A chloride channel blocker prevents the suppression by inorganic phosphate of the cytosolic calcium signals that control muscle contraction.

Abstract

Fatiguing exercise causes hydrolysis of phosphocreatine, increasing the intracellular concentration of inorganic phosphate (P_i). P_i diffuses into the sarcoplasmic reticulum (SR) where it is believed to forms insoluble Ca²⁺ salts, thus contributing to the impairment of Ca²⁺ release. Information on the P_i entrance pathway is still lacking. In amphibian muscles endowed with isoform 3 of the RyR channel, Ca²⁺ spark frequency is correlated with Ca²⁺ load of the SR and can be used to monitor this variable. We studied the effects of P_i on Ca²⁺ sparks in permeabilized fibres of the frog. Relative event frequency (f/f_ref) rose with increasing [P_i] reaching 2.54±1.6 at 5 mM and then decreased monotonically, reaching 0.09±0.03 at [P_i] = 80 mM. Measurement of [Ca²⁺]_SR confirmed a decrease correlated with spark frequency at high [P_i]. A large [Ca²⁺]_SR surge was observed upon P_i removal. Anion channels are a putative path for P_i into the SR. We tested the effect of chloride channel blocker 9-anthracenecarboxylic acid (9AC) on P_i entrance. 400 μM 9AC applied to the cytoplasm produced a non-significant increase in spark frequency and reduced the P_i effects on this parameter. Fibre treatment with 2 mM 9AC in the presence of high cytoplasmic Mg²⁺ suppressed P_i effects on [Ca²⁺]_SR and spark frequency up to 55 mM P_i. These results suggest that chloride channels (or transporters) provide the main pathway of inorganic phosphate into the SR and confirm that it impairs Ca²⁺ release by accumulating and precipitating with Ca²⁺ inside the SR, thus contributing to myogenic fatigue.

Introduction

Fatigue is a complex phenomenon, involving multiple processes that lead to a reduction in skeletal muscle performance. It can be central, with effects on the nervous system -- central, nerves and terminals -- or peripheral, with changes in muscle excitability, excitation-contraction coupling, filament interactions and other intracellular processes (Fitts & Balog, 1996). The loss of force and power, which characterizes this physiological process in normal individuals, aggravates pathological states and impairs movement in the elderly (Sundberg et al., 2019). Measurements in isolated muscles after repeated and sustained tetanic stimulation show reductions in force and shortening velocity that result from a combination of factors: the accumulation of K⁺ ions in the transverse tubular (TT) system causes membrane depolarization and consequently inactivation of the voltage sensor of excitation-contraction coupling (Fitts & Balog, 1996; Ferreira Gregorio et al., 2017) — an effect partially counteracted by the high chloride permeability of the T-tubular membrane (Pedersen et al., 2004; Dutka et al., 2008). During fatiguing activity, phosphate anions accumulate in the cytoplasm due mainly to phosphocreatine degradation. In parallel, lactic acid is produced, which lowers the pH (Allen et al., 2008; Cheng et al., 2018). It is debated whether both changes may depress velocity and force of contraction (Nelson et al., 2014; Westerblad, 2016).

The accumulation of inorganic phosphate (P_i) seems highly relevant among these factors. [P_i] can reach 40 mM during fatigue (Dawson et al., 1978; Godt & Nosek, 1989; Cady et al., 1989). Sundberg et al. (2019) showed in humans a power loss positively correlated with intracellular [P_i] during fatiguing exercise, which increased with age. P_i can affect contraction by inhibiting myofilament interaction (Hibberd et al., 1985; Fryer et al., 1995; Coupland et al., 2001) and altering intracellular Ca²⁺ release.

The Ca²⁺ release process may be altered by high P_i in various ways, including: i) direct effects on the SR Ca²⁺ release channel, ii) inhibition of Ca²⁺ uptake by the sarcoplasmic reticulum ATPase (SERCA), and iii) reduction of the free Ca²⁺ concentration in the SR, [Ca²⁺]_SR, upon precipitation of its salts. Fruen et al. (1994) showed a potentiating effect of P_i on RyRs incorporated in lipid bilayers, with an increase in open probability (p_open) of ~ 90% at 10 mM P_i. Working with Xenopus laevis, Stienen et al. (1999) reported a modest inhibition of Ca²⁺ uptake by 30 mM P_i, which increased when the pH was reduced to 6.2. Duke & Steele (2001) suggested that in fatigued conditions (absence of creatine phosphate and consequent ADP and P_i accumulation) Ca²⁺ would leak through SERCA, impairing Ca²⁺ release (Tupling, 2004). Finally, entry of P_i into the SR, where it precipitates with Ca²⁺, appears as an effective mechanism (Fryer et al., 1995, 1997; Westerblad & Allen, 1996; Posterino & Fryer, 1998; Westerblad et al., 2002; Allen et al., 2008). Kabbara & Allen (2001) showed a decrease in [Ca²⁺]_SR in fatigued toad muscle fibres and Launikonis et al. (2005a) demonstrated by confocal imaging the reduction in [Ca²⁺]_SR in the presence of high cytoplasmic P_i. In agreement with these observations, Dutka et al. (2005) showed in skinned rat muscle fibres that, in the presence of 30 mM P_i in the cytoplasm, a Ca²⁺ release-stimulating solution containing caffeine and low Mg²⁺ decreased the Ca²⁺ available for release as well as the initial rate of contraction, without reducing total Ca²⁺ content of the SR — in line with Ca²⁺-Pi precipitation in the SR.

The path through which P_i enters the SR remains in question. Anion channels constitute a putative entry pathway for P_i into the lumen of this organelle (Ide et al., 1991; Kourie et al., 1996; Laver et al., 2001). Specifically, Laver et al., 2001, assign it to a Cl⁻ channel, while results of Posterino & Fryer (1998) are inconsistent with this proposal. Defining this pathway is the main goal of the present work.

In frog skeletal muscle, spark occurrence is positively correlated with [Ca²⁺]_SR (Launikonis et al., 2006). In the present study, carried out on permeabilized frog skeletal muscle fibres, we used spark frequency as a monitor of [Ca²⁺]_SR. From these measures and direct determinations of [Ca²⁺]_SR we derived SR entry and exit movements of P_i, and obtained evidence of its precipitation with calcium inside the organelle.

Methods

Ethical approval

Adult frogs of the species Rana catesbeiana were sacrificed by rapid concussion on the head followed by rapid decapitation and pithing according to the approved protocol (No. 071140-001350-10) by the Honorary Committee for Animal Experimentation of the Universidad de la República (CHEA). 59 animals weighing between 200 and 400 g of either sex were obtained from a local commercial breeder, kept in pools with circulating water at 20 °C with a 12 hour light/dark cycle and fed daily with commercially available floating pellets. These frogs were used for spark and Fluo-5N experiments. Experiments with adult frogs of the species Rana pipiens were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals, and were approved by the Institutional Animal Care and Use Committee of Rush University (protocol No. 17–035). They were killed by a rapid blow to the head followed by rapid decapitation and pithing. Frogs were supplied by Nasco (Wisconsin, USA) and kept in a pool with circulating filtered water at 20 °C with a 12 hour light/dark cycle and fed daily with live crickets. 18 animals of either sex weighing 50 to 80 g were used for the Mag-Indo-1 experiments.

Single fibre preparation.

Single fibres were manually dissected from the semitendinosus muscle in Ringer solution (110 mM NaCl, 5 mM KCl, 2 mM MgCl₂, 5 mM HEPES). The solution was changed to a “relaxing” solution containing 115 mM K-Glutamate, 1 mM EGTA, 10 mM HEPES, 1 mM MgCl₂ and 0 Ca²⁺ where the fibres were cut and transferred to a small-volume, glass-bottom chamber. The fibres were permeabilized in the same relaxing solution with 0.05% saponin added for 1.5 minutes. After saponin treatment the fibres were incubated for 30 minutes in a “recovering” solution, containing 110 mM K-Glutamate, 5 mM ATP, 1 mM EGTA, 10 mM glucose, 5 mM phosphocreatine, 10 mM HEPES, 8% dextran with 0.9 mM free Mg²⁺ and 100 nM free Ca²⁺ to restore intracellular Ca²⁺ levels (Zhou et al., 2004). After transferring the chamber to the stage of a confocal microscope (Leica SP5 or SP2), the solution was changed to a “reference” solution, same composition as the recovering solution but with 0.3 or 0.4 mM free Mg²⁺ and 100 μM Fluo-3 or 50 μM Rhod-2. This same reference solution was used to record sparks and in the experiments designed to measure [Ca²⁺]_SR. The free Ca²⁺ and Mg²⁺ in intracellular solutions was calculated using Maxchelator (http://maxchelator.stanford.edu) and adjusted to be in the range of 50–100 nM and 0.9 mM respectively. The permeabilization solution and all solutions used thereafter contained 8% dextran to reduce cell swelling (Godt & Maughan, 1977). pH was adjusted to 7 and osmolarity to 240 mOsm/l. To suppress contraction, BTS (N-benzyl-p-toluenesulfonamide) was used at 100–200 μM. A 50 mM 9AC stock solution was prepared in DMSO. From this stock, 9AC was diluted to the final concentration in the corresponding experimental solutions. DMSO concentration in the reference solution was always the same as in the corresponding drug containing solution and never larger than 1%. When used at 2 mM 9AC was diluted directly at the final concentration in the experimental solution at pH 9 or 10 before final pH adjustment.

P_i containing solutions of the indicated concentration were prepared using the dipotassium salt. K-Glutamate content was calculated to keep osmolarity within the reference value except for the 80 mM P_i solution in which a higher osmolarity (280 mOsm/l) had to be admitted. Free Mg²⁺ and Ca²⁺ concentration were recalculated for each P_i concentration used and finally osmolarity and pH adjusted to reference values.

The SO₄^2- solution contained 80 mM K₂SO₄, 5 mM ATP, 1 mM EGTA, 10 mM glucose, 5 mM phosphocreatine, 10 mM HEPES, 8% dextran with 0.9 mM free Mg²⁺ and 100 nM free Ca²⁺. The 50 mM EGTA solution used was made replacing K-Glutamate iso-osmotically. This solution was expected to have a free [Ca²⁺]<1 nM. All experiments were performed at room temperature.

Confocal image acquisition.

Ca²⁺ sparks were detected and measured automatically on xt images (line scans) of fluorescence of Rhod-2 excited at 543 nm with a HeNe laser, or Fluo-3 excited at 488 nm with an Argon laser. Fluorescent emission was measured between 500 and 600 nm for Fluo-3 and between 570 nm and 620 nm for Rhod-2. Images were obtained scanning parallel to the fibre axis with a 63x, 1.2 N.A. water-immersion objective (Leica), at 0.160 μm per pixel. Time elapsed during acquisition of one 512 × 512 frame was 1.280 s. The images were processed using ImageJ (http://rsbweb.nih.gov/ij/) and custom routines written in the IDL environment (Harris Geospatiale, France). From the obtained image, F/F_o was calculated and sparks were automatically identified by the Sparkmaster analysis routine (Picht et al., 2007). Briefly, after image smoothing and normalization, mean intensity and standard deviation (SD) is obtained. Potential sparks are identified when pixel intensity exceeds 2X the SD above the mean value. A new mean and SD is computed excluding these areas from the image. Finally sparks are identified when the areas above 2 SD are larger than the new SD multiplied by a threshold factor set by the user. We used the recommended value of 3.8 for this factor, which according to the authors, allows detection of 90 % of the sparks (Picht et al., 2007).

The following parameters were measured for each spark: amplitude (maximum of F/F_o), duration at half maximum (FDHM) and width at half maximum (FWHM). Frequency of events was determined for sets of 10 consecutive images acquired every 2.8 s. To evaluate [Ca²⁺]_SR, fibres were loaded for 2 hours at room temperature with the fluorescent dyes Fluo-5N AM or Mag-Indo-1 AM (Invitrogen) (Kabbara & Allen, 2001; Launikonis et al., 2005b). Partially dissected whole muscles or single fibres were incubated at room temperature for 2.5 hours in Ringer or relaxing solution containing 10 μM of the AM form of the indicator and 0.05% Pluronic (Molecular Probes). After incubation, the fibres were dissected and transferred to the recording chamber, where they were permeabilized. Images of fluorescence of Fluo-5N or Mag-Indo-1 were obtained in xy mode (512 × 512 or 1024 × 1024 pixels). Fluo-5N was excited at 488 nm and fluorescence measured in the 500 and 600 nm range. To monitor [Ca²⁺]_SR ratiometrically by the SEER method (Launikonis et al., 2005b), Mag-Indo-1 was excited at 351 and 364 nm using an Argon Ion laser and the emission recorded line-interleaved at two fluorescence emission ranges: 390–440 nm and 465–535 nm. [Ca²⁺]_SR was computed from the fluorescence ratio (R) following the SEER procedure (Launikonis et al., 2005b). Briefly, fluorescence emission recorded on the fibre image in each range was averaged and ratioed. [Ca²⁺]_SR was computed according to

$${\left[{\text{Ca}}^{2+}\right]}_{SR}=K_d\gamma \frac{R-R_{min}}{R_{max}-R}$$

with K_dγ = 802 μM, R_max = 5.08 and R_min = 0.41 (Launikonis et al., 2005b).

Sparks and Fluo-5N experiments were performed on the Confocal Leica SP-5 microscope of the Unidad de Microscopía Confocal y Epifluorescencia at the Facultad de Medicina. Mag-Indo-1 experiments were performed on a Leica SP-2 confocal microscope at Rush University.

Statistics

Values in text correspond to mean ± S.D. Vertical bars in graphics, depict mean ± S.D. The significance of unpaired differences was quantified by the two-tailed Student’s t-test or one way analysis of variance (Multiple comparisons versus control group by the Holm-Sidak method). Differences were considered significant if P of the null change was less than 0.05 in the t-test or the ANOVA.

Results

Elevated cytoplasmic [P_i] reduces spark frequency.

To explore the effects of high cytoplasmic P_i, permeabilized fibres were exposed to different P_i concentrations. [P_i] was increased substituting iso-osmotically P_i for K-Glutamate while maintaining the same free [Ca²⁺] and [Mg²⁺]. Images were acquired regularly in sets of 10 and spark frequency was calculated for every set. Spark frequency in consecutive sets is illustrated in Figure 1; the fibre, immersed initially in reference solution, was exposed to various [P_i], with intervals in reference solution. Two of the P_i concentrations applied in the example resulted in a reduction of spark frequency that reverted when P_i was washed out. All results in experiments of this type are summarized in Figure 2. In the range up to 10 mM, P_i exposure increased the probability of spark occurrence. The maximum effect was observed at 5 mM P_i, a concentration slightly above physiological resting [P_i] in fast fibres (Meyer et al., 1985; Kushmerick et al., 1992). Table I summarizes effects on spark “morphometric” parameters. Except for spark frequency, all parameters remained almost unchanged up to concentrations as high as 30 mM. Because the effects on the frequency of sparks were not statistically significant up to 30 mM P_i and to offset eventual effects of increased calcium transport into the SR (see below), we evaluated the effects of higher concentrations of the anion. At very high [P_i] (55 and 80 mM), amplitude and width decreased, while duration and FDHM increased. Long lasting events, observed at these high concentrations, might account for the increase in duration parameters. The dual effects of P_i suggest the operation of two mechanisms, for instance, a direct potentiating effect on the ryanodine receptor and inhibition resulting from precipitation of Ca²⁺ within the SR. The increased ionic strength must also be taken into account at the high concentrations tested.

Figure 1. High P_i reduces spark frequency.

A, representative xt images in a typical experiment. Reference images are shown before and after perfusing the fibre with solutions containing different P_i concentrations. Some sparks are identified by an arrow. B, summary of the same experiment, showing Ca²⁺ spark frequency at various [P_i]. Each bar represents average frequency in 10 consecutive images. The fibre was exposed to three P_i concentrations, intercalated with reference solution as indicated by the horizontal bars. Symbols represent averages in each solution. Error bars represent S.D.

Figure 2. Effect of P_i on Ca²⁺ spark frequency.

Spark frequency was measured at different [P_i] as described in Methods and normalized to the value in reference solution. Bars represent mean value, circles represent individual measurements. Note opposite effects of P_i at low or high concentrations. Other spark morphology parameters are listed in Table I. The effect is statistically significant for [Pi] 2, 5, 7.5, 30, 55 and 80 mM. Number of fibres in each concentration is indicated in table I. Error bars correspond to S.D.

Table I. Averages of spark parameters at different P_i concentrations.

The experiments were conducted as described in figure 1B, P_i was washed out with reference glutamate solution after each P_i application. The parameters were normalized to the reference value in glutamate solution. Statistical significance was decided by a one way ANOVA analysis. Significant differences are indicated by **, (P<0.05). N represents the number of fibres studied. Figures in brackets correspond to S.D.

[Pi] (mM)	N	Frequency	Amplitude	FWHM	FDHM
0	45	1.00	1.00	1.00	1.00
2	6	1.55 ** (0.68)	1.03 (0.24)	1.07 (0.12)	1.05 (0.10)
5	9	2.54 ** (1.59)	1.00 (0.21)	1.08 (0.09)	0.93 (0.18)
7.5	3	1.50** (0.17)	1.10 (0.47)	0.97 (0.05)	1.02 (0.07)
10	4	0.86 (0.08)	1.02 (0.10)	0.96 (0.06)	0.94** (0.04)
20	3	0.63 (0.14)	0.97 (0.03)	0.95 (0.05)	1.00 (0.02)
30	13	0.43 ** (0.29)	1.03 (0.25)	0.84 ** (0.11)	0.86 ** (0.11)
55	4	0.07 ** (0.06)	0.85 (0.32)	0.71** (0.16)	0.77 ** (0.02)
80	3	0.09 ** (0.03)	0.33 ** (0.21)	0.72 ** (0.10)	1.70 ** (0.28)

High cytoplasmic P_i reduces [Ca²⁺]_SR through Ca-Pi precipitation inside the SR.

To test directly whether Ca-Pi precipitation occurs in the SR and examine how this could affect [Ca²⁺]_SR, which might explain the spark results, we carried out experiments with Ca²⁺-sensitive dyes trapped within the SR. We used the ratiometric dye Mag-Indo-1 or the intensiometric indicator Fluo-5N. In a first set of experiments [Ca²⁺]_SR was measured using the ratiometric dye Mag-indo-1 in SEER mode (Launikonis et al., 2005b). Fibres were loaded as described in Methods and then permeabilized. Fluorescence was followed during P_i perfusion and washout. Figure 3 illustrates two experiments in which two P_i concentrations were tested. In both cases, the fibres were kept in relaxing solution until time 0 when the solution was changed to reference. A slow rise of [Ca²⁺]_SR is observed reflecting a reloading of the SR after saponin exposure in relaxing solution. In Fig 3A perfusion with 20 mM P_i produced a small increase on [Ca²⁺]_SR, but perfusion of 80 mM P_i caused [Ca²⁺]_SR to decrease (Fig. 3B). A large [Ca²⁺]_SR surge occurred after removing P_i in both cases. This transient increase in [Ca²⁺]_SR during washout is expected if CaHPO₄ precipitation and SR Ca²⁺ uptake lead to an augmented Ca²⁺ content of the SR that is followed by cation re-solubilisation as P_i leaves the compartment. When images were acquired at high rate, a [Ca²⁺]_SR overshoot was observed regularly (Table II). The insets in the figure show on an expanded time scale the time course of this overshoot. It occurs immediately after the washout of P_i, with a time constant of a few seconds. Most remarkably, in the experiment in Fig. 3A, where the exposure to 20 mM P_i slightly increased [Ca²⁺]_SR, a prominent overshoot took place. To rule out an increase in Ca²⁺ transport from the cytoplasm into the SR, in both experiments P_i was washed out with a solution containing 50 mM EGTA and 0 Ca²⁺. This result strongly suggests that the [Ca²⁺]_SR surge is due to Ca²⁺ freed as P_i diffuses out of the SR and the precipitate dissolves. The time course of the [Ca²⁺] surge was variable and in some cases much slower. Similar result were obtained when the fibres were exposed to an 80 mM SO₄^2- containing solution (Fig. 3C and D), indicating that precipitation occurs with other anions when the concentration pass the solubility product with Ca²⁺. In 12 fibres exposed to SO₄^2- the [Ca²⁺]_SR reached a 52±52% increase with reference to the initial value upon washout with glutamate solution. When SO₄^2- was removed with the 50 mM EGTA 0 Ca²⁺ solution the increase was 39±30% (n=9). In this set of fibres where the SO₄^2- effects were studied we evaluated the time to peak of the [Ca²⁺] overshoot. When P_i was removed with the glutamate solution the time to peak was 189±126 s (n=9), whereas with the 50 mM EGTA solution it was 30±25 s (n=7). This significant difference (P <0.05) indicates a faster P_i efflux from SR in the 0 Ca²⁺ high EGTA solution. We do not have an explanation for this observation but it may be due to non-specific factors such as the different ionic strengths of the solutions.

Figure 3. The effect of Pi on [Ca²⁺]_SR.

Mag-Indo-1 [Ca²⁺]_SR measurements in two permeabilized fibres exposed to different [P_i]. At time 0 the fibres were perfused with a 100 nM Ca²⁺ reference glutamate solution. In A the reference solution was changed to a [P_i] = 20 mM and thereafter washed with a 50 mM EGTA solution with 0 Ca²⁺. In B a similar experiment as in A but [P_i] was 80 mM. Insets A and B show [Ca²⁺]_SR time course after changing to the EGTA solution in an expanded time scale. C and D show two examples of [Ca²⁺]_SR recorded in permeabilized muscle fibres exposed sequentially to glutamate, 80 mM SO₄^2- and glutamate. Note the prominent Ca²⁺ surge upon returning to the glutamate reference solution. In all cases the measurements correspond to average signals recorded from the whole fibre.

Table II.

[Ca²⁺]_SR computed from Mag-Indo-1 ratio measurements at different [P_i] show the [Ca²⁺]_SR overshoot after P_i removal. Experiments were started in glutamate reference solution that was replaced to the corresponding P_i containing solution or directly in the P_i containing solution (indicated by absence of reference measurement). The ratio was measured after its stabilization in reference and P_i. The P_i solution was washed out by glutamate reference with 1 mM EGTA and 100 nM Ca²⁺ or by a 50 mM EGTA solution with 0 Ca²⁺ as indicated in the last column. Peak [Ca²⁺]_SR computed during washout is tabulated. Average [Ca²⁺]_SR in reference was 202 μM.

		[Ca2+]SR (μM)
Fibre	[Pi] (mM)	Reference	Pi	Washout	EGTA (mM)
011504c	5	151	197	227	1
011504b	5	-	375	372	1
113004d	20	156	134	216	50
113004b	20	173	189	323	50
011604a	50	233	123	398	1
111604b	50	-	500	790	50
110904a	50	-	150	165	50
110904c	50	-	195	209	50
113004a	80	265	178	396	50
113004c	80	235	132	309	50

Figure 4 illustrates the result obtained in an experiment in which Fluo-5N was used. The fluorescence images had a characteristic striated pattern indicative that the dye was in the SR. This was confirmed by caffeine treatment that reduced the pattern to a minimum, recovering after washout of the drug. Images, always from the same region, were obtained in reference solution and after the fibre was exposed to 50 mM P_i. The fluorescence intensity was rapidly reduced but recovered after washout of P_i with fresh reference solution. An increase in florescence in fibres loaded with Rhod-2 when perfusing with solutions containing 50 mM P_i was never observed, which suggests that Ca²⁺ is not released but precipitates with P_i in the SR. As indicated before, the band pattern was completely and reversibly suppressed by 10 mM caffeine, which confirms that the signal monitors [Ca²⁺]_SR.

Figure 4. P_i reduces free Ca²⁺ in the SR.

A. Fluorescence images in a Fluo-5N-loaded fibre. A permeabilized fibre was exposed to 50 mM P_i or 10 mM caffeine (horizontal bars). P_i and caffeine were removed by perfusion of reference glutamate solution. The typical pattern corresponding to SR is depicted in white on each image (see text). Arrows indicate the sequence of the solution changes. B. Average fluorescence intensity measured in sequentially obtained images is plotted as a function of time. Grey bars correspond to power of the fundamental Fourier component of the profiles represented on the images in A.

The time course of these changes is depicted in part B of the same figure. For a quantitative measure of intensity adequate to the band pattern, an intensity profile was derived by averaging over y the fluorescence F(x,y) of a 12 × 7 μm region, with x parallel to the fibre axis. x direction profiles of the same region of the fibre are plotted on each image in Figure 4A. A. The power of the fundamental Fourier component of these profiles is represented by bars in Figure 4B. The Fourier quantification yields sharper changes, confirming that the pattern is suppressed by P_i and recovers after washout. Note also that this measure not only recovers but overshoots its reference after P_i washout — in line with the overshoot observed in the Mag-Indo-1 experiments.

Figure 5 presents [Ca²⁺]_SR–related fluorescence at various [P_i], namely Fluo-5N fluorescence above the basal level set in high caffeine, normalized by the signal in reference solution. 30 mM P_i reduced significantly this measure to 29 ± 0.24% of the initial value, an effect magnified at higher [P_i].

Figure 5. SR fluorescence as a function of [P_i] in the cytosol.

[Ca²⁺]_SR-related fluorescence in Fluo-5N-loaded fibres, evaluated by subtraction of the fluorescence remaining in 10 mM caffeine, is plotted as a function of P_i. Number of fibres at each concentration tested is indicated in brackets, total N was 17 obtained from 9 animals. Bars correspond to mean values, error bars to ± S.D; circles represent individual measurements. Asterisks indicate significant differences (P<0.05).

The effects of P_i on spark frequency and [Ca²⁺]_SR are positively correlated

Exposure to high [P_i] resulted in marked reduction in frequency of Ca²⁺ sparks. In previous work we reported a positive correlation between [Ca²⁺]_SR and spark frequency (Zhou et al., 2004; Launikonis et al., 2006). The reduction in [Ca²⁺]_SR observed here might explain the observed effects of P_i on spark frequency. The hypothesis was tested monitoring simultaneously [Ca²⁺]_SR and spark occurrence. Fibres loaded with Fluo-5N were bathed in an internal solution containing Rhod-2. F(x,y) and sets of 10 xt images at the corresponding wavelengths were obtained alternately to record [Ca²⁺]_SR and sparks. In figure 6 Fluo-5N fluorescence intensity images F(x,y) and Rhod-2 xt images are shown in reference solution, after changing to 50 mM P_i and after washout with reference solution (Fig. 6). As already described, SR fluorescence fell upon exposure and recovered on washout. The fluorescence intensity profile measured in a small area of the fibre (rectangle in Fig. 6Aa) in all conditions plotted in B further illustrates this effect. The xt images show that spark frequency changed in parallel with the fluorescence changes reported by Fluo-5N, suggesting that the reduction in [Ca²⁺]_SR is the main factor that determines the drop in frequency. The average result of similar experiments is represented in figure 6C, where the normalized averages obtained from four fibres are plotted. The putative association is strengthened by the experiments that follow.

Figure 6. Ca²⁺ spark frequency and [Ca²⁺]_SR.

A, [Ca²⁺]_SR-related fluorescence measured with Fluo-5N (a), and sparks recorded with Rhod-2 in the same fibre (b) . Fluorescence images and sparks were recorded alternately, first in K-glutamate solution (reference), then in 50 mM P_i, after washout of P_i and after perfusion of 10 mM caffeine. B, Fluorescence profile plotted from the same region of the fibre (rectangle in first image) in the four conditions: a, reference; b, 50 mM P_i; c, washout and d, 10 mM caffeine. C, normalized Fluo-5N fluorescence after subtraction of fluorescence in 10 mM caffeine (symbols) and normalized spark frequency (bars) averages obtained from similar experiments in four fibres. Error bars correspond to ± S.D. Significant differences from reference (P<0.05) are indicated by the stars for sparks and the asterisk for fluorescence.

9-anthracenecarboxylic acid (9AC) reduced the effects of P_i on Ca²⁺ spark frequency and [Ca²⁺]_SR.

Chloride channel blockers were introduced in the cytosol to curtail the entrance of P_i into the SR. DIDS (4’4’-Diisothiocyanatostilbene-2’2’-di-sulfonic acid) and SITS (4-acetamido-4’-isothiocyanato-stilbene-2,2’-disulfonic acid)^’’, tried initially, induced an increase in spark frequency, attributable to their opening effect on ryanodine receptors (O’Neill et al., 2003). 9AC was used instead, because its direct effects on spark frequency were minor. Sarcolemmal chloride channels are blocked by 9AC in the range of 200–400 μM. As summarized in Fig. 7, increasing 9AC concentration up to 800 μM slightly augmented spark frequency but the differences were not statistically significant. At concentrations beyond 800 μM, 9AC reduced spark frequency and promoted the occurrence of long-lasting channel openings. At these high concentrations there was a cell-wide increase in fluorescence, suggestive of Ca²⁺ leak from the SR. The effects of 9AC at concentrations greater than 800 μM were not explored further.

Figure 7. Effect of chloride channel blocker 9AC on spark frequency.

Plot of mean relative spark frequency as a function of 9AC concentration. Bars correspond to mean values, symbols to individual measurements. Numbers in brackets correspond to the number, n, of fibres studied for the corresponding concentration. Reference value was obtained from n = 10. Error bars correspond to ±S.D. Differences are not statistically significant (One way ANOVA, P=0.442).

To reduce P_i entry into the SR, 9AC was applied initially in the 200–400 μM range. At 400 μM, 9AC partially suppressed P_i effects on spark frequency (hatched bars in Figure 8) except at the two highest concentrations tested where still P_i had significant effects. Remarkably, the stimulatory effects at [P_i] < 10 mM were also suppressed, although the difference is only significant for 7.5 mM P_i. The effects of 9AC at [P_i] > 20 mM, which may include the partial block of chloride channels, suggest that these are a putative pathway of P_i entry to the SR. The membrane side from which 9AC blocks plasmalemmal chloride channels (CLC-1) is not completely settled. For intracellular anion channels (CLIC) which are also blocked by 9AC (Valenzuela et al., 2000; Rao et al., 2018), the location of the effector site is also uncertain. Considering that 9AC diffuses slowly through membranes, if a putative blocking site were accessible only from the lumen of the SR, a cytosolic concentration of 400 μM might be insufficient to reach a full blocking concentration within the time span of an experiment. In addition, CLC or CLIC channels isoforms present in the SR may be less sensitive to 9AC blockade (Clark et al., 1998). Based on this rationale and the fact that high concentrations of 9AC have no effect on excitation-contraction coupling (Shirokova et al., 1995), we designed a separate protocol to attain high intra-SR concentration. To reduce stimulation of RYR1 by 9AC we exposed the permeabilized fibre to 5 mM Mg²⁺, a concentration that nearly suppresses RyR activity (Laver, 2018). Under these conditions, spark frequency was drastically reduced, from 19.21±1.26 /100μm/sec in the control to 0.44±0.16 /100μm/sec (n = 4). After 5 minutes in the high [Mg²⁺] solution, the fibre was incubated for 20 minutes in one that also contained 2 mM 9AC, followed by 5 minutes in 400 μM 9AC, still with high [Mg²⁺]. Finally the fibre was bathed with the 0.4 mM Mg²⁺ reference solution used for spark recording, with the addition of 400 μM 9AC. In this final condition, sparks showed normal morphology and frequency within 10% of control values. No long duration events were observed during incubation in high 9AC or thereafter. The relationship between Ca²⁺ spark frequency and [P_i] in fibres pre-exposed to high 9AC is compared in figure 9A with that in fibres not exposed. P_i at concentrations up to 55 mM had no effect, and P_i at 80 mM concentration caused a much smaller reduction of spark frequency than it did when applied in the reference condition. The same protocol applied to Fluo-5N-loaded fibres showed high SR fluorescence even in the presence of 80 mM P_i (Figure 9B). These changes in [Ca²⁺]_SR-monitoring fluorescence follow the same dependence on P_i concentration as Ca²⁺ spark frequency, which confirms the correlation between this variable and [Ca²⁺]_SR reported above. The linear regression computed from the data in figure 9 A and B, plotted in Fig. 9C, corresponds to a r = 0.990. It should be noted that this correlation was established from measurements made in different fibres. The two variables could not be determined simultaneously. When both were measured in the same fibre (Fig. 6), the time it took to change the solution did not allowed the detection of the Ca²⁺ overshoot or the variation in spark frequency putatively associated with it.

Figure 8. 9AC suppresses P_i potentiation in the low concentration range but reduces P_i effects at higher [P_i] concentrations.

Normalized mean spark frequency as a function of [P_i] (filled bars, same data as in figure 2) compared with data for the same [P_i] in the presence of 400 μM 9AC (hatched bars). Numbers in brackets correspond to n at each [P_i] for the fibres in 9AC; symbols correspond to individual measurements in 9AC. Asterisks indicate significant differences at the same [P_i] with or without 9AC, error bars correspond to ±S.D.

Figure 9. Simultaneous exposure to 9AC and high [P_i].

A, mean relative frequency of sparks vs. [P_i] in the absence of 9AC (filled bars, data from table I) and after incubation in 2 mM 9AC (hatched bars). B, Mean [Ca²⁺]_SR-related fluorescence vs. [P_i] (filled bars, same data from figure 5) and after incubation with 2 mM 9AC (hatched bars). Symbols in A and B correspond to individual measurements; all differences in A and B at the same [P_i] are significant, error bars correspond to ±S.D. Number of fibres in each condition is indicated in brackets. C, normalized fluorescence vs. spark frequency in the same fibres. Solid line: best fit y= b*x + a. b=0.898±0.065, a=0.067±0.048 (value ± S.E. of fit). Standard error of estimate = 0.0676. Dotted lines 95% confidence interval. r=0.990. Error bars correspond to ± S.D.

Discussion

The present experiments seek to define the pathway used by inorganic phosphate to enter the SR of skeletal muscle and the effects of P_i when it is inside the SR. To probe these processes we used the frequency and morphometric features of Ca²⁺ sparks as established readouts of release channel flux and readiness to open. The observations strengthen the existing evidence that the intracellular accumulation of inorganic phosphate reduces Ca²⁺ release from the SR. They also confirm that P_i potentiates Ca²⁺ release, most probably by a direct effect on the RyR, evidenced only at low concentrations of the anion. The main outcome of the present experiments is that a chloride channel blocker suppresses both effects of P_i. The experiments also demonstrate large changes in [Ca²⁺]_SR caused by SR entry or exit of P_i.

Cytosolic effects of P_i on spark frequency.

Inorganic phosphate adopts mainly the HPO₄⁼ form at pH 7. Precipitation of Ca²⁺ salts (mostly CaHPO₄ at this pH) does not occur at normal [Ca²⁺]_cyto, even at [P_i] as high as 50 mM. By contrast, within the SR — where [Ca²⁺]_SR is between 200 and 1000 μM — dissolved P_i at low concentration coexists with the precipitate. Fryer et al. (1995) calculated a K_sp ~ 6 mM² for CaHPO₄. If it is assumed that [P_i] in the cytosol is equal to that inside the SR, precipitation as CaHPO₄ could occur when [P_i]_cyto is in the range of 6 to 30 mM. Therefore, any increase in [P_i]_cyto above these values could result in decrease of free [Ca²⁺]_SR. Assuming that the precipitate operates as an infinite sink in the time spans of interest, [Ca²⁺]_SR should not increase even if the intra-SR precipitation enhances Ca²⁺ uptake into the SR. In our experiments [Ca²⁺]_SR was in the range of 200 to 600 μM, setting the threshold for CaHPO₄ precipitation between 10 and 30 mM. In the experiment in figure 3A [Ca²⁺]_SR does not decrease in the presence of 20 mM P_i but after washout of the anion a large surge is observed indicating P_i precipitation balanced by increased Ca²⁺ uptake. We found that cytosolic P_i concentrations of 30 mM or greater reduced spontaneous Ca²⁺ spark frequency below reference values (figure 2 and table I), which suggests that intra-SR Ca²⁺ precipitated with P_i and reduced [Ca²⁺]_SR.

By contrast, at P_i concentrations of 7.5 mM or less, spontaneous Ca²⁺ spark frequency increased, a facilitation of RyR channel opening that is probably a direct effect of P_i. A direct promotion of RyR opening would be consistent with earlier observations. Fruen et al. (1994) reported a large increase in p_open of bilayer-incorporated RyR1 in the presence of 3 – 30 mM P_i on the cytosolic side. These authors speculated that the higher p_open could explain the increased Ca²⁺ release that occurs early in fatigue, possibly counteracting the decreased Ca²⁺ sensitivity of the myofilaments. Similarly, Balog et al. (2000) concluded that p_open increased in the presence of 100 mM P_i, based on flux measurements with SR vesicles.

In our preparation, 20–30 mM P_i caused a steep inhibition of spark frequency, which is consistent with a large Ca²⁺-precipitating effect of P_i at these concentrations, overcoming any potentiation of the RyR. Alternatively, the effects of P_i at high concentration could be due to an increase in the ionic strength of the solutions used. However, studies carried out in RyR incorporated in vesicles found a higher activity of the channel when the ionic strength was increased in contrast to what was found in the present study (Meissner et al., 1997). On the other hand, Ogawa & Harafuji (1990) found no effect of ionic strength on ryanodine binding to RyR.

High P_i reduces free [Ca²⁺]_SR.

By monitoring [Ca²⁺]_SR with ratiometric and intensiometric indicators we confirmed that P_i enters the SR. A reduction of [Ca²⁺]_SR in the presence of P_i was first evidenced when [Ca²⁺]_SR was quantified using the ratiometric SEER technique (Launikonis et al., 2005b). In the presence of 50 mM P_i in the cytosol, [Ca²⁺]_SR was ~ 120 μM while the value measured with the same procedure in reference solution under the same conditions was ~ 500 μM (Launikonis et al., 2005b) and ~ 200 μM in the present study (table II). After returning to the reference or 0 Ca²⁺ solution, [Ca²⁺]_SR greatly overshot its basal value in most cases (Fig. 3 and table II). The increase can be explained if the precipitates re-dissolve and Ca²⁺ joins the luminal solution at a faster rate than it can leave the SR. To exclude an increased Ca²⁺ uptake during washout as a possible mechanism to explain the observed overshoot, P_i was removed with a 0 Ca²⁺ high EGTA solution. Under these conditions, as predicted by Fryer et al. (1995), the only possible Ca²⁺ source resides within the SR. We also obtained similar results after replacing glutamate in the cytoplasmic solution by SO₄^2-, another divalent cation with similar K_sp to P_i (Launikonis et al., 2005a). A large overshoot was also observed when the SO₄^2- solution was washed out with reference solution or a 50 mM EGTA (0 Ca²⁺) solution (see Fig. 3 C and D). In work consistent with our present interpretation, Dutka et al. (2005) showed in skinned muscle fibres exposed to 30 mM P_i that Ca²⁺ release was significantly reduced. By lysing the SR they were able to demonstrate that total SR Ca²⁺ content remained unchanged in this condition. Taken together, all these findings are consistent with the precipitation of Ca²⁺ with P_i inside the SR.

The correlation between spark frequency and [Ca²⁺]_SR was previously observed by Launikonis et al. (2006). These authors found that spark frequency followed [Ca²⁺]_SR variations, although they concluded that this parameter is not a major determinant of spark frequency. Here we showed a similar relationship and used it as a monitor of [Ca²⁺]_SR. In the present experiments, however, spark frequency can only be taken as an indication of the steady [Ca²⁺]_SR; because we could not record both parameters simultaneously, the indicator does not apply to dynamic situations occurring during solution changes, including those that lead to an overshoot in [Ca²⁺]_SR.

The nature of the pathway of P_i entry into the SR.

It is generally accepted — but not demonstrated — that P_i enters the SR via anionic channels. In experiments carried out in bilayers it has been shown that P_i and SO₄^2- permeate Cl⁻ selective channels (Kourie et al., 1996). These channels are blocked by DIDS and PFA (phosphonoformic acid). In experiments carried out with concentrations between 50 and 200 μM of DIDS in the intracellular solution, we found a strong activation of the Ca²⁺ release channel in the permeabilized fibres. Other blockers such as SITS had similar effects.

Among the blockers studied, 9AC caused no potentiation of the RyR. The small average increase in spark frequency observed in figure 7 at 200 and 400 μM is not statistically significant. At this concentration in the cytosolic medium 9AC was unable to suppress the high [P_i] effects (Fig. 8), probably because the 9AC effector site is located within the SR (Pusch et al., 2002) or the SR membrane potential is too low to drive the blocker to a putative binding site within the channel (Ta et al., 2016). To reach this site we designed a protocol to increase the concentration of the drug in the SR lumen. We found that high cytosolic [Mg²⁺] offsets the stimulating effects of 9AC, so that in the presence of 5 mM Mg²⁺, 9AC could be applied at 2 mM without causing RyR stimulation. After 15 to 20 minutes in this solution, a luminal [9AC] in the range of 400 μM was likely reached because the effects of P_i up to 55 mM were completely suppressed. Neither the spark frequency nor the [Ca²⁺]_SR-monitoring Fluo-5N fluorescence changed significantly. In these 9AC pre-treated myofibers, reduction in both parameters by 50% required exposure to 80 mM P_i.

These results apparently contradict the study that led Posterino & Fryer (1998) to conclude that P_i does not share an entry pathway with Cl⁻. These authors reported no change in the recovery of force inhibited by high [P_i] when (rat) fibres were exposed to 9AC or DIDS. The discrepancy may be explained by the low 9AC concentration used in those experiments (100 μM). Our results are consistent instead with the proposal by Laver et al. (2001) that the P_i pathway into the SR lumen is a Cl⁻ channel.

The pathway remains to be defined further. Any consideration of the possibilities should take into account that 9AC and SITS block Cl⁻ channels as well as many transporters. Moreover, the molecular nature and unitary properties of some of the so-called Cl⁻ channels discussed here has not been established— they may be transporters rather than channels. Elucidation of the actual transport mechanism –whether diffusive or conformational– must await direct electrophysiological single channel or liposome flux measurements of purified protein.

9AC was used in these experiments at high doses; although no effects on Ca²⁺ release, charge movement or tubular Ca²⁺ current have been reported at 1 mM concentration on muscle fibres, we cannot rule out other possible effects of this compound. Bearing this in mind, we can speculate on possible routes for P_i entry into the SR. Candidates that have not been identified as molecular species include the high conductance chloride channel (BCl), which opens with a high probability and is not inhibited by physiological ATP (Kourie, 1997), and the low-conductance SCl. SCl is the better candidate because it is activated by changes occurring in muscle fatigue (including increased cytosolic Ca²⁺, oxidants and reduced ATP concentration; Posterino & Fryer, 1998). Among identified proteins, one to consider is the CLC-4 chloride channel/transporter of skeletal muscle SR (Slegtenhorst et al., 1994; Okkenhaug et al., 2006); however, detailed information on the role and localization of this channel is lacking (Jentsch & Pusch, 2018). Also in this group are the Intracellular Chloride Channels (CLIC) (Littler et al., 2010), which are blocked by 9AC (Ponnalagu & Singh, 2017) CLIC-1, CLIC-5A and 5B, are present in skeletal muscle fibres. Whether these proteins form channels in the SR is not clear, but CLIC-2, which also expresses in skeletal muscle (Board et al. 2004), is known to interact with RyR1 ((Dulhunty et al., 2005; Abdellatif et al., 2007; Meng et al., 2009). The CLIC proteins show sequence homology with the glutathione S-transferases (GST) and are able to insert into intracellular organelle membranes to form functional ionic channels as oligomers (Peter et al., 2014). Immunocytochemistry in progress in our laboratories reveals the presence of different isoforms of CLIC in mouse skeletal muscle, with a pattern compatible with SR localization.

9AC and P_i operated as opposing effectors on the spark readout. In the presence of the blocker, the inorganic anion was unable to potentiate the ryanodine receptor, which suggests that both bind to the same site or to closely interacting ones. Meissner et al. (1997) reported that the skeletal muscle ryanodine receptor is modulated by chloride anions, which “widen” the pCa dependence of RyR p_open (i.e., separate the [Ca²⁺] ranges that induce activation and inhibition of channel opening). More recently, Kimlicka et al. (2013) showed that the chloride binding site is located in the cytoplasmic N terminal region of RyR2. Assuming that RyR1 has a similar site, the cytosolic location was instrumental to the separation of the 9AC effects on RyRs and Cl^- channels in the present study. An alternative hypothesis to explain the lack of potentiation of the RyR in the presence of 9AC is to assume that this effect is mediated by the increased concentration of P_i within the SR acting on the luminal side of the receptor. Under this assumption, the blocking effect of the drug would prevent the entry of P_i into the SR and the concomitant stimulation of RyR. This mechanism is difficult to reconcile with the report by Balog et al. (2000), who showed that in single release channels incorporated into lipid bilayers, P_i exerts its potentiating effect only from the cis side (cytoplasmic) with no effect when applied from the luminal side even at very high concentrations.

Completing the identification of the conduits through which P_i moves would open a path for the design of strategies that oppose the processes involved in muscle fatigue. The interventions, for example, blocking drugs that prevent or slow entry of P_i into the SR, should be especially valuable in conditions further compromised by muscle fatigue, including disease and aging.

Key points.

• Accumulation of inorganic phosphate, P_i, may contribute to muscle fatigue by precipitating calcium salts inside the sarcoplasmic reticulum, SR. Direct demonstration of this process nor definition of the entry pathway of P_i into SR are fully established.

• We show that P_i promoted Ca²⁺ release at concentrations below 10 mM and decreased it at higher concentrations. This decrease correlated well with that of [Ca²⁺]_SR.

• Pre-treatment of permeabilized myofibers with 2 mM Cl^- channel blocker 9-anthracenecarboxylic acid (9AC) inhibited both effects of P_i.

• The biphasic dependence of Ca²⁺ release on [P_i] is explained by a direct effect of P_i acting on the SR Ca²⁺ release channel, combined with the intra-SR precipitation of Ca²⁺ salts. 9AC effects demonstrates that P_i enters the SR via a Cl^- pathway of as yet undefined molecular nature.