PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of elifeeLifeRecent contentAbout eLifeFor authorsSign up for alerts
 
eLife. 2017; 6: e29953.
PMCID: PMC5736350

The CDK-PLK1 axis targets the DNA damage checkpoint sensor protein RAD9 to promote cell proliferation and tolerance to genotoxic stress

Marcus B Smolka, Reviewing Editor
Marcus B Smolka, Cornell University, United States;

Abstract

Genotoxic stress causes proliferating cells to activate the DNA damage checkpoint, to assist DNA damage recovery by slowing cell cycle progression. Thus, to drive proliferation, cells must tolerate DNA damage and suppress the checkpoint response. However, the mechanism underlying this negative regulation of checkpoint activation is still elusive. We show that human Cyclin-Dependent-Kinases (CDKs) target the RAD9 subunit of the 9-1-1 checkpoint clamp on Thr292, to modulate DNA damage checkpoint activation. Thr292 phosphorylation on RAD9 creates a binding site for Polo-Like-Kinase1 (PLK1), which phosphorylates RAD9 on Thr313. These CDK-PLK1-dependent phosphorylations of RAD9 suppress checkpoint activation, therefore maintaining high DNA synthesis rates during DNA replication stress. Our results suggest that CDK locally initiates a PLK1-dependent signaling response that antagonizes the ability of the DNA damage checkpoint to detect DNA damage. These findings provide a mechanism for the suppression of DNA damage checkpoint signaling, to promote cell proliferation under genotoxic stress conditions.

Research organism: Human

Introduction

To proliferate and maintain the integrity of biological systems, cells possess the ability to overcome environmental stresses. Genomic DNA damage is a type of environmental stress, and when cells are exposed to genotoxic stress, cell cycle progression is delayed by a mechanism called the DNA damage checkpoint. Although the delay increases the time available for DNA repair processes to perform damage recovery (Elledge, 1996; Weinert and Hartwell, 1988), the degree of the checkpoint response must be tightly regulated, depending on the proliferative status of the cells. Notably, if cells must proliferate, then the checkpoint response should be down-regulated to tolerate the cellular DNA damage stresses. This may be accomplished via the interplay between the pro-mitotic activity and the DNA damage checkpoint in the cell, although the precise mechanism remains enigmatic (Serrano and D'Amours, 2014).

Among the pro-mitotic activities, CDKs are the primary driving forces that order and coordinate cell cycle events (Swaffer et al., 2016). Activation by CDKs not only initiates M-phase events, but also controls temporally ordered programs throughout the cell cycle, such as replication origin firing upon initiation of DNA replication, which potentially support proliferation (Stern and Nurse, 1996; Tanaka et al., 2007; Trovesi et al., 2013; Zegerman and Diffley, 2007). However, problems may arise once DNA replication is initiated. The DNA replication forks can encounter damaged DNA or environmental stress-mediated nucleotide depletion, and such events can stall replication polymerases at the fork (Jossen and Bermejo, 2013). The stalled replication forks expose single-stranded DNA, which is recognized by checkpoint signaling complexes that activate the S-phase checkpoint (Giannattasio and Branzei, 2017; Katou et al., 2003; Sogo et al., 2002; Zou and Elledge, 2003; Zou et al., 2003). When the DNA damage response is dominant, checkpoint signaling reduces the rate of S-phase progression by repressing late origin firing or by slowing the progression of individual replication forks. Together, these events can reduce genomic instability in proliferating cells, and thereby provide tumor-suppressor functions (Iyer and Rhind, 2013; Zegerman and Diffley, 2009).

Interestingly, under some circumstances, pro-mitotic activity dominates DNA damage checkpoint signaling, and cells commence proliferation even in the presence of damaged DNA. The key protein involved in the pro-mitotic activity to override DNA damage checkpoint signaling is PLK1. This phenomenon was initially reported as adaptation, and was originally described in a budding yeast system in which cells overrode a G2 DNA damage checkpoint via a CDC5 (budding yeast PLK1 orthologue)-dependent mechanism, after prolonged G2 arrest induced by irreparable DNA double-strand breaks (DSBs) (Pellicioli et al., 2001; Sandell and Zakian, 1993; Toczyski et al., 1997). Adaptation was also reported in a Xenopus system, in which a prolonged aphidicolin-induced replication block was compromised by a Plx1 (Xenopus PLK1 orthologue)-dependent process (Yoo et al., 2004), and an analogous phenomenon has also been reported in mammalian cells. However, mammalian cells only activate the G2/M DNA damage checkpoint when a certain amount of DNA damage (e.g. ~20 DSBs) is present. Indeed, mammalian cells can enter mitosis even in the presence of DNA damage signals, such as γ-H2AX foci (Deckbar et al., 2007; Ishikawa et al., 2010; Syljuåsen et al., 2006). In this sense, mammalian cells can somehow repress a DNA damage checkpoint mechanism to tolerate the DNA damage response in order to drive proliferation, and PLK1 exhibits the key pro-mitotic activity for this purpose. When the cellular PLK1 activity reaches a certain level, the cells can re-enter mitosis upon recovery from G2 checkpoint arrest (Liang et al., 2014). However, in the case of the DNA damage checkpoint in S-phase, the crosstalk between PLK1 and the DNA damage response becomes more complicated. In fact, PLK1 functions not only upon mitotic commitment, but also during S phase or a related DNA damage response, thereby facilitating DNA metabolism in support of rapid cell proliferation (Moudry et al., 2016; Yata et al., 2012). Importantly, the PLK1 protein contains a Polo-Box Domain (PBD) that recognizes a pre-phosphorylated substrate (Elia et al., 2003). The substrate is phosphorylated by a kinase, typically a CDK, and the PBD preferentially recognizes phospho-serine/threonine (-S-pS/pT-P-)-containing peptides (van Vugt et al., 2010).

In this study, to clarify how proliferating cells override the DNA damage checkpoint in S-phase, we revisited the CDK-dependent phosphorylation in the checkpoint processes, using human cells. We found that CDK targeted RAD9, a subunit of the DNA clamp complex 9-1-1, which acts as a DNA damage sensor protein complex in checkpoint signaling (Delacroix et al., 2007; Furuya et al., 2004; Kumagai et al., 2004; Lee et al., 2007; Liu et al., 2012; Navadgi-Patil and Burgers, 2009; St Onge et al., 2003; Takeishi et al., 2010). If CDK failed to phosphorylate RAD9, then the cells accumulated in S phase in the presence of low doses of hydroxyurea (HU). The CDK-dependent phosphorylation of RAD9 at residue 292 (pThr292) created a binding site for PLK1, and the bound PLK1 was then activated to further phosphorylate RAD9 on Thr313 and Ser326. We also found that these CDK- and PLK1-dependent phosphorylations of RAD9 were required to maintain the rate of S-phase progression in cells exposed to low-dose HU, and this correlated with the mutant phenotype in which increased amounts of active checkpoint complexes accumulated on the damaged chromatin. Additionally, mutant RAD9 proteins that could not be phosphorylated by CDK-PLK1 were also associated with the damaged chromatin. These results implied that the CDK-/PLK1-dependent control of DNA damage detection by RAD9 plays a crucial role in minimizing the DNA damage checkpoint response, thus helping to drive proliferation under conditions of DNA replicative stress.

Results

CDK-dependent phosphorylation of RAD9 at Thr292

To understand how CDK controls the DNA damage checkpoint, we focused on the human checkpoint protein RAD9, which forms a PCNA-like hetero-trimeric checkpoint clamp complex. This complex (designated as 9-1-1) functions as a DNA damage sensor by binding to damaged DNA ends (Caspari et al., 2000; Majka et al., 2006; Thelen et al., 1999), and this function must be precisely regulated to control the degree of checkpoint activation. Indeed, using the fission yeast (Schizosaccharomyces pombe) system, we previously identified multiple phosphorylation events on the S. pombe Rad9 homologue (spRad9) that regulate the checkpoint activation and the spRad9 release from damaged chromatin (Furuya et al., 2010, 2004). In the present study, we focused on Thr292 (-His-Ser-292Thr-Pro-) of the human RAD9 homologue (Figure 1A), because the residue resembles Thr321 (-His-Ser-Ser-321Thr-Pro-) of SpRad9, which when phosphorylated promotes the release of SpRad9 from DNA damage sites.

Figure 1.
CDK phosphorylates threonine 292 of RAD9.

The phosphorylation of RAD9 Thr292 was originally identified by Davey’s group (St Onge et al., 2003). As this Thr residue is followed by a proline residue and the region resembles a putative CDK phosphorylation site, we first tested whether the phosphorylation of Thr292 depended on CDK. RAD9 with a C-terminal myc-His (mH) tag was transiently expressed in HEK293A cells, and the cell lysates were subjected to fractionation by centrifugation in sucrose-containing buffer (Figure 1—figure supplement 1A)(Ohashi et al., 2014; Zou et al., 2002). Western blotting was performed using an antibody that recognizes phosphorylated Thr292 (pThr292). We detected the pThr292 signal primarily in the supernatant fraction (Figure 1—figure supplement 1A). The pThr292 band corresponded to the uppermost band of the RAD9-myc signal, as previously reported (Figure 1—figure supplement 1A) (St Onge et al., 2003). The pThr292 signal was diminished in roscovitine (CDK inhibitor)-treated cells, although the band was still detectable in RO-3306 (CDK1 inhibitor)-treated cells (Meijer et al., 1997; Vassilev et al., 2006). These results suggested that any CDK would be able to phosphorylate Thr292 of RAD9. Importantly, the pThr292 signal was diminished when a mutant RAD9, with Thr292 substituted with alanine (T292A), was expressed in place of wild-type RAD9 (Figure 1—figure supplement 1A).

We further tested the ability of CDK to phosphorylate RAD9 Thr292 in an in vitro kinase assay, using a recombinant CDK2-cyclinA2 complex. The C-terminal region of RAD9 (a.a. 266–391) was fused to GST, purified from E. coli cells by glutathione affinity chromatography, and assayed to detect CDK-dependent phosphorylation in vitro. pThr292 was detected as efficiently as pSer277 by western blotting (Figure 1B) (St Onge et al., 2003). We also confirmed that the phosphorylation at Thr292 was not dependent on other CDK phosphorylation sites (Ser277, Ser328, Ser336) that are responsible for the major bandshift of GST-RAD9 in vitro (St Onge et al., 2003) (data not shown).

Next, to assess the behavior and effects of the CDK-dependent phosphorylation of RAD9 in vivo, we constructed stable HEK293A cell lines that harbor a CMV-tetO-RAD9-mH construct (genomically integrated at the FRT locus) and therefore express wild-type or Thr292-mutated (T292A) RAD9-mH. For the in vivo cell line experiment, we used RAD9-S291A/T292A in place of RAD9-T292A, and these two mutant proteins were treated similarly throughout the manuscript. RAD9-mH was expressed when doxycycline was added to the medium (Figure 1—figure supplement 1B). Although RAD9-mH was expressed at a level approximately five times higher than endogenous RAD9, we assumed that this increased level of ectopic expression did not affect the protein’s usual cellular activity. Extra copies of RAD9 reportedly associate with, and are possibly sequestered by, the CAD (carbamoyl-phosphate synthetase) protein, which does not associate with RAD1-HUS1 (Lindsey-Boltz et al., 2004), and thus the expression of the mutant RAD9 would replace the endogenous RAD9 as a component of the 9-1-1 complexes.

A thymidine block and release was performed to synchronize the cells in G1/S, and the cell cycle profile of pThr292 was monitored. After the cells were released from the thymidine-induced G1/S block, the cells that expressed RAD9-mH from either CMV-tetO-RAD9-mH or CMV-tetO-RAD9-S291A/T292A-mH were harvested and subjected to a western blotting analysis, using anti-pThr292 (pT292: Figure 1C). The phosphorylation of Thr292 was observed weakly from mid S phase (6–8 hr in Figure 1C) and strongly at the G2/M transition, a pattern that correlated with the CDK activity. In combination with the above results, these experiments confirmed that CDK phosphorylates the Thr292 residue of RAD9 both in vitro and in vivo.

Checkpoint signaling is enhanced upon RAD9-T292A expression

We observed a slight delay in the mitotic commitment, as detected by the delayed appearance of the phosphorylation of histone H3 Ser10, in cells expressing T292A-mutated RAD9 (Figure 1C, bottom). This implied a role of the CDK-dependent phosphorylation (pThr 292) in inhibiting the RAD9 activity for the DNA damage checkpoint.

Next, we sought to determine how the CDK-dependent phosphorylation of RAD9 affected the checkpoint-activating processes at the molecular level. To do so, we monitored the amount of chromatin-bound checkpoint proteins in HEK293A-T-REx cells expressing either wild-type or T292A-mutated RAD9 under replicative stress (1.5 mM hydroxyurea), by performing a fractionation assay followed by a western blotting analysis (Figure 1D) (Ohashi et al., 2014; Zou et al., 2002). Tubulin and H2AX were specifically detected in the Triton X-100 soluble (S1) and insoluble (P1/P2) fractions, respectively, suggesting that the fractionation functioned appropriately. The majority of the RAD9-myc was found in the S1 fraction, as previously reported, and approximately 10% of the RAD9-myc was found in the insoluble fraction. The proportion of wild-type RAD9 (WT) in the insoluble fraction decreased after the cells were exposed to HU (see ‘+HU’ in Figure 1D). In contrast, the T292A-mutated RAD9 (T292A) remained in the insoluble fraction when the cells were exposed to HU. A previous report showed that upon the activation of the checkpoint, the 9-1-1 complex is loaded onto DNA damage sites by the RAD17 clamp loader complex and the DNA damage signal is transmitted to the CHK1 kinase (Zou et al., 2002). Indeed, higher levels of RAD17 and the Ser345-phosphorylated CHK1 accumulated in the insoluble fraction of HU-treated cells expressing the T292A-mutated RAD9 (+HU, compare WT and T292A in Figure 1D). The increased amounts of checkpoint components in the chromatin fraction suggested that the checkpoint is enhanced in cells expressing the T292A-mutated RAD9.

S-phase progression is slowed in cells exposed to low-dose HU when CDK fails to phosphorylate RAD9 Thr292

The enhanced association of RAD9 and phosphorylated CHK1 with chromatin in the presence of HU suggested the ectopic activation of the intra-S-phase checkpoint, when CDK failed to phosphorylate RAD9 at Thr292. The intra-S-phase checkpoint is activated when cells are under dNTP-depletion stress or when replication forks encounter DNA damage (Paulovich and Hartwell, 1995). The activation of this checkpoint delays the timing of late origin firing, which is a CDK-dependent process (Dimitrova and Gilbert, 2000; Santocanale and Diffley, 1998; Shirahige et al., 1998; Tanaka and Araki, 2010; Yekezare et al., 2013).

To detect S-phase progression in cells, we used U2OS-T-REx cells stably expressing wild-type RAD9 or T292A-mutated RAD9 from the FRT locus (Figure 1—figure supplement 1C). We performed these experiments in the presence of 0.2 mM HU, a lower concentration of HU that allowed the control U2OS cells to grow with a mild S-phase delay. Here, we assessed the cell cycle delay by flow cytometry, to measure the cellular DNA content (Figure 1E). In the absence of HU, the U2OS cells that expressed wild-type RAD9 (WT) and those that expressed the T292A-mutated RAD9 (T292A) behaved similarly to the parental U2OS-T-REx cells. However, in presence of 0.2 mM hydroxyurea, although the cells expressing wild-type RAD9 showed similar profiles to the parental U2OS cells, the cells expressing the T292A-mutated RAD9 exhibited an increased population of S-phase cells (Figure 1E, WT: 35.8%, T292A: 43.2%).

The accumulation of S-phase cells could be due to a slower rate of DNA synthesis. Therefore, we quantified the rate of dNTP incorporation by EdU (ethynyl-deoxy-uridine)-labeling (Salic and Mitchison, 2008). Without HU treatment, both cell lines (expressing wild-type or T292A-mutated RAD9) showed similar amounts of EdU incorporation during S phase. However, when these cell lines were treated with 0.2 mM HU for 24 hr, the amount of EdU incorporation in the cells expressing the T292A-mutated RAD9 was reduced to 60–70% of that in the HU-exposed cells expressing wild-type RAD9 (details shown later in Figure 4B, compare WT and T292A). This result suggested that the higher rate of genomic DNA replication under dNTP depletion stress was maintained through the CDK-dependent phosphorylation of Thr292 of RAD9.

PLK1 binds to, and is activated by, the phosphorylated Thr292 of RAD9

Next, we explored the biochemical consequences of the CDK-dependent phosphorylation of Thr292 (pThr292) of RAD9. We noticed that the peptide sequence around pThr292 (289-Ala-His-Ser-pThr-Pro-293) resembles the cognate sequence for PLK1 binding (Elia et al., 2003). PLK1 is a kinase with a C-terminal phospho-peptide-recognizing motif called the PBD, which is known to bind to a Ser-pSer/pThr-Pro consensus motif. Indeed, we found that the PBD of PLK1, but not those of PLKs2-5, efficiently bound to the C-terminus of RAD9 in a yeast two-hybrid assay (Figure 2—figure supplement 1A). We further demonstrated that this binding decreased when Thr292 of RAD9 was substituted with alanine (Figure 2—figure supplement 1A).

We then performed an in vitro ‘pull-down’ experiment using recombinant GST-RAD9 protein (C-terminus: a.a. 266–391) and PLK1. In the course of the assay, the GST-RAD9 protein was phosphorylated by a recombinant CDK2-cyclinA2 complex. PLK1 co-precipitated efficiently using GSH-beads, but only when CDK2-cyclinA2 was present (Figure 2A and B). Notably, this experiment employed a kinase buffer containing ATP, and we observed a bandshift of GST-RAD9 only when PLK1 was present (lane three in Figure 2A, lane one in Figure 2B). Consistent with the notion that PLK1 binds to and phosphorylates pre-phosphorylated substrates, we observed enriched Thr292 phosphorylation (pT292) in the bandshifted form of GST-RAD9 (Figure 2A). We should also note that hereafter, we used GST-RAD9 in which S277, S328 and S336 were each substituted with alanine; the resulting GST-RAD9 (3A) construct facilitated the recognition of PLK1-induced phosphorylation by the bandshift (Figure 2—figure supplement 1B).

Figure 2.
CDK-dependent phosphorylation of threonine 292 of RAD9 accommodates and activates PLK1.

If PLK1 binds to Thr292-phosphorylated RAD9, then the binding should be competed with a phospho-peptide that contains pThr292. Therefore, we incubated PLK1 and GST-RAD9 together with various phospho-peptides, and monitored the association of PLK1 with the GSH-bead fraction. PLK1 remained in the supernatant fraction when incubated with a pThr292-containing peptide (pT292), but not when incubated with other peptides (Non-P, pS291, pS291/pT292) or in the absence of a competing peptide (Figure 2C; compare Sup and Beads).

To further confirm the specific interaction between pThr292 and PLK1, we tested whether a pThr292-containing peptide could activate PLK1. The PBD can reportedly act as an auto-inhibitor of the PLK1 kinase domain, and the binding of a substrate to the PBD counteracts this auto-inhibitory function (Jang et al., 2002). Here, we used GST-RAD9 (3A)-T292A (which PLK1 cannot bind) as the substrate for kinase assays using PLK1 (Figure 2D). The bandshift in the anti-RAD9 western blot was only observed in the presence of a pThr292-containing peptide, and the use of other peptides in the reaction did not yield a bandshift (Figure 2D). These data suggested that a pThr292-containing phospho-peptide binds to and ‘activates’ PLK1, to promote the further phosphorylation of RAD9.

The association of PLK1 with RAD9 was confirmed by the cell line experiments. Using anti-myc-coated agarose beads, co-immunoprecipitation (co-IP) experiments were performed using cell lysates from the RAD9-mH-expressing HEK293A-T-REx cell line. First, we tested the co-IP of transiently-expressed PLK1-HA (Figure 2E). The co-IP of PLK1-HA was observed regardless of UV damage stress in cells that expressed wild-type RAD9-mH, and was greatly diminished in cells that expressed the T292A-mutated RAD9-mH (Figure 2E). We further confirmed the co-IP of endogenous PLK1. The co-IP of PLK1 was also observed in cells that expressed wild-type RAD9-mH, but was decreased in cells that expressed the T292A-mutated RAD9-mH (Figure 2F). Together, these results confirmed that PLK1 specifically targets RAD9 phosphorylated (at Thr292) by CDK, and this targeting was demonstrated to occur in vitro and possibly in vivo. In summary, the CDK-dependent phosphorylation of Thr292 of RAD9 is expected to provide a docking site on RAD9, and to act as a local activator for the further phosphorylation of RAD9 by PLK1.

PLK1 phosphorylates RAD9 at Thr313 and Ser326

Having established the conditions under which PLK1, but not PLK1-KR (kinase-deficient), phosphorylates CDK-pretreated RAD9 (Figure 3A), we sought to identify the RAD9 sites that are phosphorylated by PLK1. The phosphorylated bands of RAD9 were subjected to an LC-MS/MS analysis. As shown in Figure 3B, significant amounts of pThr313 and pSer326 were detected in GST-RAD9, when it was incubated with both CDK2-CycA2 and PLK1. pThr312 was relatively weakly detected, and its increase upon a co-incubation with PLK1 was not as dramatic as those observed for pThr313 and pSer326. Furthermore, the Thr313 and Ser326 residues of RAD9 are followed by hydrophobic amino acids, and thus Thr313 and Ser326 match the previously reported PLK1 phosphorylation consensus sequence (Nakajima et al., 2003). These results indicated that RAD9 Thr313 and Ser326 constitute sites that are phosphorylated by PLK1.

Figure 3.
PLK1 phosphorylates RAD9 on Thr 313 and Ser 326 in vivo and in vitro.

The region around the Thr313 phosphorylation site is conserved among the vertebrate RAD9A proteins (In this manuscript, RAD9 is used synonymously with RAD9A unless otherwise stated.) (Figure 3—figure supplement 1A). The RAD9A proteins are thought to be the mitotic paralogues of RAD9 that diverged from an ancestral Rad9. Interestingly, the RAD9B proteins, which are paralogues that are thought to function in testis and/or to have distinct functions (Dufault et al., 2003; Leloup et al., 2010; Pérez-Castro and Freire, 2012), lack the PLK1 phosphorylation sites identified here (Figure 3—figure supplement 1B). These observations suggested that the interaction with CDK-PLK1 may provide the critical difference in the regulation between these RAD9 paralogues.

To further confirm the PLK1-dependent phosphorylation, we raised antibodies against pThr313- and pSer326-containing peptides (Figure 3—figure supplement 1C). These antibodies, α-pThr313 (pT313) and α-pSer326 (pS326), detected GST-RAD9 that had been incubated with PLK1 in kinase buffer (Figure 3—figure supplement 1D), and the western blotting signals were largely diminished when the mutant GST-RAD9 proteins, harboring either the T313A or S326A mutation, were used as substrates in the PLK1 kinase assay (Figure 3—figure supplement 1E). To confirm whether the phosphorylation at Thr313 and Ser326 occurred by PLK1 binding to RAD9, through the CDK-dependent phosphorylation of Thr292, the PLK1 kinase assay with GST-RAD9 was performed in the absence or presence of CDK2-cyclinA2. As demonstrated by western blotting, the appearance of both the pThr313 and pSer326 signals depended on the presence of CDK2-cyclinA2 (Figure 3C, quantification in Figure 3—figure supplement 1F). Furthermore, large decreases in the pThr313 and pSer326 western blot signals were observed when GST-RAD9-T292A was used as the substrate in the PLK1 kinase assay (Figure 3D, quantification in Figure 3—figure supplement 1F).

PLK1 phosphorylates Thr313 and Ser326 of RAD9 in vivo

Next, we confirmed the PLK1-dependent phosphorylation of Thr313 and Ser326 of RAD9 in cells. First, to emphasize the PLK1 action on RAD9, PLK1 was ectopically expressed in asynchronously growing HEK293A cells that harbored a CMV-tetO-RAD9-mH construct (encoding RAD9-mH controlled by a doxycycline-inducible promoter) at the FRT locus. In addition, the phosphatase inhibitor calyculin A was added to the cells, prior to harvesting. The treatment with calyculin A forces the cells to enter into and arrest in M phase, a condition that should also activate PLK1 (Coco-Martin and Begg, 1997; Ishihara et al., 1989). The cell lysates were subjected to IP using α-myc antibody-coated agarose beads, and the immunoprecipitates were analyzed by western blotting. When the western blot was probed using the antibody against pThr313, the signals were seen with the expression of PLK1 (Figure 3—figure supplement 2A, lane 2). These signals completely disappeared when the PLK1-KR mutant was expressed and the cells were treated with the PLK1 inhibitor BI2536 (Figure 3—figure supplement 2A, lane 3). Together, these results suggested that PLK1 mediates the phosphorylation of Thr313 of RAD9. When the blots were instead probed with the antibody against pSer326, again the signals were detected when PLK1 was expressed, but these signals were decreased only slightly when PLK1-KR was expressed and the cells were treated with BI2536 (Figure 3—figure supplement 2A, compare lanes 2 and 3). This suggested that Ser326 can be phosphorylated by kinases other than PLK1. In this experiment, we also observed that the levels of both pThr313 and pSer326 were only slightly, if at all, decreased in the cells that expressed the RAD9-S291A/T292A mutant (Figure 3—figure supplement 2A, lane 4), in contrast to the results obtained in the in vitro experiment. This distinction may be due to the strong activation and over-expression of PLK1 in calyculin A-treated cells, which may obviate the need for Thr292 phosphorylation to locally activate PLK1 for RAD9 phosphorylation.

Next, to determine whether endogenous PLK1 could phosphorylate RAD9, we employed asynchronously growing HEK293 cells harboring the CMV-tetO-RAD9-mH construct at the FRT locus. These cells were exposed to UV irradiation, in the presence or absence of the PLK1 inhibitor BI2536, before harvesting, and cell lysates were prepared. RAD9-mH was immunoprecipitated with α-myc antibody-coated agarose beads, and monitored by western blotting using α-RAD9, α-pThr292, α-pThr313, and α-pSer326 antibodies (Figure 3E). The signal recognized by α-RAD9 remained unchanged regardless of the UV irradiation dose, BI2536 treatment, or T292A mutation. However, the pThr292 signal, which was relatively unchanged in the UV-irradiated cells, was increased in the BI2536-treated cells (Figure 3E, IP; lanes 5 and 6), implying that this CDK-dependent phosphorylation event (pThr292) occurred upstream of the PLK1-dependent phosphorylation. We observed diminished pThr313 levels when the cells were treated with BI2536 (Figure 3E, lanes 5 and 6), or when T292A-mutated RAD9 was expressed (Figure 3E, lanes 7, 8 and 9) instead of wild-type RAD9 (WT). These results indicated that the generation of pThr313 is dependent on PLK1, which (as demonstrated above) is bound to and activated by pThr292 of RAD9. Interestingly, the pThr313 signal increased slightly when the cells were UV-irradiated (Figure 3E, lanes 3 and 4). We also observed that the pSer326 signal increased upon UV irradiation, and a slight decrease was detected when the cells were treated with BI2536 (Figure 3E, lanes 5 and 6). However, we did not see a decrease of pSer326 in the T292A-mutated RAD9 (Figure 3E, lanes 7, 8 and 9). Although (as shown above) pSer326 can be phosphorylated by PLK1, we postulate that other kinases can substitute for PLK1 for this function, permitting the phosphorylation of Ser326 without the prerequisite phosphorylation of Thr292.

Next, we analyzed whether DNA replicative stress could also affect the PLK1-dependent phosphorylation of RAD9. HEK293A cells harboring the CMV-tetO-RAD9-mH construct at the FRT locus were synchronized by a double-thymidine block, and the released cells were treated with 0.2 mM HU. The cell lysates were subjected to IP to monitor the phosphorylation of RAD9 (Figure 3—figure supplement 2B). Although the total amounts of (RAD9), pThr313 (pT313), and pSer 326 (pS326) were relatively unchanged by the HU exposure, the amount of pThr292 (pT292) was somewhat decreased. These data suggested that pThr292 efficiently triggers PLK1-dependent phosphorylation upon UV- or HU-induced stress.

We tested the effects of other types of DNA damage on the PLK1-dependent phosphorylation of RAD9. Treatments with camptothecin (CPT: topoI inhibitor), ionizing radiation (IR), or high doses of methylmethanesulfonate (MMS: 0.03%) or mitomycin C (MMC: 1 µg/ml) yielded slight increases in the pSer326 phosphorylation (Figure 3—figure supplement 3A and B). Aphidicolin (Aph: a DNA polymerase inhibitor) treatment reduced the levels of both pThr313 and pSer 326; however, pThr292 was also decreased, and thus these results suggested that the efficiency of PLK1-dependent phosphorylation is either maintained or slightly enhanced in response to most types of genotoxic stress (Figure 3—figure supplement 3).

PLK1-dependent phosphorylation of RAD9 Thr313 is required to suppress S-phase checkpoint response

If PLK1 acts on RAD9 downstream of CDK in vivo, then the mutation of PLK1-dependent phosphorylation sites should affect cells in a manner similar to that seen with RAD9-T292A. We therefore constructed U2OS or HEK293A cell lines that encoded RAD9 with the T313A or S326A mutation, and expressed the mutant proteins under the control of the CMV promoter from a construct located at the FRT locus. The expression of RAD9-mH in U2OS cells is shown in Figure 4—figure supplement 1.

First, we assessed whether the PLK1-dependent phosphorylation affected the retention of the mutated RAD9 proteins on damaged chromatin. In a fractionation assay, we observed the increased accumulation of T313A-mutated RAD9 in the insoluble fraction upon HU treatment (Figure 4A, Triton-X insoluble, +HU). We also observed the accumulation of Ser345-phosphorylated CHK1 in the insoluble fraction of HU-treated cells that expressed the T313A-mutated RAD9 (Figure 4A, Triton-X insoluble, +HU). It should be noted that lower levels of the S326A-mutated RAD9 accumulated in the insoluble fraction (compared to the accumulation of RAD9-T313A under the same conditions). These results indicated that the PLK1-dependent phosphorylation of the Thr313 residue of RAD9 phenocopies the T292A-mutated RAD9, and suggested that PLK1 acts downstream of CDK to control the dissociation of RAD9 from chromatin upon genotoxic stress. Consistent with this interpretation, we also observed that the inhibition of the PLK1 activity by BI2536 also caused the accumulation of endogenous RAD9 in the insoluble fraction, when the cells were treated with HU (Figure 4—figure supplement 2). Since both TopBP1 and CLASPIN, which act as mediators for checkpoint signaling, also accumulated in the insoluble fraction upon the inhibition of PLK1 in HU-treated cells, we concluded that a PLK1-dependent phosphorylation event controls the association of the checkpoint machinery with chromatin.

Figure 4.
Enhanced checkpoint signaling when PLK1 fails to phosphorylate RAD9.

Next, we analyzed whether the PLK1-dependent phosphorylation (pThr313) of RAD9 follows the CDK-dependent phosphorylation (pThr292), in terms of controlling the rate of DNA replication. To do so, we quantitated the rate of dNTP incorporation in the U2OS cells. The cells that expressed the T313A- or S326A-mutated RAD9 were exposed to 0.2 mM HU for 24 hr, and then to EdU for 1 hr prior to harvest. The incorporated EdU was fluorescently labeled by the click reaction, and the cells were subjected to a flow cytometric analysis. The cells that expressed the T313A-mutated RAD9 showed a significantly lower level of EdU incorporation, as compared to the cells that expressed wild-type RAD9 (as seen in T292A-RAD9 expressing cells) (Figure 4B, Figure 4—figure supplement 3). The cells that expressed the S326A-mutated RAD9 and those expressing wild-type RAD9 incorporated similar levels of EdU. These results suggested that pThr313, as well as pThr292, of RAD9 is responsible for maintaining the rate of DNA replication upon genotoxic stress. Together, these experiments indicated that the CDK-PLK1 axis targets RAD9 to antagonize the checkpoint response upon replicative stress.

We also knocked down the endogenous RAD9 (Figure 4—figure supplement 4A), to determine whether the cellular effect observed upon the expression of the T313A- or S326A-mutated RAD9 is enhanced. In cells that expressed the T313A-mutated RAD9, the transfection of siRAD9 yielded an enhancement in the phosphorylation of Ser345 of CHK1 (Figure 4—figure supplement 4B) and a slight reduction in the rate of DNA replication upon HU treatment, as compared to cells that were not treated with siRAD9 (Figure 4—figure supplement 4C). These results indicated that, although the expression of RAD9 under the control of the CMV promoter was sufficient to confer the phenotype, the knockdown of the endogenous RAD9 could enhance the cellular effect to a certain extent. In cells that expressed the S326A-mutated RAD9, the silencing did not result in an obvious enhancement of the phosphorylation of Ser345 of CHK1 or a decrease in the DNA replication rate (Figure 4—figure supplement 4D,E). Furthermore, we also confirmed the reduction in the DNA replication rate in telomerase-positive HEK293A cells (Bryan et al., 1995) expressing either T292A- or T313A-mutated RAD9 (Figure 4—figure supplement 5A and B). This result excluded the possibility that the slowing of the DNA replication rate was caused by an atypical mechanism for checkpoint control, which may occur in U2OS cells via Alternative Lengthening of Telomeres (ALT) (Figure 4—figure supplement 5A and B).

A slower DNA replication rate was also observed in response to other types of DNA damage. Specifically, aphidicolin treatment also reduced the rate of EdU incorporation, and MMC treatment caused more cells to accumulate in early S phase when the T313A-mutated RAD9 was expressed, although the latter treatment did not yield an obvious decrease in the EdU incorporation (Figure 4—figure supplement 6). The observed phenotype may have been specific to the direct inhibition of DNA replication by some agents (HU, a dNTP synthesis inhibitor; aphidicolin, a DNA polymerase inhibitor), but not by others (MMC, a DNA damaging agent that crosslinks DNA strands).

We next performed a DNA combing assay, to determine which steps of DNA replication contribute to slowing down bulk DNA synthesis. The DNA damage checkpoint involves controlling either the dormant/late replication origin firing or the velocity of individual replication forks. The DNA combing technique allows the measurement of individual replication fork rates, and of the IOD (Inter-Origin Distance) values that may reflect the efficiency of dormant origin firings. At the end of the HU treatment, the WT- or T313A-mutated-RAD9 expressing U2OS cells were pulse-labeled by 5-iodo-2'-deoxyuridine (IdU), preceded by 5-chloro-2'-deoxyuridine (CldU) (Figure 5A). The DNA synthesis on DNA fibers was visualized (Figure 5B), and the IOD values and the fork velocities were measured (Figure 5C,D,E). In the absence of HU, the rates of individual replication forks in cells that expressed RAD9-WT were 0.85 ± 0.36 kb/minutes, and those of the RAD9-T313A expressing cells were slower (0.66 ± 0.27 kb/minutes, p-value: 5.8 × 10−7). The IOD values were 131 ± 62 kb in RAD9-WT expressing cells, and 107 ± 61 kb in RAD9-T313A expressing cells. This shorter IOD distance in Rad9-T313A expressing cells implied a higher efficiency of origin firing, which could result in a similar rate of bulk DNA synthesis to that seen in RAD9-WT expressing cells under non-stressed conditions. We then measured the fork rates under the HU stress, and observed that the rates of the DNA replication forks became slower in both RAD9-WT and RAD9-T313A expressing cells, with similar velocities (WT: 0.37 ± 0.16 kb/minutes, T313A: 0.37 ± 0.19 kb/minutes). Under these conditions, the IOD of the RAD9-T313A expressing cells was slightly longer than that of RAD9-WT expressing cells (WT: 61 ± 38 kb, T313A: 69 ± 41 kb), although the difference is marginally significant (p-value: 0.058).

Figure 5.
The effect of origin firing and DNA replication fork integrity of PLK1-dependent phosphorylation on RAD9.

The replication fork velocities under non-stressed and stressed conditions were confirmed in an independent combing assay. Samples for either non-treated or aphidicolin-treated (0.4 µM) cells (RAD9-WT and -T313A expressing cells) were applied. A slower replication fork progression rate was observed in the non-treated RAD9-T313A expressing cells (Figure 5—figure supplement 1), consistent with the data obtained in the former assay (Figure 5), although the absolute values of the fork progression rate were smaller under the latter conditions, perhaps due to the different method used to prepare the fibers on the slides (see Materials and methods). When the cells were treated with aphidicolin, the replication fork rate was similar to that under the conditions with 0.2 mM HU, again suggesting that the fork progression rate did not cause the difference in the slow S-phase progression rate. In summary, we concluded that the reason for the slower rate of S-phase progression is not due to the reduction of the replication fork rate in the RAD9-T313A expressing cells. Further analysis is needed to determine whether the increase in the distance between firing origins contributes to the slow S-phase. It should also be noted that the replication fork rate was already slower in the RAD9-T313A expressing cells under non-stressed conditions, which may suggest that the checkpoint response is enhanced in the RAD9-T313A expressing cells in the normal cell cycle (see discussion).

Finally, we observed that the T292A- or T313A-mutated RAD9 co-immunoprecipitated with increased amounts of CLASPIN upon HU treatment ((Figure 5—figure supplement 2), see ‘HU’ in IP). Together, these results suggested that the S-phase checkpoint response is enhanced when PLK1 fails to phosphorylate RAD9.

CDK- and PLK1-dependent phosphorylation of RAD9 allows cells to grow under genotoxic stress

If the collaboration of CDK and PLK1 in phosphorylating RAD9 is required for decreasing RAD9’s S-phase checkpoint function, then cells expressing the T292A- or T313A-mutated RAD9 should be growth-compromised under genotoxic stress. Indeed, we observed that cells expressing the T292A- or T313A-mutated RAD9 exhibited a slower growth rate in the presence of HU (For U2OS cells, Figure 6A; for HEK293A cells, Figure 6—figure supplement 1). Additionally, as compared to the cells that expressed wild-type RAD9, the cells that expressed T313A-mutated RAD9 exhibited a significant decrease in colony-forming efficiency (as assessed by a cell viability assay) upon exposure to 0.2 mM HU (Figure 6B). Together, these results imply a role for the CDK- and PLK1-dependent phosphorylation of RAD9 in facilitating growth under genotoxic stress.

Figure 6.
PLK1-dependent phosphorylation of RAD9 is required for proliferation under replicative stress.

Discussion

Cells must overcome environmental stresses that have the potential to disturb their proliferation. In the case of genotoxic stress, DNA damage checkpoints are activated to restore cellular viability. However, such checkpoints can also serve as anti-proliferation signals. Thus, the proliferation status of cells must be coupled to the activities of the checkpoint machineries. Here, we demonstrated the collaborative action of CDK and PLK1 in phosphorylating the checkpoint protein RAD9, thereby minimizing RAD9’s checkpoint activating function to maintain the rate of S-phase, and in turn to maintain the higher cellular growth rate (Figure 7). The CDK- and PLK1-dependent phosphorylation of RAD9 reduces the binding of RAD9 to damaged chromatin, suggesting the dynamic control of active checkpoint complexes at DNA damage sites to regulate the degree of DNA damage checkpoint signaling. Our results have elucidated the mechanism by which the cell cycle status controls the DNA damage sensor complex to ensure proliferation under genotoxic stress, thus improving our understanding of the balance between anti- and pro-proliferation signals.

Figure 7.
A schematic of PLK1-dependent damage tolerance.

CDK and PLK1 target RAD9 to modulate DNA damage detection

When cells that were defective in the CDK- or PLK1-dependent phosphorylation of RAD9 (demonstrated by the expression of T292A- or T313A-mutated RAD9) were grown under replicative stress (0.2 mM HU), more cells accumulated in S phase, and the incorporation of dNTPs during S phase was greatly decreased, as compared to that observed in cells expressing wild-type RAD9. These results suggested that CDK plays a role in maintaining a high rate of S-phase progression, via the facilitation of the PLK1-dependent phosphorylation of RAD9.

The region where CDK phosphorylates RAD9 (on Thr292) has a typical consensus sequence for a PBD binding site (van Vugt et al., 2010). Using a peptide competition assay, we showed that PLK1 bound RAD9 phosphorylated at Thr292. Furthermore a peptide phosphorylated on Thr292 activated PLK1. Since PBD acts as an auto-inhibitory domain, with inhibitory activity relieved when the domain is bound to a substrate (Xu et al., 2013), we concluded that RAD9 phosphorylated on Thr292 provides a genuine binding site for PLK1.

The interaction between a PBD and the phospho-peptide is reportedly strong, with a dissociation constant in the range of a few hundred nanomolar (van Vugt et al., 2010). This high affinity may allow a minimal amount of PLK1 to target the Thr292-phosphorylated RAD9, and would suit binding by PLK1, which exhibits rather repressed expression during interphase (Winkles and Alberts, 2005). In this scenario, once Thr292 is phosphorylated, PLK1 easily targets and binds RAD9, leading to the local activation of PLK1 and the phosphorylation of the Thr313 residue of RAD9.

What is the molecular impact of the PLK1-dependent phosphorylation of RAD9? RAD9, as a component of the PCNA-like 9-1-1 complex, is loaded onto 5' recessed DNA ends at damage sites, and also accommodates TopBP1 and CLASPIN to facilitate signaling through checkpoint kinases (Delacroix et al., 2007; Furuya et al., 2004; Kumagai et al., 2006; Lee et al., 2007; Liu et al., 2012; Majka et al., 2006; Navadgi-Patil and Burgers, 2009; Takeishi et al., 2010). We speculate that RAD9’s function as a DNA damage sensor is the target of PLK1-dependent regulation, given that we observed increased amounts of RAD9 on the chromatin when PLK1 failed to phosphorylate RAD9. The PLK1-dependent phosphorylation of RAD9 would minimize the amount of RAD9 bound at sites of stalled replication forks or replicative DNA damage, possibly by giving dynamism to the 9-1-1 complex from the DNA ends, and may in turn reduce the number of active checkpoint complexes. In RAD9, the sites of PLK1-dependent phosphorylation are located within the tail region. This region of RAD9 is expected to help the protein associate with DNA damage sites. For example, the region phosphorylated by PLK1 is close to the region required for the association with the single-stranded-DNA binding protein RP-A (Figure 3—figure supplement 1A)(Xu et al., 2008). Furthermore, we did not observe a decrease in the amount of the RAD9-bound clamp-loader RAD17 in the co-IP experiment when T292 was mutated, even under genotoxic stress conditions (data not shown). Thus, the PLK1-dependent phosphorylation may affect the affinity of the 9-1-1 complex itself to the DNA damage sites. This regulation is influenced by the balance between the cellular pro-mitotic activity and the DNA damage checkpoint activity, which may be determined by the amounts of PLK1, the CDK activity, and the degree of DNA damage stress. In other words, the cellular proliferation status may influence the cell’s ability to detect DNA damage during checkpoint activation.

Phosphorylation of the checkpoint sensor protein by CDK and PLK1 is required for maintaining the rate of DNA replication upon genotoxic stress

In eukaryotic cells, the rate of S-phase progression is related to the number of available replication origins (Yekezare et al., 2013). Upon growth under genotoxic stress, the replication rate can be controlled by S-phase checkpoint signaling, and when the signaling is dominant, it inhibits the firing of the late-replication origins (Santocanale and Diffley, 1998; Shirahige et al., 1998). This inhibition of origin firing might prevent the generation of additional replicative stress in un-replicated regions of the genome, and would suppress genomic instability (Boos et al., 2011; Petermann et al., 2010; Yekezare et al., 2013). Alternatively, the slow S phase progression rate could be caused by the reduction in the progression rate of individual replication forks, which is reportedly dependent on CHK1 activation upon camptothecin-induced stress, and may allow the fork to mount a more robust response to the damage (Iyer and Rhind, 2013; Seiler et al., 2007). It has also been proposed that the reduction of the fork speed results in the activation of dormant origins proximal to the stalled replication fork, a process that may compensate for decreases in the local rate of DNA replication (Yekezare et al., 2013).

The DNA fiber/combing assay under the HU treated conditions revealed that the velocities of the replication forks were the same between the WT- and T313A-RAD9 expressing cells. This could indicate that the slower S-phase progression seen in the T313A-mutated cells was not caused by the slower velocity of the DNA replication forks. We observed a slight increase in the IOD value in the T313A-RAD9 expressing cells, which implies the inefficient firing of dormant origins in these cells. However, its marginal significance (p-value=0.058; WT vs. T313A) may indicate another reason, such as the contribution of the inhibition of late origin firing to the slower progression of S-phase in T313A-expressing cells under HU-induced stress. Perhaps more significantly, even under non-stressed conditions, we observed a slower DNA replication fork velocity in the T313A-RAD9 expressing cells, as compared to that in WT-RAD9 expressing cells. This could indicate that a moderate activation of the S-phase checkpoint exists even in the normal cell cycle process in the T313A-RAD9 expressing U2OS cells. In this case, the efficient firing of dormant origins may compensate for the overall S-phase rate; however, a more detailed analysis (e.g., shorter pulse-label duration) should be performed to arrive at a definite conclusion. In the T313A-expressing cells, further exposure to HU-induced stress would activate an excessive DNA damage response and could cause a reduction in the number of firing origins, which would result in a slower rate of S-phase progression. In this scenario, PLK1 may raise the threshold of the DNA damage level, which would require more DNA damage stress for the checkpoint activation during S-phase. In other words, the PLK1-dependent phosphorylation of RAD9 may suppress the activation step of the checkpoint, rather than diminishing the activated DNA damage checkpoint processes, and this process could act as a novel DNA damage tolerance mechanism to facilitate cell proliferation. The dynamics between the proliferation control and the DNA damage sensors at DNA damage sites should be clarified in the future, to facilitate our understanding of the ability of cells to drive proliferation under environmental stresses.

PLK1 mediates tolerance to genotoxic stress

PLK1 is a pro-mitotic kinase, and its main function is to facilitate the mitotic process. However, PLK1 also functions in DNA metabolism, and probably coordinates cellular proliferation and genomic stability. In addition to its role in antagonizing the DNA damage checkpoint activity, PLK1 reportedly promotes DNA replication by phosphorylating Orc2, an event that may be critical to maintain genomic DNA replication under replicative stress (Song et al., 2011). The PLK1-mediated phosphorylation of Ser188 of Orc2 is responsible for the resistance of pancreatic tumor cells to the chemotherapy drug gemcitabine, an agent that inhibits DNA replication (Song et al., 2013). This phosphorylation event probably stabilizes the pre-replicative complexes to ensure that the cells continue to progress through S phase despite the genotoxic stress, possibly by keeping dormant origins active (Song et al., 2013). Combined with our finding of the PLK1-dependent phosphorylation of RAD9, these results suggest that PLK1 employs multiple targets to drive S-phase progression in cells under stress, thus facilitating tolerance to genotoxic stresses.

For cancer cells, counteracting these checkpoint-related processes is expected to be a favorable tactic for driving cellular proliferation. Indeed, the ectopic expression of PLK1 is associated with a wide range of tumor types. PLK1 is also known to be involved in promoting resistance to chemotherapeutic regimens with drugs such as doxorubicin (a DNA intercalating compound) or gemcitabine (a DNA analogue that serves as an inhibitor of DNA replication) (Gutteridge et al., 2016). Thus, an improved understanding of DNA-related metabolism involving PLK1 will be met with keen interest, as such information could facilitate the effective use of PLK1 inhibitors during cancer therapies.

In summary, we demonstrated that CDK and PLK1 target the DNA damage sensor RAD9 to minimize the checkpoint response, and revealed an underlying mechanism whereby cells can tolerate genotoxic stress and proceed with proliferation. Thus, a complex balancing mechanism exists to coordinate proliferation and the stress response, and more systematic investigations of this mechanism should be conducted in the future.

Materials and methods

Plasmids and primers

The RAD9A cDNA was purchased from OriGene. The gene encoding C-terminally myc-His6 (mH)-tagged RAD9 was constructed in pCDNA3.1 or pCDNA5/FRT. The cloned RAD9 was mutagenized for siRAD9 (s11719: Ambion) resistance. The mutations in RAD9 were introduced using KOD polymerase (TOYOBO, Japan) and mutagenized primers. The primers used to mutagenize RAD9 are presented in the supplementary table (Supplementary file 1). The primers to construct the resistant RAD9 against siRAD9 (s11719: Ambion) were #2147/#2150, and are also included in Supplementary file 1. The gene encoding GST-tagged RAD9 used for bacterial expression was constructed on the pGEX-6P plasmid. Primers #1663/#1664 (Supplementary file 1) were used for the subcloning. For the two-hybrid assay, a pLex-A fusion system and pACT2 (or pGADT7) were used to construct the bait and prey plasmids, respectively (Bartel and Fields, 1995; Tanaka et al., 2013). The regions encoding the polo-box domains of PLK1-5 were amplified from a cDNA library (a kind gift from Dr. Seiji Tanaka, Kochi University of Technology), using the primers listed in Supplementary file 1. Other plasmids used to express the PLK1, pCDNA-flag-PLK1 and pGEX-6P-PLK1 (wt and K82R) proteins were kind gifts from Dr. M. Osugi (Tokyo University). For the expression of the active CDK protein, the plasmid that expresses the CIV1 (CAK in S. cerevisiae), CDK2, and Cyclin E proteins, controlled by a dual promoter system, was provided by Dr. Tatsuro Takahashi (Kyushu University), and the plasmid that expresses the active CDK2-CyclinA2 protein is a derivative of that plasmid.

Cell culture and transfection

HEK293A cells (purchased from Invitrogen) and HEK293A T-REx cells were cultured in Dulbecco’s modified Eagle’s medium, supplemented with 10% fetal bovine serum. U2OS cells (purchased from IDAC, Tohoku University) and U2OS T-REx cells were cultured in RPMI medium, supplemented with 10% fetal bovine serum. Cell lines used were exposed to plasmocin to avoid possible contamination of mycoplasma and tested for the TAKARA PCR mycoplasma kit. Plasmid transfection was performed using the Attracten transfection reagent, according to the manufacturer’s protocol (Qiagen). For the induction from the tetO-harboring plasmid, doxycyclin (0.5 µg/ml) was used. For the knockdown of RAD9, siRAD9 (Ambion: s11719) was used. The transfection of the siRNA was performed using the Dharmafect one transfection reagent, according to the manufacturer’s protocol (Dharmacon). The control siRNA was purchased from QIAGEN. To inhibit cellular kinases, a PLK1 inhibitor (BI2536: AdooQ Bioscience, USA) and a CDK2 inhibitor (RO-3306: Santa Cruz, USA) were used. To perturb cell cycle progression, hydroxyurea (Sigma-Aldrich) was used. For the double thymidine block, first, 2.5 mM thymidine was added to the media, followed by 10 hr of incubation without thymidine. Then, a second thymidine block was performed for 20 hr before the release.

Cell growth or viability assay

To monitor cell growth, the Presto-Blue Cell Viability Reagent (Thermo Fisher) was used according to the manufacturer’s protocol. Briefly, 1 × 104 cells / 150 µl / well of U2OS cells that stably express the RAD9 gene were cultured in 96-well dishes with hydroxyurea. Every 24 hr, the 96 well dish was monitored for cell growth. Presto-Blue (17 µl) was added to each well, followed by an incubation in a CO2 incubator at 37 ˚C for 10 min. Subsequently, 75 µl of 10% SDS was added to stop the reaction, and the plate was placed in an ARVO-X3 plate reader (Perkin Elmer). A 531 nm excitation filter and a 616 nm emission filter were used, and for each cell line and hydroxyurea condition, triplicate samples were measured and the average values were plotted in the graph. To monitor the cell viability upon hydroxyurea exposure, a colony formation assay was performed. First, U2OS cells that express the RAD9 protein were treated with 0.2 mM hydroxyurea for 36 hr. The cells were collected by trypsinization, and 500 cells / 10 ml were plated onto 10 cm dishes, followed by an incubation at 37 ˚C in a CO2 incubator for 11 days. The plates were stained with crystal violet/methanol.

Antibodies and western blotting

Western blot analyses were performed with anti-γ-H2AX (Millipore), anti-H2AX (Sigma-Aldrich), anti-RAD9 (Santa Cruz or Bethyl), anti-RAD17, anti-RAD1, anti-TopBP1, anti-CLASPIN (Santa Cruz), anti-tubulin, anti-CHK1, anti-CHK1-pS345, and anti-CHK2-pT68 (Cell Signaling Technology) antibodies. The antibodies against the phospho-peptides (pThr292: LQAHSpTPHPDC, pThr313: CAMETpTIGNEG, pSer326: CEGSRVLPSIpS) were raised by MBL (Medical and Biological Laboratories (MBL), Japan). HRP-conjugated antibodies were used as secondary antibodies, the ECL signal was incorporated using a Licor Odyssey Fc imager system, and the quantification was performed with the Image Studio software.

Flow cytometry and EdU incorporation assay

To measure the incorporation rate of deoxyribonucleotides, the amount of incorporated EdU in cells was measured by a flow cytometer using the reported labeling protocol (Jia et al., 2015). Cells were incubated with 5 µM of EdU (for the non-stressed cells: 30 min, for the HU-stressed cells: 60 min) before fixation by formalin (3.7%) for 10 min on ice. The PBS-washed cells were subjected to methanol fixation. The fixed cells were washed twice with PBS-Triton X-100 (0.1%), and subjected to the Click reaction for one hour at room temperature. The reaction mixture included 20 mM Tris (pH 7.4), 1/1,000 Alexa Fluor 488-azide (Molecular Probes), 0.1% CuSO4, and 1 mg/ml of ascorbate. The labeled cells were washed with PBS-Triton X-100 and treated with a PI/RNaseA solution for 30 min, followed by a wash with PBS-Triton X-100. The cell cycle and the fluorescently labeled EdU were measured using a Guava easyCyte (Millipore) or FACSCalibur flow cytometer.

In vitro assays

The proteins were purified through GSH-beads (4 ˚C, 1 hr in TEG buffer (Tris-Cl, pH 7.5, 50 mM, NaCl 150 mM, EDTA-Na 10 mM, glycerol 10%, Triton X-100 0.5–1%)). The GST-RAD9 proteins were eluted by 10 mM of glutathione (15 min, 3 times), and the CDK complex and PLK1 proteins were eluted by PreScission Protease (GE Healthcare, 15 min, 3 times). After the elution of the CDK complex and PLK1, the proteins were further incubated with GSH-beads to eliminate the residual GST-tagged proteins. All of the purified proteins were subjected to dialysis in 50% glycerol, 10 mM Tris-Cl (pH 8.0). The in vitro kinase assay using the recombinant active CDK complex or PLK1 was performed in kinase buffer (50 mM HEPES, pH 7.4, 10 mM MgCl2, 1 mM DTT) supplemented with 1 mM ATP. The GSH pull down assay using GST-RAD9 was performed in kinase buffer supplemented with 0.1% Triton X-100.

Immunoprecipitation

Cell lysates were prepared in HEPES-based buffer (20 mM HEPES, pH 7.4, 150 mM NaCl, 10% glycerol, 0.1 % NP40, 0.1% Triton X-100, 1 mM DTT, 0.5 mM PMSF). The lysates were sonicated and the soluble fractions were quantified using a protein assay kit (Bio-Rad), and adjusted to 2 µg/µl. The immunoprecipitation was performed using a c-Myc tagged Mild Purification Kit (MBL).

Yeast two-hybrid assay

To quantify the beta-galactosidase activity, an ONPG (SIGMA-Aldrich) assay was performed, according to the Yeast Protocol Handbook (Clontech).

Liquid chromatography-tandem mass spectrometry (LC-MS/MS)

Coomassie Brilliant Blue-stained protein bands were excised from the polyacrylamide gel and subjected to in-gel tryptic digestion. The digestion was performed using an In-Gel Tryptic Digestion Kit (Thermo Fisher Scientific, Waltham, MA), according to the manufacturer’s instructions. The digestion products were reconstituted in 0.1% formic acid and injected for separation into a NanoLC-Ultra 2D-plus HPLC equipped with a cHiPLC Nanoflex system (Eksigent, Dublin, CA), which was used in the trap and elute mode with a trap column (200 μm x 0.5 mm, ChromXP, C18-CL, 3 μm, 120 Å, Eksigent) and an analytical column (ChromXP, C18-CL, 3 μm, 120 Å, Eksigent). The separation was performed with a binary gradient of solvent A (98% water, 2% acetonitrile, 0.1% formic acid) and solvent B (20% water, 80% acetonitrile, 0.1% formic acid). The gradient program was 2% to 40% solvent B for 125 min, 90% solvent B for 5 min, and 2% solvent B for 19 min, at a flow rate of 300 nL/min. The eluates from the Nano-LC system were infused on-line to the TripleTOF 5600 + mass spectrometry system (SCIEX, Framingham, MA) equipped with a NanoSpray III ion source and a heated interface. The data sets were acquired with the information-dependent acquisition method.

Identification of the phosphorylated peptides and quantitative estimation of the specific phosphorylation

The identification of the phosphorylated peptides was accomplished using the ProteinPilot software (version 4.5 Beta, SCIEX). The database was UniProtKB/Swiss-Prot (June 2014, Human), appended with a known contaminant database (SCIEX) and the GST-Rad9C fusion protein sequence. The special factors were gel-based ID and phosphorylation emphasis. Biological modifications were taken into account. The quantitative estimation of the specific phosphorylation was performed by the spectral counting method. The number of reliable MS/MS spectra (confidence ≥95%) corresponding to the peptide groups derived from the GST-Rad9 C-terminus, which had phosphorylation on the same site, irrespective of the other co-existing post-transcriptional modifications, were counted in each group and used for the quantitative estimation.

DNA molecular combing assay

The DNA fiber assay was performed as described (Conti et al., 2007; Kuriya et al., 2015; Michalet et al., 1997). Briefly, cells were labeled with 100 µM of IdU or CldU for 30 min and harvested by trypsinization. The harvested cells were embedded in agarose plugs and treated overnight with proteinase K (2 mg/ml) in ESP buffer (1% N-L-sarcosyl in 0.5 M EDTA). The prepared samples were subjected to DNA fiber assays to monitor the DNA replication fork rate (stretching factor upon combing was 1.95 kb/µm). For the DNA combing assay to monitor the Inter-Origin-Distance, the samples were sent to GENOMIC VISION (France) for the analysis. The samples were prepared as advised, and two slides per sample were analyzed (stretching factor upon combing is approx. 2 kb/µm). The Inter-Origin-Distance was calculated using Fiber Studio (GENOMIC VISION). We noticed that the values of the replication fork tracks were longer in the samples obtained from GENOMIC VISION. These were partly caused by the difference in the buffer or glass that was used to stretch the DNA fibers. Thus, we emphasize the relative differences in values within the samples obtained from same series of experimental preparations.

Acknowledgements

We thank all of members of the Radiation Biology Center for helpful comments and suggestions. Especially, we thank Ms. Mika Gunji for technical contributions. We thank Dr. Miho Ohsugi (University of Tokyo) for providing the PLK1 plasmids and the tips for PLK1 purification, Dr. Tatsuro Takahashi for providing the CDK plasmid, and Dr. Seiji Tanaka (Kohchi University of Technology) for providing the yeast two hybrid system and advice. This work was supported by KAKENHI (Grant-in-Aid for Scientific Research) for Young Scientists (A): (24687023 to KF), Grants-in-Aid for Scientific Research on Innovative Areas (23131515, 25131707 to KF), (C): 17K07284 (KF), The Uehara Memorial Foundation Grant (KF) and The Takeda Science Foundation grant for life science (KF). This work was also partly supported by NIG-JOINT (National Institute of Genetics, 2011–2017-A, (KF)), the Cooperative Research Project Program of Joint Usage/Research Center at the Institute of Development, Aging and Cancer, Tohoku University (2017, (KF)) and the joint travel program of the Institute of Molecular Embryology and Genetics, Kumamoto University (2017, (KF)).

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Funding Information

This paper was supported by the following grants:

  • http://dx.doi.org/10.13039/501100001700Ministry of Education, Culture, Sports, Science, and Technology to Kanji Furuya.
  • http://dx.doi.org/10.13039/100008732Uehara Memorial Foundation to Kanji Furuya.
  • http://dx.doi.org/10.13039/100007449Takeda Science Foundation to Kanji Furuya.
  • Cooperative Research Project Program of Joint Usage/Research Center at the Institute of Development, Aging and Cancer to Kanji Furuya.
  • Institute of Molecular Embryology and Genetics to Kanji Furuya.
  • KAKENHI (Grant-in-Aid for ScientificResearch) for Young Scientists 24687023 to Kanji Furuya.
  • KAKENHI Grants-in-Aid for Scientific Research on Innovative Areas 23131515 to Kanji Furuya.
  • KAKENHI Grants-in-Aid for Scientific Research on Innovative Areas 25131707 to Kanji Furuya.
  • KAKENHI 17K07284 to Kanji Furuya.
  • National Institute of Genetics 2011-2017-A to Kanji Furuya.

Additional information

Competing interests

No competing interests declared.

Author contributions

Conceptualization, Data curation, Formal analysis, Investigation.

Investigation, Methodology.

Data curation, Investigation, Methodology.

Resources, Data curation, Software, Formal analysis, Investigation, Methodology.

Investigation, Methodology.

Investigation, Methodology.

Supervision.

Resources, Methodology.

Conceptualization, Supervision, Methodology, Writing—original draft, Writing—review and editing.

Conceptualization, Resources, Data curation, Software, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing.

Additional files

10.7554/eLife.29953.023

Supplementary file 1.

Primers used in this study:

Primers used in this study was listed.

10.7554/eLife.29953.024

Supplementary file 2.

Statistic data to construct Figures 4B and 6A,B:

An excel formatted file was presented.

10.7554/eLife.29953.025

Supplementary file 3.

The expanded photos of DNA fibers used in Figure 5B:

The field in white squares are region presented in Figure 5B.

10.7554/eLife.29953.026

Transparent reporting form

References

  • Bartel PL, Fields S. Analyzing protein-protein interactions using two-hybrid system. Methods in Enzymology. 1995;254:241–263. [PubMed]
  • Boos D, Sanchez-Pulido L, Rappas M, Pearl LH, Oliver AW, Ponting CP, Diffley JF. Regulation of DNA replication through Sld3-Dpb11 interaction is conserved from yeast to humans. Current Biology. 2011;21:1152–1157. doi: 10.1016/j.cub.2011.05.057. [PubMed] [Cross Ref]
  • Bryan TM, Englezou A, Gupta J, Bacchetti S, Reddel RR. Telomere elongation in immortal human cells without detectable telomerase activity. The EMBO Journal. 1995;14:4240–4248. [PubMed]
  • Caspari T, Dahlen M, Kanter-Smoler G, Lindsay HD, Hofmann K, Papadimitriou K, Sunnerhagen P, Carr AM. Characterization of Schizosaccharomyces pombe Hus1: a PCNA-related protein that associates with Rad1 and Rad9. Molecular and Cellular Biology. 2000;20:1254–1262. doi: 10.1128/MCB.20.4.1254-1262.2000. [PMC free article] [PubMed] [Cross Ref]
  • Coco-Martin JM, Begg AC. Detection of radiation-induced chromosome aberrations using fluorescence in situ hybridization in drug-induced premature chromosome condensations of tumour cell lines with different radiosensitivities. International Journal of Radiation Biology. 1997;71:265–273. doi: 10.1080/095530097144148. [PubMed] [Cross Ref]
  • Conti C, Saccà B, Herrick J, Lalou C, Pommier Y, Bensimon A. Replication fork velocities at adjacent replication origins are coordinately modified during DNA replication in human cells. Molecular Biology of the Cell. 2007;18:3059–3067. doi: 10.1091/mbc.E06-08-0689. [PMC free article] [PubMed] [Cross Ref]
  • Deckbar D, Birraux J, Krempler A, Tchouandong L, Beucher A, Walker S, Stiff T, Jeggo P, Löbrich M. Chromosome breakage after G2 checkpoint release. The Journal of Cell Biology. 2007;176:749–755. doi: 10.1083/jcb.200612047. [PMC free article] [PubMed] [Cross Ref]
  • Delacroix S, Wagner JM, Kobayashi M, Yamamoto K, Karnitz LM. The Rad9-Hus1-Rad1 (9-1-1) clamp activates checkpoint signaling via TopBP1. Genes & Development. 2007;21:1472–1477. doi: 10.1101/gad.1547007. [PubMed] [Cross Ref]
  • Dimitrova DS, Gilbert DM. Temporally coordinated assembly and disassembly of replication factories in the absence of DNA synthesis. Nature Cell Biology. 2000;2:686–694. doi: 10.1038/35036309. [PMC free article] [PubMed] [Cross Ref]
  • Dufault VM, Oestreich AJ, Vroman BT, Karnitz LM. Identification and characterization of RAD9B, a paralog of the RAD9 checkpoint gene. Genomics. 2003;82:644–651. doi: 10.1016/S0888-7543(03)00200-3. [PubMed] [Cross Ref]
  • Elia AE, Cantley LC, Yaffe MB. Proteomic screen finds pSer/pThr-binding domain localizing Plk1 to mitotic substrates. Science. 2003;299:1228–1231. doi: 10.1126/science.1079079. [PubMed] [Cross Ref]
  • Elledge SJ. Cell cycle checkpoints: preventing an identity crisis. Science. 1996;274:1664–1672. doi: 10.1126/science.274.5293.1664. [PubMed] [Cross Ref]
  • Furuya K, Miyabe I, Tsutsui Y, Paderi F, Kakusho N, Masai H, Niki H, Carr AM. DDK phosphorylates checkpoint clamp component Rad9 and promotes its release from damaged chromatin. Molecular Cell. 2010;40:606–618. doi: 10.1016/j.molcel.2010.10.026. [PubMed] [Cross Ref]
  • Furuya K, Poitelea M, Guo L, Caspari T, Carr AM. Chk1 activation requires Rad9 S/TQ-site phosphorylation to promote association with C-terminal BRCT domains of Rad4TOPBP1. Genes & Development. 2004;18:1154–1164. doi: 10.1101/gad.291104. [PubMed] [Cross Ref]
  • Giannattasio M, Branzei D. S-phase checkpoint regulations that preserve replication and chromosome integrity upon dNTP depletion. Cellular and Molecular Life Sciences. 2017;74:2361–2380. doi: 10.1007/s00018-017-2474-4. [PMC free article] [PubMed] [Cross Ref]
  • Gutteridge RE, Ndiaye MA, Liu X, Ahmad N. Plk1 Inhibitors in cancer therapy: from laboratory to clinics. Molecular Cancer Therapeutics. 2016;15:1427–1435. doi: 10.1158/1535-7163.MCT-15-0897. [PMC free article] [PubMed] [Cross Ref]
  • Ishihara H, Martin BL, Brautigan DL, Karaki H, Ozaki H, Kato Y, Fusetani N, Watabe S, Hashimoto K, Uemura D. Calyculin A and okadaic acid: inhibitors of protein phosphatase activity. Biochemical and Biophysical Research Communications. 1989;159:871–877. doi: 10.1016/0006-291X(89)92189-X. [PubMed] [Cross Ref]
  • Ishikawa A, Yamauchi M, Suzuki K, Yamashita S. Image-based quantitative determination of DNA damage signal reveals a threshold for G2 checkpoint activation in response to ionizing radiation. Genome Integrity. 2010;1:10 doi: 10.1186/2041-9414-1-10. [PMC free article] [PubMed] [Cross Ref]
  • Iyer DR, Rhind N. Checkpoint regulation of replication forks: global or local? Biochemical Society Transactions. 2013;41:1701–1705. doi: 10.1042/BST20130197. [PMC free article] [PubMed] [Cross Ref]
  • Jang YJ, Lin CY, Ma S, Erikson RL. Functional studies on the role of the C-terminal domain of mammalian polo-like kinase. PNAS. 2002;99:1984–1989. doi: 10.1073/pnas.042689299. [PubMed] [Cross Ref]
  • Jia N, Nakazawa Y, Guo C, Shimada M, Sethi M, Takahashi Y, Ueda H, Nagayama Y, Ogi T. A rapid, comprehensive system for assaying DNA repair activity and cytotoxic effects of DNA-damaging reagents. Nature Protocols. 2015;10:12–24. doi: 10.1038/nprot.2014.194. [PubMed] [Cross Ref]
  • Jossen R, Bermejo R. The DNA damage checkpoint response to replication stress: A Game of Forks. Frontiers in Genetics. 2013;4:26 doi: 10.3389/fgene.2013.00026. [PMC free article] [PubMed] [Cross Ref]
  • Katou Y, Kanoh Y, Bando M, Noguchi H, Tanaka H, Ashikari T, Sugimoto K, Shirahige K. S-phase checkpoint proteins Tof1 and Mrc1 form a stable replication-pausing complex. Nature. 2003;424:1078–1083. doi: 10.1038/nature01900. [PubMed] [Cross Ref]
  • Kumagai A, Kim SM, Dunphy WG. Claspin and the activated form of ATR-ATRIP collaborate in the activation of Chk1. Journal of Biological Chemistry. 2004;279:49599–49608. doi: 10.1074/jbc.M408353200. [PubMed] [Cross Ref]
  • Kumagai A, Lee J, Yoo HY, Dunphy WG. TopBP1 activates the ATR-ATRIP complex. Cell. 2006;124:943–955. doi: 10.1016/j.cell.2005.12.041. [PubMed] [Cross Ref]
  • Kuriya K, Higashiyama E, Avşar-Ban E, Tamaru Y, Ogata S, Takebayashi S, Ogata M, Okumura K. Direct visualization of DNA replication dynamics in zebrafish cells. Zebrafish. 2015;12:432–439. doi: 10.1089/zeb.2015.1151. [PubMed] [Cross Ref]
  • Lee J, Kumagai A, Dunphy WG. The Rad9-Hus1-Rad1 checkpoint clamp regulates interaction of TopBP1 with ATR. Journal of Biological Chemistry. 2007;282:28036–28044. doi: 10.1074/jbc.M704635200. [PubMed] [Cross Ref]
  • Leloup C, Hopkins KM, Wang X, Zhu A, Wolgemuth DJ, Lieberman HB. Mouse Rad9b is essential for embryonic development and promotes resistance to DNA damage. Developmental Dynamics. 2010;239:2837–2850. doi: 10.1002/dvdy.22415. [PMC free article] [PubMed] [Cross Ref]
  • Liang H, Esposito A, De S, Ber S, Collin P, Surana U, Venkitaraman AR. Homeostatic control of polo-like kinase-1 engenders non-genetic heterogeneity in G2 checkpoint fidelity and timing. Nature Communications. 2014;5:4048 doi: 10.1038/ncomms5048. [PMC free article] [PubMed] [Cross Ref]
  • Lindsey-Boltz LA, Wauson EM, Graves LM, Sancar A. The human Rad9 checkpoint protein stimulates the carbamoyl phosphate synthetase activity of the multifunctional protein CAD. Nucleic Acids Research. 2004;32:4524–4530. doi: 10.1093/nar/gkh789. [PMC free article] [PubMed] [Cross Ref]
  • Liu S, Song N, Zou L. The conserved C terminus of claspin interacts with Rad9 and promotes rapid activation of Chk1. Cell Cycle. 2012;11:2711–2716. doi: 10.4161/cc.21041. [PMC free article] [PubMed] [Cross Ref]
  • Majka J, Binz SK, Wold MS, Burgers PM. Replication protein A directs loading of the DNA damage checkpoint clamp to 5'-DNA junctions. Journal of Biological Chemistry. 2006;281:27855–27861. doi: 10.1074/jbc.M605176200. [PubMed] [Cross Ref]
  • Meijer L, Borgne A, Mulner O, Chong JP, Blow JJ, Inagaki N, Inagaki M, Delcros JG, Moulinoux JP. Biochemical and cellular effects of roscovitine, a potent and selective inhibitor of the cyclin-dependent kinases cdc2, cdk2 and cdk5. European Journal of Biochemistry. 1997;243:527–536. doi: 10.1111/j.1432-1033.1997.t01-2-00527.x. [PubMed] [Cross Ref]
  • Michalet X, Ekong R, Fougerousse F, Rousseaux S, Schurra C, Hornigold N, van Slegtenhorst M, Wolfe J, Povey S, Beckmann JS, Bensimon A. Dynamic molecular combing: stretching the whole human genome for high-resolution studies. Science. 1997;277:1518–1523. doi: 10.1126/science.277.5331.1518. [PubMed] [Cross Ref]
  • Moudry P, Watanabe K, Wolanin KM, Bartkova J, Wassing IE, Watanabe S, Strauss R, Troelsgaard Pedersen R, Oestergaard VH, Lisby M, Andújar-Sánchez M, Maya-Mendoza A, Esashi F, Lukas J, Bartek J. TOPBP1 regulates RAD51 phosphorylation and chromatin loading and determines PARP inhibitor sensitivity. The Journal of Cell Biology. 2016;212:281–288. doi: 10.1083/jcb.201507042. [PMC free article] [PubMed] [Cross Ref]
  • Nakajima H, Toyoshima-Morimoto F, Taniguchi E, Nishida E. Identification of a consensus motif for Plk (Polo-like kinase) phosphorylation reveals Myt1 as a Plk1 substrate. Journal of Biological Chemistry. 2003;278:25277–25280. doi: 10.1074/jbc.C300126200. [PubMed] [Cross Ref]
  • Navadgi-Patil VM, Burgers PM. The unstructured C-terminal tail of the 9-1-1 clamp subunit Ddc1 activates Mec1/ATR via two distinct mechanisms. Molecular Cell. 2009;36:743–753. doi: 10.1016/j.molcel.2009.10.014. [PMC free article] [PubMed] [Cross Ref]
  • Ohashi E, Takeishi Y, Ueda S, Tsurimoto T. Interaction between Rad9-Hus1-Rad1 and TopBP1 activates ATR-ATRIP and promotes TopBP1 recruitment to sites of UV-damage. DNA Repair. 2014;21:1–11. doi: 10.1016/j.dnarep.2014.05.001. [PubMed] [Cross Ref]
  • Paulovich AG, Hartwell LH. A checkpoint regulates the rate of progression through S phase in S. cerevisiae in response to DNA damage. Cell. 1995;82:841–847. doi: 10.1016/0092-8674(95)90481-6. [PubMed] [Cross Ref]
  • Pellicioli A, Lee SE, Lucca C, Foiani M, Haber JE. Regulation of Saccharomyces Rad53 checkpoint kinase during adaptation from DNA damage-induced G2/M arrest. Molecular Cell. 2001;7:293–300. doi: 10.1016/S1097-2765(01)00177-0. [PubMed] [Cross Ref]
  • Petermann E, Woodcock M, Helleday T. Chk1 promotes replication fork progression by controlling replication initiation. PNAS. 2010;107:16090–16095. doi: 10.1073/pnas.1005031107. [PubMed] [Cross Ref]
  • Pérez-Castro AJ, Freire R. Rad9B responds to nucleolar stress through ATR and JNK signalling, and delays the G1-S transition. Journal of Cell Science. 2012;125:1152–1164. doi: 10.1242/jcs.091124. [PubMed] [Cross Ref]
  • Salic A, Mitchison TJ. A chemical method for fast and sensitive detection of DNA synthesis in vivo. PNAS. 2008;105:2415–2420. doi: 10.1073/pnas.0712168105. [PubMed] [Cross Ref]
  • Sandell LL, Zakian VA. Loss of a yeast telomere: arrest, recovery, and chromosome loss. Cell. 1993;75:729–739. doi: 10.1016/0092-8674(93)90493-A. [PubMed] [Cross Ref]
  • Santocanale C, Diffley JF. A Mec1- and Rad53-dependent checkpoint controls late-firing origins of DNA replication. Nature. 1998;395:615–618. doi: 10.1038/27001. [PubMed] [Cross Ref]
  • Seiler JA, Conti C, Syed A, Aladjem MI, Pommier Y. The intra-S-phase checkpoint affects both DNA replication initiation and elongation: single-cell and -DNA fiber analyses. Molecular and Cellular Biology. 2007;27:5806–5818. doi: 10.1128/MCB.02278-06. [PMC free article] [PubMed] [Cross Ref]
  • Serrano D, D'Amours D. When genome integrity and cell cycle decisions collide: roles of polo kinases in cellular adaptation to DNA damage. Systems and Synthetic Biology. 2014;8:195–203. doi: 10.1007/s11693-014-9151-9. [PMC free article] [PubMed] [Cross Ref]
  • Shirahige K, Hori Y, Shiraishi K, Yamashita M, Takahashi K, Obuse C, Tsurimoto T, Yoshikawa H. Regulation of DNA-replication origins during cell-cycle progression. Nature. 1998;395:618–621. doi: 10.1038/27007. [PubMed] [Cross Ref]
  • Sogo JM, Lopes M, Foiani M. Fork reversal and ssDNA accumulation at stalled replication forks owing to checkpoint defects. Science. 2002;297:599–602. doi: 10.1126/science.1074023. [PubMed] [Cross Ref]
  • Song B, Liu XS, Davis K, Liu X. Plk1 phosphorylation of Orc2 promotes DNA replication under conditions of stress. Molecular and Cellular Biology. 2011;31:4844–4856. doi: 10.1128/MCB.06110-11. [PMC free article] [PubMed] [Cross Ref]
  • Song B, Liu XS, Rice SJ, Kuang S, Elzey BD, Konieczny SF, Ratliff TL, Hazbun T, Chiorean EG, Liu X. Plk1 phosphorylation of orc2 and hbo1 contributes to gemcitabine resistance in pancreatic cancer. Molecular Cancer Therapeutics. 2013;12:58–68. doi: 10.1158/1535-7163.MCT-12-0632. [PMC free article] [PubMed] [Cross Ref]
  • St Onge RP, Besley BD, Pelley JL, Davey S. A role for the phosphorylation of hRad9 in checkpoint signaling. Journal of Biological Chemistry. 2003;278:26620–26628. doi: 10.1074/jbc.M303134200. [PubMed] [Cross Ref]
  • Stern B, Nurse P. A quantitative model for the cdc2 control of S phase and mitosis in fission yeast. Trends in Genetics. 1996;12:345–350. doi: 10.1016/S0168-9525(96)80016-3. [PubMed] [Cross Ref]
  • Swaffer MP, Jones AW, Flynn HR, Snijders AP, Nurse P. CDK substrate phosphorylation and ordering the cell cycle. Cell. 2016;167:1750–1761. doi: 10.1016/j.cell.2016.11.034. [PMC free article] [PubMed] [Cross Ref]
  • Syljuåsen RG, Jensen S, Bartek J, Lukas J. Adaptation to the ionizing radiation-induced G2 checkpoint occurs in human cells and depends on checkpoint kinase 1 and Polo-like kinase 1 kinases. Cancer Research. 2006;66:10253–10257. doi: 10.1158/0008-5472.CAN-06-2144. [PubMed] [Cross Ref]
  • Takeishi Y, Ohashi E, Ogawa K, Masai H, Obuse C, Tsurimoto T. Casein kinase 2-dependent phosphorylation of human Rad9 mediates the interaction between human Rad9-Hus1-Rad1 complex and TopBP1. Genes to Cells. 2010;15:761–771. doi: 10.1111/j.1365-2443.2010.01418.x. [PubMed] [Cross Ref]
  • Tanaka S, Araki H. Regulation of the initiation step of DNA replication by cyclin-dependent kinases. Chromosoma. 2010;119:565–574. doi: 10.1007/s00412-010-0291-8. [PubMed] [Cross Ref]
  • Tanaka S, Komeda Y, Umemori T, Kubota Y, Takisawa H, Araki H. Efficient initiation of DNA replication in eukaryotes requires Dpb11/TopBP1-GINS interaction. Molecular and Cellular Biology. 2013;33:2614–2622. doi: 10.1128/MCB.00431-13. [PMC free article] [PubMed] [Cross Ref]
  • Tanaka S, Umemori T, Hirai K, Muramatsu S, Kamimura Y, Araki H. CDK-dependent phosphorylation of Sld2 and Sld3 initiates DNA replication in budding yeast. Nature. 2007;445:328–332. doi: 10.1038/nature05465. [PubMed] [Cross Ref]
  • Thelen MP, Venclovas C, Fidelis K. A sliding clamp model for the Rad1 family of cell cycle checkpoint proteins. Cell. 1999;96:769–770. doi: 10.1016/S0092-8674(00)80587-5. [PubMed] [Cross Ref]
  • Toczyski DP, Galgoczy DJ, Hartwell LH. CDC5 and CKII control adaptation to the yeast DNA damage checkpoint. Cell. 1997;90:1097–1106. doi: 10.1016/S0092-8674(00)80375-X. [PubMed] [Cross Ref]
  • Trovesi C, Manfrini N, Falcettoni M, Longhese MP. Regulation of the DNA damage response by cyclin-dependent kinases. Journal of Molecular Biology. 2013;425:4756–4766. doi: 10.1016/j.jmb.2013.04.013. [PubMed] [Cross Ref]
  • van Vugt MA, Gardino AK, Linding R, Ostheimer GJ, Reinhardt HC, Ong SE, Tan CS, Miao H, Keezer SM, Li J, Pawson T, Lewis TA, Carr SA, Smerdon SJ, Brummelkamp TR, Yaffe MB. A mitotic phosphorylation feedback network connects Cdk1, Plk1, 53BP1, and Chk2 to inactivate the G(2)/M DNA damage checkpoint. PLoS Biology. 2010;8:e1000287 doi: 10.1371/journal.pbio.1000287. [PMC free article] [PubMed] [Cross Ref]
  • Vassilev LT, Tovar C, Chen S, Knezevic D, Zhao X, Sun H, Heimbrook DC, Chen L. Selective small-molecule inhibitor reveals critical mitotic functions of human CDK1. PNAS. 2006;103:10660–10665. doi: 10.1073/pnas.0600447103. [PubMed] [Cross Ref]
  • Weinert TA, Hartwell LH. The RAD9 gene controls the cell cycle response to DNA damage in Saccharomyces cerevisiae. Science. 1988;241:317–322. doi: 10.1126/science.3291120. [PubMed] [Cross Ref]
  • Winkles JA, Alberts GF. Differential regulation of polo-like kinase 1, 2, 3, and 4 gene expression in mammalian cells and tissues. Oncogene. 2005;24:260–266. doi: 10.1038/sj.onc.1208219. [PubMed] [Cross Ref]
  • Xu J, Shen C, Wang T, Quan J. Structural basis for the inhibition of Polo-like kinase 1. Nature Structural & Molecular Biology. 2013;20:1047–1053. doi: 10.1038/nsmb.2623. [PubMed] [Cross Ref]
  • Xu X, Vaithiyalingam S, Glick GG, Mordes DA, Chazin WJ, Cortez D. The basic cleft of RPA70N binds multiple checkpoint proteins, including RAD9, to regulate ATR signaling. Molecular and Cellular Biology. 2008;28:7345–7353. doi: 10.1128/MCB.01079-08. [PMC free article] [PubMed] [Cross Ref]
  • Yata K, Lloyd J, Maslen S, Bleuyard JY, Skehel M, Smerdon SJ, Esashi F. Plk1 and CK2 act in concert to regulate Rad51 during DNA double strand break repair. Molecular Cell. 2012;45:371–383. doi: 10.1016/j.molcel.2011.12.028. [PMC free article] [PubMed] [Cross Ref]
  • Yekezare M, Gómez-González B, Diffley JF. Controlling DNA replication origins in response to DNA damage - inhibit globally, activate locally. Journal of Cell Science. 2013;126:1297–1306. doi: 10.1242/jcs.096701. [PubMed] [Cross Ref]
  • Yoo HY, Kumagai A, Shevchenko A, Shevchenko A, Dunphy WG. Adaptation of a DNA replication checkpoint response depends upon inactivation of Claspin by the Polo-like kinase. Cell. 2004;117:575–588. doi: 10.1016/S0092-8674(04)00417-9. [PubMed] [Cross Ref]
  • Zegerman P, Diffley JF. Phosphorylation of Sld2 and Sld3 by cyclin-dependent kinases promotes DNA replication in budding yeast. Nature. 2007;445:281–285. doi: 10.1038/nature05432. [PubMed] [Cross Ref]
  • Zegerman P, Diffley JF. DNA replication as a target of the DNA damage checkpoint. DNA Repair. 2009;8:1077–1088. doi: 10.1016/j.dnarep.2009.04.023. [PubMed] [Cross Ref]
  • Zou L, Cortez D, Elledge SJ. Regulation of ATR substrate selection by Rad17-dependent loading of Rad9 complexes onto chromatin. Genes & Development. 2002;16:198–208. doi: 10.1101/gad.950302. [PubMed] [Cross Ref]
  • Zou L, Elledge SJ. Sensing DNA damage through ATRIP recognition of RPA-ssDNA complexes. Science. 2003;300:1542–1548. doi: 10.1126/science.1083430. [PubMed] [Cross Ref]
  • Zou L, Liu D, Elledge SJ. Replication protein A-mediated recruitment and activation of Rad17 complexes. PNAS. 2003;100:13827–13832. doi: 10.1073/pnas.2336100100. [PubMed] [Cross Ref]
2017; 6: e29953.

Decision letter

Marcus B Smolka, Reviewing Editor
Marcus B Smolka, Cornell University, United States;

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "CDK-PLK1 axis to target DNA checkpoint sensor protein RAD9 to tolerate genotoxic stress for promoting cell proliferation" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Jessica Tyler as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Dana Branzei (Reviewer #1); Rodrigo Bermejo (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

The 9-1-1 clamp plays a key role in activation of the DNA damage checkpoint. 9-1-1 is loaded at sites of DNA damage or replication-induced lesions, where it recruits the ATR kinase activator TOPBP1 for initiation of DNA damage checkpoint signaling. The recruitment of TOPBP1 to 9-1-1 is dependent on the CK2 kinase, which phosphorylates the 9-1-1 subunit RAD9 on serine 387 and serine 341. The role of these RAD9 phosphorylation sites has been well documented, however, much less is understood about additional phosphorylation events on RAD9, especially in mammalian systems. The manuscript by Furuya et al. reports that human RAD9 is also phosphorylated by the CDK and PLK1 kinases, and proposes that the coordinated action of CDK/PLK1 on RAD9 is important to counteract the DNA damage checkpoint. The authors claim that phosphorylation of RAD9 by CDK on threonine 292 creates a docking site for binding to PLK1, which in turn, further phosphorylates RAD9 on threonine 313 and serine 326. These claims are supported by in vitro and in vivo experiments. In addition, the authors generated human cell lines that stably express distinct Rad9 phospho-mutants and show that expression of these mutants results in enhanced checkpoint activation, reduced rates of DNA synthesis and slight sensitivity to hydroxyurea treatment.

Overall, all reviewers agreed that this is an interesting paper and that it provides important contributions to the field. The finding that the CDK and PLK1 kinases can modulate 9-1-1 function has important implications for understanding checkpoint control and the integration of checkpoint signaling with other cellular processes. The biochemical experiments were considered strong and well executed. Several useful reagents were generated, including phospho-specific antibodies and Rad9 point mutants.

Essential revisions:

1) All reviewers were concerned with the writing quality of the manuscript. A revised version of the manuscript will need to be much improved in this regard, and we strongly encourage editing by a native English speaker. The title of the manuscript also needs to be improved for clarity.

2) A major claim made by the authors is that CDK/PLK1-mediated RAD9 phosphorylation promote DNA synthesis, and they speculate that this happens largely by deregulating origin firing. However, as pointed out by the reviewers, the observed changes in DNA synthesis could be caused by changes in fork velocity. To discriminate between origin firing and fork speed, the authors should perform a molecular combing experiment in the presence of HU, where both fork rate and inter-origin distance can be measured. Genomic Vision provides such service in case the authors do not have the technique set up.

3) It is somewhat disappointing that the effects of expressing Rad9 phospho-mutants on checkpoint signaling, cell cycle and DNA synthesis are quite subtle. This could be caused by the presence of endogenous wild-type Rad9. Therefore, it would be worth generating a system to knockdown or knockout endogenous RAD9 in cell lines expressing the mutants.

4) The effect of expressing Rad9 phospho-mutants on cell proliferation was only examined in U2OS cells, which are rather unique in the sense that they use ALT for telomere lengthening and may have atypical mechanisms for checkpoint control. It is therefore important that the authors examine the effect of expressing Rad9 phospho-mutants in cell proliferation and DNA synthesis using other cells lines that do not use ALT.

5) The paper is mostly based on experiments using hydroxyurea-induced replication stress. The effect of other genotoxic stresses, such as DNA alkylation, doubled stranded breaks and inter-strand crosslinks should be tested. Examining how drugs such as MMS, CPT and MMC (and perhaps ionizing radiation) affect RAD9 phosphorylation by CDK and PLK1 will be useful to better characterize how these phosphorylation events are modulated. Furthermore, the analysis of DNA synthesis (EdU incorporation and FACS) in cells expressing Rad9 phospho-mutants treated with at least one of these other forms of genotoxicity will be useful.

6) The paper is lacking a better discussion on the mechanism by which RAD9 phosphorylation affect checkpoint signaling. More specifically, it would be useful if the authors expand on how Rad9 phosphorylation could be controlling its chromatin localization. For example, is phosphorylation preventing 9-1-1 loading? and promoting unloading from DNA?

7) The paper would also benefit from a more in-depth discussion of the physiological impact of Rad9 phosphorylation on the control of checkpoint response and cell proliferation. It is currently unclear if the authors are claiming that the action of CDK/PLK1 on RAD9 is part of a damage tolerance response (which promotes cell proliferation even in the presence of continued damage), or if it is a mechanism to promote quicker dampening of checkpoint signaling and timely resumption of cell cycle progression.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "CDK-PLK1 axis targets DNA checkpoint sensor protein RAD9, promoting tolerance to genotoxic stress and cell proliferation" for further consideration at eLife. Your revised article has been favorably evaluated by Jessica Tyler (Senior editor), Marcus Smolka (Reviewing editor), and one of the reviewers.

The manuscript has been clearly improved on the experimental side. However, the written English was still considered poor. Therefore, you will need to significantly improve the text, perhaps by seeking help from a different company for improved writing services.

For the title, I would suggest the following:

"The CDK-PLK1 axis targets the DNA damage checkpoint sensor protein RAD9 to promote cell proliferation and tolerance to genotoxic stress"

Throughout the text, I urge you to refer to "DNA damage checkpoint" instead of "DNA checkpoint", since the latter is not very commonly used.

Please keep in mind that the written English of your whole manuscript needs to be significantly improved. For the abstract, I am pasting below a suggested version, which I have modified based on your latest abstract.

Suggested abstract:

"When proliferating cells encounter genotoxic stress, the DNA damage checkpoint is activated to assist DNA damage recovery by slowing down cell cycle progression. Thus, to drive proliferation, cells need to tolerate DNA damage and suppress the checkpoint response. However, the mechanism underlying this negative regulation of checkpoint activation is still elusive. Here we show that human Cyclin-Dependent-Kinases (CDKs) target the RAD9 subunit of the 9-1-1 checkpoint clamp on Thr292 for modulation of DNA damage checkpoint activation. Thr292 phosphorylation on RAD9 creates a binding site for Polo-Like-Kinase 1 (PLK1), which further phosphorylates RAD9 on Thr313. These CDK-PLK1-dependent phosphorylation events on RAD9 suppress checkpoint activation, therefore maintaining high rates of DNA synthesis in the presence of DNA replication stress. Our results suggest that CDK locally initiates a PLK1-dependent signaling response that antagonizes the ability of the DNA damage checkpoint to detect DNA damage. These findings provide a mechanism for how replicating cells suppress DNA damage checkpoint signaling to promote proliferation in the presence of genotoxic stress."

Essential revisions:

1) All reviewers were concerned with the writing quality of the manuscript. A revised version of the manuscript will need to be much improved in this regard, and we strongly encourage editing by a native English speaker. The title of the manuscript also needs to be improved for clarity.

Our manuscript was checked (grammar) by a professional company.

We also changed the title from “PLK1-CDK axis to target DNA checkpoint sensor protein RAD9 to tolerate genotoxic stress for promoting cell proliferation” to “CDK-PLK1 axis targets DNA checkpoint sensor protein RAD9, promoting tolerance to genotoxic stress and cell proliferation”.

2) A major claim made by the authors is that CDK/PLK1-mediated RAD9 phosphorylation promote DNA synthesis, and they speculate that this happens largely by deregulating origin firing. However, as pointed out by the reviewers, the observed changes in DNA synthesis could be caused by changes in fork velocity. To discriminate between origin firing and fork speed, the authors should perform a molecular combing experiment in the presence of HU, where both fork rate and inter-origin distance can be measured. Genomic Vision provides such service in case the authors do not have the technique set up.

Thank you for this comment. We performed DNA combing assay to look into a molecular level to assess which steps of DNA checkpoint dependent processes are affected in the RAD9 mutant (RAD9-T313A; RAD9 that is not phosphorylated by PLK1). The results (Figure 5) gave us rather unexpected observation, especially in non-stressed cells, but it fits our model. First, when cells were under replicative stresses (HU), the replication fork rates were suppressed to the same level both in WT and T313A mutant. This indicates that the reduction observed in the bulk DNA synthesis rate in T313A mutant was not due to the reduction in replication fork rate. Furthermore, we observed a slight increase in inter-origin distance, which suggests that the dormant origins were rather repressed, in RAD9-T313A mutant cells. Although the difference is marginally significant (IOD; WT: 61 ± 38 kb, T313A: 69 ± 41 kb, p-value; 0.058), it is possible that the repression of dormant origin firing could be the part of the reason for reduction in bulk DNA synthesis rate in T313A mutant.

The finding we had in untreated cells was unexpected. When the mutant cells were not under the stress, we observed a great reduction in replication fork rate in T313A mutant (WT: 0.85 ± 0.36 kb/minutes, T313A: 0.66 ± 0.27 kb/minutes, p-value: 5.8 x10-7). In contrast, the inter-origin distances (IOD) were shortened in T313A mutant (WT: 131 ± 62 kb, T313A: 107 ± 61 kb, p-value: 8.0 x 10-8). This reduction in IOD suggested more activation of dormant origins compared to the WT and could indicate that reduction in the fork rate can be compensated by increasing efficiency of firing (dormant) origins, as suggested in some reports (e.g.: Iyer and Rhind, 2013). These combing data were obtained from Genomic Vision.

In parallel, we also examined the fork rate under aphidicolin-induced stress (Figure 5—figure supplement 1) as we had already observed a slight reduction in bulk DNA synthesis rate in aphidicolin-treated cells (Figure 4—figure supplement 6). The experiments were performed by one of our authors. We saw the fork rate was also similarly suppressed both in WT and T313A cells under aphidicolin-induced stress and that again suggested that the reduction in bulk DNA synthesis rate was not due to reduction in the fork rate. [We should note that there are technical differences between Genomic Vision and our work, as absolute values of the fork rate we obtained was smaller than the value of that we obtained from Genomic Vision. This is partly caused by the different buffer, glasses or technique used upon stretching DNA fiber on the glass, between us and Genomic Vision. Furthermore, difference in evaluators for the DNA forks may pick the track in different criteria (cut off size of the minimal fork size, how to exclude the fusion forks, etc). We alluded to this issue in the text and in the Materials and methods section].

We interpreted the above data in a non-stressed condition – in HU and in aphidicolin – as follows: PLK1-dependent phosphorylation on RAD9 is functioning even during normal S phase (that supports why T292/T313 phosphorylation could be seen even in non-stressed cells), and defective phosphorylation may lead to the moderate activation of S phase checkpoint that could reduce the fork speed, and in those situation, possibly dormant (or other) origin fires, and the cells compensate bulk DNA synthesis rate. However, when they suffer from 0.2 mM HU, although WT cells somehow drive cell proliferation possibly by activating dormant or late origin firing, T313A mutants failed to proliferate because the excess level of S-phase checkpoint activation may inhibit the origin firings that should help maintaining rate of bulk DNA synthesis. Thus, we think that controlling RAD9 is important to have an appropriate level of checkpoint activation to proliferate under the genotoxic stress.

We added Figure 5 (HU 0.2 mM) and Figure 5—figure supplement 1 (Aphidicolin) to show this point. Also, because these data somewhat responded to our previous discussion, we rearranged the flow of the sections within the Discussion (although the order of the Discussion section has been rearranged, the content is unchanged). We believe that the present form of the Discussion is more fluent and easier to follow. The next text in the Results, Discussion Materials and methods sections follows:

Subsection “PLK1-dependent phosphorylation of RAD9 Thr313 is required to suppress S-phase checkpoint response”:

“We went on to perform DNA combing assay to see which steps of DNA replication contribute to slow down bulk DNA synthesis. […] Also, it should be noted that the replication fork rate was already slower in RAD9-T313A expressing cells from non-stressed condition, which may suggest the enhanced checkpoint response in RAD9-T313A expressing cells in normal cell cycle (see Discussion section).”

Subsection “Phosphorylation of the checkpoint sensor protein by CDK and PLK1 is required for maintaining the rate of DNA replication upon genotoxic stress”:

“When we performed DNA fiber/combing assay in HU treated condition, we observed that velocity of replication forks was the same in WT- and T313A-RAD9 expressing cells. […] In T313A-expressing cells, further encounter to HU-induced stress would activate excess DNA damage response and could cause a reduction in number of firing origins, which would result in slower rate of S-phase progression.”

Subsection “DNA molecular combing assay”:

DNA combing assay was performed as described (Conti et al., 2007; Kuriya et al., 2015; Michalet et al., 1997). […] Thus, we emphasize on the relative difference in values within the samples obtained from the same series.”

3) It is somewhat disappointing that the effects of expressing Rad9 phospho-mutants on checkpoint signaling, cell cycle and DNA synthesis are quite subtle. This could be caused by the presence of endogenous wild-type Rad9. Therefore, it would be worth generating a system to knockdown or knockout endogenous RAD9 in cell lines expressing the mutants.

Thank you for the comment. Our original RAD9 plasmids used to create stable cell line were actually constructed as siRNA resistant form, and we knocked down endogenous RAD9 to see whether the effect of phosphorylation defective mutants (see Figure 4—figure supplement 4). We see some difference compared to cells that did not apply siRAD9. As for the T313A-mutated RAD9, we see the enhancement of checkpoint phenotype (CHK1-pS345, Figure 4—figure supplement 4B). We also see a slight reduction of DNA replication rate assessed by EdU incorporation compared to the control cells (Figure 4—figure supplement 4C). As for S326A-mutated RAD9 expressing cells that did not show the checkpoint phenotype as we reported in this manuscript, we did not see neither increase in CHK1-pS345 phosphorylation nor decrease in DNA replication rate compared to siRNA-control treated cells (Figure 4—figure supplement 4D, E). And we concluded that S326A-mutated RAD9 cells do not show accelerated checkpoint activation phenotype. Thus, we added the text as follows:

Subsection “PLK1-dependent phosphorylation of RAD9 Thr313 is required to suppress S-phase checkpoint response”:

“We further knocked down endogenous RAD9 (Figure 4—figure supplement 4A) to see whether the observed cellular effect by the expression of T313A- or S326A-mutated RAD9 is enhanced. […] In cells that expressed S326A-mutated RAD9, silencing did not result in an obvious enhancement in phosphorylation of Ser345 of CHK1 or in a decrease in DNA replication rate (Figure 4—figure supplement 4D, E).”

4) The effect of expressing Rad9 phospho-mutants on cell proliferation was only examined in U2OS cells, which are rather unique in the sense that they use ALT for telomere lengthening and may have atypical mechanisms for checkpoint control. It is therefore important that the authors examine the effect of expressing Rad9 phospho-mutants in cell proliferation and DNA synthesis using other cells lines that do not use ALT.

Thank you for the comment; this was very helpful. We tested in HEK293A cells that was reportedly telomerase-positive (Bryan et al., 1995). As for the checkpoint phenotype using HEK293A, we have already performed in Figure 1D and Figure 4A, and showed increased chromatin-bound CHK1-pS345. Here, we have added the data to show the reduction in the rate of DNA replication could be observed in HEK293A cells in Figure 4—figure supplement 5. We also performed proliferation assay and showed that RAD9-T292A and RAD9-T313A both presented decreased ability to proliferate under HU stress (Figure 6—figure supplement 1). We added the text as follows:

Subsection “PLK1-dependent phosphorylation of RAD9 Thr313 is required to suppress S-phase checkpoint response”:

“Furthermore, we also confirmed the reduction in DNA replication rate in telomerase-positive HEK293A cells (Bryan et al., 1995) when either T292A- or T313A-mutated RAD9 was expressed (Figure 4—figure supplement 5A and B). This result excluded the possibility that the slowing of DNA replication rate was caused by the atypical mechanism for checkpoint control which may occur in U2OS cells via Alternative Lengthening of Telomeres (ALT) (Figure 4—figure supplement 5A and B).”

Subsection “CDK- and PLK1- dependent phosphorylation of RAD9 allows cells to grow under genotoxic stress”:

“Indeed, we observed that cells expressing T292A- or T313A-mutated RAD9 exhibited a slower growth rate in the presence of HU (As for U2OS cells: Figure 6A, As for HEK293A cells: Figure 6—figure supplement 1).”

5) The paper is mostly based on experiments using hydroxyurea-induced replication stress. The effect of other genotoxic stresses, such as DNA alkylation, doubled stranded breaks and inter-strand crosslinks should be tested. Examining how drugs such as MMS, CPT and MMC (and perhaps ionizing radiation) affect RAD9 phosphorylation by CDK and PLK1 will be useful to better characterize how these phosphorylation events are modulated. Furthermore, the analysis of DNA synthesis (EdU incorporation and FACS) in cells expressing Rad9 phospho-mutants treated with at least one of these other forms of genotoxicity will be useful.

Thank you for the comment. We have tested to see the phosphorylation status in response to MMS, MMC, CPT, IR or Aphidicolin (Figure 3—figure supplement 3A). Since we only see a little checkpoint response in MMS (0.001%) and MMC (0.2ug/ml), we re-did the experiment using higher doses (Figure 3—figure supplement 3B). The drugs were added in average one hour before harvest, so as not to disturb the overall cell cycle profiles. In most cases, we saw either sustained or a slight increased efficiency in PLK1-dependent phosphorylation: pT313 or pS326.

We have further tested two of these stresses (MMC and aphidicolin, Figure 4—figure supplement 6) for the EdU analysis to measure the rate of bulk DNA replication. In aphidicolin-treated cells, we observed reduction in DNA replication rate to 90% in RAD9-T313A mutant cells. In MMC-treated cells, although we did not observe the obvious reduction in the rate of DNA replication, we observed more cells accumulated in early S phase in RAD9-T313A mutant cells.

These indicated that, as for the phenotype of reduction in DNA replication rate, it could be specific to the direct DNA replication inhibitor, but not to the DNA damaging agents. Thus, we added text to explain the results for the phosphorylation status as follows (subsection “PLK1 phosphorylates Thr313 and Ser326 on RAD9 in vivo”), and the EdU assay with aphidicolin or MMC-treated cells in (subsection “PLK1-dependent phosphorylation of RAD9 Thr313 is required to suppress S-phase checkpoint response”);

Subsection “PLK1 phosphorylates Thr313 and Ser326 on RAD9 in vivo”:

“We further tested the effects of other types of DNA damage on PLK1-dependent phosphorylation of RAD9. Treatment with camptothecin (CPT: topoI inhibitor), ionizing radiation (IR), or high doses of methylmethanesulfonate (MMS: 0.03%) or mitomycin C (MMC:1 μg/ml) yielded slight increases in pSer326 phosphorylation (Figure 3—figure supplement 3A and B). Aphidicolin (Aph: a DNA polymerase inhibitor) treatment reduced the levels of both pThr313 and pSer 326, however, pThr292 was also decreased, and thus these results suggested that the efficiency of PLK1-dependent phosphorylation is either maintained or slightly enhanced in response to most types of genotoxic stresses (Figure 3—figure supplement 5).”

Subsection “PLK1-dependent phosphorylation of RAD9 Thr313 is required to suppress S-phase checkpoint response”:

“A slowing of DNA replication rate also was observed in response to other types of DNA damage. Specifically, aphidicolin treatment also reduced the rate of EdU incorporation, and MMC treatment caused more cells to accumulate in early S phase when T313A-mutated RAD9 was expressed, although the latter treatment did not yield an obvious decrease in EdU incorporation (Figure 4—figure supplement 6). The observed phenotype may have been specific to the direct inhibition of DNA replication by some agents (HU, a dNTP synthesis inhibitor; aphidicolin, a DNA polymerase inhibitor) but not by others (MMC, a DNA damaging agent that crosslinks DNA strands).”

6) The paper is lacking a better discussion on the mechanism by which RAD9 phosphorylation affect checkpoint signaling. More specifically, it would be useful if the authors expand on how Rad9 phosphorylation could be controlling its chromatin localization. For example, is phosphorylation preventing 9-1-1 loading? and promoting unloading from DNA?

Thank you for the comment. It is quite difficult to distinguish whether the chromatin localization of RAD9 is controlled via loading or unloading. At present how 9-1-1 is unloaded from DNA damage sites is not fully understood. It could be possible that the clamp loader RAD17 could act as the unloader of 9-1-1. We think that, in our case, whether 9-1-1 is loaded or unloaded on chromatin could depend on the affinity of the 9-1-1 complex itself to DNA damage site (e.g. whether RPA-binding motif on RAD9 is available or not). At least we think the phospho-dependent chromatin dissociation is not controlled via regulating the interaction to a clamp loader RAD17, as T292A mutation did not affect the binding to it (IP analysis; data not shown).

We added the text as follows (subsection “CDK and PLK1 target RAD9 to modulate DNA damage detection”:

“The sites of PLK1-dependent phosphorylation of RAD9 are located within the tail region. This region of RAD9 is expected to help the protein’s association to DNA damage sites, for example the region phosphorylate by PLK1 is close to the region required for association to the single-stranded-DNA binding protein RP-A (Figure 3—figure supplement 1) (Xu et al., 2008). Furthermore, we did not observe a decrease in the amount of the RAD9-bound clamp-loader RAD17 in co-IP experiment when T292 was mutated, even under condition of genotoxic stress (data not shown). Thus, PLK1-dependent phosphorylation may affect the affinity of the 9-1-1 complex itself to the DNA damage sites.”

7) The paper would also benefit from a more in-depth discussion of the physiological impact of Rad9 phosphorylation on the control of checkpoint response and cell proliferation. It is currently unclear if the authors are claiming that the action of CDK/PLK1 on RAD9 is part of a damage tolerance response (which promotes cell proliferation even in the presence of continued damage), or if it is a mechanism to promote quicker dampening of checkpoint signaling and timely resumption of cell cycle progression.

Thank you for the comment. Since we are presenting the data to show the increased amount of RAD9 onto the chromatin in the mutant, we propose that PLK1 antagonize the function of DNA damage sensing ability of RAD9. Thus, we think it controls at the activation step of checkpoint rather than the dampening the activated checkpoint. And, at the moment, we think this is a part of the novel form of DNA tolerance mechanism. Our results may explain why PLK1 overexpression is linked to cancer cells as we discussed in the last paragraph of the Discussion: “In this scenario, PLK1 may raise a threshold of DNA damage levels, which would require more DNA damage stress for the S phase checkpoint activation. In other words, PLK1-dependent phosphorylation of RAD9 may suppress the activation step of the checkpoint, rather than dampening the activated DNA checkpoint processes, and this process could act as a novel form of DNA damage tolerance mechanism to facilitate cell proliferation.”

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "CDK-PLK1 axis targets DNA checkpoint sensor protein RAD9, promoting tolerance to genotoxic stress and cell proliferation" for further consideration at eLife. Your revised article has been favorably evaluated by Jessica Tyler (Senior editor), Marcus Smolka (Reviewing editor), and one of the reviewers.

The manuscript has been clearly improved on the experimental side. However, the written English was still considered poor. Therefore, you will need to significantly improve the text, perhaps by seeking help from a different company for improved writing services.

Thank you for the suggestion. We sought English editing services from a native English speaker (a recognized scientific English editor), and we believe that the manuscript is now easier to read. We also changed the term “DNA checkpoint” to “DNA damage checkpoint”. The content of the manuscript remains unchanged. However, upon the initial checking process, a member of the eLife staff has suggested to re-assemble the figures, and we have swapped the order of “Figure 3—figure supplement 3F and G” and “Figure 3—figure supplement 3A and B”. Accordingly, the text has been modified.

For the title, I would suggest the following:

"The CDK-PLK1 axis targets the DNA damage checkpoint sensor protein RAD9 to promote cell proliferation and tolerance to genotoxic stress"

Throughout the text, I urge you to refer to "DNA damage checkpoint" instead of "DNA checkpoint", since the latter is not very commonly used.

Please keep in mind that the written English of your whole manuscript needs to be significantly improved. For the abstract, I am pasting below a suggested version, which I have modified based on your latest abstract.

Suggested abstract:

"When proliferating cells encounter genotoxic stress, the DNA damage checkpoint is activated to assist DNA damage recovery by slowing down cell cycle progression. Thus, to drive proliferation, cells need to tolerate DNA damage and suppress the checkpoint response. However, the mechanism underlying this negative regulation of checkpoint activation is still elusive. Here we show that human Cyclin-Dependent-Kinases (CDKs) target the RAD9 subunit of the 9-1-1 checkpoint clamp on Thr292 for modulation of DNA damage checkpoint activation. Thr292 phosphorylation on RAD9 creates a binding site for Polo-Like-Kinase 1 (PLK1), which further phosphorylates RAD9 on Thr313. These CDK-PLK1-dependent phosphorylation events on RAD9 suppress checkpoint activation, therefore maintaining high rates of DNA synthesis in the presence of DNA replication stress. Our results suggest that CDK locally initiates a PLK1-dependent signaling response that antagonizes the ability of the DNA damage checkpoint to detect DNA damage. These findings provide a mechanism for how replicating cells suppress DNA damage checkpoint signaling to promote proliferation in the presence of genotoxic stress."

The abstract has been modified. We adopted the abstract suggested in the previous letter. However, since the number of words exceeded 150 words, we asked the English editor to shorten it. Thus, the included abstract is now as follows:

“Genotoxic stress causes proliferating cells to activate the DNA damage checkpoint, to assist DNA damage recovery by slowing cell cycle progression. Thus, to drive proliferation, cells must tolerate DNA damage and suppress the checkpoint response. However, the mechanism underlying this negative regulation of checkpoint activation is still elusive. We show that human Cyclin-Dependent-Kinases (CDKs) target the RAD9 subunit of the 9-1-1 checkpoint clamp on Thr292, to modulate DNA damage checkpoint activation. Thr292 phosphorylation on RAD9 creates a binding site for Polo-Like-Kinase1 (PLK1), which phosphorylates RAD9 on Thr313. These CDK-PLK1-dependent phosphorylations of RAD9 suppress checkpoint activation, therefore maintaining high DNA synthesis rates during DNA replication stress. Our results suggest that CDK locally initiates a PLK1-dependent signaling response that antagonizes the ability of the DNA damage checkpoint to detect DNA damage. These findings provide a mechanism for the suppression of DNA damage checkpoint signaling, to promote cell proliferation under genotoxic stress conditions.”

We also made the following changes:

Change: “DNA checkpoint” => “DNA damage checkpoint” throughout the manuscript.

Change: “DNA checkpoint” => “DNA damage checkpoint in S phase” in the Introduction.


Articles from eLife are provided here courtesy of eLife Sciences Publications, Ltd