Zinc-finger nuclease mediated disruption of Rag1 in the LEW/Ztm rat

Background Engineered zinc-finger nucleases (ZFN) represented an innovative method for the genome manipulation in vertebrates. ZFN introduced targeted DNA double strand breaks (DSB) and initiated non-homologous end joining (NHEJ) after pronuclear or cytoplasmatic microinjection into zygotes. Resulting frame shift mutations led to functional gene ablations in zebra fish, mice, pigs and also in laboratory rats. Therefore, we targeted the rat Rag1 gene essential for the V(D)J recombination within the immunoglobulin production process and for the differentiation of mature B and T lymphocytes to generate an immunodeficient rat model in the LEW/Ztm strain. Results After microinjection of Rag1 specific ZFN mRNAs in 623 zygotes of inbred LEW/Ztm rats 59 offspring were born from which one carried a 4 bp deletion. This frame shift mutation led to a premature stop codon and a subsequently truncated Rag1 protein confirmed by the loss of the full-length protein in Western Blot analysis. Truncation of the Rag1 protein was characterized by the complete depletion of mature B cells. The remaining T cell population contained mature CD4+/CD3+/TCRαβ+ as well as CD8+/CD3+/TCRαβ+ positive lymphocytes accompanied by a compensatory increase of natural killer cells in the peripheral blood. Reduction of T cell development in Rag1 mutant rats was associated with a hypoplastic thymus that lacked follicular structures. Histological evaluation also revealed the near-complete absence of lymphocytes in spleen and lymph nodes in the immunodeficient Rag1 mutant rat. Conclusion The Rag1 mutant rat will serve as an important model for transplantation studies. Furthermore, it may be used as a model for reconstitution experiments related to the immune system, particularly with respect to different populations of human lymphocytes, natural killer cells and autoimmune phenomena.


Background
Several modern techniques facilitate the genetic manipulation of the rat genome in vivo. Aside from the pronuclear injection of recombinant DNA constructs, the integration of lentiviral vectors into the rat genome after injection into the perivitelline space and transposonmediated insertions have enabled studies in transgenic rats [1][2][3][4][5]. Random integration of Sleeping Beauty and PiggyBac transposons caused gene inactivation at the integration sites [6,7]. Moreover, the cultivation of rat embryonic stem cells does allow gene targeting and functional deletion through homologous recombination (HR) in vitro [8][9][10][11]. However, due to the technical complexity of these procedures only the p53 gene and the Par-2 gene have been inactivated through HR in ES cells to date [12,13]. Alternative strategies for the functional inactivation of genes use the highly efficient Zinc-finger nuclease (ZFN) [14] and transcription activator-like effector nucleases (TALEN) technologies [15], both of which offer a rapid approach in gene targeting. Zinc-finger nucleases are hybrid proteins consisting of polymeric zinc-finger proteins fused to the endonuclease domain of the restriction enzyme FokI. A pair of zinc-finger nucleases binds to two contiguous target sequences in each DNA strand separated by a 6 bp cleavage site. Subsequent dimerization of the FokI domains causes DSB and initiate endogenous repair processes. Inaccurate repairs by NHEJ in response to DNA damage introduce deletions or insertions in the spacer region. The resulting frame shift leads to a premature stop codon and the translation of a truncated, inactive protein [16,17]. Therefore, ZFN have been used for the manipulation of human and rodent cells in vitro [18][19][20] as well as for gene targeting or for homology directed repair in zebra fish, mouse, rat and pig embryos in vivo [21][22][23][24][25]. In the rat, genes such as immunoglobulin heavy chain 6 (Igh-6), immunoglobulin heavy chain (gamma polypeptide) (JH locus), transgenic eGFP, renin (Ren), interleukin 2 receptor gamma (Il2rg), ATP-binding cassette, sub-family B (MDR/TAP), member 1A (Mdr1a) and Rab38 have been successfully inactivated by ZFN mediated gene disruption in single cell embryos [26][27][28][29][30].
Here we report about generating a Rag1 knockout by ZFN in LEW rats, with the intention of establishing an immunodeficient rat model on a defined genetic background.
Rag1 as well as Rag2 are highly conserved genes located in close vicinity in head to tail arrangement on chromosome 11p13 in humans, on chromosome 2qE2 in the mouse, and at chromosome 3q31 in the rat. In humans, Rag1 mutations lead to severe combined immunodeficiency (SCID) due to B and T cell loss (T -B -SCID) or Omenn's Syndrome (T + B -SCID) with a shortened life expectancy reviewed in [31][32][33][34]. A similar phenotype was described after gene targeting of Rag1 in the mouse with an arrest of B and T precursor cells in an early developmental stage, no serum IgG and an elevated number of natural killer cells (NKC) [35][36][37][38].
The generated Rag1 mutant rat on a defined genetic background can provide a useful tool for studies of the immune system, in transplantation studies, and in autoimmune diseases.

Embryo manipulation and Rag1 gene targeting
A pair of zinc-finger nucleases consisting of five zincfinger modules recognizing 15 bp at both sides of a 6 bp cut side located close to the 5' end of Exon 2, were used to induce DSBs and NHEJ in the Rag1 gene ( Figure 1A). The injection of the Rag1 specific ZFNs into 623 zygotes harvested from 35 hyperovulated donors resulted in 444 morphologically intact embryos transferred to 19 pseudopregnant (LEW x WKY) F1 females. From the 59 pups born, 49 survived the weaning period.
Surveyor mutation detection assay was performed to identify offspring carrying a Rag1 mutation by cutting of heteroduplex DNA derived from hybridization of wild type LEW/Ztm DNA with founder DNA through the Cel-I endonuclease [39,40]. Analysis of the offspring revealed partial cleavage of the 310 bp hybrid DNA into 174 bp and 136 bp fragments in one female (denominated "founder 58"), giving a 1.7% rate of transgenesis of the Rag1 gene ( Figure 1B).

Rag1 frame shift mutation
Sequencing of exon 2 of the Rag1 gene from "founder 58" verified an A to C base exchange at position 5244 combined with a 4 bp deletion at position 5246 -5249 while mutations were absent on the second allele (Figure 2 A). The 4 bp deletion in the Rag1 gene of "founder 58" caused a frame shift mutation and generated a premature stop codon at position 667 of the Rag1 mRNA, thereby encoding a truncated 198 aa protein.
This shortened protein was homologous to the N-terminal part of wild type Rag1 but lost the zincbinding dimerization domain (ZDD) and the c-terminal Core domain of the 1040 aa full length protein [38] ( Figure 2B). No off-target mutations were seen in the 10 loci with the highest sequence homology to the zinc-finger binding and cut site in the Rag1 gene by the SURVEYOR mutation detection assay in "founder 58" (Figure 3).

Germ line transmission and genotyping
Subsequently, the heterozygous "founder 58" was mated to a wild type LEW male to evaluate germ line transmission and to establish the novel Rag1 deficient LEW/ Ztm-Rag1 em1Ztm coisogenic strain. To distinguish between wild type and heterozygous pups, we used the loss of the recognition site of the restriction endonuclease StyI caused by the 4 bp deletion in Exon 2 of Rag1.  products derived from wild type allele were cleaved into 603 bp and 449 bp fragments by StyI, mutated Rag1 sections remained uncut. Six out of 11 offspring were heterozygous, demonstrating germ line competence and confirming the heterozygosity of "founder 58" ( Figure 4A). Offspring carrying the mutation was used for further breeding. Heterozygous LEW/Ztm-Rag1 em1Ztm rats showed normal reproduction with litter sizes ranging from 6 to 12. The mutation was inherited in the expected Mendelian ratio.

Macroscopical and histological evaluation
Phenotypical characterization revealed no ontogenetic disorders despite of underdeveloped thymi of the Rag1 mutant rats ( Figure 4B and C). The mean thymus weight was significantly reduced from 0.44 g (SD ±0,04) in wild type and 0.45 g (SD ±0,05) in heterozygous animals to 0.06 g (SD ±0,01) in homozygous Rag1 mutant rats ( Figure 4D and E). The reduced size of the thymus of Rag1 mutant rats was associated with a complete loss of the follicular structure showing the absence of T lymphocytes ( Figure 5B). Histological evaluation of the spleen demonstrated the differentiation in red and white pulp containing B cells in the lymphatic follicles and T lymphocytes in the periarteriolar lymphoid sheaths (PALS) in wild type animals ( Figure 5C). Besides the central artery the follicles of the white pulp are reduced and the PALS were lost in rats with partial Rag1 inactivation ( Figure 5D). In wild type lymph nodes the cortex was characterized through trabecular structures separating follicles where B cell differentiation occurred in the germinal centers ( Figure 5E). The lymph nodes of homozygous LEW/Ztm-Rag1 em1Ztm rats appeared devoid of T and B lymphocytes and follicular structures, in which physiological B cell differentiation would occur, were absent ( Figure 5F). No further macroscopic differences in organ size and structures were found in homozygous LEW/Ztm-Rag1 em1Ztm rats, and careful histological evaluation revealed no malformation of organs outside of the lymphatic system.

Disruption of wild type Rag1
Molecular characterization by RT-PCR showed a shift from the expression of wild type Rag1 mRNA in the lymphatic organs of wild type rats to the exclusive expression of mutated Rag1 in knockout rats. In heterozygous animals wild type and mutated Rag1 were transcribed equally ( Figure 6A). Western Blot analysis confirmed the loss of full length Rag1 protein in the thymus of homozygous mutant rats, while heterozygous animals exhibited a significant reduction of the Rag1 protein concentration as compared to the strong translation of Rag1 in wild type rats ( Figure 6B).

Discussion
Low efficiency of Rag1 specific ZFNs in LEW/Ztm rats The most striking advantage of the ZFN technology is the capability to manipulate embryos of any species. Nevertheless, the efficiency of the genetic manipulation may be modulated by strain variations and species differences, as well as locus specific chromatin structure and technical setup. In LEW/Ztm 71,3% of the injected embryos were still viable and resulted in 13,3% live births. However, only 1,7% carried the expected Rag1 mutation. In contrast to this the workgroup at the SAGE Lab, which used the identical ZFN mRNAs for the disruption of Rag1 in SD rats achieved a ratio of 20% of Rag1 mutated rats (5 out of 24 pups born, personal communication). The efficiency of 3% (1 out of 31 born animals carrying the mutation) seen in the targeting of the Renin gene in inbred SS (salt-sensitive) rats [30] was at a similar low level as the ablation of Rag1 in the LEW/Ztm background. Guerts and coworkers found survival rates comparable to the LEW rat with 56% to 91% after microinjection of ZFN encoding plasmids or ZFN mRNA in the SD background. The ratio of live-born pups born (13 -20%) also correlates with our results, however, the efficiency of the manipulation was much higher in the SD rats up to 75% mutants [27] instead of 1,7% on the LEW background. Nevertheless, Mashimo and colleagues detected high average ratios of born and mutated animals with 24,3% and 24,1% after micromanipulation of zygotes from inbred F344/Stm or TM/Kyo strains to inactivate the IL2rg gene [28]. However, comparing the efficiency of the Rag1 mutation in outbred SD and inbred LEW suggests that the different genetic backgrounds are an important determinant in the ratio of transgenesis, even though it also depends on the accessibility of the gene locus due to the chromatin structure [41].
Further, the conditions of the technical engineering determined the success of the ZFN mediated gene targeting. Guerts and coworkers showed that the stepwise increase of the concentration of the injected mRNAs from 0.4 μg/μl to 2μg/μl and then to 10 μg/μl correlates with reduced ratios of born embryos. Among these the number of mutant offspring increased with the injection of higher mRNA concentrations in SD rats [27]. Therefore, the functionality of the individual pair of ZFNs and the technical setup of the experiment in combination with the stress sensitive and delicate LEW embryos might be responsible for the low outcome of Rag1 mutated rats in this study [42]. Future comparison of identical ZFN in several rat inbred strains will be necessary to validate strain dependent variations of ZFN mediated manipulation efficiencies.

High fidelity of the Rag1 ZFN
The introduction of genetic modifications in the mammalian genome harbors the risk of causing unwanted, random or off-target mutation. It is well known that the Cre recombinase widely used for the conditional deletion of genomic sequences recognizes cryptic "pseudo-loxP" sites, initiates chromosomal rearrangements and infertility in mice [43,44]. Despite the high binding specificity of ZFN to their specific target sequence additional off-target mutations at homologous loci must be excluded. Neither in our LEW-Rag1 mutant rat, nor in any other published knockout rat model created by ZFN-mediated genome modification, offtarget mutations have been detected so far, thereby demonstrating the high fidelity of ZFN binding and activity [21,[26][27][28][29][30].

Leakiness of LEW/Ztm-Rag1 em1Ztm
The phenotype of the LEW-Rag1 mutant rat was characterized by a substantial decrease of Rag1 translation in heterozygous animals and by a complete loss of the Rag1 protein in the homozygous mutant rats suggesting a biallelic expression of Rag1 in wild type rats. The absence of dose dependent effects in the heterozygous animals emphasized that the translation of Rag1 from only one allele was sufficient for a functional immune system while homozygous Rag1 mutation was associated with a reduction of T cells, 11% CD3 positive T cells and some CD45RA expressing B cells were detected. Rag1 deficiency excludes lymphocyte maturation as shown in the Rag1 knockout mouse. T cell precursors mature only up to the double negative CD4 -CD8 -IL2R-α positive stage in the absence of murine Rag1 [35,37]. Therefore, the origin of mature T lymphocytes in the Rag1 knockout rat is still unknown as extrathymic thymocyte maturation, such as intestinal intraepithelial TCRγδ + T lymphocytes (IEL) in BB rats, CD2 + CD7 + preTCRα + T cells in the adult liver or CD2 + CD7 + CD3 -T cells in the human small intestine, always require the expression of functional Rag1 [46][47][48]. A similar frame shift mutations like in the LEW/Ztm-Rag1 em1Ztm were observed in the human Rag1 gene in two patients with Omenn's syndrome carrying an adenosine deletion at position 877 of the Rag1 cDNA also introducing a premature stop codon in front of the nonamer binding region (NBR) [49,50]. Santagata and colleagues showed that in response to the frame shift mutation truncated Rag1 protein had been translated from an internal methionine without the N-terminal BIIA nuclear localization signal. Like in LEW/Ztm-Rag1 em1Ztm homozygous rats, in these patients differentiation of mature, activated but oligoclonal T cells with impaired function took place due to cytoplasmic localization of the truncated Rag1 and reduced recombinational activity [49,50]. In the future cloning of the LEW/Ztm-Rag1 em1Ztm allele followed by in vivo translation and recombination should demonstrate the translation, cellular localization and functionality of the mutated Rag1 protein in the LEW/Ztm rats.

Functional compensatory NK cell increase
The elevated NK cell number in the peripheral blood of LEW/Ztm-Rag1 em1Ztm rats correlates with the observation of an increase in NKC population in spleen, lymph nodes, lung, and liver of Rag1 knockout mice [36]. Grundy and Sentman speculate that the NKCs substitute T lymphocytes by migration along chemokine gradients into these organs. The increase in NK cell concentration in the blood of LEW/Ztm-Rag1 em1Ztm mutant rats was also associated with a reduction of thymocytes, but the molecular pathways involved in this compensatory process are still unknown.
Further characterization of the LEW/Ztm-Rag1 em1Ztm strain as well as xenotransplantation experiments of cells and tissues will be performed to establish the Rag1 mutant rat as a innovative model for studies in autoimmune diseases, transplantation approaches and carcinogenesis as well as cancer therapy.

Animals
All rats were bred and maintained at the Central Animal Facility of the Hannover Medical School, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany (subline code: Ztm: http://www.mh-hannover.de/einrichtungen/tierlabor). All handling of animals has been conducted in accordance with German law for animal protection and with the European Communities Council Directive 2010/63/EU for the protection of animals used for experimental purposes. All experiments have been approved by the Local Institutional Animal Care and Research Advisory committee and permitted by the local government (Lower Saxony State Office for Consumer Protection, Food Safety, and Animal Welfare Service -LAVES; Az.10/0051).

Husbandry
LEW, LEW/Ztm-Rag1 em1Ztm , and (LEW x WKY)F1 rats were maintained under standardized conditions at a temperature of 22 ± 2°C, relative humidity of 55±5%, and artificial light for 14h. Commercial softwood granulate bedding was sterilized (Lignocel, Altromin; Lage, Germany). They received an autoclaved commercial pelleted diet (Altromin 1314) and water ad libitum. The rats were kept as pairs or in sibling groups. Microbiological status was monitored according to FELASA recommendations [51]. The rats were positive for parvovirus and apathogenic protozoa.

Zinc-finger nucleases
Rag1 specific ZFN consisting of 5 zinc-finger modules fused to the endonuclease domain of the restriction endonuclease FokI were designed and functionality evaluated by Sigma Advanced Genetic Engineering Labs using the CompoZr W ZFN technology. The forward ZFN recognized the DNA sequence GTCTACTGCCCAAGG between positions 5233-5247, while the reverse ZFN detected the DNA segment GACCGTGGAGTGGCA located at 5253-5267 in Exon 2 of genomic Rag1 (Rattus norvegicus strain BN/SsNHsdMCW chromosome 3, Rnor_5.0: NCBI Reference Sequence: NC_005102.3). The ZFN binding sites were separated by the 6 bp cut site. ZFN mRNAs were translated in vitro using the MessageMax T7 ARCA-Capped Message Transcription Kit (#MMA60710, Epicentre Biotechnologies, Madison, WI) and the Poly (A) Polymerase tailing kit (#PAP5104H, Epicentre Biotchnologies, Madison WI) following the manufacturer's recommendations. mRNA quantity was measured using the Nanodrop spectrophotometer (Peqlab, Erlangen, Germany) and sufficient quality was confirmed by agarose gel electrophoresis.

Mutation detection
Rag1 from the putative founder animals and from wild type LEW rats was amplified with the Optimase W Polymerase (Transgenomic, Glasgow, UK) and the primer pair Cel-I forward 5'-CTCATTGCCAGAGTTTTCCG-'3 and Cel-I reverse 5'-TGCTGACCCTAGCCTGAGTT-'3 using the following PCR protocol: Performing blast search 10 genomic loci showing strong homology to the ZFN binding and cut site of Rag1 were identified. Off-target mutations at these loci were excluded by PCR amplification at an annealing temperature of 58°C followed by Cel-I digest using the SURVEYOR W Mutation Detection Assay (Transgenomic, Glasgow, UK) (Oligonucleotides see Additional file 1.

Western blot
Whole cell protein extracts were isolated from spleen, thymus, bone marrow and lymph nodes (mesenteric and cervical) from wild type, heterozygous, and Rag1 deficient rats (n=10) with NP40 buffer supplemented with Complete Mini Protease Inhibitor Cocktail Tablets (Roche, Mannheim, Germany), sodium orthovanadate (NEB, Frankfurt, Germany), PMSF (Sigma-Aldrich, Seelze, Germany) and DTT (Sigma-Aldrich, Seelze, Germany). Bradford Assay served to estimate the protein concentration. 50 μg of protein supplemented with 5x LSB loading dye was separated in a 10% gel by SDS-PAGE and transferred to Optitran BA-S85 membrane by semi-dry blotting. After blocking for 1 hour at room temperature with blocking solution (5% dry milk (Carl Roth, Karlsruhe, Germany) and 2,5% BSA, fraction V (Sigma-Aldrich, Seelze, Germany) in 3% TBS-Tween, pH 7,4) the membrane was incubated with the first antibody diluted in blocking solution at 4°C over night. Incubation with the secondary, HRP-conjugated antibodies was done for 1 hour at room temperature in 3% TBS-Tween, pH 7,4. Roti-Lumin W was used as chemoluminescence substrate. For detection of the Rag1 protein a 1:300 dilution of the RAG-1 antibody (sc-5599, SCBT, LaJolla, CA) was sufficient followed by incubation with a 1:2500 dilution of a donkey anti rabbit secondary antibody (ab16284, Abcam, Cambridge, UK). GAPDH expression was demonstrated with a 1:300 dilution of the anti-GAPDH antibody (OBT1636, AbD Serotec, Düsseldorf, Germany) in combination with a 1:5000 dilution of the rabbit anti mouse antibody (STAR13B, AbD Serotec, Düsseldorf, Germany).

Histology and thymus weight
Tissues were fixed in 4% formaldehyde solution for 3 days, paraffin-embedded and cut into 4 μm slices before staining with hematoxylin and eosin. Microphotographs were taken using a Zeiss AxioCam MRc camera and analyzed histologically. Thymus weight was measured using a LA230S scale (Sartorius, Göttingen, Germany). One-way analysis of variance (ANOVA) and Bonferroni's Multiple Comparison Test were performed using Graph Pad Prizm Software (Version 5.0a) to determine statistical significances with a threshold of P < 0.05.

FACS analysis
Blood was harvested from wild type, heterozygous and Rag1 deficient rats (n=10), and lymphocytes were isolated following standard procedures. Cells were incubated for 30 min at 4°C with a 1:10 dilution of the anti-CD45RA FITC (sc-53048, SCBT, LaJolla, CA) antibody to detect B cells, while the number of T cells and natural killer cells were determined by double staining with 1:10 dilutions of the PerCP mouse anti-rat αβ T-cell receptor (BD Pharmingen, Heidelberg, Germany) and the mouse anti rat CD161:FITC antibody (AbD Serotec, Düsseldorf, Germany). Double stainings were performed using mouse anti rat CD3:FITC/CD4:RPE dual color reagent (DC041), mouse anti rat CD3:FITC/CD8:RPE (DC042) and mouse anti rat CD4/CD25 (DC040) (Abd Sertec). For CD4/CD8 double staining the antibodies mouse anti rat CD4-Fitc (554843, BD Pharmingen) and mouse anti rat CD8A-PE (554857, BD Pharmingen) were used. Stained cells were measured together with unstained cells and isotype controls stained with 1:50 dilutions of the PerCP Mouse IgG1 κ isotype control antibody (BD Pharmingen, Heidelberg, Germany) or the Mouse IgG (FITC) antibody (ab37356, Abcam, Glasgow, UK) with the FACSCalibur flow cytometer (BD Pharmingen, Heidelberg, Germany). ARCSIN transformation, One-way analysis of variance (ANOVA) and Tukey's Multiple Comparison Test were performed using Graph Pad Prizm Software (Version 5.0a) to determine statistical significances with a threshold of P < 0.05.

Additional file
Additional file 1: Off-target gene loci:RAG1 ZFN mutagenesis. Competing interests EJW and XC are full-time of employees of SAGE Labs of Sigma-Aldrich Corporation. NHZ, SG, DW, MD and HJH declare that they have no competing interests.
Authors' contributions NHZ carried out the microinjection, genotyping, molecular biological analysis and drafted the manuscript. SG performed the histopathological examinations. DW was responsible for the FACS analysis. EJW and XC validated the Rag1 ZFNs and provided technical advice. MD and HJH participated in the study design. All authors read and approved the final manuscript.