A novel technique based on in vitro oocyte injection to improve CRISPR/Cas9 gene editing in zebrafish

Contemporary improvements in the type II clustered regularly interspaced short palindromic repeats/CRISPR-associated protein 9 (CRISPR/Cas9) system offer a convenient way for genome editing in zebrafish. However, the low efficiencies of genome editing and germline transmission require a time-intensive and laborious screening work. Here, we reported a method based on in vitro oocyte storage by injecting oocytes in advance and incubating them in oocyte storage medium to significantly improve the efficiencies of genome editing and germline transmission by in vitro fertilization (IVF) in zebrafish. Compared to conventional methods, the prior micro-injection of zebrafish oocytes improved the efficiency of genome editing, especially for the sgRNAs with low targeting efficiency. Due to high throughputs, simplicity and flexible design, this novel strategy will provide an efficient alternative to increase the speed of generating heritable mutants in zebrafish by using CRISPR/Cas9 system.

Scientific RepoRts | 6:34555 | DOI: 10.1038/srep34555 rapidly than Cas9 capped RNA 23,24 and codon-optimized Cas9 protein improved higher knockout efficiency than original one 25,26 . Nevertheless, the subsequent screening for germline transmission of induced mutations could be time-consuming and labor intensive for the sgRNAs with low targeting efficiencies 15,27 . In 2014, Dong et al. reduced the screening time for mutants by labelling the zebrafish primordial germ cells 28 . However, the improvement for the specific targeting efficiency remained below expectations.
A valuable technique of ovum preservation has been successfully applied in mammals [29][30][31] , including humans [32][33][34] . Recently, the preservation of fish oocytes in vitro was also reported 35,36 . Goetz and his colleagues used modified cortland medium to store trout eggs for at least 2 days 37 . Ubilla et al. reported that in vitro storage of rainbow trout eggs in their coelomic fluid for 96 h caused a decline in the quality of eggs and reduced their capacity to develop into larvae 38 . The preferable preservation technology of zebrafish oocytes was established in 2008 by Seki et al., which contained 90% Leibovitz L-15 medium at pH 9.0, containing 0.5 mg mL −1 BSA and 1 ug mL −1 17α -20β -Dihydroxy-4 Pregnen-3-one (DHP). In their findings, they kept the zebrafish oocytes alive from stage III to mature in vitro, nevertheless the hatching rate remained low 39,40 .
Here, we reported a modified in vitro oocyte storage strategy to increase the efficiency of genome editing by CRISPR/Cas9 system in zebrafish. This novel technique with in vitro oocyte injection prior to their storage improved the efficiencies of gene targeting and germline transmission. Contrary to the conventional genome editing approaches, this method utilized the inherent translation system to impel the oocytes to produce Cas9 proteins in advance, which greatly increased the efficiency of genome editing. Undoubtedly, it also improved the efficiency of germline transmission. Therefore, our novel technique will increase the speed of generating heritable mutants of zebrafish using CRISPR/Cas9 system.

Materials and Methods
Ethics Statement. All experiments in this study were in accordance with the "Guide for the Care and Use of Zebrafish husbandry. Adult fish and embryos of AB strains were raised and maintained as described by Westerfield et al. 41,42 . Fish were kept in the automatic water cycle system on a 14 h light: 10 h dark period. The adult female and male zebrafish were maintained at 28 °C in separate tanks and mated once a week. CRISPR guide RNAs design and Cas9 capped RNA synthesis. Cas9 target sites were selected in exons across the genome and designed with an online tool, ZIFIT Targeter (http://zifit.partners.org/ZiFiT/ ChoiceMenu.aspx). The plasmid pXT7-Cas9 was linearized by XbaI 6 . Cas9 capped RNA was synthesized using Ambion mMESSAGE mMACHINE T7 Transcription Kit (Ambion, USA). The sgRNAs transcription templates were prepared by PCR with T7-targetsite-F primers and a universal reverse primer gRNA-R (Table 1) 43 . The sgR-NAs PCR products were purified by TIANquick Midi Purification Kit (TIANGEN, China) and then transcribed with MAXIscript T7 Kit (Ambion, USA). RNAs were purified by MicroElute RNA Clean-Up Kit (OMEGA, USA). The RNA quality and concentration were analyzed by electrophoresis and the NanoDrop2000 spectrophotometer (Thermo Fisher Scientific, France). The efficiencies of sgRNAs were examined by the sgRNAs activity detection kit in vitro (Viewsolid Biotech, China).
In vitro oocyte storage. The general method of in vitro oocyte storage is adapted from Seki et al. 39 and Nair et al. 40 . The oocyte storage medium was 90% Leibovitz's L-15 medium with L-glutamine (Gibco) and 0.5 mg mL −1 bovine serum albumin (BSA, Sigma-Aldrich), and then pH was adjusted to 9.0 with 10N NaOH. Before oocyte collection, adult zebrafish females and males were maintained in separate spawn tanks overnight. Adult zebrafish females with swollen and elastic abdomen were chosen. Before squeezing the oocytes, zebrafish females were treated with 0.2 mg mL −1 ethyl 3-aminobenzoate methanesulfonate salt, then gently pressed the abdomen under the microscope. Oocytes were harvested from 1-2 females and scattered in the oocyte storage medium.
Oocyte micro-injection. The mature oocytes were collected for micro-injection by using a micro-injector (Eppendorf, Germany). The injected reagents were diluted in 0.2 M KCl. The final concentrations of Cas9 RNAs and sgRNAs were 300 ng μ L −1 and 30~50 ng μ L −1 , respectively. One nL solution containing 300 pg Cas9 capped RNA and 30~50 pg sgRNAs was micro-injected into 1-cell embryo. The oocytes were arranged along the edge of a glass slide in oocyte storage medium before micro-injection (Fig. 1). For gene knock-in experiment, Cas9 capped RNAs, mc4r sgRNAs and the single-stranded mloxP donor DNAs (Table 1) were micro-injected into zebrafish oocytes. After completion of micro-injection, the oocytes were cultured in the storage medium to avoid light for 30 min.
In vitro fertilization. Injected oocytes were fertilized after 30 min. Sperm solution was produced by shearing the testes from two AB male fish and grinding in 200~300 μ L cortland solution (NaCl 7.25 g L −1 ; KCl 0.38 g L −1 ; CaCl 2 0.162 g L −1 ; NaHCO 3 1.0 g L −1 ; NaH 2 PO 4 0.41 g L −1 ; MgSO 4 0.23 g L −1 ; glucose 1.0 g L −1 ). The quality of sperm was judged by the color of the sperm solution. The color of mature sperm was milky white. The cortland solution was kept on ice for at least 1 h before use. During this experiment, the sperm solution was kept on ice all the time. The storage medium was removed from the culture dish, and 20~30 μ L sperm solution was added directly on the top of the oocytes, followed by a few drops of fresh water with a plastic dropper. After 1~2 min, the culture dish was washed gently with fresh water.
Scientific RepoRts | 6:34555 | DOI: 10.1038/srep34555 Genomic DNAs isolation. Genomic DNAs were extracted by using the CTAB extraction method 44 . For the DNAs extraction of embryos and oocytes, one embryo and five oocytes were incubated in 40 μ L and 60 μ L lysis solution (100 mM Tris-Hcl pH 8.5, 0.5 M EDTA, 10% SDS, 5 M NaCl, 20 mg mL −1 Proteinase K) at 55 °C for 2 h. For adult zebrafish DNAs extraction, partial tails were lysed in 100 μ L lysis solution, and incubated at 55 °C for 3 h. Then, DNAs were purified by chloroform, isopropanol and 75% ethanol. Finally, 20~30 μ L double distilled water was added into each sample.

Primer Description
Sequence (5′ to 3′)  Gene editing assay and sequencing analysis. The 16 °C ramping at max rate and holding at 16 °C for 2 min) for heteroduplex formation. After annealing, products were treated with 0.5 μ L T7 endonuclease I for 15 min at 37 °C, and analyzed on 1.5% agarose gels. Gels were observed and analyzed by using a gel-imaging system (JUNYI, Beijing, China). To detect mloxP knock-in, larvae DNAs at 72 hpf were isolated and PCR was performed with mc4r T7E R matching the mc4r sequence and mloxP R matching the sequence of the mloxP donor sequence (Table 1). To further confirm the success of gene editing, about 2 μ L above all PCR products were cloned into the pMD-19T vector (Takara, Japan) for sequencing analysis.
Statistical analysis. For statistical analysis, two-tailed unpaired Student's t-test was performed by using SPSS software 45 . P values < 0.05 were considered statistically significant. Each experiment was performed independently for three times.

Results
An effective storage medium for oocyte preservation. The oocyte preservation and in vitro maturation have been investigated for many years. On the basis of previous findings, we compared the effects of three oocytes storage media: hepes-cortland medium, 90% L-15 medium at pH 9.0 with BSA, and without BSA. Due to the difficulty in obtaining the coelomic fluid of zebrafish, oocyte preservation in coelomic fluid was not suitable here. Thus, we collected oocytes and stored them in three different preservation media that are mentioned above. The results showed that the mature oocytes stored in 90% L-15 medium with BSA at pH 9.0 remained stable after 2 h. The hepes-cortland medium showed almost the same effect as the 90% L-15 medium without BSA at pH 9.0 and induced the formation of a cavity between the oocyte shell and vitelline membrane (Fig. 2).

The hatching rates in vitro.
To find a suitable storage time, the effects of storage time on hatching rates were investigated. The results showed that there was a significantly decreasing trend for oocyte hatching rates with increase in oocytes storage time in vitro (Fig. 3). The hatching rates dramatically declined, when the storage time was longer than 60 min (Fig. 3A). However, the egg hatching rates were more than 80%, if the storage time was less than 30 min (Fig. 3B). As expected, the oocytes stored in hepes-cortland medium and 90% L-15 medium without BSA failed to fertilize in vitro. In addition, a longer storage time resulted in a higher deformity rate in the 90% L-15 medium with BSA at pH 9.0 (data not shown). Our results indicated that Seki's medium was conducive to oocytes storage, but DHP was unnecessary. However, BSA was proved as a critical component for the oocytes storage medium.
Translational capacity of mature oocytes. To verify the translational capacity of mature oocytes, mCherry capped RNAs were micro-injected into mature oocytes and the oocytes were cultured in the storage medium. The red fluorescence was observed after 2 h incubation (Fig. 4). To elucidate whether the Cas9-sgRNAs complexes efficiently worked in zebrafish oocytes or not, the Cas9 capped RNAs and mc4r or mpv17 sgRNAs were micro-injected into mature oocytes and incubated in the storage medium for 30 min or more, then the PCR-based T7E1 assay was performed. However, there was no mutation detected in the oocytes genome (data not shown). Thus, we speculated that the genome DNAs of zebrafish oocytes had a high affinity between histones and DNA backbone, which prevented Cas9 protein and target sgRNAs from binding with genomic DNAs to initiate the cleavage activity of CRISPR-Cas9 complex.

Toxicity assay of Cas9 capped RNAs and sgRNAs.
To test the toxicity of Cas9 capped RNAs and sgR-NAs in the oocytes, the fertilization rate and hatching rate were statistically analyzed after injecting Cas9 capped RNAs/sgRNAs or phenol red. The results revealed that the toxicity of Cas9 capped RNAs/sgRNAs was very low, compared to phenol red injected groups. However, the fertilization and hatching rates were significantly lower than the control group ( Fig. 5 and Table S1), indicating that the injection caused mechanical damage to the oocytes. In addition, there was no significant difference observed in fertilization rates between Cas9 capped RNAs/sgRNAs groups and phenol red groups, albeit their hatching rates were significantly different ( Fig. 5 and Table S1). About eight percent of embryos were observed with developmental deformity and were unable to hatch properly after injecting Cas9 capped RNAs/sgRNAs.

Improving the efficiency of gene knock-in in zebrafish founder. Although gene knock-in was
successfully performed by ZFN, TALEN and Cas9 in zebrafish embryos 4,46,47 , the efficiency of gene knock-in was unsatisfactory. To test whether the current method of oocytes storage could enhance the efficiency of gene knock-in or not, a mloxP knock-in experiment was performed. The results were quite promising and showed that a single-stranded oligo DNA as a donor containing a mloxP site 6 and 27 nt homology arms on both ends was effectively knocked into mc4r locus (Fig. 6A and Fig. S1B). The size of PCR products was 352 bp as expected   6A). To further confirm the insertion of mloxP into mc4r locus, the PCR products were amplified from the tested embryo and sequenced. As shown in Fig. S1, mloxP were successfully knocked into mc4r locus both by the normal and oocytes storage injection. Moreover, the efficiency of gene knock-in was successfully improved by 49.6% in the oocytes storage injection groups, compared to 26% in the normal injection groups (Fig. 6B and Table S2).
Enhancing the efficiency of gene knock-out in zebrafish founder. Zebrafish mc4r, mpv17, mstna, mc3r and mrap2b founders were selected to evaluate the efficiency of genome editing by using this novel technique based on in vitro oocyte injection and storage. Five sgRNAs were designed, which targeted mc4r, mpv17, mstna, mc3r and mrap2b, respectively. The results showed that 5 target genes were efficiently knocked out by using this technique (Figs S2 and S3). The mutation rates of those 5 genes (mc4r, mpv17, mstna, mc3r and mrap2b) 6C and Table S3).

Improving the efficiencies of germline transmission in the offspring.
To test the efficiencies of germline transmission, mc4r and mpv17 genome DNAs were extracted from F 1 generation and T7E1 assay and sequencing analysis were performed. Results showed that the efficiencies of germline transmission for mc4r and mpv17 mutations were 96.7% and 91% ( Fig. 6D and Table S4), which were significantly higher than the common CRISPR/Cas9 system (70% and 35.2%).  Table S4. (C) The efficiencies of genome editing of mc4r, mpv17, mstna, mc3r and mrap2b were measured by T7E1 assay and sequencing analysis. (D) The efficiencies of germline transmission of mc4r and mpv17 were also measured by T7E1 assay and sequencing analysis in zebrafish F 1 generation. Three batches of F 1 embryos with different numbers (17~28) were performed both in the oocytes storage injection groups and in the normal injection groups. Each experiment was repeated for three times independently.

Discussion
Here, a novel technique based on in vitro oocyte injection was developed and optimized to improve the efficiencies of genome editing and germline transmission in zebrafish by using a CRISPR/Cas9 system. Previously, Nakajima and Yaoita reported that the gene editing activity of TALEN was improved by injecting TALEN capped RNAs into the oocytes of Xenopus laevis 48 . However, little is known about translational capacity of mature oocytes in zebrafish by using a CRISPR/Cas9 system. Chang et al. reported that the CRISPR/Cas9 nuclease generated site-specific cleavage (~35%) by editing three genes in zebrafish 6 . Hwang et al. reported that the mutation frequencies for ten genes induced by CRISPR/Cas9 nuclease were low in zebrafish, and the highest mutation was only 59.4% 49 . The low efficiencies of genome editing and germline transmission are the common issues confronted with many researchers. To solve these problems, researchers developed and evaluated different approaches. In this study, the rates of genome editing and germline transmission for all the replicates were more than 90%, which indicated that the micro-injection of Cas9 capped RNAs into oocytes could reduce the time for Cas9 translation, and subsequently improve the probability of mutations transferred to the next generation.
Our findings are consistent with the results reported by Fujii et al. 23 . Gagnon et al. revealed that the direct injection of Cas9 protein and sgRNAs complex into zebrafish eggs strikingly increased the mutagenic activity, but the efficiency of germline transmission was not discussed 50 . Our current findings are entirely different from the results using the direct injection of Cas9 protein/sgRNAs complex. Since the commercial Cas9 protein purified from bacteria lacked post-translational modifications, Cas9 proteins antecedently expressed in oocytes exhibited higher enzyme activity than the Cas9 protein derived from bacteria. Thus, this novel technique based on in vitro oocyte injection can improve the efficiencies of knock-out, knock-in and germline transmission by using a CRISPR/Cas9 system. It's worth mentioning that an increasing efficiency of knocking mloxP into mc4r locus was achieved in this study, but whether the knocking-in efficiency was correlated with that of germline transmission deserves further studies.
The choices of effective storage media and appropriate storage time were the critical factors to affect the probability of the successful oocytes storage in vitro. The oocytes of freshwater fish usually develop the cavity between the shell and vitelline membrane due to osmotic phenomenon. In this study, the cavity forming effects of three oocytes storage media were compared. However, 90% L-15 medium at pH 9.0 with BSA was the only medium for inhibiting the formation of cavity. The cavity between the oocyte shell and vitelline membrane can hinder the oocytes' fertilization, which was confirmed by hepes-cortland medium and 90% L-15 medium without BSA. Generally, the hatching rates of oocytes are closely associated with the storage time in vitro. Seki et al. obtained 12% hatching rates by making the zebrafish oocytes matured in vitro from stage III 35 . In this study, relatively longer storage time was resulted in higher oocytes deformity rates. Therefore, 30 min storage time was appropriate to improve the hatching rate and the translation of Cas9 protein in zebrafish oocytes.
In conclusion, a novel technique based on in vitro oocyte injection and storage was developed and evaluated in this study. The combination of these two techniques was confirmed to improve the efficiencies of genome editing and germline transmission in zebrafish by using a CRISPR/Cas9 system. This novel strategy will reduce the unnecessary labor intensive screening and will significantly increase the speed of generating heritable mutants in zebrafish.