Cellular Responses to the DNA Strand-scission Enediyne C-1027 Can Be Independent of ATM, ATR, and DNA-PK Kinases*

The current paradigm based upon ionizing radiation (IR) studies states that cells deficient in either ataxia-telangiectasia-mutated kinase (ATM) or related phosphatidylinositol 3 (PI 3) -kinases (ATR and DNA-PK) are hypersensitive to DNA strand breaks because they are unable to rapidly activate downstream effectors such as p53. Here we have contrasted cell responses to IR and C-1027, a radiomimetic antibiotic that induces DNA strand breaks. At equal levels of DNA double strand breaks, cell lines with inactive ATM or other phosphatidylinositol 3-kinases displayed classical hypersensitivity to IR but not to C-1027. Moreover, phosphorylation of p53 Ser-15 induced by C-1027 was independent of ATM, ATR, or DNA-PK function. We have concluded that the model based on IR studies cannot always be directly applied to DNA damage induced by other strand-scission agents.

DNA damage resulting from endogenous cellular processes or exogenous threats such as ionizing radiation (IR), 1 UV, or chemicals poses a challenge to a cell's genomic stability. The ability of a cell to detect and respond to DNA damage is crucial for its survival. Therefore, a complex network of DNA damage sensors, signal transmitters, and effectors (checkpoints) has evolved in all eukaryota. Checkpoint activation results in a delay in cell cycle progression, induction of DNA repair, and regulation of specific gene expression (1,2). These checkpoints prevent cells from replicating or propagating damaged DNA, thus maintaining genome stability and integrity.
It is believed that the top level of sensors/transmitters in the signal transduction cascade that responds to DNA strand breaks, or to DNA damage in general, consists of several conserved protein kinases (3). The most important among these are the members of the phosphatidylinositol 3-kinase family: ataxia-telangiectasia-mutated protein (ATM) (4,5), ATM-and Rad3-related protein (ATR) (6), and the DNA-dependent protein kinase (DNA-PK) (7). In response to DNA damage, the tumor suppressor protein p53 is activated and stabilized by post-translational modifications, including phosphorylation of Ser-15 by one or all of these kinases.
Following DNA damage induced by IR or other strand-scission agents, ATM phosphorylates p53 on Ser-15 in vitro (8) and in vivo (9,10). Cells containing non-functional ATM protein failed to phosphorylate Ser-15 and were hypersensitive to IR and radiomimetic drugs (4,11,12). In vitro studies showed that ATM is a manganese-dependent, serine/threonine kinase able to phosphorylate several cellular proteins, including itself (8,13). Less is known about ATR, a protein kinase that has been shown to phosphorylate p53 on Ser-15 and Ser-37 in vitro and on Ser-15 in vivo in response to both IR-type (breaks) and UV-type (bulky modification) DNA damage. ATR is probably responsible for the slow kinetic component of damage-induced p53 phosphorylation/accumulation (6,14). Recent work has demonstrated that overexpression of catalytically inactive ATR in human cells resulted in hypersensitivity to IR and hydroxyurea as well as in abrogation of a cell cycle checkpoint (15,16). DNA-PK is a multiprotein complex composed of a catalytic subunit (DNA-PK cs ) and a heterodimeric (Ku70 and Ku80) DNA binding component (17). In response to DNA strand breaks, the Ku70/Ku80 subunit binds free DNA ends facilitating the DNA binding and activation of DNA-PK cs (18). In vitro, the activated DNA-PK cs can phosphorylate a variety of proteins involved in DNA damage sensing and/or repair, including p53 on Ser-15 and Ser-37 (19). Despite the apparent overlap of function between ATM and ATR (and probably DNA-PK), it is widely accepted, based upon IR studies, that DNA double strand breaks are signaled to p53 primarily through ATM kinase (20).
We asked whether the ATM-dependent DNA strand break response pathway(s) is truly universal and have addressed this question by comparing cellular responses to IR and a strand break-inducing compound, C-1027, in wild-type and ATM-deficient cells. C-1027 is a member of the enediyne family of drugs, which are often classified as radiomimetic agents, i.e. DNAinteracting drugs which induce single strand (SSB) and/or double strand (DSB) breaks by free radical attacks on the deoxyribose moieties in DNA (21,22). A characteristic feature of C-1027 is its propensity to induce a higher ratio of DNA DSB to SSB as compared with IR. In addition, C-1027 biological effects are similar to IR-induced effects, including reduction of cellular DNA replication, delayed progression through S phase, and G 2 /M cell cycle block (23)(24)(25).
In this work, we have assessed the role of ATM, as well as ATR and DNA-PK, in the DNA damage responses induced by the DNA strand-scission agent C-1027 using several human cell lines with either wild-type or absent/inactive damage-activated kinases. We found that wild-type cells respond similarly to DNA damage induced by IR or C-1027. However, ATMdeficient cells were not hypersensitive to C-1027 and did not exhibit radioresistant DNA synthesis after drug treatment, indicating the presence of an active, ATM-independent S phase checkpoint. Moreover, in ATM-deficient cells, IR failed to induce p53 Ser-15 phosphorylation, whereas C-1027 stimulated efficient phosphorylation. Likewise, the effects of C-1027 on cells were not dependent on ATR or DNA-PK status.
Growth Inhibition Assay-Cells growing in 35-mm dishes were treated with drugs for 2 h or irradiated on ice at 1.25 Gy/min and then washed and incubated in fresh, drug-free medium. Following a 3-day incubation, dishes were washed and viable cells attached to the dish counted. Cell growth inhibition was calculated by comparing the number of treated cells to non-treated controls.
Incorporation of [ 3 H]Thymidine into Cellular DNA-Inhibition of thymidine incorporation was determined as described previously (26). 14 C-prelabeled cells were treated with drugs for the indicated time or irradiated on ice and allowed to recover in warm medium. [methyl-3 H]Thymidine was added to a final concentration of 0.2 Ci/ml for the last 30 min of the drug treatment. The acid-insoluble radioactivity in samples was measured using an LS-3800 liquid scintillation counter. Inhibition of thymidine incorporation into cellular DNA was calculated as the ratio of 3 H to 14 C in drug-treated samples compared with nontreated controls.
Genomic DNA Damage and Repair-Genomic DNA damage was quantitated as described previously (25) with some modifications. 14 Cprelabeled cells were treated with drugs for 2 h or irradiated on ice and harvested immediately (for DNA damage) or incubated in fresh medium for 24 h (for damage repair). Cells were washed and resuspended in 0.5% low melting point agarose, poured into a plug former (250 l per plug block), and kept at 4°C until solidified. Plugs containing approximately 3 ϫ 10 5 cells were incubated with 1 mg/ml proteinase K in lysis buffer (10 mM Tris, pH 8, 2% sodium sarcosinate, 0.5 M EDTA, pH 8.0) for 24 h at 50°C, followed by 2 h of incubation with 0.1 mg/ml RNase A at 37°C in 10 mM Tris (pH 7.6), 0.1 M EDTA. The plugs were loaded into wells of a 0.8% agarose gel. Pulse-field gel electrophoresis was conducted in 0.5ϫ Tris borate/EDTA for 68 h at 45 V with a switching time of 60 min, using a ChefDR apparatus (Bio-Rad). The 14 C DNA signal from dried gels was analyzed using a PhosphorImager scanner (Amersham Biosciences). Damage to genomic DNA was calculated as a fraction of DNA migrated from the well in treated and non-treated cells (FAR, fraction activity released; Ref. 27).
Immunoblotting Analysis of p53-Cells incubated with drugs at 37°C for 0.5-6 h or irradiated on ice and then allowed to recover for 0.5-6 h at 37°C were harvested and washed in phosphate-buffered saline. Total cellular extracts were prepared by incubating cells on ice in lysis buffer (50 mM Tris, pH 7.6, 5 mM EDTA, 5 mM EGTA, 150 mM NaCl, 0.3% Nonidet P-40, 0.2% Triton X-100, 50 mM sodium fluoride, 1 mM sodium o-vanadate, 0.1 mM ␤-glycerophosphate, 1 mM phenylmethylsulfonyl fluoride, and 10 g/ml (each) aprotinin, pepstatin, and leupeptin) for 10 min, followed by two cycles of rapid freeze-thaw in a dry ice/methanol bath. The cell lysates were cleared by centrifugation, and protein content was determined using the Bradford method (Bio-Rad). To verify phosphorylation, in some experiments the extracts were treated with alkaline calf thymus phosphatase for 1 h at 37°C according to the manufacturer's instructions (Sigma). Equal amounts of protein were electrophoresed on SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. The membranes were stained with Ponceau S to ensure equal transfer and then probed with primary antibodies, followed by secondary antibodies conjugated with horseradish peroxidase. The primary antibodies used were anti-p53 (DO-1, Santa Cruz Biotechnology, Santa Cruz, CA), anti-phospho-p53 Ser-15 (Cell Signaling Technology, Beverly, MA), and anti-␤-actin (Sigma). Protein bands were visualized by enhanced chemiluminescence, and the levels of p53 were measured using a personal densitometer (Amersham Biosciences).

IR-and C-1027-induced p53 Ser-15 Phosphorylation in ATM and p53
Wild-type Cells-The ability of C-1027 to stimulate p53 accumulation and phosphorylation was compared with that of ionizing radiation in HCT-116 cells bearing functional p53. Treatment of HCT-116 cells with either IR or C-1027 resulted in an increase in the levels of total p53 protein and p53 Ser-15 phosphorylation. As shown in Fig. 1A, the control, mocktreated cells expressed low amounts of non-phosphorylated Ser-15 p53 protein. The levels of p53 protein and Ser-15 phosphorylation were both dependent on the dose of IR or C-1027 (2.5, 7.5, 25 Gy and 0.1, 0.3, and 1 nM, respectively). Similar increases in p53 protein levels and phosphorylation were observed in kinetic studies in which cells were treated with IR (25 Gy) or C-1027 (1 nM) for 0.5, 2, and 6 h (Fig. 1B).
To confirm that the similar p53 response in HCT-116 cells is evoked by similar levels of genomic DNA damage induced by both agents, alkaline and neutral comet assays were carried out using cells incubated under the same conditions as p53 accumulation/phosphorylation studies. Assays conducted under alkaline conditions allow the detection of overall DNA damage, including double and single strand breaks, abasic sites, and alkali-labile adducts, whereas under neutral conditions only DNA double strand breaks are detectable. Cells treated with 25 Gy of IR or 1 nM C-1027 exhibited a similar level of both DNA double strand breaks and overall DNA damage (data not shown). Thus, relatively equal amounts of DNA damage induced by two different agents, IR and C-1027, caused similar increases in p53 accumulation and Ser-15 phosphorylation in cells with wild-type p53 and active kinases.
C-1027-induced p53 Ser-15 Phosphorylation Is ATM-inde- pendent-An active ATM kinase is known to be required for rapid and efficient p53 Ser-15 phosphorylation in response to IR-induced DNA damage. To test the role of ATM in C-1027induced p53 Ser-15 phosphorylation, a pair of SV40-transfected human fibroblast GM00637H (ATMϩ) and GM05849C (ATMϪ) cell lines was used. Although the presence of viral large T antigen in these cells precludes studies of p53 accumulation, Ser-15 phosphorylation can be readily measured because the level of p53 is stable (6).
Both ATMϩ and ATMϪ cell lines were mock-treated (control) or incubated with 1 nM C-1027 for 2 h (C-1027) or exposed on ice to 25 Gy of IR and postincubated in fresh, warm medium for an additional 2 h (IR). As shown in Fig. 2A, IR and C-1027 were equally efficient in inducing p53 Ser-15 phosphorylation in ATMϩ cells. However, only C-1027 was able to stimulate Ser-15 phosphorylation in ATMϪ cells, although to a lower extent than in ATMϩ cells. Calf intestinal phosphatase was used to confirm the phosphorylated state of the p53 Ser15-P band.
The level of initial DNA damage in the ATMϩ and ATMϪ cell lines was determined quantitatively using pulse-field gel electrophoresis. As presented in Fig. 2B, treatment with 25 Gy of IR or 1 nM C-1027 induced similar amounts of DNA damage in the tested cell lines. Therefore, the differences observed between C-1027-and IR-induced p53 Ser-15 phosphorylation levels in ATMϪ cells could not be explained by variances in the level of DNA damage.
ATM Deficiency Does Not Result in Hypersensitivity to C-1027-induced DNA Damage-A characteristic feature of ATM-deficient cells is their hypersensitivity to ionizing radiation. Irradiated ATMϪ cells cannot activate checkpoints and repair machinery; instead, they proceed with radioresistant DNA synthesis, which results in accumulation of unrepaired DNA, chromosomes breaks, and eventually in cell death.
To test the effects of IR and C-1027 on cell growth, ATMϩ or ATMϪ cells were treated with drugs for 2 h or irradiated on ice before being incubated in fresh warm medium for an additional 3 days. Resulting growth curves are presented in Fig. 3A. ATM-deficient cells were hypersensitive to ionizing radiation (sensitivity ratio, 5.5). In contrast, C-1027 was equally active against both cell lines (sensitivity ratio, 1.5), suggesting cellular responses to this drug are ATM-independent. Treatment of ATMϩ and ATMϪ cells with 25 Gy of IR or with 1 nM C-1027 for 2 h led to a decrease in [ 3 H]thymidine incorporation as compared with untreated controls, indicating inhibition of DNA synthesis (Fig. 3B). However, both cell lines were equally sensitive to C-1027-induced inhibition (ϳ30% of control), while DNA synthesis was more resistant to IR in ATMϪ than in ATMϩ cells (ϳ70 and 50% of control, respectively). Collectively, these findings demonstrate a marked difference in cellular responses to DNA damage induced by IR and C-1027.
The Dose and Time Course of p53 Accumulation/Phosphorylation Induced by C-1027 or IR Are Different-To further evaluate similarities and differences between IR-and C-1027-stimulated phosphorylation of p53 Ser-15, the dose and time dependence of phosphorylation were examined in ATMϩ and ATMϪ cell lines. The doses of ionizing radiation ranging from 2.5 to 75 Gy resulted in efficient phosphorylation of p53 Ser-15 in ATMϩ cells, whereas the phosphorylation was absent in ATMϪ cells even at the highest dose (Fig. 4A, left panel). It is worth noting that the level of Ser-15 phosphorylation after a 2-h postincubation remained virtually constant despite the 30fold dose range. In contrast to this, cells treated with C-1027 (0.1-3 nM) showed a gradual increase in p53 Ser-15 phosphorylation as the drug concentration increased (Fig. 4A, right  panel).
Additional differences between the abilities of IR and C-1027 to induce phosphorylation of p53 Ser-15 were observed in time course experiments. Ionizing radiation (25 Gy) induced rapid Ser-15 phosphorylation in ATM wild-type cells (ATMϩ), reaching a plateau after 0.5 h of postirradiation (Fig. 4B, left panel) and a negligible increase in phosphorylation in ATMϪ cells following up to 2 h of postincubation. However, prolonged incubation (6-h postirradiation) resulted in a p53 Ser-15 phosphorylation level in ATMϪ cells that was similar to the level in ATMϩ cells (data not shown), which is in agreement with previous studies (28,29). In C-1027-treated cells, p53 Ser-15 was phosphorylated more gradually, reaching a plateau level after 1-2 h of incubation (Fig. 4B, right panel). In addition, similar kinetics of C-1027-induced phosphorylation were observed independently of ATM status of the cells.
The Phosphorylation of p53 on Ser-15 in Response to C-1027induced DNA Damage Is Independent of ATR or DNA-PK Status-The relatively slow kinetics of p53 Ser-15 phosphorylation in response to C-1027 treatment pointed to possible involvement of ATR kinase. To investigate the effect of ATR activity on p53 Ser-15 phosphorylation in response to IR-and C-1027induced damage, the SV40-transformed human fibroblast cell line (GM00847) transfected with doxycycline-inducible ATR wild-type (ATR wt ) or ATR kinase-dead (ATR kd ) protein was employed (15).
Cells were preincubated with doxycycline for 24 h to induce overexpression (ϩ) or left untreated (Ϫ) and then exposed to 25 Gy of IR or 1 nM C-1027 (Fig. 5A). After 2 h of postirradiation or incubation with the drug, there were no detectable differences in phosphorylation levels between cells overexpressing ATR wt and cells overexpressing ATR kd . These findings suggest that ATR alone cannot be responsible for C-1027-induced p53 Ser-15 phosphorylation. To further examine the effects of ATR kinase on cellular responses to IR and C-1027, growth inhibition studies were carried out as described earlier. As shown in Fig. 5B, the overexpression of ATR kd resulted in increased sensitivity to IR (sensitivity ratio, 4.0). However, cells treated with C-1027 demonstrated equal sensitivity (sensitivity ratio, 0.9) to the drug, independent of ATR kd expression. Additionally, growth inhibition studies using cells overexpressing ATR wt protein revealed no differences in sensitivity to either IR or C-1027 (data not shown).
Because neither ATM nor ATR activity is exclusively needed for C-1027-stimulated p53 Ser-15 phosphorylation or cell growth inhibition under the conditions tested, the involvement of DNA-PK kinase, the third member of the phosphatidylinositol 3-kinase family known to phosphorylate p53 on Ser-15, was investigated. Studies were carried out using a pair of human glioblastoma cell lines, one of them deficient in active DNA-PK protein (MO59J, DNA-PKϪ) and the other with active kinase (MO59K, DNA-PKϩ). Treatment with IR (25 Gy) or C-1027 (1 nM for 2 h) stimulated p53 Ser-15 phosphorylation (Fig. 6A) independently of DNA-PK status, indicating that this kinase also is not crucial for Ser-15 phosphorylation. However, cells lacking DNA-PKϪ were significantly more sensitive to ionizing radiation than DNA-PKϩ cells (sensitivity ratio, 4.8). In contrast, there was no difference in sensitivity to C-1027 treatment (sensitivity ratio, 0.8) (Fig. 6B). ATMϩ and ATMϪ were exposed to IR (25 Gy) or C-1027 (1 nM), and the levels of p53 Ser-15 phosphorylation were determined at the indicated time points (0 -2 h).

FIG. 5. Active ATR kinase is not required for p53 Ser-15 phosphorylation in response to IR-or C-1027-induced DNA damage.
A, human fibroblast cells transfected with inducible (doxycyclineϩ or doxycyclineϪ) active (ATR wt ) or kinase-dead (ATR kd ) ATR kinase were exposed to IR (25 Gy) or C-1027 (1 nM), and the levels of p53 Ser-15 phosphorylation were determined at 2 h postirradiation or with drug treatment. B, ATR kd -transfected cells were mock-treated (closed circles) or induced with doxycycline (open circles) for 24 h and then drug-treated for 2 h or irradiated on ice at the indicated doses. Following a 3-day incubation, cells were counted, and the results were expressed as percentage of the growth of non-treated cells.

DISCUSSION
Post-translational modifications of p53 play a key role in cellular responses to DNA damage induced by ionizing radiation or radiomimetic agents (10,30). Currently, the accepted paradigm derived from IR studies is that the presence of DNA damage is signaled to p53 through a protein kinase cascade controlled by ATM and manifests in phosphorylation of Ser-15 from p53 (20). Accordingly, the disruption of ATM functionality increases cellular sensitivity to DNA strand breaks because of an inability to sense, signal, and/or repair the damage, leading to erroneous DNA replication and ultimately genomic instability and cell death.
By comparing IR and the radiomimetic enediyne antibiotic C-1027, we have demonstrated in this work that cellular response to DNA strand breaks could depend on the nature of the damage-inducing agent. In preliminary experiments, HCT-116 cells harboring active DNA damage-dependent kinases and wild-type p53 responded similarly to equal levels of strand breaks induced by IR or C-1027. Based on IR studies, the inactivation of ATM should diminish C-1027-induced p53 Ser-15 phosphorylation and increase C-1027 cytotoxicity. However, in contrast to IR, C-1027 stimulated phosphorylation of p53 Ser-15 in ATM-deficient cells (Fig. 2). Moreover, ATMdeficient cells were not hypersensitive to C-1027 nor did they exhibit radioresistant DNA synthesis after drug treatment (Fig. 3B), indicating the presence of an active and ATM-independent S phase checkpoint. The somewhat lower p53 phosphorylation level observed in ATMϪ cells could indicate that another kinase(s), which in ATMϩ cells cooperates with ATM to achieve fast and efficient p53 Ser-15 phosphorylation, is unable to complete this process alone.
It is important to stress that the differences in p53 Ser-15 phosphorylation and radioresistant DNA synthesis induction were observed at IR and drug doses that caused similar amounts of DNA double strand breaks as assessed by quantitative pulse-field gel electrophoresis analysis (Fig. 2B). In addition, because p53 is a stress response protein, the cytotoxic stress induced by IR or C-1027 was kept at comparable levels in phosphorylation experiments. We can therefore conclude that under these conditions the differences in cellular responses to IR or C-1027 were not because of differences in DNA damage or stress levels.
The time and dose dependence experiments (Fig. 4) demonstrated that the IR-inducible increase in p53 Ser-15 phosphorylation in ATMϩ cells is immediate and readily observed even at the lowest doses used. However, in ATMϪ cells, p53 Ser-15 phosphorylation was not detected even at the highest doses (75 Gy) following 2 h of incubation. In contrast, although C-1027 induction of p53 Ser-15 phosphorylation was found to be a little slower and less efficient, it was present in both cell lines (ATMϩ and ATMϪ). These findings support the theory that a kinase other than ATM participates in signaling C-1027-induced DNA. Taken together, our data show that ATM is not essential for the signaling of C-1027-induced DNA strand breaks to p53, in contrast to the ATM requirement for signaling IR-induced damage.
It has been proposed that the initial, fast stage of IR-induced p53 Ser-15 phosphorylation is controlled by ATM kinase, whereas the second and slower stage is maintained by ATR (6,15,31). The slower, gradual increase over time in Ser-15 phosphorylation in C-1027-treated ATMϩ cells, as compared with IR-treated cells, was observed. Similar kinetics of phosphorylation in the drug-treated ATMϪ cells points to ATR as a kinase involved in C-1027-stimulated p53 Ser-15 phosphorylation. However, ATR also is not sufficient to exclusively mediate the cellular responses to C-1027, since active ATR is not required to phosphorylate p53 Ser-15 in response to C-1027 (Fig. 5). Furthermore, cell lines expressing normal and kinase-dead ATR were equally sensitive to cytotoxic effects of C-1027. In contrast, ATR kd cells were more sensitive to killing by IR, as already demonstrated (15).
Because DNA-PK is also involved in DNA surveillance and repair and is known to phosphorylate p53 on Ser-15, we evaluated p53 phosphorylation and cytotoxic responses to IR and C-1027 in DNA-PK wild-type (DNA-PKϩ) and DNA-PK-deficient (DNA-PKϪ) cell lines (Fig. 6). The data show that both treatments were able to induce Ser-15 phosphorylation independently of DNA-PK status. Similar to ATM-and ATR-deficient cell lines, DNA-PKϪ cells were not hypersensitive to C-1027, although they were hypersensitive to IR, in accordance with previously published results (32).
Based on data presented here, we propose a scenario in which C-1027-induced DNA damage can activate both ATM and ATR. In wild-type cells, phosphorylation of p53 Ser-15 could occur through either of the kinases or their cooperative action. However, in cells deficient in one kinase activity, the second active kinase could conduct the phosphorylation. Studies using double-deficient cells (ATMϪ cells overexpressing ATR kd protein) would be needed to confirm this model.
The question arises whether C-1027-induced DNA damage is similar to or different from IR-induced damage and thus is signaled through the same or different pathway(s). The main mechanism of DNA scission by IR or radiomimetic drugs is based on oxidation of deoxyribose in DNA, initiated by hydrogen abstraction at C-1Ј, C-4Ј, and/or C-5Ј carbon atoms (33,34). In addition to strand breaks, IR produces a multitude of different base and nucleotide damage in DNA, including heterogeneous clustered damage (35). In contrast, radiomimetic drugs usually cause only a subset of the IR-induced DNA damage. C-1027 primarily attacks the deoxyribose C-4Ј atom, and to a lesser extent C-1Ј, and produces both single and double strand breaks by cleaving DNA strands two base pairs apart at specific sequences (5Ј-TAT and 5Ј-AGA) (21, 36). Radiomimetic drug-induced DNA damage is also more sequence-or regionspecific, as opposed to random damage induced by IR (37). It has been also shown that using strict anaerobic conditions and high micromolar concentrations, C-1027 can form DNA interstrand cross-links and adducts in vitro (38). However, these conditions are almost impossible to achieve in cells and are unlikely to contribute to increased p53 Ser-15 phosphorylation in cells treated with C-1027.
Another factor for differences in cellular responses to C-1027 as compared with IR is found in the ratios of DSB to SSB induced by a particular treatment. C-1027 is known to produce relatively high levels of double strand damage with DSB/SSB ratios ranging from 1:5 (21) to 1:2 (39). In contrast, IR and other drugs such as neocarzinostatin or bleomycin produce mostly single strand damage on DNA (22,33,40). In addition, the DSB induced by these agents are often the results of closely placed SSB, whereas for C-1027, SSB and DSB result from different binding modes of the drug to DNA (21). Despite differences in chemistry and/or localization of DNA damage between IR and radiomimetic drugs, to the best of our knowledge C-1027 is the only agent able to stimulate p53 Ser-15 phosphorylation and DNA replication inhibition in an ATMϪ cell line. Reports have been published showing that several other radiomimetic compounds (for example, bleomycin and neocarzinostatin) induced radioresistant DNA synthesis and increased cell death in ATM-deficient cells (41,42). Compared with the related enediyne neocarzinostatin, C-1027 independence on ATM activity for stimulating p53 phosphorylation is somewhat unexpected since we have shown recently that both C-1027 and neocarzinostatin induced similar changes in Replication Protein A (RPA) protein status in the 293 human embryonic kidney cell line (43). The phosphorylation of the subunit RPA32, presumably by one of the phosphatidylinositol 3-kinases, may modulate DNA replication or repair directly or indirectly through interactions with other proteins, including p53 (44,45). Cells treated with C-1027 and neocarzinostatin at doses causing similar levels of DNA damage responded with equivalent increases in RPA32 hyperphosphorylation and decreased replication activity, suggesting a common response to enediyne-induced DNA damage. However, as presented here, those similarities do not necessarily extend to other types of DNA damage responses.
In conclusion, our results indicate that the model for cellular responses to DNA double strand breaks derived from IR studies cannot necessarily be applied to DNA strand breaks induced by other agents. Future studies are needed to relate the type, chemistry, localization, and severity of DNA strand breaks to differences in cellular responses by using a variety of DNA strand-scission agents besides ionizing radiation. C-1027 may be an ideal agent to use in this study, because unlike IR or other radiomimetic drugs it induces relatively high levels of sequence-specific double strand breaks.