Ethics statement

All experiments using zebrafish were reviewed and approved by the university’s Institutional Animal Care and Use Committee (#AM23-026).

Genome editing

Scn4aa and scn4ab knockout lines were generated using CRISPR-Cas9. For the synthesis of guide RNA (gRNA), oligonucleotide pairs, 5′- TAGGTGGGACTGGCCAACGTTC-3′ and 5′-AAACGAACGTTGGCCAGTCCCA-3′ for scn4aa, and 5′-TAGGTGGCCGAAATCCTAATTA-3′ and 5′-AAACTAATTAGGATTTCGGCCA-3′ for scn4ab were annealed and ligated with DR274 (Addgene), respectively. The template DNAs for the gRNA synthesis were generated using PCR with the forward (5′-GGTCAGTATTGAGCCTCAGG-3′) and backward (5′-AAAGCACCGACTCGGTGCCA-3′) primers. The gRNA sequences were transcribed using the MEGAscript T7 Transcription Kit (Thermo Fisher Scientific). To synthesize the Cas9 mRNA, pCS2-Cas9 (Addgene) was linearized using NotI (New England Biolabs) and transcribed using the mMESSAGE mMACHINE T7 ULTRA Kit (Thermo Fisher Scientific). Cas9 mRNA (100 ng/µL) and gRNA (100 ng/µL) for scn4aa and scn4ab were co-injected into single-cell stage embryos. The CRISPR targeted sequence was GTTGGGACTGGCCAACGTTCAGG for scn4aa and GCTGGCCGAAATCCTAATTATGG for scn4ab. F1 fish were identified by sequencing DNA fragments amplified from genomic DNA extracted from the caudal fin. The amplification was performed using forward (5′- GGTGTTCAAGCTTATAGCTATGGACCCTTAC-3′) and backward (5′- TGTAGGGTATCTTCTGCAATGTTGCAG-3′) primers for scn4aa, and forward (5′- GACACGGAAAATGTGCGGAACTGGTG-3′) and backward (5′- CAACAGGATGCATATACCTACCAACTGG-3′) primers for scn4ab. These primers were designated scn4aaF, scn4aaR, scn4abF, and scn4abR, respectively.

Video recording of zebrafish behavior

Escape responses were measured and analyzed as previously described [13]. Briefly, the trunk of an embryo was mechanically stimulated with a plastic pipette to induce escape behavior. In some experiments using 2% (w/v) methylcellulose solution, the tail of fish was pinched with forceps to elicit responses. Images were obtained using a Stemi 2000-C stereomicroscope (ZEISS) at 1,000 frames per second (fps) using a FastCaM 1,024PCI high-speed camera (Photron). The data were acquired as successive images. For measurement of the neck angle, a line was manually drawn from the swim bladder to the midpoint between eye balls for each image on ImageJ (version 1.52), whose angle was measured by the “measure” function. The angle values were normalized to the first image such that the angle of the first image became zero and were plotted against time. The turn speed (degrees/ms) was calculated as the change in the angle. When recorded in methylcellulose, the obtained plot was not smooth, so smoothing manipulation was performed using the savgol_filter() function from the scipy.signal library (version 1.11.3) in Python 3.10.12 before calculating the angle.

To record coiling activity, the chorion was manually removed, and embryos obtained at 28–32 hours post-fertilization (hpf) were embedded in low melting temperature agarose (NIPPON GENE). Coiling behavior was evoked by tactile stimuli on the dorsal side. Locomotion was recorded using a high-speed camera at 500 fps, and the shutter speed was 1/3000 s under the stereo microscope. The angle of the tail tip relative to the original position was measured using ImageJ. Turn speeds (degree/ms) were calculated as the change in the angle.

RNA sequencing

Adult male zebrafish (two WT and one NaVDKO) were anesthetized, and the trunk muscles were dissected. Total RNA was extracted from the isolated tissue using ISOGEN (NIPPON GENE) following the manufacturer’s instructions, including digestion with DNase I. The RNA was sequenced in paired mode on an Illumina HiSeq6000 by Rhelixa, Inc. (Tokyo, Japan). The quality of the sequence reads was assessed using FastQC software (version 0.11.7). Low-quality bases and adaptor sequences were trimmed using the Trimmomatic software (version 0.38). The trimmed reads were aligned to the zebrafish genome GRCz11 using HISAT2 (version 2.1.0).

Ca²⁺ imaging

The promoter region of mylz2 (approximately 2 kb) was amplified from 5′-aaa-gaattc-actagt-ATTCGCCACAGAGGAATGAGCC-3′ and 5′-tttgtcgacGTGTGAAGTCTAAGAAGATCAAGAAGAGAAGTC-3′ primers and subcloned between Tol2 right and GCaMP7a sequence of the pT2RUASGCaMP7a plasmid (Kawakami lab). For in vivo Ca²⁺ imaging, the generated pT2G-mylz2pro plasmid was injected into single-cell stage WT and NaVDKO embryos. Embryos were prepared as described previously [13]. Briefly, the fish were anesthetized using N-benzyl-p-toluene sulfonamide (BTS) [14] and mounted on low-temperature melting agarose. Spontaneous activity is induced by N-methyl-d-aspartate (NMDA) [15]. Images were acquired using an SP8 confocal microscope (Leica) equipped with a 5 × /0.15 objective lens. Sixty-four pixel-square regions were scanned every 41 ms.

For Ca²⁺ imaging of the isolated myocytes, the transgenic line Tg[mylz2-GCaMP7a] was generated by injecting pT2G-mylz2pro into single-cell stage WT embryos. Isolated myocytes were transferred to the 100 µL Tyrode’s solution droplet on the coverslip. Subsequently Tyrode’s solution containing 10 μM acetylcholine (ACh, 5 μL) was gently applied to stimulate myocytes. For the tetrodotoxin (TTX) assay, TTX was added to the droplet on the coverslip (final concentration, 1 µM). The ACh solution also contained 1 µM TTX. The myocytes were used for the assay within 6 h of preparation. Images were acquired using an EOS RP camera (Canon) attached to an MVX10 microscope (Olympus). For analysis, color images were converted to 8-bit monochrome images by ImageJ (Fiji, version 1.54f). The fluorescence intensity of each pixel was indicated in colors using the “fire” look-up table in ImageJ.

Isolation of myocytes

Embryos or adult zebrafish were anesthetized using MS-222. The trunk was digested by 3 mg/mL type I collagenase (Sigma-Aldrich) in Tyrode’s solution containing 110 mM NaCl, 2.1 mM KCl, 1.25 mM MgCl₂･6H₂O, 0.28 mM NaH₂PO₄, 3.5 mM NaHCO₃, 5 mM glucose, 15 µ M CaCl₂, 10 mM HEPES, and 4 mM pyruvate; the pH was adjusted to 7.5. Myocytes were washed in Tyrode’s solution and the CaCl₂ concentration was gradually increased to 1 mM [16].

Electrophysiology

The sodium current in the isolated myocytes was recorded as previously described [17]. The pipette solution contained 58 mM CsCH₃SO₃, 32 mM CsCl₂, 10 mM EGTA, 10 mM HEPES (pH 7.1). The bath solution contained 100 mM NaCl, 2 mM KCl, 0.2 mM CaCl₂, 28 mM MgCl₂, 3 mM glucose, and 5 mM HEPES (pH 7.4). The leak currents and the current for charging the cell capacitance were subtracted using the P/-10 protocol. The sampling frequency was set at 100 kHz. The pulse interval was set at 5 s. “Gating” current of the DHPR was recorded in the presence of 1 µM TTX in the bath solution. The recording conditions, including the bath solution, pipette solution, pulse protocol, and holding potential, were identical to those used for the sodium current recording. To identify the genotype of isolated myocytes, genomic DNA was isolated from the head of the dissected fish myocytes. DNA fragments were amplified using the primers scn4aaF, scn4aaR and scn4abF, scn4abR, respectively (as detailed in the Genome Editing section), and subsequently sequenced. To record calcium currents, the pipette solution contained 100 mM CsCl₂, 20 mM CsF, 2 mM MgCl₂, 0.1 mM EGTA-Cs, 10 mM HEPES, 40 mM glucose, with the pH adjusted to 7.4 using CsOH. The bath solution contained 130 mM tetraethylammonium (TEA)-Cl, 10 mM CaCl₂, 1 mM MgCl₂, 20 mM glucose, 10 mM HEPES, with the pH adjusted to 7.4 using TEA-OH. The leak currents and the current for charging the cell capacitance were subtracted using the P/-10 protocol. The sampling frequency was set to 50 kHz. The pulse interval was set at 7 s. Recording of the miniature end-plate current (mEPC) was recorded as described previously with some modifications [13,18]. Embryos were skinned and pinned in the recording chamber and spontaneous activity was recorded in a whole-cell configuration. Membrane potential was held at −90 mV. The bath solution contained 100 mM NaCl, 2 mM KCl, 0.2 mM CaCl₂, 2.8 mM MgCl₂･6H₂O, 3 mM glucose, and 5 mM HEPES (pH 7.4). Nifedipine (Merck, 10 µM) and TTX (1 µM) were added before recording. The pipette solution contained 120 mM KCl, 5 mM EGTA, and 5 mM HEPES (pH 7.2). The sampling frequency was set at 100 kHz. The series resistance was lower than 10 MΩ and compensated 65%–80%. Miniature events with amplitudes larger than 0.1 nA and coefficients of determination higher than 0.97, as determined by exponential fitting, were analyzed. Data were acquired using an EPC10 patch-clamp amplifier with PatchMaster software (version 2 × 73, HEKA Electronik) and analyzed using custom-made programs in Python 3.10.12.

Simulation of muscle membrane potential

The simulation was performed using the NEURON software (https://www.neuron.yale.edu/neuron/) (version 8.2.3). The myofiber was modeled as a cylinder and the potential amplitude was calculated. ACh release sites were incorporated as cylinders 1 µm in length and 16 µm in diameter. The Hodgkin–Huxley model provided in the software was used under zero sodium conductance to mimic NaVDKO. The default values were used for the other parameters, except for the resistivities of the membrane, fiber interior, and diameter of the cylinder.

Immunohistochemistry

Zebrafish embryos were anesthetized using MS-222 and fixed with 4% paraformaldehyde (PFA) for 4 h at room temperature. After rinsing with phosphate buffered saline including 0.1% Tween 20 (PBS-T) and dehydrated by washing in 100% methanol, samples were stored at −20 °C. For immunostaining, embryos were hydrated in a series of washes with 75%, 50%, and 25% methanol and PBS. The embryos were permeabilized with 1 mg/mL collagenase type I (Fujifilm Wako) in PBS for 3 h at room temperature. After rinsing with PBS-T, samples were incubated overnight at 4 °C with the anti-synaptic vesicle protein 2 (SV2) antibody (Developmental Studies Hybridoma Bank) at 1:200 dilution in blocking solution containing 5% sheep serum (Sigma-Aldrich) in PBS-T. After washing with PBS-T, samples were incubated overnight at 4 °C with goat anti-mouse IgG secondary antibody conjugated with Alexa 555 (Thermo Fisher Scientific) at 1:500 dilution and α-bungarotoxin (BTX) Alexa Fluor 488 conjugate (Molecular Probes) at 1:500 dilution in blocking solution. Images were acquired using an SP8 laser scanning confocal microscope (Leica) equipped with a 20 ×/0.75 objective lens in sequential mode.

For analysis, z-stack color images were superimposed and split into individual red, green, blue images using “split channels” in ImageJ (Fiji, version 1.54f). “R” and “G” images were adjusted using the “Threshold” tool to eliminate background signals. The areas stained with BTX and SV2 were estimated from the number of positive pixels and the matched fraction of BTX and SV2 signals was defined as follows: (number of positive pixels for both BTX and SV2)/ (number of positive pixels for either BTX or SV2). Positive pixels were counted using custom-made programs in Python 3.10.12.

For BTX staining of isolated myocytes, myocytes after dissociation were fixed in 4% PFA at room temperature for 30 min. After washing by PBS-T, myocytes were stained by BTX Alexa Fluor 488 conjugate (Molecular Probes; 1:400 with PBS-T) at room temperature for 1 h in the dark. Fluorescence was observed using a BZ- × 700 fluorescence microscope (KEYENCE).

Swimming treadmill for adult fish

The 170 mL model SW10000 swim tunnel respirometer (Loligo Systems) was used. The water velocity was calibrated according to the manufacturer’s instructions. The oximeter was calibrated using 0% and 100% O₂ saturation. Five percent (w/v) Na₂SO₃ and water aerated for 1 h were used for 0% and 100% O₂ solutions, respectively. A critical swimming speed (Ucrit) was calculated as (U + T)/ 60, where U is the actual flow rate at which fish terminated swimming and T is the time fish spent swimming at the flow rate [19,20]. Three-month-old fish were used for analysis. Water temperature ranged from 24.9–26.5 °C and the oxygen saturation exceeded 89%. In the Ucrit analysis, the regression line was estimated using the polyfit() function in the NumPy library version 1.26.0. in Python 3.10.12.

Statistics

In the bar graphs, the mean and standard error of the mean (SEM) are shown, in addition to individual dots. P-values were calculated using the Welch’s t test. Welch’s t test and Pearson’s correlation coefficient were calculated using the ttest_ind() function in the scipy.stats library version 1.11.3 and the corrcoef() function in the NumPy library version 1.26.0, respectively, in Python 3.10.12.

Generation of scn4aa and scn4ab double knockout zebrafish

We generated scn4aa and scn4ab heterogeneous zebrafish by the CRISPER-Cas9 method. Together, these two genes encode NaV1.4. Because of genome duplication during evolution, it is common for two genes in zebrafish to correspond to a single human gene, in this case, SCN4A (S1 Table). Four bases were deleted in exon 14 of scn4aa gene and 14 bases were inserted in exon 8 of the scn4ab gene (S1A1 and S1B1 Fig). These changes cause frameshifts and create premature stop codons (S1A2 and S1B2 Fig). Premature stop codons were located between S3 and S4 of domain II for scn4aa, and between S5 and S6 of domain I for scn4ab (S1C Fig). According to structural studies of NaVs [21,22], the remnant protein is expected to be nonfunctional even if it reaches the plasma membrane.

Subsequent crossing of the two lines generated a double-heterogeneous fish that was identified by sequencing. The scn4aa/scn4ab double knockout line (NaVDKO) was generated by crossing two heterogeneous fish. To verify the functional abolition of NaV1.4 proteins in muscles, we recorded the voltage-dependent sodium current from isolated trunk myofibers of embryos. Trunk muscle comprises two types of myocytes: superficial and deep muscles (i.e., slow and fast fibers). In WT animals, slow fibers do not have voltage-dependent sodium currents [23]. As positional information is lost after myocyte dissociation, we recorded currents without distinguishing between fast and slow fibers.

Current recordings from WT fibers in the presence of cesium in the pipette solution showed transient inward currents following the start and end of depolarizing pulses in 24 cells from nine fish (upper panel in Fig 1A). The inward current in the initial phase (<5 ms) of the depolarization pulse disappeared by the application of 1 μM TTX (middle panel in Fig 1A), demonstrating that the current was derived from NaVs [14]. The residual ON- and OFF-current recorded in the presence TTX is therefore the “gating” current of the DHPR, that is, voltage dependent calcium channel (middle panel in Fig 1A) [24].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Electrophysiology of isolated myocytes.

(A) Representative sodium current of WT (upper panel), WT in the presence of 1 μM TTX (middle panel) and NaVDKO (lower panel) myocytes. Top panel indicates the timing of the depolarization pulse. Traces evoked by voltages ranging from −90 to 20 mV in 10 mV increments are superimposed. (B) Comparison of the peak ON-current density between WT and NaVDKO myocytes. Currents were recorded from myocytes at 3–6 days post fertilization (dpf). (C) Representative calcium current recordings from isolated myocytes. Top panel indicates the timing of the test pulse. Traces evoked by voltages ranging from −90 mV to 10 mV in 20 mV increments are superimposed. (D) Current–voltage relationship of the calcium current recording. Average current amplitudes 100–102 ms after the onset of the test pulse were plotted against potentials. Data were shown as mean ± SEM (n = 6 for both WT and NaVDKO). The numerical data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003137.g001

When we recorded currents from 24 cells in seven NaVDKO fish, we observed only “gating” currents of the DHPR (bottom panel in Fig 1A and 1B). These results indicate that NaVDKO fish lack functional expression of NaVs in their myocytes.

Caenorhabditis elegans does not have a gene encoding NaV in its genome. The roles of NaV are substituted by Ca²⁺-dependent spikes in excitable cells [25–27]. The previously described absence of calcium currents in zebrafish myocytes [17,28] was confirmed in the present study (Fig 1C and 1D). The findings indicate that NaVDKO embryos do not contain sodium APs or calcium spikes.

Locomotion of embryos

Slow muscles which lack NaV expression [23] underlie the tail coiling movement observed in 1-dpf zebrafish embryos [29]. We examined the coiling activity of NaVDKO zebrafish to determine its potential effects on slow muscles. Rostral parts of 28–32-hpf embryos were embedded in a gel and coiling behaviors were recorded (S2A Fig). The deviation of the tail in the images was measured as the angle (S2B Fig). The maximum turn speeds of individual embryos were indistinguishable between WT and NaVDKO zebrafish (S2C Fig).

As embryos develop, rapid locomotion driven by fast muscles becomes dominant. We expected that NaVDKO zebrafish would have compromised quick locomotion owing to dysfunctional fast fibers, as observed in mutants whose fast muscles do not receive inputs from motor neurons [13]. The touch responses of the embryos were recorded. Normal embryos display fast and large body bends (C bends), followed by smaller counter bends [30,31]. Unexpectedly, NaVDKO embryos exhibited robust escape responses comparable to those of WT embryos (Fig 2A and 2B and S1 and S2 Movies). The maximum turn angle of the C-bend (the first turn) and the maximum turn speed were not statistically different from those of the WT (Fig 2C and 2D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Escape response of WT and NaVDKO zebrafish.

(A) Snapshots of the representative escape responses of WT (upper) and NaVDKO (lower) zebrafish at 4 dpf in normal water. Images were captured every 2 ms. Scale bar: 1 mm. (B) Representative plots of the “neck angle” during the escape behavior of WT and NaVDKO fish. The inset shows the definition of “neck angle.” (C) Analysis of the maximum neck angle in the first turn of 4 dpf fish. Individual data points are shown as dots. Bars and error bars represent the mean and SEM, respectively (n = 10 for WT, n = 9 for NaVDKO). (D) Analysis of the maximum turn angle speed of 4 dpf fish. The maximum turn angle speed was calculated from the differential of the neck angle (n = 10 for WT, n = 9 for NaVDKO). The numerical data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003137.g002

Considering that differences can be observed under exigent conditions, we tested the escape response of embryos in a viscous solution containing 2% methylcellulose, in which embryos required more power for escape. After pinching the tail, the embryos displayed smaller turn angles than those in water (S3A, S3B and S3C Fig and S3 and S4 Movies). The maximum angle of the turn calculated from the smoothed curve (S3B Fig) was again indistinguishable between WT and NaVDKO (S3C Fig). The maximum turn speed of the NaVDKO was comparable to that of the WT zebrafish (S3D Fig). The collective findings indicated no significant difference in the escape response between WT and NaVDKO zebrafish.

Lack of compensatory mechanisms in NaVDKO

The normal locomotion of NaVDKO (Figs 2, S2 and S3) prompted us to examine whether the upregulation of other NaVs compensated for the deleted NaVs. RNA sequencing of trunk muscles from WT and NaVDKO zebrafish (S2 and S3 Tables). In the analysis of differentially expressed genes (DEG), among eight scn genes of zebrafish, only scn8aa and scn12aa showed positive logFC values (0.173 and 0.158 for scn8aa and scn12aa, respectively; S4A Fig and S4 Table). However, the respective p-values of 0.881 and 0.935 for scn8aa and scn12aa did not reach level of significance (S4 Table). Moreover, the transcript per million (TPM) values for scn8aa and scn12aa, were negligible in both the WT and NaVDKO embryos compared to those for scn4aa and scn4ab (S4 Table and S4B Fig). The lack of robust upregulation among scn genes in NaVDKOs is compatible with the absence of Na⁺ currents in the myocytes of NaVDKOs (Fig 1A and 1B).

We also examined alpha1, a pore-forming subunit of calcium channels. The only gene that displayed significant upregulation was cacna1ia, which encodes CaV3.3 (S5 and S6 Table). We examined whether upregulation of cacna1ia leads to the appearance of calcium currents in myocytes. Calcium currents were not recorded in myocytes of NaVDKO or WT zebrafish (Fig 1C and 1D). The negligible level of TPM of the cacna1ia was consistent with the current recordings. In summary, the sodium or calcium channels in NaVDKO do not compensate for the loss of NaV1.4.

We also examined the possibility that the deletion of NaV1.4 is compensated by remodeling of the NMJ. Pre-and post-synapses were stained with antibodies against SV2 and alpha-BTX conjugated with Alexa488, respectively. At 28 hpf, nicotinic ACh receptors (nAChRs) and SV2 were observed in the middle body segments (arrowhead in S5A₁ Fig) and spinal cord (arrow in S5A₁ Fig) [32]. Areas with signals above the thresholds (S5A₂ Fig) were measured. Areas of positive signals were not different between WT and NaVDKO embryos (S5A₃ Fig), and the double-positive fraction for BTX and SV2 signals was not distinct between the two (S5A₄ Fig). In 3 dpf embryos, BTX and SV2 signals were observed in both types of NMJs. Signals at the boundaries between body segments belong to slow muscles (arrow in S5B₁ Fig) and punctate signals in body segments were NMJs of fast muscles (arrow head in S5B₁ Fig) [12,33]. The signal-positive area and double-positive fraction did not differ between WT and NaVDKO zebrafish (S5B₂, S5B₃ and S5B₄ Fig). In 5 dpf embryos, the number of NMJs increased compared with 3 dpf embryos (S5B₁ and S5C₁ Fig), whereas no difference was observed between WT and NaVDKO embryos (S5C₂, S5C₃ and S5C₄ Fig). These data suggest that the knockout of NaV1.4 is not compensated by the NMJ remodeling.

Ca²⁺ imaging of muscle fibers

To explore NaVDKO embryos at the cellular level, we examined changes in intracellular calcium ion concentration ([Ca²⁺]_i) in the trunk muscles of WT and NaVDKO embryos. [Ca²⁺]_i was visualized by transient expression of the genetically encoded Ca²⁺ indicator GCaMP [34]. GCaMP expression was driven by the fast fiber-specific mylz2 promoter (S6A Fig) [35]. Notably, GCaMP was ectopically expressed in some slow fibers that could be anatomically distinguished from fast fibers. The embryo was anesthetized and spontaneous activity was induced by the application of NMDA. We found a transient increase in [Ca²⁺]_i in the WT (S6B Fig, upper panels and S5 Movie). These events concurrently occurred in fast and slow fibers (S6C Fig). Increase in [Ca²⁺]_i was also observed in NaVDKO embryos, which was not distinct from that in WT embryos (S6B and S6D Fig and S6 Movie).

The results presented thus far indicate that NaVs are dispensable for muscle contraction. If this is the case, WT muscle fibers are expected to contract even in the presence of TTX. To ensure penetration of TTX into fast muscle fibers, myofibers were dissociated using collagenase. To identify fast fibers, transgenic zebrafish expressing GCaMP7a driven by the mylz2 promoter were generated.

First, we stimulated myofibers by 10 μM ACh in the absence of TTX. For S7A₁ Fig, ACh was applied along the indicated directions. The timing of the fluorescence change varied depending on the location of the cell relative to the ACh application (Figs 3B, S7A₁ and S7A₂ and S7 Movie).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Ca²⁺ imaging of isolated myocytes in the absence and presence of TTX.

(A) Expression of GCaMP7a was driven by the mylz2 promoter. (B1, B2, and B3) Representative Ca²⁺ imaging of WT myocytes at 3 dpf embryos. Numbers below the images indicate time from the beginning of the recording. Fluorescent intensity is shown in color. (B4) Maximum fluorescence intensity in the region of interest (ROI) (B1, B2 and B3) was plotted against time. (C1, C2, and C3) Ca²⁺ imaging of WT myocytes at 5 dpf in the presence of 1 μM TTX. (C4) Maximum fluorescence intensity of ROI shown in C1, C2 and C3 were plotted against time. The numerical data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003137.g003

Next, Ca²⁺ imaging was performed in the presence of 1 μM TTX. The application of ACh and TTX induced a robust increase in fluorescence (Figs 3C and S7B and S8 Movie). NaV currents were completely blocked by 1 μM TTX (Fig 1A, upper and middle panels), these data again supported the notion that NaVs are dispensable for the myofiber contraction in zebrafish embryos.

Swimming capability of adult NaVDKO

After concluding that NaVs are dispensable for embryonic locomotion, we hypothesized that NaVs play an essential role in adults. Accordingly, we assessed the swimming capability of the adult fish. Escape behavior was evoked by pinching the tail in methylcellulose in the same manner as in the embryos. The locomotion of adult NaVDKO and WT fish was compared (S8A and S8B Fig and S9 and S10 Movies), and no significant difference was found in the maximum angles of the first turn (S8C Fig). The maximum turning speed calculated from the smoothed curves did not differ (S8D Fig). The findings do not indicate compromised locomotion in adult NaVDKO zebrafish.

To examine another aspect of locomotion, adult NaVDKO zebrafish were examined using a swimming treadmill. In the swim tunnel, adult fish swim in a cylindrical tube in which the flow rate is controlled by the rotation speed of the screw located at the end of the cylinder (S9A Fig). The flow rate was maintained at 15 cm/s for the first 1 min as the fish acclimated to the tube. Measurement of the Ucrit started at 20 cm/s. The flow rate was increased by 1 cm/s every minute until the fish stopped swimming (S9B Fig).

While a weak negative correlation was found between Ucrit and body mass index (BMI), Ucrit was not associated with the fish genotype (S9C and S9D Fig). Therefore, the swimming ability of adult NaVDKO was comparable to that of WT fish.

Simulation of muscle excitation without voltage-dependent sodium conductance

The EPP activates NaVs, and the generated AP spreads by activating neighboring NaVs until depolarization reaches the DHPR in the T-tubule. The amplitude of the EPP is generally smaller than that of the AP, which is insufficient for the stimulation of DHPR [36]. Therefore, it was unexpected that the NaVDKO zebrafish would show normal locomotion.

DHPRs in zebrafish myofibers lack Ca²⁺ conductance (Fig 1C and 1D) [17,28]. Moreover nAChRs in fast muscles has a low Ca²⁺ permeability [13,37]. Therefore, the sole source of Ca²⁺ for contraction is the SR, for which DHPR activation is crucial. One possible mechanism for the contraction of NaVDKO fibers is the direct activation of DHPRs by EPPs.

To examine whether sufficiently large depolarization can be transmitted along the plasma membrane without NaVs, we performed simulations using NEURON software. The muscle fiber was modeled as a cylinder for simplicity and ease of computation (Fig 4A); a short cylindrical section was inserted as neuromuscular synapses, and the synaptic currents were injected in this area (dark gray cylinder in Fig 4A). Depolarization was evoked by current injection at the ACh release site and was conducted in the light gray section, where the time-dependent change in the potential was calculated. According to the cable theory [38], the parameters of membrane resistivity and fiber interior are key factors for attenuation. However, to the best of our knowledge, the resistivity of zebrafish myofibers has not been reported. Therefore, we used an extensor longus digit. IV for the European frog [39]. The parameters used in the simulation are summarized in S7 Table.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Mathematical simulation of depolarization in embryonic myocytes.

(A) A single synapse model with a length of L. The ACh release site was modeled as a 1 μm-long cylinder (dark gray). The current was injected at the ACh release site and potentials at 0.1, 0.5, and 0.9 L away from the ACh release site were calculated. (B) Membrane potentials for L = 100 (B1) and 200 μm (B2). Potentials at 0.1, 0.5, and 0.9 L from the ACh release site were calculated. Three colored traces (0.1, 0.5, and 0.9 L) overlapped. Square pulses above the plots indicate the timing of the synaptic input. (C) BTX staining of myofibers in 4 and 6 dpf fish. Myofibers were from WT (C1) and NaVDKO (C2) fish. Upper panels show magnified views. Scale bars: 50 μm. (D) Representative miniature end-plate currents (mEPC) of WT (left) and NaVDKO (right). (E) Representative histogram of the mEPC amplitude in WT (dark gray) and NaVDKO (light gray). The curves were fit to a Gaussian function. (F) Geometry of the multiple-synapse model. Thirteen 1 μm-long cylinders (dark gray) and 12 7 μm-long cylinders (light gray) were placed alternately. Synaptic currents were injected into individual ACh release sites. The 7 μm-long cylinders are designated as region #1 through #12. (G) Potentials at the center of region #6. The numerical data presented in this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003137.g004

We initially employed the one-synapse model shown in Fig 4A to focus on the attenuation of depolarization along the fiber length in the absence of NaV. Potentials at positions 0.1, 0.5, and 0.9 L from the ACh release site (L = length of the cylinder) in the light gray cylinder (Fig 4A) were calculated when the current was injected at the ACh release site. The current amplitude and duration were set as 15.6 nA and 0.6 s, respectively. When the fiber length was 100 μm (L = 100 μm), the maximum amplitude of depolarization was around 80 mV, which was comparable between 0.1 and 0.9 L (Fig 4B₁), indicating the negligible attenuation of the potential. When the fiber length was changed from 100 to 200 μm, the membrane depolarization was reduced to approximately 0 mV, whereas attenuation was not significant (Fig 4B₂). Smaller depolarization and noticeable attenuation were evident when the length was set to 500 or 1,000 μm (S10A and S10B Fig).

Muscle fibers in zebrafish embryos were 83.1 ± 5.3 and 101.0 ± 5.8 μm in length for WT and NaVDKO, respectively (data are expressed as mean ± SEM, with 25 cells for WT and 28 cells for NaVDKO) (Fig 4C₁ and 4C₂). The findings indicate that attenuation of the membrane potential is not significant in fibers at the embryonic stages.

Next, we simulated the change in membrane voltage in a fiber with a current amplitude and synapse distribution closer to the biological situation. A previous report showed that the mean quantal content of the neuromuscular junction (all ACh release sites) in zebrafish embryos is approximately 12 [18]. Quantal content is defined as the number of presynaptic vesicles released into the synaptic cleft(s) of a single motor neuron by the AP. We counted the number of ACh release sites per fiber by BTX staining; 13.1 ± 0.9 for WT (n = 27) and 12.4 ± 0.9 for NaVDKO (n = 19) (Fig 4C₁ and 4C₂). These numbers are close to the quantal content, which may indicate that one synaptic vesicle was released at a single ACh release site.

We recorded the mEPC (Fig 4D). Mean amplitude was estimated by fitting the histogram plot using Gaussian distribution (Fig 4E). The mean amplitude value was 1.2 ± 0.1 nA for WT and 1.3 ± 0.2 nA for NaVDKO (n = 12 for both). The decay constant of mEPC obtained by fitting to the exponential equation was 0.58 ± 0.03 ms for WT (n = 4,538 events) and 0.61 ± 0.04 ms for NaVDKO (n = 6,127 events). To simulate currents with these parameters, the synaptic input was modeled as a square pulse with the amplitude and the width set to 1.2 nA and 0.6 ms at individual release sites. ACh release sites were diffusely disseminated in fast muscle fibers (Fig 4C). For simulation, 13 release sites were inserted at equal intervals (Fig 4F). The myofiber cylinder was separated into 12 regions (1–12), where the membrane potential was calculated. Assuming that the release probability of all ACh release sites is one, which means that a 1.2 nA current was simultaneously injected at all sites, the maximum depolarization in the middle of individual regions was approximately 80 mV (Figs 4G and S11A). Recordings of the “gating” current from isolated myofibers revealed that the DHPR was fully activated when the membrane was depolarized at approximately 80 mV (S11B Fig), suggesting that DHPRs can be fully activated by the EPP alone.

To examine the dependency of the obtained results on the employed parameters, we performed a simulation using the resistivities of axons in squid giant axon [40], lobster leg axon [38] and goldfish Mauthner axon in goldfish [41] (S7 Table). In the geometry shown in Fig 4F, robust depolarization exceeding +30 mV was observed for all parameters (S12 Fig). These estimates suggest that DHPR is fully activated even in the absence of NaVs in the myofibers of embryos.

Our analyses revealed the normal locomotion of adult NaVDKO fish (S8 and S9 Figs). The adult myofibers were much longer than those of the embryos (1,102 ± 87 µm, mean ± SEM, n = 16; S13A Fig). To examine whether EPP can directly stimulate DHPR in adults, we employed another set of parameters to simulate adult muscle fibers. The resistivities and amplitude of the synaptic current did not change in the embryos because the current could not be recorded by conventional patch-clamp techniques owing to the size of adult fibers.

We could not determine the number of ACh release sites per fiber because some were linear rather than punctate (S13A Fig). To reconcile this observation with the simulation, a range of ACh release site numbers per fiber were applied: 11, 51, 101, and 139 (S13B Fig). The potentials in the middle of this region are shown in S13C Fig. At sites 11 and 51, the maximum potentials were below 0 mV, whereas for the 101- and 139-site models, the maximum potentials were greater than 0 mV. The results of the 101- and 139-site models suggest that myofibers may be able to contract without NaVs, in which case the disseminated distribution of ACh release sites is the reason why NaVs are dispensable in adult fibers.

Excitation of fast fibers without NaV1.4s

Contrary to the established view, this study showed that NaVs are dispensable for muscle contraction in zebrafish embryos and adults. Skeletal muscle fibers are roughly divided into fast and slow fibers, and NaVs are expressed only in fast fibers [23]. We previously knocked out the gamma and epsilon subunits of nAChRs specifically expressed in fast fibers, generating zebrafish lacking functional fast fibers [13]. These embryos were observed to swim much slower than WT fish. In sharp contrast, the swimming speed of the NaVDKO fish examined in this study was comparable to that of the WT fish (Figs 2, S2, S3, S8 and S9). These findings support the idea that fast fibers function normally in NaVDKO fish.

Ten genes encoding NaVs exist in human chromosomes, while the zebrafish genome contains eight α-subunit genes according to the genome assembly, GRCz11 (S1 Table). In situ hybridization of α-subunit genes in zebrafish embryos showed that scn4aa was expressed in the head, pharyngeal muscle, and pectoral fin at 72 hpf, whereas scn4ab was expressed in the trunk somites at 24 and 60 hpf [42]. Current recordings from isolated myofibers (Fig 1A and 1B) confirmed that NaV1.4, likely scn4ab, underlies the dominant sodium conductance in the myofibers of zebrafish.

Physiological roles of NaV1.4s in zebrafish

Mutations in human NaV1.4, which alter skeletal muscle excitability, cause myotonia, periodic paralysis, congenital myopathy, and myasthenic syndrome. However, knockout of NaV1.4 did not affect zebrafish locomotion. Synaptic connections in mammalian skeletal muscles are located at the center of the myocytes (the innervation zone). In addition, mammalian myofibers are generally longer than zebrafish myofibers, extending more than a few centimeters. As shown in S10B Fig, the membrane depolarization of a 1 mm-long fiber was attenuated in the absence of sodium conduction. Therefore, the distribution of ACh release sites and length of the fiber would be determinant factors for the physiological significance of NaV1.4.

APs have been reported in muscles of more primitive organisms than in fish. The amphioxus genome contains five genes encoding NaVs and the muscle contains sodium-dependent APs [43]. Muscle fibres in the ascidian Halocynthia aurantium have Ca²⁺-dependent AP [44]. Five NaV1s and one NaV2 that are permeable to both sodium and calcium are found in the genome of the lamprey [45,46], among which NaV1γ expresses in muscles [47]. Despite the conservation of APs in the process of evolution from common ancestors of chordates, APs may not be necessary for muscle contractions per se in these organisms because the simulation performed in the present study suggested that the potential was not attenuated in short fibers. AP may be a redundant mechanism for depolarizing the membrane, which is sufficient for the full activation of DHPR. One possibility is that animals can save energy by using APs for muscle contraction. In zebrafish, individual muscle fibers receive input from a single primary motor neuron and up to three secondary motor neurons [48]. Because NaVs amplify EPP, the excitation of a single motor neuron may be sufficient for the excitation of myofibers. In contrast, excitation of all motor neurons may be necessary in the absence of APs. The present demonstration of the dispensability of NaV in the zebrafish muscle fibres sheds new light on the evolutionary roles of APs.

Although NaV2 has been lost in the zebrafish genome, NaV1.4s remains functional throughout evolution without becoming a pseudogene. This strongly suggests that NaV plays an essential role in zebrafish survival.

NaV channels are expressed in non-excitable cells such as immune cells, cancer cells, and osteoblasts in mammals [49]. The scn8aa gene in zebrafish encodes NaV1.6 that is expressed in motor neurons [50]. NaV1.6 is involved in the locomotor activity, development, innervation, and regeneration of fins [51–53]. In some electric fish, scn4ab1, a variant of scn4ab, is expressed in the electric organs [54]. In the sea urchin Strongylocentrotus purpuratus, NaV2 is expressed in the ventrolateral ectoderm during early developmental stages and plays a role in normal skeletal patterning [55]. The sea anemone Nematostella vectensis expresses NaV1 and NaV2 [56,57], which are not expressed in muscle fibres [57]. NaV1.4s in zebrafish may play a physiological role outside the skeletal system.