FEN1 Functions in Long Patch Base Excision Repair Under Conditions of Oxidative Stress in Vertebrate Cells

Base excision repair (BER) is the major system for repairing oxidized, alkylated, and deaminated DNA bases in the genomic DNA (1). BER consists of two major subpathways known as single-nucleotide (SN)-BER and long patch (LP)-BER that are distinguished by their repair patch sizes and by the enzymes involved. BER is initiated by a DNA glycosylase that cleaves the glycosidic bond between the modified base and the deoxyribose phosphate backbone (2). This process, however, generates mutagenic and toxic BER intermediates, such as abasic (AP) sites and strand breaks (3). In one form of SN-BER, AP endonuclease 1 (APE1) incises the DNA backbone 5′ to the abasic site sugar, and DNA polymerase β (POLβ) incorporates the appropriate nucleotide into the gap and eliminates the 5′ AP site through its 5′-deoxyribose-phosphate lyase activity (4, 5). Finally, DNA ligase I or III seals the nick in the lesion-containing DNA strand to complete SN-BER. In addition to formation during BER, ∼9,000 AP sites are generated spontaneously in each mammalian cell per day, and these sites also are normally repaired by SN-BER and LP-BER (6, 7). Furthermore, reactive oxygen species oxidize deoxyribose in the DNA backbone resulting in single-strand breaks (SSB; refs. 8, 9). The 5′-oxidized deoxyribose lesions are resistant to the 5′-deoxyribose-phosphate lyase activity of POLβ but can be excised by the LP-BER subpathway (10). In the LP-BER subpathway, DNA polymerase β/δ/ϵ incorporates 2 to 15 nucleotides into the repair patch and displaces the lesion-containing strand to generate a 5′-flap that is then excised by flap endonuclease 1 (FEN1; refs. 11, 12) and, finally, a DNA ligase completes the LP-BER process. In addition, oxidized bases in DNA are excised by bifunctional glycosylases, such as NTH1, OGG1, NEIL1, NEIL2, and NEIL3 (13). These enzymes are capable of AP lyase activity that nicks the DNA strand in association with the base removal reaction (14, 15). This type of bifunctional glycosylase activity leaves the 3-α,β–unsaturated aldehyde blocked, and APE or polynucleotide kinase is required to remove the 3′-α,β–unsaturated aldehyde blocking group before the gap-filling DNA synthesis (16) and completion of SN-BER or LP-BER. Although DT40 cells lacking POLβ are not hypersensitive to H₂O₂ in cell survival assays, after exposure to H₂O₂ in POLβ null DT40 cells, we consistently detected a higher accumulation of SSBs than in wild-type DT40 cells (17, 18). These results suggested that under oxidative stress conditions, POLβ may be involved in gap-filling DNA synthesis for SN-BER initiated by bifunctional DNA glycosylases. In the absence of POLβ, other DNA polymerases could back up this DNA synthesis role with relatively high efficiency. Because of the constant generation of endogenous mutagens, AP sites and their oxidation leading to SSBs are believed to be the most common type of DNA damage in cells (19, 20).

Loss of BER genes [e.g., APE1, POLβ, FEN1, and X-ray repair cross-complementing protein 1 (XRCC1)] leads to embryonic lethality in mice (21-25), suggesting a critical need for the repair of BER intermediates during mouse embryogenesis. The importance of POLβ in SN-BER has also been shown by using POLβ-deficient mouse embryonic fibroblasts (26). POLβ-deficient mouse embryonic fibroblasts are hypersensitive to methyl methanesulfonate (MMS), among other alkylating agents. The hypersensitivity to MMS seems to be due to the loss of 5′-deoxyribose-phosphate lyase activity of POLβ, but not from a deficiency in polymerase activity (5). However, the absence of POLβ in mouse embryonic fibroblasts does not alter cell doubling times nor make them strongly sensitive to H₂O₂, but it does make them hypersensitive to alkylating agents. The lack of strong sensitivity toward H₂O₂-induced damage may be due to polymerases λ, ι, or Q being used as backup enzymes in the repair of oxidized bases (17, 18, 27) or to an alternative base incision repair pathway (28).

In eukaryotic cells, FEN1 is believed to play critical roles in DNA metabolism, including removal of the RNA primer segment at the 5′ end of Okazaki fragments during replication (6, 9, 29). Upon deletion of the FEN1 homologue (Rad27) in Saccharomyces cerevisiae, cells are viable and exhibit an elevated spontaneous mutational frequency, indicating that the processing of Okazaki fragments during replication can still occur, but at the expense of genome integrity. The processing of Okazaki fragments in the absence of Rad27 is thought to be done by the exonuclease activity of DNA polymerase δ (POLδ; refs. 30-33). S. cerevisiae Rad27 mutant cells also exhibit hypersensitivity to MMS. This is not surprising, as the only form of robust BER present in S. cerevisiae is LP-BER (6, 34-37), and thus, the loss of Rad27 attenuates the LP-BER of the damage caused by MMS. DT40 FEN1 knockout cells also show a slow rate of cell proliferation and hypersensitivity toward MMS and are hypersensitive to H₂O_2, as well (4). The reason for their sensitivity toward MMS and H₂O₂ is not known, but it could be the result of a deficiency in LP-BER. It is important to note that both S. cerevisiae and DT40 cells are believed to use homologous recombination to a greater extent than many other eukaryotic cell types (1). Therefore, homologous recombination could play a role to back up the BER pathway in S-G₂ cells, particularly in the absence of Rad27/FEN1-dependent LP-BER (38, 39).

In the present study, we used DT40 cells and cell extracts to investigate the effect of FEN1 gene deletion on DNA repair capacity, cell viability, and genomic DNA synthesis. We found that extract from FEN1 null cells exhibited lower uracil-DNA BER capacity than extract from wild-type cells, and using an assay specific for the LP-BER subpathway, the reduction in BER capacity seemed to be due to a deficiency in the LP-BER subpathway. FEN1 null cells showed an increased level of apoptosis compared with the wild-type cells after exposure to H₂O₂. Using a DNA fiber–labeling strategy that allows one to visualize replication on a large number of individually labeled replication units, we found that FEN1 null cells had a slower replication rate than wild-type cells. In addition, a higher proportion of replication forks stalled in FEN1 null cells after DNA damage, suggesting that FEN1-mediated LP-BER serves an essential function in preventing DNA replication forks from prematurely terminating at oxidative DNA damage sites. These observations represent enhanced understanding of the importance of FEN1 in the LP-BER subpathway in vertebrates. By making use of LP-BER assays of extracts, our results confirm and extend the results reported by Matsuzaki et al. (4).

Wild-type DT40 and DT40-derived mutant cells (4, 40) were cultured as a suspension in a humidified atmosphere with 5% CO₂ at 39.5°C. The culture medium consisted of RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum (Sigma), 1% chicken serum (Sigma and Invitrogen), 100 μg/mL penicillin, and 100 μg/mL streptomycin (Invitrogen).

The DNA substrate for the 7,8-dihydro-8-oxoguanine (8oxoG)-DNA BER assay was a 100-bp duplex DNA constructed by annealing two synthetic oligodeoxyribonucleotides (Operon Biotechnologies, Inc.) to introduce a 8oxoG:C bp at position 23: 5′-CGAGTCATCTAGCATCCGTATCXACTCGTTACGTGATCGTGTACTGCATGTGTATGTCGTATGATGTCTATGTCTCGAACTACGTAGACTTACTCATTGC-3′ and 5′-GCAATGAGTAAGTCTACGTAGTTCGAGACATAGACATCATACGACATACACATGCAGTACACGATCACGTAACGAGTCGATACGGATGCTAGATGACTCG-3′, in which X represents 8oxoG. The DNA substrate for the uracil-DNA BER assay was a 35-bp duplex DNA constructed by annealing two synthetic oligodeoxyribonucleotides (Oligos Etc., Inc.) to introduce a G:U bp at position 15: 5′-GCCCTGCAGGTCGAUTCTAGAGGATCCCCGGGTAC-3′ and 5′-GTACCCGGGGATCCTCTAGAGTCGACCTGCAGGGC-3′.

Cell extracts were prepared as previously described (41). Briefly, cells were washed twice with PBS at 25°C, collected by centrifugation, and resuspended in an equal volume of Buffer I [10 mmol/L Tris-HCl (pH 7.8), 200 mmol/L KCl, and protease inhibitor cocktail (Boehringer Mannheim)]. An equal volume of Buffer II [10 mmol/L Tris-HCl (pH 7.8), 200 mmol/L KCl, 2 mmol/L EDTA, 40% glycerol, 0.2% Nonidet P-40, 2 mmol/L DTT, and protease inhibitor cocktail] was then added. The suspension was rotated for 1 h at 4°C, and the resulting extract was clarified by centrifugation at 14,000 rpm at 4°C. The supernatant fraction was used in the in vitro BER assays. The protein concentration of extracts was determined by the Bio-Rad protein assay using bovine serum albumin as the standard.

The “total BER” assay (final volume 10 μL) was done with the 100-bp 8oxoG containing oligonucleotide DNA substrate or 35-bp uracil-containing oligonucleotide DNA substrate as described above. The duplex oligonucleotide (50 nmol/L) was incubated with 10 μg of DT40 whole-cell extract prepared from wild-type or FEN1 null cells in 50 mmol/L Tris-HCl (pH 7.5), 5 mmol/L MgCl₂, 20 mmol/L NaCl, 0.5 mmol/L DTT, and 4 mmol/L ATP. For the assay of total BER, incubation was conducted in the presence of 20 μmol/L each of dATP, dCTP (for 8oxoG BER)/dGTP (for uracil-BER), and dTTP, and 2.3 μmol/L [α-³²P]dGTP for 8oxoG BER analysis or [α-³²P]dCTP for uracil-BER analysis at 37°C for 2, 5, 10, and 20 min.

For the 8oxoG LP-BER assay, 8oxoG-containing oligonucleotide was 5′-end labeled by incubation with OptiKinase (USB Corp.) in the presence of [γ-³²P]ATP for 30 min at 37°C. The labeled DNA was then annealed with the complementary strand, and the DNA was purified with a MicroSpin G-25 column (GE Healthcare). The ³²P-labeled duplex oligonucleotide substrate was subjected to a similar incubation in the presence of 2 μmol/L dGTP and 50 μmol/L ddCTP as previously described (41). For the uracil LP-BER assay, nonend-labeled duplex uracil containing the DNA substrate was incubated in the presence of 2.3 μmol/L [α-³²P]dCTP and 50 μmol/L dideoxythymidine 5′-triphosphate (ddTTP), with or without 20 and 100 nmol/L purified FEN1 at 37°C for 1, 5, and 10 min. In all incubations, LP-BER is quantified by measuring the amount of two-nucleotide intermediate formed, and SN-BER is quantified by measuring the amount of ligated BER product formed.

The reaction products were analyzed by 15% polyacrylamide denaturing gel electrophoresis. The initial rate of the in vitro BER reaction was determined from a plot of PhosphorImager radiolabel signal in ligated BER product (representing total BER or SN-BER) or in two-nucleotide addition product (representing the intermediate of LP-BER) against each time point. The linear relationship between activity and time of incubation for each extract was obtained. [γ-³²P]ATP (7,000 Ci/mmol) was from MP Biomedicals. [α-³²P]dGTP (6,000 Ci/mmol) and [α-³²P]dCTP (3,000 Ci/mmol) were from Perkin-Elmer Life Sciences and GE Healthcare, respectively. Purified human FEN1 was prepared as described (42).

To construct chicken FEN1 expression vectors, full-length chicken FEN1 cDNAs were inserted into the multicloning site of the pCR3-loxP-MCS-loxP expression vector (43). The expression plasmid (pCR3-loxP-FEN1/IRES-EGFP-loxP) was linearized and electroporated with a Gene Pulser apparatus (Bio-Rad) at 550 V and 25 μF. Subsequently, the cells were transferred into 20 mL fresh medium and incubated for 24 h. Cells were resuspended in 80 mL of medium containing G418 (2 mg/mL, Sigma) and divided into four 96-well plates. After 7 to 10 d, drug-resistant colonies in the 96-well plates were transferred to 24-well plates. The drug-resistant colonies stably expressing chicken FEN1 were used for cell survival assays and molecular fiber combing analysis.

Colony formation was assayed in a medium containing methylcellulose as previously described (44). To determine the sensitivity to H₂O₂, cells were treated with various concentrations of H₂O₂ for 30 min at 39.5°C in complete medium and then washed twice with PBS. After serial dilution, the cells were plated in triplicate in the medium described above in 6-cm plates and were incubated at 39.5°C for 6 to 7 d. Colonies were counted, and percent survival was determined relative to the number of colonies obtained from nontreated cells.

Cells (3 × 10⁵) were harvested, and DNA was purified using the AquaPure DNA extraction kit (Bio-Rad). The extracted DNA was resuspended in 50 μL of 10 mmol/L Tris buffer with 1 mmol/L EDTA (pH 8.0) and analyzed by electrophoresis on a 2.0% agarose gel in TAE buffer. DNA bands were visualized under UV light after ethidium bromide staining.

For apoptosis analysis, cells were washed twice in cold PBS and resuspended in binding buffer [10 mmol/L HEPES/NaOH (pH 7.4), 140 mmol/L NaCl, and 2.5 mmol/L CaCl₂] at a concentration of 10⁶ cells/mL. One hundred microliters of cells were transferred to a FACScan tube with the addition of 0.1 μg fluorescein-conjugated Annexin V (IQ Corp) and 0.5 μg propidium iodide (PI). Cells were mixed gently and incubated in the dark at room temperature for 15 min. Then, 400 μL of binding buffer was added to the sample, and the sample was analyzed by flow cytometry within 1 h of staining. The data from 1 × 10⁴ cells were collected and analyzed using the LYSIS II software. The signals of green fluorescence (FL1; Annexin V) and orange fluorescence (FL2; PI) were measured by logarithmic amplification.

Cells were treated with H₂O₂ for 10 min, followed by its quenching with catalase (18 U/mL). Cells were further incubated with iododeoxyuridine (IdU) at 100 μmol/L for 20 min. After centrifugation, cells were washed with cold PBS and stored at −80°C until DNA isolation. DNA was isolated from cultured cells using the PureGene DNA extraction kit (Gentra Systems, Inc.). Briefly, cell pellets were thawed and lysed in lysis buffer supplemented with 20 mmol/L 2,2,6,6-tetramethylpiperidinoxyl (TEMPO, Aldrich). Proteins were precipitated and removed by centrifugation, and the DNA/RNA in the supernatant was precipitated with isopropyl alcohol. The DNA/RNA pellet was resuspended in a lysis buffer containing 20 mmol/L TEMPO and incubated with RNase A (100 mg/mL) at 37°C for 30 min, followed by protein and DNA precipitation. The DNA pellet was resuspended in sterilized distilled water with 2 mmol/L TEMPO, and DNA samples were stored at −80°C until analysis. Experimental samples (1 μg) or a reference DNA containing a high level of IdU were denatured in 220 μL TE buffer at 100°C for 10 min. After addition of an equal volume of 2 mol/L ammonium acetate, denatured DNA was applied to nitrocellulose filters using a Minifold II apparatus (Schleicher & Schuell). The filters were soaked in 5× SSC [0.75 mol/L NaCl, 75 mmol/L sodium citrate (pH 7.0)] for 15 min, dried, and baked in a vacuum oven for 2 h at 80°C. Filters were incubated for 2 h at 37°C with blocking buffer [20 mmol/L Tris-HCl (pH 7.5), 0.1 mol/L NaCl, 1 mmol/L EDTA, 0.5% casein, and 0.1% deoxycholic acid] followed by a 2-h incubation at 37°C in a blocking buffer containing anti-bromodeoxyuridine antibody (Becton Dickinson, 1:10,000). The filters were washed extensively in washing buffer [20 mmol/L Tris-HCl (pH 7.5), 0.26 mol/L NaCl, 1 mmol/L EDTA, and 0.1% Tween 20] and treated with a labeled polymer (peroxidase-labeled polymer conjugated to goat anti-rabbit and goat anti-mouse immunoglobulins, Daco) diluted 1:1,000 in a hybridization buffer. After rinsing the membrane, the enzymatic activity on the membrane was visualized by enhanced chemiluminescence reagents (Amersham). The membrane was exposed to X-ray film, and the developed film was analyzed using a Kodak image analysis system.

DNA fiber combing analysis was conducted using logarithmically growing DT40 cells as described (45). Cells were pulsed with IdU, exposed to H₂O₂, and then pulsed with CldU. The cells were harvested and washed. A portion of the cells was lysed on a glass slide and the DNA fibers were straightened (combed) and fixed. The presence of IdU or CldU was detected by immunostaining with red (AlexaFluor 594) and green (AlexaFluor 488) fluorescent antibodies, respectively.

The availability of DT40 cells with homozygous deletion of the FEN1 gene enabled us to investigate the effect of FEN1 on cell tolerance toward oxidative stress. In a cell survival assay, FEN1 null cells were found to be hypersensitive to H₂O₂ (Fig. 1A). The hypersensitivity was complemented by the ectopic expression of chicken FEN1 (Fig. 1B). These results are consistent with the idea that FEN1 plays a role in the repair of H₂O₂-induced DNA damage and that the FEN1 deficiency causes an accumulation of toxic BER intermediates under conditions of H₂O₂-induced stress. Because the FEN1 null cells showed considerable viability in the presence of H₂O₂, FEN1-independent protective pathways are present in these cells.

Because FEN1 functions in in vitro BER (11, 42, 46) in mammalian cells, the requirement of FEN1 in BER was addressed in wild-type DT40 cells and their isogenic counterparts deficient in FEN1. First, to confirm the deletion of FEN1 in DT40-derived FEN1-deletion cells, immunoblotting analysis was done using a specific antibody against FEN1. As expected, FEN1-deletion cells were completely devoid of FEN1 expression (Fig. 2A). To examine a possible role of FEN1 in DT40 cell BER, an extract-based BER assay was done using an oligonucleotide substrate containing 8oxoG as the oxidative DNA lesion initiating BER (Fig. 2B). First, we measured the total BER capacity of wild-type DT40 and FEN1 null cell extracts. The initial rate determination of 8oxoG-DNA repair with FEN1 null cell extract showed 44% reduction in BER activity, compared with wild-type cell extract (Fig. 2C and D). Next, because it is expected that FEN1 is involved in the LP-BER subpathway, the requirement of FEN1 in LP-BER was examined in the DT40 cell extracts. The discrimination between the LP-BER and SN-BER subpathways was achieved using an assay that depends on chain termination after the insertion of the first nucleotide in the LP-BER gap filling. This was accomplished by making use of ddCTP instead of dCTP in the reaction mixture (Fig. 3A). In the SN-BER pathway, dGMP replaces the 8oxo-dGMP lesion, and the resulting BER intermediate is ligated to complete the repair. In contrast, in the LP-BER pathway, ddCTP incorporation follows dGMP insertion, and the LP-BER intermediate accumulates because it cannot be ligated. After separation by 15% PAGE, the SN-BER product is observed as a 100-bp ligated molecule, whereas the LP-BER product is observed as the 24-bp intermediate molecule. Results of the in vitro LP-BER assay for 8oxoG DNA lesion showed that the amount of LP-BER product was reduced (by 65%) in the FEN1 null cell extract (Fig. 3B and C), indicating that FEN1 plays a role in the LP-BER subpathway, but that FEN1 is not essential for LP-BER. This result correlates with the H₂O₂-sensitive phenotype in FEN1 null cells (Fig. 1).

To confirm a possible involvement of FEN1 on the DT40 cell LP-BER, in vitro uracil BER analysis was conducted (Fig. 4A). The in vitro BER analysis is typically evaluated by making use of uracil-containing DNA substrate, revealing a protein involvement in BER (47). This in vitro analysis was applied to the DT40 cell system and revealed the involvement of Pol Q in LP-BER (18). In total BER analysis, the initial rate determination of uracil-DNA repair in FEN1 null cell extract showed a modest reduction (40%) in BER activity, compared with wild-type cell extract (Fig. 4B and C). The in vitro LP-BER analysis was accomplished by the same experimental strategy in the case of 8oxoG LP-BER, except we made use of ddTTP instead of ddCTP and the uracil lesion instead of the 8oxoG lesion (Fig. 5A). Results of the in vitro uracil LP-BER assay confirmed the results of 8oxoG LP-BER assay, indicating that the amount of LP-BER product was reduced in the FEN1 null cell extract (Fig. 5B to D); the amount of SN-BER in wild-type and FEN1 null cell extracts was similar (Fig. 5B). The deficiency in LP-BER of FEN1 null extract was fully complemented by the addition of purified FEN1 (Fig. 5C and D). These results confirm that FEN1 plays a role in the LP-BER subpathway, but indicate that FEN1 is not essential for LP-BER. Based on studies with purified enzymes, the role of FEN1 is to excise the second nucleotide in the lesion-containing strand allowing gap-filling for the two-nucleotide long repair patch (11). When FEN1 is absent, gap filling stalls after one nucleotide is inserted into the gap.

To better understand the mechanism by which FEN1 null cells are hypersensitive to H₂O₂ in the cell survival assay, apoptosis was assessed by DNA fragmentation and flow cytometric analyses. As early as 3 hours after exposure to H₂O₂, fragmentation was detected in the DNA extracted from FEN1-deficient cells. This was completely blocked by BAF, a pan-caspase inhibitor (Fig. 6A). Next, DT40 cells treated with H₂O₂ were stained with a combination of Annexin V and PI, and were analyzed by flow cytometry. By 3 hours after H₂O₂ exposure, an increase in the percentage of early-stage apoptotic cells was observed (Fig. 6B). Taken together, these results indicate that H₂O₂ treatment caused apoptosis in FEN1 null cells within 3 hours of exposure.

To further characterize the effect of H₂O₂ treatment of FEN1 null cells, we developed a quantitative DNA synthesis assay based on IdU immuno-slot-blotting. This assay monitors DNA synthesis by measuring the amount of IdU incorporated into cellular DNA. A standard curve confirmed the ability of this immuno-slot-blotting assay to measure IdU-DNA over a range of IdU levels. In the absence of H₂O₂ exposure, wild-type DT40 cells showed greater IdU incorporation than FEN1 null cells (Fig. 7A and B). Both wild-type and FEN1 null cells showed an inhibition of IdU incorporation after H₂O₂ exposure; the FEN1 null cells showed a reduction of IdU incorporation with H₂O₂ exposure (Fig. 7A and B).

The DNA labeling technique (fiber combing) allows one to distinguish between ongoing replication forks (elongation), newly fired origins (initiation), and fork stalling events by identifying regions that have both labels, the second label only, or the first label only, respectively (Fig. 8A and B). At its level of sensitivity, this technique does not allow detection of tracks shorter than 500 bp. Therefore, any tracks incorporated during LP-BER would be completely undetected by fluorescence. The relative proportions of tracks with different label combinations can be used to evaluate interference with replication fork progression after DNA damage (Fig. 8C; ref. 48). As can be seen in Table 1 and Fig. 9, exposure of wild-type DT40 cells to H₂O₂ did not increase the relative percentage of IdU-only (red) tracks, suggesting that H₂O₂-induced damage was not a strong enough impediment to stop replication (i.e., cause the replication fork to terminate prematurely). However, H₂O₂ exposure did reduce the frequency of CldU-only (green) tracks (i.e., the firing of new origins). In contrast, exposure of FEN1 null cells to H₂O₂ both increased termination of replication and decreased replication initiation. This indicated that after H₂O₂-induced DNA damage, the signals for checkpoint activation were greater in the FEN1 null cells, causing an increase in prematurely terminated replication forks and blocking initiation of forks. The ectopic expression of FEN1 reduced the amount of prematurely terminated replication forks after H₂O₂ exposure (Fig. 9A). We conclude that FEN1 was required to sustain replication fork progression under conditions of H₂O₂-induced oxidative stress.

The importance of FEN1 in LP-BER has been investigated in vivo using S. cerevisiae lacking Rad27: the deletion of the Rad27 gene caused hypersensitivity to MMS in cell survival assays. This hypersensitivity was reversed almost completely by the disruption of the APN1 gene, a major AP endonuclease in yeast, indicating a significant role of FEN1 in the yeast BER pathway (49). However, from these experiments, it is difficult to address the role in higher eukaryotes because the enzymes and subpathways of BER seem to be significantly different from those in yeast (50). DT40 cells, on the other hand, have both SN- and LP-BER subpathways, express poly ADP ribose polymerase isoforms and other BER cofactors, and thus represent a valuable model for dissecting the roles of specific BER proteins in higher eukaryotes.

Using cell extract from wild-type DT40, we found that both LP-BER and SN-BER subpathways could repair our BER substrate (∼45% of the substrate was repaired by LP-BER and ∼55% by SN-BER), and these results were similar to those observed with mammalian cells (51). The absence of FEN1 caused the rate of overall BER to decrease by 65% for oxidative DNA damage (Fig. 2D) and 40% for the uracil lesion (Fig. 4C), and this seemed to be due to a reduction in the LP-BER subpathway (Figs. 3 and 5). Addition of purified FEN1 complemented the deficiency in LP-BER capacity in the FEN1 null cell extract (Fig. 5B). We also found that LP-BER can occur even in the absence of FEN1 for both types of DNA damage (Figs. 3C and 5C), indicating that there is an alternate pathway that can assume the function of FEN1. This may explain why FEN1 null DT40 cells are viable.

Our results showed that although FEN1 is dispensable in DT40 cells under physiologic conditions, cells need FEN1 to survive increased oxidative stress. To gain insight into this hypersensitivity phenotype, we studied global levels of DNA replication. We observed that replication in wild-type DT40 cells was ∼40% lower under oxidative stress and was depressed even further in FEN1 null cells. This was consistent with a linkage between the repair role of FEN1 and DNA replication in the presence of oxidative stress. This linkage most likely was through the removal of oxidative damage in the replicating template. To gain further insight on this point, we used the fiber combing technique (Table 1; Figs. 8 and 9) to visualize and characterize the stage of replication that was affected after cells were subjected to oxidative stress. We found that fork progression in FEN1 null cells was slower than in the wild-type DT40 cells (data not shown). This reduction of fork movement was not severe enough to increase the number of premature fork terminations, indicating that DNA replication could proceed, albeit more slowly. When wild-type cells were exposed to H₂O₂, the number of newly fired origins was reduced, suggesting that H₂O₂ exposure elicited a checkpoint response (52, 53). FEN1 null cells exhibited a much stronger decrease in origin initiation, suggesting a stronger checkpoint response to oxidative stress.

H₂O₂ can initiate the generation of highly reactive hydroxyl radicals through the transition metal–catalyzed Haber-Weiss reaction (8, 54). We previously showed the biphasic induction of oxidative DNA damage by H₂O₂ with a nearly linear increase in DNA damage between 60 and 600 μmol/L H₂O₂ in HeLa cells (8). Concentrations of H₂O₂ that are able to induce oxidative DNA damage can be different among cell lines, most likely because of differences in antioxidant capacity. Oxidants, in general, are known to induce both apoptosis and necrosis in cells (55, 56), with the concentrations required being dependent on the cell type under study. For example, in response to 50 μmol/L H₂O₂, Jurkat T-lymphocytes underwent apoptosis within 6 hours, as measured by flow cytometry (57), and the phenomenon was associated with caspase activation. Human Burkitt lymphoma (B-cell origin, Raji) cells were more resistant to H₂O₂ compared with Jurkat T-lymphocytes: Raji cells underwent apoptosis after treatment with 0.5 mmol/L H₂O₂ for 24 hours (58), and there was no detectable cleavage of caspase-3 in Raji cells treated with 0.05 and 0.5 mmol/L H₂O₂ for up to 9 hours (59). In the present study, caspase-dependent apoptosis was observed in FEN1 null cells by 3 hours after exposure to 10 μmol/L H₂O₂. Because we made comparisons in isogenic DT40 cell lines, we conclude that the hypersensitivity of FEN1 null cells to H₂O₂-induced apoptosis was due to the lack of FEN1 function.

FEN1-dependent LP-BER is believed to be the predominant BER pathway for removing a variety of 5′-end lesions resistant to the β-elimination by POLβ. Deoxyribose structures, including chemically reduced abasic sites, abasic site analogues (e.g., tetrahydrofuran), and sugar lesions arising from either C-1′ or C-2′ oxidation of deoxyribose that have been initially cleaved on the 5′ side by APE1 are refractory to the excision by POLβ (60, 61). Because the chemical structures of these 5′-end lesions prevent β-elimination by POLβ, they will be substrates for FEN1-dependent repair by LP-BER (62). Additionally, deoxyribose lesions oxidized at the C-5′ position, which still retain the nucleobase (63), also are substrates for FEN1-dependent LP-BER. In the absence of FEN1, abasic sites from C-1′–oxidized deoxyribose at the 5′-end of an SSB may lead to the formation of DNA-protein cross-links especially between POLβ and C-1–oxidized abasic sites (60, 62). Such DNA-protein cross-links in cells could cause chromosomal aberrations and cell death. Our results, obtained from in vitro experiments, clearly support the idea that FEN1-dependent LP-BER plays a significant role in repairing oxidized DNA sugar damage in the context of the cell. We previously showed that endogenous abasic sites likely include oxidized deoxyribose (7). The present study suggests an important role of FEN1 in the repair of oxidized abasic sites resistant to β-elimination by POLβ

In conclusion, using the DT40 cell system, we showed that cell extract from FEN1 null cells was normal in SN-BER activity but was reduced in the LP-BER subpathway activity; some FEN1-independent LP-BER activity exists in DT40 cells; FEN1 deficiency reduced DNA replication in cells under physiologic conditions, as well as increased the amount of stalled replication forks under oxidative stress; and in the absence of FEN1, oxidative stress induced apoptotic cell death. These results support the hypothesis that FEN1 plays an important role in the LP-BER subpathway in higher eukaryotes.

No potential conflicts of interest were disclosed.

We thank Sankar Mitra, Bruna Brylawski, Shuji Yonei, Tadayoshi Bessho, Brian Pachkowski, William Kaufmann, and Keith Caldecott for critically reading the manuscript, and Bonnie E. Mesmer and Noriko Tano for their expert secretarial assistance.

Grant Support: Center for Environmental Health and Susceptibility Pilot Project Program, P42-ES05948, ES11746, P30-CA16086, P30-ES10126 from the NIH, by the Intramural Research Program of the NIH, and NIEHS (Z01-ES050158 and Z01-ES050159), by a grant from the National Cancer Institute (R01-CA084493), and by Grants-in-Aid for Scientific Research from the Ministry of Education of Japan (K. Tano).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.