TIMELESS Suppresses the Accumulation of Aberrant CDC45·MCM2-7·GINS Replicative Helicase Complexes on Human Chromatin*

The replication licensing factor CDC6 recruits the MCM2-7 replicative helicase to the replication origin, where MCM2-7 is activated to initiate DNA replication. MCM2-7 is activated by both the CDC7-Dbf4 kinase and cyclin-dependent kinase and via interactions with CDC45 and go-ichi-ni-san complex (GINS) to form the CDC45·MCM2-7·GINS (CMG) helicase complex. TIMELESS (TIM) is important for the subsequent coupling of CMG activity to DNA polymerases for efficient DNA synthesis. However, the mechanism by which TIM regulates CMG activity for proper replication fork progression remains unclear. Here we show that TIM interacts with MCM2-7 prior to the initiation of DNA replication. TIM depletion in various human cell lines results in the accumulation of aberrant CMG helicase complexes on chromatin. Importantly, the presence of these abnormal CMG helicase complexes is not restricted to cells undergoing DNA synthesis. Furthermore, even though these aberrant CMG complexes interact with the DNA polymerases on human chromatin, these complexes are not phosphorylated properly by cyclin-dependent kinase/CDC7-Dbf4 kinase and exhibit reduced DNA unwinding activity. This phenomenon coincides with a significant accumulation of the p27 and p21 replication inhibitors, reduced chromatin association of CDC6 and cyclin E, and a delay in S phase entry. Our results provide the first evidence that TIM is required for the correct chromatin association of the CMG complex to allow efficient DNA replication.

The replication licensing factor CDC6 recruits the MCM2-7 replicative helicase to the replication origin, where MCM2-7 is activated to initiate DNA replication. MCM2-7 is activated by both the CDC7-Dbf4 kinase and cyclin-dependent kinase and via interactions with CDC45 and go-ichi-ni-san complex (GINS) to form the CDC45⅐MCM2-7⅐GINS (CMG) helicase complex. TIMELESS (TIM) is important for the subsequent coupling of CMG activity to DNA polymerases for efficient DNA synthesis. However, the mechanism by which TIM regulates CMG activity for proper replication fork progression remains unclear. Here we show that TIM interacts with MCM2-7 prior to the initiation of DNA replication. TIM depletion in various human cell lines results in the accumulation of aberrant CMG helicase complexes on chromatin. Importantly, the presence of these abnormal CMG helicase complexes is not restricted to cells undergoing DNA synthesis. Furthermore, even though these aberrant CMG complexes interact with the DNA polymerases on human chromatin, these complexes are not phosphorylated properly by cyclin-dependent kinase/CDC7-Dbf4 kinase and exhibit reduced DNA unwinding activity. This phenomenon coincides with a significant accumulation of the p27 and p21 replication inhibitors, reduced chromatin association of CDC6 and cyclin E, and a delay in S phase entry. Our results provide the first evidence that TIM is required for the correct chromatin association of the CMG complex to allow efficient DNA replication.
During cell growth, faithful DNA duplication is essential for correctly transmitting genetic material to daughter cells during cell division, and defects in this process can result in genome instability, tumorigenesis, or cell death. During DNA replication, double-stranded DNA must be unwound by the MCM2-7 a hexameric helicase to generate ssDNA, 2 which is then used as the template for DNA synthesis by DNA polymerase. However, even though there is an abundance of MCM2-7 protein on the chromatin throughout the cell cycle, most of the MCM2-7 is kept inactive, and only a small population is recruited and associated with active replicons to participate in DNA synthesis (1)(2)(3). During early G 1 phase, the MCM2-7 protein is recruited to the active replication origins, where it forms part of the pre-replication complex, and its recruitment requires the CDC6 and CDT1 licensing factors (4,5). To ensure the accurate timing of replication initiation, the activity of the chromatin-bound MCM2-7 protein is tightly regulated both by post-translational modifications and protein-protein interactions. Specifically, the catalytic activity of the MCM2-7 helicase is suppressed during G 1 phase and can only be activated both via the interaction with GINS and CDC45, which forms the CMG complex, and via phosphorylation by DDK and CDK during replication initiation (6,7).
Mammalian TIM is an essential protein for embryonic development (8). TIM and its interacting protein, TIPIN, are highly conserved replication factors that directly interact with the MCM2-7 helicase (9). TIM and TIPIN form a stable heterodimer, and this interaction is crucial for the stability and nuclear localization of both proteins in vivo (10 -12). They are the mammalian homologs of Saccharomyces cerevisiae Tof1 and Csm3, respectively (13,14). Tof1 and Csm3 are part of the replication progression complex that couples DNA unwinding and DNA synthesis activities and stabilizes replication forks at pause sites (15)(16)(17)(18). Tof1 also plays a role in activating the DNA damage response pathway during S phase (19,20). The functions of Tof1 and Csm3 are conserved in their vertebrate homologs, TIM and TIPIN (21,22). For example, when cells encounter DNA damage during S phase, TIM-TIPIN dimers promote phosphorylation of CHK1, which activates the intra-S phase checkpoint response and arrests replication forks. In the absence of TIM-TIPIN, cells continue to synthesize damaged DNA, leading to catastrophic consequences, as demonstrated by increased cell death (21,22). In undamaged cells, TIM dysfunction decreases the rate of replication fork progression and uncouples the DNA polymerase and MCM2-7 helicase activity (21). TIM-TIPIN also facilitates the loading of cohesin subunits to establish sister chromatid cohesions (23,24). The role of TIM-TIPIN in cohesion establishment is consistent with the discovery of Csm3 and Tof1 mutations in genetic screens for chromosome segregation defects (14,25).
Here we report a novel function of human TIM for the correct association of the CMG complex on chromatin. We found that TIM-TIPIN interacts with MCM2-7 not only during S phase but also throughout the whole cell cycle. Human cell lines treated with TIM siRNAs contain elevated amounts of the p21 and p27 replication inhibitors, and this phenotype coincides with a delay in S phase entry and decreased association of CDC6 and cyclin E with chromatin. As a consequence, there is reduced recruitment of MCM2-7 to the active replication origin. Unexpectedly, despite the inefficient recruitment of MCM2-7 to the active replication origin during G 1 phase in TIM-deficient cells, the levels of chromatin-bound CMG complexes remain unchanged, and the presence of these CMG complexes on the chromatin is no longer restricted to S phase. Although these CMG complexes interact with DNA polymerases, the MCM4 subunit has an altered phosphorylation pattern at the DDK-and CDK-dependent PG sites, which are important for efficient DNA replication (26,27). Our data unveil a novel role for TIM in preventing the accumulation of aberrant CMG complexes on the chromatin outside of S phase. We propose that the presence of these non-S phase CMG complexes with altered post-translational modifications acts as a false negative feedback signal to prevent CDC6 and cyclin E from binding to DNA, thereby hindering DNA replication in TIM-deficient cells.

Results
TIM Deficiency Leads to Inefficient S Phase Entry-Mammalian TIM is a component of the replication fork progression complex and is required for the efficient progression of replication forks during S phase (21,22,28). In addition, TIM promotes the sister chromatid cohesion necessary for proper chromosomal segregation during mitosis (23,24). Reduced levels of cohesin complexes during early G 1 phase can also lead to slow replication progression and can lengthen S phase by limiting the number of replication origins that fire (29). It is therefore expected that TIM deficiency would lead to the accumulation of S phase cells. To test this, we depleted TIM using two separate siRNAs in HEK293 cells (Fig. 1A, left panel) and human osteosarcoma U2OS cells (Fig. 1B, left panel). Consistent with a published report (11), we observed a destabilization of TIPIN in TIM knockdown cells using either of the TIM siRNAs (Fig. 1, A and B, left panels, second blots). The percentage of the asynchronized cell population undergoing DNA synthesis was monitored by measuring DNA content (Fig. 1, A and B, right panels, x axis) and BrdU incorporation by pulse-labeling the cells with BrdU prior to analysis (Fig. 1, A and B, right panels, y axis). Unexpectedly, BrdU incorporation analysis of the TIM siRNA-treated HEK293 (Fig. 1A, right panel) and U2OS (Fig.  1B, right panel) cells showed that the S phase population was reduced compared with the control cells.
An accelerated replication fork progression rate could lead to a shorter S phase and, hence, a reduction in the S phase population. Alternatively, a decrease in the number of cells undergoing DNA synthesis may result from inefficient S phase entry.
To test these possibilities, we evaluated the cell cycle progression of HEK293 cells that were transfected with either control or TIM siRNAs for 48 h and then synchronized to G 2 /M phase by nocodazole treatment for an additional 20 h. After nocodazole release, cell cycle progression was monitored by flow cytometry. Within the first 4 -6 h after release, the number of G 1 phase cells increased in both the control and TIM siRNA knockdown cultures (Fig. 1C, first and second columns). 12 h after nocodazole release, the majority of the control knockdown cells showed positive BrdU staining (Fig. 1C, fourth column). However, the initiation of DNA synthesis was delayed at 12 h after nocodazole release in the majority of TIM-depleted cells that had completed mitosis and entered G 1 phase. Overall, these cell cycle analyses revealed an unexpected defect in S phase entry caused by TIM depletion.
High Levels of the Replication Inhibitors p21 and p27 Are Present in TIM-depleted Cells without Evidence of DNA Damage-Consistent with the observation of a reduced S phase population, elevated levels of p21 and p27 were present in TIM knockdown HEK293 and U2OS cells compared with control cells (Fig. 1, A and B, left panels, third and fourth blots). Expression of the SV40 large T antigen has been reported to stimulate the initiation of host DNA replication and reduce the expression levels of p21 and p27 (30,31). However, HEK293T cells, which contain the SV40 large T antigen, also accumulated p21 and p27 when TIM was depleted ( Fig. 2A, third and fourth panels). p21 and p27 block DNA replication by inactivating CDK2cyclin E, which is important for G 1 /S phase transition (32). For cells to enter S phase, p21 and p27 must be ubiquitinated and degraded during late G 1 phase to allow the activation of CDK2cyclin E for replication initiation. Therefore, we asked whether the delay in S phase entry in TIM knockdown cells was associated with the persistent presence of p21 and p27. Indeed, when we synchronized HEK293T cells with nocodazole to various cell cycle stages (Fig. 2B), we found that, unlike in control knockdown cells, both p21 and p27 failed to be degraded in TIM-depleted cells 11 h after nocodazole release, when most of the control knockdown cells were in S phase (Fig. 2, B and C).
TIM is important for replication fork stability and efficient DNA replication progression (21,22). Therefore, we tested the possibility that the delay in S phase entry observed in TIMdepleted G 1 cells was due to the presence of unreplicated or damaged DNA inherited from the previous cell cycle. Because DNA damage activates the ATR-CHK1-dependent checkpoint response during G 1 phase (33, 34), we analyzed the level of p-CHK1 in TIM knockdown cells. Consistent with a previous report that TIM facilitates efficient CHK1 activation (35), we found that UV irradiation induced CHK1 phosphorylation in both control and TIM knockdown cells. However, the amount of UV-induced p-CHK1 was reduced in TIM knockdown cells (Fig. 2D). As anticipated, TIM depletion alone did not activate CHK1 in the absence of UV treatment (Fig. 2D).
To further confirm that TIM depletion does not stimulate DNA damage, we asked whether ssDNA accumulates during G 1 phase. This ssDNA accumulation can occur because of either unreplicated or unresolved damaged DNA inherited from the previous cell cycle in TIM-depleted cells. To test this hypothesis, we carried out immunofluorescence microscopy using an anti-BrdU antibody to detect ssDNA foci formed under non-denaturing conditions in cells precultured with BrdU for 18 h. Under non-denaturing conditions, only the BrdU at ssDNA regions is recognized by the anti-BrdU antibody. ssDNA foci were normally visible only in cells synchronized to S phase (13 h) but not in those synchronized to G 1 phase (6 h; Fig. 2, E-G). Even though ssDNA-positive cells were also observed during G 1 phase when TIM was depleted (Fig.  2H), the number of ssDNA-positive G 1 cells remained low (Ͻ7%). These results indicate that the inefficient S phase entry seen in the majority of the TIM-depleted cells (Fig. 1C) is not due to the presence of unreplicated or damaged DNA. In addition, even though at 13 h post nocodazole release ssDNA foci were detected in TIM-depleted cells, the number of ssDNA- Right panels, flow cytometry to determine cell cycle distribution of asynchronized HEK293 (A) and U2OS (B) cells treated with control or TIM siRNAs. The cells were labeled with BrdU and co-stained for BrdU incorporation (y axis) and DNA content (propidium iodide, x axis). Percentages of cells in S phase are shown in black, and the differences in the percentages between control and TIM knockdown cells are shown in red. C, flow cytometry analysis of HEK293 cells treated with control or TIM siRNAs after release from nocodazole blocking at the indicated time points. 2C and 4C represent cells containing one or two copies of each chromosome, respectively. All data are representative of a minimum of two independent experiments. All Western blots in each subfigure were from the same lysate or experiment.
positive cells (Fig. 2E) and the number of ssDNA foci per cell were reduced in these cells compared with control knockdown cells (Fig. 2, compare I with G). Together, these results are consistent with reduced efficiency of DNA replication initiation rather than DNA damage.
Reduced Chromatin Association of CDC6 and Cyclin E in TIM-depleted Cells-We further determined the cause for the persistent presence of p21 and p27 in TIM-depleted cells. The degradation of p21/p27 during late G 1 phase is mediated by SCF SKP2 -dependent ubiquitination (36). We found that the SKP2 protein level was reduced in HEK293T, U2OS, and HT1080 cells (Fig. 2J). This reduction is most likely due to the fact that even though SKP2 was visible during M phase, it failed to accumulate during G 1 and S phases in TIM knockdown cells (Fig. 2C, fourth panel). APC/C is known to degrade SKP2 (37). However, the level of cyclin A, which is an APC/C substrate (38), remained similar to the level in control cells (Fig. 2C, fifth panel). Therefore, the low level of SKP2 in the absence of TIM is unlikely to result from increased APC/C activity.
In addition to APC/C-dependent degradation of SKP2, the cellular level of SKP2 is also controlled by E2F-dependent transcription (39 -41). Prior to replication initiation, E2F activity is FIGURE 2. TIM-deficient cells accumulate high levels of replication inhibitors without accumulating DNA damage. A, the protein levels of TIM, TIPIN, p21, and p27 were analyzed by SDS-PAGE and Western blotting of WCE from HEK293T cells, which contain the SV40 large T antigen, treated with either control or TIM siRNAs. Actin was used as the loading control. B, flow cytometry of HEK293T control (left panel) and TIM knockdown (right panel) cells after release from nocodazole block at the indicated time points. C, the protein levels of TIM, p21, p27, SKP2, and cyclin A were analyzed on Western blots of WCEs that were prepared from the synchronized control and TIM knockdown HEK293T cells shown in B. Actin was used as the loading control for SDS-PAGE. noc, nocodazole. D, the protein levels of TIM, CHK1, and p-CHK1 were analyzed by Western blotting using WCE prepared from control and TIM knockdown HEK293T cells with or without UV irradiation. E, quantification of the percentage of HEK293T cells (Ͼ200 cells/sample) containing ssDNA, based on positive staining of pulselabeled BrdU under non-denaturing conditions 2, 6, and 13 h after release from nocodazole treatment. Error bars represent S.D. F-I, flow cytometry (left panels) and representative immunofluorescent staining of the pulse-labeled BrdU (right panels) in control siRNA-treated HEK293T cells (F and G) and TIM siRNA-treated cells (H and I) shown in E. J, the protein levels of TIM, SKP2, and p27 were analyzed by SDS-PAGE and Western blotting of WCE from HEK293T, U2OS, or HT0180 cells treated with either control or TIM siRNAs. Actin was used as the loading control. 2C and 4C represents cell containing one or two copies of each chromosome, respectively. All data are representative of a minimum of two independent experiments. All Western blots in each subfigure were from the same lysate or experiment.
suppressed by the binding of Rb at promoters (42)(43)(44)(45). During late G 1 phase, cyclin E-CDK2 is recruited to the chromatin to phosphorylate Rb and disrupt the Rb-E2F interaction at promoters, allowing the activation of E2F-dependent transcription (42,46). Therefore, we analyzed the levels of chromatin-bound cyclin E and found that the amount of cyclin E on the chromatin was substantially reduced in TIM knockdown cells (Fig. 3A). Furthermore, the level of chromatin-bound CDC6, which is required for chromatin recruitment and the activation of cyclin E-CDK2 activity (47)(48)(49), was also reduced in TIM knockdown cells (Fig. 3A).
The chromatin-bound CDC6 is necessary for the recruitment of MCM2-7 to the active origins during G 1 phase and the subsequent interaction with CDC45 and GINS to form the CMG complex and initiate DNA replication (4, 6, 7). Therefore, in CDC6 siRNA knockdown cells, chromatin recruitment of the components of the CMG complex, such as SLD5 and MCM7, was reduced (Fig. 3B). Because the amount of chromatin-bound CDC6 was reduced in TIM-depleted cells, we expected that the recruitment of MCM2-7 helicase to the active origin would be compromised in TIM knockdown cells. Indeed, the amount of MCM5 found at the active lamin B2 origin was reduced in TIM knockdown cells compared with control knockdown cells synchronized to G 1 phase (Fig. 3, C-E). On the other hand, there was no significant change in the amount of origin-bound ORC ( Fig. 3C-E, ORC2), whose association with the origin is independent of CDC6 (50). In summary, the reduced levels of chromatin-bound CDC6 and cyclin E are consistent with a low level of SKP2, reduced recruitment of MCM2-7 helicase to the replication origin, the accumulation of p21/p27, and the inefficient S phase entry observed in TIM knockdown cells. and GINS (represented by SLD5) in the cytoplasmic (Cyt), nucleoplasmic (Nuc), and CB fractions prepared from HEK293T cells treated with control or CDC6 siRNA. C, Western blotting analysis of MCM5, ORC2, and active RNAPII isolated from ChIP in formaldehyde-cross-linked control and TIM knockdown cells. D and E, real-time PCR analysis of the relative abundance of MCM5, ORC2, and RNAPII bound to the lamin B2 origin in HEK293 cells (D) and HEK293T cells (E) 6 h after nocodazole release, when cells were predominantly in G 1 phase. ChIP using IgG alone was used as the background level, and the relative abundance of MCM5, ORC2, and RNAPII shown was after subtracting the background signal from IgG ChIP. Each value represents the mean Ϯ S.D. calculated from triplicate PCRs per one representative experiment. All data are representative of a minimum of two independent experiments. All Western blots in each subfigure were from the same lysate or experiment.
CMG Helicase Complexes Accumulate on Human Chromatin in Non-S Phase, TIM-depleted Cells-Surprisingly, even though the levels of the chromatin-bound CDC6 and the origin-bound MCM2-7 in TIM knockdown cells were reduced, we found that components of the CMG complex, including the GINS (SLD5 and PSF2) and CDC45, were still found on the chromatin (Fig. 3, A, fifth through seventh panels, and B). To further determine whether these chromatin-bound GINS and CDC45 proteins were part of CMG complexes, we purified the CMG complexes using FLAG-CDC45 from U2OS (Fig. 4A) and HEK293 (Fig. 4B) cells that were treated with either control or TIM siRNA. Unexpectedly, the protein-protein interactions among the components of the CMG complex in TIM knockdown cells were still detected (Fig. 4, A and B, compare the first and second columns). Given that CDC6 is required to establish the pre-replication complexes and subsequently assemble the CMG complexes in normal cells (5), we tested whether depleting CDC6 in TIM knockdown cells would eliminate these CMG complexes. When the cells were treated with CDC6 siRNA for 60 h, we found that all of the CMG complexes were abolished in these cells (Fig. 4, A and B, compare the first and third columns). However, depletion of CDC6 in TIM knockdown cells (e.g. TIM-CDC6 double knockdown) failed to eliminate the chromatin-bound CMG complexes (Fig. 4, A and B, compare the third and fourth columns). To exclude the possibility that the chromatin-bound CMG complexes in the CDC6-TIM double knockdown cells were due to an increase in the S phase population compared with CDC6-deficient cells, we analyzed the cell cycle distribution of HEK293 knockdown cells by flow cytometry. We found that the S phase population was decreased in TIM, CDC6, and TIM-CDC6 knockdown cells compared with control cells (Fig. 4C). In addition, even though CMG complexes were formed in TIM-CDC6 knockdown cells, these cells still showed severely disrupted cell cycle distributions similar to CDC6 knockdown cells (Fig. 4C). To further confirm that the accumulation of these CMG complexes on the chromatin were caused by TIM deficiency, we exogenously expressed an siRNA-resistant TIM construct in CDC6-TIM double knockdown cells. We found that the exogenously expressed TIM proteins stabilized endogenous TIPIN (Fig. 4B, second panel, right lane) and suppressed the CMG complexes (Fig. 4B, bottom four panels, compare the fifth and sixth lanes).
Because the accumulation of CMG complexes in TIM knockdown cells cannot be eliminated by CDC6 depletion, we  OCTOBER 21, 2016 • VOLUME 291 • NUMBER 43 next asked whether the chromatin association of these CMG complexes still takes place only in S phase. When we purified chromatin-bound FLAG-CDC45 from control knockdown HEK293T cells synchronized to either early G 1 (4 h) or S phase (12 h) after nocodazole release (Fig. 5A), we found that, as expected, MCM2-7 and GINS were co-purified with CDC45 during S phase but not during early G 1 phase (Fig. 5B). However, CMG complexes were detected in an abundant amount in both G 1 phase and S phase cells, when TIM was depleted (Fig. 5,  A and B), suggesting that the presence of CMG is no longer restricted to S phase when TIM is absent. Previously, we showed that human RECQ4 DNA helicase forms a chromatinspecific complex with MCM10, MCM2-7, CDC45, GINS, and TIM-TIPIN during S phase to promote efficient DNA replication initiation (51). Here we found that the interaction of RECQ4 and MCM10 with CMG also failed to be eliminated by CDC6 depletion in TIM knockdown cells (Figs. 4B and 5, C and  D). Importantly, in control knockdown cells, RECQ4 only interacted with CMG in cells harvested 11 h after nocodazole release, when a substantial portion of the cells were in S phase (Fig. 5, E, bottom panel, and F, fifth lane). However, in TIM knockdown cells, we found that an increasing amount of CMG was co-purified with FLAG-RECQ4 from the human chromatin as early as 2 h after nocodazole release, when cells were undergoing mitosis (Fig. 5, E, top panel, and F, second lane). Together, these results provide evidence that, in TIM knockdown cells, the CMG complexes are present on the chromatin even when they are not undergoing DNA synthesis.

TIMELESS and the DNA Replication Complex
The CMG Complexes in TIM-depleted Cells Exhibit Altered Patterns of MCM4 Phosphorylation-Despite the increased amount of the RECQ4-CMG helicase complex (Fig. 5F), TIM knockdown cells showed inefficient DNA synthesis. We hypothesized that the RECQ4-CMG replicative helicase complexes that accumulate in TIM knockdown cells could be catalytically defective. Indeed, when the helicase activity of equal amounts of CMG-containing FLAG-RECQ4 helicase complex purified from the TIM knockdown and control cells was compared (Fig. 5G), the amount of single-stranded dissociation product in the presence of the TIM knockdown RECQ4-CMG complex was significantly reduced compared with that of the control cell complex (Fig. 5H).
We next wished to determine the cause of the decreased catalytic activity of the CMG-containing helicase complex in TIM knockdown cells. It has been established that both CDK and DDK are required for the assembly and activation of the CMG complex at active replication origins and that MCM4 is a direct target of DDK and CDK during replication initiation (27,(52)(53)(54)(55). Indeed, chromatin-bound MCM4 was present in multiple modified forms that exhibited slower mobility on SDS-PAGE, and the presence of this modified MCM4 protein was abolished by treatment with inhibitors of either the CDC7 component of DDK or CDK (Fig. 6A). Given that, in the absence of TIM, the interactions of CDC45 and GINS with the MCM2-7 helicase are no longer restricted to S phase (Fig. 5, B and F), we next determined whether the CMG complexes that formed in TIM knockdown cells contained the expected post-translational modifications. First, we measured the levels of phosphorylated MCM4 by SDS-PAGE mobility shift in the purified chro-matin-bound FLAG-MCM7 complex. We found a reduction in phosphorylated MCM4 when TIM was depleted (Fig. 6B, fourth  panel). We used an antibody against phosphorylated Ser/Thr residues to further evaluate the presence of phosphorylated MCM4. Consistent with the observed change in MCM4 mobility, we also observed reduced levels of phosphorylated polypeptides with SDS-PAGE mobility between 100 and 150 kDa, where phosphorylated MCM4 was found in these purified complexes (Fig. 6B, fifth panel).
We next tested whether the altered phosphorylation pattern of the chromatin-bound MCM4 is due to the accumulation of stalled replication forks. It has been shown that replication fork arrest by hydroxyurea or APH leads to ATR-induced MCM4 hyperphosphorylation as part of the replication block checkpointresponse (56).Indeed,weobservedanincreaseinhyperphosphorylated MCM4 in APH-treated cells (Fig. 6C, compare the first and third lanes). However, this APH-induced hyperphosphorylation of MCM4 is distinct from the MCM4 phosphorylation pattern found in TIM knockdown cells (Fig. 5C, compare the second and third lanes). This APH-induced MCM4 hyperphosphorylation is not dependent on TIM, as this phenomenon was also observed in TIM knockdown cells treated with APH (Fig. 6C, fourth lane). This result indicates that the reduced phosphorylation of MCM4 on chromatin in the absence of TIM is not likely to be a consequence of replication fork arrest.
In yeast, multiple DDK-dependent phosphorylation sites, which are important for DNA replication, have been identified within the N terminus of MCM4 (27). In humans, the MCM4 N terminus contains three intrinsic DDK target sites (Ser or Thr followed by Asp or Glu (Ser/Thr-Asp/Glu)), and they are Ser 26 , Ser 131 , and Ser 141 (Fig. 6D, underlined residues). We found that a phosphomimetic mutant containing Ser-to-Glu mutations at the three intrinsic DDK sites (Fig. 6E, 3 E-D/E) failed to rescue the alternation of MCM4 phosphorylation levels in TIM knockdown cells (Fig. 6E, compare the four left lanes), suggesting that the phosphorylation of these intrinsic DDK target sites is not affected by TIM. In addition to the intrinsic DDK target sites, DDK can also target the first Ser/Thr of the MCM4 PG sites (phosphorylated Ser/Thr-Ser/Thr-Pro) when the second Ser/ Thr is already phosphorylated by CDK (26,27,57). Human MCM4 contains six PG sites within the N terminus: Ser 2 , Ser 6 , Ser 31 , Thr 53 , Ser 70 and Ser 87 (Fig. 6C, residues in bold). We found that phosphomimetic mutations at these PG sites (e.g. Fig. 6E, 6E-EP) or phosphodefective mutations (e.g. Fig. 6F, 6A-AP) eliminated the difference in MCM4 mobility on SDS-PAGE between control and TIM knockdown cells, suggesting that the chromatin-bound MCM2-7 helicase in the absence of TIM may not contain properly phosphorylated PG sites.
The PG phosphorylation events on MCM4, which are necessary for CMG assembly and DNA synthesis, are tightly regulated during the cell cycle (26,27). Restricting CDC45 and GINS to only interacting with the PG-phosphorylated MCM2-7 helicase ensures the correct timing of CMG activation and replication initiation in normal cells. Indeed, CDC45 normally interacts only with the MCM2-7 helicase, which contains hyperphosphorylated MCM4. This hyperphosphorylated MCM4 can be converted to unphosphorylated MCM4 by phosphatase, as demonstrated by FLAG-CDC45 immunopre-  (SLD5 and PSF2). G, Western blotting analysis of FLAG-RECQ4 complexes immunopurified from the chromatin fractions prepared from control cells and TIM siRNA knockdown cells using anti-FLAG antibody. H, helicase activity of equal amounts of FLAG-RECQ4 complex purified from control cells or TIM knockdown HEK293T cells assayed using 32 P-labeled splayed-arm substrates. The 32 P-labeled ssDNA products were visualized by autoradiography following neutral PAGE. All data are representative of a minimum of three independent experiments. All Western blots in each subfigure were from the same lysate or experiment. 2C and 4C represents cell containing one or two copies of each chromosome, respectively. cipitation (Fig. 6G, compare the first and third lanes). Because there is an excessive amount of inactive MCM2-7 protein on the chromatin (1, 2), the majority of the chromatin-bound MCM2-7, which was purified using FLAG-MCM7, was not hyperphosphorylated (Fig. 6G, second lane). Hence, the CDC45-associated, hyperphosphorylated MCM2-7 was a minor species of the total MCM2-7 helicases (Fig. 6G, compare the first and second lanes). When we analyzed the phosphorylation states of the CMG complexes purified from HEK293T, U2OS, and HEK293 cells stably expressing FLAG-CDC45, we found that the complexes purified from TIM knockdown cells contained differently modified forms of MCM4 with faster  (Fig. 6, H and I). Even though these CMG complexes were capable of interacting with DNA polymerases such as Pol ␦ (Fig. 6I, fifth panel), these CMG helicase complexes were not likely to support efficient DNA replication because of their altered phosphorylation states. Although we cannot exclude the possibility that other post-translational modifications on CMG complexes may be affected by TIM depletion, our data indicate that TIM is at least important for preventing the accumulation of CMG complexes that are not phosphorylated correctly at the PG sites.
The Interaction of TIM with MCM2-7 Is Independent of the Cell Cycle-The ability of TIM to prevent the accumulation of the aberrant CMG complexes in non-S phase cells suggests that TIM may associate with the MCM2-7 helicase independent of DNA synthesis. However, TIM is only known to be a component of the replication fork progression complex (22). Therefore, we looked at the kinetics of the interaction between TIM and MCM2-7 throughout the cell cycle. First, we found that the stable interaction of TIM-TIPIN with MCM2-7 takes place on the chromatin (Fig. 7A). Interestingly, when we examined a HEK293T cell line stably expressing FLAG-MCM7, we found that, although CDC45 and GINS primarily associate with the MCM2-7 helicase during S phase (11 h), TIM-TIPIN interacts with MCM2-7 on the chromatin throughout the cell cycle (Fig.  7, B and C). On the other hand, TIM was primarily enriched in the FLAG-CDC45 complex purified from S phase cells (11 h) compared with the other phases (Fig. 7, D and E), and this result further supports the interaction of TIM with MCM2-7 prior to the interaction between MCM2-7 and CDC45 to form CMG. Consistent with this finding, we showed that CDC6 was also co-purified with TIM in the CB fraction (Fig. 7A). In addition, MCM2-7 could still be co-purified with TIM and TIPIN in the absence of CDC6, RECQ4, or MCM10 (Fig. 7, F and G), all of which are proteins that interact with MCM2-7 during G 1 phase and G 1 /S transition to facilitate CMG assembly in human cells (51,58). Therefore, we concluded that the interaction of TIM-TIPIN with MCM2-7 is independent of the cell cycle.

Discussion
This is the first report that demonstrates a function for human TIM in ensuring the correct chromatin association of the CMG complexes during the cell cycle (Fig. 8). During the normal cell cycle, CMG is only present on the chromatin during S phase to initiate DNA replication and allow cells to progress through S phase (5). We showed that TIM and TIPIN stably interact with the MCM2-7 helicase throughout the cell cycle (Fig. 7) and that TIM is important for preventing aberrant CMG complexes from accumulating on the chromatin when cells are not in S phase (Fig. 5, B and F). As a consequence, in the absence of TIM, aberrant CMG complexes containing MCM4, with altered phosphorylation at PG sites (Fig. 6), accumulate on the chromatin even in non-S phase cells (Fig. 8). Despite their abundance in TIM knockdown cells, these abnormal CMG complexes are likely to be catalytically altered (Fig. 5H).
What is the molecular mechanism by which TIM prevents the accumulation of aberrant CMG complexes in non-S phase cells? After DNA synthesis is completed, CMG is disassembled by CDC48-dependent ubiquitination (59). It is possible that The purified complexes were analyzed with the indicated antibodies to detect the presence of MCM2-7 (MCM4 and MCM7), GINS (PSF2 and SLD5), and TIM. F, the protein levels of RECQ4, MCM10, and CDC6 were analyzed by Western blotting using WCE prepared from control, RECQ4, MCM10, and CDC6 knockdown cells stably expressing FLAG-TIM. G, Western blotting analysis of FLAG-TIM complexes immunopurified from the soluble chromatin fractions of the corresponding knockdown cells described in A to detect the presence of TIM (FLAG), TIPIN, MCM2-7 (MCM7), MCM10, and RECQ4. All data are representatives of a minimum of two independent experiments. All Western blots in each subfigure were from the same lysate or experiment. 2C and 4C represents cell containing one or two copies of each chromosome, respectively. OCTOBER 21, 2016 • VOLUME 291 • NUMBER 43 TIM has a role in the disassembly or dissociation of the CMG complexes from the chromatin by facilitating or preserving this modification (Fig. 8). In TIM-deficient cells, the persistent chromatin association of the CMG complexes after DNA replication may continue to the next cell cycle, leading to the appearance of these CMG complexes on the chromatin prior to the next round of DNA replication. Because CDC6 is not known to be involved in the disassembly of the CMG complex, CDC6 depletion fails to eliminate these aberrant CMG complexes accumulated on the DNA in TIM knockdown cells (Fig.  4). Even though these CMG complexes remain on the DNA, they might still be targeted for dephosphorylation, leading to the altered phosphorylation pattern observed on the CMG complexes (Fig. 6). Alternatively, we cannot exclude the possibility that, in the absence of TIM, the accumulation of the CMG complex in the non-S phase cells may be a consequence of aberrant MCM2-7 chromatin loading or CMG assembly. In cells, CDC6 has been implicated in loading MCM2-7 at both active origins and the excess dormant origins, and this may be achieved by CDC6 continuously loading MCM2-7 at the active origin and releasing the loaded MCM2-7 to allow the helicase to slide away from the active origin (60,61). However, in vitro, purified MCM2-7 helicase alone is capable of binding to DNA (53,62). Furthermore, interactions among MCM2-7, CDC45, and GINS can form in the absence of CDC6 in vitro using recombinant proteins (53,63). Most likely, negative regulatory mechanisms to minimize CDC6-independent DNA binding of MCM2-7 and CMG assembly may exist in cells. Indeed, multiple phosphorylation events that inhibit MCM2-7 binding to DNA have been reported (64,65). In addition, in yeast, the SLD2 and SLD3 initiation factors interact with MCM2-7 helicase to block its interaction with CDC45 and GINS, and phosphorylation of SLD2 by CDK removes the block, thereby allowing the CMG complex to assemble (63,66,67). It is possible that TIM interacts with MCM2-7 prior to S phase to ensure that the MCM2-7 helicase can only be loaded onto the DNA via a CDC6-dependent pathway to form proper CMG complexes. In addition, mammalian TIM-TIPIN can interact with the MCM3, 4, 5, 6, and 7 subunits (9), whereas CDC45 and GINS interact with MCM2, 3, and 5 (68). It is plausible that TIM-TIPIN behaves similarly to yeast SLD2 and SLD3 and that the interactions of MCM2-7 with TIM-TIPIN and CDC45/GINS are normally mutually exclusive but can simultaneously interact with the MCM2-7 helicase after post-translational modification when cells enter S phase. Hence, in the absence of TIM, the interactions of MCM2-7 with CDC45 and GINS are not regulated and may form prior to DNA synthesis (Fig. 8). If so, it will be interesting to determine whether the interface of the MCM2-7 helicase that interacts with TIM-TIPIN can be modified at active replicons by CDK or DDK in a CDC6-dependent manner to allow the simultaneous interaction of CDC45 and GINS with MCM2-7 when TIM-TIPIN is present.

TIMELESS and the DNA Replication Complex
We further demonstrated that, in addition to preventing the accumulation of aberrant CMG complexes in non-S-phase cells, TIM contributes to the stable chromatin association of CDC6 and cyclin E (Fig. 3A). We propose the following possible mechanism by which TIM promotes the association of CDC6 with chromatin. CDC6 binding to DNA is tightly regulated to avoid re-replication during each round of the cell cycle. Possible mechanisms to prevent CDC6 from binding to DNA after rep- In normal cells, TIM-TIPIN is known to travel with the CMG complex to maintain proper replication fork progression (left). However, TIM-TIPIN also interacts with MCM2-7 outside of S phase. We suggest that, after completion of DNA synthesis, the interaction of TIM-TIPIN with MCM2-7 may be responsible for the disassembly and dissociation of the CMG from the chromatin prior to mitosis or entry to the next round of the cell cycle. In the absence of TIM, aberrant CMG complexes are present on the chromatin during G 1 phase (right). The presence of these abnormal CMG complexes may provide false signals to prevent CDC6-dependent CMG assembly at active origins and initiate DNA replication. lication initiation include nuclear export and cyclin A-dependent phosphorylation (4). It is possible that the assembly of CMG complexes may also provide a signal to release CDC6 from DNA for nuclear export to avoid re-replication. If so, the presence of the aberrant CMG complexes on the chromatin during G 1 phase could provide false signals that either prevent CDC6 from binding to chromatin or weaken its DNA binding (Fig. 8).
When cells enter G 1 phase of the cell cycle, CDC6 is recruited by ORC to the origin of replication (69), and this is followed by the chromatin recruitment of the MCM2-7 helicase and CDK2cyclin E (Fig. 8) (70, 71). The chromatin binding and activation of CDK2-cyclin E provide a positive feedback loop that further enriches CDC6 on the DNA. This enrichment of CDC6 on the DNA can be achieved by protecting CDC6 from APC/C-dependent degradation (72) and activating the E2F transcription factor (73), which further amplifies the expression of CDC6 (74,75). E2F-dependent transcription is also necessary for accumulation of SKP2, a component of the SCF SKP2 ubiquitin ligase, which promotes replication initiation by degrading p27 and p21 (36) during late G 1 phase (39 -41, 76). The reduced chromatin loading of CDC6 observed in TIM knockdown cells may explain why these cells also exhibited decreased levels of DNAbound cyclin E (Fig. 3A), reduced recruitment of MCM2-7 helicase to the active replication origin (Figs. 3D and 8), reduced MCM4 phosphorylation (Fig. 6), low levels of SKP2 in G 1 and S phases (Fig. 2C), high levels of p21 and p27 throughout the cell cycle (Figs. 1, A and B, and 2A), and a delay in S phase entry (Figs. 1C and 2B). The presence of these altered CMG helicase complexes may also explain the slow rate of replication fork progression that is observed in TIM knockdown cells when the cells start DNA synthesis (12).
Cell Culture, siRNA, Cell Fractionation, and Immunopurification-The human cell lines HEK293T and HT1080 were generous gifts from Dr. Stephen West (The Francis Crick Institute). The human cell lines HEK293 and U2OS were obtained from Dr. Jeremy Stark (City of Hope). The cells were cultured in DMEM supplemented with 10% fetal bovine serum at 37°C in a 5% CO 2 incubator. Plasmid transfection, chromatin fractionation, and immunopurification were performed as described previously (51). The siRNA sequences for RECQ4, MCM10, and CDC6 were also described previously (51). The TIM stealth siRNAs #1 (5Ј-GGUUCGAGAGAUGACUGAGG TIMELESS and the DNA Replication Complex OCTOBER 21, 2016 • VOLUME 291 • NUMBER 43 GCUAU-3Ј (11)) and #2 (5Ј-UCCAGGGUAGCUU AGUC-CUUUCAAA-3Ј) were purchased from Invitrogen. The siRNAs were transfected into the cells using DharmaFECT 1 transfection reagent (Thermo Scientific) according to the protocol of the manufacturer. For TIM siRNA or TIM-CDC6 double siRNA knockdown, the cells were treated with 25 nM siRNA for 24 h followed by a second 25 nM siRNA treatment. The cells were harvested 36 h after the second siRNA transfection. For cell cycle synchronization, the cells were cultured with nocodazole-containing medium 24 h after the second siRNA transfection. For UV treatment, cells in log phase were irradiated with 30 J/m 2 UV light and harvested after 1 h of incubation. For APH treatment, the cells were cultured with 1 g/ml of APH for 24 h prior to harvest. For CDC7 and CDK inhibitor treatments, PHA 767491 (5 M, Santa Cruz Biotechnology) and roscovitine (20 M, Sigma) were added to the tissue culture media to a final concentration as indicated for 24 h prior to harvest.
Cell Cycle and ssDNA Analyses-The cells were synchronized at G 2 /M phase with 50 ng/ml nocodazole in complete medium for 20 h. The cells were released by washing twice with complete DMEM. For DNA content analysis, the cells were stained with propidium iodide (50 g/ml) and analyzed in a CyAn ADP Analyzer (Beckman Coulter). For the BrdU incorporation assay, the cells were pulse-labeled with 10 M BrdU (Sigma) for 30 min. The BrdU-positive cells were detected using a rat anti-BrdU antibody followed by a goat anti-rat DyLight 488-conjugated secondary antibody. The cell cycle profile distributions were determined with the FlowJo software package (Tree Star Inc.). ssDNA detection was performed as described previously (78). Briefly, the siRNA-transfected cells were grown on coverslips and labeled with 10 M BrdU for 18 h, followed by nocodazole synchronization for 18 h. After nocodazole release, the cells were fixed with 3% paraformaldehyde containing 0.2% Triton X-100 for 15 min, followed by blocking with blocking buffer (3% BSA and 0.5% Tween 20 in PBS) for 30 min. The coverslips were then incubated overnight with a rat anti-BrdU antibody (1:500), diluted in blocking buffer overnight at 4°C, washed with PBS, stained with anti-rat IgG conjugated with Rhodamine (1:1000) for 2 h at room temperature, washed with PBS, stained with DAPI, and mounted in anti-fade mounting medium.
Phosphatase Treatment-The co-purified FLAG-CDC45 and FLAG-MCM7 complexes were eluted with 3ϫ FLAG peptide in Elution-A buffer (10 mM HEPES (pH 7.9), 0.2 M NaCl, 0.2 mM EDTA, 0.05% Triton X-100, and 10% glycerol). The eluates were then dialyzed against dialysis buffer (10 mM Tris (pH 7.5), 100 mM NaCl, 0.2 mM EDTA, and 10% glycerol). An aliquot was incubated in 30 l of the buffer supplied by the manufacturer with or without 400 U of protein phosphatase (New England Biolabs). The samples were separated by SDS-PAGE and detected on Western blots using an anti-MCM4 antibody.
ChIP and Real-time PCR-The cells were cross-linked with 1% formaldehyde for 10 min at room temperature, and the reac-tion was stopped by adding glycine to a final concentration of 0.125 M for 5 min at room temperature. The fixed cells were rinsed twice with PBS and lysed with cytoplasmic lysis buffer (10 mM Tris (pH 8.0), 0.34 M sucrose, 3 mM CaCl 2 , 2 mM MgCl 2 , 0.1 mM EDTA, 1 mM DTT, and 0.5% Nonidet P-40) plus protease inhibitors. Lysates were pelleted by centrifugation at 5000 rpm for 5 min at 4°C. The pellets were resuspended in ChIP lysis buffer (1.0% SDS, 10 mM EDTA, and 50 mM Tris (pH 8.0)) plus protease inhibitors, and chromatin was sheared by sonication to generate DNA fragments of Ͻ1 kb. Chromatin was diluted 10 times in ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris (pH 8.0), and 167 mM NaCl) plus protease inhibitor and precleared with protein A/G beads (Thermo Scientific) for 1 h at 4°C. Antibodies (ORC2, MCM5, and RNA polymerase II) were used for immunoprecipitation. A rabbit IgG (Santa Cruz Biotechnology) was used as a negative control. ChIP complexes were collected with protein A/G beads and washed sequentially twice with low-salt buffer (0.1% SDS, 1.0% Triton X-100, 2 mM EDTA, 20 mM Tris (pH 8.0), and 0.15 M NaCl), high-salt buffer (0.1% SDS, 1.0% Triton X-100, 2 mM EDTA, 20 mM Tris (pH 8.0), and 0.5 M NaCl), LiCl buffer (0.25 M LiCl, 1.0% Nonidet P-40, 1% sodium deoxycholate, 1 mM EDTA, and 10 mM Tris (pH 8.0)), and Tris-EDTA buffer. The ChIP complexes were eluted with 300 l of elution buffer (0.1 M sodium bicarbonate and 1.0% SDS) at room temperature for 15 min. The chromatin was reverse cross-linked by adding 20 l of 5 M NaCl and incubated at 65°C overnight. The DNA was digested with RNase A and proteinase K and purified by phenol-chloroform and ethanol precipitation. Real-time PCR was performed to amplify the human Lamin B2 origin region (B48) (79) with an ABI 7500 fast system using SYBR Green fluorescence. Enrichment was calculated using the comparative Ct method.
Author Contributions-X. X. conducted most of the experiments and analyzed the results. J. T. W. contributed to the analysis of CMG complex formation. M. L. contributed to the cell cycle analysis. Y. L. conceived the idea for the project, analyzed the results with X. X., and wrote the paper.