Gene Targeting of Mouse Tardbp Negatively Affects Masp2 Expression

Amyotrophic Lateral Sclerosis (ALS) is a devastating adult onset neurodegenerative disease affecting both upper and lower motor neurons. TDP-43, encoded by the TARDBP gene, was identified as a component of motor neuron cytoplasmic inclusions in both familial and sporadic ALS and has become a pathological signature of the disease. TDP-43 is a nuclear protein involved in RNA metabolism, however in ALS, TDP-43 is mislocalized to the cytoplasm of affected motor neurons, suggesting that disease might be caused by TDP-43 loss of function. To investigate this hypothesis, we attempted to generate a mouse conditional knockout of the Tardbp gene using the classical Cre-loxP technology. Even though heterozygote mice for the targeted allele were successfully generated, we were unable to obtain homozygotes. Here we show that although the targeting vector was specifically designed to not overlap with Tardbp adjacent genes, the homologous recombination event affected the expression of a downstream gene, Masp2. This may explain the inability to obtain homozygote mice with targeted Tardbp.


Introduction
Amyotrophic Lateral Sclerosis (ALS) is an adult-onset and invariably fatal neurodegenerative disease characterized by the loss of motor neurons in the brain and spinal cord. The most common neuropathological feature of ALS is the presence of ubiquitinated inclusions (UBIs) in the cytoplasm of affected motor neurons [1,2]. The nature of the inclusions remained largely unknown until a remarkable discovery identified the TAR DNA binding protein (TDP-43) as a major component in UBIs of ALS and also of Frontotemporal Lobar Degeneration (FTLD-TDP) [3,4]. TDP-43 positive inclusions are present in both sporadic and familial forms of ALS and FTLD [3,4] and have also been identified in other neurodegenerative diseases, such as, Parkinson [5], Alzheimer [6], Huntington [7] and inclusion body myopathies [8]. Although mutations in the gene encoding TDP-43, TARDBP, are relatively infrequent (,1% of total ALS cases), TDP-43 pathology is common to .90% of ALS cases, including those caused by mutations in the recently identified gene, C9orf72 [9].
TDP-43 is ubiquitously expressed and is normally localized to the nucleus where it exerts multiple functions in transcription and pre-mRNA splicing [10]. However, the cytoplasmic location of pathological TDP-43 suggests that disease might be caused by a loss of nuclear function. To this end, Tardbp knockout mice have been generated to investigate this loss-of-function hypothesis, however homozygote mice were embryonic lethal, dying between 3.5 to 8.5 days of embryonic development, while heterozygotes did not display any motor neuron pathology [11,12]. To circumvent this problem we attempted to generate Tardbp conditional knockout mice using the classical Cre/loxP system, flanking exons 2 and 3 with loxP sites. Although we successfully generated viable and fertile heterozygote mice with the targeted allele, we were unable to obtain homozygotes. This was not due to effects of the targeting event on reducing TDP-43 expression, thereby mimicking a knockout mouse, but instead we show that the targeting event affected the expression of a downstream gene, Masp2. This effect on Masp2 could underlie the inability to obtain homozygous mice with targeted Tardbp.

Generation of Tardbp targeting vector
To generate a conditional knockout of the Tardbp gene a mouse bacterial artificial chromosome (BAC) clone containing Tardbp (RP23-29102) was obtained from the Children's Hospital Oakland Research Institute (CHORI, https://bacpac.chori.org/) and modified by recombineering. First, RP23-29102 was made proficient for recombination by electroporation of PSC101gbaA plasmid encoding the recombination machinery. Following integration of a Zeo cassette in intron 1 of the Tardbp gene, a loxP site was exchanged with the Zeo cassette and placed as the most 59 loxP recombination site, upstream of exon 2. A neomycin resistance cassette flanked by FLP recognition target (FRT) sites was amplified by Polymerase Chain Reaction (PCR) and integrated downstream of Tardbp exon 3. Finally, the modified gene was transferred to a plasmid containing the diphtheria toxin cassette to generate the targeting construct ( Figure 1).

Screening of ES clones and mouse genotyping
The Tardbp targeting vector was electroporated into C57BL6/ 129 embryonic stem (ES) cells at the Toronto Centre for Phenogenomics (http://www.phenogenomics.ca/). Initial identification of positive ES cell clones was performed by ethanol precipitation of genomic DNA (gDNA) and PCR amplification using primers specific for the most 59 loxP site. We found 8 positive ES clones which were subsequently expanded in 24 well plates and screened for recombination of the 59 and 39 homology arms by sequencing and Southern blot analyses, respectively. The sequence of the 59 and 39 FRT sites flanking the neomycin cassette on all 8 ES clones was verified. For Southern blots, 15 mg of gDNA was digested overnight with HindIII, run overnight on 0.8% agarose gels and transferred by capillarity to a Hybond-N+ nylon membrane (GE Healthcare). Prehybridization for 2 hours with ULTRAhyb Ultrasensitive Hybridization Buffer (Ambion) was followed by hybridization overnight using a radioactively labeled probe for detection of the endogenous Tardbp gene.

Mouse Breeding
All protocols were conducted in accordance with the Canadian Council on Animal Care and approved by the University of Toronto Faculty of Medicine and Pharmacy Local Animal Care Committee as well as the University of Toronto Animal Care Committee. ES clones with the correct recombination event were used to obtain germline-transmitting chimeras by morula aggregation at the Toronto Centre for Phenogenomics (http://www. phenogenomics.ca/). Deletion of the neomycin selection cassette was achieved by breeding the F1 Tardbp targeted mice with Actin-FLP mice (Jackson laboratory, #003800). Excision of the neomycin cassette was verified using PCR flanking primers.

Quantitative Real Time PCR
Mice at 5 months of age were euthanized in a CO 2 chamber and tissues were dissected and immediately snap-frozen in liquid nitrogen. Total RNA extraction was performed using RNAeasy kit (Qiagen); 2 mg of total RNA was used to synthesize cDNAs with oligodT and Superscript III Reverse Transcriptase (Life Technol- ogies). Real Time PCR was performed in 20 ml reactions using SYBR Green ER qPCR SuperMix Universal (Life Technologies) according to manufacturer's instructions for the ABI7500 instrument (Applied Biosystems). cDNA was diluted 1/10 and 1/20 and all the reactions were performed in triplicates. Tardbp transcripts were measured using primers specific for exons 5 (forward) and 6 (reverse) respectively. TDP_RT_F1: 59GTGGTAGATGTCTT-CATTCCCAAACC; TDP_RT_R1: GATCAAATCCTCTCCA-CAAAGAGACTG. Masp2 transcripts were measured with primers that discriminate between MAp19 and Masp2 splicing variants. The sequence of the primers is as follows: Masp_ex3_F: CTCCACCAGAACAAGCACAC; Masp_ex3_3UTR_R: CTGACTGGTATCTGACTCACTGG for MAp19 splicing variant; and Masp_ex6_F: GTGGAGACGCACCCTGAAGCC; Masp_ex7_R: AGCGTCTTCCCACAAAATGGGCC for Masp2 splicing variant. Actb gene, encoding for b-Actin, was used as a reference gene to normalize Tardbp, MAp19 and Masp2 transcript quantification, b-Actin_RT_F: GAACCCTAAGGC-CAACCGT;b-Actin_RT_R: GAGGCATACAGGGACAGCA-CAG. Statistical differences between quantitative real time PCR samples were assessed by GraphPad software (www.graphpad. com).

Results
To generate a conditional knockout of Tardbp we used the classical loxP-Cre recombination system. Using BAC recombineering technology, two loxP sites were introduced flanking exons 2 and 3 ( Figure 1A). Deletion of exons 2 and 3 prevents translation from both the first translation start site in exon 2 and an alternate translation start site located in exon 3.
Homologous recombination between the targeting vector and the Tardbp locus in ES cells generated the Tardbp targeted allele. The targeting event was verified in ES clones surviving neomycin selection by Southern blot and sequencing ( Figure 1B). ES clone 3G was empirically selected to generate germ-line transmitting chimeras which were crossed to C57BL/6 mice. In order to remove the flipped neomycin selection cassette, F1 Tardbp targeted mice were crossed to Actin-FLP mice and neomycin excision was verified by PCR amplification using primers flanking the FRT sites ( Figure 1C). Heterozygote mice bearing floxed exons 2 and 3 at the Tardbp locus were designated as Tardbp +/2lox .
Since it has previously been shown that homozygous Tardbp knockout mice are embryonic lethal [11][12][13], it is possible that the inability to obtain homozygote Tardbp 2lox/2lox mice was due to an inadvertent effect of the gene targeting event on the expression of Tardbp. Analysis of Tardbp mRNA transcripts in lumbar spinal cord revealed no diminishment in TDP-43 expression between Tardbp +/2lox mice and wildtype littermates and this was verified at the protein level by immunoblot analysis of brain, spinal cord and liver tissue (Figure 2A, 2B). However, since TDP-43 can autoregulate its own expression levels the possibility that the targeting event disrupted Tardbp expression cannot be excluded [14,15].
An alternate explanation for the absence of Tardbp 2lox/2lox homozygote mice could be an indirect effect occurring as a consequence of the homologous recombination event at the Tardbp locus. Even though the homology arms of the targeting vector were specifically designed not to overlap with adjacent gene sequences, it is possible that the homologous recombination event affected the expression of neighboring genes.
In the mouse genome, the 39 untranslated region (UTR) of the Tardbp gene overlaps with the 2 last exons of the Masp2 gene. The Masp2 gene encodes for two isoforms. MAp19 is the short isoform (19 KDa), with the last exon located 14.5 Kb downstream of Tardbp 39UTR. A longer isoform (74 KDa) called Masp2, has the last two exons overlapping 3.5 Kb with Tardbp 39UTR.
To investigate this hypothesis we used qRT-PCR to analyze the expression levels of Masp2 and MAp19 transcripts in the liver, where it is highly expressed, of Tardbp +/2lox and WT littermates ( Figure 2C). A significant reduction in expression of the Masp2 transcript was observed in Tardbp +/2lox mice compared to WT littermates, and this correlated with a ,50% reduction in Masp2 protein levels, as revealed by immunoblot analyses of replicate tissue samples (Figures 2C and 2D). A marginal but not significant reduction in MAp19 transcripts was also observed ( Figure 2C), that correlated with a slight reduction in MAp19 protein levels ( Figure 2D). These results show that targeting of the Tardbp locus affected expression from the adjacent gene, Masp2.

Discussion
In this investigation we generated heterozygous mice for a conditional knockout of the Tardbp gene by flanking exons 2 and 3 with loxP recombination sites. We found that while heterozygote Tardbp +/2lox mice were viable and fertile, we were unable to obtain homozygotes. Since Tardbp knockout mice are embryonic lethal [11][12][13], we checked whether there was a loss of TDP-43 expression from the targeted allele. There was no apparent reduction in TDP-43 expression at either the mRNA or protein level, which might be real or a result of auto-regulation of the Tardbp wildtype allele [14,15], as has been reported in other Tardbp heterozygote knockout mice [11][12][13]. We also considered whether the targeting event disrupted adjacent genes. In silico design of the targeting vector is one of the most important steps in generating gene knockouts in mice. Since homologous recombination of a targeting vector is a rare event [16], our targeting vector was designed with long homology arms to increase the frequency of targeted integration at the Tardbp locus [17], while avoiding overlap with genes located upstream and downstream of Tardbp.   4). B) Immunoblot analysis of TDP-43 protein expression in mouse brain, spinal cord, and liver shows no reduction in expression in Tardbp +/2lox mice compared to wildtype littermates. Actin was used to normalize the amount of protein loaded. C) Quantification of Masp2 splicing variants mRNA in mouse liver by quantitative Real-Time PCR. Masp2 long variant was significantly reduced in Tardbp +/2lox compared to WT (p = 0.03, n = 3). No significant difference was found for MAp19 transcript levels between WT and Tardbp +/2lox mice. D) Analysis of total protein extracts from mouse liver by Western blot. Masp2 is clearly reduced in Tardbp +/2lox mice livers when compared to WT, while MAp19 expression is marginally reduced. doi:10.1371/journal.pone.0095373.g002 expected to affect expression of this gene. Nevertheless, we found a significant reduction in both mRNA and protein expression of the long isoform of Masp2 in heterozygote Tardbp +/2lox mice compared to their non-transgenic littermates, indicating that the targeting did affect Masp2 and that this could underlie the inability to obtain homozygous Tardbp 2lox mice. Of note, the greater effect on Masp2 compared to MAp19 might relate to the overlap of the last 2 exons of Masp2 with the 39UTR of Tardbp.
The reduced expression of Masp2 could be the result of an introduction of unexpected mutations in the chromosome by the recombination event at the Tardbp locus. Even though homologous recombination generally occurs with high fidelity [18], many investigations have reported the incorporation of mutations upon targeted homologous recombination [19][20][21]. Moreover, the fidelity of homologous recombination can be compromised when the sequences of the targeting vector and the chromosome are not isogenic [18]. In our study, the RP23-29102 BAC used to make the targeting vector had a C57BL/6J background while the mouse Embryonic Stem cells were from a hybrid 129/C57BL6 background. It is possible that the homologous recombination event at Tardbp introduced mutations that altered regulatory sequences governing the expression of the Masp2 gene. Even though regulatory sequences have been identified upstream of the Masp2 gene [22,23], the regulation of this gene remains poorly understood. Some of the Masp2 regulatory enhancers could reside downstream of the gene, as has been reported for other genes [24,25].
Masp2 deficiency in mice confers protection from gastrointestinal ischemia/reperfusion injury and are viable in a homozygous state [26], which would seem contradictory to our inability to obtain homozygote Tardbp 2lox/2lox Masp2 deficient mice. However, this discrepancy could be due to the mouse strain background of the previously reported Masp2 deficient mouse, which was established in 129/Sv embryonic stem cells, whereas Tardbp +/2lox mice were in a hybrid C57BL6/129 strain background. It is well recognized that the strain background has a relevant influence in phenotype of transgenic mice [27,28].
There are relatively few publications on the generation and use of Tardbp conditional knockout mice for the study of ALS [29][30][31]. Chiang et al., 2010 floxed Tardbp exon 3 and found that ubiquitous deletion of Tardbp leads to a metabolic phenotype in embryonic stem cells. Wu et al., 2012 showed that deletion of Tardbp exons 2 and 3 in heterozygote mice leads to an age dependent progressive motor dysfunction. However, to date neither group has reported on homozygote mice. Iguchi et al., 2013 obtained homozygote mice with a conditional deletion of Tardbp exon 2, after backcrossing for at least 5 generations, and showed an age dependent motor dysfunction. Thus, generation of TDP-43 deficient mice has been a challenging task and in this investigation we show for the first time an effect of targeting Tardbp on downstream gene Masp2.