Localized TPC1-mediated Ca²⁺ release from endolysosomes contributes to myoseptal junction development in zebrafish

In zebrafish embryos, many aspects of the development and differentiation of the trunk musculature are well documented (Devoto et al., 1996; Blagden et al., 1997; Cortés et al., 2003; Ono et al., 2015). Teleost skeletal muscles are organized as a series of somite-derived myotome blocks, which extend the length of the trunk (Fig. 1A; Fig. S1A). Each myotome block is separated by a vertical myoseptum, which is functionally homologous to mammalian tendons (Gembella and Vogel, 2002). The myoseptum contains a collagen fibril-rich extracellular matrix (ECM), to which the actin cytoskeleton of individual myofibers is attached via multi-protein complexes forming myoseptal junctions (MJs; Charvet et al., 2011). The MJs are the site of force transmission at the onset of prehatching trunk coiling contractions (Saint-Amant and Drapeau, 1998) and during subsequent swimming behavior (Naganawa and Hirata, 2011). In each myotome block there are two main types of skeletal muscle cell, slow muscle cells (SMCs), the focus of this report, and fast muscle cells (FMCs; Fig. 1A; Fig. S1A–C) (Devoto et al., 1996). The SMCs arise from adaxial cells, which originate in a medial position adjacent to the notochord (Blagden et al., 1997), and these can be classified into two main types, superficial SMCs, and muscle pioneer cells (Fig. S1B). Both types of progenitor cell elongate in a rostral–caudal direction in each myotome block and migrate as mononucleate cells through the paraxial mesoderm (composed mainly of FMC precursors) until they reach the periphery of the trunk myotome (Devoto et al., 1996; Cortés et al., 2003; Stellabotte and Devoto, 2007; Ono et al., 2015). Following elongation, the medial surface of the muscle pioneer cells remains adjacent to the notochord, distinguishing them from the superficial SMC population. On reaching the myotome periphery, dynamic interactions occur between the terminal sarcolemma of each superficial SMC and the ECM in the vertical myoseptum, which results in a molecular coupling to prevent detachment of the superficial SMCs when contraction begins at ∼17 h post fertilization (hpf) (Henry et al., 2005; Charvet et al., 2011; Wood and Currie, 2017). The attachment of the superficial SMCs at the MJ involves component delivery and assembly of transmembrane protein complexes such as the dystrophin-associated protein complex (DAPC), which links the actin cytoskeleton in the SMCs to laminins in the myoseptal ECM via the large rod-shaped protein, dystrophin and the laminin receptor dystroglycan (Fig. S1D; Basset et al., 2003; Guyon et al., 2003).

It is recognized that zebrafish skeletal muscle differentiation is regulated by multiple signaling networks (Jackson and Ingham, 2013), some of which are better understood than others. Here, we investigated the regulatory role played by the ubiquitous second messenger, Ca²⁺ (Berridge et al., 2003). Distinct, highly reproducible patterns and frequencies of propagating pan-cellular Ca²⁺ transients have been reported to be required for early myofibrillogenesis in zebrafish SMCs (Brennan et al., 2005; Cheung et al., 2011; Kelu et al., 2015, 2017), but not for the migration or elongation of these cells (Brennan et al., 2005). This highly reproducible series of propagating Ca²⁺ transients are first generated between ∼17–20 hpf in both the cytoplasmic and nuclear domains of differentiating SMCs. These are suggested to be associated with excitation–transcription coupling and subsequent myofibrillar organization (Brennan et al., 2005; Cheung et al., 2011), and they coincide with early spontaneous coiled contraction movements of the embryonic trunk (Saint-Amant and Drapeau, 1998). A second phase of SMC-generated pan-cellular Ca²⁺ transients, which starts at ∼23 hpf, is restricted to the cytoplasmic domain and is proposed to be associated mainly with excitation–contraction coupling (Cheung et al., 2011). These later transients coincide with the initiation of the evoked swimming behavior that begins at ∼26 hpf (Saint-Amant and Drapeau, 1998). The bulk of the Ca²⁺ generating these early myogenic transients is released mainly from the sarcoplasmic reticulum (SR) of SMCs via a combination of inositol (1,4,5)-trisphosphate receptors and ryanodine receptors (Brennan et al., 2005; Cheung et al., 2011). However, it has recently been reported that these Ca²⁺ signals are initiated via the action of two-pore channel (TPC) type 2 (TPC2, encoded by tpcn2 in zebrafish) (Kelu et al., 2015, 2017), perhaps via a proposed triggering mechanism (Zhu et al., 2010a). Indeed, the knockdown or pharmacological inhibition of TPC2 completely eliminates both phases of the propagating Ca²⁺ transients in SMCs and results in a disruption of myofibrillogenesis and subsequent motility (Kelu et al., 2017).

TPCs are members of the voltage-gated ion channel superfamily (Calcraft et al., 2009; Zhu et al., 2010a; Galione, 2019). To date, three TPC isoforms (TPC1, TPC2 and TPC3) have been identified and these are localized on different components of the endolysosomal trafficking pathway. TPC1 (encoded by tpcn1 in zebrafish) and TPC3 (encoded by tpcn3 in zebrafish) are predominantly (but not exclusively) expressed on endosomes, whereas TPC2 is localized primarily on lysosomes (Calcraft et al., 2009). TPCs are described as being master regulators at the intersection of endolysosomal membrane trafficking (Marchant and Patel, 2015; Grimm et al., 2017; Vassileva et al., 2020). However, compared to TPC2-mediated Ca²⁺ signaling, less is known about the regulatory role played by TPC1 during skeletal muscle myogenesis. Addressing this question was the rationale for this work, which used zebrafish as a familiar and well-suited model system. It is known that heterotypic organelle interactions take place between endolysosomes and other membrane-bound organelles, including the endoplasmic reticulum (ER) (Kinnear et al., 2004; Kilpatrick et al., 2013), nucleus (Chaumet et al., 2015), and trans Golgi-network (Jung et al., 2012) as well as the plasma membrane itself (Klumperman and Raposo, 2014). In addition, homotypic and heterotypic interactions occur between endolysosomes, and these are suggested to be responsible for key processes associated with trafficking, the maintenance of cellular homeostasis, autophagy and the regulation of developmental events (Dowling et al., 2008; Lin et al., 2015; Grimm et al., 2017). Reports indicate that TPC-mediated Ca²⁺ signaling might play a significant role in regulating these various organellar membrane interactions (Zhu et al., 2010b; Kilpatrick et al., 2017; Davis et al., 2020). The importance of TPC activity in these events is clearly indicated via attenuation of their function, which results in a diverse range of disease phenotypes (Patel and Kilpatrick, 2018; Jin et al., 2020; Moccia et al., 2021).

As TPC-related isoform-specific roles have previously been reported from experiments using Xenopus oocytes and cultured mammalian cells (Lin-Moshier et al., 2014), we hypothesized that Ca²⁺ release from TPC1 might have independent and distinct developmental and/or regulatory functions from TPC2 during myogenesis in SMCs. Thus, here we used a combinatorial approach (by applying genetic, molecular, pharmacological, and various live- and fixed-imaging techniques in both intact embryos and cultured SMCs) to explore the role of TPC1-mediated Ca²⁺ signaling during SMC myogenesis at ∼24 hpf and ∼48 hpf. We report that unlike the pan-cellular Ca²⁺ transients triggered by TPC2, TPC1 generates localized non-propagating (LNP) Ca²⁺ transients that regulate essential endolysosomal trafficking and organellar membrane contact site activity at the sarcolemma–vertical myoseptal interface. These events appear to be essential for forming and then maintaining the MJs once the SMCs begin to contract, as well as for terminating the elongation of these cells, which prevents them from crossing the myoseptal boundaries. As attenuation of TPC1-mediated Ca²⁺ release results in phenotypes that resemble various dystroglycanopathies, we discuss how TPC1 might represent a possible target for therapeutic intervention.

To investigate the potential role of TPC1 in SMC attachment and stabilization at the MJ (Fig. S1), we attenuated the expression of tpcn1 via the use of morpholino oligonucleotides (MOs; Figs S2,S3) and by generating a mutant line via CRISPR/Cas9 mutagenesis (Fig. S4). Splice- and translation-blocking tpcn1 MOs (tpcn1-S-MO and tpcn1-T-MO, respectively) were designed, and their efficacy characterized by reverse transcription polymerase chain reaction (RT-PCR) analyses of tpcn1 transcripts and the attenuation of TPC1–EGFP fluorescence, respectively (Figs S2,S3). Our MO results were compared with p53 (also known as tp53)- and standard control (SC) MOs, and rescue experiments were performed for the tpcn1-T-MO using a tpcn1 mRNA construct (tpcn1-mRNA). The latter indicated the specificity of MO-based knockdown.

We characterized two gene loss-of-function models by generating a stable tpcn1 mutant knockout line carrying a 16 bp insertion in exon 5 (named tpcn1^dhkz101 in accordance with the ZFIN zebrafish nomenclature guidelines; http://zfin.org), and a first generation (F₀) tpcn1 mosaic mutant (hereafter referred to as a tpcn1 ‘E8 crispant’) using a single guide RNA targeting exon 8 (Fig. S4). The knockdown of tpcn1 transcripts in homozygous tpcn1 mutants and the effective generation of indels in tpcn1 were validated by quantitative RT-PCR (qRT-PCR), Sanger sequencing and high-resolution melting analysis (HRMA; Fig. S4).

We investigated whether tpcn1 knockdown or knockout might affect any gross aspects of SMC development. Fig. 1B–H shows the gross morphology of [and the localization of myosin heavy chain (MyHC) and F-actin in] the skeletal musculature of tpcn1 morphants, mutants and E8 crispants at ∼24 hpf (in somites 7 and 8) and ∼48 hpf (in somites 7 to 11). [Note, the specific isotypes of myosin heavy chain recognized by the antibody F59, as used here, in zebrafish are unknown; however, this antibody is widely used to label SMCs in early embryos (Devoto et al., 1996; Kelu et al., 2015).] Quantification of the morphological changes observed at these times is shown in Fig. 2.

At ∼24 hpf, the tpcn1 morphants exhibited a significant decrease in the number of SMCs and in the SMC myofiber width and length when compared with the various controls (untreated, SC-MO or p53-MO alone; Figs 1Bai,Cai,Dai and 2Aa–Ac). The morphants also exhibited more U-shaped somites instead of the usual chevron-shaped somites seen in the controls (compare Fig. 1Cai and Dai with Fig. 1Bai, and see Fig. 2Ad). The somite shape was quantified by measuring the somite intersection angle. This and the other dimensions measured are shown in Fig. 2Aa*. The abnormalities observed in the tpcn1-T-MO (plus p53-MO) group were partially rescued when embryos were co-injected with tpcn1-T-MO (plus p53-MO) and tpcn1-mRNA, such that the SMC myofiber length and width, and somite intersection angle were all partially restored (Figs 1Eai,Eaii and 2A). Similarly, injection of tpcn1-T-MO (plus p53-MO) also significantly inhibited the spontaneous coiling frequency of embryos between ∼17.5 hpf and ∼23.5 hpf, when compared with the p53-MO controls, whereas co-injection of tpcn1-T-MO (plus p53-MO) and tpcn1-mRNA partially rescued the inhibitory effect of the tpcn1-T-MO (Fig. S5).

At ∼48 hpf, the tpcn1 morphants exhibited more pronounced abnormalities at the MJ. For example, SMC boundary crossings (in magenta in Fig. 1Cbi,Dbi and quantified in Fig. 2B) were observed in ∼40–52% of the morphants. SMC detachment from the MJ was also observed in the morphants (Figs 1Cbii,Dbii and 2C). This is despite the somite boundaries forming normally in the tpcn1 morphants, as shown in the representative examples at ∼17 hpf (Fig. S6A). At ∼48 hpf, in somites 7–11, SMC detachment ranged from 1 to 11 detachments per somite (Fig. 2C), occurring at approximately equivalent frequencies from the posterior and anterior ends, and in the dorsal and ventral regions of the somite block (Fig. S6B). Co-injection of tpcn1-T-MO (plus p53-MO) and tpcn1-mRNA also partially rescued embryos at ∼48 hpf, as a lower incidence of myotome boundary crossings and SMC detachments from the MJ were observed (Fig. 2B,C).

In the heterozygous and homozygous tpcn1 mutant embryos, there was some recapitulation of the MO-induced phenotypes at ∼24 hpf, such that there was a decrease in the number of SMCs, and in the SMC myofiber length and width, as well as an increase in the somite intersection angle (Figs 1Fai,Faii,Gai,Gaii and 2A–C). However, the number of embryos displaying boundary crossings or SMC detachments at ∼48 hpf was considerably lower, such that only 1 of 14 tpcn1^+/− embryos produced an SMC myotome boundary-crossing phenotype and only 2 of 15 tpcn1^−/− embryos showed SMC detachments (Fig. 1Fbi,Gbi). The function of tpcn1 was also investigated in F₀tpcn1 E8 crispants (Fig. 1H). At ∼24 hpf, there was a decrease in the number of SMCs, and in the SMC myofiber length (but not width), as well as an increase in the somite intersection angle (Fig. 2A). In addition, at ∼48 hpf, ∼19% of the tpcn1 E8 crispants exhibited SMC myotome boundary crossing defects (Fig. 2B), and ∼38% exhibited between 1 and 9 SMC detachments per somite in somites 7–11 (Fig. 2Cd).

Fig. 3 shows the localization of acidic organelles (including endolysosomes) in the trunk musculature of intact embryos stained with LysoTracker Red (referred to hereafter as LysoTracker) at ∼24 hpf. Distinct accumulations of LysoTracker-labeled puncta were shown to be localized adjacent to the MJ (Fig. 3Aa). Co-labeling embryos with LysoTracker and BODIPY FL (to reveal the SMC sarcolemma) showed that LysoTracker-labeled puncta were localized at the ends of SMCs adjacent to the MJ (Fig. 3Ab). When embryos were injected with tpcn1–tdTomato mRNA and then labeled with BODIPY FL, TPC1-decorated puncta were observed to be closely associated with the MJ (Fig. 3Ba,Bb). In addition, when embryos were injected with tpcn1–EGFP mRNA and then labeled with LysoTracker, some LysoTracker-labeled puncta also exhibited TPC1–EGFP fluorescence, as seen by the yellow labeling (Fig. 3Ca,Cb).

TPC1–EGFP-expressing embryos labeled with LysoTracker were also used to investigate the dynamics of the TPC1-decorated vesicles in the vicinity of the MJ (Fig. 3D; Movie 1). TPC1–EGFP and LysoTracker-labeled puncta were tracked for 2 min using confocal microscopy (Fig. 3Da–Dcii). The TPC1–EGFP-labeled puncta were shown to move in both anterograde and retrograde directions, to and from the sarcolemma at the MJ. Although most of the puncta traveled distances of 0–0.72 µm, a small proportion of the TPC1–EGFP puncta traveled up to ∼1.44 µm within this time-lapse series (Fig. 3Ddi). In addition, most of the puncta traveled at velocities of 0–0.8 µm/s (Fig. 3Ddii) although some moved more quickly, at velocities greater than 3.0 µm/s (Fig. 3Db,Dc). For example, as shown in Fig. 3Dbi and Dbii, a TPC1-decorated vesicle (black arrowhead) traveled in a retrograde direction at ∼3.48 µm/s to an acidic organelle (white arrowhead), at which time an overlap in their fluorescence (white arrow) was observed. Later in this time-lapse series, another TPC1–EGFP-bearing vesicle was observed ‘pinching off’ from a cluster of TPC1–EGFP- and LysoTracker-positive organelles and moving in an anterograde direction at ∼2.15 µm/s toward the MJ (Fig. 3Dci,Dcii). The schematic in Fig. 3E summarizes the reported differences in pH, localization of TPC1 (and TPC2) and luminal Ca²⁺ concentration in the different components of the endolysosomal pathway relative to the MJ, as well as the distribution of cortical F-actin, polarity of microtubules and direction of travel of microtubule motors [image modified with permission from John Wiley and Sons (Morgan and Galione, 2014)]. Fig. 3F shows the localization of cortical F-actin (white arrowheads) but not sarcomeric MyHC in the vicinity of the MJ.

To identify which members of the endolysosomal system were decorated with TPC1 in SMCs, embryos were co-injected at the one-cell stage with tpcn1–EGFP mRNA and a plasmid encoding mKate2-tagged Rab5c, Rab11ba, Lamp1 (also known as Lamp1b) or Rab32b (Fig. 4Aai–Adii; Hall et al., 2020). Rab proteins are known to regulate key steps during membrane trafficking, including the formation, transport and fusion of vesicles (Hall et al., 2020). In addition, different Rabs (along with Lamp1) are expressed by (and can therefore be used to identify) specific components of the endolysosomal system. For example, Rab5c is present on (and can be used to help identify) early endosomes, Rab11ba is present on recycling endosomes, Rab32b is present on late endosomes and Golgi-derived vesicles, and Lamp1 is present on lysosomes and late endosomes (Hall et al., 2020). Some embryos were co-injected with tpcn1–EGFP and tpcn1–tdTomato mRNAs as positive controls. Primary cultured cells were prepared at ∼48 hpf, and the colocalization of TPC1–EGFP with these endolysosomal markers (or with TPC1–tdTomato) at the MJ of SMCs was determined. SMC primary cultures were used for several experiments because cells grown in culture have a better fluorescence signal-to-noise ratio than intact embryos. We showed that Lamp1 and Rab5c were the markers that correlated most strongly with TPC1 distribution at the end of cells, whereas Rab11 and Rab32 overlapped with TPC1 fluorescence to a lesser extent (Fig. 4Ae). By co-injecting embryos with tpcn1–EGFP mRNA and mKate2–KDEL DNA (as a marker of the SR) prior to primary cell culture, we showed that TPC1-decorated organelles were also closely associated with the SR throughout cultured SMCs (Fig. 4Ba), including the end regions that would have formed MJs in intact embryos (white arrowheads, Fig. 4Bb).

We also showed close associations between TPC1-decorated organelles and microtubules (Fig. 4C). Embryos were injected at the one-cell stage with tpcn1–tdTomato mRNA, and then the primary cultured SMCs were incubated with Tubulin Tracker Green. The localization of TPC1 puncta in relation to the distribution of microtubules is shown in a representative SMC (Fig. 4Ca,Cb). Treatment of TPC1–EGFP-expressing SMCs with nocodazole, which disrupts the polymerization of microtubules, led to a reduction in the velocity of the TPC1 puncta at the cell end (Fig. 4Cc–Cf).

Imaging SMCs expressing TPC1–tdTomato at super resolution allowed us to resolve the pleiotropic morphology of the dynamic TPC1-decorated endolysosomes (Fig. 4Da; Movie 2). These endolysosomes moved displacement distances of up to ∼1.8 µm to and from the cell ends (Fig. 4Dbi) and exhibited a range of areas up to ∼0.95 µm², although most were less than ∼0.2 µm² (Fig. 4Dbii). In the example shown (Fig. 4Da), endolysosomes appear as small, large or ramified puncta. The time series in Fig. 4Daii–Daix shows a small endolysosome traveling in an anterograde direction toward the end of the cell; as this endolysosome moved, it became larger. Over the 4 s duration of the time series, the vesicle traveled with a mean velocity of ∼1.38 µm/s (Fig. 4Dax); however, the velocity decreased as the vesicle approached the end of the cell (Fig. 4Daxi).

We investigated the effect of tpcn1 knockdown on the distribution and morphology of endolysosomes at the MJ in intact ∼24 hpf embryos by visualizing LysoTracker fluorescence (Fig. 5). In embryos injected with tpcn1-S-MO or tpcn1-T-MO (both plus p53-MO; Fig. 5Ca,Da), and in tpcn1^−/− mutant embryos (Fig. 5Ga), the normal distribution and size of LysoTracker-stained puncta were compromised when compared with those in control (untreated, p53-MO-injected or wild-type ABTU strain) embryos (Fig. 5Aa,Ba,Fa). Line-scan analyses of LysoTracker fluorescence across the MJ (yellow lines in Fig. 5Aa–Ga) showed that in the controls, most of the LysoTracker-labeled puncta resided within ∼10 µm on either side of the MJ (Fig. 5Ab,Bb,Fb). In contrast, in the tpcn1 morphants and mutants, the puncta were far more diffusely distributed (Fig. 5Cb,Db,Gb) and considerably larger (see yellow arrowheads in Fig. 5Ca,Da,Ga). However, a similar distribution and size of puncta as the controls was observed when tpcn1-T-MO plus p53-MO embryos were co-injected with the rescue tpcn1-mRNA (Fig. 5Ea,Eb). We quantified the distribution of the puncta by calculating the ratio of the number of puncta concentrated at the MJ compared to those in the total imaged area (as shown in Fig. 5Hai). The data indicate that in the tpcn1-T-MO-injected embryos the ratio was significantly reduced when compared with the p53-MO controls, but that these effects were reversed by the injection of the rescue mRNA (Fig. 5Haii). We also quantified the area of individual MJ-associated puncta (Fig. 5Hb) and showed that MO-mediated knockdown of tpcn1 led to these puncta being significantly larger, and co-injection with tpcn1-mRNA resulted in them resembling the puncta in the controls. In addition, the MJ-associated puncta were significantly larger in the tpcn1^−/− mutants compared with those in the wild-type ABTU controls.

Fig. 6 shows the localization of β-dystroglycan (encoded by dag1) at the MJ in intact tpcn1 morphants, E8 crispants or following treatment with the vacuolar proton pump (V-H⁺-ATPase) inhibitor bafilomycin A1 at ∼24 hpf. The images in Fig. 6A–I are projected confocal stacks showing lateral views of somites 8 and 9, and the plots in Fig. 6A–I show the changes in fluorescence intensity along line-scans taken in the dorsal, midline and ventral regions of these somites, as indicated by the yellow lines in Fig. 6Aa. The results indicate that β-dystroglycan was still partially localized in the dorsal, midline and ventral regions of MJs of the tpcn1 morphants but, compared with the untreated and p53-MO controls, the relative level of MJ localization was reduced (compare Fig. 6C,D with Fig. 6A,B). However, when tpcn1-T-MO (plus p53-MO) was co-injected with the rescue tpcn1-mRNA, the level of β-dystroglycan localization more closely resembled that of the controls (compare Fig. 6E with Fig. 6A,B). The tpcn1 E8 crispants and bafilomycin A1-treated embryos also exhibited a disrupted localization of β-dystroglycan at the MJs when compared with the Cas9 and DMSO vehicle controls, respectively (Fig. 6F–I). The number of puncta within a region of interest placed in the medial region of somite 8 was quantified in the morphants, E8 crispants and drug-treated embryos (Fig. 6Ja,Jb). Our data shows that the tpcn1 morphants, E8 crispants and bafilomycin A1-treated embryos all exhibited a significant increase in the percentage area of β-dystroglycan-labeled puncta in a mid-somite region of interest when compared with their respective controls. However, when tpcn1-T-MO (plus p53-MO) was co-injected with the rescue tpcn1-mRNA, the percentage area was significantly lower than when tpcn1-T-MO (plus p53-MO) was injected alone (Fig. 6Jc).

We investigated whether TPC1 might play some role in generating the pan-cellular Ca²⁺ signals that occur during early zebrafish development. Therefore, we assessed the effect of MO-mediated tpcn1 knockdown on the three sequential SMC-generated Ca²⁺ signaling phases: signaling period 1 (SP1); quiet period (QP); and signaling period 2 (SP2; Cheung et al., 2011). We used transgenic Tg(α-actin:aeq) zebrafish embryos, which express apoaequorin (the protein component of the bioluminescent Ca²⁺ reporter aequorin) specifically in the skeletal muscle cells. Active aequorin was then reconstituted by incubating the embryos in a solution containing the apoaequorin cofactor, coelenterazine (Shimomura, 1997; Cheung et al., 2011). In these experiments, Tg(α-actin:aeq) embryos were injected at the one- to four-cell stage with p53-MO alone, or with p53-MO and either tpcn1-T-MO or tpcn1-S-MO. As shown in Fig. 7, when tpcn1 expression was knocked down, distinct patterns of SP1, QP and SP2 were still observed, although there was a delay (∼60–90 min) in the start of SP1 (compare the signal start times in Fig. 7Ba and Ca with Fig. 7Aa). However, when the data were adjusted for the start of the SP1, tpcn1 knockdown had no significant effect on the pattern or frequency of the pan-cellular Ca²⁺ signals when compared with the p53-MO controls (Fig. 7Ab,Bb,Cb).

We investigated TPC1-mediated Ca²⁺ release by fusing TPC1 in-frame with a low-affinity, high dynamic range, genetically-encoded Ca²⁺ indicator, G-GECO1.2, to generate TPC1–G-GECO (TPC1–GG; Fig. 8A). The tpcn1–GG mRNA was injected into embryos at the one-cell stage, and primary cultures were then prepared at ∼48 hpf. Co-injection of tpcn1–GG mRNA with mKate2–Lamp1 DNA allowed us to simultaneously assess late endosome and lysosome motility, morphology and interactions alongside Ca²⁺ signaling events in cultured SMCs (Fig. 8Ba). A time series of confocal images of a representative cultured SMC expressing TPC1–GG and mKate2–Lamp1 demonstrated LNP Ca²⁺ events associated with endolysosomal dynamics that occur during possible membrane contact events (Fig. 8Bb–Bg). To confirm the validity of our TPC1–GG reporter construct, SMCs were exposed to extracellular Ca²⁺ via treatment with ionomycin and CaCl₂ (black arrows in Fig. 8Ca, and Fig. 8Cb–Cd) such that the TPC1–GG sensor was saturated, as seen by the yellow color in Fig. 8Cd, and via the elevated relative G-GECO1.2 signal intensity (%G-GECO1.2_max) shown in the line graph (Fig. 8Ca). The number of detectable TPC1–GG-generated LNP Ca²⁺ fluorescence events was quantified at the cell ends as a percentage of the total number of TPC1–GG detected events recorded during the entire time-lapse series (i.e. both before and after the addition of ionomycin and CaCl₂; Fig. 8Da). The Ca²⁺-specificity of the TPC1–GG sensor was demonstrated by buffering Ca²⁺ levels in SMCs using cell-permeable BAPTA-AM or dimethyl-BAPTA-AM (DMB-AM), whereas the nicotinic acid adenine dinucleotide phosphate (NAADP)-specificity of the sensor was demonstrated via treatment with the NAADP antagonist, trans-Ned-19. Our data showed that pretreatment of cells with BAPTA-AM, DMB-AM or trans-Ned-19 significantly attenuated the number of TPC1–GG puncta at the cell ends when compared with the number of puncta in untreated cells (Fig. 8Da). However, significantly more puncta were observed following treatment with DMB-AM (K_d=0.44 µM) than with BAPTA-AM (K_d=0.70 µM; Pethig et al., 1989). Application of a mutated TPC1–GG (TPC1^L264P–GG) also significantly reduced the number of puncta recorded. Furthermore, the mKate2–Lamp1-labeled puncta were significantly larger following treatment with trans-Ned-19, BAPTA-AM or DMB-AM compared with those in untreated cells. However, in this case there was no significant difference when comparing the two Ca²⁺ buffers (Fig. 8Db). TPC1^L264P–GG also induced slight enlargement of these puncta, but the data were not significantly different from those obtained for the TPC1–GG control.

Considerable interest is building with regards to the intracellular signals generated by Ca²⁺ release from endolysosomes via the various Ca²⁺-release channels (including TPCs, MCOLN and P24X channels), that decorate their surface (Dong et al., 2010; Morgan and Galione, 2014; Huang et al., 2014; Marchant and Patel, 2015; Vassileva et al., 2020). This is in part due to the role these channels play in the regulation of endolysosomal trafficking and membrane contact events within cells, which in turn contributes to a plethora of cellular functions, including differentiation (Kelu et al., 2017), cell division (Horton et al., 2015) and membrane repair (Reddy et al., 2001). These Ca²⁺-release channels are distributed heterogeneously among different endolysosomes. They can also be activated by family-specific endogenous agonists, generate distinct Ca²⁺ signaling signatures, possess varying domains of cytoplasmic influence, and have either specific or general targets (Raffaello et al., 2016; Yang et al., 2019). As such, they present a complex signaling interactome with multiple regulatory functions during development, physiology, homeostasis and disease (Zhu et al., 2010b; Marchant and Patel, 2015; Jin et al., 2020; Vassileva et al., 2020).

Embryos with attenuated tpcn1 expression appear to phenocopy reported zebrafish muscle disease models (Goody et al., 2017) with regards to (1) SMC detachment from the MJ (Fig. S1) and (2) SMC boundary crossing (Figs 1,2). In tpcn1 morphants, these myoseptal and SMC abnormalities were most obvious at ∼48 hpf, suggesting that the integrity of the MJs might be progressively challenged during myotome development in these embryos. We propose that following the onset of more intense contractile activity during the hatching phase starting at ∼48 hpf (Saint-Amant and Drapeau, 1998), the forces generated might expose weaknesses in the SMC myoseptal attachments. This results in the abnormal SMC detachment phenotypes we observed in the morphants, mutants and E8 crispants. Our tpcn1 morphant phenotypes resemble other zebrafish models following knockdown or knockout of cell–ECM adhesion components such as integrin α7 or dystroglycans (Gupta et al., 2011; Goody et al., 2012). Given that endolysosomal trafficking in ECM formation and remodeling is well documented (Moreno-Layseca et al., 2019), this suggests a possible role for TPC1-mediated Ca²⁺ release in regulating the trafficking of these essential ECM components to the forming MJ. This is followed by the maintenance of junction integrity and stability once trunk contractions begin.

The myotome boundary-crossing phenotype we and others have observed in zebrafish morphants and mutants, resembles that of various related human myopathies (Henry et al., 2005; Goody et al., 2017). Some of these mutated genes express key components of the ECM and the DAPC that are essential for the formation of the MJ. We suggest, therefore, that Ca²⁺ released via TPC1 might be involved in regulating membrane contact events essential for trafficking and depositing required components to the MJ, and as a result contribute to the termination of SMC elongation, described as fiber ‘capture’ (Goody et al., 2017). We noted a paucity of abnormal phenotypes in the tpcn1 mutants compared with the morphants. Similar discrepancies between other mutants and morphants have been reported before in zebrafish (Rossi et al., 2015). However, we demonstrated a closer reproducibility in the phenotypes of the tpcn1 F₀ E8 crispants and morphants (Figs 1,2).

The vertical myosepta form from the early somite boundaries, which develop in the paraxial mesoderm during the segmentation stage (Kimmel et al., 1995; Henry et al., 2000). We found that these somite boundaries still developed in the tpcn1 morphants (Fig. S6A). These observations are supported by a previous report showing that no regular Ca²⁺ transients are generated during the formation of the rostral–caudal somite boundary (Leung et al., 2009). Furthermore, local uncaging of a photolabile Ca²⁺ buffer (Diazo-2) in the paraxial mesoderm just prior to the onset of somite formation does not inhibit the formation of the rostral–caudal somite boundary, although the extension of the lateral somite boundary is abnormal compared with that of controls (Leung et al., 2009). These data suggest that rostral–caudal somite boundary formation does not require a Ca²⁺ transient generated by release via any Ca²⁺ channel (including TPC1). We suggest, therefore, that the myotome boundary-crossing phenotype we observed at 48 hpf when tpcn1 expression was attenuated is not due to the absence of a rostral–caudal somite boundary. Rather, we propose that the lack of capture of SMCs is due to the absence of an essential MJ component (or components) that are delivered to the forming MJ in a TPC1-dependent manner.

We have previously shown that in tpcn2 morphants and mutants, despite the pan-cellular SP1 and SP2 Ca²⁺ transients being completely inhibited and the myofibrillar structure disrupted, SMCs remain attached to the MJ and no SMC myotome boundary crossings are observed (Kelu et al., 2017). However, a somewhat opposite situation was observed in the tpcn1 morphants, mutants and E8 crispants, where the most prominent abnormal features were the detachment of some SMCs at the MJ and myotome boundary crossover by others. Furthermore, myofibrillogenesis was also relatively normal (Fig. 1). A somewhat similar situation has been reported following complete inhibition of pan-cellular cytosolic Ca²⁺ signals in macrophages; localized TPC-dependent Ca²⁺ nanodomains are still generated and the cells are able to phagocytose IgG-coated fluorescent beads (Davis et al., 2020). This supports the suggestion that TPC-isoform-specific Ca²⁺ signals regulate distinct cellular activities.

In addition to TPCs, other Ca²⁺-release channels, including MCOLN and P2X4 channels, are expressed in endolysosomes (Dong et al., 2010; Cheng et al., 2014; Huang et al., 2014; Li et al., 2017), and orthologs of these genes have been found in zebrafish (Diaz-Hernandez et al., 2002; Li et al., 2017). However, when two mcoln1 genes were knocked out in zebrafish, although there was a progressive abnormal accumulation of autophagosomes in the skeletal muscle fibers, no effect was reported regarding sarcomeric assembly and myofibrillogenesis, or the localization of vinculin (a key cell–matrix adhesion protein) at the MJ (Li et al., 2017). Furthermore, no SMC detachment or myotome boundary crossing was reported from the mcoln1 mutants. This is unlike our tpcn1 morphants and mutants, where SMC detachments and boundary crossing were observed (Fig. 1). The findings reported by Li et al. (2017) indicate that in mcoln1 mutants some aspects of endolysosomal trafficking and membrane contact activity are normal whereas others are not. These results suggest that the Ca²⁺-signaling signature generated via MCOLN1 might be distinct and thus have different functions than Ca²⁺ released via TPC1 in SMCs.

The localization of endolysosomes near the SMC MJs in intact zebrafish (Fig. 3A,B), resembles that reported in isolated rat myotubes (Kaisto et al., 1999). The range of velocities recorded for TPC1-labeled vesicles undergoing transport away from or towards the MJ (Fig. 3D) is within the range of velocities previously reported for dynein-mediated retrograde vesicle transport (Granger et al., 2014) and kinesin motor family-mediated anterograde vesicle transport (Arpağ et al., 2014), respectively.

Microtubules (in conjunction with plus-end- or minus-end-directed motors) are known to play a major role in endolysosomal trafficking (Granger et al., 2014). The visualization of microtubules in cultured SMCs and treatment with nocodazole showed that this is also the case with regards to TPC1-labeled vesicles (Fig. 4C). Our results indicate that at the ends of SMCs, TPC1-labeled vesicles move with variable velocity, but slow as they approach the MJ (Fig. 4D). In other systems it has been proposed that this might be due to elements of the cytoskeleton impeding the movement of trafficking vesicles, resulting in them slowing and/or pausing at microtubule intersections (Zajac et al., 2013; Verdeny-Vilanova et al., 2017). Furthermore, in the cortical region of cells, the anterograde microtubule-based transport system transfers trafficking vesicles to the actin cytoskeleton for final delivery to the plasma membrane (Granger et al., 2014). This same process then acts in reverse to pass endocytosed vesicles to the microtubule system for retrograde transport. In SMCs, elements of the cortical actin cytoskeleton are also responsible for attaching terminal sarcomeres (via dystrophin) to the DAPC and ultimately to laminins in the ECM (Ervasti and Campbell, 1993). This terminal actin structure can be clearly seen in Figs 1Baii and 3F. We suggest, therefore, that at the ends of the SMCs near the MJ, a combination of vesicle size, intersecting microtubules and the cortical actin cytoskeleton act together to reduce the velocity of vesicles in either direction.

The introduction of tpcn1-S-MO or tpcn1-T-MO had a significant deleterious effect on both the distribution (Fig. 5Haii and size (Fig. 5Hb) of endolysosomes. However, in the case of the tpcn1-T-MO, both aspects could be rescued via the introduction of a tpcn1-mRNA, indicating the specificity of the MO-based knockdown. Our observations regarding the distribution and size of endolysosomes are consistent with other reports where the expression of tpcn1 was attenuated (Kilpatrick et al., 2017), and the morphology and localization of endolysosomes were disrupted (de Araujo et al., 2020). Thus, our new data in zebrafish support a previously proposed role for TPC1 in the normal functioning of the endolysosomal trafficking network (Vassileva et al., 2020).

We co-expressed zebrafish TPC1 with a variety of endolysosomal markers (Castonguay et al., 2017; Hall et al., 2020) and showed that at the end of SMCs near the MJ, TPC1 appeared to populate the membranes of early endosomes and of late endosomes and lysosomes (Fig. 4Ae). These findings support previous reports indicating the broad distribution of TPC1 in the endolysosomal system (Morgan and Galione, 2014). In addition, as reported from other cell types (Raiborg et al., 2015), TPC1-labeled endosomes were closely associated with the SR throughout SMCs. They were especially prevalent around the nucleus of primary cultured SMCs (Fig. 4Ba) but were also found at the ends of these cells, which would normally form the MJ in vivo (Fig. 4Bb). Such endolysosomal–ER and endolysosomal–SR contact sites are currently the subject of considerable interest with regards to the maturation of endolysosomes and intraorganellar signaling (Cremer et al., 2020; Huang et al., 2020). Furthermore, it has been suggested that localized TPC-mediated Ca²⁺ release might play a key role in regulating such membrane contact site activity (Kilpatrick et al., 2017; Vassileva et al., 2020).

As attenuation of tpcn1 expression resulted in SMC detachment from the vertical myosepta (Fig. 1), we hypothesized that vesicle cargoes transported by endolysosomes to the MJ might include key components of the DAPC and/or the ECM. We chose, therefore, to explore what the disruption of TPC1-mediated Ca²⁺ signaling activity might have on the distribution of β-dystroglycan, a key transmembrane component of the DAPC (Parsons et al., 2002). The mislocalization of β-dystroglycan following TPC1 attenuation in ∼24 hpf embryos (Fig. 6), is comparable to that observed in zebrafish models of dystroglycanopathies (Gupta et al., 2011). As the successful capture and attachment of SMCs at the MJ likely involves trafficking and localization of additional DAPC components, as well as other transmembrane linkage protein complexes and ECM constituents (Goody et al., 2012, 2017), further investigation is required to identify which of these might also be regulated by TPC1-generated LNP Ca²⁺ transients.

We have previously published a series of papers showing that Ca²⁺ released via TPC2 acts to trigger a series of pan-cellular Ca²⁺ signals that are essential for specific myogenic events during the differentiation of SMCs in zebrafish embryos (Kelu et al., 2015, 2017, 2019). When the activity of TPC2 is attenuated in morphant or mutant embryos, or in wild-type embryos treated with trans-Ned-19 or bafilomycin A1, these pan-cellular signals are significantly inhibited and myofibrillar organization is extensively disrupted (Kelu et al., 2015, 2017). These observations suggest that following attenuation of TPC2 activity, there is no compensation by other Ca²⁺-release channels resident in the endolysosomal or other organellar membranes (El-Brolosy and Stainier, 2017). The fact that the pan-cellular Ca²⁺ transients showed no significant attenuation following MO-mediated tpcn1 knockdown (Fig. 7) supports the proposition that TPC isoforms have distinct signaling functions. Thus, although TPC2 generates pan-cellular propagating Ca²⁺ signals via a triggering process, TPC1-mediated Ca²⁺ release generates LNP Ca²⁺ transients similar to those described in other cell types (Ruas et al., 2010; Zhu et al., 2010a; Davis et al., 2020; Jin et al., 2020). Therefore, both TPCs can function within the endolysosomal system of the same SMC to regulate different myogenic processes that respond to distinct Ca²⁺ signaling signatures. Although attenuation of tpcn1 expression had no significant effect on the SP1 and SP2 signaling signature once SP1 was initiated, there was a substantial delay of ∼1 h before the onset of SP1 in the morphants compared with the control embryos (Fig. 7). As both the tpcn1-T- and tpcn1-S-MO were injected into embryos at the one- to four-cell stage, the attenuation of tpcn1 expression might have affected an earlier developmental event requiring TPC1-mediated signaling, resulting in the SP1 and SP2 delay observed. As the delay was not the focus of this paper, however, this suggestion remains essentially speculative and requires additional experimentation to resolve its cause.

Given that the attenuation of tpcn1 expression did not appear to affect the pan-cellular SMC Ca²⁺ signals (Fig. 7), we utilized TPC1–GG (Fig. 8; Zhao et al., 2011), which revealed that TPC1 generated LNP Ca²⁺ transients in SMCs. These findings were further confirmed via the use of TPC1^L264P–GG (to attenuate TPC1 activity) or via treatment of cultured SMCs with the NAADP antagonist trans-Ned-19, or with the Ca²⁺ buffers BAPTA-AM or DMB-AM, all of which significantly reduced the number of GG-detected LNP Ca²⁺ events at the end of cultured SMCs (Fig. 8Da). The fact that treatment with trans-Ned-19 also resulted in a significant increase in the mean area of individual Lamp1-decorated vesicles (Fig. 8Db), supports our finding that following tpcn1 knockdown there is an increase in the area (and altered morphology) of LysoTracker-stained puncta (Fig. 5Hb). Such a change in vesicle size indicates abnormal endolysosomal homeostasis (de Araujo et al., 2020). When comparing the effect of the same concentration of DMB-AM and BAPTA-AM, which have de-esterified dissociation constants (K_d) of 0.44 µM and 0.70 µM, respectively (at pH 7.0, 1 mM Mg²⁺ and 300 mM KCl; Pethig et al., 1989), the former had less of an effect on reducing the percentage of localized Ca²⁺ signaling events than the latter (Fig. 8Da). This suggests that the changes observed with BAPTA-AM were due to the faster buffering of the localized Ca²⁺ transients rather than a possible inhibitory side-effect of introducing the Ca²⁺ buffer. Similar results regarding the effectiveness of Ca²⁺ buffers with respect to their dissociation constants have been reported from several systems where a localized elevation of cytosolic Ca²⁺ has been found to be a required intracellular signal (Speksnijder et al., 1989; Davis et al., 2020; Guo et al., 2020). These findings suggest that LNP Ca²⁺ release by TPC1 might contribute to required endolysosomal dynamics and membrane contact events in the vicinity of the MJ in SMCs.

Previous reports provide clear evidence of a role for TPCs in endolysosomal trafficking (Marchant and Patel, 2015; Vassileva et al., 2020), and analysis of the TPC interactome has revealed an association between TPCs and Ca²⁺-sensitive proteins including annexins, Rab proteins and SNARE proteins (Moccia et al., 2021). All of these are known mediators of membrane fusion, which regulates contact sites between endocytic vesicles and the ER (Eden et al., 2016). Several other studies have used BAPTA to show that localized TPC-evoked Ca²⁺ signals are likely mediators of endolysosomal trafficking events (Lin-Moshier et al., 2014; Grimm et al., 2017), and it has been proposed that Ca²⁺ release via TPCs provides a localized Ca²⁺ signal that regulates the fusion of closely apposed endolysosomal structures (Marchant and Patel, 2015).

We suggest that LNP TPC1-generated Ca²⁺ release helps regulate (with higher fidelity than pan-cellular TPC2-mediated Ca²⁺ release) key localized aspects of endolysosomal trafficking and membrane contact activity, which plays a critical role in SMC capture and attachment at the MJ. The generation of similar LNP Ca²⁺ events by TPCs has recently been reported to contribute to phagocytosis in macrophages (Davis et al., 2020). This suggests that LNP TPC-generated Ca²⁺ signals at membrane contact sites might be a common feature among different cell types and in different species. It has been proposed that endolysosomes constitute an active signaling interactome required to regulate intracellular trafficking (Marchant and Patel, 2015). We suggest that by expressing a diverse population of Ca²⁺-release channels – each with their own family-specific endogenous agonists, Ca²⁺ signaling signatures and Ca²⁺-sensitive targets – endolysosomes provide an additional level of regulatory complexity with regards to intracellular Ca²⁺ signaling. Thus, extending the elegant ‘cellular Ca²⁺ signaling toolkit’ proposition (Berridge et al., 2003), we suggest that endolysosomes possess their own ‘organellar Ca²⁺ signaling toolkit’, which interacts discerningly with other signaling elements to regulate a variety of cellular processes required for the development, maintenance and function of SMCs.

The AB and ABTU wild-type zebrafish strains were obtained from the Zebrafish International Resource Center (University of Oregon, Eugene, OR, USA) and Prof. Han Wang (Soochow University, Suzhou, China), respectively, whereas the α-actin-apoeaquorin-IRES-EGFP [Tg(α-actin:aeq)] line was previously developed in the Miller laboratory (Cheung et al., 2011). The fish were kept in AHAB systems (Pentair Aquatic Eco-Systems, Apopka, FL, USA) and their fertilized eggs were obtained and maintained at ∼28°C using well-established protocols (Westerfield, 2000). All the procedures used in this study were conducted in accordance with the guidelines and regulations outlined by the Animal Ethics Committee of HKUST, and the Department of Health, Hong Kong.

To detect endogenous Ca²⁺ signals generated during embryogenesis, Tg(α-actin:aeq) embryos were incubated with f-coelenterazine (NanoLight^® Technologies, Pinetop, AZ, USA) to reconstitute active aequorin (Shimomura, 1997). Coelenterazine was prepared and incubated with the embryos as described previously (Kelu et al., 2015). The embryos were placed in grooves made in 1% agarose within individual custom-built imaging chambers and then transferred to a Photon Multiplier Tube (PMT)-based system (Science Wares, Inc., MA, USA) for luminescence detection, as described previously (Cheung et al., 2011; Kelu et al., 2015).

All the MOs (Gene Tools LLC, Philomath, OS, USA) were prepared at a stock concentration of 1 mM in Milli-Q water and stored at room temperature. The expression of TPC1 was attenuated using splice (S)- or translation (T)-blocking tpcn1 MOs. The tpcn1-S-MO was designed by us (with advice from Gene Tools), whereas the tpcn1-T-MO was designed by Dr. John Parrington (University of Oxford, UK). To evaluate the specificity of the tpcn1-T-MO by visualization of TPC1–EGFP fluorescence, embryos were co-injected with 2 ng tpcn1-T-MO plus 2 ng p53-MO into the yolk and ∼150 pg of tpcn1–EGFP mRNA into the blastodisc. The design and synthesis of this fluorescently tagged TPC1 construct is described in the section below titled ‘Design, cloning and injection of fluorescently tagged TPC1 constructs’.

Information about the protocols used to assess the knockdown efficiency of the tpcn1-S-MO is provided in the following sections (see ‘Total RNA isolation and cDNA synthesis’ and ‘Total RNA isolation and RT-PCR to assess the knockdown efficiency of the tpcn1-S-MO’).

To avoid possible off-target effects of the tpcn1 MOs on p53 activity, each was co-injected with a p53-MO (Robu et al., 2007). The p53-MO (injected alone) and a standard control MO were also used as specificity controls (Kelu et al., 2015). The MOs were prepared and injected into AB wild-type embryos using methods described previously (Webb and Miller, 2013; Kelu et al., 2015). The sequences of the standard control MO and p53-MO are as described previously (Kelu et al., 2017), whereas those for the tpcn1 MOs were as follows: tpcn1-S-MO: 5′-GTAGCGGATCAGATAATGACCTGCA-3′ and tpcn1-T-MO: 5′-TCCGCCATCCCGCAGAACCAAGACT-3′. The optimal injection dose of each MO was selected by evaluating the phenotypic dose response. Therefore, ∼1 nl of the standard control MO (∼5 ng), p53-MO (∼5 ng), tpcn1-S-MO plus p53-MO (at a 1:1 ratio of each of ∼2.5 ng, ∼5 ng or ∼8 ng), or tpcn1-T-MO plus p53-MO (at a 1:1 ratio of each of ∼1 ng, ∼2 ng or ∼5 ng) was injected into the yolk of embryos at the one- to four-cell stage. In some experiments, embryos were co-injected with 2 ng tpcn1-T-MO plus 2 ng p53-MO and 50 ng or 100 pg tpcn1-mRNA.

Embryos were dechorionated manually at ∼24 hpf or ∼48 hpf using a pair of watchmaker's forceps (No. 5; Regine Switzerland SA, Morbio Inferiore, Switzerland). To prepare cDNA, total RNA was extracted by homogenizing 40 embryos in TRIzol reagent (Ambion, Invitrogen Corp., CA, USA) by trituration through a P200 pipette tip. Samples were kept at −80°C until RNA isolation using the protocol provided with the TRIzol reagent. In brief, RNA was extracted with chloroform and precipitated using isopropanol. To isolate RNA, samples were centrifuged (at 12,000 g for 10 min at 4°C), and the resulting pellets were washed with 75% ethanol, after which they were centrifuged again (at 7500 g for 5 min at 4°C) and then resuspended in nuclease-free water. Subsequently, RNA was subjected to RT-PCR using random primers and the High-capacity cDNA Reverse Transcription kit (Applied Biosystems, Waltham, MA, USA) to synthesize cDNA.

RNA was isolated from 40 whole AB wild-type embryos at ∼24 hpf that were either untreated or injected with the standard control MO, p53-MO, or tpcn1-S-MO plus p53-MO, and cDNA was synthesized as described above. PCR was performed on cDNA using the following primer pairs to assess the outcome of steric inhibition of the tpcn1-S-MO on the pre-mRNA transcript of tpcn1:

(1) A forward (splice-F) and reverse (splice-RI) primer pair was designed to amplify a 319 bp amplicon from wild-type tpcn1-mRNA. If the tpcn1-S-MO led to exon 2 being skipped (which we considered to be the most likely outcome produced by splice-blocking MOs), then the size of this amplicon was expected to be reduced to 123 bp. Splice-F, 5′-CTTATCCTGACCTGGGACGATG-3′ (targeting a region on exon 1); splice-RI, 5′-AGAGGGACAGGAGCATGAGTAAC-3′ (targeting a region on exon 3).

(2) Another reverse primer, targeting a region on exon 2 (splice-RII), was designed to verify the splice-F+splice-RI outcome. If exon 2 was skipped in the aberrantly spliced tpcn1 mRNA, then a reduction in the level of the amplicon (178 bp) produced by the splice-F+splice-RII primer pair would be the expected result. Splice-RII, 5′-GGTATATTGCTGCTTCTTGGAAGTT-3′.

(3) Given that the splice-F+splice-RI primer pair produced a band of greater size by agarose gel electrophoresis, and because the splice-F+splice-RII primer pair generated equivalently strong bands in all the samples, it was deduced that exon skipping did not occur upon injection of the tpcn1-S-MO. For this reason, another reverse primer was designed to detect intron inclusion: splice-RIII, 5′-CCTGAACTGAGTGTGTGTTGG-3′. Splice-RIII targeted a region on intron 2, and the amplicon produced by splice-F+splice-RIII was expected to be 255 bp.

β-actin primers were used as an internal control in these experiments. The PCR cycling parameters comprised 30 s denaturation at 95°C, 30 s annealing at 58°C and 30 s elongation at 72°C. All the amplicons were analyzed using 1% agarose gel electrophoresis.

To validate the effects of the tpcn1-T-MO, embryos were injected with a rescue construct comprising a tpcn1-202 mRNA that lacked the MO recognition site (Eisen and Smith, 2008). To prepare the rescue mRNA, tpcn1-202 was cloned from cDNA using custom primers designed to incorporate silent mutations to avoid MO recognition at the start codon. The primer sequences used for PCR amplification were as follows: rescue-F, 5′-AATTGGTACCATGGCAGATGGGGACGACGAC-3′; rescue-R, 5′-GGCCCTCGAGTTAGTTAATGCTGTTTGTGGCACC-3′. The leader sequences (to optimize restriction enzyme digestion) are underlined; restriction enzyme sites for KpnI and XhoI (New England Biolabs Inc., MA, USA) are shown in bold, and silent mutations are shown in italics. PCR was performed using the Phusion^® High-Fidelity DNA Polymerase (New England Biolabs Inc.) with an initial 30 s denaturing step at 98°C, followed by 35 cycles of a 5 s denaturation step at 98°C and a 75 s annealing and extension step at 72°C. This was followed by a final elongation step at 72°C for 5 min. The amplicons were analyzed via 1% agarose gel electrophoresis. They were then purified using a PCR-M Clean Up System (Viogene BioTek Corp., Taipei, Taiwan) and cloned into the pSP64TNE expression vector (Addgene, MA, USA) using a T4 DNA ligase (Thermo Fisher Scientific, Inc., CA, USA). The results were confirmed by Sanger Sequencing (BGI, Shenzhen, China).

To prepare mRNA, the pSP64TNE-tpcn1-202 plasmid was linearized with SmaI (New England Biolabs Inc.), and in vitro transcription was performed with a mMESSAGE mMACHINE SP6 transcription kit (Ambion). The mRNA was then purified using phenol-chloroform extraction (Sambrook and Russell, 2006) and diluted to ∼200 ng/µl in RNase-free water for storage at −80°C. During rescue experiments, embryos at the one-cell stage were injected in the yolk with ∼2 ng tpcn1-T-MO+∼2 ng p53-MO, after which they were injected in the blastodisc with ∼50 ng or ∼100 pg tpcn1-mRNA.

To generate a tpcn1 knockout zebrafish line and tpcn1 E8 crispants, CRISPR/Cas9-mediated targeted mutagenesis was performed by injection of Cas9 mRNA and sgRNAs against the tpcn1 sequence following established protocols (Varshney et al., 2016). Two sgRNA sequences, targeting regions in exon 5 and exon 8, were chosen using the ZebrafishGenomics UCSC Genome Browser track (http://genome.ucsc.edu/cgi-bin/hgHubConnect) developed previously (Varshney et al., 2015). These targets are known to be common to all three of the known tpcn1 splice variants and are predicted to have as few off-targets as possible. These exons lie within the first of two homologous six-transmembrane subunits (IS1–IS6 and IIS1–IIS6) that comprise the TPC dimer. Specifically, they encode the third transmembrane helices (IS3) within the voltage-sensing domain and a pore domain (IP), respectively.

To synthesize sgRNAs, an oligo-based method was used such that the target-specific oligos and an 80 nucleotide (nt) ‘generic’ DNA oligo were annealed and extended to create a template for in vitro transcription (Varshney et al., 2015). The ‘generic’ DNA oligo encodes the chimeric guide RNA (crRNA:tracrRNA) needed to generate the ‘stem-loop’ structure that is bound by the Cas9 nuclease (Cong et al., 2013).

The sequences of the target-specific oligos are as follows: sgRNA-E5, 5′-GATCACTAATACGACTCACTATAggCTCTTGTGACGCGCAGGTGTTTTAGAGCTAGAAATAGC-3′; sgRNA-E8, 5′-GATCACTAATACGACTCACTATAggCGTATTCGAAGAATCGTGTTTTAGAGCTAGAAATAGC-3′.

An sgRNA targeting the tyrosinase (tyr) gene was used as a positive control to assess the sgRNA design and the quality of the Cas9 mRNA (Varshney et al., 2016); its sequence is as follows: sgRNA-tyr-ctrl, 5′-GATCACTAATACGACTCACTATAggACTGGAGGACTTCTGGTTTTAGAGCTAGAAATAGC-3′.

For these three sequences, protective bases are shown in non-bold italics; the T7 promoter (17 nt) and minimal sequence, ‘gg’, for initiation are in bold; the target sites that precede an NGG protospacer adjacent motif (PAM; 15–18 nt) sequence are underlined (NGG not included in the actual sequence); and a 20-nt sequence complementary to the ‘generic’ guide RNA is in bold italic font. The sequence of the ‘generic’ DNA oligo/universal bottom-strand ultramer (Integrated DNA Technologies, IA, USA) is as follows, with the sequence complementary to the target-specific oligos shown in italics: 5′-AAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAACGGACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAAC-3′.

To synthesize the sgRNAs, an oligo assembly reaction that contained 10 µM of either of the target-specific oligos and the ‘generic’ DNA oligo was annealed and extended using the Phusion^® High-Fidelity DNA polymerase (New England Biolabs Inc.). The cycling conditions were 2 min denaturation at 98°C, followed by a 10 min annealing step at 50°C and a 10 min extension step at 72°C. The annealed oligos were then used as templates for in vitro transcription using the mMESSAGE mMACHINE™ T7 Transcription kit (Ambion), and RNA was purified using phenol-chloroform (as described above). The sgRNAs were run on a 2.5% agarose gel to verify successful RNA synthesis, and then diluted to ∼200 ng/µl in RNase-free water for storage at −80°C.

To synthesize Cas9 mRNA, the pT3TS-nls-zCas9-nls plasmid (Jao et al., 2013) was first linearized using the XbaI restriction enzyme (New England Biolabs Inc.). Digestion was then verified by gel electrophoresis with a 1% agarose gel and purified using the Monarch^® PCR & DNA Cleanup kit. An in vitro transcription reaction was then carried out using the mMessage mMachine T3 Transcription kit (Ambion), and the Cas9 mRNA was purified using phenol-chloroform (as described above). Purified RNA was run on a 1% agarose gel, which produced a band at ∼1.6 kb, indicative of successful Cas9 mRNA synthesis. The RNA was diluted to ∼500 ng/µl in RNase-free water and divided into aliquots for storage at −80°C. Then, ∼100 pg sgRNA per target and 300 pg Cas9 mRNA were injected into the blastodisc of ABTU wild-type embryos at the one-cell stage. Following injection with the tyr sgRNA, the reduced pigmentation in embryos at ∼48 hpf was indicative of successful mutagenesis.

A combination of HotSHOT raw genomic extraction (Meeker et al., 2007) and HRMA (Samarut et al., 2016) was used at several stages in the workflow to establish the tpcn1 mutant line (called tpcn1^dhkz101 in accordance with the ZFIN zebrafish nomenclature guideline; http://zfin.org). These are described as ‘stages 1 to 4’ in the following text. By modifying the source of genomic DNA, HRMA was used to rapidly assess mutagenic activity of the Cas9 mRNA and the tpcn1 sgRNAs in F₀ embryos (stage 1), as well as to identify CRISPR-induced indels in the F₁ adults (stage 2) and to perform genotyping in the F₂ embryos (stage 3) and adults (stage 4).

Genomic DNA was extracted from tpcn1 mutant intact embryos and the caudal fin of adults. For the former, embryos were randomly selected and euthanized at ∼24 hpf, whereas for the latter a small region of the caudal fin (hereafter called a fin clip) was excised from adult fish after anesthetization in Danieau's solution (17.4 mM NaCl, 0.21 mM KCl, 0.12 mM MgSO₄·7H₂O, 0.18 mM Ca(NO₃)₂·4H₂O and 1.5 mM HEPES; pH 7.2) containing ∼0.02% ethyl 3-aminobenzoate methanesulfonate (MS-222) (Westerfield, 2000). The embryos and fin clips were incubated in 50 mM NaOH for 20 min at 95°C, vortexed briefly and cooled to 4°C before the addition of 1/10 volume 1 M Tris-HCl (pH 8.0). To extract genomic DNA from the embryos (at stages 1 and 3), each embryo was placed in the well of a 384-well plate (LightCycler 480 Multiwell Plate 384, white; Roche Molecular Systems, Inc., CA, USA) containing 20 µl NaOH and 2.5 µl Tris-HCl, and then heated to 95°C in a Heratherm™ oven (Thermo Fisher Scientific Inc.). To extract genomic DNA from the fin clips (at stages 2 and 4), these were transferred to 1.5 ml microcentrifuge tubes (Axygen, Corning Inc., NY, USA) and heated to 95°C in a Thermolyne dry bath (Thomas Scientific, NJ, USA). The extracted DNA was either used immediately for HRMA or stored at 4°C. To perform HRMA, the genomic DNA of the sgRNA-injected embryos or adults was used as a template for an EvaGreen-based PCR reaction using the Precision Melt Supermix (Bio-Rad) and a LightCycler 480 instrument (Roche, Basel, Switzerland). Primers to amplify a ∼150–200 bp region surrounding the sgRNA target site of exon 5 or exon 8 were designed using the qPCR settings of Primer3 (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi). The primer sequences are as follows: HRMA-E5-F, 5′-TTGTGTTTGGATTTGGTGCAC-3′; HRMA-E5-R, 5′-CTCCACAGTAACGGCAGTCC-3′; HRMA-E5-F and HRMA-E5-R were expected to produce an amplicon of 192 bp. HRMA-E8-F, 5′-TATCAGCATGCCCCTGTTTG-3′; HRMA-E8-R, 5′-AGCTCGATGGACAGATACACG-3′; HRMA-E8-F and HRMA-E8-R were expected to produce an amplicon of 169 bp. To amplify these loci for HRMA at various steps in the generation of the stable tpcn1 mutant line or tpcn1 E8 crispants, a 10 µl reaction mixture was prepared for each sample, which consisted of 5 µl Precision Melt Supermix, 0.5 µl of each primer (10 µM), 2 µl genomic DNA and 2 µl MilliQ water. The reaction protocol for both PCR and HRMA on the LightCycler 480 instrument was as follows: initial denaturation for 2 min at 95°C, and then 45 cycles comprising a 10 s denaturation step at 95°C, a 30 s annealing step at 60°C, and a 20 s extension step. At the end of these 45 amplification cycles, there was a plate-read step at 72°C. Heteroduplex formation was performed with a 30 s denaturation step at 95°C and a 1 min annealing step at 60°C. Finally, melting and plate-read steps (20 readings/°C) were performed using a temperature ramp from 70°C to 95°C at 0.025°C/s, before finishing with a cooling step to 40°C for 30 s. Melt curve data of all the samples (including ABTU wild-type controls) were exported from the ‘Melt Curve Genotyping’ mode of the LightCycler^® 480 SW 1.5.1 software (Roche, Basel, Switzerland), and uploaded to uAnalyze v2.1 (https://dna-utah.org/ua/uanalyze.html; Dwight et al., 2012). This software was used to normalize the data and plot difference curves for genotyping. Fish with CRISPR-induced indels in exon 5 were identified by analyzing shifts in their melting curves from the wild-type samples. The genotyping results were confirmed by Sanger sequencing the F₁ progeny. To prepare samples for sequencing, PCR was performed on genomic DNA using the Rapid Taq Master Mix (Vazyme, Nanjing, China) and the following primers: HRMA-E5-F, 5′-TTGTGTTTGGATTTGGTGCAC-3′; HRMA-E5-genotyping-R, 5′-TGTTACCAACATCCGGTGAA-3′. These primers were used to amplify a 431 bp region surrounding the sgRNA target site of exon 5. PCR was performed with an initial 3 min denaturing step at 95°C, followed by 35 cycles comprising a 15 s denaturation step at 95°C, a 15 s annealing step at 51°C and a 10 s extension step at 72°C. These cycles were followed by a final elongation step at 72°C for 5 min. The amplicons were analyzed via 1% agarose gel electrophoresis. They were then purified using a Monarch^® PCR and DNA Cleanup kit (New England Biolabs Inc.) and sequenced using the forward primer (HRMA-E5-F).

To identify indels from the sequencing data, the known genomic sequences of the wild-type samples and the potential mutants (containing double peaks in the chromatogram) were analyzed using Poly Peak Parser (http://yosttools.genetics.utah.edu/PolyPeakParser/; Hill et al., 2014). The selected tpcn1 line containing a frameshift mutation (+16 bp) in exon 5 was then bred to generate a homozygous population.

To generate F₀tpcn1 E8 crispants, ∼100 pg E8 sgRNA and ∼300 pg Cas9 mRNA were injected into ABTU embryos at the one-cell stage. At ∼24 hpf and ∼48 hpf, the E8 crispants were fixed, immunolabeled and imaged for phenotypic analysis. Subsequently, genomic DNA was extracted from individual embryos at ∼48 hpf and HRMA was conducted to identify the presence of indels at the tpcn1 locus. The cycling conditions used for HRMA and the subsequent analysis of melt curves were performed as described above.

RNA was isolated from ABTU wild-type embryos as well as from tpcn1 heterozygous and homozygous embryos at ∼48 hpf. Experiments were conducted in triplicate, using 40 randomly selected embryos from one clutch for each. The extracted RNA was treated with TURBO DNase (Invitrogen) for 20 min at 37°C, and then ∼1000 ng was reverse transcribed to produce cDNA, as described above. This was used as a template for amplification by PCR with the Rapid Taq Master Mix (Vazyme). The primers used are as follows: tpcn1-RTPCR-F, 5′-CGCAGAAACTTGAGGCAGATA-3′; tpcn1-RTPCR-R, 5′-CGTGTATCTGTCCATCGAGCT-3′. The amplicons were then separated by gel electrophoresis using 2% agarose gels.

RNA was extracted and treated with TURBO DNase, and then cDNA was synthesized, as described above. qRT-PCR was performed using the iTaq Universal SYBR Green Supermix (Bio-rad Laboratories, Inc., CA, USA) according to the manufacturer's instructions with three biological replicates and three technical replicates using 384-well plates. Two primers against tpcn1 were designed, one targeting the exon 2–exon 3 junction, and the other spanning exons 22 and 23. Gene expression was analyzed using the ΔΔCt method (Livak and Schmittgen, 2001), and the relative level of tpcn1 expression was compared to four reference genes: actb2 (Tang et al., 2007), mob4, lsm12b (Hu et al., 2016) and elfa (also known as eef1a1l2; McCurley and Callard, 2008). The qRT-PCR primers used are as follows: tpcn1-qRTPCR-E2&3-F, 5′-GGAGGGAGAGAATAACGACAAG-3′; tpcn1-qRTPCR-E3-R, 5′-CAGAAGCTCCACCACATAGAA-3′; tpcn1-qRTPCR-E22-F, 5′-ACTGGAGCCGCCTTTATTT-3′; tpcn1-qRTPCR-E23-R, 5′-CTCTGTTCTTCCGGCTGTAAT-3′; actb2-F, 5′-CGAGCTGTCTTCCCATCCA-3′; actb2-R, 5′-TCACCAACGTAGCTGTCTTTCTG-3′; mob4-F, 5′-CACCCGTTTCGTGATGAAGTAC-3′; mob4-R, 5′-CAGCTCAGGCATCGCTTTC-3′; lsm12b-F, 5′-GTTGTCCCAAGCCTATGCAATC-3′; lsm12b-R, 5′-TGATGTTCTTCTCCTGCCATTTAC-3′; elfa-F, 5′-CTTCTCAGGCTGACTGTGC-3′; elfa-R, 5′-CCGCTAGCATTACCCTCC-3′.

Fluorescently tagged TPC1 constructs were generated following existing protocols (Brailoiu et al., 2009; Castonguay et al., 2017), in which TPC1 was tagged at the C terminus with either an enhanced green fluorescent protein (EGFP) or tdTomato to avoid interfering with NAADP-sensitive motifs at the N terminus (Churamani et al., 2013). To generate these constructs, the full length tpcn1 coding sequence was amplified from the pSP64TNE-tpcn1-202 plasmid using a primer pair that omitted the stop codon and included a short linker. The resulting amplicon was cloned in-frame into the pSP64TNE-EGFP or pSP64TNE-tdTomato plasmids (Cheung et al., 2006) at the KpnI and AgeI site. The primer sequences used are as follows: forward primer tpcn1-fullcds-F, 5′-TAAGCAGGTACCAGACTATCCGTGTCCCCCAA-3′; reverse primer tpcn1-w/oSC-R, 5′-TGCTATACCGGTccGCCTCCGGAGTTAATGCTGTTTGTGGC-3′. The leader sequences are underlined; restriction sites for KpnI and AgeI (New England Biolabs Inc.) are in bold; and (in the reverse primer alone) the two additional bases included to ensure in-frame fusion are in lower case. When these two bases are combined with the bases shown in italics (i.e. ccGCCTCCGGA), they encode a short linker between TPC1 and the downstream fluorophore. mRNA encoding TPC1–EGFP or TPC1–tdTomato was synthesized as described above, and ∼150 pg of this mRNA was injected into AB wild-type embryos at the one-cell stage.

To monitor highly localized endolysosomal-generated cytoplasmic Ca²⁺ domains, a low-affinity, high dynamic range genetically encoded Ca²⁺ indicator (G-GECO1.2, K_d 1.2 µM; Zhao et al., 2011) was tethered to the C terminus of TPC1. The CMV-GEM-GECO1.2 plasmid containing the G-GECO1.2 coding sequence, was obtained from Addgene (plasmid #32446). The coding sequence was fused in-frame behind previously generated pSP64TNE expression vectors containing wild-type tpcn1 using Gibson assembly. A dominant-negative ‘pore-dead’ zebrafish TPC1 construct was also designed to reproduce the pore-dead TPC1 mutant generated by Brailoiu et al. (2009), which contains the L273P mutation in human TPC1. The protein sequences of human and zebrafish TPC1 were aligned using the constraint-based alignment tool (COBALT; Papadopoulos and Agarwala, 2007) and it was deduced that Leu264 in the zebrafish TPC1 protein sequence is the equivalent of Leu273 in the human sequence. Thus, a construct in which Leu264 of zebrafish TPC1 was substituted for proline was generated using mutagenic and outer primers. The mutagenic primer sequences are as follows:

tpcn1-L264P-F, 5′-ATTGTGAGTCCTTTCGTCCTC-3′; tpcn1-L264P-R, 5′-GAGGACGAAAGGACTCACAAT-3′. The codon that substitutes leucine for proline is shown in italics.

The outer primer sequences (TPC1-fullcds-F and TPC1-w/oSC-R) are as described above. These primers were used to generate fragments with overlapping ends to be cloned into the pSP64TNE expression vector to create the pSP64TNE-L264P-tpcn1 plasmid. Subsequently, the following primers were designed to amplify the G-GECO1.2 coding region from the CMV-GEM-GECO1.2 plasmid with overhangs for homology-based annealing: forward primer GECO-F, 5′-CTCCGGAGGCGGACCGGTCGCCACCatggtcgactcatcacgt-3′; reverse primer GECO-R, 5′-GGTAACCAGATCCGAATTCTttacttcgctgtcatcatttgtac-3′. Sequences that are complementary to the G-GECO1.2 coding sequence are in lowercase; those that overlap with the linearized pSP64TNE-tpcn1 vectors are in uppercase, and the short linker sequence between TPC1 and G-GECO1.2 (in the forward primer) is shown in italics. PCR was performed using the Phusion^® High-Fidelity DNA polymerase (New England Biolabs Inc.) with an initial 30 s denaturing step at 98°C, followed by 35 cycles of a 5 s denaturation step at 98°C, a 20 s annealing step at 61°C and a 20 s extension step at 72°C. This was followed by a final elongation step at 72°C for 8 min. The amplicons were analyzed by 1% agarose gel electrophoresis. They were then purified using a Monarch^® DNA gel extraction kit. Fragments containing tpcn1 or mutagenic tpcn1 were amplified from the pSP64TNE-tpcn1 or pSP64TNE-L264P-tpcn1 plasmid using the following primers to create overlapping ends with the G-GECO1.2 fragment: forward primer SP6-F: 5′-AGAATTCGGATCTGGTTACCACTAAACC-3′; reverse primer tpcn1-w/oSC-R, 5′-ACCGGTccGCCTCCGGAGTTAATGCTGTTTGTGGC-3′. The two additional bases included to ensure in-frame fusion are indicated in lowercase. PCR was performed using the Phusion^® High-Fidelity DNA polymerase (NEB, MA, USA) using the same settings described above except that the 20 sec annealing step was at 67°C and the extension step was for 2 min at 72°C. After the final elongation step at 72°C for 8 min, the amplicons were analyzed and purified as described above.

The results of these reactions were two amplicons as follows: (1) a G-GECO1.2 fragment and (2) a linearized TPC1 vector, which were annealed in a Gibson assembly reaction using the Gibson Assembly^® master mix (New England Biolabs Inc.). The results were confirmed by Sanger Sequencing. mRNA encoding TPC1–G-GECO1.2 or TPC1^L264P–G-GECO1.2 was synthesized using the mMESSAGE mMACHINE SP6 transcription kit as described above, and ∼150 pg of the mRNA was injected into the blastodisc of AB wild-type embryos at the one-cell stage.

Primary cultured SMCs used for immunolabeling were prepared and fixed via well-established methods (Kelu et al., 2015, 2017). To prepare primary cultured SMCs for live-cell imaging, several minor modifications were made to this protocol following recommendations reported by Sassen et al. (2017). Importantly, the cells were incubated in L-15 medium (Gibco Life Technologies, MA, USA) containing 10% fetal bovine serum (Gibco Life Technologies); 1× glutamine (Invitrogen Corp., Eugene, OR, USA); and 1.2% penicillin-streptomycin (10,000 units/ml; Gibco Life Technologies) for all the steps leading up to plating, and a plating time of 20 min was used. In addition, for experiments in which the nuclei were to be labeled, cells were resuspended in culture medium containing 1 µg/ml Hoechst 33258 (Molecular Probes, Thermo Fisher Scientific, Inc.) prior to plating.

Cells expressing TPC1–G-GECO1.2 (TPC1-GG) were plated for ∼45 min, as described above, and washed three times with Ca²⁺-free solution comprising 145 mM NaCl, 5 mM KCl, 3 mM MgCl₂, 10 mM glucose, 1 mM EGTA and 20 mM HEPES (pH 7.4). Cells were then maintained in 2 ml of this solution for the start of image acquisition. Halfway through imaging (i.e. ∼2 min into the 4 min time-lapse series), 500 µl of a solution containing 10 µM ionomycin (Sigma-Aldrich, Merck Group, MI, USA) and 50 mM CaCl₂ was added to the cells, to provide final concentrations of 2 µM and 10 mM, respectively. The ionomycin and CaCl₂ step was used to permeabilize the cells and saturate any of the remaining TPC1-GG that was not already fluorescing. In some experiments, TPC1-GG cells were treated with 25 µM BAPTA-AM [1,2-Bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetrakis(acetoxymethyl ester); Thermo Fisher Scientific, Inc.] or 25 µM dimethyl-BAPTA-AM (Thermo Fisher Scientific) for 45 min at ∼28°C during plating in the presence of 0.03% w/v Pluronic F127 (Sigma-Aldrich), followed by a 15 min de-esterification. In another series of experiments TPC1-GG cells were treated with 10 µM trans-Ned-19 (Enzo Life Sciences, Inc., NY, USA) for 45 min at ∼28°C during plating. All experiments were carried out within 45 min to 2 h after plating, and the temperature was maintained at ∼28°C throughout imaging. Images were acquired as described below.

Time-lapse image files were imported into ImageJ (https://imagej.nih.gov/ij/; National Institutes of Health, MD, USA) and analyzed with the TrackMate plugin (Tinevez et al., 2017). Images were split into separate channels and subjected to the ‘smoothened’ function once, before proceeding with particle tracking of the labeled puncta. The Laplacian of Gaussian (LoG) segmentation detector feature of the plugin was selected, and an estimated particle diameter of 0.5 µm and a threshold of 0.5 were used. The median filter of the plugin was disabled, and subpixel localization was enabled. Thresholding of the particles resolved by the LoG detector was performed using the ‘Quality’ values, such that those that formed the first histogram peak with low quality values were removed. The remaining particles were analyzed using the Linear Assignment Problem (LAP) tracker, with a linking maximum distance of 0.5 µm; gap closing distance of 0.5 µm over 2 frames; and with track splitting and merging enabled. The velocities of the individual TPC1 puncta were taken as a close approximation only, as we did not take into consideration any displacements that might have occurred within the z-plane. A value of 0.2 µm/s was selected as a threshold value to exclude background noise during data analysis.

To quantify observed differences in β-dystroglycan labeling patterns between controls, tpcn1 morphants, E8 crispants and bafilomycin A1-treated embryos, the percentage area of β-dystroglycan labeling within a region of interest (ROI) of 23 µm×23 µm was calculated. This value was obtained from confocal images of the ROI, which were acquired of the dorsal region of somite 8 and processed using ImageJ as follows: confocal z-stacks of β-dystroglycan fluorescence images were maximally projected, after which they were subjected to the ‘smoothened’ and ‘threshold’ functions in the default settings. The ‘analyze particles’ function was then used to calculate the total percentage area of β-dystroglycan within the ROI.

To quantify the number of TPC1-GG-recorded Ca²⁺ events at the ends of isolated SMCs, a ROI spanning one-fifth of the entire length of each SMC was used. The TrackMate plugin of ImageJ was used to quantify the number of TPC1-GG puncta within this ROI before the addition of ionomycin and CaCl₂, as well as the number of TPC1-GG puncta in the entire cell both before and after the addition of ionomycin and CaCl₂. A filter was set to exclude puncta with ‘quality values’ <3. To produce traces of all the TPC1-GG puncta over time against percentage maximal GG fluorescence, TPC1-GG puncta in the entire SMC were tracked using the LoG detector of the TrackMate plugin, as described above. A table containing ‘spot statistics’ was exported to Microsoft Office Professional Plus Excel 2013 (Microsoft Corp., WA, USA), and values for ‘Mean intensity’, ‘Track ID’ and ‘frame’ were subsequently exported to Minitab 17.3.1 (Minitab Inc., State College, PA, USA) for graph plotting.

A stock solution of nocodazole (Abcam, Cambridge, UK) was prepared at 5 µM in DMSO and stored at −20°C. In one series of live-cell imaging experiments, primary cultured cells were imaged for ∼2 min in culture medium on a heated stage (at ∼28°C). Nocodazole was then added to the cells to a final concentration of 1 µM. Image acquisition was halted, but the imaging chamber was kept on the heated stage for ∼15 min, after which, the same cell was imaged for a further ∼2 min. The TrackMate plugin of ImageJ was used to quantify the dynamics of TPC1–EGFP puncta before and after treatment with nocodazole, as described above. Values for velocity were extracted from the ‘spot statistics’ function for each time-lapse series and exported to Minitab 17.3.1 for graph plotting.

Stock solutions of bafilomycin A1 (1 mM; Tocris Bioscience, Bristol, UK) and trans-Ned-19 were prepared as previously described (Kelu et al., 2015). As both compounds are membrane permeable (Yoshimori et al., 1991; Naylor et al., 2009), dechorionated AB wild-type embryos were incubated at ∼28°C with 1 µM bafilomycin A1 or 100 µM trans-Ned-19 (both diluted in Danieau's solution) using methods described previously (Kelu et al., 2015).

Embryos at ∼24 hpf were dechorionated (as described above) and fixed with phosphate-buffered saline (PBS) containing 4% paraformaldehyde overnight at 4°C. After fixation, excess paraformaldehyde was removed and the embryos were washed first with PBS and then with PBS containing 0.1% Triton X-100 (PBST), after which they were permeabilized with PBST containing 1% DMSO (PBSTD) for 1 h. Permeabilized embryos were then incubated with blocking buffer (PBST containing 10% goat serum and 1% bovine serum albumin; Sigma-Aldrich) for 2 h, after which they were further incubated with blocking buffer containing F59 mouse anti-myosin heavy chain (used at 1:20; F59 Developmental Studies Hybridoma Bank, IA, USA) or anti-β-dystroglycan (used at 1:50; 43DAG1/8D5 Leica Biosystems, Wetzlar, Germany) at ∼4°C overnight. To label embryos with anti-TPC1 (used at 1:25; A15847 ABclonal, MA, USA), slight modifications were made such that 0.2% Triton X-100 was used in place of 0.1% Triton X-100, and the antibody incubation time was increased to 2 days. After the primary antibody incubation step, embryos were then washed extensively before being incubated with an Alexa Fluor 488-conjugated goat anti-rabbit IgG (H+L) antibody or Alexa Fluor 488-conjugated goat anti-mouse IgG (H+L) antibody (both used at 1:200 and both from Molecular Probes) at ∼4°C overnight in the dark. The embryos were then incubated with Alexa Fluor 568–phalloidin (used at 1:50; Molecular Probes) and 1 µg/ml Hoechst 33258 (Thermo Fisher Scientific, Inc.) at room temperature for 1 h in the dark to label F-actin and the nuclei, respectively. The embryos were washed extensively with wash buffer (a 1:10 dilution of blocking buffer) between each incubation step. At the end of these final wash steps, the labeled embryos were rinsed with PBST and then with PBS. To improve the incubation and washing efficiencies, gentle shaking was applied throughout using a mini gyro-rocker (Techne SSM3, Cole-Parmer, Stone, UK). Working dilutions of the antibodies and phalloidin were prepared in blocking buffer followed by centrifugation at 16,873 g for 5 min to pellet any unwanted debris.

AB wild-type embryos were sequentially injected into the blastodisc at the one-cell stage with ∼150 pg tpcn1–EGFP or tpcn1–tdTomato mRNA and ∼50 pg actc1b-mKate2–rab5c, –rab11ba, –rab32b or –lamp1 DNA (Addgene, plasmid numbers 109651, 109596, 109622 and 109507, respectively; Hall et al., 2020). At ∼48 hpf, muscle cell cultures were prepared as described above. Two reporter channels, one for TPC1–EGFP and one for the appropriate intracellular endolysosomal marker (or TPC1–tdTomato as a positive control), were used for analysis. Pearson's correlation analysis was performed using the ImageJ plugin JaCoP (Bolte and Cordelières, 2006) on single optical sections. To label the SR, AB wild-type embryos were injected sequentially with ∼150 pg tpcn1–EGFP and ∼50 pg actc1b-mKate2–KDEL DNA (Addgene, plasmid number 109494) (Hall et al., 2020) at the one-cell stage and muscle cell cultures were prepared and imaged as described above. To label the microtubules, AB embryos were injected with ∼150 pg tpcn1–tdTomato mRNA, and then the cells were stained with Tubulin Tracker Green (Thermo Fisher Scientific, Inc.) for 45 min at ∼28°C during plating in the presence of 0.03% w/v Pluronic F127.

To co-label cell membranes and the acidic organelles in intact AB wild-type embryos, embryos at ∼75% epiboly (∼8 hpf) were incubated overnight in Danieau's solution (Westerfield, 2000) containing 25 µM BODIPY^® FL C5-ceramide (Thermo Fisher Scientific, Inc.). They were then bathed in Danieau's solution containing 10 µM LysoTracker^® Red DND-99 (Thermo Fisher Scientific, Inc.) for 45 min at ∼28°C prior to imaging via confocal microscopy at ∼24 hpf.

LysoTracker fluorescence was quantified by analyzing confocal z-stacks taken through the SMCs in the dorsal region of somite 8. The LysoTracker fluorescence channel was maximally projected, after which a threshold of default settings was applied, and the puncta were quantified using the ‘analyze particles’ function of ImageJ with the following parameters: size=10–400 pixel² and circularity=0.00–1.00. To quantify the organization of LysoTracker-stained organelles at the MJ, the BODIPY^® FL C5-Ceramide fluorescence channel was used to define the location of the MJ, and then the ‘polygon select tool’ was used to draw a rhombus with a base of 20 µm centered on either side of it. This rhombus was used to create an ROI that was overlaid onto the corresponding LysoTracker fluorescence images. The organization of the puncta within this region was quantified by calculating the ratio of the number of particles within the ROI (which comprised two somite boundaries within the confocal image, i.e. somites 7 to 8, and somites 8 to 9), and the total number of puncta in the image.

Fixed embryos and primary cultured cells were imaged using a TCS SP8 laser scanning confocal microscope (Leica Microsystems GmbH, Wetzlar, Germany) with HC PL APO 40×/1.30 and HC PL APO 60×/1.40 oil objective lenses, respectively. Regarding the embryos, unless otherwise stated, the trunk was imaged from a lateral view at (and posterior to) somite 7. The respective wavelengths (excitation/detection) of the various fluorophores used are as follows: Alexa Fluor 488, 488 nm/519 nm; Alexa Fluor 568, 568 nm/600 nm; GFP, 488 nm/509 nm; tdTomato, 554 nm/581 nm; mKate2, 588 nm/635 nm; LysoTracker^® Red DND-99, 576 nm/590 nm; and DAPI and Hoechst 33258, 350 nm/461 nm. The temperature was maintained at ∼28°C for all live (cell and embryo) imaging experiments using a heated imaging chamber (TOKAI HIT Stage Top Incubator^®, Shizuoka-ken, Japan). Live embryos were anesthetized by immersion in Danieau's solution containing ∼0.02% MS-222 ∼5 min before the start of imaging. For time-lapse imaging of whole embryos and cultured cells, single confocal images were acquired at 867 ms and 655 ms intervals, respectively. For other experiments, images were acquired as a confocal z-stack by sequential scanning between the different channels to minimize fluorescence crosstalk. In some experiments, single optical sections alone were used for data analysis, whereas in others they were reconstructed using ImageJ.

Live imaging of primary cultured cells was also performed using a Zeiss LSM 980 with Airyscan 2 in conjunction with a 561 nm DPSS laser and Airyscan GaAsP detector. Images were acquired using a 63×/1.4 oil objective lens and the ZEN3.1 software (Zeiss, Germany).

To assess TPC1–EGFP fluorescence with or without injection of the tpcn1-T-MO, AB wild-type embryos at ∼2.5 hpf were manually dechorionated as described above and placed in individual wells in a 96-well flat bottom cell culture plate (SPL Life Sciences, Gyeonggi, South Korea). Bright-field and EGFP fluorescence images were acquired using exposure times of 4 ms and 1.5 s, respectively, using a Nikon Digital Sight DS-5Mc digital camera mounted on a Nikon AZ100 microscope (Nikon Instruments Inc., Melville, NY, USA) with a Nikon AZ Plan Apo 5×/0.5 NA objective lens. EGFP fluorescence was observed with a Nikon Intensilight C-HGFI mercury light source and the GFP/FITC single band filter cube, using excitation wavelengths of 465–495 nm and emission wavelengths ≥515 nm. All images were acquired using the NIS-Elements Advanced Research software (Nikon Instruments Inc.).

All image measurements were carried out using ImageJ. Numerical data were exported to Microsoft Office Professional Plus Excel 2013 (Microsoft Corp., WA, USA) for basic descriptive statistics and to GraphPad Prism (GraphPad Software, San Diego, CA, USA) for graph plotting. Note, in the box plots in Fig. 4Ae, the line in the middle indicates the mean and not the median. Data were also imported into Minitab 17.3.1 to run one-way ANOVA and calculate significance with Tukey's multiple-comparison test or Fisher's least significant difference (LSD) test. Figures were prepared using CorelDRAW version X8 (Corel Corp., Ottawa, ON, Canada).

We thank Dr John Parrington (University of Oxford, UK) for designing and providing the tpcn1-202-T-MO, and Dr Jeffrey J. Kelu and Ms Ching Man Chan (Division of Life Science, HKUST) for their technical support.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.259564.