Zebrafish stm is involved in the development of otoliths and of the fertilization envelope

Summary Using an in vivo assay, we selected 11 genes that were highly upregulated during the induction of ovulation in zebrafish using microarray analysis and RNA sequencing. The starmaker gene (stm) was one of these genes. Although stm has been previously reported to be involved in otolith formation during the early development of zebrafish, we detected its expression in eggs and showed that stm was related to fertilization by establishing an stm gene knockout strain using the CRISPR/Cas9 system. Further phenotypic analysis of stm knockout fish was conducted in this study. With a higher nonfertilization rate, the stm mutant strain showed an extremely low survival rate. Otoliths of stm homozygous mutant zebrafish showed abnormal morphology in embryos and adult fish. However, fish did not show any abnormalities in swimming behaviour in either embryos or adults. Stm proteins were detected on the chorion of ovulated eggs before spawning. Fibre-supported knob-like structures on the fertilization envelope (FE) also showed abnormal structures in stm mutants. The Stm protein is necessary for otolith formation, and a lack of Stm causes abnormal otolith formation. The partial defect of otolith formation does not cause defects in swimming behaviour. The Stm protein is expressed in the chorion and is responsible for the formation of fibre-supported knob-like structures on the FE. It was suggested that a lack of Stm caused a lower fertilization rate due to inadequate formation of the FE. Lay summary In zebrafish, the protein Starmaker (Stm) was identified as having a role in ovulation. Stm is also known to be required for the formation of ear stones (otoliths) which are needed to keep the body in balance. Zebrafish lacking Stm were produced by genome editing. As expected, Stm-deficient fish formed abnormal otoliths. To investigate the role of Stm in ovulation, fertilization and early development, we tried mating of Stm mutants and observed their juveniles. Although no problem found in ovulation, we found low fertilization rate and abnormal structure of knob-like structure (small pit) on the egg membrane. Survival rate of embryos with abnormal egg membrane was extremely low. It was demonstrated that Stm protein is necessary to form the functional egg membrane to protect embryos from the outside environment.


Introduction
By using an in vivo induction method for oocyte maturation and ovulation in zebrafish by adding compounds to water (Tokumoto et al. 2011), we identified 11 genes that were highly upregulated during the induction of ovulation (Klangnurak et al. 2018). Then, we determined the functions of these genes by establishing gene knockout zebrafish strains of these genes. We have already established genome-modified fish for the starmaker (stm) gene. The Starmaker (stm) gene has been reported to be responsible for the formation of otoliths in zebrafish (Sollner et al. 2003).
Otoliths are important structural organs for equilibrium maintenance. Teleost fish, including zebrafish, possess a set of three types (conglomerate: Lapillus, flat stone: Sagitta, stellate: Asteriscus) of otoliths on each side of the body. Thus, there are six otoliths in each fish body (Platt 1993). Otoliths contain calcium carbonate and some matrix proteins, among which Stm is known to play an important role in changing the crystal structure of calcium carbonate from calcite to aragonite (Sollner et al. 2003). It is known that there are three types of crystal structures of calcium carbonate, calcite, aragonite, and vaterite, and calcite is the most stable crystal state at normal temperature and pressure. In mammals, calcite is the main component of otoliths, but in fish, aragonite is the main component. The stm protein is an acidic protein containing a large amount of aspartic acid, and some of the serine and threonine residues contained therein are phosphorylated. It is believed that this causes strong binding to calcium ions, which in turn causes a change in crystal structure from calcite to aragonite (Kalka et al. 2019).
Fish eggs develop a fertilization envelope (FE) that surrounds embryos and protects them from direct exposure to the outside water environment (Laale 1980, Grierson & Neville 1981, Harvey et al. 1983, Cameron & Hunter 1984. A perivitelline space is formed after fertilization by the cortical reaction on the vitelline membrane (Ohta & Nashirozawa 1996). The formation of this perivitelline space is related to blastodisc formation. The liquids between the vitelline membrane and the FE protect the egg against physical impacts with the external environment and are responsible for a gas exchange through diffusion (Donovan & Hart 1986). There have been no reports on the relationship between Stm and FE formation. In this study, we demonstrated that Stm is responsible for the proper formation of FE.
Although the fertilization rate decreased, ovulation could be induced in our stm homozygous mutant fish, which demonstrated that stm was not an ovulationinducing gene (Klangnurak et al. 2018). In this study, further analysis of the stm mutant strain was conducted. Abnormal otolith shapes similar to the previous report of the knockdown experiment with antisense morpholino oligo was observed (Sollner et al. 2003). However, abnormal otolith morphology, which is important in fish equilibrium, had no effect on swimming behaviour. We found abnormal formation of fibre-supported knob-like structures on the FE that might be responsible for the hardening of the FE.

Materials
Zebrafish (roy) were cultivated following the standard protocol (Westerfield 1995) and were kept in a flow system maintained at 28.5°C with a 14:10 light/dark cycle. Zebrafish were fed Paramecium spp. in the larval period for approximately 1 month, after which they were fed live brine shrimp in the morning and instant food (Tetra Guppy, Tetra GmbH, Melle, Germany) in the evening. In this study, the use of zebrafish and the experimental protocol for use were approved (approval no. 2019F-5 and 2020F-4) by the Institutional Ethics Committee of Shizuoka University, Japan.

Mutant line generation and phenotype observation
The stm genome-edited fish were established using CRISPR/Cas9 (Klangnurak et al. 2018). The F0 zebrafish were investigated for mutations using heteroduplex mobility assay (HMA) (Foster et al. 2019). Heterozygous mutant zebrafish of stm (stm +/− ) were established by pairing F0 mutant zebrafish and WT zebrafish. The same strain of stm +/− was inbred to produce a homozygous mutant (stm −/− ) in the F2 generation. The fertility of stm −/− was investigated by pairing the zebrafish with each other or with WT fish. In the evening, stm −/− males and females were placed in the breeding tank, which has a small cage inside the large tank. The bottom of the small tank has a net through which the fertilized eggs can pass through to the floor to prevent the adult fish from eating them. The next morning, the embryos were collected in the petri dish from the large tank. The embryos were counted, and the morphological characteristics of unfertilized, abnormal and surviving embryos were checked under a stereomicroscope. The fry (5 dpf) were released to the small tank and fed Paramecium spp. for at least 2 weeks. Then, the fish were fed live brine shrimp in the morning and instant food (Tetra Guppy, Tetra GmbH, Melle, Germany) in the evening. Strains of the stm −/− mutant were propagated by pairing zebrafish with the same genotypes with each other. However, when no fertile females were obtained, heterozygous or WT females were used for pairing to continue the strain.

Immunohistochemistry
Before spawning, female zebrafish containing ovulated eggs on the posterior side of the abdomen were sacrificed by cervical spine destruction and fixed in 4% paraformaldehyde (PFA) at 4°C overnight and soaked with 30% sucrose for 3 h or until the samples sunk. The fixed fish body was transferred to an embedding chamber and embedded with Tissue-tek O.C.T. compound. The embedding chamber was dipped into liquid N 2 . Samples were cut into 10 μm thick sections on a cryostat microtome (CryoStar NX70, Thermo Fisher Scientific) at −20°C (Westerfield, 1995). The cut samples were transferred to slides and a PAP pen (Daido Sangyo Co. Ltd., Tokyo, Japan), which disperses a hydrophobic material, was used to circumscribe the section. The slides were gently washed three times with TPBS (0.1% Tween 20 in PBS solution; 0.8% NaCl, 0.02% KCl, 0.02 M PO 4 , pH 7.3) for 2 min. The solution was removed and replaced with blocking solution (5% nonfat milk in PBS buffer; 0.8% NaCl, 0.02% KCl, 0.02 M PO 4 , pH 7.3) for 30 min under a dark cover with distilled water-moistened paper. The slides were washed with TPBS for 5 min three times and incubated with anti-zebrafish starmaker MAB (Abmart Inc. Shanghai, China) diluted 100-fold in PBS at 4°C overnight in the dark. The samples were washed with TPBS for 10 min three times and the antibody was replaced with Alexa Fluor 555-conjugated anti-mouse immunoglobulin (Cell Signaling Technology) for 1 h at room temperature in the dark. The experimental slides were washed with TPBS 10 min three times, treated with prolonged reagent, and observed by confocal laser microscopy (LSM700, Carl Zeiss).

Stereoscopic observation
A stereomicroscope (Olympus SZX12) and a microscope camera (Olympus DP70) were used for observation and photography. The otoliths of WT and stm mutant fry were observed and photographed under a microscope. Otoliths (pebble stones: Lapillus, flat stones: Sagitta, star stones: Asteriscus) were excised from the WT and stm mutant adult fish and photographed.

Scanning electron microscope observation
Otoliths (pebble stones: Lapillus, flat stones: Sagitta, stellates: Asteriscus) were extracted from adult WT and stm −/− mutants. The FEs were removed from eggs in water and washed with ultrapure water. Eggs with FEs were washed with ultrapure water. Samples were washed with ultrapure water and then critical point dried using a freeze dryer (Aqua FD-6500; SUN Technologies, Kyoto, Japan). Dried samples were coated with 5 nm platinum using an autofine coater (JEC-3000FC; JEOL, Tokyo, Japan). Subsequently, the FEs were observed with s.e.m. (JSM-6510LV; JEOL).

Micro-CT imaging
WT and stm mutant adult zebrafish were fixed with 70% ethanol and stored in 70% ethanol. The heads of zebrafish were scanned using an X-ray micro-CT device (ScanX-mate-E090S105, Comscantechno Co., Ltd., Japan) at a tube voltage peak of 60 kVp and a tube current of 100 μA. The sample was rotated 360° in steps of 0.24°, generating 1500 projection images of 992 × 992 pixels. The micro-CT data were reconstructed at an isotropic resolution of 13.3 × 13.3 × 13.3 μm. Three-dimensional tomographic images were obtained using OsiriX MD software (version 9.0, Pixmeo, SARL, Switzerland) and Imaris software (version 9.1, Carl Zeiss Microscopy Co., Ltd.). Supplementary video (Supplementary video 1, see section on supplementary materials given at the end of this article) was edited using Adobe Premiere Pro CC (Adobe Systems Co., Ltd., Japan).

Statistical analysis
Summary data are presented as the mean ± s.d. Student's t test was used to determine the statistical significance of the difference in data. Data were considered significant at *P < 0.05 or **P < 0.001.

Results
A mutant strain of the stm gene with deletion of 14 bps and insertion of 1 bp within exon 7 was found (Fig. 1A). While the WT zebrafish Stm protein is 613 amino acids in length, it was presumed that the stm −/− mutant expressed the peptide of 98 amino acids in length with the same sequence of the first 76 amino acids as the WT (Fig. 1B). However, significantly lower expression of stm mRNA was detected in stm −/− mutants by qPCR analysis (Fig. 2A). Thus, it was suggested that even truncated form of Stm protein was not expressed in stm −/− mutants.
Because we selected stm gene as an ovulationrelating gene that showed significant increase of mRNA expression during ovulation. Thus, it is highly possible that mutations in stm gene show maternal effects. Then we used embryos and fishes obtained from paring of stm −/− mutants in all the phenotypic analysis below. The survival rate of embryos was significantly low in embryos from the pairing of stm −/− mutants due to the high nonfertilization rate (Fig. 2B). Two otoliths were observed in both WT and mutant embryos (Fig. 2C). In the mutants, star-shaped otoliths, which gave rise to the gene name (starmaker), were observed in all the embryos examined. In the adult fishes, three types of otoliths, pebbles (Lapillus), flat stones (Sagitta), and stellates (Asteriscus), developed in the stm −/− mutant and the WT fish. However, the shape of these otoliths was abnormal in all the stm −/− mutants fishes examined (more than five of females and males) (Fig. 3). The shape of the Lapillus in the stm −/− mutant was significantly different from that in the wild type. Lapillus in the WT showed a smooth surface but the surface in the stm −/− mutant was rough. Although relatively similar shapes were observed in the case of Sagitta and Asteriscus, the surfaces of these stones were also uneven in the stm −/− mutant compared with the wild type. The differences were more clearly observed using scanning electron microscopy (s.e.m.) (Fig. 4). Pebbles appeared as crystal clusters in the stm −/− mutant. Sagitta and Asteriscus showed layered structures in the stm −/− mutant. In contrast, the surfaces of all three stones were smooth in the wild type.
Then, we tried to observe otoliths in vivo by micro-CT scanning (Fig. 5). Although the shapes of the three types of otoliths were different, the placement of otoliths was similar between the WT and the stm −/− mutant. The three- dimensional arrangement of otoliths in the WT and the stm −/− mutant can be compared in the Supplementary movie (sMovie 1). As mentioned earlier, the morphology of otoliths in the stm −/− mutant was abnormal, as expected. However, the stm −/− mutant zebrafish did not show any abnormality in behaviour (Supplementary videos 2 and 3). We tried to check the movement using tactile stimulation in embryos (Supplementary video 2) (Yang et al. 2011). Additionally, we checked female chasing behaviours during mating in adulthood (Supplementary video 3). We could not detect any difference in movement between WT and the stm −/− mutants. In tactile stimulation in embryos, all the embryos examined (total 70 embryos) from three different pairs of stm −/− mutants showed no abnormal behaviour. In adult, 7 of stm −/− male showed normal female chasing behaviours.
In addition to the morphology of otoliths, we had already found that eggs from stm −/− mutants showed a high nonfertilization rate (Klangnurak et al. 2018). By paring stm −/− mutants in later generations (F3 to F5), we obtained high percentages of unfertilized eggs. Thus, only a few juvenile survivors were obtained from the F3 generation and later generations. The same results were confirmed in the F6 generation (Fig. 2B). However, few survivors showed fertility, and we could continue the strain. To address the reason for unsuccessful fertilization, we conducted immunohistochemical staining of oocytes and eggs. In the sections of whole body sectioning of fish before spawning, which contained ovulated eggs and immature oocytes, strong Stm signals were detected on the chorion of ovulated eggs in WT (Fig. 6). In the sections of immature oocytes that were located in the anterior position in the ovary, the nuclei of follicular cells were stained with DAPI, but no Stm protein signal  was observed. In the sections of ovulated eggs that were located in the posterior position in the ovary, signals of the binding of anti-Stm antibody were observed on the chorion. The absence of DAPI signals around the chorion confirmed the removal of follicular cells by ovulation. Contrary no signal of binding of anti-Stm antibody were observed on the chorion of ovulated eggs in stm −/− mutants. The results suggested that Stm was expressed during ovulation and accumulated on the surface of ovulated eggs. Furthermore, it was confirmed that the absence of Stm in stm −/− mutants.
From the expression of Stm on the chorion, it was expected that Stm plays a role in the formation of the FE. Thus, we observed the FE. On the surface of the FE, many crystal-like structures were found on the WT and only an extremely small number of structures were found on the FE of the stm −/− mutants by microscopic observation (Fig.  7). Additionally, there were many cracks in the FE of stm −/− mutants but not in the WT. For a more in-depth look at the crystal-like structures, we observed the FE by s.e.m. It was confirmed that the crystal-like structures on the surface of the FE were fibre-supported knob-like structures reported before (Joo & Kim 2013), and these structures were present on both sides of the FE (Fig. 8). In stm −/− mutants, knoblike structures appear to be crushed and spread. Presence or absence of Stm in knob-like structures in WT or stm −/− mutants were confirmed by the immunohistochemical staining of FE. Signals of the binding of anti-Stm antibody were observed in knob-like structures on the FE of WT (Fig. 9). On the contrary, no signal of binding of anti-Stm  antibody were observed on the FE of stm −/− mutants. It was indicated that Stm is necessary for the formation of fibresupported knob-like structures on FE.

Discussion
In this study, phenotype analysis was conducted on a stm mutant strain with a different sequence from 77 aa to 98 aa. We observed structures that closely resembled the shape of otoliths in the fry of our mutant strain, which is similar to that of stm knockdown individuals injected with moderate amounts of MO (Sollner et al. 2003). Particular morphology of the star-like structure was found in Lapillus, the most anterior otolith among the three otoliths (Baxendale & Whitfield 2016). The location of otoliths in stm −/− mutants was close to that in WT. The stm −/− mutants did not show any deficiency in movements from embryo to adulthood. Thus, morphological changes in otoliths caused by the lack of Stm do not cause the problems in maintaining balance.
In this study, we found strong expression of Stm on the chorion. Additionally, we found that Stm is  necessary for the formation of fibre-supported knob-like structures on the surfaces of the FE (Joo & Kim 2013). It was suggested that the formation of fibre-supported knoblike structures is necessary for a fully functional FE and that improper FE formation results in nonfertilization. We judged abnormal formation of single cells on the animal pole as unfertilized eggs. However, it is possible that the phenotype is also caused by the insufficient formation of the FE. These results suggested that Stm is necessary not only for otolith formation but also for the formation of the FE. The microscopic observation showed cracks in the FE. During the preparation of samples for s.e.m. observation, decompression treatment caused wrinkles in the FE of stm mutant eggs, whereas the FE of WT remained smooth.
These results suggested that Stm may have a role in the hardening of the FE.
As the FE is a barrier for embryos from the outside environment, its permeability to substances has been analysed for a long time. As a result, it is generally accepted that the FEs shield the zebrafish embryo against solutes in the exposure solution, as it possesses pores of an average diameter of 1.5 μm that occur at distances ranging from 1.5 to 3 μm (Laale 1977, Rawson et al. 2000, Lee et al. 2007. Additionally, transparencies of toxic compounds were tested, and the most significant example was atrazine. Wiegand et al. have shown in uptake studies using 14 C-labelled atrazine that atrazine penetrates the FE very quickly and reaches the zebrafish embryo within the first 10 s after the onset of exposure (Wiegand et al. 2000). Creton showed the passage of 3 kDa fluorescently labelled dextrans across the FE but a restriction of 10 kDa fluorescent dextrans (Creton 2004). More recently, Lee et al. established an in vivo imaging method to observe the entry of silver nanoparticles (Ag NPs) through the FE (Lee et al. 2007(Lee et al. , 2012. They showed that single Ag NPs (30−72 nm diameters) passively diffused into  the embryos through pores in FE via random Brownian motion and stayed inside the embryos throughout their entire development and caused dose-and size-dependent toxic effects. These results indicated that the FE is a sieve with an approximate pore size of 3000 Da. The most remarkable report concerning the function of the FE was the absorption of Cd 2+ . Burnison et al. demonstrated that 61% of the total absorbed Cd 2+ was bound to the FE and that only a small proportion of 1% was found in the zebrafish embryo during a 4-h exposure by using 109 Cd (Burnison et al. 2006). In our stm knockout zebrafish, the formation of fibre-supported knob-like structures was disrupted. From the known functions of Stm as a Ca 2+ -binding protein, it can be hypothesized that a lack of Stm caused a spreading form of knob-like structures. This caused a need to change the function of FE and resulted in changes in the components or conditions of the fluid. Further analysis of the composition of fluid and the strength and substance permeability of FE in stm −/− mutants are necessary for the future.
In stm −/− zebrafish, mutants showed abnormal otolith structures. However, no obvious deficiencies in the movement were observed. We showed for the first time that Stm is involved in the formation of fibre-supported knob-like structures on FE. Although the functions of this structure remain to be solved, altering the strength and substance permeability of the FE might be the cause of unfertilized eggs.

Supplementary materials
This is linked to the online version of the paper at https ://do i.org /10.1 530/ R AF-20 -0040 .

Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this study.

Funding
This study was supported by a grant from JSPS KAKENHI (Grant Number 16K07419 and 20K06719) (to T T). This work was also supported by Grants-in-Aid for Scientific Research in Priority Areas from the Ministry of Education, Culture, Sports, Science and Technology of Japan.
Author contribution statement T P drafted the manuscript. K W conducted genome editing. K W, T F, Md. R, T P, Md. M R and Md. H A established the strains and maintained them. T F and T P performed immunofluorescent observation. C Y, T P and T P conducted electron microscope observations. A M performed micro CT scan observations. T T participated in the study design, supervised the study and wrote the paper. All authors read and approved the final manuscript.