A simple strategy for heritable chromosomal deletions in zebrafish via the combinatorial action of targeting nucleases

Precise and effective genome-editing tools are essential for functional genomics and gene therapy. Targeting nucleases have been successfully used to edit genomes. However, whole-locus or element-specific deletions abolishing transcript expression have not previously been reported. Here, we show heritable targeting of locus-specific deletions in the zebrafish nodal-related genes squint (sqt) and cyclops (cyc). Our strategy of heritable chromosomal editing can be used for disease modeling, analyzing gene clusters, regulatory regions, and determining the functions of non-coding RNAs in genomes.


Background
Genome editing tools such as transcription activator-like effector nucleases (TALENs) and zinc finger nucleases (ZFNs) have revolutionized the fields of biotechnology, gene therapy and functional genomic studies in many organisms [1][2][3][4]. However, engineering large chromosomal deletions in vertebrates has been largely restricted to mice, where the typical strategy used is gene targeting, and subsequently, specific regions are engineered by site-specific recombination systems such as the Cre/Lox or Flp/FRT systems [5]. Although site-specific recombination has been used successfully to analyze the functions of genes involved in embryonic development, cancer and other diseases, this strategy is time, labor and resource intensive. Hence, rapid and facile methods to engineer chromosomes are of immense value.
Analysis of regulatory elements in the genome, and determining the activity and functions of gene clusters require generation of chromosomal lesions. Although large chromosomal lesions have been generated by gamma ray treatment and other methods, these lesions are often accompanied by complex rearrangements affecting several loci, which is a limitation for precise analysis of specific genomic regions or regulatory elements. In addition, the size and position of the re-arrangements cannot be predetermined by these methods [6,7]. Therefore, precise and easy techniques to generate segmental mutations at desired locations on chromosomes would be useful for analysis of gene clusters and large regulatory regions in the genome.
Recent genome-wide transcriptome analyses in cells and organisms have identified several non-coding and novel coding RNAs. However, determining their functions requires the generation of RNA-null alleles [8]. The TALEN and ZFN technologies have been used successfully in many organisms to generate small insertion and deletion mutations at target sites of specific genes [3,[9][10][11]. Large chromosomal deletions and inversions have been shown in cell lines using ZFNs, and deletions using two pairs of TALENs have been generated in silkworm, swine fibroblasts and zebrafish [12][13][14][15][16]. So far, however, heritable chromosomal deletions that specifically abolish expression of a transcript have not been reported with these nucleases in any organism. Thus, rapid and easy methods to generate whole-locus, element-specific or transcript-specific deletions would greatly facilitate functional genomic studies.
Here, we report a simple, effective and rapid strategy to generate a whole locus deletion in zebrafish, by the simultaneous use of two pairs of TALENs or TALEN pairs in conjunction with ZFN pairs, that we used successfully to precisely delete the nodal-related gene sqt and generate sqt RNA-null alleles. We also report targeted deletions in a second zebrafish nodal gene, cyclops (cyc), for which gamma ray-and chemically induced chromosomal rearrangements and point mutations were reported, but a precise locus-specific deletion was not available [6,[17][18][19][20][21][22][23]. Our strategy of heritable chromosomal editing can potentially be applied for functional genomic studies in a variety of organisms.

Results
To test if large deletions can be generated efficiently by using multiple TALENs, we first targeted the reporter gene encoding enhanced green fluorescent protein (EGFP). We designed and synthesized two TALEN pairs spaced approximately 600 bp apart to target egfp sequences [9,24] (black arrows in Figure 1a). Each TALEN pair was tested individually and in combinations, at various doses, by injecting transgenic zebrafish embryos (Tg (Ds DELGT4) sg310 ) with ubiquitous and robust EGFP expression ( Figure 1b). Injected embryos were assessed for abnormalities or lethality, and for EGFP expression. The cutting efficiency of each TALEN pair was estimated by T7 endonuclease I (T7E1) assay on ten individual injected embryos (Table 1) and calculated by sequencing pooled PCR amplicons from the embryos ( Figure S1A,D,H in Additional file 1). Higher doses of egfp TALEN pairs resulted in increased numbers of abnormal embryos and lethality ( Table 2). Loss of EGFP expression was observed at 30 hours post-fertilization (hpf) in sectors of the eyes and neural tube of egfp TALEN-injected embryos ( Figure 1c). Thus, the injected egfp TALEN pairs induce mutations in egfp and effectively disrupt EGFP expression in embryos. PCR with primers flanking the TALEN sites (Figure 1a, blue and magenta triangles) shows an approximately 250 bp fragment in all injected embryos (n = 23), compared to a 854 bp wild-type egfp fragment, indicating excision of intervening sequences in some cells of injected embryos (Figure 1d). Sequencing of PCR products from individual embryos showed large as well as small deletions, likely due to mosaicism of the injected nuclease RNA pairs, and non-homologous end joining events (Figure 1e; Figure S1B,C,E in Additional file 1). Comparison of sequences of single TALEN versus double TALEN pair injections shows lower deletion frequency with single TALEN pair injections, presumably because small deletions induced by single nuclease pairs are repaired more efficiently than the larger lesions induced by multiple TALEN pairs ( Figure S1A-E,H in Additional file 1). Moreover, the majority of mutant alleles from double TALEN injections showed complete deletions ( Figure S1B,C,E in Additional file 1). These results show that defined large deletions that disrupt target gene expression can be generated easily via the combinatorial action of multiple TALENs.
Next, to determine if large deletions in endogenous loci and element-specific deletions can be generated effectively, we designed and synthesized one TALEN pair towards sequences approximately 230 bp upstream of the predicted transcriptional start site (TSS), and a second pair targeting sequences within cyc exon 1 (cyc5TAL, chr12: 49,427,780-49,427,835; cyc3TAL, chr12: 49,428,165-49,428,221), spanning a genomic region of approximately 380 bp that encompasses the TSS (Figure 2a). Similarly, to target sqt, we generated one TALEN pair specific to the 5' sequences upstream of the TSS, and a second pair targeting sequences in the 3' UTR of sqt (sqt5TAL, chr21: 19,838,838,767;sqt3TAL,chr21: 19,840,840,929; zebrafish genome assembly Zv9). The TALEN target sites span a chromosomal region of approximately 2.16 kb, encompassing the sqt gene. We also used sqt5TAL in conjunction with a ZFN pair targeting sqt exon 1 (sqtZFN2, Figure 2a), to delete a 98 bp genomic region surrounding the TSS (sqtZFN2, chr21: 19,838,838,934).
To determine the optimal dosage, we microinjected various concentrations of sqt TALEN and ZFN pairs, or cyc TALEN pairs into one-cell zebrafish embryos individually and in combinations, and assessed the cutting efficiency, phenotypes and survival at 24 hpf (Table 3; Figures S1A-D, F-G, I-J and S2 in Additional file 1). Cyclopia and midline defects, phenotypes that are visible in cyc and sqt mutant embryos [23,25], were found at frequencies ranging from 13 to 40% for cyc and 15 to 25% for sqt, indicating bi-allelic mutations in a proportion of injected F0 embryos (Figure 2b,c, Table 3; Figure S2 in Additional file 1). The efficacy of deletion mutations was assessed by PCR and sequencing from individual 1 dpf embryos (Figure 2d-f; Figure S1A-D,F-G,I-J in Additional file 1). Alignment to wild-type sqt genomic sequences showed that each TALEN and ZFN pair by itself generated small insertions and deletions of varying lengths ( Figure S1D,I in Additional file 1), consistent with previous reports using single ZFN or TALEN pairs [9,10,26]. PCR performed on embryos injected with combinations of cyc5TAL and cyc3TAL, sqt5TAL and sqt3TAL, or sqt5TAL and sqtZFN2 showed both small and large fragments, including some of the size expected by targeted deletion of the intervening sequences (approximately 400 bp for cyc, approximately 220 bp for sqt whole locus deletion, approximately 300 bp for sqt TSS deletion), although the sqt5TAL+sqtZFN2 pair was not as efficient as the other double nuclease pair injections (Figure 2d-f; Figure S1B,C,F,G in Additional file 1). Sequencing of the PCR amplicons and alignment to wild-type cyc and sqt sequences shows that large deletions can be accompanied by insertions at both 5' and 3' targeting sites, indicative of non-homologous end joining events ( Figure 2G,H; Figure  S1B, C, F, G in Additional file 1). These results show that    To determine the germ-line transmission frequency of the deletion mutations, we raised sqt and cyc nucleaseinjected embryos to adulthood, and screened their progeny by PCR with primers spanning the targeting sites (colored triangles in Figure 2a; Table S2 in Additional file 1). For cyc, we observed deletions in 4.5 to 23% F1 progeny of 10/ 36 F0 founders. Of these, 9/10 founders yielded embryos with complete TSS deletions, 1/10 showed a smaller deletion near the cyc3TAL target site, and the same founder also transmitted a second mutation comprising a 187 bp deletion near the cyc5TAL target site, together with a 174 bp inversion and a 14 bp insertion near the 3' cyc3TAL target site (Table 4; Figure S3 in Additional file 1). In 3/10 founders, we also observed multiple mutation events ( Figure S3 in Additional file 1). We identified sqt wholelocus deletions in 3.3 to 9.5% F1 progeny of 6/56 F0 founders injected with the sqt5TAL and sqt3TAL pairs. The smaller 5' TSS deletions generated with the sqt5TAL and sqtZFN2 pairs were observed in 3.3 to 6.7% F1 embryos from 2/28 F0 fish (Table 3; Figure S4A,B in Additional file 1). However, of the two founders with the sqt TSS deletion, only one appears to have been due to targeting by both sqt5TAL and sqtZFN2 pairs, whereas the other is likely from activity of the sqt5TAL alone (Figures S4 and S5 in Additional file 1). These results suggest that cyc is targeted at higher efficiency than sqt by the nucleases. For sqt, the efficiency and the germ-line transmission frequency of the larger sqt whole locus deletions are not substantially different from that of the smaller sqt TSS deletions.
We then examined embryos obtained by mating fish heterozygous for the ZFN-induced sqt sg7 , sqt5TAL/ sqt3TAL-induced sqt sg32 whole-locus or sqt5TAL/ sqtZFN2-induced sqt sg27 TSS deletion mutations with sqt cz35 insertion mutant carriers [25], and found embryos that manifest cyclopia and deficiencies in midline structures such as the notochord (Figure 3a-j). Therefore, the sqt TALEN and ZFN-induced lesions do not complement the sqt cz35 insertion mutant phenotypes.
We then ascertained if the sqt TSS and whole locus deletions actually abolish sqt RNA expression in mutant embryos. We also determined if adjacent genomic regions and elements were affected by the sqt deletions, by examining expression of neighboring genes (eif4ebp1, htr1ab, and rnf180; Figure 3k) at appropriate stages. RT-PCR analyses to detect expression of immediate flanking loci show that their transcription is unaffected in the sqt sg27 TSS deletion mutant embryos (Figure 3l). By contrast, expression of sqt RNA is significantly reduced in embryos heterozygous for the sqt sg27 TSS deletion mutation, and is not detected in homozygous sqt sg27 embryos (Figure 3l). Similarly, sqt RNA expression is not detected in homozygous sqt sg32 whole-locus deletion mutant embryos, whereas flanking gene expression is unaffected (Figure 3m). Thus, our sqt deletions do not affect neighboring transcriptional units and these deletions are bona fide sqt RNA-null alleles.

Discussion
Our method demonstrates the ease of generating heritable whole locus deletions by the combinatorial action of multiple targeting nucleases. The ability to easily create targeted, heritable deletions in animal models such as zebrafish will greatly facilitate generation and analysis of humanized deletion mutations such as those observed in patients with hereditary neuropathies or polydactyly [27,28]. Therefore, we believe our strategy can be applied in a variety of organisms, including those (for example, the mouse) in which current methods for chromosomal engineering employ the labor-, time-and resource-intensive strategy of first generating targeted insertions by homologous recombination, and then editing via Cre/Lox or Flp/FRT systems. Recently, Zu et al. [29] described a method using TALENs to make precise modifications by homologous recombination in zebrafish. This is an immensely valuable method for directed genome engineering, and could potentially also be used to generate precise segmental deletions by engineering loxP sites at desired locations on chromosomes [29,30]. However, the   germ-line transmission efficiency of targeted homologous recombination by this method is currently approximately 1.5%. In contrast, our method to generate deletions by direct targeting of chromosomal segments using multiple targeting nucleases is efficient, and does not require introduction of Cre recombinase or breeding for additional generations (Table 5). Therefore, combinations of targeted nucleases can be used to rapidly generate chromosomal deletions at predetermined locations.
Chromosomal deletions can be used for analyzing gene clusters and regulatory regions, and for determining the functions of non-coding as well as coding RNAs in the genome. In support of this possibility, our sqt sg27 TSS deletion that is predicted to excise the TSS elements and sqt sg32 whole-locus deletion indeed result in mutant embryos that are sqt RNA-null. Furthermore, zygotic sqt sg27 and sqt sg32 deletion mutant embryos manifest phenotypes that are consistent with the previously identified sqt mutations. Thus, this strategy can be used effectively to investigate the roles of all 'functional' RNAs in the genome.
The various targeting nucleases have different constraints pertaining to target sites. For instance, TALENs prefer a 5' T nucleotide, whereas CRISPR/Cas9 requires a GG dinucleotide for targeting. The spacer requirements for the various nucleases are also different, and targetable sites for the different nucleases likely occur at different frequencies in genomes [31,32]. Therefore, using the combinatorial action of various nucleases can facilitate generation of defined deletions at desired locations with higher efficacy. Moreover, some TALEN and ZFN sites (for instance, our sqt ZFN target sites) are just not targeted efficiently for reasons that are still unclear. Hence, the ability to use multiple targeting nucleases in various combinations offers additional flexibility and alternative approaches to engineer chromosomes than is possible with individual nuclease pairs. The efficiency and precision of the deletion events can be improved further by using nuclease variants such as the 'GoldyTALEN' system [30].
Our simple, facile and efficient strategy is largely PCRbased, and, therefore, can be used with modest resources to generate deletion mutants for investigating functional elements in the genome. Finally, this approach of generating large, defined heritable deletions by simultaneously targeting two discrete regions on a single chromosome can potentially also be deployed with RNA-guide mediated or other emerging DNA cleavage methods [32][33][34] to enhance the toolkit for heritable chromosomal engineering in a variety of organisms.

Conclusions
Targeted and heritable chromosomal deletions can be rapidly generated in a whole organism by using the combinatorial action of targeting nucleases. Multiple nuclease pairs are apparently more effective than single nuclease pairs in generating targeted deletions. Whole-locus as well as TSS element-specific deletions were generated efficiently by this method, and stably transmitted through the germ-line. The deletion mutations result in transcript-null alleles that manifest embryonic mutant  x Figure 3 Heritable deletions in the sqt locus that result in RNA-null alleles. (a) PCR on single wild-type or sqt deletion mutant embryos (using primers indicated in Figure 2a) shows a 220 bp fragment in a sqt sg32 locus-deletion embryo, and a 380 bp fragment in TSS deleted sqt sg27 mutant embryo. Sometimes a larger approximately 500 bp fragment is observed in sqt sg27 /+ heterozygous embryos, but the sequence is identical to the 478 bp product. (b) Percentage of embryos with sqt mutant phenotypes in sqt cz35/+ , sqt sg27/+ , sqt sg32/+ and sqt sg7/+ in-crosses and mating of sqt cz35/+ with sqt sg27/+ , sqt sg32/+ and sqt sg7/+ . The cz35 allele is an approximately 1.9 kb insertion in sqt exon 1; the sg27 allele is a 98 bp deletion of sqt TSS sequences; sg32 allele is a whole locus deletion of sqt; the sg7 ZFN allele harbors a GGCC insertion in sqt exon 2. (cj) DIC images of 24 h wild-type (c), sqt cz35/cz35 (d), sqt sg27/cz35 (e), sqt sg32/cz35 (f), sqt sg7/cz35 (g), sqt sg27/sg27 (h), sqt sg32/sg32 (i), and sqt sg7/sg7 (j) embryos; scale bar in (c), 100 μm. (k) UCSC genome browser view of the sqt locus and neighboring genomic region. (l,m) RT-PCR with primers to detect expression of sqt RNA and transcripts of neighboring genes, eif4ebp1, rnf180, and htr1ab, shows lack of sqt RNA expression in sqt sg27/ sg27 (l) and sqt sg32/sg32 (m) embryos whereas all neighboring gene transcripts are expressed at wild-type levels. Actin (act) expression was used as control. In contrast, both un-spliced and spliced sqt RNA is detected in wild-type and heterozygous embryos. phenotypes, demonstrating functional consequences of the chromosomal lesions. This simple, facile and efficient strategy can be used with modest resources. Thus, our strategy can be used to generate disease models, and for analysis of gene clusters, regulatory regions and functional RNAs in the genomes of a variety of organisms.

Generation of plasmids encoding TALENs and ZFNs
The egfp, sqt and cyc TALENs target sites were designed using an online tool [35]. To check for unique targeting sites, BLAST and UCSC BLAT search was performed with the zebrafish genome assembly (Zv9) using the target site sequences. The TAL effector repeats were constructed from four TAL effector single unit vectors (pA, pT, pG NN and pC) using the 'unit assembly' method [9]. Plasmids encoding sqtZFN1 and sqtZFN2 nuclease pairs were obtained from ToolGen, Inc. (Seoul, South Korea). The TALEN and ZFN target sites for egfp, cyc and sqt are shown in Figures 1a and 2a and sequences are listed in Table S1 in Additional file 1.

TALEN and ZFN capped mRNA synthesis
Using the Ambion ® SP6 mMESSAGE mMACHINE kit (Life Technologies, Carlsbad, California, United States of America), capped TALEN mRNAs were transcribed in vitro from 1.0 µg of the respective NotI linearized TALEN expression vectors. To synthesize capped ZFN mRNAs, sqt-specific ZFN plasmids were linearized with XhoI and transcribed using T7 RNA polymerase (Promega, Fitchburg, Wisconsin, United States of America). RNA was purified by phenol-chloroform precipitation and dissolved in RNase-free water.

Microinjection of capped TALEN and ZFN mRNA into zebrafish embryos
All experiments using animals were performed in accordance with institutional animal care and use guidelines. For sqt and cyc experiments, embryos from wild type (AB) fish were used for injections. For egfp TALEN experiments, embryos from Tg (Ds DELGT4) sg310 homozygous males mated with wild-type AB females were used. Tg (Ds DELGT4) sg310 transgenic fish harbor a Ds transposon-mediated enhancer trap insertion on chromosome 21. Various dosages and combinations of nuclease RNAs were tested to determine the toxicity, and the maximum dose that yielded less than 50% lethality was used (Table 2). For testing single TALEN or ZFN pairs, either 12.5, 25 or 50 pg of each mRNA was injected into one-cell stage zebrafish embryos. Higher lethality rates and abnormal embryos were observed with the cyc TALEN pairs and, therefore, cyc5TAL and cyc3TAL mRNAs were introduced at 6.25, 12.5 or 25 pg doses per embryo. For double TALEN pair or TALEN +ZFN experiments, a cocktail of either 12.5 or 25 pg of each mRNA was injected into one-cell stage zebrafish embryos. Injected embryos were examined at prim-5 stage under a Leica MZ12.5 stereomicroscope. PCR products from individual embryos injected with each single nuclease pair were tested by the T7E1 assay and sequencing to assess the efficacy of each nuclease pair. Ten embryos that were morphologically normal were selected and processed for PCR and sequencing. The remaining embryos were raised to adulthood to determine the germ-line transmission rates.

PCR and sequence analyses
To detect deletions in founder (F 0 ) embryos, at least 10 TALEN-and ZFN-injected embryos were individually lysed at 24 hpf in 20.0 µl of DNA extraction buffer (10 mM Tris pH 8.2, 10 mM EDTA, 200 mM NaCl, 0.5 % SDS, 100 µg/ml proteinase K) for 5 h at 55°C, followed by heat inactivation of proteinase K at 65°C for 10 minutes. Genomic DNA was diluted five-fold using 1× TE Buffer (pH 8.0), and 2 µl aliquots were used in 20 µl PCR reactions. For single nuclease pair experiments, fragments containing 100 to 150 bp upstream and downstream of the expected target sites were amplified with Go Taq polymerase (Promega). For double TALEN or TALEN+ZFN experiments, primers annealing to regions 100 to 150 bp  Table S2 in Additional file 1). Five microliter aliquots of products from ten single embryo PCRs were pooled, gel purified to remove primer dimers and cloned into either Promega pGEM ® -T easy TA cloning vector or Fermentas pJET1.2 blunt end cloning vector, and transformed using XL1-blue heat-shock competent bacterial cells. At least 48 bacterial colonies were picked for screening by PCR. PCR products were diluted three-fold, and 1 μl was used directly for sequencing using the same primer pairs. Sequences were analyzed by comparison to the Zv9 Zebrafish Genome Assembly.

T7E1 assay to detect indels induced by single nuclease pairs
Five microliter aliquots of single embryo PCR products were diluted to 20 μl in 1× NEB Buffer 2, denatured at 95°C for 5 minutes, slowly cooled to room temperature to allow annealing and formation of hetero-duplexes. The individual preps were then treated with 5 units of T7E1 (New England Biolabs) for 30 minutes at 37°C. Digested products were separated on a 3.5% agarose/ 1×TBE gel and band intensity analyzed using ImageJ (NIH) to calculate mutation frequencies [36].

Genotyping of F1s
To assess the germ-line transmission rates, injected F 0 fish were raised to adulthood, and mated either with siblings or wild-type fish to obtain F 1 progeny. For genotyping sqt nuclease-or cyc TAL-injected embryos, PCR was performed using primers listed in Table S2 in Additional file 1, and Taq polymerase (Promega). PCR amplicons were electrophoresed on a 2% agarose gel. To screen for germ-line transmission events at the endogenous sqt locus, we analyzed progeny from pairwise mating of founders. Single embryos from six founder fish (three pairs) were screened per 96-well plate. At least 30 embryos (24 hpf) from each mating were collected, lysed and analyzed by PCR using the same primer pairs as used for the transient assays. This number allowed efficient detection of germ-line transmission events (whose frequency ranged from 3 to 10%), and recovery of the mutation. Bands of aberrant sizes were either sequenced directly or after cloning into the pGEM ® -T easy vector system. F 1 progeny of positive F 0 s were raised to adulthood, and heterozygous carriers for the deletions were identified by fin-clipping and routine genotyping PCR analysis, using primers listed in Table S2 in Additional file 1. The sqt sg7 ZFN1-induced allele harbors a 4 bp insertion in exon2 (chr21: 19839892-19839896; Figure S5 in Additional file 1). The sg7 mutation is predicted to result in a frame-shift after amino acid 143 in Sqt protein and premature termination after amino acid residue 146. Homozygous sqt sg7 embryos express sqt RNA [37]. The sqt sg27 mutants harbor an indel (chr21: 19838727-1983870; Figure S5 in Additional file 1) and lack the transcriptional start sequences, and the lesion in sqt sg32 is a whole locus deletion of 2.1 kb on chromosome 21 ( Figure S5 in Additional file 1). For analyzing germ-line transmission rates of cyc deletions, we collected progeny from pairwise mating of founders in pools of five embryos since the somatic mutation frequency for the cyc TALENs was higher than that for sqt. At least ten pools from each successful mating were collected and used in PCRs to ensure that founders with mutant clone sizes larger than 2% were identified. Subsequently, founders that yielded mutations were mated with wild-type (AB) fish. At least 22 single embryos from each mating were collected for PCR and sequencing to confirm and determine the germ-line transmission rate. (For a list of primers, see Table S2 in Additional file 1.)

Semi-quantitative RT-PCR
Using TRIzol reagent (Invitrogen, Carlsbad, California, United States of America), both genomic DNA and total RNA were extracted from single 30% epiboly stage and 2 dpf (for htr1ab expression) embryos obtained from heterozygous sqt sg27/+ and sqt sg32/+ crosses. For genotyping, 50 ng of genomic DNA was used as template in 20 µl PCR reactions. For first-strand cDNA synthesis, 250 ng of total RNA was used in a pdN6-primed reaction using SuperScript ® II Reverse Transcriptase (Life Technologies). First-strand cDNA (1 µl) was used in 20 µl PCR reactions to detect expression of sqt, ring finger protein (rnf180), 5-hydroxytryptamine (serotonin) receptor 1A b (htr1ab), eukaryotic translation initiation factor 4E binding protein 1 (eif4ebp1) and control actin (act), using the primers listed in Table S3 in Additional file 1.

Additional material
Additional file 1: Figures S1 to S5 and Tables S1 to S3. Figure S1:  Figure S2: representative phenotypes observed at 24 h in embryos injected with sqt or cyc TALENs. At high doses, the proportion of abnormal embryos increases. Figure S3: (A) Alignment of cyc sequences showing TSS deletions in embryos from F 0 founders, compared to wild-type cyc. Insertions are indicated in red, and gaps are shown by dashed lines. Letter suffixes (for example, 1A and 1B) represent different alleles from the same founder. Eight founders injected with cyc TALENs transmitted complete deletion of the intervening sequences, some of which also show insertion events. Founder F0-4, shows a larger deletion that extends beyond the 3' end of the cyc3TAL target site. Genomic coordinates on chromosome 12 for wild-type cyc are indicated for the regions shown. (B) Alignment of cyc sequence of F0-10 to wildtype cyc shows a deletion (dashed lines), accompanied by an inversion (yellow highlight) and insertion (red font). Figure S4: (A) Alignment of sqt sequences showing the whole-locus and TSS deletions in embryos from F 0 founders, compared to wild-type sqt. Insertions are highlighted in red font, and gaps are indicated with dashed lines. All six founders for sqt5TAL/sqt3TAL showed embryos with the 2.1 kb whole locus deletion.
(B) For sqt5TAL/sqtZFN2, founder 1 (F0-1) showed embryos with the intervening sequences excised, whereas founder 2 (F0-2) sequences indicate that only sqt5TAL was active. Genomic coordinates on chromosome 21 for wild-type sqt are indicated for the regions shown. TALEN and ZFN target sites are indicated. Figure S5: alignment of sqt sequences showing the sqt sg27 TSS deletion (A), sqt sg32 whole-locus deletion (B), and sqt sg7 ZFN (C) mutations in comparison to wild-type sqt. Insertions are highlighted in red font, and gaps are indicated with dashed lines. TALEN and ZFN target sites are indicated. Table S1: list of target sites for TALENs and ZFNs. Table S2: list of primers for genotyping, sequencing, and T7E1 assays.