Induction of a Caffeine-Sensitive S-Phase Cell Cycle Checkpoint by Psoralen Plus Ultraviolet a Radiation

Home , S phase

Oncogene (2003) 22, 6119–6128 & 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc ORIGINAL PAPERS Induction of a caffeine-sensitive S-phase cell cycle checkpoint by psoralen plus ultraviolet A radiation

Christoph Joerges1, Inge Kuntze1 and Thomas Herzinge*,1

1Klinik und Poliklinik fu¨r Dermatologie und Allergologie, Ludwig-Maximilians-University, Frauenlobstr. 9-11, D-80337 Munich, Germany

Induction of interstrand crosslinks (ICLs) in chromosomal vitro and in vivo (reviewed in Luftl et al., 1998). ICLs are DNA is considered a major reason for the antiproliferative considered to be the main reason for the antiprolifera- effect of psoralen plus ultraviolet A (PUVA). It is unclear tive effects of PUVA. However, the exact molecular as to whether PUVA-induced cell cycle arrest is caused by mechanism by which ICLs inhibit cell growth is not ICLs mechanically stalling replication forks or by known. triggering cell cycle checkpoints. Cell cycle checkpoints Cell cycle checkpoints serve to maintain genomic serve to maintain genomic stability by halting cell cycle stability by halting cell cycle progression in response to progression to prevent replication of damaged DNA incomplete replication, incorrect chromosome segrega- templates or segregation of broken chromosomes. Here, tion or chromosomal damage (Nurse, 1997; Weinert, we show that HaCaT keratinocytes treated with PUVA 1998). In mammalian cells, upstream signaling elements arrest with S-phase DNA content. Cells that had in these responses are the phosphatidylinositol 3-kinase completed DNA replication were not perturbed by PUVA family members ataxia telangiectasia mutated (ATM) and passed through mitosis. Cells treated with PUVA and ATM- and Rad3-related protein (ATR) (Bentley during G1-phase continued traversing G1 until arresting in et al., 1996; Cimprich et al., 1996). ATM and ATR early S-phase. PUVA induced rapid phosphorylation of phosphorylate and thereby activate the Chk1 and Chk2/ the Chk1 checkpoint kinase at Ser345 and a concomitant Cds1 checkpoint kinases (Matsuoka et al., 1998; Liu decrease in Cdc25A levels. Chk1 phosphorylation, de- et al., 2000), which in turn phosphorylate key promoters crease of Cdc25 A levels and S-phase arrest were of the core cell cycle machinery, the Cdc25 dual- abolished by caffeine, demonstrating that active check- specificity phosphatases. There are three Cdc25 proteins point signaling rather than passive mechanical blockage in mammalian cells. While Cdc25C and Cdc25B by ICLs causes the PUVA-induced replication arrest. promote progression from G2-phase into mitosis (Sadhu Overexpression of Cdc25A only partially overrode the et al., 1990; Galaktionov and Beach, 1991; Sebastian S-phase arrest, suggesting that additional signaling events et al., 1993), Cdc25A regulates entry into S-phase implement PUVA-induced S-phase arrest. (Hoffmann et al., 1994; Jinno et al., 1994). Phosphor- Oncogene (2003) 22, 6119–6128. doi:10.1038/sj.onc.1206613 ylation by Chk1 or Chk2/Cds1 renders Cdc25 phosphatases inactive by inhibiting their phosphatase activity Keywords: cell cycle; keratinocytes; psoralens; replica- (Blasina et al., 1999a; Furnari et al., 1999), sequestration tion checkpoint; ultraviolet radiation in the cytoplasm (Peng et al., 1997; Sanchez et al., 1997) or targeted degradation via the proteasome (Mailand et al., 2000; Zhao et al., 2000). Here, we have investigated the effect of PUVA on cell cycle progression in HaCaT keratinocytes. We present data to suggest that Introduction PUVA arrests HaCaT cells exclusively during DNA replication, and that this arrest is not caused by passive Upon activation by ultraviolet A light (UVA), psoralens stalling of replication forks by ICLs but rather by active form mono- or bifunctional adducts with pyrimidine checkpoint signaling involving Chk1 and Cdc25A. bases. Bifunctional adducts result in covalent cross- linking of opposite strands of DNA (interstrand crosslinks ICLs) (Gasparro, 1988). In medicine, the ¼ Results combination of psoralens and ultraviolet A irradiation (PUVA) is used for the treatment of several hyperpro- PUVA arrests cells at various stages of S-phase liferative skin diseases, most notably psoriasis. PUVA inhibits the proliferation of various cell types both in To test for the effect of PUVA on the cell cycle, HaCaT keratinocytes were pulse labeled with bromo-deoxy- uridine (BrdU) prior to PUVA treatment. Doses were *Correspondence: T Herzinger; chosen in a range equaling that used in vivo during E-mail: [email protected] PUVA therapy. If not indicated otherwise, the concen- PUVA-induced S-phase checkpoint C Joerges et al 6120 tration of 8-methoxy-psoralen (MOP) was constant in less than twofold of that of G1 cells. Most of these near- all experiments (0.1 mg/ml) and corresponded to the 4n cells were BrdU positive while the fraction of BrdU MOP serum level found in patients after ingestion of negative cells was significantly lower than in untreated MOP (Ljunggren et al., 1980). The dose of UVA was cells. In response to increasing doses of PUVA, cells increased from 0.05 to 1.0 J/cm2, the latter representing a arrested at successively earlier stages of S-phase (Figure dose applied early in the course of PUVA therapy. At 1c-f). After 1.0 J/cm2 (Figure 1e), the cell cycle profile 12 h following treatment with a low dose (0.05 J/cm2), was reminiscent of that of asynchronous cells harvested cells seemed to accumulate in the G2/M peak of the PI immediately after a BrdU pulse (Figure 1(g), except that histogram (Figure 1b). However, mathematical model- G2 cells were absent after PUVA. This indicated that ing using ModFit cell cycle analysis software revealed most of the cells had not progressed in the cell cycle that most cells in that peak had a DNA content slightly during the 12 h following PUVA treatment (Figure 1e).

Figure 1 PUVA arrests cells in S-phase. HaCaT cells were incubated with MOP and BrdU for 20 min prior to irradiation with the indicated doses of UVA and MOP. After 12 h, cells were ﬁxed and stained for FACS analysis with FITC-anti-BrdU-antibody (FL-1) and PI (FL-2). Note that BrdU-positive cells do not represent cells that incorporate BrdU by the time of harvest, but rather are cells that were in S-phase by the time of PUVA treatment (a–f). The last panel (g) shows asynchronous cells that were ﬁxed immediately after a 20 min BrdU pulse. Note that in this panel BrdU-positive cells are cells that are in S-phase by the time of harvest

Oncogene PUVA-induced S-phase checkpoint C Joerges et al 6121 To further increase the PUVA dose, the concentration be disturbed by PUVA-induced damage, and can pass of MOP was increased rather than the UVA dose, since freely into the subsequent G1-phase. UVA doses higher than 1.0 J/cm2 caused cell cycle effects by themselves (not shown) (Figure 1f). Still. G2/M cells Cells are not arrested in G2 by PUVA were absent after PUVA. Cell cycle effects of PUVA were identical irrespective of whether cells were pulsed To determine whether cells that are in G2 by the time of with BrdU or not (data not shown). The response to PUVA treatment can enter mitosis, cells were blocked in PUVA clearly differed from the one observed after other metaphase with nocodazole and the number of mitotic cells DNA damaging agents such as ultraviolet or ionizing was counted at various time points after PUVA. During radiation (Orren et al., 1995; Herzinger et al., 1995; the ﬁrst 4 h following PUVA, cells were entering mitosis at Tobey, 1975). There were two possible explanations for the same rate as mock-treated controls, indicating that G2 this observation: PUVA-induced damage could be cells are not arrested at the G2/M border by PUVA detrimental to cells in G2-phase, thus eliminating these (Figure 2A). After 6 and 8 h, no more cells entered mitosis, cells from our analysis. Alternatively, G2 cells might not suggesting that all cells had left G2 by that time.

Figure 2 Cells are not blocked at G2/M by PUVA. (a) After PUVA or mock treatment, cells were arrested in metaphase by nocodazole. At the indicated time points, mitotic ﬁgures were counted. (b) Schematic diagram of the pulse and pause procedure employed in the experiment shown in panel C. (c) Pulse and pauses analysis of G2/M progression after PUVA: HaCaT cells were pulse labeled with BrdU and treated with PUVA (0.1 mg/ml MOP, 1.0 J/cm2 UVA) or UVB (10 mJ/cm2) after an interval of 1 or 2 h. Cells were processed for FACS analysis 12 h after irradiation

Oncogene PUVA-induced S-phase checkpoint C Joerges et al 6122 To specifically trace cells that were in G2 during rather than replication, cells were arrested in G1 with PUVA treatment, cells were pulsed with BrdU and the Cdk-inhibitor roscovitine after PUVA. In the PUVA-treated after a pause of 1 or 2 h. With increasing absence of roscovitine, cells incorporated BrdU during the interval between the BrdU pulse and PUVA the 16 h following PUVA but did not show any increase treatment, an ever larger part of BrdU-positive cells in DNA content (from hours 8 to 24, Figure 3c, panel 3). can be expected to represent cells that had completed When treated with roscovitine after PUVA, cells did not DNA replication and were already in G2 by the time of incorporate BrdU (Figure 3c, panel 5), demonstrating PUVA treatment (Figure 2B). Cells were harvested and that BrdU incorporation does not reflect repair DNA processed for FACS analysis 12 h after PUVA or mock synthesis. treatment. By that time, most of the mock-treated BrdU-labeled cells had entered the subsequent G1, and some were already beginning to enter S-phase again PUVA-induced replication block is mediated (Figure 2C e, f, open arrows). Most of the PUVA- by a caffeine-sensitive checkpoint signal treated cells were still trapped in S-phase (Figure 2C a, The S-phase arrest induced by PUVA might be caused b). However, a small proportion of BrdU-positive cells by a checkpoint signal sensing DNA damage or had entered G1 (Figure 2C a). This population was not incomplete replication. Alternatively, it can be explained found in cells that were irradiated immediately after as consequence of a passive mechanical block of BrdU labeling (compare Figure 1e), suggesting that this replication forks by ICLs. In the latter case, agents represents cells that had already entered G2 by the time interfering with checkpoint signaling should not have an of PUVA treatment. However, other than the mock- effect on the S-phase arrest. Caffeine is known to treated controls, these cells did not enter S-phase. Upon override replication or DNA damage checkpoints, likely doubling the interval between pulse labeling and by inhibiting the ATM and/or ATR checkpoint kinases irradiation, this population increased twofold (Figur- (Blasina et al., 1999b; Esashi and Yanagida, 1999; e(2C b, arrow), reflecting the twofold larger proportion Sarkaria et al., 1999; Tominaga et al., 1999; Zhou et al., of BrdU-labeled G2 cells present at the time of PUVA 2000). To test whether active checkpoint signaling is treatment. In contrast, cells treated with UVB at a dose involved in the PUVA-induced cell cycle arrest, we equally effective in inducing cell cycle arrest were still incubated cells with caffeine prior to PUVA treatment trapped in G2, indicated by the absence of BrdU- and for the subsequent 12 h. Most cells were arrested in positive cells with G1 DNA content (Figure 2C c, d) and S-phase by PUVA in the absence of caffeine (Figure 4a). the presence of a significant BrdU-negative 4n-DNA In contrast, upon addition of caffeine, cells passed content population (Figure 2C c, d, rectangles). These through S-phase and mitosis, and entered the subse- findings demonstrate that the DNA damage induced by quent G1-phase (Figure 4b, BrdU-positive cells, upper PUVA is not sensed as a relevant signal for preventing left), similar to mock-treated controls (Figure 4d). For cells to enter mitosis, while UVB-induced lesions unknown reasons, addition of caffeine caused a slight apparently are. slowdown of cell cycle progression in mock-treated controls (compare panels c and d). Cells are not arrested in G1 by PUVA To determine whether PUVA treatment during G1- PUVA-induced replication block involves Chk1 and phase can prevent cells from entering S-phase, cells were Cdc25A synchronized in G0 by serum depletion for 48 h. followed by restimulation with 10% FCS. In the absence These findings suggested that PUVA induced a check- of PUVA. cells commenced DNA synthesis approxi- point signal in HaCaT cells. Chk1 is one of the mately 16 h following readdition of serum, and had checkpoint kinases acting directly downstream of traversed about 2/3 of S-phase by 24 h (Figure 3a, panels ATM/ATR. In response to DNA damage or incomplete 1–5). When treated with PUVA 8 h after readdition of replication, Chk1 is activated by phosphorylation at serum, cells still initiated replication at 16 h (Figure 3a, Ser345 (Liu et al., 2000; Zhao and Piwnica-Worms, panel 8, arrow) suggesting that passage through G1 was 2001). Using an antibody specific for the Ser345- unperturbed by PUVA. However, once having entered phosphorylated form of Chk1, we found that PUVA S-phase cells were arrested and did not continue DNA phosphorylates Chk1 (Figure 5, lanes 2 and 7). The replication (Figure 3a, panel 9). Since traversion overall level of Chk1 remained unaltered. Increase of through G1-phase cannot be monitored by changes in Chk1 phosphorylation occurred within 1 h following DNA content, we used phosphorylation of the retino- PUVA and was therefore unlikely to be caused by a blastoma protein pRb as a biochemical marker for G1 positional effect due to synchronization of cells. progression. Being hypophosphorylated in quiescent Phosphorylation of Chk1 was mirrored by a decrease cells. pRb is increasingly phosphorylated as cells in the level of Cdc25A, a dual-specificity phosphatase approach S-phase (Figure 3b). Cells treated with PUVA known to promote S-phase entry and progression at 8 h continued to phosphorylate pRb like untreated (Figure 5, lanes 2 and 7). Addition of caffeine prevented controls (Figure 3b), showing that PUVA does not the PUVA-induced phosphorylation of Chk1 and the arrest cells in G1. To exclude the possibility that BrdU decrease in Cdc25A level (Figure 5, lanes 3 and 8), incorporation after PUVA might reflect DNA repair suggesting that caffeine allowed cells to by-pass the S-

Oncogene PUVA-induced S-phase checkpoint C Joerges et al 6123

Figure 3 Cells are not arrested in G1 by PUVA. (a) After 48 h of serum depletion (panel 1), HaCaT cells were restimulated with medium containing 10% FCS. After 8 h, cells were either mock (panels 2–5) or PUVA (panels 6–9) treated. Cells were labeled with BrdU 15 min prior to processing for FACS analysis at the indicated time points. (b) In parallel to FACS analysis, cell lysates were analysed by immunoblot for pRb-phosphorylation using a phospho-speciﬁc antibody. At 8 h following restimulation with serum, cells were either mock (lanes 2–5) or PUVA (lanes 6–9) treated. Blotting with a-tubulin served as a control for equal loading. (c) Cells were treated as above except that in some samples (panels 4 and 5) the Cdk2-inhibitor roscovitine was added by the time of PUVA treatment (8 h)

Oncogene PUVA-induced S-phase checkpoint C Joerges et al 6124

Figure 4 Caffeine overrides PUVA-induced cell cycle block. HaCaT cells pulsed with BrdU were treated with PUVA (0 1 mg/ml MOP, 1.0 J/cm2 UVA) or mock in the presence or absence of 2 mm caffeine and processed for FACS 12 h later

phase checkpoint by interfering with ATM or ATR lane 3) Like endogenous Cdc25A, exogenous Cdc25A signaling and Chk1 phosphorylation. was reduced in response to PUVA, but still expressed at much higher than endogenous levels (Figure 6a, lane 4). Cells overexpressing Cdc25A were no longer arrested at Overexpression of Cdc25A partially overrides the beginning of S-phase by PUVA but rather pro- PUVA-induced replication arrest gressed to about mid-S-phase (Figure 6b, panel 4). If decrease in Cdc25A levels was essential for the However, unlike with caffeine (compare Figure 4), PUVA-induced S-phase arrest, overexpression of exo- abrogation of the arrest was incomplete since cells genous Cdc25A should interfere with such an arrest. To failed to enter the subsequent G1-phase. This suggests test this, we employed a rat ﬁbroblast cell line that stably that downregulation of Cdc25A is a necessary, yet not expresses a Flag-tagged human Cdc25A under the sufﬁcient event in the induction or maintenance of control of a tetracycline repressible promoter (Blomberg PUVA-induced S-phase arrest, implicating other caf- and Hoffmann. 1999). Withdrawal of tetracycline for feine-sensitive signaling events acting in parallel to 24 h caused a strong induction of Cdc25A (Figure 6a, Cdc25A.

Oncogene PUVA-induced S-phase checkpoint C Joerges et al 6125 induce DNA double-strand breaks. However, DNA double-strand breaks may result from repair of ICLs. In yeast, ICLs have been shown to be repaired by homologous recombination repair, thus transiently giving rise to double-strand breaks (Jachymczyk et al., 1981; Magana-Schwencke et al., 1982; de Andrade et al., 1989). Increased rates of homologous recombination have also been found after ICL inducing agents in mammalian cells (Wang et al., 1988). However, no signs of chromosomal breakage could be found during mitosis in cells that had been PUVA treated during G2, leading to the suggestion that homologous recombination repair might be restricted to cell cycle phases other than G2 (Akkari et al., 2000). Repair of ICLs has been estimated to occur at a rate of 11 ICLs per genome (Meniel et al., 1995) per hour, which is rather slow when compared to the approximately 96 000 ICLs induced per genome by 1 J/cm2UVA/0.1 mg/ml MOP (Gunther et al., 1995). The presumable lack of a relevant amount of strand breaks in G2 cells after PUVA could be explained either by the slow rate or the complete lack of recombination repair during G2-phase. Another reason for the lack of a checkpoint response outside S-phase might be that, as opposed to other ICL inducing agents such as platinum compounds or nitrogen mustard. PUVA-induced ICLs cause only minor distortion to the DNA double helix Figure 5 PUVA-induced replication block involves Chk1 and (Dronkert and Kanaar, 2001). Cdc25A. HaCaT cells were treated with PUVA (0.1 mg /ml MOP, 1.0 J/cm2 UVA) or mock in the presence or absence of 2 mm Eucaryotic DNA replication is initiated in a precise caffeine and harvested 1 or 12 h later. Cell lysates were analysed temporal order from many origins (Diffley, 1998). by immunoblotting using antibodies to Cdc25A, phosphorylated Stalling of replication forks by ICLs interrupts DNA Chk1 (Ser345) and total Chk1. Blotting with a-tubulin served as a synthesis from the corresponding origin. At a constant control for equal loading concentration of MOP, the number of ICLs increases almost linearly with the UVA dose applied (Gasparro, 1988). With increasing the number of ICLs, DNA Discussion synthesis should become less and less complete. Total block of replication can be expected if the number of Our findings provide evidence that PUVA induces a ICLs exceeds that of origins of replication. However, an checkpoint signal resulting in S-phase arrest and a estimate of the ICLs induced in our experiments makes decrease of Cdc25A levels. Mechanical blockage of this unlikely to be the case here: 1 J/cm2 UVA/ 0.1 mg/ml replication forks by ICLs is not sufficient to prevent MOP induces approximately 1.6 ICLs per mbp while continuation of DNA replication since the replication there are an estimated five origins per mbp in the human arrest can be overridden by caffeine. The fact that genome (Hamlin, 1992). The number of relevant ICLs PUVA-induced cell cycle arrest is restricted to S-phase per origin might still be lower since part of the ICLs will suggests further that ICLs do not trigger a DNA be located in DNA that has already been replicated. damage checkpoint signal per se but rather require the Direct evidence against a pure mechanical block is context of ongoing replication to do so. provided by the finding that caffeine can override the The lack of a G2/M-arrest following PUVA was PUVA-induced replication arrest. Many replication- unexpected since many other DNA damaging agents and DNA-damage checkpoints are sensitive to caffeine, such as UV or ionizing radiation have been shown to resulting in inappropriate cell cycle progression (Busse cause a G2/M arrest (Tobey, 1975; Herzinger et al., et al., 1977; Busse et al., 1978; Schlegel and Pardee, 1995; Orren et al., 1995). Induction of a 4n-DNA peak 1986). ATM and ATR have been identified as in vivo by low doses of PUVA has also been interpreted to targets of caffeine (Blasina et al., 1999b; Hall-Jackson represent a G2-arrest, but this was not distinguished et al., 1999). ATM and ATR are upstream regulators of from a late S-phase arrest (Cohen et al., 1981; Varga the Chk1 and Chk2/Cds1 checkpoint kinases. Over- et al., 1982). Using presynchronized cells, Akkari et al. expression of a dominant-negative ATR protein in (2000) recently showed that PUVA-induced ICLs cause human cells has been shown to abrogate phosphoryla- cell cycle arrest in S-phase, but not in G2-phase. DNA tion of Chk1 (Liu et al., 2000). In Xenopus extracts, double-strand breaks represent the most relevant lesion depletion of ATR prevented both phosphorylation of for the G2/M checkpoint since they may lead to Chk1 and cell cycle arrest in response to HU or UVC inaccurate chromosome segregation during mitosis. (Guo et al., 2000; Hekmat-Nejad et al., 2000). Similar to Other than ionizing radiation, PUVA does not directly our observations with PUVA, caffeine has been reported

Oncogene PUVA-induced S-phase checkpoint C Joerges et al 6126

Figure 6 Overexpression of Cdc25A partially overrides PUVA-induced replication arrest. (a) Rat ﬁbroblasts expressing Flag-tagged human Cdc25A under the control of a tetracyline-regulated promoter were PUVA (lanes 2 and 4) or mock (lanes 1 and 3) treated 24 h after growing in the presence (promoter suppression) or absence (promoter induction) of 10 mg/ml tetracycline. Cell lysates were harvested 12 h later, and expression levels of Cdc25A were analysed by immunoblot using antibodies to Cdc25A or the Flag-tag, respectively. Blotting with a-tubulin served as a control for equal loading. (b) Cells were treated as in (a), pulsed with BrdU prior to PUVA or mock treatment, and analysed by FACS 12 h later

to abrogate Chk1 phosphorylation in response to UV C overexpression of Cdc25A. However abrogation by radiation or the replication inhibitor hydroxyurea Cdc25A was less complete than with caffeine, allowing (Mailand et al., 2000; Molinari et al., 2000; Feijoo cells only to precede to about mid-S-phase and not et al., 2001). Like PUVA, UVC and HU also result in a further. This finding suggests that Cdc25A downregula- rapid decrease of Cdc25A abundance (Mailand et al., tion is necessary, yet not sufficient, for the PUVA- 2000; Molinari et al., 2000). Conversely, overexpression induced S-phase arrest. Other caffeine-sensitive signal- of Cdc25A accelerates S-phase entry (Blomberg and ing pathways seem to operate in parallel. Evidence for Hoffmann, 1999; Sexl et al., 1999) and interferes with such a parallel, yet ATM-dependent, pathway involving the replication checkpoint in response to hydroxyurea the Nbs1/Mre11/Rad50 complex has been provided (HU) by inducing premature chromosome condensation recently (Falck et al., 2002). (Molinari et al., 2000). In our experiments, the PUVA- The existence of a DNA replication checkpoint in induced S-phase arrest was partially abrogated by mammalian cells has first been demonstrated by cell

Oncogene PUVA-induced S-phase checkpoint C Joerges et al 6127 fusion experiments. Fusion of S-phase cells with G2 cells 20 min prior to irradiation until harvest for FACS or delays entry into mitosis, suggesting that the ongoing immunoblotting, respectively. UVB irradiation was carried DNA synthesis generates a checkpoint signal (Rao and out in a parallel bank of three Philips TL20/12 fluorescent Johnson, 1970). Damage to DNA by ionizing radiation tubes. during S-phase results both in a block of replication initiation and chain elongation. Cells lacking functional Cell cycle analysis ATM are defective in this checkpoint and continue After disaggregation with trypsin/EDTA solution cells were DNA synthesis in the presence of damage (radio- fixed in ice-cold 70% ethanol. Fixed cells were stained with a resistant DNA synthesis ¼ RDS) (Painter and Young, fluorescein isothiocyanate-conjugated a-BrdU antibody and 1980). Prevention of RDS is controlled by the ATM- propidium iodide (PI) (Sigma) as described (Herzinger and Chk2/Cds1-Cdc25A pathway and does not involve Reed, 1998), and cell cycle phase distribution was analysed Chk1(Falck et al., 2001). In contrast, arrest in response using a Becton Dickinson FACScan flow cytometer. to replication blocking agents such as HU has been For determination of the mitotic index, cells were seeded on shown to be caffeine sensitive yet independent of ATM. glass coverslips 24 h prior to PUVA or mock treatment. In addition, resumption of bulk DNA synthesis after Nocodazole (0.1 mg/ml) was added after treatment. After 1–8 h release from replication block correlates with inactiva- cells were fixed in 70% ethanol and nuclei were stained with DAPI. tion of Chk1 but not Chk2 (Chaturvedi et al., 1999; Feijoo et al., 2001). Therefore, DNA damage and stalled replicons seem to generate distinct checkpoint signals Cell extracts and immunoblotting during S-phase. The cell cycle arrest induced by PUVA Cells were lysed in ice-cold RIPA buffer (1% deoxycholate, shares the distinct features of a replication checkpoint: it 1% Triton X-400, 50 mm Tris-HCl pH 7.5,. 150 mm NaCl, is restricted to S-phase, it is sensitive to caffeine, and 0.1% SDS) containing 0.1 mm sodium orthovanadate, 1.0 mm it involves phosphorylation of Chk1. We therefore sodium fluoride and Completet proteinase inhibitor cocktail speculate that a few stalled replication forks generate (Roche). Total cellular protein (10 mg) was boiled in 2 Â a checkpoint signal resulting in the halt of bulk sample buffer (100 mm Tris-HCl pH 6.8, 200 mm dithiothreitol, replication. 4% SDS, 20% glycerol, 0.001% bromphenol blue) and resolved on a 8% SDS/polyacrylamide gel. Proteins were transferred to polyvinylidene difluoride membranes (Immobi- lon, Millipore) by semidry electroblotting for 30 min at 15 V Materials and methods constant voltage. After transfer, membranes were stained with amido black (0.1% in 45% methanol/10% acetic acid) to Reagents, cell lines control for equal loading. Membranes were then blocked in HaCaT cells were a kind gift of N Fusenig (Boukamp et al., Milk Diluent (Kirkegaard & Perry Laboratories) overnight at 1988) and grown as described previously (Herzinger et al., 41C. Antibody incubations were carried out at room tempera- 1995). Rat fibroblasts overexpressing Cdc25A under the ture. Dilutions were as follows: a-phospho-Chk1 (Ser345) control of tetracycline were kindly provided by I Hoffmann 1 : 1000; a-Cdc25A 1 : 100; a-Chk1 1 : 200; a-tubulin 1 : 10 000; anti- (Blomberg and Hoffmann, 1999). Antibodies were from Cell Flag 1:330; HRP-conjugated secondary antibodies 1 : 10 000). Signaling Technology (a-phospho-Chk1 (Ser345), a-phospho- Antigen/antibody complexes were visualized using ECL pRb (Ser807/811)). Santa Cruz (a-Cdc25A, a-Chk1), Sigma Western blotting detection reagents (Amersham Biosciences). (anti-a-tubulin, anti-Flag M5), Promega (horseradish perox- idase (HRP)-conjugated secondary antibodies) or Becton Dickinson (a-BrdU-FITC). Roscovitine was from Calbio- chem. Abbreviations UVA, ultraviolet A; ICLs, interstrand crosslinks; PUVA, psoralen plus ultraviolet A; MOP, 8-methoxy-psoralen; PBS, Cell irradiation phosphate-buffered saline; BrdU, bromo-deoxy-uridine; PI, propidium iodide; HU, hydroxyurea; RDS, radio-resistant A total of 100 000 cells were seeded per 6 cm well. After 24 h, DNA synthesis.. cells were incubated with MOP (0.1 mg/ml) in PBS for 20 min prior to irradiation with UVA (PUVA 200, Waldmann, Germany, equipped with 14 SYLVANIA F8T5 fluorescent Acknowledgements tubes). For labeling of S-phase cells, BrdU (10 mm) was added We thank L Hengst. B Lu¨ scher. J Lukas and DA Wolf for during the preincubation period in some experiments. After insightful discussions and technical advice, and I Hoffmann irradiation, cells were washed twice with warm phosphate- and N Fusenig for providing cell lines. This work was buffered saline (PBS) and fed with complete medium (DMEM supported by a grant from the Deutsche Forschungsge- with 10% FCS). When indicated, caffeine (2 mm) was added meinschaft (He2518/2-1).

References

Akkari YM, Bateman RL, Reifsteck CA, Olson SB Blasina A, de Weyer IV, Laus MC, Luyten WH, Parker AE and Grompe M. (2000). Mol. Cell. Biol., 20, and McGowan CH. (1999a). Curr. Biol., 9, 1–10. 8283–8289. Blasina A, Price BD, Turenne GA and McGowan CH. Bentley NJ, Holtzman DA, Flaggs G, Keegan KS, DeMaggio (1999b). Curr. Biol., 9, 1135–1138. A, Ford JC, Hoekstra M and Carr AM. (1996). EMBO J, Blomberg I and Hoffmann I. (1999). Mol. Cell. Biol., 19, 15, 6641–6651. 6183–6194.

Oncogene PUVA-induced S-phase checkpoint C Joerges et al 6128 Boukamp P, Petrussevska RT, Breitkreutz D, Hornung J, Donehower LA and Elledge SJ. (2000). Genes Dev., 14, Markham A and Fusenig NE. (1988). J. Cell Biol., 106, 1448–1459. 761–771. Ljunggren B, Carter DM, Albert J and Reid T. (1980). J. Busse PM, Bose SK, Jones RW and Tolmach LJ. (1977). Invest. Dermatol., 74, 59–62. Radiat. Res., 71, 666–677. Luftl M, Rocken M, Plewig G and Degitz K. (1998). J. Invest. Busse PM, Bose SK, Jones RW and Tolmach LJ. (1978). Dermatol., 111, 399–405. Radiat. Res., 76, 292–307. Magana-Schwencke N, Henriques JA, Chanet R and Mous- Chaturvedi P, Eng WK, Zhu Y, Mattern MR, Mishra R, tacchi E. (1982). Proc. Natl. Acad. Sci. USA, 79, 1722–1726. Hurle MR, Zhang X, Annan RS, Lu Q, Faucette LF, Scott Mailand N, Falck J, Lukas C, Syljuasen RG, Welcker M, GF, Li X, Carr SA, Johnson RK, Winkler JD and Zhou BB. Bartek J and Lukas J. (2000). Science, 288, 1425–1429. (1999). Oncogene, 18, 4047–4054. Matsuoka S, Huang M and Elledge SJ. (1998). Science, 282, Cimprich KA, Shin TB, Keith CT and Schreiber SL. (1996). 1893–1897. Proc. Natl. Acad. Sci. USA, 93, 2850–2855. Meniel V, Magana-Schwencke N and Averbeck D. (1995). Cohen SR, Burkholder DE, Varga JM, Carter DM and Mutat. Res., 329, 121–130. Bartholomew JC. (1981). J. Invest. Dermatol., 76, 409–413. Molinari M, Mercuric C, Dominguez J, Goubin F and Draetta de Andrade HH, Marques EK, Schenberg AC and Henriques GF. (2000). EMBO Rep., 1, 71–79. JA. (1989). Mol. Gen. Genet., 217, 419–426. Nurse P. (1997). Cell, 91, 865–867. Difﬂey JF. (1998). abc. Curr. Biol., 8, R771–R773. Orren DK, Petersen LN and Bohr VA. (1995). Mol. Cell. Biol., Dronkert ML and Kanaar R. (2001). Mutat. Res., 486, 15, 3722–3730. 217–247. Painter RB and Young BR. (1980). Proc. Natl. Acad. Sci. Esashi F and Yanagida M. (1999). Mol. Cell, 4, 167–174. USA, 77, 7315–7317. Falck J, Mailand N, Syljuasen RG, Bartek J and Lukas J. Peng CY, Graves PR, Thoma RS, Wu Z, Shaw AS and (2001). Nature, 410, 842–847. Piwnica-Worms H. (1997). Science, 277, 1501–1505. Falck J, Petrini JH, Williams BR, Lukas J and Bartek J. Rao PN and Johnson RT. (1970). Nature, 225, 159–164. (2002). Nat. Genet., 30, 290–294. Sadhu K, Reed SI, Richardson H and Russell P. (1990). Proc. Feijoo C, Hall-Jackson C, Wu R, Jenkins D, Leitch J, Gilbert Natl. Acad. Sci. USA, 87, 5139–5143. DM and Smythe C. (2001). J. Cell Biol., 154, 913–923. Sanchez Y, Wong C, Thoma RS, Richman R, Wu Z, Piwnica- Furnari B, Blasina A, Boddy MN, McGowan CH and Russell Worms H and Elledge SJ. (1997). Science, 277, 1497–1501. P. (1999). Mol. Biol. Cell, 10, 833–845. Sarkaria JN, Busby EC, Tibbetts RS, Roos P, Taya Y, Karnitz Galaktionov K and Beach D. (1991). Cell, 67, 1181–1194. LM and Abraham RT. (1999). Cancer Res., 59, 4375–4382. Gasparro FP. (1988). Psoralen DNA Photobiology, Vol. 1. Schlegel R and Pardee AB. (1986). Science, 232, 1264–1266. Gasparro FP (ed.).. CRC Press: Boca Raton, pp. 5–36. Sebastian B, Kakizuka A and Hunter T. (1993). Proc. Natl. Gunther EJ, Yeasky TM, Gasparro FP and Glazer PM. Acad. Sci. USA, 90, 3521–3524. (1995). Cancer Res., 55, 1283–1288. Sexl V, Diehl JA, Sherr CJ, Ashmun R, Beach D and Roussel Guo Z, Kumagai A, Wang SX and Dunphy WG. (2000). MF. (1999). Oncogene, 18, 573–582. Genes Dev., 14, 2745–2756. Tobey RA. (1975). Nature, 254, 245–247. Hall-Jackson CA, Cross DA, Morrice N and Smythe C. Tominaga K, Morisaki H, Kaneko Y, Fujimoto A, Tanaka T, (1999). Oncogene, 18, 6707–6713. Ohtsubo M, Hirai M, Okayama H, Ikeda K and Nakanishi Hamlin JL. (1992). BioEssays, 14, 651–659. M. (1999). J. Biol. Chem., 274, 31463–31467. Hekmat-Nejad M, You Z, Yee MC, Newport JW and Varga JM, Wiesehahn G, Bartholomew JC and Hearst JE. Cimprich KA. (2000). Curr. Biol., 10, 1565–1573. (1982). Cancer Res., 42, 2223–2226. Herzinger T, Funk JO, Hillmer K, Eick D, Wolf DA and Kind Wang YY, Maher VM, Liskay RM and McCormick JJ. P. (1995). Oncogene, 11, 2151–2156. (1988). Mol. Cell. Biol., 8, 196–202. Herzinger T and Reed SI. (1998). J. Biol. Chem., 273, Weinert T. (1998). Cell, 94, 555–558. 14958–14961. Zhao H and Piwnica-Worms H. (2001). Mol. Cell. Biol., 21, Hoffmann I, Draetta G and Karsenti E. (1994). EMBO J., 13, 4129–4139. 4302–4310. Zhao S, Weng YC, Yuan SS, Lin YT, Hsu HC, Lin SC, Jachymczyk WJ, von Borstel RC, Mowat MR and Hastings Gerbino E, Song MH, Zdzienicka MZ, Gatti RA, Shay JW, PJ. (1981). Mol. Gen. Genet., 182, 196–205. Ziv Y, Shiloh Y and Lee EY. (2000). Nature, 405, Jinno S, Suto K, Nagata A, Igarashi M, Kanaoka Y, Nojima 473–477. H and Okayama H. (1994). EMBO J., 13, 1549–1556. Zhou BB, Chaturvedi P, Spring K, Scott SP, Johanson RA, Liu Q, Guntuku S, Cui XS, Matsuoka S, Cortez D, Tamai K, Mishra R, Mattern MR, Winkler JD and Khanna KK. Luo G, Carattini-Rivera S, DeMayo F, Bradley A, (2000). J. Biol. Chem., 275, 10342–10348.

Oncogene