Regulation and roles of Cdc7 kinase under replication stress

Abstract

Cdc7 (cell division cycle 7) kinase together with its activation subunit ASK (also known as Dbf4) play pivotal roles in DNA replication and contribute also to other aspects of DNA metabolism such as DNA repair and recombination. While the biological significance of Cdc7 is widely appreciated, the molecular mechanisms through which Cdc7 kinase regulates these various DNA transactions remain largely obscure, including the role of Cdc7-ASK/Dbf4 under replication stress, a condition associated with diverse (patho)physiological scenarios. In this review, we first highlight the recent findings on a novel pathway that regulates the stability of the human Cdc7-ASK/Dbf4 complex under replication stress, its interplay with ATR-Chk1 signaling, and significance in the RAD18-dependent DNA damage bypass pathway. We also consider Cdc7 function in a broader context, considering both physiological conditions and pathologies associated with enhanced replication stress, particularly oncogenic transformation and tumorigenesis. Furthermore, we integrate the emerging evidence and propose a concept of Cdc7-ASK/Dbf4 contributing to genome integrity maintenance, through interplay with RAD18 that can serve as a molecular switch to dictate DNA repair pathway choice. Finally, we discuss the possibility of targeting Cdc7, particularly in the context of the Cdc7/RAD18-dependent translesion synthesis, as a potential innovative strategy for treatment of cancer.

Background

The serine–threonine kinase Cdc7 is essential for chromosomal DNA replication in a wide spectrum of species, being conserved from yeast to human.^1-4 The activity of the catalytic subunit Cdc7 is positively regulated by a complex formation with its activation subunit, the Dbf4 protein, also known as ASK (activator of S-phase kinase in human).^5-10 The Cdc7–ASK/Dbf4 complex is therefore also commonly referred to as the Dbf4-dependent kinase, or DDK in short. DDK regulates the timing of DNA replication origin firing throughout S phase mainly by phosphorylation of MCM proteins, the major components of replicative helicase.^11-13

When the replication fork stalls, due either to endogenous obstacles or upon exposure of a cell to various genotoxic insults, the DNA replication checkpoint pathway becomes activated¹⁴ and suppresses further origin firing or slows down the fork progression. In budding yeast Saccharomyces cerevisiae, Rad53 checkpoint kinase (an ortholog of human Chk2) directly phosphorylates Dbf4 upon fork stalling to block late origin firing, most likely through downregulation of the DDK activity,^15-18 although this is not proven yet. Similarly, an in vitro study using Xenopus egg extracts revealed that DDK activity is inhibited upon DNA damage caused by etoposide treatment in an ATR kinase-dependent manner.¹⁹ However, later work using a similar experimental system presented a contradictory result, namely preserved DDK activity upon DNA damage.²⁰ It is also controversial whether the human DDK is inactivated under genotoxic stress conditions. Previously, it was reported that Cdc7 kinase activity in budding yeast is downregulated in response to hydroxyurea treatment that arrests cells in early S phase through inhibition of ribonucleotide reductase.¹⁸ Earlier work with human cells also concluded that the DDK activity is inhibited upon etoposide treatment through the ATR-dependent inhibition of formation of the human DDK complex.²¹ However, more recent reports proposed that the human DDK is not inactivated under various genotoxic treatments.^20,22 Recent reports showed that the human ASK/Dbf4 is a direct substrate of the upstream DNA damage response kinases ATM and/or ATR, and that such phosphorylation of ASK/Dbf4 is required for activation of the so-called intra-S phase checkpoint that suppresses DNA replication upon exposure of S-phase cells to ionizing radiation.²³ This is consistent with the possibility that the human DDK function is downregulated upon genotoxic stress, which would lead to inhibition of the late origin firing. Interestingly, however, the authors found that the phosphorylation of ASK/Dbf4 did not affect the kinase activity of Cdc7, indicating that this regulatory phosphorylation may impact a currently unknown mechanism, such as regulation of some protein–protein interactions that may facilitate suppression of late origin firing in human cells.²³ It should be also noted that the human DDK activity is required for Chk1 activation, most probably through phosphorylating Claspin, an adaptor protein essential for ATR-dependent phosphorylation of Chk1,^24,25 consistent with the idea that the DDK is active under replication stress.

Taken together, such divergent and partly contradictory reports highlight the important questions to be addressed regarding the human DDK: (1) Is the human DDK activity preserved under genotoxic/replication stress or not? (2) If it does remain active, is it only a certain subset of Cdc7 that is active, and what is the role of the DDK activity under such conditions?

In addition, although cell cycle-dependent regulation of transcription was reported for the components of the mammalian DDK,^4,26,27 it has been unclear whether and how the protein abundance of the DDK subunits is regulated upon replication block, and whether there are any differences in such regulatory mechanisms between unperturbed conditions vs. replication stress. Physiological and pathological roles of the DDK activity under genotoxic stress should also be addressed, considering that DDK has been proposed as a promising target of anticancer treatment.^28-35

Stabilization of DDK on Chromatin in Response to Replication Stress

Recently, fresh insights have been provided into these important issues. New results showed that under replication stress conditions, human DDK accumulates as an active complex on chromatin, and that such stabilization of the DDK protein components is mediated by the ATR–Chk1 kinase pathway-dependent inactivation of the anaphase-promoting complex/cyclosome-Cdh1(APC/C^Cdh1).³⁶ The APC/C^Cdh1 ubiquitin ligase is one of the major cell cycle-regulatory machines, which becomes activated in later mitosis, remains active throughout the following G₁ phase, and then is switched off by the enhanced CDK kinase activity (CDK1 and CDK2 in human cells) from the G₁/S boundary until early M phase.^37-40 Since the replication stress-responsive ATR–Chk1 signaling inhibits CDK activity (including CDK2 in S phase) via checkpoint pathways, such silencing of CDK could, in principle, lead to re-activation of the APC/C^Cdh1 at a “wrong” cell cycle phase, for example in S phase, thereby grossly deregulating the DNA replication factories and other S-phase factors. As the recent work clarified,³⁶ this conundrum is solved through concomitant Chk1-medietad inactivation of the APC/C^Cdh1 complex, through promoting autodegradation of Cdh1 under replication stress conditions. Thus, the ATR–Chk1 axis silences the CDKs via checkpoints to delay replication, and at the same time ensures that the APC/C^Cdh1- mediated proteolysis is not inappropriately activated in S phase.

Interestingly, and in contrast to replication stress response in S phase, the APC/C^Cdh1 pathway becomes active in response to DNA double-strand breaks induced by X-ray irradiation or doxorubicin treatment, and this plays an important role in G₂ checkpoint activation through degradation of PLK1, an essential kinase that regulates mitotic entry.^41,42

It should be noted that DNA repair factors such as 53BP1 and Rap80 are among candidate substrates targeted for degradation by the APC/C^Cdh1 ubiquitin ligase,^43,44 suggesting that the APC/C^Cdh1 pathway may be also involved in regulation of homologous recombination (HR), given a role of these proteins in restricting DNA end resection, an essential step in HR.^45,46

Furthermore, the DDK activity itself is required for the stabilization of the human Dbf4/ASK protein, suggesting that one of the important roles of the DDK activity under replication stress is to stabilize ASK/Dbf4, and thereby the DDK complex.³⁶

Central Role of DDK in TLS-Mediated DNA Damage Repair: Concept of the DDK-RAD18 Interaction as a Molecular Switch in DNA Repair Pathway Choice

So, what could be the biological advantage for the active DDK to be stabilized on chromatin under replication stress? We believe, based on our own data³⁶ and work done mainly on yeast models (see below), that the answer is in the choice of DNA repair pathway, in order to preserve genomic integrity while preventing collapse of the assembled replication fork machineries under conditions of replication stress. Previous studies in yeast strongly suggested that the DDK plays a crucial role in translesion synthesis (TLS),^47,48 a mechanism that enables continued DNA replication across the encountered DNA lesion. More recently, human DDK was shown to phosphorylate a cluster of serine residues on RAD18, and this positively regulates interaction between RAD18 and DNA polymerase η (Pol η).^49,50 Furthermore, motif-C, a conserved C2H2-type zinc finger domain of the ASK/Dbf4 protein, interacts with the N-terminal region of RAD18, the ubiquitin ligase essential for TLS.³⁶ This ASK/Dbf4–RAD18 interaction is required for chromatin binding of RAD18 and RAD18-mediated recruitment of Pol η, another key factor for TLS.³⁶ Therefore, these results connect the ATR–Chk1 checkpoint pathway to the TLS pathway via the DDK. In addition, the complex formation of RAD18-Pol η plays a key role in the recruitment of the complex to PCNA, through interaction between the PIP box motif of Pol η and PCNA.⁵¹ Based on these studies, we propose here a conceptual model for how the checkpoint signaling regulates the TLS pathway in human cells. Thus, the checkpoint kinases help to stabilize the active DDK on chromatin to protect the replication fork from collapse and promote the TLS under replication stress caused by DNA lesions (Fig. 1).

Figure 1. Stabilization of chromatin-bound DDK upon replication stress and induction of TLS through DDK-mediated recruitment of RAD18–Pol η complex. (A) Replication forks encounter a DNA lesion. (B) Stalled forks activate the replication checkpoint, which stabilizes DDK on chromatin through inactivation of the APC/C^Cdh1 pathway. DDK phosphorylates RAD18 and facilitates interaction between RAD18 and Pol η. RAD18 monoubiquitinates PCNA, independently of DDK. (C) The DDK–RAD18-Pol η complex is loaded on chromatin to promote TLS.

Moreover, the observed interaction between the N-terminal region of RAD18, containing the RING domain, and the ASK/Dbf4³⁶ raises an attractive hypothesis about the regulation of pathway choice under conditions of DNA damage/replication stress (Fig. 2). Intriguingly, the RING domain of RAD18 also interacts with Rad51C, an essential factor for HR, and this interaction is required for the HR process to function.⁵² It should also be considered that while the TLS pathway is one of the major pathways to continue DNA replication past the DNA lesion, the HR is an alternative pathway to restart DNA replication at the site of stalled fork, functioning predominantly during S phase and G₂ phases.⁵³ Overall, the mechanistic basis of a molecular switch between TLS and HR, for which we can now suggest a plausible scenario, has been poorly understood until now, certainly for human cells. The interaction between ASK/Dbf4 and RAD18 may suppress HR by preventing RAD18–Rad51C complex formation. Further enhancing the flexibility of this candidate regulatory “switch” role of the RING domain of RAD18, this domain also interacts with FANCD2, a protein essential for repair of DNA inter-strand crosslinks by the Fanconi anemia (FA) pathway.⁵⁴ We propose that the choice of the binding partner for RAD18’s RING domain may provide the long-thought switch mechanism, and determine which DNA repair pathway should be selected in face of a particular genotoxic insult or type of DNA lesions (Fig. 2). This hypothesis is also consistent with the observation that both chromatin binding and the foci formation of Rad51 and monoubiquitination of FANCD2 are intact in ASK/Dbf4-depleted cells upon cisplatin treatment,³⁶ suggesting that the interaction between ASK/Dbf4 and RAD18 specifically regulates the RAD18-dependent TLS, but not HR or FA pathways.

Figure 2. A model for RAD18-dependent DNA repair pathway choice in response to genotoxic insults. RAD18 may function as a molecular switch for DDR pathway choice through alternating partner factors that bind to its RING domain.

In order to sustain DNA replication across the DNA lesion, replisome (the replication fork complex) must be stabilized; otherwise the stalled fork would collapse, and, subsequently, a DNA double-strand break would be generated, which may undermine genomic integrity. In yeast, DDK plays a crucial role in fork stabilization upon genotoxic insults.^55,56 DDK-mutant yeast cells feature spontaneous Rad52 foci, suggesting the ongoing rescue of collapsed forks by HR.⁵⁶ An analogous role was also proposed for mammalian DDK,^23,57 hence this function of DDK in replisome stabilization to continue DNA replication across lesions appears to be conserved in all eukaryotes. Given that excessive HR can also compromise genome integrity, we speculate that by regulating pathway choice, DDK might also help to actively prevent untimely HR, a process that could be accomplished by modulating helicase activities of PARI, FBH1, or RTEL1, for example.^58-60

Suppression of Late Origin Firing in Response to Replication Stress

At this moment, it is still unclear whether the entire pool of DDK stays active after exposure to replication stress. Although the bulk activity of DDK may be unaltered after the replication stress, it is possible that checkpoint kinase-mediated phosphorylation may inactivate or promote dissociation of at least a certain subset of Cdc7–ASK/Dbf4 complexes or those associated with the early-firing pre-RC, and thus render it unavailable for activation of late dormant origins.

If the human DDK remains active under replication stress conditions, how can the firing of late replication origins be suppressed without inhibiting the DDK function? Recently, a potential relevant role of a phosphatase was reported in Xenopus. In the Xenopus cell-free system, protein phosphatase 1 (PP1)-mediated dephosphorylation of MCM proteins, which is triggered by the checkpoint kinases, counteracts the DDK-dependent origin firing without modulating the DDK kinase activity.⁶¹ Thus, it might be an intriguing possibility that the checkpoint kinases activate PP1, which then counteracts the DDK-dependent phosphorylation of the MCM complex to prevent late origin firing. However, there is currently no evidence either in Xenopus egg extracts or in any other systems that depletion or inactivation of phosphatases override the checkpoint-mediated inhibition of late origin firing. Nor is there any evidence that PP1 is indeed recruited to late origins upon replication stress. These issues need to be examined in the future to test this hypothesis.

Recently, Rif1, a telomere-binding factor in yeasts, was found to be a major regulator of replication timing in both yeasts and mammals.^62-64 In the absence of Rif1, replication timing changes dramatically both in yeast and mammalian cells during normal cell cycle progression. In both species, timing and efficiency of pre-RC assembly is unaffected, but that of Cdc45 changes in the absence of Rif1.^62,63 In fission yeast, Rif1 binds close to the late/dormant origins and suppresses the firing of nearby origins⁶² (Kanoh et al. unpublished data). It is unknown whether mammalian Rif1 shows a similar binding preference. Nevertheless, the similarity is striking, in that the absence of Rif1 causes general loss of replication timing regulation throughout the genome.⁶⁵ It is unlikely that Rif1 is involved in cellular responses to replication stress, since rif1-null yeast cells and Rif1-depleted human cells are resistant to HU.^62,65 However, it is of interest that PP1 interacts with Rif1, and that this interaction may play a role in Rif1-mediated regulation of replication timing in budding yeast.⁶⁶

DDK and Cancer

Last but not least, we wish to discuss DDK in the context of tumorigenesis, and also as a candidate target for innovative cancer therapy. The major reason why the DDK-mediated response to replication stress is highly relevant for cancer is that activation of diverse oncogenes and loss of some tumor suppressors evoke replication stress and consequent DNA damage that triggers the checkpoint responses of the ATR–Chk1 and ATM–Chk2 signaling cascades, as exemplified in cell culture experiments and documented by analyses of clinical specimens from a range of human malignancies.^67-75 This anti-cancer barrier provided by the DNA damage checkpoints precedes the activation of the ARF tumor suppressor, the second anti-cancer barrier,^76-80 and helps delay or prevent tumor progression by triggering senescence or cell death among the nascent cancer cells.^67,68,70,71 However, in those tumors that eventually do progress, the ATM-Chk2-p53 and other checkpoints are commonly bypassed or inactivated,^67,72,73 while the ATR–Chk1 axis, which generally helps cells to cope with replication stress (as discussed above), remains operational, as cancer cells need to deal with the omnipresent challenge of enhanced replication stress.⁸¹ In such scenario, and given the above discussed roles of DDK under replication stress, the DDK activity (including the ASK/Dbf4-RAD18 role in TLS) may provide one of the stress-support functions on which cancer cells depend more than their normal counterpart cells.

In addition, we suggest that the occurrence of much greater endogenous replication stress in cancer cells would provide opportunities for therapeutic intervention by DDK inhibitors. Indeed, several such scenarios could be proposed, as would be rationalized on the basis of the synthetic lethality principle, best documented by the selective sensitivity of HR- (e.g., BRCA1/2-) defective tumors to PARP inhibitors.^82-85 Since the HR pathway is impaired in BRCA-deficient tumors, small-molecule inhibitors of Cdc7 kinase activity (and/or inhibitors of RAD18 ubiquitin ligase, yet to be developed) could be useful as an alternative of PARP inhibitors. Indeed, it has been shown that the HR pathway is required for the viability of rad18∆ yeast cells upon chronic low-dose UV light exposure.^86,87 It was also reported that double mutants for RAD18 and RAD54, a gene required for HR, exhibit synthetic lethality in chicken DT40 cells, in contrast to only modest impact of either mutation alone.⁸⁸ Moreover, human cells deficient in the TLS pathway, including RAD18 deficiency, are hypersensitive to cisplatin treatment,⁸⁹ whose tolerance is also dependent on the HR pathway. Since the FA pathway is proposed to be involved in HR,^90-95 the Cdc7 kinase inhibitors could be also effective for tumors in which the FA pathway is impaired.

In conclusion, targeting the DDK activity under conditions of enhanced replication stress and the nodal point of RAD18 interactions in DNA repair pathway choice may inspire novel treatment strategies in oncology. Such DDK/RAD18-antagonizing drugs could be used as single treatment in subsets of patients selected based on suitable biomarker information on the DNA repair pathway status and extent of replication stress.^{67,70,81,96,97} Alternatively, they could be combined with standard-of-care genotoxic chemotherapeutics or ionizing radiation to obtain higher sensitization of cancer cells.

Despite the undisputable progress in recent years, a great deal remains to be done to mechanistically understand the cellular responses to replication stress in general, and the role(s) of the DDK in checkpoint responses and DNA repair in particular. Given the numerous emerging fertile avenues of this research, it is clear that this fascinating field will move even more rapidly, and into the forefront of basic and likely also translational biomedical research.

Glossary

Abbreviations:

APC/C

anaphase promoting complex/cyclosome

ATR

ataxia telangiectasia and Rad3-related protein

CDK

cyclin-dependent kinase

DDK

Dbf4-dependent kinase

DDR

DNA damage response

FA

Fanconi anemia

HR

homologous recombination

PIP box

PCNA-interacting protein box

RING domain

really interesting new gene domain

TLS

translesion synthesis