SV40 large T antigen transactivates the human cdc2 promoter by inducing a CCAAT box binding factor.

Cyclin-dependent protein kinases (Cdks) play a key role in the cell division cycle of eukaryotic cells. Cdc2, the first mammalian Cdk that was discovered, is expressed in S phase and functions in the G2 to M phase transition. By transfecting segments of the human cdc2 promoter linked to a reporter gene into monkey kidney (CV-1) cells, we identified the region containing the Sp1, E2F, and two CCAAT box binding sites as essential and sufficient for basal transcription. SV40 large T antigen (SV40-LT) is a viral oncoprotein that transactivates viral and cellular promoters and induces DNA synthesis in quiescent cells. SV40-LT transactivated wild-type cdc2 promoter/reporter constructs in a dose-dependent manner, coinciding with an increase in endogenous cdc2 mRNA. A mutant promoter from which the two CCAAT box motifs were deleted was 8-fold less sensitive to SV40-LT. Activation by SV40-LT did not require its ability to bind the retinoblastoma or p53 tumor suppressor proteins. SV40-LT induced a specific CCAAT box-binding factor (CBF) in CV-1 and COS-7 cells, as judged by gel shift and Southwestern analyses. Similar results were obtained in human fibroblasts expressing a conditional SV40-LT. The SV40-LT-inducible CBF appears to be novel and differs from the CBF that activates heat shock protein 70 gene expression.

The cyclin-dependent protein kinases (Cdks) 1 play an important role in the control of cell division in eukaryotic cells. The first Cdks identified were the products of the cdc2 gene in the fission yeast Schizosacchomyces pombe and the CDC28 gene of Saccharomyces cerevisiae. The mammalian homologue of cdc2 and CDC28 encodes a catalytic subunit, p34 cdc2 , which is now known to be an important regulator of cell cycle progression (Nurse and Bissett, 1981;Beach et al., 1982;Reed et al., 1985;Simanis and Nurse, 1986;Draetta et al., 1987;Lee and Nurse, 1987) (for reviews, see Nurse (1990), Draetta (1990), and Pines and Hunter (1990)). p34 cdc2 , like other Cdks, requires association with a cyclin regulatory subunit for kinase activity (reviewed by Morgan (1994) and Pines (1994)). p34 cdc2 is primarily required for progression from G 2 to M (Riabowol et al., 1989;Fang and Newport, 1991;Murray, 1992), whereas other mammalian Cdks (e.g. p33 cdk2 ) have been implicated in the initiation of DNA replication (Fang and Newport, 1991;Tsai et al., 1991;Pagano et al., 1993) (for reviews, see Morgan (1994), Pines (1994), and references therein).
Expression of the mammalian cdc2 mRNA, as well as p34 cdc2 protein, depends on the cellular growth state, declining significantly when cells undergo growth arrest, differentiation, or development (Draetta et al., 1987;Lee et al., 1988;Krek and Nigg, 1989;D'Urso et al., 1990;McGowan et al., 1990;Wang et al., 1991;Dalton, 1992), and increasing when quiescent cells enter the cell cycle. The human cdc2 promoter contains consensus binding sites for several cellular transcription factors, including ATF, c-Myb, Sp1, E2F, and CCAAT box binding factor (Dalton, 1992;Ku et al., 1993). The product of the c-myb protooncogene (Ku et al., 1993) and the Rb tumor suppressor protein (Dalton, 1992;Yamamoto et al., 1994) have been shown to regulate the cdc2 promoter through c-myb and E2F binding elements, respectively. Some members of the E2F transcription factor family interact with the Rb protein, whereas other members interact with the Rb-related protein p107 and p130 (Bagchi et al., 1991;Chittenden et al., 1991;Bandara and La Thangue, 1991;Chellappan et al., 1992;Shirodkar et al., 1992;Lees et al., 1993;Beijersbergen et al., 1994;Ginsberg et al., 1994;Hijmans et al., 1995;Sardet et al., 1995). Since E2F is important for the transcription of several cell cycle-regulated genes such as c-myc (Hiebert et al., 1989;Thalmeier et al., 1989) and genes expressed at the G 1 /S boundary (De Gregori et al.1995), the Rb-E2F interaction is an important negative regulator of the expression of these genes. The interaction of Rb or the related proteins, p107 and p130, with E2F is disrupted by several viral oncoproteins. These include SV40-LT, adenovirus E1A, and papilloma virus E7 (Whyte et al., 1988;De Caprio et al., 1988;Dyson et al., 1989;Chellappan et al., 1992). This is consistent with the finding that the Rb-mediated repression of cdc2 promoter can be reversed by coexpression of adenovirus E1A protein (Dalton, 1992).
SV40-LT induces DNA synthesis in human fibroblasts that are quiescent and/or senescent (Tsuji et al., 1983;Gorman and Cristofalo, 1985;Wright et al., 1989;Shay et al., 1991;Sakamoto et al., 1993;Dimri and Campisi, 1994), and express low or undetectable levels of cdc2 mRNA and p34 cdc2 protein (Richter et al., 1991;Stein et al., 1991). These findings suggest that SV40-LT might activate, directly or indirectly, one or more * This research was supported in part by National Institutes of Health Grants CA33099 (to R. P.) and AG09909 and AG11658 (to J. C.), and by National Science Foundation Grant MCB-9507034 (to R. P.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. § Supported in part by a postdoctoral fellowship from the Marrion Merrell Dow Foundation.
growth regulatory genes such as cdc2. Here, we examine whether SV40-LT activates the human cdc2 promoter and, if so, whether the E2F binding sites are required for this activation. We show that, for basal activity of the human cdc2 promoter in monkey kidney (CV-1) cells, the region containing the Sp1, E2F, and two CCAAT box binding elements (between Ϫ171 and Ϫ28) are essential and sufficient. This activity was transactivated severalfold by SV40-LT, and only the two CCAAT box binding motifs were required for optimal transactivation. In addition, we show that the ability of SV40-LT to bind and inactivate the retinoblastoma (Rb) or p53 tumor suppressor proteins is not necessary for this transactivation. Finally, we show that SV40-LT induces a novel CCAAT box binding factor in cycling CV-1 cells and SV40-transformed COS-7 cells, and in human fibroblasts that are reversibly senescent due to a conditional SV40-LT.

EXPERIMENTAL PROCEDURES
Materials-All enzymes involved in the cloning procedures were purchased from New England Biolabs, Inc. Dulbecco's modified Eagle's medium, Trizol, and Lipofectin were obtained from Life Technologies, Inc. Defined/supplemented bovine calf serum and fetal bovine serum were purchased from Hyclone Laboratories (Logan, UT) or Sigma. Antibiotics, dexamethasone, CsCl, ethidium bromide, n-butyryl-coenzyme A, and o-nitrophenyl-␤-D-galactopyranoside were purchased from Sigma. Bradford reagent for protein estimation was from Bio-Rad. Baker Si500F TLC plate-silica gel was from J.T. Baker, Inc. [ 14 C]Chloramphenicol (49 mCi/mmol) was from Moravek Biochemicals, Inc.
Construction of cdc2 Promoter Deletions-Methods for plasmid constructs were carried out according to standard procedures (Sambrook et al., 1989). To construct cdc2 promoter deletions, cdc2-PstI, and cdc2-7 clones (kindly provided by Dr. Bruno Calabretta; Ku et al., 1993) were used. The 431-bp TaqI fragment of cdc2-7 containing the promoter sequence from nucleotide Ϫ359 to nucleotide ϩ64 was ligated to the AccI site of pCAT-Basic (Promega) to obtain cdc2-TaqI (Fig. 1). cdc2-TaqI was digested by SmaI and ApaI to remove a 81-bp fragment, and the vector fragment was blunt-ended by treatment with Escherichia coli DNA polymerase (Klenow), which upon religation gave rise to cdc2-Taq⌬SA. cdc2-TaqI was digested by SmaI and BamHI, and the SmaI-BamHI fragment was ligated to HindIII-digested pCAT-Basic, bluntended by DNA polymerase (Klenow), and then digested by BamHI to yield cdc2-SmaI. To construct cdc2-Pst⌬NA, cdc2-PstI was digested by NarI, blunt-ended by DNA polymerase (Klenow), and then digested by HindIII to isolate the 766-bp fragment. This fragment was cloned into cdc2-TaqI, which had been digested with ApaI, blunt-ended by DNA polymerase (Klenow), digested with HindIII, and then gel-purified to remove the HindIII-ApaI fragment. To obtain pTI-CCAAT-CAT, the pTI-CAT plasmid containing a basic promoter (Shi et al., 1991) was digested with BglII, blunt-ended by DNA polymerase (Klenow), and ligated to the SmaI-ApaI fragment containing the two CCAAT box binding sites of the cdc2 promoter. In all these constructs, the cdc2 promoter fragments were fused to the reporter gene, CAT, as shown in Fig. 1.
Cell Culture, DNA Transfections, and CAT Assays-Monkey kidney cells (CV-1) and COS-7 cells (constitutively expressing SV40-LT) were cultured at 37°C, 5% CO 2 in Dulbecco's modified Eagle's medium containing 50 g/ml each of streptomycin and penicillin, and 10% defined/supplemented bovine calf serum. IDH4 cells were cultured in Dulbecco's modified Eagle's medium, 10% fetal calf serum, and 10 Ϫ6 M dexamethasone as described . IDH-4 cells were plated at 3-5 ϫ 10 3 /cm 2 in 100-mm or 35-mm dishes. Twenty-four hours after plating, dexamethasone-deprived cells were washed and shifted to medium lacking dexamethasone and phenol red but containing charcoal-stripped serum. Cells were grown for 7 days, and dexamethasoneadded cells were grown in the continual presence of dexamethasone. For dexamethasone-added/deprived cells, the cells were first dexamethasone-deprived for 7 days and then grown in the presence of dexamethasone for 24 h. Cells were harvested and fractionated into cytoplasmic and nuclear fractions as described (Chen et al., 1994). Plasmid DNAs were purified by CsCl-ethidium bromide equilibrium density gradients (Sambrook et al., 1989). For transfection, cells were grown to 50 -70% confluence and plasmid DNAs were transfected using Lipofectin as follows. A cdc2 promoter-CAT reporter plasmid (3 g) and the RSV-␤-gal plasmid (1 g;, E. coli lacZ under control of the RSV-LTR) as an internal control, were transfected into CV-1 cells. Cells were harvested 45 h post-transfection, and cell extracts were prepared by three cycles of freeze-thawing. CAT assays were carried out at 37°C for indicated intervals in a total volume of 125 l containing 250 mM Tris-HCl (pH 7.8), 25 g of butyryl-coenzyme A, and 0.25 Ci of [ 14 C]chloramphenicol (Moravek Biochemicals, Inc.) according to the protocol supplied by Promega and as described by Gorman et al. (1982). Acetylated [ 14 C]chloramphenicol derivatives separated by thin-layer chromatography were scraped and counted in a liquid scintillation counter. Each experiment was repeated at least twice with different DNA preparations. E. coli ␤-galactosidase activity was determined using the substrate o-nitrophenyl-␤-D-galactopyranoside as described (Craven et al., 1965). Transfection experiments with reporter plasmids were carried out in duplicate, and results with more than 30% variation between duplicates were discarded and the experiments repeated.
Mutagenesis of CCAAT Box Binding Sites in cdc2-TaqI Plasmid-Two overlapping oligodeoxynucleotides (shown below) containing CCAAT 3 GGCCT mutations in the consensus CCAAT box binding sites in the SmaI-ApaI fragment were annealed.

5Ј-GGGAAGCCTACCCAGCGTAGCTGGGCTCTGAggccCTGCTTTGAAAG-3Ј
5Ј-CCGGATTCAggccTCGGGTAGCCCGTAGACTTTCAAAGCAGggccTC-3Ј (Lowercase letters indicate the mutated CCAAT box, and the underline indicates the overlap between the oligonucleotides.) The annealed oligonucleotides were treated with E. coli DNA polymerase (Klenow) to fill-in the 5Ј-protruding ends and cloned into cdc2-TaqI, which had been digested with SmaI and ApaI, blunt-ended, and gel purified to remove the SmaI-ApaI fragment, to yield the plamid cdc2-TaqMutCCAAT.
Northern (RNA) Analysis-Total RNA was isolated using Trizol, fractionated on a 0.8% agarose-formaldehyde gel by electrophoresis and transferred to a GeneScreen membrane following the manufacturer's protocol. Antisense cdc2 mRNA probe was prepared from a cDNA clone using the SP6 RNA polymerase for in vitro transcription as described (Sambrook et al., 1989). The 18 and 28 S RNAs were visualized by staining the membrane with 0.02% methylene blue and 0.5 M NaAc (pH5.2), and destaining in 20% methanol. After photography, the methylene blue was removed from the gel by soaking in a solution containing 0.2 ϫ SSPE (Sambrook et al., 1989) and 1% SDS. The membrane was prehybridized at 65°C in hybridization solution containing 1 M NaCl, 0.05 g/ml dextran sulfate, 50 g/ml yeast tRNA, 40% (v/v) formamide, and 1% SDS for 3 h and then hybridized overnight at 65°C in hybridization solution containing 32 P-labeled cdc2 RNA probe. The membranes were rinsed twice with 1 ϫ SSC, 0.1% SDS and washed in this buffer at 65°C for 1 h. The blots were then rinsed twice with 0.3 ϫ SSC, 0.1% SDS at room temperature, and once in this buffer at 65°C for 1 h. The blots were subjected to autoradiography (NEF-495 x-ray film) at Ϫ70°C.
Gel Mobility Shift Assay-cdc2-TaqI or cdc2-TaqMutCCAAT was digested by SmaI and XbaI. The 180-bp SmaI-XbaI fragment containing the wild-type or mutated CCAAT box binding sites was dephosphorylated, and labeled with [␥-32 P]ATP by T4 polynucleotide kinase. Gel mobility shift assays were carried out as described (Chen et al., 1994). Briefly, nuclear extracts were mixed with labeled probe in 20 l of DNA-binding reaction buffer containing 25 mM HEPES-KOH (pH 7.9), 5 mM KCl, 0.5 mM EDTA, 1 g/ml bovine serum albumin, 10% glycerol, 0.25 mM dithiothreitol, and 0.1 g/ml poly(dI-dC) and incubated at 37°C for 30 min. The reaction was stopped by adding 2 l of stop solution containing 50 mM EDTA, 0.05% bromphenol blue, 0.05% xylene cyanole, and 5% glycerol. The DNA-binding complexes were separated on a 4% polyacrylamide gel. The gel was dried and exposed to NEF-495 film at Ϫ70°C.
Southwestern Analysis-The interaction between the CCAAT box motif and nuclear proteins was analyzed essentially as described (Silva et al., 1987). Nuclear proteins were separated on a 8% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. After renaturing, the proteins on the membrane were probed with radiolabeled DNA fragment containing CCAAT box binding sites. The membrane was then washed to remove unbound probe and exposed to NEF-495 film at Ϫ70°C.

Characterization of cdc2
Promoter-An analysis and characterization of control elements in the human cdc2 promoter that bind the c-Myb protein as well as the transcription factor E2F has been described (Dalton, 1992;Ku et al., 1993;Yamamoto et al., 1994). However, the functions of other control elements have not been studied. To further characterize the human cdc2 promoter, we constructed a series of deletions in the cdc2 promoter linked to the reporter gene CAT (Fig. 1). Monkey kidney (CV-1) cells were transfected with these reporter plasmids along with RSV-␤-gal, which encodes the E. coli ␤-galactosidase and serves as an internal control for transfection efficiency. CAT and ␤-galactosidase activities were determined in extracts prepared 45 h post-transfection.
As shown in Fig. 2 (A and B), the cdc2-PstI (parent) plasmid exhibited significant promoter activity, as reported previously (Ku et al., 1993). In contrast, cdc2-Pst⌬NA, having a deletion between the NarI and ApaI sites (Ϫ171 and Ϫ28 nucleotides upstream of the transcription start site) showed only 2% of the parent promoter activity. This result indicates that the Sp1, E2F, and the two inverted CCAAT box binding elements are essential for basal cdc2 promoter activity. Deletion of sequences upstream of the TaqI site (Ϫ359; cdc2-TaqI; Fig. 1) reproducibly increased promoter activity over that shown by the parent cdc2-PstI plasmid (see Fig. 2, A and B). This finding suggests that there may be a negative regulatory element upstream of the TaqI site. Deletion of sequences between the SmaI and ApaI sites, which contain the two CCAAT box motifs (cdc2-Taq⌬SA in Figs. 1 and 2 (A and B)), decreased promoter activity by 5-fold. When sequences upstream of the SmaI site (Ϫ109) were deleted (cdc2-SmaI, containing only the two CCAAT box binding sites, shown in Figs. 1 and 2 (A and B)), promoter activity was reduced about 2.5-fold compared with the cdc2-TaqI construct. These results suggested that although the Sp1, E2F, and the CCAAT-box motifs all contribute to cdc2 promoter activity, the two CCAAT box motifs play an essential and major role.
To explore the activity of the two CCAAT box binding sites in the cdc2 promoter, the SmaI-ApaI fragment was cloned upstream of the TATA box (BglII site) in the basic promoter/ reporter plasmid pTI-CAT (Shi et al., 1991). The resultant plasmid, pTI-CCAAT-CAT, was transfected into CV-1 cells and CAT activity was determined. Fig. 3 (A and B) shows that pTI-CAT, which contains only the TATA element, displayed low promoter activity, as reported (Shi et al., 1991). By contrast, when the two cdc2 CCAAT box binding sites were present upstream of the TATA element in pTI-CAT, there was a 14-fold increase in CAT activity. These data further demonstrate that the CCAAT box motifs in the cdc2 promoter are functional, not only in the cdc2 promoter, which lacks a TATA element, but also in a basic promoter containing only a TATA element.
Transactivation of cdc2 Promoter by SV40 T Large Antigen through CCAAT Box Binding Sites-We next sought to deter- A, autoradiography of a thin-layer chromatography plate from a representative CAT assay (Gorman et al., 1982) using extracts prepared from CV-1 cells transiently transfected with a cdc2 promoter/CAT construct shown in Fig. 1 (3 g each) and RSV-␤-gal plasmid (1 g; Zhao and Padmanabhan (1988)) as internal control. The CAT activity assay was carried out by incubation of the extracts with labeled [ 14 C]chloramphenicol at 37°C for 2 h. B, the results in A were quantitated by liquid scintillation counting of the radioactivity in each sample, and are presented as a bar graph. mine whether the cdc2 promoter is a target for transactivation by SV40-LT. CV-1 cells were cotransfected with a cdc2-promoter construct and a SV40-LT expression plasmid pCMV-TAg (Sakamoto et al., 1993;Dimri and Campisi, 1994;Hara et al., 1996). In control experiments, an equal amount of the pCMV vector plasmid (without an insert) was included instead of pCMV-TAg. The results shown in Fig. 4 (A and B) indicate that expression of SV40-LT efficiently induced transcription from the parent cdc2-PstI plasmid (compare lanes with and without SV40-LT). Deletion of sequences upstream of the TaqI site (Ϫ359 as in cdc2-TaqI; compare lanes with and without SV40-LT) or upstream of the SmaI site (Ϫ109 as in cdc2-SmaI; compare lanes with and without SV40-LT) did not appreciably affect the level of transactivation by SV40-LT.
These results indicate that the CCAAT box binding sites alone contribute 73% of the level of SV40-LT-mediated transactivation achieved with the cdc2-TaqI construct. This construct contains the ATF, c-Myb, Sp1, and E2F, as well as the CCAAT box motifs. Deletion of sequences between the SmaI (Ϫ109) and ApaI (Ϫ28) sites, which contains the two CCAAT box binding sites, dramatically reduced the SV40-LT-mediated transactivation (Fig. 4, cdc2-Taq⌬SA, lanes with and without SV40-LT). The cdc2-Pst⌬NA promoter showed low promoter activity, which could not be transactivated by SV40-LT (compare lanes with and without SV40-LT). Thus, the elements upstream of the Sp1 site, notably the ATF and c-myb binding sites, were unresponsive to SV40-LT-mediated transactivation. The low level of transactivation by SV40-LT observed with this construct is consistent with an earlier study showing that SV40-LT can activate transcription from a promoterless reporter gene (Rice and Cole, 1993). Taken together, these results show that the CCAAT box binding motifs are the major target for SV40-LT-mediated activation.
One mechanism by which SV40-LT may activate the cdc2 promoter is by disrupting transcriptionally inactive E2F-Rb or Rb-related protein complexes. SV40-LT interacts directly with Rb or Rb-related proteins, thereby releasing transcriptionally active E2F, which participate in the transcription of E2F-dependent genes (Chellappan et al., 1992;Dyson et al., 1990;Ludlow et al., 1989; for a review, see Nevins (1992)). Our results indicate that the E2F binding site in the human cdc2 promoter does not contribute significantly to SV40-LT-mediated activation (compare cdc2-TaqI, cdc2-Taq⌬SA, and cdc2-SmaI in Fig. 4). Alternatively, SV40-LT, which also interacts with p53, may relieve p53-mediated repression of cdc2 promoter. To investigate these possibilities and determine whether the Rb-or p53-binding regions of SV40-LT are required for activation of the cdc2 promoter, CV-1 cells were cotransfected with the cdc2-TaqI promoter containing the E2F binding sites along with pCMV-TAg (wild-type SV40-LT), pCMV-TAg-Rb Ϫ or pCMV-TAg-p53 Ϫ (mutant SV40-LT defective in Rb-or p53-binding, respectively; see Tevethia et al. (1988), Sakamoto et al. (1993), Dimri and Campisi (1994), and Hara et al. (1996)). The results of these experiments shown in Fig. 5 indicated that transactivation of the cdc2 promoter by SV40-LT was independent of its ability to bind to Rb or p53.
To confirm that activation of cdc2 promoter by SV40-LT was

FIG. 3. Activation of a heterologous basic promoter/reporter pTI-CAT by the CCAAT box binding motifs of the cdc2 promoter.
The CCAAT box binding elements of the cdc2 promoter were excised and cloned into the basic promoter pTI-CAT as described under "Experimental Procedures." Transfection and CAT activity assay were carried out as described in Fig. 2. Panels A and B represent the autoradiography and the bar graph as in Fig. 2.

FIG. 4. Transactivation of various deletion mutants of cdc2
promoter by SV40 T antigen. CV-1 cells were transfected with 3 g of a cdc2 promoter-CAT reporter plasmid ( Fig. 1) with or without 2 g of pCMV-TAg (Sakamoto et al., 1993). In the absence of pCMV-TAg plasmid, an equal amount of pCMV vector plasmid was used for transfections to equalize the total amount of plasmid DNA in each transfection. The CAT activity assay was carried out at 37°C for 1 h. Panels A and B represent the autoradiography and the bar graph as in Fig. 2. mediated through CCAAT box binding sites, increasing amounts of the SV40-LT expression plasmid (pCMV-TAg) and fixed amounts of two cdc2 promoter constructs, cdc2-TaqI and cdc2-Taq⌬SA, were cotransfected into CV-1 cells. These plasmids have the same upstream sequences except that cdc2-Taq⌬SA lacks the two CCAAT box binding sites. The results are shown in Fig. 6. As the amount of pCMV-TAg was increased, CAT expressed from the cdc2-TaqI plasmid increased correspondingly (up to 1 g). By contrast, CAT expressed from the cdc2-Taq⌬SA plasmid increased only moderately (up to 0.01 g) with increasing dosage of pCMV-TAg. Higher amounts of pCMV-Tag (1 and 2 g) had no effect on activation. The peak level of activation obtained with cdc2-TaqI (containing wildtype CCAAT box motif) was 8-fold higher than that obtained with the mutant cdc2-Taq⌬SA construct (in which the two motifs were deleted; Fig. 6).
It still remained possible that differences in the reporter expression driven by various cdc2 promoter segments were due to differences in the distance between the transcription factor binding sites and the transcription start sites created by the deletions we introduced in the promoter. To rule out this possibility, and further confirm that the CCAAT box binding motifs were the prime targets for SV40-LT-mediated transactivation, we introduced point mutations into the two CCAAT box binding sites in cdc2-TaqI by site-directed mutagenesis. The mutant cdc2-TaqMutCCAAT construct was transfected with or without pCMV-TAg into CV-1 cells, and cell extracts were assayed for CAT activity. The results, shown in Fig. 7, demonstrate that the response of the mutant cdc2-TaqMutCCAAT promoter to SV40-LT activation was significantly reduced, although it retained detectable basal promoter activity. These results presented in Figs. 6 and 7, taken together, support our conclusion that the two CCAAT box binding sites in the human cdc2 promoter contribute significantly to transactivation by SV40-LT.
The results of the above experiments measured SV40-LT transactivation of the cdc2 promoter in transient transfection assays. To determine whether SV40-LT also activates the en-dogenous cdc2 gene, we analyzed cdc2 mRNA levels in CV-1 cells mock-transfected or transfected with increasing amounts of pCMV-TAg plasmid, as well as the level of mRNA transcribed from the endogenous cdc2 gene in SV40-transformed COS-7 cells. At 45 h post-transfection, total RNA were extracted from the cells. As shown in Fig. 8, Northern hybridization using an antisense cdc2 RNA as a probe indicated that the endogenous cdc2 mRNA levels increased in a dose response manner with increasing amounts of SV40-LT (lanes 1-7). The highest level of cdc2 mRNA induced by SV40-LT was about 3-fold greater than the level in mock-transfected cells (compare lane 1 with lane 6). Because transfection is well below 100% efficient, these results suggest that SV40-LT can induce substantial levels of the endogenous cdc2 mRNA.
The results thus far indicate that SV40-LT activation of the cdc2 promoter is mediated by the CCAAT box binding motifs. We therefore asked whether SV40-LT induces and/or interacts with a CCAAT box-binding factor (CBF). We performed electrophoretic mobility shift assays (EMSA) to analyze the DNAprotein complexes formed from such interactions using radiolabeled wild-type and mutant CCAAT box binding elements in the cdc2 promoter. As shown in Fig. 9, nuclear extracts from CV-1 cells transfected with pCMV-TAg (lane 2) or from SV40-LT-expressing COS-7 cells (lane 3) formed a specific DNAprotein complex, whereas extracts from mock-transfected CV-1 cells did not form this complex (lane 1). Furthermore, when a mutant CCAAT box motif was used for EMSA, this specific DNA-protein complex was not formed in all three extracts (lanes 4 -6). These results demonstrate that expression of a specific CBF was augmented by SV40-LT in CV-1 cells in FIG. 5. Transactivation of cdc2 promoter by wild-type and mutant SV40-LTs lacking pRb or p53 binding domains. The plasmids encoding wild-type or mutant SV40-LTs were individually cotransfected with cdc2-PstI into CV-1 cells, and the CAT activity was measured as described in Fig. 4, except that the enzyme reaction was carried out at 37°C for only 30 min.

FIG. 6. Dose response of CCAAT box binding motifs to transactivation by SV40 T antigen.
CV-1 cells were cotransfected with 3 g of cdc2-TaqI or cdc2-Taq⌬SA promoter-CAT reporter plasmid (Fig.  1) and with increasing amounts of pCMV-TAg plasmid. CAT activity assay was carried out at 37°C for 1 h. Panels A and B represent the autoradiography and the bar graph as in Fig. 2. The bar graph shows the CAT activities (cpm) versus the amount of pCMV-TAg plasmid used for transfections (g). transient transfection assays and in stably transformed COS-7 cells. Lum et al. (1990) reported that a specific CBF of 114 kDa binds to the CCAAT box motif of the HSP70 promoter and is responsible for the activation of HSP70 gene expression. To determine whether the CBF that binds the CCAAT box binding motif in the cdc2 promoter is the same as that involved in the activation of the HSP70 promoter, an antibody to the HSP70-CBF was used for supershift analysis in EMSA. As shown in Fig. 9 (lanes 7-9), the HSP70-CBF antibody had no effect on the mobility of the SV40-LT-inducible cdc2 promoter CCAAT box motif-specific DNA protein complex. To further confirm this result, we asked whether the HSP70-CBF can activate the cdc2 promoter in a transient transfection assay. As shown in Fig. 10, expression of HSP70-CBF in CV-1 cells activates the HSP70 promoter as expected (Agoff et al., 1993). However, HSP70-CBF had no effect on the cdc2 promoter. These results suggest that a novel CBF, which is involved in activation of cdc2 promoter, may be induced by SV40-LT.
To further characterize the CBF that interacts with the cdc2 promoter, Southwestern analysis was performed. Nuclear extracts from mock-transfected cells, pCMV-TAg-transfected CV-1 cells, and COS-7 cells were fractionated by SDS-PAGE. The proteins were transferred to a nitrocellulose membrane, renatured, and probed with a 32 P-labeled DNA fragment containing the CCAAT box binding motif. Proteins bound to the labeled probe were detected by autoradiography. The results, shown in Fig. 11, indicate that a protein of about 110 kDa specifically bound to the CCAAT box binding sites in the cdc2 promoter. In mock-transfected CV-1 cells, the level of this protein was low (lanes 1-3), whereas SV40-LT, whether transiently expressed in CV-1 cells (lanes 4 -6) or constitutively expressed in stably transformed COS-7 cells (lanes 7 and 8), greatly enhanced the expression of this CBF. Probing the membrane with the anti-HSP70/CBF antibody showed that the HSP70/CBF protein migrated more slowly than the radiolabeled SV40-LT-inducible protein (data not shown).
We also tested the ability of SV40-T to induce a 110-kDa CBF/cdc2 protein in human fibroblasts that are reversibly immortalized by an inducible SV40-LT (IDH4 cells; Wright et al. (1989)). IDH4 cells were grown in the absence or presence of dexamethasone (Fig. 12, Ϫdex, ϩdex, and Ϫ/ϩdex), and extracts were prepared 7-8 days later for Western and Southwestern analyses. Under these conditions, the cells expressed high levels of SV40-LT in the presence of dexamethasone (relative to the tubulin control), and very low levels of SV40-LT in the absence of dexamethasone (Fig. 12). The same extracts also showed high and low levels of the 110-kDa CBF/cdc2 protein, respectively (Fig. 13). When dexamethasone-deprived cultures expressing low levels of T antigen and CBF/cdc2 were restimulated with dexamethasone (Fig. 12, Ϫ/ϩdex), both T antigen and CBF/cdc2 were induced substantially within 24 h (Figs. 12 and 13), and the CBF/cdc2 was enriched in the nuclear fraction (Fig. 13). These results support our data on the induction of a CBF/cdc2 protein by SV40-LT in CV-1 and COS-7 cells. Together, the data suggest that SV40-LT induces the expression of a potentially novel 110-kDa CBF/cdc2 in monkey kidney cells as well as in human fibroblasts.

FIG. 7. SV40-LT mediated transactivation of the wild-type and mutant CCAAT box binding motifs.
The CCAAT box binding motifs in the cdc2-TaqI promoter/reporter plasmid were mutated specifically by point mutagenesis as described under "Experimental Procedures." Transfections and CAT assays were carried out as described in Fig. 4, except that the enzyme reaction was carried out at 37°C for only 30 min to ensure that the substrates were not used up at the end of the incubation period. FIG. 9. DNA-binding activity of nuclear extracts from mocktransfected and pCMV-TAg-transfected CV-1 cells and COS-7 cells. Nuclear extracts were prepared and gel mobility shift assays were performed as described under "Experimental Procedures." Lanes 1, 4, and 7, extracts from mock-transfected CV-1 cells. Lanes 2, 5, and 8, extracts from pCMV-TAg-transfected CV-1 cells. Lanes 3, 6, and 9, extracts from COS-7 cells. Lanes 1-3 and 7-9, extracts were probed with the wild-type CCAAT box binding site. Lanes 4 -6, extracts were probed with the mutant CCAAT box binding site. Lanes 7-9, monoclonal antibody against HSP70-CBF (Agoff et al., 1993) was added.

DISCUSSION
This study further dissects the cis-regulatory elements in the human cdc2 promoter that contribute to basal transcription, and reveals the CCAAT box binding motifs as major targets for transactivation mediated by SV40-LT. Studies carried out in other laboratories on the human cdc2 promoter revealed multiple regulatory elements, suggesting complex regulation of cdc2 gene expression. The human cdc2 promoter does not contain a classical TATA-like element, although two TTTGAAA elements, located between 66 and 10 bp upstream of the transcription start site, are present (Dalton, 1992;Ku et al., 1993). We found that the basal activity of cdc2-TaqI (Ϫ359) was reproducibly higher than that of cdc2-PstI (Ϫ722) in cycling CV-1 cells. This suggests that the region upstream of Ϫ359 nucleotides may have a negative regulatory element such as the one that is highly homologous to the yin-yang-1 (YY-1; Shi et al., 1991) binding site (AAAATGT). The YY-1 site at position Ϫ593 to Ϫ587 in the cdc2 promoter lies in a region that is responsive to cooperative induction by c-Myc and Ha-Ras v-12 (Born et al., 1994). Further mutational analysis of the cdc2 promoter in our study revealed that, within the Ϫ359 region, the Sp1, E2F, and two CCAAT box binding motifs are important for overall basal activity.
Growth-dependent expression of the human cdc2 gene suggests that it is a likely target for tumor suppressor proteins Rb and p53 which are negative regulators of proliferation. It is also a likely target for positive regulators of proliferation such as cellular protooncogene products. Rb or the Rb-related proteins p107 and p130 interact physically with the components of the transcription factor E2F (Bagchi et al., 1991;Chellappan et al., 1991;Chittenden et al., 1991;Nevins, 1992). The Rb-E2F interaction inactivates the transcription of E2F-dependent genes, many of which are needed for DNA synthesis. These genes include those encoding dihydrofolate reductase, thymidine kinase, c-Myc, DNA polymerase ␣, and Cdc2 kinase.
A major breakthrough regarding the function of Rb in the control of proliferation came from the discoveries that the transforming proteins of certain DNA tumor viruses target Rb by physical interaction and inactivate its function. For example, SV40-LT (De Caprio et al., 1988), adenovirus E1A (Whyte et al. 1988), and HPV-E7 (Dyson et al., 1989) proteins bind and inactivate Rb. E1A induces cell proliferation and the endogenous cdc2 mRNA and p34 cdc2 protein (Draetta et al., 1988;Wang et al., 1991;Dalton, 1992). Since the E1A protein encoded by the 12 S mRNA has no transactivation function (for a review, see Moran and Matthews (1987)), but forms a stable complex with Rb (Whyte et al., 1988) and disrupts Rb-E2F complexes in vitro (Bagchi et al., 1990;Bandara and La Thangue, 1991), it was thought that activation of cdc2 promoter by E1A occurred via disruption of Rb-E2F complexes (Dalton, 1992). SV40-LT and HPV-E7 proteins also act via E2F binding sites in promoters, and the Rb-binding domains of SV40-LT and HPV-E7 are critical for this activity (Phelps et al., 1988;Loeken and Brady, 1989). These results suggest that E1A, SV40-LT, and HPV-E7 proteins may have a common FIG. 12. Conditional expression of SV40-LT in human diploid fibroblasts. IDH-4 cells were grown in the absence and presence of dexamethasone (Ϫdex and ϩdex, respectively) and the lysates were prepared for Western blot analyses as described under "Experimental Procedures." After 7 days, cells grown in the absence of dexamethasone were induced for 24 h with dexamethasone (Ϫ/ϩdex). The extracts were fractionated by SDS-PAGE (8%) and subjected to Western blotting using a mouse monoclonal antibody against SV40-LT or against ␤-tubulin as control. The secondary antibody (horseradish peroxidase-conjugated sheep anti-mouse IgG from Amersham Corp.) binding was detected by a chemiluminescence kit as described under "Experimental Procedures."  Agoff et al. (1993)) and cdc2-PstI ( Fig. 1) or HSP70-CAT and the CAT activity assays were carried out as described in Fig. 4, except that the enzyme reaction was carried out at 37°C for 2 h.
FIG. 11. Southwestern analysis of binding of nuclear proteins from CV-1 cells to CCAAT box motif of cdc2 promoter. Nuclear proteins were separated on a 8% SDS-PAGE and transferred to nitrocellulose membrane. The proteins were renatured and probed with 32 P-labeled CCAAT box motif as described under "Experimental Procedures." Lanes 1-3, 1, 2, and 3 g of total protein from mock-transfected CV-1 cell lysate. Lanes 4 -6, 1, 2, and 3 g of proteins from pCMV-TAgtransfected CV-1 cell lysate. Lanes 7 and 8, 1.4 and 2.8 g of COS-7 cell lysate, respectively. pathway in transactivation of promoters containing E2F binding sites.
Our results indicate that the two CCAAT box binding motifs in the human cdc2 promoter not only contribute significantly to the overall basal activity of the promoter but also are the major target for transactivation by SV40-LT. Contrary to our expectations, our results also clearly show that the region containing the Sp1 and E2F binding sites are not responsive to a significant extent to transactivation by SV40-LT in cycling CV-1 cells. Furthermore, we found little difference between the wild-type and a mutant SV40-LT defective in Rb binding in their ability to stimulate cdc2 promoter constructs containing E2F binding sites. At least one other study showed that E2F-1 failed to activate a cdc2 promoter-reporter (CAT) gene in human glioblastoma (T98G) and osteosarcoma (SAOS-2) cells (Sala et al., 1994). Moreover, the Rb binding domain of SV40-LT was not required to immortalize rat lung epithelial cells or elevate p34 cdc2 and cyclin A levels (Oshima et al., 1993). Our results show that neither the Rb or p53 binding domains are critical for SV40-LT activation of the cdc2 promoter.
Previous studies have characterized a nuclear factor that bound the two inverted CCAAT box binding elements (Ϫ146 to Ϫ121 and Ϫ82 to Ϫ56) present in the Rous sarcoma virus long terminal repeat (RSV-LTR) promoter, and addition of serum to serum-deprived rat 3Y1 fibroblast cells increased the level of this factor (Dutta et al., 1990;Maity et al., 1988). This nuclear factor also bound the CCAAT box elements in the promoters of the human heat shock protein 70 (HSP70), c-Ha-ras, and histone H1 genes, but not the CCAAT box in the adenovirus type 2/5 origin of DNA replication (which is recognized by NFI/CTF family; see Santoro et al. (1988)), or the SV40 enhancer core sequence (recognized by CCAAT enhancer-binding protein C/EBP; Landschultz et al., 1988). The CCAAT box binding factor (CBF) that controls transcription from the HSP70 promoter has been cloned; it encodes a 114-kDa polypeptide that binds the CCAAT box element (Lum et al., 1990). From the results of Dutta et al. (1990), it appears that the serum-induced factor that recognizes the CCAAT box elements in RSV-LTR, c-Ha-ras, and histone H1 promoters is either related to, or likely to be the same as, the 114-kDa CBF that binds the HSP70 promoter (CBF/HSP70).
Our evidence suggests that CBF/cdc2 differs from CBF/ HSP70. On the other hand, CBF/cdc2 may be related to the factor (CCAAT box-binding protein; CBF/tk) that is induced by serum in presenescent human diploid fibroblast (IMR-90) cells but absent in senescent cells (Pang and Chen, 1993). CBF/tk recognizes two inverted CCAAT box motifs in the human thymidine kinase (tk) gene (Pang and Chen, 1993). The biochemical properties of this factor (CBF/tk), such as sensitivity to heat, metal chelating agents, and protein synthesis inhibitors, suggest that it is a distinct member of the CCAAT box-binding protein family that could play a role in the cell cycle-and senescence-dependent regulation of TK gene expression (Pang and Chen, 1993). Interestingly, SV40-LT extends the limited proliferative life span of human diploid fibroblasts, and induces DNA synthesis in quiescent and senescent cells without inducing mitosis (Gorman and Cristofalo, 1985;Shay and Wright, 1989;Wright et al., 1989) (for reviews, see Goldstein (1990), Peacocke and Campisi (1991), Campisi et al. (1996), and references therein). SV40-LT has also been shown to induce DNA synthesis, p34 cdc2 kinase, and Rb phosphorylation in quiescent baby rat kidney cells (Wang et al., 1991). The results of our study indicate that SV40-LT mediates its activation of the human cdc2 gene through induction of a specific CBF (CBF/ cdc2) in immortal monkey kidney cells (CV1, COS) and reversibly immortal human fibroblasts (IDH4), which is distinct from the CBF/HSP70. Therefore, it will be interesting to determine the levels of the CBF/cdc2 in senescent and quiescent human diploid fibroblasts.