Genome Editing in the Olive Flounder (Paralichthys olivaceus) Using CRISPR/Cas9 and a Simple Microinjection System

TAN Xungang, WANG Ling, WU Zhihao, JIAO Shuang,WANG Lijuan, ZOU Yuxia, JIANG Jingteng, and YOU Feng, *

Genome Editing in the Olive Flounder () Using CRISPR/Cas9 and a Simple Microinjection System

TAN Xungang1), 2), #, *, WANG Ling1), 2), 3), #, WU Zhihao1), 2), JIAO Shuang1), 2),WANG Lijuan1), 2), ZOU Yuxia1), 2), JIANG Jingteng4), 5), and YOU Feng1), 2), *

1),,,,266071,2),,266071,3),100049,4),264200,5),.,264200,

The whole-genome sequence of the olive flounder () provides a basis for gene functional analyses, which is important for the aquaculture industry. Understanding gene function will help us to select better economic traits such as fast growth and better culture conditions, which further will increase the aquaculture output. Gene knockout is an important reverse genetics approach forstudies of gene function. In this study, the CRISPR/Cas9 genome editing method with a microinjection sys- tem using a simple braked needle was employed in olive flounder. After injection in embryos, green fluorescent protein expression was detected in 40% of larvae. The proportion of normal-hatched larvae was approximately 50%. Different mutations, including short indels and fragment deletions, were found in our test genesand. Additionally, we detected more than one mutation in a single larva. In summary, our microinjection technique and CRISPR/Cas9 can be applied to study gene functions in olive flounder.

CRISPR/Cas9;; microinjection;; needle with brake; olive flounder

1 Introduction

Overexpression and knockdown/knockout are important approaches forstudies of gene function in many organisms. The key step in these methods is the deliveryof a small number of macromolecules, such as DNA, RNA,or protein, into cells or embryos. Many techniques have been applied to this step, such as microinjection, electro- poration, and transfection.

Microinjection can directly transfer macromolecules into individual eggs or embryos, so it is widely used in many species, such as mouse (Palmiter and Brinster, 1986),(Kobayashi and Satou, 2018), chicken, and(Moody, 2018).With genetic modification by generation of transgenics or gene knockouts, the gene function can be studied in these species (Du., 1992; Davis., 2000; Jao., 2013).

However, microinjection into fish eggs faces various chal-lenges, including its inability to optically locate the nucleus in fertilized eggs, the hard chorion of some fish species, and internal pressure in the egg. To overcome these difficulties, special methods and tools have been developed formicroinjection into fish eggs (Zhu., 1985, 1986; Cho- urrout., 1986; Ozato., 1986; Brem., 1988; Fletchr., 1988; McEvoy., 1988; Stuart., 1988; Goto., 2015, 2019; Kim., 2019). In marine flat- fish that spawn pelagic eggs, the hard and elastic chorion of marine pelagic egg, and the high internal osmotic pres- sure should be considered, as they can bring more chal- lenges to the microinjection. Microinjection can damage the eggs, which will reduce the larval survival rate. The status of the injected larvae is closely related to the survival rate during the metamorphosis of the larvae. During metamorphosis, culture conditions must be tightly control- led to achieve a high survival rate. Microinjection condi- tions should be carefully controlled too.Microinjection has been applied in marine flatfish with pelagic eggs, such as Chinese tongue sole () (Cui., 2017) and olive flounder () (Kim., 2019). For marine pelagic eggs, such as those of Nibe croaker () and red sea bream ()(Yamamoto., 2011; Kishimoto., 2018), microinjection was performed with an oil-driven micro- injector and a specialized needle equipped with a brake. For the Chinese tongue sole, microinjection was accom- plished with a simple needle and an air pressure-driven microinjector (Cui., 2017). For olive flounder (Kim., 2019), microinjection was performed with an air pres- sure-driven microinjector and specialized brake needle,which has also been used in sea bream (Kishimoto., 2018) and barfin flounder () (Goto., 2015, 2019). When using an oil-driven microinjector, re- searchers require much time to become familiar with the technique (Yamamoto., 2011; Kishimoto., 2018). For the preparation of a needle with a brake, specialized equipment, technique, and additional time are needed (Ki- shimoto., 2018; Goto., 2019). Therefore, a me- thod involving simple needle preparation and easy hand- ling of the microinjector should be developed for pelagic eggs.

The development of new technologies, including zinc- finger nucleases, TALEN (transcription activator-like effec- tor nuclease), and CRISPR/Cas9 (clustered regularly in- terspaced short palindromic repeats/CRISPR-associated proteins 9), has extended the application of genome edit- ing from the mouse model (Doetschman., 1987; Tho-mas., 1987; Thompson., 1989) to non-rodent mo- dels, including fish species, and promoted efficient genefunctional research (Gaj., 2013). In recent years, CRSIPR/Cas9 has been used more and more extensively in many fishes, such as zebrafish () (Hwang., 2013), Atlantic salmon (L)(Edvardsen., 2014), tilapia () (Zhang., 2014), Atlantic killifish ()(Aluru., 2015), medaka ()(Inoue., 2017),rainbow trout ()(Cleveland., 2018), pufferfish () (Kato-Unoki., 2018), fathead minnow ()(Maki., 2020), and Arabian Killifish ()(Hamied., 2020). However, genome editing has been success- ful only in three marine fish species: Chinese tongue sole using TALEN technology (Cui., 2017), red sea bream using CRISPR/Cas9 (Kishimoto., 2018), and olive flounder (Kim., 2019) using CRISPR/Cas9.

Growth of fish is important in aquaculture. For oliveflounder, fast growth might be achieved by improving mu- scle development,or by culturing more female fishas the female flounders grow faster than the males. However, mu- scle development and gonadal differentiation mechanisms are not very clear in olive flounder., a muscle- specific gene, is essential for myoblast fusion and muscle hypertrophy (Landemaine., 2014). Gonadal soma-de- rived factor (gsdf), a gonad-specific growth factor, plays an important role during early germ cell and testicular dif- ferentiation (Jiang., 2016). The functions of these two genes in olive flounder have not yet been characterized.

The olive flounder () is an economically im-portant fish in East Asia. A comprehensive characterization of gene function in this species will provide insight into its unique characters, such as the morphological change from bilaterally symmetrical to asymmetric after metamorpho- sis (Shao., 2017) and the behavior change from pe- lagic to benthic (Fuiman., 1997; Jiao., 2019). It will also improve the flounder industry by providing a ba- sis for genetic improvements, such as genome editing, trans- genesis, selection for critical alleles, and so on. Though genome editing with CRISPR/Cas9 has been successful in olive flounder, the microinjection system is too complex with a special technique (Kim., 2019). In this study, a new microinjection method using a simple needle was developed to introduce exogenous Cas9 and gRNA into fer- tilized eggs of olive flounder, which can be employed to edit the gene with a high larval survival rate.

2 Materials and Methods

2.1 Ethics Statement

Experiments with olive flounder (adult and larvae) were performed according to the regulations of local and cen- tral governments and were approved by the Institutional Animal Care and Use Committee of the Institute of Ocea- nology, Chinese Academy of Sciences.

2.2 Fish and Embryo Culture

Adult flounders were cultured at Shenghang fish farm in Rongcheng city, Shandong province, China. They were more than three years old and weighed more than 1.5kg per fish.They were kept at a stocking density of appro- ximately 50–60 fish/pond (16m2) (female:male=2:1), and were fed with fresh fish twice per day. Fertilized eggs were collected from the flounder spawning pond. Eggs were cultured in boiled seawater without aeration at 15℃±1℃ in a glass or plastic beaker.

2.3 Microinjection

The micropipette (Figs.1A, B) was prepared from glass capillaries (GD-1 or G-1; Narishige, Tokyo, Japan) using the HEKA pipette puller PIP5 and a micro grinder (EG- 400; Narishige). The micropipette was pulled in two steps. Firstly, the needle was pulled without breakage at a high temperature (from 1000 to 1200℃, depending on the room temperature and glass capillaries). Then, it was broken after pulling at low temperatures (from 700 to 900℃, depend- ing on the room temperature and glass capillaries). For the needle with a modified brake, the second heated site was moved a little apart from the first. Two different needles were pulled, and one of them was prepared with a brake. Both types of needles could be used for microinjection. If the sites for the first and second heating were the same, two of the same needles without brakes were prepared. The first three pairs of pulled needles were checked to ensure that they met the requirements under a dissecting micro- scope each time. Then, a batch of needles could be prepared all at once. Eggs were injected using a microinjector (Pli- 100; Harvard Apparatus, Holliston, MA, USA).

Fig.1 Microinjection of fertilized flounder eggs. A, normal needle for injection; B, needle with a brake for injection; Scale bar: 50μm; C, un-injected embryo; D, inject- ed embryo; Scale bar: 1mm.

Fertilized eggs (=10–20) were aligned in a groove of a 1.5% agar plate (filled with 1/3 seawater) (Goto., 2015,2019). The plasmid myod-green fluorescence protein (GFP) (Zhang., 2006) was dissolved in distilled H2O to a final concentration of 50μgmL?1with 0.01% phenol red.Approximately 1–2nL of myod-GFP plasmid with 0.0125%dextran rhodamine B (Thermo Scientific, Waltham, MA,USA) or CRISPR/mRNA (250ngmL?1) and gRNA(100ngmL?1) was injected into each embryo. Injectionswere performed at the protoplasmic-streaming-formation stage to one-cell stage (one-cell was not included, stage I), one-cell to two-cell stage (two-cell not included, stage II), and two-cell to four-cell stage (four-cell not included, stage III). Injected or un-injected (control) embryos were cultured in boiled seawater with 0.0002% methylene blue and 1ppm florfenicol without aeration at 15℃±1℃ in a 1-L glass or plastic beaker. Before hatching, newboiled seawater with florfenicol was added every day afterdead embryos were removed and counted. After they hatched, all larvae were anesthetized with 0.02% MS222, fixed in 4% paraformaldehyde, and observed under a Nikon fluores- cence microscope for analysis of GFP expression in- GFP-injected embryos. Moreover, immunohistochemistry was performed to confirm GFP expression. Normal and ab-normal larvae were counted. All injected embryos were reared in a secured land-based hatchery to prevent the es- cape of genetically modified organisms into the natural en- vironment after CRISPR/Cas9 or plasmid injection.

2.4 Cas9-Capped mRNA and gRNA Synthesis

Olive flounder(GeneBank Accession: NW_ 017861210.1) exon 1 and(GeneBank Accession No. KY703432) exon 1 were cloned using gene-specific pri- mers (YPMM-E1-F and YPMM-E1-R for; YPgsdf-E-F and YPgsdf-E-R for) and sequenced to confirm the gRNA target sequence. Detection primers were designed using the online tool (https://www.ncbi.nlm.nih. gov/tools/primer-blast/index.cgi?LINK_LOC=BlastHome). Cas9-capped mRNAs were synthesized using the ZCas9 plasmid (Liu., 2014) and MESSAGE mMACHINE T7 (Thermo, Foster City, CA, USA) and purified using MEGA clear (Thermo, USA). gRNA was predicted using online tools (http://zifit.partners.org/ZiFiT/CSquare9Nu- clease.aspxs), synthesized using the MAXIscript T7 Kit (Thermo, USA), and purified using the mirVana miRNA Isolation Kit (Thermo, USA).

Two gRNAs targeting distinct sites in exon 1 of the floun- der myomaker gene were synthesized (Fig.2). The two tar- get sites were named YPMMgRNA-1 (5’-GGCGATGAA CGCTCCCATGGTGG-3’), and YPMMgRNA-2 (5’-GGCGGCGACGCTGGCTGCAGGGAGG-3’). The first GG arerequired for correct T7 RNA polymerase transcription. ThePAM sequences (NGG) are underlined. For gRNA synthe- sis, a DNA template was generated by PCR on the pMD- 18T plasmid containing the gRNA scaffold (Chang., 2013), thatwas the template,-specific forward primers (YPMMgRNA1, YPMMgRNA2), and gRNA re- verse primer (Table 1).

Two gRNAs targeting distinct sites of the flounderwere on the reverse strand of exon 1, which were named YPgsdfgRNA-1 and YPgsdfgRNA-2 (Fig.2). The-spe-cific forward primers (YPgsdfgRNA1, YPgsdfgRNA2) were used for gRNA synthesis (Table 1).

2.5 DNA Isolation from Fish Embryos, Mutation Detection, and Genotyping

Larvae at the hatching and swimming stages were an- esthetized with 100mgL?1MS222 (Tricaine methanesul- fonate). Five larvae were collected per tube as a group. The larvae were treated as described by Wang. (2014). PCR was performed in a total reaction volume of 25μL with 2μL of template, using gene-specific primers (YPMM- E1-F and YPMM-E1-R for; YPgsdf-E-F and YPgsdf-E-R for; Table 1). PCR products were sub- jected to direct sequencing for mutation detection.

Fig.2 Design of the target site of gRNAs in myomaker and gsdf. A, genome structure of myomaker; B, genome structureof gsdf. The gRNAs were designed in exon 1. The gRNA-1 site was targeted to the start codon.The gRNA sequence is underlined. Red letters represent the NGG site. ATG in boldface represents the start codon.

Table 1 Primers for gRNA synthesis and PCR products analyses

For larval genotyping, genomic DNA was isolated fromrandomly selected groups, which contained five larvae. The PCR products were purified and ligated into the Peasy-T3 vector (Transgene, China). After the ligated products weretransformedinto.and identified with blue-white screening, single white colonies were selected, and their plasmid inserts were purified and sequenced. For each band obtained in the original PCR amplification, at least 10 se- quences that matched the target sequence were obtained. These sequences were compared with the wild-type se- quence using the online software ClustalW (https://embnet. vital-it.ch/software/ClustalW.html.).

2.6 Immunohistochemistry

The larvae injected with-GFP were fixed in 4% paraformaldehyde (PFA) and whole-mount immunohisto- chemistry was performed to confirm GFP expression us- ing GFP antibody according to the method of Tan and Du (2002) with some modifications. After treatment in cold acetone, embryos were washed in PBST (1× PBS with 0.1%Tween 20). Then, the larvae were incubated with 10% goat serum in PBST for 30min at room temperature. After wash-ed with PBST, the larvae were incubated with anti-GFP antibody (Molecular Probe, 1:1000) in PBST containing 1% goat serum at 4℃overnight. Then, the larvae were wash- ed with PBST, followed by incubation with HRP-labeled anti-rabbit secondary antibody (Molecular Probe, 1:1000 in PBST) for 1h. After washed with PBST, the larvae were incubated in DAB staining solution (Tiangen, China) at room temperature. The staining was monitored and stop- ped by washing with PBST. Embryos were photographed in 50% glycerol under a Leica DM LB2 microscope (Lei- ca, Germany).

3 Results

3.1 Test of the Microinjection Method with a Modified Needle

To test whether the injections had been successful in flounder eggs using our simply prepared needle (Fig.1), dextran rhodamine B and flounderpromoter-driven GFP (-GFP) plasmid (Zhang., 2006) were injected together into the eggs, and the injected eggs were cultured to the swimming stage. Among the hatched larvae, more than 80% (18/22) were normal and showed red fluorescence (Fig.3B), while about 40% (8/22) show GFP expression (Fig.3D). Immunohistochemistry analysis con- firmed the identity of the expressed protein as GFP (Fig.3F).

Fig.3 Rhodamine dye and GFP expression in hatched and free-swimming larvae after injection.A and C, Control; B, Rhodamine dye in injected larvae; D, GFP expression; E, Immunohistochemistry analysis of GFP expression in control larvae; F, Immunohistochemistry analysis of GFP expression in myod-GFP injected larvae. Scale bar: 100μm.

There was no obvious difference in the effect of the two different needles on the injected embryos. However, the needle without the modified brake was not easily blocked during microinjection, and usually one needle with a mo- dified brake could replace three needles without brakes in terms of the degree of clogging. Therefore, the use of a needle with a modified brake could increase the injection speed.

3.2 Transient Expression of GFP in Injected Olive Flounder Larvae

Usually microinjection is performed at the one-cell to two-cell stage. To determine the optimal stage for injection, the plasmid-GFP was injected at three stages,., protoplasmic streaming to one-cell (not included, stage I), one-cell to two-cell (visible cleavage furrow, cell has not divided, stage II), and two-cell to four-cell (visible cleavage furrow, cell has not divided, stage III) stages. The nor- mal, hatched with GFP expression larvae accounted for about 40% (44.4%, 41.2%, and 38.2%, respectively) of thetotal hatched and free-swimming larvae (Table 2) for all three stages. However, the percentages of normal hatched larvae for stage I and stage II were higher (57.1% and 56.8%) than that for stage III (42.0%). Abnormal, hatched, and liv- ing larvae accounted for approximately 11.1% (7/63), 9.8% (5/51), and 27.6% (21/76) (Table 2), and GFP expression was detected in 3.2%, 11.8%, and 25.0% (Table 2) of the to- tal hatched larvae, respectively.

3.3 Gene Modification in Olive Flounder Larvae

To determine whether CRISPR/Cas9 could be applied to olive flounder embryos, two different genes,and, were selected. Two gRNAs were designed (Table 2) to test site mutations and fragment deletions. One month after injection, the survival rate was approximately 2%,while it was about more than 80% for un-injected embryos.For each gene, the PCR products were first direct sequenced to identify mutations (Fig.4). Then, one tube containing five larvae was randomly selected for single-clone sequencing and genotype analyses for each gene. The single-clone mu- tation rate was approximately 45% (14/31 forand 5/11 for) (Table 3). There were 10 distinctgenotypes (Fig.5, Table 3) and fourgenotypes (Fig.6, Table 3). For, most mutations were frag-ment deletions from 30bp to 170bp, which was between the gRNA-1 and gRNA-2 sites. The other mutations were at the gRNA1 site, which also contained one insertion mu- tation. For, the mutations were found only at the site around gRNA-1. For themutations, 4 of 10 wereinducedmicrohomology-mediated joining (Ansai.,2014), whereas for themutations, this propor- tion was three of four (Fig.5 and Fig.6).

Table 2 Stage effect on hatch and GFP expression after microinjection

Note: The normal hatching rates in control (without injection) were >90%.

Fig.4 Direct sequencing chromatogram of PCR products.A, myomaker control embryos; B, Cas9 mRNA and myomaker gRNA injected embryos; C, gsdf control embryos; D, Cas9 mRNA, and gsdf gRNA injectedembryos. Red box indicated the gRNA site.

Table 3 Type and frequency of myomaker and gsdf mutations in larvae

Note: ?/+, deletion/insertion number.

Fig.5 Different flounder myomaker mutation types.A, Large fragment deletion (≥90bp) of myomaker; B, Fragment deletion (<90bp) of myomaker; C, Deletion at the first gRNA site of myomaker. Red letter, gRNA site; WT, wild-type sequence; M?/+XX. Deletion/insertion sequence length; lower-case letters, insertion; Underlined text indicates microhomology.

Fig.6 Different flounder gsdf mutation types.Red letter, gRNA site; WT, wild-type sequence; M?/+XX. Deletion/inser- tion sequence length; lower-case letters, insertion; Underlined text indicates microhomology.

4 Discussion

In this study, we successfully established a microinjec- tion and CRISPR/Cas9 genome editing method for olive flounder. Some noteworthy aspects of this process are dis- cussed below.

The needle was an important factor for successful micro- injection. Olive flounder embryos have osmotic pressure as well as hard and elastic chorion. To penetrate the cho- rion and block suck-back, a needle with a brake has been used in pelagic eggs, such as the eggs of barfin flounder (Goto., 2015, 2019) and olive flounder (Kim., 2019). However, the preparation of such a needle requires special equipment and is complex and time-consuming (Goto., 2019). It requires approximately five minutes to prepare a needle with a brake using special equipment for a skilled person. Also, the needle is easily blocked du- ring injection, as the brake region is too narrow (appro- ximately 3–5μm) (Goto., 2019). To resolve the issues of osmotic pressure and complex needle preparation, we designed two different methods. First, we prepared a needle with a modified brake through a puller. Two-step pul- ling was performed at different temperatures, first at a high temperature, then at a low temperature, which hasbeen described in detail inMaterial and Methods. The time needed to prepare such a needle was only one minute, and no special instrument was required. Our experiment show- ed that the needle with a modified brake was more effi- cient and less often blocked. In the second method, to ad- dress the high osmotic pressure, we used a needle without a brake and adjusted the in/out air balance. The balance in the injector was adjusted to a position that could induce slight leakage of the inner solution when the needle was in the water in the injection plate. There was no leakage when the needle was out. The balance pressure was usu- ally 1.5–2 psi, and the injection pressure was about 2–3psi.Using these two different strategies, the plasmid could be successfully injected into the flounder eggs with a high survival rate of the hatched larvae.

From egg to adult, all fishes of Pleuronectiformes had to overcome the important stage metamorphosis, a high-death- rate stage (Thorisson, 1994). Therefore, a larger number of hatched larvae would increase the quantity of adult fish. In tongue sole (Cui., 2017), the hatch rate was only 0.1%, while in barfin flounder (Goto., 2015), a high hatch rate was achieved using special needles (Goto., 2019). In our method, the hatch rate was approximately 50% for plasmid, which is higher than that achieved intongue sole and lower than that attained in barfin flounder. However, in our method, the needle preparation required only the incorporation of a puller, whereas the method used in barfin flounder required not only a puller but also other specialized instruments (Goto., 2019). Specialized ins- truments require special skills for their operation, which brings more difficulties to the application. Thus, our nee- dle preparation method can increase the efficiency and suc- cess of the experimnt.

The injection stage was also important for the efficien- cy of exogenous macromolecules. Most injections are conducted at the blastodisc to two-cell stage, such as injections in zebrafish, medaka, Nibe croaker (Yamamoto., 2011), red sea bream (Kishimoto., 2018), and olive flounder (Kim., 2019). In olive flounder, the in- jection efficiency was lower than that inzebrafish due to differences in chorion strength and cell penetration ability. In our experiments, GFP was expressed after injection at all three stages, with substantial differences in its expres- sion levels. The embryos injected at the earlier two stages showed low GFP expression levels with high percentages of normally developed larvae, whereas embryos of the later stage showed high GFP expression with a low percentage of normally developed larvae. It is possible that methyla- tion levels differed among these stages (Goll and Halpern, 2011). These results suggest that plasmid injections to test gene function should be performed at the two-cell to four- cell stage.

Our CRISPR/Cas9 experiments successfully induced mu- tations in two different genes in olive flounder. Four dif- ferent mutations were found in thegene, and 10 dif- ferent mutations were found in. Interestingly, these 10 mutations were detected in only five larvae in a single tube. In red sea bream and tongue sole, different mu- tations have also been detected within a single fish (Cui., 2017; Kishimoto., 2018; Kim., 2019). This may be attributed to the injection stage (two-cell) and va- riation in genome editing efficiency among cells. It is also possible that mutations occurred at the one-cell stage but underwent homologous repair in different cells during cell proliferation.

The mutation rates differed between the gRNA-1 and gRNA-2 sites. Two gRNAs targetingandwere designed to delete a fragment. For, the probabilities of being digested by Cas9 were equal for the DNA sequences at the two gRNA sites. Thus, most mutations were fragment deletions between the two gRNA sites. In addition to the fragment deletion, some nucleotides were also deleted at one gRNA site, mostly at the gRNA-1site for. It is possible that the mutation ratewas higher at the 5’ end of the gene than at the 3’ end. For, all of the mutations were near the gRNA-1 site. It is possible that the gRNA-2 ofhas no function at theearly developmental stages. Another possible reason is thatthe sequence was correctly repaired although it was digest- ed by Cas9. Hence, two gRNAs for each gene will increase the probability of obtaining a mutant.

5 Conclusions

In conclusion, a microinjection system with a simple needle preparation method was developed in this research. It has been successfully applied in olive flounder with a high survival rate. Short sequence mutations and fragment deletions in olive flounder can be obtained by employing the CRISPR/Cas9 methodology. This system can be used for other pelagic eggs. In the future, the function ofandin flounder can be analyzed using the knock-out fish developed in this study. The functions of other genes can be studied with the similar strategy.

Acknowledgements

This work was supported by the National Natural Sci- ence Foundation of China (Nos. 31672636, 31772834, and 31972774), the National Key R&D Program of China (Nos. 2018YFD0901202 and 2018YFD0900202), and the Key Research and Development Program of Shandong Pro- vince, China (No. 2019GHY1120070). The authors declare no financial or commercial conflict of interest.

Aluru, N., Karchner, S. I., Franks, D. G., Nacci, D., Champlin, D., and Hahn, M. E., 2015. Targeted mutagenesis of aryl hydro- carbon receptor 2a and 2b genes in Atlantic killifish ()., 158: 192-201.

Ansai, S., Inohaya, K., Yoshiura, Y., Schartl, M., Uemura, N., Ta-kahashi, R.,., 2014. Design, evaluation, and screening me- thods for efficient targeted mutagenesis with transcription ac- tivator-like effector nucleases in medaka., 56: 98-107.

Brem, G., Brenig, B., H?rstgen-Schwark, G., and Winnacker, E. L., 1988. Gene transfer in tilapia ()., 68: 209-219.

Chang, N., Sun, C., Gao, L., Zhu, D., Xu, X., Zhu, X.,., 2013. Genome editing with RNA-guided Cas9 nuclease in zebrafish embryos., 23 (4): 465-472.

Chourrout, D., Guyomard, R., and Houdebine, L. M., 1986. High efficiency gene transfer in rainbow trout () by microinjection into egg cytoplasm., 51: 143-150.

Cleveland, B. M., Yamaguchi, G., Radler, L. M., and Shimizu, M., 2018. Editing the duplicated insulin-like growth factor bind- ing protein-2b gene in rainbow trout ()., 8 (1): 16054.

Cui, Z., Liu, Y., Wang, W., Wang, Q., Zhang, N., Lin, F.,., 2017. Genome editing reveals dmrt1 as an essential male sex- determining gene in Chinese tongue sole ()., 7: 42213.

Davis, B. R., Brown, D. B., Prokopishyn, N. L., and Yannariel- lo-Brown, J., 2000. Micro-injection-mediated hematopoietic stemcell gene therapy., 2 (4): 412-419.

Doetschman, T., Gregg, R. G., Maeda, N., Hooper, M. L., Melton,D. W., Thompson, S.,., 1987. Targetted correction of a mu-tant HPRT gene in mouse embryonic stem cells., 330: 576-578.

Du, S. J., Gong, Z. Y., Fletcher, G. L., Shears, M. A., King, M. J., Idler, D. R.,., 1992. Growth enhancement in transgenic At- lantic salmon by the use of an ‘a(chǎn)ll fish’ chimeric growth hor- mone gene construct., 10 (2): 176-181.

Edvardsen, R. B., Leininger, S., Kleppe, L., Skaftnesmo, K. O., and Wargelius, A., 2014.Targeted mutagenesis in Atlantic sal- mon (L.) using the CRISPR/Cas9 system induces complete knockout individuals in the F0 generation., 9 (9): e108622.

Fletcher, G. L., Shears, M. A., King, M. J., Davies, P. L., and Hew, C. L., 1988. Evidence for antifreeze protein gene trans- fer in Atlantic salmon ()., 45 (2): 352-357.

Fuiman, L. A., 1997. What can flatfish ontogenies tell us about pelagic and benthic lifestyles?,37(3-4): 257-267.

Gaj, T., Gersbach, C. A., and Barbas III, C. F., 2103. ZFN, TALEN,and CRISPR/Cas-based methods for genome engineering., 31: 397-405.

Goll, M. G., and Halpern, M. E., 2011. DNA methylation in ze- brafish., 101: 193-218.

Goto, R., Saito, T., Kawakami, Y., Kitauchi, T., Takagi, M., Todo,T.,., 2015. Visualization of primordial germ cells in the fer- tilized pelagic eggs of the barfin flounder., 59 (10-12): 465-470.

Goto, R., Saito, T., Matsubara, T., and Yamaha, E., 2019. Micro- injection of marine fish eggs: Methods and protocols. In:. Liu, C. Y., ed., Huma- na Press, New York, 475-487.

Hamied, A., Alnedawy, Q., Correia, A., Hacker, C., Ramsdale, M., Hashimoto, H.,., 2020. Identification and characteriza- tion of highly fluorescent pigment cells in embryos of the Ara- bian killifish ()., 23 (11): 101674.

Hwang, W. Y., Fu, Y., Reyon, D., Maeder, M. L., Tsai, S, Q., Sander, J. D.,., 2013. Efficient genome editing in zebra- fish using a CRISPR-Cas system., 31 (3): 227-229.

Inoue, D., Stemmer, M., Thumberger, T., Ruppert, T., B?renz, F., Wittbrodt, J.,., 2017. Expression of the novel maternal centrosome assembly factor Wdr8 is required for vertebrate em-bryonic mitoses., 8: 14090.

Jao, L. E., Wente, S. R., and Chen, W., 2013. Efficient multiplex biallelic zebrafish genome editing using a CRISPR nuclease system., 110 (34): 13904-13909.

Jiang, D. N., Yang, H. H., Li, M. H., Shi, H. J., Zhang, X. B., and Wang, D. S., 2016.is a downstream gene ofthat functions in the male sex determination pathway of the Nile tilapia., 83 (6): 497-508.

Jiao, S., Tan, X., Sui, Y., and You, F., 2019. Muscle fibre type com- position in the lateral muscle of olive flounder., 121 (1): 1-6.

Kato-Unoki, Y., Takai, Y., Kinoshita, M., Mochizuki, T., Tatsu- no, R., Shimasaki, Y.,., 2018. Genome editing of puffer- fish saxitoxin- and tetrodotoxin-binding protein type 2 in., 153: 58-61.

Kim, J., Cho, J. Y., Kim, J. W., Kim, H. C., Noh, J. K., Kim, Y. O.,., 2019. CRISPR/Cas9-mediateddisruption en- hances muscle mass in the olive flounder., 512: 734336.

Kishimoto, K., Washio, Y., Yoshiura, Y., Toyoda, A., Ueno, T., Fukuyama, H.,., 2018. Production of a breed of red sea breamwith an increase of skeletal muscle mass and reduced body length by genome editing with CRISPR/ Cas9., 495: 415-427.

Kobayashi, K., and Satou, Y., 2018. Microinjection of exogenous nucleic acids into eggs,species., 1029: 5-13.

Landemaine, A., Rescan, P. Y., and Gabillard, J. C., 2014. Myo- maker mediates fusion of fast myocytes in zebrafish embryos., 451 (4): 480-484.

Liu, D., Wang, Z., Xiao, A., Zhang, Y., Li, W., Zu, Y.,., 2014. Efficient gene targeting in zebrafish mediated by a zebrafish- codon-optimized cas9 and evaluation of off-targeting effect., 41 (1): 43-46.

Maki, J. A., Cavallin, J. E., Lott, K. G., Saari, T. W., Ankley, G. T., and Villeneuve, D. L., 2020. A method for CRISPR/Cas9 mu- tation of genes in fathead minnow ()., 222: 105464.

McEvoy, T., Stack, M., Keane, B., Barry, T., Sreenan, J., and Gan- non, F., 1988. The expression of a foreign gene in salmon em- bryos., 68: 27-37.

Moody, S. A., 2018. Microinjection of mRNAs and oligonucleo- tides., 2018 (12): pdb.prot097261.

Ozato, K., Kondoh, H., Inohara, H., Iwamatsu, T., Wakamatsu, Y.,and Okada, T. S., 1986. Production of transgenic fish: Introduc-tion and expression of chicken d-crystallin gene in medaka em- bryos., 19 (4): 237-244.

Palmiter, R. D., and Brinster, R. L., 1986. Germ-line transforma- tion of mice., 20: 465-499.

Shao, C., Bao, B., Xie, Z., Chen, X., Li, B., Jia, X.,., 2017. The genome and transcriptome of Japanese flounder provide insights into flatfish asymmetry., 49 (1): 119- 124.

Stuart, G. W., McMurray, J. V., and Westerfield, M., 1988. Repli- cation, integration and stable transmission of foreign sequences injected into early zebrafish embryos., 103 (2): 403-412.

Tan, X., and Du, S. J., 2002. Differential expression of twogenes in fast and slow muscles of gilthead seabream ()., 212 (5): 207-217.

Thomas, K. R., and Capecchi, M. R., 1987. Site-directed muta- genesis by gene targeting in mouse embryo-derived stem cells., 51 (3):503-512.

Thompson, S., Clarke, A. R., Pow, A. M., Hooper, M. L., and Melton, D. W., 1989. Germ line transmission and expression of a corrected HPRT gene produced by gene targeting in em- bryonic stem cells., 56 (2): 313-321.

Thorisson, K., 1994. Is metamorphosis a critical interval in the early life of marine fishes?, 40: 23-36.

Wang, Q., Tan, X., Jiao, S., You, F., and Zhang, P. J., 2014. Ana- lyzing cold tolerance mechanism in transgenic zebrafish ()., 9 (7): e102492.

Yamamoto, Y., Kabeya, N., Takeuchi, Y., Higuchi, K., Yatabe, T., Tsunemoto, K.,., 2011. Establishment of a stable trans- genic strain in a pelagic egg spawning marine teleost, Nibe croaker., 313: 42-49.

Zhang, X., Wang, H., Li, M., Cheng, Y., Jiang, D., Sun, L.,., 2014. Isolation of doublesex- and mab-3-related transcription factor 6 and its involvement in spermatogenesis in tilapia., 91 (6): 136.

Zhang, Y., Tan, X., Zhang, P. J., and Xu, Y., 2006. Characteriza- tion of muscle-regulatory gene,, from flounder () and analysis of its expression patterns du- ring embryogenesis., 8 (2): 139- 148.

Zhu, Z., Li, G., He, L., and Chen, S., 1985. Novel gene transfer into the fertilized eggs of goldfish (L1758)., 1 (1): 31-34.

Zhu, Z., Xu, K., Li, G., Xie, Y., and He, L., 1986. Biological effects of human growth hormone gene microinjected into the fertilized eggs of loach(Cantor)., 31 (14): 988-990.

September 19, 2020;

November 10, 2020;

March 16, 2021

? Ocean University of China, Science Press and Springer-Verlag GmbH Germany 2021

#The two authors contributed equally to this paper.

E-mail: tanx@qdio.ac.cn

E-mail: youfeng@qdio.ac.cn

(Edited by Qiu Yantao)