Evidence for Repressional Role of an Inverted CCAAT Box in Cell Cycle-dependent Transcription of the Human DNA Topoisomerase IIα Gene*

Expression of DNA topoisomerase IIα (topo IIα) is cell cycle-regulated at both the transcriptional and the post-transcriptional levels. In order to identifycis-acting elements responsible for transcriptional regulation during the cell cycle, we investigated NIH/3T3 cells stably transfected with luciferase reporter plasmids containing various lengths of the human topo IIα gene promoter. Serum-deprived cells expressed low levels of luciferase, and following serum-induced cell cycle re-entry luciferase levels were gradually elevated 2-fold. During S phase, a steep 3-fold increase in luciferase activity was seen, reaching its maximum approximately 22 h after serum addition. This pattern was observed with both a full-length (nucleotides (nt) −295 to +90] and a deletion (nt −90 to +90) promoter construct. In contrast, when testing a deletion construct (nt −51 to +90) lacking the first inverted CCAAT box (ICB1) the S phase-specific induction was absent. Mutation of ICB1 revealed that it had a repressive character, since luciferase levels rose rapidly to maximal levels immediately following serum addition. Furthermore, electrophoretic mobility shift assays demonstrated a marked decrease in ICB1 binding activity following serum addition. Together, this suggests a role of ICB1 in cell cycle-dependent repression of topo IIα transcription.

Type II DNA topoisomerases are essential nuclear enzymes that regulate DNA topology by creating a transient double strand break through which a second intact double helix is passed (reviewed in Refs. 1 and 2). They are also the key cellular targets for a number of clinically important anticancer agents such as etoposide and the anthracyclines, doxorubicin and daunorubicin (3).
There exist two isoforms of topoisomerase II (topo II) 1 in higher eukaryotes, namely the ␣ (170 kDa) and the ␤ (180 kDa) isoform encoded by separate genes (4,5). The ␣ isoform is primarily required for chromosome (de)condensation and sister chromatid segregation during mitosis (6,7), but roles for topo II␣ in replication and transcription have also been proposed (1). In contrast, a precise biological function of topo II␤ has not yet been established, although recent evidence has emerged for a role during differentiation (reviewed in Ref. 8).
A major difference between the ␣ and ␤ isoform of topo II is their expression. The topo II␤ gene is constitutively expressed in proliferating as well as differentiated tissue (9), and its transcription rate is more or less constant throughout the cell cycle. In contrast, topo II␣ activity is primarily associated with proliferating cells and progressively decreases as cells are induced to differentiate or deprived of serum (10). Levels of topo II␣ enzyme also changes within a single cell cycle, with low levels during G 0 /G 1 and accumulation during S and G 2 to reach maximal levels during mitosis (11). This accumulation is obtained by enhancement of both transcriptional activity and mRNA stability during S phase (12,13). This cell cycle-dependent pattern of expression is conserved among mammals, including mice and humans (14). However, no specific cis-acting DNA elements and/or trans-acting factors have been associated with this regulation.
Topo II activity is an important determinant for the cytotoxic effect of the topo II-targeting drugs. Thus, a direct correlation between topo II levels and drug sensitivity of cells is a common finding. Indeed, reduction in the amount of topo II enzyme is one mechanism by which cancer cells acquire drug resistance (reviewed in Ref. 15). For this reason, considerable effort has been made to determine the mechanisms regulating topo II expression. These investigations have revealed that topo II␣ transcription is highly susceptible to environmental stimuli such as heat shock, growth arrest, and drug treatment. For instance, when Swiss 3T3 cells are confluence-arrested topo II␣ mRNA is rapidly down-regulated in response to reduced binding of the transcription factor NF-Y to an ICB (16). Another ICB has been associated with both transcriptional activation and repression of the topo II␣ promoter in response to heat shock and p53 expression, respectively (17,18). During phorbol ester and sodium butyrate induced monocytic differentiation of promyeloid HL-60 cells the topo II␣ promoter is trans-activated during the early stages, whereas it is down-regulated at later stages (19). Recently, the transcription factors c-MYB, B-MYB, NF-M, and Sp3 have been identified as trans-activators of the topo II␣ gene promoter in various cellular backgrounds (20 -22).
We have investigated human topo II␣ gene regulation during the cell cycle. Using NIH/3T3 cells stably transfected with luciferase reporter plasmids containing various promoter segments we show that induction of topo II␣ promoter activity was sensitive to treatment with aphidicolin and roscovitine. Furthermore, S phase-specific induction can be partially reversed by treatment with the same drugs. In addition, we demonstrate that an ICB within the minimal promoter is required for G 0 / * This work was supported in part by the Danish Cancer Society. 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.
¶ Present address: Institute of Cancer Biology, Danish Cancer Society, DK-2100 Copenhagen, Denmark.

Plasmids-Human
Cell Culture-The mouse fibroblast cell line NIH/3T3 (gift from C. Caradelli) was maintained in Dulbecco's modified Eagle's medium, supplemented with 10% fetal calf serum, 100 units/ml penicillin G, and 100 g/ml streptomycin sulfate (Life Technologies, Inc.), in an incubator at 37°C in a humidified atmosphere containing 5% CO 2 . NIH/3T3 cells were stably transfected by cotransfecting the Ϫ562TOP2LUC, Ϫ384T-OP2LUC, Ϫ295TOP2LUC, Ϫ219TOP2LUC, Ϫ148TOP2LUC, Ϫ90T-OP2-LUC, Ϫ51TOP2LUC, Ϫ295(mSP1)TOP2LUC, Ϫ295(mMYB)T-OP2LUC, and Ϫ295(mICB1)TOP2LUC constructs together with the pTK-Hyg vector (CLONTECH) 10:1 (w/w) by the calcium phosphate method (23). Stable transfectants were selected for 4 weeks with 200 g/ml hygromycin B (Calbiochem). Subsequently, cells were maintained in DMEM containing 200 g/ml hygromycin B. As we only investigated the relative changes of a given promoter construct, the entire population of resistant cells rather than individual clones was pooled, thus eliminating effects due to copy number and site of integration. Synchronization of stably transfected NIH/3T3 cells by serum starvation was performed by incubating cells for 72 h in DMEM containing 0.5% fetal calf serum.
Thymidine Incorporation and Luciferase Assay of NIH/3T3 Cells-NIH/3T3 stable topoisomerase II␣ promoter transfectants were seeded in six-well clusters (approximately 2 ϫ 10 4 cells/well) and serumstarved to obtain synchronized populations. Cells were released into complete medium, and cell lysates were prepared every 2 h for a 22-h period. Thirty minutes prior to harvest, cells were incubated in prewarmed and CO 2 -equilibrated complete medium containing 5 Ci/ml [methyl-3 H]thymidine (47.0 Ci/mmol; Amersham Pharmacia Biotech). Subsequently, cells were washed twice with phosphate-buffered saline and then lysed with reporter lysis buffer (Promega) and lysates were submitted to luciferase assay and quantitation of thymidine incorporation. Briefly, 150 l of cold 20% trichloroacetic acid was added to an equal volume of cell lysate, vortexed, and incubated 15 min on ice. The lysates were then filtered through Whatman GF/C glass fiber filters, washed, air-dried, and incorporated radioactive thymidine quantitated by liquid scintillation counting. Luciferase assays were performed with 25 l of cell lysate using the Luciferase Assay System (Promega) according to the manufacturer's instructions. Total light production was measured with a Turner Designs model TD-20/20 luminometer.
Drugs-Aphidicolin (Sigma) was used at 3 M for complete inhibition of DNA synthesis. Roscovitine (Calbiochem), an inhibitor of cyclin-dependent kinases CDC2, CDK2, and CDK5 (36), was used at 25 M for inhibition of CDK2-mediated S phase entry. For inhibition of RNA and protein synthesis, cells were treated with 0.1 g/ml actinomycin D and 10 g/ml cycloheximide (Sigma), respectively.
Preparation of Nuclear Extracts-NIH/3T3 cells (ϳ4 ϫ 10 7 ) were collected by exposure to trypsin, washed once in phosphate-buffered saline, and resuspended in 200 l NE buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 0.75 mM spermidine, 0.15 mM spermine, 0.2 mM EDTA, 0.2 mM EGTA, 0.5 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride) and incubated 15 min on ice. Cells were then lysed by passage 10 times through a 25-gauge needle, and the lysate was centrifuged for 30 s in a microcentrifuge at 10,000 ϫ g. The resulting nuclear pellet was resuspended in 100 l of NESG buffer (20 mM HEPES, pH 7.9, 0.4 M NaCl, 25% (v/v) glycerol, 0.75 mM spermidine, 0.15 mM spermine, 0.2 mM EDTA, 0.2 mM EGTA, 0.5 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride), incubated 30 min on ice with frequent mixing, and centrifuged for 20 min at 20,000 ϫ g at 4°C to remove insoluble material. The supernatant (nuclear extract) was stored at Ϫ80°C, and its protein concentration was determined by the Bio-Rad protein assay according to the manufacturer's instruction.
Electrophoretic Mobility Shift Assay-To detect sequence-specific protein-DNA interactions, EMSAs were performed. Briefly, 15 g of nuclear extract protein was incubated for 30 min at room temperature in a final volume of 20 l containing 10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol, 8% (v/v) glycerol, 0.1 g of poly(dI-dC) DNA (Amersham Pharmacia Biotech), and 2 ϫ 10 4 cpm of 32 P-labeled double-stranded oligonucleotide probe in the absence or presence of various competitors as indicated. The reaction mixtures were then applied to a nondenaturating 5% polyacrylamide gel and separated by electrophoresis at 100 V for 3 h in a buffer containing 25 mM Tris, 190 mM glycine, and 1 mM EDTA. Following electrophoresis, the gel was dried, exposed, and then recorded using a PhosphorImager. The following oligonucleotides (together with complementary ones) were used for EMSAs (5Ј to 3Ј): GAGTCAGGGATTGGCTGGTCTGCT-TCGGGC (ICB1) and GAGTCAGGGAAAAACTGGTCTGCTTCGGGC (mICB1). All oligonucleotides were purchased from Life Technologies, Inc. End labeling of oligonucleotides was done using the Ready-To-Go TM T4 polynucleotide kinase (Amersham Pharmacia Biotech) and [␥-32 P]ATP (3000 Ci/mmol; Amersham Pharmacia Biotech) according to the manufacturer's instructions. Specific activity of the labeled oligonucleotides were determined by Cerenkov counting.

Promoter Activity in Stable Cell Lines Exhibits Wild Type
Behavior-The topo II␣ transcription pattern is characterized by an increase during S phase of the cell cycle. To examine this pattern in detail, cell lines stably transfected with a range of different promoter constructs were prepared. The feasibility of this system required that a stably transfected full-length promoter construct would mimic the previously described expression pattern of topo II␣ mRNA (12). NIH/3T3 cells stably transfected with the Ϫ562TOP2LUC construct were synchronized by serum starvation, and luciferase activity and incorporation of radioactively labeled thymidine was measured every 2 h for a period of 22 h following serum addition. Low levels of luciferase activity and thymidine incorporation were observed in serum starved cells (Fig. 1A). Upon serum addition, when the cells treated with Me 2 SO re-entered the cell cycle, an immediate elevation of luciferase activity was detected reaching a plateau after approximately 4 -6 h (Fig. 1A). Subsequently, as cells entered S phase, seen as a steep increase in thymidine incorporation (12-14 h after serum addition), a concomitant increase in luciferase activity was seen that continued throughout the entire time course (Fig. 1A). Together, this fitted the expected expression pattern of topo II␣ mRNA and, as an identical pattern, was obtained in untreated cells (data not shown; Fig. 3A); this approach was accordingly regarded as feasible.
S Phase Entry Is Necessary for Induction of Topo II␣ Gene Transcription-It has been demonstrated that the accumulation of topo II␣ mRNA during late S phase can occur independently of DNA synthesis and S phase entry (12). Specifically, treatment of synchronized HeLa cells with the DNA polymerase ␣ inhibitor aphidicolin only slightly delayed topo II␣ mRNA accumulation. In order to investigate if the transcriptional induction of the topo II␣ gene promoter also was insensitive to DNA synthesis inhibition, cells synchronized by serum starvation were treated with 3 M aphidicolin or 25 M roscovitine from 1 h after serum addition and throughout the 22-h time course. When treated only with solvent (Me 2 SO) a normal expression pattern was observed (Fig. 1A). In contrast, treatment with aphidicolin or roscovitine abolished DNA synthesis and hence entry into S phase (Fig. 1, B and C). Accordingly, no S phase-specific induction of luciferase expression was seen (Fig. 1, B and C). However, an induction to almost wild type levels (Fig. 1D, black bars) could be re-established in aphidicolin-and roscovitine-treated cells when the drug was removed 16 h after serum addition and the cells incubated for another 8 h in the absence of drug (Fig. 1D, middle panel), indicating that the induction indeed was S phase-specific.
Interestingly, an approximately 2.5-fold increase in luciferase activity occurred in aphidicolin-treated cells 4 -6 h following serum addition (Fig. 1B). In roscovitine-treated cells, an approximately 2-fold increase in luciferase was also observed, although with different kinetics (Fig. 1C). This immediate elevation of luciferase expression was also found in Me 2 SOtreated cells (Fig. 1A), implying either the presence of a specific serum-responsive cis-acting element in the promoter or a "global" effect of serum addition.
Induction of the Topo II␣ Promoter Is Dependent on Active S Phase Progression-In order to investigate if S phase entry per se was sufficient to mediate the observed S phase-specific induction of luciferase activity, or if ongoing DNA synthesis was required, cells were treated with aphidicolin and roscovitine 14 h after serum addition when DNA synthesis had started (see Fig. 1A). In addition, cells were treated with actinomycin D or cycloheximide to inhibit RNA or protein synthesis, respectively. As expected, treatment with solvent (Me 2 SO) had no effect on luciferase activity, whereas drug treatment more or less abolished the S phase-specific induction. In the case of actinomycin D, luciferase levels remained almost constant (Fig.  2), presumably due to contributions from already transcribed luciferase mRNAs and active luciferase protein. As seen, treatment with cycloheximide caused a steady decrease in luciferase activity (Fig. 2), reflecting the luciferase protein half-life of approximately 3 h (24). Although greatly impairing the S phase-specific induction, treatment with roscovitine or aphidicolin did not completely abrogate transcription of luciferase mRNA, as reflected by the higher levels of luciferase activity compared with the actinomycin D-treated cells (Fig. 2).
Sequences between bp Ϫ90 and Ϫ51 Are Required for Proper S Phase Induction-The expression pattern mediated by the full-length promoter is, as mentioned, characterized by a steep increase in transcriptional activity coinciding with the onset of S phase (see Fig. 1A). This results in an approximately 7-fold induction relative to promoter activity in serum-starved cells, while the induction is only about 3-fold compared with the activity levels reached at the immediate plateau at 6 -12 h (Figs. 1A and 3A). All deletion constructs tested displayed similar thymidine incorporation patterns, although with some variations in incorporated counts ( Fig. 3; data not shown). When deleting sequences between bp Ϫ562 and Ϫ90, no significant changes were observed in the ability of the constructs to mediate the S phase-specific induction of luciferase activity (Fig. 3, A-C), and induction levels ranged from 4-to 7-fold when comparing maximal observed luciferase activities with those in serum-starved cells (Table I). Interestingly, when comparing the induction from the immediate plateau levels to maximal levels, the increase only ranged between approximately 2.5and 3-fold (Table I). Thus, although important for basal pro-  moter activity per se, the sequences between bp Ϫ562 and Ϫ90 are dispensable with regard to mediating S phase-specific transcriptional induction in a heterologous context. In contrast, this induction was greatly impaired when deleting sequences between bp Ϫ90 and Ϫ51 (Fig. 3D). Although capable of an approximately 3-fold total induction, the S phase-specific induction was almost absent (Fig. 3D; Table I).
The sequences between bp Ϫ90 and Ϫ51 harbor only one cis-acting element, namely an ICB. Since the Ϫ51TOP2LUC construct had a low basal activity (not shown), this could also influence the expression. To demonstrate a possible role of this ICB (termed ICB1) in the S phase-specific induction of luciferase activity, the Ϫ295TOP2LUC construct was mutated in its ICB1. Additional mutations were also made in the MBS and the GC-box (both contained in the Ϫ51TOP2LUC construct), and the expression patterns of the three mutated constructs were examined together with the Ϫ295TOP2LUC construct. All three mutations lead to a small reduction in the maximal obtainable luciferase levels compared with the wild type promoter (Fig. 4). The Ϫ295(mSp1)TOP2LUC changed expression pattern in a modest manner, as loss of the GC-box seemed to reduce the immediate increase observed in all other constructs, but not, as mentioned, the ability to mediate S phase-specific induction ( Fig. 4B; Table I). Nevertheless, the constructs carrying mutations in the MBS or the GC-box retained the ability of S phase induction. This was not the case for the Ϫ295(mICB1)TOP2LUC construct, as its expression differed markedly from the other three constructs (Fig. 4D). Most notable was the sharp and rapid increase in luciferase activity immediately after serum addition, reaching a plateau after approximately 6 h (Fig. 4D). From this plateau the levels did not increase significantly as the cells progressed into S phase ( Fig. 4D; Table I). Together this suggests an important repressive role of ICB1 in the S phase-specific induction of the topo II␣ gene.
ICB1 Binding Activity Decreases in Response to Cell Cycle Re-entry-Since the abovementioned data indicated that ICB1 might have a role in regulating S phase-specific induction of human topo II␣ promoter activity, we decided to investigate if a cell cycle-dependent ICB1 binding activity was present in NIH/3T3 cells. In order to do so, nuclear extracts were prepared from synchronized cells at different time points following serum addition and tested for ICB1 binding activity in EMSAs. When examining exponentially growing NIH/3T3 cells with a oligonucleotide probe harboring ICB1, they contained a probe binding activity (Fig. 5A, lane 2). Serum-starved cells (t ϭ 0 h; Fig. 5A, lane 3) contained large amounts of probe binding activity compared with exponentially growing cells (Fig. 5A,  lane 2). As cells were stimulated to cell cycle re-entry by serum

TABLE I Maximal and S phase-specific induction of topo II␣ promoter deletion
and mutation constructs Experiments similar to the one described in the legend to Fig. 3 were performed. Maximal induction (LU max /LUt ϭ 0) was determined by comparing the maximal observed values (in light units, LU) with the values at t ϭ 0 h, and the S phase-specific induction (LU max /LUt ϭ 6) was determined by comparing the maximal observed values with the immediate plateau levels (t ϭ 6 h). Values are averages of two to eight independent experiments (ϮS.E. when applicable).
addition (t ϭ 0 h), probe binding activity decreased as cells progressed through the cell cycle (Fig. 5A, lanes 4 -9). Most notable are the very reduced levels of probe binding activity in nuclear extracts from cells 12-20 h post-stimulation (Fig. 5,  lanes 6 -8), when cells were in S phase (Fig. 1A). Twenty-four hours after serum addition, a slight increase in binding activity was observed (Fig. 5A, lane 9), presumably coinciding with entry into mitosis.
To demonstrate that the probe binding activity was dependent on ICB1, a series of EMSA experiments with various competitors were conducted. Nuclear extracts from exponentially growing cells were incubated with labeled ICB1 probe alone or in combination with 100-fold excess of unlabeled ICB1 or mICB1 probe. Coincubation with a probe containing a mutation in ICB1 (mICB1) did not reduce the amount of probeprotein complex (Fig. 5B, lane 4), whereas the ICB1 probe was an efficient competitor when added in 100-fold excess (Fig. 5B, lane 3).

DISCUSSION
Cell cycle-dependent regulation of human DNA topo II␣ gene expression is controlled both at the transcriptional and post-transcriptional levels. To investigate the transcriptional changes in topo II␣ expression during the cell cycle, a panel of NIH/3T3 cell lines stably transfected with luciferase reporter plasmids containing various lengths of the human topo II␣ promoter was constructed. The entire population of resistant cells, rather than individual clones, was pooled in order to eliminate unwanted effects due to copy number and site of integration. Initial experiments indicated that this was a feasible approach, as the luciferase activity pattern in synchronized cells mimicked previously observed topo II␣ expression patterns (12,14). Furthermore, the 2-3-fold S phase-specific induction levels observed were comparable with previously reported levels (12)(13)(14). Interestingly, transcriptional activation could be blocked in a reversible manner by treatment with the DNA synthesis inhibitor aphidicolin or the CDK inhibitor roscovitine. In contrast, aphidicolin does not interfere with the timing of changes in topo II␣ mRNA stability in HeLa cells (12). A similar discrepancy is also observed when cells are heat shocked: topo II␣ mRNA stability decreases upon heat shock, whereas the topo II␣ promoter FIG. 5. ICB1 binding activity is cell cycle-dependent. A, radioactive labeled ICB1 probe was incubated with (lanes 2-9) or without (lane 1) 15 g of nuclear extract from synchronized (lanes 3-9) or exponentially growing (lane 2) NIH/3T3 cells. Following serum starvation, serum was added (t ϭ 0 h) and nuclear extracts were prepared every 4 h for a period of 24 h. Probe/protein complexes were analyzed by EMSA and the specific ICB1-binding protein complex is indicated by an arrow. B, radioactive labeled ICB1 probe was incubated with (lanes 2-4) or without (lane 1) 15 g of nuclear extract from exponentially growing NIH/3T3 cells. In addition, unlabeled competitor probes were added in excess as indicated. Probe-protein complexes were analyzed by EMSA, and the specific ICB1-binding protein complex is indicated by an arrow. The asterisk indicates nonspecific probe-binding protein complexes. is activated (12,17). Supposedly, these differences reflect different regulatory mechanisms responsible for controlling topo II␣ gene expression at the transcriptional and posttranscriptional levels, respectively.
Experiments with synchronized NIH/3T3 stable transfectants indicated that an ICB situated between bp Ϫ90 and Ϫ51 in the topo II␣ promoter (ICB1) played an important role in the S phase-specific induction of the promoter. Specifically, loss of ICB1 function resulted in abrogation of the S phasespecific induction by prematurely inducing the promoter during the G 1 phase of the cell cycle. In addition, EMSA experiments demonstrated a cell cycle-dependent ICB1 binding activity in NIH/3T3 nuclear extracts. Binding to ICB1 was most pronounced in serum-starved cells, and as cells entered S phase, ICB1 binding was markedly decreased. Together, these data suggest a role for ICB1 in transcriptional repression of the topo II␣ gene.
The presence of an ICB1 binding activity has been described in several studies (17,25); however, none of these have addressed its occupation as a function of growth state. Nevertheless, ICB1 has been proposed to mediate repression of topo II␣ transcription. In vivo footprint analysis of the human topo II␣ promoter revealed that promoter activation by heat shock was accompanied by reduced protein binding to ICB1 (17). Interestingly, ICB1 has previously been associated with p53-mediated repression of the topo II␣ promoter (18). The CDK inhibitor p21 Waf1/Cip1 is transcriptionally activated by p53 in response to DNA damage, which among other things leads to inhibition of DNA synthesis (reviewed in Ref. 26). Our data suggest that transcriptional activation of the topo II␣ gene can be suppressed by inhibition of ongoing DNA synthesis and CDK activity. Thus, p53-mediated repression of the topo II␣ promoter may be an indirect effect of DNA synthesis and CDK inhibition as a consequence of p21 Waf1/Cip1 activation. This is supported by the fact that the topo II␣ promoter does not contain a bona fide p53 binding site.
ICBs are predominantly bound by the transcriptional activator NF-Y (27,28). However, evidence has accumulated for a more complex role for ICBs, other than mediation of transcriptional activation (reviewed in Ref. 29). For instance, occupation of ICBs by the transcriptional repressor CDP/cut has been reported in several cases (30 -32). Furthermore, the binding specificity of CDP/cut for individual ICBs is determined by their surrounding sequences, and especially GCboxes seem to enhance the ICB-specific binding by CDP/cut (33). There is some support for a role of CDP/cut in the transcriptional regulation of the topo II␣ promoter. First, ICB1 can mediate repression and is situated next to a GC-box (37). Second, CDK2 and/or CDC2 activity seems to be important for transcriptional activation of the topo II␣ promoter, and both CDK2 and CDC2 have been shown to be present in a high molecular weight protein complex containing CDP/cut (32,34). Alternatively, recent reports indicate that NF-Y activity also can be modified via interactions with other factors (35,36). If this is the case in terms of topo II␣ promoter regulation, such interactions would have to affect the DNA binding activity of NF-Y, since protein levels of the A subunit of NF-Y are unaffected by the growth state of NIH/3T3 cells (data not shown). Although such an alteration of NF-Y DNA binding activity has been proposed to mediate the downregulation of the topo II␣ promoter in confluence-arrested cells (16), this is probably not the case in the cell cycle-dependent repression of the topo II␣ promoter reported here, as an antibody against NF-Y did not produce a supershift of the ICB1-probe-protein complex (data not shown).
In summary, we have shown that the transcriptional induction of the topo II␣ gene requires S phase entry, ongoing DNA synthesis, and CDK2 and/or CDC2 activity. Furthermore, we have demonstrated that ICB1 plays an important repressive role in the transcriptional regulation of topo II␣. The challenge now is to identify an ICB1-binding protein complex and deliniate the biochemical changes responsible for its differential occupation of ICB1.