Minichromosome Maintenance 2 Bound with Retroviral Gp70 Is Localized to Cytoplasm and Enhances DNA-Damage-Induced Apoptosis

The interaction of viral proteins with host-cellular proteins elicits the activation of cellular signal transduction pathways and possibly leads to viral pathogenesis as well as cellular biological events. Apoptotic signals induced by DNA-damage are remarkably up-regulated by Friend leukemia virus (FLV) exclusively in C3H hosts; however, the mechanisms underlying the apoptosis enhancement and host-specificity are unknown. Here, we show that C3H mouse-derived hematopoietic cells originally express higher levels of the minichromosome maintenance (MCM) 2 protein than BALB/c- or C57BL/6-deriverd cells, and undergo more frequent apoptosis following doxorubicin-induced DNA-damage in the presence of the FLV envelope protein gp70. Dual transfection with gp70/Mcm2 reproduced doxorubicin-induced apoptosis even in BALB/c-derived 3T3 cells. Immunoprecipitation assays using various deletion mutants of MCM2 revealed that gp70 bound to the nuclear localization signal (NLS) 1 (amino acids 18–24) of MCM2, interfered with the function of NLS2 (amino acids 132–152), and suppressed the normal nuclear-import of MCM2. Cytoplasmic MCM2 reduced the activity of protein phosphatase 2A (PP2A) leading to the subsequent hyperphosphorylation of DNA-dependent protein kinase (DNA-PK). Phosphorylated DNA-PK exhibited elevated kinase activity to phosphorylate P53, thereby up-regulating p53-dependent apoptosis. An apoptosis-enhancing domain was identified in the C-terminal portion (amino acids 703–904) of MCM2. Furthermore, simultaneous treatment with FLV and doxorubicin extended the survival of SCID mice bearing 8047 leukemia cells expressing high levels of MCM2. Thus, depending on its subcellular localization, MCM2 plays different roles. It participates in DNA replication in the nucleus as shown previously, and enhances apoptosis in the cytoplasm.

Viral infections are known to modify cellular processes related to DNA-damage responses or DNA synthesis [14][15][16]. We have previously shown that Friend leukemia virus (FLV) infection markedly enhances the IR-induced apoptosis of hematopoietic cells in C3H mice via P53, ATM, and DNA-PK [17]. Mice infected with FLV and then treated with a low dose of total body irradiation (TBI) exhibit severe anemia. However, p53 knockout mice, Atm knockout mice, and DNA-PK-deficient SCID mice with a C3H background do not exhibit this phenotype. A comparison of the apoptotic signals after FLV infection, TBI, or FLV+TBI treatment of these mice revealed that ATM is necessary for the general signal transduction of TBI-induced apoptosis [18], while DNA-PK plays a specific role in enhancing p53-dependent apoptosis following FLV infection [19,20].
The enhancement of p53-dependent apoptosis occurs almost exclusively in the C3H strain of mice [21]. In relation to this hostspecific apoptosis-enhancement, we have previously demonstrated that the FLV-derived envelope protein gp70 enhances cellular apoptotic signaling in association with host-specific overexpressed proteins, including the minichromosome maintenance (MCM) 2 protein, resulting in the activation of DNA-PK, which phosphorylates P53 [22]. MCM2 is one of a set of 6 proteins (MCM complex; MCM2-7) that play essential roles in DNA replication [23]. The MCM complex associates with the origins of DNA replication to form part of the pre-replicative complex (preRC) [24]. Activation of the MCM complex by cyclin-dependent kinases leads to the initiation of DNA synthesis and MCM proteins also act as a replicative helicase to unwind DNA at replication forks during DNA synthesis [25,26]. The MCM complex contains a nuclear localization signal (NLS) and a nuclear export signal (NES) [27]. The NLS is split between MCM2 and MCM3 and the NES is located in MCM3 adjacent to the NLS sequence. The transport of all MCM proteins is interdependent, suggesting that nuclear import requires the formation of the hexameric complex, which would result in the assembly of a complete NLS [28,29]. MCM proteins are expressed in cycling cells but are down-regulated and dissociated from the chromatin in quiescent cells [30]. Thus, detection of MCM proteins has emerged as a method for evaluating the proliferative state and growth fraction in dynamic cell populations. Indeed, elevated expression of several members of the MCM complex has been reported in various malignant tumors [31,32]. Furthermore, studies with human samples have indicated the utility of MCM2 as a proliferation marker, and a high level of MCM2 expression in malignant tumors has been associated with several clinicopathological parameters, such as advanced tumor grade, advanced stage, and poor prognosis [33][34][35][36]. Thus, MCM2 usually acts to support cellular proliferation. However, as described above, MCM2 enhances TBI-induced apoptosis in the presence of gp70. To determine importance of such contradictory functions of the MCM2 protein in the regulation of cellular dynamics, the molecular mechanisms underlying MCM2-induced apoptosis and MCM2-gp70 interaction need to be elucidated. An understanding of the overall functions of MCM2 would enable the molecular targeting of specific functions possibly to regulate cellular proliferation/apoptosis in a cell type-specific manner and develop a novel strategy to control tumor cell growth.

Doxorubicin-induced Apoptosis of FLV-infected Cells Correlates with High Levels of Mcm2 in Vivo
In previous studies, TBI caused prominent apoptosis in the bone marrow cells of FLV-infected C3H mice, but not FLV-infected BALB/c and C57BL/6 mice [17]. From a therapeutic perspective, systemic distribution of the effects of DNA-damage would be more easily achieved by chemical agents than IR. Therefore, to determine whether DNA-damaging agents enhanced apoptosis to similar extents in FLV-infected mice of different strains, uninfected or FLV-infected BALB/c, C57BL/6, and C3H mice were intraperitoneally administered with a low dose of doxorubicin or PBS, and the apoptotic cell ratio was measured in the bone marrow and spleen. In FLV-infected BALB/c and C57BL/6 mice, the apoptotic cell ratios after treatment with doxorubicin were similar to the ratios in uninfected mice ( Figure 1A, B). On the other hand, FLV-infected doxorubicin-treated C3H mice exhibited significantly higher ratios with uninfected mice ( Figure 1C). Thus, we could generalize as to the effects of DNA-damage by IR and chemical agents on the enhancement of apoptosis by FLVinfection in hematopoietic organs.
Next, we examined the expression of Mcm2 mRNA in the bone marrow and spleen of uninfected and FLV-infected BALB/c, C57BL/6, and C3H mice. Mcm2 levels were significantly higher in the bone marrow cells of C3H mice than in BALB/c and C57BL/ 6 mice ( Figure 1D). Spleen Mcm2 levels were also higher in C3H mice than in BALB/c and C57BL/6 mice. Furthermore, in C3H mice, the spleen Mcm2 levels were elevated by FLV-infection ( Figure 1E). Similar trends were observed across all the inbred strains tested. These results suggest that doxorubicin treatment induces significant apoptosis in FLV-infected C3H mice in association with higher levels of Mcm2. Moreover, we performed a comparative GeneChip analysis using RNA isolates from mouse spleen and identified several genes that exhibited various expression patterns in the different mouse strains ( Figure 1F-L).
Mcm2 expression was higher in C3H mice than in C57BL/6 mice, and Mcm2 expression was elevated by FLV-infection ( Figure 1G). Genes that exhibited expression patterns similar to that of Mcm2 are listed in Table S1.
Dual Transfection with Mcm2/gp70 Enhances DNAdamage-induced Apoptosis in BALB/c-derived 3T3 Cells To investigate whether apoptosis enhancement was related to the high levels of Mcm2 in FLV-infected cells, we analyzed doxorubicin-induced apoptosis sensitivity in Mcm2 and/or gp70transfected 3T3 cells. First, the expression of Mcm2 was analyzed in each mouse cell line. BALB/c-derived 3T3 cells and primary cultured BALB/c-fibroblasts expressed low levels of Mcm2 compared to C3H-derived 8047 cells, 32D cells and primary cultured C3H-fibroblasts ( Figure 2A).
Next, the viability and apoptotic cell ratios of 3T3 cells were evaluated after doxorubicin treatment. Gp70 plus Mcm2-transfected 3T3 cells exhibited a significant decrease in viability and an increase in apoptotic cell ratio compared to control cells, whereas cells transfected with gp70 or Mcm2 exhibited no significant change in viability and apoptotic cell ratio ( Figure 2B, C). Gp70 and/or MCM2 protein levels following gp70and/or Mcm2-transfection were similar in all the experimental groups ( Figure 2D). Next, we knocked down the expression of Mcm2 in BaF3 and 32D cells using siRNA. The 32D cell line, with a high level of endogenous gp70 expression, was established from FLV-infected C3H mouse bone marrow [37] ( Figure 2E). Mcm2 knockdown significantly reduced Mcm2 mRNA expression and apoptotic cell ratio of 32D cells treated with doxorubicin in contrast to the non-remarkable change in the apoptotic cell ratio of BaF3 cells ( Figure 2F). These results suggest that the host-specific enhancement of DNA-damageinduced apoptosis is associated with the higher level of Mcm2 expression in C3H-derived cells.

Gp70 Directly Binds to the N-terminal Portion of MCM2
To examine the molecular interactions between MCM2 and gp70, immunoprecipitation experiments were performed. We generated plasmids encoding HA-tagged full-length MCM2 (MCM2-FL) and various deletion mutants: MCM2-DC, MCM2-DN, MCM2-N and MCM2-C ( Figure 3A). Each of these plasmids was transfected into 3T3 cells along with FLAG-tagged gp70. Irrespective of doxorubicin treatment, gp70 interacted with MCM2-FL, MCM2-DC, and MCM2-N, but not with MCM2-DN or MCM2-C ( Figure 3B, C). These results indicate that gp70 associates with the N-terminal portion of MCM2. Gp70 binding inhibited the formation of the MCM complex ( Figure S1). As shown in Figure 3B and 3C, the size of MCM2-N was larger than the expected size. Generally, phosphorylated proteins are sometimes larger than their unphosphorylated counterparts [38,39]. Indeed, the N-terminal portion of MCM2 possesses many phosphorylation sites [40]. Therefore, the apparent molecular weight of MCM2-N may be higher than expected. Further, MCM2-C does not have as many phosphorylation sites [40]. As a result, MCM2-N may appear larger than MCM2-C.
We also generated plasmids encoding a FLAG-tagged gp70 deletion mutant ( Figure S2A) and performed a similar pull-down assay after co-transfection with HA-tagged Mcm2-FL. MCM2 bound to the middle portion of gp70 ( Figure S2B, C) and enhanced apoptosis in response to doxorubicin ( Figure S2D, E). The C-terminal Portion of MCM2 is Essential for the Enhancement of Doxorubicin-induced Apoptosis Next, to identify the functional domain of MCM2 essential for apoptosis enhancement following DNA-damage, a functional analysis was performed using MCM2 deletion mutants. First, Mcm2-FL or the deletion mutant were introduced into 3T3 cells with or without gp70. After the transfection, 3T3 cells were treated with doxorubicin, and cell viability and apoptotic cell ratio were measured. 3T3 cells, transfected with gp70 and the Mcm2-FL exhibited a significant decrease in viability and an increase in apoptotic cell ratio compared to cells transfected with the negative control ( Figure 4A, B). Surprisingly, cells transfected with gp70 and Mcm2-DNor Mcm2-C, which did not interact with gp70, also exhibited a significant decrease in viability and an increase in apoptotic cell ratio relative to the negative control ( Figure 4A Previous studies have shown that MCM2 is essential for DNA replication [23,25], and its expression is up-regulated in proliferating cells [41]. Mcm2-transfected 3T3 cells exhibited no significant change in cell count during the early stage ( Figure  S3A, B). However, at a later-stage (96 h), the cell count was significantly higher in Mcm2-transfected 3T3 cells than in the control ( Figure S3C, D).
We next examined the protein levels of DNA-PK, phospho-DNA-PK (pS2053), P53, phospho-P53, and cleaved caspase-3 in  Figure 4F). These results indicate that not only the binding of MCM2 with gp70 but also deletion of the N-terminal portion enhances DNA-damage-induced apoptosis via the activation of P53 by DNA-PK. Furthermore, MCM2 lacking the C-terminal portion did not induce apoptosis even with gp70 coexpression indicating that the C-terminal portion of MCM2 was essential for the enhancement of DNA-damage-induced apoptosis.
DNA-PK is robustly activated by auto-phosphorylation at Ser 2056 (S2053 in mouse) in apoptotic cells [42], while phosphorylation at Thr 2609 is associated with non-homologous end joining [43]. Therefore, to examine whether DNA-PK was exclusively required for the enhancement of apoptosis, we inhibited DNA-PK activity using NU7026 in the presence ( Figure 4G) or absence of gp70 ( Figure 4H). Inhibition of DNA-PK activity by NU7026 substantially reduced the level of phospho-DNA-PK (pS2053) and completely abolished apoptosis enhancement in cells expressing the Mcm2 mutants ( Figure 4G, H). These results and knockdown experiments ( Figure S4) indicate that DNA-PK activation is necessary for the enhancement of doxorubicin-induced apoptosis.
The Gp70-MCM2 Complex Binds to PP2A and Causes Hyperphosphorylation of DNA-PK To determine the mechanism by which the gp70-MCM2 complex activated DNA-PK to enhance apoptosis, we sought to identify the upstream regulatory factors of DNA-PK. We focused on protein phosphatase 2A (PP2A), because this molecule has been shown to dephosphorylate DNA-PK and control its function [44][45][46]. 3T3 cells were transfected with Mcm2-FL or Mcm2 deletion mutants with or without gp70 and treated with doxorubicin. In the absence of gp70, PP2A did not interact with MCM2-FL or the mutants ( Figure 5A, left). In gp70-transfected cells, PP2A coprecipitated with MCM2-FL, MCM2-DN, and MCM2-C, but not with MCM2-DC or MCM2-N ( Figure 5A, right). Thus, PP2A interact with the C-terminal portion of MCM2 in gp70-transfected 3T3 cells.
To determine whether the enhanced apoptosis was caused by the inactivation of PP2A, the PP2A-specific inhibitor okadaic acid (OA) was added to 3T3 cells that were treated with doxorubicin. As expected, the OA-treated 3T3 cells exhibited a significant increase in apoptotic cell ratio compared to the control ( Figure 5B). Furthermore, NU7026 treatment abrogated the doxorubicininduced apoptosis enhancement in OA-treated 3T3 cells ( Figure 5B). The expression of phospho-DNA-PK (pS2053) was upregulated in OA-treated 3T3 cells after doxorubicin treatment ( Figure 5C). These results suggest that the gp70-MCM2 complex binds to and inhibits PP2A. Consequently, DNA-PK is hyperphosphorylated and doxorubicin-induced apoptosis is enhanced via the P53/cleaved caspase-3 pathway.
Data represent the mean and 95% confidence intervals (CI) of 3 independent experiments. (D) Quantitative RT-PCR analysis of Mcm2 mRNA expression in the bone marrow of uninfected and FLV-infected BALB/c, C57BL/6, and C3H mice. The bone marrow cells of the C3H strain exhibit higher levels of Mcm2 in all groups compared to the corresponding groups of BALB/c and C57BL/6 mice (*p,0.01, for each group). (E) Quantitative RT-PCR analysis of Mcm2 mRNA expression in the spleen of uninfected and FLV-infected BALB/c, C57BL/6, and C3H mice. Spleen Mcm2 expression is higher in the ''FLV (+), Doxorubicin (2)'' and ''FLV (+), Doxorubicin (+)'' C3H mice than in the corresponding groups of BALB/c and C57BL/6 mice (*p,0.01 and # p,0.01, respectively). In C3H mice, FLV-infection induces higher levels of Mcm2 expression compared to the expression in uninfected mice. Data represent the mean and 95% CI from 5 mice in each group and are representative of 2 independent experiments. The GeneChip data for Mcm-associated and apoptosis-associated genes were analyzed using the Percellome method. Forty-eight male C57BL/6 and C3H mice were divided into 16 groups of 3 mice each. Uninfected or FLV-infected C57BL/6 and C3H mice were administered (i.p.) with 15 mg/kg (high dose) or 1.5 mg/kg (low dose) of doxorubicin, and the spleen was sampled 0, 1, 6, and 24 h after administration. The spleen transcriptome was measured using the Affymetrix Mouse 430-2 GeneChip. (F) The Percellome data were plotted on 3-dimensional graphs for average, +1 SD, and 21 SD surfaces as demonstrated in the left schema. The scale of expression (vertical axis) is the copy number per cell. The x-axis of the 3-dimensional graph shows the experimental groups, including the C3H and C57BL/6 mice with doxorubicin treatment (high and low doses) with or without FLV-infection. The y-axis shows the time course (0, 1, 6, and 24 h) after treatment with doxorubicin and the z-axis (vertical) indicates the intensity of mRNA expression of each gene. The data of each point are connected to form a surface illustration. The expression patterns of genes are compared using the surface images.   The gp70-MCM2 Complex is Localized in the Cytoplasm The MCM2 protein contains an NLS in the N-terminal portion. Thus, MCM2 localizes to the nucleus when expressed in HeLa cells [47]. To investigate the cellular localization of MCM2, immunofluorescence was performed on 3T3 cells transfected with Mcm2-FL or mutated Mcm2, with or without gp70 and treated with doxorubicin. In 3T3 cells singly transfected with Mcm2-FL or the mutants, MCM2-FL as well as MCM2-DC and MCM2-N were localized in the nucleus ( Figure 6A). By contrast, MCM2-DN and MCM2-C lacking the NLS were localized in the cytoplasm ( Figure 6A). In cells transfected with gp70 plus Mcm2-FL or gp70 plus mutated Mcm2, MCM2-FL and all the MCM2 deletion mutants were detected in the cytoplasm ( Figure 6B). These results indicate that gp70 binding inhibits the nuclear translocation of MCM2 and show that MCM2 lacking an NLS remains in the cytoplasm. We confirmed that overexpression of gp70 and/or MCM2-FL or the mutants did not cause any significant changes in the cell-cycle profile of the transfected cells ( Figure S5). Furthermore, the transfected gp70 induced the cytoplasmic localization of DNA-PK as well as MCM2 ( Figure S6).
MCM2 has 2 NLS domains, NLS1 and NLS2. NLS2 but not NLS1 is required for the nuclear localization of mouse MCM2 [47]. Thus, to further examine the gp70-mediated inhibition of MCM2 nuclear translocation, we generated plasmids encoding HA-tagged MCM2 NLS deletions; deletion of NLS1 (MCM2-DNLS1), deletion of NLS2 (MCM2-DNLS1), and deletion of NLS1 to NLS2 (MCM2-DNLS1-2) ( Figure 6C). 3T3 cells were transfected with these mutants and treated with doxorubicin, and apoptotic cell ratios were determined. The ratio was significantly increased in Mcm2-DNLS2and Mcm2-DNLS1-2-transfected cells compared to the negative control. By contrast, Mcm2-DNLS1transfected cells exhibited no increase in the number of apoptotic cells ( Figure 6D). Furthermore, MCM2-DNLS1 was localized in the nucleus, whereas MCM2-DNLS2 and MCM2-DNLS1-2 were detected in the cytoplasm ( Figure S7). These results indicate that deletion of NLS2 alters the subcellular localization of MCM2 and the apoptosis enhancement seen in the presence of the gp70-MCM2.

Induction of Leukemia cell Apoptosis by DNA-damage in FLV-infected Hosts
To determine whether C3H-derived leukemia cells exhibited enhanced apoptosis in response to gp70 and DNA-damage in vivo, SCID mice were intravenously transplanted with 8047 cells, inoculated with FLV, and treated with doxorubicin. As expected, the 8047 cell-containing liver samples from FLV-infected mice exhibited a stronger expression of gp70 than those from uninfected mice ( Figure 7A). Treatment with a low dose of doxorubicin caused significant enhancement of apoptosis in FLV-infected SCID mice but not in uninfected mice ( Figure 7B, C). These results indicate that gp70 overexpression and DNA-damage induction elicit significant apoptosis of C3H-derived leukemia cells in vivo.
Next, to investigate the subcellular localization of MCM2 in the transplanted 8047 cells from hepatic nodules, immunohistochem- istry was performed. MCM2 was localized in the nucleus of 8047 cells in uninfected SCID mice ( Figure 7D, top), whereas some 8047 cells exhibited cytoplasmic MCM2 in the FLV-infected mice ( Figure 7D, bottom). Furthermore, the number of cells with cytoplasmic MCM2 was remarkably increased in FLV-infected doxorubicin-treated mice compared to FLV-infected PBS-treated mice ( Figure 7D, bottom right and E).
A survival analysis was performed on mice treated with PBS or doxorubicin twice a week. FLV-infected and doxorubicin-treated mice exhibited a significant improvement in survival compared to the other groups ( Figure 7F). These results suggest significant effects of cytoplasmic MCM2 on apoptosis induction in leukemia cells in the in vivo model. Although not so remarkable, FLVinfection alone prolonged the survival of 8047 cell-transplanted mice. The phenomenon may be caused by intrinsic host defense mechanisms such as innate immunity systems and inflammatory reactions by natural killer cells, neutrophils, monocyte/macrophages etc., against leukemia cells. The reactions may include reactive oxygen species or other stress signaling pathways associated with DNA-damage induction. Thus, the circulating leukemia cells may differ from the leukemia cells used in vitro experiments without any stimulation for DNA-damage.

Discussion
A novel strategy for controlling tumor cell growth is to target regulators of cellular proliferation/apoptosis. However, the cellular dynamics of non-tumor cells should not be influenced by these treatments. This is very difficult, but infection with certain types of viruses elicits tumor cell-specific changes in cellular dynamics [48]. Thus, virus-host interaction may provide clues to develop a novel strategy for tumor therapy. Our previous study has shown that FLV infection strongly enhances radiation-induced apoptosis in the hematopoietic cells of C3H mice, although the response is not uniform among the host strains [17]. Elucidation of the molecular mechanisms underlying this host-and cell typespecificity may provide an effective means to induce tumor cellspecific apoptosis in host tissues.
Regarding host specificity, MCM2 was identified as a C3Hspecific protein that enhances DNA-damage-induced apoptosis in association with the envelope protein of FLV, gp70. However, MCM2 is part of a conserved set of MCM proteins (MCM2-7), with essential roles in the regulation of DNA replication: functioning as license components for S-phase initiation and further acting as a helicase to unwind DNA at replication forks [25,26,49]. Indeed, MCM proteins are frequently overexpressed in a variety of cancer or pre-cancerous cells [31][32][33][34][35][36]. In this study, MCM2 has several functional domains [50]. However, there are no reports on its functions in apoptosis. Our study demonstrated that a novel functional domain in the C-terminal portion of MCM2 plays a role in apoptosis enhancement under specific conditions in conjunction with gp70 ( Figure 8A).
MCM2 is known to interact with various types of molecules, including protein PP2A [51]. PP2A is one of the major Ser/Thr phosphatases implicated in the regulation of cellular processes such as cell cycle progression [52], apoptotic cell death [53][54][55], and DNA replication and DSB repair [45,52,53]. In the GeneChip assay of the present study, Ppp2ac exhibited an expression pattern similar to that of Mcm2 in the in vivo experiments (correlation coefficient .90%; Figure 1L, Table S1). Furthermore, our results suggest that PP2A dephosphorylates DNA-PK and regulates its function, as described previously [44][45][46]. Depletion of PP2A by RNAi has been shown to induce hyperphosphorylation of DNA-PK and suppression of DNA end-joining followed by enhanced cytogenetic abnormalities including chromosomal and chromatid  The MCM complex (MCM2-7) contains an NLS. MCM2 has 2 NLS domains and histone-binding sites in the N-terminal portion, and therefore deletion of the N-terminal portion resulted in the inhibition of nuclear translocation. NLS2 but not NLS1 is required for the nuclear localization of mouse MCM2 [47]. In the present study, nuclear translocation of MCM2 was inhibited by the binding of gp70 to NLS1, and that the cytoplasmic MCM2 enhanced DNA-damage-induced apoptosis.
In conclusion, we identified a novel function of MCM2: the enhancement of DNA-damage-induced apoptosis. This function occurred in association with gp70, an FLV-derived envelope protein. Gp70 directly bound to the N-terminal portion of MCM2 and inhibited its translocation. The cytoplasmic MCM2-gp70complex induced an interaction of MCM2 with PP2A, thereby interfering with the PP2A-DNA-PK interaction and leading to enhanced DNA-damage-induced apoptosis via the activation of P53 by DNA-PK ( Figure 8B). These results suggest that regulation of the molecular dynamics of MCM2 may be a novel apoptosisinducing therapeutic method to specifically target malignant tumors that express higher levels of MCM2 than normal tissues.

Ethics Statement
Animal experiments were conducted and carried out in strict accordance with the Act on Welfare and Management of Animals of the government of Japan and the Guidelines for the Care and Use of Laboratory Animals of the Tokyo Medical and Dental University. All experiments were approved by the Animal Experiment Committee of the Tokyo Medical and Dental University (No. 100115). All efforts were made to minimize suffering in animal experiments.

Mice and Cell Lines
Eight to 10-week-old male C3H/HeJ mice (H22 k ) raised under specific-pathogen-free conditions were purchased from Japan SLC, Inc. (Shizuoka, Japan) with the permission of Dr. Yoshiya Shimada of the National Institute of Radiological Sciences in Chiba. Specific-pathogen-free C57BL/6J mice (H22 b ) and BALB/c mice (H22 d ) aged 8-10 weeks were also purchased from Japan SLC, Inc. Six-week-old male specific-pathogen-free SCID mice (C.B.17 scid/scid , H22 d ) were purchased from CLEA Japan Inc. (Tokyo, Japan).
The mouse fibroblast cell line 3T3 and the mouse acute myeloid leukemia cell line, BaF3, both derived from BALB/c mice, and the C3H mouse bone marrow cell-derived 32D cells were purchased from the RIKEN Cell Bank (Tsukuba, Ibaraki, Japan). The radiation-induced myeloid leukemia cell line from C3H mice, 8047, was established at the National Institute of Radiological Sciences in Chiba [22]. The cells were cultured in RPMI-1640 medium (Sigma, St. Louis, MO, USA). Primary cultured fibroblasts were derived from the lungs of BALB/c and C3H mice and cultured in DMEM (Sigma). The medium was supplemented with 10% fetal calf serum (FCS), penicillin (50 units/mL) (Invitrogen, Carlsbad, CA, USA), and streptomycin (50 mg/mL) (Invitrogen) and the cells were cultured at 37uC in a humidified atmosphere of 5% CO 2 in air.

Viral Infection and DNA-damage Induction
The NB-tropic FLV complex, originally provided by Dr. C. Friend, was prepared as described previously [56]. Eight-to 10week-old BALB/c, C57BL/6, and C3H mice were inoculated intraperitoneally (i.p.) with FLV at a highly leukemogenic dose of 10 4 PFU/mouse [57]. On day 7 after the infection with FLV, BALB/c, C57BL/6, and C3H mice were administered (i.p.) with 1.5 mg/kg of doxorubicin hydrochloride. In experiments in vitro, 3T3 cells were treated with 1 mM doxorubicin to induce apoptosis.

Detection of Apoptotic Cells
To determine the apoptotic cell ratios in mouse bone marrow and spleen cells after treatment with 1.5 mg/kg of doxorubicin for exhibit higher levels of gp70 than those from uninfected mice (*p,0.01). Data represent the mean and 95% CI of from 10  To determine the apoptotic cell ratios in 32D, BaF3, and 3T3 cells after treatment with 1 mM doxorubicin for 24 h, samples were collected from each experimental group and washed with ice-cold PBS. These samples were stained with propidium iodide (PI), incubated with FITC-labeled anti-annexin V antibody, and Normally, MCM2 is recruited into the nucleus for participation in DNA replication. As a result, cellular proliferation is upregulated (proliferation signal). However, when gp70 is present in the cytoplasm, it binds to MCM2 and inhibits its nuclear entry. Furthermore, cytoplasmic gp70-MCM2-complex interacts with PP2A and inhibits its interaction with DNA-PK. Consequently, hyperphosphorylated DNA-PK enhances DNA-damage-induced apoptosis via a P53-related pathway (apoptosis signal). doi:10.1371/journal.pone.0040129.g008 analyzed on a FACScan flow cytometer. For detecting apoptotic cells in tissue sections, the terminal deoxy-transferase (TdT)mediated dUTP nick-end labeling (TUNEL) method was used as previously described [58]. Briefly, tissue sections were deparaffinized and incubated with proteinase K (prediluted, DAKO Cytomation, Glostrup, Denmark) for 15 min at room temperature. After the tissues were washed, TdT, FITC-dUTP and -dATP (BoehringerMannheim, Mannheim, Germany) were applied and the sections were incubated in a moist chamber for 60 min at 37uC. Anti-FITC-conjugated antibody-peroxidase (POD converter, Boehringer Mannheim) was employed to detect FITC-dUTP labeling, and color development was achieved with 3,39-diaminobenzidine (DAB) solution containing 0.3% hydrogen peroxide. The sections were then observed under a microscope and the proportion of TUNEL-positive cells was determined by dividing the number of positively stained cells by the total number of cells.

Sybr Green Real-time RT-PCR
RNA was extracted from the bone marrow and spleen cells of BALB/c, C57BL/6, and C3H mice, 8047, 32D, BaF3, and 3T3 cell lines, and primary cultured fibroblasts using Trizol (Invitrogen) according to the manufacturer's instructions. Briefly, the liquid phase was incubated with chloroform for phase separation. Total RNA was finally extracted using one isopropanol precipitation step and one ethanol wash. The RNA pellet was diluted in RNase-and DNase-free water (Qiagen, Hilden, Germany). Then cDNA was generated from RNA using TaqManH Reverse Transcription Reagents (Applied Biosystems, Foster, CA, USA) and quantitative RT-PCR was performed. For quantitative RT-PCR, specific primers were used with the Lightcycler Sybr Green master mix (Roche, Basel, Switzerland). The sequences of the primers are as follows: for Mcm2, GAGGATGGAGAGGAACTCATTG and ATCTTCCTCGCTGCTGTCA; for Dna-pk, GAATTCACCA-CAACCCTGCT and GCTTTCAGCAGGTTCACACA; for Atm, CCTTTGTCCTTCGCGATGTTA and GCTGTATGA-CAAACTCGACTTAATAGGT; and for gp70, AAGGTCCAG CGTTCTCAAAAC and AGGTGGCGTTAGCTGTTTGT. The PCR product was detected using the ABI Prism 7900HT Sequence Detection System (Applied Biosystems [ABI], Carlsbad, CA, USA). The primers and TaqMan probes for Gapdh were purchased from ABI. Mcm2, Dna-pk, Atm, and gp70 RNA levels were normalized to that of Gapdh.

GeneChip Analysis
FLV-infected or uninfected C57BL/6 and C3H mice were administered (i.p.) with 15 mg/kg (high dose) or 1.5 mg/kg (low dose) of doxorubicin, and the spleen was sampled at 0, 1, 6, and 24 h after administration. Total RNA was isolated using RNeasy mini kit (Qiagen), according to the manufacturer's instructions. First-strand cDNAs were synthesized by incubating 5 mg of total RNA with 200 U of SuperScript II reverse transcriptase (Invitrogen) and 100 pmol of the T7-(dT) 24 primer [59-GGCCAGT-GAATTGTAATACGACTCACTATAGGGAGGCGG-(dT) 24 -39]. After second-strand synthesis, the double-stranded cDNAs were purified using a GeneChip Sample Cleanup Module (Affymetrix, Santa Clara, CA, USA), according to the manufacturer's instructions. Double-stranded cDNAs were labeled by in vitro transcription using a BioArray High Yield RNA transcript labeling kit (Enzo Diagnostics, Farmingdale, NY, USA). The labeled cRNA was then purified using a GeneChip Sample Cleanup Module (Affymetrix) and treated with 16 fragmentation buffer (40 mM acetate, 100 mM KOAc, 30 mM MgOAc) at 94uC for 35 min. For hybridization to a GeneChip Mouse Genome 430 2.0 Array (Affymetrix), 15 mg of fragmented cRNA probe was incubated with 50 pM control oligonucleotide B2, 16 eukaryotic hybridization control (1.5 pM BioB, 5 pM BioC, 25 pM BioD and 100 pM Cre), 0.1 mg/mL herring sperm DNA, 0.5 mg/mL acetylated BSA and 16 manufacturer-recommended hybridization buffer in a 45uC rotisserie oven for 16 h. Washing and staining were performed with GeneChip Fluidic Station (Affymetrix) using the appropriate antibody dilution, washing and staining protocol. The phycoerythrin-stained arrays were scanned as digital image files and the scanned data were analyzed with GeneChip Operating Software (Affymetrix) [59]. All data are available online (http://www.nihs.go.jp/tox/TtgSubmitted.htm) from the National Institute of Health Sciences. Twenty-four hours after the transfection, the cells were treated with 1 mM doxorubicin in culture medium for 24 h. Then 10 mL of MTT labeling reagent was added to each well and incubation continued for 4 h at 37uC. Next, 100 mL of solubilization solution was added to each well and incubation was continued overnight at 37uC. Absorbance was determined at 560 nm using a microplate reader (Bio-Rad, Hercules, CA, USA).

RNA Interference
Small-interfering RNA (siRNA) was used to silence the Mcm2 gene. The sequence of siRNA used was CAGGTGACAGACTT-TATCAAA. An irrelevant siRNA (GCACACAGACTGCAAT-CACAGGTTA) that did not lead to specific degradation of any cellular mRNA was used as a negative control. BaF3 and 32D cells (2610 5 cells) were transfected with 120 pmol of Mcm2 or control siRNA using AmaxaH Cell Line NucleofectorH Kit V (LONZA, Basel, Switzerland) according to the manufacturer's instructions. The oligonucleotides used for cloning short hairpin RNA (shRNA)-encoding sequences targeting DNA-PK and ATM into the pSUPER vector (Oligoengine, Seattle, WA, USA) were as follows: Sh-Dna-pk; GATCCCCAGGGCCAAGCTATCATTCT-ttcaagagaAGAATGATAGCTTGGCCCTTTTTTA; and sh-Atm; GATCCCCCATACTAAAGACATTttcaagagaAATGTCTTTGA-GTAGTATGTTTTTA. The annealed oligonucleotides were subcloned into BglII and HindIII sites. These constructs were transfected into 3T3 cells (2610 5 cells) using Hily Max Transfection Reagent (Nippon Gene). The controls were generated by mock-transfected with a sh-empty vector.

Chromatin Loading Assay
Chromatin loading of MCM2 was performed as described previously [60]. Briefly, 3T3 cells were harvested using trypsin, and the cell pellets were lysed by incubating in complete cytoskeleton (CSK) buffer (20 mM HEPES, 100 mM NaCl, 3 mM MgCl 2 , 300 mM sucrose, and 0.1% NP-40) for 15 min on ice. Cytoplasmic fractions were obtained as supernatants after low speed centrifugation (3,0006g) at 4uC. Pellets were rinsed with complete CSK buffer for 10 min on ice and recentrifuged to obtain a chromatin-enriched fraction. Pellets were then sonicated for 5 s in CSK buffer and subjected to high-speed centrifugation (16,0006g). The post-sonication supernatant was designated as the chromatin-bound fraction.

Analysis of Cell Cycle Distribution
Cell cycle distribution was monitored by quantifying the cellular DNA content after staining with PI. Cells were fixed with ethanol for 20 min at 220uC. After centrifugation, cells were suspended in PBS containing PI (50 mg/mL) and RNase (0.2 mg/mL), incu-bated at room temperature for 30 min, and analyzed on a FACScan flow cytometer (BD FACSCanto TM Flow Cytometer).

Transplantation of MCM2-expressing Leukemia Cells into SCID Mice and Apoptosis Induction
The 8047 cells (1 x 10 5 cells) derived from C3H mice were transplanted intravenously into SCID mice via the tail vein. Two weeks after the transplantation, FLV was injected (i.p.) into SCID mice at a dose of 10 4 PFU/mouse. Then, 7 days after FLV inoculation, the mice were treated twice a week with 1.5 mg/kg of doxorubicin.

Immunohistochemistry
Formalin-fixed paraffin-embedded tissue sections (4 mm thick) of the liver from 8047-transplanted SCID mice were de-waxed in xylene and re-hydrated through graded alcohol to water. Antigen retrieval was achieved with a 10-min autoclave treatment in 0.1 M citrate buffer (pH6.0). Endogeneous peroxidase activity was inhibited by dipping the slides in 0.3% hydrogen peroxide in methanol for 30 min and non-specific protein binding was blocked by incubation with normal horse serum (Vector Laboratories, Burlingame, CA, USA). Sections were then treated with anti-MCM2 mouse monoclonal antibody (BD Biosciences) (1:2,000) overnight at 4uC. Detection was achieved using the streptavidinbiotin-peroxidase complex technique (Vector Laboratories) with DAB as the chromogen.

Statistical Analysis
Statistical significance was determined using a two-tailed Student's t-test. For Kaplan-Meier analysis of SCID mice transplanted with 8047 cells, a log-rank test was performed.  Figure S4 Knockdowns of Dna-pk and Atm in gp70 plus Mcm2-transfected cells using the pSUPER shRNA system. The expression of Dna-pk (A) and Atm (B) mRNAs and DNA-PK (C) and ATM (D) proteins were examined by quantitative RT-PCR and western blotting, respectively. Cell survival (E, G) and apoptotic cell ratio (F, H) were determined with the MTT assay and annexin V-staining, respectively, after treatment with 1 mM doxorubicin for 24 h. Note the apoptosis-abrogating effects of sh-Dna-pk (E, F). Asterisks (*) indicate significant differences between sh-Dna-pk-treated and sh-Control-treated cells (*p,0.01). However, Atm knockdown causes no remarkable change in viability or apoptotic cell ratio relative to that of cells treated with sh-Control Table S1 Identification of genes with expression patterns similar to that of Mcm2 using the GeneChip assay. Gene expression patterns were determined by the GeneChip assay in FLV-infected or un-infected C3H/C57BL/6 mice after treatment with doxorubicin. A part of genes exhibited similar expression patterns with Mcm2. The similarity in gene expression patterns was evaluated with the Percellome system using a Pearson product-moment correlation coefficient. (DOCX)