Efficient Gene Knock-out and Knock-in with Transgenic Cas9 in Drosophila

Bacterial Cas9 nuclease induces site-specific DNA breaks using small gRNA as guides. Cas9 has been successfully introduced into Drosophila for genome editing. Here, we improve the versatility of this method by developing a transgenic system that expresses Cas9 in the Drosophila germline. Using this system, we induced inheritable knock-out mutations by injecting only the gRNA into embryos, achieved highly efficient mutagenesis by expressing gRNA from the promoter of a novel non-coding RNA gene, and recovered homologous recombination-based knock-in of a fluorescent marker at a rate of 4.5% by co-injecting gRNA with a circular DNA donor.

knock-out knock-in Cas9 gRNA Drosophila The RNA-guided CRISPR/Cas9 system has been successfully used for the purpose of genome editing in many organisms, including Drosophila (Golic 2013). Cas9 nuclease is guided to its target DNA by a small guide RNA (gRNA), which shares a homologous region of approximately 20 nucleotides (nt) with the cleavage site. By direct injection of synthesized Cas9 mRNA and customized gRNA into wild-type Drosophila embryos, we previously achieved a mutational rate of up to 100% in the germline (Yu et al. 2013). We considered that at least two aspects of our previous technique could be improved upon. First, the generation and handling of in vitro synthesized RNA, particularly the relatively large Cas9 mRNA, could be challenging to inexperienced researchers, which could result in inconsistencies in mutational efficiency. Second, the utility of Cas9 in generating knock-in mutations has not been adequately demonstrated in Drosophila. Here, we report efficient knock-out and knock-in manipulations of the Drosophila genome with a Cas9 transgenic system. We note that several groups have independently developed tools to conduct mutagenesis using transgenic lines expressing Cas9 in the germline (Sebo et al. 2014;Ren et al. 2013;Kondo and Ueda 2013) and to generate knock-in mutations (Baena-Lopes et al. 2013;Gratz et al. 2014). In addition, an efficient knock-in system using TALEN has been developed (Katsuyama et al. 2013).

MATERIALS AND METHODS
Fly stocks Several different wild-type stocks such as w 1118 , y w and Oregon-R (red eye) were used in this study. All balancers were obtained from the Bloomington Stock Center (Indiana). Flies were cultured at 25°.
Donor plasmid pHisc-RA-RFP construction: To generate the pHisc-RA-RFP donor plasmid, a synthetic DNA fragment containing the attP, FRT, and 3·P3RFP sequences was cloned into pEASY-T to generate vector pEASY-RFP (available upon request). Subsequently, 1.5-kb and 1.2-kb sequences, respectively, corresponding to the 59 and 39 ends of the histone cluster right-arm (Hisc-RA) were inserted flanking the RFP cassette to obtain the final construct pHisc-RA-RFP.
gRNA injections DNA templates were PCR-amplified from gRNA scaffold vectors for in vitro gRNA transcription (Yu et al. 2013). Transcription was performed with the RiboMAX Large Scale RNA Production Systems-T7 Kit (Promega, USA) according to the manufacturer's protocol. Purified gRNA (0.2 mg/ml) was injected into vasa-Cas9 fly embryos either directly or after mixing with purified donor plasmid (0.8 mg/ml) according to standard procedures. All primers are listed in Table S1.

Fly transformation and genetics
To obtain transgenic vasa-Cas9 flies, the PiggyBac-vasa-cas9 vector was mixed with helper vector PiggyBac-transposase (Thibault et al. 2004) and injected into w 1118 fly embryos. Flies carrying vasa-Cas9 were identified by GFP expression in the eye when viewed under a fluorescence stereomicroscope. Mutation position was mapped to chromosome 2 after crossing with balancer flies, and homozygous vasa-Cas9 flies were established by subsequent crossing.
To obtain transgenic gRNA flies, the five pUAST-U6B/CR7T-gRNA vectors were separately injected into w 1118 fly embryos, and at least three independent lines per transgenic gRNA were identified through observation of the RFP eye marker. The vectors targeted the following loci: pUAST-U6B-k81-gRNA, the ms(3)k81 locus; pUAST-U6B-y1-gRNA and pUAST-CR7T-y2-gRNA, the yellow locus; and pUAST-CR7T-w1-gRNA and pUAST-CR7T-w2-gRNA, the white locus. After transgenic gRNA flies were crossed with the vasa-Cas9 flies, F 0 flies expressing both fluorescent markers were individually collected and were crossed to wildtype flies to generate F 1 germline transformants. F 0 flies and F 1 progeny were phenotypically and genotypically assessed. For the vasa-Cas9 fly embryos directly injected with gRNA, F 0 and F 1 were tested; however, for embryos receiving coinjections of gRNA and donor plasmid, only F 1 flies with visible RFP eye markers were used for PCR analysis and DNA sequencing.
Screening and analysis of mutations formed via nonhomologous end joining F 0 (from injected embryos) and F 1 (progeny of F 0 ) ms(3)k81-disrupted flies that resulted from direct injection of synthetic gRNA into vasa-Cas9 fly embryos were recovered and analyzed as described previously (Yu et al. 2013). All five transgenic pUAST-gRNA lines were crossed with vasa-Cas9 flies to induce F 0 mutations at the three loci of interest (ms(3)k81, white, and yellow). Each F 0 fly containing transgenic vasa-Cas9 and gRNA was then crossed to wild-type flies to recover F 1 progeny with mutations resulting from germline transmission. For the ms(3)k81 locus, mutations of each F 0 fly containing transgenic vasa-Cas9/ pUAST-U6B-k81-gRNA and its F 1 offspring were recovered and analyzed as described above. In case of the yellow locus, pUAST-U6B-y1-gRNA and pUAST-CR7T-y2-gRNA were used. The target yellow genomic region was amplified by PCR from a single transgenic vasa-Cas9/pUAST-gRNA F 0 fly using appropriate primers (Table S3). The corresponding PCR products were digested with T7 endonuclease I (New England Biolabs, USA) and analyzed by gel electrophoresis (Kondo and Ueda 2013). The mosaic yellow phenotype evaluation in F 0 transgenic vasa-Cas9/pUAST-gRNA and indel scoring in F 1 progeny were performed according to previous methods (Liu et al. 2012). For the white locus, two pUAST-gRNA lines (pUAST-CR7T-w1-gRNA and pUAST-CR7T-w2-gRNA) were used to induce the germline transmission mutations. First, a vasa-Cas9 fly expressing the white gene on the X-chromosome was identified from the progeny of Oregon-R flies and vasa-Cas9 flies by screening for red eyes with GFP expression. The female red-eye vasa-Cas9 flies were crossed to pUAST-CR7T-gRNA males to generate male F 0 flies, which were identified by screening for the GFP and RFP eye markers. The male F 0 flies were then crossed to wild-type flies to obtain F 1 progeny with germline-transmitted mutations. The mosaic eye phenotype of F 0 transgenic vasa-Cas9/pUAST-gRNA was investigated and indel scoring in F 1 progeny was performed and analyzed as above. Primers used for PCR and sequencing are listed in Table S3.

Recovery and analysis of germline mutants via homologous recombination
Hisc-RA-gRNA and the donor pHisc-RA-RFP plasmid were coinjected into vasa-Cas9 fly embryos, and germline mutations containing the attP-FRT-RFP fragment were recovered by collecting F 1 progeny with visible RFP eye expression. Genomic DNA was prepared from individual heterozygous and some homozygous candidates. The target genome region was amplified using three pairs of primers, including one pair that flanked the locus (His-F01 and His-R02) ( Figure 1A). The corresponding products were analyzed by gel electrophoresis and sequenced ( Figure 1B). Homozygous knock-in stocks were established after crossing to balancer flies. Primers used for PCR and sequencing are listed in Table S3.

RESULTS AND DISCUSSION
Cas9 knock-out by gRNA injection into Cas9-producing embryos To eliminate the need to synthesize Cas9 mRNA for injection, we set out to achieve Cas9 expression in the germline by driving the cas9 gene with the vasa promoter (Lasko and Ashburner 1988). We recovered a single piggyBac-mediated insertion of this transgene in chromosome 2. Transgenic Cas9 flies are homozygous viable and fertile. To investigate whether these transgenic flies can be used to efficiently produce site-specific mutations, we set out to knock-out ms (3)k81 using a previously designed gRNA (Yu et al. 2013). Mutations in ms(3)k81 cause male sterility (Fuyama 1986). When we injected Cas9 mRNA and ms(3)k81 gRNA simultaneously, all surviving males were sterile and all harbored somatic ms(3)k81 mutations (Yu et al. 2013). We hypothesized that in these males, both copies of the ms(3) k81 gene might have been mutated in male germ cells. Many surviving females were fertile, and all of their progeny were mutants for ms (3) k81, consistent with our proposition that both copies of ms(3)k81 had been mutated as a result of the repair of a Cas9-induced DNA break.
Here, we injected only the ms(3)k81 gRNA into vasa-cas9 embryos. We detected somatic ms(3)k81 mutations in 70% of the surviving males (12/17). All 12 males were fertile and produced a ms(3)k81 mutational rate of 25% in the germline (Table 1). Surviving females produced a similar mutational rate in their germline (Table 1). Therefore, our transgenic system can efficiently induce mutations in the germline ( Figure S1A).

Knock-out by transgenic expression of both gRNA and Cas9
To completely eliminate the need to inject RNA, we set out to establish a transgenic gRNA system to combine with germline Cas9 expression. The U6B promoter has commonly been used for gRNA expression (Kondo and Ueda 2013;Ren et al. 2013;Sebo et al. 2014). Here, we were interested in testing the efficacy of using the promoter of another non-coding RNA gene to drive gRNA expression. We chose the non-coding CR34335 gene, which is ubiquitously expressed at high levels based on genome-wide RNA-seq data (Graveley et al. 2010), and established a gRNA expression system that is driven by the CR34335 promoter, which we named CR7T. gRNA for various targets were put under the control of the U6B or CR7T promoters and introduced into the fly genome by P-element-mediated transformation.
As a test of the functionality of our U6B expression system, we generated flies with vasa-Cas9 and U6B-k81, which expresses a gRNA designed for the ms(3)k81 used previously. All males (n = 26) were sterile, suggesting that both ms(3)k81 copies had been mutated in male germ cells. Twenty-two females were fertile, from which we obtained a ms(3) k81 mutational rate of 14.3% (Table 1 and Figure S1B). Therefore, we observed a lower efficiency using the U6B promoter in the female vs. the male germline. We tested the same system on the yellow locus and observed a similar effect (y1 gRNA in Table 1). In contrast, for the y2 gRNA against yellow, we achieved equally high (more than 90%) frequencies from both germlines using the CR7T promoter (Table 1 and Figure  S2). We further confirmed the utility of CR7T in driving gRNA expression by using it in the generation of mutations at two target locations in white, both with reasonably high efficiencies (Table 1 and Figure S3).
In summary, we successfully generated knock-out mutations to three genes by transgenically expressing both Cas9 and gRNA. The germline mutational rate ranged from 14% to 95%. Additionally, we developed a new gRNA expression system using the CR7T promoter.

Knock-in by Cas9-induced homologous recombination
The ability to introduce predefined genetic modifications is highly desirable in genome editing. The knock-in approach utilizes the ability of the cell to incorporate exogenously supplied DNA into the host genome by homologous recombination (HR). Previously, a 50-nt modification (an attP attachment site for the phiC31 integrase) was successfully introduced into the yellow locus by Cas9-induced HR (Gratz et al. 2013). This modification was accomplished by co-injection of Cas9 mRNA, gRNAs against yellow, and single-stranded oligos as donor DNA for recombination [single-strand oligodeoxynucleotide (ssODN)]. The ssODN had two 60 nt homology arms flanking attP.
For the Cas9-induced knock-in method to become routine practice in Drosophila, at least two issues need to be further examined. The first issue is whether this approach can be applied to a general locus where an easily screenable phenotype (such as yellow body color in yellow-mutant flies) is not available. The second issue is whether double-stranded donor DNA can be used when the size of the desired modification precludes the use of oligos. As an attempt to address the first question, we used ssODN to introduce an FRT site along with an attP to the left side (centromere-distal) of the histone locus on chromosome 2. The oligo is 135 nt in total length with a 24-nt homology arm on the left side and a 27-nt homology arm on the right side. Using PCR as a screening method, we successfully recovered precise knock-in events in 10 of the 112 F 1 progeny screened, suggesting that the ssODN approach can, in principle, be applied without the necessity of a screenable phenotype in the target locus. PCR followed by sequencing confirmed the precise integration of the FRT and attP elements ( Figure S5). However, a similar approach applied to the right (centromere-proximal) side of the histone locus failed, with the oligo being 87 nt in total length, with a 26-nt homology arm on left side and a 27-nt homology arm on the right side. The reason for this is unclear.
As an attempt to address the second question, we modified the histone cluster locus using a plasmid as the HR donor. The overall scheme of this targeting experiment is shown in Figure 1A. Our goal was to knock-in an element with attP and FRT along with a 3·P3RFP marker gene to the right side of the histone locus. In the donor plasmid, these three DNA elements were flanked by a 1.5-kb homology arm to the left and a 1.2-kb arm to the right ( Figure 1A). The inclusion of these elements into the donor also disrupted a Cas9 cleavage site so that Cas9 will only cut at the chromosome but not the donor. The in vitro synthesized gRNA molecules and donor DNA were injected into vasa-Cas9 embryos. We set-up 89 fertile crosses from flies that survived the injection process. Using the RFP marker, we recovered at least four independent events with a total of 49 F 1 progeny showing RFP expression. PCR amplification followed by sequencing confirmed that in all of these 49 F 1 individuals, a single copy of the 2-kb RFP cassette was precisely targeted to the desired location ( Figure 1B). Therefore, we have demonstrated the targeted insertion of a large piece of heterologous DNA (2 kb in total) aided by the use of n a visible marker in Drosophila. Using this RFP-marked insertion and the previously generated attP-FRT insertion to the left of the histone locus, we succeeded in deleting the entire 125-kb histone locus using FLP-mediated recombination (data not shown).