Cadmium down-regulates human OGG1 through suppression of Sp1 activity.

Cadmium is a well known human and animal carcinogen and is a ubiquitous contaminant in the environment. Although the carcinogenic mechanism of cadmium is a multifactorial process, oxidative DNA damage is believed to be of prime importance. In particular, cadmium suppresses the capacity of cells to repair oxidative DNA damage. In this study, cadmium treatment led to a significant increase in gamma-ray-induced 8-oxoguanine (8-oxoG) formation. Western blotting and semiquantitative reverse transcription-PCR revealed that cadmium treatment caused a decrease in the expression level of human OGG1 (8-oxoguanine-DNA glycosylase-1; hOGG1) in human fibroblast GM00637 and HeLa S3 cells. In addition, the cadmium-mediated decrease in hOGG1 transcription was the result of decreased binding of the transcription factor Sp1 to the hOGG1 promoter. Finally, we show that an increase in the functional hOGG1 expression level could inhibit the cadmium-mediated increase in gamma-ray-induced 8-oxoG accumulation as well as in gamma-radiation-induced mutation frequency at the HPRT (hypoxanthine-guanine phosphoribosyltransferase) gene locus. These results suggest that cadmium attenuates removal of gamma-ray-induced 8-oxoG adducts, which in turn increases the mutation frequency, and that this effect might, at least in part, result from suppression of hOGG1 transcription via inactivation of Sp1 as a result of cadmium treatment.

One possible mechanism of cadmium-induced carcinogenicity includes the mediation of promutagenic DNA damage, as demonstrated by chromosomal aberrations, DNA-protein crosslinks, DNA strand breaks, and 8-oxoguanine (8-oxoG) 1 measurements (6 -11). Although cadmium-induced mutagenicity and carcinogenicity are a multifactorial process, oxidative mechanisms are believed to be of prime importance (12). Therefore, several reports have indicated that the mutagenicity of cadmium is associated with intracellular reactive oxygen species (ROS) (13)(14)(15). However, cadmium does not directly catalyze the formation of ROS via the Fenton reaction (16,17), and cadmium compounds are not bacterial mutagens and are only weakly mutagenic in mammalian cells (18,19). Therefore, cadmium-induced DNA lesions may be indirectly caused by cadmium, e.g. via inhibition of oxidative DNA repair systems.
ROS have a variety of ill effects on DNA, including oxidized bases, abasic (apurinic/apyrimidinic) sites, strand breaks, and DNA-protein cross-links (20). Among the oxidative lesions, 8-oxoG is one of the major base lesions formed after oxidative attack on the DNA (21). Relatively large quantities of 8-oxoG are produced in mammalian cells either as a byproduct of normal oxidative metabolism or as a result of exogenous sources of ROS, such as ionizing radiation, single oxygen sensitizer dyes, and redox-active organic molecules (22). To prevent the mutagenic effect of 8-oxoG, the bacterium Escherichia coli contains a "GO system" (23). The bacterial GO system consists of three proteins: MutM (also known as Fpg protein), a DNA glycosylase/lyase that recognizes 8-oxoG:C and catalyzes the excision of 8-oxoG (24,25); MutY, a DNA glycosylase that recognizes 8-oxoG:A and catalyzes the excision of A (26); and MutT, a specific phosphatase that cleaves 8-oxo-dGTP (27). In mammalian cells, OGG1 (8-oxoguanine-DNA glycosylase-1), which is structurally unrelated but functionally similar to Fpg (formamidopyrimidine glycosylase) in E. coli (28 -30), is the main defense enzyme against the mutagenic effects of 8-oxoG in the cellular DNA. Inactivation of the OGG1 gene in yeast and mice leads to an increase in mutation frequency in the cells (31,32). The human OGG1 (hOGG1) gene is found on chromosome 3p26.2, and allelic deletions in this region frequently occur in a variety of human cancers (33). In addition, the hOGG1 gene is somatically mutated in some cancer cells and is highly polymorphic among human population groups (34,35).
Recently, several groups have demonstrated that cadmium suppresses the capacity of cells to repair oxidative DNA damage (36,37). Moreover, Potts et al. (38) have clearly shown that cadmium inhibits oxidative DNA repair by the down-regulation of OGG1 in rat lung epithelial cells and that this down-regulation of OGG1 is related to cadmium-mediated decreases in OGG1 mRNA. Therefore, there has been considerable interest in determining whether this phenomenon also occurs in human cells and, if so, which transcription factor mediates the downregulated OGG1 expression in cadmium-treated cells. However, up to now, the mechanism by which cadmium causes these effects has not been known. The hOGG1 promoter region was recently cloned (39), and sequence analysis of the promoter region revealed two CpG islands and no TATA or CAAT boxes, suggesting that the hOGG1 gene is essentially a housekeeping gene, which is believed to be expressed in a constitutive way. However, the activity of the housekeeping gene can be modulated under certain conditions, such as oxidative stress (40). In human tissues, the level of OGG1 expression is much higher in the thymus, testis, kidney, intestine, brain, cerebrum, and germinal center of B cells than in other tissues (28,42,43). Moreover, a specific increase in the 8-oxoG incision activity occurs in the mitochondria with age (44), suggesting that OGG1 expression may be modulated by a variety of stimuli. The presence of an Nrf2-binding site, which contributes to regulating the antioxidant responses and coordinating induction of the genes coding the detoxifying enzymes (45), within the hOGG1 promoter region suggests that hOGG1 expression is modulated by oxidative stress. In support of this hypothesis, several groups have demonstrated that hOGG1 is inducible by ROS at the mRNA level in several experimental systems. For example, diesel exhaust particles, an environmental pollutant, are known to generate ROS and to induce lung cancer in experimental animals. Treatment of rats with diesel exhaust particles results in doseand time-dependent changes in the levels of 8-oxoG as well as induction of hOGG1 mRNA (46). In addition, hOGG1 gene expression levels are elevated after rats are exposed to crocidolite asbestos, which is associated with epithelial cell injury in the process of carcinogenesis via oxidative stress (47). Moreover, the DNA-alkylating agent methyl methanesulfate can up-regulate hOGG1 expression through induction of the transcription factor NF-YA (48). Therefore, the activity of certain transcription factors can be important to regulation of hOGG1 expression under some circumstances.
It has been proposed that zinc finger transcription factors, such as Sp1 (specificity protein-1), may be targets for the toxic action of cadmium and other heavy metals (49). Sp1 is a zinc finger transcription factor known to play a role in eukaryotic gene expression (50 -52). Recent studies have demonstrated that cadmium significantly suppresses Sp1 DNA binding activity (53,54). Moreover, Sp1 DNA binding activity is negatively regulated by phosphorylation in cadmium-exposed alveolar epithelial cells (55). Accordingly, we investigated whether cadmium indeed down-regulates hOGG1 mRNA in human cells and whether cadmium-induced modulation of Sp1 transcriptional activity is responsible for suppression of hOGG1 expression in human cells. We used semiquantitative reverse transcription-PCR and Western blotting to measure the alterations in hOGG1 mRNA and protein in human fibroblast GM00637 cells and human cervix cancer HeLa S3 cells exposed to cadmium, respectively. We then examined the effect of cadmium on the putative Sp1-binding sites of the hOGG1 promoter by electrophoretic mobility shift assay (EMSA) and chromatin immunoprecipitation (ChIP). The level of 8-oxoG accumulation was determined by HPLC using an electrochemical detector, and mutation frequency was determined by measuring the mutation frequency at the HPRT (hypoxanthine-guanine phosphoribosyltransferase) gene locus. Our results confirm that cadmium suppresses the mRNA and protein expression of OGG1. Subsequent experiments focused on elucidating the mechanisms responsible for this suppression of OGG1 expression. We found that the decreased Sp1 transcriptional activity as a result of cadmium treatment had a profound negative impact on hOGG1 transcription. Finally, we found that suppression of OGG1 expression by cadmium led to increased accumulation of 8-oxoG and that this increased the frequency of mutations caused by ␥-irradiation.

EXPERIMENTAL PROCEDURES
Chemicals-Cell culture media, fetal bovine serum, penicillin, and streptomycin were purchased from Invitrogen. Chloroform and isopropyl alcohol were purchased from Calbiochem. All other chemicals were supplied by Sigma.
Cell Line Maintenance-Human fibroblast GM00637 cells (Coriell Institute for Medical Research) were maintained in Earle's minimal essential medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Human cervix cancer HeLa S3 cells (American Type Culture Collection) were maintained in Dulbecco's minimal essential medium, 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. All cells were maintained at 37°C in an atmosphere containing 5% CO 2 . Drosophila SL2 cells (American Type Culture Collection) were cultured in Grace's insect cell culture medium at 25°C.
Preparation of Constructs and Transfection-The constructs of hOGG1 and the hOGG1 promoter (p(Ϫ700)hogg1luc) are described elsewhere (48). The Drosophila Sp1 expression vector pPac-Sp1 and the empty vector pPac were kindly provided by Dr. Robert Tjian (University of California, Berkeley, CA). The luciferase reporter construct containing Sp1-responsive elements (Sp1-luciferase) was kindly provided by Dr. Toshiyuki Sakai (Kyoto Prefectural University of Medicine, Kyoto, Japan). To create stable hOGG1-expressing cells, GM00637 and HeLa S3 cells were transfected with the hOGG1 expression vectors using the Lipofectamine method (Promega Corp., Madison, WI) according to the manufacturer's instructions. After transfection, the cells were incubated with complete medium containing 400 g/ml G418 for 5 weeks. The cell clones resistant to G418 were isolated and analyzed. For Drosophila SL2 cells, transfection was carried out by calcium phosphate precipitation (Promega Corp.). Briefly, 1 ϫ 10 6 cells were plated 24 h before transfection, and the complexes were incubated with the cells for 4 -6 h. The pPac plasmid was used as a negative control for Sp1, and either pPac or pCMV was added to equalize the amount of DNA transfected in each well whenever necessary.
Luciferase Activity Assay-The luciferase activity after reporter gene transfection was determined as described previously (48). Briefly, the reporter plasmids were transiently transfected into cells using the Lipofectamine method (Promega Corp.) according to the manufacturer's protocol. Transfection efficiency was determined using the Renilla luciferase gene-containing pRL-CMV plasmid (Promega Corp.). HCT116 cells were transiently transfected with the reporter plasmid and then treated with 0-10 M CdCl 2 . 24 h after treatment, the transfected cells were washed twice with phosphate-buffered saline and lysed in passive lysis buffer (Promega Corp.) with gentle shaking at room temperature for 20 min. The cell lysate was centrifuged at 13,000 rpm for 2 min to pellet the cell debris. The supernatants were transferred to a fresh tube, and the dual-luciferase activity in the cell extracts was determined according Promega Corp. Briefly, each assay mixture contained 2 l of cell lysate and 10 l of firefly luciferase-measuring buffer (LAR II®, Promega Corp.). The firefly luciferase activity was measured using a luminometer (programmed to perform a 2-s pre-measurement delay, followed by a 10-s measurement period for each reporter assay). After measuring the firefly luciferase activity (Stop & Glo®, Promega Corp.), Renilla luciferase-measuring buffer was added, and the Renilla luciferase activity was measured. Each transfection was performed in duplicate, and all were repeated at least three times.
Treatment of Cultures-The cells were seeded into 100-mm diameter tissue culture dishes and allowed to attach for a period of 16 -24 h at 37°C. The cells were treated with different concentrations of CdCl 2 (Sigma) for 24 h to determine the CdCl 2 -induced cytotoxicity and the increase in the accumulation of 8-oxoG in the DNA by CdCl 2 . In ␥-irradiation studies, the cells were preincubated with 10 M CdCl 2 for 20 h and irradiated either with 10 -50-gray (Gy) ␥-rays (for 8-oxoG assay) or with 0.5-5-Gy ␥-rays (for clonogenic cell survival assay) using a ␥-cell irradiator (Clonac 600C, Linac Systems, LLC) at a dose rate of 0.96 Gy/min. The control cells were subjected to a similar treatment, but without CdCl 2 .
Cytotoxicity Assay-The extent of cell death was assayed by trypan blue exclusion. Cell viability was determined as the percentage of the total cell number that remained unstained.
Clonogenic Cell Survival Assay-Cells were seeded at 4 ϫ 10 5 /25-cm culture flask and incubated at 37°C in a 5% CO 2 atmosphere. The cells were preincubated without or with 10 M CdCl 2 for 20 h and irradiated with 0.5-5-Gy ␥-rays. Cells were then treated with cisplatin or UVC light, washed twice with phosphate-buffered saline, trypsinized, and resuspended in fresh medium. They were counted with a Coulter counter, and the number of cells required for plating was obtained by successive dilutions in fresh complete medium. The cells were plated in triplicate in 100-mm Petri dishes over a homogeneous feeder layer formed 24 h previously by plating 5 ϫ 10 4 irradiated cells. Cells were then allowed to grow at 37°C in a 5% CO 2 atmosphere for 14 days. Fresh medium was added on day 7. On day 14, cultures were fixed with methanol and stained with Giemsa. The number of colonies exceeding 50 cells was counted with a binocular lens. The survival fraction was determined as the ratio of the number of colonies observed after treatment to the number of cells seeded, adjusted to the plating efficiency.
Western Blotting-The cells were washed with phosphate-buffered saline and lysed at 0°C for 30 min in 20 mM HEPES (pH 7.4), 2 mM EGTA, 50 mM glycerol phosphate, 1% Triton X-100, 10% glycerol, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 /ml aprotinin, 1 mM Na 3 VO 4 , and 5 mM NaF. The protein content was measured using a Bio-Rad dye binding microassay, and 20 g of protein/lane was electrophoresed on 10% SDS-polyacrylamide gels after boiling for 5 min in Laemmli sample buffer. The proteins were blotted onto Hybond ECL membranes (Amersham Biosciences). Markers (Fermentas) were used as the size standards. After electroblotting, the membranes were blocked with 10 mM Tris-HCl (pH 7.4), 150 mM NaCl, and 0.1% Tween 20 containing 5% milk and incubated with anti-hOGG1, anti-Sp1, and anti-␣-tubulin antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) diluted in blocking buffer for 4 h. The primary antibody dilutions were those recommended by the manufacturer. The membranes were washed, incubated with the appropriate secondary antibodies (1:4000; Santa Cruz Biotechnology, Inc.) in blocking buffer for 2 h, and washed repeatedly. The blotted proteins were detected using an enhanced chemiluminescence detection system (iN-tRON Biotech, Seoul, Korea).
Semiquantitative Reverse Transcription-PCR-RNA extraction was carried out using RNA STAT-60 TM (Tel-Test, Inc., Friendswood, TX) according to the manufacturer's instructions. Briefly, after homogenizing the cells in RNA STAT-60, the homogenate was mixed with chloroform (5:1, v/v), shaken vigorously for 15 s, and then centrifuged at 13,000 rpm for 15 min at 4°C. The RNA in the upper colorless aqueous phase was precipitated by adding isopropyl alcohol, washed twice with 70% ethanol, and air-dried for 10 min. The RNA was then resuspended in diethyl pyrocarbonate-treated water. 10-l RNA aliquots were prepared and stored at Ϫ70°C until needed. 2 g of the total RNA was reverse-transcribed using a Moloney murine leukemia virus cDNA synthesis system (Promega Corp.), and the reverse-transcribed DNA was subjected to PCR. The profile of the replication cycles was as follows: denaturation at 94°C for 50 s, annealing at 58°C for 50 s, and polymerization at 72°C for 1 min. In each reaction, the same amount of glyceraldehyde-3-phosphate dehydrogenase was used as an internal control. The primers used for PCR were as follows: hOGG1, 5Ј-CTG CCT TCT GGA CAA TCT TT-3Ј (forward) and 5Ј-TAG CCC GCC CTG TTC TTC-3Ј (reverse; designed to amplify a 551-bp region); and glyceraldehyde-3-phosphate dehydrogenase, 5Ј-CCA TGG AGA AGG CTG GGG-3Ј (forward) and 5Ј-CAA AGT TGT CAT GGA TGA CC-3Ј (reverse; designed to amplify a 194-bp region) (total number of cycles ϭ 26). The PCR products were resolved on 1% agarose gels, stained with ethidium bromide, and then photographed.
EMSA-The nuclear protein was extracted from the GM00637 and HeLa S3 cells as described previously (48). Briefly, the cells were washed with phosphate-buffered saline and pelleted at 4°C for 5 min at 2500 rpm. The upper supernatant was carefully removed, and the pellet was resuspended in 1 ml of ice-cold cell lysis buffer containing 10 mM HEPES (pH 7.9), 1.5 mM MgCl 2 , 0.5 mM phenylmethylsulfonyl fluoride, 1 mM aprotinin, and 1 mM leupeptin. The suspension was incubated on ice for 15 min with occasional mixing by inversion. Subsequently, 24 l of 10% Nonidet P-40 was added; the cells were mixed by inversion; and the nuclei were immediately pelleted at 4°C for 5 min at 3000 rpm. The supernatant was discarded, and the pellet containing the nuclei was resuspended by adding dropwise 50 l of ice-cold nuclear extraction buffer containing 30 mM HEPES, 0.3 mM EDTA, 1.5 mM MgCl 2 , 0.2 mM dithiothreitol, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 mM aprotinin, 1 mM leupeptin, and 450 mM KCl. The mixture was then incubated on ice for 30 min, and the nuclear debris were pelleted at 15,000 rpm for 20 min. The protein concentration of the nuclear extracts was determined using a BCA protein assay kit (Pierce) according to the manufacturer's instructions. The nuclear extracts were stored at Ϫ70°C until needed. EMSA binding reactions were incubated for 20 min at room temperature in a 20-l final volume containing 5 g of nuclear extract, 20 fmol of end-labeled double-stranded oligonucleotide, and 1 g of poly(dI-dC)⅐poly(dI-dC) (Amersham Biosciences). The binding buffer contained 30 mM HEPES, 0.3 mM EDTA, 1.5 mM MgCl 2 , 0.2 mM dithiothreitol, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1 mM aprotinin, and 1 mM leupeptin. Supershift assays were performed by adding 1 l of anti-Sp1 polyclonal antibody (Santa Cruz Biotechnology, Inc.). The complexes were resolved on a 5% nondenaturing polyacrylamide gel in 0.5ϫ Tris borate/EDTA by electrophoresis at 200 V for 1 h. The gels were dried and exposed overnight at Ϫ70°C to Eastman Kodak x-ray film.
ChIP-ChIP was performed as described (56). The region amplified was the promoter region of human hOGG1 (Ϫ535 to Ϫ321 from the transcription start site). The DNA sequence of the 5Ј-primer was 5Ј-GCCGGCGGGACGACAATCCG-3Ј, and that of the 3Ј-primer was 5Ј-TAGGCGTTCGCCCTCCTTGG-3Ј. The PCRs were repeated using varying cycle numbers and different amounts of templates to ensure that results were within the linear range of PCR.
8-OxoG Glycosylase Activity Assay-Cells at the exponential phase were centrifuged at 800 ϫ g for 5 min. Cell pellets (10 6 cells/assay) were then suspended in 2 volumes of homogenization buffer (50 mM Tris-HCl, 50 mM KCl, 1 mM EDTA, 5% glycerol, and 0.05% 2-mercaptoethanol (pH 7.5) and homogenized. The homogenates were mixed with streptomycin (final concentration of 1.5%) to remove the nucleic acids. The supernatants obtained by centrifugation were dialyzed extensively against homogenization buffer and used as the cell extracts for the endonuclease nicking assay. An 8-oxoG-containing 21-mer with the sequence 5Ј-CAGC-CAATCAGTXCACCATTC-3Ј (where X is 8-oxoG) along with its complementary strand (Midland Certified Reagent Co., Midland, TX) were used. The oligonucleotides were 3Ј-end-labeled using terminal transferase and [␣-32 P]dideoxy-ATP (3000 Ci/mmol; Amersham Biosciences). The endlabeled oligomer was annealed with its complementary oligonucleotide, and the resulting duplex DNA was used as the assay substrate. The duplex substrate DNA (20 pmol) was incubated with the cell extracts (10 g of protein) at 37°C for 1 h in 1 ml of reaction mixture containing 50 mM Tris-HCl, 50 mM KCl, and 1 mM EDTA (pH 7.5). The reaction was terminated by heating at 90°C for 3 min. The reaction products were electrophoresed on 7 M urea-containing 20% denaturing polyacrylamide gels (DNA sequencing gel). The gels were wrapped in Saran Wrap and exposed to Kodak film for visualization.
Analysis of 8-OxoG in the Cellular DNA-The 8-oxoG levels in the DNA of the cells were measured using a slight modification of a method described previously (29). Briefly, the cellular DNA was isolated using a DNA extractor WB kit (Wako, Osaka, Japan). 50 g of the isolated DNA was digested with 2 units of P1 nuclease (Roche Applied Science) in a 100-l solution containing 1 mM EDTA and 10 mM sodium acetate (pH 4.5). The nucleotide solution was filtered through an Ultrafree Probind filter (Millipore Corp., Bedford, MA), and a 20-l aliquot of the sample was injected into an HPLC column (YMC-Pack TM ODS-AM, 5 m, 4.6 ϫ 300 mm) equipped with an electrochemical detector (Coulochem II, ESA Biosciences, Inc., Chelmsford, MA) at a flow rate of 1 ml/min. The mobile phase consisted of 12.5 mM citric acid, 25 mM sodium acetate, 30 mM NaOH, and 10 mM acetic acid containing 3% methanol. Deoxyguanine (dG, Sigma) and 8-oxoG (Cayman Chemical Co., Inc., Ann Arbor, MI) were used as standards. The 8-oxoG level in the DNA is expressed as 8-oxoG/10 7 dG.
HPRT Mutation Assay-The method used to determine the frequency of the ␥-ray-induced HPRT mutants has been described (57). Briefly, cells were plated at 5 ϫ 10 4 /100-mm diameter dish and cultured in Earle's minimal essential medium, 10% fetal bovine serum, 100 M hypoxanthine, 0.4 M aminopterin, and 16 M thymidine (Invitrogen) for at least 14 days. The cells were incubated with 10 M CdCl 2 for 6 h and then exposed to 0.5-and 1-Gy ␥-rays. After ␥-irradiation, the cells were allowed an additional 12-14 days to express the HPRT mutants before being plated in selective medium containing 30 M 6-thioguanine (Sigma). A separate set of cells plated at a cloning density in medium lacking 6-thioguanine was used to determine the cloning efficiency of the cells at the time of selection. The mutant colonies were stained with 0.5% crystal violet in 50% methanol and scored after 12-14 days. For each dose and for the untreated control, the frequency of the mutants was calculated from the number of 6-thioguanine-resistant colonies formed divided by the number of cells selected. This frequency was corrected for the cloning efficiency of the cells at the time of selection. The induced frequencies were calculated by subtracting the background frequencies in the control population.
Data Analysis-Data in all experiments are represented as the mean Ϯ S.D. Statistical comparisons were carried out using the unpaired t test. p values Ͻ0.05 were considered to be statistically significant.

Cadmium Increases the Level of ␥-Ray-induced 8-OxoG-
Cadmium-induced cytotoxicity was initially evaluated to examine the effect of cadmium on ␥-ray-induced 8-oxoG accumulation. GM00637 and HeLa S3 cells were cultured in different cadmium concentrations. After treatment, the adherent and non-adherent cells were pooled, and cell viability was determined using a trypan blue assay. As shown in Fig. 1A, exposure of the cells to up to 20 M CdCl 2 for 24 h did not cause any significant level of cell death compared with the untreated cells. Treatment of the cells with 50 M CdCl 2 caused a significant reduction in viability (38ϳ47% of the untreated cells). The radiation dose survival curves were obtained by colony forming assay. GM00637 and HeLa S3 cells were preincubated without or with 5 or 10 M CdCl 2 for 20 h and then irradiated with 0.5-5-Gy ␥-rays. The fraction of cells that survived the exposure to ␥-rays showed that the colony-forming ability was only slightly decreased at doses of up to 1-Gy ␥-rays; thereafter, it abruptly dropped (Fig. 1B). Based on the cytotoxic effects of cadmium and/or ␥-ray exposure, 1-20 M CdCl 2 concentrations were chosen for the subsequent exper- iments, and 0.5-1-Gy ␥-rays were used for HPRT mutant assay.
The effect of cadmium on the formation of 8-oxoG adducts, which are one of the major base lesions formed after oxidative attack on DNA, was investigated next. GM00637 and HeLa S3 cells were treated without or with 1, 5, or 10 M CdCl 2 . 24 h after treatment, the amount of 8-oxoG in the genomic DNA was then measured by HPLC using an electrochemical detector. As shown in Fig. 1C, the cells treated with CdCl 2 did not show any significant increase in 8-oxoG formation. However, the cells pretreated with CdCl 2 had significantly higher levels of ␥-rayinduced 8-oxoG (Fig. 1D). These results suggest that, whereas cadmium is too weak to generate appreciable amounts of 8-oxoG in GM00637 and HeLa S3 cells, cadmium might sensitize the cells to an increase in ␥-ray-induced 8-oxoG accumulation.
Effect of Cadmium on hOGG1 Expression-Experiments were conducted to confirm that cadmium suppresses hOGG1 expression in human fibroblasts and cervix cancer cells, as has been documented previously for rodent lung epithelial cells (38). GM00637 and HeLa S3 cells were treated without or with 10 M CdCl 2 , and hOGG1 protein levels were determined in the attached cells after a 24-h CdCl 2 treatment. SDS-PAGE was used to separate the whole cell extracts of the protein from the untreated cells and from the CdCl 2 -treated cells. As shown in Fig. 2A, Western blotting with a specific antibody against hOGG1 showed that treating the cells with 10 M CdCl 2 caused a decrease in hOGG1 protein levels.
To determine the functional significance of the CdCl 2 -mediated decrease in hOGG1 expression, nuclear extracts from the CdCl 2 -treated cells were prepared and examined for their ability to cleave 8-oxoG using a 21-mer oligonucleotide containing a single 8-oxoG. As shown in Fig. 2B, treating the cells with 10 M CdCl 2 for 24 h caused a significant decrease in the ability of the nuclear extracts to cleave the 8-oxoG:C substrate. This suggests that CdCl 2 treatment suppresses hOGG1 expression and leads to a decrease in the hOGG1 repair activity.
We further examined whether or not the cadmium-mediated decrease in hOGG1 expression occurs at the transcriptional level. GM00637 and HeLa S3 cells were treated without or with 10 M CdCl 2 , and the hOGG1 mRNA levels were determined in the attached cells 12 h after CdCl 2 treatment. The PCR exponential phase was obtained, and the optimal number of PCR cycles, 26, was determined (data not shown). Semiquantitative reverse transcription-PCR analysis showed that the hOGG1 mRNA in the cells treated with 10 M CdCl 2 was reduced by Ͼ80% compared with the untreated cells (Fig. 3A). To examine the effect of CdCl 2 on the transcriptional regulation of hOGG1, we constructed plasmids (p(Ϫ700)hogg1luc) in which the expression of the luciferase coding sequence was driven by the hOGG1 promoter (bp Ϫ700 to ϩ93). GM00637 and HeLa S3 cells were transfected with p(Ϫ700)hogg1luc and then treated with various CdCl 2 doses. The CdCl 2 -treated cells showed a dose-dependent decrease in the hOGG1 promoter activity (Fig.  3B). Treatment with 5 and 10 M CdCl 2 resulted in a 40ϳ55 and 75ϳ85% reduction in the hOGG1-luciferase activity, respectively, compared with the untreated cells.
Involvement of the Sp1 Transcription Factor in the Regulation of hOGG1 by Cadmium-The potential mechanisms responsible for the observed transcriptional inhibition of hOGG1 by CdCl 2 were examined next. Because the hOGG1 promoter contains several putative Sp1-binding sites (Fig. 4A) and CdCl 2 has been shown to interact with Sp1 and to affect its DNA binding activity (55), this study focused on the transcription factor Sp1. To determine whether CdCl 2 modulates the transcriptional activity of Sp1, cells were transiently transfected with an Sp1-luciferase reporter construct and then treated without or with 1, 5, or 10 M CdCl 2 for 24 h. As shown in Fig.  4B, cadmium treatment caused a dose-dependent inhibition of the Sp1-mediated luciferase activity in both GM00637 and HeLa S3 cells.
To determine whether the Sp1 protein binds to the hOGG1 promoter and to determine whether CdCl 2 contributes to the Sp1 binding activity of the hOGG1 promoter, the Sp1 binding activity in the GM00637 and HeLa S3 cells after CdCl 2 treatment was analyzed by EMSA employing the oligonucleotides containing the hOGG1-specific Sp1-binding sites. The cells treated with CdCl 2 showed a dose-dependent decrease in Sp1 DNA binding activity (Fig. 4C). To confirm the composition of the Sp1 complexes formed at putative Sp1-binding sites, specific antibodies against the Sp1 transcription factor were added to the DNA-protein binding assay. As shown in Fig. 4C, supershifted DNA-protein complexes were observed after adding the anti-Sp1 antibodies to the DNA binding reaction performed with the oligonucleotides containing the hOGG1-specific Sp1binding sites.
Experiments were subsequently performed to evaluate the in vivo effects of cadmium on the interaction of Sp1 with the hOGG1 promoter. To accomplish this, a ChIP assay was performed using cells that had first been transfected with the hOGG1 promoter fragment and then subsequently treated for 12 h without or with 10 M CdCl 2 . The results of this experiment demonstrated that a complex containing the Sp1 protein and the hOGG1 promoter sequence could be precipitated by anti-Sp1 antibodies (Fig. 4D). Moreover, the amount of this ChIP complex was dramatically reduced by CdCl 2 treatment. A negative control using a nonspecific antibody (anti-mouse IgG) was included to establish that the binding was specific for Sp1 (Fig. 4D). The Western blots of total Sp1 in CdCl 2 -treated and untreated cells were similar (Fig. 4D) and thus verified that the difference in the amount of complex recovered from these cells was due to differential DNA binding as opposed to a difference in antibody binding. The overall results of the ChIP assay support the conclusion that deceased interaction of Sp1 with Sp1-binding sites in the hOGG1 promoter contributes to CdCl 2mediated suppression of hOGG1 transcriptional activity in GM00637 and HeLa S3 cells.
Basal Transcriptional Regulation of hOGG1 by Sp1-To confirm a possible role of Sp1 in CdCl 2 -mediated transcriptional repression of hOGG1, the Drosophila cell line SL2, which has an Sp1-deficient background, was used. We confirmed that the Sp1 expression vector used could produce the Sp1 protein detectable by Western blot analysis (Fig. 5A). We found that the basal hOGG1 promoter activity in the SL2 cells was similar to the background expression of the control vector. However, Sp1 expression significantly increased the hOGG1 promoter activ-ity, whereas CdCl 2 treatment decreased the hOGG1-luciferase activity by ϳ80% (Fig. 5B).
Silencing of Sp1 Expression in GM00637 and HeLa S3 Cells via siRNA Leads to a Decrease in the hOGG1 Promoter Activity-To further confirm that Sp1 is indeed important for the CdCl 2 -mediated decrease in hOGG1 expression in GM00637 and HeLa S3 cells, siRNAs in the form of 21-bp RNA duplexes, which target Sp1, were used in an attempt to inhibit its expression. GM00637 and HeLa S3 cells were transfected with the mock, control siRNA, or Sp1-specific siRNA oligonucleotides. The cells were harvested 48 h after transfection, and their protein expression levels were determined. Western blot analysis revealed that the Sp1-specific siRNA oligonucleotide levels decreased by Ͼ80% in terms of their overall Sp1 protein expression level compared with the mock-transfected or control siRNA-transfected cells (Fig. 6A). The hOGG1 promoter activity following Sp1 siRNA transfection was next examined, and we found that cells with reduced Sp1 levels had a significantly lower level of hOGG1 promoter activity compared with the mock-or control siRNA-transfected cells (Fig. 6B). This demonstrates that Sp1 contributes to the basal transcriptional regulation of hOGG1 in GM00637 and HeLa S3 cells.
Expression of hOGG1 Suppresses ␥-Ray-induced 8-OxoG Accumulation and ␥-Ray-induced Mutation Frequency of the HPRT Gene in Cadmium-treated Cells-We examined whether or not hOGG1 expression could inhibit ␥-ray-induced 8-oxoG accumulation in CdCl 2 -treated cells. hOGG1 was subcloned into the pcDNA3 vector to form pcDNA3-hOGG1. This construct was transfected into GM00637 and HeLa S3 cells. Several stably transfected cell lines were established after selection using G418 for 5 weeks. Western blot analysis revealed that the hOGG1 expression levels in CdCl 2 -treated hOGG1expressing cells were significantly higher than in CdCl 2treated empty vector-transfected cells (Fig. 7A).
To examine the effect of hOGG1 expression on ␥-ray-induced 8-oxoG accumulation in CdCl 2 -treated cells, 20 h after pretreatment with 10 M CdCl 2 , the empty vector-and hOGG1transfected cells were exposed to 50 Gy of ␥-irradiation; and 3 h FIG. 3. Effect of CdCl 2 on hOGG1 mRNA and the hOGG1 promoter activity. A, upper panels, subconfluent GM00637 (left) and HeLa S3 (right) cells were treated with 10 M CdCl 2 , and the total RNA was isolated 12 h later. The hOGG1 mRNA level was evaluated by semiquantitative reverse transcription-PCR using the hOGG1-specific primers over 26 cycles. Lower panel, the levels of hOGG1 mRNA relative to those of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were determined. The results correspond to three independent experiments (ϮS.D.). *, p Ͻ 0.05. B, GM00637 and HeLa S3 cells were transfected with a reporter plasmid containing the hOGG1 promoter sequences that drive the expression of the luciferase (Luc) gene and then treated without or with 5 or 10 M CdCl 2 . The cells were harvested 24 h after treatment, and the luciferase assay was performed as described under "Experimental Procedures." The luciferase activities in the untreated cells were arbitrarily set to 100%, and the relative luciferase activities in the CdCl 2 -treated cells were calculated accordingly. Values are presented as the mean Ϯ S.D. from four separate experiments. *, p Ͻ 0.05. after irradiation, the 8-oxoG level in the genomic DNA was measured. We found that the amount of 8-oxoG in the genomic DNA from the CdCl 2 -treated empty vector-transfected cells exposed to 50 Gy of ␥-radiation was 73.1 Ϯ 9.6 to 82.4 Ϯ 10.2/10 7 dG (Fig. 7B). However, hOGG1 expression led to a significant decrease in 8-oxoG accumulation. Therefore, the ␥-ray-induced mutation frequencies in the CdCl 2 -treated empty vector-and hOGG1-transfected cells were evaluated to determine whether hOGG1 expression could inhibit the ␥-rayinduced mutation in CdCl 2 -treated cells. We investigated the ␥-irradiation-induced mutagenesis of the HPRT gene; a mutation of this gene leads to 6-thioguanine resistance of the mutant cells. The cells were incubated with 10 M CdCl 2 for 20 h and exposed to 0.5-and 1-Gy ␥-rays, and the ␥-ray-induced mutant frequencies of the HPRT gene were then measured. As shown in Table I, the empty vector-transfected cells showed a significantly higher mutation frequency compared with the hOGG1-transfected cells. These results suggest that, when the cells were irradiated with ␥-rays, the CdCl 2 -mediated suppression of hOGG1 expression contributed, at least in part, to the CdCl 2 -induced increase in 8-oxoG accumulation and mutation frequency. DISCUSSION In this study, we have demonstrated that exposure of human fibroblasts and cervix cancer cells to cadmium led to down- FIG. 4. CdCl 2 suppresses Sp1 activity. A, shown is the localization of the Sp1 elements in the hOGG1 promoter (p(Ϫ700)hogg1luc). B, GM00637 and HeLa S3 cells were transfected with the Sp1-luciferase (Luc) reporter plasmid and then treated without or with 1, 5, or 10 M CdCl 2 . The cells were harvested 24 h after treatment, and the luciferase assay was performed as described under "Experimental Procedures." The luciferase activities in the untreated cells were arbitrarily set to 100%, and the relative luciferase activities in the CdCl 2 -treated cells were calculated accordingly. Values are presented as the mean Ϯ S.D. from four separate experiments. *, p Ͻ 0.05. C, the 35-bp oligonucleotide probes containing the Sp1 motifs of the hOGG1 promoter (Ϫ474 to Ϫ431) were incubated with nuclear extracts isolated from GM00637 and HeLa S3 cells. The nuclear extracts were prepared from cells treated without or with 5, 10, or 20 M CdCl 2 . For supershift assays, anti-Sp1 antibodies were added to the reaction mixtures and incubated for 30 min prior to separating the DNA-protein complexes. The DNA-protein complexes were run on neutral polyacrylamide gels and visualized by autoradiography. The results shown are representative of at least three independent experiments. D, GM00637 and HeLa S3 cells transfected with the hOGG1 promoter (p(Ϫ700)hogg1luc) were treated with (ϩ) or without (Ϫ) 10 M CdCl 2 for 12 h, cross-linked with formaldehyde, and then sonicated. The fragmented chromatins were immunoprecipitated by adding either the anti-Sp1 antibody or anti-mouse IgG (nonspecific). The samples were then amplified by PCR using the primers flanking the Sp1-binding sites, and the PCR products were resolved on a 1.5% agarose gel. To compare the relative amounts of chromatin used in immunoprecipitation, the input chromatin from the untreated and treated cells was also amplified by PCR (Input). Western blot experiments were performed with the anti-Sp1 antibody and the anti-␣-tubulin antibody as a control for equal loading. The results are representative of several experiments. regulation of hOGG1 expression. Using EMSA and ChIP analyses, we have shown that the cadmium-mediated decrease in hOGG1 expression resulted from decreased binding of the transcription factor Sp1 to the hOGG1 promoter. Subsequent experiments revealed that suppression of hOGG1 by cadmium caused an increased accumulation of 8-oxoG and that this increased the frequency of mutations caused by ␥-irradiation.
Cadmium has been suggested to play a role in the induction of various genotoxic effects, such as DNA strands breaks and oxidative DNA base modifications. However, cadmium cannot form stable DNA adducts (58) and is not a directly redox-active metal (59), which is a potential mechanism of the oxidative DNA damage-induced genotoxicity of some heavy metals. Although cadmium-induced DNA damage, including DNA singleand double-strand breaks, chromosomal aberrations, sister chromatid exchange, and DNA-protein cross-links, have been observed in different mammalian cells, these effects are restricted to cytotoxic concentrations (6 -11). At nontoxic concentrations, cadmium exerts pronounced co-genotoxic effects when combined with various types of mutagens such as ultraviolet light, ionizing radiation, and other DNA-damaging agents (60). Besides the direct genotoxic effect of cadmium, the indirect genotoxic effects of cadmium are believed to be involved in cadmium-induced mutagenesis and carcinogenesis. One mechanism frequently proposed is that the increase in oxidative DNA lesions is attributable to cadmium exposure, which is mediated by interference with oxidative DNA repair processes (36,37).
OGG1 serves to remove oxidized bases such as 8-oxoG, which is the most stable and critical mutagenic lesion after oxidation of DNA. Recently, several lines of evidence have shown that cadmium suppresses the activity of OGG1 in mammalian cells. For example, low concentrations of cadmium, which do not generate oxidative base modifications, have been demonstrated to inhibit the repair of oxidized damage in mammalian cells (18). In addition, under GSH-depleted conditions, cadmium inhibits the 8-oxoG repair activity in rat testis, and 8-oxoG accumulates in the DNA (61). Similarly, Potts et al. (62) have shown that hydrogen peroxide-induced 8-oxoG accumulation is much higher in cadmium-adapted epithelial cells compared with non-adapted cells. Moreover, Potts et al. (38) reported that cadmium is effective as an inhibitor of OGG1 expression both in vivo in rats exposed to aerosols of cadmium acetate and in vitro in a rat lung epithelial cell line. This down-regulation was observed at both the mRNA and protein levels, and it was produced by 10 M cadmium, which caused no loss in the viability of the cell culture. Furthermore, at low nontoxic concentrations, cadmium attenuated removal of the hydrogen peroxide-induced 8-oxoG adducts when it was added to the human-hamster hybrid cells after exposure to hydrogen peroxide (63). In our system, the accumulation of 8-oxoG was not increased after treatment with cadmium at concentrations up to 5 M. With 10 M cadmium, there was a slight but insignificant increase (Fig. 1C), suggesting that cadmium might be too weak to generate appreciable amounts of 8-oxoG in intact human fibroblasts and cervix cancer cells. However, the 8-oxoG levels in the 10 M cadmium-treated cells after exposure to ␥-irradiation, which is expressed as the ratio of 8-oxoG to dG, were significantly higher than in untreated cells (Fig. 1D). In addition, we also found that exposure of human fibroblasts and HeLa S3 cells to cadmium led to a significant decrease in hOGG1 protein and mRNA levels (Figs. 2 and 3). These findings indicate that cadmium suppresses hOGG1 expression in human cells.
To investigate the potential mechanisms responsible for the observed transcriptional inhibition of hOGG1 by cadmium, we focused on the transcription factor Sp1 for three reasons. First, cadmium inhibits expression of hOGG1 mRNA (38); second, cadmium has been shown to interact with Sp1 and to affect its DNA binding activity (53)(54)(55); and third, the hOGG1 promoter contains several putative Sp1-binding sites (Fig. 4A). Sp1 is a member of a family of transcription factors with zinc fingertype DNA-binding domains and binds to GC-or GT-rich DNA sequences (64,65). Sp1 is important for the expression of many cellular genes, particularly housekeeping genes (50). Although Sp1 has generally been considered to constitutively regulate gene expression, its activity and cellular content have been shown to be regulated during a variety of cellular processes (51,52). Recent studies using purified preparations of Sp1 or nuclear extracts containing Sp1 have demonstrated that cadmium significantly suppresses Sp1 DNA binding activity (53)(54)(55). Sp1 is often essential for initiation of TATA-less promoter transcription (66 -68). In mammalian cells, promoters lacking TATA boxes generally have several Sp1 elements (69). Recently, the hOGG1 promoter region was cloned (39,48,70), but little is known about which transcription factors actually participate in OGG1 regulation. Sequence analysis of the promoter region revealed two CpG islands, no TATA or CAAT boxes, and several putative NF-Y-and Sp1-binding sites (39,48). The results of this study suggest that the Sp1 transcription factor contributes to cadmium-mediated hOGG1 inhibition. Using in vitro EMSA experiments and an in vivo ChIP assay, we demonstrated that Sp1 interacted with the putative Sp1-binding sites of the hOGG1 promoter and that this interaction was markedly suppressed by cadmium treatment (Fig. 4). These results suggest that the Sp1 transcription factor is a specific target for cadmium, resulting in the impairment of its DNA binding properties.
To confirm whether Sp1 is indeed important for the regulation of OGG1 expression, we used the Drosophila cell line SL2, which has an Sp1-deficient background. These results clearly show that the cadmium-induced hOGG1 inhibition was Sp1dependent (Fig. 5). In addition, the Sp1-targeted siRNA oligonucleotides caused a reduction in the hOGG1 promoter activity (Fig. 6), suggesting that Sp1 plays an important role in the basal transcriptional activity of hOGG1. An important feature of the transcriptional regulation by the Sp1 protein is its requirement for cofactors, including Ets, p53, and histone deacetylases, along with their synergistic interaction with other members of the Sp family (67,68,71). Ongoing studies will determine how Sp1 interacts in the regulation of hOGG1.
The classification of cadmium as a human carcinogen is supported by strong evidence from animal experiments. Various cadmium compounds produce adenocarcinomas of the lung in rats after inhalation (72,73). Prostate and pancreatic tumors were induced by subcutaneous injection of cadmium chloride in rats; testicular tumors were induced in rats by oral exposure; and cadmium produced local tumors at various sites of injection, typically sarcomas, in rats and mice (74). Even though the carcinogenic mechanisms of cadmium are still far from being elucidated completely, one possible explanation is the induction of promutagenic DNA damage. Cadmium is predominantly a non-genotoxic carcinogen. It is essentially nonmutagenic in bacterial tests and is only weakly mutagenic in mammalian cells in vitro (18,19). However, cadmium compounds enhance the mutagenicity and carcinogenicity of directly acting genotoxic agents, and this property has been explained by the inhibition of DNA repair processes by cadmium (75). 8-OxoG preferentially mispairs with adenosine during replication, which gives rise to G:C-to-T:A transversion mutations (76,77). 8-OxoG is believed to represent a major source of ROS-induced mutagenesis in all aerobic cells because of their persistent generation, relative abundance, and potent mutagenicity. An Ogg1 deficiency in yeast, as well as an Fpg deficiency in bacteria, results in a mutator phenotype (31,78). Similarly, Ogg1-deficient mice show an abnormal accumulation of 8-oxoG in their genome and exhibit a significantly higher mutation rate (79). In addition, inactivation of the OGG1 gene in human cells leads to an increase in the mutation frequency (33-35, 80, 81). Therefore, the cadmium-mediated decrease in hOGG1 expression might contribute to cadmiuminduced mutagenesis. To test this hypothesis, we investigated the effect of hOGG1 expression on the accumulation of 8-oxoG and the induced mutagenesis. The coding region of hOGG1 was cloned and permanently transfected into GM00637 and HeLa S3 cells. As shown in Fig. 7, the 8-oxoG levels in the cadmiumtreated hOGG1-expressing clones after exposure to ␥-irradiation were significantly lower than in the parental and empty vector-expressing cells. Moreover, we found that, when combined with ␥-rays, cadmium markedly increased the mutation rate in GM00637 and HeLa S3 cells. Furthermore, this increased mutagenesis was significantly suppressed by hOGG1  7. hOGG1 expression leads to a decrease in 8-oxoG accumulation after ␥-radiation in CdCl 2 -treated GM00637 and HeLa S3 cells. A, hOGG1-expressing cells (Cd-hOGG1) and pcDNA3-transfected cells (Cd-pcDNA3) were treated with 10 M CdCl 2 . The untreated parental cells (None) were used in the control experiment. The cells were harvested 24 h after treatment, and the hOGG1 expression level was analyzed by Western blotting using the anti-hOGG1 antibody. For equal loading in the control experiment, the membranes were reprobed with the anti-␣-tubulin antibody. B, empty vector-or hOGG1-transfected cells were treated without or with 10 M CdCl 2 for 20 h. The untreated cells (None), the 10 M CdCl 2 -pretreated pcDNA3-expressing cells (Cd-pcDNA3), and the 10 M CdCl 2 -pretreated hOGG1-expressing cells (Cd-hOGG1) were exposed to 50 Gy of ␥-radiation. The amount of 8-oxoG in the DNA were measured 3 h after treatment as described under "Experimental Procedures." Values are presented as the mean Ϯ S.D. from four separate experiments. *, p Ͻ 0.05. expression (Table I). These results strongly suggest that the cadmium-mediated down-regulation of hOGG1 expression contributes to an increased mutation frequency induced by ␥-irradiation.
In summary, we have presented evidence that cadmium causes decreased levels of hOGG1 expression through decreased binding of the transcription factor Sp1 to the hOGG1 promoter. The suppression of hOGG1 expression via cadmium leads to an increase in 8-oxoG accumulation and mutation frequency induced by ␥-irradiation. Moreover, we have shown that an increase in the functional hOGG1 expression level can inhibit the cadmium-mediated increase in ␥-ray-induced 8-oxoG accumulation as well as the increase in the ␥-irradiation-induced mutation rate. Because inactivation of the OGG1 gene in mammalian cells leads to a mutator phenotype, it is expected that cells suppressing OGG1 expression due to cadmium exposure could have an enhanced probability of undergoing malignancy (41). Therefore, the down-regulation of hOGG1 via cadmium might, at least in part, contribute to cadmium-mediated mutagenesis and carcinogenesis. c Cd-pcDNA3, 10 M CdCl 2 -pretreated pcDNA-expressing cells; Cd-hOGG1, 10 M CdCl 2 pretreated hOGG1-expressing cells.
d The average mutant frequency was significantly lower in hOGG1transfected cells than in pcDNA3-transfected cells (p Ͻ 0.05)