Targeted disruption of Tbc1d20 with zinc-finger nucleases causes cataracts and testicular abnormalities in mice

Background Loss-of-function mutations in TBC1D20 cause Warburg Micro syndrome 4 (WARBM4), which is an autosomal recessive syndromic disorder characterized by eye, brain, and genital abnormalities. Blind sterile (bs) mice carry a Tbc1d20-null mutation and exhibit cataracts and testicular phenotypes similar to those observed in WARBM4 patients. In addition to TBC1D20, mutations in RAB3GAP1, RAB3GAP2 and RAB18 cause WARBM1-3 respectively. However, regardless of which gene harbors the causative mutation, all individuals affected with WARBM exhibit indistinguishable clinical presentations. In contrast, bs, Rab3gap1-/-, and Rab18-/- mice exhibit distinct phenotypes; this phenotypic variability of WARBM mice was previously attributed to potential compensatory mechanisms. Rab3gap1-/- and Rab18-/- mice were genetically engineered using standard approaches, whereas the Tbc1d20 mutation in the bs mice arose spontaneously. There is the possibility that another unidentified mutation within the bs linkage disequilibrium may be contributing to the bs phenotypes and thus contributing to the phenotypic variability in WARBM mice. The goal of this study was to establish the phenotypic consequences in mice caused by the disruption of the Tbc1d20 gene. Results The zinc finger nuclease (ZFN) mediated genomic editing generated a Tbc1d20 c.[418_426del] deletion encoding a putative TBC1D20-ZFN protein with an in-frame p.[H140_Y143del] deletion within the highly conserved TBC domain. The evaluation of Tbc1d20ZFN/ZFN eyes identified severe cataracts and thickened pupillary sphincter muscle. Tbc1d20ZFN/ZFN males are infertile and the analysis of the seminiferous tubules identified disrupted acrosomal development. The compound heterozygote Tbc1d20ZFN/bs mice, generated from an allelic bs/+ X Tbc1d20ZFN/+ cross, exhibited cataracts and aberrant acrosomal development indicating a failure to complement. Conclusions Our findings show that the disruption of Tbc1d20 in mice results in cataracts and aberrant acrosomal formation, thus establishing bs and Tbc1d20ZFN/ZFN as allelic variants. Although the WARBM molecular disease etiology remains unclear, both the bs and Tbc1d20ZFN/ZFN mice are excellent model organisms for future studies to establish TBC1D20-mediated molecular and cellular functions.

Mouse models of human genetic disorders are excellent resources for elucidation of the molecular and cellular disease etiologies. Recently, we reported that blind sterile (bs) mice, initially identified over 30 years ago as a spontaneous autosomal recessive mouse mutation exhibiting cataracts [11,12] and male infertility [13,14], carry a loss of function mutation in the Tbc1d20 gene [5]. The bs mice recapitulate the lens and testicular phenotypes observed in the WARBM4 children, although no morphological brain abnormalities were noted [5]. Rab3gap1 -/mice do not exhibit any morphological abnormalities of the eyes, brain, or genitalia, but exhibit synaptic exocytosis abnormalities [15]. Recently, it was shown that Rab18 -/mice exhibit cataracts, atonic pupils, and progressive hind limb weakness associated with accumulations of neurofilament and microtubules in the synaptic terminals [16]. This phenotypic variability between mice with disrupted WARBM genes has been previously attributed to genespecific and species-specific compensatory mechanisms present in mice [4,5].
Rab3gap1 -/and Rab18 -/mice are mouse models that were genetically engineered using standard approaches [15,16]. In contrast, the Tbc1d20 mutation in the bs mouse arose spontaneously [11]. Our genetic analysis of the bs mice identified a 416 kb genomic region in linkage disequilibrium within the bs locus [5]. The analysis of the bs critical region identified 16 RefSeq candidate genes and further evaluation of the candidate genes focused on the sequencing of the exons and exon/intron boundaries as well as RT-PCR analysis and subsequent sequencing of the open reading frames [5].
This approach identified a c.[691 T > A; 692_703del] mutation in the Tbc1d20 gene as causing the bs phenotype; subsequent functional analysis of the TBC1D20-bs protein determined that the bs mutation results in the loss of TBC1D20 functional [5]. Given that we did not sequence the entire 416 kb bs critical region, we cannot eliminate the possibility that another mutation not residing within the exon/intron regions or open reading frames of the 16 candidate genes, but resides within the bs linkage disequilibrium region, may be contributing to the phenotypic differences between the bs, Rab3gap1 -/-, and Rab18 -/mice.
As a part of this study, we set out to unequivocally establish the phenotypic consequences caused by the disruption of the Tbc1d20 gene. We utilized the zinc-finger nuclease (ZFN)-mediated genomic editing approach to generate the Tbc1d20 ZFN/ZFN mice. Our results show that the Tbc1d20 ZFN/ZFN mice exhibit cataracts and testicular phenotypes indistinguishable from the cataract and testicular phenotypes identified in the bs mice. Additionally, the complementation analysis confirmed that the bs and Tbc1d20 ZFN/ZFN mice are allelic variants.

Results and discussion
ZFN-mediated disruption of the Tbc1d20 locus The ZFN mediated targeting of the Tbc1d20 gene (NM_024196) was designed to cut a 6 bp region within exon 4 (see Methods). This approach generated 3 Tbc1d20 ZFN founder mice with a 9 bp c.[418_426del] deletion ( Figure 1A). The Tbc1d20 ZFN transcript encodes a putative TBC1D20-ZFN protein with an in-frame 3 amino acid deletion p.
[H140_Y143del] within a highly evolutionarily conserved TBC domain ( Figure 1B). TBC1D20 is an ER associated protein that functions as a GTPase activating protein (GAP) enhancing the GTP hydrolysis rate when bound to RAB1 or RAB2 [5,17,18]. It was shown previously that overexpression of mouse or human TBC1D20-WT protein results in the disruption of Golgi structures [5,17]. It was also shown that overexpression of the catalytically inactive mouse or human TBC1D20 proteins did not have an effect on the Golgi morphology [5,17]. Therefore, we proceeded to evaluate the effects of overexpression of the FLAG-tagged TBC1D20-WT and TBC1D20-ZFN proteins of Golgi structures in the HeLa cells. FLAG immunostaining confirmed the ER pattern of expression for both TBC1D20-WT and TBC1D20-ZFN proteins ( Figure 1C-D). HeLa cells overexpressing of the FLAG-tagged TBC1D20-WT protein exhibited disrupted Golgi structures and only residual GM130 immunostaining ( Figure 1C). In contrast, both untransfected ( Figure 1E) and HeLa cells overexpressing the FLAGtagged TBC1D20-ZFN protein exhibited similar GM130 immunostaining pattern ( Figure 1D) suggesting that TBC1D20-ZFN did not disrupt Golgi structures. Therefore, these findings suggested that TBC1D20-ZFN catalytic function was disrupted.
Eye, testicular, and brain phenotypes in Tbc1d20 ZFN/ZFN mice The Tbc1d20 ZFN/+ heterozygote mice did not phenotypically differ from the WT mice. The het to het breedings A B C D E Figure 1 The evaluation of the Tbc1d20 ZFN allele. ZFN-mediated genomic editing resulted in the Tbc1d20 ZFN transcript characterized by a 9 bp c.
[418_426del] deletion (A). The Tbc1d20 ZFN allele encodes the TBC1D20-ZFN mutant protein with an in-frame 3 amino acid p.[H140_Y143del] deletion within a highly evolutionarily conserved TBC domain. Missing amino acids are depicted in red (B). (C) Overexpression of FLAG-tagged TBC1D20-WT (green) led to a disruption of the Golgi as evident by the punctate GM130 immunostaining (red). (D) Overexpression of the FLAG-tagged TBC1D20-ZFN protein (green) did not disrupt GM130 immunostaining of the Golgi and did not differ from GM130 immunostaining of the untransfected HeLa cell (E). DNA was stained with DAPI (blue). Scale bars = 5 μm.
of the Tbc1d20 ZFN/+ mice recovered Tbc1d20 +/+ (n = 13), Tbc1d20 ZFN/+ (n = 27), and Tbc1d20 ZFN/ZFN (n = 10) progeny and these ratios did not significantly differ, following a chi-squared test, from expected ratios for a Mendelian autosomal recessive locus. Following the eyelid opening around postnatal day P14, clinical eye evaluation identified nuclear cataracts only in Tbc1d20 ZFN/ZFN that by P28 progressed to total cataracts characterized by vacuoles present throughout the entire lens (not shown). Histological analysis of Tbc1d20 ZFN/ZFN eyes confirmed severely disrupted vacuolated lenses with ruptured lens capsule and lenticular material in the vitreal cavity ( Figure 2B) although some lenticular material was also present in the anterior chamber ( Figure 2F). Lens epithelial cells did not appear to exhibit any gross morphological abnormalities whereas cortical and nuclear fiber cells were severely shortened and disorganized ( Figure 2D). Although retinal dismorphology and rosetting were evident in Tbc1d20 ZFN/ZFN eyes ( Figure 2B), the retina was laminated suggesting that rosetting may have been caused by the lens rupture and not by a defect in retinal development. Tbc1d20 ZFN/ZFN eyes also exhibited thickened pupillary sphincter muscle ( Figure 2F) that was not previously identified in bs eyes [5] suggesting that this TBC1D20-associated phenotype may be influenced by genetic modifiers.
Evaluation of the Tbc1d20 ZFN/ZFN brains did not identify any gross morphological abnormalities (not shown). Collectively these findings indicated that in Tbc1d20 ZFN/ZFN mice eye and testicular phenotypes are fully penetrant without any brain morphological abnormalities consistent with findings previously reported for bs mice [5].

Cellular phenotypes of Tbc1d20 ZFN/ZFN MEFs
An accumulation of enlarged lipid droplets (LDs) following oleic acid supplementation was the only cellular abnormality in the skin-derived TBC1D20-deficient fibroblasts from a WARBM4 patient [5]. Primary bs MEFs also exhibit an accumulation of enlarged LDs following treatment with oleic acid, but additionally the bs MEFs also exhibited enlarged Golgi structures [5]. Therefore, we proceeded to evaluate the LD and Golgi morphology in control and Tbc1d20 ZFN/ZFN MEFs. Our analysis confirmed a significant accumulation of enlarged LDs in the Tbc1d20 ZFN/ZFN MEFs ( Figure 4B) when compared to the LDs in the MEFs from the control mice ( Figure 4C) 24 h following oleic acid treatment and subsequent staining with the neutral lipid dye BODIPY 493/503. However, we did not observe any difference in the Golgi structures between control and Tbc1d20 ZFN MEFs following immunostaining with GM130 ( Figure 4D and F). Western blot analysis confirmed there was no difference in levels of GM130 protein in control and Tbc1d20 ZFN MEF cell lysates (not show). Although bs MEFs exhibited enlargement of Golgi structures, Golgi structures in the TBC1D20-deficient skin fibroblasts from a WARBM4 patient did not differ from Golgi structures in control skin fibroblasts [5]. However, thickened Golgi ribbons were observed in HeLa cells following shRNA mediated TBC1D20 knock-down [17]. Collectively these findings indicate that a spectrum of Golgi phenotypes is associated with TBC1D20 functional deficiency indicating that this phenotype is most likely influenced by genetic modifiers.

Conclusions
In mice, the disruption of Tbc1d20 results in vacuolated cataracts and a defect in acrosomal formation resulting in male infertility. At the cellular level, disruption of Tbc1d20 resulted in an accumulation of LDs. Thickening of the pupillary sphincter muscle eye phenotypes and aberrant Golgi cellular phenotypes were not penetrant on all genetic backgrounds suggesting that these phenotypes, caused by disruption of Tbc1d20, may be influenced by genetic modifiers. Although molecular and cellular disease etiology caused by TBC1D20 functional deficiency in mice and humans remains unclear, bs and Tbc1d20 ZFN/ZFN mice are allelic variants and as such are excellent model organisms for future studies focusing on elucidating TBC1D20 function.

Mice
To target the mouse Tbc1d20 (NM_024196.3) gene, ZFN plasmid design, assembly, validation and mRNA was done by the CompoZr Custom ZFN Service (Sigma). The ZFNs were designed to cut the c.[419ACTACT424] sequence within exon 4. The Tbc1d20 targeting ZFN mRNA was injected into the B6D2F1/Crl (F1 het from C57BL/6 N and DBA2 strains) embryos, which were implanted into pseudo-pregnant females. Pups were genotyped using standard conditions with ZFN-F 5′CTGGGTGTCATG AGCAATGT3′ and ZFN-R 5′AGGAGGCTGAGGAGTG ACCT3′ primers, electrophoresed, gel purified using the QIAquick Gel Extraction Kit (Qiagen), and screened for mutations using the Cel1 nucleotide mismatch assay (Sigma). The founders were confirmed by Sanger sequencing (Retrogen). Tbc1d20 ZFN/+ did not differ phenotypically from Tbc1d20 +/+ mice and both genotypes were used as controls. RNA was isolated from spleen, kidney, liver, and testes and the Tbc1d20 transcript was reverse transcribed, PCR-amplified and sequenced as previously described [5]. Comparative sequence analysis was performed using DNAStar software. Allelic breedings utilized bs/+ mice previously obtained from Jackson Laboratories and the bs allele was genotyped as previously described [5]. The treatment and use of all animals in this study was compliant with all protocols and provisions approved by the Institutional Animal Care and Use Committee (IACUC) at the Medical College of Wisconsin.

Clinical evaluations, histology, and immunohistochemistry
Mouse eyes were examined with a Topcon SL-D8Z slit lamp biomicroscope with a Nikon SLR-based Photo Slit Lamp imaging system following mydriasis with 1% Atropine Sulfate (Bausch & Lomb). Eyes, brains, and testes were collected at 8 weeks of age. Eyes and testes were fixed in 4% paraformaldehyde (PFA), paraffin embedded and H&E stained as previously described [5]. Brains were fixed at 4°C for 24 h in 4% PFA followed by 30% sucrose for 24-72 hrs. Brains were then sectioned at 30 μm on a sliding microtome (Leica) and stained with DAPI to label all nuclei. Immunostaining was done with TRA54 (B-Bridge) as a primary antibody and DyLight 488 goat anti-rat (Abcam) as a secondary antibody following the manufacturer's recommendations. PNA staining was performed utilizing the Lectin PNA-Alexa-488 conjugate (Life Technologies) according to the manufacturer's recommendations. Slides were DAPI stained according to the manufacturer's recommendations (Life Technologies), mounted using Fluoromount-G (Southern Biotech), and imaged using a Nikon DS-Fi1 camera on a Nikon Eclipse 80i microscope using NIS-Elements software (Nikon).
HeLa cells were cultured in DMEM containing 10% fetal bovine serum at 37°C and 5%CO 2 . For transfections, HeLa cells were grown on glass slides in 12-well plates and transfected with Lipofectamine LTX (Life Technologies) following the manufacturer's recommendations. Following transfections, the coverslips were washed with 1XPBS, then fixed with 4% PFA in PBS pH7.4 for 15 minutes at room temperature, washed with ice cold 1XPBS, permeabilized with 0.25% Triton X-100 in PBS (PBST), and then washed with 1X PBS for 3X5 minutes. The coverslips were immunostained with FLAG (Sigma) and GM 130 (Abcam) antibodies overnight at 4°C and for 1 hr at RT, with Alexa 488 and 546-conjugated (Life Technologies) secondary antibodies following the manufacturer's recommendations. The coverslips were stained with DAPI for 5 min, washed with 1XPBS, mounted onto glass slides with Fluoromount-G mounting medium, and photographed with a Nikon DS-Fi1 camera on a Nikon Eclipse 80i microscope.

Mouse embryonic fibroblasts (MEFs)
MEFs were isolated from the E13.5 mouse embryos (from the Tbc1d20 ZFN/+ X Tbc1d20 ZFN/+ cross) that genotyped either Tbc1d20 ZFN/ZFN or Tbc1d20 +/+ and were maintained as previously described [5,21]. Lipid droplets were evaluated as described previously utilizing media supplemented with 400 μM oleic acid (Sigma Aldrich) for 24 h and stained with 1 μg/μL BODIPY 493/503 (Life Technologies) [5]. All slides were mounted using Vectashield with DAPI (Vector Labs). Imaging was done with a Nikon DS-Fi1 camera on a Nikon Eclipse 80i microscope using NIS-Elements software (Nikon). Quantification of the lipid droplets was performed as previously described [22] using ImageJ (US National Institutes of Health) and NIS-Elements software. For each analysis, at least 20 cells per genotype were evaluated and statistical significance was determined by a t-test (Graphpad Prism) where p < 0.05 was treated as significant. For Golgi analysis, the control and Tbc1d20 ZFN/ZFN MEFs were immunostained using GM130 (Abcam) primary antibody and Alexa 488-conjugated secondary antibody (Life Technologies) following manufacturers' recommendations. Western blots were run using cell lysates generated from control and Tbc1d20 ZFN/ZFN MEFs following lysis with RIPA buffer supplemented with a protease inhibitor cocktail (Sigma). Cell lysates were immunoblotted with GM130 (BD Biosciences) primary antibody and HRP-conjugated secondary antibody (Abcam) following the manufacturer's recommendations as previously described [5]. Even loading was established following immunoblotting with β-actin HPR conjugated antibody (Abcam). The detection was performed using the ECL Western Blot Analysis System (Amersham) following the manufacturer's instructions.