DNA Replication Determines Timing of Mitosis by Restricting CDK1 and PLK1 Activation

1Cell proliferation is critical to the propagation of life and in eukaryotes is based on mitotic cell division. Mitotic cell division requires two basic steps: first, cells need to replicate their DNA, and second, cells need to divide the replicated genome to daughter cells. Strict separation of these steps in time is critical as division of partially duplicated chromosomes causes genome instability and cell death. Furthermore, activation of mitotic factors during DNA replication causes replication fork instability and DNA damage. How cells coordinate genome duplication and mitosis thus constitutes a fundamental question in cell cycle control, which has major consequences for both proliferation and genome stability.

2Almost 30 years ago, Hartwell and Weinert introduced the concept of a checkpoint and described it as a safety mechanism that prevents cell division if DNA replication is not complete. However, empirical data supporting an intrinsic DNA replication checkpoint proved hard to obtain, as treatments to block DNA replication fork progression are prone to cause DNA damage, evoking secondary responses that halt mitotic progression. The concept of a checkpoint soon expanded and became increasingly used to depict a cell cycle block in response to DNA damage and other cellular stresses. Stress-induced checkpoints have been widely studied, which has led to the identification of well-conserved signaling kinases, such as CHK1 and p38. Meanwhile, the discovery of autonomous biochemical oscillators in yeast cells and de-nucleated Xenopus egg extracts provided alternative models that could explain orderly cell cycle progression without the need of a DNA replication checkpoint. That is, self-regulating oscillators could in principle ensure sufficient time to finish DNA replication before mitosis is triggered.

3The key to understand how DNA replication and cell division are separated is to understand how cell division is initiated. Mitotic entry is driven by the kinases CDK1 and PLK1, which through multiple feedback loops enhance each other’s activity. Activation of CDK1 depends on CDK2-mediated accumulation of mitotic inducers, such as PLK1, which at a threshold level can induce a switch-like activation and mitotic entry. The threshold is defined by factors opposing CDK1 activation, including PP1 and PP2A phosphatases and Wee1/Myt1 kinases. In addition, checkpoint kinases as CHK1 and p38-MK2 can raise the activation threshold through phosphorylation of multiple CDK regulators, effectively suppressing CDK activity in the presence as well as absence of exogenous DNA damage. However, if and how unperturbed DNA replication is integrated into this regulatory circuit remains unknown. We previously showed that both CDK1 and PLK1 are activated when the bulk of DNA replication is completed at the S/G2 transition. These observations resurged the prospect of the original checkpoint model and prompted us to test whether DNA replication restricts mitotic kinase activation (Figure 1A).

4Here, we generated a double-degron system to rapidly deplete the essential DNA replication-initiation factor CDC6 and show that untransformed human cells shorten the cell cycle and prematurely enter mitosis in the absence of DNA replication. Using RNAi and inhibitors to independently target DNA replication licensing or firing, we find similar results in cancer cells. We also find that abrogating CHK1 activity in transformed cells, or CHK1 and p38 activity in untransformed cells, enhances CDK activation specifically upon G1/S transition, supporting the notion of a DNA replication checkpoint in human cells. While checkpoint inhibition causes premature activation of CDK2, PLK1, and CDK1, we find the latter most harmful and causing severe DNA replication stress.

5Our results indicate that DNA replication is not only an output but also an integral signaling component of the cell cycle oscillator and is a major determinant for the timing of mitosis. DNA replication, through checkpoint kinase signaling, coordinates the cell cycle by allowing a slow buildup of CDK2 activity while suppressing CDK1 and PLK1 activation until S phase completion.

6In order to simultaneously monitor DNA replication and mitotic kinase activation, we generated untransformed but immortalized retinal pigment epithelial (RPE) cell lines expressing a marker for DNA replication (proliferating cell nuclear antigen [PCNA]-cb) and a sensor for PLK1 target phosphorylation (PLK1-FRET). In agreement with our previous findings in U2OS cancer cells, PLK1 target phosphorylation remained undetectable during active DNA replication and only appeared when the bulk of PCNA foci disappeared (Figure 1B). The close correlation between PCNA foci disappearance and Plk1-mediated phosphorylation was retained when completion of DNA replication was delayed by addition of thymidine (Figures 1C, S1A, and S1B). To test PLK1-dependent phosphorylation of an endogenous target, we next assessed TCTP-pS46 staining by quantitative immunofluorescence. To estimate temporal kinetics from fixed asynchronous cell populations, we sorted single cells based on cyclin A2 content, DNA content, and nuclear size. DNA replication was assessed by measuring EdU nucleotide incorporation immediately before fixation, and PLK1 activity was determined by PLK1-dependent TCTP-pS46 staining using a low dose of the PLK1 inhibitor BI2536 (Figure S2A). We note that there is a close correlation between the disappearance of EdU staining and the appearance of BI2536-sensitive TCTP-pS46 staining in U2OS, RPE, and early-passage primary fibroblasts (Figures 1D, S2A, and S2B). The robust coupling between completion of S phase and activation of PLK1 as assessed by both fixed and live readouts suggests that PLK1 activation can function as a marker for the S/G2 transition in human cells.

7To assess whether PLK1 activation is restricted by DNA replication, we sought to analyze cells that progressed through the cell cycle without replicating DNA. In yeast, mutants in the DNA replication licensing factor Cdc6 can efficiently suppress DNA replication without blocking cell cycle progression. We therefore set out to generate a genetic system that would allow rapid degradation of CDC6 in untransformed human cells. We used p53−/− RPE cells to avoid a licensing checkpoint and labeled both endogenous CDC6 alleles with an optimized double-degron tag. The double degron is based on two protein degradation tags: mAID and SMASh, allowing rapid depletion of endogenously tagged CDC6d (Figures 2A, 2B, S3A, and S3B; STAR Methods). CDC6d degradation using both mAID and SMASh prevented EdU incorporation more efficiently than using mAID or SMASh alone, showing an additive effect of both degradation tags on suppressing DNA replication (Figures 2C and S3B). After 48 hr CDC6d degradation, cyclin A2 expression and CDK-mediated phosphorylation of lamin A/C pS22 was not reduced, indicating that cells remained engaged in the cell cycle (Figures S3C and S3D). Whereas mitotic cells were abundant, they showed half the DNA content of their control counterparts and few anaphase figures, as expected for cells entering mitosis without sister chromatids (Figures S3E–S3G). Taken together, our results show that the CDC6d double-degron system can be used to suppress DNA replication.

8We next pursued to study cell cycle progression in the absence of DNA replication. We aimed to prevent initiation of replication, as perturbation of ongoing replication leads to DNA damage and cell cycle arrest. Meanwhile, to avoid secondary effects due to continued proliferation with incomplete DNA replication, we sought to detect the immediate effects of suppressed initiation of DNA replication. To this end, we combined CDC6d degradation and a CDC7 inhibitor, targeting both origin licensing and origin firing, the two main events necessary for initiation of DNA replication. DNA replication licensing by CDC6 can in principle start as soon as CDK activities are reset at anaphase. Cells depleted of CDC6d before mitotic exit should thus be impaired in DNA replication initiation. In contrast, G1 or S phase cells already harbor licensed origins and thus should remain DNA replication competent despite CDC6d degradation. To test this prediction, we monitored EdU incorporation between 10 and 12 hr after targeting CDC6, CDC7, or both CDC6 and CDC7. We reasoned that, at this time, G2 cells would have progressed through mitosis and recently entered S phase, whereas G1 cells would have progressed to late S phase (Figures 2D and 2E). To independently monitor DNA replication and cell cycle position, we plotted EdU incorporation versus cyclin A2 content of single cells. The gradual accumulation of cyclin A2 levels from the G1/S border allows identification of cell cycle position within S phase. As predicted, EdU incorporation was effectively abrogated in early S phase cells, whereas it was sustained in late S phase cells (Figures 2D and 2E). This shows that our setup efficiently prevents initiation of DNA replication, while allowing ongoing DNA replication to continue.

9To study the direct consequence of failed DNA replication initiation, we used time-lapse microscopy to follow single cells that entered mitosis shortly after destruction of CDC6d and inhibition of CDC7 (Figures 3A and S4; Video S1). In parallel, we treated cells with EdU and stained for cyclin A to detect efficiency of DNA replication in S phase (Figures 3A–3C). Combined CDC6d depletion and CDC7 inhibition decreased median cell cycle duration by ∼4 hr (−25%) and allowed cells to re-enter mitosis as early as 6–8 hr after cell division (Figure 3D). Initiation of DNA replication thus constitutes a major brake on cell cycle progression.

10We find that only a limited amount of DNA replication is sufficient to delay the cell cycle. After CDC6d depletion or CDC7 inhibition alone, EdU incorporation is suppressed but still detectable in early S phase cells, while double treatment effectively abolished new synthesis of DNA (Figures 3B and 3C). The need for combined CDC6d depletion and CDC7 inhibition to efficiently suppress DNA replication could in part be due to compensatory mechanisms that enhance firing in genomic regions where origins are sparse or due to the fact that CDK activity can complement CDC7 to fire a restricted set of origins. Importantly, CDC6d depletion or CDC7 inhibition alone postponed mitotic entry, increasing median cell cycle duration by ∼10 hr (+70%; Figure 3D). The limited availability of licensed or fired origins is expected to delay DNA replication completion, which in turn could postpone mitotic entry. Our results are in agreement with a threshold model in which low levels of DNA replication are sufficient to prolong the cell cycle while a complete block of DNA replication leads to premature mitosis.

11To study the impact of DNA replication on activation of PLK1, we targeted origin licensing (by depleting licensing factors CDC6 and CDT1) or origin firing (by combined CDC45 depletion and CDC7 inhibition) in U2OS PLK1-FRET PCNA-cb cells. We find that, after interfering with either DNA replication licensing or origin firing, a subset of cells showed PLK1 activation 4–6 hr after mitosis (Figures 4A, 4B, and S5A). Similarly, some cells entered mitosis 6–8 hr after cell division, in contrast to control cells, in which the shortest time between two cell divisions was 15 hr. Premature PLK1 activation correlated to the absence of PCNA-cb foci and abolished EdU incorporation in S phase cells, indicating that premature PLK1 activation depends on suppression of DNA replication (Figures 4A–4D and S5B). In agreement with a threshold model, intermediate reduction of EdU incorporation or PCNA foci delayed PLK1 activation compared to control cells (Figure 4C). Our data suggest that DNA replication restricts activation of the mitotic kinase PLK1 and that completion of DNA replication functions as a trigger for PLK1 activation.

12The function of replication intermediates as direct inhibitors of mitosis via constitutive checkpoint activation is a long-standing yet unproven hypothesis in the cell cycle field. Our finding that DNA replication initiation postpones cell cycle progression predicts the need for factors that can sense ongoing DNA replication and suppress the activity of mitotic kinases. Good candidates for such signaling factors are DNA damage checkpoint kinases, as they are known to control DNA replication fidelity during unperturbed growth and inhibit PLK1 activity upon ectopic DNA damage. Indeed, we find that two structurally unrelated CHK1 inhibitors cause premature PLK1 activation in replicating U2OS cells (Figure S6A). Notably, CHK1 inhibition did not affect progression through G1 phase, but uncoupled nuclear PLK1 target phosphorylation from on-going DNA replication, causing interphase cells to reach high PLK1-FRET levels in the presence of PCNA-cb foci (Figures 5A–5D). Similarly, a live-cell CDK1/2-activity sensor rapidly became saturated in U2OS cells containing PCNA-cb foci (Figures S6B and S6C). CHK1 can function synergistically with the p38-MK2 axis, and combined CHK1/p38 inhibition increased CDK1/2-mediated phosphorylation of lamin A/C in RPE cells (Figure 5E). These data support the notion that mitotic kinases are actively suppressed during S phase.

13Loss of CDK1/2 regulation is highly genotoxic and causes DNA replication stress. Accordingly, inhibition of CHK1/p38 leads to phosphorylation of the DNA damage marker H2AX specifically in replicating cells (Figures 5E, 5F, and S6D). A short pulse of CDK activity is likely sufficient to induce DNA damage, as contrary to CDK inhibition 1 hr after CHK1/p38 inhibition, simultaneous addition of all inhibitors abolished H2AX phosphorylation (Figures 5E, 5G, S6E, and S6F). Whereas addition of a CDK2 inhibitor decreased EdU incorporation in S phase cells, inhibition of CDK1 suppressed phosphorylation of H2AX (Figures 5G and S6F). Although this does not rule out overlapping functions of CDK1 and CDK2, it suggests that CDK1 activity is the major toxic component for replicating cells. Thus, CDK2 activity stimulates DNA replication, which in turn activates checkpoint kinases to restrict CDK1 and PLK1 activity. As a consequence, DNA replication is part of a feedforward system that restricts mitotic kinase activation to prevent replication stress.

14A prediction of such a system is that DNA replication continuously restricts CDK activity. We therefore sought to monitor CDK activity as cells entered S phase. A live sensor for CDK1/2 activity shows a gradual increase, starting from before S phase is initiated (; Figure 6A). In the presence of CHK1 inhibition, we detect no difference in CDK1/2 sensor localization through G1 phase (Figures 6A–6C). However, we detect enhanced CDK activity from the moment cells enter S phase as estimated by appearance of PCNA-cb foci (Figures 6A and 6B) or by expression of an anaphase-promoting complex (APC)/CCdh1 substrate reporter (; Figure 6C). Similarly, blocking DNA replication licensing or DNA replication firing causes elevated CDK activity upon S phase entry (Figures 6D–6F). We also detect elevated CDK1/2-mediated phosphorylation of Lamin A/C pS22 in cells in which DNA replication initiation is prevented by targeted destruction of CDC6d (Figures S7A–S7C). Thus, rather than merely a consequence of cell cycle signaling, DNA replication coordinates gradual CDK activation throughout S phase with activation of mitotic kinases PLK1 and CDK1 after completion of S phase (Figure 7).

15Separation of DNA replication and cell division lies at the heart of the cell cycle. Several not mutually exclusive models for how cells ensure completion of DNA replication before mitosis have been proposed (reviewed in ). These models include (1) a slow buildup of CDK activity coupled to different thresholds for starting DNA replication and mitosis, (2) different cyclin proteins appear at given times and cellular locations, (3) transcriptional oscillator networks (; but see also ), and (4) a checkpoint that restricts mitosis before DNA replication is completed. Our results suggest that, in human cells, DNA replication is a major determinant of cell cycle duration and timing of mitosis. We propose that DNA replication functions as a continuous brake with the possibility to separate and coordinate different cell cycle activities.

16We find that severe suppression of DNA replication initiation is required to prematurely activate mitotic kinases and to advance mitotic entry. For instance, CDC7 inhibition or CDC6 degradation alone postpones mitotic entry, while combined treatment advances mitotic entry (Figure 3). This suggests that a threshold of replication activity exists, under which mitotic entry is not prevented. DNA replication in human cells is performed by thousands of replication forks. How many replication forks are required to reach this threshold is not clear. However, it is likely that a single replication fork is not sufficient to block mitotic entry as Plk1 activation is observed in the presence of occasional weak PCNA-cb foci and DNA synthesis at fragile sites has been detected in mitotic cells.

17While complete DNA replication is critical for proper chromosome segregation in mitosis, DNA replication per se is not a prerequisite for mitotic entry. Yeast mutants lacking CDC6, CDT1, ORC2, RFC5, CDC45, or CDC7/DBF4 enter mitosis with unreplicated chromatids. Whether mitotic entry in these studies depends on gene-specific functions independent of DNA replication initiation has been debated. Given that all these proteins have in common to affect initiation of DNA replication, we consider that the most likely explanation involves initiation of DNA replication. In support of this notion, we detect premature mitotic kinase activation and shorter cell cycles in the absence of DNA replication origin licensing and/or origin firing in human cells.

18Once initiated, CDK2 activity will promote both DNA replication and stimulate CDK1 and PLK1 activity. However, at the same time, CDK1 and PLK1 are restricted by DNA replication, constituting an incoherent feedforward loop (Figure 7). Given that CDK1 activation causes DNA damage during DNA replication, there is a possibility that the feedforward loop is complemented by feedback between CDK1 and DNA replication, ensuring a continuous low level of checkpoint activation. We note that a feedforward system not only ensures that mitosis is avoided until DNA replication is completed but also can ensure that mitotic kinase activation commences immediately after S phase, ensuring a tight coupling between S phase and mitosis.

19The feedforward loop between CDK2, DNA replication, and mitotic kinase activation contributes to genome stability in two ways. First, it ensures that mitosis is delayed until the bulk of DNA is duplicated. Second, it prevents DNA replication stress and CDK1-mediated DNA damage. In this sense, complete deregulation of the feedforward loop likely is lethal, whereas slight perturbations have the possibility to contribute to CDK-dependent DNA lesions in S phase and might promote the occurrence of underreplicated DNA in mitosis. Such perturbations may not only have consequences for mitosis but also for G1 duration and cell fate of daughter cells.

20Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Arne Lindqvist (arne.lindqvist@ki.se).

21Source of cell lines used in this study is reported in the Key Resources Table.

22Human hTERT-RPE1 (hereafter referred to as RPE), U2OS and BJ cells were cultured in an ambient-controlled incubator at 37°C and 5% CO2. Early passage BJ cells were purchased from ATCC (CRL-2522) and examined prior to population doubling 10. RPE and U2OS cells were a kind gift from Dr. René Medema. All cell lines were regularly tested for mycoplasma. RPE cells were cultured using DMEM-F12 GlutaMAX (Invitrogen) supplemented with 1% Pen/Strep (HyClone) and 10% heat-inactivated FBS (HyClone). U2OS and BJ cells were cultured using DMEM GlutaMAX (Invitrogen) supplemented with 1% Pen/Strep (HyClone) and 6% or 10% heat-inactivated fetal bovine serum (FBS, HyClone), resp.

23Reporter plasmids pCCC-TagRFP (Chromotek, PCNA-cb), pCSII-EF-Cdk2probe or pRetroX-SG2M-Red (Clontech, 631465) were used to obtain stable cell lines by random plasmid integration. Plasmid transfection was performed using LipofectAMINE 2000 Reagent (Life Technologies) or standard calcium phosphate precipitation. Stable polyclonal populations were selected by flow cytometry using BD-FACSAria II and correct expression patterns were validated by live-cell microscopy.

24TP53 knockout RPE cells were from. Stable OsTIR expression was established by CRISPR-mediated integration: Transient GFP+ cells were sorted 3 days post transfection of pROSA26-DV1_OsTIR and PX458_Cas9_GFP (Addgene 48138) and subsequently subjected to bleomycin selection. For inducible expression of OSTir1 we used a previously described construct that was combined with a bleomycin selection cassette and cloned into a Rosa26 targeting construct to generate pROSA26-DV1_OsTIR. To target the human Rosa26 locus we used the following gRNA sequence: 5′ GACCTGCTACAGGCACTCGT 3′ cloned into pPV576_gRNA_ROSA26. Subsequent double degron tagging at the human CDC6 locus was established through CRISPR-mediated targeting: First a CDC6 guide sequence was cloned into PX458_Cas9_GFP (Addgene 48138) using BbSI sites and two annealed oligos CACCGC CAGCTG AATACT TTCGGG and AAACCC CGAAAG TATTCA GCTGGC. Second, a tagging vector was generated named hCDC6_mAID_SMASh_T2A_Neo, by cloning CDC6 homology arms (gBlock IDT) into a mAID_SMASh_T2A_Neo backbone using SbfI and NotI sites. The mAID_SMASh_T2A_Neo backbone was constructed via Gibson assembly cloning.

25The resultant PX458_Cas9_GFP_hCDC6 and hCDC6_mAID_SMASh_T2A_Neo vectors were sequence-verified and transfected into OsTIR-RPEP53KO cells. Three days post transfection GFP+ cells were sorted and subjected to Geneticin selection (G418; GIBCO). Clonal cell lines were established by limiting dilution in a 96-well plate. Successful Cdc6 degron-tagging in single clones was verified by western blot and degron-induced growth arrest.

26Knockdown studies were performed using SMARTpool scrambled control siRNA (Dharmacon), SMARTpool ON-TARGET plus CDC6 siRNA (Dharmacon), SMARTpool ON-TARGET plus CDC45 siRNA (Dharmacon), FlexiTube Hs_CDT1_7 (QIAGEN) and FlexiTube Hs_CDT1_8 (QIAGEN). siRNAs were transfected at 20nM concentration using HiPerFect reagent (QIAGEN) and OptiMEM (Invitrogen). The following small-molecule inhibitors were used at the indicated final concentrations: 50 nM CHIR-124 (CHK1 inhibitor; Selleck Chemicals), 30nM UCN-01 (CHK1 inhibitor; Sigma), 5 uM RO-3306 (Cdk1 inhibitor; Calbiochem), 10 uM NU6140 (Cdk2 inhibitor; Calbiochem), 25uM Roscovitine (Cdk1/2 inhibitor; Selleck Chemicals), 3uM CDK1/2 inhibitor III (Millipore),10 uM SB202190 (p38 inhibitor; Selleck Chemicals), 50 nM BI2536 (PLK1 inhibitor, Selleck Chemicals), 200nM GSK461364 (PLK1 inhibitor; Selleck Chemicals), 2.5 mM Thymidine (Sigma Aldrich), 15 uM XL-413 (CDC7 inhibitor; Selleck Chemicals), 100 nM BMS-650032 (Asunaprevir; Adooq Bioscience), 50 uM 3-Indoleacetic acid (Auxin; Sigma Aldrich), 1ug/ml Doxycycline (Sigma Aldrich).

27For time-lapse microscopy of RPE CDC6d cells (Figure 2D), transmitted light images were acquired every 10 min using a Leica DMI6000 Imaging System with 20x (NA 0.4) air objective at 37°C and 5% CO2. Images were analyzed using ImageJ and cells were tracked manually. For live cell fluorescence imaging experiments, 6000-9000 cells were seeded in 96-well imaging plates (BD Falcon) in DMEM with 1% Pen/Strep and 6%–10% FBS and at least 6 hr prior to imaging the medium was changed to Leibowitz-15 (Invitrogen) with 1% Pen/Strep and 6%–10% FBS. Images were acquired at 37°C at 20-45min intervals using ImageXpress microscope (Molecular Devices) 20X objective (NA 0.45) and analyzed using custom ImageJ or MATLAB scripts. PLK1 FRET was measured using ImageJ, using rolling ball background subtraction, 3x YFP signal multiplication and RatioPlus plugin (background 0 and clipping values 25-50). Normalized PCNA-cb intensity variation (i.e., standard deviation of nuclear RFP signal divided by median nuclear RFP intensity) is used as a surrogate measurement for DNA replication; the last few bright PCNA foci at the end of S-phase cause a characteristic peak in PCNA-cb intensity variation (Figures 1C and 4C). CDK1/2 activity probe was analyzed using ImageJ; nuclear ROIs were selected manually and ratio between nuclear YFP intensities and cytoplasmic YFP intensities was determined using custom ImageJ macro’s and Excel (see Data S1 for example macro). For quantitative IF experiments 6000-9000 cells were seeded in 96-well imaging plates (BD Falcon) in DMEM with 1% Pen/Strep and 6%–10% FBS and after treatment washed in DPBS (GIBCO), fixed in formalin (Sigma Aldrich) for 7 min, permeabilised in cold methanol (Sigma Aldrich) for 2 min and finally incubated in blocking media (TBS supplemented with 0.1% Tween20 and 2% bovine albumin serum) for 1 hr at room temperature. For pulse-chase experiments or DNA replication measurements 10mM EdU (5-ethynyl-2′-deoxyuridine, Molecular probes) was added to live cells prior to fixation. Fixed samples were incubated with primary antibodies in blocking media overnight at 4°C, washed in DPBS (GIBCO) and incubated with secondary antibodies and DAPI for 1h at room temperature. Samples were stored in PBS and in case of DNA replication measurements EdU-Click chemistry was performed by incubation in 100mM Tris, 1mM CuSO4, 100mM ascorbic acid and fluorescent dye azide (#A10277, Invitrogen) for 1h at room temperature. Images were acquired at room temperature using ImageXpress microscope (Molecular Devices) 20X objective (NA 0.45) and analyzed using custom Cell profiler pipelines. To assess kinetics and cell cycle stages from quantitative IF data, cells were ordered and distributed linearly based on increasing nuclear Cyclin A2 intensities and normalized DAPI levels. For RPE and BJ cells in Figure 1C, cells containing low DAPI levels (corresponding to under mid S-phase levels) were sorted based on the ratio between nuclear size and roundness, and remaining cells were sorted by Cyclin A2 content.

28The following antibodies were used for IF and diluted in TBS supplemented with 0.1% Tween20 and 2% BSA: CyclinA2 (1:400; #sc-751; Santa Cruz), CyclinA2 (1:400; #4656; Cell Signaling), PLK1 (1:400; ab14210; Abcam), pTCTP (1:400; #5251; Cell Signaling), pH2AX (1:800; #2577; Cell Signaling), pCyclinB1 (1:400; #ab55184; Abcam), pLaminA/C (1:400; #2026; Cell Signaling), Alexa Fluor 488-Goat anti-Rabbit (1:2000; #A11008 Life Technologies) and Alexa Fluor 555-Goat anti-Mouse (1: 2000; #A21422 Life Technologies). The following antibodies were used for WB and diluted in TBS supplemented with 0.1% Tween20 and 5% milk powder: CDC6 (1:400; #sc-9964 Santa Cruz), CDT1 (1:400; #sc-28262 Santa Cruz), GAPDH (1:20000; #G8795; Sigma), anti-Rabbit HRP (1:2000; #ab6721; Abcam) and anti-Mouse HRP (1:2000; #ab97023; Abcam).

29Statistical analysis was performed using the Student’s t test. Quantitative IF studies included >200 cells per conditions and represent at least two independent experiments. Qualitative example images are representative for >10 independent biological replicates. Information about the statistical tests and the number of cells analyzed is provided in the figure legends. For all experiments, samples were not randomized and the investigators were not blinded to the group allocation during experiments and outcome assessment.

30Original imaging data have been deposited to Mendeley Data and are available at https://doi.org/10.17632/47s44zbm7c.1.