CRISPR/Cas9 mediated genome editing in ES cells and its application for chimeric analysis in mice

www.nature.com/scientificreports OPEN received: 11 May 2016 accepted: 21 July 2016 Published: 17 August 2016 CRISPR/Cas9 mediated genome editing in ES cells and its application for chimeric analysis in mice Asami Oji1,2,3,*, Taichi Noda1,3,*, Yoshitaka Fujihara1, Haruhiko Miyata1, Yeon Joo Kim1,†, Masanaga Muto1,2,3, Kaori Nozawa1,3,4, Takafumi Matsumura1,2, Ayako Isotani5 & Masahito Ikawa1,2,4,5 Targeted gene disrupted mice can be efficiently generated by expressing a single guide RNA (sgRNA)/ CAS9 complex in the zygote. However, the limited success of complicated genome editing, such as large deletions, point mutations, and knockins, remains to be improved. Further, the mosaicism in founder generations complicates the genotypic and phenotypic analyses in these animals. Here we show that large deletions with two sgRNAs as well as dsDNA-mediated point mutations are efficient in mouse embryonic stem cells (ESCs). The dsDNA-mediated gene knockins are also feasible in ESCs. Finally, we generated chimeric mice with biallelic mutant ESCs for a lethal gene, Dnajb13, and analyzed their phenotypes. Not only was the lethal phenotype of hydrocephalus suppressed, but we also found that Dnajb13 is required for sperm cilia formation. The combination of biallelic genome editing in ESCs and subsequent chimeric analysis provides a useful tool for rapid gene function analysis in the whole organism. The genome sequencing projects revealed that approximately 20,500 protein-coding genes are encoded in the 3.3 billion base-pairs of DNA sequence in both humans and mice (http://www.genome.gov/12011238). In addition to the protein coding regions, flanking regions play important roles in regulating gene expression. Recent studies also highlighted the importance of non-coding RNAs such as miRNAs for transcriptional and post-transcriptional control1–4. Overall, the genetic sequence of the genome is the blueprint of all living organisms. Gene-manipulated animals have been a mainstay in studying physiological functions. The adaptation of gene targeting in ESCs and subsequent chimeric mouse production and intercrosses enables us to generate homozygous knockout (KO) mice and study gene function in mammals5–7. However, only a few laboratories can take advantage of these gene targeting strategies because these experiments are time-consuming, laborious, and costly. In 2006, the International Knockout Mouse Consortium started to generate a predesigned knockout mouse library to accelerate physiological studies5–7. The targeting strategy is well designed so that researchers can observe LacZ reporter gene expression in the heterozygous state, analyze phenotype in the homozygous state, and generate conditional KOs with promoter driven Cre lines in most cases. With this technology, nearly 20,000 protein coding genes have been knocked out in ESCs to date. However, their use is limited by the inefficient germline transmission of the targeted ESCs and the complicated processes of serial mating. More importantly, as we know more about gene features, researchers require a more finely-tuned gene modification system for gene functional analysis as well as making disease model animals. Thus, an efficient system for generating these animals on demand is necessary. 1 Research Institute for Microbial Diseases, Osaka University, Suita, Osaka 5650871 Japan. 2Graduate School of Pharmaceutical Sciences, Osaka University, Suita, Osaka 5650871 Japan. 3Research Fellow of Japan Society for the Promotion of Science (JSPS), Tokyo 1020083, Japan. 4Graduate School of Medicine, Osaka University, Suita, Osaka 5650871 Japan. 5Immunology Frontier Research Center, Osaka University, Suita, Osaka 5650871 Japan. † Present address: Biomedical Sciences, St. George’s Medical School, London SW17 0RE, UK. *These authors contributed equally to this work. Correspondence and requests for materials should be addressed to M.I. (email: ) Scientific Reports | 6:31666 | DOI: 10.1038/srep31666 1 www.nature.com/scientificreports/ Figure 1. Experimental designs of the pronuclear-injection and the ESC-transfection. (a) The sgRNA/ CAS9 expressing plasmid was injected into pronuclei of fertilized eggs. The induction of monoallelic or biallelic mutations at the one cell stage results in mutant mice carrying one or two different mutations in all cell types. On the other hand, the introduction of a mutation after the 2-cell stage results in mosaic mice in the founder generation. “X” indicates the mutated sites. (b) Mouse ESCs with GFP fluorescence were transfected with sgRNA/CAS9 expressing plasmids. The ESC clone with the desired mutation can be expanded and injected into WT (+/+) 8-cell embryos to generate chimeric mice. The ESC-derived mutant cells can be identified by fluorescence (green). (a,b) The mosaic mice in (a) may carry multiple cell types with unidentified mutations whereas chimeric mice in (b) carries WT (+/+) host cells and cells with identified mutations (−/−). The emergence of the CRISPR/Cas9 system opened a new era for mammalian genome editing8,9. The sgRNA/ CAS9 complex recognizes a target site in the genome by 20 nts upstream of the protospacer adjacent motif (PAM) to cause a blunt end double-strand break (DSB) between 3rd and 4th nucleotides upstream of the PAM. After DSB formation, non-homologous end-joining (NHEJ) causes indels that disrupt gene functions. If a reference DNA strand is co-injected, designed mutations and transgenes are efficiently incorporated by homology-dependent repair (HDR) following DSBs. The CRISPR/Cas9 system works efficiently in mammalian cells and the targeted gene-manipulated mice can be generated by simply injecting sgRNA/CAS9 complex into the pronuclei of fertilized eggs with or without the reference DNA, depending on the desired mutation10–12. Therefore, sgRNA/ CAS9-mediated genome editing in fertilized eggs is now widely used in many laboratories and animal facilities to generate gene-manipulated mice. However, some factors remain to be optimized. For targeted gene disruption, small indels are easily introduced by single sgRNA/CAS9 complexes, but removal of the entire gene by two sgRNA/CAS9 complexes becomes less efficient as the distance between the two sgRNAs increases. The efficiency of HDR-mediated gene manipulation in mouse oocytes varies among not only target genes but also among laboratories12–15. More importantly, the mice generated by sgRNA/CAS9 complexes were thought to be homozygous in the founder generation. However, later experiments demonstrated the high prevalence of mosaicism in most founder animals which obfuscate genotyping and phenotyping analyses (Fig. 1)16,17. Subsequent matings are required to generate fully homozygous mutant animals, making the experimental process similar to conventional ESC-mediated gene manipulation. In the present study, we examined complicated genome editing in mouse ESCs (Fig. 1). The indels and large deletions were efficiently obtained compared to fer (...truncated)