Targeted Gene Deletion of miRNAs in Mice by TALEN System

Mice are among the most valuable model animal species with an enormous amount of heritage in genetic modification studies. However, targeting genes in mice is sometimes difficult, especially for small genes, such as microRNAs (miRNAs) and targeting genes in repeat sequences. Here we optimized the application of TALEN system for mice and successfully obtained gene targeting technique in mice for intergenic region and series of microRNAs. Microinjection of synthesized RNA of TALEN targeting each gene in one cell stage of embryo was carried out and injected oocytes were transferred into pseudopregnant ICR female mice, producing a high success rate of the targeted deletion of miRNA genes. In our condition, TALEN RNA without poly(A) tail worked better than that of with poly(A) tail. This mutated allele in miRNA was transmitted to the next generation, suggesting the successful germ line transmission of this targeting method. Consistent with our notion of miRNAs maturation mechanism, in homozygous mutant mice of miR-10a, the non- mutated strand of miRNAs expression was completely diminished. This method will lead us to expand and accelerate our genetic research using mice in a high throughput way.


Introduction
To explore the region-specific genome functions in vivo, generation of gene-targeting animals is a powerful strategy. In this regard, enormous amount of endeavors was used for generating knockout mice, since embryonic stem cells (ES cells) for homologous recombination or nuclear transfer technologies are established firstly in mice species. It should be also noted that compared with other model animals, mice have several strong advantages, such as in the size of body and the fertilization cycle, genetically close relationship to human genome, exist of inbred strains. In fact, as post genome study, worldwide consortiums are trying to generate almost all the genes targeting ES lines [1]. However, recent drastic progress of molecular biology, including non-coding RNA discovery and spatial-temporal enhancer research, ask us for creating even further amount of and complicated genome deletion mice, including large scale generation of microRNA (miRNA) targeting mice [2], which are sometimes difficult to generate with conventional methods. Recently developed Zinc-finger nuclease (ZFN) and transcription activator-like effector nuclease (TALEN) are powerful tools to generate knockout animals in several model species including zebrafish, Xenopus and rats [3][4][5][6][7][8][9]. One of the merits of this system is that they are free from using ES cells. The genome of fertilized zygotes was directly modified by injection of artificial endonuclease, such as ZFNs and TALENs. ZFNs and TALENs are fusion proteins containing DNA-binding domain and FokI endonuclease at the C-termini. The recognition sequences of DNAbinding domain can be designed with some restrictions such as ZFNs require (GNN) n sequences and TALENs T at 39 end of the recognition sequences. When two ZFNs or two TALENs bind to their target sequences in a genome in an appropriate orientation and distance, FokI part of the proteins are dimerized and cause double-strand break, resulting in stimulation of the breakage repair mechanisms in cell nuclei [10][11][12]. The breakage is repaired by non-homologous end-joining (NHEJ), which frequently results in small insertions or deletions [13,14].
Although ZFNs and TALENs are basically compatible for application, TALENs have some advantages over ZNFs; TALENs are more flexible for designing and assembling recognition sequences [15], cost and time are efficient than ZNFs and construction for TALENs is freely available from Addgene (http://www.addgene.org). Actually, TALENs have been applied for making knockout mice for several genes [16][17][18]. Also some genes for miRNAs have been targeted in human cell line using TALENs [19].
Taken together, application of TALENs to mice genetic studies are urged and in this regard, here we report optimized highefficient genome targeting protocol in mice by TALEN system.

Genotyping
Genomic DNA was extracted from york sac of E13.5 embryos or tail tips of pups. PCR was performed as above. PCR products obtained were treated with ExoSAP-IT and used as templates for sequencing and T7 endonuclease I (T7EI) assay. Sequencing primers used were coR for the gene desert on mouse chromosome 11, miR-146a-R for mmu-mir-146a, 10aF for mmu-mir-10a and 10bF for mmu-mir-10b. To confirm mutant alleles in nucleotide levels, independent PCR and direct sequencing was carried out and the PCR product was cloned and sequenced as above. For T7EI assay, denatured and re-annealed PCR products were digested with T7EI and then analyzed by agarose gel electrophoresis. Band intensities were quantified by MultiGauge version 3.2 software (Fujifilm Corporation).

Quantitative real-time RT-PCR
Total RNA was purified from kidneys of 4-week-old mice using ISOGEN (Nippongene) according to the manufactures instruction. Reverse transcription (RT) and PCR were carried out using TaqMan MicroRNA RT Kit (Life Technologies), TaqMan Universal PCR Master Mix II, w/o UNG (Life Technologies) and TaqMan micoRNA assay (Life Technologies). As an internal control TaqMan against U6 (RNU6B) was used. nnCt method was adopted for calculation of relative gene expression levels. Expression levels were obtained from average of duplicate trial.

Evaluation of TALENs activities in cultured cell line
The aim of this study is generation of knockout mouse deficient for a gene desert and miRNAs using TALEN through microinjection of fertilized oocytes. As a first step, we examined if TALEN technology is useful for introduction of mutation on a gene desert and miRNA loci on mouse genome. To test this, we ordered construction of TALENs for specific gene desert of mouse genome, which is located on chromosome 11, mmu-mir-146a, mmu-mir-10a and mmu-mir-10b from Cellectis bioresearch. The activity of the TALENs were evaluated by the company using yeast single strand annealing (SSA) assay, which is a method for measuring genome cleavage activity ranged 0-1, where high number indicates high activity and value of 0.45 or higher means good cutter. The value of SSA assay for TALEN pairs we obtained were 0.86 for the gene desert, 0.64 for mmu-mir-146a, 0.92 for mmu-mir-10a and 0.75 for mmu-mir-10b, showing that all TALEN have cleavage activity in yeast.
We examined the genome editing activities of TALEN pairs for the gene desert, mmu-mir-146a, mmu-mir-10a and mmu-mir-10b in NIH3T3 cultured cell line. Since the TALEN plasmids does not contain drug resistant gene for mammalian cell line, pIRES2-EGFP, which contains neomycine resistant gene, was co-transfected with TALEN plasmids to NIH3T3 to increase TALEN introduced cell populations by G418 selection. To evaluate efficiencies of genome editing, nucleotide sequences of TALEN cleavage sites were identified by PCR amplification followed by cloning of PCR products from TALEN transfected cells. Sequencing analysis showed that the genome editing activities were confirmed at rate of 2.5% (2 mutated sequences out of 79 sequences) for the gene desert on chromosome 11, 8.2% (4 mutated sequences out of 48 sequences) for mmu-mir-146a and 1.2% (1 mutated sequence out of 85 sequences) for mmu-mir-10b while no mutated sequence was identified from 207 sequences for mmu-mir-10a (Fig. 1), indicating that the results are not always parallel between this method and yeast SSA assay in that TALEN pair for mmu-mir-146a shows the highest activity in former assay whereas lowest in latter among four TALEN pairs. This result implying that at least three TALEN pairs, for the gene desert, mmu-mir-146a and mmu-mir-10b, could be useful for making knockout mouse.

Optimization of TALEN mRNA injection method
To make knockout mouse using TALEN technology, we optimized TALEN mRNA microinjection method in mouse fertilized oocytes using TALEN pairs for the gene desert with  concentration of RNA for injection into cytoplasm of pronuclearstage oocytes. We injected TALEN mRNAs without poly (A) tail at 10 ng/ml to 190 oocytes. One hundred sixty nine embryos entered to 2-cell stage the next day and transferred into oviduct of pseudopregnant female (Table 1). Embryos were collected at E13.5 and genotypes were determined using DNAs prepared from tails with PCR direct sequencing. As a result, no mutant was obtained. It may be possible that RNA concentration was not enough for inducing mutation using TALEN RNAs, therefore 50 and 100 ng/ml of TALEN RNAs without poly (A) tail were used for microinjection. 38 and 40 oocytes were injected, 22 and 23 embryos entered to 2-cell stage and 17 and 20 of E13.5 embryos were obtained using 50 and 100 ng/ml of TALEN RNAs, respectively (Table 1). Genotyping showed that no mutant was produced. We next tried to use several concentrations of TALEN mRNAs with poly (A) tail. TALEN RNAs at the concentration of 10, 50 and 100 ng/ml were injected to 52, 58 and 58 oocytes, and 41, 20 and 40 embryos developed to 2-cell stage and 26, 16 and 30 of E13.5 embryos were obtained, respectively (Table 1). Genotyping showed that no mutant was produced. It may be that TALEN RNAs at the concentration of 100 ng/ml is not still sufficient for producing mutant mice or the number of microinjection is too small to obtain a mutant because the efficiency of TALEN RNAs at that concentration is too low for inducing mutation. Finally we injected TALEN mRNAs with or without poly (A) tail at the concentration of 100 ng/ml. Microinjection was carried out using 155 and 196 oocytes, 82 and 138 embryos developed 2-cell stage and 56 and 73 embryos were collected at E13.5 (Table 1). Genotyping showed that no mutant was obtained from TALEN mRNAs with poly (A) tail injected embryos whereas 5 embryos with different mutations (out of 56 embryos; success rate is 8.9%) were identified from TALEN mRNAs without poly (A) tail injection (Fig. 2). Mutations were further re-confirmed by sequencing after independent round of PCR. From this result TALEN mRNAs without poly (A) tail seems to work better for inducing mutation than TALEN mRNAs with poly (A) tail. We used TALEN mRNAs without poly (A) tail at concentration of the 100 ng/ml was used for following experiment.

Production of mmu-mir-146a knockout mouse
Utilizing above condition, we set out to make a knockout mouse deficient in series of miRNAs; mmu-mir-146a, mmu-mir-10a, and mmu-mir -10b. Genotyping of pups derived from the injected oocytes by PCR direct sequencing showed 8 out of 27 individuals (29.6%) contain wild type and mutated alleles in miR-146a (no biallelically modified mice were obtained so far) (Fig. 3B). One type of mutated allele was identified in four pups respectively (Mutant1-4, Fig. 3B), whereas two types of mutated alleles were identified in three pups respectively (Mutant5-7, Fig. 3B) and four were identified in one pup (Mutant8, Fig. 3B), implying that the TALEN protein worked at one-cell stage embryos in the former four cases and at two-cell stage or after in the latter four cases.

Production of mmu-mir-10a and -10b knockout mice
We next tried to generate mmu-mir-10a and -10b deficient mice. To this end, mRNAs without polyA synthesized from TALEN plasmids for mmu-mir-10a and -10b were mixed and injected into fertilized oocytes at the concentration of 100, 200 and 400 ng/ml each for single TALEN pair. Genotyping of obtained pups showed that there was not a mutant mouse by microinjection at the concentration of 100 and 200 ng/ml each for single TALEN pair, whereas two mutant mice were obtained from injection at 400 mg/ ml (Table 1). Both mutants have mutated mmu-mir-10b and wild type mmu-mir-10a alleles (Fig. 2B). Higher concentration of RNA mixture for microinjection enabled us to produce mmu-mir-10b, prompted us to try microinjection of TALEN mRNAs at the concentration of 500 ng/ml for generating mmu-mir-10a deficient mice. Genotyping of the pups thus obtained showed that 3 out of 87 (3.4%) in total contained mutated alleles (Table 1, Fig. 3B), suggesting that TALEN activity is better in fertilized oocyte than NIH3T3 because we could not identify any mutated sequence by TALEN transfected NIH3T3 genome. Taken together, the optimum concentrations of RNA derived from TALEN plasmids could be different from design to design, probably because efficiencies of TALENs are different among target sequences of TALENs.

Transmission of the mutated allele to the next generation
To examine if the mutated alleles induced by microinjection of TALEN RNAs can be transmitted to next generation or not, two independent heterozygous mice deficient for mmu-mir-10a and -10b were mated with wild type C57BL/6 strain. Genotyping of the F1 pups was performed using T7EI assay. As shown in Fig. 3C and 3D, mutated mmu-mir-10a and -10b allele were identified. Ratios of the intensities of digested/undigested bands were quantified, showing that lane 1, 9 and 10 of Figure 3C (mmu-mir-10a mutant 3) are similar (0.39, 0.39 and 0.37, respectively), and lane 2 and 7 in Figure 3D are similar (0.19 and 0.24, respectively). Since F1 sequence of mutated allele is indicated with black line at the bottom. Deleted sequence is shown with a dark orange box with a white waved line. (B) PCR genotyping of the miRNA mutants. Genomic sequence is shown at the top with the target sequences of a pair of TALEN colored in green. Blue lines and boxes indicate pre-miRNAs and mature-miRNAs sequences, respectively. Deleted and inserted nucleotides are indicated with red dashes and blue letters, respectively. The sizes of deletions and insertions are shown to the right of mutated alleles with n and +, respectively. Alleles without mutation are indicated with wt (wild type). (C) Genotyping of F1 offspring of mmu-mir-10a KO (Mutant3) mouse by T7EI assay. Arrows indicate the digested PCR products containing mutation. DNA size marker is shown at left.  pups should be heterozygous, it seems likely that founder of mmumir-10a mutant 3 and mmu-mir-10b mutant 1 are heterozygous mutants without mosaicism. This results indicate that mutations induced by TALENs are transmittable to next generations, suggesting that TALEN mediated mutagenesis of mouse genome can be useful for making homozygous knockout mice by mating F1 heterozygous mice.

Expression levels
TALENs for miRNA knockout we used cleaved one of mature-miRNA sequence on miRNA precursor. It should be reasonable that mutated mature-miRNA itself, if not all alleles, does not function because several bases were deleted from short sequence of mature-miRNA. Also mutation on mature-miRNA may disturb precursor-miRNA to be processed to mature-miRNA because precursor-miRNA could not form hairpin structure. In this case remaining intact mature-miRNA located the other side of mutated mature-miRNA on the precursor molecule can not be expressed. We tested this possibility by RT-PCR analysis of mir-10a* using 6base deleted mir-10a mutant mouse. As shown in Fig. 4, the expression level was reduced in wt/mutant hetero mouse and deleted in mutant homo mouse, suggesting that mutations introduced by TALEN at one of mature-miRNA is enough for deplete function for both mature-miRNAs on a precursor.

Discussion
Our results demonstrate that TALENs can be applicable for generating knockout mice for gene desert and miRNA loci without using ES cells in high throughput way and in a short period of time, by optimizing RNA concentration for microinjection for each set of TALEN design. Also we showed that functions for both mature-miRNAs on a precursor can be depleted by introduction of mutations by TALEN at one of mature-miRNA.
To use ES lines to make conventional KO mice, the current homologous recombination system utilizes positive selection with Neo cassette. Sometimes this Neo gene effects other gene expressions nearby but deletion of this gene will need additional mating with Cre-expressing mice, which takes extra 2-3months. TALEN system does not need this step.
Although gene targeting of mice by TALENs has been reported as an alternative of ES method, application of TALENs to mice genetic studies should provide us a huge advantage to overcome difficulties on generating KO mice as follows.
miRNAs are often located in the intron region of host genes. Current technique using ES cells and recombination will leave lox genes which may affect host gene expression level. In this regard, application of TALEN for targeting miRNAs that provides to very short deletion in genome without any additional sequences should be very useful without hurting backbone gene expression.
Furthermore, in general, inter gene regions, including enhancer sequences, and some specific regions of genes such as genes in Y chromosome may contain relatively high repeat sequences, which sometimes leads very low efficiency of homologous recombination. TALEN system is targeting very specific but short genome sequences, thus, this problem can also be solved.
TALEN application will be useful in other aspects of genetic researches using mice. It will be also applicable to make domain specific dominant negative (or active) deletion mice, when inframe mutation is introduced in the specific region of gene. In addition, to make double knockout mice of mutually closely located two genes used to be a painful work, as cross of each KO mice cannot generate double KO (DKO) and thus need to establish new ES lines from one of KO mice. Application of TALEN system may overcome this problem.
TALEN method will provide us the chances to expand our genetic research using mice in a high throughput way.