Correction of a CFTR/G542X Mutation Using CRISPR/Cas9 Editing in Ovine-bovine Interspecies Embryos

Cystic brosis (CF) is a human genetic disease caused by mutations in the CF transmembrane conductance regulator (CFTR) gene. Correction of CFTR mutations at embryo stage could be a permanent solution to cure this disease. To assess the eciency of CFTR/G542X mutation correction in vitro by CRISPR/Cas9, we utilized embryos generated by ovine-bovine interspecies somatic cell nuclear transfer (iSCNT) due to a limited access to sheep oocytes. First, we evaluated the developmental capacity of reconstructed iSCNT embryos. These embryos were able to develop to 16-cell stage, allowing for individual embryo genotyping. Then, we optimized the concentrations of Cas9:gRNA ribonucleoprotein (RNP) for 1-cell stage embryo injection. Genotyping results showed that we achieved high eciencies (88.9–100%) of indel mutations at the target locus after injection of different concentrations of RNP. When an RNP (0.09 µg/µl:2.3 µM) was co-injected with a ssODN (18 µM), the G542X mutation was corrected via the homology-directed repair in 11.1% (1/9) of iSCNT embryos. Taken together, we developed an effective strategy to correct the CFTR/G542X mutation in ovine-bovine iSCNT embryos by CRISRP/Cas9. Our strategy overcomes the limitation of oocyte source and provides the opportunity of mimicking the editing of any other genes in one-cell embryos of different species. First, we investigated the developmental potential of ovine-bovine iSCNT embryos. Based on the developmental results of reconstructed embryos, we optimized the concentrations of Cas9/gRNA for cytoplasmic injection and developed a strategy that allowed the correction of the CFTR/G542X mutation by CRISPR/Cas9 editing at embryonic stage. G542 locus of wild-type sheep CFTR A HpyAV enzyme identied in gRNA_1, which indel mutation detection. Three days after injection of different concentrations of Cas9:gRNA_1 RNP, mutations in target loci were analyzed in each of iSCNT embryos using PCR/RFLP assays and Sanger sequencing (Fig. and 2c). The detection results showed that different mutations, including nucleotide deletion, insertion, and replacement, were introduced at target loci of the injected iSCNT embryos. We achieved high biallelic mutation eciencies ranging from 88.9-100% after injection of Cas9:gRNA_1 RNP at concentrations of 0.02 µg/µl:0.57 µM to 1.4 µg/µl:36 µM The mutation eciency reduced to 54.5% when the concentration decreased to 5.3 ng/µl:0.14 µM and no biallelic mutations were found using 1.3 ng/µl:36 nM of Cas9:gRNA_1 RNP for injection. Mosaicism in iSCNT embryos was detected when lower concentrations of RNP (64- to 1024-fold dilution) were used (Supplementary Fig. 1). The cleavage rates of injected iSCNT embryos varied from 77.8 to 100% and no signicant difference (P > 0.05) was observed among different concentration groups. ovine-bovine interspecies by generated ovine-bovine cloned embryos using eGFP ovine broblasts as nuclear donors, which allowed us to accurately differentiate cloned embryos from those parthenogenetically activated and evaluate the development capacity of reconstructed embryos. The ovine-bovine embryos successfully developed to 16-cell stage but arrested prior to the formation of morulae or blastocysts. Our results are consistent with those published by Lagutina et al., who reported the development of iSCNT embryos to 12- to 16-cell stage 22 . Hua et al. observed 24.6% of blastocyst development with 117 ± 13 cells per blastocyst from such iSCNT embryos 23 . It has been reported in many studies that the bovine oocytes as recipient cytoplasm support the development at least to 10- to 16-cell of iSCNT embryos produced using nuclear donor cells from other species, including buffalo, sheep, mouse, cat, rabbit, horse 22 . The morula or blastocyst development of iSCNT embryos has been observed even when nuclear donor cells have a distant taxonomical relationship with bovine recipient oocytes, as in the case of interclass chicken-cow, interorder pig-, dog-, monkey- and human-cow 21 . The gaur-bovine iSCNT embryos were able to implant and develop to term 24 . Though term development of iSCNT embryos only limited in several closely related species, short-term in vitro developmental capacity of these embryos can an ecient window stage and enough genome materials for gene and mutation analysis. compound has been shown to be effective for gene editing of embryos in large animals 9,25 . Assessing mutation eciency of Cas9/gRNA compound in embryos prior to embryo transfer is critical because it is a time-saving and cost-effective strategy to generate CRISPR-mutated large animals. Our results showed that the bases encoding G542 amino acid of sheep CFTR gene in genome of ovine-bovine iSCNT embryos can be disrupted with a high eciency of 88.9–100% through injection of Cas9/gRNA RNP at different concentrations. Typically, higher RNP concentrations is more ecient, while also lead to increased embryonic toxicity 26,27 . However, we observed no difference in embryo cleavage rates when different concentrations of RNP were used, thus, a wide concentration range is available to test and optimize reagents for ecient gene editing in embryos. their when using 20 of their correction eciency in ovine-bovine embryos when Cas9/gRNA was co-injected with a 101nt donor ssODN. Using similar strategy and size of ssODN, Miao et al. conducted the correction of a missense mutation, Fgfr3/G374R and achieved 22.5% correction in newborn mice 33 . In our study, we rst utilized CRISPR editing to correct a CFTR mutation in ovine-bovine embryos. Our results support the hypothesis that precise genome editing in embryos by CRISPR could be a true cure for certain genetic diseases. The high degree of sequence similarity between farm animals and humans should allow for accurate modeling of single nucleotide correction in human embryos compared to other trials in mouse embryos or in transfected cell lines. In addition, it is noticeable that we did not detect mutagenesis induced by off-targeting in these mutated and corrected embryos, increasing the odds to utilize this strategy for clinical applications. cytoplasmic injection of Cas9:gRNA RNP results in highly ecient mutation induction in genome of ovine-bovine iSCNT embryos. The CFTR/G542X mutation can be corrected via co-injection of CRISPR/Cas9 RNP and donor ssODN compound into iSCNT embryos, which could be a potential strategy to correct genetic defect in genome at embryo stage. The use of iSCNT embryos for gene editing overcomes the limitation of oocyte source and provides the opportunity of mimicking the editing of any other genes in embryos of different species. Despite the limited developmental competency, the ovine-bovine iSCNT embryos were able to progress to the 16-cell stage that is sucient for gene-editing optimization and assessment of its eciency. Specic were to the CFTR and used to amplify and parts of introns anking it. Based on the sequencing results, we designed gRNAs by searching for the (N) 20 NGG motifs using the Benchling software (https://benchling.com/academic). gRNA_1 was chosen to induce DSB at G542 locus in exon 12 of the CFTR gene for mutation eciency detection. A restriction enzyme site was identied in gRNA_1 to facilitate indel mutation eciency detection using PCR/Restriction Fragment Length Polymorphism (RFLP) assays. gRNA_2 and ssODN were designed to correct the G542X mutation of the CFTR in iSCNT embryos, which were generated by using G542X sheep fetal broblasts as nuclear donors. The size of ssODN was 101nt and the mutated nucleotide, ‘G’, located in the middle of it was anked with 50 nucleotides at each side. The gRNAs were synthesized by Synthego with chemical modications (2’-O-methyl 3’-phosphorothioate modications in the rst and last 3 nucleotides). The TrueCut™ Cas9 protein V2 was purchased from Thermo Fisher Scientic. The ssODN was synthesized as the Ultramer → DNA oligonucleotides with chemical modications by Integrated DNA Technologies. and 100 U/ml penicillin/streptomycin (15140-122, Gibco). The bovine and ovine broblasts expressing the enhanced green uorescent protein (eGFP) were constructed and cryopreserved in our lab. The CFTR/G542X sheep fetal broblasts were isolated from a cloned fetus collected on day 45 of gestation and cryopreserved in our lab. The broblasts at passage 2 were used for second-generation cloning. Both bovine and ovine broblasts were grown to 80% − 90% conuence and used as nuclear donor cells for SCNT after 24 hours of serum starvation (0.5% FBS in DMEM). The cloned embryos were cultured in caprine SOF medium for 7 days 34 . The in vitro development of cloned embryos was observed under a stereo microscope and eGFP expression photographed by a uorescent microscope (Observer Z1, Zeiss).


Introduction
Cystic brosis (CF) is an autosomal recessive genetic disease caused by mutations in the CF transmembrane conductance regulator (CFTR) gene. Mutations in this gene result in either a lack of CFTR protein or a defective CFTR protein that cannot perform its key function in the cell. CFTR protein is an anion-selective channel that allows primarily chloride ion but also bicarbonate movement across epithelial cell membranes. Aberrant CFTR thus results in dysregulation of epithelial uid transport in the lung, pancreas, and other organs 1 . Common pathologies of CF include recurrent lung infection and in ammation, pancreatic insu ciency, which if untreated gives rise to malnutrition, intestinal obstruction, and reduced growth 2 . There are approximately 70,000 individuals living with CF worldwide, with a frequency of 1 in 2,000-3,000 among Caucasian newborns.
To date, over 2000 CFTR mutations have been identi ed and categorized into 6 types based on their functional effects including protein translation, cellular processing, and channel gating 3 . G542X is a primary mutation in the nonsense mutation category, the second most common mutation type, accounting for 4.6% of CF patients. Nonsense mutations generate a premature stop codon, leading to early termination of CFTR translation, mRNA decay through the surveillance pathway, and ultimately, lack of CFTR function 4 . Some drugs, such as gentamicin (G418) and ataluren (PTC124) has been developed as potential treatments for CF patients with nonsense mutations, which can promote ribosomal readthrough of premature termination codons. However, the use of G418 was limited due to its severe renal toxicity and ototoxicity 5 and ataluren showed no statistically signi cant difference compared to placebo in phase 3 clinical trials 6,7 . In addition, it has been proved that inhibiting the nonsense-mediated mRNA decay pathway with antisense oligonucleotides can upregulate target RNA and functional protein expression in cells homozygous for the CFTR/W1282X mutation, which is the second most common nonsense mutation that causes CF 8 . However, no further clinical trials using this strategy have been reported. Therefore, development of a safe and effective treatment for CF with nonsense mutations remains elusive.
Genome editing technologies, especially those based on CRISRP/Cas9, have been successfully applied in genome manipulation in different animals 9,10 . Recent successes in precise genome editing trials in early-stage or cloned human embryos have suggested a potentially true cure for genetic diseases 11,12 . However, genome editing of human embryos causes huge concerns because of ethical issues and technical uncertainties regarding the e ciency, mosaicism, off-target effect, and unintended homologous recombination. Genome editing by CRISPR/Cas9 generates double-strand breaks (DSBs) that evoke the error-prone non-homologous end joining (NHEJ) DNA repair pathway, which might cause off-target mutagenesis 13 . During the S to G2 phases of cell circle, DSBs can also activate an alternative DNA repair pathway, the homology-directed repair (HDR), that repairs the DSB sites using a template from the homologous chromosome or an introduced exogenous DNA molecule, leading to the introduction of desired mutations 14 .
Presently, CRISPR/Cas9 is predominantly employed to generate gene knockouts through NHEJ repair pathway 15 . Because HDR e ciency is relatively low, applications of HDR strategy for gene therapy, especially in human germline, have been limited 16 . To address the concerns raised by germline editing, especially in humans, extensive experiments should be conducted, thus, increasing the demand for oocytes and embryos, of which the source is rare, and the use strictly regulated in some species.
Interspecies somatic cell nuclear transfer (iSCNT) is a cloning technique that utilizes recipient oocytes and donor cells derived from two different species. The iSCNT experiments primarily focused on applications of preservation/rescue of endangered species and establishment of embryonic stem cell lines for regenerative medicine research. Reprogramming of human somatic cells by iSCNT has been conducted in many labs without generating the ethical issues surrounding the use of human oocytes [17][18][19] . It has been proved that iSCNT embryos between two species that are closely related can develop to term and lead to healthy cloned animals 20 . The in vitro development of morulae or blastocysts has been achieved when nuclear donor cells and recipient oocytes had a very distant taxonomical relation, as in the case of interfamily bovine -pig, interorder cat and pandarabbit, human -rabbit, -cattle, -sheep, and -goat, though remains ineffective and results are not always reproducible 21 . In the present study, we hypothesized that 1) iSCNT could be an alternative approach to produce embryos for testing gene-editing at embryonic stage in certain species, where oocytes are rare or di cult to obtain, such as sheep in this study, and 2) CRISPR/Cas9-mediated HDR could be used for the correction of the G542X mutation in CF. First, we investigated the developmental potential of ovine-bovine iSCNT embryos. Based on the developmental results of reconstructed embryos, we optimized the concentrations of Cas9/gRNA for cytoplasmic injection and developed a strategy that allowed the correction of the CFTR/G542X mutation by CRISPR/Cas9 editing at embryonic stage.

Results
In vitro development of cloned ovine-bovine embryos The in vitro development results of cloned ovine-bovine embryos and control bovine-bovine embryos is summarized in Table 1. eGFP bovine and ovine broblast cells were used as nuclear donors for SCNT. During the SCNT procedure, we observed no signi cant difference (P > 0.05) in fusion rates (92.3% vs. 81.3%) between the two groups, indicating that our protocol for cloning of bovine embryos was suitable to produce ovine-bovine iSCNT embryos. Moreover, no difference (P > 0.05) was observed in cleavage rates (85.5% vs. 86.3%) between the groups. The in vitro development of iSCNT embryos was arrested at the 16-cell stage and no morulae and blastocysts were observed (Fig. 1); however, 19.5% of embryos in the control group developed normally to blastocyst stage.

CRISPR/Cas9-introduced indel mutations in genome of iSCNT embryos
Speci c PCR primers (F: ATGGTTGCATTTGAAGGA, R: CCCTGTGCTAGGCAGAAT, 495 bp) were designed and used to amplify exon 12 of sheep CFTR gene and partial intron sequences. Based on the sequencing results, the guide RNA sequence, gRNA_1: AACATAGTTCTTGGAGAAGG, was used to target the G542 locus of wild-type sheep CFTR gene (Fig. 2a). A HpyAV enzyme site was identi ed in gRNA_1, which could be employed to facilitate indel mutation detection. Three days after injection of different concentrations of Cas9:gRNA_1 RNP, mutations in target loci were analyzed in each of iSCNT embryos using PCR/RFLP assays and Sanger sequencing ( Fig. 2b and 2c). The detection results showed that different mutations, including nucleotide deletion, insertion, and replacement, were introduced at target loci of the injected iSCNT embryos. We achieved high biallelic mutation e ciencies ranging from 88.9-100% after injection of Cas9:gRNA_1 RNP at concentrations of 0.02 µg/µl:0.57 µM to 1.4 µg/µl:36 µM (Fig. 2d). The mutation e ciency reduced to 54.5% when the concentration decreased to 5.3 ng/µl:0.14 µM and no biallelic mutations were found using 1.3 ng/µl:36 nM of Cas9:gRNA_1 RNP for injection. Mosaicism in iSCNT embryos was detected when lower concentrations of RNP (64-to 1024-fold dilution) were used ( Supplementary Fig. 1). The cleavage rates of injected iSCNT embryos varied from 77.8 to 100% and no signi cant difference (P > 0.05) was observed among different concentration groups.
Correction of the CFTR/G542X mutation in genome of iSCNT embryos The G542X fetal broblasts were isolated from a cloned G542X fetus produced in our laboratory (unpublished) and subsequently used as nuclear donor cells for the establishment of cloned ovine-bovine embryos. We genotyped G542X broblasts by PCR ampli cation and Sanger sequencing. The results veri ed the presence of homozygous G542X mutations in this cell line, in which the rst nucleotide 'G' at G542 site of CFTR was replaced by a 'T', resulting in the formation of a stop codon 'TGA' (Fig. 3a and 3c). Based on the sequencing results, a gRNA_2: AACATAGTTCTTTGAGAAGG, and a ssODN, 5'-T*T*TGATAATAGGACATCTCCAAGTTTTCAGAGAAAGACAACATAGTTCTTGGAGAAGGTGGAATCACATTGAGTGGAGGTCAGCGAGCAAGAATTTCTT*T*A-3', were utilized to correct the G542X mutation in the iSCNT embryos (Fig. 3a). After injection of Cas9/gRNA_2 and ssODN compound, the correction of the G542X mutation in injected embryos was con rmed by using PCR and Sanger sequencing ( Fig. 3b and 3c Off-target (OT) detection of mutated and corrected iSCNT embryos We chose 11 potential OT sites in sheep genome related to the gRNA_1 and 6 OT sites for the gRNA_2 with the highest sequence homology (scores > 1) (Supplementary Table 1). Speci c PCR primers were designed to amplify DNA fragments ranging from 300 bp to 900 bp spanning each potential OT locus (Supplementary Table 2). Genomic DNAs derived from whole genome ampli cations of 5 mutated and 4 G542X-corrected iSCNT embryos were used as the templates for PCR ampli cation. Sequencing results of PCR products veri ed that none of these embryos had mutations in the analyzed OT sites (Supplementary Fig. 2).

Discussion
The generation of interspecies embryos by cloning, reported more than 30 years ago, uncovered an distinctive fact that they could be the most useful applications for researching the mechanism of nuclear-cytoplasm interactions and conservation and rescue of endangered or extinct species 20 . Here, we expand its applications to the evaluation of embryo gene-editing in species where oocytes are scarce or an access to them is limited. For instance, our lab has a limited access to sheep oocytes while we can routinely obtain a large quantity of cattle oocytes. Our data indicate that ovine-bovine interspecies embryos can be effectively gene-edited by CRISPR/Cas9 approach.
We generated ovine-bovine cloned embryos using eGFP ovine broblasts as nuclear donors, which allowed us to accurately differentiate cloned embryos from those parthenogenetically activated and evaluate the development capacity of reconstructed embryos. The ovine-bovine embryos successfully developed to 16 23 . It has been reported in many studies that the bovine oocytes as recipient cytoplasm support the development at least to 10-to 16-cell of iSCNT embryos produced using nuclear donor cells from other species, including buffalo, sheep, mouse, cat, rabbit, horse 22 . The morula or blastocyst development of iSCNT embryos has been observed even when nuclear donor cells have a distant taxonomical relationship with bovine recipient oocytes, as in the case of interclass chicken-cow, interorder pig-, dog-, monkey-and human-cow 21 . The gaur-bovine iSCNT embryos were able to implant and develop to term 24 . Though term development of iSCNT embryos only limited in several closely related species, short-term in vitro developmental capacity of these embryos can provide an e cient window stage and enough genome materials for gene and mutation analysis.
Injection of Cas9/gRNA compound has been shown to be effective for gene editing of embryos in large animals 9,25 . Assessing mutation e ciency of Cas9/gRNA compound in embryos prior to embryo transfer is critical because it is a time-saving and cost-effective strategy to generate CRISPRmutated large animals. Our results showed that the bases encoding G542 amino acid of sheep CFTR gene in genome of ovine-bovine iSCNT embryos can be disrupted with a high e ciency of 88.9-100% through injection of Cas9/gRNA RNP at different concentrations. Typically, higher RNP concentrations is more e cient, while also lead to increased embryonic toxicity 26,27 . However, we observed no difference in embryo cleavage rates when different concentrations of RNP were used, thus, a wide concentration range is available to test and optimize reagents for e cient gene editing in embryos.
An important factor to be considered in gene editing of embryos using the RNP injection approach is the occurrence of mosaicism in edited embryos, which results from indel introduction after the rst round of DNA replication. We observed embryos with mosaic mutations when lower concentrations of RNP were used. Our results are consistent with ndings in porcine zygotes editing experiments, which reported that lowering Cas9 protein concentrations from 100 ng/µl to 20 ng/µl for cytoplasmic injection decreased both mutation e ciency and the rates for biallelic mutations 28 . The minor discrepancy between our and their results is that we achieved higher biallelic mutation e ciency when using 20 ng/µl of Cas9 protein for injection compared to no biallelic mutations in analyzed porcine blastocysts in their study. Differences in embryos, genes, reagents, and injection methods may account for this discrepancy. Different concentrations of RNP used for embryonic injection in large animals has been reported in other experiments with variable outcomes [29][30][31] . Therefore, we suggest that it is necessary to optimize injection conditions before starting to generate mutate embryos for transfer.
The G542X mutation is the most common nonsense mutation in CFTR. Currently, there are no available treatments for CF nonsense mutations that are both safe and clinically effective 32 . In this study, we achieved up to 11.1% of biallelic correction e ciency in ovine-bovine embryos when Cas9/gRNA was co-injected with a 101nt donor ssODN. Using similar strategy and size of ssODN, Miao et al. conducted the correction of a missense mutation, Fgfr3/G374R and achieved 22.5% correction in newborn mice 33 . In our study, we rst utilized CRISPR editing to correct a CFTR mutation in ovinebovine embryos. Our results support the hypothesis that precise genome editing in embryos by CRISPR could be a true cure for certain genetic diseases. The high degree of sequence similarity between farm animals and humans should allow for accurate modeling of single nucleotide correction in human embryos compared to other trials in mouse embryos or in transfected cell lines. In addition, it is noticeable that we did not detect mutagenesis induced by off-targeting in these mutated and corrected embryos, increasing the odds to utilize this strategy for clinical applications.
In conclusion, cytoplasmic injection of Cas9:gRNA RNP results in highly e cient mutation induction in genome of ovine-bovine iSCNT embryos. The CFTR/G542X mutation can be corrected via co-injection of CRISPR/Cas9 RNP and donor ssODN compound into iSCNT embryos, which could be a potential strategy to correct genetic defect in genome at embryo stage. The use of iSCNT embryos for gene editing overcomes the limitation of oocyte source and provides the opportunity of mimicking the editing of any other genes in embryos of different species. Despite the limited developmental competency, the ovine-bovine iSCNT embryos were able to progress to the 16-cell stage that is su cient for gene-editing optimization and assessment of its e ciency.

Methods gRNAs and ssODN Design.
Speci c primers were designed according to the sheep CFTR genome sequences (GenBank, NC_019461.2) and used to amplify the exon 12 of CFTR and parts of introns anking it. Based on the sequencing results, we designed gRNAs by searching for the (N) 20 NGG motifs using the Benchling software (https://benchling.com/academic). gRNA_1 was chosen to induce DSB at G542 locus in exon 12 of the CFTR gene for mutation e ciency detection. A restriction enzyme site was identi ed in gRNA_1 to facilitate indel mutation e ciency detection using PCR/Restriction Fragment Length Polymorphism (RFLP) assays. gRNA_2 and ssODN were designed to correct the G542X mutation of the CFTR in iSCNT embryos, which were generated by using G542X sheep fetal broblasts as nuclear donors. The size of ssODN was 101nt and the mutated nucleotide, 'G', located in the middle of it was anked with 50 nucleotides at each side. The gRNAs were synthesized by Synthego with chemical modi cations (2'-O-methyl 3'phosphorothioate modi cations in the rst and last 3 nucleotides). The TrueCut™ Cas9 protein V2 was purchased from Thermo Fisher Scienti c. The ssODN was synthesized as the Ultramer→ DNA oligonucleotides with chemical modi cations by Integrated DNA Technologies. SCNT.
Both bovine SCNT and ovine-bovine interspecies SCNT were performed as described by Fan et al. for goats 34 , with modi cations wherein an aspiration technique was used for oocyte recovery instead of a slicing technique and bovine oocyte maturation medium instead of caprine medium. The bovine oocyte maturation medium consists of TCM-199 supplemented with 10% fetal bovine serum (FBS, SH30070.03, HyClone) 5 µg/ml luteinizing hormone (L9773, Sigma-Aldrich), 0.5 µg/ml follicle stimulating hormone (F2293, Sigma-Aldrich), and 100 U/ml penicillin/streptomycin (15140-122, Gibco). The bovine and ovine broblasts expressing the enhanced green uorescent protein (eGFP) were constructed and cryopreserved in our lab. The CFTR/G542X sheep fetal broblasts were isolated from a cloned fetus collected on day 45 of gestation and cryopreserved in our lab. The broblasts at passage 2 were used for second-generation cloning. Both bovine and ovine broblasts were grown to 80% − 90% con uence and used as nuclear donor cells for SCNT after 24 hours of serum starvation (0.5% FBS in DMEM). The cloned embryos were cultured in caprine SOF medium for 7 days 34 . The in vitro development of cloned embryos was observed under a stereo microscope and eGFP expression photographed by a uorescent microscope (Observer Z1, Zeiss).
Cytoplasmic Injection of iSCNT Embryos.
The iSCNT embryos, collected 4 h after activation, were subjected to cytoplasmic injection. Prior to injection, 2 µl of 100 µM gRNA was incubated with 1.5 µl of 5 µg/µl Cas9 protein for 10 min at room temperature to form a ribonucleoprotein (RNP) complex. The RNP was then incubated with 2 µl of H 2 O (for KO) or 200 µM ssODN (for KI) for 5 min and the resulting mixture was used for cytoplasmic injection. A group of 30-50 embryos were transferred to a 50 µl of HSOF drop covered with mineral oil 34 and 5-10 pl of Cas9:gRNA or Cas9:gRNA:ssODN compound was injected into the cytoplasm of an embryo. The inter diameter of an injection pipette was 5-7 µm and the injection for each group of embryos was completed within 20 min using a Piezo micromanipulator (PiezoDrill, Burleigh). The injected embryos were temporarily cultured in SOF for 10 min to assess postinjecting survival rate, and then the survival embryos were transferred to a new 30 µl SOF drop for culture.

Mutation analysis of injected embryos.
The embryos were collected 3 days after cloning and injection procedure for mutation detection. Crude DNA from each embryo was prepared according to the method described by Sakurai et al. with some modi cations 35 . Brie y, under a stereo microscope, 1-2 µl of SOF medium containing 1 embryo was transferred to 10 µl embryo lysis buffer (200 µg/ml proteinase K, 100 mM Tris-HCl (pH 8.3), 100 mM KCl, 0.02% gelation, 0.45% Tween 20) in a 0.2 ml PCR tube. The mixture was vortexed for 5 s and then incubated at 56 °C for 1 h followed by 95 °C for 10 min in a PCR machine. The resulting crude DNA solution was stored at -20 °C until use. The whole genome ampli cation was conducted using a REPLI-g Mini Kit (Cat. 150023, Qiagen) according to the manufacturer's protocol. The products from whole genome ampli cation of embryos were used for mutation analysis and off-target detection. The mutations of embryos were detected with PCR/RFLP assays and Sanger sequencing as described elsewhere 36 . The number of embryos with bi-allelic mutations was calculated and used for indel mutation and correction e ciency analysis.
Off-target analysis.
We conducted a BLAST search of the sheep nucleotide sequence database with two gRNA sequences as the queries to nd the genomic sequences with the highest homology using Crispor online software (Version 4.97). We chose potential off-target (OT) sites with high sequence homology (scores above 1) to the seed sequences. Speci c PCR primers were designed to amplify DNA fragments of approximate 500 bp spanning each potential OT locus. PCR ampli cation and Sanger sequencing were used for the OT analysis of the genomic DNAs derived from whole genome ampli cations of mutated and G542X-corrected iSCNT embryos.

Statistical analysis.
The experiments for the generation of bovine-bovine and ovine-bovine cloned embryos were repeated for three times. The data for fusion and in vitro development experiments were subjected to arcsine square root transformation and analyzed by unpaired t-test. The data for cleavage rates collected during the production of mutated and G542X-corrected embryos were analyzed by x 2 test. A difference was considered signi cant when the P value was less than 0.05. Figure 1 In vitro development of cloned bovine-bovine and ovine-bovine embryos. Cleaved embryos and weak eGFP expression were observed in both groups on Day 2 after activation. The in vitro development of iSCNT embryos was arrested at the 16-cell stage and eGFP expression was weaker in ovinebovine group than in the control on Day 4. The embryos at morula and blastocyst stages with strong eGFP expression were found in the control group, but not in iSCNT group, when observed on Day 6 and 7. WL, white light.