Suppression of replication fork progression in low-dose ... - Nature

Oncogene (2006) 25, 5921–5932

& 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc

ORIGINAL ARTICLE

Suppression of replication fork progression in low-dose-specific p53-dependent S-phase DNA damage checkpoint T Shimura1, M Toyoshima1, SK Adiga1, T Kunoh2, H Nagai1, N Shimizu3, M Inoue4 and O Niwa1 1

Department of Late Effect Studies, Radiation Biology Center, Kyoto University, Yoshida Konoe, Sakyo-ku, Kyoto, Japan; Department of Radiation System Biology, Radiation Biology Center, Kyoto University, Yoshida Konoe, Sakyo-ku, Kyoto, Japan; 3 Graduate School of Biosphere Science, Hiroshima University, Hiroshima, Japan and 4Medical Research Institute, Kanazawa Medical University, Uchinada-machi, Kahoku-gun, Ishikawa, Japan 2

The S-phase DNA damage checkpoint is activated by DNA damage to delay DNA synthesis allowing time to resolve the replication block. We previously discovered the p53-dependent S-phase DNA damage checkpoint in mouse zygotes fertilized with irradiated sperm. Here, we report that the same p53 dependency holds in mouse embryonic fibroblasts (MEFs) at low doses of irradiation. DNA synthesis in p53 wild-type (WT) MEFs was suppressed in a biphasic manner in which a sharp decrease below 2.5 Gy was followed by a more moderate decrease up to 10 Gy. In contrast, p53/ MEFs exhibited radioresistant DNA synthesis below 2.5 Gy whereas the cells retained the moderate suppression above 5 Gy. DNA fiber analysis revealed that 1 Gy irradiation suppressed replication fork progression in p53 WT MEFs, but not in p53/ MEFs. Proliferating cell nuclear antigen (PCNA), clamp loader of DNA polymerase, was phosphorylated in WT MEFs after 1 Gy irradiation and redistributed to form foci in the nuclei. In contrast, PCNA was not phosphorylated and dissociated from chromatin in 1 Gy-irradiated p53/ MEFs. These results demonstrate that the novel low-dosespecific p53-dependent S-phase DNA damage checkpoint is likely to regulate the replication fork movement through phosphorylation of PCNA. Oncogene (2006) 25, 5921–5932. doi:10.1038/sj.onc.1209624; published online 8 May 2006 Keywords: p53; PCNA; ATM; S-phase DNA damage checkpoint; replication fork progression

Introduction The p53 tumor suppressor gene plays prominent roles for the maintenance of genomic integrity and its functions in G1/S and G2/M checkpoints have been well documented (Levine, 1997; Fei and El-Deiry, 2003). Correspondence: Dr T Shimura, Laboratory of Molecular Pharmacology, Center for Cancer Research, NCI, NIH Bldg 37, Rm 5056 37 Convent Dr Bethesda, MD 20892-4255, USA. E-mail: [email protected] Received 27 June 2005; revised 8 March 2006; accepted 13 March 2006; published online 8 May 2006

In contrast, the role of p53 in the S-phase DNA damage checkpoint was reported only in a few studies in which the checkpoint was induced in tissue culture cells by depletion of the nucleotide substrates or in mouse zygotes by fertilization with X-ray-irradiated sperm (Agarwal et al., 1998; Shimura et al., 2002a). p53 was shown to respond to the replication block induced by aphidicolin (Gottifredi et al., 2001; Nayak and Das, 2002) and colocalized with BLM, Rad51 and 53BP1 at the sites of stalled DNA replication fork (Sengupta et al., 2003, 2004). p53 also accumulated during S-phase in Cdc7/ ES cells with defective replication origin firing (Kim et al., 2002). These studies suggest that p53 is likely to play a certain role in the S-phase DNA damage checkpoint in mammalian cells. The S-phase DNA damage checkpoint is activated by DNA damage to delay DNA synthesis, allowing time to resolve the replication block (Boddy and Russell, 2001; Osborn et al., 2002). Previously, the S-phase DNA damage checkpoint in tissue culture cells was studied at radiation doses higher than 5 Gy; mostly 10–20 Gy which were supra-lethal doses to the cells. The colony survival of Chinese hamster ovary (CHO) cells, a standard tissue culture cell line, decreases to 103 for 10 Gy and to 107 for 20 Gy (Elkind and Sutton, 1959). However, the S-phase DNA damage checkpoint is induced in normal human fibroblast (NHF) at a dose as low as 1 Gy (Brown et al., 2003) which introduces around 40 double strand breaks in the whole genome but affects the colony-forming survival of repair proficient cells minimally. Therefore, it is desirable to use lower radiation doses of a more biological significance for the study of the molecular mechanism of the S-phase DNA damage checkpoint. Indeed, we have described a novel p53-dependent S-phase DNA damage checkpoint in mouse zygotes fertilized with X-irradiated sperm. The S-phase DNA damage checkpoint in mouse zygotes operated in a biphasic manner such that 3H-TdR uptake was sharply suppressed at sperm doses below 2 Gy but less sharply at higher doses and the first portion was found to be dependent on p53 (Shimura et al., 2002a). The S-phase DNA damage checkpoint is known to operate through two mechanisms; by suppressing new origin firing and replication fork progression (Boddy

p53-dependent S-phase DNA damage checkpoint T Shimura et al

5922

and Russell, 2001; Osborn et al., 2002). Suppression of late origin firing in response to DNA damage is dependent on Mec1 and Rad53 in yeast (Santocanale and Diffley, 1998; Shirahige et al., 1998), and on their respective homologues ATM and Chk2 in mammalian cells (Larner et al., 1999; Falck et al., 2001). Suppression of replication fork progression was demonstrated in a classic work of fiber autoradiography analysis of irradiated cells (Watanabe, 1974). As p53 was reported not to be required for the suppression of origin firing (Xie et al., 1998; Falck et al., 2001; Merrick et al., 2004), the p53-dependent S-phase DNA damage checkpoint is likely to act by suppressing replication fork progression. The replication fork consists of a template and a replication complex in which proliferating cell nuclear antigen (PCNA) plays a pivotal role. Proliferating cell nuclear antigen assembles as a trimer in S-phase with the help of replication factor-C (RF-C) and binds to DNA in an ATP-dependent manner. The function of PCNA is to recruit DNA polymerase d and e onto the replication site during S-phase, and to recruit repair factors and translesional DNA polymerases to the site of DNA damage (Tsurimoto, 1999; Stelter and Ulrich, 2003). Proliferating cell nuclear antigen is found in a soluble form in nuclei of quiescent cells and a detergentinsoluble trimeric form in S-phase cells (Bravo and Macdonald-Bravo, 1987). The latter form is in stable association with the replication fork and forms focus in cells treated with UV and alkylating agents (Miura et al., 1996; Savio et al., 1998). In this paper, we have studied the role of p53 in the S-phase DNA damage checkpoint in MEFs exposed to low-dose gamma-irradiation. Furthermore, we examined the damage response of PCNA to clarify the molecular mechanism of p53-dependent suppression of replication fork progression.

Results Low-dose-specific p53-dependent S-phase DNA damage checkpoint in mouse embryonic fibroblasts The p53-dependent S-phase DNA damage checkpoint was studied in MEFs (Figure 1a). A biphasic dose– response was noted for the suppression of DNA synthesis in wild-type (WT) MEFs. When analysed 4 h after irradiation, a sharp decrease in the rate of DNA synthesis was noted at doses below 2.5 Gy which was followed by a more moderate decrease at higher doses. As expected, radioresistant DNA synthesis (RDS) was evident in ATM/ MEFs where both the low-dose sharp decrease and the high-dose moderate decrease were absent. p53/ MEFs exhibited clear RDS at doses below 2.5 Gy, whereas its DNA synthesis was suppressed above 5 Gy. Interestingly, we repeatedly observed that the rate of DNA synthesis increased slightly at low dose(s) of irradiation in p53/ and ATM/ MEFs. This slight increase was consistent with that reported in cells from A–T patients (Falck et al., 2001). The low-dose-specific RDS in p53/ Oncogene

MEFs was not owing to abrogation of a p53/p21 regulated G1/S-phase checkpoint in the cells, as suppression of DNA synthesis in p21/ MEFs exhibited the same low-dose-specific sharp decrease as in WT MEFs in addition to the high-dose range moderate decrease (Figure 1a). The RDS analysis in Figure 1a by pulse labeling with 3 H-thymidine measures overall DNA synthesis of an asynchronous cell population. It is therefore desirable to test the effect of irradiation specifically on S-phase cells. We now performed sequential pulse labeling of the cells with two nucleotide analogs, 5-iodo-20 -deoxyuridine (IdU) and 5-chloro-20 -deoxyuridine (CldU). DNA labeled with these analogs can be differentially stained by two distinct antibodies to two colors; IdU as red (Cy3) and CldU as green (Alexa488). Cells were pulse labeled for 10 min with IdU before irradiation to specifically label the S-phase cells. They were then irradiated, incubated for 1 h and pulse labeled for 15 min with CldU to examine the effect of radiation on DNA synthesis (Figure 1b). The staining pattern of replication foci was reported to change during S-phase (Jackson and Pombo, 1998; Ma et al., 1998; Dimitrova and Gilbert, 2000; Leonhardt et al., 2000). Therefore, we focused specifically on cells with the homogeneous IdU staining characteristic to the early S-phase, to examine the changes in the distribution of replication foci 1 h after irradiation. In the unirradiated WT, p53/ and ATM/ MEFs, the staining pattern of CldU was homogeneous and overlapped with that of IdU. Upon 1, 2.5 and 5 Gy irradiations, however, the second CldU labeling took foci-like staining in WT MEFs and the uptake of CldU was clearly suppressed as judged from the decreased intensity of the green fluorescence. In contrast, p53/ MEFs exhibited the same homogeneous staining for CldU even after exposures to doses below 2.5 Gy, in addition to the unchanged intensity of the green fluorescence. The intensity in p53/ MEFs was clearly decreased when the dose was 5 Gy. The staining pattern and the intensity of CldU were unaffected in ATM/ MEFs by any doses of radiation in agreement with the RDS phenotype of ATM/ cells (Painter and Young, 1980). Thus, the p53-dependent S-phase DNA damage checkpoint found originally in sperm irradiated mouse zygotes appears to operate also in MEFs at low doses of gamma-irradiation. In order to determine whether p53 is activated during S phase after gamma-irradiation, double immunostaining for 5-bromo-20 -deoxyuridine (BrdU) and p53phosphoserine-15 (phospho-p53) was performed in WT and ATM/ MEFs. As expected, cells labeled with BrdU were almost completely devoid of phosphop53 in unirradiated WT and ATM/ MEFs. Within 1 h after 1 Gy irradiation, the phosphorylated form of p53 was detected in BrdU-positive WT MEFs (Figure 1c). This indicated that p53 was activated by a low dose of irradiation in S-phase cells. In contrast, p53 was not phosphorylated after 1 Gy irradiation in ATM/ MEFs (Figure 1c). This suggested that ATM is essential for activation of p53 after a low-dose irradiation in S-phase cells.


5923

Figure 1 Low-dose-specific p53-dependent S-phase checkpoint in mouse embryonic fibroblasts (MEFs). (a) Suppression of DNA synthesis after irradiation in MEFs of different genetic background. The rate of DNA synthesis was calculated from the ratio of 3 H-/14C-thymidine uptake in irradiated cells divided by the ratio of 3H-/14C-thymidine uptake in unirradiated control of wild-type (WT), p53/, p21/ and ATM/ MEFs 4 h after irradiation. Standard deviations are in parentheses. (b) DNA replication was visualized by the uptake of 5-iodo-20 -deoxyuridine (IdU) and 5-chloro-20 -deoxyuridine (CldU) in WT, p53/ and ATM/ MEFs. Cells were labeled with IdU (Cy3; red color) for 10 min, irradiated, further incubated for 1 h and labeled with CldU (Alexa488; green color) for 15 min. (c) Wild-type MEFs and ATM/ MEFs were labeled with 5-bromo-20 -deoxyuridine for 1 h before irradiation to identify cells in S phase and irradiated with 1 Gy. After further incubation for 1 h, cells were immunostained with antibodies to p53-phosphoserine-15 (phospho-p53) and against BrdU. Nuclei were stained with 4, 6-diamidino-2-phenylindole.

p53-dependent slowing down of replication fork progression after low-dose irradiation A DNA fiber assay was performed to study the mode of action of the p53-dependent S-phase DNA damage checkpoint (Merrick et al., 2004). This assay allows us to analyse the rate of replication fork progression and the initiation of DNA replication after irradiation. Cells were pulse labeled with IdU (red signal; R), irradiated, labeled with CldU (green signal; G) for 20 min, and then DNA fiber preparations were made (Figure 2a). The experiments revealed five patterns of fiber staining, R-G, G-R-G, R-G-R, R-only and G-only tracks (Figure 2a). The R-G track, a DNA fiber with

unidirectional red and green staining, corresponds to continuous elongation of replication fork during the first and the second labeling periods, respectively. Similarly, the G-R-G track is for the initiation of DNA replication in the first labeling period with continuation of bidirectional replication fork progression during the second labeling period. In these events, the rate of replication fork progression after irradiation is represented by the length of green fibers. The R-G-R track and R-only track correspond to the termination and termination and/or stalled fork progression, respectively. The G-only track is for a new initiation of replication in the second labeling period. Oncogene


5924

As shown in Figure 2b, irradiation suppressed the new origin firing (the frequency of the G-only track) to about a half the level at 1 and 5 Gy in WT and p53/

Oncogene

MEFs, whereas it was not in ATM/ MEFs. The G-only tracks can arise from the mechanical or radiation induced fragmentation of R-G, G-R-G and R-G-


5925

R tracks. Therefore, irradiation and subsequent labeling of the cells with IdU followed by CldU were tested. The new initiation as represented by G-R-G tracks in this procedure was also suppressed in WT and p53/ MEFs, suggesting the inhibition of initiation by irradiation (Supplementary Figure 1). Again, initiation of DNA replication was not inhibited in irradiated ATM/ MEFs. The increase in the frequency of the R-only track after gamma-irradiation indicated that replication fork stalled in WT and p53/ MEFs upon DNA damage (Figure 2c). The frequency of the R-only track is significantly higher in unirradiated p53/ MEFs than in the corresponding WT MEFs for yet unknown mechanism. Consequentially, radiation induced stalling of replication forks was relatively less in p53/ MEFs than in WT MEFs. In contrast, replication forks did not stall after irradiation of ATM/ MEFs. The majority of labeled DNA fibers (80%) in the WT, p53/ and ATM/ MEFs were those of the R–G tracks indicating that DNA replication continued even after irradiation (Figure 2d), although the rate of fork progression was reduced dose dependently as judged by the shorter length of green portion of the fibers (see Figure 2a for 1Gy and 5 Gy-irradiated WT MEFs, and 5 Gy-irradiated p53/ MEFs). We next examined the rate of fork progression by measuring the length of red and green portions in R-G tracks (ongoing replication forks). The average extension rate was 1.74 mm/min during both labeling periods in unirradiated WT, p53/ and ATM/ MEFs, whereas the rate dropped to 0.80 and 0.60 mm/min after 1 and 5 Gy of irradiation in WT MEFs (Figure 2e). This suppression did not occur in p53/ MEFs at low doses of 1 and 2.5 Gy (extension rate; 1.66 mm/min and 1.25 mm/min), but did take place at a higher dose of 5 Gy (extension rate; 0.72 mm/min). The extension rate of replication fork in ATM/ MEFs was unaffected by irradiation (Figure 2a, e). As for the validity of the fork progression rate analyses, one has to keep in mind that each of the replication forks may well differ in its speed of progression. In order to normalize for this difference, the ratio of the R-track length (the first labeling period) and the G-track length (the second labeling period) of R-G tracks was analysed following

the procedure by Henry-Mowatt et al. (2003). The distribution of the ratio is shown in Figure 2f. The results again demonstrate that the low-dose-specific p53-dependent S-phase checkpoint functions in suppression of replication fork progression after irradiation. ATM is essential for this slowing of replication fork in irradiated cells. p53-dependent focus formation of proliferating cell nuclear antigen after irradiation Proliferating cell nuclear antigen is a replication fork processivity protein at the ongoing replication fork during normal S-phase. In order to analyse the mechanism of p53-dependent suppression of replication fork progression, spatial distribution of PCNA was examined 1 h after irradiation. The cells in S-phase were pulse labeled with IdU for 10 min, irradiated, incubated for another 1 h and then fixed and stained for further analyses. As stated before, the staining pattern of replication foci in cells was reported to change during S-phase. Cells in early S-phase displayed numerous replication foci located in the internal, euchromatic region of the nucleus whereas cells in the late S-phase showed small numbers of relatively large foci in the heterochromatin region (Dimitrova and Gilbert, 2000; Leonhardt et al., 2000). Again, we focused only on cells with the homogeneous staining pattern of IdU characteristic to early S-phase to examine the redistribution of PCNA after irradiation. As shown in Figure 3a, IdU and PCNA signals were homogeneously distributed in nuclei of unirradiated WT, p53/ and ATM/ MEFs. Irradiation of WT MEFs with 1 and 5 Gy triggered redistribution of PCNA from the homogeneous staining to the focal staining (Figure 3a). To our surprise, PCNA disappeared in p53/ MEFs 1 h after 1 Gy irradiation whereas the large foci were clearly detected when the dose was as high as 5 Gy (Figure 3a). Disappearance of PCNA was also observed when the dose was 2.5 Gy (Figure 3b). Thus, the lack of suppression of DNA synthesis at low-dose irradiated p53/ MEFs paralleled with the absence of PCNA in the nucleus. Conversely, suppression of DNA synthesis paralleled with the focus like clustering of PCNA in low- and high-

Figure 2 p53-dependent suppression of replication fork progression after irradiation. DNA fiber assay was performed in wild-type (WT), p53/ and ATM / mouse embryonic fibroblasts (MEFs). Cells were labeled with 5-iodo-20 -deoxyuridine (IdU) (Cy3; red color) for 10 min, irradiated, and then labeled with 5-chloro-20 -deoxyuridine (CldU) (Alexa488; green color) for 20 min. DNA fiber staining was performed as described in the Materials and methods and the each experiment was repeated twice independently. The frequency of new origin firing, stalled replication forks and ongoing replication forks was estimated from the number of indicated tracks divided by total labeled tracks. The data presented in the panels b–d are the summary of the scoring of 100 tracks. Error bars represent the corresponding standard deviations; in some cases the standard deviations were too small to be visible on the histogram. (a) Images of labeled tracks are shown in control, 1- and 5 Gy-irradiated WT, p53/ and ATM/ MEFs. Typical red-green (R-G) tracks, red (R)-only tracks, green (G)-only tracks, green-red-green (G-R-G) tracks and red-green-red (R-G-R) tracks are indicated by arrow. (b) The frequency of new origin firing (G-only tracks) was studied in WT, p53/ and ATM/ MEFs. (c) The frequency of stalling of replication forks (R-only tracks) was studied after irradiation of WT, p53/ and ATM/ MEFs. (d) The frequency of ongoing replication forks (R-G tracks) was studied after WT, p53/ and ATM/ MEFs. (e) The length of fork extension were studies during first (red IdU, 10 min) and second (green CldU, 20 min) labeling period. The results of the analyses on 50 tracks are shown with the standard deviations in parentheses. (f) The ratio of the lengths of red and green portions of the R-G tracks was analysed. The measurements were done on 50 R-G tracks on WT, p53/ and ATM/ MEFs with and without irradiation (1, 2.5 and 5 Gy). Oncogene


5926

Figure 3 p53-dependent focus formation of proliferating cell nuclear antigen (PCNA) in S-phase after irradiation. Cells were labeled with 5-iodo-20 -deoxyuridine (IdU) for 10 min to identify S-phase cells and fixed for immunostaining for PCNA 1 h later after irradiation. Caffeine (3 mM) was added in medium 30 min before irradiation. 5-iodo-20 -deoxyuridine incorporation was detected by Cy3-conjugated antibodies (red) and PCNA distribution was detected by immunofluorescence with Alexa488-conjugated antibodies (green). (a) Staining pattern for IdU and PCNA are shown for wild-type (WT), p53/ and ATM/ mouse embryonic fibroblasts (MEFs) after irradiation with and without caffeine. (b) A quantitative analysis of the effect of irradiation on PCNA focus formation during S-phase in WT, p53/, p21/ and ATM/ MEFs. The staining pattern of PCNA was scored only in cells with that were in early S-phase at the time of irradiation as judged from the homogeneous IdU staining pattern. Each data point represents the distribution of 100 nuclei. The experiment was repeated three times with independent samples. Error bars represent the corresponding standard deviations; in some cases the standard deviations were too small to be visible on the histogram. (c) The staining pattern of PCNA, g-H2AX and ATM-phosphoserine-1981 in WT MEFs. Cells were immunostained with anti-PCNA (detected by Cy3, red in g-H2AX staining or Alexa488, green in anti-ATM-phosphoserine-1981 staining), g-H2AX (detected by Alexa488, green) and antiATM-phosphoserine-1981 (detected by Cy-3, red) after irradiation of WT MEFs.

dose irradiated WT MEFs, and high dose irradiated p53/ MEFs. In the case of ATM/ MEFs, the homogeneous staining pattern of PCNA was unchanged by either 1 or 5 Gy of irradiation. Figure 3b shows the dose responses of PCNA focus formation in IdUpositive MEFs of different genetic backgrounds. The observations can be summarized as follows: a dosedependent increase of the frequency of cells with PCNA foci for WT and p21/ MEFs; disappearance of foci for p53/ MEFs below 2.5 Gy whereas an increase in the frequency to the wild-type levels at 5 Gy; unchanged homogeneous staining in ATM/ MEFs even after Oncogene

5 Gy without any focal staining. In order to determine whether or not ATM regulates p53-dependent focus formation of PCNA at low dose of irradiation, caffeine, an inhibitor of ATM and ATR kinase, was added to WT and p53/ MEFs cultures before irradiation. As shown in Figure 3a, PCNA staining stayed homogeneous in caffeine-treated WT and p53/ MEFs after irradiation. These results suggest that ATM mediates the p53-dependent focus formation of PCNA after irradiation. When DNA is damaged, DNA repair proteins accumulate at DNA double strand beaks to form foci (DNA


5927

repair foci), which include PCNA (Karmakar et al., 2001; Solomon et al., 2004). The relationship between gamma-irradiation induced PCNA foci and DNA repair foci was studied by analyzing focus formation of gH2AX and phospho-ATM which accumulate at DNA double strand beaks (Bakkenist and Kastan, 2003; Pilch et al., 2003; Sedelnikova et al., 2003). Interestingly, PCNA foci did not co-localize with g-H2AX foci and phospho-ATM foci after low-dose irradiation of WT MEFs (Figure 3c). These results demonstrate that PCNA foci, the sites of the slowly moving replication fork, are distinct from DNA repair foci. Therefore, slowing down of replication fork movement and the resulting suppression of DNA synthesis are the consequence of cellular response to DNA damage, rather than the result of the mechanical block of replication fork movement at the sites of DNA double strand breaks. A very similar conclusion was reached in our previous study of mouse zygotes fertilized by irradiated sperm in which the suppression of DNA synthesis was observed even in the unirradiated female pronuclei (Shimura et al., 2002a). p53-dependent chromatin binding of proliferating cell nuclear antigen after low-dose irradiation Disappearance of PCNA in 1 Gy-irradiated p53/ MEFs could result from a decrease in the amount of PCNA proteins in the cells. However, the total amounts of PCNA protein in both WT and p53/ MEFs were unchanged by gamma-irradiation at doses up to 10 Gy (Figure 4a). The chromatin bound form of PCNA was therefore investigated using the method of Balajee and Geard (2001). The chromatin-bound PCNA in detergent-insoluble fraction was found to be less abundant in p53/ MEFs than in WT MEFs. The amount decreased dramatically after exposure to 1 Gy, but stayed unchanged when the dose was 5 Gy (Figure 4b). Disappearance of chromatin-bound PCNA in detergentinsoluble fraction of 1 Gy-irradiated p53/ MEFs was not owing to degradation of proteins, because PCNA was present in detergent-soluble nucleoplasmic fraction. Phosphorylation and nuclear focus formation of proliferating cell nuclear antigen after irradiation Phosphorylation was shown to be required for the binding of PCNA to chromatin in UV-irradiated cells (Prosperi et al., 1993). Therefore, 32P-labeling experiments were carried out to examine whether or not this was also the case for gamma-irradiated cells. A significant uptake of 32P was observed in 1 Gy-irradiated WT MEFs when the cells were incubated with 32P-orthophosphate (Figure 5a). In contrast, the uptake of 32P was not observed in 1 Gy-irradiated p53/ MEFs and 5 Gyirradiated ATM/ MEFs. Therefore, it is possible that phosphorylation is required for the chromatin binding of PCNA after gamma-irradiation. It is also possible that the lack of chromatin binding of PCNA in 1 Gyirradiated p53/ MEFs could be related to its absence of phosphorylation (Figure 4b). To test these possibilities, we utilized lambda-phosphatase to remove

Figure 4 p53-dependent chromatin binding of proliferating cell nuclear antigen (PCNA) after low-dose irradiation. (a) Whole cell lysates were prepared 1 h after irradiation. Immunoblot analyses were performed with antibodies against PCNA and b-tubulin in wild-type (WT) and p53/ mouse embryonic fibroblasts (MEFs). (b) Detergent-soluble and detergent-insoluble fractions of cell lysates were prepared 1 h after irradiation as described in Material and methods. Immunoblot analyses were performed with antibodies against PCNA and b-tubulin in WT and p53/ MEFs.

the phosphate residue(s) of PCNA. Treatment of the chromatin fraction of cell extract with lambda-phosphatase decreased the amount of chromatin-bound PCNA for 5 Gy-irradiated WT MEFs, but not for unirradiated WT MEFs (Figure 5b). In addition, immunostaining analysis revealed that PCNA foci were completely removed when 5 Gy-irradiated WT MEFs were treated with lambda-phosphatase whereas the same treatment did not affect other proteins such as histone H3 (Figure 5c and d). These results indicate that the phosphorylation on PCNA by ATM or ATM-dependent kinase is essential for the chromatin binding of PCNA after irradiation. In addition, phosphorylation of PCNA requires the functional p53 at the low-dose range such as 1 Gy. Activation of ATM in wild-type mouse embryonic fibroblasts and p53/ mouse embryonic fibroblasts Phosphorylation on PCNA was defective in p53/ MEFs after 1 Gy irradiation and this could be owing to the lack of, or weak activation of ATM in p53/ MEFs when DNA damage is limited. Therefore, activation of ATM was investigated by analyzing the active form of posphorylated ATM. Anti-ATM-phosphoserine-1981 antibody detected phospho-ATM foci at 1 h after irradiation in PCNA-positive cells and the number of these foci increased dose dependently both in WT and p53/ MEFs (Figure 6). These results suggest that activation of ATM is normal in p53/ MEFs Oncogene


5928

Figure 5 p53 assisted ATM-mediated phosphorylation of proliferating cell nuclear antigen (PCNA) after low-dose gammairradiation. (a) Phosphorylation of PCNA after irradiation. Cells were incubated with 0.2 mCi/ml 32P-orthophosphate for 1 h, irradiated and further incubated for another 1 h in the presence of 32P-orthophosphate. The chromatin-bound fraction was obtained from wildtype (WT), p53/ and ATM/ mouse embryonic fibroblasts (MEFs), immunoprecipitated with anti-PCNA antibody and evaluated by autoradiography. (b) Phosphatase treatment of the chromatin fraction from irradiated WT MEFs. The chromatin-bound fraction of cell lysates from WT MEFs was incubated with lambda-phosphatase (1.6 m/ml) and then immunobloted with antibodies against PCNA and b-tubulin. (c) Proliferating cell nuclear antigen staining of irradiated WT MEFs after phosphatase treatment. The cells on coverslips were treated with a hypotonic lysis solution. They were then treated with lambda-phosphatase (l-ppase þ in the lower panel) and immunostained for PCNA (detected by Cy-3, red) and acetyl-histone H3 (detected by Alexa488, green). (d) Cells with PCNA foci were scored and the frequencies were calculated by dividing the number of cells with PCNA foci with that of total cells. The frequency decreased after treatment with lambda-phosphatase.

even though the cells lack ATM mediated phosphorylation of PCNA after irradiation. Discussion Low-dose-specific p53-dependent S-phase DNA damage checkpoint In the present study, the ATM-dependent S-phase DNA damage checkpoint was shown to consist of low-dose and high-dose components. The low-dose component operates at doses below 2.5 Gy and is dependent on the functional p53. p53 was reported not to be required in the S-phase DNA damage checkpoint (Xie et al., 1998; Falck et al., 2001; Merrick et al., 2004). However, these studies used gamma-ray doses higher than 5 Gy, which are supra-lethal to the cells. In fact, the use of high doses was a common practice from the time of the discovery of radiation induced suppression of DNA synthesis in 1974 (Watanabe, 1974) and the use was continued even after the discovery of radioresistant DNA synthesis in AT cells (Painter and Young, 1980). The doses below 5 Gy was not in general, used for the analysis of suppression of DNA synthesis, despite 1 Gy was shown to suppress DNA synthesis in normal human fibroblasts Oncogene

(Brown et al., 2003). (Laderoute 1996) reported that p53 was not essential for the suppression of DNA synthesis at 2 Gy in HL-60 cells expressing mutant p53. However, one cannot draw such conclusion on the role of p53, as ATM protein, a key factor in S-phase DNA damage checkpoint, was undetectable in HL60 (Gately et al., 1998). Low dose specificity of the p53-dependent S-phase DNA damage checkpoint was found previously in p53 þ / þ and p53/ mouse zygotes fertilized with irradiated sperm (Shimura et al., 2002a). Now the same checkpoint is demonstrated in mouse fibroblasts. Thus, this p53 dependency is supposed to be a general trait of the S-phase DNA damage checkpoint in mammalian cells exposed with low doses of radiation. Low-dose-specific suppression of replication fork progression by p53 The effect of ionizing radiation on DNA synthesis in eukaryotic cells was well documented (Rowley et al., 1999). When the size distribution of 3H-labeled DNA was monitored in alkaline sucrose gradients, small DNA fragments of the nascent strands of newly initiated replication were reduced whereas large fragments of the progressing replication forks stayed constant after


5929

found radiation-induced suppression of chain elongation in p53 proficient E-11 normal human diploid cells whereas the same authors found no such suppression in p53 mutated cell line as discussed above (Painter and Young, 1975). Our present study clearly demonstrated that p53 is required in suppression of replication fork progression after 1 Gy irradiation of MEFs and this is the true checkpoint response rather than the mechanical interference of replication fork movement by the damage on template strand of DNA.

Figure 6 Activation of ATM in S phase wild-type (WT) and p53/ mouse embryonic fibroblasts (MEFs) after irradiation. Staining patterns for proliferating cell nuclear antigen (PCNA) and ATM-phosphoserine-1981. Cells were immunostained with antiPCNA (detected by Alexa488, green) and anti-ATM-phosphoserine-1981 antibodies (detected by Cy-3, red) at 1 h after irradiation in WT and p53/ MEFs. Phospho-ATM foci were evident after irradiation during S phase (PCNA-positive cells) of WT and p53/ MEFs.

moderate radiation doses of 1–5 Gy in CHO and Hela cells (Painter and Young, 1975; ). It was also found that ionizing radiation induced a transient block of origin firing but did not affect progression of replication fork in Hela cells (Merrick et al., 2004). Suppression of replication fork progression was shown to occur only at extremely high doses in mouse leukemic L5178Y (Watanabe, 1974; Makino and Okada, 1975). These reports concluded that moderate doses of ionizing radiation inhibit initiation of DNA synthesis rather than suppression of replication fork progression. However, all of the cell lines used in these studies were either p53 deficient or p53 mutated (Storer et al., 1997; Hu et al., 1999; Bohnke et al., 2004). Therefore, the lowdose specificity and p53 dependency of the suppression of replication fork progression may well have been overlooked. In fact, 1.5 Gy gamma-radiation was reported to induce suppression of replication fork progression in p53 proficient normal human fibroblasts (Heffernan et al., 2002). Furthermore, (Painter and Young, 1980)

p53 dependence of ATM-mediated phosphorylation of proliferating cell nuclear antigen In the present analysis, we have discovered a novel role of p53 in the S-phase DNA damage checkpoint. The data suggest that ATM or ATM-dependent kinase phosphorylates PCNA in the p53 proficient cells. Phosphorylated PCNA binds to chromatin and forms foci in irradiated cells. This PCNA focus formation was absolutely dependent on ATM, and was associated with the suppression of replication fork progression. The phosphorylation and focus formation of PCNA, and the suppression of DNA synthesis were p53-dependent at doses below 2.5 Gy, but not at higher doses, suggesting the damage level dependence of the S-phase DNA damage checkpoint. In fact, the S-phase DNA damage checkpoint was reported to require a threshold level of damage to activate the checkpoint response (Shimada et al., 2002). Full activation of Chk2, an effector of the S-phase DNA damage checkpoint, was reported to require radiation doses above 2 Gy (Buscemi et al., 2004), indicating the presence of a threshold for the induction of the whole battery of S-phase DNA damage responses. Our data suggested that 1 Gy of gammairradiation is enough to activate p53 to execute the p53dependent suppression of replication fork progression in MEFs through phosphorylation of PCNA. Two of the targets of p53, Gadd45 and p21, are known to interact with PCNA (Smith et al., 1994; Waga et al., 1994), but the significance of the Gadd45/PCNA interaction is not at all clear (Kearsey et al., 1995). p21 cannot be the mediator of the p53-dependent focus formation of PCNA and the subsequent S-phase DNA damage checkpoint because these can occur in cells deficient in p21 as shown by the data of Figure 1a and Figure 3c. In addition, p53 mutants that lack the transcriptional activation function are nonetheless capable of activating the S-phase DNA damage checkpoint in mouse zygotes (Toyoshima et al., 2005). These findings indicate that the protein interaction function of p53, but not the transactivation function, is integral to the prosecution of the low-dose-specific p53-dependent S-phase DNA damage checkpoint. A novel role of proliferating cell nuclear antigen and its phosphorylation in regulating the replication fork progression Proliferating cell nuclear antigen redistributed from the homogeneous staining to the focal staining in irradiated Oncogene


5930

WT MEFs. However, PCNA dissociated from chromatin in low dose irradiated p53/ MEFs whereas the PCNA staining pattern was unchanged in ATM/ MEFs at any dose. Furthermore, caffeine, an inhibitor of ATM and ATR, abrogated disassembly of PCNA from chromatin after 1 Gy irradiation of p53/ MEFs. These suggest that focus formation and chromatin binding of PCNA at low radiation doses required both ATM and p53. ATM is known to be essential in the S-phase DNA damage checkpoint (Painter and Young, 1980) and radiation is well known to suppress replication fork progression for many years (Watanabe, 1974). However the precise mechanism of downregulating the replication fork movement has not been clarified. Based on the present study, a working hypothesis can be formulated on the molecular mechanism of the p53-dependent S-phase DNA damage checkpoint which is summarized as follows. After low-dose irradiation, ATM or ATM-dependent kinase is activated to phospholyrate various target proteins. Proliferating cell nuclear antigen is also phosphorylated but only with the aid of p53 and the modified PCNA then binds to the replication complex on the chromatin. The complex with phosphorylated PCNA moves more slowly, allowing time to repair the damage, thus reducing the risk of colliding onto the damage along the template strand. In the absence of p53, ATM-mediated phosphorylation of PCNA does not take place even though low doses of gamma-irradiation activate ATM. Non-phosphorylated PCNA in the presence of activated ATM dissociates from the chromatin, leading to DNA synthesizes without PCNA in p53/ MEFs. Under this condition, replication of damaged DNA templates proceeds at a normal rate which is likely to results in a higher rate of chromosome aberrations. Indeed, we have reported previously the induction of numerous chromosome aberrations in p53/ mouse zygotes fertilized with irradiated sperm (Shimura et al., 2002b). In the absence of ATM, PCNA binds with chromatin without ATM mediated phosphorylation of PCNA, however PCNA dose not form foci after irradiation as in WT MEFs. Replication forks continue progression without recognition of DNA damage after irradiation in ATM/ MEFs. Therefore, PCNA did not change the localization before and after irradiation in ATM/ MEFs. In our present study, a new role of PCNA in regulating replication fork progression is excavated where p53 assisted phosphorylation of PCNA by ATM is essential. Whether or not this new pathway of the S-phase DNA damage checkpoints regulates other cellular activities such as resolution of and recombination at the collapsed replication forks remain to be elucidated (Janz and Wiesmuller, 2002). Materials and methods Cells and culture conditions Cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM glutamine, and 100 U/ml penicillin/streptomycin (GIBCO, Grand Island, NY, USA). Primary MEFs (WT, p53/, Oncogene

ATM/, p21/) were derived from 13.5-day-old mouse embryos of the ICR strain. Following removal of the head and internal organs, each embryo was rinsed with phosphatebuffered saline (PBS), minced with scissors and cultured onto 60 mm tissue culture dishes. ATM/ MEFs were kindly provided by Proffessor Komatsu (Radiation Biology Center, Kyoto University) and were those immortalized by the human TERT-gene. The cultures were maintained at 371C in a humidified 5% CO2 atmosphere. Irradiation Gamma-irradiation was performed at room temperature using a 137Cs source at a dose rate of 1.25 Gy/min (Gammacell 40 exactor, MDS Nordion, Ottawa, Canada). After irradiation, the medium was replaced and cells were incubated for further analyses. Radio-resistant DNA synthesis assay The rate of DNA synthesis was measured by the 14C-/3Hthymidine double labeling method. Briefly, cells were prelabeled with 0.37 kBq/ml 14C-thymidine for 2 days, irradiated and further incubated for 4 h. Cells were then pulse labeled with 3H-thymidine at 37 kBq/ml for 15 min. Cells were washed with PBS(), trypsinized, trapped onto glass microfiber filters (Whatman GF/C), rinsed twice with an ice-cold 10% trichloroacetic acid, then with ice-cold ethanol and acetone, and air-dried. The radioactivity of each sample was quantified by a liquid scintillation counter (Packard TriCarb 2200CA) and the 3H/4C ratio was taken as the rate of DNA synthesis. DNA fiber analysis Cells were labeled with 20 mM IdU for 10 min and then labeled with 100 mM CldU for 20 min. Cells were trypsinized and resuspended in PBS at 1 106 cells/ml. 2.5 ml of cells were mixed with 7.5 ml of lysis buffer (0.5% SDS in 200 mM TrisHCl, pH 7.4, 50 mM ethylene diaminetetraacetic acid) on a glass slide. After 8 min, DNA spreads were fixed in 3:1 methanol/acetic acid, and stored in 70% ethanol at 41C. 5-chloro-20 -deoxyuridine and IdU staining was performed according to the protocol described by others (Dimitrova and Gilbert, 1999). Immunofluorescence Cells were seeded onto glass coverslips placed in 10 mm tissue culture dishes. After irradiation with gamma-ray, cells were incubated for indicated times. The coverslips were treated with a hypotonic lysis solution (10 mM Tris-HCl pH 7.4, 2.5 mM MgCl2, 1 mM phenylmethylsulfonyl fluoride and 0.5% Nonidet P-40) for 8 min on ice. This step was omitted for immunostaining of p53-phosphoserin-15. The coverslips were washed twice with PBS, fixed with ice-cold acetone (5 min) and then with ice-cold methanol (5 min). After rehydration with PBS, the cells were permeabilized and blocked for 30 min at room temperature in 1% bovine serum albumin (BSA) and 0.1% Triton X-100 in PBS. Anti-PCNA antibody (Mouse IgG, PC10; Oncogen, Darmstadt, Germany) (Rabbit IgG, SC-7007; Santa Cruz, U.S.A.), anti-p53-phosphoserine-15 (Rabbit IgG, SC11764-R; Santa Cruz, U.S.A.), anti-acetylhistone H3 (upstate), anti-ATM-phosphoserine 1981 (Rockland, Gilbertsville, U.S.A.) and anti-g-H2AX (Trevigen, Gaithersburg, U.S.A.) were diluted in PBS with 0.5% BSA and incubated with the coverslips for 1 h at 371C. The coverslips were then washed three times with 0.1% Triton X-100 in 0.1 PBS, incubated for 1 h at 371C with secondary antibodies conjugated with fluorescein isothiocyanate (fluorescein isothiocyanate, Santa Cruz for rabbit IgG), Alex 488


5931 (Molecular Probes, Carlsbad, U.S.A., for mouse IgG) or Cy-3 (Jackson Immuno Research Laboratories, Inc., West Grove, U.S.A., for mouse IgG). The coverslips were washed three times with 0.1% Triton X-100 in 0.1 PBS, counterstained for DNA with 4, 6-diamidino-2-phenylindole (DAPI) (4 mg/ml prepared in Vectashield mouting medium; Vector laboratories, Burlingame, U.S.A.). In the case of double immunostaining for p53-phosphoserin15/BrdU and PCNA/IdU, cells were incubated with 10 mM BrdU or IdU before irradiation. After immunostaining with anti-PCNA antibody (Santa Cruz), samples were fixed with 4% paraformaldehyde for 5 min at room temperature, washed with PBS, and then DNA was denatured in 2N HCl, 0.5% Triton X-100, for 10 min at 371C. After washing with PBS, the coverslips were incubated with anti-BrdU antibody (Becton Dickinson, Franklin Lakes, U.S.A.) diluted 1:100 in PBS with 0.5% BSA for 1 h at 371C. The coverslips were washed with 0.1% Triton X-100 in 0.1 PBS and incubated for 1 h at 371C with secondary antibody conjugated with Cy-3. After washing three times with 0.1% Triton X-100 in 0.1 PBS, the coverslips were stained for DAPI. Western blot analysis Soluble and insoluble proteins were prepared using the protocol of (Balajee and Geard, 2001). Proteins were separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (PAGE) gel and transferred electrophoretically to a nitrocellulose membrane (Hybond-ECL, Amersham Pharmacia Biotech, Piscataway, U.S.A.). The membrane was blocked with 5% non-fat milk. It was thereafter incubated with mouse anti-PCNA antibody (Mouse IgG, PC10; Oncogene, Darmstadt, Germany) or mouse anti-b-Tubulin antibody (SigmaAldrich, St. Louis, U.S.A.) for 1 h at room temperature, then

with peroxidase-conjugated anti-mouse IgG antibody (Santa Cruz), and developed with the ECL Western blotting detection system (Amersham Pharmacia Biotech). Phosphorylation studies Cells were plated in two 100 mm tissue culture dishes and were washed with phosphate-free DMEM (GIBCO). Plates were incubated with phosphate-free DMEM (GIBCO) containing 10% FCS and 0.2 mCi/ml 32P-orthophosphate (Amersham) for 1 h. Cells were irradiated and reincubated in the presence of 0.2 mCi/ml 32P-orthophosphate for an additional 1 h and then washed with ice-cold PBS. The cell extracts of chromatinbound fraction were prepared by the methods described above. Extracts were centrifuged for 10 min at 10 000 g. Total protein from the supernatant was incubated for 1 h at 41C with 15 ml of protein A/G agarose beads (Santa Cruz), after which the material was centrifuged. The supernatant was incubated for 3 h at 41C with anti-PCNA antibody (Santa Cruz) and 30 ml of protein A/G agarose beads, after which, the beads were collected and washed three times with binding buffer (50 mM Tris-HCl pH 8.0, 500 mM NaCl, 0.5% deoxycholate, 0.1% SDS and 1% Nonidet P-40). After immunoprecipitation, PCNA was analysed by SDS–PAGE and autoradiography. Acknowledgements We thank Drs Donald MacPhee, Martin Levin, Yosef Shiloh and Mirit I Aladjem for critical reading of the manuscript. This work was supported by a Grant-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan and by a grant from Nuclear Safety Research Association.

References Agarwal ML, Agarwal A, Taylor WR, Chernova O, Sharma Y, Stark GR. (1998). Proc Natl Acad Sci USA 95: 14775–14780. Bakkenist CJ, Kastan MB. (2003). Nature 421: 499–506. Balajee AS, Geard CR. (2001). Nucleic Acids Res 29: 1341–1351. Boddy MN, Russell P. (2001). Curr Biol 11: R953–R956. Bohnke A, Westphal F, Schmidt A, El-Awady RA, Dahm-Daphi J. (2004). Int J Radiat Biol 80: 53–63. Bravo R, Macdonald-Bravo H. (1987). J Cell Biol 105: 1549–1554. Brown KD, Rathi A, Kamath R, Beardsley DI, Zhan Q, Mannino JL et al. (2003). Nat Genet 33: 80–84. Buscemi G, Perego P, Carenini N, Nakanishi M, Chessa L, Chen J et al. (2004). Oncogene 23: 7691–7700. Dimitrova DS, Gilbert DM. (1999). Mol Cell 4: 983–993. Dimitrova DS, Gilbert DM. (2000). Nat Cell Biol 2: 686–694. Elkind MM, Sutton H. (1959). Nature 184: 1293–1295. Falck J, Mailand N, Syljuasen RG, Bartek J, Lukas J. (2001). Nature 410: 842–847. Fei P, El-Deiry WS. (2003). Oncogene 22: 5774–5783. Gately DP, Hittle JC, Chan GK, Yen TJ. (1998). Mol Biol Cell 9: 2361–2374. Gottifredi V, Shieh S, Taya Y, Prives C. (2001). Proc Natl Acad Sci USA 98: 1036–1041. Heffernan TP, Simpson DA, Frank AR, Heinloth AN, Paules RS, Cordeiro-Stone M et al. (2002). Mol Cell Biol 22: 8552–8561. Henry-Mowatt J, Jackson D, Masson JY, Johnson PA, Clements PM, Benson FE et al. (2003). Mol Cell 11: 1109–1117.

Hu T, Miller CM, Ridder GM, Aardema MJ. (1999). Mutat Res 426: 51–62. Jackson DA, Pombo A. (1998). J Cell Biol 140: 1285–1295. Janz C, Wiesmuller L. (2002). Oncogene 21: 5929–5933. Karmakar P, Balajee AS, Natarajan AT. (2001). Mutagenesis 16: 225–232. Kearsey JM, Shivji MK, Hall PA, Wood RD. (1995). Science 270: 1004–1005; author reply 1005–1006. Kim JM, Nakao K, Nakamura K, Saito I, Katsuki M, Arai K et al. (2002). EMBO J 21: 2168–2179. Laderoute MP. (1996). Anticancer Res 16: 2825–2830. Larner JM, Lee H, Little RD, Dijkwel PA, Schildkraut CL, Hamlin JL. (1999). Nucleic Acids Res 27: 803–809. Leonhardt H, Rahn HP, Weinzierl P, Sporbert A, Cremer T, Zink D et al. (2000). J Cell Biol 149: 271–280. Levine AJ. (1997). Cell 88: 323–331. Ma H, Samarabandu J, Devdhar RS, Acharya R, Cheng PC, Meng C et al. (1998). J Cell Biol 143: 1415–1425. Makino F, Okada S. (1975). Radiat Res 62: 37–51. Merrick CJ, Jackson D, Diffley JF. (2004). J Biol Chem 279: 20067–20075. Miura M, Sasaki T, Takasaki Y. (1996). Radiat Res 145: 75–80. Nayak BK, Das GM. (2002). Oncogene 21: 7226–7229. Osborn AJ, Elledge SJ, Zou L. (2002). Trends Cell Biol 12: 509–516. Painter RB, Young BR. (1975). Radiat Res 64: 648–656. Painter RB, Young BR. (1976). Biochim Biophys Acta 418: 146–153. Painter RB, Young BR. (1980). Proc Natl Acad Sci USA 77: 7315–7317. Oncogene


5932 Pilch DR, Sedelnikova OA, Redon C, Celeste A, Nussenzweig A, Bonner WM. (2003). Biochem Cell Biol 81: 123–129. Prosperi E, Stivala LA, Sala E, Scovassi AI, Bianchi L. (1993). Exp Cell Res 205: 320–325. Rowley R, Phillips EN, Schroeder AL. (1999). Int J Radiat Biol 75: 267–283. Santocanale C, Diffley JF. (1998). Nature 395: 615–618. Savio M, Stivala LA, Bianchi L, Vannini V, Prosperi E. (1998). Carcinogenesis 19: 591–596. Sedelnikova OA, Pilch DR, Redon C, Bonner WM. (2003). Cancer Biol Ther 2: 233–235. Sengupta S, Linke SP, Pedeux R, Yang Q, Farnsworth J, Garfield SH et al. (2003). EMBO J 22: 1210–1222. Sengupta S, Robles AI, Linke SP, Sinogeeva NI, Zhang R, Pedeux R et al. (2004). J Cell Biol 166: 801–813. Shimada K, Pasero P, Gasser SM. (2002). Genes Dev 16: 3236–3252. Shimura T, Inoue M, Taga M, Shiraishi K, Uematsu N, Takei N et al. (2002a). Mol Cell Biol 22: 2220–2228.

Shimura T, Toyoshima M, Taga M, Shiraishi K, Uematsu N, Inoue M et al. (2002b). Radiat Res 158: 735–742. Shirahige K, Hori Y, Shiraishi K, Yamashita M, Takahashi K, Obuse C et al. (1998). Nature 395: 618–621. Smith ML, Chen IT, Zhan Q, Bae I, Chen CY, Gilmer TM et al. (1994). Science 266: 1376–1380. Solomon DA, Cardoso MC, Knudsen ES. (2004). J Cell Biol 166: 455–463. Stelter P, Ulrich HD. (2003). Nature 425: 188–191. Storer RD, Kraynak AR, McKelvey TW, Elia MC, Goodrow TL, DeLuca JG. (1997). Mutat Res 373: 157–165. Toyoshima M, Shimura T, Adiga SK, Taga M, Shiraishi K, Inoue M et al. (2005). Oncogene 24: 3229–3235. Tsurimoto T. (1999). Front Biosci 4: D849–D858. Waga S, Hannon GJ, Beach D, Stillman B. (1994). Nature 369: 574–578. Watanabe I. (1974). Radiat Res 58: 541–556. Xie G, Habbersett RC, Jia Y, Peterson SR, Lehnert BE, Bradbury EM et al. (1998). Oncogene 16: 721–736.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene