Introduction

Genetically modified pigs have many applications in biomedical and agricultural research. They have been commonly used to model various human diseases and develop therapeutic strategies. CRISPR/Cas9 system as a new gene-editing technique, found in bacteria and archaea, has been successfully used to create many gene-targeted animals, such as mouse, rat, and monkey, because of its high efficiency and availability [1–8]. Its worth noting that these animals have been generated by co-injection of Cas9 mRNA and sgRNA into one-cell stage embryos. Double-strand break and non-homologous end joining mediated by Cas9 mRNA and sgRNA occur in the target locus probabilistically; as a result, mosaic mutation and numerous alleles in target site are formed, which would complicate subsequent studies. Therefore, many of the resulting founder animals are chimeric with multiple mutations. To establish animals with single mutation, one or two rounds of further breeding and selection need to be performed among the offspring, but this procedure is not easily applicable in large animals, such as pigs, because they have long gestation cycles and entail high recipient costs. In addition, the founders without germline mutation or mosaicism would be useless in further breeding and therefore related research. Hence, we aimed to target genes in somatic cells by applying CRISPR/Cas9 technology and then perform SCNT to produce knockout pigs in this study. CRISPR/Cas9 system-mediated gene editing in porcine fetal fibroblasts (PFFs) has been reported by Tan et al. [9]. The efforts that used mutant cells as donor nuclei to generate gene-targeted pigs carrying stable and identical mutations have not been followed.

Here, we aimed to investigate the feasibility of generating gene-targeted pigs by targeting genes in somatic cells via CRISPR/Cas9 system and SCNT approach. We initially targeted a gene, TYR, encoding tyrosinase, which is a key enzyme involved in melanin biosynthetic pathway. The genetic deficiency of TYR gene in human can cause oculocutaneous albinism type 1 (OCA1), an autosomal recessive disorder characterized by reduced melanin pigment in the hair, skin, and eyes [10, 11]. CRISPR/Cas9 system can also be used to target multiple genes in cells and embryos with high efficiency [2, 5, 12]. As such, we also aimed to target two genes in the same somatic cell simultaneously. We chose PARK2 and PINK1 genes as the loci of interest. PARK2 gene encodes a protein called parkin, a component of multiprotein E3 ubiquitin ligase complex. PINK1 gene encodes PTEN-induced putative kinase 1, a mitochondrial serine/threonine-protein kinase. The mutation of either of these two genes would lead to autosomal recessive early-onset Parkinson’s disease in humans [13, 14].

Materials and methods

Animals

The animal use protocol complied with the guidelines of Institutional Animal Care and Use Committee of Guangzhou Institute of Biomedicine and Health, Chinese Academy of Sciences (Animal Welfare Assurance #A5748-01). All surgical procedures were performed under anesthesia, and all efforts were made to minimize animal suffering.

Construction of Cas9 and sgRNA vectors

CMV promoter-driven Cas9 (41815) and Cas9-nickase (Cas9n, 41816) plasmids were purchased from Addgene. U6-sgRNA cloning vector was constructed by introducing 2 BbsI restriction sites to the downstream region of U6 promoter of plasmid gRNA_GFP-T1 (41819, Addgene). Targeting sgRNA was designed by GN20GG rule [15, 16]. Two complementary oligo DNA of sgRNA were synthesized and then annealed to a double-strand DNA, ligated to the BbsI sites of U6-sgRNA cloning vector to form sgRNA expressing plasmid. These constructs were confirmed by sequence analysis (BGI, Guangzhou, China).

PFFs cell culture, transfection and selection

Porcine fetal fibroblasts were isolated from 35-day-old fetuses of two Chinese indigenous strains, called the Banna mini-pig and Bama mini-pig. The fetuses, removed heads, tails, limbs and viscera, were digested in cell culture medium containing 0.32 mg/mL collagenase IV and 2,500 IU/mL DNase I for 3 h at 39 °C. Isolated PFFs were then cultured in 10 cm dishes for 12 h and frozen in fetal bovine serum (FBS) containing 10 % dimethylsulfoxide. A day before transfection, PFFs were thawed and cultured to sub-confluency. Then, approximately 1 × 106 PFFs were electroporated with vectors of Cas9 and sgRNA at 230 μV and 500 μF using a Gene Pulse Xcell electroporator (Bio-Rad, Hercules, CA, USA). After 24 h of recovery, the electroporated cells were selected with G418 at appropriate concentration for about 10 days. Individual cell colonies were picked up and cultured in 48-well plates. When confluency was reached, the cell colonies were subcultured and 10 % of each colony was lysed individually in 15 μL of lysis buffer (0.45 % NP-40 plus 0.6 % Proteinase K) for 90 min at 56 °C and then 10 min at 95 °C. The lysate was used as template for PCR. The primers were designed to amplify across the target sites. For TRY gene, the forward primer was 5′-CCTGATGGAGAAGGAATGCTGC-3′ and reverse primer was 5′-TTGGCCATAGGTGCCTGTG-3′. For PINK1 and PARK2 gene, PCR were performed using PINK1 and PARK2 specific primers, respectively, forward primer 5′-GTTTGCCTGGTAGCCAAGGA-3′ and reverse primer 5′-CTTCACTCGTCAGCGACCTCA-3′ for PINK1; forward primer 5′-GGTGAGGGGTAAATCAGTTCA-3′ and reverse primer 5′-CAGCAGAAGCAGCATAAGCA-3′ for PARK2. The PCR conditions were 94 °C for 5 min; 94 °C for 30 s, 60 °C for 30 s, 72 °C for 30 s, for 35 cycles; 72 °C for 5 min; hold at 16 °C. The PCR products were sequenced to identify the existence of the mutation. Some PCR products were selected to ligate into a pMD18-T vector (Takara, Dalian, China) and sequenced to determine the exact mutant sequences. The positive cell colonies were expanded and then cryopreserved. These cells were thawed and cultured to reach sub-confluency in a 24- or 12-well plate before SCNT.

SCNT and production of mutant piglets

Methods used for porcine oocyte collection, in vitro maturation and SCNT were similar to our previous studies [17–19]. Briefly, cumulus oocyte complexes (COCs) were selected and cultured in maturation medium for 42–44 h at 39 °C. Matured oocytes were enucleated by aspirating the first polar body and adjacent cytoplasm with a glass pipette in a manipulation medium of HEPES-buffered M199 plus cytochalasin B (7.5 μg/mL). The cells identified as gene knockout by sequencing were used as donor cells to be injected into the perivitelline space of the oocytes. Fusion and activation were performed with two successive DC pulses at 1.2 kV/cm for 30 μs using an electrofusion instrument. The reconstructed embryos were cultured in PZM3 at 39 °C overnight and surgically transferred to the oviducts of the surrogates that were observed in estrus the day before. The pregnancy status was monitored by ultrasonography 24 days after embryo transfer. Once detected, the pregnancy status was monitored weekly until delivery. The cloned piglets were delivered by natural birth. The genomic DNA that was isolated from the ear skin biopsy of the newborn cloned piglets was assessed for mutagenesis at the targeted site by PCR-based assays.

Histology and immunofluorescence staining

The tissues of skin and eyes from TYR −/− pigs, brains from PARK2 −/−/PINK1 −/− pigs were fixed with 4 % buffered paraformaldehyde for 48 h. The entire tissue was then dissected, embedded in paraffin wax, and exhaustively sectioned at 3 μM to a slide. Sections were deparaffinized in xylene and rehydrated in a graded series of alcohol, followed by dH2O. Sections of skin and eyes from TYR −/− pigs were stained with hematoxylin and eosin, then differentiated, and coverslipped. Immunofluorescence staining was performed on sections of brain cortex from PARK2 −/−/PINK1 −/− pigs. Briefly, antigen retrieval was performed in a 1 M citrate buffer (pH 6.0) bath for 20 min. Tissues were incubated with primary antibodies (rabbit anti-Parkin polyclonal, AB9244, Millipore, Temecula, CA, USA; anti-PINK1 antibody, ab23707, Abcam, Cambridge, UK) overnight in a humid chamber at 4 °C. The tissues were washed thrice with PBS and incubated for 1 h with a Alexa Fluor 594-labeled goat anti-rabbit IgG (H + L) secondary antibody (A-11012, Invitrogen, Eugene, OR, USA) and the slides were analyzed by fluorescence microscopy (Olympus, Tokyo, Japan).

Off-target analysis

Off-target cleavage sites were predicted and searched as previously described [2]. All potential off-target sites (OTS) with homology to 23 bp sequence (sgRNA + PAM) were retrieved by a base-by-base scan of the whole pig genome, allowing for ungapped alignment with up to 6 mismatches in the sgRNA target sequence. All the potential OTS were PCR amplified and sequenced to confirm the off-target effects, respectively. The primers for amplifying the off-target sites are listed in Supplemental Table 1.

Results

CRISPR/Cas9-mediated gene targeting in PFF cells

For PARK2 and PINK1 genes, sgRNAs were designed based on the original Cas9 approach [15, 16]. PARK2-sgRNA and PINK1-sgRNA target the first exon region following the start codon, respectively (Fig. 1a). For the TYR gene, a method based on modified Cas9 called Cas9-nickase was used to design sgRNAs [20]. Cas9-nickase harbors a D10A mutation and can only nick one strand of double-stranded DNA. Double nicks are created when a pair of sgRNA functions with Cas9-nickase, consequently cutting the double strands of the target DNA. Cas9-nickase-based approach could reduce the risk of off-target and inflict less damage to the integrity of a genome [20]. Figure 1a shows the sgRNA pair targeting the TYR gene exon 1 region.

Fig. 1
figure 1

CRISPR/Cas9 mediates gene targeting in PFFs. a Schematic of sgRNAs targeting TYR, PARK2 and PINK1 loci. The target loci are located in the first exon following the start codon of these three genes. sgRNA targeting sites are highlighted in red. PAM are highlighted in blue and underlined. b Sanger sequencing of the targeting sites in mutant colonies used in SCNT. For each gene, the wile-type sequence is shown at the top with the target sites highlighted in red. The net change in length caused by each mutation is to the right of each sequence (+, insertion; ∆, deletion). Lowercase letters denote inserted base pairs. PAM is highlighted in blue. Indel insertion or/and deletion

Full size image

To target PARK2 and PINK1 genes, we co-transfected Cas9, PARK2-sgRNA, and PINK1-sgRNA expressing vectors into PFFs derived from a 35-day-old fetus by electroporation. To target the TYR gene, we co-transfected Cas9-nickase and two TYR-sgRNAs encoding vectors into PFFs. After viable cells were selected using G418 for approximately 10 days, surviving cell colonies were collected and then analyzed individually by sequencing PCR products covering the target locus. The mutation rates among the selected cell colonies at the target loci were 49.4, 66.7, and 69.9 % for TYR-sgRNAs, PARK2-sgRNA and PINK1-sgRNA, respectively (Table 1). Cas9 and Cas9-nickase can both be used efficiently to generate KO cell colonies. Cas9-nickase was slightly less efficient in targeting genes possibly because of the lower co-transfection efficiency of three plasmids when compared with using Cas9. Mutation patterns remarkably varied among the positive colonies (data not shown). The percentage of cell colonies containing double homozygous indels of sequences in PARK2 and PINK1 genes was up to 38.1 % (4 mutant alleles shown in Table 1).

Table 1 CRISPR/Cas9-mediated gene targeting in PFFs
Full size table

Generation of gene-targeted pigs via SCNT

Six TYR −/− cell colonies and 15 PARK2 −/−/PINK1 −/− double-KO cell colonies, containing deletions from 1 to 60 bp and insertions from 1 to 102 bp (Fig. 1b), were chosen as donors for SCNT. We usually pooled 3–6 colonies in one nuclear transfer and the reconstructed embryos from the pooled cells were transferred to a surrogate simultaneously (Table 2). A total of 1,705 reconstructed embryos derived from TYR −/− cells were introduced to 7 surrogate mothers. Among these surrogate mothers 4 developed to term and gave birth to 18 female cloned piglets (Fig. 2a; Table 2). Among the 18 piglets, 8 were weak at birth and died in 2 weeks. The remaining piglets appeared healthy (Table 3). A total of 1,729 reconstructed embryos derived from PARK2 −/−/PINK1 −/− double-KO cells were introduced to 10 surrogate mothers. Among these surrogate mothers, 4 developed to term and gave birth to 20 cloned piglets (Fig. 2a; Table 2). Among these 20 piglets, 2 were stillborn, 7 were weak at birth and died in 2 weeks. Five piglets were born healthy but also died 1 month after birth. The remaining piglets developed normally (Table 3).

Table 2 Somatic cell nuclear transfer results for the generation of knockout cloned pigs
Full size table
Fig. 2
figure 2

Generation of gene-targeted piglets via SCNT. a Newborn piglets carrying TYR gene mutation (left panel) and PARK2/PINK1 double mutations (right panel). The left panel shows four newborn piglets from the same litter; the white piglets are TYR deficient exhibiting a typical albinism phenotype (asterisk) and the black piglets are wild-type littermates. The bottom right box shows the light pink-tinted iris and pupil of the TYR KO piglets (double asterisk) compared with the dark iris of WT piglets. The right panel shows PARK2/PINK1 double-KO piglets. b Sanger sequencing of the target sites in all cloned pigs. For each gene, the wild-type (WT) sequence is shown at the top in which the target sites are highlighted in red. Right-most column shows the indel of each sequence. TYR-8, 9 and 18 are WT. Deletion and insertion are denoted as “∆” and “+” plus the number of base pairs. PAM is highlighted in blue and lowercase letters indicate inserted base pairs

Full size image
Table 3 Summary of knockout pigs for their health status and origin donor cells
Full size table

DNA sequencing analysis results revealed that 15 of the TYR −/− newborns and 20 PARK2 −/−/PINK1 −/− newborns carried the expected homozygous mutations at the target loci (Fig. 2b). Among the TYR −/− piglets and PARK2 −/−/PINK1 −/− piglets, various types of mutations corresponding to the donor cells were found at the target loci, including deletions from 1 to 60 bp and insertions from 1 to 102 bp (Figs. 1b, 2b). The mutation pattern in four of the piglets was not found among the donor cell lines, which may be caused by the mixed cells (Table 3). As the same reason, there were three unexpected WT pigs due to impure colonies that contain few WT cells (Table 2).

Phenotype identification of gene-targeted pigs

We further examined whether gene mutations cause a depletion of protein expression or related phenotypes. The TYR biallelic KO pigs showed typical albinism. The dark pigment in skin, hair, and eyes were completely lost in these TYR biallelic KO pigs; whereas the wild-type (WT) littermates were black in the same tissues (Fig. 2a). Melanin was present in WT piglets but was confirmed absent by histologically analyzing the sections of the skins and irides of dead TYR biallelic KO piglets with albinism (Fig. 3a, b). These results indicated that tyrosinase and its downstream products were eliminated completely in TYR biallelic KO pigs. For PINK1 and PARK2 double-KO piglets, immunofluorescence analysis was performed to determine the presence of the two proteins. PINK1 and parkin proteins were expressed in the brain neurons of WT pigs but were undetectable in PARK2 −/−/PINK1 −/− pigs. This result confirmed that parkin and PINK1 were depleted at a protein level (Fig. 3c). Thus far, typical symptoms of Parkinson’s disease, such as shaking, rigidity, slowness of movement, and difficulty in walking, have not been observed in 7-month-old live mutant pigs in this study.

Fig. 3
figure 3

Phenotype identification of TYR −/− and PARK2 −/−/PINK1 −/− piglets. a H&E staining of the skin of WT and TYR KO piglets. The arrowheads indicate the melanin in the basal layer of the epidermis. The arrows point to the cross-sections of hair strands with different colors from TYR KO pig and WT pig. b H&E staining of irides of WT and TYR KO piglets. The arrowheads indicate the melanin in the iris of the WT piglets. c Parkin and PINK1 immunofluorescence analysis in the brain cortex sections of WT and PARK2/PINK1 double-KO piglets. The cortex neurons of KO pigs are negative for anti-parkin and anti-PINK1 staining; by contrast, the WT controls show positive staining in the brain neurons. Scale bars 50 μm

Full size image

Off-target analysis of gene-targeted pigs

Off-target effect is a major drawback concern of the CRISPR/Cas9 system. To test whether off-target occurred in these genetically modified piglets, we screened the pig genome and predicted 16, 37, 10, and 24 potential OTS for TYR-sgRNA1, TYR-sgRNA2, PARK2-sgRNA and PINK1-sgRNA sites, respectively (Supplemental Table 2). Genomic DNA from the ear tissues of the 35 mutant pigs was used to perform off-target analysis. The fragments around the potential off-target loci were amplified by PCR and then sequenced individually. As a result, none of the sequencing reads exhibited mutations, suggesting that no off-target occurred at these 87 potential OTS in any of the 35 piglets derived from Cas9 and Cas9-nickase targeting.

Discussion

In this study, we used Cas9 and Cas9-nickase both and we found the two methods were both efficient in generating knockout pigs. Cas9-nickase showed a bit lower efficiency in targeting that may be caused by lower co-transfection efficiency of three plasmids. Although we did not detect potential off-targeted mutagenesis in our mutant pigs, it surely exists and caused unexpected side effect in many other studies [21–23]. We can consider using Cas9-nickase in case of higher off-target site effect of Cas9, since there is no obvious difference in targeting efficiency between Cas9 and Cas9-nickase. Recently, there was also a report about using truncated sgRNAs (17 or 18 bp) to enhance targeting specificity of Cas9 [24]. These improved CRIPSR/Cas9 versions can reduce off-target mutagenesis in somatic cells and therefore help to increase cloning efficiency.

Previously, generation of gene-targeted animals with CRISPR/Cas9 system was realized through co-injection of Cas9 mRNA and sgRNA together into one-cell stage embryos. Instead, in this study we used plasmids of eukaryotic expressing CRISPR/Cas9 to mutate the genome of somatic cells. Constitutive expression of CRISPR/Cas9 components can lead to higher on-target editing efficiencies. However, extended persistence of these components in the cell might also lead to increased frequencies of off-target mutations, which has been previously reported with ZFNs [25]. Severe off-target side effect would impair the integrity of genome of cell and damage its development in embryo manipulation. Therefore, the clonability of the mutant cells achieved with constitutive expression of CRISPR/Cas9 components needs to be determined. To address this concern, we transferred the mixed embryos derived from different mutant cell lines (3–6) to a single recipient. Table 2 shows that 4 of 6 TYR KO cell lines and 12 of 15 PARK2/PINK1 KO cell lines can develop and form cloned pigs; this result indicated that the cloning efficiency of mutant cells was not compromised substantially. We also observed that the number of piglets developed from some particular KO cell colonies (such as T5) was evidently more than that of the other KO cell colonies; hence, these particular KO cell colonies may exhibit better competency to develop to term. All of the fibroblasts subjected to mutation were derived from the same fetus; as such, failure to produce cloned piglets by using some cells may be caused by off-target. Therefore, mixed embryos derived from multiple cell lines could be transferred to a single surrogate, and this procedure could be a practical strategy to increase the chance to achieve cloned gene-targeted pigs via SCNT.

In summary, our study showed that gene-targeted (single- or double-gene targeted) pigs can be effectively achieved by using the CRISPR/Cas9 system combined with SCNT without detectable off-target effects. This gene-editing system provides a convenient and efficient approach to generate genetically modified pigs with agricultural and biomedical applications.