Highly efficient generation of biallelic reporter gene knock-in mice via CRISPR-mediated genome editing of ESCs

Targeted gene knockout and knock-in mice are valuable tools for elucidating the function of genes in vivo (Capecchi, 2001). Recently, the Cas9 endonuclease from Streptococ-cus pyogenes type II CRISPR system has been demonstrated as a powerful tool for gene targeting. Under the guidance of a synthetic 20-nucleotide single guide RNA (sgRNA), Cas9 protein can bind to specific genome locus and generate targeted double-stranded break (DSBs) to facilitate efficient genome editing (Cong et al., 2013; Mali et al., 2013). A one-step method has been reported to generate gene-targeted knock-in mice by injecting Cas9 mRNA in combination with a sgRNA and single-stranded DNA oligo complex/construct into the cytoplasm of a zygote (Hai et al., 2014; Wang et al., 2013; Yang et al., 2013). However, it is inefficient to generate knock-in mice that carry a reporter gene. Using Cas9 protein combined with mRNA of dual-crRNA and tracrRNA has greatly increased the efficiency (Aida et al., 2015), but in most mice resulted in the modification of a single allele. There were few reports on the generation of reporter knock-in mice from embryonic stem cells (ESCs) that were genetically modified by the CRISPR/Cas9 system. Here, we report an effective method, combining the CRISPR/Cas9 system and eight cell-stage embryo injection technology, leading to rapid generation of biallelically modified reporter knock-in mice within one month. We chose Tbx3 as the candidate gene to demonstrate the feasibility of this method because we have reported that Tbx3 can improve the quality of induced pluripotent stem cells. To protect the integrity of Tbx3 function, we designed sgRNA targeting site just upstream the stop codon of Tbx3 and used the self-cleaving 2A peptide for mediation of the fusion between Tbx3 and GFP (Tbx3-2A-GFP) (Fig. 1A). A plasmid containing Cas9 and sgRNA was transfected into ES cells (G4 cell line), and T7 endonuclease I (T7EI) assay verified a target site efficiency of 46% (Fig. 1B). For generating Tbx3-2A-GFP ESCs, we transfected ESCs with the CRISPR expression vector combined with the Tbx3-2A-GFP donor plasmid. Tbx3 expression pattern in ESCs is heterozygous and most cells are Tbx3-negative. Addition of two small-molecule inhibitors (2i) of mitogen-activated protein kinase (MAPK) and glycogen synthase kinase 3 (GSK3) pathways facilitated a homogenous Tbx3 positive ESC stage (Fig. S1A), which were helpful for sorting correct knock-in ESCs. Therefore, we supplemented the media with 2i the day following transfection. GFP expression was detected as early as 48 h post-transfection (Fig. …


Targeting vector
To construct donor vector, we amplified chromosome sequence flanking the stop codon of Tbx3 locus as homology arms. When designing primers for amplifying 5'-homologous arm, we introduced degenerate bases to remove CRISPR targeting site within the donor vector. Through fusion PCR, eGFP was fused with 5'-homologous arm preceded by a 2A self-cleavage peptide.
For differentiation, we use standard embryonic bodies (EB) formation method. ESCs were cultured in non-adherent conditions at a density of 1×10 5 cells/ml for 2 days on a rotator with 40 rpm. The formed EBs were cultured on plated coated with gelatin for another 5 days before detection lineage markers with immunofluorescence staining.
Electrotransfection was performed using Nucleofector Kits for Mouse Embryonic Stem Cells (Lonza) following manufacturer's protocol.

T7 endonuclease I assay and indel rate analysis
G4 cells were transfected with plasmid DNA as described above and harvested 2 days post transfection. Genomic DNA of 2×10 6 cells were extracted using DNeasy Blood & Tissue Kit (Qiagen) following manufacturer's protocol.
Genomic region flanking targeting site was amplified using high-fidelity DNA polymerase (TransGen Biotech), and the product was purified with QIAquick PCR Purification Kit (Qiagen) following manufacturer's protocol. A total of 400ng of the purified PCR product was mixed with 2µl NEBuffer 2 (NEB) and ultrapure water was added to a final 20µl volume. Heteroduplex were formed using a Thermocycler in following program: 95º C for 10min, 95º C to 85º C ramping at -2º C /s, 85º C for 1min, 85º C to 75º C ramping at -0.3º C /s, 75º C for 1min, 75º C to 65º C ramping at -0.3º C /s, 65º C for 1min, 65º C to 55º C ramping at -0.3º C /s, 55º C for 1min, 55º C to 45º C ramping at -0.3º C/s, 45º C for 1min, 45º C to 35º C ramping at -0.3º C/s, 35º C for 1min, 35º C to 25º C ramping at -0.3º C/s, 25º C for 1min. After heteroduplex formation, add 0.5ul T7 Endonuclease I (NEB) and hold at 37º C for 1 hour. The digested products were analyzed by 10% polyacrylamide gel run in 1×TBE. The indel rate was calculated with ImageJ software.

Embryonic Microinjection
To guarantee the injection of ES cells into eight-cell stage embryo before compaction, we collected the two-cell stage embryo and cultured Injected embryos were cultured to develop into blastocyst in vitro as described above.

Embryo transfer
We used CD1 females mated with vasectomized CD1 males as recipients for injected embryos. We usually prepared two kinds of recipients which were 2.5dpc and 0.5dpc, depending on the development stage of injected embryos. For blastocyst stage, twelve to fifteen embryos were transferred into uterus of 2.5dpc pseudopregnan CD1 females and morula stage embryos into oviduct of 0.5dpc recipients.

Southern blot analysis
Genomic DNA of cells and tails of mice were extracted using the method of phenol-chloroform extraction. A total of 5µg genomic DNA was separated on a 0.7% agarose gel after digested by BglII (NEB), and then transferred to a nylon membrane (Roche) and hybridized with PCR based Dig(Roche)-labeled probes.

Immunohistochemistry
After fixation in 4% PFA, the embryo was dehydrated in a series of grade ethanol, embedded in paraffin and cut into section. Heat induced epitope retrieval was performed using citrate buffer (pH 6.0). For detection of GFP, we use anti-GFP rabbit monoclonal antibody (2956S, Cell Signaling) at 1:200 dilution rate.

Real-time (quantitative) PCR
Total RNA was extracted using RNeasy Mini Kit (Qiagen) in accordance with the mannufacturer's protocol. cDNA were synthesized with oligo-dT primer by M-MLV Reverse Transcriptase Kit (Promega).
Q-PCR reactions were performed using the SYBR Green I Master Mix and LightCycler 480 (Roche). Gene-specific primers for Q-PCR can be found at previous study (Han et al., 2010).

Prediction and detection of potential off targets
We screened potential off target sites around the mouse genome (mm10) with CasOT software (Xiao et al., 2014) based on the role: screened potential target sites allowing for up to three base pair mismatches compared with sgRNA and the other three different PAM in the first base pair . Genomic DNA regions around potential off target sites were amplified, purified and analyzed by T7EN I analysis as described above.

Statistical analyses
Student's t-tests were used to compare differences between any two groups. Table S1. Off-target Analysis, Related to Figure 1 and S3.
Mismatches between potential off-target site and on-target sequence are shown in lower-case, boldface and underlined. Sequence of PAM and sgRNA is separated by hyphen. Coordinate shows the location of potential off-target site in mice genome. Indel mutation frequencies in targeted mice were calculated by T7EI assay. OT indicates off-target; /, not tested.