Mechanistic insights into store-operated Ca2+ entry during excitation-contraction coupling in skeletal muscle

BBA - Molecular Cell Research 1866 (2019) 1239–1248

Contents lists available at ScienceDirect

BBA - Molecular Cell Research journal homepage: www.elsevier.com/locate/bbamcr

Mechanistic insights into store-operated Ca2+ entry during excitationcontraction coupling in skeletal muscle

T

⁎

Xaver Koeniga, , Rocky H. Choib, Klaus Schickera, Daniel P. Singhb, Karlheinz Hilbera, Bradley S. Launikonisb a b

Center for Physiology and Pharmacology, Medical University of Vienna, Schwarzspanierstrasse 17, 1090 Wien, Austria School of Biomedical Sciences, The University of Queensland, Brisbane, QLD 4072, Australia

A R T I C LE I N FO

A B S T R A C T

Keywords: Store-operated calcium entry Skeletal muscle Confocal microscopy Calcium stim1 orai1 Excitation-contraction coupling Electrical field stimulation Skinned fibre

Skeletal muscle fibres support store-operated Ca2+-entry (SOCE) across the t-tubular membrane upon exhaustive depletion of Ca2+ from the sarcoplasmic reticulum (SR). Recently we demonstrated the presence of a novel mode of SOCE activated under conditions of maintained [Ca2+]SR. This phasic SOCE manifested in a fast and transient manner in synchrony with excitation contraction (EC)-coupling mediated SR Ca2+-release (Communications Biology 1:31, doi: https://doi.org/10.1038/s42003-018-0033-7). Stromal interaction molecule 1 (STIM1) and calcium release-activated calcium channel 1 (ORAI1), positioned at the SR and t-system membranes, respectively, are the considered molecular correlate of SOCE. The evidence suggests that at the triads, where the terminal cisternae of the SR sandwich a t-tubule, STIM1 and ORAI1 proteins pre-position to allow for enhanced SOCE transduction. Here we show that phasic SOCE is not only shaped by global [Ca2+]SR but provide evidence for a local activation within nanodomains at the terminal cisternae of the SR. This feature may allow SOCE to modulate [Ca2+]SR during EC coupling. We define SOCE to occur on the same timescale as EC coupling and determine the temporal coherence of SOCE activation to SR Ca2+ release. We derive a delay of 0.3 ms reflecting diffusive Ca2+equilibration at the luminal ryanodine receptor 1 (RyR1) channel mouth upon SR Ca2+-release. Numerical simulations of Ca2+-calsequestrin binding estimates a characteristic diffusion length and confines an upper limit for the spatial distance between STIM1 and RyR1. Experimental evidence for a 4- fold change in t-system Ca2+permeability upon prolonged electrical stimulation in conjunction with numerical simulations of Ca2+-STIM1 binding suggests a Ca2+ dissociation constant of STIM1 below 0.35 mM. Our results show that phasic SOCE is intimately linked with RyR opening and closing, with only μs delays, because [Ca2+] in the terminal cisternae is just above the threshold for Ca2+ dissociation from STIM1 under physiological resting conditions. This article is part of a Special Issue entitled: ECS Meeting edited by Claus Heizmann, Joachim Krebs and Jacques Haiech.

1. Introduction Skeletal muscle fibres support a robust store-operated calcium entry (SOCE) mechanism that is present at the level of every sarcomere [1–3]. Stromal interaction molecule 1 (STIM1) and calcium release-activated calcium channel 1 (ORAI1) are considered the molecular correlate of SOCE [4–11]. STIM1 proteins are located at the sarcoplasmic reticulum (SR) membrane and sense the luminal depletion of calcium (Ca2+) to trigger the opening of coupled Ca2+-selective ORAI1 channels in the tubular (t)-system membrane [4,12]. STIM1 and ORAI1 proteins are expressed in mature, healthy skeletal muscle [13,14] and deficiency of

⁎

either protein abolishes SOCE and results in the development of skeletal myopathy, developmental defects and fibre-type shifts [9,15–17]. The canonical activation of SOCE, as reported from non-excitable cells, involves STIM1 puncta formation and translocation, and alignment of SR- and plasma-membrane junction sites [18–21]. Skeletal muscle fibres are remarkable in that such junction sites between the tsystem and SR membranes pre-exist. At the ‘triads’, the t-system, a complex network of plasma membrane invaginations, couples to the SR to provide the platform for excitation contraction (EC) coupling. Action potentials (APs) propagate into the t-system to activate the voltage sensors (L-type Ca2+ channels) and trigger Ca2+-release from the SR

Corresponding author. E-mail address: [email protected] (X. Koenig).

https://doi.org/10.1016/j.bbamcr.2019.02.014 Received 30 November 2018; Received in revised form 20 February 2019; Accepted 22 February 2019 Available online 27 February 2019 0167-4889/ © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

through the ryanodine receptor 1 (RyR1; [22]). Early studies showed SOCE to manifest upon complete depletion of sarcoplasmic reticulum (SR) Ca2+ [1,9,10,23–26]. This tonic form of SOCE in Ca2+ depleted muscle is non-physiological as SR Ca2+ levels are largely maintained during muscle activity and are immediately restored at the cessation of tetani [1,27–29]. This is accomplished by the high capacity of the SR Ca2+-ATPase (SERCA) that pumps Ca2+ back into the SR and calsequestrin that strongly buffers Ca2+ inside the SR ([Ca2+]SR; [30]). Recently, we demonstrated physiological stimulation of the muscle to activate SOCE acutely. In fast twitch skeletal muscle fibres SOCE appeared in a fast and transient manner during individual action potentials and after each AP in a train of stimulations [29]. This phasic mode of SOCE coincided with SR Ca2+ release into the cytosol and occurred at a millisecond timescale. STIM1 and ORAI1 are positioned at the terminal cisternae (TC) and the t-system, respectively [9,31], and neither SR remodelling nor STIM1 redistribution is required for the activation of SOCE [4]. SOCE occurs at the transverse but not the longitudinal elements of the tsystem [32]. A long splice variant of STIM1 (STIM1L), predominantly expressed in skeletal muscle and brain [33], forms permanent clusters with ORAI1 and allows immediate activation of SOCE [14,33]. Accordingly, a compelling model to explain phasic SOCE depicts STIM1 and ORAI1 to pre-position, and maybe even physically precouple, within the triads [10,29]. However, it remains unresolved how SOCE is activated rapidly during EC coupling without average [Ca2+]SR falling to levels that are expected to activate SOCE [34,35]. Measurements of [Ca2+]SR transients during EC coupling have been limited to reporting the globally averaged [Ca2+]SR, not that within the TC [27,36,37]. There, STIM1 and ORAI1 may be able to respond almost without delay to the opening and closing of RyRs to provide a phasic SOCE flux. If physiological SOCE is dependent on the gradient of Ca2+ within the SR during EC coupling, we can make a number of predictions. The activation of SOCE can be manipulated by changing the Ca2+ content of the SR; a globally averaged [Ca2+]SR should not correlate with SOCE flux; and the kinetic delay between Ca2+ release and STIM1 activation should be similar to that of action potential stimulation and Ca2+ release during EC coupling. We examined these predictions using Ca2+sensitive dyes trapped in the t-system, SR or within the open cytoplasmic compartment of skinned skeletal muscle fibres and fast tracking of changes in Ca2+-dependent fluorescence during normal EC coupling [2,24,29,38].

Fig. 1. Free [Ca2+]SR as set by different free [Ca2+]cyto in skinned rat EDL fibres. [Ca2+]SR was derived from the calibrated fluo-5N signal using an in situ derived kD,Ca of 0.4 mM (see Methods). Steady-state [Ca2+]SR was determined at different free cytosolic Ca2+ concentrations ([Ca2+]cyto) of 28, 67, 200 and 1342 nM Ca2+ and 1 mM free [Mg2+] to load the SR with Ca2+. Values are derived from 6 to 11 fibres per data point and given as mean ± SEM.

0.71 ± 0.08 mM (mean ± SEM). Note that under the condition used ([Ca2+]cyto buffered by 50 mM EGTA), that the [Ca2+]SR could rise from below 0.05 mM in the presence of caffeine and zero Ca2+ to the steady states displayed in Fig. 1 upon washout of caffeine and addition of Ca2+ in < 10 s (not shown). We reasoned that if SOCE is indeed activated by Ca2+ dissociation from STIM1 during opening of the RyR1, [Ca2+]SR would be a key factor for determining the activation threshold. To this end we returned to our technique for measuring phasic SOCE, where we combined the recording of rhod-5N trapped in the t-system and of fluo-4 from the open cytosol of skinned rat EDL fibres (see Methods; [29]). According to Fig. 1 we varied [Ca2+]SR by manipulating [Ca2+]cyto. Note that [Ca2+]cyto was buffered by 10 mM EGTA, which is necessary to inhibit Ca2+ uptake by the t-system during EC coupling, allowing SOCE flux to dominate the change in [Ca2+]t-sys [29]. Note that the cytoplasmic [Ca2+] transients remain observable under 10 mM EGTA, which mark the successful stimulation of AP-induced Ca2+ release. A calibrated signal of both fluorescent signals during electrical field stimulation can be seen in Fig. 2a. Phasic SOCE, as reflected by the sharp step-wise depletions of [Ca2+]tsys, was synchronously activated with every AP during the entire train of stimulation. For lower [Ca2+]cyto, 28 and 67 nM, the summation of individual depletion steps led to a gradual depletion of overall [Ca2+]tsys, as reported previously [29]. Importantly, phasic SOCE showed a clear dependence on [Ca2+]cyto. At low concentrations (28, and 67 nM) phasic SOCE was readily seen during a 2 Hz train of APs, and repeated activation reduced steady-state [Ca2+]tsys significantly. At higher concentrations of [Ca2+]cyto, and thus also at higher concentrations of [Ca2+]SR (compare Fig. 1), phasic SOCE appeared of reduced amplitude at 200 nM [Ca2+]cyto, and was apparently absent at 1342 nM [Ca2+]cyto. A summary of steady-state [Ca2+]tsys values derived at the end of stimulation train is depicted in Fig. 2b. Note the basal increase of [Ca2+]cyto for 200 and 1342 nM upon prolonged stimulation due to a reduction in buffering power at higher [Ca2+]cyto (Because [EGTA]total is the same across the range of [Ca2+]cyto solutions, [EGTA]free is reduced as [Ca2+]cyto increases). Phasic SOCE could depend on [Ca2+]cyto in three ways. (i) Higher 2+ [Ca ]cyto would allow the extrusion of Ca2+ at higher rates via the Ca2+ extrusion machinery of the t-system, the plasma membrane Ca2+ATPase (PMCA) and the Na+/Ca2+-exchanger (NCX). These higher extrusion rates and thus faster recovery of [Ca2+]t-sys after depletion would compete for the observed step-wise depletion. (However, note that EGTA was present in mM levels to blunt the uptake of Ca2+ by the t-system.) (ii) Higher [Ca2+]cyto increases [Ca2+]SR (Fig. 1) to affect SOCE activation. Thus, the same amount of Ca2+ released during a

2. Results 2.1. Cytosolic Ca2+ concentrations shape phasic SOCE by changing intra SR free Ca2+ levels First, we wanted to know if free intra SR Ca2+ concentrations ([Ca2+]SR; here and in the following [Ca2+] will always denote free Ca2+ concentrations) would affect the activation of phasic SOCE. For our previous measurements of phasic SOCE we relied on the use of skinned fibres of rat extensor digitorum longus (EDL) muscle [29], which allows for a precise manipulation of the fibre's cytosol. Because a direct manipulation of [Ca2+]SR is not possible with this approach, we determined the dependence of [Ca2+]SR on [Ca2+]cyto beforehand, as changing of [Ca2+]cyto is expected to affect the steady level of [Ca2+]SR reached. [Ca2+]SR under each condition was assessed with the Ca2+sensitive dye fluo-5N loaded into the membrane bound compartments of skinned fibres using a modified version of the fluo-5N AM technique [39–41]. The presence of Carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP) limited mitochondrial accumulation of the dye without compromising SR loading (see Methods). The determined levels of [Ca2+]SR changed significantly when we raised [Ca2+]cyto from 28 to 1342 nM, a summary of which is given in Fig. 1. At 67 nM [Ca2+]cyto, which is close to expected endogenous levels, [Ca2+]SR was 1240

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

a

c

b

d

Fig. 2. Phasic SOCE and [Ca2+]cyto. Skinned rat EDL muscle fibres were loaded with rhod-5N trapped in the sealed t-system and fluo-4 in the open cytosol and electrically field stimulated during confocal imaging (xyt scans). Mean fluorescence intensities, recorded at 55 frames s−1, were monitored over time and calibrated to yield [Ca2+]tsys and [Ca2+]cyto, respectively (see Methods). (a) [Ca2+]tsys (left axes) and [Ca2+]cyto (right axes) during a train of 2 Hz stimulation started shortly after beginning of the recording. Fibres were bathed in solutions containing different levels of free [Ca2+]cyto, 28, 67, 200, and 1342 nM. Decline of [Ca2+]tsys upon stimulation was fit with a mono-exponential function (green). (b) Summary of stead-state values of [Ca2+]tsys as induced by the train of stimulation and derived from the mono exponential fits for the different levels of [Ca2+]cyto tested. Data are from n = 4 fibres and given as box and whiskers with min and max values. (c) Overlay of three recordings (aligned to the beginning of stimulation, grey arrow) of [Ca2+]t-sys as derived from the same fibre; in the presence of 67 nM [Ca2+]cyto (black trace), in the presence of 200 nM [Ca2+]cyto (dark blue trace), or after preloading in 200 nM for 2 min but in the presence of 67 nM [Ca2+]cyto, as before (light blue). End of stimulation for individual traces is indicated by an arrow in respective colour. (d) Typical recordings of [Ca2+]tsys and [Ca2+]cyto over time during prolonged electrical stimulation. Fibre was first bathed in 67 nM [Ca2+]cyto and [Ca2+]tsys (black) and [Ca2+]cyto (grey). After stimulation fibre was allowed to recover and then solution was exchanged to one free of Ca2+ (Ca2+-free) but preserved EGTA (10 mM). Again [Ca2+]tsys (olive) and [Ca2+]cyto (light olive) was monitored; the figure shows an overlay of both conditions. (For interpretation of the references to colour in this figure legend, the reader is referred to the online version of this chapter.)

appeared similar to the recordings in 200 nM but then deviated and approached the recording in 67 nM [Ca2+]cyto. The blunted activation of SOCE under conditions when the SR was loaded to near maximal levels (dark blue trace) compared to control (light blue), despite identical recording conditions (67 nM [Ca2+]cyto), suggested the enhanced [Ca2+]SR as the responsible factor. Our data are not consistent with differences in t-system Ca2+ re-uptake countering phasic SOCE and also suggest no prominent role of CDI to inhibit phasic SOCE up to 200 nM [Ca2+]cyto. At 1342 nM [Ca2+]cyto CDI might contribute significantly but the current approach does not allow to us to separate this from high [Ca2+]SR maintaining STIM1 bound with Ca2+ during AP-induced Ca2+ release. Again, it is important to note that during EC coupling in the presence of 67 nM Ca2+, regardless of the initial [Ca2+] level (whether the SR was loaded at 67 or 200 nM Ca2+) the [EGTA]free was the same in both cases, providing identical Ca2+-buffering power to dampen t-system Ca2+ uptake during EC coupling. In a second set of experiments we repeated to record activation of phasic SOCE at 67 nM [Ca2+]cyto (Fig. 2d, black trace). This time, however, after allowing the fibre to recover, we exchanged to a Ca2+ free solution before challenging it again with a 2 Hz train (Fig. 2d, green trace). Importantly, for both conditions SR loads are the same at

single AP would only activate a much smaller fraction of STIM proteins at high [Ca2+]cyto compared to conditions of low [Ca2+]cyto. (iii) As for many Ca2+ channels, Orai1 channels exhibit a Ca2+-dependent inactivation (CDI) [42–44] that could contribute to phasic SOCE inhibition at high [Ca2+]cyto. To discriminate between these options, we recorded phasic SOCE during a train of APs under control conditions, i.e. at 67 nM [Ca2+]cyto (Fig. 2c, black trace). Then we equilibrated the fibre in the same solution but increased [Ca2+]cyto to 200 nM for 2 min. Compared to keeping the fibre in 67 nM [Ca2+]cyto, this loads the SR to near maximal levels (about 1 mM; Fig. 1). Repeating the stimulation showed that under these conditions [Ca2+]t-sys level was higher (consistent with [24]) and phasic SOCE was significantly reduced (Fig. 2c, dark blue trace; also compare with Fig. 2a top right and bottom left). Finally, we loaded the fibre again in 200 nM for 2 min but then exchanged the solution back to the original one (67 nM) and started stimulation immediately thereafter (Fig. 2c, light blue trace). Note that [Ca2+]SR continues to stay high (and so does [Ca2+]t-sys) after switching back to low [Ca2+]cyto [45] and that increasing the Ca2+ load of the SR does not change the amount of Ca2+ released with a single action potential [46]. From the overlay of all three traces it can be seen that SOCE first 1241

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

the solution was then switched to one in which Ca2+ was omitted (Ca2+ free) and which contained cyclopiazonic acid (CPA) to inhibit SERCA function. A concentration of 100 μM CPA fully inhibited Ca2+ re-uptake into the SR in our preparation (Suppl. Fig. 1) and resulted in a gradual depletion of the fully loaded [Ca2+]SR via constant leakage through the RyR, as [Ca2+]SR is no longer replenished by the SR Ca2+ pump (Fig. 3b, c & Suppl. Fig. 3). Similar results were obtained by application of caffeine or reduction in [Mg2+]cyto (Suppl. Fig. 2), two additional conditions known to deplete [Ca2+]SR, also via the RyR [24,47]. In all these measurements SOCE flux increased to a maximal level right after application of CPA (within a few seconds) and then gradually reduced because of the diminishing Ca2+-driving force with the progressive depletion of [Ca2+]t-sys. Thus, although bulk [Ca2+]SR is still high, as reported by the fluo-5N signal, SOCE can be activated under conditions of significant RyR leak [70]. Note that a blockade of the RyR with tetracaine [70] during the application of CPA does not allow SOCE to activate (Suppl. Fig. 3). It is important to note that fluo-5N monitors the globally averaged [Ca2+]SR. The [Ca2+] near the STIM1 proteins near the RyRs will be less than the average [Ca2+]SR because the open RyRs are the source of Ca2+ exit from the SR in CPA, low Mg2+ or caffeine. Also, prolonged depletion of the SR upon application of CPA (in the order of several minutes) may lead to the recruitment of SOCE complexes along different pathways that rely on the migration of additional pools of STIM1 and might involve other Ca2+ channels than ORAI1 [31,33,48]. These aspects, however, haven't been addressed here.

the start of stimulation. An overlay of both traces reveals that the depletion pattern under Ca2+ free conditions exactly follows the depletion pattern when [Ca2+]cyto was 67 nM at the beginning of stimulation and the re-uptake of Ca2+ in between stimuli appears similar. However, with continuing stimulation [Ca2+]t-sys eventually deviated and became fully depleted when exposed to Ca2+ free conditions. Note that there is no more reuptake of Ca2+ when train of stimulation was stopped at the end of the recording. Thus, from the initial [Ca2+]SR, only phasic SOCE is activated in either 67 nM or a Ca2+ free solution. The bathing solution is strongly buffered with 10 mM EGTA, which competes with the SR Ca2+ pump for the Ca2+ that is released during an AP. After a few releases of Ca2+ (because of the action of EGTA sequestering Ca2+, preventing it from completely refilling the SR after each AP) this critically depleted the SR in the zero Ca2+ solution to chronically activate a subset of SOCE channels independently of further RyR openings. Overall, Fig. 2 shows that SOCE is sensitive to the [Ca2+]SR and that CDI may contribute at the highest [Ca2+]cyto tested. Whereas SOCE is not easily observed at relatively high [Ca2+]SR, phasic SOCE operates at “endogenous” [Ca2+]SR as set by normal resting [Ca2+]cyto. Phasic SOCE continues to be activated during EC coupling when [Ca2+]SR reaches low levels but is joined by a chronic activation of SOCE. We note that this nature of SOCE activation will lead to a depletion of the SR Ca2+ stores when levels are high (removal of Ca2+ from the fibre via NCX and PMCA upon each AP) and a convergence towards endogenous store levels (when the extrusion of Ca2+ by NCX and PMCA is balanced by the influx through SOCE).

2.3. The temporal delay between SR Ca2+ release and SOCE 2.2. SOCE is activated prior to depletion of bulk [Ca2+]SR It was now of interest to us to gain a better understanding of the underlying dynamics between SR Ca2+release and the activation of SOCE because, if SOCE is activated following APs, the activation and deactivation of SOCE needs to occur within the time between APs. We could previously show recordings of phasic SOCE at high temporal resolution to demonstrate that phasic SOCE activates on a timescale of only several milliseconds [29]. Here we went further, in that we wanted to determine the precise temporal coherence between SR Ca2+ release

Next, we wanted to directly assess how [Ca2+]SR affects the activation of SOCE. To this end we performed experiments in which we combined the loading of the t-system (with rhod-5N) by the additional loading of the SR with fluo-5N [39–41] (see Methods). A typical experiment is shown in Fig. 3. The fibre was bathed in a solution that fully loaded the SR to near maximal levels (200 nM free Ca2+). After a steady-state was reached,

a TL

rhod-5N

fluo-5N

50 µm

b

c

Fig. 3. Activation of SOCE during the depletion of [Ca2+]SR with CPA (a) Transmitted light image (left), fluo-5N fluorescence (middle), and rhod-5N fluorescence (right) of a typical recording. (b) Time course of the fluorescence intensity as reported by rhod-5N trapped in the t-system and fluo-5N loaded into the SR. fluorescence recorded at a sampling rate of 1 frame−1 s was averaged across the fibre area. Applied [Ca2+]cyto levels are given in nM. Minimal (Fmin) and maximal (Fmax) fluorescence determined in the presence of an ionophore at the end of the recording allowed a calibration of the signal (see Methods). Light grey bars indicate the times of solution exchange. (c) Calcium concentration in the SR and t-system ([Ca2+]SR and [Ca2+]t-sys) during the application of cyclopiazonic acid (CPA) as derived from the calibrated signal shown in 3b (left axis). t-system SOCE Ca2+-flux as derived from the time derivative of [Ca2+]t-sys (right axis). Representative example of n = 5 independent experiments. 1242

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

of phasic SOCE, as derived from 5 fibres amounted to 0.29 ± 0.06 ms (mean ± SEM; Fig. 4d). Fluo-4 exhibits a fast binding and unbinding of Ca2+, but the kinetics is still too slow to accurately follow the fast release of Ca2+ from the SR (see e.g. [49]). To check how our measured fluorescence signals would be affected by the chosen experimental conditions (respective dye and EGTA concentrations) we simulated the effects of our buffering system using a one compartment model on a given hypothetical ‘true’ [Ca2+]cyto transient to understand the imposed effects on the obtained fluorescent recordings (see Methods). It can be seen in Fig. 4e that the recorded fluo-4 signal traces behind the ‘true’ [Ca2+]cyto transient, but that the point of inflection (at the very beginning of Ca2+ release) is accurately reported. Thus, our simulation suggests that the derived difference in the intercept values is not affected by the kinetics of fluo4.

a fluo-4

rhod-5N

b

c

2.4. t-system permeability changes during prolonged field stimulation

d

In a last set of experiments we used the same experimental technique as employed in Fig. 2, but significantly extended the period of APs stimulation from < 10 s to > 20 s (Fig. 5). As mentioned earlier, the step-wise changes in [Ca2+]tsys reflect phasic activation of SOCE. We corrected for the change in driving force during the opening of SOCE channels caused by the drops in [Ca2+]t-sys (Δ [Ca2+]t-sys) to derive respective values for the transient change in t-system membrane Ca2+-permeability (Fig. 5b). A quantification of relative permeability values for succeeding stimulation pulses is shown in Fig. 5b. It becomes apparent that relative permeability increases about 4-fold from the beginning to the end of stimulation.

e

2.5. Estimating the operating point of phasic SOCE About 2 mM of Ca2+ is released during a single AP in EDL fibres (per SR volume, [29,46]). Taking into account that Ca2+ is released solely from the terminal cisternae (TC), the amount corresponds to a loss of 9.3 mM of total Ca2+ within the TC volume (TC and SR cover 2 and 9.3% of fibre volume, respectively; [51,52]). To understand how this change in total [Ca2+]TC would affect free [Ca2+]TC we simulated Ca2+-binding to calsequestrin (CASQ), the predominant Ca2+-buffer in the SR (see Methods). Our numerical calculations showed that upon a single AP total [Ca2+]TC drops from 13 mM (endogenous levels [30]) to 3.7 mM, forcing a reduction in free [Ca2+]SR from 0.7 mM to 0.46 mM (Fig. 6a). We further asked in what way this drop in [Ca2+]TC would affect the 2+ Ca -occupancy of STIM1 during an AP. Thus, we simulated Ca2+binding to STIM1 with reported high cooperativity and a range of different Kd values (Fig. 6b; see Methods; [53,54]). Ca2+ release during an action potential is a very stereotyped event. Regardless of the [Ca2+]SR, Ca2+-dependent inactivation (CDI) of the cytoplasmic face of the RyR closes the channel after the same amount of Ca2+ has been released at low AP frequencies [46]. CDI thus provides a means to maintain constant Ca2+ release quanta for the entire time period of stimulation (an overestimate as Ca2+ release quanta will certainly diminish in the course of stimulation as SR gradually depletes at 2 Hz stimulation at 67 nM [Ca2+]cyto). Based on the observed 4-fold increase in relative permeability (Fig. 5b), it follows that the maximal change in STIM1 occupancy induced by the first AP must not exceed 25% (because it needs to be able to increase 4-fold during the course of stimulation). As can be seen in Fig. 6b this holds true for Kd values up to 0.3 mM, but already clearly fails at a value of 0.4 mM. Thus, the combination of experimental data and numerical simulations suggest an upper limit of about 0.3 mM for the Kd of STIM1 in skeletal muscle. Taken together, under physiological resting conditions (67 nM [Ca2+]cyto) when [Ca2+]SR is 0.7 mM (Fig. 1), STIM1 is almost fully bound by Ca2+, and therefore SOCE is shut. However, because [Ca2+]SR is close to the Kd-value of STIM1, AP-induced SR Ca2+-release

Fig. 4. Temporal coherence of SR Ca2+ release and activation of phasic SOCE. (a) Original recordings of dual xt-linescan imaging of rhod-5N and fluo-4 fluorescence in the t-system and cytosol, respectively. Rat EDL fibres were stimulated at 0.2 Hz and acquired images were averaged over at least 25 consecutive stimulation pulses. (b) The averaged and normalised (see Methods) fluorescence of rhod-5N (t-system, black) and fluo-4 (cytosol, yellow) during a single stimulation pulse was fit with a segmented equation consisting of an initial constant baseline followed by a single exponential rise (blue) and an alpha function (red), respectively. (c) Zoom in of (b) into the region of intercept. (d) Spaghetti plot for the intercept times of the t-system and cytosol signal from a total of 5 different fibres. (e) Numerical simulation on how accurately fluo-4 fluorescence follows a ‘true’ [Ca2+]cyto transient (see Methods). Inset shows a close-up for the initial rising phase of both signals. The intercept time of calibrated fluo-4 signal is not influenced by the dye kinetic as it matches the intercept time of the ‘true’ Ca-transient (blue). (For interpretation of the references to colour in this figure legend, the reader is referred to the online version of this chapter.)

and phasic activation of SOCE. To this end we again combined fluorescent measurement of rhod-5N loaded into the t-system with fluo-4 present in the cytosol to obtain simultaneous recordings of [Ca2+]t-sys and [Ca2+]cyto (see Methods). The necessary temporal resolution was obtained with confocal linescans (133 μs line−1). An example of such an experiment can be seen in Fig. 4. Stimulation pulses were delivered at a frequency of 0.2 Hz to guarantee a full recovery of the signals in between pulses. To increase the signal to noise ratio multiple stimulation pulses within a recording were averaged using a custom written software routine (see Methods). Fig. 4b shows the result of this analysis after normalisation of the data. Fitting of the data allowed to derive intercept times and to determine the temporal delay between the onset of SR Ca2+-release and depletion of the t-system by phasic SOCE (Fig. 4b, c). The mean difference in intercept values, i.e. the delay between SR Ca2+ release and activation 1243

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

a

b

Fig. 5. Change of relative Ca2+-permeability during prolonged electrical stimulation (a) Spatially averaged and calibrated rhod-5N fluorescence ([Ca2+]tsys) during prolonged 2 Hz electrical field stimulation (at 67 nM [Ca2+]cyto). Individual t-system depletion steps (Δ[Ca2+]tsys) at each stimulation pulse were calculated as the difference of [Ca2+]tsys before and immediately after stimulation. Δ[Ca2+]tsys reflects a change in the Ca2+-permeability (PCa) of the t-system membrane, which we estimated by the first derivative divided by the driving force (t-system Ca2+ flux follows the Goldman-Hodgkin-Katz flux equation. Under the assumption of a constant membrane voltage and that [Ca2+]cyto ≪ [Ca2+]tsys, Δ[Ca2+]tsys/Δt is proportional to PCa* [Ca2+]t-sys; [50]). (b) A plot of the relative permeability values for each pulse during prolonged stimulation. Blue line represents best fit with an exponential function reaching a plateau at 0.23. Data are derived from n = 4 fibres, and values are given as mean ± SEM. (For interpretation of the references to colour in this figure legend, the reader is referred to the online version of this chapter.)

causes significant Ca2+ dissociation from STIM1 and activation of SOCE.

persistent SOCE when [Ca2+]cyto was buffered to levels below 10 nM in 10 mM EGTA (“zero Ca2+”) that more readily caused the SR to deplete during trains of action potentials, as less Ca2+ is available to refill the SR after each AP-induced Ca2+ release. Our results suggest that the prepositioning of STIM1 at the terminal cisternae (TC) of the SR allows STIM1 to constantly detect Ca2+ levels near the luminal face of the RyRs, to make immediate adjustments to the permeability of the tsystem SOCE channels during muscle activity. This feature of skeletal muscle allows SOCE, in conjunction with t-system Ca2+ extrusion, to work together to promote the endogenous [Ca2+]SR during EC coupling. We acknowledge that at higher [Ca2+]cyto (> 200 nM), where phasic SOCE appeared to be blunted by high [Ca2+]SR, a second possibility, Ca2+-dependent inactivation (CDI) of the Orai1 channel, could be occurring. It is not simple to distinguish between these two effects in skinned fibres. However, our manipulation of the cytoplasm, to load at 200 nM to a high [Ca2+]SR and then track t-system and cytoplasmic Ca2+ during stimulation after returning the preparation to 67 nM [Ca2+]cyto did provide a control for the effect a high steady [Ca2+]cyto on Orai1 CDI (Fig. 2). A characterization of Orai1 CDI in skeletal muscle fibres is currently lacking and required to resolve these issues. While STIM1 at the TC does not respond directly to the global averaged [Ca2+]SR, relatively high levels of bulk [Ca2+]SR, for example, will equilibrate with the TC to affect the likelihood of Ca2+-STIM1 remaining bound during an AP-induced Ca2+ release. The higher the

3. Discussion Recently, we used a fluorescence-based technique that allowed for a direct measurement of SOCE during EC-coupling in skeletal muscle fibres [29]. Using this approach we demonstrated the presence of a phasic mode of SOCE that is activated in a fast and transient manner during each, individual action potential at least up to 50 Hz stimulation. This provided a basis for understanding the physiological role of SOCE in skeletal muscle. In the present study we define the timescale of SOCE activation is in the < 1 ms range, matching that of EC coupling. This rapid response to local Ca2+ depletions near the luminal mouth of open RyRs is consistent with the relative independence of SOCE activation on globally averaged [Ca2+]SR following fibre stimulation. However, activation of phasic SOCE was shifted when steady-state levels of [Ca2+]SR were raised or lowered, as a consequence of likely altered near membrane [Ca2+] levels. In our experiments we modified [Ca2+]SR by manipulating 2+ [Ca ]cyto, cytoplasmic Ca2+-buffering and the SR Ca2+ pump function (Fig. 2). In these experiments we observed phasic SOCE during EC coupling at normal resting [Ca2+]cyto and [Ca2+]SR; blunted SOCE when [Ca2+]SR was raised to levels greater than endogenous, even in the presence of normal resting [Ca2+]cyto; and activation of phasic and

Fig. 6. Numerical simulations for the binding of Ca2+ to calsequestrin and STIM1. STIM1 - Ca binding (a) Numerical simulation of Ca2+-binding to calsequestrin and how this affects the dependence of total to free [Ca2+]SR. Ca2+-calsequestrin binding was modelled with no or a physiologically high cooperativity, using a Hill slope (nH) of 1 and 3.42, respectively [41]. 13 mM total endogenous [Ca2+]SR (dotted line top, [30]) before stimulation and corresponding 0.7 mM [Ca2+]SR (dotted black line right) drop to respective values of 3.8 and 0.46 mM during an AP, respectively. (b) Numerical simulations of Ca2+-binding to STIM1, assuming highly cooperative binding (Hill slope of 4.7; [53]), for different values of the apparent dissociation constant (Kd = 0.2, 0.3, 0.35, 0.4, 0.6 mM). Black dotted lines indicate the region of the physiological drop in free [Ca2+]SR upon a single AP as derived in (a). Under resting conditions, 0.7 mM of [Ca2+]SR entails almost full occupancy of STIM1, which drops by 25% during and AP (blue dotted line) assuming a Kd value of 0.35 mM. (For interpretation of the references to colour in this figure legend, the reader is referred to the online version of this chapter.)

a

b

2+

1244

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

free Ca2+ the fewer STIM1 will unbind Ca2+ and thus less SOCE activation will occur upon Ca2+ release. We have also shown activation of persistent SOCE during EC coupling by restricting the availability of Ca2+ to the SR Ca2+ pump, to cause a slow depletion of the SR during a train of APs. The lower levels of [Ca2+]SR achieved in our experiments in the presence of 10 mM EGTA may not occur physiologically because of the high Ca2+-buffering power of the SR Ca2+ pump [1,28]. It is important to note that when [Ca2+]SR varies from the endogenous level that this does not affect the AP-induced Ca2+ release, unless [Ca2+]SR gets very low [46]. Therefore, modulation of [Ca2+]SR by SOCE during EC coupling is unlikely to be important for the immediate action of EC coupling or “fatigue resistance” because Ca2+ release can continue normally over a broad range of [Ca2+]SR [17]. Ca2+ release during prolonged muscle use is more significantly restricted by the build-up of cytoplasmic metabolites that interfere with the activation of RyR opening [55] than any consequence of SOCE activity [28]. Furthermore, we should point out that for intact muscle in the body the type of stimulation for the muscle to produce a tetanic force response will be high frequency bursts of APs, in the range 20–100 Hz, depending on the muscle fibre type and animal species [55]. At the higher frequencies of stimulation, subsequent releases of Ca2+ from the SR are affected by the previous releases. The initial AP in a high frequency train releases the largest amount of Ca2+ and the quickly following APs release typically ~20–25% of the Ca2+ released by the first AP [46,56,57]. The reduction in Ca2+ release in the following APs in a train is due to Ca2+-dependent inactivation (CDI) of the RyR occurring more quickly than following the first AP because cytoplasmic Ca2+binding sites are saturated by the Ca2+ released by the first AP, allowing [Ca2+] to build-up more rapidly at the RyR cytoplasmic face following subsequent APs [57]. The smaller release of Ca2+ from the SR during subsequent APs in a train will be accompanied by smaller fluxes of SOCE, as we have demonstrated [29]. Note that this behaviour keeps the t-system Ca2+ extrusion fluxes and SOCE fluxes balanced during a tetanus. Importantly, steady-levels of bulk [Ca2+]SR do not drop significantly during AP-induced Ca2+ release [27,36,37], indicating that the activation of SOCE does not require global changes of bulk [Ca2+]SR. The depletion must be restricted to a Ca2+ nanodomain behind the RyR channel pore. Thus, SOCE can be active despite a full SR given that the RyR is sufficiently opened to cause significant depletion within the local Ca2+ nanodomain. Because [Ca2+] varies more at the luminal face of the RyR than deep in the SR, STIM1 can be a more sensitive detector of Ca2+ at the TC than at the more longitudinal components of the SR where [Ca2+] levels are more static, away from the RyRs. The present study and those performed previously [10,29,58] suggest that the machinery of SOCE and EC coupling co-share the triadic junctions and that both signals are transmitted by direct protein-protein interactions, despite being diametrically directed. The “in-step” activation of SOCE with each action potential [29] shows that both signals must be transmitted on the same timescale. The delay of EC coupling, reflecting the movement of the voltage-sensors to the active release of Ca2+ via the RyR, was determined to lie in the μs range at physiological temperatures [59–61]. More recently, in mouse intact fibres, fast imaging of [Ca2+]cyto transients showed distinct responses of Ca2+ release and inactivation to twin field pulses at 1 kHz [57]. This shows RyR channel activation and inactivation occurs on a μs timescale. The expected high off-rate of STIM1 for Ca2+ also indicates μs kinetics [62] in conjunction with the reported high cooperativity among the 5–6 luminal Ca2+ ions required to bind STIM1 and keep SOCE inactive [53]. Thus, it is expected that the direct coupling of STIM1 and ORAI1 takes place at a μs timescale at physiological temperatures as well. The observed delay of 0.3 ms between SR Ca2+ release and activation of phasic SOCE (Fig. 4) must therefore stem from diffusional delays within the SR lumen. Considering a luminal SR Ca2+ diffusion (10 μm2 s−1;

[63]), we estimate a characteristic diffusion length, x = sqrt(2Dt), of about 80 nm to correspond to the observed temporal delay. RyRs are among the largest known proteins with a lateral physical dimension of 28 × 28 nm [64–66]. STIM1 proteins on the other hand are much smaller, occupying only about 1.5 × 2.5 nm [67]. In the closest conceivable arrangement STIM1 would thus position about 15 nm from the RyR channel pore. However, RyRs cluster to form ‘tetrads’ at the TC of skeletal muscle fibres, which could sterically obstruct a close assembly and significantly increase the mean distance between the luminal site of SR Ca2+-release and the Ca2+-sensor STIM1. Our results and these considerations put STIM1 proteins into the nearest neighbourhood of the RyR receptors, some 15 to 80 nm away from the channel pore. In all our experiments involving electrical field stimulation to activate phasic SOCE we have relied on the use of high Ca2+-buffering by 10 mM EGTA. This experimental condition reduced immediate t-system Ca2+ re-uptake, as carried by PMCA and NCX, and thus allowed to isolate phasic SOCE. Conversely, reducing EGTA to levels comparable to the fibre's endogenous buffering capacity apparently increased the rates of t-system Ca2+ uptake and masked the activation of phasic SOCE [29]. Notably, high EGTA does not affect the intrinsic properties of the Ca2+-release machinery [41]. The high EGTA will also compete with SERCA function and reduce Ca2+-uptake into the SR to slowly deplete [Ca2+]SR upon prolonged repetitive field stimulation. As such the observed saw-tooth like decline in [Ca2+]tsys (Fig. 2, Fig. 5) reflects phasic activation of SOCE at the beginning but, as [Ca2+]SR levels fall with progressive stimulation, gradually involves tonic activation of SOCE. This is reflected in our numerical simulations (Fig. 6b). At the start of stimulation [Ca2+]SR is 0.7 mM and STIM1 is fully Ca2+-bound (Ca2+occupancy of 1). Quantal SR Ca2+ release upon stimulation reduces STIM1 Ca2+ occupancy and activates phasic SOCE. Upon repetitive stimulation [Ca2+]SR declines and converges closer to the Kd of STIM1 for Ca2+, moving leftward along the Ca2+-occupancy curve (Fig. 6b). The same quantal release upon stimulation consequently causes larger amounts of Ca2+ to dissociate from STIM1 and a stronger activation of phasic SOCE as reflected by the observed 4-fold change in t-system permeability (Fig. 5b). On the other hand it entails the tonic activation of SOCE at the same time. In this study we further define the properties of the activation of phasic SOCE during EC coupling, which is seemingly driven by a Ca2+nanodomain around the luminal RyR channel mouth. While global [Ca2+]SR is not directly responsible for phasic SOCE activation, the steady state averaged [Ca2+]SR shapes the response of the Ca2+-nanodomain during quantal Ca2+-release to facilitate or hinder Ca2+dissociation from STIM1 and activation of phasic SOCE. This may be important in modulating [Ca2+]SR during EC coupling. Our kinetic results put phasic SOCE and EC coupling on the same time scale and thus reinforce the notion that both machineries work side by side within the triads of skeletal muscle.

4. Methods 4.1. Animal model and skinned fibre preparation Male Wistar and Oncins France Strain A rats in an age range between 3 and 8 months were used for the present study. Animals were killed by asphyxiation in CO2 and the extensor digitorum longus (EDL) muscle was rapidly excised. Isolation of single fibres and mechanical ‘skinning’ of the fibres was performed as previously described [24,29]. Briefly, EDL muscles were placed in a Petri dish filled with paraffin oil and covered with a layer of Sylgard. Segments of individual fibres were isolated and the sarcolemma was mechanically removed with forceps. The skinned fibres were mounted in a custom built chamber on top of a 1.5 coverslip and filled with physiological salt solution.

1245

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

4.2. Solutions

4.4. Electrical field stimulation

Skinned fibres were bathed in a physiological cytosolic salt solution (internal solution) based on previous work [24,29]. The internal solution consisted of (in mM): 90 HEPES, 10 EGTA, 40 HDTA (1,6-Diaminohexane-N,N,N′,N′-tetraacetic acid), 8.77 MgO, 1.94 CaCO3, 8 ATP*2Na, 10 creatine phosphate*2Na. pH was adjusted to 7.1 ± 0.1 with KOH. Calculated free Ca2+ and Mg2+ concentrations amounted to 67 nM and 1 mM, respectively, with a pre-determined Kd of EGTA for Ca2+ of 200 nM [24]. To change [Ca2+]cyto in the solution to free concentrations of 28, 200, or 1342 nM, the amount of CaCO3 was lowered as in [24], considering the use of 10 mM EGTA. In experiments where we used a Ca2+-free or zero Ca2+ solution no CaCO3 was added. Small impurities of the used chemicals or from contamination by the used glassware will be buffered by 10 mM EGTA to an estimated free Ca2+ concentration below 10 nM.

Electrical field stimulation was delivered by two platinum electrodes running in parallel to the fibre long axis. Individual pulses were applied at a frequency of 2 Hz by rectangular voltage steps of up to 100 V in amplitude and 2–4 ms in duration. The experimental chamber was mounted above a 20× air or 40× water immersion objective on a laser scanning confocal microscope system (Nikon A1R+ or Leica TCS SP8). Images were acquired by using a 12 kHz resonant scanner. Rhod5N was excited by a 561 nm laser line while fluo-4 and fluo-5N were excited by a 488 nm laser line. Emitted light was collected at 40 nm bands with high-sensitivity GaAsP detectors. Independent experiments were performed on individual fibres, which in most cases, because of the experimental demand and low success rate, were also derived from individual animals. 4.5. Linescan analysis

4.3. Loading of fluorescent dyes

Linescan data were analysed by custom written routines using Python 3.6.2. First, fluorescent values for fluo-4 and rhod-5N were spatially averaged along the imaging line. Peaks of Ca2+ transients in the fluo4 channels were detected using the PEAKUTILS package setting the threshold of detection to 5 times the signal root mean square values. From the entire train of transients at 0.2 Hz stimulation, single transients were extracted in windows ranging from 100 data points before to 1000 data points after the detected peak. Sampling time in the recordings was 133 μs line−1. To avoid jitter in the peak positions resulting from temporal undersampling and variations in field stimulus generation, the resulting traces were fit with a combination of an ‘activating’ and an ‘inactivating’ sigmoidal function A ∗ (1/(1 + exp (k1 ∗ (t-shift))) ∗ 1/(1 + exp(k2 ∗ (t-shift)))) using the SCIPY package for python. A, amplitude; t, time; shift, offset time from zero; rate constants k1 and k2, one positive one negative. The time points of the maxima of these fits were then used to accurately align individual transients. Finally, these aligned signals were averaged and normalised for both channels (Fig. 4). Normalised t-system and cytosol signals were then fit with a nonlinear regression to a segmented equation consisting of an initial constant baseline followed by a single exponential rise,

The Ca -sensitive fluorescent dye rhod-5N was trapped in the sealed t-system following our previous work [24,29]. Briefly, small bundles of fibres embedded in paraffin oil were exposed to a bubble of Ringer solution containing 2.5 mM of the Ca2+-sensitive dye rhod-5N. An incubation time of at least 10 min allowed the dye to diffuse into the t-system. Thereafter individual fibres were mechanically skinned using fine forceps. The ‘skinning’ procedure leads to removal of the sarcolemma and seals the entry points of the t-tubules, thus trapping rhod-5N in the enclosed t-system. Calibration of rhod-5N fluorescence and conversion to [Ca2+]tsys was performed with a Kd of about 0.8 mM as in [24]. The Ca2+-sensitive fluorescent dye fluo-4 was added at a concentration of 10 μM to the respective physiological cytosolic salt solutions. Calibration and conversion to [Ca2+]cyto was performed as in [29]. To load the sarcoplasmic reticulum with the Ca2+-sensitive fluorescent dye fluo-5N we followed [39–41] with minor modifications. Individual mechanically skinned fibres were mounted in an experimental chamber and bathed in 100 nM [Ca2+]cyto internal solution (same formulation as above) with 10 μM fluo-5N acetoxymethyl (AM) ester. 10 μM carbonilcyanide p-triflouromethoxyphenylhydrazone (FCCP) and 0.05% Pluronic F-127 detergent were added to decouple mitochondria and to help disperse the AM ester, respectively. Fibres were incubated for 1 h at 30 °C. Thereafter the solution was exchanged to the same internal solution but without fluo-5N AM. Fibres were then incubated for an additional 1 h at room temperature to allow for complete hydrolysis of the acetyl moiety. For SR experiments with fluo5N, fibres were bathed in solutions that contained 50 mM EGTA, with free [Ca2+]cyto calculated in the same manner as for 10 mM EGTA. A detailed description of the formulation of solutions is given in [24]. fluo-5N in the SR was calibrated with [Ca2+]SR using the same methods applied to calibrated fluo-5N and rhod-5N in the t-system of skinned fibres [24]. Briefly, fibres with SR-loaded fluo-5N were exposed to 25 μM Ionomycin and 25 μM of the Ca2+ ionophore A23187 to make all membranes permeable to Ca2+ in the cytoplasmic bathing solution. Cytoplasmic solutions with 50 mM EGTA and no added Ca2+ (0 Ca2+), 300 μM Ca2+ buffered by 10 mM trisodium nitrilotriacetate monohydrate (NTA) and 2 and 5 mM Ca2+ in Na+-based solutions were applied to the fibre while imaging on the confocal microscope. When calibration failed in some fibres, due to vacuole formation, respective mean SR values for a given [Ca2+]cyto (Fig. 1) were taken as a reference value. Florescence signals were allowed to reach a steady state at each [Ca2+]. The fluorescence signals determined at known [Ca2+] allowed determination of the in situ kD,Ca of fluo-5N to be 0.4 mM. This is similar to a value determined inside the SR of cardiomyocytes [68]. 2+

y=c

x < x0

y = c + a ∗ (1 − exp(− k/(x − x0)))

x ≥ x 0,

or an alpha-function,

y=c

x < x0

y = c + a ∗ (x − x0)/(1 + exp((x − x0)/k ))

x ≥ x 0,

respectively. c, constant (baseline); a, amplitude; k, slope factor; x0, point of intercept. 4.6. Numerical simulations 4.6.1. Ca2+-EGTA-Fluo4-system Computational modelling was performed using the IQM Tool box (IntiQuan IQM Tools Repository, https://iqmtools.intiquan.com) and MATLAB 2017b. A single compartment model was defined describing the binding of Ca2+ to EGTA and to fluo-4 as second-order reactions with a 1:1 binding stoichiometry,

[EGTA] + [Ca] ↔ [CaEGTA] [Fluo4] + [Ca] ↔ [CaFluo4] Rate constants were taken from Manno et al. 2013 and Kong et al. 2013 [41,69]. On- and off-rates are given in μM−1 ms−1 and ms−1, respectively. 1246

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

kon_Fluo4 = 0.3 koff_Fluo4 = 0.3 kon_EGTA = 0.015 koff_EGTA = 0.001

Acknowledgements Funding: This work was supported by “Österreichische Muskelforschung” (ÖMF) supported by the “Harley Davidson Charity Fonds” (AP00845OFF/KP00845OFF to X. Koenig) and an Australian Research Council (ARC) Discovery Project Grant (DP180100937) (to B.S.L.). B.S.L. was a Future Fellow (FT140101309) of the ARC.

The change in free cytosolic Ca2+ concentration ([Ca2+]cyto) was described by the following differential equation describing the binding and unbinding to Fluo-4 and EGTA, respectively.

Appendix A. Supplementary data

d/dt[Ca]cyto = −kon_Fluo4 ∗ [Fluo4] ∗ [Ca] + koff_Fluo4 ∗ [CaFluo4] − kon_EGTA ∗ [EGTA] ∗ [Ca]+ koff_EGTA ∗ [CaEGTA]

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.bbamcr.2019.02.014.

The model was solved numerically with the following initial conditions

References

Fluo4(t = 0) = 10 μM CaFluo4(t = 0) = 0 EGTA(t = 0) = 10000 μM CaEGTA(t = 0) = 0 Ca(0) = 0

[1] N. Kurebayashi, Y. Ogawa, Depletion of Ca2+ in the sarcoplasmic reticulum stimulates Ca2+ entry into mouse skeletal muscle fibres, J. Physiol. 533 (2001) 185–199. [2] B.S. Launikonis, M. Barnes, D.G. Stephenson, Identification of the coupling between skeletal muscle store-operated Ca2+ entry and the inositol trisphosphate receptor, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 2941–2944. [3] B.S. Launikonis, R.M. Murphy, J.N. Edwards, Toward the roles of store-operated Ca2+ entry in skeletal muscle, Pflugers Arch. Eur. J. Physiol. 460 (2010) 813–823. [4] R.T. Dirksen, Checking your SOCCs and feet: the molecular mechanisms of Ca2+entry in skeletal muscle, J. Physiol. 587 (2009) 3139–3147. [5] M. Vig, et al., CRACM1 is a plasma membrane protein essential for store-operated Ca2+ Entry. Science, 312 (2006) 1220–1223. [6] M. Vig, et al., Defective mast cell effector functions in mice lacking the CRACM1 pore subunit of store-operated calcium release-activated calcium channels, Nat. Immunol. 9 (2008) 89–96. [7] J. Stiber, et al., STIM1 signalling controls store-operated calcium entry required for development and contractile function in skeletal muscle, Nat. Cell Biol. 10 (2008) 688–697. [8] A.D. Lyfenko, R.T. Dirksen, Differential dependence of store-operated and excitation-coupled Ca2+ entry in skeletal muscle on STIM1 and Orai1, J. Physiol. 586 (2008) 4815–4824. [9] L. Wei-Lapierre, E.M. Carrell, S. Boncompagni, F. Protasi, R.T. Dirksen, Orai1-dependent calcium entry promotes skeletal muscle growth and limits fatigue, Nat. Commun. 4 (2013) 1–12. [10] J.N. Edwards, et al., Ultra-rapid activation and deactivation of store-operated Ca2+entry in skeletal muscle, Cell Calcium 47 (2010) 458–467. [11] S. Feske, et al., A mutation in Orai1 causes immune deficiency by abrogating CRAC channel function, Nature 441 (2006) 179–185. [12] A. Michelucci, M. García-castañeda, S. Boncompagni, R.T. Dirksen, Role of STIM1/ ORAI1-mediated store-operated Ca2+ entry in skeletal muscle physiology and disease, Cell Calcium 76 (2018) 101–115. [13] J.N. Edwards, et al., Upregulation of store-operated Ca2+ entry in dystrophic mdx mouse muscle, Am. J. Physiol. Cell Physiol. 299 (2010) C42–C50. [14] B. Darbellay, S. Arnaudeau, C.R. Bader, S. Konig, L. Bernheim, STIM1L is a new actin-binding splice variant involved in fast repetitive Ca 2+ release, J. Cell Biol. 194 (2011) 335–346. [15] C.A. McCarl, et al., ORAI1 deficiency and lack of store-operated Ca2+ entry cause immunodeficiency, myopathy, and ectodermal dysplasia, J. Allergy Clin. Immunol. 124 (2009). [16] T. Li, et al., STIM1-Ca2+ signaling is required for the hypertrophic growth of skeletal muscle in mice, Mol. Cell. Biol. 32 (2012) 3009–3017. [17] E.M. Carrell, A.R. Coppola, H.J. McBride, R.T. Dirksen, Orai1 enhances muscle endurance by promoting fatigue-resistant type I fiber content but not through acute store-operated Ca2+entry, FASEB J. 30 (2016) 4109–4119. [18] Putney, J. & Box, Q. O. James 14. Putney, Jr. (1986). [19] H.E. Collins, X. Zhu-Mauldin, R.B. Marchase, J.C. Chatham, STIM1/Orai1-mediated SOCE: current perspectives and potential roles in cardiac function and pathology, Am. J. Physiol. Heart Circ. Physiol. 305 (2013) H446–H458. [20] R.S. Lacruz, S. Feske, Diseases caused by mutations in ORAI1 and STIM1, Ann. N. Y. Acad. Sci. 1356 (2015) 45–79. [21] J. Avila-Medina, et al., The complex role of store operated calcium entry pathways and related proteins in the function of cardiac, skeletal and vascular smooth muscle cells, Front. Physiol. 9 (2018). [22] W. Melzer, The role of Ca2+ ions in excitation-contraction coupling of skeletal muscle fibres, Biochim. Biophys. Acta 1241 (1995) 59–116. [23] C. Vandebrouck, D. Martin, M.C. Van Schoor, H. Debaix, P. Gailly, Involvement of TRPC in the abnormal calcium influx observed in dystrophic (mdx) mouse skeletal muscle fibers, J. Cell Biol. 158 (2002) 1089–1096. [24] T.R. Cully, J.N. Edwards, R.M. Murphy, B.S. Launikonis, A quantitative description of tubular system Ca2+handling in fast- and slow-twitch muscle fibres, J. Physiol. 594 (2016) 2795–2810. [25] X. Zhao, et al., Compromised store-operated Ca2+ entry in aged skeletal muscle, Aging Cell 7 (2008) 561–568. [26] F.-X. Boittin, et al., Ca2+−independent phospholipase A2 enhances store-operated Ca2+ entry in dystrophic skeletal muscle fibers, J. Cell Sci. 119 (2006) 3733–3742. [27] M. Canato, et al., Massive alterations of sarcoplasmic reticulum free calcium in

A hypothetical ‘true’ [Ca2+]cyto transient rise was given as ‘input’, modelled as an exponential function with a rise time of 50 μs. The simulation returned the concentration of Ca2+-bound Fluo-4 ([CaFluo4]), which is directly proportional to the measured fluorescence intensity. 4.6.2. Ca2+-calsequestrin system We describe Ca2+-calsequestrin binding in the SR of skeletal muscle fibres using a simplified description with a Hill equation based on fit values derived in Manno et al. (Kd of 0.46 mM and a Hill slope, nH = 3.42; [41]). This formalism assumes multiple Ca2+-binding sites with perfect cooperativity following the reaction scheme

CASQ + Ca ↔ CaCASQ The concentration of Ca2+-bound calsequestrin is then calculated as

[CaCASQ] = [CASQ]t ∗ [Ca]n /([Ca]n + K d n) With [Ca], [CaCASQ], n, and Kd being the free Ca2+ concentration in the SR, the Ca2+-bound calsequestrin concentration, the Hill coefficient, and the dissociation constant, respectively. Using the conservation laws:

[CASQ]t = [B] + [CanCASQ] [Ca]t = [Ca] + n ∗ [CanCASQ] One derives the following equation for [Ca]t as a function of [Ca]

[Ca]t = [Ca] ∗ (1/n ∗ [Ca]n + K d + [Ca](n − 1) ∗ [CASQ]t ) /(1/n ∗ [Ca]n + K d) The total amount of calsequestrin and total endogenous levels of Ca2+ have been determined in rat EDL fibres to be 387 μM and 13 mM, both with respect to SR volume, respectively [30]. To obtain a resting free Ca2+concentration in the SR of 0.7, as determined experimentally (Fig. 1), the number of Ca2+ ions that bind to each calsequestrin molecule was assumed to be 60. This number, matches previously published estimates (binding of 30–80 Ca2+ ions), and implies a total of 23 mM of calsequestrin Ca2+-binding sites in the SR. Conflict of interest The authors declare no conflict of interest. Transparency document The Transparency document associated with this article can be found, in online version. 1247

BBA - Molecular Cell Research 1866 (2019) 1239–1248

X. Koenig, et al.

[28] [29] [30]

[31]

[32]

[33] [34] [35] [36] [37]

[38]

[39]

[40]

[41]

[42] [43] [44] [45]

[46]

[47]

[48]

skeletal muscle fibers lacking calsequestrin revealed by a genetically encoded probe, Proc. Natl. Acad. Sci. 107 (2010) 22326–22331. T.R. Cully, B.S. Launikonis, Store-operated Ca2+ entry is not required for store refilling in skeletal muscle, Clin. Exp. Pharmacol. Physiol. 40 (2013) 338–344. X. Koenig, R.H. Choi, B.S. Launikonis, Store-operated Ca2+ entry is activated by every action potential in skeletal muscle, Commun. Biol. 1 (31) (2018). R.M. Murphy, N.T. Larkins, J.P. Mollica, N.A. Beard, G.D. Lamb, Calsequestrin content and SERCA determine normal and maximal Ca2+ storage levels in sarcoplasmic reticulum of fast- and slow-twitch fibres of rat, J. Physiol. 587 (2009) 443–460. S. Boncompagni, A. Michelucci, L. Pietrangelo, R.T. Dirksen, F. Protasi, Exercisedependent formation of new junctions that promote STIM1-Orai1 assembly in skeletal muscle, Sci. Rep. 7 (2017) 1–12. T.R. Cully, et al., Human skeletal muscle plasmalemma alters its structure to change its Ca2+-handling following heavy-load resistance exercise, Nat. Commun. 8 (2017) 1–10. S. Sauc, et al., STIM1L traps and gates Orai1 channels without remodeling the cortical ER, J. Cell Sci. 128 (2015) 1568–1579. J. Soboloff, B.S. Rothberg, M. Madesh, D.L. Gill, STIM proteins: dynamic calcium signal transducers, Nat. Rev. Mol. Cell Biol. 13 (2012) 549–565. M. Prakriya, R. Lewis, Store-operated calcium channels, Physiol. Rev. 4 (2015) 1383–1436. B.S. Launikonis, et al., Depletion ‘skraps’ and dynamic buffering inside the cellular calcium store, Proc. Natl. Acad. Sci. 103 (2006) 2982–2987. R. Rudolf, P.J. Magalhães, T. Pozzan, Direct in vivo monitoring of sarcoplasmic reticulum Ca2+and cytosolic cAMP dynamics in mouse skeletal muscle, J. Cell Biol. 173 (2006) 187–193. G.D. Lamb, D.G. Stephenson, Measurement of force and calcium release using mechanically- skinned fibers from mammalian skeletal muscle, J. Appl. Physiol. 125 (2018) 1105–1127. A.A. Kabbara, D.G. Allen, The use of the indicator flou-5N to measure sarcoplasmic reticulum calcium in single muscle fibres of the cane toad, J. Physiol. 534 (2001) 87–97. A.P. Ziman, C.W. Ward, G.G. Rodney, W.J. Lederer, R.J. Bloch, Quantitative measurement of Ca2+in the sarcoplasmic reticulum lumen of mammalian skeletal muscle, Biophys. J. 99 (2010) 2705–2714. C. Manno, M. Sztretye, L. Figueroa, P.D. Allen, E. Ríos, Dynamic measurement of the calcium buffering properties of the sarcoplasmic reticulum in mouse skeletal muscle, J. Physiol. 591 (2013) 423–442. A.B. Parekh, J.R. Putney, J. W, Store-operated calcium channels, Physiol. Rev. 85 (2005) 757–810. A. Zweifach, R.S. Lewis, Rapid inactivation of depletion-activated calcium current (ICRAC) due to local calcium feedback, J. Gen. Physiol. 105 (1995) 209–226. L. Fierro, A.B. Parekh, Membrane biology fast calcium-dependent inactivation of calcium release-activated calcium current, Currents 17 (1999) 9–17. B.S. Launikonis, J. Zhou, D. Santiago, G. Brum, E. Ríos, The changes in Ca2+ sparks associated with measured modifications of intra-store Ca2+ concentration in skeletal muscle, J. Gen. Physiol. 128 (2006) 45–54. G.S. Posterino, G.D. Lamb, Effect of sarcoplasmic reticulum Ca2+ content on action potential-induced Ca2+ release in rat skeletal muscle fibres, J. Physiol. 551 (2003) 219–237. G.D. Lamb, M.A. Cellini, D.G. Stephenson, Different Ca2+releasing action of caffeine and depolarisation in skeletal muscle fibres of the rat, J. Physiol. 531 (2001) 715–728. F. Antigny, et al., TRPC1 and TRPC4 channels functionally interact with STIM1L to promote myogenesis and maintain fast repetitive Ca2 +release in human myotubes, Biochim. Biophys. Acta - Mol. Cell Res. 1864 (2017) 806–813.

[49] A.J. Bakker, S.I. Head, D.G. Stephenson, Time course of calcium transients derived from Fura-2 fluorescence measurements in single fast twitch fibres of adult mice and rat myotubes developing in primary culture, Cell Calcium 21 (1997) 359–364. [50] B.S. Launikonis, D.G. Stephenson, O. Friedrich, Rapid Ca2+ flux through the transverse tubular membrane, activated by individual action potentials in mammalian skeletal muscle, J. Physiol. 587 (2009) 2299–2312. [51] T. Yoshioka, A.P. Somlyo, Calcium and magnesium contents and volume of the terminal cisternae in caffeine-treated skeletal muscle, J. Cell Biol. 99 (1984) 558–568. [52] B.R. Eisenberg, Quantitative ultrastructure of mammalian skeletal muscle, Handb. Physiol. Skelet. Muscle (1983) 73–112, https://doi.org/10.1002/cphy.cp100103. [53] A. Gudlur, et al., Calcium sensing by the STIM1 ER-luminal domain, Nat. Commun. 9 (2018) 4536. [54] P.B. Stathopulos, G.Y. Li, M.J. Plevin, J.B. Ames, M. Ikura, Stored Ca2+depletioninduced oligomerization of stromal interaction molecule 1 (STIM1) via the EF-SAM region: an initiation mechanism for capacitive Ca2+entry, J. Biol. Chem. 281 (2006) 35855–35862. [55] D.G. Allen, G.D. Lamb, H. Westerblad, Skeletal Muscle Fatigue: Cellular Mechanisms, Physiol. Rev. 88 (2008) 287–332. [56] S.M. Baylor, S. Hollingworth, Sarcoplasmic reticulum calcium release compared in slow-twitch and fast-twitch fibres of mouse muscle, J. Physiol. 551 (2003) 125–138. [57] A.J. Bakker, T.R. Cully, C.D. Wingate, C.J. Barclay, B.S. Launikonis, Doublet stimulation increases Ca 2+ binding to troponin C to ensure rapid force development in skeletal muscle, J. Gen. Physiol. 149 (2017) 323–334. [58] B.S. Launikonis, E. Ríos, Store-operated Ca2+entry during intracellular Ca2+release in mammalian skeletal muscle, J. Physiol. 583 (2007) 81–97. [59] P.H. Zhu, I. Parker, R. Miledi, Minimal latency of calcium release in frog twitch muscle fibres, Proc. R. Soc. London - Biol. Sci. 229 (1986) 39–46. [60] J. Vergara, M. Delay, A transmission delay and the effect of temperature at the triadic junction of skeletal muscle, Proc. R. Soc. London - Biol. Sci. 229 (1986) 97–110. [61] E. Ríos, G. Pizarro, Voltage sensor of excitation-contraction coupling in skeletal muscle, Physiol. Rev. 71 (1991) 849–908. [62] R.M. Luik, B. Wang, M. Prakriya, M.M. Wu, R.S. Lewis, Oligomerization of STIM1 couples ER calcium depletion to CRAC channel activation, Nature 454 (2008) 538–542. [63] G. Smith, N. MacQuaide, Cytoplasmic versus intra-SR: the battle of the Ca2+diffusion coefficients in cardiac muscle, Biophys. J. 95 (2008) 1005–1006. [64] J.T. Lanner, D.K. Georgiou, A.D. Joshi, S.L. Ryanodine Receptors Hamilton, Structure, expression, molecular details, and function in calcium release, Cold Spring Harb. Perspect. Biol. 2 (2010) 1–21. [65] R. Zalk, et al., Structure of a mammalian ryanodine receptor, Nature 517 (2015) 44–49. [66] A. des Georges, et al., Structural basis for gating and activation of RyR1, Cell 167 (2016) 145–157.e17. [67] P.B. Stathopulos, L. Zheng, G.Y. Li, M.J. Plevin, M. Ikura, Structural and mechanistic insights into STIM1-mediated initiation of store-operated calcium entry, Cell 135 (2008) 110–122. [68] T.R. Shannon, T. Guo, D.M. Bers, Ca2+ scraps local depletions of free [Ca2+] in cardiac sarcoplasmic reticulum during contractions leave substantial Ca2+ reserve, Circ. Res. 93 (2003) 40–45. [69] C.H.T. Kong, D.R. Laver, M.B. Cannell, Extraction of sub-microscopic ca fluxes from blurred and Noisy fluorescent Indicator images with a detailed model fitting approach, PLoS Comput. Biol. 9 (2013). [70] T.R. Cully, et al., Junctional membrane Ca 2+ dynamics in human muscle fibers are altered by malignant hyperthermia causative RyR mutation, Proc. Natl. Acad. Sci. 115 (2018) 8215–8220.

1248