Induction of the Human Oxidized Base-specific DNA Glycosylase NEIL1 by Reactive Oxygen Species

NEIL1, mammalian DNA glycosylase and ortholog Escherichia Nei/Fpg, is involved in the repair of oxidatively damaged bases in mammalian cells. Exposure of HCT116 human colon carcinoma cells to reactive oxygen species, generated by glucose oxidase (GO), enhanced the levels of NEIL1 mRNA and polypeptide by 2–4-fold by 6 h after GO treatment. A similar oxidative stress-in-duced increase in human NEIL1 ( hNEIL1 ) promoter-dependent luciferaseexpressioninHCT116cellsindicatesthatreactiveoxygen species activates NEIL1 transcription. The transcriptional start site of hNEIL1 was mapped, and the upstream promoter sequence was characterized via luciferase reporter assay. Two identical CRE/AP-1-bindingsiteswereidentifiedinthepromoterthatbindstranscrip- tion factors c-Jun and CREB/ATF2. This binding was significantly enhancedinextractsofcellstreatedwithGO.Furthermore,asimul-taneous increase in the level of phosphorylated c-Jun suggests its involvement in up-regulating the NEIL1 promoter. Oxidative stress-induced activation of NEIL1 appears to be involved in the feedback regulation of cellular repair activity needed to handle an increase in the level of oxidative base damage. for luciferase activity using a dual-luciferase reporter assay system (Promega) as per the manufacturer’s instructions. All experiments were performed at least in triplicate, and the data were normalized to co-transfected Renilla luciferase (con-trol reporter) to correct for variation in transfection efficiency. Co-ex-Oxidative

ROS, 2 generated endogenously and as a result of environmental insult, induce oxidative DNA damage, which could affect the integrity of cellular genomes. Oxidative DNA damage, including a multitude of damaged bases, abasic (activator protein (AP)) sites, and DNA strand breaks, is believed to be responsible for sporadic mutations leading to cancer and other pathological as well as age-related syndromes (1)(2)(3)(4). Most of the oxidative damage in DNA is repaired via the base excision repair pathway, in which oxidized and altered bases are excised by DNA glycosylases in the first step of repair (5). Base excision repair is highly conserved among all organisms ranging from Escherichia coli to humans. In E. coli, three oxidatively damaged base-specific glycosylases, endonuclease III (Nth), formamidopyrimidine-DNA glycosylase (Fpg), and endonuclease VIII (Nei), excise modified DNA bases with overlap-ping specificity. Based on structural similarity and reaction mechanism, these enzymes are classified in two families with Nth in one and Nei and Fpg in the other (6 -8). In mammalian cells, only two DNA glycosylases, both belonging to the Nth family, were identified earlier for excision of oxidized bases. NTH1 removes mostly oxidized pyrimidines, e.g. thymine glycol, whereas 8-oxoguanine DNA glycosylase (OGG1), functionally similar to Fpg but mechanistically analogous to Nth, is most active with 8-oxoguanine and formamidopyrimidine-guanine substrates (9,10). Most surprisingly, both NTH1-and OGG1-null mice are viable and show no major phenotype despite the known mutagenic and toxic effects of their substrate lesions (11,12). These results suggested the presence of additional DNA glycosylases in mammals that could serve as back-up enzymes for NTH1 and OGG1.
We and subsequently several other groups identified and characterized two additional human DNA glycosylases specific for oxidatively damaged bases and named them NEIL1 and NEIL2 (13)(14)(15)(16)(17). Both of these enzymes belong to the Fpg/Nei family; NEIL1 and NEIL2 share weak sequence similarity but have common characteristics, including overlapping substrate ranges (15)(16)(17)(18)(19). However, NEIL1 differs from NEIL2 because of its S-phase-specific activation, whereas NEIL2 expression is independent of the cell cycle (13,14). We have shown recently that both NEIL1 and NEIL2 possess an unusual ability to excise lesions from the DNA bubble or single-stranded structures that could not be carried out by OGG1 or NTH1 (20). This novel substrate structure preference of NEILs raises the possibility of their specialized in vivo functions for repair of oxidized bases in transient bubbles formed during replication and/or transcription.
Taken together, these results support the scenario that OGG1 and NTH1 are involved primarily in repair of oxidative damage formed in inactive sequences constituting the bulk of the genome. In contrast, NEIL1 and NEIL2 may be preferentially involved in repair of oxidized bases from the active sequences during replication or transcription, which should be more urgent than the repair of inactive sequences. In this report, we examined the possibility of modulation of NEIL1 activity, and we observed that human NEIL1 (hNEIL1) is activated by ROS. We then characterized the hNEIL1 promoter and identified a pair of CRE/ AP-1 sequences involved in the oxidative stress response.

EXPERIMENTAL PROCEDURES
Cell Culture and Chemicals-The human colorectal carcinoma line HCT116 (with wild type p53), a gift of B. Vogelstein (The Johns Hopkins University), was cultured in McCoy's 5A modified medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, penicillin (100 units/ml), and streptomycin (100 g/ml; Invitrogen). At 70% confluence, the cells were treated with glucose oxidase (GO; Roche Applied Science) at 100 ng/ml for 1 h, followed by washing with and incubation in fresh medium. This treatment did not affect cell viability, as judged by the rate of cell growth.
Measurement of Intracellular ROS Level-Changes in intracellular ROS level after GO treatment were determined as described previously (21). Briefly, the cells were suspended in phosphate-buffered saline and then treated with 5 M 5-(and-6) carboxy-2Ј,7Ј-dichlorodihydrofluorescein diacetate (H 2 DCF-DA; Molecular Probes) for 15 min at 37°C. In control experiments, we determined that cellular uptake of H 2 DCF-DA reached a plateau at about 18 min when 0.6% of DCF, the oxidized form of H 2 DCF, was released in the medium (data not shown). As a control, we also treated the cells with DCF, and we observed a small increase in intracellular fluorescence (Ͻ5%) after glucose oxidase treatment (data not shown). The change in DCF fluorescence of treated versus mocktreated cells was determined by flow cytometry (BD Biosciences FAC-Scan). The mean fluorescence for 12,000 cells, from three or more independent experiments, was analyzed, after correction for DCF extrusion in the medium, and expressed as ϮS.E. (Fig. 1).
Western Analysis-Cells were washed with ice-cold phosphate-buffered saline, lysed on ice for 30 min in a lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 1 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture; Roche Applied Science), and then microcentrifuged. Western analysis of the supernatant was carried out with antibodies to hNEIL1 (Alpha Diagnostics, San Antonio, TX), CREB, ATF2, c-Jun, and phospho-c-Jun (Santa Cruz Biotechnology), using enhanced chemiluminescence assay (ECL kit; Amersham Biosciences); the signals were quantitated by densitometric scanning.
RT-PCR Analysis-RNA extraction was carried out with the RNeasy mini kit (Qiagen) according to the manufacturer's instructions. Total RNA (2-5 g) was reverse-transcribed using a Titan one-step RT-PCR kit (Roche Applied Science), followed by a limiting cycle amplification protocol using the following primer combinations: hNEIL1 forward, 5Ј-CGG CGG CTG CGT GGA GAA GTC-3Ј, and reverse, 5Ј-GTC CCA GCG GCC GAA CCG GCG-3Ј, designed to amplify a 300-bp region; GAPDH forward, 5Ј-GTG AAG GTC GGA GTC AAC-3Ј, and reverse, 5Ј-GGT GAA GAC GCC AGT GGA-3Ј, designed to amplify a 294-bp region (total number of cycles: 24 for hNEIL1 and 20 for GAPDH). The products of the semiquantitative RT-PCR were resolved in a 1.8% agarose gel, stained with ethidium bromide, and photographed.
For quantitative, real time RT-PCR, one-step RT-PCR was performed with 100 ng of cellular RNA for both the target gene and an endogenous control in single-plex tubes using TaqMan one-step RT-PCR master mix reagent kit (P/N 4309169). The cycling parameters (for 40 cycles) in an ABI7000 thermal cycler were as follows: reverse transcription at 48°C for 30 min, AmpliTaq activation 95°C for 10 min, denaturation 95°C for 15 s, and annealing/extension 60°C for 1 min. Duplicate C T values were calculated using Microsoft Excel and using a comparative C T (⌬⌬C T ) method as described by the manufacturer (Applied Biosystems). The amount of target (2 Ϫ⌬⌬CT ) was calculated after normalization to the 18 S RNA as an endogenous reference, and relative to a calibrator (one of the experimental samples).
Electrophoretic Gel Mobility Shift Analysis (EMSA)-Nuclear extracts from HCT116 cells were prepared as described earlier (22). For EMSA, 5 g of nuclear extract was incubated with 2 pmol of 5Ј-32 Plabeled duplex oligonucleotides and 1 g of poly(dI-dC)⅐(dI-dC) in a binding buffer containing 4% glycerol, 1 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM dithiothreitol, 50 mM NaCl, and 10 mM Tris-HCl, pH 7.5, for 30 min at room temperature, followed by electrophoresis in 6% nondenaturing polyacrylamide gel in 0.5 ϫ Tris borate/EDTA (TBE) at 200 V for 1 h. The gels were dried and subjected to PhosphorImager analysis. Competition of binding was performed in the presence of a 50-or 100-fold excess of the unlabeled oligonucleotide and incubating for 20 min at room temperature before adding the probe. Supershift assays were performed by preincubating the assay mixture with 4 g of antibody against CREB/ATF2 or c-Jun for 1 h at 4°C before the addition of 32 P-labeled oligonucleotide duplex probe.
Genomic Cloning of the 5Ј-Upstream Region of hNEIL1-An NdeI/ XhoI fragment of the hNEIL1 cDNA was used to screen a BAC DNA library (Resgen) under stringent conditions. Restriction mapping and Southern blot analysis of BAC clones identified 18-kb HindIII, 13-kb EcoRI/HindIII, and 12-kb BamHI/HindIII fragments that hybridized to the probe. The 13-kb EcoRI/HindIII fragment was further subcloned into pBluescript and sequenced. The sequences of the genomic clone were confirmed by matching with the published NEIL1 gene sequence (GenBank TM accession number AAH10876). Analysis of candidate ciselements to identify relevant transcription factors was performed using the programs version 2.2. Parameters for MatInspector were set for 1.0 core similarity (a 4-nucleotide highly conserved sequence) and 0.85 matrix similarity, employing the vertebrate matrix group.
Primer Extension Analysis-An antisense primer, P2 (5Ј-GGCT-GAGGCAGGAGAATC-3Ј), complementary to the 5Ј-untranslated region of the hNEIL1 cDNA, near the beginning of exon 1, was 5Ј-32 Plabeled using [␥-32 P]ATP and T4 polynucleotide kinase and annealed to 50 g of total RNA at 58°C for 20 min. The primer extension was performed with the annealed RNA DNA duplex, using 5 units of avian myeloblastosis virus-reverse transcriptase (Promega) at 42°C for 1 h in the supplied extension buffer, and the products were resolved in a 6% polyacrylamide sequencing gel. A control reaction was set up with a second antisense oligo P1, further downstream to the P2 sequence. Sequencing reactions, generated by Sanger's dideoxy termination method, using primer P2 annealed to the hNEIL1 plasmid, were run in parallel to determine the exact length of the primer extension products.
Plasmids, Transfection, and Luciferase Assay-A 2.1-kb fragment spanning the upstream sequence of hNEIL1, including part of the first exon, was subcloned from the 13-kb genomic clone at the XhoI/HindIII sites of the promoterless luciferase vector pGL2-basic (Promega), and the recombinant plasmid was named p(Ϫ2100) hNEIL1 luc. Starting with this plasmid, we generated several deletion constructs by PCR amplification, with sense primers representing different segments of the 5Ј regions of hNEIL1; the 5Ј deletion fragments were subsequently fused upstream of the luciferase gene in the expression vector. The sequences of sense primers to generate these reporter plasmid constructs are given in TABLE ONE. All of these were used in combination with a common antisense primer, 5Ј-CTT CCC GGG TTC AAG TGA TTC TCC TGC CTC-3Ј. Point mutations were generated in p(Ϫ1800/ϩ40)NEIL1 luc with the QuikChange site-directed mutagenesis kit (Stratagene) per the manufacturer's instructions. The following oligo sets used are as follows (mutated nucleotides are underlined): distal CRE/AP-1 site, 5Ј-GGC AGG CAG ATC ACC TGA GGT TGG GAG TTT GAG ACC AGC-3Ј and 5Ј-GCT GGT CTC AAA CTC CCA ACC TCA GGT GAT CTG CCT GCC-3Ј; proximal CRE/AP-1 site, 5Ј-GGC GGG CAG ATT ACT TGA GG T TGG GAG TTC AAG ACC AGC-3Ј and 5Ј-GCT GGT CTT GAA CTC CCA ACC TCA AGT AAT CTG CCC GCC-3Ј. The nucleotide sequences of both deletion and mutant constructs were verified by sequencing. The HCT116 cells were transfected with the reporter plasmids using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. The cells were harvested 36 h after transfection, and their extracts were assayed for luciferase activity using a dual-luciferase reporter assay system (Promega) as per the manufacturer's instructions. All experiments were performed at least in triplicate, and the data were normalized to co-transfected Renilla luciferase (control reporter) to correct for variation in transfection efficiency. Co-ex-pression experiments were performed using expression vectors containing full-length cDNAs of the transcription factor c-Jun.
Statistical Analysis-In all experiments, including densitometric scans of Western blots or RT-PCR analysis and luciferase assays, results were calculated as the means (ϮS.D.) of three independent experiments. Statistical analyses were performed using one-way analysis of variance for comparisons of multiple groups. p values of 0.05 or less were deemed to be statistically significant. All statistical tests were done using SigmaStat software (Jandel Scientific).

RESULTS
Oxidative Stress Transiently Increases NEIL1 Protein and mRNA Levels in HCT116 Cells-To explore the effects of oxidative stress on NEIL1 expression, we treated HCT116 cells for 1 h with a subtoxic dose (100 ng/ml) of the ROS generator GO, and we then analyzed NEIL1 protein and mRNA levels at various times. GO treatment caused 2.4 Ϯ 0.3-fold increase in the intracellular ROS level in these cells after 1 h (Fig. 1). Mock and GO-treated cells were harvested at 0, 3, 6, 9, 12, and 20 h. Western analysis of whole cell and more significantly nuclear extracts showed an increase in the NEIL1 protein level within 3 h of GO treatment, reaching up to 2-4-fold over the control at 6 -9 h, and before returning to the base line by 20 h (Fig. 2). The OGG1 polypeptide level remained unchanged over a similar time course of GO treatment ( Fig.  2A). The GO treatment similarly increased the NEIL1 mRNA level in HCT116 cells (up to 4-fold at 6 -9 h), as analyzed by both semi-quantitative (Fig. 3A) and quantitative RT-PCR (Fig. 3B). The GO treatment was performed in parallel in MRC5, human primary fibroblasts. Real time RT-PCR analysis of RNA revealed a similar induction of the transcript level upon exposure to ROS (Fig. 3B).
Identification and Characterization of cis-Elements in the hNEIL1 Promoter-To analyze further the basis for ROS-induced activation of the NEIL1 gene, we characterized its cis-elements. We cloned a 13-kb EcoRI/HindIII genomic fragment containing the complete transcription unit of the NEIL1 gene into pBluescript, together with a significant part of the 5Ј-flanking region. Comparison of the genomic sequences of hNEIL1 to its cDNA showed that the transcribed sequence spans ϳ8 kb and contains 10 exons. The transcriptional start site was determined by primer extension analysis; the longest product observed extended 40 bp from the primer, P2, indicating a G in the sense strand as the site most frequently used for transcription (Fig. 4). Primer P1 failed to show any primer extension product. The initiation codon of the NEIL1 polypeptide was located in the 2nd exon, exon 1 (1 kb) being the untranslated region. Transfac software analysis of a 2.1-kb fragment spanning upstream sequences, including part of the 1st exon, identified several putative cis-elements for transcription factors. It is interesting to note the presence of binding sites for several b-Zip family proteins, including c-Jun, CREB, and NF-E2/AP-1-related factors (Fig. 5A).  To test for basal promoter activity of the upstream sequences, a NEIL1 promoter luciferase reporter construct p(Ϫ2100/ϩ40)NEIL1 luc, containing NEIL1 genomic sequence Ϫ2100 to ϩ40 bp, was used for transient expression in transfected HCT116 cells, as described under "Experimental Procedures." A series of nested 5Ј deletion mutants of the NEIL1 promoter was subsequently generated in pGL2-basic (as shown in TABLE ONE and Fig. 5B), and the luciferase activity was measured. Deletion of the region between Ϫ1800 and Ϫ900 bp caused a 2-3-fold decrease of luciferase activity. Additional deletion of the downstream sequence to Ϫ600 or Ϫ500 bp did not further decrease promoter activity, whereas deletion of the region spanning Ϫ300 and Ϫ100 bp resulted in another 3-fold decrease in promoter activity, indicating that this region is the major contributor to NEIL1 expression. Furthermore, the minimal promoter activity could be detected in the plasmid containing the Ϫ50/ϩ40 sequences.
Oxidative Stress Transactivates the NEIL1 Gene Promoter-To identify the promoter region essential for transcriptional up-regulation of NEIL1 during oxidative stress, HCT116 cells were transfected with three different NEIL1 reporter constructs: p(Ϫ1200) NEIL1 luc (bearing both proximal and distal CRE/AP-1 elements); p(Ϫ900) NEIL1 luc (bearing only proximal CRE/AP-1); and p(Ϫ50) NEIL1 luc (lacking both CRE/AP-1 elements). The luciferase activity was measured in extracts of transfected cells after GO treatment (Fig. 6A). A Ͼ2-fold increase in luciferase activity was observed with the p(Ϫ1200) NEIL1 luc construct within 3 h after exposure to GO. This up-regulation reached its peak at 6 h (Ͼ3.5-fold) and returned to the basal level at 20 h. Similarly, an ϳ3-fold induction was observed with p(Ϫ900) NEIL1 luc after exposure to GO. However, the promoter-reporter construct p(Ϫ50) NEIL luc, containing the minimal promoter sequence, showed no up-regulation in luciferase activity under these conditions. These results indicate that the cis-acting elements responsible for oxidative stress-induced NEIL1 promoter activity are located 60 -1200 bp 5Ј to the transcriptional start site.
Oxidative Stress Activates the NEIL1 Promoter via a Pair of Proximal and Distal CRE/AP-1 Sequences-Examination of the sequence within the Ϫ60to Ϫ1200-bp region of the NEIL1 promoter revealed two copies of putative elements at positions Ϫ61 and Ϫ1018, both with sequence similarity to the palindromic consensus CRE (TGACGTCA) and AP-1 (TGACTCA). To test whether these CRE-like sites are required for oxidative stress-induced expression of NEIL1, we mutated these sites either individually or simultaneously and assayed the promoter activity, following GO treatment (Fig. 6B). Mutation of the proximal CRE (Mut1), showed an ϳ4-fold decrease in oxidative stress-induced promoter activation, whereas mutation of distal CRE/AP-1 (Mut2) showed a 2-fold decrease compared with the wild type. Moreover, mutation at both CRE/AP-1 sites (Mut3) further decreased the GO-induced promoter activation by ϳ6-fold. These observations suggest that the two CRE/AP-1 sites are independently involved in oxidative stress-mediated activation of the NEIL1 promoter.

Oxidative Stress Stimulates Binding of CREB/c-Jun to NEIL1
CRE/AP-1-To identify the transcription factors interacting with NEIL1-CRE/AP-1, we incubated nuclear extracts with 32 P-labeled oligo duplexes of sequences corresponding to the proximal and distal NEIL1-CRE sites, respectively. As shown in Fig. 7, two closely spaced complexes were observed in both instances (indicated by arrows) that could be eliminated by addition of 100-fold molar excess of unlabeled NEIL1specific or consensus AP-1 oligo. Lack of competition with a similar excess of mutated AP-1 oligonucleotides established the binding spec-ificities of the proximal as well as distal NEIL1-CRE/AP-1 sequences. Similar results were obtained when the CRE consensus and mutant oligonucleotides were used as competitors in EMSA (data not shown). Additionally, no specific binding was observed with the mutated NEIL1-CRE/AP-1 oligo (Fig. 7, last lane in panel i).
We then investigated whether oxidative stress leading to activation of the NEIL1 promoter was associated with increased binding of the CRE/ AP-1 complex to the cis-element. Treatment of HCT116 with GO resulted in a significant increase in the binding of the nuclear extract to the NEIL1 CRE/AP-1 (Fig. 7, A and B, panel ii), as early as 30 min after treatment, which returned to the basal level within an hour. To identify the proteins present in the binding complexes, supershift analyses were performed with antibodies specific for various members of the CREB/ ATF and AP-1 families (Fig. 7, A and B, panel iii). Both anti-Jun (c-Jun/ AP-1) and CREB-1 antibodies decreased the amount of the specific DNAprotein complexes. A pronounced supershifted complex was generated with the phospho-c-Jun antibody (KM1). Incubation of nuclear extracts with ATF1 antibody did not generate a supershifted band, whereas ATF2 antibody slightly decreased the level of the specific complex. With extracts of stressed cells, the CRE complex was completely supershifted with addition of phosphorylated c-Jun-specific antibody (KM1), suggesting that under oxidative stress the majority of c-Jun bound to the NEIL1-CRE was in the phosphorylated form. We have further concluded that NEIL1-CRE complex responsive to oxidative stress consists of c-Jun and CREB proteins.
GO Treatment Stimulates Phosphorylation of c-Jun-We then investigated whether oxidative insult by GO treatment, which induced NEIL1 promoter activity and CRE protein binding, increased the levels of CREB, c-Jun, and ATF2 polypeptides. Western blot analysis of GOtreated cells showed that the CREB and ATF2 levels were slightly enhanced, but a marked increase in the level of phosphorylated c-Jun was observed within 30 min of exposure to GO (Fig. 8A). These results indicate that CRE/AP-1 activation was correlated with oxidative stressinduced up-regulation of NEIL1 expression.   -GGT TAC CCA TGT TGT TGC ATG TGG CAG-3Ј  p(Ϫ1800/ϩ40) NEIL1 luc  1200-luc sense  5Ј-GAA CCA TAT GTT TTC TAA TGC AGA GG-3Ј p(Ϫ1200/ϩ40) NEIL1 luc 900-luc sense 5Ј-CTA CTC AGG AGG CTG AGG CAA GAG-3Ј p(Ϫ900/ϩ40) NEIL1 luc 600-luc sense 5Ј-GTG ATG CAG CAT CTA TCT CAA ACA-3Ј p(Ϫ600/ϩ40) NEIL1 luc 300-luc sense 5Ј-CTA CTT GGG AGG CTG AGG CTG AGG-3Ј p(Ϫ300/ϩ40) NEIL1 luc 100-luc sense 5Ј-CTC ACT TCT GTA ATC CCA GCA CTT-3Ј p(Ϫ100/ϩ40) NEIL1 luc 50-luc sense 5Ј-GGT CAG GAG TTC AAGACC AGC CTG-3Ј p(Ϫ50/ϩ40) NEIL1 luc c-Jun Induces Both Basal and Oxidative Stress-induced Promoter Activity of NEIL1-The EMSA studies suggested that the oxidative stress-induced enhancement of the affinity of nuclear extracts for the NEIL1 promoter was due to activation of c-Jun. We directly tested this possibility by examining the effect of c-Jun overexpression on NEIL1 promoter activity. HCT116 cells were co-transfected with a c-Jun expression plasmid and p(Ϫ1800/ϩ40)NEIL1 luc containing both copies of CRE/AP-1. c-Jun dose-dependent increase in luciferase activity was observed, and GO treatment further increased the activity by about 2-fold (Fig. 8B). Stimulation of activity due to c-Jun overexpression was significantly attenuated when both CRE/AP-1 were mutated in the NEIL1 promoter construct, Mut3 (Fig. 8B). This confirmed that promoter activation was mediated by the CREs.

DISCUSSION
We observed oxidative stress-dependent activation of hNEIL using sublethal GO treatment for sustained ROS generation. An increase in the levels of both NEIL1 mRNA and polypeptide suggested that ROS enhances NEIL1 transcription. To elucidate the molecular basis of NEIL1 induction, we first characterized the hNEIL1 promoter. Func-tional analysis of nested deletions of the NEIL1 gene upstream of the coding region revealed that the basal promoter is localized in a 50-bp segment 5Ј to the transcription start site. However, robust promoter activity induced by the Ϫ1200 construct indicated the presence of major cis-elements in the distal region of the promoter. Analysis of the sequence spanning Ϫ1800 to ϩ40 bp of the promoter suggested the presence of cis-elements specific for various members of the b-Zip family of transcription factors. The b-Zip proteins are characterized by the presence of a basic domain required for interaction with DNA and an adjacent leucine zipper domain that allows dimerization between the family members (24). We identified a pair of nearly identical DNA sequences that differ by a single nucleotide from the palindromic consensus CRE, a target for binding of the b-Zip proteins CREB and ATF2 (25)(26)(27). Additionally, this CRE-like sequence is similar to the AP-1 element (28), and so could serve as a target for the binding of c-Jun and or c-Fos transcription factors, which also belong to the b-Zip family. To identify the specific trans-acting elements, we performed EMSA with a NEIL1 CRE oligo and nuclear extracts. Supershift of the specific complex with anti-c-Jun antibody and decrease in binding of nuclear extract in the presence of anti-CREB and anti- ATF2 antibodies confirmed the presence of these proteins in the binding complex.
Further analysis of promoter reporter constructs containing both CRE/AP-1 sequences in the NEIL 5Ј region showed a significant increase in luciferase activity relative to that with a deletion construct containing only the proximal CRE/AP-1. Oxidative stress activated both, whereas the minimal promoter lacking the CREs was not activated by oxidative stress. Mutational analysis further showed the requirement of key bases in both CRE/AP-1 for response to oxidative stress. These findings confirm that GO induces hNEIL1 expression via transactivation of the NEIL1 promoter. To the best of our knowledge, this is the first evidence for a role of CRE/AP-1 in NEIL1 expression.
In addition to their role in the regulation of a variety of genes, the CRE/AP-1 family of transcription factors are among the most prominent regulators of the oxidative stress response. Furthermore, the consensus sequence recognized by the classic CRE/AP-1 family members, 5Ј-TGA(G/C)T(C/A)A-3Ј, is similar to the antioxidant-response element core sequence 5Ј-GTGACNNNGC-3Ј. The presence of similar sequences in the 5Ј region of the NEIL1 promoter strongly suggests that NEIL1 induction upon cellular exposure to ROS is regulated by binding of the members of the CREB/c-Jun family of transcription factors similar to these antioxidant-response element-like upstream elements. Our studies also confirmed that c-Jun and the other related transcription factors are induced by GO treatment. Further studies will be needed to determine the mechanism by which GO promotes CREB/c-Jun binding in order to identify other members of the signaling pathway.
Several other mammalian repair genes are inducible by oxidative/ genotoxic stress. The O 6 -methylguanine-DNA-methyltransferase (MGMT) gene is transcriptionally activated by different types of genotoxic agents, including alkylating drugs and ionizing radiation (29,30). The MGMT promoter contains two AP-1-binding sites, which play a key role in MGMT basal regulation (31). However, the elements required for MGMT induction remain to be identified. DNA polymerase ␤ is induced by simple methylating agents, such as MNNG which is also mediated through a CREB (32). The polymerase ␤ expression level is also enhanced by oxidative treatments (33). Similarly expression of mammalian APE1 (AP endonuclease 1) was induced by oxidative stress (34,35). The promoter of the human APE1 contains consensus sequences for binding NF-B, CRE, and AP-1. Similar to our observation for NEIL1, transcriptional activation of the APE1 gene was shown to require CREB/c-Jun (36). The human OGG1 promoter was recently shown to contain a pair of inverted CCAAT motifs that are crucial to its induction by the DNA-alkylating agent methylmethane sulfonate (37). Furthermore, the transcription factor NF-YA was shown to interact with the methylmethane sulfonate regulatory elements. Earlier the yeast NTH1 gene was shown to contain multiple stress-response elements in its promoter (38). However, we have shown here that OGG1 was not activated by oxidative stress. The teleological basis for activation of OGG1 and NTH1 by nonoxidants is not clear.
In summary, our results demonstrate that ROS up-regulate hNEIL1 expression through activation of the transcription factors, CREB/c-Jun. An important question that remains to be answered is the biological role of hNEIL1 activation in response to oxidative stress. We hypothesize that endogenous ROS induce DNA damage at a basal, steady-state level in unstressed cells. However, exogenous oxidative stress, although increasing the genomic damage level, simultaneously activates the cellular response to repair the additional load of DNA damage. NEIL1 activation is a part of this coordinated cellular response. This scenario implies that NEIL1 or other proteins activated by oxidative stress are limiting in the repair pathway. It is possible that NEIL1 has a unique role among the oxidized base-specific mammalian DNA glycosylases, including OGG1 and NTH1, by preferentially repairing damage in active genes. Lack of OGG1 activation by ROS underscores the key difference in regulation of NEIL1 and OGG1. This is further supported by the observations that OGG1-or NTH1-null mouse fibroblasts are no more sensitive to ROS than the parental wild type cells (12,39). In contrast, down-regulation of NEIL1 expression by small interfering RNA in mouse embryonic stem cells sensitizes these cells to ␥-irradiation (40). Thus the fine-tuning of NEIL1 expression, both in response to cell cycle or external oxidative stress, may be more critical than of