Cdc7 Is an Active Kinase in Human Cancer Cells Undergoing Replication Stress*

Cdc7 kinase promotes and regulates DNA replication in eukaryotic organisms. Multiple mechanisms modulating kinase activity in response to DNA replication stress have been reported, supporting the opposing notions that Cdc7 either plays an active role under these conditions or, conversely, is a final target inactivated by a checkpoint response. We have developed new immnunological reagents to study the properties of human Cdc7 kinase in cells challenged with the ribonucleotide reductase inhibitor hydroxyurea or with the DNA topoisomerase II inhibitor etoposide. We show that Cdc7·Dbf4 and Cdc7·Drf1 complexes are stable and active in multiple cell lines upon drug treatment, with Cdc7·Dbf4 accumulating on chromatin-enriched fractions. Cdc7 depletion by small interfering RNA in hydroxyurea and etoposide impairs hyper-phosphorylation of Mcm2 at specific Cdc7-dependent phosphorylation sites and drug-induced hyper-phosphorylation of chromatin-bound Mcm4. Furthermore, sustained inhibition of Cdc7 in the presence of these drugs increases cell death supporting the notion that the Cdc7 kinase plays a role in maintaining cell viability during replication stress.

Cdc7 kinase promotes and regulates DNA replication in eukaryotic organisms. Multiple mechanisms modulating kinase activity in response to DNA replication stress have been reported, supporting the opposing notions that Cdc7 either plays an active role under these conditions or, conversely, is a final target inactivated by a checkpoint response. We have developed new immnunological reagents to study the properties of human Cdc7 kinase in cells challenged with the ribonucleotide reductase inhibitor hydroxyurea or with the DNA topoisomerase II inhibitor etoposide. We show that Cdc7⅐Dbf4 and Cdc7⅐Drf1 complexes are stable and active in multiple cell lines upon drug treatment, with Cdc7⅐Dbf4 accumulating on chromatin-enriched fractions. Cdc7 depletion by small interfering RNA in hydroxyurea and etoposide impairs hyper-phosphorylation of Mcm2 at specific Cdc7-dependent phosphorylation sites and drug-induced hyper-phosphorylation of chromatin-bound Mcm4. Furthermore, sustained inhibition of Cdc7 in the presence of these drugs increases cell death supporting the notion that the Cdc7 kinase plays a role in maintaining cell viability during replication stress.
Cdc7 is an evolutionary conserved kinase that promotes DNA replication origin firing by phosphorylating one or more subunits of the MCM complex, the putative replicative DNA helicase (1)(2)(3)(4)(5). Cdc7, like cyclin-dependent kinases, is activated by the binding of regulatory subunit Dbf4 (6). Whereas only one regulatory subunit exists in lower eukaryotes, a second related protein Drf1/ASK1L/Dbf4B was recently identified in human and Xenopus (7)(8)(9). Both Dbf4 and Drf1 appear to be important for the replication of the human genome (8), whereas it is the Cdc7⅐Drf1 complex that plays the major role in Xenopus egg DNA replication (10,11). Drugs that disrupt the synthesis of nascent DNA strands, either by blocking dNTP supply, such as hydroxyurea (HU), 3 or that delay fork progression, such as topoisomerase inhibitors, activate a cellular response pathway. This pathway, known as the S-phase checkpoint, coordinates the activation of DNA replication origins and the stabilization of stalled replication forks, with the action of DNA repair proteins and the regulation of the cell cycle (12)(13)(14)(15)(16). This pathway involves several protein kinases including ATM, ATR, Chk1, and Chk2, and it is important for the maintenance of genome stability and for determining cell fate upon insults (12,17,18).
There is much evidence to indicate that Cdc7 kinase function and activity may be modulated by the S-phase checkpoint pathway (19). In budding yeast, it was reported that the Cdc7 regulatory subunit, Dbf4, is phosphorylated by the Chk2 homologue Rad53 kinase, it dissociates from chromatin and kinase activity is down-regulated (20 -22). However, in an earlier study it was reported that high levels of Cdc7⅐Dbf4 kinase were present in S-phase cells arrested with HU (23). Upon release from HU block, Cdc7 is required for the firing of late replication origins, although, in the absence of Cdc7 activity, duplication of the genome is most likely completed from the resumption of stalled replication forks (24). In the presence of DNA damage caused by UV radiation or by the alkylating drug methyl methanesulfonate, Cdc7 was shown to be important for induced mutagenesis (5,25). Finally, deletion analysis has defined an N-terminal domain within the Dbf4 protein, which is dispensable for DNA replication, but is required for maintaining cell viability in HU, possibly by targeting the kinase to stalled forks (26,27). Similarly, a growing body of genetic evidence indicates that HSK1, the fission yeast Cdc7 homologue is important for HU and alkylating agent-induced checkpoint responses (28 -31).
In higher eukaryotes, the question is further complicated by the presence of at least two Cdc7 complexes. In the Xenopus cell-free DNA replication system, the Cdc7⅐Dbf4 complex is inhibited by dissociation of the Dbf4 regulatory subunit after treatment with etoposide (32). However, the Cdc7⅐Drf1 complex is not affected by aphidicolin, an inhibitor DNA polymerases, and the overall Cdc7 kinase activity does not vary in the presence of DNA with double strand breaks (9,33). Furthermore, in aphidicolin the Drf1 subunit accumulates on chromatin in an ATR-and claspin-dependent manner (9). Similarly to Xenopus Cdc7⅐Dbf4, a recent study also reported that human Dbf4 dissociates from Cdc7 after treatment of Bcr-Abl negative, ATR-proficient leukemia cells with etoposide (34).
We have recently mapped specific Cdc7-dependent phosphorylation sites in the Mcm2 protein (35). In this previous analysis we noticed that these sites are normally phosphorylated in a cell cycle-dependent manner from the beginning of S-phase until mitosis. Moreover, arresting the cells in S-phase with either HU or etoposide resulted in hyper-phosphorylated Mcm2 at Cdc7-dependent sites, suggesting that human Cdc7 kinase is active under these conditions. Here, to explore a potential role of human Cdc7 kinase during replication stress, we have studied the biochemical properties of Cdc7 complexes in cells challenged with either HU or etoposide and characterized the phenotypes caused by Cdc7 depletion when cells are undergoing replication stress caused by these agents.
Cell Cycle and Apoptosis Analysis-For analysis of DNA content cells were prepared and analyzed with a FACScan (BD Biosciences) as previously described (36). TUNEL assay was performed using the APO-BRDU kit from Sigma according the manufacturer's instructions and analyzed with FACScan.
Protein Manipulations-Purification and characterization of recombinant Cdc7⅐Dbf4 kinase and Mcm2 N-terminal fragment (amino acids 10 -294) were previously described (7). Soluble cytoplasmic/nucleoplasmic proteins were prepared in CSK buffer (100 mM NaCl, 10 mM PIPES, pH 6.8, 300 mM sucrose, 1 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, 0.5% Triton X-100) with protease and phosphatase inhibitors (Sigma). Protein extracts were then centrifuged for 5 min at low speed (600 ϫ g). The supernatant was collected and labeled as soluble fraction. The pellet was then resuspended in the same buffer with 300 mM NaCl for 15 min on ice and then centrifuged at high speed thus obtaining high salt and insoluble fractions. Insoluble proteins were then further extracted in 5% SDS. For micrococcal nuclease digestion, after the first extraction in 100 mM NaCl CSK buffer, the pellet was washed once, resuspended in the same buffer with 1 mM CaCl 2 , and incubated for 5 min at 37°C with 100 units of micrococcal nuclease (U. S. Biochemical Corp.).
For co-immunoprecipitation experiments, whole cell lysates, prepared in 300 mM NaCl containing CSK buffer, were incubated with 4 g of anti-Cdc7, anti-Dbf4, anti-Drf1, or normal mouse IgG (VECTOR) antibodies and with 30 l of protein G-Sepharose beads (Amersham Biosciences) for 3 h at 4°C with end-over-end mixing. Beads were washed with NET buffer (100 mM NaCl, 1 mM EDTA, 50 mM Tris-HCl, pH 7.5, 0.1% Triton X-100) and proteins were recovered in 1ϫ Laemmli buffer. Kinase assays were performed as described (7). 250 g of protein extract was used for assessing Cdc7 activity, whereas 1 mg was used for Dbf4 and Drf1.

RESULTS
Characterization of Anti-Cdc7 12A10 mAb-The biochemical and physiological characterization of human Cdc7 kinase has been limited to date by the lack of appropriate reagents. We have recently generated an anti-Cdc7 mAb (mAb 12A10) that immunoprecipitates a cell cycle-regulated kinase activity that correlates with the expression of the Cdc7 polypeptide, strongly suggesting that mAb 12A10 specifically immunoprecipitates an active Cdc7 kinase complex (35).
We have further characterized the specificity of our mAb 12A10 in parallel with a commercially available anti-Cdc7 mAb (DCS-341) and polyclonal rabbit and goat antibodies. We found that when these antibodies were used at the same concentration of 0.2 g/ml in Western blot, both mAbs were able to recognize up to 1 ng of recombinant Cdc7 kinase (Fig. 1A), goat N19 only very weakly detected 25 ng of protein, whereas goat polyclonal C-20 or rabbit polyclonal H110 antibodies failed to recognize the same amount of protein (supplementary Fig. 1A). Specificity of 12A10 and DCS-341 mAbs was confirmed by probing membranes in which whole cell extracts prepared from HeLa cells, growing normally, treated with HU, or depleted of Cdc7 by siRNA, were blotted. Both mAbs recognized a single polypeptide of ϳ60 kDa consistent with Cdc7 molecular mass. In the HU-treated sample an approximate 2-fold increase in signal was observed, consistent with the previous finding that Cdc7 polypeptide accumulates in S-phase cells (35) and that HU causes cells to accumulate in S-phase of the cell cycle ( Fig.  1B, lanes 2 and 5). The immunoreactive band was instead strongly reduced in cells transfected with Cdc7-specific siRNA (Fig. 1B, lanes 3 and 6), demonstrating that both mAbs indeed specifically recognize Cdc7 polypeptide. In contrast C20, N19, and H110 antibodies recognized multiple bands whose levels were not modulated by either HU or Cdc7 siRNA transfection and are most likely nonspecific (supplementary Fig. 1B). We then tested these antibodies for their ability to immunoprecipitate Cdc7 using an immunoprecipitation/Western blot assay. We found that Cdc7 was highly enriched in the immunoprecipitate obtained with both 12A10 and DC-341 mAbs, however, when the same samples were probed with anti-Dbf4 antibodies, Dbf4 was only present in the 12A10 immunoprecipitate and not in the DCS-341 immunoprecipitate (Fig. 1C). It is possible that under non-denaturing conditions DCS-341 mAb either causes dissociation of Dbf4 from Cdc7 or it recognizes an epitope masked by the presence of Dbf4. Instead, 12A10 mAb specifically recognizes Cdc7 protein and can co-immunoprecipitate Dbf4 and a Cdc7-associated kinase activity (35). In a similar approach we found that anti-Dbf4 8F4 mAb and polyclonal H300 antibodies used in this study recognize Dbf4 in Western blot and are capable of immunoprecipitating a Cdc7⅐Dbf4 complex (supplementary Fig. 2). Characteristics of anti-Drf1 mAb were previously described (7). We conclude that multiple reagents are now available for studying Cdc7 function in different conditions.
Human Cdc7⅐Dbf4 and Cdc7⅐Drf1 Complexes Remain Stably Associated upon Replication Stress-We next used anti-Cdc7 12A10 mAb, anti-Drf1 mAb, and anti-Dbf4 mAb to immunoprecipitate Cdc7 complexes from protein extracts prepared from HeLa cells, which were either mock treated or treated with HU or etoposide. We found that both Dbf4 and Drf1 co-immunoprecipitate with Cdc7 using the anti-Cdc7 mAb, and that Cdc7 polypeptide was present in the immunoprecipitates obtained with both anti-Dbf4 and anti-Drf1 antibodies independently of the treatment of the cells (Fig. 2A). We also observed that the amount of Cdc7⅐Dbf4 and Cdc7⅐Drf1 complexes recovered increased ϳ2-fold after drug treatments, correlating with accumulation of these proteins in S-phase (7,35,37) and indicating that both complexes were stable under these conditions. Finally, we observed that Drf1 was not present in the Dbf4 immunoprecipitation and vice versa indicating that Cdc7, Dbf4, and Drf1 do not form ternary complexes upon stress and this is consistent with our previous finding using overexpressed proteins (7).
These results indicate that in HeLa cells Cdc7 complexes do not dissociate in HU or etoposide. These findings are different from a previous report by Dierov et al. (34) suggesting dissociation of Cdc7 from Dbf4 in etoposide-treated hemapoietic cells. To address these discrepancies, we tested Cdc7 association with Dbf4 in an enlarged panel of human cell lines including two hemapoietic derived cell lines either expressing or not expressing the Bcr-Abl oncogene (Fig. 2C). Using the anti-Cdc7 12A10 mAb, we again observed that an intact Cdc7⅐Dbf4 complex was recovered from extracts prepared from both cell lines either untreated or also after etoposide treatment (Fig. 2B). Similar results were obtained using HCT-116 colon cancer cells (Fig. 4A, lower panel) and using the L-363 plasma cell leukemiaderived cell line (supplementary Fig. 3) suggesting that lack of Cdc7⅐Dbf4 dissociation may be a general feature in etoposidetreated human cells. Although we do not have access to the cells derived from the same AML patients that were used by Dierov et al. (34), we observed that they used a commercial anti-Cdc7 antibody that in our laboratory failed to efficiently immunoprecipitate a native Cdc7⅐Dbf4 complex (supplementary Fig. 4), suggesting that the quality of antibodies might have affected the previous Cdc7⅐Dbf4 stability study. Therefore, we conclude that in multiple human cell lines both Cdc7⅐Dbf4 and Cdc7⅐Drf1 complexes are stable upon treatment with genotoxic agents.
Cdc7 Accumulates on Chromatin in HU and Etoposide-Previous work has shown that human Cdc7 and Dbf4 polypeptides can bind to chromatin in a cell cycle-dependent manner (38).  Importantly, both in budding yeast and Xenopus, changes in the association of Cdc7 subunits with chromatin were suggested to modulate the kinase functions in response to genotoxic treatments occurring in S-phase (9,21,22). To investigate this regulation in human cells, we began to study Cdc7 fractionation properties. In untreated cells ϳ60% of the Cdc7 is extracted in a low salt CSK buffer, whereas almost all remaining Cdc7 is subsequently extracted by increasing the ionic strength to 300 mM NaCl. Upon drug treatment, the overall level of Cdc7 increases, but, whereas an increase of 1.5-fold was observed in the low salt extraction, a 3-fold increase was measured in the fraction extracted with high salt buffer, indicating that both the levels and the distribution of Cdc7 are affected by the drugs (Fig. 3A). Under the same conditions Mcm2 was mostly found in the low salt fraction and drug treatment did not obviously change its distribution. Consistent with previous observations (39), phosphorylated RPA was preferentially retained on chromatin and was largely eluted by increasing ionic strength (Fig. 3A).
To examine Cdc7 association with chromatin, the pellet obtained from a low salt extraction was treated with micrococcal nuclease. DNA digestion removed a considerable portion of Cdc7 from the insoluble fraction, indicating that at least 50% of low saltresistant Cdc7 is indeed chromatin bound (Fig. 3B). Immunoprecipitation of Cdc7 complexes from sequentially extracted samples indicated that Cdc7⅐Dbf4 complex is found mostly in the high salt fraction where it accumulates upon genotoxic treatments. Instead, although Cdc7⅐Drf1 levels rise upon treatment, the complex is equally distributed in the two fractions (Fig. 3C).
In addition, single cell analysis by immunofluorescence, indicated that in HU and etoposide, Cdc7 was in the nucleus of every single cell, and that non-ionic detergent washes, a treatment that can discriminate between nuclear soluble and chromatin-engaged proteins (40 -42), did not release Cdc7 from the nuclei (Fig. 3D). Thus, in contrast to budding yeast Cdc7⅐Dbf4 and Xenopus Dbf4 (20, 32), in human cells Cdc7⅐Dbf4 and Cdc7⅐Drf1 do not dissociate from chromatin-enriched fractions after HU and etoposide treatment.
Cdc7 Is an Active Kinase in Perturbed S-phase-With the aim of assessing if Cdc7 activity is affected by drug treatment, we measured the levels of kinase activity present in the Cdc7, Dbf4, and Drf1 immunoprecipitates by adding radiolabeled ATP and a recombinant Mcm2 N-terminal fragment as a substrate. In all cases, the kinase activity recovered was increased in response to HU and etoposide treatments both in HCT-116 (Fig. 4A, top panel) and HeLa (data not shown) cells, and appears to correlate with the amount of Cdc7 subunit present in the immunoprecipitate (Fig. 4A, bottom panel), suggesting that no obvious change in the specific activity of the complexes took place upon drug treatment. In these experiments, we normalized the activity versus the amount of Cdc7 co-immunoprecipitated with the three different mAbs and we observed that Cdc7⅐Dbf4 was A, sequential extraction of Cdc7, Mcm2, and Rpa2 proteins from drug or mock-treated cells in CSK buffer containing 100 mM (low salt), 300 mM NaCl (high salt), and 5% SDS containing buffer. Relative amounts of Cdc7 in the different fractions were determined by densitometric analysis of the film (see text). B, Cdc7 is partially extracted after micrococcal nuclease digestion. Cells were mock or HU-treated and lysed in low salt buffer. Proteins were separated into soluble (S) and insoluble (P) fractions. Pellets were then subjected to nuclease digestion (p ϩ M) before further separation into soluble (S1) and pellet (P1) fractions. C, immunoprecipitation of Cdc7 complexes from sequentially extracted protein samples. Cells were drug or mock-treated, extracts prepared by sequential extraction in low (L) and high (H) salt buffer as in B, and immunoprecipitations (IP) were performed with the indicated antibodies. Cdc7 protein in the immunoprecipitates was detected by Western blot. D, nuclear detection of Cdc7 in HeLa cells either not treated (Ϫ) or treated with HU or etoposide (Eto). Cells were stained with 4Ј,6-diamidino-2-phenylindole (DAPI) to visualize the nuclei and by indirect immunofluorescence using anti-Cdc7 DCS-341 mAb. In the lower panels cells were extracted with Triton-X-100 before fixation to remove soluble nuclear proteins. more efficient at phosphorylating Mcm2 than Cdc7⅐Drf1 (Fig.  4, A and B). This may reflect intrinsic properties of these complexes or may simply be due to steric hindrance of the anti-Drf1 antibody.
To understand if Cdc7 was active in vivo we coupled drug treatment with Cdc7 depletion by siRNA. We have previously shown that Cdc7 depletion by siRNA in HeLa does not immediately affect incorporation of nucleotides into the DNA, at least for the first 24 h, but upon prolonged inhibition (ϳ48 -60 h post-transfection), cells slow down DNA synthesis and then enter apoptosis (36). We therefore exploited this temporal window, in which Cdc7 siRNA does not impair viability and the bulk of DNA synthesis, to assess the effects of its down-regulation in the presence of inhibitors of DNA replication elongation. HeLa cells were transfected with either luciferase-or Cdc7-specific siRNA and treated with HU. Under these experimental conditions we first observed by FACS analysis that both Cdc7 and control siRNA-transfected cells were blocked by HU (Fig. 5B) and that siRNA transfections did not alter the kinetics of S-phase progression in a parallel sample not treated with HU throughout the experiment (supplementary Fig. 5). We observed that Cdc7 protein levels that normally accumulate in HU were reduced by Cdc7-specific siRNA (Fig. 5A). Importantly, we found that Mcm2 phosphorylation, which causes the protein to migrate faster in SDS-PAGE (3,35,36), was strongly affected in HU-treated Cdc7-depleted cells (Fig. 5A). Identical findings were observed when cells were challenged with etoposide instead of HU (Fig. 5, C and D (35,43,44). We have previously shown that in growing cells, phosphorylation at Ser 53 and Ser 40 (that can be detected with anti-pSer53, anti-pSer40/41, and anti-pSer41 antibodies) is fully dependent upon Cdc7 activity (35). To gain further insight into the functions of Cdc7⅐Dbf4 and Cdc7⅐Drf1 complexes, we set up a siRNA procedure that allowed us to efficiently deplete Dbf4, Drf1, or both from HeLa cells. Fig. 6A shows that upon Dbf4 siRNA transfection, the amount of Cdc7⅐Dbf4 kinase that can be immunoprecipitated with anti-Dbf4 antibody is drastically decreased. Similarly, upon Drf1 siRNA transfection, Cdc7⅐Drf1 levels are strongly decreased. Finally, after transfection with a mixture of Dbf4 and Drf1 siRNAs, the Cdc7 catalytic subunit cannot be recovered using either anti-Dbf4 or anti-Drf1 antibodies.
Next, HeLa cells were transfected with luciferase, Cdc7, Dbf4, and/or Drf1 siRNA and then treated with HU or etoposide. After 48 h cells were collected and Mcm2 phosphorylation at specific phospho sites was analyzed with specific anti-phospho-Mcm2 antibodies. We observed that the levels of phospho-Ser 40 and phospho-Ser 53 Mcm2 were decreased by Cdc7 depletion both in HU and etoposide. Phosphorylation levels at Ser 53 were decreased by Dbf4 but not by Drf1 depletion, whereas phosphorylation at Ser 40 was only partially decreased by Dbf4, unaffected by Drf1, and strongly down-regulated by simultaneous depletion of both regulatory subunits (Fig. 6B). Phosphorylation at Ser 108 , a site that is also phosphorylated by ATR (45,46), was induced by both HU and etoposide and not obviously affected by Cdc7 depletion under these conditions. Finally, none of these depletions affected overall levels of the Mcm2 protein or its phosphorylation at Ser 41 , which is in agreement with our previous findings (35), although this is not consistent with another study (44), and indicates that Mcm2 phosphorylation at this site is not only Cdc7 independent but also Dbf4 and Drf1 independent (Fig. 6B).
These results indicate that either in the presence or absence of genotoxic drugs, both Cdc7⅐Dbf4 and Cdc7⅐Drf1 can phosphorylate Mcm2 at Ser 40 but only Cdc7⅐Dbf4 is responsible for Ser 53 phosphorylation. Because of this complexity we decided to focus on analyzing the effects of overall Cdc7 activity by manipulating the expression of the Cdc7 catalytic subunit.
Cdc7 Is Required for Optimal Drug-induced Mcm4 Phosphorylation-Drug-induced replication stress specifically causes hyper-phosphorylation of a chromatin-bound fraction of the Mcm4 subunit of the MCM complex. Mcm4 hyper-phosphorylation occurs at multiple sites, it possibly affects helicase activity and, unlike hyper-phosphorylation of the Mcm2 subunit, it is detectable as reduced mobility in SDS-PAGE (3,47). We observed that upon Cdc7 depletion, the drug-induced mobility shift of chromatin-bound Mcm4 was abolished both in HU and etoposide (Fig. 7). These results are consistent with previous findings that active Cdc7 kinase accumulates on chromatin in HU, that human Cdc7 can phosphorylate Mcm4 in vitro (3), and that Xenopus Cdc7 also phosphorylates chromatin bound Mcm4 (33). Lack of a drug-induced phosphorylation of Mcm4 was not caused by a general impairment of the response to DNA replication stress as the kinetics of histone H2AX phosphorylation at Ser 139 were not affected by Cdc7 depletion (Fig.   5, A and B). Altogether these data indicate that upon druginduced replication stress Cdc7 activity is required for the full phosphorylation of at least two subunits of the replicative MCM helicase.
Increased Cell Death in HU and Etoposide upon Sustained Inhibition of Cdc7 Kinase-To understand whether the maintenance of high levels of Cdc7 kinase activity in cells treated with genotoxic agents is physiologically relevant, we analyzed cell survival after prolonged exposure to the drugs in the presence or absence of Cdc7. HeLa cells and A2780 ovary carcinoma cells were transfected with luciferase or Cdc7-specific siRNA and then either mock treated or incubated with HU or etoposide. Cell samples were collected 48 h later and analyzed for apoptosis using a TUNEL assay. In HeLa cells, at this early time post-transfection and consistent with previous results (36), cells positive in the TUNEL assay were a small fraction of the total, however, we observed that lack of Cdc7 exacerbated the effects caused by the two drugs (p value of 0.012 in HU and 0.02 in etoposide) (Fig. 8A). Additivity in causing cell death, by the combination of Cdc7 inhibition with HU or etoposide treatment, was more evident in the A2780 cell line, where the combination caused a 3-5-fold increase in the percentage of TUNEL-positive cells in HU and etoposide, respectively (p value 0.017 and 0.011) (Fig. 8B).
We then repeated Cdc7 down-regulation in the presence or absence of HU or etoposide and pan-caspase inhibitor Z-VAD. Cells were collected after 48 h and analyzed by FACS. Accumulation of cells with sub-G 1 DNA content was almost completely abolished by addition of Z-VAD (Fig. 8C), further confirming that cell death was caused by the apoptotic pathway in the A2780 cell line. We conclude that in these cellular models Cdc7 function is required to survive drug-induced replication stress.

DISCUSSION
Active Cdc7 Kinase Complexes during Perturbed S-phase-In this work we have developed immunological reagents that allowed  us to characterize, in parallel, the activity of the two known human Cdc7 kinase complexes, Cdc7⅐Dbf4 and Cdc7⅐Drf1, in cells undergoing replication stress caused by dNTPs supply inhibition (HU), or by topological constraints caused by topoisomerase inhibition (etoposide). Our data indicate that neither Cdc7⅐Dbf4 nor Cdc7⅐Drf1 human complexes are down-regulated by genotoxic agents. Under these stress conditions, the complexes do not dissociate and the amount of kinase activity that can be measured after immunoprecipitation is increased ϳ2-fold, correlating with the amount of the Cdc7 complexes present in the extracts and the amount of protein immunoprecipitated. Furthermore, biochemical fractionation experiments indicate that Cdc7 complexes are not released from chromatin in drug-treated cells, thus suggesting that they are correctly positioned to regulate proteins required for replication of the genome.
siRNA-mediated depletion of Cdc7 subunits further supports the biochemical evidence that both Cdc7 complexes are active in cells exposed to genotoxic drugs. Indeed, we found that both Cdc7⅐Dbf4 and Cdc7⅐Drf1 complexes contribute, although not equally, to maintain high levels of Mcm2 phosphorylated at Cdc7-dependent phosphorylation sites Ser 40 and Ser 53 , and that drug-induced phosphorylation of chromatin-bound Mcm4 is defective upon Cdc7 depletion.
Role of Cdc7 during Stress-Our results provide compelling evidence that in human cancer cell lines Cdc7 activity is required for protecting cells from apoptosis during replication stress caused by HU and etoposide. This observation is reminiscent of the fact that specific Cdc7 and Dbf4 mutations that alter sensitivity to HU and other genotoxic agents can be isolated in both budding (5,25,27) and fission yeast (28 -31).
At the molecular level the reasons for the maintenance of high levels of Cdc7 kinase activity in a perturbed S-phase are not clear. It is unlikely that Cdc7 is required for activating FIGURE 8. Sustained inhibition of Cdc7 in the presence of genotoxic agents increases cell death. A, HeLa cells were transfected with either luciferase-(white bars) or Cdc7 (black bars)-specific siRNA. Four hours after transfection, cells were either not treated (NT) or treated with HU or etoposide for the 48 h prior to collection. The percentage of TUNEL positive cells and S.D. from three different experiments are indicated. Statistical significance between luciferase and Cdc7 siRNA-transfected samples in HU and etoposide (ETO) was determined using the paired, one-tail distribution Student's t test (p Ͼ 0.05). B, as in A but A2780 ovary carcinoma cells were used (p Ͼ 0.05). C, apoptotic cell death caused by a combination of Cdc7 inhibition and genotoxic drugs can be prevented by a pan-caspase inhibitor. A2780 cells were transfected with either luciferase-(Luc) or Cdc7 (Cdc7)-specific siRNA. Four hours after transfection, cells were either not treated or treated with HU or etoposide for 48 h in the absence or presence of caspase inhibitor Z-VAD. DNA content of cell populations was analyzed by FACS after staining with propidium iodide. Samples were acquired on logarithmic scale to better appreciate the amount of cells with sub-G 1 DNA content. new replication origins as these do not normally fire in the presence of replication stress and DNA damage (14, 48,49). It is also unlikely that stalled replication forks collapse in the absence of Cdc7 because, at least in budding yeast, the bulk of DNA synthesis resumes from previously arrested replication forks (24). It is, however, possible that Cdc7 may regulate some subtle events that occur when replication forks are arrested, such as fine-tuning of MCM helicase activity or some protein-protein interactions at the replisome. In this respect it was recently reported that the Cdc7⅐Dbf4 kinase also phosphorylates the p150 subunit of the chromatin assembly factor 1, thus promoting its binding to proliferating cell nuclear antigen (50). It is likely that an active Cdc7 kinase may also be required to control chromatin assembly factor 1 activity and histone deposition at delayed replication forks in etoposide and HU. One intriguing hypothesis is that a Cdc7-dependent event, that may or may not involve MCM phosphorylation, may be required for the recruitment of factors that deal with DNA damage in the proximity of a replication fork such as translesion DNA polymerases, thus explaining the hyper-mutagenic phenotype of certain budding yeast Cdc7 mutants (25,51). Finally, at this stage we cannot formally exclude that Cdc7 may also be involved in an as yet unidentified survival pathway that prevents the death of S-phase cells and that is independent from stalled replication forks. In this case the increased cell death observed upon drug treatment could be simply related to the fact that these cells spend an increased time in this phase of the cell cycle.
Importantly, in human cancer cell lines, down-regulation of Cdc7 in the presence of genotoxic drugs increases the number of cells that enter the apoptotic pathway. This suggests that a combination of drugs that impair the elongation reaction of DNA synthesis, together with specific Cdc7 inhibitors, may prove a viable strategy to tackle tumor cells. In the future it will be interesting to verify if this additive cell killing is generally observed by combining inhibitors of elongation with inhibition of other factors involved in DNA replication initiation.