All procedures involving mice complied with the Guide for the Care and Use of Laboratory Animals (NIH publication No. 85–23, revised 1996). Experimental protocols were approved by the Institutional Animal Care and Use Committee of the University of Maryland School of Medicine.

Dysferlin-null (A/J) and control (A/JCr, C57Bl/6) mice were obtained from the Jackson Laboratory (A/J, C57Bl/6) or the National Cancer Institute, Frederick, MD (A/JCr) or bred at the University of Maryland, Baltimore (C57Bl/6). Mice were anesthetized with 2.5–4.5% isoflurane vaporized in oxygen and euthanized by cervical dislocation. Mice were 12–16 weeks of age at the time their tissues were studied.

mVenus-dysferlin (N-terminal Venus) (Addgene plasmid 29,768) (Covian-Nares et al., 2010) was provided by The Jain Foundation (www.jain-foundation.org). The constructs carrying Venus-dysferlin and Venus-dysferlin missing the C2A domain (residues 1–107) have been reported (Kerr et al., 2013; Muriel et al., 2022). For insertion of the GCaMP6f_u sequence (AddGene), we deleted the Venus moiety from Venus-dysferlin missing the C2A domain with NheI and KpnI and replaced it with GCaMP6f_u using the same restriction sites. This placed the GCaMP6f_u sequence in frame with the rest of the dysferlin ORF.

In vivo gene transfer via electroporation into FDB fibers was adapted from published methods (DiFranco et al., 2009), as described (Kerr et al., 2013; Lukyanenko et al., 2017; Muriel et al., 2022). Venus-dysferlin (V-Dysf), GCaMP6f_u and their variants were visualized in cultured myofibers (see below) with a Zeiss Duo Laser Scanning Confocal System (Carl Zeiss, Thornwood, NY), equipped with a C-Apochromat 40×/1.20 W Korr objective. Fluorescence excitation and emission were at 488 and >505 nm, respectively, with the laser intensity attenuated to 1%.

Mice were anesthetized and FDB muscles from both feet were harvested. A 2 week period was allowed after electroporation. Single myofibers were prepared in DMEM with 2% (wt/vol) BSA, 1 μl/ml gentamicin, and 2 mg/ml type II collagenase (Gibco, ThermoFisher, Waltham, MA) for 2 h at 37°C. Myofibers were kept for 12–14 h at 37°C and plated on 96-well plates coated with laminin (Sigma-Aldrich, St. Louis, MO) 1 h before experimentation. Fibers were then washed in normal Tyrode’s solution, pH 7.4, containing 140 mM NaCl, 5 mM KCl, 0.5 mM MgCl₂, 0.3 mM NaH₂PO₄, 5 mM HEPES, 5.5 mM glucose, 1.8 mM CaCl_2.

Isolated FDB fibers were loaded for 45 min at 37°C with 4.4 µM Rhod-2AM in the presence of 0.25% Pluronic F-127 (both from ThermoFisher), diluted in culture medium, then washed with Tyrode’s solution. When BAPTA-AM, Fluo-4AM or EGTA-AM was used, it was included in the same solution at concentrations from 0 to 250 nM.

Rhod-2 was excited with the 560 nm laser line with the intensity attenuated to 0.5% and emission was monitored at >575 nm with a LP 575 filter. Fluo-4 or GCaMP6f_u were excited by light at 488 nm (25 mW argon laser, intensity attenuated to 1%) and fluorescence was measured at wavelengths of >515 nm. Trains of voltage pulses transients were induced by field stimulation (1 Hz for 10 s) every 1 min for 5 min, as reported (Lukyanenko et al., 2017; Muriel et al., 2022). Perfusions and imaging were done in the dark.

Line-scan images were taken in the middle of myofibers at 1.9 ms per line at maximal aperture. ImageJ 1.31v (NIH, Bethesda, United States) averaged the profiles for every pixel over time and took the maximal value for each of 10 voltage pulses to calculate the mean maximal value of the Ca²⁺ transients, which were typically 225 pixels in width. The difference between maximal fluorescence intensity (F _max) and background fluorescence (F _o ), normalized to F _o . is reported.

As the amplitudes of the Ca²⁺ transients in electroporated myofibers are higher than those in controls (Lukyanenko et al., 2017), we analyzed results from electroporated and non-electroporated samples separately, with differences analyzed with the paired Student t-test for the former, to compare regions that expressed the transgene with regions that did not, and a simple t test for the latter.

Osmotic shock injury (OSI) was induced as described (Lukyanenko et al., 2017). In brief, cultured FDB fibers were superfused with normal Tyrode’s solution and then for 60 s with a hypotonic Tyrode’s solution containing 70 mM NaCl. Cells were then perfused with isotonic Tyrode’s solution for 5 min. Experiments were performed at room temperature (21–23°C). Data were collected from myofibers cultured from 2 or more mice.

Calibration was done in the glass-bottom chambers used for culturing myofibers. Solutions containing known concentrations of Fluo-4 were added in the solution (mM): 10 HEPES, 264.2 KCl, 5 EGTA, 2.7 CaCl₂ (to yield [Ca²⁺]_free = 100 nM), 1.1 MgCl₂ (to yield [Mg²⁺]_free = 1 mM), pH 7.4. The concentrations were calculated with WEBMAXC STANDARD. Line-scan images through the solution were taken at 1.9 ms per line at maximal aperture. Averaged data from 5 experiments were used to build the calibration curve. We used identical conditions to scan fibers preloaded with 10 nM Fluo-4AM.

Immunoblotting for RyR used antibodies specific for RyR1 (ThermoFisher/Invitrogen), RyR2 (ProteinTech) and RyR3 (Millipore, ThermoFisher/Invitrogen), diluted 1:500. Gels were 3–8% Tris-Acetate with Tris-acetate-SDS running buffer (NuPAGE/Invitrogen). Secondary antibodies were HRP conjugates (anti-rabbit from Invitrogen; anti-mouse from Jackson ImmunoResearch), used at 1:10,000. Blots were visualized with SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher) and imaged with a BioRad ChemiDoc MP instrument.

Quantitative data are shown as mean ± SE. Statistical significance was determined with Student’s t test and Χ ² analysis. A value of p < 0.05 was considered statistically significant.

BAPTA-AM and EGTA-AM were from MilliporeSigma. Fluo-4AM was from Invitrogen. Unless specified otherwise, all other chemicals were from Sigma-Aldrich.

We first examined the effects of incubating dysferlin-null A/J myofibers with increasing concentrations of BAPTA-AM in the presence of 4.4 µM Rhod-2AM. We elicited Ca²⁺ transients with field stimulation and recorded the increase in Rhod-2 fluorescence, which tracks [Ca²⁺]_i (Lukyanenko et al., 2017; Muriel et al., 2022). Figures 1, 2 show that incubation with ≥100 nM BAPTA-AM reduces the apparent amplitude of the Ca²⁺ transient, perhaps because the BAPTA generated by cleavage of the -AM moieties in the sarcoplasm accumulates to levels high enough to compete with the Rhod-2 and reduce its fluorescence during the transient. By contrast, concentrations of BAPTA-AM of ≤50 nM increase the amplitude of the Ca²⁺ transient by ∼15% (Figures 1C, 2A), to levels seen in dysferlin-positive, control A/JCr myofibers. This difference is significant (p < 0.05). DMSO, the vehicle, had no effect in the absence of the chelator (2.40 ± 0.10, n = 138 vs. 2.57 ± 0.07, n = 280, A/J fibers with DMSO present vs. DMSO absent, respectively, p = 0.09). These results show that low concentrations of BAPTA-AM enhance the amplitude of Ca²⁺ transients in A/J muscle fibers, restoring them to control levels.

We next examined the effects of BAPTA-AM on the Ca²⁺ transients of dysferlin-null A/J myofibers after a brief osmotic shock injury (OSI). As above, we recorded the transients via Rhod-2 fluorescence in response to electrical stimulation. As we reported earlier (Kerr et al., 2013; Lukyanenko et al., 2017; Muriel et al., 2022), OSI significantly decreases the amplitude of Ca²⁺ transients of A/J fibers 5 min after injury, and the transients that appear are frequently accompanied by Ca²⁺ waves and spontaneous transients (Figures 1, 2B–D). In fibers incubated in 10 nM BAPTA-AM, however, the amplitudes of the transients were not reduced after osmotic shock injury (Figure 2B), and waves and spontaneous transients were reduced to control levels (Figures 1, 2C,D). These changes did not occur with DMSO alone (Figures 3B–D, A/J alone), suggesting that BAPTA mediates these effects.

We studied the concentration dependence of the inhibition of abnormal Ca²⁺ signaling as a function of BAPTA-AM concentration. We found that 5 nM BAPTA-AM was not sufficient to protect the transient against loss of amplitude after OSI, but that concentrations of 10 and 50 nM were effective (Figures 2A,B). These results suggest that the presence of the chelator was sufficient to maintain normal Ca²⁺ signaling following OSI, mimicking the activity of dysferlin in this assay.

We next measured the effects of the -AM derivative of Fluo-4, a variant of BAPTA that fluoresces upon binding of Ca²⁺. Although Fluo-4AM partitions into the sarcoplasm and is cleaved by intracellular esterases much like BAPTA-AM (Paredes et al., 2008; Smith et al., 2018), ten-fold higher concentrations were required to mimic the effect of BAPTA-AM on the recovery of the Ca²⁺ transient after OSI and to reduce the appearance of Ca²⁺ waves (Figures 3B–D). Even at 250 nM, Fluo-4AM failed to reduce Ca²⁺ waves to the levels seen with 10 nM BAPTA-AM (compare Figures 3D–2D). Thus, Fluo-4AM is less effective than BAPTA-AM, perhaps because its affinity for Ca²⁺ is ∼2-fold poorer (Paredes et al., 2008). Fluo-4 is also larger, carries a negative charge and has a dielectric constant that is 20-fold higher than BAPTA’s, and so may accumulate in the sarcoplasm less efficiently than BAPTA.

Despite the differences between Fluo-4AM and BAPTA-AM, we used the former to approximate the intracellular concentration of BAPTA needed to reduce the abnormalities in Ca²⁺ signaling that we routinely assay. We measured the fluorescence intensity of intracellular Fluo-4 in myofibers incubated with 10 nM Fluo-4AM under conditions identical to those we used with BAPTA-AM and compared it to a standard curve, with the assumption that the concentration of free intracellular Ca²⁺ is ∼100 nM (e.g., López et al., 1983; Harkins et al., 1993; Head 1993; Pressmar et al., 1994; Konishi and Watanabe 1995; Baylor and Hollingworth 2007). For the standard curve, we measured the fluorescence intensities at different concentrations of the K⁺ salt of Fluo-4 in the presence of 100 nM Ca²⁺ and under identical confocal imaging conditions (see Methods). The results indicate that Fluo-4 reached concentrations in the sarcoplasm of ∼60 nM, or about 6 times higher than its concentration in the bath (Supplemental Figure S1). Given the significant differences between the effects of BAPTA-AM and Fluo-4AM, BAPTA may accumulate to levels significantly higher than 60 nM, thereby buffering sarcoplasmic free [Ca²⁺] after OSI to ≤100 nM.

Although the affinities of BAPTA and Fluo-4 for Ca²⁺ (in Mg²⁺-free conditions) are ∼160 and ∼370 nM, respectively, they both have relatively high “on” and “off” rates for Ca²⁺ and thus are able to chelate Ca²⁺ rapidly. The affinity of EGTA for Ca²⁺ is close to that of BAPTA, but its “on” and “off” rates for binding Ca²⁺ are ∼100-fold slower than BAPTA’s. We used EGTA-AM to test the idea that rapid, transient changes in the concentration of Ca²⁺, rather than changes in the equilibrium concentration alone, play a role in the abnormal Ca²⁺ signaling that we observe in A/J myofibers before and after OSI. Our results (Figure 4) show that incubating myofibers with concentrations of EGTA-AM up to 25-fold higher than the effective concentration of BAPTA-AM fails to protect the transient against disruption by OSI, although low concentrations are able to enhance the amplitude of the transient before injury. This suggests that the dysregulation of Ca²⁺ that alters the stability, but not the initial amplitude, of the Ca²⁺ transient in A/J fibers is due to rapid, transient changes in Ca²⁺ rather than the overall levels of Ca²⁺ in the cytoplasm.

Loading myofibers with BAPTA, Fluo-4 or EGTA, as we have done, can reduce the resting and peak levels of free Ca²⁺ in the sarcoplasm, which may only indirectly alter levels at the triad junction. We have postulated that abnormal Ca²⁺ signaling that follows injury of A/J myofibers is due to changes in Ca²⁺ at the triad junction that destabilize the LTCC-RyR1 couplons there, reducing normal Ca²⁺ release and promoting CICR. Here we test if suppression of local increases in Ca²⁺ at the triad junction is indeed sufficient to protect the Ca²⁺ transient from injury.

For these studies, we placed GCaMP6f_u at the N-terminus of Venus-dysferlin lacking the C2A domain (GCaMP6f_u-DYSF-ΔC2A; Figure 5A). We chose this GCaMP variant because, like the C2A domain (Abdullah et al., 2014; Wang et al., 2021), GCaMP6f_u binds Ca²⁺ rapidly and with high affinity (Helassa et al., 2016). We found that, similar to the Venus construct of DYSF-ΔC2A, GCaMP6f_u-DYSF-ΔC2A traffics normally to membranes at the level of the A-I junction (Figure 5B), concentrating in transverse tubules of the triad junction like both Dysf-ΔC2A and WT dysferlin (Muriel et al., 2022). By contrast, GCaMP6f_u expressed as a Venus fusion protein (Figure 5A) distributes much more uniformly in the sarcoplasm and, like GFP itself, only accumulates at the level of Z-disks (Figure 5B). When A/J fibers expressing GCaMP6f_u in the sarcoplasm or at the triad junction are loaded with Rhod-2 and electrically stimulated, both the GCaMP6f_u moiety and Rhod-2 register the changes in Ca²⁺ concentration. The signals generated by Rhod-2 were brighter than those generated by GCaMP6f_u, but we were generally able to use either for measurements of transient amplitudes and waves.

Remarkably, fibers expressing GCaMP6f_u-DYSF-ΔC2A show an increase in the amplitude of the Ca²⁺ transient similar to that observed with BAPTA and that seen when WT dysferlin is restored to dysferlin-null fibers (Figures 6A,B). (Please note that the amplitudes of the transients in uninjured myofibers subjected to electroporation are higher than those studied without electroporation, as reported Lukyanenko et al., 2017). Moreover, the transient after OSI remains at control levels, similar to fibers expressing WT dysferlin (Figures 6A,C). In addition, fibers expressing GCaMP6f_u-DYSF-ΔC2A rarely show Ca²⁺ waves (Figure 6D). These results are consistent across a wide range in the level of expression of the fusion protein (Figure 6E), suggesting that it is acting in a limited volume, i.e., the triad junction.

By contrast, GCaMP6f_u expressed as a Venus fusion protein fails to increase the amplitude of the transient in uninjured fibers (Figure 7B), and is somewhat less protective of the Ca²⁺ transients after OSI (e.g., Figures 7A,C). More strikingly, however, it fails to suppress the appearance of Ca²⁺ waves (Figure 7D). The differences with GCaMP6f_u-DYSF-ΔC2A in amplitude and wave frequency after recovery from OSI were both statistically significant (p < 0.05). These results were far less consistent as a function of concentration than those of GCaMP6f_u-DYSF-ΔC2A (Figure 7E; data obtained with Rhod-2, only). These results suggest that GCaMP6f_u in the sarcoplasm acts as a weak Ca²⁺ chelator and thus is considerably less effective than GCaMP6f_u-DYSF-ΔC2A in the triad junction.

Our findings suggest that GCaMP6f_u localized by DYSF-ΔC2A to the triad junction, but not as a cytoplasmic protein, can serve the function of the C2A domain of dysferlin in Ca²⁺ signaling. Thus, one of dysferlin’s likely roles in maintaining the health of skeletal muscle is to bind Ca²⁺ at triad junctions and thereby protect the muscle from injury.

As the presence of non-junctional RyR isoforms could account for CICR in A/J muscle fibers, we used immunoblotting of muscle extracts to determine if RyR2 and RyR3 were expressed together with RyR1. The blots of extracts of Tibialis anterior muscles from A/J mice, which like FDB fibers are primarily fast twitch, showed RyR1 to be present at high levels but RyR2 and RyR3 to be undetectable (Figure 8). They were readily detectable in heart and brain, however (not shown). This is consistent with several reports in GEO Profiles which show much lower levels of RyR2 and RyR3 mRNAs than RyR1 mRNA in Tibialis anterior muscles (see also Futatsugi et al., 1995; Marziali et al., 1996). Thus, RyR2 and RyR3 are unlikely to contribute significantly to abnormal Ca²⁺ signaling in dysferlinopathic muscle.