Induction of OGG1 Gene Expression by HIV-1 Tat*

To identify the cellular gene target for Tat, we performed gene expression profile analysis and found that Tat up-regulates the expression of the OGG1 (8-oxogua-nine-DNA glycosylase-1) gene, which encodes an enzyme responsible for repairing the oxidatively damaged guanosine, 8-oxo-7,8-dihydro-2 (cid:1) -deoxyguanosine (8-oxo-dG). We observed that Tat induced OGG1 gene expression by enhancing its promoter activity without chang-ing its mRNA stability. We found that the upstream AP-4 site within the OGG1 promoter is responsible and that Tat interacted with AP-4 and removed AP-4 from the OGG1 promoter by in vivo chromatin immunoprecipitation assay. Thus, Tat appears to activate OGG1 expression by sequestrating AP-4. Interestingly, although Tat induces oxidative stress known to generate 8-oxo-dG, which causes the G:C to T:A transversion, we observed that the amount of 8-oxo-dG was reduced by Tat. 1.2 and as a final The hybridized at 10- (cid:3) m resolution on an Affymetrix 428 array scanner. Analysis of differential expression of each performed using Version 4.2 computer software (Bio-Discovery Ltd.). Normalization of hybridized was performed by global scaling.

Tat is an essential transactivator of human immunodeficiency virus (HIV) 1 gene expression and viral replication (1). Tat stimulates viral gene expression by directly binding to the characteristic RNA stem-loop-bulge structure called the transactivation response region located within the long terminal repeat (2,3) and enhancing the processivity of RNA polymerase II (4,5). The transcriptional activity of Tat is supported by interaction with cellular factors such as positive transcription elongation factor-b (6 -8) and histone acetyltransferase (9). Cyclin T1, a regulatory subunit of the positive transcription elongation factor-b complex, binds to the activation domain of Tat and facilitates the hyperphosphorylation of the C-terminal domain of RNA polymerase II at the vicinity of HIV genes. Thus, Tat makes RNA polymerase II highly competent for the transcription elongation and productive expression of HIV genes (10).
Although much of the efforts in Tat studies have focused on its transcriptional activation from the HIV provirus, the actions of Tat on cellular genes have also been revealed. For example, Tat is known to promote cellular transformation (11), to induce oxidative stress (12,13), and to elicit inflammatory reactions (14,15). Choi et al. (16) observed that Tat transgenic mice exhibit decreased gene expression of the ␥-glutamylcysteine synthetase regulatory subunit and decreased GSH content in tissues. These biological actions of Tat are considered to cause activation of nuclear factor-B, AP-1 (activating protein-1), and mitogen-activated protein kinase (13,17). These findings prompted us to search for cellular target genes of Tat, either up-regulated or down-regulated, using a gene expression profile analysis.
In addition to the very high efficiency of the viral replication rate that is mainly ascribable to Tat action, HIV owes its morbidity to its high mutation rate, leading to the emergence of drug resistance and escape from the host immune response. In fact, the high frequency of G:C to A:T and G:C to T:A mutations was previously observed in HIV-1 and other lentiviruses (18 -21). Recent studies (22)(23)(24)(25) have deciphered one such mechanism that involves the HIV-encoded virion infectivity factor blocking the enzymatic activity of cytidine deaminase CEM15 (also known as APOBEC3G for apolipoprotein B mRNA-editing enzyme, catalytic polypeptide-like 3G), which induces G:C to A:T hypermutation in newly synthesized DNA. Another type of mutation, G:C to T:A transversion, is mediated by the generation of 8-oxo-7,8-dihydro-2Ј-deoxyguanosine (8-oxo-dG) by radical oxygen species (ROS) and occurs at the DNA level (26,27). The oxidatively damaged guanosine, 8-oxo-dG, is widely accepted as a pre-mutagenic lesion because of its potential to mispair with adenine, thus generating the G:C to T:A transversion. This type of mutation is often found in tumor suppressor genes and oncogenes, such as p53 and K-ras, in mammalian cells (28,29). The OGG1 (8-oxoguanine-DNA glycosylase-1) enzyme is responsible for the excision/repair of this oxidatively damaged DNA by excising 8-oxo-dG (30 -32). In fact, OGG1 gene knockout actually shows accumulation of such a mutation (33,34).
In this study, we demonstrate the up-regulation of OGG1 by Tat and provide evidence that this effect of Tat is through the sequestration of the negative transcription factor AP-4 for the expression of OGG1. We examine the effect of Tat on the actual levels of 8-oxo-dG in the presence and absence of small interfering RNA (siRNA) against OGG1 mRNA.

EXPERIMENTAL PROCEDURES
Plasmids-The cDNA of wild-type Tat (101 amino acids) originating from HIV-1 was amplified by PCR with the oligonucleotide primer pair 5Ј-CGC GGA TCC GCG CCA CCA TGG ATT ACA AGG ATG ACG ACG ATA AGA TGG AGC CAG TAG ATC CTA GAC TAG AGC CCT GG-3Ј (forward; containing an EcoRI site and a FLAG epitope) and 5Ј-CCG GAA TTC CGG CTG ATG GAC CGG ATC TGT CTC-3Ј (reverse; containing a BamHI site). The amplified DNA fragment was digested with EcoRI and BamHI and ligated in-frame into the pIND-V5 expression vector (Invitrogen), thus generating pIND-Tat. As a control, we employed mutant Tat (mTat) lacking transcriptional activity because of the absence of binding activity with cyclin T1 or the transactivation response region (6 -8). The plasmid expressing mutant Tat (pIND-mTat) in which Cys 30 and Lys 41 were substituted with Ala was generated using a QuikChange site-directed mutagenesis kit (Stratagene) with the following mutagenic oligonucleotide primer pairs: 5Ј-CTA TTG TAA AAA GGC CTG CTT TCA TTG CC-3Ј (forward) and 5Ј-GGC AAT GAA AGC AGG CCT TTT TAC AAT AG-3Ј (reverse) or 5Ј-GTT TCA CAA CAG CCG CCT TAG GCA TC-3Ј (forward) and 5Ј-GAT GCC TAA GGC GGC TGT TGT GAA AC-3Ј (reverse). To generate the mammalian expression plasmid for AP-4, AP-4 gene was amplified by PCR with the oligonucleotide primer pair 5Ј-CGC GGA TCC GCG CCA CCA TGG ATT ACA AGG ATG ACG ACG ATA AGA TGG AGC CAG TAG ATC CTA GAC TAG AGC CCT GG-3Ј (forward; containing an EcoRI site) and 5Ј-CCG GAA TTC CGG CTG ATG GAC CGG ATC TGT CTC-3Ј (reverse; containing a BamHI site). The amplified DNA fragment was digested with EcoRI and BamHI and ligated in-frame into the pcDNA-Myc expression vector (Invitrogen). The construction of pCD12luc, containing the HIV-1 long terminal repeats V3 and R linked to the luciferase gene, and pcDNA-Tat was described previously (35,36). Human OGG1 promoter-luciferase fusion constructs, including pPR116, pPR128, pPR130, and pPR143, were kindly provided by Dr. J. P. Radicella (Radiobiologie Moleculaire et Cellulaire, CNRS-CEA, Fontenay aux Roses, France) (37). The mutant pPR128-luc reporter constructs were generated using a QuikChange site-directed mutagenesis kit. The mutant sequences (sense strand) utilized were as follows: 5Ј-AP-4 site mutant (m5ЈAP-4), GAC GGC AGG CAG tcg cga TGG CGG CCG GCG; 3Ј-AP-4 site mutant (m3ЈAP-4), GGG AAA GGC GAG tcg cga GCA GAG AGC CCA G; GA TA site mutant (mGATA), CTT GCA GCC Tct TAG TTA AGA TAC AGC; and AP-2 site mutant (mAP-2), CAG CTG TGG CGG CCa ttC GGG ACG ACA ATC (with consensus binding sites underlined and mutated sequences in lowercase letters). The construct containing mutations in both the 5Ј-and 3Ј-AP-4 sites (mwAP-4) was generated by two successive PCRs using the m5ЈAP-4 and m3ЈAP-4 mutant sequences. The control luciferase reporter plasmid pGL3-Basic vector was purchased from Promega. All constructs were confirmed by dideoxynucleotide sequencing using an ABI PRISM TM dye terminator cycle sequencing ready kit (PerkinElmer Life Sciences) on an Applied Biosystems 313 Automated DNA Sequencer.
Cell Lines That Inducibly Express Tat, mTat, and LacZ-HEK293-EcR cells, stably transfected with pVgRXR expressing the ecdysone receptor, were purchased from Invitrogen and transfected with pIND-Tat and pIND-mTat to establish Tat/293 and mTat/293 cells, respectively. The control cell line (LacZ/293) was a gift from Dr. L. Naumovski (Stanford University, Stanford, CA) (38). Expression of these genes is under the stringent control of a homolog of the insect hormone 20-OHecdysone, ponasterone A (PonA; Invitrogen). Cells containing these plasmids were selected by 500 g/ml G418 and 450 g/ml Zeocin. Cell clones were singly isolated by two successive rounds of limiting dilution of cells and were screened for the expression and transcriptional activity of Tat proteins.
Preparation of mRNA-Total cellular RNA was prepared from each cell clone using RNeasy (Qiagen Inc.). Purification of polyadenylated mRNA was carried out using an Oligotex-dT30 super RNA purification kit (Takara, Ohtsu, Japan) as described previously (39,40). The mRNA samples were digested with RNase-free DNase, ethanol-precipitated, and further purified through Microcon YM-100 columns (Amicon Inc.). The quantity and quality of mRNA were assessed by capillary electrophoresis using an Agilent 2100 bioanalyzer.
Generation of Fluorescently Tagged cDNA and Gene Expression Profile Analysis-Gene expression profiles were examined as described (39,40) using the human 3K DNA CHIP TM (Takara) containing 2600 human genes of known functions. Briefly, fluorescently labeled cDNA was synthesized from 1-g aliquots of purified mRNA by oligo(dT)-primed polymerization using SuperScript II reverse transcriptase (Invitrogen). The pool of nucleotides in the labeling reaction contained 0.5 mM each dGTP, dATP, and dTTP; 0.3 mM dCTP; and 0.1 mM fluorescent nucleotide (Cy3-or Cy5-labeled dCTP, Amersham Biosciences). Fluorescently labeled cDNA was purified by chromatography through Microcon YM-20 columns (Amicon Inc.). The microarray slide was hybridized to combined Cy5-dCTP-and Cy3-dCTP-labeled cDNA probes for 14 h in hybridization solution (6ϫ SSC and 0.2% SDS with 5ϫ Denhardt's solution and carrier DNA) at 65°C under coverslips. After hybridization, the microarray slide was washed twice with 1.2ϫ SSC and 0.2% SDS at 55°C for 5 min, with 1.2ϫ SSC and 0.2% SDS at 65°C for 5 min, and with 0.05ϫ SSC at room temperature as a final wash. The hybridized array was scanned at 10-m resolution on an Affymetrix 428 array scanner. Analysis of differential expression of each gene was performed using ImaGene Version 4.2 computer software (Bio-Discovery Ltd.). Normalization of hybridized signals was performed by global scaling. These experiments were repeatedly performed: we performed comparative microarray analyses three times (24 h after PonA stimulation in 293/Tat and 293/LacZ cells) and two times (12 h after PonA stimulation in 293/Tat and 293/LacZ cells).
Co-immunoprecipitation and Immunoblot Assay-The experimental procedures have been described previously (41). Briefly, cells were harvested with lysis buffer (25 mM HEPES-NaOH (pH 7.9), 150 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA, 0.3% Nonidet P-40, 1 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride). The lysates were cleared by centrifugation, and the supernatants were incubated overnight with the indicated antibodies at 4°C. For immunoprecipitation with the FLAG epitope, anti-FLAG antibody M2 affinity gel beads (Sigma) were used. The immune complexes were washed three times with 1 ml of lysis buffer, and the antibody-bound proteins were dissolved by boiling in 2ϫ Laemmli sample buffer. After centrifugation, the supernatant proteins were separated by SDS-PAGE and transferred to nitrocellulose membrane (Hybond-C, Amersham Biosciences). The membrane was probed with antibodies, including anti-cyclin T1 and anti-AP-4 (Santa Cruz Biotechnology Inc.), anti-FLAG (Sigma), and anti-Myc (Invitrogen) antibodies; and immunoreactive proteins were visualized by enhanced chemiluminescence (SuperSignal, Pierce). To evaluate the level of OGG1 protein, cells were similarly treated with lysis buffer, and the cell lysate was analyzed by Western blotting using anti-OGG1 antibody (Novus Biologicals, Inc.).
Transfection and Luciferase Assay-Cells were transfected using Fu-GENE 6 transfection reagent (Roche Applied Science) as described (36). Jurkat cells were transiently transfected by electroporation (42). Briefly, cells (2 ϫ 10 7 /ml) were electroporated in the presence of 2 g of pcDNA-Tat or control plasmid (pcDNA3.0, Invitrogen) in 400 l of serumfree RPMI 1640 medium using the Electro Cell Manipulator 600 apparatus (BTX) at 260 V/1050 microfarads. For the internal control, we employed pRL-TK, expressing Renilla luciferase, which is not modified by Tat action. The transfected cells were harvested, and the extracts were subjected to luciferase assay using the Luciferase Assay System TM (Promega). The luciferase activity was normalized to Renilla luciferase activity as an internal control to assess the transfection efficiency. The data are presented as the -fold increase in luciferase activities (means Ϯ S.D.) relative to the control from three independent transfections.
Reverse Transcription (RT)-PCR-For cDNA synthesis, 1 g of purified total RNA were reverse-transcribed using oligo(dT) primer and SuperScript II reverse transcriptase. The cDNA was then amplified from each RNA sample with Taq PCR Master Mix (Qiagen Inc.) and gene-specific primers designed using Oligo Version 4.0 software (Molecular Biology Insights). The primer sequences for each amplified gene were as follows: TFPI2 (tissue factor pathway inhibitor-2), 5Ј-CAG GAG CCA ACA GGA AAT AAC-3Ј (forward) and 5Ј-GAA TAC GAC CCC AAG AAA TGA-3Ј (reverse); OGG1, 5Ј-GCG TGC GCA AGT ACT TCC AGC-3Ј (forward) and 5Ј-CCA GTG ATG CGG GCG ATG TTG-3Ј (reverse); OGG1 type 1, 5Ј-GCG TGC GCA AGT ACT TCC AGC-3Ј (forward) and 5Ј-TAA AGG GAA GAT AAA ACC ATC-3Ј (reverse); OGG1 type 2, 5Ј-GCG TGC GCA AGT ACT TCC AGC-3Ј (forward) and 5Ј-GCA TCA CAT GAC CAA TTA CTG-3Ј (reverse); MEN1␤2 homolog Demand TM ). Quantitative measurements of each mRNA level were performed in triplicate. The accuracy of mRNA quantitation for each gene was confirmed by measurement of serially diluted control mRNA samples and comparison of the fluorescent intensity from a standard carve of the mRNA levels. Amplification of ␤-actin mRNA was performed with all samples to control the variation in mRNA levels. The gene expression levels were normalized to ␤-actin levels for each mRNA preparation, and the -fold increase in an individual gene was calculated by comparison with the result obtained without PonA stimulation. The non-template control was incubated in each amplification reaction to exclude the contaminating template.
HIV Infection and p24 Determination-PBMCs were stimulated with phytohemagglutinin for 48 h and infected with HIV-1 MN . HIV-1 MN was challenged at 100 TCID 50 (where TCID 50 is the tissue culture infectious dose resulting in 50% infected cells)/2.0 -2.5 ϫ 10 6 PBMCs or Jurkat cells for 1 h (43). These cells were washed twice with phosphatebuffered saline (PBS), and HIV-1 p24 antigen concentration in the culture supernatant was measured using a commercial kit (ZeptoMetrix Corp., Buffalo, NY).
mRNA Stability Assay-Total cellular RNA was serially prepared from Tat/293 and mTat/293 cells treated with PonA for 24 h, followed by the treatment with 2 g/ml actinomycin D (Sigma). The amount of OGG1 mRNA was analyzed by real-time RT-PCR as described above.
Detection of ROS-The detection of ROS was measured by flow cytometry with a FACScan (BD Biosciences) equipped with argon ion laser delivering 200 megawatts of power at 488 nm, and the results were analyzed using CellQuest TM software (BD Biosciences). After treatment of cells with PonA (10 M) to induce Tat or mTat, the oxidation-sensitive fluorescent probe 2Ј,7Ј-dichlorofluorescein diacetate (Molecular Probes, Inc.) was added to the culture at a final concentration of 5 M, followed by incubation at 37°C for 30 min. The cells were washed twice with PBS, resuspended in PBS, and then subjected to the fluorescence-activated cell sorter detection of 5,6-carboxy-2Ј,7Ј-dichlorofluorescein (green) at 530 nm.
Measurement of Intracellular Reduced GSH and GSSG Contents-The total cellular GSH and GSSG concentrations were measured using a glutathione quantification kit (Dojindo). Briefly, each cell culture (5 ϫ 10 6 cells) was centrifuged at 2000 ϫ g for 5 min, and the cell pellet was resuspended in 100 l of PBS. The cell suspension was treated with 80 l of 10 mM HCl, crashed by freeze/thawing repeated twice, and further treated by adding 20 l of 5% sulfosalicylic acid to precipitate the proteins. The supernatant was obtained by centrifugation at 10,000 ϫ g for 10 min, and the total GSH and GSSG concentrations were determined. The sample was incubated with GSH reductase and NADPH at 30°C for 5 min, and the total GSH concentration thus generated was measured by reaction with 5,5Ј-dithiobis(2-nitrobenzoic acid) at 30°C for 10 min. Spectrophotometric detection of 5-mercapto-2-nitrobenzoic acid was performed at 415 nm. The GSSG concentration was determined by performing the same procedure, but GSH was masked by treatment with 2-vinylpyridine and triethanolamine prior to the reaction with GSH reductase and NADPH. The concentration of reduced GSH was determined by subtracting the GSSG concentration from the total GSH concentration. Each determination was performed in triplicate, and experiments were repeated at least twice.
Measurement of Manganese Superoxide Dismutase Activity-The enzymatic activity of manganese superoxide dismutase was measured using a WST-1 TM superoxide dismutase assay kit (Dojindo) with slight modifications. Briefly, equal numbers of cells (1.5 ϫ 10 7 ) were washed twice with PBS, and the cell lysates were extracted by freeze/ thawing. The manganese superoxide dismutase activity in the supernatant protein lysate was determined by incubation with WST-1, xanthine, and xanthine oxidase at 37°C for 20 min. To mask the copper superoxide dismutase and zinc superoxide dismutase activities, the protein lysate was treated with 1 mM KCN. The inhibitory action of manganese superoxide dismutase contained in each cell lysate was assessed by the spectrophotometric determination of WST-1 formazan at 450 nm. Quantitation was achieved by comparison with the absorption of standard manganese superoxide dismutase (Sigma) with known concentrations.
Electrophoretic Mobility Shift Assay-The experimental procedure was carried out as described previously (36). The AP-4 consensus sequence was taken from the 5Ј-AP-4 site in the OGG1 promoter. The wild-type and mutant oligonucleotide sequences (sense strand) utilized were as follows: wild-type, 5Ј-CGG CAG GCA GCA GCT GTG GCG G-3Ј; and mutant, 5Ј-CGG CAG GCA GTC GCG ATG GCG G-3Ј. These oligonucleotides were labeled using a 5Ј-end labeling kit (Takara) in the presence [␥-32 P]dATP (Amersham Biosciences). DNA binding reactions were performed at 4°C for 20 min in a 10-l reaction volume. Analysis of binding complexes was performed by electrophoresis on 6% polyacrylamide gels with 0.5ϫ Tris borate/EDTA buffer, followed by autoradiography. The specificity of DNA binding was assessed by preincubating extracts with anti-AP-4 antibody and competitor at room temperature for 10 min.
Chromatin Immunoprecipitation (ChIP) Assay-ChIP assay was performed according to the recommendations of Upstate Biotechnology, Inc., with some modification. Briefly, cells were cross-linked with 1% formaldehyde for 10 min at room temperature, washed twice with ice-cold PBS, and lysed for 10 min at 2 ϫ 10 6 cells in 200 l of SDS lysis buffer. The chromatin was sheared by sonication 13 times for 10 s at one-third of the maximum power with 20 s of cooling on ice between each pulse. Cross-linked released chromatin fractions were precleared with salmon sperm DNA and protein A-agarose beads for 1 h, followed by immunoprecipitation overnight with the desired antibodies at 4°C. The immunoprecipitates were sequentially washed once with lysis buffer, twice with high salt buffer, twice with low salt buffer, and twice with Tris/EDTA buffer. After the wash, immune complexes were collected with salmon sperm DNA and protein A-agarose beads at room temperature for 1 h and extracted with 1% SDS and 0.1 M NaHCO 3 . The eluted samples were reverse-cross-linked with proteinase K at 65°C for 6 h and treated with RNase at 37°C for 1 h. DNA was recovered by phenol/chloroform and chloroform extraction and ethanol precipitation. Finally, DNA was dissolved in 30 l of Tris/EDTA buffer and subjected to PCR. The primer sequences were as follows: OGG1 promoter (Ϫ615 to Ϫ450), 5Ј-CAA ACG TCC CAT TCC GAG GAA AG-3Ј (forward) and 5Ј-GGC CTT TAG GCG TCC TCT GAG A-3Ј (reverse); and ␤-actin promoter (Ϫ980 to Ϫ915; used as a control), 5Ј-TGC ACT GTG CGG CGA AGC-3Ј (forward) and 5Ј-TCG AGC CAT AAA AGG CAA-3Ј (reverse). The number of PCR cycles was as follows: 33 PCR cycles for all ChIP experiments and 25 PCR cycles for the input samples, in which PCR amplification was performed under the linear range of AP-4 binding to the OGG1 promoter. For each reaction, 10% of the cross-linked released chromatin was saved and reversed by proteinase K digestion at 65°C for 6 h, followed by DNA extraction; and the recovered DNA was used as input control.
Measurement of 8-Oxo-dG-The amounts of 8-oxo-dG in the cellular DNA were measured using a high performance liquid chromatography (HPLC)-electrochemical detector (ECD) system, which is highly selec-tive, with sensitivity at the femtomole level, as described previously (26,44). Briefly, cellular DNA was isolated using a DNA extractor WB kit (Wako Pure Chemical Industries, Osaka, Japan). The isolated DNA was digested with P1 nuclease (Wako Pure Chemical Industries) to obtain 8-oxo-dG in the nucleoside form (8-hydroxydeoxyguanosine). The nucleoside solution was filtered with an Ultrafree Probind filter (Millipore Corp.) and injected into an HPLC column (CAPCELL PAK C 18 MG, Shiseido, Tokyo, Japan) equipped with an ECD (Coulochem II, ESA, Inc.) at a flow rate of 0.8 ml/min with the mobile phase consisting of 10 mM Na 2 HPO 4 and 8% methanol. The 8-oxo-dG value in the DNA was calculated as the number of 8-oxo-dG residue/10 6 dG residues.

Establishment and Characterization of Ecdysone-inducible Cell Lines Expressing Tat and mTat-Because
Tat is known to impair cell viability and its activity in long-term maintenance of cells expressing Tat may preselect certain cell types and preclude exploration of the Tat-mediated alteration of cellular functions, we adopted a stringent ecdysone-inducible system using PonA, a phytoecdysteroid that is inert in mammals (45). To generate Tat-or mTat-expressing cells upon treatment with PonA, we transfected the pIND-Tat or pIND-mTat plasmid into HEK293-EcR cells stably transfected with pVgRXR expressing the receptor for ecdysone (PonA). These cells were singly isolated by two successive rounds of limiting dilution and screened for expression of Tat proteins and their transcriptional activity in stimulating HIV gene expression.
As shown in Fig. 1A, the expression of Tat and mTat proteins was detected after 12 h of PonA treatment and was maintained for at least an additional 60 h. (PonA dose-dependent Tat expression is also shown.) In Fig. 1B, Tat-mediated HIV-1 transactivation was examined. As expected, Tat (but not mTat) augmented HIV-1 gene expression in a PonA dose-dependent manner. Fig. 1C shows that Tat (but not mTat) bound to endogenous cyclin T1 in cells as reported previously (6 -8).
These results indicate that both Tat/293 and mTat/293 cells inducibly express Tat proteins and that the functional integrity of Tat is maintained in Tat/293 cells. Thus, we explored the gene expression profile in these cells.
Gene Expression Profile Analysis of Cells Expressing Tat-To identify genes either up-regulated or down-regulated by Tat in the newly established ecdysone-inducible cell lines, the gene expression profiles were compared with and without Tat expression. The mRNA was isolated from Tat/293 and Lac Z/293 cells without (control) and with PonA treatment. The cDNA probes were synthesized from each mRNA, labeled with Cy5 (for Tat-or LacZ-expressing cells) or Cy3 (for non-expressing cells), and hybridized to a gene chip (human 3K DNA CHIP TM ). Representative results are shown in Fig. 2A, where we compared genes expressed in Tat-expressing (Cy5-labeled) and non-expressing control (Cy3-labeled) 293/Tat cells. Similar comparisons were made with LacZ-expressing cells (data not shown).
As shown in Fig. 2A, 12 genes, including TFPI2, stanniocalcin-1, SEPP1, and OGG1, were up-regulated by Ͼ2-fold by Tat after 24 h of induction upon PonA treatment. Five of these 12 genes were up-regulated by Ͼ2-fold even after 12 h of Tat induction. The details of these genes are summarized in Table  I. In control LacZ/293 cells, expression of the TFPI2 gene was up-regulated by 2.0-fold when cells were treated with PonA, suggesting nonspecific stimulation by PonA (data not shown).
On the other hand, eight genes were down-regulated to Ͻ60% after 24 h of Tat induction (summarized in Table II). Downregulation of these genes was not observed in LacZ-expressing cells (data not shown).
To confirm these results, we carried out RT-PCR analysis to examine the mRNA levels of Tat-regulated genes before and after Tat induction. We also examined the effect of mTat to further confirm the specificity of Tat action. Fig. 2B shows the results of eight genes up-regulated by Tat by Ͼ2.3-fold (stanniocalcin-1, SEPP1, OGG1, MEN1␤2 homolog FLJ23538/clone 137308, ETV5, SLC20A1, integrin ␣7, and ENPP2; excluding TFPI2). Among these genes, stanniocalcin-1, SEPP1, and ETV5 were also up-regulated by induction of mTat or even LacZ, suggesting a nonspecific effect of PonA. Thus, we concluded that five genes (OGG1, MEN1␤2 homolog FLJ23538/clone 137308, integrin ␣7, SLC20A1, and ENPP2) are specifically up-regulated by Tat because these genes were not up-regulated by either mTat or LacZ. Whereas ENPP2 was up-regulated after 12-24 h of Tat induction and subsequently down-regulated, the other four genes were constitutively up-regulated.
Similarly, RT-PCR analysis was performed with the eight genes down-regulated by Tat (Fig. 2C). Two of the genes (SLC1A3 and LTA1) were down-regulated by mTat or LacZ. The other six genes were down-regulated by Tat, but not by mTat or LacZ. Among the genes down-regulated by Tat, NDRG1, RGS16, and hexokinase-2 are known to be under the transcriptional control of p53 (46 -48), an observation consistent with previous reports of Tat down-regulating the action of p53 (49,50).
Induction of OGG1 Gene Expression by Tat-Because Tat up-regulated the OGG1 gene the most, we further analyzed the effect of Tat on OGG1 mRNA expression and stability. The human OGG1 gene encodes two isoforms, type 1 (a and b) and type 2 (a, b, and c), resulting from alternative splicing of the single OGG1 precursor mRNA (51). As shown in Fig. 3A, Tat induced all types of OGG1 mRNA. We carried out real-time RT-PCR to determine more precisely the mRNA levels of OGG1 before and after Tat induction. As shown in Fig. 3B, OGG1 gene expression was up-regulated by 3.6-, 6.7-, 9.8-, and 8.2-fold upon Tat expression after 6, 12, 24, and 48 h of PonA treatment, respectively. mTat did not affect OGG1 gene expression (Fig. 3B). A similar extent of stimulation was observed for OGG1 protein levels as revealed by Western blotting (Fig. 3C). No induction of OGG1 protein was observed when mTat was induced. Induction of OGG1 protein by Tat (but not mTat) was confirmed in Jurkat T cells, a natural host of HIV-1 infection (Fig. 3C, right panels). Furthermore, the effect of HIV-1 infection on OGG1 gene expression was examined with PBMCs isolated from two individuals and Jurkat cells. These cells were infected with HIV-1 MN , and the OGG1 mRNA level was quantified by real-time RT-PCR together with the amounts of HIV-1 produced in the culture supernatant. As shown in Fig. 3D, up-regulation of OGG1 mRNA levels upon HIV-1 infection was observed and was associated with elevation of viral p24 antigen levels. Mock infection did not induce OGG1 expression (data not shown).
Furthermore, we examined the effect of Tat on the stability of OGG1 mRNA using PonA-inducible cells. After 24 h of Tat or mTat induction, cells were treated with actinomycin D, and total RNA samples were obtained after 1, 3, 5, and 7 h of actinomycin D treatment to determine the level of OGG1 mRNA. As shown in Fig. 3E, the decay profiles of OGG1 mRNA were similar in cells expressing Tat and mTat (control), with half-lives of 5.0 and 4.3 h, respectively. These findings indicate that the positive effects of Tat on OGG1 gene expression are at the level of transcription.

Induction of OGG1 Expression Is Not through the Oxidative Stress Induced by Tat-Because
Tat is known to induce oxidative stress (12,15,17), it is possible that OGG1 induction might be an indirect effect of Tat, although it is not yet known whether oxidative stress induces OGG1 gene expression. We first examined whether Tat induces oxidative stress in Tat/293 cells. Fluorescence-activated cell sorter analysis with the oxidation-sensitive fluorescent probe 2Ј,7Ј-dichlorofluorescein diacetate showed that Tat (but not mTat) increased the intracellular ROS levels (Fig. 4A). We measured the intracellular levels of GSH and GSSG. As expected, the content of GSSG, the oxidized form of GSH, was increased (ϳ2.3-fold) in Tat-expressing cells in contrast to control mTat-expressing cells (Fig. 4B). The GSH (reduced form) level in Tat-expressing cells was decreased to 57% of that in control cells. No significant reduction in GSH was detected in mTat-expressing cells. Furthermore, the manganese superoxide dismutase activity was down-regulated by Tat as reported previously (12). After 24 h of Tat induction, the manganese superoxide dismutase activity was decreased to 59% (Fig. 4C).
These results led us to examine whether Tat-induced OGG1 gene expression is attributable to the oxidative stress associated with Tat action. However, treatment with antioxidants did not block Tat-mediated OGG1 induction (Fig. 4D). 293/Tat cells were pretreated with antioxidants, including pyrrolidine dithiocarbamate, N-acetyl-L-cysteine, epigallocatechin gallate (a phenolic antioxidant), and Trolox (a water-soluble vitamin E analog), prior to Tat induction (Fig. 4D, left panel). When the OGG1 mRNA was measured by real-time RT-PCR, Tat-induced OGG1 expression was not affected by the antioxidants. In  (for control 293 cells). These probes were hybridized, in combination, to a gene chip (human 3K DNA CHIP TM ). The signal intensity of each gene on the microarray chip was addition, H 2 O 2 and oxidative stress inducers such as inflammatory cytokines (tumor necrosis factor-␣ and interleukin-1␤) and lipopolysaccharide could not up-regulate OGG1 gene expression (Fig. 4D, right panel). In support of these findings, when we performed transient luciferase assay using the OGG1 promoter construct, no OGG1 induction by these stimuli was observed (data not shown), consistent with a previous report by Dhénaut et al. (37). Therefore, it is unlikely that Tat induces OGG1 expression through ROS production.
Transactivation of OGG1 by Tat-These observations prompted us to examine the possibility that Tat-mediated OGG1 expression is the direct effect of Tat. We thus examined the effect of Tat on OGG1 promoter activity. The transient luciferase assay was performed on various regions of the OGG1 promoter linked to the luciferase reporter gene (Fig. 5). As shown in Fig. 5A, Tat stimulated the transcriptional activity of the reporter constructs containing the sequence upstream from position Ϫ472. Whereas the sequence upstream from position Ϫ945 was dispensable for Tat-mediated transactivation, no such effect was observed when the region spanning positions Ϫ945 to Ϫ472 was deleted (Fig. 5A, lower left panel). In 293 cells expressing mTat, no induction of OGG1 promoter activity (other than the constitutive transcriptional activity) was observed. We also examined the Tat-mediated transactivation of the OGG1 promoter in Jurkat cells and obtained essentially the same results (Fig. 5A, lower right panel).
Thus, Tat appears to transactivate OGG1 expression through transcription factors located within the OGG1 promoter region from positions Ϫ945 to Ϫ472. To further elucidate the mechanism by which Tat induces OGG1 transcription, we created specific mutants lacking binding sites for GATA, AP-4, and/or AP-2 located in this region. When AP-4 sites were mutated, Tat no longer augmented the promoter activity (Fig. 5B,  black bars). However, no reduction in Tat-mediated transactivation was observed when other sites were mutated. In addition, the basal promoter activity was augmented when the 5Ј-AP-4 site was mutated, whereas mutation of the GATA, AP-2, and 3Ј-AP-4 sites had little effect on the basal OGG1 promoter activity (Fig. 5B, hatched bars), indicating that AP-4 acts as a transcriptional repressor of OGG1 expression. These results suggest that AP-4 sites are required for Tat-induced OGG1 gene expression and that the 5Ј-AP-4 site negatively regulates OGG1 gene expression. In fact, overexpression of AP-4 inhibited both the Tat-stimulated and basal levels of OGG1 gene expression without affecting the level of Tat expression (Fig. 5C).
Tat Interacts with AP-4 and Removes It from the OGG1 Promoter-To further investigate the mechanism by which Tat stimulates OGG1 gene expression, we first examined the effect of Tat on AP-4 DNA binding by electrophoretic mobility shift assay. As shown in Fig. 6A, constitutive AP-4 DNA binding was observed in the cells, and a significant reduction in AP-4 DNA binding was observed when Tat was induced by PonA treatment. No such effect was observed when mTat was expressed. We then examined whether Tat associates with AP-4 in cultured cells by co-immunoprecipitation with either Tat (FLAG epitope-tagged) or AP-4 (Myc epitope-tagged). As shown in Fig.  6B, when Tat was immunoprecipitated with anti-FLAG antibody, endogenous AP-4 protein was detected within the immune complex. No AP-4 was co-immunoprecipitated with mTat. Conversely, when AP-4 was immunoprecipitated with plotted. Genes with a signal intensity Ͻ200 U (of Cy5 and Cy3) were excluded from further analysis. Solid and dashed lines indicate the upper and lower boundaries of 1.5-and 2.0-fold changes, respectively. B and C, confirmation of genes up-or down-regulated by Tat using RT-PCR. B, genes up-regulated by Tat. Up-regulation of the stanniocalcin-1, SEPP1, and ETV5 genes observed in cDNA array analysis appeared to be unspecific. C, genes down-regulated by Tat. Down-regulation of the SLC1A3 and LTA1 genes was considered nonspecific. RT-PCR analysis was performed with gene-specific primers and total RNA prepared from Tat/293, mTat/293, and LacZ/293 cells. Each cell culture was treated with PonA (10 M) for the indicated periods of time. N.D., not determined.
a Data from single determinations 12 h after Tat induction. b Data from triplicate determinations 24 h after Tat induction (expressed as means Ϯ S.D.).
c Nonspecifically up-regulated by PonA treatment (see Fig. 2 for the results of RT-PCR).
Furthermore, the ChIP assay was performed to examine whether the inhibition of AP-4 DNA binding by Tat occurs at the endogenous OGG1 gene promoter. Tat/293 and mTat/293 cells were transfected with plasmids expressing Myc-tagged AP-4 or Myc-tagged LacZ (control), stimulated with PonA to express Tat or mTat, treated with formamide, and sonicated, and the cross-linked protein⅐DNA complex was immunoprecipitated with specific antibodies recognizing the Myc (AP-4) or V5 (Tat) epitope. The immunoprecipitated DNA was analyzed by PCR using primer pairs for the AP-4-binding sites within the OGG1 promoter (Ϫ615 to Ϫ450). As demonstrated in Fig. 6C, a significant reduction in AP-4 bound to the OGG1 promoter DNA was observed, and the extent of reduction was proportionate to the amount of Tat expressed (corresponding to the timedependent expression of Tat in Fig. 1). No such effect was observed when mTat was expressed. The antibody to Tat precipitated the OGG1 promoter (Fig. 6C, left panels), indicating that the Tat⅐AP-4⅐DNA ternary complex may be transiently formed. These observations indicate that Tat directly activates OGG1 gene expression through sequestrating AP-4, a negative transcriptional regulator of OGG1 expression.
Reduction of 8-Oxo-dG Levels by Tat and Effect of OGG1 Knockdown-Because OGG1 is responsible for the excision/repair of the oxidation-damaged DNA by excising 8-oxo-dG (30) and because Tat induces expression of OGG1, we measured the amounts of 8-oxo-dG in the cellular DNA by the HPLC-ECD method (26). Fig. 7A shows the levels of 8-oxo-dG before and after Tat expression. In control cells, the level of 8-oxo-dG was 8.7 Ϯ 0.34 residues/10 6 dG residues (Fig. 7B). However, upon expression of Tat, the 8-oxo-dG levels were reduced to 7.7 Ϯ 1.6 (0.89fold), 5.6 Ϯ 0.30 (0.64-fold), and 4.5 Ϯ 1.2 (0.52-fold) residues/10 6 dG residues after 6, 12, and 24 h of Tat induction, respectively. Statistically significant reduction was observed after 12 and 24 h of Tat induction. No significant reduction in 8-oxo-dG levels was observed when mTat was expressed. Taken together, these observations indicate that Tat prevents accumulation of 8-oxo-dG by directly up-regulating OGG1 expression.
To confirm these findings, we adopted a siRNA technique to specifically knock down OGG1 mRNA and examined the effect of Tat on the level of 8-oxo-dG when endogenous OGG1 was

DISCUSSION
In this study, we have explored the biological effects of Tat using gene expression profile analysis. We found that (i) Tat induces the OGG1 gene and that (ii) Tat down-regulates the NDRG1, RSG16, and hexokinase-2 genes. The latter genes are known to be under the transcriptional control of p53 (46 -48). Interestingly, Li et al. (49) observed the repression of p53 mRNA by Tat. Moreover, Tat was shown to directly inactivate p53 by protein-protein interaction (50,52). Thus, the Tat-mediated down-regulation of these genes is consistent with previ-ous findings. However, Tat-mediated OGG1 induction has not been reported. Thus, in this study, we analyzed the mechanism by which Tat induces OGG1 gene expression.
We found that Tat-mediated OGG1 induction is not through stabilization of the OGG1 mRNA. In addition, Tat-mediated OGG1 induction was not reversed by treatment with antioxidants, indicating that Tat-mediated OGG1 induction could not be attributable to oxidative stress induced by Tat. By performing transient luciferase assay using the reporter plasmid containing various regions of OGG1, we found that Tat induces OGG1 gene expression through the central AP-4 site (located at positions Ϫ545 to Ϫ540) in its upstream region. Although AP-4 is known to activate expression of the SV40 (53) and transforming growth factor-␤ (54) genes, it is also known to repress expression of the human angiotensinogen (55) and HIV-1 (56) genes, although the mechanism of AP-4 action has not been clarified. Our experiments have revealed that AP-4 negatively regulates OGG1 gene expression via binding to the central promoter constructs in the presence (black bars) or absence (hatched bars) of pcDNA-Tat. Cells were harvested, and the luciferase (luc) activity was measured. The data are presented as the -fold increase in luciferase activity (means Ϯ S.D.) relative to the transcriptional activity of the full-length OGG1 promoter (pPR116) from three independent experiments. B, involvement of AP-4 in Tatmediated OGG1 expression. Various mutations in the binding sites for GATA, AP-4, and AP-2 were introduced into a wild-type OGG1 promoter (pPR128-luc), and the effects of Tat were examined. Schematic representations of wild-type and mutant pPR128-luc constructs are shown (upper panel). Tat/293 cells were transfected with these reporter constructs, and the effects of Tat were similarly examined (lower panel). Note the increase in basal promoter levels with mutants containing a substitution in the 5Ј-AP-4 site and the lack of Tat response in mutants containing 5Ј-and/or 3Ј-AP-4 sites. C, repression of OGG1 expression by AP-4. Tat/293 cells were transfected with pPR116-luc together with various amounts of plasmid expressing Myc-AP-4, incubated for 24 h, and treated with PonA (10 M) for an additional 24 h to induce Tat expression (upper panel). The dosedependent expression of AP-4 was revealed by Western blotting with anti-Myc tag antibody (lower panels). Tat protein expression was monitored using anti-FLAG antibody.
AP-4 site and that Tat activates OGG1 promoter activity by sequestrating AP-4 from the OGG1 promoter. Thus, the positive effect of Tat on OGG1 gene expression appears to be a direct effect.
Intriguingly, we observed that the extent of oxidation-induced guanosine modification (8-oxo-dG) was reduced, although Tat induced oxidative stress as revealed by the increase in ROS and GSSG and the decrease in manganese superoxide dismutase activity. When OGG1 expression was knocked down by siRNA, the amount of 8-oxo-dG was increased, suggesting that the Tat-induced reduction of 8-oxo-dG requires OGG1 gene expression and that the Tat-mediated induction of OGG1 appears to be independent of Tat pro-oxidant action.
Thus, in addition to its crucial role in viral replication, Tat appears to play a role in maintenance of the genetic integrity of proviral and host cell DNAs. Although various conditions associated with HIV infection and replication are pro-oxidant (12,13,(57)(58)(59), the observed mutations accumulated within the HIV genome have been revealed to be in favor of the G:C to A:T transition (18 -22) rather than the G:C to T:A transversion mediated by the oxidative modification. This is in contrast with most of the mutations associated with human cancers, where the G:C to T:A transversion is predominant (28,29). If induction of OGG1 were through oxidative stress associated with Tat actions, the level of 8-oxo-dG should have been higher in Tatexpressing cells than in control cells. These findings indicate that Tat-mediated OGG1 induction is more than a feedback action. Although additional studies are needed, such as the FIG. 6. Mechanism by which Tat induces OGG1 gene expression: Tat interacts with AP-4 and removes AP-4 from the OGG1 promoter. A, effect of Tat on AP-4 DNA binding. Nuclear extracts were prepared from Tat/293 and mTat/293 cells, and AP-4 DNA binding was examined by electrophoretic mobility shift assay using AP-4 or mutant AP-4 probes. To verify the AP-4⅐DNA complex, nuclear extracts were incubated with anti-AP-4 antibody or excess amounts of competitor oligonucleotides (10-or 50-fold). The positions of the specific protein⅐DNA and supershifted (S.S.) complexes (arrowheads) are indicated. B, interaction between Tat and AP-4 in vivo. Upper panel, cell lysates were prepared, and immune complexes containing Tat or mTat were immunoprecipitated (IP) with anti-FLAG antibody (detecting Tat). The immunoprecipitates were separated by SDS-PAGE, followed by Western blotting (WB) with anti-AP-4 antibody. One-tenth of each protein lysate used in each reaction was loaded as the input control. Lower panels, Tat/293 and mTat/293 cells were transfected with plasmid expressing Myc-AP-4, and expression of Tat proteins was induced by PonA (10 M). The cell lysates were immunoprecipitated with anti-Myc antibody (detecting AP-4), and the immune complex was analyzed for the presence of Tat by Western blotting with anti-FLAG antibody. C, ChIP assay. Upper panel, the OGG1 promoter region amplified by the primer pairs in ChIP assay is illustrated. Arrows indicate the positions of PCR primers. Lower panels, cell lysates were prepared from Tat/293 and mTat/293 cells that were transfected with plasmid expressing Myc-AP-4 or Myc-LacZ (control) and treated with PonA (10 M) for expression of Tat and mTat. Cross-linked chromatin fragments were prepared, and the association of AP-4, LacZ, Tat, and OGG1 promoter DNA was analyzed by ChIP assay. The recovered DNA was amplified by PCR with promoter-specific primers and analyzed on a 2% agarose gel. DNAs isolated from sonicated cross-linked chromatin fragments were used as inputs. The ␤-actin promoter DNA was similarly analyzed as a control. effect of OGG1 mutation on the extent of mutation of viral and cellular genomes during chronic HIV infection, the Tat-mediated induction of OGG1 could be viewed as a regulated "feedforward" mechanism. FIG. 7. Reduction in the levels of 8-oxo-dG by Tat. A, electrochemical chromatographs of 8-oxo-dG. The HPLC patterns were traced using an ECD. The 8-oxo-dG peaks (ECD response; shaded areas in the upper panels) are indicated by arrows. The dG peaks (absorption at 290 nm; shaded areas in the lower panels) are also indicated. Cells were treated for the indicated periods of time with PonA (10 M) to express Tat or mTat. Nuclear DNA samples were prepared and measured for 8-oxo-dG levels. DNA was digested to obtain deoxynucleosides and analyzed with an HPLC-ECD system as described under "Experimental Procedures." B, levels of 8-oxo-dG in DNA expressing Tat or mTat. The amount of 8-oxo-dG is expressed as the number of 8-oxo-dG residues/10 6 dG residues. The results represent the means Ϯ S.D. from four independent experiments. *, p Ͻ 0.005; **, p Ͻ 0.001. C, OGG1 knockdown by siRNA. Tat/293 cells were transfected with 100 nM siRNAs directed against various portions of OGG1 (Nos. 1-3) or green fluorescent protein (control) mRNAs. After 72 h of transfection, cells were lysed, and OGG1 protein levels were assessed by Western blotting with anti-OGG1 antibody (upper panel). The blot was stripped and reprobed with anti-␣-tubulin antibody (lower panel). D, effects of OGG1 depletion and expression of Tat and AP-4 on the levels of 8-oxo-dG. First bar, control Tat/293 cells (no treatment); second bar, Tat/293 cells transfected with plasmid expressing AP-4; third and fourth bars, Tat/293 cells transfected with siRNA against green fluorescent protein (siRNA control) or OGG1 (No. 1), respectively, and incubated for 72 h; fifth bar, Tat/293 cells transfected with siRNA against OGG1 (No. 1) for 48 h and treated with PonA to induce Tat expression for an additional 24 h. At 72 h post-transfection, nuclear DNA samples were extracted, and the levels of 8-oxo-dG were measured similarly as described for A. The levels of 8-oxo-dG are shown as the -fold increase compared with the level of the no-treatment sample (first bar). *, p Ͻ 0.05; **, p Ͻ 0.01.