Two zebrafish cacna1s loss-of-function variants provide models of mild and severe CACNA1S-related myopathy

Abstract

CACNA1S-related myopathy, due to pathogenic variants in the CACNA1S gene, is a recently described congenital muscle disease. Disease associated variants result in loss of gene expression and/or reduction of Cav1.1 protein stability. There is an incomplete understanding of the underlying disease pathomechanisms and no effective therapies are currently available. A barrier to the study of this myopathy is the lack of a suitable animal model that phenocopies key aspects of the disease. To address this barrier, we generated knockouts of the two zebrafish CACNA1S paralogs, cacna1sa and cacna1sb. Double knockout fish exhibit severe weakness and early death, and are characterized by the absence of Cav1.1 α1 subunit expression, abnormal triad structure, and impaired excitation-contraction coupling, thus mirroring the severe form of human CACNA1S-related myopathy. A double mutant (cacna1sa homozygous, cacna1sb heterozygote) exhibits normal development, but displays reduced body size, abnormal facial structure, and cores on muscle pathologic examination, thus phenocopying the mild form of human CACNA1S-related myopathy. In summary, we generated and characterized the first cacna1s zebrafish loss-of-function mutants, and show them to be faithful models of severe and mild forms of human CACNA1S-related myopathy suitable for future mechanistic studies and therapy development.

excitation contraction coupling, congenital myopathy, Cav1.1, zebrafish

Introduction

The L-type calcium channel in skeletal muscle (Cav1.1), also known as the dihydropyridine receptor (DHPR), plays an essential role in converting electrical information encoded in an action potential in the surface membrane to trigger Ca²⁺ release from the sarcoplasmic reticulum and thus drive muscle contraction during a process referred to as excitation-contraction coupling (ECC). During ECC, action potentials generated at the neuromuscular junction rapidly propagate down the sarcolemma and into invaginations of the surface membrane called transverse tubules (or T-tubules). Depolarization of the T-tubule membrane potential causes voltage-driven conformational changes in the Cav1.1 complex that are mechanically linked to the opening of type-1 ryanodine receptor (RyR1) Ca²⁺ release channels located in the adjacent sarcoplasmic reticulum (SR) membrane. This interaction activates RyR1 to release Ca²⁺ stored in the SR in a process referred to as depolarization-induced Ca²⁺ release (DICR) [1, 2]. STAC3 (SH3 and cysteine-rich domain 3), an adaptor protein that interacts with cytoplasmic domains of Cav1.1, is also a key component for ECC and DICR, and is essential for Cav1.1 voltage-sensing and the conformational coupling between Cav1.1 and RyR1 [3, 4]. Defects in the α-subunit of Cav1.1 (Cav1.1α1S, encoded by the CACNA1S gene) are associated with a heterogeneous class of muscle diseases, including periodic paralysis [5], malignant hyperthermia [6], and myotonic dystrophy [7].

Recently, a congenital myopathy linked to pathogenic variants in CACNA1S has been elucidated (OMIM #620246) [8, 9]. Clinically, individuals with CACNA1S-related myopathy exhibit neonatal onset hypotonia, delayed motor development, progressive muscle weakness with facial involvement, respiratory disturbances ranging from normal to severe, and mild dysphagia. Histological examination of muscle biopsies shows changes generically classified as “core myopathy,” including core-like/mini-core structures, morphological alterations in the inter-myofibrillar network, and centralized/internal nuclei. Molecular studies indicate that both dominant and recessive cases are caused by a reduction in Cav1.1 protein due to missense or frameshift variants that reduce protein translation or stability, ultimately leading to impaired DICR [8]. However, the exact pathomechanisms of CACNA1S-related myopathy remain unclear, which has made it difficult to fully comprehend the disease etiology and develop effective treatments.

To elucidate disease pathophysiology and develop therapies, it is essential to establish model organisms that display similar phenotype(s) as patients and that can be studied efficiently and extensively both in vivo and in vitro. To date, five mouse models associated with Cacna1s mutations have been reported: dysgenic (mdg) [10], Cav1.1-R528H [11], Cav1.1-N617D [12], Cav1.1-Δe29 [13], and Cav1.1-E1014K [14, 15] mice. The mdg mice carry a spontaneous null-mutation of Cacna1s, which results in a lack of Cav1.1 expression, leading to completely paralyzed muscles and respiratory failure, causing the mice to die at birth [10, 16, 17]. While useful for studying fetal myogenesis in severe CACNA1S-related myopathy, mdg mice are not ideal for assessing the overall picture of the disease during development and adulthood. The Cav1.1-R528H mice were developed by introducing the R528H mutation, the most common cause of familial hypokalemic periodic paralysis (HypoPP) in humans [11]. These mice exhibit physiologic abnormalities consistent with human HypoPP, but show no difference in Cav1.1 expression [11], indicating that the pathogenesis associated with Cav1.1-R528H is different from that of CACNA1S-related myopathy. Cav1.1-N617D and Cav1.1-E1014K mice have mutations that abolish inward Cav1.1 Ca²⁺ currents. These mice showed no differences in muscle force, endurance, or fatigue [12, 15], suggesting that they do not exhibit a congenital myopathy. Cav1.1-Δe29 mice were engineered to constitutively lack a developmentally regulated exon (exon 29) that suppresses Cav1.1 Ca²⁺ conduction in adult muscle. Cav1.1-Δe29 mice exhibit increased calcium influx during ECC, altered calcium homeostasis, a fiber type shift toward slow fibers, and reduced mitochondrial content and oxidative enzyme activity, but do not have an overt myopathy [13].

Given the absence of models that phenocopy CACNA1S-related myopathy, development of new animal models of this disorder is thus needed. To that end, we generated and characterized two zebrafish lines (cacna1s-dKO and cacna1s-aKO-bHet) that carry null mutations in the paralogs of the CACNA1S gene. Using a variety of methods, we demonstrate that cacna1s-dKO fish lack Cav1.1α1S expression and DICR, leading to complete paralysis and death by 14 days post-fertilization (dpf). These fish have normal early muscle development and sarcomeric structure, allowing for studies of triad function in muscle fibers in the absence of secondary damage. In contrast, cacna1s-aKO-bHet fish grow to adulthood, but exhibit mild and non-progressive muscle weakness. We observed core-like structures in muscle fibers of adult cacna1s-aKO-bHet fish, which have not previously been reported in zebrafish muscle sections. Interestingly, reduced muscle activity in cacna1s-aKO-bHet affected craniofacial development, resulting in facial deformities reminiscent of human facial myopathic changes. Our findings indicate that cacna1s-dKO fish represent a model of severe CACNA1S-related myopathy, while cacna1s-aKO-bHet fish model a milder form of core myopathy.

Results

Generation of cacna1sa and cacna1sb mutant zebrafish

To study the impact of loss-of-function mutations in zebrafish cacna1s, we used the CRISPR/Cas9 system to target the two paralogs of human CACNA1S, cacna1sa and cacna1sb, generating multiple genotypes in combination with mutations in cacna1sa and cacna1sb. We designed guide RNAs at exon 14 in cacna1sa and at exon 16 in cacna1sb, and the two single guide RNAs (sgRNAs) were co-injected with Cas9 mRNA into one-cell-stage embryos (Fig. 1A). High-resolution melting (HRM) analysis and Sanger sequencing revealed F1 carriers with two mutations: a 5 bp deletion in cacna1sa (Δ5: c.2024_2028del, p.Ala675Glyfs^*20), which introduces a premature stop codon, and an insertion-deletion mutation in cacna1sb (Indel: c.2271_2287 + 8del, c.2271_2287 + 8ins, p.Pro758Leufs), which leads to multiple abnormal splicing variants and introduces premature stop codons (Fig. 1B, Supplementary Figs S1–S3). We next crossed cacna1sa^Δ5/+ and cacna1sb^Indel/+ to create the trans-heterozygous mutant line cacna1sa^Δ5/+; cacna1sb^Indel/+ (cacna1s-dHet) (Fig. 1A). We then inter-crossed cacna1s-dHets and obtained eight mutant lines according to combinations of mutations harbored in cacna1sa and cacna1sb alleles. In this study, we present the data focusing on the following mutant lines: cacna1sa^Δ5/Δ5 (cacna1s-aKO), cacna1sb^Indel/Indel (cacna1s-bKO), cacna1sa^Δ5/Δ5; cacna1sb^Indel/+ (cacna1s-aKO-bHet) and cacna1sa ^Δ5/Δ5 cacna1sb ^Indel/Indel (cacna1s-dKO) (Fig. 1C).

Figure 1

Generation of cacna1s mutant fish. (A) Schematic workflow of establishing cacna1s trans heterozygous mutants. sgRNA-1 was used for cacna1sa gene editing and sgRNA-2 for cacna1sb. The sequences of the gRNAs are described in the methods and Supplementary Fig. S1. (B) Schematics showing domains of Cacna1sa and Cacna1sb proteins and the obtained mutations. Yellow: transmembrane domain, Green: EF-hand domain, Blue (Ca chan IQ): voltage gated calcium channel IQ domain. (C) Progeny achieved by crossing cacna1s-dHet mutants. (D) mRNA levels of cacna1sa and cacna1sb in 6 dpf fish pools (WT; n = 5, aKO; n = 4, bKO; n = 4, dKO; n = 5). Values are Mean ± SEM (fold); cacna1sa: WT (1.00 ± 0.09), aKO (0.20 ± 0.01), bKO (0.59 ± 0.07), dKO (0.20 ± 0.04); cacna1sb: WT (1.00 ± 0.14), aKO (1.08 ± 0.10), bKO (0.13 ± 0.01), dKO (0.15 ± 0.03). Statistical analysis was performed using one-way ANOVA followed by Dunnett’s multiple comparisons test with P < 0.01 (^**), P < 0.0001(^****) or non-significance (ns).

To examine the level of cacna1s transcripts in the mutants, we performed quantitative reverse transcription PCR (RT-qPCR) on total RNA extracted from 6 dpf fish. A significant and large magnitude decrease in cacna1sa mRNA levels was observed in cacna1s-aKO and cacna1s-dKO fish, with a mild but significant reduction seen in cacna1s-bKO. Marked reductions in cacna1sb mRNA were observed in cacna1s-bKO and cacna1s-dKO Fish compared to their WT siblings (Fig. 1D). These results show that the transcripts are likely subject to nonsense-mediated mRNA decay (NMD) by both Δ5 and Indel mutations, which is consistent with the creation of null alleles for each gene.

Cacna1s-dKO zebrafish are paralyzed and lack skeletal muscle contractile function

To characterize phenotypes of the generated mutant zebrafish, we assessed their overall development, morphology, and behavior. No overt phenotypic differences were observed between WT, cacna1s-aHet, cacna1s-bHet, cacna1s-dHet, cacna1s-aKO, cacna1s-bKO, cacna1s-aKO-bHet, and cacna1s-aHet-bKO up to 7 days. No clear difference in touch-evoked escape response was observed in any genotype other than cacna1s-dKOs, though some cacna1s-bKO and cacna1s-aHet-bKO embryos exhibited circular movement just post-hatching (2 dpf), which completely resolved by the day following. By contrast, cacna1s-dKO fish exhibited no spontaneous tail coiling with subsequent bending in body axis at post-hatching (Fig. 2A), indicative of an early onset deficiency in motor function.

Figure 2

Characterization of cacna1s mutant larvae. (A) Representative bright field images of 4 dpf embryos. A total of 10 embryos per group were imaged. cacna1s-dKO larvae exhibited curved body axis. Scale bars = 1 mm. (B) Representative snapshots of touch-evoked response of 4 dpf larvae. A total of five fish per group were captured for video. WT zebrafish swim away rapidly after touch stimulation, while cacna1s-dKO showed no response or movement (see also Supplement Movies). (C) Quantification of motor ability of mutant fish (6 dpf). Total distance traveled during photochemical activation was normalized to the average of WT siblings and plotted (WT; n = 48, dHet; n = 136, aKO; n = 42, aKO-bHet; n = 69, bKO; n = 43, dKO; n = 28). Values are Mean ± SEM (mm); WT (1.00 ± 0.04), dHet (0.92 ± 0.02), aKO (1.02 ± 0.04), aKO-bHet (0.89 ± 0.03), bKO (0.74 ± 0.04), dKO (0.00 ± 0.00). Statistical analysis by one-way ANOVA followed by Dunnett’s multiple comparisons tests where P < 0.05 (^*) and P < 0.0001 (^****). (D) Kaplan-Meier curve showing reduced survival in cacna1s-dKO (n = 21, median survival = 9 days) and cacna1s-bKO (n = 19, median survival = 9 days, gray), but not in cacna1s-dHet (n = 23), cacna1s-aKO-bHet (n = 19), cacna1s-aKO (n = 14) or WT siblings (n = 26). Log-rank test: WT vs dKO; P < 0.0001, WT vs bKO; P < 0.0001, WT vs aKO; P = 0.2830, WT vs dHet; P = 0.1171, WT vs aKO-bHet; P = 0.1213.

To further assess the motor performance of cacna1s-dKO, we conducted a touch-evoked response test on 4 dpf larvae (Fig. 2B, Supplementary Movies S1 and S2). WT sibling displayed a rapid escape reaction by a mechanosensory stimulation on their tail, whereas dKO did not show any escape response or even partial body movement. The complete paralysis of dKO muscle suggested that homozygous KO of cacna1sa and cacna1sb abolishes Cav1.1 function, thus resulting in a complete lack of muscle contractile function.

Cacna1sb, responsible for Cav1.1α1S expression in fast-twitch muscle, is vital for swimming and survival

The two cacna1s genes are differentially expressed in zebrafish: cacna1sa is expressed only in superficial slow-twitch muscle, whereas cacna1sb is mainly expressed in deep fast-twitch muscle and to a lesser extent in slow-twitch muscles [18]. To investigate the difference in muscle performance between fish carrying mutations in cacna1sa and/or cacna1sb, we performed a photochemical movement assay on all cacna1s mutants (Fig. 2C). Notably, cacna1s-aKO did not display differences compared to their WT siblings, whereas cacna1s-bKO showed a marked decrease in motor behavior. This indicates that cacna1sa deficiency in slow-twitched muscle is not sufficient to impair larval movement, either because of the preserved activity of the fast fibers, or due to compensation in slow muscles by cacna1sb. Interestingly, cacna1s-aKO-bHet exhibited reduced swimming performance as compared to WT. In addition, all cacna1s-dKO larvae traveled zero distance, consistent with their paralyzed phenotype (Fig. 2C). By contrast, none of the remaining other genotypes, including cacna1s-dHet, demonstrated differences compared to WT.

To evaluate the impact of mutations on long-term survival, we assessed their survival rate. Cacna1s-bKO and cacna1s-dKO did not live longer than 14 days (Fig. 2D). Fish of all other genotypes grew to adulthood and became fertile. There was no statistical difference in survival curve between WT, cacna1s-aKO, cacna1s-aKO-bHet and cacna1s-dHet. These data indicate that dysfunction of fast-twitch muscle due to cacna1sb deficiency results in swimming defects and early lethality. Of note, despite having similar survival rates, cacna1s-bKO continue to be retain swim activity (though significantly impaired) until death (which is likely due to impaired feeding and oxygenation), while cacna1s-dKO are paralyzed at all larval ages. Further, the cacna1s-dKO are grossly edematous starting from 6 dpf and morphologically distinct from the cacna1s-bKO.

Cacna1s-dKO zebrafish completely lack Cav1.1 expression

To determine the expression and localization of Cav1.1 and two other essential components of the ECC apparatus, RyR1 and STAC3, we conducted immunofluorescence (IF) staining on whole-mount larvae (4 dpf) and isolated myofibers from 5 dpf larvae. The expression of Cav1.1 in cacna1s-dKO was greatly reduced in both slow-twitch and fast-twitch muscle as expected, whereas alignments of myofibers and striations of sarcomeres remained well-organized (Fig. 3A). cacna1s-bKO also exhibited significantly reduced Cav1.1 expression in fast-twitch muscle, while cacna1s-aKO showed protein expression in both slow and fast-twitch muscles (Fig. 3A), consistent with the known distribution of cacna1sa to slow fibers and cacna1sb to both fiber types [18].

Figure 3

Expression and localization of Cav1.1 and other ECC components in cacna1s mutant Fish. (A) Whole-mount immunofluorescence of 4 dpf larvae. In cacna1s-dKO, Cav1.1 was not detected in either slow or fast muscle, while myosin striations were normally observed. The cacna1s-bKO did not express Cav1.1α1S in fast-twitch muscle, whereas cacna1s-aKO expressed Cav1,1 in both fast and slow twitch muscle. Fiber types were distinguishable by their orientations; superficial slow-twitch fibers aligned parallel to the anterior-posterior axis, while deep fast-twitch fibers angled slightly toward the midline. Scale bars = 20 μm. A total of 8 zebrafish per group were analyzed. (B) Representative images of IF staining on isolated myofibers from 5 dpf larvae. In cacna1s-dKO, not only Cav1.1 but also STAC3 expression was diminished, while RyR1 was detected at the SR cisternae with slightly reduced expression. In ryr1-dKO, co-expression of Cav1.1 and STAC3 was observed with fainter signals than that of WT, and no RyR1 expression was detected. Scale bars = 5 μm. A total of 10 myofibers per group were imaged.

For the examination of other ECC proteins, we focused analyses on the cacna1s-dKOs. In cacna1s-dKO myofibers, STAC3 expression was greatly reduced and RyR1 expression was modestly diminished (Fig. 3B). We additionally analyzed ryr1 null zebrafish (ryr1-dKO due to mutations in both ryr1a and ryr1b) [19] as a reference since RyR1 is physically and functionally paired with Cav1.1, and found that Cav1.1 and STAC3 remained co-localized within the triad in ryr1-dKO fish, despite reduced expression compared to WT (Fig. 3B). Western blot analysis corroborated these changes in expression, with significantly reduced levels of both Cav1.1 and STAC3 in cacna1s-dKO fish and a more modest reduction of both proteins in ryr1-dKO fish (Supplementary Fig. S4). RyR1 levels by Western blot appeared absent in ryr1-dKO and present in cacna1s-dKO, though blotting of WT samples showed variability, thus making interpretation of these blots challenging (Supplementary Fig. S4).

Cacna1s-dKO zebrafish exhibit morphological alterations in SR/T-tubule junctions

To further study the impact on muscle structure of cacna1s-dKO, we performed electron microscopy (EM) analyses using 5 dpf WT, cacna1s-dKO, and ryr1-dKO larvae. There were no differences in size and appearance of sarcomeres between groups. At the location of the ECC apparatus (i.e. the triad), we observed that longitudinal SRs in ryr1-dKO were frequently enlarged (Fig. 4A). Qualitatively, T-tubule lumens in cacna1s-dKO (and to a lesser extent in ryr1-dKO) were dilated, and T-tubule and SR membranes were located close together (Fig. 4B).

Figure 4

Ultrastructural analysis of cacna1s-dKO muscle. (A) Representative images of electron micrographs of WT, cacna1s-dKO, and ryr1-dKO at 5 dpf. Myofibrils were well-organized and triads were aligned regularly at the sarcomere Z-lines in WT and both mutants. Arrows show enlarged longitudinal SR in ryr1-dKO. Scale bars = 500 nm. A total of 4 larvae per group were analyzed. (B) Representative images of triad structure at higher magnification. In cacna1s-dKO, T-tubules were enlarged and the gaps between T-tubule and SR were extremely narrow compared to WT. Scale bars = 100 nm. (C) A diagram showing the elements measured. d; minor axis of T-tubule luminal dimension, e; major axis of T-tubule luminal dimension, f; area of T-tubule luminal dimension, g; distance between SR terminal cisternae and T-tubule, h; Distance between SR terminal cisternae, i; width of SR terminal cisternae. (D–K) lengths and areas indicated in (C) were measured and plotted, revealing T-tubules in both cacna1s-dKO and ryr1-dKO expanded toward SR in triads. Numbers of triad evaluated in (D–K) were as following: WT; n = 91, cacna1s-dKO; n = 56, and ryr1-KO; n = 56. All values are Mean ± SEM (see Supplementary Table S1). Differences were analyzed by Tukey’s multiple comparisons test and considered to be statistically significant at P < 0.05 (^*), P < 0.01 (^**), P < 0.001 (^***) or P < 0.0001(^****).

To further evaluate the morphological changes in triads, we also quantified T-tubule dimensions, terminal SR length, and junctional gap width in WT, cacna1s-dKO, and ryr1-dKO fish (Fig. 4C). Specifically, T-tubules of cacna1s-dKO and ryr1-dKO fish were significantly extended in the minor axis (Fig. 4D), while slightly shortened along the major axis (Fig. 4E), resulting in increased cross-sectional area (Fig. 4F) and a more circular aspect ratio compared to that of WT (Fig. 4J). Additionally, in cacna1s-dKO and ryr1-dKO fish, the distance between the SR cisternae and T-tubule was markedly decreased (Fig. 4G), with the gap between two cisternae across the triad increased (Fig. 4H), and SR cisternae width reduced (Fig. 4I), resulting in no net change in ratio between T-tubule and SR cisternae lengths (Fig. 4K). These results suggest that the loss of physical interactions between Cav1.1 and RyR1 proteins within the junction result in significant morphological changes in the triad.

Cav1.1 deficiency in cacna1s-dKO zebrafish abolishes electrically-evoked Ca²⁺ release and enhances caffeine-induced Ca²⁺ release

To examine the impact of Cav1.1 deficiency on ECC and RyR1 function, we quantified the peak magnitude of electrically-evoked and caffeine-induced (10 mM) Ca²⁺ release in single myofibers from wild type (WT) and cacna1s-dKO zebrafish loaded with fluo-4 (Fig. 5). While electrically-evoked Ca²⁺ release was abolished in myofibers from cacna1s-dKO zebrafish (Fig. 5A and B), peak caffeine-induced Ca²⁺ release was significantly increased (Fig. 5C and D). These results are consistent with Cav1.1 both being required for ECC and being a regulator of caffeine activation of RyR1 Ca²⁺ release, as reported previously [20, 21].

Figure 5

Electrically-evoked and caffeine-induced Ca²⁺ release in fibers from cacna1s-dKO zebrafish. (A) Representative diagram of Ca2+ transient traces with a 10 Hz stimulation on wild type (WT) and cacna1s-dKO myofibers dissociated from 5 dpf larvae. (B) Histogram showing that average (±SE) peak change in relative fluo-4 fluorescence (ΔF/F) during 10-Hz stimulation is reduced in cacna1s-dKO fibers compared to that of WT fibers (t-test, P = 0.016). WT: fiber average across n = 4 zebrafish; cacna1s-dKO: fiber average across n = 5 zebrafish. (C) Representative diagram of Ca2+ transient traces with an application of 10 mM caffeine on myofibers dissociated from 5 dpf larvae. (D) Histogram showing that average (±SE) peak change in relative fluo-4 fluorescence (ΔF/F) is increased in cacna1s-dKO fibers in response to exposure to 10 mM caffeine compared with that of WT fibers (t-test, P = 0.018). WT: fiber average across n = 4 zebrafish; cacna1s-dKO: fiber average across n = 5 zebrafish. Intracellular Ca²⁺ dynamics were quantified from relative peak change of fluo-4 fluorescence from baseline (ΔF/F). Data are represented as mean ± SEM. The differences were analyzed by unpaired t-test and considered to be statistically significant at P < 0.05 (^*), or P < 0.001 (^***).

Adult cacna1sa null fish exhibit reduced body size

Unlike cacna1s-bKO and cacna1s-dKO zebrafish, cacna1s-aKO and cacna1s-aKO-bHet Fish survive into adulthood, thus affording the opportunity to study the impact of these mutations on later development and adult phenotypes. To characterize adult mutant zebrafish, we observed their morphology and behavior. Qualitatively, both cacna1s-aKO and cacna1s-aKO-bHet fish exhibit reduced body size compared to WT fish, with cacna1s-aKO-bHet also showing morphological changes including mouths constantly opened and protruding jaws (Fig. 6A). Such changes in body size and facial appearance were observed consistently in all cacna1s-aKO-bHet Fish, whereas body size was more variable with cacna1s-aKO, and the mouth and jaw phenotypes seen only infrequently (in approximately 50% fish in the third generation, and not seen in earlier generations). Quantitatively, both cacna1s-aKO and cacna1s-aKO-bHet mutants showed reductions in body size, as shown by measuring standard length, total length, and body depth (Fig. 6B and C). These results indicate that loss of cacna1sa expression (either alone or in combination with reduced cacna1sb levels) impairs body growth, possibly because slow twitch muscle activity contributes to feeding and/or muscle mass growth.

Figure 6

Characterization of adult cacna1s mutants. (A) Representative appearance of 7 months-old fish. cacna1s-aKO and cacna1s-aKO-bHet fish were smaller than WT, with cacna1s-aKO-bHet showing protruding jaw (arrow). Scale bar = 5 mm. (B) Schematic showing standard length, total length, and body depth measurements. (C) Standard length, total length, and body depth were measured and plotted; WT siblings (n = 15), dHet (n = 10, green), aKO (n = 11) and aKO-bHet (n = 13). All values are mean ± SEM (see Supplementary Table S1). Differences were analyzed by Tukey’s multiple comparisons test and considered to be statistically significant at P < 0.05 (^*), P < 0.01 (^**) P < 0.001 (^***) or P < 0.0001(^****).

Cacna1s-aKO-bHet zebrafish are model of core myopathy

Due to the full penetrance of phenotypes in the cacna1s-aKO-bHet as compared to the cacna1s-aKO alone, we focused subsequent analyses exclusively on the cacna1s-aKO-bHet fish. Given the morphological distinctions observed in these fish, we examined cacna1sa/cacna1sb transcript levels and expression of key ECC proteins (Cav1.1, STAC3, RyR1, and Junctophilin-1; Jph1). By conducting RT-qPCR with mRNA isolated from (mixed fast and slow) skeletal muscle of 4-month-old fish, we confirmed that cacna1sa transcripts were markedly decreased in cacna1s-aKO-bHet Fish as expected, while cacna1sb transcript levels were not significantly altered (Fig. 7A). Western blot analyses revealed that expression levels of Cav1.1, RyR1, STAC3, and Junctophilin-1 (Jph1), which is known to directly interact with Cav1.1 [22], were not significantly differently between skeletal muscle of WT and cacna1s-aKO-bHet fish (Fig. 7B and C, Supplementary Fig. S5), potentially reflecting the persistence of cacan1sb expression, particularly in fast fibers.

Figure 7

Molecular analyses of cacna1s-aKO-bHet zebrafish mutants. (A) mRNA levels of cacna1sa were markedly reduced but cacna1sb were not significantly different in adult cacna1s-aKO-bHet at 4 months of age. A total of 6 biological independent samples per group were analyzed. Values are Mean ± SEM (fold); cacna1sa: WT (1.00 ± 0.22), aKO-bHet (0.15 ± 0.02); cacna1sb: WT (1.00 ± 0.17), aKO-bHet (0.66 ± 0.09). Statistical analysis was performed using by unpaired t test where P < 0.01 (^**) or non-significance (ns). (B) Representative Western blot for Cav1.1, STAC3, RyR1, Jph1 with myosin heavy chain as a myofibril marker and GAPDH as an internal control. A total of 8 samples (4 months old) per group were used for analyses. One blot was used for probing Cav1.1α1S, Jph1 and GAPDH, and the other blot was used for STAC3 and myosin with stripping between each process. RyR1 was processed in parallel with the same samples as used on the other two blots. A total of 8 samples per group were used for analyses and full images of the probed blots are provided in Supplementary Fig. S5. (C) Quantifications of protein expression levels, normalized to total protein (loading control) and myosin (muscle mass control), and plotted by fold changes relative to the average of WT on the same blots. WT (n = 7) vs cacna1s-aKO-bHet (n = 8). All values are Mean ± SEM (fold); Cav1.1: WT (1.00 ± 0.08), aKO-bHet (0.98 ± 0.06); STAC3: WT (1.00 ± 0.11), aKO-bHet (0.91 ± 0.05); RyR1: WT (1.00 ± 0.12), aKO-bHet (1.00 ± 0.10); Jph1: WT (1.00 ± 0.12), aKO-bHet (0.69 ± 0.09). Statistical analysis was performed using unpaired t-test and P values were noted on graphs.

We next examined muscle morphology by histological analyses. We focused investigations on tail muscle at 16 months of age, applying the common muscle histological stains hematoxylin and eosin (H&E) and toluidine blue (TB). By H&E staining, decreased myofiber size was observed in deep muscle of cacna1s-aKO-bHet fish compared to that of WT fish, while no such difference was observed for superficial muscle (Fig. 8). Of note, apparent core-like structures were detected by TB staining of deep muscle (but not superficial muscle) from male cacna1s-aKO-bHet fish, but not in either age-matched female cacna1s-aKO-bHet or WT fish (Fig. 8A). Interestingly, fibers from male cacna1s-aKO-bHet fish with core-like structures exhibited visibly darker and more granular toluidine blue staining than that of WT Fish, potentially representing higher ATP activity (and shift to a more type 1 histotype), and consistent with the distinctive pattern of the inter-myofiber reticulum reported previously in human CACNA1S-myopathy patients [8]. Overall, based on reduced motor performance (Fig. 2C) and muscle mass (Fig. 6A), as well as the presence of core structures, cacna1s-aKO-bHet zebrafish phenocopy human core myopathy.

Figure 8

Histopathological analyses on cacna1s-aKO-bHet. (A) In cacna1s-aKO-bHet at 16 months of age, there were occasional hypotrophic fibers in the fast twitch muscle compartment, while all slow twitch muscle was relatively uniformly sized. Toluidine blue (TB) staining revealed striking core structures (arrows) in fast twitch fibers in male mutants with dusty stained sarcoplasm. Scale bars = 20 μm. A total of 4 samples (2 female, 2 male) per group were used for analyses. (B) cacna1s-aKO-bHet showed a marked reduction in deep myofiber size as measured by minimum Feret’s diameter. WT (n = 4, 1 section per individual) and cacna1s-aKO-bHet (n = 4, 1 section per individual) at 16 months of age were analyzed and total fibers measured across four different sections per group were as follows; WT: n = 837 (217 + 214 + 222 + 184), aKO-bHet: n = 1408 (310 + 323 + 360 + 415). Mean minimum Feret’s diameter in each section was calculated and plotted. Values are mean ± SEM (μm); WT (30.96 ± 0.67), aKO-bHet (21.36 ± 0.50). Statistical analysis was performed using unpaired t-test and considered to be statistically significant at P < 0.0001 (^****). (C) Percentage distribution of minimum Feret’s diameter in deep muscle. The pattern of percentage distribution in aKO-bHet is left-shifted (i.e. shifted toward smaller fiber size) compared to that in WT. Sample size is the same as (B). Values are Mean ± SEM. % myofiber size distribution ± SEM for <10 μm/10–20 μm/20–30 μm/30–40 μm/40–50 μm/50–60 μm/>60 μm bins are: WT; 0.25 ± 0.14/8.45 ± 2.42/35.55 ± 2.17/42.49 ± 4.50/12.85 ± 2.53/0.27 ± 0.27/0.14 ± 0.14, aKO-bHet; 7.46 ± 1.38/39.25 ± 2.60/36.68 ± 2.55/15.09 ± 3.07/1.52 ± 0.54/0.00 ± 0.27/0.00 ± 0.00.

The craniofacial deformities of cacna1s-aKO-bHet Fish mirror myopathic facial changes in congenital myopathy

Individuals affected with CACNA1S-related congenital myopathy exhibit an abnormally high-arched palate [8]. Moreover, most congenital myopathy patients manifest facial dysmorphisms such as elongated-face, dolichocephaly, and the mouth being held in an open position [23]. Thus, we assessed cranio-facial development and structure of cacna1s-aKO-bHet zebrafish.

To probe the relationship between facial dysmorphology and cartilage/bone development, we visualized cartilage structure with alcian blue staining in 7 dpf and 10 dpf fish. At 7 dpf, no differences were found in cranial structures between WT and cacna1s-aKO-bHet fish (Supplementary Fig. S6A). However, at 10 dpf, all cacna1s-aKO-bHet larvae could be distinguished by exhibiting a protruding jaw, fixed open mouth, disorientation of Meckel’s cartilage, and smaller-angled Ceratohyal cartilage (Fig. 9A and Supplementary Fig. S6B). We next examined calcified bone structure of 28 dpf fish with alizarin red staining. Interestingly, there is a striking jaw deformity in cacna1s-aKO-bHet fish, including protrusion and fixed in an open position, that is associated with disorientation of the jaw joint that is shifted upward and inward (Fig. 9B).

Figure 9

Morphological analysis on cacna1s-aKO-bHet. (A) At 7dpf, bright field (left) and alcian blue staining images (middle and right) showed no obvious difference between WT and aKO-bHet mutants in mandible development. At 10 dpf, aKO-bHet mutants could be recognized by protruding jaw (arrow). Meckel’s cartilage of mutants were dislocated as dropping in the lateral view of alcian blue staining, and ceratohyal angle (asterisk) of mutants was narrower than that of WT. The angle between the two ceratohyal elements is highlighted (right most panels, chevron shaped line). Abbreviations: m, Meckel’s cartilage; Ch, ceratohyal bone. Scale bars: 200 μm. A total of 8 larvae per group were imaged. (B) Representative imaging of WT and aKO-bHet stained with alizarin red at 28 dpf. Lateral view of mutant revealed the jaw joint (arrow) positioned upward therefore the dentary bones downturned. The ossification of the craniofacial bones in upper jaw and skull (dashed lines in lateral and dorsal views) were severely decreased in aKO-bHet mutants compared to WTs. The ceratohyal angle (asterisk) of mutants was narrower than that of WT, and non-progressive compared to that of 10 dpf. There was no difference in spinal formation and mineralization between WT and mutants, while vertebrae and fin rays of mutants were thinner than that of WT in the longitudinal view. Abbreviations: den, dentary bone. Scale bars: 500 μm. A total of 8 samples per group were imaged. (C) The micro-CT reconstruction images of WT and aKO-bHet mutant at the age of 14 months. Lateral view of mutant exhibited that the dentary bone was dislocated with increased calcification and irregular margins, and that upper jaw elements were smaller than that of WT. Ventral view of mutant showing significantly narrowed snout, mouth, and jaw due to the deformity of mandibular bones. Dorsal view of mutant’s skulls displays a lower density compared to WT. The vertebrae and spine of mutant were thin and shortened, corresponding to the body size. Abbreviations: den, dentary bone. Scale bars: 1 mm. A total of 2 samples per group were analyzed.

Compared to age-matched WT fish, cacna1s-aKO-bHet fish also exhibited a delay in the ossification of craniofacial bones in the upper jaw and skull, whereas there no differences in bone formation or mineralization were observed in other skeletal elements (Fig. 9B). Using high resolution micro-CT and three-dimensional skeletal reconstructions, we found that this pattern of jaw deformity persisted from juvenile through adulthood (14 months). This included dislocation of the jaw joint and down turning of the lower jaw, and an open-fixed mouth, mild hypoplasia of the upper jaw, narrowing of the lower jaw, a smaller snout, and thinned and deformed skulls (Fig. 9C). Collectively, our results indicate that cacna1s-aKO-bHet zebrafish exhibit dysmorphic cranio-facial features consistent with the facial involvement observed in congenital myopathy patients. Of note, ryr1-dKO fish are also reported to show abnormal craniofacial morphology [19], though the premature death of these Fish precludes assessing this phenotype longitudinally or with imaging techniques such as micro-CT.

Discussion

The aim of the study was to develop and characterize novel zebrafish models of CACNA1S-related myopathy. We achieved this by generating single knockouts of each zebrafish CACNA1S paralog (cacna1s-aKO and cacna1s-bKO), as well as both double mutants (cacna1s-dHet and cacna1s-aKO-bHet) and complete double knockouts (cacna1s-dKO). Cacna1s-dKO Fish completely lack Cav1.1 expression and exhibit a severe disease presentation, while cacna1s-aKO-bHet Fish exhibit a more mild “core myopathy” version of the disorder, enabling the first full characterization in zebrafish of facial dysmorphism from a muscle mutant. Consistent with prior studies in mice [17, 24], we demonstrate the absolute requirement of intact Cav1.1 for muscle EC coupling and confirm that triad localization of STAC3 requires Cav1.1 but not RyR1. In total, we generated novel models of CACNA1S-related myopathy suitable for future studies of disease pathomechanisms and therapy identification and translation.

The zebrafish genome contains two distinct genes orthologous to human CACNA1S; cacna1sa and cacna1sb [18]. Both genes code for intact zebrafish Cav1.1 α1 subunits (zf-α1S-a and zf-α1S-b) that contain all the molecular characteristics required for functional interaction with other essential proteins of ECC including Cav1.1 β1 subunit [25], RyR1 [26], and STAC3 [27], and thereby, to support voltage sensor function in DICR. The two zf-α1S isoforms share an amino acid homology of 73%, with zf-α1S-a expressing primarily in superficial slow muscle whereas zf-α1S-b expresses mainly in deep fast muscle [18]. We found that complete knockout of zf-α1S-a did not significantly affect motor function or survival, while complete knockout of zf-α1S-b led to larval death due to weakness of fast-twitch muscle leading to a marked swimming defect. Cacna1s-dKO zebrafish, which lack expression of zf-α1S in all muscle, have impaired DICR and complete paralysis, analogous to the severe phenotypic subtype of CACNA1S-related myopathy with fetal akinesia [28]. In contrast, cacna1s-aKO-bHet fish exhibit a non-progressive muscle weakness with facial involvement and core structures in muscle fibers, modeling a mild form of human core myopathy.

Cacna1s-dKO fish exhibit morphologically mature skeletal muscle with a similar overall architecture to that of WT fish; e.g. well-organized myofibrils with normal sarcomeres, T-tubules in transverse orientation, normal shaped triads, and no obvious morphological changes in the nucleus and mitochondria. In contrast, mdg mice exhibit significant muscle degeneration from early embryonic stages, loose appositions of swollen SR and T-tubule, vacuolization and contraction clots in myoplasm, as well as abnormal swollen and irregularly shaped nuclei [10, 29]. It is not feasible to use mdg mice for in vivo experiments, as they die immediately following birth. Nevertheless, myotubes differentiated from primary myoblast cultures isolated from newborn mdg/mdg mice have been utilized in multiple elegant Cav1.1 structure/function studies [17, 30–32]. However, myotubes grown under 2D cell culture conditions are limited with regard to the degree of structural maturation, thereby limiting their usefulness for studying dynamic processes of differentiated skeletal muscles. Zebrafish skeletal muscles, in contrast, achieve a high level of differentiation within only a few days after birth [25] and their optical clarity allows dynamic imaging of all the cells in live larvae. Thus, cacna1s-dKO fish provide a powerful new experimental tool to assess the impact of ECC ablation in more differentiated and striated myofibers with minimal level of secondary morphological damage, which is likely to enable new insights into ECC via both in vivo and in vitro experiments.

The frequency of triads in mouse diaphragm muscle is reduced to ~50% in both Ryr1 (“dyspedic”) and Cacnals null (mdg/mdg) mice compared to that observed in WT mice [33]. Additionally, the orientation of the SR/T-tubule junctions relative to the fiber long axis is random, and T-tubules stay predominantly longitudinal when newly formed, as the transverse orientation of T-tubules occurs during the postnatal period in mice [34]. In contrast, electron microscopy observations of myofibers from cacna1s-dKO and ryr1-KO fish at 5 dpf revealed that SR/T-tubule junctions are present and regularly positioned at all Z lines, and the orientation of all T-tubules were transverse. As a result, in contrast to that reported in mice, triads assemble efficiently in cacna1s- and ryr1-deficient zebrafish muscle.

The normal formation of SR/T-tubule junctions in cacna1s-dKO and ryr1-dKO zebrafish is consistent with previous findings that SR-T tubule membrane docking is independent of the presence of Cav1.1 and RyR1 [33]. However, the junctional gap between SR and T-tubule is significantly narrower in cacna1s-dKO and ryr1-dKO, suggesting that Cav1.1 and RyR1 are involved in maintaining the architecture of SR/T-tubule junctions. Indeed, the junctional gap between SR and T-tubules is also smaller in mdg/mdg mouse diaphragm at E18 [33] and hindlimb muscle at E21 [35] and in Cav1.1β1-null mouse hindlimb muscle at E18 [36], while the difference in the junctional distance is indistinguishable between normal and dysgenic diaphragm at E18 [33]. In cacna1s and ryr1 dKO zebrafish, T-tubule membranes shift closer to SR cisternae during the maturation of triads, likely after initial SR-T tubule docking, at the point when Cav1.1 and RyR1 are normally physically and/or functionally connected. Of note, we did not focus on junctional feet in our electron microscopic analyses; junctional feet have previously been shown to be less distinct and less frequent in mdg/mdg mouse myotubes and myofibers (as compared to WT littermates) and will be of interest to study in the future in our models [37, 38]. In total, our results indicate that Cav1.1-RyR1 physical and functional interactions are required to maintain the mature structure of SR/T-tubule junctions.

The cacna1s-aKO-bHet Fish developed in this study provide a first zebrafish model of a congenital myopathy associated with mutations in genes involved in ECC that survives into adulthood. Prior zebrafish models with impaired DICR reported to date include cacnb1-null [25], ryr1-dKO [19], ryr1b-null [39], and stac3-null [27], which are all larval lethal. Heterozygote mutants from these prior lines do not exhibit phenotypes related to muscle disorders. On the other hand, our cacna1s-aKO-bHet Fish exhibit mild and non-progressive muscle weakness with a normal life span. One of the advantages of acquiring adult fish for muscle study is that relatively large masses of skeletal muscle can be obtained, which allows for comprehensive histological analyses in differentiated muscle. Indeed, in cacna1s-aKO-bHet Fish, we observed cores by optical microscopy, which are histopathological lesions common in human congenital myopathies but that have not previously been reported in zebrafish. Access to sufficient muscle material also enables preparation of sections from frozen samples for enzyme histochemistry, which combines biochemistry and morphology as a gold standard for pathological analyses of muscle diseases. The availability of large quantities of muscle and cost advantages over mice should also facilitate improved throughput and a more comprehensive understanding of the mechanisms that underlie the formation of cores, which remains elusive.

In our study, we found cores only in male cacna1s-aKO-bHet Fish, suggesting a potential sex bias in phenotype. Core-like structures visualized by light microscopy are not a frequent observation in human CACNA1S-related myopathy (more common is a trabecular appearance on SDH/NADH stains) [8], though have been reported in one adult male. In general, phenotypic differences between sexes have been observed in both patients and animal models of RYR1 and CACNA1S myopathies. For example, a mouse model of central core disease (CCD) and malignant hyperthermia susceptibility caused by Ryr1 mutation demonstrates more severe response in male mice to both heat and anesthesia [40]. Moreover, incomplete penetrance and later disease onset has been observed in woman with CACNA1S-related HypoPP with R528H mutation [41], and the phenotype of the corresponding Cav1.1-R528H mouse was more severe in male [11]. Further work will be required to better elucidate the role of biologic sex in the phenotypic presentation of CACNA1S-related myopathy.

The craniofacial dysmorphisms observed in cacna1s-aKO-bHet fish resemble the “myopathic facies” seen in patients with congenital myopathies, which include long (and narrow) face, high-arched palate, and open mouth. The narrow snout and mandible of cacna1s-aKO-bHet fish is a comparable phenotype, in spite of differences in facial structure and craniofacial development between fish and humans [42]. The characteristic craniofacial dysmorphism in congenital myopathies is believed to be primarily caused by pronounced facial weakness of the lower face and mouth [23]. Indeed, reduced loads in mouse facial muscle by altered diet consistency or by genetic abnormalities causing muscular dystrophy (mdx mice) have been shown to affect craniofacial development and shaping [43]. In zebrafish, removal of muscle activity through anaesthetization or genetic manipulation (myod^fh261) cause a change in the shape of jaw joint [44]. Additionally, paralyzed ryr1-KO zebrafish exhibit abnormal protruding Meckel’s cartilage, which is attributed to strains in jaw joint due to diminished muscle activity [19]. Therefore, the dysmorphic lower jaw phenotype of cacna1s-aKO-bHet fish likely involves a similar mechanism. Further, the open-mouth phenotype may be a manifestation of the broad extent of morphological changes in multiple craniofacial elements, as jaw movements are thought to be critical to the proper development of zebrafish snout and skulls [45]. To determine whether this facial dysmorphism is specific to defects in the ECC apparatus, or are instead a common manifestation of congenital myopathy models in fish, assessment of facial phenotypes in other genetic fish models are needed.

In conclusion, this work describes the generation and characterization of the first zebrafish models of CACNA1S-related myopathy. Two distinct zebrafish models mirror multiple features of severe and mild CACNA1S-related myopathy and illustrate that changes in Cav1.1 expression levels in different muscle compartments impact disease severity. These new zebrafish models are suitable for further defining disease pathomechanisms and identifying therapies for CACNA1S-related myopathy.

Materials and Methods

Zebrafish strains and husbandry

Zebrafish (AB strain) were raised and maintained at the Zebrafish Facility at the Hospital for Sick Children in strict accordance with the Animals for Research Act of Ontario and the Guidelines of the Canadian Council on Animal Care. All experiments in this study were approved by an institutionally approved animal use protocol (#1000052731). Genotyping of Fish was performed using genome DNA extracted by fin clipping.

Generation of cacna1sa and/or cacna1sb mutant lines

CRISPR/Cas9-mediated mutagenesis in AB zebrafish was performed as previously described [46]. Guide RNAs for cacna1sa (ZFIN:ZDB-GENE-090514-1, Ensembl:ENSDARG00000029457, NCBI_Gene: 570683) and cacna1sb (ZFIN:ZDB-GENE-051227-1, Ensembl: ENSDARG00000042552, NCBI_Gene: 405791) were designed using the online tool Chopchop (https://chopchop.cbu.uib.no) [47]. The in vivo cutting efficiency of each sgRNAs was evaluated as previously described [48] and sgRNA-1; 5|${}^{\prime }$|-AGCGGTGGATAATCTGGCCGAGG-3|${}^{\prime }$| in cacna1sa at exon 14 and sgRNA-2; 5|${}^{\prime }$|-TGGAAAGTCAGCAGGAGGAAAGG-3|${}^{\prime }$| in cacna1sb at exon 16 were selected (Supplementary Fig. S1). These two sgRNAs had no predicted off-targets. One-cell-stage WT embryos were injected with 50 pg sgRNA-1, 100 pg sgRNA-2 and 150 ng of Cas9 mRNA and raised to adulthood. The potential founder (F0) was outcrossed to WT zebrafish and F1 embryos were genotyped using high resolution melting (HRM) analysis and sanger sequencing. F1 Fish carrying mutation cacna1sa ^Δ5/+ or cacna1sb ^Indel/+ were further outcrossed to WT to avoid inbreeding. To generate cacna1s-dHet (cacna1sa ^Δ5/+ cacna1sb ^Indel/+), single-heterozygous Fish of F2 or later generations were crossed (Fig. 1A). Primers used for HRM to detect cacna1sa Δ5 were as follows: forward 5|${}^{\prime }$|-GAATGATCAACCGCAGACATCCT-3|${}^{\prime }$| and reverse 5|${}^{\prime }$|-GCTTCTTCTCCTCTGCTTTTTCC-3|${}^{\prime }$|⁠. Primers used for PCR and sanger sequencing were as follows: cacna1sa Δ5 forward 5|${}^{\prime }$|-GCGAAGGCTGGAACTATGTC-3|${}^{\prime }$|⁠, reverse 5|${}^{\prime }$|-GCAGGCAGGCACTCATTTAG-3|${}^{\prime }$|⁠; cacna1sb Indel forward 5|${}^{\prime }$|-AGGAGGAGAAAGCCCTGATG-3|${}^{\prime }$|⁠, reverse 5|${}^{\prime }$|-GCCAGCAGTGTGGTTTAAAAT-3|${}^{\prime }$|⁠. Institutional line designation for cacna1s Δ5 and cacna1sb Indel are hsc198 and hsc199, respectively.

RT-qPCR

A pool of five embryos (6 dpf) with removed heads and internal organs were collected as a biological independent sample (5 pools per group). Adult fish (n = 6 per group, 4.5 months old) were euthanized, and tissues were collected indicated in Supplementary Fig. S7A. Total RNA was isolated using the RNeasy mini kit (QIAGEN) and reverse-transcribed using iScript™ cDNA Synthesis Kit (Bio-Rad) using a total of 20 ng of RNA. Quantitative PCR was performed using SYBR Green (Thermo Fisher, A25741) on the Applied Biosystems StepOne RealTime PCR System (Thermo Fisher Scientific) in triplicates. mRNA levels of cacna1sa and cacna1sb were analyzed using the 2-ΔΔCT method with eef1a1l1 as an endogenous control gene. Primers used were: cacna1sa forward 5|${}^{\prime }$|-GCTGTCAATGGGAATGGAAT-3|${}^{\prime }$|⁠, cacna1sa reverse 5|${}^{\prime }$|-GCCTTTGGCTCGATTGATAG-3|${}^{\prime }$|⁠, cacna1sb forward 5|${}^{\prime }$|-TCTGCCGAAACTCCTTCAAC-3|${}^{\prime }$|⁠, cacna1sb reverse 5|${}^{\prime }$|-CCCTCAACACCCTGAGAATC-3|${}^{\prime }$|⁠, eef1a1l1 forward 5|${}^{\prime }$|-CTGTTACCTGGCAAAGGGGA-3|${}^{\prime }$|⁠, eef1a1l1 reverse 5|${}^{\prime }$|-CGTGGC CAATAACCACGATG-3|${}^{\prime }$|⁠. Statistical analysis was performed using one-way ANOVA with Dunnett’s multiple comparisons test or unpaired t-test in GraphPad Prism version 9.3.1 (GraphPad Software).

Swimming behavior analysis

To examine evoked escape response, 4 dpf fish were touched with a fine needle at the tail region, and movement of the response was recorded using an Olympus SZX7 Stereo Microscope with Firefly color camera (FMVU-03MTC) at 30 frames per second. To quantify swimming performance, 6 dpf larvae from cacna1s-dHets crossing were individually transferred to a 96-well plate and incubated with 10 μM optovin 6b8 (ChemDiv) at 28.5°C for 5 min prior to the photoactivation as previously described [49]. Motor activity of the embryos was recorded and analyzed using ZebraBox (Viewpoint Behaviour Technology) with a photoactivating protocol; three 10-second lighting cycles with a 1-min break (dark period) in between. After recording, genomic DNAs of individual larvae were extracted for genotyping. Total distance traveled, including during the dark periods, was normalized to the average distance of WT siblings on the same plate. The normalized data from three independent experiments was plotted and statistically analyzed using one-way ANOVA with Dunnett’s multiple comparisons test in GraphPad Prism version 9.5.1 (GraphPad Software).

Survival analysis

Larvae laid by one pair of cacna1s-dHets were fin clipped at 4 dpf, grouped by genotype and transferred to the system for feeding at 6 dpf. Survival counts and health checks were performed daily until the Fish reached 60 days old. Kaplan–Meier survival plot was generated and statistically analyzed using Log-rank test in GraphPad Prism version 9.5.1 (GraphPad Software). Two biological independent experiments were conducted.

Immunofluorescence (IF) assay

Whole-mount IF preparations were processed as previously described [50]. Zebrafish embryos were fixed with ice-cold methanol (Electron Microscopy Sciences, Hatfield, PA, USA), left overnight in −20°C, followed by 15 min treatment with pre-chilled acetone at −20°C. Then, embryos were blocked with 2% goat serum, 1% BSA, 1% DMSO in PBST (0.1% tween in PBS) for 2 h at room temperature (RT). Embryos were next co-incubated with primary antibodies anti-CACNA1S [1A] (1:100, abcam ab2862) and anti-myosin (1:100, DSHB A4.1025) overnight at 4°C, and then secondary antibodies Alexa Fluor 488 goat anti-mouse IgG1(α1) (1:300, Invitrogen) and Alexa Fluor 555 goat anti-mouse IgG2 (γ2a) (1:300, Invitrogen) overnight at 4°C. The embryos were mounted on glass slides with ProLong Gold Antifade Mountant with DAPI (ThermoFisher Scientific). Images were taken using a Leica SP8 Lightning Confocal microscope.

Myofiber isolation from 5 dpf larvae was processed as previously described [51]. Myofibers were fixed with 4% paraformaldehyde and blocked with 10% goat serum and 1% BSA in PBST. Primary antibodies anti-CACNA1S [1A] (1:200, abcam ab2862), anti-RyR (1:100, DSHB 34C), and anti-STAC3 (1:200, gifted by Dr Kuwada at Michigan University) were used overnight at 4°C. Secondary antibodies Alexa Fluor 488 goat anti-rabbit IgG (1:1000, Invitrogen), Alexa Fluor 647 goat anti-mouse IgG (1:1000, Invitrogen), and Rhodamine Phalloidin (1:1000, Invitrogen) were used for incubation for 2 h at room temperature. Images were taken using the Quorum Spinning Disk Confocal microscope.

Western blotting

A pool of 60 larval tails (6 dpf) were lysed with 180 μl lysis buffer (Cell Signaling #9803) containing protease inhibitor (Roche #11836170001) and phosphatase inhibitor (Millipore #524625) (WT; 7 pools, cacna1s-dKO; 5 pools, ryr1-KO; 4 pools). Four-months-old zebrafish were dissected as indicated in Supplementary Fig. S7A, and then lysed (WT; n = 8, cacna1s-Δ5Indel; n = 8). Larvae were homogenized with a microfuge pestle and adult samples were homogenized with QIAGEN TissueLyser II. Protein concentration was measured using the Pierce BCA protein assay kit (ThermoFisher Scientific #23225). Aliquots containing 40 μg of protein were resolved by SDS-PAGE and transferred onto a PVDF membrane (BioRad #1620174) according to standard procedures. Total protein stains were conducted using Revert 700 Total Protein Stain (LI-COR Biosciences) followed by imaging with the Odyssey Fc Imager (LI-COR). Primary antibodies used were: anti-CACNA1S [1A] (1:1000, abcam ab2862), anti-RyR (1:500, DSHB 34C), anti-STAC3 (1:1000, gifted by Dr Kuwada at Michigan University), anti-junctophilin1 (1:1000, Invitrogen #40-5100), anti-myosin (1:1000, DSHB, A4.1025), and anti-GAPDH (1:5000 Origene TA802519). Anti-rabbit or anti-mouse IgG-HRP conjugate (1:5000, Bio-Rad) were used for secondary antibody. Finally, detection was performed on membranes with ECL Western Blotting Substrate (Perkin Elmer NEL103001EA) and imaged on a ChemiDoc MP (Bio-Rad).

Ultrastructural analysis

Electron microscopy (EM) on 5 dpf larvae was performed as previously described [52]. For the measurement of triad elements, four embryos of each genotype were fixed and processed for imaging. At least three different locations per sample were analyzed. For triad measurements, the images were taken at 62 kx magnification in a plane perpendicular to the orientation of a T-tubule, and with the entire shape of the T-tubule and paired SR cisternae clearly detected. Length and area of interest were assessed in a blind manor; measurements were conducted by a person not involved in EM imaging, with all sample information concealed and using the open-source software Image J. Statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparisons test in GraphPad Prism version 9.3.1 (GraphPad Software).

Measurement of electrically-evoked Ca²⁺ release in dissociated fibers

Zebrafish myofibers were dissociated with collagenase as previously described [51]. Myofibers were loaded with 5 μM fluo-4AM (Molecular Probes) for 45 min at room temperature in a normal rodent Ringer’s solution consisting of (in mM): 145 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES, pH 7.4, followed by de-esterification in normal rodent Ringer’s solution supplemented with 25 μm N-benzyl p-toluene sulfonamide (BTS) for 20 min at room temperature to inhibit contractions. Fluo4-loaded myofibers were excited at 480 ± 15 nm and fluorescence emission detected at 535 ± 20 nm was collected at 10 kHz using a photomultiplier system. Myoplasmic Ca²⁺ transients in fluo4-loaded dissociated fibers were electrically stimulated using a glass electrode filled with 200 mM NaCl placed adjacent to the cell of interest. This stimulation protocol consisted of five twitch stimulations delivered at 1 Hz followed by a 10 Hz stimulation and then finally 30 s application of 10 mM caffeine. Caffeine was applied to the fiber using a rapid (response time < 5 s) local perfusion system (Warner Instrument Corporation, Hamden, CT). Peak Ca²⁺ transient amplitude was expressed as (F_max-F₀)/F₀ (or ΔF/F), where F_max is the maximum fluorescence and F₀ is the fluorescence at time 0 (prior to stimulation).

Body size measurement of adult fish

Progeny from cacna1s-dHets were raised in one tank, fin clipped at 7 months, and grouped according to the genotype. Fish were anesthetized with 0.02% tricaine, transferred onto a measuring board, quickly photographed, and returned to a tank. Measurements were performed using image J and data were plotted and statistically analyzed using one-way ANOVA with Tukey’s multiple comparisons test in GraphPad Prism version 9.3.1 (GraphPad Software).

Histopathological analysis

At 16 month of age, four cacna1s-aKO-bHet mutants (two female and two male) and four WT siblings (two female and two male) were euthanized and then fixed with 4% PFA for 48 h at 4°C. After rinsing with PBS, Fish were incubated in 14% EDTA (pH 7.4) in PBS on a rocker at room temperature for 4 days, with daily solution changes. Paraffin embedded samples were cross-sectioned perpendicular to body axis with 4 μm thickness (Supplementary Fig. S7B). Slides were stained with Mayer’s H&E or toluidine blue following standard protocols. Micrographs were captured with an Infinity1 camera (Lumenera Corp.) through an Olympus BX43 light microscope. Immunohistochemical (IHC) assays with anti-SERCA1 antibody was applied to distinguish type 1 (slow-twitch) and type 2 (fast-twitch) fibers (Supplementary Fig. S7C). IHC was performed using standard protocol for paraffin sections. Slides were deparaffinized, rehydrated and processed heat-induced epitope retrieval with citrate-based antigen unmasking solution (Vector Laboratories H-3300-250). Sections were blocked in 10% horse serum and then incubated with anti-SERCA1 (1:500, abcam ab2819) for 1 h at room temperature. Biotinylated secondary antibody (ABC-HRP kit, Vector Laboratories PK-6101or PK-6102) was applied for 30 min, and incubated in the Avidin-Biotin detection system for 30 min. Signal was detected using DAB Substrate Kit (Vector Laboratories), counterstained in hematoxylin, and mounted with Permount (Thermo Fisher Scientific). Myofiber size were measured as previously described [53].

Skeletal staining

Cartilage staining with Alcian blue was conducted in accordance with a modification of the protocol previously described [54]. Zebrafish larvae at 7 and 10 dpf were fixed in 4% PFA overnight at 4°C, and then separated head and body. Bodies were used for genotyping. Heads were washed with PBS and dehydrated with 50% ethanol, and then immersed in alcian blue solution (0.02% Alcian Blue, 50 mM MgCL2, 70% ethanol) overnight at room temperature. After staining, specimens were washed with water and bleached in a solution of 1.5% H202 and 1% KOH until pigmentation was removed. Finally, specimens were cleared with a solution of 20% glycerol with 0.5% KOH for 1 h, and then transferred into a solution of 50% glycerol with 0.25% KOH. Samples were imaged with a Zeiss Axio Zoom V16 microscope.

Bone staining with Alizarin red was conducted according to the protocol previously described [55]. Zebrafish at 28 dpf were fixed in a fixative solution (5% formalin, 5% Triton X-100, 1% KOH) for 24 h at 42°C, and then immersed for clearing in an enhancement solution (20% ethylene glycol, 5% Triton X-100, 1% KOH) for 48 h at 42°C. Then specimens were immersed in alizarin red solution (0.05% alizarin red S, 20% ethylene glycol, 1% KOH) overnight at room temperature. After staining, fish the specimens thoroughly washed in a prewarmed clearing solution (20% Tween 20, 1% KOH) at 42°C until background staining was removed. Finally, the specimens were moved through graded series of glycerol (30%, 50%, 70%, 90%) and store in 100% glycerol. Samples were imaged with a Zeiss Axio Zoom V16 microscope.

Micro-computed tomography (micro-CT)

The preparation and procedure for micro-CT scanning on 14-month-age zebrafish were according to the protocol described previously [56]. Briefly, Fish were fixed in neutral buffered 10% formalin (Sigma) overnight at 4°C and then mounted in 2% low melt agarose (BioShop) in a plastic vial. Specimens were scanned for 1 h using SkyScan 1275 high-speed Micro-CT scanner (Bruker) with the X-ray power at 45 kVp and 200 μA. All three-dimensional Micro-CT datasets were reconstructed with 17.7–22.8 μm isotropic resolution. The images were analyzed using CTVox software (Bruker).

Statistical analysis

All statistical analyses were conducted using GraphPad Prism version 9.5.1 (GraphPad Software). Details of statistic testing used in each experiment have been described in the methods section and figure legends. All significance is reported at P < 0.05 and all values are expressed as mean ± SEM. All sample data points were included in each analysis unless stated otherwise.

Acknowledgements

We thank Dr John Kuwada at University of Michigan for offering the anti-STAC antibody, and Vanessa Schartner for discussion. We thank Dr Kuwada and Dr Jonah Grunwald for originally supplying ryr1 KO zebrafish. We acknowledge the Imaging Facility, the Nanoscale Biomedical Imaging Facility, and the Zebrafish Facility at the Hospital for Sick Children for their technical assistance.

Funding

This work was supported by an RDMM grant from the Canadian Institutes of Health Research (CIHR), a project scheme grant from CIHR to J.J.D. (148603), AFM-Téléthon (22734) to J.L., and through an NIH R01s to J.J.D. and R.J.D. (R01AR078000 and R01082209).

Data availability

All data relevant to the study are provided in the manuscript. Source images and data are stored at the Hospital for Sick Children and are available upon request.