Defect of Mitotic Vimentin Phosphorylation Causes Microophthalmia and Cataract via Aneuploidy and Senescence in Lens Epithelial Cells*♦

Background: Vimentin, an intermediate filament (IF) protein, is phosphorylated in mitosis. Results: Disruption of vimentin phosphorylation during cell division leads to chromosomal instability (CIN) and premature aging in mouse lens tissue. Conclusion: Our data document the first physiological importance of vimentin phosphorylation during mitosis for organogenesis and tissue homeostasis. Significance: Our data suggest a possible causal relationship between CIN and premature aging.


Vimentin, a type III intermediate filament (IF) protein, is phosphorylated predominantly in mitosis. The expression of a phosphorylation-compromised vimentin mutant in T24 cultured cells leads to cytokinetic failure, resulting in binucleation (multinucleation)
. The physiological significance of intermediate filament phosphorylation during mitosis for organogenesis and tissue homeostasis was uncertain. Here, we generated knock-in mice expressing vimentin that have had the serine sites phosphorylated during mitosis substituted by alanine residues. Homozygotic mice (VIM SA/SA ) presented with microophthalmia and cataracts in the lens, whereas heterozygotic mice (VIM WT/SA ) were indistinguishable from WT (VIM WT/WT ) mice. In VIM SA/SA mice, lens epithelial cell number was not only reduced but the cells also exhibited chromosomal instability, including binucleation and aneuploidy. Electron microscopy revealed fiber membranes that were disorganized in the lenses of VIM SA/SA , reminiscent of similar characteristic changes seen in age-related cataracts. Because the mRNA level of the senescence (aging)-related gene was significantly elevated in samples from VIM SA/SA , the lens phenotype suggests a possible causal relationship between chromosomal instability and premature aging.
Intermediate filaments (IFs), 3 together with microtubules and actin filaments, form the cytoskeletal framework in the cytoplasm of various eukaryotic cells. IF proteins are divided into six groups related to their primary sequence characteristics, their tissue-specific expression profiles, and differentiation status (1)(2)(3). For example, GFAP and desmin are type III IF proteins that are expressed specifically in astroglial and muscular cells, respectively (4 -6). However, vimentin, another type III IF protein, is expressed in all mesenchymal cells with the eye lens being the tissue with by far the highest levels (7,8).
There is increasing evidence that IF disassembly is regulated by phosphorylation of Ser/Thr residues in the amino-terminal head domain on IF proteins (9,10). The first direct evidence was obtained by an in vitro study that treated polymerized vimentin filaments with a purified protein kinase to cause their disassembly (11). Site-and phosphorylation state-specific antibodies that can recognize a phosphorylated residue and its flanking sequence (12,13) are powerful tools able to resolve the spatial and temporal details of IF phosphorylation in cells (14,15). Using antibodies raised against four distinct phosphorylation sites in GFAP, we showed that mitotic IF phosphorylation is regulated by different protein kinases in a spatiotemporal manner (13,16). Further detailed studies revealed that several mitotic kinases, such as Aurora-B (17,18), Cdk1 (19,20), Plk1 (21), and Rho kinase (22)(23)(24), participate in the phosphorylation of type III IF proteins during cell division. By transient expression of type III IF proteins mutated at these mitotic phosphorylation sites to Ala, we found that preventing the phosphorylation of IFs during cell division inhibited cytokinesis by the retention of an IF bridge that connected the two daughter cells (17,18,21,25,26). These findings indicated that mitotic IF phosphorylation is essential for the efficient separation of the two daughter cells. The presence of the IF bridge presented the cell pair with two options. In the first option, the IF bridge could be torn apart presumably by cell adhesion-dependent traction force (27) to allow the completion of cytokinesis (26). In the second option, cytokinesis failed, resulting in the formation of binucleate (multinucleate) cells (21). The physiological significance of IF phosphorylation during cell division for organogenesis and tissue homeostasis, however, has yet to be determined.
We have generated knock-in mice to produce VIM alleles in which the serine residues in vimentin that are phosphorylated in mitosis have been replaced with alanine residues (indicated in Fig. 1A). Homozygotic mice bearing such mutations (VIM SA/SA ) presented with microophthalmia and lens cataract as phenotypes. The heterozygotes (VIM WT/SA ) showed no obvious phenotype. Compared with WT (VIM WT/WT ) and VIM WT/SA mice, the epithelial cell number was significantly reduced in lenses from VIM SA/SA mice. We observed not only binucleate (multinucleate) cells but also aneuploid cells in VIM SA/SA lens, whereas these abnormal cells were not detected in VIM WT/WT or VIM WT/SA lens. In addition, the mRNA level of the senescence (aging)-related gene was significantly elevated in the lens of VIM SA/SA .

EXPERIMENTAL PROCEDURES
Generation of Vimentin Knock-in Mice-A bacterial artificial chromosome clone containing the 129Sv mouse vimentin genome locus was purchased (Mouse bMQ bacterial artificial chromosome library, The Wellcome Trust Sanger Institute). The targeting vector was designed to substitute Ser to Ala in the vimentin head domain at the indicated positions (Fig. 1, A and  B). The 5Ј and 3Ј homology arms were amplified by PCR. Sitedirected mutagenesis was introduced on exon 1 for the substitution. Then they were ligated into the 5Ј and 3Ј side of the PGKneobpA-loxP-positive selection marker cassettes, with an MC1DTpA-negative selection marker cassette at the 5Ј homologous arm to enrich for homologous recombinants. The targeting vector was electroporated into EB3 ES cells with 129 background, as described previously (28,29).
G418-resistant clones were initially screened by Southern blotting analysis with a unique external probe following EcoRI digestion (see Fig. 1, B and C). Correctly targeted ES clones were microinjected into C57BL/6 blastocysts, which were in turn transferred into foster mothers to obtain chimeric mice. Following germ line transmission, homozygous mutant (VIM SA/SA ) mice were produced by intercrossing heterozygous (VIM WT/SA ) mice (Fig. 1, D and E). Mutations were verified by PCR followed by DNA sequencing (Fig. 1F). The mutant mice were backcrossed for more than 10 generations onto a C57BL/6 background. Genotyping was performed by PCR ( Fig. 1E)  Immunoblotting-Samples preparation was described previously (30). After incubation with the desired antibody, PVDF membranes (Immobilon-P, Millipore) were developed using the following detection reagents (Western Lightning Chemiluminescence Reagent plus, PerkinElmer Life Sciences, or Super-Signal West Femto Maximum Sensitivity Substrate, or Thermo Fisher Scientific).
Histology and Immunohistochemistry-Fixation and H&E staining of histological sections was performed according to standard protocols. For immunohistochemistry, 3,3Ј-diaminobenzidine staining was performed with a kit (Dako EnVisionϩ System-HRP Labeled Polymer, Dako), according to the manufacturer's protocol.
Immunostaining-Paraformaldehyde-fixed tissue slides were deparaffinized, microwaved in 10 mM citrate buffer, pH 6.0, for 15 min, and then blocked in 5% (v/v) donkey serum in TBST for 1 h. Sections were incubated with primary antibodies overnight at 4°C followed by TBST wash and then incubated with appropriate secondary antibodies (Invitrogen) for 1 h at room temperature.
Cultured cells were grown on coverslips (Iwaki Glass Co., Ltd.) and immunostained with the following modifications: fixation, with 1% (w/v) formaldehyde in PBS for 15 min at room temperature followed by permeabilization with 0.2% (v/v) Triton X-100 for 15 min or with 100% methanol at Ϫ20°C for 10 min; blocking, with 1% (w/v) BSA/PBS for 15 min; primary antibodies, for 1 h at room temperature; secondary antibodies for 30 min. DNA was also stained with 1 g/ml DAPI.
FISH-Mouse eyes were fixed with 4% (w/v) paraformaldehyde in PBS overnight at 4°C and embedded in paraffin. 5-mthick sections were prepared and mounted. Slides were deparaffinized and then treated with 0.1% (w/v) pepsin, 0.1 N HCl for 15 min. FISH probes were mounted with specimen, denatured for 10 min at 90°C, and then hybridized overnight at 37°C. After hybridization, slides were washed stringently in 50% formamide, 2ϫ saline sodium citrate (SSC), and 1ϫ SSC. DNA was also stained with 1 g/ml DAPI.
Fluorescence image was captured with a fluorescence microscope (Leica CW-4000). The mouse two-color FISH probe (Chromosome 12 probe; labeled with Spectrum Green, chromosome 19 probe; labeled with Cy3) was purchased from Chromosome Science Labo Inc. (Sapporo, Japan).
For the cultured cells, some modifications were made. In brief, cells were fixed with Carnoy's fixative (3:1 methanol/glacial acetic acid) at room temperature for 5 min with three changes before FISH probes were hybridized to the samples.
Primary Cell Culture-Primary lens epithelial cell cultures were established by microdissecting the anterior lens capsule and then culturing on collagen I-coated plastic dish (BD Biosciences) with Dulbecco's modified Eagle's medium supplemented with 10% (v/v) FCS and antibiotics. Sequentially, they were replaced on collagen I-coated coverslips (Iwaki Glass Co., Ltd.) and cultured for a total of 7 days from the harvest, followed by the fixation.
Transmission Electron Microscopy of Lens Fiber Cells-The fixation was based on the following standard method with some modifications for the equatorial section. Lens was fixed with 2% fresh formaldehyde and 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.4, for 2 h at room temperature. After washing with 0.1 M cacodylate buffer, pH 7.4, they were postfixed with ice-cold 1% OsO 4 in the same buffer for 2 h. The samples were rinsed with distilled water, stained with 0.5% aqueous uranyl acetate for 2 h or overnight at room temperature, dehydrated with ethanol and propylene oxide, and embedded in Poly/Bed 812 (Polysciences, Inc.). Then ultrathin sec-tions were cut through the lens equator, double stained with uranyl acetate and Reynolds's lead citrate, and viewed with a transmission electron microscope (JEM-1010; JEOL) with a charge-coupled device camera (BioScan model 792; Gatan, Inc.) at an accelerating voltage of 100 kV.
Statistical Analyses-Mean values Ϯ S.E. were calculated and used to test for significance between treatments in each dataset using the two-tailed Student's t test (Graph Pad software). Differences were considered significant when *, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001.

RESULTS
Lens Disorder in VIM SA/SA Mice-Knock-in mice were generated that only expressed vimentin mutated at Ser-6, Ser-24, Ser-38, Ser-46, Ser-55, Ser-64, Ser-65, Ser-71, Ser-72, Ser-82, and Ser-86 to Ala (Fig. 1, A-F). These serine residues in mouse vimentin have been identified as phosphorylation sites during mitosis. Western blot analyses of embryos showed no significant difference in vimentin protein levels between WT (VIM WT/WT ), heterozygous (VIM WT/SA ), and homozygous (VIM SA/SA ) mice (Fig. 1F). Next, we compared the expression patterns of vimentin and desmin in VIM WT/WT and VIM SA/SA embryos. There were only marginal differences in vimentin and desmin expression, as judged by in situ hybridization of mouse whole embryos at E8.5 ( Fig. 2A) and E9.5 (Fig. 2B). These data collectively indicate the successful generation of vimentin mutant mice.
VIM SA/SA mice were viable, but the eyes were microophthalmic, and the lenses were smaller than those in littermate controls (Fig. 2C). The gross morphology and histology of the eye tissues were unchanged for VIM WT/WT and VIM WT/SA mice (Fig. 2C). Lens cataract was observed in VIM SA/SA mice at 11 months old, whereas VIM WT/WT and VIM WT/SA mice all had clear lenses (Fig. 2D). As shown in Fig. 2E, we observed no redundant protein expression of desmin, GFAP, and nestin (with which vimentin can form heteropolymeric filaments) (3) in VIM SA/SA lens. The fact that the lens is the tissue where vimentin is the most abundantly expressed (Fig. 2F) likely explains the cataract phenotype in VIM SA/SA mice.
Binucleation and Aneuploidy in VIM SA/SA Lens-A single layer of epithelial cells covers the anterior hemisphere of the lens. The cells in the central zone of the epithelium (anterior region) rarely undergo cell division, but at the lens equator there is a band of dividing cells termed the germinative zone in the epithelium (equatorial region). The progeny of these cells in the germinative zone of the equatorial region differentiate into lens fiber cells, a process that starts in the transitional zone of the equatorial region (34). The immunohistochemical analyses with anti-PCNA (which stains active DNA replication sites in cells; Fig. 3A) identified the germinative zone (34).
Histological changes were observed in both the anterior (Fig.  3, B and C) and equatorial regions (Fig. 4, A and B) of VIM SA/SA lenses. The number of lens epithelial cells was significantly decreased in VIM SA/SA mice, compared with VIM WT/WT or VIM WT/SA mice. In contrast, there was no significant difference between similar regions in VIM WT/WT and VIM WT/SA mice. Because binucleation (multinucleation) was observed in tissue culture cells expressing phosphorylation-compromised vimen-tin (21), lenses from VIM SA/SA were stained with E-cadherin, ␤-catenin (the markers of cell-cell boundaries), and DAPI (nucleus). In both anterior (Fig. 3, D and E) and equatorial (Fig.  4, C and D) regions of VIM WT/WT or VIM WT/SA lens epithelium, each cell had one nucleus. In contrast, binucleate (multinucleate) cells were observed in VIM SA/SA lens epithelium. Approximately 13 and 1% of cells each had two nuclei in the anterior (Fig. 3, D and E) and equatorial (Fig. 4, C and D) regions, respectively. FISH analyses using mouse chromosome 12 (green) and 19 (red) probes were used to determine chromosomal instability (CIN) in VIM SA/SA lens. Approximately 3 and 17% of nuclei in the anterior (Fig. 3, F and G) and equatorial (Fig. 4, E and F) regions, respectively, of VIM SA/SA lens epithelial cells had more than two FISH signals per chromosome. Aneuploid cells were not detected in VIM WT/WT or VIM WT/SA lens epithelial cells.
To analyze the mechanism by which CIN occurs in lens epithelial cells, we analyzed primary cultures of lens epithelial cells. VIM SA/SA -derived primary culture cells exhibited not only binucleation (Fig. 5A) but also aneuploidy (Fig. 5B). These phenomena were rarely observed in VIM WT/WT -or VIM WT/SA -derived primary culture cells. The binucleate cells in primary cell cultures derived from VIM SA/SA lenses possessed four centrosomes (Fig. 5A) consistent with cytokinetic failure causing binucleation (multinucleation) just like T24 cultured cells  DECEMBER 13, 2013 • VOLUME 288 • NUMBER 50

JOURNAL OF BIOLOGICAL CHEMISTRY 35629
that express similar phosphorylation-compromised vimentin mutants (21). Does this account for aneuploidy in VIM SA/SA lens epithelial cells? Extra centrosomes have been reported to be sufficient to induce chromosome missegregation (such as lagging chromosomes) during cell division (35). Interestingly, tetraploid cells with four centrosomes showed a high frequency of chromosome missegregation whereas tetraploid cells with two centrosomes decreased to a level observed in diploid cells with two centrosomes (35). These observations led us to propose the following model. Lens epithelial cells in VIM SA/SA mice form an IF bridge, which induces binucleation (multinucleation) due to cytokinetic failure. The binucleation (multinucleation) results in abnormal numbers of centrosomes per cell, which in turn promotes aneuploidy. Given the regional differences in the frequency of cell division rates in the anterior and equatorial regions of the lens, this could explain why there is more binucleation (early event) in the central region and more aneuploidy (later event) in the equatorial zone (Fig. 3 versus Fig. 4).
Premature Aging in VIM SA/SA Lens-How CIN induces microophthalmia and cataract in VIM SA/SA mice is an important question. As judged by TUNEL staining, it is less likely that CIN promotes apoptosis in VIM SA/SA lens. 4 Table 1   tions suggested that lens cataract followed after CIN in VIM SA/SA lens.
To examine lens fiber morphology in detail, lens epithelial cells were costained with anti-AQP0 (also called membrane intrinsic protein (MIP)) and anti-vimentin. Lens fiber cells were disorganized at 4 months as seen by the AQP0 staining (Fig. 6, A  and B) after birth when CIN was observed in all VIM SA/SA mice (also see Table 1). The vimentin network was also disorganized, but no vimentin aggregates were seen (Fig. 6, A and B).
Inter-fiber spaces were also increased in VIM SA/SA lens (Fig.  6, A and B), and in the bow region of the lens, the characteristic distribution of the fiber cell nuclei was disrupted (Figs. 4A and 6B). Vacuoles (area devoid of cellular materials; Fig. 4A) were  also apparent in VIM SA/SA lenses. Such observations are consistent with other mouse models that develop lens cataract (36 -39).
The cross-sectional profiles of fiber cells at the lens equator were analyzed by transmission electron microscopy (40). The profiles of both VIM WT/WT or VIM WT/SA lens fiber cells were very regular maintaining their characteristic hexagonal geometry over 11 months (Fig. 6C). VIM SA/SA lens fiber cells were, however, irregular both in size and shape. Size variation was very apparent as too was the disorganized membrane morphology manifested in, for example, irregular membrane processes (Fig. 6C). In samples from 11-month-old VIM SA/SA mice, the fiber cell boundaries were apparently lost in some areas (see a right lowest photograph in Fig. 6C), which we interpret as a degenerative process (41). Even with the increased resolution of the electron microscope, no electron dense aggregates were observed in the VIM SA/SA lenses. The observed changes in the plasma membranes of the VIM SA/SA lenses resemble those seen in age-related cataracts (41).
Given this comparison, real time PCR analyses to monitor mRNA levels of several senescence (aging)-related genes (p21, Ndrg2, p16 Ink4a , and p19 Arf ) were performed on lens samples. The induction of p21 was recently reported to promote senescence in lens epithelial cells leading to cataract formation (42). N-myc downstream regulated gene 2 (Ndrg2) is considered to be a lens-specific senescent marker (43). Two aging biomarkers of various human and rodent tissues, p16 Ink4a and p19 Arf (44), were also reported to be induced in eye tissue of hypomorphic BubR1 (BubR1 H/H ) progeroid mice (45). The elevation of p21 promotes senescence of lens epithelial cells and cataract formation, whereas the later elevation of p16 Ink4a and p19 Arf is related to age-related tissue deterioration (42).
As shown in Fig. 6D, mRNA expression of p21 or Ndrg2 was significantly elevated in lens of 4-month-old VIM SA/SA mice, compared with the littermate control. 12 months after birth, transcript levels of not only p21 and Ndrg2 but also p16 Ink4a and p19 Arf were significantly higher in VIM SA/SA lens than in VIM WT/WT or VIM WT/SA lens (Fig. 6E). These data suggest that premature cataract formation might represent a progeroid-like lens phenotype in VIM SA/SA mice.

DISCUSSION
In this study, we have analyzed the physiological effects of the expression of a vimentin mutant that will remain unphosphorylated during cell division by replacing with alanine all the ser-  ine residues known to be phosphorylated in mitosis. Binucleate (multinucleate) epithelial cells were observed in the lenses of VIM SA/SA mice (Figs. 3 and 4), mimicking the observations made with T24 cultured cells that expressed phosphorylationcompromised vimentin. Pairs of cells remained connected by an IF bridge, which subsequently induced cytokinetic failure (21). We also found aneuploid lens epithelial cells in VIM SA/SA mice (Figs. 3 and 4), which was unexpected. Perhaps this is due to the presence of additional centrosomes in the binucleated cells (as discussed above). CIN accelerates premature aging, and in VIM SA/SA mice this was manifested as cataracts, the classic age-related phenotype of the eye lens. Congenital cataracts can be caused by vimentin mutations, for example, the E151K (VIM WT/E151K ) mutation in the vimentin rod domain (46). Magin and co-workers (47) also reported lens cataract formation in the transgenic mice with vimentin carrying R113C mutation (VimR113C). The cataract pathology observed for both these mutations in the vimentin rod domain is different from that reported here in several respects. Both are autosomal dominant mutations (46,47). In contrast, no cataract pathology was observed in VIM WT/SA mice. The inheritance is therefore recessive and not dominant in character. Rod domain vimentin mutants also formed protein aggregates, and the endogenous vimentin network was disrupted (46,47). Such consequences were rarely, if at all, observed in both VIM WT/SA and VIM SA/SA mice (Fig. 6). This observation is consistent with our previous finding that SA mutants can normally form vimentin filament networks in interphase cells (25). Finally, the expression of HSP70 was elevated in response to abnormal vimentin structures such as protein aggregates in transgenic VimR113C mice (47), but such a phenomenon was not observed in the lenses of VIM SA/SA mice (Fig. 2E). Taken together, it is less likely that the expression of phosphorylationcompromised vimentin may induce abnormal vimentin filament structure and subsequently lens cataract.
In conclusion, this study documents for the first time the physiological importance of vimentin phosphorylation during cell division in the lens. It also paves the way for future studies evaluating the relationship between CIN and premature aging.