ATR suppresses apoptosis after UVB irradiation by controlling both translesion synthesis and alternative tolerance pathways

Ultraviolet light (UV) damages DNA through the induction of pyrimidine dimers (Beukers and Berends, 1960; Setlow and Carrier, 1964), which stall high-fidelity polymerases (Kanazir and Errera, 1954; Masters and Pardee, 1962). Cells avoid replication arrest by removing photoproducts through nucleotide excision repair (NER) or by switching replication to translesion synthesis (TLS) polymerases to bypass these lesions. Defects in any of these pathways can lead to the rare hereditary disorder xeroderma pigmentosum (XP), characterized by extreme sun sensitivity and increased risk of skin cancer. Whereas the seven classical complementation groups (XP-A to XP-G) exhibit defective NER, the XP variant group (XP-V) results from mutations in the TLS DNA polymerase η (DiGiovanna and Kraemer, 2012, Menck and Munford, 2014).

Despite a normal capacity to remove UV-induced DNA lesions, XP-V cells exhibit a defect in their ability to replicate damaged DNA (Lehmann et al., 1975) and present increased mutagenesis after UV irradiation (Maher et al., 1976; Wang et al., 1993). Interestingly, the slower recovery of DNA synthesis in these cells is further affected by caffeine (Cleaver et al., 1979). Primary XP-V cell lines are only slightly more sensitive to UV after caffeine; however, after p53 inactivation – either by transformation with SV40 or HPV16 genes – the sensitivity dramatically increases (Cleaver et al., 1999). DNA-damage-induced loss of the G₁/S checkpoint is a common feature in cancer cell lines and enables genome instability (Hanahan and Weinberg 2011) because it promotes DNA replication in the presence of damage. Thus, cells immortalized with SV40 depend even more on Pol η to bypass UV lesions, as XP-V SV40 cells exhibit more asymmetric replication forks (Despras et al., 2010), increased ATR kinase activation (Bomgarden et al., 2006) and sister chromatid exchange (Cleaver et al., 1999).

ATR kinase acts as a master regulator of the UV damage response and controls cell cycle progression and DNA replication and repair (Cimprich and Cortez, 2008). Activation starts with single-stranded DNA (ssDNA) coated with RPA proteins, a signal that arises from DNA replication and repair. Thus, UV damage activates ATR throughout the cell cycle through NER intermediates (Hanasoge and Ljungman, 2007) or through stalled replication forks during S-phase, as a consequence of helicase and polymerase uncoupling. The ATR DNA damage response involves at least 570 phosphorylation targets (Stokes et al., 2007) including DNA repair substrates in NER (XPA) and recombination (BRCA1, WRN and BLM), as well as checkpoint effectors, such as CHK1, that help stabilize replications forks, inhibit origin firing and induce the G₂/M checkpoint (Cimprich and Cortez, 2008). Recent work has shown that inhibiting ATR activates origin firing; under replicative stress, this exhausts RPA and generates fork collapse and DNA double-stranded breaks (DSBs), resulting in death signals (Toledo et al., 2013).

The same signal that induces the ATR checkpoint – ssDNA – also activates the recruitment of TLS polymerases through PCNA monoubiquitylation (Chang et al., 2006). However, it remains unclear whether these processes are independent and whether lesion bypass requires the replication checkpoint because ATR is not necessary for PCNA ubiquitylation (Niimi et al., 2008). If ATR is depleted from XP-V cells, they should be more sensitized to UV light, as they would be susceptible to mitotic catastrophe owing to their lack of a replication checkpoint (Nghiem et al., 2001). If TLS after UV irradiation depends on ATR signaling, cells proficient in TLS – even in the absence of NER – would behave similarly to XP-V cells after ATR knockdown. Thus, our aim was to investigate the specific roles of ATR in cell cycle control after UVB irradiation and the contribution of ATR to Pol η translesion synthesis. To this end, we specifically depleted ATR using siRNA in wild-type, XP-V or XP-C fibroblasts immortalized with SV40 to assess the impact of ATR in cells with different DNA damage response backgrounds.

We depleted ATR in SV40-immortalized fibroblasts using two different small interfering RNA (siRNA) sequences (siATR; Fig. 1). Both sequences sensitized cells to UVB exposure. Sequence I produced more efficient silencing at 72 h after transfection, with ATR mRNA levels <20% (Fig. 1B) and protein levels reduced to less than 30% compared with control cells (Fig. 1C), and was thus selected for the subsequent experiments.

We evaluated the effect of ATR deficiency on the sensitivity of these fibroblasts to UVB light by measuring clonogenic survival, which detects cells that have escaped from all types of cell death and senescence. After ATR silencing, wild-type cells (MRC5), cells deficient in Pol η (XP-V) and XPC (XP-C) each became more sensitive to UVB light (Fig. 1D). However, ATR depletion induced stronger sensitization in cells lacking Pol η, including after low UVB doses. Importantly, ATR silencing alone did not alter the cell growth of these cells in the short-term experiments (up to 120 h after the first transfection, Fig. 1E)

To more precisely evaluate how ATR depletion sensitizes cells to UVB irradiation, we quantified the sub-G₁ cell population, which represents cells with fragmented nuclei, a late step of apoptosis. Again, after ATR silencing, the three cell lines became hypersensitive to UVB at 48 h post irradiation (hpi), although the effect was more pronounced in XP-V cells (Fig. 2A).

To confirm the role of ATR in protecting cells from apoptosis after UVB irradiation, we quantitated the level of active caspase-3 in the silenced cells. Depleting ATR kinase increased caspase-3 activity in all three cell lines; however, caspase-3 activation started earlier (24 hpi) in XP-C and XP-V cells (Fig. 2B,C) compared with only 48 hpi for the wild-type cell line MRC5 (Fig. 2C). The fold induction reflects the increase in positive cells for this effector caspase, which was observed mainly in S-phase or in G₂/M cells (Fig. 2C).

Replication arrest after UVC irradiation induces the phosphorylation of H2AX (denoted γH2AX), and this phosphorylation is further increased in SV40-immortalized cells, especially in XP-V cells (Laposa et al., 2007; Limoli et al., 2002). Given the role of ATR in the replication fork and in the induction of caspase-3 activity during S-phase, we quantified the level of γH2AX after UVB irradiation to evaluate whether sensitivity was related to replication stress.

The levels of γH2AX at 6 hpi increased in both ATR-depleted XP-C and XP-V cells after UVB irradiation, although there seemed to be a saturation level at 600 J/m² for XP-V cells, probably because all S-phase cells were positive for γH2AX staining (this saturation is not observed in MRC5 or XP-C cells). For the wild-type cells, this increase in the number of γH2AX-labeled cells was observed mainly after very high UVB doses (1200 J/m²) and only after ATR silencing (Fig. 3A). There was also an corresponding increase in the percentage of γH2AX-positive cells (Fig. 3B), and cell cycle analyses indicated that over 70% of these γH2AX-positive cells are in S-phase. Thus, the lack of ATR most likely leads to an increase in the pan-nuclear γH2AX staining (Fig. 3C) of replicating cells (Fig. 3B) and confirms that ATR protects from S-phase stress after UVB exposure. Thus, these results indicate that mutations in either TLS or NER will lead to more DNA polymerase blockage across pyrimidine dimers, which depends on the ATR damage response, even after irradiation with low UVB doses.

SV40-immortalized cell lines are defective in the G₁ checkpoint and will proceed to replication even in the presence of DNA damage. After UVB irradiation, ATR will activate intra-S-phase and G₂/M-phase checkpoints (Cimprich and Cortez, 2008). Therefore, we would expect that there would be a loss of DNA damage checkpoints after ATR silencing, such that cells would proceed to mitosis even in the presence of DNA damage. Therefore, we analyzed the cell cycle progression of ATR-depleted cells to determine how they respond to S-phase stress after UVB exposure.

In the first 24 h after UVB irradiation, the cell lines accumulated in S-phase. However, they overcame replication stress and activated the G₂ checkpoint after 48 h (Fig. 4A, upper panels), with the exception of the XP-V cells, which still accumulated in S-phase at higher UVB doses at 48 hpi, as is expected because of their deficiency in TLS. All ATR-depleted cells showed similar cell cycle profiles at 24 hpi when compared with their respective controls, though the profiles already indicated a stronger S-phase blockage. These cells were unable to recover at 48 hpi and continued displaying a profile similar to that observed at 24 hpi (Fig. 4A, lower panels), including S-phase accumulation (Fig. 4B). The effect is again very high for XP-V cells, even at a lower UVB dose, with most of the cells unable to overcome the G₁/S phase transition, whereas for MRC5 cells, this S-phase blockage in depleted cells was evident only after higher UVB doses.

ATR-depleted cells accumulate in S-phase after UVB irradiation independently of their genetic background. However, cell populations could simply accumulate in S-phase if they die in G₂/M by mitotic catastrophe. To investigate whether the cells were in fact reaching mitosis and dying through possible mitotic catastrophe due to loss of the DNA damage checkpoint, we blocked cells with nocodazole – a known inhibitor of the mitotic spindle – to accumulate cells in metaphase.

As expected, the non-irradiated cells accumulated in G₂/M after 24 h nocodazole treatment (Fig. 5A, upper panels, 0–24 h) or and upon 24 h nocodazole treatment after 24 h of culture (Fig. 5A, lower panels, 24–48 h). After UVB irradiation, the ATR-proficient cells still accumulated in S-phase at 24 hpi but accumulated in G₂/M at 48 hpi, independently of their genetic background, indicating that cells that do not die by apoptosis can recover from damage and pass through S-phase. However, upon UVB irradiation of ATR-depleted cells, no accumulation of cells in G₂/M was observed, as these cells were mostly blocked in S-phase (Fig. 5A). Thus, ATR depletion prevents recovery from replication blockage, even in cells proficient in NER and TLS.

We then questioned whether the ATR-depleted cells were attempting to replicate their damaged genomes or whether they had arrested prior to replication. Cells were labeled with a 30-minute pulse of BrdU at 24 or 48 hpi. All ATR-depleted cells tested still incorporated BrdU at 24 hpi; however, the labeling was more heterogeneous and was mostly present for cells that were in S-phase in the wild-type cell line (MRC5), whereas the XP-V and XP-C cell lines showed greater decreases in positive cells even at lower doses of UVB irradiation (Fig. 5B). The silenced cells incorporate even less BrdU at 48 hpi (Fig. 5B), confirming that siRNA against ATR sensitizes cells by increasing S-phase stress and that these cells are unable to achieve mitosis.

In this work, we observed that ATR knockdown increased replication stress and promoted the early induction of apoptosis following UVB irradiation in SV40-immortalized human cells. As expected, this effect was more pronounced at lower UVB doses in cells deficient in global excision repair (XP-C) or in translesion synthesis (XP-V), although similar results could be observed at higher UVB doses in wild-type human cells (proficient for NER and TLS). However, we found that following UVB irradiation – unlike what would be expected after the loss of a DNA checkpoint – that ATR-silenced cells were unable to achieve mitosis, as these cells were not able to complete S-phase. Whereas cells with ATR recovered and accumulated in G₂/M at 48 h after UVB exposure, cells with ATR knockdown persisted with the same cell cycle profile, namely, they were unable to progress through S-phase. Even MRC5 fibroblasts, which are proficient in NER and TLS, were unable to progress under replication stress in the absence of ATR because no silenced cells accumulated in G₂/M after nocodazole and UVB treatment. Thus, ATR also regulates recovery of replication arrest following UV irradiation independently of Pol η in SV40-immortalized cell lines.

Our results confirm reports from previous studies about the role of ATR in suppressing caspase-3-dependent cell death following low UV doses (Al Khalaf et al., 2012; Myers et al., 2009). The UV sensitivity after ATR knockdown correlates with increased levels of γH2AX (phosphorylation at S139), which is a marker of DSBs after ionizing radiation (Rogakou et al., 1998). However, following UV irradiation, a pan-nuclear γH2AX staining was observed, rather than the discrete foci observed after ionizing radiation. In fact, this early marker does not correspond to DSBs after UV irradiation (de Feraudy et al., 2010), and its role is still not clear given that knocking down H2AX does not affect UV sensitivity (Revet et al., 2011). However, this chromatin remodeling histone can be used as a marker of DNA damage responses. H2AX is known to be an ATR target after replication stress (Ward and Chen, 2001); however, we observed increased H2AX phosphorylation following UV irradiation after ATR was knocked down in the cell lines tested. Ward and Chen (Ward and Chen, 2001) used kinase-dead ATR, whereas we used siRNA and performed the experiments in SV40-immortalized cells, which are known to have a defective G₁ checkpoint. If γH2AX increases dramatically in S phase, we would expect to observe less phosphorylation as a consequence of cell cycle change in p53-proficient cells, which normally arrest at the G₁ checkpoint. We also cannot exclude the possibility that the residual (30%) levels of ATR present in our experiments could still phosphorylate H2AX; however, evidence suggests that ATM phosphorylates H2AX following UV irradiation, and kinase-dead ATR could interfere with the pan-nuclear phosphorylation of S139 by related kinases during S-phase (de Feraudy et al., 2010).

The level of p53 alters the sensitivity of cells to UV light. XP-V cells immortalized with HPV E6/E7 – a process that degrades p53 – are more sensitive to UV light when compared with the same XP-V cells immortalized with SV40 – a process that stabilizes p53 but inhibits its transcription transactivation activity (Cleaver et al., 2002). p53 also inhibits TLS, and SV40-immortalized cell lines bypass cyclobutane pyrimidine dimer lesions with twice the efficiency of wild-type cells (Hendel et al., 2008). Even the SV40 XP-V cell line bypasses UV-induced lesions more efficiently – through other TLS polymerases – although with a higher error rate (Hendel et al., 2008). Nevertheless, these SV40 XP-V cell lines still exhibit more UV-induced sister chromatid exchange (Cleaver et al., 1999) and Mre11 foci (Limoli et al., 2002), indicating the additional need to activate alternative fork recovery pathways when p53 is not functional.

A previous report has shown that the inhibitors caffeine, UCN or even siRNA against Chk1 sensitize XP-V but not wild-type cell lines to UV irradiation (Despras et al., 2010). The use of siRNA against ATR in the present work is more specific than these inhibitors, and we confirmed that XP-V was further sensitized to UV light; however, even NER-proficient (MRC5) and deficient (XP-C) cells were also sensitized by ATR depletion (Fig. 1C, Fig. 2A, Fig. 3B). Thus, we cannot rule out the possibility that ATR protection involves kinase-independent mechanisms because siRNA against ATR affects the whole protein and not just kinase activity. Remarkably, ATR-silenced cells were still unable to resume replication following UVB irradiation, even in cells proficient in DNA repair and TLS, albeit at higher UVB doses. Pol η is also an ATR phosphorylation target (Chen et al., 2008; Göhler et al., 2011), and ATR depletion could impair TLS in wild-type cells; however, we observed that S-phase progression was impaired regardless of genetic background in SV40-immortalized cells following ATR depletion and UVB irradiation. Thus, other DNA damage tolerance pathways – such as template switch and homologous recombination – which are essential in SV40-immortalized cells owing to their G₁ checkpoint defect – must also be compromised in ATR-silenced cells to explain these results.

Defective DNA damage checkpoints could increase death by mitotic catastrophe as a result of persistent DNA synthesis in the presence of replication stress and the induction of chromosomal breaks, as suggested by a previous study with cells deficient in ATR activity and the G₁ checkpoint (Nghiem et al., 2001). However, in ATR-depleted SV40-immortalized cell lines we observed an increase in UV-induced sensitivity that correlated with that amount of S-phase stress (Fig. 3) and no accumulation of cells on G₂/M phase after nocodazole treatment (Fig. 5A). We also did not observe nuclei with premature chromatin condensation when these cells were irradiated with UVB doses as low as 100 J/m² (data not shown). These cells also incorporated less BrdU (Fig. 5B), confirming that they were unable to complete DNA synthesis and that cell death occurred before mitosis. The study by Nghiem et al., 2001 revealed nuclei with premature chromatin condensation in cells defective in the G₁ checkpoint, because of overexpression of MDM2, cyclin E or HPV E6 (which is different from p53 inactivation by the SV40 T antigen), and the addition of caffeine or kinase-dead ATR, which affects only the kinase activity, could explain the different results.

If cells do not die from catastrophic mitosis, then how can we explain our observation that ATR silencing still induces cell death in S phase? ATR is vital for replication even without exogenous damage (Petermann and Caldecott, 2006) because it has roles in maintaining the stability of fragile sites in chromosomes, and preventing breakage and deletions or chromosomal rearrangements (Casper et al., 2002). Thus, ATR signaling is implicated in tumorigenesis by imposing a barrier to cancer progression (Fang et al., 2004). Moreover, ATR acts locally, by stabilizing the arrested fork, and globally, by inhibiting new replication origins (Sørensen and Syljuåsen, 2012). RPA-coated ssDNA, which occurs during replication stress regardless of the source, signals the activation of ATR. However, cells replicating with DNA damage increase the amount of exposed ssDNA and deplete RPA, leading to fork collapse and replication catastrophe (Toledo et al., 2013). Given that we found that ATR-depleted wild-type SV40 cells are unable to resolve replication stress and die before achieving mitosis, even with wild-type Pol η, but are still sensitized when this DNA polymerase is mutated, we propose that efficient DNA TLS depends on both ATR-dependent and -independent pathways (Fig. 6). Furthermore, to avoid fork stall in cells with increased genome instability – such as SV40-immortalized cells – multiple bypass pathways would work in synchrony. If stall persists, the collapsed fork would produce a double-stranded end, a structure that differs from a double-strand break in that it has no second end to rejoin (Shrivastav et al., 2008, Batista et al., 2009). Nevertheless, homologous recombination proteins, which are also controlled by ATR (Sørensen et al., 2005), would drive replication restart at these ends, avoid caspase-3 activation, and cells would progress into mitosis.

Investigating how DNA damage checkpoints regulate DNA repair and tolerance pathways is vital to understanding how genomic stability is sustained in wild-type cells or lost in neoplastic cells. The ATR-mediated DNA damage response not only activates cell cycle checkpoints but also regulates replication and different tolerance pathways, and ATR is a vital kinase for protecting the genome against DNA damage that blocks replication. Inhibiting ATR has the potential to protect against UVB-induced carcinogenesis (Conney et al., 2007) and is growing as a potential sensitizer of chemotherapy or radiation, with selective killing of p53-deficient cancer cells (Reaper et al., 2011; Peasland et al., 2011; Prevo et al., 2012; Sultana et al., 2013). Thus, knowledge of the ATR-mediated DNA damage response can also aid in the design of more efficient cancer treatment protocols.

Wild-type (MRC5-V1), XP-V (XP30RO) and XP-C (XP4PA) SV40-immortalized fibroblasts were maintained in Dulbecco's modified Eagle's medium (LGC Biotecnologia, SP, Brazil) supplemented with 10% fetal bovine serum (Cultilab, Campinas, SP, Brazil) and antibiotics (0.1 mg/ml penicillin, 0.1 mg/ml streptomycin and 0.25 mg/ml fungizone – Life Technologies, CA) at 37°C in a humidified 5% CO₂ atmosphere. The XP30RO cell line has a homozygous deletion of the POLH gene, which leads to an inactive truncated protein lacking 42 amino acids at the N terminus. The XP4PA cell line has a homozygous deletion of two nucleotides in the XPC gene (TG at nucleotides 1744–1745), leading to a truncated non-functional protein.

Transient downregulation of ATR expression was achieved by transfecting cells with a specifically designed short interfering RNA (siRNA) oligonucleotide (ATR-HSS100878 Stealth; 5′-UUAACAUGUUCUUACCCUCAGGUGG-3′) obtained from Invitrogen Life Technologies, CA. A scrambled siRNA duplex from Qiagen, Netherlands (catalog number 1027310: 5′-UUCUCCGAACGUGUCACGUdTdT-3′) was used as a negative control. Cells were transfected with a final concentration of 40 nM siRNA using Oligofectamine (Invitrogen-Life Technologies, Carlsbas, CA), as described by the manufacturer, and were treated at 72 h after transfection.

Total RNA was extracted using TRIzol reagent (Invitrogen-Life Technologies), according to the manufacturer's protocol, and were treated with the DNase prior to RT-PCR (with a kit from Promega, WI). cDNA was prepared using the High Capacity cDNA Archive kit (Applied Biosystems-Life Technologies). Gene expression was determined by real-time PCR. Briefly, 1 µl of diluted cDNA, 10 µl of SYBR green master mix, 10 µM forward and reverse primers and nuclease-free water were used in a combined total volume of 20 µl for each reaction. Q-PCR was carried out using the 7500 Real-Time PCR System (Applied Biosystems-Life Technologies). The relative expression levels of the genes of interest were calculated using the relative standard curve method, based on the individual Q-PCR primer efficiencies, and the quantified values were normalized against the housekeeping gene encoding β-actin. The primer sequences used were as follows: 5′-CCCATCTGTCAGTTTTTGGTAGAA-3′ (forward) and 5′-CTGGCATGGAGTATTCGGAAGT-3′ (reverse) for ATR; 5′-CCAGCTCACCATGGATGATG-3′ (forward) and 5′-ATGCCGGAGCCGTTGTC-3′ (reverse) for β-actin.

Cells were irradiated with UVB at the indicated doses in 1 ml of PBS buffer by using a UVB lamp emitting mainly at 312 nm with a dose rate of 3.5 J/m²/s. The UVB dose was monitored by an VLX 3W radiometer (Vilber Lourmat, France). After UVB irradiation, the PBS was discarded, and the cells were maintained in culture medium.

Cells were harvested by trypsinization, washed once with ice-cold PBS, and kept on ice for 20 min after resuspention in RIPA buffer (10 mM Tris-HCl pH 7.4, 5 mM EDTA, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS and 1∶500 of protease inhibitor cocktail set III by Calbiochem-Merck Millipore, Billerica, MA). The remaining cell debris was removed by centrifugation at 18,000 g for 30 min, and supernatants were collected. Quantification was performed using the Pierce BCA Protein Assay kit (Thermo Scientific, Waltham, MA, USA).

Total protein (50 µg) from cell extracts was separated on a 7.5% SDS polyacrylamide gel. Thereafter, proteins were blotted onto a nitrocellulose transfer membrane (Amersham, GE Healthcare, Little Chalfont, United Kingdom) for 1.5 h. Membranes were blocked for 1 h in 5% (w/v) milk powder in PBS, incubated overnight at 4°C with primary antibody against ATR (Santa Cruz Biotechnology, Dallas, TX), washed three times with PBS-Tween 20, and incubated for 1 h with horseradish-peroxidase-coupled secondary antibody (1∶2000; Sigma-Aldrich, St Louis, MO). After a final wash with PBS-Tween 20 (three times for 10 min each), the blots were developed using an enhanced chemiluminescence detection system (Immobilon Western, Merck Millipore).

Cells were seeded on 6-cm dishes at 1×10³ cells per plate. After 16 h, cells were irradiated as described above. After 7–10 days, cultures were fixed with 10% formaldehyde for 10 min and stained with 1% Crystal Violet for 5 min. Colonies (defined as 15 cells or more) were counted using a Nikon SMZ 2B stereomicroscope. The survival frequency was calculated as the number of colonies counted on treated plates relative to that on untreated plates.

Cells were seeded on 3.5-cm dishes at 10⁵ cells per plate. After 24 h, cells were irradiated as described above and incubated at 37°C for 48 h. Then, both adherent and detached cells were harvested and centrifuged at 450 g for 3 min. Cell pellets were resuspended in 1 ml of cold 70% ethanol and kept at −20°C for at least 2 h. After rinsing with PBS, cells were resuspended in 200 µl of a propidium iodide solution (200 µg/ml RNase A, 20 µg/ml propidium iodide, plus 0.1% Triton X-100 in PBS) and kept in the dark for 30 min. The propidium iodide fluorescence of each sample was measured on a flow cytometer (Guava, Merck Millipore,). Guava Express Plus software was used to quantify subdiploid nuclei (sub G₁), and cell cycle analysis was performed using ModFit LT software, excluding cells with DNA content greater than 4N.

Cell pellets (10⁵ cells) were fixed with 2% formaldehyde, kept on ice for 15 min, and then fixed with 1 ml of cold 70% ethanol and kept at −20°C for at least 2 h. Cells were then resuspended in permeabilization and blocking solution (3% BSA; 0.2% Triton X-100 in PBS) and incubated for 1 h with primary antibody (1∶10, rabbit anti-active-caspase-3 antibody by BD, Pharmingen, NJ; 1∶500, mouse anti-γH2AX Ser139 antibody by Millipore). After washing, cells were incubated for 1 h with the corresponding FITC-coupled secondary antibody (1∶200, Sigma-Aldrich,). Then, cells were incubated in propidium iodide solution for 30 min. The fluorescence signal from each sample was measured on a flow cytometer (Guava, Merck Millipore) and analyzed by Guava Express Plus software, excluding cells with DNA content greater than 4N.

Cells were plated onto coverslips 1 day before treatment. They were fixed with 100% methanol for 10 min and then washed with PBS. Samples were blocked in 5% BSA for 30 min. After washing once with PBS, cells were incubated with primary antibodies for 60 min. Coverslips were washed with PBS and then incubated with secondary antibodies for another 60 min. They were then washed with PBS and stained with DAPI (4′,6′-diamidino-2-phenylindole) for 1 min. Images were obtained on a Zeiss Axiovert 200 microscope equipped with a Zeiss AxioCam MRm camera using the Axiovision 4.5 software.

Before harvesting, cells were incubated for 24 h with 100 ng/ml nocodazole (Sigma-Aldrich) added to the medium immediately after UVB irradiation (0–24 h) or at 24 hpi (24–48 h). Then, cells were fixed with 70% ethanol for cell cycle analysis by flow cytometry. Because nocodazole inhibits the mitotic spindle, treated cells are trapped in metaphase and the population accumulates in G₂/M.

A final concentration of 10 µM BrdU (Sigma-Aldrich) was added to the medium 30 min before harvesting (24 or 48 h after UVB irradiation). Cell pellets (10⁵ cells) were fixed overnight at −20°C with cold 75% ethanol. Cells were then resuspended in 2 ml pepsin solution (Pepsin 14 µM, 1.5% HCl 2 M) for 20 min at 37°C and then in 1.5 ml of 2 M HCl for 20 min at room temperature. Cell pellets were subsequently permeabilized and blocked (20 mM Hepes pH 7.5, 0.5% FBS, 0.5% Tween 20 in PBS), followed by incubation for 1 h in primary antibody (1∶100, mouse anti-BrdU M0744 by Dako, Denmark). After washing, cells were incubated for 45 min with anti-mouse IgG-FITC antibody secondary antibody (1∶200, Sigma). Then, cells were incubated in propidium iodide solution for 30 min. The fluorescence signal from each sample was measured on a flow cytometer (Guava, Merck Millipore) and analyzed using Guava Express Plus software. Cells with DNA content greater than 4N were excluded.

Statistics were performed using GraphPad Prism v5.01 (GraphPad). Graphs were built plotting mean±s.d. and tests of significance were by means of a two-way ANOVA with Bonferroni corrections post-test. Significance was assumed when P ≤0.05.