When human cells enter S-phase, overlapping differential inhibitory mechanisms downregulate the replication licensing factors ORC1, CDC6 and Cdt1. Such regulation prevents re-replication so that deregulation of any individual factor alone would not be expected to induce overt re-replication. However, this has been challenged by the fact that overexpression of Cdt1 or Cdt1+CDC6 causes re-replication in some cancer cell lines. We thought it important to analyze licensing regulations in human non-cancerous cells that are resistant to Cdt1-induced re-replication and examined whether simultaneous deregulation of these licensing factors induces re-replication in two such cell lines, including human fibroblasts immortalized by telomerase. Individual overexpression of either Cdt1, ORC1 or CDC6 induced no detectable re-replication. However, with Cdt1+ORC1 or Cdt1+CDC6, some re-replication was detectable and coexpression of Cdt1+ORC1+CDC6 synergistically acted to give strong re-replication with increased mini-chromosome maintenance (MCM) loading. Coexpression of ORC1+CDC6 was without effect. These results suggest that, although Cdt1 regulation is the key step, differential regulation of multiple licensing factors ensures prevention of re-replication in normal human cells. Our findings also show for the first time the importance of ORC1 regulation for prevention of re-replication.
In eukaryotic cells, the periodic assembly and disassembly of essential pre-replication complexes (pre-RC) at replication origins ensure one and only one chromosomal DNA replication per single cell cycle (reviewed by Bell and Dutta, 2002; Diffley, 2004; Fujita, 2006). The pre-RC assembly reaction involves loading of a presumptive replicative helicase, the mini-chromosome maintenance (MCM) 2-7 complex, onto chromatin by an origin recognition complex (ORC), CDC6 and Cdt1, which only occurs during the low cyclin-dependent kinase (Cdk) period from late mitosis through G1-phase. At the onset of S-phase, Cdk activity is regained, which activates the MCM complex to initiate replication and simultaneously prohibits reassembly of pre-RC by suppressing MCM loaders. Indeed, in mammalian cells, Cdk1 inactivation results in MCM reloading and subsequent re-replication (Fujita et al., 1998; Itzhaki et al., 1997). One of the suppression mechanisms is phosphorylation of CDC6, leading to degradation in yeast (Drury et al., 1997; Jallepalli et al., 1997) or nuclear export in mammalian cells (Fujita et al., 1999; Jiang et al., 1999; Petersen et al., 1999; Saha et al., 1998). In human cells, ORC1 and Cdt1 are degraded after S-phase through phosphorylation by Cdks and subsequent ubiquitylation by SCFSkp2 ubiquitin ligase (Fujita et al., 2002; Liu et al., 2004; Méndez et al., 2002; Sugimoto et al., 2004). In budding yeast, the function of ORC2 is suppressed through Cdk phosphorylation (Nguyen et al., 2001). The MCM complex is also phosphorylated by Cdks (Bell and Dutta, 2002; Diffley, 2004; Fujita, 2006). Cdt1 is more strictly regulated by two other mechanisms after S-phase: geminin binding (Lee et al., 2004; McGarry and Kirschner, 1998; Tada et al., 2001; Wohlschlegel et al., 2000) and replication-coupled proteolysis mediated by the cullin4-DDB1Cdt2 ubiquitin ligase and proliferating cell nuclear antigen (Arias and Walter, 2006; Higa et al., 2006; Hu and Xiong, 2006; Jin et al., 2006; Nishitani et al., 2006; Senga et al., 2006).
Because these multiple MCM loaders are strictly regulated by overlapping and differential mechanisms, as mentioned above, it is probable that deregulation of individual components would fail to induce re-replication. Indeed, in budding yeast, only when MCM, CDC6 and ORC are simultaneously deregulated does overt re-replication occur without lowering Cdk activity (Nguyen et al., 2001). Also, in mammalian cells, several studies have shown that overexpression of CDC6 or ORC1 protein alone does not cause overt re-replication (Peterson et al., 1999; Saha et al., 2006; Tatsumi et al., 2006; Vaziri et al., 2003). Because Cdt1 activity is very strictly inhibited by multiple mechanisms after S-phase, it is possible that its deregulation is a more deleterious insult than with other initiation proteins. In fact, excess Cdt1 can induce re-replication in Xenopus egg extracts (Arias and Walter, 2005; Li and Blow, 2005; Maiorano et al., 2005; Yoshida et al., 2005). This could be caused by large amounts of maternal initiation proteins in the eggs. In addition, it has been reported that overexpression of Cdt1 is sufficient to induce re-replication during Drosophila development, although the relative expression levels of exogenous to endogenous Cdt1 were not precisely mentioned (Thomer et al., 2004). In several mammalian cell lines, it has been demonstrated that overexpression of Cdt1 alone can induce overt re-replication and that addition of CDC6 further enhances the re-replication (Liu et al., 2007; Nishitani et al., 2004; Sugimoto et al., 2008; Vaziri et al., 2003). However, such Cdt1-induced overt re-replication has been observed only in certain cancer-derived cell lines and not in normal mammalian cells (Liu et al., 2007; Tatsumi et al., 2006; Vaziri et al., 2003). One reason why Cdt1 alone is able to induce overt re-replication in some cancer-derived cells might be that they constitutively overexpress replication initiation factors such as ORC1, CDC6, Cdt1 and MCM (Tatsumi et al., 2006). Therefore, it is clearly important to investigate licensing controls in non-transformed normal human cells. Recently, it was suggested that, in non-transformed cell lines, the ATR-mediated S-phase checkpoint is activated at the onset of re-replication upon Cdt1 overexpression and that this ATR-mediated S-phase checkpoint inhibits further re-replication (Liu et al., 2007). However, it is also possible that resistance to Cdt1-induced re-replication is the result of overlapping inhibitory pathways involving multiple licensing factors. In addition, it has remained unclear whether deregulation of ORC1 can contribute to re-replication induction. Therefore, we here investigated whether concomitant deregulation of ORC1, CDC6 and Cdt1 can induce overt re-replication in human non-cancerous cells that are resistant to Cdt1-induced re-replication.
HEK 293T and HFF2/T cells are more resistant to Cdt1-induced re-replication than HeLa cells
Previously, we have shown that retrovirus-mediated Cdt1 overexpression can induce overt re-replication in cancer-derived HeLa cells (Sugimoto et al., 2008). Here, we first examined whether retroviral overexpression of Cdt1 induces overt re-replication in HFF2/T cells [normal human fibroblasts immortalized by telomerase (Tatsumi et al., 2006)] and in low tumorigenic HEK 293T kidney cells immortalized by adenoviral E1A and E1B (Numa et al., 1995). HeLa cells were also studied for comparison. All these cells were infected with retroviruses expressing wild-type, T29A or Cy+D1m T7-Cdt1 or control retroviruses [multiplicity of infection (MOI)=1 for HeLa and HEK 293T; MOI=10 for HFF2/T] and selected with hygromycin B. The T29A mutation impairs Cdk phosphorylation and SCFSkp2-mediated Cdt1 destabilization, and Cy+D1m mutations impair both SCFSkp2- and APC/CCdh1-mediated destabilization (Sugimoto et al., 2008; Takeda et al., 2005). At 4 days after infection, cells were collected and analyzed. Immunoblot analyses confirmed expression of the introduced proteins (Fig. 1B). The levels of the T7-Cdt1 T29A were higher than the wild type, and those of the T7-Cdt1 Cy+D1m were further increased (Fig. 1B), as previously described (Sugimoto et al., 2008). Flow cytometry analyses revealed overexpression of wild-type Cdt1 to induce detectable re-replication (defined by a DNA content higher than 4N) in HeLa cells (∼1.5% in the control case versus 8% with T7-Cdt1); T7-Cdt1 T29A (∼19%) and T7-Cdt1 Cy+D1m (∼40%) further enhanced the re-replication (Fig. 1A) (Sugimoto et al., 2008). It was previously reported that adenovirus-mediated overexpression of wild-type Cdt1 cannot induce detectable re-replication in the normal human fibroblasts IMR90 and WI38 (Liu et al., 2007). In agreement, we observed little difference between control virus-infected and T7-Cdt1-infected HFF2/T cells (∼1% for control versus 3.5% for T7-Cdt1) (Fig. 1A). However, even in HFF2/T cells, re-replication was clearly detected with T7-Cdt1 T29A and the percentage was further increased with Cy+D1m (∼7.7% for T29A and 14% for Cy+D1m). However, HFF2/T cells were significantly more resistant to Cdt1-induced re-replication than HeLa cells (e.g. in the case of T7-Cdt1 Cy+D1m, ∼40% in HeLa versus ∼14% in HFF2/T). Several previous reports provided evidence that re-replication can be induced in HEK 293T cells with Cdt1 overexpression (Nishitani et al., 2006; Teer and Dutta, 2008). In these reports, cells were co-transfected with Cdt1 and GFP, and selected for GFP-positive cells to enrich cells expressing Cdt1 at high levels. Under our experimental conditions, overexpression of wild-type T7-Cdt1 induced detectable re-replication in HEK 293T cells, but to a lesser extent than in HeLa (∼4.5% versus ∼8%). Also in HEK 293T cells, re-replication was elevated upon T29A overexpression and was further increased by Cy+D1m (Fig. 1A). Nevertheless, as with HFF2/T cells, there was still significant resistance to Cdt1-induced re-replication (e.g. in the case of Cdt1 Cy+D1m, ∼40% in HeLa versus ∼19% in HEK 293T). In all three cell lines, the extent of re-replication appeared to increase in parallel with the steady-state levels of Cdt1 proteins (Fig. 1).
Previously, we demonstrated that in HeLa cells in which DNA content is increased by retroviral Cdt1 overexpression for 4 days, the percentage of cells with two or more centrosomes is not remarkably increased and 8N ploidy cells do not accumulate, indicating that the increase in DNA content (>4N) actually results from genuine re-replication within a single S-phase, rather than from mitotic failure (Sugimoto et al., 2008). Also, in the present studies, there was no accumulation of 8N ploidy cells in Cdt1-overexpressing HeLa, HEK 293T or HFF2/T cells (Fig. 1A). In addition, we sought to examine the effects of changing the Cdt1 expression levels on re-replication induction in these three cell lines by altering the MOI for retrovirus infection. Cells were infected with the retroviruses expressing Cdt1 (wild-type, T29A or Cy+D1m) at different MOI (0.1, 1 and 10, respectively). Because of the cytotoxicity of vesicular stomatitis virus glycoprotein (VSVG), which was used for production of pseudotype recombinant retroviruses in our system, and limitations in preparation of high-titer retroviruses (in the case of Cdt1 retroviruses, a maximum of ∼106 colony-forming units per ml in HeLa cells), it was impossible to further elevate the MOI. In addition, for HeLa cells, even infection of retroviruses at MOI=10 caused cell toxicity and, therefore, the data were excluded. Overall, the extent of re-replication was increased in parallel with the increase in Cdt1 protein level (supplementary material Fig. S1). Taken together, these data suggest that (1) HFF2/T and HEK 293T cells are resistant to induction of re-replication by wild-type Cdt1, and (2) nevertheless, overt re-replication can be induced even in such cells when licensing control is robustly deregulated by Cdt1 mutants stabilized in S-phase.
As to the reason why normal human fibroblasts are resistant to Cdt1-induced re-replication, it was previously suggested that although reloaded MCM re-unwinds DNA, resultant inappropriately re-unwound DNA activates the ATR checkpoint, which inhibits further DNA synthesis (Liu et al., 2007). We therefore investigated whether Cdt1-induced re-replication is affected by shRNA-mediated silencing of ATM and/or ATR. HFF2/T cells were co-infected with retroviruses expressing wild-type T7-Cdt1 (MOI=10) and shRNAs for ATM (MOI=1) and/or ATR (MOI=1) and selected with both hygromycin B and puromycin. As shown in supplementary material Fig. S2B, expression of ATM protein was specifically reduced by ∼80% with the shRNA expression, and ATR protein was reduced by ∼50%. In our system, enhancement of Cdt1-induced re-replication was not observed when ATM or ATR were silenced (supplementary material Fig. S2A). Co-silencing of both could not enhance Cdt1-induced re-replication (supplementary material Fig. S2A). Similar results were obtained with HEK 293T, HeLa and T98G cells (data not shown).
Simultaneous deregulation of Cdt1+CDC6 or Cdt1+ORC1 synergistically induces re-replication, and triple overexpression leads to strongest re-replication in HEK 293T cells
Using HEK 293T cells that are resistant to Cdt1-induced re-replication as above, we examined whether simultaneous deregulation of multiple licensing factors (i.e. ORC1, CDC6 and Cdt1) synergistically induces re-replication. In these experiments, FLAG-ORC1, 2HA-CDC6 and/or T7-Cdt1 were introduced by transfection, together with GFP to survey transfection efficiency; 48 hours after transfection, the cells were analyzed by immunoblotting and flow cytometry. Overexpression of the introduced proteins was confirmed (Fig. 2C). Compared with the results shown in Fig. 1A, no significant re-replication was induced by Cdt1 overexpression alone (Fig. 2A,B; supplementary material Fig. S3A), probably because of low transfection efficiency compared with retroviral infection. Also, no significant re-replication was induced by ORC1 or CDC6 overexpression alone. However, coexpression of Cdt1+ORC1 or Cdt1+CDC6, but not ORC1+CDC6, induced detectable re-replication (∼7.6% and ∼4.2%, respectively) (Fig. 2A,B; supplementary material Fig. S3A). Interestingly, coexpression of Cdt1+ORC1+CDC6 synergistically induced the strongest re-replication (∼11.6%) (Fig. 2A,B; supplementary material Fig. S3A).
To test whether levels of chromatin-bound MCM are increased in cells with re-replicated DNA, whole-cell extracts and nuclear fractions containing chromatin- and nuclear-matrix-bound proteins were prepared from asynchronously growing transfected HEK 293T cells (Fujita et al., 1997). As expected, cells expressing Cdt1+ORC1 or Cdt1+CDC6 possessed more chromatin-bound MCM7 and MCM3 proteins than control cells (∼1.7-fold; P<0.05) (Fig. 2D). Coexpression of Cdt1+ORC1+CDC6 further increased MCM loading (∼twofold; the difference between Cdt1+ORC1 and Cdt1+ORC1+CDC6 was significant at P<0.05). MCM proteins are gradually detached from chromatin as replication proceeds and levels of chromatin-bound MCM proteins are lowest when cells reach the G2-M phase (Fujita, 2006; Fujita et al., 1998). Therefore, it could be that an increase in chromatin-bound MCM proteins owing to MCM reloading might be more clearly observed when cells are synchronized in late S-phase to G2-M phase. However, unfortunately, we found that cell-cycle progression of HEK 293T cells overexpressing Cdt1, ORC1 and/or CDC6 is significantly delayed compared with control transfectants when treated with nocodazole or hydroxyurea and then released for synchronization (data not shown). Because asynchronous cell growth was not remarkably affected by overexpression of these proteins (except for ORC1+CDC6+Cdt1 triple overexpression, which significantly inhibits the growth) (supplementary material Fig. S3B), such delay might be due to the combined cytotoxicity of the overexpression and drugs.
It has been demonstrated that Cdt1-induced re-replication damages DNA and thereby activates ATR and ATM pathways (Liu et al., 2007; Vaziri et al., 2003), and it has also been suggested that Cdt1 overexpression can activate ATM independent of re-replication (Tatsumi et al., 2006). We therefore examined the ATM checkpoint activation in the current experimental setting. As we reported previously (Tatsumi et al., 2006), the levels of phospho-ATM and phospho-Chk2, the presence of which reflects activation of the checkpoint, increased significantly in cells overexpressing Cdt1; this was not observed when CDC6, ORC1 or CDC6+ORC1 were overexpressed (Fig. 2C). Importantly, Cdt1-induced activation of the ATM-Chk2 pathway was not further augmented by overt re-replication elicited by coexpression of ORC1, CDC6 or ORC1+CDC6 (Fig. 2C). The data provide further strong support for the notion that Cdt1 has an ability to activate the ATM-Chk2 pathway independent of the function to induce re-replication. Elucidation of the mechanism is an interesting research challenge.
A CDC6 AAA mutant deficient in Cdk phosphorylation evokes stronger re-replication than wild type when coexpressed with Cdt1 in HEK 293T cells
Overexpression of a phospho-deficient Cdt1 T29A mutant induces re-replication more efficiently than wild type (Fig. 1) (Sugimoto et al., 2008; Takeda et al., 2005), demonstrating that negative regulation of Cdt1 by Cdk phosphorylation is crucial for prohibition of re-replication. CDC6 is also a substrate of Cdks and the phosphorylation leads to nuclear export in mammalian cells (Fujita et al., 1999; Jiang et al., 1999; Petersen et al., 1999; Saha et al., 1998). Although it has been speculated that this pathway also contributes to prohibition of re-replication, direct evidence is lacking for human cells that phospho-deficient CDC6 mutants induce stronger re-replication than the wild type (Jiang et al., 1999; Liu et al., 2007; Petersen et al., 1999; Vaziri et al., 2003). In the Xenopus egg extract system, phospho-deficient CDC6 mutants do not cause re-replication (Pelizon et al., 2000). On the other hand, in fission yeast, phosphorylation leads to CDC6 degradation, and overexpression of unphosphorylatable CDC6 causes a strong re-replication (Jallepalli et al., 1997). Taking advantage of our experimental system, in which coexpression of Cdt1 and CDC6 induces detectable re-replication in HEK 293T cells, we examined whether Cdk phosphorylation of CDC6 is actually involved in prevention of re-replication. For this purpose, we employed a phospho-deficient CDC6 AAA mutant (S54A, S74A, S106A) and a phospho-mimic CDC6 EEE mutant (S54E, S74E, S106E) (Jiang et al., 1999; Petersen et al., 1999). We transfected HEK 293T cells with wild-type, AAA, or EEE 2HA-CDC6 expression vectors, with or without wild-type T7-Cdt1 (Fig. 3). The levels of ectopically expressed 2HA-CDC6 AAA were lower than the wild type and, conversely, those of 2HA-CDC6 EEE were elevated (Fig. 3C). This might be because phosphorylation of CDC6 by Cdks interferes with APC/CCdh1-mediated destabilization of CDC6 (Mailand and Diffley, 2005). Next, we examined the subcellular localization of ectopically expressed CDC6 (Fig. 3D). Consistent with previous findings (Fujita et al., 1999), endogenous CDC6 proteins were detected both in soluble and in the chromatin- and nuclear-matrix-bound fractions. Wild-type 2HA-CDC6 proteins were also distributed in both fractions but predominantly in the soluble fraction compared with endogenous CDC6. Thus, it is possible that the N-terminal 2HA-tag somewhat affects CDC6 structure and function. Almost all CDC6 AAA was bound to chromatin or the nuclear matrix and most CDC6 EEE was found in the soluble fraction, as expected (Jiang et al., 1999; Petersen et al., 1999). When we examined DNA content, the CDC6 AAA or EEE mutant alone had only a limited effect on re-replication, as with the wild-type CDC6 (Fig. 3A,B; supplementary material Fig. S4). When coexpressed with Cdt1, CDC6 AAA induced stronger re-replication than the wild type, in spite of lower levels (∼4.2% for Cdt1+CDC6 wild type verses ∼6.6% for Cdt1+CDC6 AAA; P<0.05) (Fig. 3A,B; supplementary material Fig. S4). No significant difference was seen between the Cdt1+CDC6 wild-type and Cdt1+CDC6 EEE. We also examined whether levels of chromatin-bound MCM are increased in cells with Cdt1+CDC6 AAA in comparison with Cdt1+CDC6 wild-type or Cdt1+CDC6 EEE. Whole-cell extracts and nuclear fractions containing chromatin- and nuclear-matrix-bound proteins were prepared from asynchronously growing transfected HEK 293T cells and examined, as above. Cells expressing Cdt1+CDC6 AAA possessed more chromatin-bound MCM7 and MCM3 proteins than Cdt1+CDC6 wild-type or Cdt1+CDC6 EEE (P<0.05) (Fig. 3E). Together, these results provide evidence that the phospho-deficient CDC6 mutant can induce stronger re-replication than the wild type, and thus that Cdk phosphorylation of CDC6 is important for prevention of re-replication in human cells.
Simultaneous deregulation of Cdt1+CDC6, Cdt1+ORC1 or Cdt1+ORC1+CDC6 synergistically induces re-replication even in normal human cells
The above data demonstrate that even in HEK 293T cells that are relatively resistant to Cdt1-induced re-replication, simultaneous deregulation of Cdt1+CDC6 or Cdt1+ORC1 synergistically induces overt re-replication, and the triple overexpression leads to strongest re-replication. However, HEK 293T cells are immortalized with adenoviral E1A and E1B, which inactivate Rb and p53, respectively. To determine whether similar findings might be obtained with normal human cells, we co-transduced Cdt1, CDC6 and/or ORC1 into HFF2/T cells using retrovirus vectors (pCLMSCVhyg-T7-Cdt1, pCLHCX-ORC1, pQCXIP-2HA-CDC6, or their control empty vectors). In the case of `Cdt1', the cells were simultaneously infected with the three retroviruses produced with pCLMSCVhyg-T7-Cdt1, pCLHCX and pQCXIP, respectively. The cells were then co-selected with both hygromycin B and puromycin, and collected and analyzed at 4 days after infection. Immunoblot analyses with whole-cell extracts confirmed overexpression of the introduced proteins (Fig. 4C). Cdt1, ORC1 or CDC6 overexpression alone did not induce any detectable re-replication in HFF2/T cells (Fig. 4A,B; supplementary material Fig. S5). However, in agreement with the data for HEK 293T cells, re-replicated DNA was detected with coexpression of Cdt1+ORC1 or Cdt1+CDC6 (∼7% and ∼5.5%, respectively). Again, the strongest re-replication (∼11%) was induced when the three MCM loaders were overexpressed simultaneously.
Redundant and differential inhibitory pathways for the multiple licensing factors ensure prevention of re-replication
When human cells enter S-phase, they employ redundant and differential inhibitory mechanisms to downregulate the multiple licensing factors, ORC1, CDC6 and Cdt1 (Bell and Dutta, 2002; Diffley, 2004; Fujita, 2006). Such regulations might operate to prevent re-replication and thus it is conceivable that deregulation of the individual factor alone is insufficient to induce overt re-replication. However, this notion has been challenged by the fact that Cdt1 overexpression alone can induce overt re-replication in some cancer-derived cell lines (Liu et al., 2007; Nishitani et al., 2004; Sugimoto et al., 2008; Vaziri et al., 2003). Here, we demonstrated that, although overexpression of each of ORC1, CDC6 or Cdt1 alone was insufficient to induce overt re-replication in non-cancerous cell lines, including normal human fibroblasts, co-overexpression of Cdt1+ORC1 or Cdt1+CDC6 caused detectable re-replication, and triple deregulation synergistically induced the strongest re-replication. These data provide strong support for the notion that overlapping differential regulation of multiple licensing factors ensures prevention of re-replication in human normal somatic cells, similar to the case in budding yeast (Nguyen et al., 2001). Because coexpression of ORC1 and CDC6 could not induce detectable re-replication, Cdt1 regulation must also be the key step in normal human cells.
Previously, Liu et al. reported that, although adenoviral overexpression of wild-type Cdt1 alone does not cause overt re-replication in normal human fibroblasts, Cdt1 overexpression in combination with ATR silencing induces re-replication (Liu et al., 2007). On the basis of these findings, they suggested that although overexpressed Cdt1 can reload MCM onto chromatin and lead to origin re-unwinding, the inappropriately unwound DNA activates the ATR checkpoint, which inhibits further DNA synthesis. Our present data indicate that if MCM is sufficiently reloaded by overexpression of Cdt1 mutants resistant to SCFSkp2-mediated S-phase degradation or by simultaneous overexpression of Cdt1 plus ORC1 or CDC6, then detectable re-replication is induced even in normal human fibroblasts. In addition, in our retrovirus-mediated system, ATR silencing by shRNA expression did not further enhance Cdt1-induced re-replication in HFF2/T, HEK 293T, HeLa and T98G cells, seemingly conflicting with the previous findings. One possible reason for the apparent discrepancy is that we performed the assays 4 days after retrovirus infection rather than 48 hours after. Adenovirus vectors express many adenoviral proteins, including E4 protein, which could also have affected the results. For example, it should be noted that E4 inactivates the Mre11-Rad50-NBS1 complex (Stracker et al., 2002).
ORC1 regulation after S-phase plays a role in prevention of re-replication in human cells
Human ORC1 is known to be downregulated after S-phase, probably through Cdk phosphorylation and SCFSkp2-mediated degradation (Fujita et al., 2002; Méndez et al., 2002). However, to our knowledge, it remained unclear whether such ORC1 regulation actually contributes to prevention of re-replication in human cells. Here, we showed that although overexpression of ORC1 alone does not induce detectable re-replication, overt re-replication does occur in combination with Cdt1 overexpression. These data demonstrate for the first time that regulation of ORC1 is an important mechanism for maintaining genome integrity. Because Cdk phosphorylation sites in human ORC1 have yet to be determined, we could not prepare phospho-deficient mutants. Thus, it remains to be elucidated whether ORC1 phosphorylation by Cdks is actually involved.
Phosphorylation by Cdks downregulates CDC6 function to prevent re-replication in human cells
Human CDC6 is a substrate of Cdks, and its phosphorylation leads to nuclear export (Fujita et al., 1999; Jiang et al., 1999; Petersen et al., 1999; Saha et al., 1998). Thus, it has been speculated that this pathway also contributes to prohibition of re-replication. However, although it has been clearly shown that deregulation of CDC6 by overexpression leads to re-replication in combination with Cdt1 overexpression, it remains to be elucidated whether phospho-deficient CDC6 mutants indeed cause stronger re-replication than the wild type in human cells (Jiang et al., 1999; Liu et al., 2007; Petersen et al., 1999; Vaziri et al., 2003). Here, we provide evidence that this is indeed the case. In this regard, it was reported that coexpression of phospho-deficient CDC6 and non-degradable Cdt1 induces re-replication in Caenorhabditis elegans (Kim et al., 2007).
Previously, it was shown that a phospho-mimic CDC6 DDD mutant drives pre-RC assembly more efficiently than the wild type in G0 or Cdk-inhibited G1 cells where APC/CCdh1 is active (Mailand and Diffley, 2005). This is because CDC6 phosphorylation by Cdk interferes with APC/CCdh1-mediated ubiquitylation (Mailand and Diffley, 2005). The data might appear to contradict our finding that the CDC6 AAA mutant induces stronger re-replication than the wild type when coexpressed with Cdt1 in HEK 293T cells. However, re-replication induced by Cdt1+CDC6 might occur mainly in the S-, G2- and M-phases where Cdk activity is high and, conversely, APC/CCdh1 activity is low. Thus, our data suggest that CDC6 phosphorylation by Cdks contributes to prevention of re-replication in the S-, G2- and M-phases. Nuclear export might be one mechanism of Cdk phosphorylation-mediated inhibition of CDC6 function, as suggested previously (Fujita et al., 1999; Jiang et al., 1999; Petersen et al., 1999; Saha et al., 1998). However, under our experimental conditions, although the levels of chromatin- and nuclear-matrix-bound CDC6 EEE were higher than those of CDC6 AAA, CDC6 EEE induced less re-replication than CDC6 AAA when coexpressed with Cdt1 (Fig. 3), suggesting that Cdk phosphorylation could directly inhibit CDC6 activity.
Materials and Methods
HFF2/T, HEK 293T, and HeLa cells were grown in Dulbecco's modified Eagle's medium with 8% fetal calf serum.
The retrovirus vector pCLMSCVhyg-T7-Cdt1 and the mammalian expression vector pcDNA3.1-zeo-Flag-ORC1 were as described previously (Sugimoto et al., 2008; Tatsumi et al., 2006). For retroviral expression of ORC1, pCLHCX-ORC1 was constructed by inserting ORC1 cDNA into pCLHCX prepared from pCLNCX (Imgenex) by replacing the neomycin-resistance gene with a hygromycin-resistance gene. Potential polyadenylation signals in ORC1 cDNA were eliminated without changing amino acids by oligonucleotide-directed mutagenesis. The retrovirus vector pQCXIP-2HA-CDC6 was constructed as follows. Plasmid pBS2HA-CDC6 was prepared as described previously (Fujita et al., 2002) and the resultant 2HA-CDC6 was cloned into pENTR4 (Invitrogen). Then, using LR Clonase (Invitrogen), the fragment was recombined from pENTR4 into pQCXIP with puromycin-resistance gene (Clontech). The cDNAs encoding Cdk phosphorylation–deficient CDC6 AAA (Ser54, Ser74, and Ser106 were replaced with alanines) and phosphorylation-mimic EEE (Ser54, Ser74, and Ser106 were replaced with glutamates) mutants (Jiang et al., 1999) were gifts from Wei Jiang, and were excised and inserted into pENTR4. For silencing ATM or ATR, the sequences corresponding to ATM or ATR cDNA (underlined) were expressed as shRNAs as follows. The oligonucleotides for the sense strand (for ATM, 5′-GATCCCCGCTGATTGTAGCAACATACTTCAAGAGAGTATGTTGCTACAATCAGCTTTTTGGAAA-3′; and for ATR, 5′-GATCCCCGGCAGTGCCACACCAGAGGAATATAATACAGTTCAAGAGACTGTATTATATTCCTCTGGTGTGGCACTGCCTTTTTGGAAA-3′) and the complement oligonucleotides were annealed and introduced into the pCL-SI-MSCVpuro-H1R retroviral expression vector as previously described (Haga et al., 2007; Tatsumi et al., 2006). pAcGFP1-Mem was purchased from Clontech.
HEK 293T cells (4×105 in six-well culture plates) were co-transfected with a mixture (total ∼1 μg) of three different plasmids (pCLMSCVhyg-T7-Cdt1, pcDNA3.1-zeo-Flag-ORC1, pQCXIP-2HA-CDC6) or their empty vectors, and with pAcGFP1-Mem to monitor transfection efficiencies using TransIT-293 reagents (Mirus, Madison, WI). For example, in the case of `Cdt1 alone', the cells were transfected with a mixture of the following plasmids: 0.3 μg of pCLMSCVhyg-T7-Cdt1, 0.3 μg of pcDNA3.1-zeo, 0.3 μg of pQCXIP and 0.03 μg of pAcGFP1-Mem. Forty-eight hours after transfection, cells were lysed in 1×SDS sample buffer (62.5 mM Tris-HCl pH 6.8, 2% SDS, 5% β-mercaptoethanol, 10% glycerol, 0.01% bromophenol blue) containing multiple protease and phosphatase inhibitors or were subjected to FACS analyses.
Recombinant retroviruses were produced as described previously (Haga et al., 2007; Tatsumi et al., 2006). For the studies shown in Fig. 1, HFF2/T, HEK 293T, and HeLa cells were infected with retroviruses expressing wild-type, T29A or Cy+D1m T7-Cdt1 or with control retroviruses (MOI=1 for HeLa and HEK 293T, and MOI=10 for HFF2/T), and selected with 200 μg/ml hygromycin B.
For the studies shown in Fig. 4, HFF2/T cells were co-infected with three different recombinant retroviruses (MOI=1 for each), each of which was produced with retroviral vectors pCLMSCVhyg-T7-Cdt1, pCLHCX-ORC1 or pQCXIP-2HA-CDC6, or with their control empty vectors. For example, in the case of `Cdt1 alone', the cells were simultaneously infected with the following retroviruses: pCLMSCVhyg-T7-Cdt1, pCLHCX and pQCXIP. The cells were then selected with both 200 μg/ml hygromycin B and 0.5 μg/ml puromycin.
Detection of re-replication and immunoblot analyses were carried out at 4 days post-infection.
Cells were treated with a CycleTEST PLUS DNA reagent kit (Becton Dickinson) for propidium iodide staining, and then analyzed with a Becton Dickinson FACS Calibur. Statistical analysis was performed with the two-tailed Student's t-test to validate the data. P values of <0.05 were considered statistically significant.
Immunoblotting and antibodies
Immunoblotting and quantification of the band signals were performed as described previously (Sugimoto et al., 2008). Preparation of polyclonal rabbit antibodies against human ORC1, CDC6, Cdt1, MCM7 and MCM3 was as detailed earlier (Fujita et al., 1998; Sugimoto et al., 2004; Tatsumi et al., 2006).
Other antibodies were purchased from different companies: actin (AC-15, Sigma), GFP (46-0092, Invitrogen), Thr68-phosphorylated Chk2 (number2661, Cell Signaling), ATR (sc-1887, Santa Cruz), ATM (2C1, Gene Tex), Ser1981-phosphorylated ATM (200-301-400, Rockland), Lamin A/C (number2032, Cell Signaling), Ras (Ab-4, Oncogene Research Products).
We thank Wei Jiang (The Burnham Institute, CA) for plasmids. We are also grateful to Satoko Yoshida for technical and Akiko Noguchi for secretarial assistance. This work was supported in part by a grant to M.F. from the Ministry of Education, Culture, Sports, Science and Technology of Japan. N.S. was supported by a JSPS research fellowship.