have been shown to lead to a decrease of up to 18 mV of the resting membrane potential in mouse extensor digitorum longus muscle after 1 min of repetitive tetanic stimulation (Juel, 1986). Our study indicates that such changes in the resting membrane potential should produce SR Ca²⁺ depletion, contributing to the decrease of contractile performance. Muscle fatigue caused by SR Ca²⁺ depletion is thus expected to occur not only during muscle activity (see below) but also during resting intervals before recovery of electrolyte changes takes place. Additionally, our results emphasize that the control of internal membrane potential is an absolute prerequisite when investigating Ca²⁺ signaling in skeletal muscle, because uncontrolled changes in the membrane potential even of a few millivolts may produce alteration of the SR Ca²⁺ content that cannot be detected with cytosolic Ca²⁺ indicators and affect in this way the amount of Ca²⁺ released.

In physiological conditions, the excitatory signal consists of a train of action potentials. From frog to human skeletal muscle, it has long been known that the contractile response declines during continuous high frequency stimulation, caused in part by the decline of the Ca²⁺ signal that initiates contraction (Allen et al., 2008). In this study, we showed that the cytosolic Ca²⁺ signal declined during a 1-min train of action potentials, but our SR Ca²⁺ measurements clearly indicated that SR Ca²⁺ depletion was maintained during the entire duration of the stimulation, giving evidence that Ca²⁺ release channels do not close. We concluded from these experiments that the cytosolic Ca²⁺ signal decline during a sustained train of action potentials exclusively arose from SR depletion. The mechanisms involved in the Ca²⁺ signal decline are consequently different when the fiber is stimulated by long-lasting depolarizing pulses of high voltage or by a train of action potentials, as SR Ca²⁺ depletion together with inactivation occurs in the first situation, whereas only SR Ca²⁺ depletion takes place in the second. Comparing the decay of Ca²⁺ signals in response to a train of action potentials (52% after 50 s) and in response to a single voltage pulse to +20 mV (84% after 50 s) led us to estimate that during a voltage step, 52% of the decay may be caused by depletion, whereas 32% may be caused by inactivation. Our last series of experiments gives an explanation for this discrepancy by demonstrating that the depolarization produced by an action potential is likely too short to turn on inactivation processes and should last longer than 10 ms to induce Ca²⁺ release channel closure. These results have important physiological consequences. The physiological excitatory stimulus of muscle fibers, which consists of a train of action potentials, is indeed optimal; the tetanic rate of depolarization induced by a train of action potentials allows the SR to reach its maximal Ca²⁺ release capacity (Ziman et al., 2010; Robin et al., 2012) but without producing voltage-dependent inactivation of Ca²⁺ release channels. This guarantees that maximal contractile force can be generated. However, in return, SR Ca²⁺ depletion cannot be prevented and should thus predominantly contribute to the decline of Ca²⁺ release and force during long-lasting stimulation of muscle leading to fatigue.

We thank Vincent Jacquemond for critical comments on the manuscript.

This work was supported by grants from the Université Lyon 1, the Centre National de la Recherche Scientifique, and the Association Française contre les Myopathies.

Edward N. Pugh Jr. served as editor.