Rad9 Protects Cells from Topoisomerase Poison-induced Cell Death*

Previous studies have suggested two possible roles for Rad9 in mammalian cells subjected to replication stress or DNA damage. One model suggests that a Rad9-con-taining clamp is loaded onto damaged DNA, where it participates in Chk1 activation and subsequent events that contribute to cell survival. The other model suggests that Rad9 translocates to mitochondria, where it triggers apoptosis by binding to and inhibiting Bcl-2 and Bcl-x L . To further study the role of Rad9, parental and Rad9 (cid:1) / (cid:1) murine embryonic stem (ES) cells were treated with camptothecin, etoposide, or cytarabine, all prototypic examples of three classes of widely used anticancer agents. All three agents induced Rad9 chromatin binding. Each of these agents also triggered S-phase checkpoint activation in parental ES cells, as indicated by a caffeine-inhibitable decrease in [ 3 H]thymidine incorporation into DNA and Cdc25A down-regulation. In-terestingly, the ability of cytarabine to activate the S-phase checkpoint was severely compromised in Rad9 (cid:1) / emission filter. Alternatively, cleavage of caspase substrates was assessed by immunoblotting. cell lysates prepared in M guanidine hydro- chloride under reducing conditions as described previously (43). Aliquots containing 50 (cid:1) g of total cellular protein subjected to SDS- PAGE on gels 5–15% acrylamide gradients, to nitrocellulose, previously

Recent studies have suggested two possible roles for Rad9 after DNA damage. The first involves participation of Rad9 in a heterotrimeric clamp that is assembled on chromatin following replication stress or other types of DNA damage. When replication forks stall, the single-stranded DNA-binding protein replication protein A binds areas of single-stranded DNA (1)(2)(3)(4). Bound protein replication protein A facilitates the bind-ing of the kinase ATR 1 and its binding partner ATR-interacting protein to chromatin (4). At the same time, a preassembled complex of Rad9, Hus1, and Rad1 (5,6), known as the 9-1-1 complex, is loaded onto the damaged chromatin by a clamp loader consisting of Rad17 and the four replication factor C small subunits (7,8). In a manner that remains poorly understood, the chromatin-bound 9-1-1 complex facilitates ATR-mediated phosphorylation and activation of Chk1 (reviewed in Ref. 9). Chk1 then phosphorylates (10 -13) and causes proteolytic destruction of (14) Cdc25A, thereby abrogating the ability of this phosphatase to activate Cdk2/cyclin E complexes and drive S-phase progression (reviewed in Ref. 15). Chk1-mediated phosphorylation of Cdc25A and Cdc25C also prevents the activating dephosphorylation of Cdc2/cyclin B complexes and progression through G 2 phase (reviewed in Ref. 16). Based on these data, along with the observation that disruption of Chk1 function sensitizes cells to a wide range of genotoxic stresses (17)(18)(19)(20)(21)(22), one would predict that cells lacking Rad9 would be highly sensitive to replication stress and certain types of DNA damage.
Additional studies have suggested a potential second function for Rad9. Based on the observation that a nine-amino acid sequence near the N terminus of Rad9 is similar to the BH3 domain of proapoptotic Bcl-2 family members, it has been proposed that Rad9, like other BH3-only proteins (23), might traffic to mitochondria, bind to antiapoptotic Bcl-2 family members, and facilitate apoptosis (24). Consistent with this possibility, exogenous epitope-tagged Rad9 has been observed to translocate to mitochondria after DNA damage (24,25), interact with Bcl-2 (24) and Bcl-x L (25), and induce apoptosis (24,25). In agreement with this proposed role, Rad9 antisense oligonucleotides also reportedly diminish methyl methanesulfonate-induced apoptosis (24). Collectively, these results suggest that Rad9 plays a critical role in the induction of apoptosis after replicative stress or genotoxic damage and lead to the prediction that Rad9 deletion might result in resistance to apoptosis.
Among the agents that activate Chk1 are drugs that target DNA topoisomerases. Etoposide, a prototypic topoisomerase II poison (26), inhibits the religation steps in the topoisomerase II catalytic cycle, thereby stabilizing covalent DNA-topoisomerase II complexes that interact with helicases to produce frank DNA double-strand breaks (27). Camptothecin, the founding member of a large class of topoisomerase I poisons (26,28), inhibits the religation step of topoisomerase I (29) to produce covalent topoisomerase I-DNA complexes (30,31) that likewise yield DNA double-strand breaks (32)(33)(34) after interaction with advancing replication forks (33,35). More recently, it has been reported that cytarabine, a nucleoside analogue that is widely used to treat acute leukemias and lymphomas (36), might also act by trapping topoisomerase I-DNA covalent complexes after incorporation of the analogue into DNA (37,38). Direct or indirect evidence has indicated that each of these prototypic anticancer agents also causes activation of Chk1 (19 -21, 39), which ordinarily occurs in an ATR-dependent manner (39,40). Because the ATR/Chk1 pathway is usually implicated in replicative stress, and whereas topoisomerase poisons induce DNA strand breaks, the exact roles of various DNA damage-activated signal transduction components in response to these agents remain to be more fully defined (41,42).
To further evaluate the role of Rad9 in the response to topoisomerase poisons and replicative stress, we have now examined the effects of camptothecin, etoposide, and cytarabine in Rad9 Ϫ/Ϫ ES cells and their wild-type counterparts. Results of this analysis failed to demonstrate any apoptotic defect in Rad9-deficient cells. To the contrary, deletion of this critical component of the checkpoint machinery sensitized cells to all three of the agents.
Cell Lines and Tissue Culture-Rad9 Ϫ/Ϫ ES cells, derived by targeted interruption of both murine Rad9 alleles, were generated and characterized by Kevin Hopkins, Wojtek Auerbach, Xiang Yuan Wang, M. Prakash Hande, Haiying Hang, Debra J. Wolgemuth, Alexandra L. Joyner, and Howard B. Lieberman, as described previously (44). Stable Rad9 Ϫ/Ϫ ES cell clones expressing human wild-type Rad9 or a Rad9 mutant in which eight phosphorylation sites were converted to alanine (Rad9-8A) were derived by transfection and selection as described (44). ES cells were propagated on gelatinized tissue culture plates in knockout Dulbecco's modified Eagle's medium containing 15% ES cell qualified fetal bovine serum (Cell and Molecular Technologies, Inc, Phillipsburg, NJ), 0.1 mM nonessential amino acids, 2 mM L -glutamine, 10 Ϫ4 M 2-mercaptoethanol, and 10 3 units/ml leukemia inhibitory factor.
To perform clonogenic assays, aliquots containing 400 ES cells were plated on gelatinized 60-mm dishes in triplicate and allowed to adhere for 4 h. Drugs or diluent were then added at the indicated final concentrations. After a 24-h incubation, drugs were removed, plates were washed, drug-free medium was added, and the cells were cultured for 7 days. Colonies were then stained with Coomassie Blue and counted. Survival was calculated as the percentage of colonies in dishes treated with drug compared with diluent.
Chromatin Binding-Retention of 9-1-1 complexes was analyzed as described previously (45). In brief, cells were treated with the indicated drug for 6 h. Alternatively, cells were irradiated with 15 J/m 2 UV irradiation at 254 nm 6 h prior to harvest. After release by trypsinization, cells were sedimented at 1000 ϫ g for 5 min, washed once with ice cold phosphate-buffered saline, and incubated for 10 min at 4°C in low-salt permeabilization buffer consisting of 10 mM HEPES, pH 7.4, 10 mM KCl, 1 mM MgCl 2 , 0.1% Triton X-100, 10 g/ml aprotinin, 5 g/ml pepstatin, 5 g/ml leupeptin, 20 nM microcystin-LR, 1 mM Na 3 VO 4 , and 10 mM ␤-glycerophosphate. Nuclei were sedimented at 1000 ϫ g for 5 min, washed with permeabilization buffer, and incubated for 10 min at 4°C with high-salt nuclear extraction buffer consisting of 1% Triton X-100, 50 mM HEPES, pH 7.4, 150 mM NaCl, 30 mM Na 4 P 2 O 7 , 10 mM NaF, 1 mM EDTA, 10 g/ml aprotinin, 5 g/ml pepstatin, 5 g/ml leupeptin, 20 nM microcystin-LR, 1 mM Na 3 VO 4 , and 10 mM ␤-glycerophosphate, a buffer that has a conductivity equivalent to a 250 mM NaCl solution. After centrifugation at 21,000 ϫ g for 5 min to remove insoluble material, clarified lysates were either used for immunoprecipitation as indicated or mixed with 1 ⁄3 volume 4ϫ SDS-PAGE sample buffer and boiled for 10 min. Aliquots derived from equal amounts of protein as assayed by the method of Bradford (46) were separated by SDS-10% PAGE. After polypeptides were transferred to Immobilon-P membranes (Millipore) and subjected to immunoblotting as described (38), bound antibodies were visualized by chemiluminescence (Super-Signal, Pierce).
Checkpoint Assays-To assay the S-phase checkpoint, 1-2 ϫ 10 4 cells were plated onto gelatinized 96-well plates and allowed to adhere for 16 h. After cells were incubated with the indicated drugs or diluent for 60 min at 37°C, [methyl-3 H]thymidine (TRA120, 5 Ci/mmol, Amersham Biosciences) was then added at 2 Ci/well. Cells were labeled for 20 min in the continued presence of drug or diluent, released by trypsinization, and harvested by transferring to glass filters using a Skatron semiautomatic harvester (Skatron Instruments AS, Lier, Norway). The filter-bound cells were lysed with distilled water. Filterbound radioactivity was determined by liquid scintillation counting. [ 3 H]Thymidine incorporation was calculated as the ratio of treated to control samples.
Apoptosis Assays-For morphological analyses, adherent cells were released by trypsinization, pooled with nonadherent cells, sedimented at 200 ϫ g for 10 min, washed with ice-cold phosphate-buffered saline, and fixed for a minimum of 12 h in 3:1 methanol:acetic acid. Samples were then deposited on glass coverslips, air-dried, stained with 1 g/ml Hoechst 33258, and examined by fluorescence microscopy as described previously (49). A minimum of 400 cells/sample were scored for apoptotic changes (chromatin condensation or nuclear fragmentation). Samples were photographed using a Zeiss Axioplan microscope equipped with a numerical aperture 1.40 ϫ 63 objective, a 365-nm excitation filter, and a 420-nm emission filter.
Alternatively, cleavage of caspase substrates was assessed by immunoblotting. Whole cell lysates were prepared in 6 M guanidine hydrochloride under reducing conditions as described previously (43). Aliquots containing 50 g of total cellular protein were subjected to SDS-PAGE on gels with 5-15% (w/v) acrylamide gradients, transferred to nitrocellulose, and probed as described previously (43,50).

Drug-induced Rad9
Chromatin Binding-In light of previous studies showing that the ATR/Chk1 pathway plays a critical role in survival of cells after many genotoxic stresses, including topoisomerase poisons (reviewed in Refs. 22, 39, 40; see also Refs. 21, 51, 52), we assessed the role of the Chk1 regulator Rad9 in the response to camptothecin, etoposide, cytarabine, and, as a control, UV light. To evaluate recruitment of Rad9 to chromatin after treatment with these agents, Rad9 chromatin binding was assessed by differential extraction followed by immunoblotting (44) in K562 human leukemia cells, which were extensively utilized in previous studies of 9-1-1 chromatin binding (45), and murine ES cells. As shown in Fig.  1, submicromolar concentrations of each agent induced in- were treated for 6 h with diluent, 100 -500 nM CPT, 10 -1000 nM cytarabine, or 100 -500 nM etoposide, as indicated. Alternatively, cells were exposed to 15 J/m 2 UV light and then incubated for 6 h. Cells were released by trypsinization, and chromatin-bound and released Rad9 was immunoprecipitated. The immunoprecipitates were separated by SDS-PAGE and immunoblotted for Rad9. creased binding of Rad9 to chromatin. It is worth noting that Rad9 chromatin binding occurs within 6 h of treatment with drug concentrations that kill less than 1 log of cells (see below). Moreover, these submicromolar drug concentrations are readily achievable in the clinical setting (28,36,53). 2 Drug-induced Phosphorylation of Chk1 and Chk2 in Parental and Rad9 Ϫ/Ϫ ES Cells-Rad9 is required for Chk1 activation induced by ionizing radiation, UV light, and the ribonucleotide reductase inhibitor hydroxyurea (44). To determine whether Rad9 also participates in Chk1 activation induced by the anticancer drugs used in Fig. 1, we assessed Chk1 phosphorylation on Ser 345 , a modification that is catalyzed by ATR and is required for Chk1 activation (40,54). These studies examined parental ES cells, which express Rad9, and Rad9 Ϫ/Ϫ ES cells lacking Rad9 ( Fig. 2A). Like UV light, all three chemotherapy agents induced robust Chk1 phosphorylation in the parental ES cells (Fig. 2B). In contrast, induction of Chk1 phosphorylation by cytarabine was almost abolished in Rad9 Ϫ/Ϫ cells, and induction of Chk1 phosphorylation by etoposide or camptothecin was markedly diminished (Fig. 2B).
To evaluate whether Rad9 is also required for activation of Chk2 by these agents, we examined a Chk2 mobility shift that is due to regulatory phosphorylation and is tightly correlated with Chk2 activation (55,56). Fig. 2B shows that Chk2 was weakly phosphorylated after treatment of parental cells by camptothecin, etoposide, or UV light (Fig. 2B, lanes 3-5). This phosphorylation was more prominent after treatment of Rad9 Ϫ/Ϫ cells with the same genotoxins (Fig. 2B, lanes 8-10), perhaps as a consequence of defective Chk1 activation. In contrast, Chk2 phosphorylation was not detectable after cytarabine treatment in either parental or Rad9 Ϫ/Ϫ cells (Fig. 2B, lanes 2 and 7), indi-cating a dichotomy in cellular responses triggered by cytarabine and the classical topoisomerase poisons.
S-phase Checkpoint Activation in Rad9 Ϫ/Ϫ ES Cells-Previous studies using Chk1-deficient cells and pharmacologic Chk1 inhibitors have shown that Chk1 plays an essential role in the activation of the S-phase checkpoint by ionizing radiation and agents that disrupt DNA replication (20, 56 -58). Correspondingly, Rad9 and Hus1, which are essential for optimal Chk1 activation, are also required for activation of the S-phase checkpoint by agents that disrupt DNA replication (44,59). Paradoxically, however, neither Rad9 nor Hus1 is required for activation of the S-phase checkpoint triggered by ionizing radiation (44,60). To determine whether Rad9 participates in activation of the S-phase checkpoint by topoisomerase poisons, which, like ionizing radiation, also induce double-strand breaks, parental and Rad9 Ϫ/Ϫ cells were treated with drugs for 1 h and pulsed with radiolabeled thymidine in the continued presence of the drugs for 20 min. Treatment of parental ES cells with 50 nM cytarabine, 500 nM camptothecin, or 500 nM etoposide diminished thymidine incorporation into DNA by 35-45% (Fig. 3A, white bars). Pretreatment of parental cells with 3 mM caffeine, a well characterized checkpoint disrupter (18,61), prior to addition of the anticancer drugs, restored  thymidine incorporation almost to control levels (Fig. 3A, black  bars). These results indicate that the decreased thymidine incorporation represented checkpoint activation rather than physical blockage of replication forks. Analysis of thymidine incorporation in Rad9 Ϫ/Ϫ ES cells revealed that the cytarabine-induced S-phase checkpoint was absent in cells lacking Rad9 (Fig. 3B, white bar). In contrast, in Rad9 Ϫ/Ϫ cells treated with camptothecin or etoposide, the S-phase checkpoint response was nearly identical to that observed in parental cells.
Consistent with the results of these thymidine incorporation assays, immunoblotting demonstrated that Cdc25A depletion after a 3-h cytarabine treatment or UV irradiation was Rad9dependent, whereas Cdc25A depletion after a 3-h camptothecin or etoposide treatment was not (Fig. 3C). Collectively, these results demonstrate that Rad9 does not play a role in the S-phase checkpoint activated by the topoisomerase inhibitors, similar to the results seen with ionizing radiation. Moreover, they provide a second difference between the effects of topoisomerase poisons and cytarabine.
Drug-induced Apoptosis in Rad9 Ϫ/Ϫ Cells-Because defects in the Chk1 signaling pathway decrease survival after genotoxic stress (22), we predicted that the Chk1 activation defect in Rad9 Ϫ/Ϫ cells might sensitize cells to genotoxins. On the other hand, recent studies suggesting that Rad9 is a BH3-only protein involved in the mitochondrial release of cytochrome c after DNA damage (24,25) would predict a decrease in DNA damage-induced apoptosis in Rad9 Ϫ/Ϫ cells. To distinguish between these alternatives, parental and Rad9 Ϫ/Ϫ cells were treated with camptothecin, etoposide, or cytarabine and analyzed for apoptotic morphological changes at various time points. Results of this analysis demonstrated that chromatin condensation and nuclear fragmentation occur after drug treatment in the absence of Rad9 (Fig. 4A), ruling out the possibility that this polypeptide plays an essential role in the induction of apoptosis by these agents. To the contrary, quantitation of the morphological apoptotic changes (Fig. 4, B and C) demonstrated that apoptosis occurred earlier (Fig. 4B) and at lower camptothecin concentrations (Fig. 4C) (p ϭ 0.012 by paired t test). Similar results were obtained with etoposide (Fig. 4D), which induced apoptosis in 50 Ϯ 2% of the Rad9 Ϫ/Ϫ cells but only 24 Ϯ 11% of the parental cells during a 48-h exposure to 125 nM drug (n ϭ 4, p ϭ 0.026), and cytarabine (Fig. 4E), which induced apoptosis in 43 Ϯ 11%  2-5, respectively). Aliquots containing 50 g of total cellular protein were subjected to SDS-PAGE followed by blotting with reagents that detect the indicated polypeptides. Fragments corresponding in size to caspase digestion products reported previously (72) are indicated with arrowheads.
of the Rad9 Ϫ/Ϫ cells but in only 20 Ϯ 11% of the parental cells during a 48-h exposure to 500 nM drug (n ϭ 4, p Ͻ 0.05).
To provide independent support for the hypothesis that apoptosis can be triggered in the absence of Rad9, whole cell lysates prepared from Rad9 Ϫ/Ϫ cells treated with varying concentrations of camptothecin (Fig. 4F) or etoposide 3 were examined by immunoblotting. In both instances, caspase activation was readily detected, as indicated by the disappearance of procaspase-3 (Fig. 4F, top) and cleavage of multiple caspase-3 substrates to their signature fragments (Fig. 4F, arrowheads). Taken together, these results demonstrate that Rad9 is not essential for anticancer drug-induced apoptosis. Instead, this polypeptide, likely by means of activating the Chk1 signaling pathway, seems to play a role in protecting cells from these apoptotic stimuli.
Effect of Rad9 on Clonogenic Survival-To rule out the possibility that the increased apoptosis in Rad9 Ϫ/Ϫ cells merely reflected enhanced kinetics without altering long-term survival, we assessed the ability of individual cells to proliferate using colony-forming assays. As illustrated in Fig. 5A, the IC 90 for camptothecin was 2.9 Ϯ 0.7-fold lower in Rad9 Ϫ/Ϫ cells than in parental cells (n ϭ 7, p ϭ 0.001). Likewise, Rad9 Ϫ/Ϫ cells displayed diminished colony-forming ability after treatment with SN-38, the active metabolite of the camptothecin analog irinotecan (Fig. 5B). Taken together, these results demonstrate that Rad9 enhances the survival of cells treated with these agents.
To rule out the possibility that these results reflected an alteration other than Rad9 gene deletion, Rad9 Ϫ/Ϫ cells expressing wild-type human Rad9 or the Rad9-8A mutant, in which eight phosphorylation sites have been mutated to alanines (44), were treated with camptothecin. As shown in Fig.  5C, wild-type human Rad9 enhanced the ability of cells to withstand a 24-h camptothecin treatment. In contrast, Rad9-8A, which activates Chk1 poorly (44), did not enhance colony survival after camptothecin treatment.
Rad9 deletion also sensitized cells to other drugs. As illustrated by the experiments in Fig. 5, D and E, Rad9 Ϫ/Ϫ cells also displayed a 3.2 Ϯ 1.5-fold decrease in the IC 90 for etoposide (n ϭ 3, p ϭ 0.02) and a 2.0 Ϯ 0.5-fold decrease in the IC 90 for cytarabine (n ϭ 4, p ϭ 0.02). Nonetheless, this sensitization was limited to agents that induce DNA damage or replicative stress. When cells were treated with paclitaxel, which binds ␤-tubulin, inhibits microtubule dynamics, and activates the mitotic checkpoint (62)(63)(64)(65), there was no difference in drug sensitivity between parental and Rad9 Ϫ/Ϫ cells (Fig. 5F).

DISCUSSION
The results of the present study provide new insight into the potential roles of Rad9 after treatment with agents that cause 3  replication stress or DNA double-strand breaks. In particular, we observed that Rad9 disruption decreased the ability of cytarabine and topoisomerase poisons to induce Chk1 activation. In addition, Rad9 deficiency diminishes S-phase checkpoint activation by cytarabine but not by the topoisomerase poisons. Despite these differences in checkpoint response, Rad9 Ϫ/Ϫ cells are hypersensitive to both antimetabolites and topoisomerase poisons. These results have potentially important implications for the current understanding of the mechanisms of action of the agents studied as well as the functions of Rad9.
Previous studies using Rad9-and Hus1-deficient cells demonstrated that these cell lines had defects in the S-phase checkpoint activated by UV light and benzo(a)pyrene dihydrodiol epoxide, two agents that introduce bulky DNA lesions capable of stalling replication forks (44,59). In contrast, neither Rad9 nor Hus1 was required for the S-phase checkpoint activated by ionizing radiation, which produces double-stranded breaks (44,59). These results suggest that the lesions produced by UV light and ionizing radiation initially activate different signaling pathways that then lead to S-phase arrest.
Analysis of the chemotherapeutic agents used in this study revealed a similar dichotomy. As was observed previously with UV light (44), the ability of cytarabine to trigger S-phase checkpoint activation and Cdc25A degradation required Rad9 (Fig.  3). In contrast, topoisomerase poisons (Fig. 3), like ionizing radiation (44), activated these checkpoint events even in the absence of Rad9. These results indicate that the cellular responses produced by cytarabine and UV light are similar to each other and are somewhat different from the responses produced by ionizing radiation and the topoisomerase poisons.
The lesions that activate the S-phase checkpoint in a 9-1-1dependent manner have in common the fact that they introduce replication stress (44,59). The fact that cytarabine also requires Rad9 to induce S-phase arrest is consistent with earlier reports showing that cytarabine induces replication stress after its incorporation into DNA (66 -69). On the other hand, recent studies have suggested that cytarabine might act like a topoisomerase I poison (37,38). Evidence to support this claim has included (i) the demonstration that the religation step of topoisomerase I is slowed ϳ5-fold when an oligonucleotide containing cytidine arabinoside in place of cytidine at a topoisomerase I cleavage site is exposed to purified enzyme in vitro; (ii) the observation that covalent topoisomerase I-DNA complexes are increased in neoplastic cell lines several hours after the addition of cytarabine to cultures; and (iii) the demonstration that camptothecin-selected P388/camptothecin (CPT) murine lymphoma cells, which lack detectable topoisomerase I and are 4000-fold more resistant to CPT, are 5-fold resistant to cytarabine. In contrast, our observations indicate that the cell cycle response of ES cells to cytarabine is quite different from the response to the topoisomerase I poison CPT. We cannot at present rule out the possibility that our results reflect differences between ES cells and lymphoma cells. Likewise, we cannot exclude the possibility that cytarabine causes the accumulation of topoisomerase I-DNA complexes over time, possibly as a consequence of biochemical changes that contribute to apoptosis. Nonetheless, the present study suggests that the cell cycle response to cytarabine is most consistent with the action of a classical antimetabolite, as originally envisioned by earlier authors (36).
In contrast to the results seen with cytarabine, Rad9 is not required for activation of the S-phase checkpoint by topoisomerase I or topoisomerase II poisons (Fig. 3). This observation is similar to results seen previously with ionizing radiation. Because topoisomerase poisons and ionizing radiation all introduce double-stranded DNA breaks, it is likely that this lesion is the trigger for the Rad9-independent S-phase checkpoint described in Fig. 3 and our previous study (44). Additional work has also indicated that the topoisomerase I poison topotecan induces activation of the S-phase checkpoint in SV40-transformed human fibroblasts (39). Importantly, expression of a kinase dead ATR allele in these cells diminishes topotecan-induced S-phase slowing (39), suggesting a critical role for ATR in S-phase checkpoint activation. Consistent with this observation, the ATR substrate Chk1 is also required for activation of S-phase checkpoint and Cdc25A degradation after treatment with ionizing radiation in other cell types (11,13). In contrast, Rad9 deficiency does not affect that S-phase checkpoint in ES cells treated with the topoisomerase poisons (Fig. 3) or ionizing radiation (44). Although the reasons for this discrepancy are not currently known, Chk2 is hyperactivated when Rad9 Ϫ/Ϫ ES cells are treated with the topoisomerase poisons (Fig. 2B). Because both Chk1 and Chk2 phosphorylate overlapping sites on Cdc25A that participate in degradation of this phosphatase (13), one possible explanation for our findings is that the hyperactivation of the Chk2 signaling pathway contributes to topoisomerase poison-induced S-phase checkpoint activation in the Rad9 Ϫ/Ϫ ES cells.
Even though Rad9 Ϫ/Ϫ cells have no reproducible defects in their ability to activate the S-phase checkpoint after treatment with topoisomerase poisons, Rad9 deletion nonetheless sensitizes ES cells to camptothecin and etoposide (Figs. 4 and 5). These observations confirm and extend recent data showing that Rad9 Ϫ/Ϫ ES cells are also hypersensitive to ionizing radiation in the absence of an S-phase checkpoint defect (44). In the present study, these differences in drug sensitivity were observed when apoptosis was quantified and when long-term antiproliferative effects of the various agents were assessed using colony forming assays. These results raise the possibility that Rad9 plays critical roles in other aspects of the response to DNA damage, in addition to its role in checkpoint activation. Further studies are required to identify these additional roles that contribute to survival of cells with wild-type Rad9 function.
Finally, recent studies have raised the possibility that Rad9 might also contribute to the DNA damage response by acting as a BH3-only polypeptide and trafficking to mitochondria, where it facilitates apoptosis by binding to and inhibiting the antiapoptotic polypeptides Bcl-2 and Bcl-x L . A previous study from our laboratory (45), however, failed to detect endogenous Rad9 in mitochondria by immunoblotting or immunofluorescence after ionizing radiation but could not rule out the possibility that the amount of mitochondrial Rad9 was below the limit of detection with the reagents utilized. Consistent with these earlier results, Rad9 Ϫ/Ϫ cells displayed readily detectable apoptosis after a variety of DNA damaging treatments (Fig. 4), indicating that trafficking of Rad9 to mitochondria cannot be required for induction of apoptosis in ES cells. To the contrary, Rad9 Ϫ/Ϫ cells were hypersensitive to camptothecin and etoposide as assessed by apoptotic and clonogenic assays (Figs. 4 and 5). We cannot at present rule out that mitochondrial trafficking of Rad9 makes a minor and nonessential contribution to induction of apoptosis in ES cells, nor can we eliminate the possibility that mitochondrial trafficking of Rad9 plays a more important role in other cell lines. Nonetheless, the present results suggest that the predominant role of Rad9 in ES cells relates to protection from agents that induce DNA double-strand breaks or replication stress.