CRISPR/Cas9-mediated gene manipulation to create single-amino-acid-substituted and floxed mice with a cloning-free method

Clustered regulatory interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) technology is a powerful tool to manipulate the genome with extraordinary simplicity and speed. To generate genetically modified animals, CRISPR/Cas9-mediated genome editing is typically accomplished by microinjection of a mixture of Cas9 DNA/mRNA and single-guide RNA (sgRNA) into zygotes. However, sgRNAs used for this approach require manipulation via molecular cloning as well as in vitro transcription. Beyond these complexities, most mutants obtained with this traditional approach are genetically mosaic, yielding several types of cells with different genetic mutations. Recently, a growing body of studies has utilized commercially available Cas9 protein together with sgRNA and a targeting construct to introduce desired mutations. Here, we report a cloning-free method to target the mouse genome by pronuclear injection of a commercial Cas9 protein:crRNA:tracrRNA:single-strand oligodeoxynucleotide (ssODN) complex into mouse zygotes. As illustration of this method, we report the successful generation of global gene-knockout, single-amino-acid-substituted, as well as floxed mice that can be used for conditional gene-targeting. These models were produced with high efficiency to generate non-mosaic mutant mice with a high germline transmission rate.

Scientific RepoRts | 7:42244 | DOI: 10.1038/srep42244 To overcome the target vector construction and in vitro RNA transcription required for sgRNA, a recent study used crRNA and tracrRNA in combination with commercially available Cas9 protein and a regular gene targeting vector, to generate knock-in mice 2 . In our current study we report a method that capitalized on a cloning-free CRISPR/Cas9 system using commercial Cas9 protein combined with chemically synthesized crRNA, tracrRNA, and single-strand oligodeoxynucleotides (ssODNs). This system allowed us to consistently obtain high-efficiency editing of multiple genes in the mouse, and successfully generate non-mosaic mutant mouse models with a variety of editing schemes, including, frame-shift indel mutations, single-amino-acid substitutions, and LoxP-inserted conditional alleles. Taken together, our method is a simple, cloning-free, and highly efficient technique, which accelerates our ability to produce rapid mouse genome editing in vivo.

Results
Generation of gene-knockout mice. The double-strand breaks (DSBs) induced by the CRISPR/Cas9 system stimulate DNA repair by at least two distinct mechanisms, non-homologous end joining (NHEJ) and homology-directed repair (HDR) 13,14 . NHEJ is error-prone and introduces unpredictable patterns of insertions and deletions, which can lead to disruption of the protein-coding capacity of a defined locus. To investigate whether gene-knockout mice can be efficiently generated by direct pronuclear injection of a Cas9 protein:crRNA:tracr-RNA complex into mouse zygotes (Fig. 1), we designed crRNA targeting alpha kinase 2 exon 3, which contains the start codon (Fig. 2a). Using the T7 endonuclease 1 (T7E1) assay to detect on-target CRISPR/Cas9 events, we found that six out of 16 (37.5%) mice contained indel mutations (Fig. 2b). Subsequent sub-cloning of flanking regions surrounding the crRNA targeting site identified three mice with frameshift mutations (Fig. 2c). Each of these mice appears to be non-mosaic as they contain only wildtype and one type of mutant allele. Generation of single amino-acid-substituted mice. Single amino-acid substitutions frequently cause genetic diseases in man and thus recapitulation of these variants is used to produce mouse models of disease. To investigate whether mice with single amino-acid-substitutions could be generated easily using the same approach noted above, we aimed to generate founder mice harboring an arginine (257) to histidine amino acid substitution in the Nesprin-1α 2 (SYNE1) gene (Figs 1 and 3a). We injected a mixture of 30 ng/μ L Cas9 protein, 0.6 pmol/μ L SYNE1-R257H crRNA, 0.6 pmol/μ L tracrRNA, and 20 ng/μ L ssODN into the pronuclei of zygotes. In order to quickly screen for correctly targeted mice, we designed mutation-specific primers with a site-specific variant sequence at the 3′ terminus of the forward primer, as shown in Fig. 3b. We found six of 17 newborn mice contained the correct gene variant allele (Fig. 3c). The HDR efficiency of the targeted mutation by Cas9 protein:crRNA:tracrRNA complex injection was thus 35%. The genotypes of PCR-positive mice were confirmed by sequencing the six individual sub-clones (Fig. 3d). We found that the six correctly targeted mice contained only two types of alleles, compared with multiple alleles as reported by Cas9 mRNA/sgRNA injection 15 , indicating that the founders are heterozygous mice rather than chimeras.
To investigate germline transmission of the variant alleles to the F1 generation, we backcrossed F0 knock-in mice with wildtype C57/B6 mice. We tested three different variant lines using PCR and subsequent sequencing, and found that all F0 mice showed successful germline transmission with an average efficiency of 52.8% (ranging from 36% to 71% among different founders). A representative genotyping screen of offspring from one founder (designated #13) is provided (Fig. 3e). Consistent with only two types of alleles in the founders, these results indicate that the founders are heterozygous mice rather than chimeras. We have so far successfully generated eight different variant mice using this method (Table 1).
Finally, we investigated off-target cleavage in the CRISPR/Cas9 generated mice produced with our methodology. This has been a serious problem associated with CRISPR/Cas9-mediated genome editing techniques in the past 11,16,17 . Here we tested the top five predicted off-target sites for mutagenesis using the CRISPR Design (MIT) website 18 (Fig. 4a). A region flanking each predicted off-target site was amplified by PCR from wildtype and two mutant founders (Fig. 4b). The same molecular size band resulted from all samples, indicating there were no large deletions flanking the sites. Furthermore, we confirmed the sequence flanking the predicted off-target sites by    sequencing six individual clones of each founder (Fig. 4c). The sequencing data confirmed no off-target cleavage, suggesting that genome editing by Cas9 protein:crRNA:tracrRNA complex injection is highly specific.
Generation of conditional knockout mice. We further investigated whether mice could be generated using this one-step approach by insertion of two LoxP sites into the same allele of the paxillin gene (Fig. 1). Use of this facilitated methodology would greatly accelerate the time to produce a model that could be used to conditionally manipulate a gene of choice. We designed a crRNA targeting the paxillin gene intron 5 (left) and a crRNA targeting paxillin intron 6 (right), as well as the corresponding LoxP site oligos with 60 bp homology sequences on either side surrounding each Cas9-mediated DSB (Fig. 5a). Mutation-specific PCR analysis identified that two out of 12 (17%) mice had either left (#10) or right (#2) LoxP insertion (Fig. 5b). One pup (#5) however, was shown to have successful insertion of both LoxP sites in the paxillin allele (Fig. 5b). The precise integration of LoxP sites into the same allele was confirmed by sequencing PCR products generated from DNA from this model (Fig. 5c). To demonstrate that the floxed allele is indeed functional, we crossed the paxillin floxed mice with germline deleter Cre mice (Sox2-Cre) 19 . As shown in the Fig. 5d, the DNA between the two LoxP sites was deleted by Cre recombinase.

Discussion
The CRISPR/Cas9 system is proving to be a powerful yet simple tool to manipulate the genome for the generation of genetically modified animals. In this study, we report the successful generation of both single amino-acid-substituted and floxed alleles, using a simplified, cloning-free CRISPR/Cas9-mediated method. Our system offers an advantage over more traditional methods since it generates non-mosaic mutant mouse models, and does so without CRISPR/Cas9 vector construction or in vitro RNA transcription manipulation.
In each of our mutant mice lines generated by pronuclear injection of a Cas9 protein:crRNA:tracrRNA complex, with or without ssODN, only two types of alleles resulted with a high frequency of germline transmission (~50%), supporting the notion that founders are germline heterozygous, and not chimeric. In our study, we also did not observe any detectable off-target effects in the top five predicted off-target sites for mutagenesis using Cas9 protein:crRNA:tracrRNA-mediated genome editing, which is consistent with other methods 2 . This suggests that it is feasible to achieve low off-target effects through injection of Cas9 protein into the pro-nuclei of mouse zygotes. No chimerism and detectable off-target effects observed using our approach may reflect that the Cas9 protein has a relatively shorter half-life in embryos than Cas9 mRNA 20 . This phenomenon is also consistent with the finding that a Cas9 protein-RNA complex was rapidly degraded in cultured cells 21,22 . The rapid degradation of Cas9 protein may be advantageous in avoiding side effects of Cas9 nuclease activity. Another possibility is that it takes time for Cas9 mRNA to translate into Cas9 protein 23 . When Cas9/sgRNA complexes translocate to the nucleus with the function of the nuclear localization signal, the donor DNA may be transferred in a delayed fashion 24 . In this scenario, repair of DSBs by NHEJ could already be in process prior to the arrival of donor DNA at the target site.
In summary, we report a novel method for the highly efficient generation of non-mosaic mutant mouse models with indel mutations, single-amino-acid substitutions, and LoxP-inserted alleles, through one-step injection of commercially available Cas9 protein combined with chemically synthesized crRNA and tracrRNA. This simple, customizable and ready-to-use genome editing system accelerates our ability to produce rapid mouse genome editing in vivo.

G T T G G A G A T T T C A G A A A G G A C G G
OT #1 Methods CRISPR target sequence design. Guide sequences for CRISPR/Cas9 injection were designed as previously detailed 18 . The target sequence preceding the protospacer adjacent motif (PAM) was obtained from the exon region of the indicated genes.

T C T C G A G G T T T C A G A A A G G A A G G G C T G G A G G T A G C A G A A A G G A C A G G A T G C A G A T T T C A G A A A G G A C A G A T G G A G T T T C C A G A A A G G A A A G G
Chemical synthesis of crRNA, tracrRNA, and ssODN. The crRNA and tracrRNA were designed as previously reported 2 , chemically synthesized, and RNase-Free HPLC purified by Integrated DNA Technologies (Coralville, IA, USA). Single-strand ODN was chemically synthesized and standard desalted by Integrated DNA Technologies (Coralville, IA, USA). All sequences are listed in Table 2.    PCR screening and sequence analysis. To prepare genomic DNA, mouse tails were incubated in 300 μ L of 50 mM NaOH at 98 °C for 30 minutes, then 50 μ L of 1 M Tris-HCl (pH 8.0) was added to the resultant solution. Genomic fragments at targets sites were amplified by PCR with Taq (NEB # E5000S) and two pairs of primers. All primers used in this study are listed in Table 3. The PCR conditions were as follows: 30 cycles of 94 °C for 20 seconds, 60 °C for 20 seconds and 72 °C for 30 seconds. The resultant PCR amplicons were analyzed by electrophoresis on a 2% agarose gel. For sequencing, PCR products were further sub-cloned using a Zero Blunt TOPO PCR cloning kit (Life Technologies, Grand Island, NY). The plasmid DNAs containing the genomic fragments were prepared from individual colonies. Six individual clones from each PCR reaction were sequenced.
T7 endonuclease I (T7E1) assay. The DNA fragment that overlaps the crRNA target site was amplified by PCR using the following primers: forward primer 5′ -AGGCTATTCTGTAGCTTCGCTC-3′ and reverse primer 5′ -CATTATCTGTCATTCAAGCCGTAAG-3′ The purified PCR product (~250 ng) was denatured and reannealed in NEB Buffer 2 (NEB) to form heteroduplex DNA, which was subsequently digested with T7E1 (NEB, M0302L) at 37 °C for 15 minutes and analyzed on a 2% agarose gel.