A Missense Mutation in Rev7 Disrupts Formation of Polζ, Impairing Mouse Development and Repair of Genotoxic Agent-induced DNA Lesions*

Background: Rev7 encodes a subunit of Polζ for translesion DNA synthesis (TLS). Results: We found a Rev7 mutation in mice that causes developmental defects and increases susceptibility for genotoxicity. Conclusion: Rev7 is essential for mouse development through its function in cell proliferation. Significance: These findings demonstrate a unique function of Polζ in development that is absent in other TLS polymerases. Repro22 is a mutant mouse produced via N-ethyl-N-nitrosourea-induced mutagenesis that shows sterility with germ cell depletion caused by defective proliferation of primordial germ cells, decreased body weight, and partial lethality during embryonic development. Using a positional cloning strategy, we identified a missense mutation in Rev7/Mad2l2 (Rev7C70R) and confirmed that the mutation is the cause of the defects in repro22 mice through transgenic rescue with normal Rev7. Rev7/Mad2l2 encodes a subunit of DNA polymerase ζ (Polζ), 1 of 10 translesion DNA synthesis polymerases known in mammals. The mutant REV7 did not interact with REV3, the catalytic subunit of Polζ. Rev7C70R/C70R cells showed decreased proliferation, increased apoptosis, and arrest in S phase with extensive γH2AX foci in nuclei that indicated accumulation of DNA damage after treatment with the genotoxic agent mitomycin C. The Rev7C70R mutation does not affect the mitotic spindle assembly checkpoint. These results demonstrated that Rev7 is essential in resolving the replication stalls caused by DNA damage during S phase. We concluded that Rev7 is required for primordial germ cell proliferation and embryonic viability and development through the translesion DNA synthesis activity of Polζ preserving DNA integrity during cell proliferation, which is required in highly proliferating embryonic cells.

The arrest of DNA replication after nucleotide damage disrupts the cell cycle and causes genome instability and cell death. Cells can relieve arrest through translesion DNA synthesis (TLS) 5 that provides tolerance to DNA damage (1,2). In processing damaged DNA through this mechanism, a nucleotide opposite the lesion is incorporated to bypass the damage but at the cost of increased rates of mutation (3,4). Unlike classic DNA polymerases, many TLS DNA polymerases are present in most human cells (5,6). One, DNA polymerase (Pol), is a member of the B family of DNA polymerases and consists of a catalytic and an accessory subunit encoded by Rev3 (official name, Rev3l) and Rev7 (official name, Mad2l2 for mitotic spindle assembly checkpoint protein MAD2B), respectively. Pol has been documented to play a particularly important role in TLS, which is required for DNA repair, recombination, and chromosome stability (1,(7)(8)(9)(10).
In yeast, a strong epistatic relationship occurs between Rev3 and Rev7 mutations, enhancing susceptibility to genotoxic agents. Although these functions are conserved in vertebrates, studies using cultured cells with defects of these genes document greater complexity (11)(12)(13). In addition, a null Rev3 mutation in mouse results in severe early developmental lethality, leaving its function, especially its involvement in various cell lineages, largely unknown (14 -16). A more restricted conditional null mutation of Rev3 in mouse embryonic fibroblast cells and adult B cells documents a need for Pol, without which increased chromosomal aberrations lead to enhanced tumorigenesis (17,18), reduced somatic mutations, and cell proliferation (3,19). Although overexpression of REV7 is associated with high mortality in patients with colon cancer (20), relatively little is known about the role of REV7 and its regulation of Pol function in TLS.
REV7/MAD2L2 is also homologous to the cell cycle checkpoint protein MAD2, and the two proteins share common HORMA domains. The MAD2 protein is a key component of the mitotic spindle assembly checkpoint as an inhibitor of anaphase promoting complex (APC), a surveillance mechanism that delays anaphase until all chromosomes are properly aligned on the metaphase plate (21,22). Several reports have described the function of Mad2l2 in cell cycle regulation, and mammalian epithelial cells infected by Shigella show cell cycle arrest secondary to disruption of Mad2l2 with APC cdh1 (23), which suggests that Rev7/Mad2l2 might regulate the cell cycle by interacting with APC.
Herein, we report a unique function of Rev7 in mouse development and DNA repair using a mouse model of infertility produced by a project known as ReproGenomics at The Jackson Laboratory (24). Repro22 mice are useful for studying the mechanism of mouse primordial germ cells (PGCs), as both males and females are sterile with a lack of PGCs. Interestingly, the mutant mice display severe embryonic lethality and reduced embryonic body weight. Using positional cloning, we identified a missense mutation in the N terminus of Rev7 (C70R) that disrupts Rev7 interaction with Rev3. In the absence of Pol activity, cells become hypersensitive to interstrand cross-link (ICL) damaging agents and accumulate doublestranded breaks (DSBs) in their DNA. We next showed that Rev7 function is essential to resolving the replication stall in S phase of the cell cycle. Our results show that Rev7 is required for PGC proliferation and embryonic viability and development through the TLS activity of Pol to preserve DNA integrity during cell proliferation, which is required in highly proliferating embryonic cells.

EXPERIMENTAL PROCEDURES
Mice and Histological Preparations-The heterozygous repro22/ϩ mice were obtained from The Jackson Laboratory, and the line was maintained through sib mating. The JF1/Ms mice were obtained from the National Institute of Genetics, Mishima, Japan. The repro22/ϩ heterozygous mice were crossed to JF1/Ms mice to generate F 1 mice. The F 1 mice carrying the mutant allele were intercrossed to generate F 2 progeny. Congenic mouse lines were constructed by backcrossing heterozygous mice to C3H mice for 9 generations. Affected and normal mice were classified by the weight of their testes and ovaries. Mice were euthanized, and tissues were excised and fixed by immersion in Bouin's fluid. After dehydration, tissues were embedded in paraffin wax and sectioned. After deparaffination in xylene, sections were rehydrated and stained with Mayer's hematoxylin and eosin. Experimental procedures were approved by the animal care committee of Okayama University.
Alkaline Phosphatase Staining-The concepti used in this study were obtained by mating repro22/ϩ male and female mice. Noon on the day of vaginal plug identification was defined as embryonic day 0.5 (E0.5), and pregnant females were euthanized on days E7.5, E8.5, E9.5, E10.5, E11.5, E12.5, and E13.5. For whole-mount preparations, embryos were removed from decidua in phosphate-buffered saline (PBS) and fixed in 4% paraformaldehyde in PBS for 2 h at 4°C, rinsed in PBS, placed in 70% ethanol for 2-3 h at 4°C, washed with distilled water, and processed as whole mounts for alkaline phosphatase staining (28,29). The embryos were stained at room temperature for 10 min with Fast Red TR salt Sigma; F8764) and -naphthyl phosphate (Sigma; 71090) and ␣-naphthyl phosphate containing 5% borax Sigma; F8764) and 10% MgCl 2 in the staining solution. To obtain serial sections, we fixed the embryos in 4% paraformaldehyde and dehydrated them ethanol and xylene before embedding them in paraffin. The sections (7 m) containing gonads were stained with NBT-BCIP (Boehringer DIG labeling detection kit).
Linkage Mapping and Mutation Detection-Genomic DNA was isolated from the tails of 1511 progeny with phenol/chloroform extraction, and fine mapping of the repro22 locus was performed using polymorphic microsatellite markers on mouse chromosome 4 (Table 1). These markers were genotyped as described previously (25).
For mutation detection, total RNA of the affected and normal animals was extracted from liver, testis, lung, spinal cord, and inner ear tissues using TRIzol reagent (Invitrogen) according to manufacturer instructions. After complementary DNA (cDNA) synthesis using random hexamers and Superscript III reverse transcriptase (Invitrogen), the entire coding regions of five candidate genes were amplified from the cDNAs of affected and normal mice. Primers designed to amplify the open reading frame of Rev7 messenger RNA are listed in Table 1. Purified polymerase chain reaction (PCR) products were cloned into pGEM-T Easy vector (Promega, Madison, WI) and sequenced using the dye-terminator method with an ABI 3100 sequencer (Applied Biosystems, Foster City, CA). The DNA sequences of normal and affected mice were compared. After identifying the missense mutation in Rev7, we designed primers to amplify a 239-bp fragment that contained it (see Table 1) to genotype the mutation with BstUI PCR-restriction fragment length polymorphism (RFLP).
Transgenic Rescue Experiments-To identify conserved genomic regions of Rev7 that may include potential regulatory regions of the gene, we performed a comparative analysis between mice and humans using VISTA (26). The BAC clone RP24 -193K16 containing Rev7 and three other genes was digested using ClaI restriction enzyme to isolate a 24.8-kb fragment containing the entire Rev7 gene including 5 kb of the potentially regulatory 5Ј region and 3 kb of the 3Ј untranslated region of the gene and part of the BAC vector. Transgenic mice were generated via intracytoplasmic sperm injection-mediated transgenesis (27). Oocytes were obtained through mating (C57BL/6J ϫ C3H/HeJ) of F 1 females and C3H/HeJ males and injected with the BAC fragment containing the intact copy of normal Rev7. Transgenic mice were identified with PCR using primers specific to the BAC vector sequence in the fragment (ATCCGGCGCGCCAATAGTCATGCC and CCCTATAGT-GAGTCGTATTAGC) that amplified a 130-bp fragment. The transgenic founders with multiple copies of the normal Rev7 transgene (Rev7 TG ) were crossed with repro22/ϩ females, and the resulting Rev7 TG/ϩ ;repro22/ϩ offspring were crossed to establish Rev7 TG/ϩ ;repro22/repro22 mouse lines. The genotype of the endogenous Rev7 was determined using genotypes of two flanking microsatellite markers (D4Mok8 and D4NIMr1) as well as PCR-RFLP for the mutation site. The presence of Rev7 TG was determined with PCR of the BAC vector sequence and Southern hybridization.
Expression Analysis of Rev7 in Developing PGCs-We examined the expression of Rev7 in PGCs isolated from mouse embryos via fluorescence-activated cell sorting (FACS) of ␤-galactosidase-expressing cells (30). Total RNA was extracted using TRIzol reagent (Invitrogen) and treated with DNase I. The first-strand cDNA was synthesized from the RNA template with reverse transcription using Superscript III reverse transcriptase with oligo-dT primers (Invitrogen). The expression of Rev7 was examined using semiquantitative reverse transcription-PCR.
In Vitro Interaction Assay-cDNA encoding human REV7 and REV3 fragments (residues 1847-1898) was inserted into a pETDuet-1 vector (EMB Millipore, Billerica, MA). The resultant plasmids encoded REV7 with an N-terminal hexameric His tag and the REV3 fragment (31). The C70R substitution of REV7 was introduced via the QuikChange protocol (Agilent, Santa Clara, CA). For the in vitro interaction assay between REV7 and the REV3 fragments (32), His-tagged REV7 or Rev7 C70R mutant protein was co-expressed with the REV3 fragments in Escherichia coli BL21(DE3). The cell lysate was applied to Ni 2ϩ -Sepharose resin (GE Healthcare) in buffer (50 mM HEPES-NaOH, pH 7.4, 500 mM NaCl, and 20 mM imidazole). After washing with buffer (50 mM HEPES-NaOH, pH 7.4, 1.5 M NaCl, and 20 mM imidazole) bound proteins were analyzed using sodium dodecyl sulfate polyacrylamide gel electrophoresis with Coomassie Brilliant Blue staining.
Assay for DNA Damage Sensitivity-Mouse embryonic fibroblasts (MEFs) were isolated from E14.5 embryos obtained from interbred repro22/ϩ mice. After removal of the head and intestinal organs, each embryo was minced and digested with trypsin (PBS, 0.05% trypsin, 0.04% EDTA) at 37°C. The cells were seeded in 100-mm culture dishes (passage 0). MEFs were cultured in 5% CO 2 at 37°C in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum (FBS; Hyclone Laboratories, Logan, UT), and experiments were performed with early passage MEFs. Cytotoxic and cytostatic effects of DNA damage were assayed via examination of the viability rate of MEFs or the proliferation activity in microculture tetrazolium assays using a CellTiter 96 Aqueous One Solution Assay (Promega) according to manufacturer instructions. Briefly, early passage (2-3) MEFs were seeded into 96-multiwell plates (5 ϫ 10 3 cells/well) and cultured in DMEM with 10% FBS. Three hours later, the percent of viable MEFs in the reference wells of each genotype was determined as the 0-h time point. After an additional 24 h, cells were treated with the indicated concentrations of mitomycin C (Sigma), and their proliferation activity was determined 48 h later.
Flow Cytometric Analysis of Mitomycin C-treated Cells-Rev7 C70R/C70R , Rev7 C70R/ϩ , and normal cells (3 ϫ 10 5 cells) were cultured in 10 ml of DMEM containing 10% FBS to adhere in a 10-cm dish. On day 1 of the culture, adhered cells were treated with 1.0 g/ml mitomycin C (Sigma) for 18 h. After washing with PBS 3 times to remove mitomycin C, cells were cultured for an additional 48 h in 10 ml of fresh culture medium. For estimating the frequency of phosphorylated histone H2AX (␥H2AX)-positive cells, MEFs that were cultured for 0 or 48 h after mitomycin C treatment were harvested using 0.025% trypsin, 0.04% EDTA in PBS. These cells were stored in 70% ethanol at Ϫ20°C until staining for FACS analysis.
For flow cytometric analysis, cells were fixed with 3.7% formalin in PBS for 10 min, permeabilized with 0.1% Triton-X, 1% bovine serum albumin (BSA) in PBS for 20 min, and then blocked with 3% BSA in PBS for 30 min at room temperature. The cells were stained with anti-␥H2AX (clone: JBW301, Millipore) followed by goat anti-mouse IgG-Alexa-488 (Invitrogen) in PBS and 3% BSA. After washing in PBS twice, the cells were counterstained with propidium iodide (50 g/ml; Sigma) in PBS containing RNase (10 g/ml). The frequency of ␥H2AXpositive cells in each population was estimated on FACS Calibur (BD Biosciences) and Flowjo software (Tree Star, Ashland, OR) according to Huang and Darzynkiewicz (33).
Spindle Checkpoint Assembly Function Assessment-Rev7 C70R/C70R mutant, Rev7 C70R/ϩ , and normal primary MEFs were treated with colcemid and harvested hourly for 8 h. The cells were fixed with 70% ethanol permeabilized with 1% BSA and 0.1% Triton X-100 in PBS containing 0.1% mg/ml RNase. The cells were then stained with Alexa Fluor 488 antibody recognizing phosphorylated histone H3 (Cell Signaling Technology, Beverly, MA) and stained with propidium iodide as described above. The ratio of phosphohistone H3 (Ser-10)-positive cells to total cells was recorded as the mitotic index.
Terminal Deoxyribonucleotidyltransferase-mediated 2Ј-Deoxyuridine,5Ј-triphosphate-digoxigenin Nick End Labeling (TUNEL) and Proliferating Cell Nuclear Antigen (PCNA) Assay-Rev7 C70R/C70R cells were seeded in a 16-well glass chamber slide at 5 ϫ 10 3 cells/well. After 24 h of seeding, cells were treated with the indicated concentrations of mitomycin C for 18 h. Cells were washed with PBS and further cultured in fresh medium for 48 h. MEFs were then fixed with 3.7% formaldehyde in PBS for 5 min and permeabilized with 0.25% Triton-X in PBS for 5 min at room temperature. Apoptotic cells were detected with TUNEL assay using an In Situ Cell Death Detection kit, POD (Roche Diagnostics). After the TUNEL assay, preparations were counterstained with 4Ј,6-diamidino-2-phenylindole (DAPI) in VECTASHIELD Mounting Medium, H-1200 (Vector Laboratories, Burlingame, CA). A TUNEL assay for histological section was carried out using paraformaldehyde-fixed mouse tissues.
For the PCNA assay, tissue sections were incubated with PBS containing 0.3% H 2 O 2 for 30 min at room temperature. The sections were then incubated with biotin-conjugated anti-PCNA antibody (1:1000) with 3% BSA in PBS overnight at 4°C. The sections were then incubated in series with streptavidin-HRP diluted with 3% BSA in PBS (1:1000) and 3,3Ј-diaminobenzidine tetrahydrochloride (DAB) choromogen (Dako, Kyoto, Japan). The sections were counterstained with Mayer's hematoxylin and examined by light microscopy.
Statistical Analysis-Data are expressed as the means Ϯ S.E. of the mean. Comparisons were made using the two-tailed Stu-dent's test (t test), ⌾ 2 , or analysis of variance and considered highly significant with p values of Յ0.01 and significant with p values of Յ0.05. The p values and significance levels are described in the figure legends.

Growth Retardation and Increased Embryonic Lethality of Homozygous repro22
Mice-During our initial analyses of homozygous repro22 mice, we found decreased size and body weight in adult mice of both genders compared with those of normal littermates. These differences were evident as early as E7.5 and E8.5 and persisted from birth into adulthood, suggesting that the genetic defect in repro22 mice causes severe growth retardation in prenatal development (Fig. 1A). We also found that a significant number of homozygous mutant animals died during embryogenesis. Analysis of 1511 F 2 offspring from intercrosses of repro22/ϩ mice indicated a dramatically reduced number of affected mice. The distortion from Mendelian ratio with no sex ratio distortion suggested an autosomal single recessive mode of inheritance with partial lethality. The prog-eny data indicated that only ϳ40% of affected animals survived after weaning (Fig. 1B), and both male and female mice were strikingly sterile.
Defective Development of PGCs in repro22 Mice-At 8 weeks of age, the male and female gonads of repro22 mice were ϳ10% the size of those of normal mice (Fig. 1C). Examining gonadal histology, we observed complete absences of germ cells in male and female mice. Although testicular tubules were formed, they lacked germ cells, and neither follicles nor germ cells were present in adult ovaries (Fig. 1D). The documented absence of germ cells in adults of both sexes suggested a common defect in prenatal germ cell development. In mice, PGCs first appear in the developing gonad at E7.25 and are defined by their large size and strong alkaline phosphatase activity (28,34). Therefore, we examined the developmental process of PGCs in affected and normal mouse embryos from E7.5 to E13.5 using alkaline phosphatase staining of gonadal sections and whole-mount embryos. The presence of pre-migratory PGCs in the hindgut pocket of affected and normal mice at E8.5 suggested that specification and initial formation of PGCs were not significantly affected in repro22 mutant mice ( Fig. 2A). In addition, PGCs were not significantly decreased in repro22 embryos at E7.25 compared with those in normal embryos ( Fig. 2A), suggesting normal PGC specification in repro22 mice. However, the number of PGCs in the affected mice was approximately half that of normal mice at E8.5, and the differences became more pronounced in the hindgut mesentery at E9.5 and in the genital ridges at E11.5 (Fig. 2B). Notably, although fewer in number, PGCs in repro22 mice did successfully arrive and colonize the genital ridges at E11.5, suggesting that the defect in the mutant mouse is not in migration but in proliferation and survival of PGCs or both ( Fig. 2A).
The widening difference in cell number persisted in E12.5 gonads, and at E13.5, when normal gonads were fully populated, few or no PGCs were detected in the mutant gonad either by alkaline phosphatase staining or PCAM1 immunostaining (Fig. 2C). We observed no differences in testicular cord formation, which was comparable with that of normal male gonads. Taken together, these observations during PGC development suggested that both defects in proliferation and survival of PGCs during PGCs development account for germ cell depletion in affected adult male and female mice. These findings in PGC development as well as growth retardation and partial lethality of the mutant mouse suggested that the mutated gene in repro22 mice has an essential role in proliferation and survival of both somatic and germ cells during development.
Fine Mapping and Positional Cloning of the repro22 Locus-The repro22 mutant locus had been mapped initially to a 13-Mb region between D4Mit57 and D4Mit42 on the distal end of chromosome 4 (The Jackson Laboratory). To refine the location of the mutation, we determined the genotypes of microsatellite markers (Table 1) in this region in 160 affected mice (320 meiotic events) of 1511 F 2 offspring from intercrosses of C57BL/6 and JF1/Ms mice as well as C57BL/6 and C3H/HeJ strains. No recombination between repro22 and D4Mok5 or D4Mok6 was observed, but one to four recombination events between repro22 and the other eight microsatellite markers were present. These data refined the critical region to a 70-kb interval between D4Mok3 and D4Mok7 markers (Fig. 3A), which included a cluster of three F-box genes (Fbox2, Fbox4, and Fbox6), one EST(2610109H07Rik), and Rev7 (Fig. 3 B).
We determined the coding sequence of all of these genes in affected and normal mice by sequencing cDNA derived from various tissues. No functional mutation or nucleotide difference specific to the affected animals was observed in Fbox2, Fbox4, Fbox6, or 2610109H07Rik genes. However, in Rev7, we identified a single nucleotide substitution of T to C at nucleotide 208 in the coding region, which resulted in an amino acid substitution of Cys to Arg at codon 70 (C70R) in a highly conserved region of the HORMA domain, which is implicated in chromatin binding (Fig. 3, C and D). The complete association between the missense mutation and the repro22 phenotype in the F 2 progeny was observed with PCR-RFLP analysis using BstuI endonuclease, and this allele was absent in various inbred strains (Fig. 3E). These findings indicated that the C70R muta-  Transgenic Rescue-To confirm that the Rev7 C70R mutation was responsible for germ cell depletion in Rev7 C70R/C70R mice, we performed genetic complementation by establishing transgenic mice with a genomic fragment containing the normal Rev7 gene. A 24.8-kb genomic fragment containing the entire coding region and the predicted 5Ј regulatory region of Rev7 was isolated from a BAC clone and injected into mouse oocytes via intracytoplasmic sperm injectionmediated transgenesis (Fig. 4, A and B). Mice with the Rev7 transgene were crossed with repro22/ϩ mice and subsequently intercrossed with Rev7 TG/ϩ ;Rev7 C70R/ϩ mice to pro-duce Rev7 TG/ϩ ;Rev7 C70R/C70R transgenic mice (Fig. 4C). Analysis of developing PGCs at E12.5 documented that the genital ridges of the Rev7 TG/ϩ ;Rev7 C70R/C70R mice were as full of alkaline phosphatase-positive PGCs as those in normal control mice (Fig. 4D). Fertility was fully restored in both male and female Rev7 TG/ϩ ;Rev7 C70R/C70R mice, and the size of their testes and ovaries were normal. On histological examination, the seminiferous tubules of the Rev7 TG/ϩ ; Rev7 C70R/C70R mice were filled with germ cells including spermatogonia, spermatocytes, spermatozoa, and mature spermatids, and the ovaries contained multiple follicles with oocytes at various stages of development as well as corpora lutea (Fig. 4E). No histological difference was observed  between the testes or ovaries of normal and Rev7 TG/ϩ ; Rev7 C70R/C70R mice. Therefore, the repro22 phenotypes were completely rescued by Rev7 TG , demonstrating that the loss of function mutation of Rev7 C70R is responsible for germ cell depletion in Rev7 C70R/C70R mice. Expression Analysis of Rev7 in PGCs and Adult Tissues-Based on alkaline phosphatase staining in Rev7 C70R mutant mice, defects in PGCs were first apparent in the gonad at E8.5 and culminated with near complete loss of germ cells by E13.5. We, therefore, assessed the expression of the Rev7 gene during the proliferative stage of PGC development using FACS sorting of ␤-galactosidase-tagged PGCs (30). Rev7 expression in both male and female gonads was detected from E7.5 to E13.5 but not at E14.5 when the proliferative phase of PGCs was complete in male and female mice. Although a single transcript predominated during embryonic development, three additional isoforms of Rev7 were present; particularly at E11.5, two variant transcripts specific to female PGCs were observed (Fig. 5 A). We confirmed that these variants are alternatively spliced isoforms caused by skipping of exons 2 and/or 4 by sequencing of these transcripts.
We also examined expression of Rev7 in various tissues of adult mouse. As shown in Fig. 5B, Rev7 expressed ubiquitously in many tissues and highest expression was observed in testis. It The clone was digested using ClaI restriction enzyme to isolate a 24.8-kb fragment containing the entire Rev7 gene including 5 kb of the potential regulatory 5Ј region and a 3-kb 3Ј untranslated region of the gene. B, comparative analysis of mouse and human Rev7 genomic regions using the VISTA program (26) to identify potential regulatory regions of the gene (pink columns marked with asterisks). C, genotyping of Rev7 TG /ϩ;Rev7 C70R/C70R rescue mice using two microsatellite markers flanking the repo22 region, T208C allele RFLP, and a 130-bp PCR fragment to determine the Rev7 TG allele. D, alkaline phosphatase (ALP) staining of XX gonads at E12.5 in Rev7 TG /ϩ and Rev7 C70R/C70R ;Rev7 TG mice documents complete rescue of PGC proliferation. E, hematoxylineosin staining of testis and ovary from Rev7 TG /ϩ and Rev7 C70R/C70R ;Rev7 TG mice (7 weeks old) confirms that the Rev7 transgenic mouse can fully rescue Rev7 C70R/C70R mice. FEBRUARY 7, 2014 • VOLUME 289 • NUMBER 6

JOURNAL OF BIOLOGICAL CHEMISTRY 3817
is to be noted that the expression of the Rev7 transcripts is also observed in tissues of Rev7 C70R/C70R mice (Fig. 5B), indicating that the C70R mutation does not affect the level of Rev7 expression. The observed elimination of Rev7 expression in the mutant testis could be due to the absence of the germ cells in the mutant testis.
Effects of the Rev7 Mutation on Somatic Cells-The ubiquitous expression of Rev7 suggested that that REV7 functions not only germ cells but also in various types of somatic cells, and the defects of REV7 could result in defective proliferation of somatic cells as observed in the germ cells. We, therefore, examined growth of MEFs isolated from normal, Rev7 C70R/ϩ , and Rev7 C70R/C70R mouse embryos. As shown in Fig. 6A, the growth ratio C70R/C70R MEF was significantly lower than that of normal MEFs at 48 h of culture. These findings indicated that the defect of REV7 affects the proliferation of somatic cells.
Next, we examined in vivo effects of the Rev7 mutation on somatic cells. Because the effects of the Rev7 mutation are expected to be more apparent in rapidly proliferating cells, we assessed ratios of growth and apoptosis in intestinal epithelial cells. As shown in Fig. 6, B and C, the numbers of PCNA-positive S-phase cells were not significantly different between normal and Rev7 C70R/C70R mice. However the number of TUNELpositive cells in Rev7 C70R/C70R mice was significantly increased, and the number of epithelial cells in each villus was significantly reduced in Rev7 C70R/C70R mice. These findings indicated that the Rev7 mutation results in growth defects of intestinal epithelial cells by an increased number of apoptotic cells.
Defects in Spindle Assembly Checkpoint Function Do Not Account for the repro22 Phenotype-REV7/MAD2L2 is homologous to the MAD2 protein, which is a key component of the mitotic spindle assembly checkpoint. Thus, Rev7/Mad2l2 might also regulate cell cycle progression as an arbiter of the spindle assembly checkpoint and could contribute to the growth retardation and germ cell depression observed in Rev C70R mutant mice. To test this possibility, we treated Rev7 C70R/C70R MEF with colcemid and harvested the cells hourly over 8 h. The cells were fixed and stained with phosphohistone H3 (Ser-10) antibody, a marker of mitosis, and the mitotic index was measured for each genotype (11,35). We observed increased cell arrest at mitosis in the presence of colcemid in all genotypes (Fig. 7A) without significant difference. We conclude that that Rev7 C70R is unlikely to be involved in the regulation of the spindle assembly checkpoint.
Rev7 C70R Disrupts Interaction with REV3-Therefore, we focused on the possibility that the Rev7 C70R mutation affects  interactions with REV3, the catalytically active subunit of DNA Pol, and affects TLS activity of the cell. The crystal structure of human REV7 in complex with the REV3 fragment documents that Cys-70 contacts Ile-163, Ile-166, and Trp-171 of REV7, which is adjacent to Pro-1880 of REV3 (Fig. 7B). Thus, the C70R mutation of REV7 would significantly perturb the interaction between Trp-171 of REV7 and Pro-1880 of REV3, which is crucial for REV7-REV3 interaction, and the C70R mutation could disrupt the formation of DNA Pol. To confirm whether C70R disrupts REV7-REV3 interactions, we examined the interaction of REV7 with the REV3 fragment using an in vitro binding assay (Fig. 7B). The mutant REV7 protein with C70R bound little if at all to the REV3 1847-1898 fragment, whereas normal REV7 protein showed a strong signal for binding to the REV3 1847-1898 fragment. Thus, we conclude that the Rev7 C70R mutation significantly impairs REV3 interactions that form the DNA Pol complex, resulting in defective TLS activity.
Rev7 C70R Mutation Sensitizes Embryonic Cells to ICL Agents-The Rev family of proteins is uniquely associated with protection against cisplatin-and mitomycin C-induced DNA damage and repair of ICL in human cells as well as in chicken DT40 cell lines (11,36). We, therefore, investigated whether the Rev7 C70R mutation sensitizes cells to a particular type of DNA damage similar to that observed in siRNA-directed knockdown of REV7 in human (36) or nasopharyngeal carcinoma cells lines (37). Relative cell survival rates of normal, Rev7 C70R/ϩ , and Rev7 C70R/C70R MEFs were measured with microculture tetrazolium assays after treatment with various concentrations of mitomycin C. The cytotoxicity of mitomycin C is ascribed largely to its capability to generate ICLs in DNA (38). Unlike normal controls, Rev7 C70R mutant cells were hypersensitive to ICL-inducing mitomycin C (Fig. 7C). These findings indicate that the disruption of REV7/REV3 interaction caused by the Rev7 C70R mutation increases the susceptibility of cells to agents that cause ICL damage.
Analysis of ␥H2AX and 53BP1 Localization in Response to ICL Agents-Because Rev7 C70R/C70R cells showed significant sensitivity to DNA-damaging agents, we next investigated the efficiency of DNA repair systems in mutant cells under induced DNA damage. DNA damages, including DSB, can be visualized as subnuclear foci of ␥H2AX and/or 53BP1 protein. To determine whether the REV7 mutation induces the formation or persistence of DNA damages, we stained Rev7 C70R/C70R and normal cells with antibodies to ␥H2AX and 53BP1. The ␥H2AX foci were dramatically increased in Rev7 C70R/C70R cells. Cells were classified according to the number of ␥H2AX foci per nucleus, and when uniform staining of the entire nucleus without foci was observed, the cells were classified as displaying pan-nuclear (PN) staining (Fig. 8A). Although Ͻ20% of normal and Rev7 C70R/ϩ cells had PN staining, Ͼ60% of Rev7 C70R/C70R cells had PN staining (Fig. 8B), indicating accumulated DNA damage in the mutant cells after mitomycin C treatment. As a result of increased ratio of PN staining, Ͻ20% of Rev7 C70R/C70R nuclei showed 0 -30 ␥H2AX foci, whereas ϳ50% of normal and Rev7 C70R/ϩ cells showed 0 -30 ␥H2AX foci (Fig. 8B). Interestingly, when the cells were double-stained with antibodies to ␥H2AX and 53BP1, a significant portion of the ␥H2AX foci in nucleus with low density foci co-localized with 53BP1, but the PN-stained cells showed no 53BP1 colocalization (Fig. 8C). Because ␥H2AX is accumulated at the site DNA damages, including DSB and ICL, which cause stalling of replication fork, 53BP1 protein is mainly enriched at sites of DSB (39). Therefore, the PN staining of ␥H2AX FIGURE 7. Rev7 C70R mutation disrupts DNA polymerase/(Pol) formation and sensitizes cells to ICL agents. A, mitotic index in normal, Rev7 C70R/ϩ , and Rev7 C70R/C70R cells at different time points after treatment with colcemid showed no significant differences among the genotypes. B, in vitro interaction assay between REV7 and REV3 fragments. In contrast to normal REV7, the C70R substitution of REV7 impairs REV3 binding. The detailed structure around Cys-70 of REV7 in the REV7-REV3 complex in which REV7 is blue and green (safety belt) and the REV3 fragment is yellow indicates that Cys-70 interacts with Ile-163, Ile-166, and Trp-171 of the safety belt, which are adjacent to Pro-1880 of REV3. The figures were produced by the structural data of REV7-REV3(1847-1898) complex (PDB ID 3ABD) (31). C, reduced growth rate of Rev7 C70R/C70R mutant cells compared with Rev7 C70R/ϩ or normal cells after treatment with mitomycin C (0.25-1.50 m). hRev7-His, human Rev7-His. Values represent the means Ϯ S.E. (n ϭ 6) and were statistically different at p Ͻ 0.02. without 53BP1 focus observed in the present study is suggested to represent replication fork stalling caused by ICL rather than DSB.
Rev7 Is Required for S Phase Repair of ICL Damage-To determine the cell cycle phase that requires DNA Pol, we performed cytometry analysis in normal, Rev7 C70R/ϩ , and Rev7 C70R/C70R cells. MEFs were stained with ␥H2AX antibody and counterstained with propidium iodide to determine the cell cycle stage 16 and 48 h after mitomycin C treatment (33). Control cells from each genotype without mitomycin C treatment had weak or negative signals. However, a striking difference was observed after treatment with mitomycin C. At 16 h, a large number of cells with strong positive signals for ␥H2AX corresponding to PN staining was observed in all genotypes. Fortyeight hours after mitomycin C treatment, the number of strongly positive cells decreased dramatically in normal and Rev7 C70R/ϩ cells, indicating DNA repair and cell cycle resumption. However, Rev7 C70R/C70R cells remained strongly positive for ␥H2AX staining, and the portion of cells in S phase was dramatically increased compared with that in normal cells (Fig.  9, A and B), indicating stalling of the cell cycle at S phase. An increased number of TUNEL-positive cells was also observed in Rev7 C70R/C70R MEFs after mitomycin C treatment (Fig. 9C). Thus, in the absence of functional DNA Pol, Rev7 C70R/C70R cells do not undergo ICL repair, increasing apoptosis after mitomycin C treatment and suggesting hypersensitivity to ICLinducing agents.

DISCUSSION
Homozygous reprod22 mice display growth retardation, germ cell depletion, and sterility. Using positional cloning, we characterized repro22 as a C70R missense mutation in a highly conserved region of the HORMA (DNA binding) domain of the Rev7 gene. Mouse Rev7 has significant homology with yeast rev7, a member of the yeast rev family of genes including rev1, rev3, and rev7. The encoded REV7 protein interacts with REV3 to form DNA Pol, a critical component of TLS (7,40). The structure of human REV7 in complex with a REV3 fragment has been defined (32), and replacement of Cys-70 with Arg introduces an elongated and positively charged side chain predicted to interfere with normal interactions within the REV7 safety belt and prevent binding to REV3. This outcome was confirmed experimentally by the observation that REV3 bound to REV7, but not to Rev7 C70R , when co-expressed in bacteria. The REV7-REV3 interaction is indispensable for REV3 function (32). Thus, the Rev7 C70R mutation is predicted to affect Pol catalytic activity and prevent the TLS activity required to repair DNA damage and necessary to ensure cell cycle progression and proliferation during development. Beyond TLS activity, Rev7 is widely accepted to be a multifunctional protein acting as an adaptor in various biological pathways. Rev7 interacts with various proteins, including the Shigella effector IpaB in bacterial invasion of host cells, and is an activator of APC by interacting with Cdh1, Cdc20, or both (23). We examined whether Rev7 C70R mutation displays defects in spindle assembly checkpoint function by treating the MEFs with colcemid and observed no difference between normal and Rev7 C70R/C70R MEFs. Because colcemid is a microtubule-depolymerizing drug and could trigger the spindle assembly checkpoint, these data indicated no defect related to APC function of REV7. A, levels of ␥H2AX staining were classified based on the number of ␥H2AX foci per nucleus and pan-nuclear staining. B, significant increase of pan-nuclear staining in Rev7 C70R/C70R cells, which display replication-stalled nuclei, after treatment with mitomycin C. C, co-localization of ␥H2AX (green) and the DNA damage response protein 53BP1 (red) in Rev7 C70R/C70R cells. The cells were treated with mitomycin C. Whereas a significant portion of the ␥H2AX foci in nucleus with low density foci co-localize with 53BP1, those with pan-nuclear staining are completely free of 53BP1 staining and represent replication-stalled cells at S phase.
A significant number of DNA lesions occur daily in every cell, and the majority of them stall DNA polymerases, resulting in DNA replication stress. As a curative measure, canonical polymerases can be replaced by TLS polymerases that are adapted to correct replicative defects including those arising as interstrand adducts. This formulation is consistent with our experimental observation that reduction of Pol function in Rev7 C70R/C70R cells leads to increased susceptibility to the DNA-damaging agent mitomycin C, which catalyzes ICL. Therefore, we conclude that the inability of Rev7 C70R/C70R cells to correct DNA lesions resulted in DNA replication stress.
The repair of ICL is not well understood in mammals but involves a recombination-independent pathway in the G 1 phase and a replication-and recombination-dependent pathway in the S phase of the cell cycle (41,42). The Fanconi anemia (FA) pathway plays a pivotal role in resolving ICL at S phase and acts upstream of TLS DNA polymerases (Fig. 8). The FA core complex contains an E3 ubiquitin ligase that is recruited to the DNA lesion. Mono-ubiquitination of FANCD2 and FANCI (FA II) coordinates endonucleolytic cleavage of the blocked replication fork followed by unhooking that leaves the cross-linked nucleotide tethered to the complementary strand that is then bypassed by TLS DNA polymerase (41)(42)(43). Pol, Pol, and Pol have been implicated in these pathways (42), and reconstitution of ICL repair in vitro has documented that depletion of Pol impairs the process in Xenopus laevis egg extracts (44). We found that Rev7 C70R/C70R cells are hypersensitive to ICL agents, displaying increased DSBs and replication stalling in S phase. Involvement of Pol in ICL repair has also been reported in HeLa cells using small interfering RNA technology (36). The processing of ICLs in G 1 phase involves nucleotide excision repair and endonucleolytic cleavage of ICL. In addition to the FA core complex, PCNA becomes mono-ubiquitinated in response to DNA damage, which allows loading of y-family TLS polymerases and bypass of the lesion followed by nucleotide excision repair of the damaged DNA (Fig. 8A) (43,45). Whether Rev7 C70R mutant cells bypass the G 1 phase of ICL repair is unknown. However, because Pol is known as a primary participant in G 1 phase ICL DNA repair (46), Rev7 C70R mutant cells are unlikely to have G 1 phase ICL repair defects caused by a lack of Pol function in replication-independent ICL repair.
Among genetically targeted mice with germ cell depletion, mutations affecting the FA complex proteins FANCL, FANCC, and FANCA show phenotypes similar to those of Rev7 C70R mice, with drastic reduction of germ cells proliferation and infertility (47)(48)(49)(50)(51)(52). In particular, FancL mutant mice (51) and Usp1 ؊/؊ Fancd2 Ϫ/Ϫ double-mutant mice (52) show the most severe phenotypes similar to those observed in Rev7 C70R mutant mice. Unlike Rev7 C70R , mutations in other TLS polymerases do not affect PGC development (53)(54)(55)(56)(57)(58). In addition, similar to cells with the Rev7 C70R mutation, cells derived from FA and Usp1 null mutants are sensitive to ICL-inducing agents. Our data show that the lack of REV7 function results in accumulation of ␥H2AX without 53BP1 colocalization at S phase, which could represent unresolved replication-fork stalling caused by DNA damages (39). These findings indicated that S phase ICL repair requires Rev7 and suggest that DNA Pol functions downstream of the FA pathway. The more severe phenotypes of Rev7 C70R mutant mice show that although some redundancy might occur in the FA and Usp1 families of genes, the TLS polymerases downstream of these pathways are restricted to control by Pol. Interestingly, other TLS polymerases, including Pol, the primary participant in G 1 phase ICL DNA repair, showed neither defects related to fertility nor developmental defects.
A recently reported targeted mutation of Rev7 (Rev7 KO ) (59) shows phenotypes similar to those of Repro22 mice, with growth retardation and embryonic loss of germ cells. That report and our own show lost PGCs during development from E8.5 to E13.5, reduced body size, and embryonic lethality. Moving beyond the common phenotypes observed in both cases and taking advantage of a point mutation in Rev7, we performed additional functional analysis of Rev7, the findings of which are summarized as below.
We studied formation of PGCs at E7.5 and observed that it is unaffected by Rev7 C70R mutation, which demonstrates that Rev7 is not involved in PGC formation. In addition, a previous study (59) could not show any clear expression of Rev7 in PGCs during development. Instead, we isolated PGCs from both male and female embryos during development from E7.5to E14.5 FIGURE 10. Model describing requirement of Rev7 in ICL repair. A, processing of ICLs in the G 1 phase of the cell cycle involves nucleotide excision repair (NER) and endonucleolytic cleavage of ICL. PCNA becomes mono-ubiquitinated by E2 ubiquitin ligases in response to DNA damage, which allows loading of TLS polymerases and bypass of the lesion followed by nucleotide excision repair of the damaged DNA. B, in the S phase of the cell cycle, ICL repair requires an upstream FA pathway coupled to Pol (left panel). FA core proteins are recruited by FANCM (M) to the ICL (red bar), which monoubiquitinates the FA-II complex followed by DNA incision and unhooking of the cross-linked DNA. On the incised strand, TLS polymerase is recruited to bypass the unhooked cross-link followed by nucleotide excision repair. However, the Rev7 C70R mutation disrupts formation of Pol, leading to incomplete DNA replication and persistence of S phase-stalled cells with double-strand breaks (right panel). Thus, Pol is the TLS polymerase downstream of the FA pathway required to resolve ICL and relieve DNA replication arrest in S phase. and clearly show that Rev7 is expressed during that period and that the expression was absent on E14.5 coincident with a well known PGC proliferation stage.
PGC development has three important steps, namely formation, migration, and proliferation. We found that the Rev7 C70R/ C70R cells are proliferation-deficient through three pieces of evidence obtained in the present study. First, accumulated cells arrested at S phase of the cell cycle and strongly expressed ␥H2AX and displayed widespread apoptosis. Second, although Rev7 is expressed ubiquitously in various tissues, quiescent cells are not affected by the Rev7 mutation; it affects mainly proliferative cells during embryonic development. Third, the expression pattern in PGC development is consistent with the well known period of proliferation between E7.5 and E13.5. The expression ceased on E14.5, when the proliferation of PGCs stopped in both male and female mice. Rather than the proposed defect in germ cell maintenance (34), our results with Rev7 C70R mice indicate a defect in proliferation secondary to absent TLS activity. Consistent with our results, conditional targeting of Rev3 in skin epithelia (60) does not affect survival but is essential for responses requiring rapid proliferation of epithelial cells. Also notable is that the present findings indicated decreased growth ratio of the Rev7 C70R MEFs and increased apoptosis in Rev7 C70R intestinal epithelial cells, which are among the most rapidly proliferating cells. These findings also support our hypothesis that the rev7 mutation mainly affects rapidly proliferating cells and accounts for growth retardation and increased embryonic lethality observed in Rev7 C70R/C70R mice.
The point mutation of the Rev7 C70R disrupts its interaction with Rev3, therefore, impairing Pol formation. We next excluded the possibility that this point mutation might affect another proposed function of Rev7 as an APC activator. We found that Rev7 C70R mutant cells are hypersensitive to ICL agents and that ICL repair is abrogated at S phase in Rev7 C70R/C70R cells, suggesting that ICL repair at this stage requires Rev7 and Pol. Interestingly, the phenotypes in repro22 mice were similar to those of mice with mutations in the FA pathway. This pathway is well known for its function in replication-dependent ICL repair during S phase, which acts upstream of a TLS polymerase. We, therefore, show that the TLS polymerase required for replication-dependent ICL repair is Pol.
Using the results of the present study, we proposed a model (Fig. 10) in which Pol is required for ICL repair at S phase of the cell cycle and acts in concert and downstream of the FA pathway. Rev7 C70R/C70R mice will provide an excellent model for investigation of the function of Pol in genome stability in both somatic and germ cells.