Role of somite patterning in the formation of Weberian apparatus and pleural rib in zebrafish

Abstract

In the vertebrate body, a metameric structure is present along the anterior–posterior axis. Zebrafish tbx6^−/− larvae, in which somite boundaries do not form during embryogenesis, were shown to exhibit abnormal skeletal morphology such as rib, neural arch and hemal arch. In this study, we investigated the role of somite patterning in the formation of anterior vertebrae and ribs in more detail. Using three‐dimensional computed tomography scans, we found that anterior vertebrae including the Weberian apparatus were severely affected in tbx6^−/− larvae. In addition, pleural ribs of tbx6 mutants exhibited severe defects in the initial ossification, extension of ossification, and formation of parapophyses. Two‐colour staining revealed that bifurcation of ribs was caused by fusion or branching of ribs in tbx6^−/−. The parapophyses in tbx6^−/− juvenile fish showed irregular positioning to centra and abnormal attachment to ribs. Furthermore, we found that the ossification of the distal portion of ribs proceeded along myotome boundaries even in irregularly positioned myotome boundaries. These results provide evidence of the contribution of somite patterning to the formation of the Weberian apparatus and rib in zebrafish.

X‐ray micro‐CT scans revealed the role of somite patterning in the formation of Weberian apparatus. In addition, myotome boundaries may provide a positional cue for the ossification of pleural ribs in zebrafish larvae.

Introduction

The vertebral column is a defining feature observed in the body of the vertebrates and comprises repetitive units of vertebrae along the anterior–posterior axis. This periodic structure of vertebrae originates from embryonic segments, somites. Somites are bilateral repetitive units of paraxial mesoderm that are sequentially segregated from the anterior‐most region of presomitic mesoderm during embryogenesis. Somite segmentation plays a crucial role in the proper formation of vertebrae and ribs in vertebrates. Spondylocostal dysostosis, a rare human genetic disorder, characterized by vertebral segmentation defects and malformations of ribs, is caused by autosomal recessive mutations of several genes related to somite segmentation such as DLL3, HES7, MESP2, LFNG, TBX6 and RIPPLY2 (Turnpenny et al. 2003; Whittock et al. 2004; Sparrow et al. 2010; Sparrow et al. 2013; McInerney‐Leo et al. 2015; Otomo et al. 2019). In mice too, loss‐of‐function mutations of these orthologous genes recapitulate the phenotype observed in spondylocostal dysostosis (Kusumi et al. 1998; Bessho et al. 2001; Serth et al. 2003; White et al. 2003; Morimoto et al. 2007; Makino et al. 2013). After somite segmentation in avian embryos, some of the somitic cells undergo an epithelial‐to‐mesenchymal transition that gives rise to sclerotome cells. Subsequently, calcification of vertebrae is mediated by the endochondral ossification in which the cartilage is eventually replaced with mineralized bone.

Somite patterning also plays an important role in the proper formation of vertebrae in zebrafish, although, in contrast to the endochondral ossification in higher vertebrates, the ossification of vertebrae and ribs in zebrafish is mediated by the intramembranous ossification in which mesenchymal cells condense and differentiate directly into osteoblasts (Bird & Mabee, 2003). fused somites (fss)/tbx6 mutants, which apparently lack all the somite boundaries during embryogenesis, exhibited abnormal morphology of vertebrae, neural arches which extend toward the dorsal, and hemal arches which extend toward the ventral (van Eeden et al. 1996; Nikaido et al. 2002; Fleming et al. 2004). In previous studies, skeletal analysis using specimens stained by alizarin red did not investigate anterior vertebrae (van Eeden et al. 1996; Fleming et al. 2004), thus, the contribution of somite patterning to the formation of anterior vertebrae remains unknown in zebrafish. In anterior vertebrae, Otophysan fishes including zebrafish possess the specialized structure of axial skeleton, Weberian apparatus, which mechanically conveys auditory information and pressure change from the swim bladder to the inner ear (Sanger & McCune, 2002; Grande & Young, 2004). In the present study, using X‐ray micro‐computed tomographty (CT) scans, we show that specific bones, including Weberian ossicles, are significantly affected in tbx6 homozygous mutants. In addition, we successively compared the ossification of ribs between zebrafish wild‐type and tbx6^−/− larvae in more detail and show that pleural ribs in tbx6 mutants exhibit severe defects in the initial ossification, extension of ossification, and formation of parapophyses. Furthermore, we show that myotome boundaries may provide a positional cue for the ossification of pleural ribs in zebrafish larvae.

Materials and methods

Ethics statement

All the experiments using live zebrafish were approved by the Committee for Animal Care and Use of Saitama University and were conducted in accordance with regulations.

Husbandry of zebrafish

All the experiments except mutant analysis were performed using Riken WT, which were maintained at 27°C with a 14‐h light/10‐h dark cycle. Embryos were obtained from natural spawning and the larvae were maintained at 28.5°C. Post‐embryonic stages were determined by standard length, which corresponds to the distance from the snout to the caudal peduncle (Parichy et al. 2009). The allele of fss/tbx6 mutants used in the present study was fss^ti1, which behaves as a null mutation of a T‐box gene, tbx6 (previously designated tbx24) (van Eeden et al. 1996; Nikaido et al. 2002). The genotype of tbx6 was determined as previously described (Kinoshita et al. 2018).

X‐ray micro‐computed tomography analysis

Adult male fish (ca. 1 year 4‐month old) were fixed in 4% paraformaldehyde in phosphate‐buffered saline (PBS) at 4°C overnight and were transferred to 70% ethanol. Using an X‐ray micro‐CT (ScanXmate‐E090S105; Comscantechno), these fixed specimens were scanned at a tube voltage peak of 85 kV and a tube current of 90 μA. For scanning of the Weberian apparatus, samples were rotated 360 degrees in steps of 0.24 degrees, generating 1500 projection images of 992 × 992 pixels. These micro‐CT data were reconstructed with coneCTexpress software (Comscantechno) at an isotropic resolution of 4.2 µm. For the whole scanning of the zebrafish with higher resolution, the entire body was scanned in three parts for one specimen. For each part, samples were rotated 360 degrees in steps of 0.3 degrees, generating 1200 projection images of 992 × 992 pixels. The micro‐CT data of each part were reconstructed and stored as a dataset with an isotropic resolution of 10 μm. Finally, the data of three sites were combined to generate the dataset of the whole specimen. Using Imaris software (Carl Zeiss), the dataset of each specimen was saved as TIFF volume data. These data were used for the next three‐dimensional (3D) image analysis. Volume rendering images of skeletons of each specimen were created using OsiriX MD software (Pixmeo). From the same dataset, 3D free‐move and free‐rotate surface models were created using Imaris and CINEMA 4D (MAXON Computer). These models were deposited to Sketchfab (https://sketchfab.com) and available for the following link: https://skfb.ly/6PtNw for whole‐body bone structure of sibling zebrafish; https://skfb.ly/6PtNv for whole‐body bone structure of zebrafish tbx6^−/−; https://skfb.ly/6PtNu for the Weberian apparatus of sibling zebrafish; ://skfb.ly/6NEZz for the Weberian apparatus of zebrafish tbx6^‐/‐.

Skeletal staining of live zebrafish by calcein

Skeletons of zebrafish larvae were visualized by a fluorescent chromophore, calcein (Dojindo), essentially as described previously (Du et al. 2001). Briefly, prior to calcein staining, fish were fasted for 24 h to avoid background signals in the digestive systems. Then, the fish were stained with 0.2% calcein in 1/3 Ringer’s solution (39 mm NaCl, 0.97 mm KCl, 0.6 mm CaCl₂, 1.67 mm HEPES, pH 7.2) for 10 min. After washing five times with 1/3 Ringer’s solution for 10 min each time, the larvae were anaesthetized with tricaine methanesulfonate (MS222; Sigma Aldrich) and mounted on glass slides with 2% methyl‐cellulose in distilled water. The stained larvae were observed under a fluorescence stereomicroscope (Leica MZ FLIII) with a digital camera (Leica DFC350F). After taking photos, the stained fish were carefully washed with fresh water at least four times to remove methyl‐cellulose and then separately maintained in each well of a six‐well plate at 28.5°C. After a few days, calcein staining and observations of zebrafish larvae were performed again as described above.

Successive staining of skeletons by calcein and alizarin red S

For successive staining, skeletons of live larvae were first visualized by calcein as described above. After 3 days, the larvae were stained with 0.01% alizarin red S (Wako) in 1/3 Ringer’s solution for 60 min as previously described (Bensimon‐Brito et al. 2016). After washing five times with 1/3 Ringer’s solution for 10 min each time, ribs of the anaesthetized larvae were observed under a fluorescence stereomicroscope.

Calcein staining and immunostaining of myotomes

Skeletons of live zebrafish larvae were first stained with 0.2% calcein in 1/3 Ringer’s solution for 10 min. After observation of the calcein signal, larvae showing the ossification of ribs were selected for immunohistochemistry. After washing five times with 1/3 Ringer’s solution for 10 min each time, the anaesthetized larvae were fixed in 4% paraformaldehyde in PBS at 4°C overnight. The fixed larvae were then dehydrated with methanol and stored at −20°C for at least 3 h. After rehydration, the fixed specimens were reacted with MF20 mouse monoclonal antibody (Hybridoma Bank) at a 1:4 dilution in 2% bovine serum albumin in PBS. After washing six times with PBS/0.25% TritonX‐100 for 15 min each time, the larvae were incubated with Alexa647‐conjugated anti‐mouse IgG secondary antibody (Abcam) at a 1:250 dilution in 2% bovine serum albumin in PBS. After washing six times with PBS/0.25% TritonX‐100 for 15 min each time, specimens were mounted in 1% low‐melting agarose gel (Sigma Aldrich). For the detection of calcein and MF20‐positive signals, fluorescent images were captured by a confocal microscope (Olympus, FV1000).

Laser ablation of the myotome boundary

The myotome boundary was injured using the infrared laser equipped with IR‐LEGO systems (Sigma Koki). Zebrafish larvae (ca. 1 month post‐fertilization) that did not yet exhibit the calcification of pleural ribs by calcein staining, were used. Briefly, zebrafish larvae anaesthetized with MS222 were mounted on glass slides with 2% methyl‐cellulose in distilled water and were placed on the microscope. Irradiation was carried out with an infrared laser (80 mW for 1 s) at room temperature. After the injury of the myotome boundary, zebrafish larvae were maintained at 28.5°C. After several days, calcein staining was carried out to visualize calcified ribs in the treated larvae as described above.

Results

Previous studies using specimens stained by alizarin red did not describe anterior vertebrae (van Eeden et al. 1996; Fleming et al. 2004). To investigate the role of somite segmentation in the zebrafish vertebrae, we performed X‐ray micro‐CT scanning for adult wild‐type and tbx6 ^{− / −} fish (Fig. 1 and Videos S1 and S2). In the anterior region of zebrafish, the Weberian apparatus is a bone structure that is composed of the anterior‐most four vertebrates and four ossicles: tripus, intercalarium, scaphium and claustrum (Sanger & McCune, 2002; Grande & Young, 2004). Among them, we found that tripus, comprising fan‐shaped bones located on both sides of ventral to the third centrum, was severely reduced and irregularly twisted in tbx6 ^{− / −} fish (Fig. 1B and Video S2). By contrast, the other Weberian ossicles (intercalarium, scaphium and claustrum), which are located to the lateral or dorsal sides of the first to third centra were not apparently affected in tbx6 ^{− / −} fish. Among other bones except for the Weberian apparatus, lateral processes, which are bilaterally extending arched ossicles from the first and second vertebrae, were found to be significantly shortened in tbx6 ^{− / −} fish. Although the transformer process of vertebra 4 and the os suspensorium are both ventrally extending ossicles from lateral sides of the fourth centrum, we found that the ventral region of os suspensorium was severely shortened whereas the transformer process of vertebra 4 was not significantly altered and was properly bifurcated at the tip in tbx6 ^{− / −} fish. These results suggest that the somite segmentation of zebrafish embryos selectively contributes to the formation of ossicles in anterior vertebrae.

Figure 1.

Micro‐computed tomography (CT) scan analysis of sibling and tbx6 ^{− / −}adult fish. As a control, tbx6 heterozygous sibling male fish, which showed a phenotype indistinguishable from wild‐type fish, were used. Two adult male fish for each genotype were analyzed by micro‐CT scan. More detailed data are provided with Videos S1 and S2 and are also deposited in the web sites (See Materials and methods). The background of tbx6 mutant used in micro‐CT scan analysis was TL2 (Kishimoto et al. 2004). Scale bars are 3 mm in (A) and 0.5 mm in (B). (A) Micro‐CT scan images of bone structures of the whole body in sibling and tbx6 ^{− / −}adult fish. (B) Micro‐CT scan analysis showing abnormal morphology of anterior vertebrae in tbx6^‐/‐adult fish. Scaphium (sc), intercalarium (ic), supraneurals (sn), tripus (tri), os suspensorium (os), transverse process of vertebra 4 (tp4), lateral process (lp), neural spine (ns), claustrum (cl) and pleural rib (pr) are indicated in panels.

We next focused our analysis on abnormal development of ribs in tbx6 ^{− / −} fish. To successively compare the ossification of ribs between wild‐type and tbx6 ^{− / −} larvae, we stained calcified bones using calcein (Fig. 2). Around 5.5 mm standard length, the ossification of ribs occurred slightly distal to the centra in wild‐type larvae (arrowhead in Fig. 2A) as previously described (Bird & Mabee, 2003; Parichy et al. 2009). In contrast, initial ossification of ribs in tbx6 ^{− / −} larvae was detected in several lateral pleural regions; the ossification of ribs in some cases was initiated distal to the centra (arrow in Fig. 2D) and that of other ribs was initiated slightly distal to the centra (arrowhead in Fig. 2D), as was observed in wild‐type larvae (Fig. 2A). Subsequently, the ossification of ribs progressively extended toward the ventral region to form the rib cage and also extended slightly toward the centra in wild‐type larvae (Fig. 2A–C) (Bird & Mabee, 2003). In tbx6 ^{− / −} larvae, however, ossification of some ribs was significantly delayed in comparison with that in wild‐type fish (arrowhead in Fig. 2A–F). In addition, we found that ossification of ribs often proceeded in an irregular direction, unlike the almost straight extension of rib ossification in wild‐type larvae (Fig. 2A–F). As was described previously (van Eeden et al. 1996; Fleming et al. 2004), we frequently observed bifurcation of the ribs in tbx6 ^{− / −} larvae; at least one bifurcation of ribs was detected on either the left or right side of the ribs (100%; n = 30). In some cases, a calcified sheet‐like structure was formed between ribs of tbx6 ^{− / −} larvae (asterisk in Fig. 2F).

Figure 2.

Abnormal skeletal patterning of pleural ribs in somite segmentation mutants. (A–F) Successive calcein staining revealed the ossification process of ribs in wild‐type and tbx6 ^{− / −} larvae. Live zebrafish larvae (wild‐type, 5.8–7.5 mm standard length; tbx6 ^{− / −}, 6.3–7.7 mm standard length) were stained by calcein at an interval of a few days. For the arrow, arrowhead and asterisk, see the Results section. Anterior is to the left and dorsal is to the top in all lateral views. Scale bar, 250 µm

Bifurcation of ribs can occur by the fusion of two ribs or the branching of a rib. To examine how ribs are bifurcated in tbx6 ^{− / −} larvae, we traced the ossification of ribs using two different staining reagents. In a few days after calcified ribs had been first stained with calcein, the calcified ribs were labelled again with alizarin red S. Two‐colour staining of ribs in wild‐type juvenile showed that the ossification occurred in a straight fashion (Fig. 3A–C; n = 7). By contrast, two‐colour staining of ribs in tbx6 ^{− / −} larvae clearly revealed that the fusion of two adjacent ribs occurred by irregular extension of ossified ribs (Fig. 3D–F; n = 3). In addition, we observed a split of the rib in tbx6 ^{− / −} larvae (Fig. 3G–I; n = 3); thus, our results revealed that both fusion and branching of ribs lead to the bifurcation in tbx6 ^{− / −} juvenile fish.

Figure 3.

Fusion and branching of pleural ribs in tbx6 ^{− / −} larvae. (A–I) Visualization of the ossification process of ribs in wild‐type (A–C) and tbx6 ^{− / −} (D–I) larvae by calcein (green) and alizarin red S (red) staining. The ribs of live larvae were first stained by calcein. In 3 days, ribs were stained again by alizarin red S. Fluorescent signals were photographed after alizarin Red S staining. In merged images, the already ossified region of ribs is indicated by a yellow signal and the newly calcified region of ribs is indicated by a red signal. Arrowheads in (D–F) show fusion of ribs and arrowheads in (G–I) show the branching of ribs in tbx6 ^{− / −} larvae. The fluorescence in the digestive system is a background signal (asterisk). The standard lengths of larvae are 6.7 mm (A–C), 6.6 mm (D–F), and 6.7 mm (G–I). Lateral view and anterior is to the left. Scale bar, 250 µm

The parapophysis, which is a cup‐like joint attached to the rib and centrum, also showed abnormal morphology in tbx6 ^{− / −} larvae (Fig. 4). In wild‐type larvae, the calcified parapophysis was formed adjacent to the anterior‐most region of each centrum (Fig. 4A; Bird & Mabee, 2003). In contrast, the positioning of the parapophysis to the centrum was considerably varied in tbx6 ^{− / −} larvae (Fig. 4B–D); 51% of parapophyses were restricted to the anterior region, as was observed in wild‐type larvae, but 18% of them were located in the middle region and 31% were located in the posterior region (n = 67). In addition, parapophyses in tbx6 ^{− / −} larvae appeared to be larger than those in wild‐type larvae. Parapophyses were normally attached to the fully extended dorsal ribs (Fig. 4A). However, in tbx6 ^{− / −} larvae, we often observed that the parapophysis exhibited a prickle‐like structure (Fig. 4B). The two adjacent parapophyses were jointed to one rib in tbx6 ^{− / −} larvae (Fig. 4C). Furthermore, the fusion of two adjacent parapophyses was seen in some tbx6 ^{− / −} larvae (Fig. 4D). These observations suggest that the somite patterning regulates proper positioning and formation of the parapophysis in zebrafish embryos.

Figure 4.

Oblique views of parapophyses in wild‐type (A) and tbx6 ^{− / −} (B–D) larvae. Skeletons of wild‐type and tbx6 ^{− / −} larvae were stained by calcein and regions surrounding the parapophyses were observed. The standard lengths of stained larvae are 7.3 mm (A), 7.1 mm (B), 6.9 mm (C), and 7.0 mm (D). Arrowheads in (A) represent normal parapophyses in a wild‐type larva. Arrowheads in (B) show parapophyses joined to a prickle‐like short rib in a tbx6 ^{− / −} larva. Arrowhead in (C) shows the fusion of two parapophyses. Arrowhead in (D) shows two parapophyses attached to one short rib. Scale bar, 250 µm

Next, we observed the skeletal patterning of pleural ribs in zebrafish (Fig. 5). In the zebrafish body, a repetitive structure of the myotome, which is derived from somites, is already evident along the anterior‐to‐posterior axis prior to the ossification of ribs. By comparing the calcein signal with the signal visualized by the MF20 antibody, which recognizes myosin heavy chains, we found that the calcification of distal ribs occurred along the myotome boundaries in wild‐type larvae, though ribs proximal to centra did not appear to overlap (Fig. 5A,C,E). Although zebrafish tbx6 ^{− / −} embryos lack somite boundaries (van Eeden et al. 1996; Nikaido et al. 2002), tbx6 ^{− / −} larvae possessed irregular but robust myotome boundaries (van Eeden et al. 1996; Windner et al. 2012; Windner et al. 2015; Ban et al. 2019), which are thought to be recovered by the medial‐to‐lateral migration of slow muscle progenitor cells (Devoto et al. 1996; Henry et al. 2005). By comparing the ossified ribs with the myotomes, we found that the irregular patterning of distal ribs in tbx6 ^{− / −} larvae coincided well with the irregular myotome boundaries (Fig. 5B,D,F). Interestingly, the bifurcation of ribs overlapped with the region where the bifurcation of myotome boundaries was observed (arrowhead in Fig. 5B,D,F).

Figure 5.

Osiffication of ribs occurs along myotome boundaries. (A–F) Wild‐type (A, C, E) and tbx6 ^{− / −} (B, D, F) larvae were stained to visualize calcified ribs with calcein (green; A, B, E, F) and myotomes by MF20 monoclonal antibody (magenta; C, D, E, F). A representative image for each genotype is shown. The standard lengths of wild‐type and tbx6 ^{− / −} larvae shown in this figure are 7.9 and 7.8 mm, respectively. Arrowheads in (B, D, F) indicate the bifurcation of ribs along bifurcated myotome boundaries. (G, H) Laser‐induced injury of the myotome boundary resulted in the abnormal ossification of ribs. The infrared laser was spotted on the tailfin and the myotome boundary in control (n = 5) and injured larvae (n = 6), respectively. The fluorescent signal was imaged by a fluorescence stereomicroscope. Lateral view and anterior is to the left. Scale bar, 250 µm.

To examine whether the myotome boundary indeed affects the ossification of pleural ribs, wild‐type myotome boundaries prior to the ossification of ribs were injured using an infrared laser. In control‐treated larvae, the ossification of pleural ribs proceeded in a straight fashion as was observed in wild‐type larvae (Fig. 5G). In contrast, the larvae in which the myotome boundary corresponding to the first pleural rib was injured by a laser exhibited irregular ossification of pleural ribs (Fig. 5H). These results suggest that the myotome boundary provides an important positional cue for the extension of the ossification of ribs in zebrafish.

Discussion

In the present study, we performed micro‐CT scan analysis focusing on anterior vertebrae including the Weberian apparatus. Interestingly, tripus, lateral processes and transformer process of vertebra 4 were specifically influenced in tbx6 ^{− / −} larvae, although the other skeletal elements were not significantly altered (Fig. 1). In mice and chicken, cells in the rostral and caudal halves within a somite are known to subsequently differentiate into different regions of the vertebral column (Aoyama & Asamoto, 2000; Saga & Takeda, 2001). In tbx6‐deficient zebrafish embryos, the rostral half property is lost, and instead, the caudal half property is expanded (caudalization) (Nikaido et al. 2002; Ban et al. 2019). Although subsequent differentiation of rostral and caudal somitic cells into the future vertebrae is not strictly lineage‐restricted in zebrafish (Morin‐Kensicki et al. 2002), our results suggest that the rostral region of a somite preferentially contributes to the formation of anterior vertebrae including tripus, lateral processes and transformer process of vertebra 4, and further suggest the differential contribution of the rostral and caudal compartments of a somite to the anterior vertebrae in zebrafish.

We examined in detail the ossification process of ribs in juvenile tbx6 ^{− / −} mutants, which were shown to exhibit abnormal skeletal patterning of ribs (van Eeden et al. 1996; Fleming et al. 2004). As was the case in wild‐type larvae, we found that the irregular extension of calcified distal ribs overlapped well with the severely disorganized myotome boundaries in tbx6 ^{− / −} larvae (Fig. 5). In addition, we showed that injury of the myotome boundary was able to alter the path of ossification of ribs. We also observed that the bifurcation of ribs occurs in the region where the myotome boundaries are already bifurcated. Although the molecular mechanisms underlying this process remain unknown, we presume that the myotome boundary or cells located at the junction of myotomes attract the ossification of pleural ribs. Alternatively, as was observed in the segmental formation of intersomitic vessels in zebrafish (Torres‐Vazquez et al. 2004), the myotome may exert a repulsive force to the ossification of ribs. Interestingly, a recent study showed that zebrafish scleraxis encoding a bHLH transcription factor is expressed in the region adjacent to developing ribs and the disruption of scleraxis leads to the lack of rib mineralization (Kague et al. 2019). Combined with that study, the present observations suggest that myotome boundaries or adjacent cells provide positional information for proper extension of the calcified ribs in zebrafish.

Since ribs of zebrafish are thought to be of somitic origin, as is the case for avian embryos (Aoyama & Asamoto, 2000; Huang et al. 2000; Olivera‐Martinez et al. 2000), tbx6 ^{− / −} progenitor cells of ribs may possess some cell‐autonomous defects. Indeed, by calcein staining, we often observed that the ossification of ribs was hampered in tbx6 ^{− / −} larvae (Fig. 2). Severe defects in the formation of parapophyses and the positioning of parapophyses were found in tbx6 ^{− / −} larvae (Fig. 4). Thus, our results suggest that the deformities of ribs in tbx6 ^{− / −} larvae are not only caused by disorganized boundaries between myotomes, which are derived from the absence of somite boundary formation during embryogenesis, but are also attributable to the intrinsic abnormalities of rib progenitor cells and the deformation of parapophyses. Finally, it should be noted that neural and hemal arches, which also extend from the centra, were also severely disorganized in tbx6 ^{− / −} larvae (van Eeden et al. 1996; Fleming et al. 2004). We observed that myotome boundaries did not overlap with the neural and hemal arches in wild‐type and tbx6 ^{− / −} larvae, suggesting that another mechanism is involved in the patterning of neural and hemal arches in zebrafish.

Conflict of interest

None declared.

Supporting information

Video S1 . Whole‐body bone structure in sibling and tbx6^−/− adult zebrafish.

Video S2 . Weberian apparatus of sibling and tbx6^−/− zebrafish.