Enhanced expression of the human vacuolar H+-ATPase c subunit gene (ATP6L) in response to anticancer agents.

We have isolated two overlapping genomic clones that contain the 5'-terminal portion of the human vacuolar H(+)-ATPase c subunit (ATP6L) gene. The sequence preceding the transcription initiation site, which is GC-rich, contains four GC boxes and one Oct1-binding site, but there is no TATA box or CCAAT box. In vivo footprint analysis in human cancer cells shows that two GC boxes and the Oct1-binding site are occupied by Sp1 and Oct1, respectively. We show here that treatment with anticancer agents enhances ATP6L expression. Although cisplatin did not induce ATP6L promoter activity, it altered ATP6L mRNA stability. On the other hand, the DNA topoisomerase II inhibitor, TAS-103, strongly induced promoter activity, and this effect was completely eradicated when a mutation was introduced into the Oct1-binding site. Treatment with TAS-103 increased the levels of both Sp1/Sp3 and Oct1 in nuclear extracts. Cooperative binding of Sp1 and Oct1 to the promoter is required for promoter activation by TAS-103. Incubation of a labeled oligonucleotide probe encompassing the -73/-68 GC box and -64/-57 Oct1-binding site with a nuclear extract from drug-treated KB cells yielded higher levels of the specific DNA-protein complex than an extract of untreated cells. Thus, the two transcription factors, Sp1 and Oct1 interact, in an adaptive response to DNA damage, by up-regulating expression of the vacuolar H(+)-ATPase genes. Furthermore, combination of the vacuolar H(+)-ATPase (V-ATPase) inhibitor, bafilomycin A1, with TAS-103 enhanced apoptosis of KB cells with an associated increase in caspase-3 activity. Our data suggest that the induction of V-ATPase expression is an anti-apoptotic defense, and V-ATPase inhibitors in combination with low-dose anticancer agents may provide a new therapeutic approach.

Tumor cells possess high glycolytic activity, and rapid growth produces acidic metabolites. Moreover, tumor cells often exist in an hypoxic microenvironment lower in pH than that of surrounding normal cells. Hence, proton extrusion may be up-regulated to protect tumor cells from acidosis. Four major types of pH regulators have been identified in tumor cells as follows: sodium-proton exchangers, bicarbonate transporters, proton-lactate symporters, and proton pumps. The vacuolar H ϩ -ATPase (V-ATPase) 1 is ubiquitously expressed in eukaryotic cells (1)(2)(3)(4)(5)(6), not only in vacuolar membranes but also in plasma membrane (7)(8)(9). It is a multisubunit enzyme composed of a membrane sector and a cytosolic catalytic sector (10); it pumps protons from the cytoplasm to the lumen of the vacuole and also regulates cytosolic pH. V-ATPase is active in the plasma membrane of human tumor cells (11), and V-ATPase genes are considered "housekeeping genes." However, cytosolic pH is critical for the cytotoxicity of anticancer agents (12), and cellular acidosis is thought to be a trigger for apoptosis and to play a role in drug resistance. Therefore, understanding the mechanisms regulating tumor acidity is important for developing new approaches to cancer chemotherapy.
By using differential display, we have shown that one of the proton pump subunit genes, ATP6L (subunit c), is induced by cisplatin (13), and several V-ATPase subunit genes are upregulated in drug-resistant cell lines (13,14). Interaction of the V-ATPase c subunit with ␤ 1 integrin has been reported (15,16), and ␤ 1 integrin-mediated signaling prevents lung cancer cells from drug-induced apoptosis. The level of the V-ATPase c subunit may be critical for V-ATPase activity. In order to study transcriptional regulation of the c subunit at the molecular level, we have identified its promoter sequences and characterized the transcription factors that regulate its expression in cancer cells. We hypothesized that V-ATPase expression is up-regulated in response to cellular acidosis and show that c subunit promoter activity is activated by treatment with anticancer agents, especially the DNA topoisomerase II inhibitor, TAS-103 (17,18), which can induce cellular acidosis (19). We show also that the levels of two transcription factors, Sp1 and Oct1, increase in response to genotoxic stress and that V-ATPase inhibition strongly enhances TAS-103-induced apoptosis.

Isolation of V-ATPase Subunit c (ATP6L) Genomic Clones and DNA
Sequencing-ATP6L genomic clones were isolated from a human placental genomic library in EMBL3 by screening with cDNA. All positive phage were mapped with EcoRI and SalI. Several genomic fragments were also used as hybridization probes to confirm the overlapping regions. Two genomic DNA fragments around the first exon were subcloned into pUC18 (Fermentas AB, Lithuania) and sequenced with an Automated sequencer 377 (PE Applied Biosystems).
Primer Extension Analysis-The primer, 5Ј-GTCACATGACCT-GGGCCCCG-3Ј, derived from the first exon of ATP6L, was labeled at its 5Ј end and hybridized with poly(A) RNA from KB cells in 80% formamide, 0.4 M NaCl, 40 mM PIPES (pH 6.4), and 1 mM EDTA for 4 h at 52°C. The primer-RNA hybrid was precipitated and resuspended in reverse transcriptase mixture (Invitrogen). After 1 h of incubation at 42°C, the reaction was terminated by making the solution 20 mM in EDTA. The RNA was hydrolyzed with 0.125 M NaOH for 1 h at 65°C, the reaction neutralized, and the extended DNA then precipitated with alcohol. The DNA was analyzed on a 7 M urea, 6% polyacrylamide gel to determine the size of the extended product. Sequencing reactions using the same primer were similarly analyzed. Cell Culture and Antibody-Human epidermoid cancer KB cells (20), human prostate cancer PC3 cells (21), and human breast cancer MCF7 cells (22) were cultured in Eagle's minimal essential medium (Nissui Seiyaku Co., Tokyo, Japan) or Dulbecco's modified Eagle medium (Nissui Seiyaku Co., Tokyo, Japan) containing 10% fetal bovine serum, 0.292 mg/ml L-glutamine, 100 units/ml penicillin, and 100 g/ml kanamycin. The anti-Sp1 (catalogue number sc-420 for supershift assay, sc-59 for chromatin immunoprecipitation assay, and Western blotting), anti-Sp3 (sc-644), anti-Oct1 (sc-232), and anti-Oct2 (sc-233) antibodies were purchased from Santa Cruz Biotechnology. Antiserum to V-ATPase subunit E was generated by multiple immunization of a New Zealand White rabbit with synthetic peptides as described (23). The sequence of the synthetic peptides is ALFGANANRKFLD.
Northern Blot Analysis-Total RNA from KB cells was isolated using Sepasol reagent (Nacalai Tesque, Kyoto, Japan). RNA samples (20 g/lane) were separated on a 1% formaldehyde-agarose gel and transferred to a Hybond Nϩ filter (Amersham Biosciences) with 10ϫ SSC. Prehybridization and hybridization were performed as described (24). For analysis of stability of V-ATPase subunit transcripts by cisplatin or TAS-103, KB cells were treated with actinomycin D (1 g/ml) and cisplatin (10 M) or TAS-103 (4 M) for 6 h. Cisplatin was purchased from Sigma, and TAS-103 was kindly provided from Taiho Pharmaceutical Co., Ltd. (Tokyo, Japan).
Separation of Membrane Fractions-KB cells were treated with or without TAS-103 for 12 h. Briefly, cells were homogenized in 0.25 M sucrose, and the homogenates were centrifuged at 3,000 rpm for 10 min. The supernatant was centrifuged at 15,000 rpm for 30 min. The pellets were resuspended to 0.25 M sucrose. The resuspension was overlaid with 2.10 and 1.25 M sucrose cushions and centrifuged at 24,000 rpm for 12 h. The membrane fractions at the 0.25-1.25 M sucrose interface were collected and used for Western blotting.
Immunoprecipitation Assay-For metabolic labeling, KB cells in a 100-mm tissue culture dish were cultured in Dulbecco's methionine and cysteine-free modified Eagle's medium (Invitrogen) supplemented with 1% dialyzed fetal calf serum and were labeled with 50 Ci/ml [ 35 S]methionine and -cysteine labeling mixture (Amersham Biosciences) with or without 4 M TAS-103 for 12 h. After washing the cells twice with ice-cold phosphate-buffered saline (PBS), cells were lysed in RIPA buffer (50 mM Tris-HCl (pH 7.5), 1 mM EDTA, 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 0.5% sodium deoxycholate, 1 mM PMSF). After centrifugation at 3,000 rpm for 5 min at 4°C, 2 mg of supernatant (cellular fraction) were incubated with antiserum to V-ATPase subunit E or preimmune binding to 15 l of protein A/G-agarose. The mixtures were incubated for 12 h at 4°C and washed three times with RIPA buffer. Immunoprecipitation samples and 1% of preimmunoprecipitation samples (input) were simultaneously separated on a 15 or 10% SDS-PAGE and autoradiography.
Construction of ATP6L Promoter-Luciferase Reporter and Expression Plasmids-The EcoRI and NotI fragment (nt Ϫ1627 to nt ϩ194) of the ATP6L gene into the SmaI site of basic vector 2 (Nippon Gene, Tokyo) was designed pV-ATPase c Luc1. For the construction of deletion constructs, it was digested with PstI (pV-ATPase c Luc2), BshTI (pV-ATPase c Luc6), and NarI (pV-ATPase c Luc7). The digestion products were self-ligated. Other constructs (pV-ATPase c Luc3, -4, and -5) were constructed by PCR (Fig. 7A). The pV-ATPase c Luc3m1, Luc3m2, Luc3m3, Luc3SR, and Luc3-5bp were constructed by PCR-based method using mutated oligonucleotides (Figs. 7B and 9B). ATP6E genomic clones were isolated from a human placental genomic library, and the 5Ј-flanking region of the ATP6E (nt Ϫ715 to ϩ132) gene was subcloned in basic vector 2 (pV-ATPase E Luc1). Sp1 cDNA (encoding amino acids 30 to C-terminal) was kindly provided by Dr. Robert Tjian (University of California, Berkeley), and the XhoI fragment added start codon at the N-terminal start codon was ligated in pcDNA3 vector (Amersham Biosciences). For expression plasmids, full-length cDNA fragments of human Sp3, Oct1, and Oct2 were generated by reverse transcription-PCR using total RNA from KB cells and cloned into the pcDNA3 vector. The following oligonucleotides were used for cDNA constructions: Sp3, 5Ј- ATGGCTGCCTTGGACGTGGATAGC-3Ј and 5Ј-TTACTCCATTGTCTCATTTCCAGAAAC-3Ј; Oct1, 5Ј-ATGAACAATC-CGTCAGAAACCAGTAAACC-3Ј and 5Ј-TCACTGTGCCTTGGAGGCG-GTGGTGG-3Ј; Oct2, 5Ј-ATGGTTCACTCCAGCATGGGGGC-3Ј and  5Ј-TTACCCCGTGCTGGGGTTCAGG-3Ј. Transient Transfection-Cells were seeded into 12-well tissue culture plates at a concentration of 4 ϫ 10 4 KB cells, PC3 cells, and MCF7 cells. On the following day, cells were transfected with 0.4 g of luciferase reporter plasmid DNA using 2 l of Superfect reagent (Qiagen, Germany) according to the manufacturer's instructions. The ␤-galactosidase reporter gene (pSV-␤-gal, Nippon Gene, Tokyo) was co-transfected as an internal control. After transfection for 12 h, the cells were washed, incubated at 37°C for 12 h in fresh medium or in medium containing either TAS-103 (4 M) or cisplatin (10 M), and then harvested. For co-transfection experiments with Sp1, Sp3, Oct1, and Oct2 expression plasmids, PC3 cells were transfected with 0.2 g of luciferase reporter plasmid (pV-ATPase c Luc3) and 0.4 g of expression plasmid. After transfection for 12 h, the cells were incubated at 37°C for 24 h in fresh medium and then harvested.
Luciferase Assay-Lysed cells were assayed for luciferase activity using a Picagene kit (Toyoinki, Tokyo, Japan); the light intensity was measured for 15 s with a luminometer (Dynatech ML1500, JEOL, Japan). The ␤-galactosidase enzyme assay was performed according to the protocol of Promega.
In Vivo Footprint-KB cells and MCF7 cells were treated with dimethyl sulfate (DMS) in vivo (25). The extracted DNA was cleaved with 1 M piperidine at 90°C for 30 min. As a control guanine ladder, naked genomic DNA from KB cells was reacted with DMS in vitro and cleaved with piperidine as described above. Ligation-mediated PCR was performed as described previously (25,26). The nucleotide sequences of individual primers are as follows: 5Ј-GCCTGCAGCTTCACGCC-3Ј (Ϫ172 to Ϫ184) primer 1; 5Ј-CGCCGGGAACCCAACACCTGC-3Ј (Ϫ135 to Ϫ115) primer 2; 5Ј-CCGGGAACCCAACACCTGCAGACGACGC-3Ј (Ϫ133 to Ϫ106) primer 3 for analysis of the lower strand of the subunit c gene. Primers 1 and 2 were used for the first strand synthesis and PCR amplification, respectively. Primer 3 was labeled at the 5Ј end with [␥-32 P]ATP and used for final detection of the ladder.
Chromatin Immunoprecipitation Assay-Protein-DNA cross-linking was performed by incubating KB cells with formaldehyde at a final concentration of 1% for 10 min at room temperature. Cells were washed with PBS and collected by centrifugation at 1,200 rpm for 5 min. Cells were then lysed in buffer X (50 mM Tris-HCl (pH 8.0), 1 mM EDTA, 120 mM NaCl, 0.5% Nonidet P-40, 10% glycerol, and 1 mM PMSF) for 15 min on ice. The lysate was sonicated with 10 pulses of 10 s each at 50 -60% of maximum power with a sonicator (Taitec, Tokyo, Japan) equipped with a microtip to reduce the chromatin fragments to average sizes of less than 500 bp. Soluble chromatin was precleared by addition of 10 mg of protein A-Sepharose. An aliquot of precleared chromatin containing 1 ϫ 10 6 cells was removed and used in the subsequent PCR analysis. The remainder of the chromatin was divided, with each having 1 ϫ 10 6 cells, and diluted with buffer X. Then protein-DNA was incubated with 2 g of anti-Sp1, anti-Oct1 antibody, and normal rabbit IgG in a final volume of 800 l overnight at 4°C. Immune complexes were collected by incubation with 15 l of protein A/G-agarose for 1 h at 4°C. Protein A/G-agarose pellets were washed once with 1 ml of buffer X, once with high salt buffer X (50 mM Tris-HCl (pH 8.0), 1 mM EDTA, 500 mM NaCl, 0.5% Nonidet P-40, 10% glycerol, and 1 mM PMSF), once with LiCl buffer (10 mM Tris, 1 mM EDTA, 0.25 M LiCl, 1% Nonidet P-40, and 1% sodium deoxycholate (pH 8.0)), and twice with TE (10 mM Tris, 1 mM EDTA (pH 8.0)). Immune complexes were eluted twice with 250 l of elution buffer (0.1 M NaHCO 3 , 1% SDS). To reverse the protein-DNA cross-linking, eluted samples were incubated with 0.2 M NaCl for 4 h at 65°C. Samples were digested with proteinase K (0.04 mg/ml) for 2 h at 45°C and then with RNase A (0.02 mg/ml) for 30 min at 37°C. DNA was purified with phenol/chloroform followed by ethanol precipitation. Purified DNA was resuspended in 20 l of H 2 O. Aliquots of 1 l of serial dilution were analyzed by PCR with the appropriate primer pairs. The V-ATPase c promoter primers are as follows: 5Ј-CTGCAGACGACGCG-CAGCCGCAGAGGAGGC-3Ј and 5Ј-GCGCGAGACCGGTCCAACGCT-GCGGAGATC-3Ј, and the YB-1 promoter primers are 5Ј-AGATCTCT-ATCACGTGGCTGTTGC-3Ј and 5Ј-AAGCTTATCAGTCCTCCATTCTC-ATTGG-3Ј. Amplification was performed for a pre-determined optimal number of cycles. PCR products were separated by electrophoresis on 2% agarose gel, which were stained with ethidium bromide.
Preparation of Nuclear Extracts-Nuclear extracts using buffer S of KB cells were prepared as described (27). Briefly, 2 ϫ 10 7 cells were collected with PBS, resuspended in 1 ml of ice-cold 10 mM HEPES-KOH (pH 7.9), 1.5 mM MgCl 2 , 10 mM KCl, 0.2 mM PMSF, 0.5 mM DTT, and incubated on ice for 15 min. The cells were lysed with a dropping of 0.6% Nonidet P-40, and the lysate was centrifuged at 3,000 rpm for 10 min. The resulting nuclear pellets were resuspended in 50 l of ice-cold buffer S (20 mM HEPES-KOH (pH 7.9), 25% glycerol, 1.5 mM MgCl 2 , 10 mM KCl, 0.2 mM EDTA, 0.2 mM PMSF, and 0.5 mM DTT), and KCl was added to a final concentration of 0.4 M and incubated for 15 min on ice with frequent gentle mixing. Following centrifugation for 5 min at 4°C in a microcentrifuge to remove insoluble material, the supernatant (nuclear extract) was stored at Ϫ70°C. Nuclear extracts using buffer C were also prepared as described (28). Briefly, 2 ϫ 10 7 cells were collected with PBS, resuspended in 1 ml of ice-cold 10 mM HEPES-KOH (pH 7.9), 10 mM KCl, 0.5 mM PMSF, 1 mM DTT, 0.1 mM EDTA, 0.1 mM EGTA, and incubated on ice for 15 min. The cells were lysed with a dropping of 0.6% Nonidet P-40, and the lysate was centrifuged at 3,000 rpm for 10 min. The resulting nuclear pellets were resuspended in 50 l of ice-cold buffer C (20 mM HEPES-KOH (pH 7.9), 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, and 1 mM DTT) and incubated for 15 min on ice with frequent gentle mixing. Following centrifugation for 5 min at 4°C in a microcentrifuge to remove insoluble material, the supernatant was stored at Ϫ70°C. Its protein concentration was determined by the method of Bradford.
Western Blotting-Preparation of nuclear extracts and separation of membrane fractions were described above. Nuclear extracts (100 g of protein) of KB cells prepared with buffer C were separated on a 10% SDS-PAGE, and membrane fractions (40 g) by sucrose gradient of KB cells were separated on a 15% SDS-PAGE gel and transferred to polyvinylidene difluoride membrane (Millipore) using a semidry blotter. Prestained protein marker (Nacalai Tesque, Kyoto, Japan) was used as a molecular weight standard. Immunoblot analysis was performed with an appropriate dilution of each antibody.
DNA Fragmentation-KB cells in a 100-mm tissue culture dish were treated with TAS-103, bafilomycin A1 (Wako, Ohsaka, Japan), and a combination of these drugs for 36 h. The cells were washed twice with ice-cold PBS and then collected by centrifugation at 1,500 rpm for 10 min. The cell pellets were resuspended in 500 l of Tris-EDTA buffer (20 mM Tris-HCl (pH 8.0), 20 mM EDTA) containing 0.1% SDS and proteinase K (0.5 mg/ml) at 50°C for 2 h and then with RNase A (0.02 mg/ml) for 30 min at 37°C. DNA was purified with phenol/chloroform followed by ethanol precipitation. Purified DNA was resuspended in 100 l of H 2 O. DNA samples were separated by electrophoresis on 2% agarose gel, which were stained with ethidium bromide (19).
Measurement of Caspase-3 Activity-Cells were seeded into 12-well tissue culture plates at a concentration of 4 ϫ 10 4 KB cells and treated with TAS-103, bafilomycin A1, and a combination of these drugs for 36 h. The cells were removed by trypsinization and resuspended in hypotonic cell lysis buffer (25 mM HEPES-KOH (pH 7.5), 5 mM MgCl 2 , 5 mM EDTA, 5 mM DTT, 2 mM PMSF, 10 g/ml pepstatin A, and 10 g/ml leupeptin). Following centrifugation for 20 min at 4°C in a microcentrifuge, the supernatant fractions were collected. The fluorescence (CPP32 activity) of each sample was analyzed with the Fluorometric CaspACE TM Assay System (Promega, Madison, WI) according to the manufacturer's instructions. For caspase activity with CaspACE TM FITC-VAD-FMK in situ marker (Promega, Madison, WI), KB cells were treated with TAS-103, bafilomycin A1, and a combination of these drugs for 48 h. Cells were stained with CaspACE TM FITC-VAD-FMK in situ marker according the manufacturer's instructions and analyzed using fluorescence microscopy. For statistical analysis of each experiment, 4 fields (ϫ400) were counted per stimulation and cell type (between 300 and 400 cells in total).

RESULTS
To isolate genomic clones encoding the 5Ј region of the ATP6L gene, a human genomic library was screened with a previously isolated ATP6L cDNA clone (13). Two clones containing non-identical inserts were characterized. The restriction map of these clones is shown in Fig. 1A, and sequence analysis confirmed that they encode ATP6L. To localize the first exon more accurately, the promoter proximal plasmid was digested with restriction enzymes and analyzed by Southern blotting using cDNA. In order to determine the nucleotide sequence of the promoter region, a 1.9-kb EcoRI-SalI fragment of EMBL3 was subcloned into pUC18 (Fig. 1A), and the nucleotide sequence of the first exon and its 5Ј-flanking region were determined (Fig. 1B). This fragment contained exons with sequences identical to the 5Ј portion previously determined from cDNA.
To define precisely the transcription initiation site, we performed primer extension. The cDNA products extended from the primer were analyzed by electrophoresis and sequenced using the same primer. Two major transcription initiation sites were observed (Fig. 2). About 20% of the transcripts initiated at ϩ1 and 80% initiated at ϩ75. The transcription initiation site of the human gene is located 236 bp upstream from that of the mouse gene (29). An additional 100-bp sequence of the 5Јuntranslated region has been published. This indicates that the transcription initiation sites of the human gene are completely different from those of the mouse gene. The nucleotide sequences of the factor binding sites are also not the same, suggesting that regulation of the human gene differs from that of the mouse.
Analysis of the region upstream of the first exon failed to locate any sequence motifs such as TATA and CCAAT boxes. There were four GC boxes and one Oct1-binding site in the proximal promoter region, with one GC box on its own in the untranslated region. The GC content around the first exon was about 70%. To determine whether the region upstream of the first exon possesses promoter activity, the available restriction sites were utilized to construct a series of deletion reporter constructs. These constructs were tested by transient transfection in human cancer cell lines KB, PC3, and MCF7 cells. DNA extending only as far as Ϫ113 yielded full promoter activity, whereas the region between ϩ57 and ϩ194 retained 50% of maximum activity (data not shown).
We reported previously that V-ATPase gene expression is induced by cisplatin treatment (13) and examined further upregulation of the V-ATPase genes by anticancer agents. As shown in Fig. 3A, the steady-state mRNA levels of two V-ATPase genes, ATP6L and -6E, increased 3-5-fold when cells were treated with cisplatin and TAS-103. To determine whether c and E subunit protein levels also increased when cells were treated with TAS-103, we performed Western blotting and immunoprecipitation using antibody against the E subunit. As expected, TAS-103 increased the expression of the E subunit (Fig. 3B). Immunoprecipitation assay showed that levels of the c (16 kDa), cЉ (19 kDa), and D subunits (34 kDa) also increased (Fig. 3C). We also analyzed the levels of the higher molecular weight V-ATPase subunits in 10% SDS-PAGE. The levels of subunit a (100 -116 kDa), subunit A (70 -73 kDa), and subunit C (40 -45 kDa) were increased after TAS-103 treatment (Fig. 3D). These results indicate that TAS-103 may stimulate the expression of the V-ATPase complex as a whole. Next, we investigated whether the up-regulation of V-ATPase gene expression is because of transcriptional activation, and we found that TAS-103 could activate the promoter of the ATP6L and -6E genes but cisplatin could not (Fig. 4A). On the other hand, cisplatin and TAS-103 both reduced the rate of degradation of ATP6L and -6E mRNAs (Fig. 4B). These find-ings suggest that drug treatments can promote expression of the V-ATPase by both transcriptional and post-transcriptional mechanisms.
Many potential GC boxes or related motifs are found among the GC-rich stretches in the promoter region. In order to confirm the existence of functional GC boxes and other transcription factor binding sites, we performed an in vivo footprint experiment as shown in Fig. 5. Because this experiment was performed using primers for the lower strand, transcription factor bindings to the complementary strand of the nucleotide sequence shown in Fig. 5 were detected. Protection of four 5Ј-guanines and hypersensitivity of the 3Ј-guanine of the consensus 5Ј-GGGCGG-3Ј, which are typical Sp1 guanine binding signals, were observed at two potential GC box sequences (Ϫ73 to Ϫ68 and Ϫ36 to Ϫ31). In addition, there was slight protection of guanines at Ϫ84 and Ϫ61 on the lower strand surrounding the distal GC box (Ϫ73 to Ϫ68). Although the proximal GC box (Ϫ36 to Ϫ31) overlapped with an additional GC box (Ϫ41 to Ϫ36), a typical profile of Sp1 binding was not detected on this upstream GC box (Ϫ41 to Ϫ36). No signs of any binding to other tentative GC boxes located downstream of the transcription start site and GC box-like motif (Ϫ91 to Ϫ86) were seen in experiments with upper strand primers (data not shown).
In order to show that Sp1 (Ϫ73 to Ϫ68) and Oct1 (Ϫ64 to Ϫ57) bound specifically to the V-ATPase c promoter in vivo, we utilized the chromatin immunoprecipitation assay as shown in Fig. 6. PCR amplification of the V-ATPase c promoter was carried out with DNA extracted from the immunocomplex. Fig.  6A shows that significant levels of the V-ATPase c promoter sequence were detected as a 120-bp PCR product in the complexes immunoprecipitated with anti-Sp1 and anti-Oct1 antibody. The YB-1 promoter sequence was not detected, because there are no Sp1-and Oct1-binding sites in the YB-1 promoter, as shown in Fig. 6B. Furthermore, the V-ATPase c promoter sequence was not observed when normal rabbit IgG was used. and N, NarI. Overlapping regions within these clones were confirmed by Southern blotting. The EcoRI-SalI fragment (1.9 kb) contains the first exon and 5Ј-flanking sequence. B, nucleotide sequence of the 5Ј upstream region and the first exon of the ATP6L gene. Nucleotides are numbered relative to the transcription initiation site, determined by primer extension. The first exon is boxed, and the intron sequence is in lowercase. The sequences that serve as recognition sites for Sp1 (5Ј-GGGCGG-3Ј) and Oct1 (5Ј-ATGCAAAT-3Ј) are underlined and double underlined, respectively. The ATG initiation site is in boldface. The arrow indicates the position of the primer used for primer extension. We next examined the effect of TAS-103 on luciferase activity induced by a series of 5Ј-deleted promoter constructs assayed 12 h after transfection (Fig. 7A). The luciferase activity of Luc1-3 was increased by about 3-6-fold compared with the control by 4 M TAS-103. The activity of Luc4 was, however, not increased. These results suggest that an element responsible for V-ATPase c promoter activation by TAS-103 is located between Ϫ77 and Ϫ58. This region contains the GC box and the Oct1-binding site. To examine whether mutations of either the GC box or the Oct1-binding site affect the stimulation of Vfractions were loaded on a 15% SDS-PAGE gel and transferred to a polyvinylidene difluoride membrane. Immunoblot analysis was performed with antiserum to the V-ATPase E subunit. C and D, immunoprecipitation with antiserum to the V-ATPase E subunit. Antisera to the V-ATPase E subunit or preimmune binding protein A/G-agarose were incubated for 12 h at 4°C with 35 S-labeled protein from KB cells with or without exposure to TAS-103 (4 M). The mixtures were washed three times and separated on a 15% (C) or 10% SDS-PAGE gel (D). Molecular mass markers are indicated, as well as the positions of proteins that probably correspond to the low molecular weight V-ATPase subunits D and E, cЉ, c (C), and the high molecular weight subunits a, A, C, E (D).

. Transcriptional activity and mRNA stability of the ATP6L and -6E genes in KB cells treated with anticancer agents.
A, transcriptional activity of the ATP6L and -6E genes in response to anticancer agents. The 5Ј-flanking regions of the ATP6L (nt Ϫ1627 to nt ϩ194) and the -6E (nt Ϫ715 to nt ϩ132) genes were subcloned in basic vector 2 (B2) containing a luciferase reporter. The resulting constructs were transiently transfected into KB cells, together with a ␤-galactosidase reporter as an internal control. After transfection for 12 h, the cells were incubated for 12 h in fresh medium or in medium containing either TAS-103 or cisplatin. Luciferase activity was measured as described under "Materials and Methods." Error bars indicate S.D. B, stability of ATP6L and -6E gene transcripts in KB cells treated with anticancer agents. KB cells were incubated with actinomycin D (1 g/ml) and cisplatin (10 M) or TAS-103 (4 M) for 6 h, and the steadystate levels of ATP6L and 6E mRNAs were assayed by Northern blotting. 20 g of total RNA was loaded per lane. ATPase c promoter activity by TAS-103, we made three constructs with mutations in the promoter sequences of these binding elements (Fig. 7B). TAS-103 responsiveness was reduced when a mutation was introduced into the GC box, and mutation of the Oct1-binding sequence completely inhibited TAS-103-induced luciferase activity, suggesting that the Oct1binding site has a role in the promotion of V-ATPase c expression by TAS-103. The Luc3m1, Luc3m2, and Luc3m3 mutations had no apparent effect on basal transcriptional activity in the absence of TAS-103 (Fig. 7B).
To investigate whether V-ATPase c gene transcription activated by TAS-103 was affected by an interplay between Sp1 and Oct1, we performed EMSA on KB cells following TAS-103 treatment to study the interaction between the promoter and transcription factors using nuclear extracts made in buffer S. Six probes were utilized covering the entire core promoter, as described under "Materials and Methods." When Oligos 1, 5, and 6 were used as probes, we could not detect any retarded band (data not shown). However, a retarded band was observed when Oligos 2-4 were used, and its strength was increased by treatment with TAS-103 (Fig. 8A). GC boxes were present in each of these three oligonucleotide probes, and the intensity of the retarded band was reduced by addition of unlabeled GC box DNA (data not shown). Furthermore, the retarded bands were supershifted by antibody to Sp1 but not by antibody to Sp3 or preimmune antibody (Fig. 8B).
We noted that the band retarded by Oligo3 probe was not completely shifted by the addition of an excess of Sp1 antibody (data not shown). Oligo3 contains a GC box and an Oct1binding site. We used nuclear extracts made with two different buffers, as shown under "Materials and Methods," to examine the specificity of the DNA-protein interaction by appropriate competition assays. The band retarded by the nuclear extract prepared in buffer S was also only completely eliminated by a 25-fold excess of unlabeled Oligo3 but not by Oligo3m1, Oligo3m2, and Oligo3m3. On the other hand, the band retarded by the nuclear extract prepared in buffer C was almost completely eliminated by a 25-fold excess of unlabeled Oligo3m1 as well as Oligo3 but not by Oligo3m2 and Oligo3m3 (Fig. 8C). These results suggest that only Oct1 binds to the Oligo3 probe when the nuclear extract is prepared in buffer C but that either Sp1 or Oct1 bind to Oligo3 in a mutually exclusive manner when the nuclear extract is prepared in buffer S. We performed supershift assays to confirm these results (Fig. 8D). The retarded band was completely shifted by the addition of the Oct1-specific antibody when the nuclear extract was prepared with buffer C, suggesting that only Oct1 can bind to the Oligo3 probe under these conditions. Furthermore, the retarded band was shifted by the addition of either Sp1-or Oct1-specific antibody when the nuclear extract was prepared with buffer S. This indicates that both Sp1 and Oct1 can bind to the Oligo3 but that Sp1 and Oct1 cannot bind simultaneously to the same oligonucleotide. However, the in vivo footprint clearly showed that both Sp1 and Oct1 could bind simultaneously in vivo.
We next analyzed the DNA-protein complex in more detail after prolonged electrophoresis. Two complexes (C3 and C6) were formed with Oligo3 (Fig. 9A). Both can be supershifted with either anti-Sp1 antibody (C2 and C5) or Oct1 antibody (C1 and C4). This indicates that the slower migrating complex C3 is the complex formed with Sp1 and Oct1 simultaneously, and the faster migrating complex C6 appears to involve either Sp1 or Oct1. Similar results were obtained with the artificial oligonucleotides Oligo3SR and Oligo3-5bp. Both Sp1 and Oct1 binding and promoter activation were slightly enhanced when the phase of the Sp1-binding site was reversed. On the other hand, promoter activity was reduced when 5 bp were inserted between the Sp1-and the Oct1-binding sites (Fig. 9B).
Because both Sp1 and Oct1 binding was increased in nuclear extracts prepared from drug-treated cells, we investigated whether these transcription factors participated in the cellular response to TAS-103. As shown in Fig. 10A, the cellular levels of both Sp1/Sp3 and Oct1 increased substantially in a time-dependent manner in KB cells exposed to 4 M TAS-103. To investigate further the role of Sp1 and Oct1 in V-ATPase c promoter activity, the cDNA expression plasmids Sp1, Sp3, Oct1, and Oct2 were co-transfected into PC3 cells together with the Luc3 plasmid and promoter activity assayed 24 h later (Fig.  10B). The results are expressed relative to the activity observed after co-transfection of empty plasmid with the Luc3 reporter. As shown in Fig. 10B, co-transfection of the Oct1 expression plasmid increased Luc3 promoter activity ϳ1.7-fold, whereas co-transfection of the Sp1, Sp3, and Oct2 expression plasmids had no effect. These data suggest that an increase in Oct1 may be critical for the TAS-103-induced up-regulation of V-ATPase c subunit.
In order to better understand the mechanism underlying TAS-103-induced up-regulation of V-ATPase, we tested whether treatment with a V-ATPase inhibitor together with TAS-103 would increase apoptosis. KB cells were treated with a low dose of TAS-103 in the presence or absence of bafilomycin A1 (30,31), and apoptosis was assessed by DNA fragmentation (Fig. 11A), an increase of caspase-3 activity (Fig. 11B), and in situ labeling of activated caspase (Fig. 11C). Apoptosis was significantly stimulated by the combined treatment, suggesting that V-ATPase functions as an anti-apoptotic factor. DISCUSSION We have described the cloning and characterization of the human vacuolar H ϩ -ATPase subunit c (ATP6L) gene, and we have isolated overlapping genomic clones encompassing 10 kb of its 5Ј sequence and 14 kb of its 5Ј-flanking region (Fig. 1A). We determined the nucleotide sequence surrounding the 5Ј end of the gene that contains the initiation sites for transcription. The regulatory regions are highly GC-rich, and the CpG islands are located 5Ј to the first exon. Thus, the ATP6L promoter has structural features common to housekeeping genes. No typical TATA and CCAAT boxes exist in the region preceding the first exon (Fig. 1B), and this may account for the presence of two major transcription initiation sites (Fig. 2). Multiple GC boxes are found in the promoter and first exon. GC boxes are frequent DNA elements present in many promoters and are required for appropriate expression of many ubiquitous genes. This feature of the ATP6L 5Ј region is consistent with its ubiquitous expression in human tissues and suggests that it encodes a protein with an essential cellular function.
Recently, the nucleotide sequence of the mouse V-ATPase c subunit promoter has been reported (29). Although there is significant homology between the proximal promoter sequence of human and mouse, the GC boxes are not conserved in the mouse gene. Furthermore, it is noteworthy that the human transcription initiation site is completely different from that of the mouse. The human ATP6L cDNA sequence has been published, and based on a sequence alignment of human cDNA, human cDNA has an additional 100 bp compared with the mouse initiation site. This indicates that the human and the mouse ATP6L genes are differently regulated.
As shown in Fig. 3A, mRNA levels of two V-ATPase genes were significantly (3-5-fold) up-regulated when cells were treated with TAS-103. Quantitation by PhosphorImager indicates that the protein levels of V-ATPase subunits c and E increased only 1.5-2-fold compared with a dramatic increase of mRNA. This discrepancy is probably due to the translational or post-translational control of V-ATPase protein. Functional analysis of the promoter region in a transient expression system demonstrated significant promoter activity in human cancer cells, and this activity increased 3-6-fold in TAS-103treated cells (Fig. 4A). Furthermore, the region between Ϫ77 and Ϫ58 was required for the transcriptional up-regulation (Fig. 7A). Protein DNA interaction on the proximal promoter region was investigated by in vivo DMS footprint experiments. We detected clear evidence of Sp1 binding to two GC boxes. As FIG. 7. Functional analysis and deletion mapping of the human ATP6L promoter in KB cells. A, schematic representation of the ATP6L-luciferase reporters and the relative promoter activity of the ATP6L gene. Deletion constructs of the 5Ј-flanking region of the ATP6L gene were subcloned into basic vector 2 upstream of the luciferase reporter gene. All reporter constructs were transiently transfected into KB cells, together with a ␤-galactosidase reporter as an internal control. After transfection for 12 h, the cells were incubated for 12 h in fresh medium or in medium containing TAS-103 (4 M). The sequences that serve as the recognition sites for Sp1 are shown as black boxes, and those for Oct1 as the white box. Error bars indicate S.D. B, transcriptional activity of luciferase reporters with mutant Sp1-or Oct1-binding sites. Three constructs with mutations introduced into the Sp1-binding site (Luc3m1), the Oct1-binding site (Luc3m2), and both the Sp1 and Oct1 sites (Luc3m3) were subcloned into basic vector 2 upstream of the luciferase reporter gene. The wild-type sequence of the Sp1 and Oct1 sites is 5Ј-CCGCCCCGTATGCTAAT-3Ј (Luc3). The mutant sequences are as follows: 5Ј-CCTTCCCGTATGCTAAT-3Ј (Luc3m1), 5Ј-CCGC-CCCGTATGCTTTT-3Ј (Luc3m2), and 5Ј-CCTTCCCGTATGCTTTT-3Ј (Luc3m3). Luciferase assays were carried out as described.
shown in Fig. 5, the G residue in the octamer sequence was protected in the footprint. Also, the V-ATPase c promoter sequence that contains both Sp1-and Oct1-binding sites was recovered in complexes immunoprecipitated with either anti- Reporter plasmids with reversed GC box (Luc3SR), or a 5-bp insertion between the Sp1-and Oct1-binding site (Luc3-5bp), were constructed. The mutated sequences are 5Ј-GGCGGGCGTATGCTAAT-3Ј (Luc3SR) and 5Ј-CCGCCCCGTctagaAT-GCTTTT-3Ј (Luc3-5bp). Luciferase assay was carried out as described under "Materials and Methods." oligonucleotides as described under "Materials and Methods." The arrow indicates the principal retarded band, and F denotes free probe. B, analysis of GC box-binding proteins by supershift assay. Nuclear extracts from TAS-103 (4 M)-treated KB cells, made using buffer S, were incubated with probes and 2 l of anti-Sp1 or anti-Sp3 antibody (Ab) for 30 min at 4°C. The position of the supershifted band is indicated by the arrow. F indicates the free probe. C, ability of mutated oligonucleotides to compete for Oligo3 binding. Oligonucleotides with mutations in the GC box (Oligo3m1), in the Oct1-binding site (Oligo3m2), and in both sites (Oligo3m3) of Oligo3 were prepared as described under "Materials and Methods." 5-, 10-, and 25-fold molar excesses of unlabeled oligonucleotide were preincubated with buffer C or buffer S nuclear extracts from KB cells with or without TAS-103 (4 M) treatment, and labeled Oligo3 was then added. F, free probes. D, analysis of GC box and specific octamer sequence binding proteins by supershift assay. Nuclear extracts of KB cells treated with TAS-103 (4 M), made using buffer C or buffer S, were incubated with probe and 2 g of anti-Sp1, anti-Sp3, anti-Oct1, or anti-Oct2 antibody for 30 min at 4°C. The positions of the supershifted bands are indicated by the arrows. F is the free probe. Sp1 or anti-Oct1 antibodies in chromatin immunoprecipitation assays (Fig. 6).
In EMSA, specific promoter-DNA binding was stimulated when the cells were treated with TAS-103. When Oligo3 was used as probe, the transcription factors in the protein-DNA complexes were identified as Sp1 and Oct1. Another noteworthy observation was that Sp1 binding was observed in nuclear extracts prepared with buffer S but not with buffer C. Evidently, differences in the way the nuclear extract is prepared can significantly affect the profile of DNA-protein interaction when binding sites for two factors exist in the same probe (Fig.  8, C and D). Further study is necessary to determine the exact difference between the nuclear extracts prepared with the two different buffers.
Oct1, a member of the POU homeodomain family, is ubiquitously expressed and plays a role in activating transcription of various genes (32,33). Functional cooperation between Sp1 and Oct1 has been reported in the regulation of the human U2 small nuclear RNA promoter (34). Furthermore, the two transcription factors interact in vivo in the yeast two-hybrid system and regulate human U2 small nuclear RNA genes (35), indicating that both transcription factors are necessary for basal transcriptional activity of this gene. Recently, Oct1 has been shown to be induced after cells are treated with DNA-damaging agents and anticancer agents, including UV irradiation, cisplatin, etoposide, and camptothecin (36,37). We have confirmed that both camptothecin and etoposide induce ATP6L promoter activity (data not shown). However, we could not detect promoter activation when cells were treated with cisplatin (Fig.  4A). To our knowledge, the present study is the first to demonstrate activation of Oct1 target genes after treatment with anticancer agents.
V-ATPase subunit genes are inducible by treatment of human cancer cells with cisplatin and are up-regulated in cisplatin-resistant cell lines (13). Transient transfection of a reporter plasmid showed that promoter activity is not activated by cisplatin treatment (Fig. 4A) and is not enhanced in resistant cell lines (data not shown). One possible explanation is that posttranscriptional mechanisms, such as mRNA stabilization, may be involved in the cisplatin induction and up-regulation of this gene in drug-resistant cells. Another possibility is that the pathway signaling DNA damage to transcription factors may differ between cisplatin and other anticancer agents. Because cisplatin can block degradation of the mRNA (Fig. 4B), certain pathways signaling DNA damage in human cancer cells may increase mRNA stability.
Our data indicate that drug-induced gene expression is regulated by both transcriptional and post-transcriptional mech- The cells were harvested at 3, 6, and 12 h after treatment, and nuclear extracts were prepared. 100 g of nuclear extracts were loaded on a 10% SDS-PAGE gel and transferred to a polyvinylidene difluoride membrane. Immunoblot analysis was performed with an appropriate dilution of anti-Sp1, anti-Sp3, anti-Oct1, or anti-Oct2 antibodies. Control indicates Coomassie Brilliant Bluestained protein. B, luciferase assay with expression plasmids. Luciferase reporter plasmids (pV-ATPase c Luc3) were transiently transfected into KB cells, together with expression plasmids for Sp1, Sp3, Oct1, and Oct2. After transfection for 12 h, the cells were incubated for 24 h in fresh medium at 37°C. Luciferase activity was measured as described under "Materials and Methods." Error bars indicate S.D. anisms. We have demonstrated induction of Sp1/Sp3 and Oct1 in response to anticancer agents (Fig. 10A). In contrast to the untreated cells, induction of both Sp1 and Sp3 protein was observed to increase 5-10-fold. The level of Oct1 protein was observed to increase 3-5-fold. Sp3 might act as repressor to inhibit Sp1-dependent transcription, suggesting that Sp1-dependent transcription of ATP6L gene is reduced by the Sp3 induced by TAS-103. Thus, induction of ATP6L mRNA might be substantially affected by the induction of Oct1. The results of both Northern blot analysis and reporter assays were consistent with those of the Oct1 induction. Both co-transfection experiments and reporter assays with mutated promoters confirm that Oct1 is the main factor involved in the induction of promoter activity by TAS-103. Because both Sp1 and Oct1 are ubiquitously expressed, these housekeeping transcription factors may cooperate with other transcription factors, including basal transcription factors and cofactors, to protect cells from genotoxic stress. We have shown that Sp1 binds to the two GC boxes located in the proximal promoter region of ATP6L. Sp1 is not the only a protein acting through GC boxes; the existence of a small protein family consisting of Sp1, Sp2, Sp3, and Sp4 has been reported (38 -41). Hence the absolute levels of Sp1 and Sp3 or their nuclear ratio may be responsible for differences in promoter binding and target genes expression. The transcriptional cofactors required for Sp1 activity have been identified (42), further emphasizing that these cofactors may be involved in the activation of Sp1 target genes in response to anticancer agents. The exact mechanisms by which Sp1 and Oct1 act on the expression of target genes to promote tumorigenesis and drug resistance remain unclear.
Expression of V-ATPase could have significance for cell growth (43), cell motility (44), tumorigenesis (45), metastasis (46), and apoptosis (47). The reason for induction of V-ATPase by anticancer agents is unclear, but increased V-ATPase activity may represent a cellular anti-apoptotic response. Our results show that TAS-103 can induce apoptosis, especially in the presence of bafilomycin A1 (Fig. 11), suggesting that V-ATPase inhibits apoptosis of cancer cells by preventing cellular acidosis. It will be of considerable interest to identify the transcription factors that regulate the expression of other V-ATPase subunit genes, and studies along these lines are in progress.