Distinct Recruitment of E2F Family Members to Specific E2F-Binding

Home , E2F1, E2F4, G0 phase

Distinct recruitment of E2F family members to speciﬁc E2F-binding sites mediates activation and repression of the E2F1 promoter

Keigo Araki1, Yusuke Nakajima2, Kazuhiro Eto1 and Masa-Aki Ikeda*,1

1Section of Molecular Craniofacial Embryology, Graduate School, Tokyo Medical and Dental University, 1-5-45, Yushima, Bunkyo-ku, Tokyo 113-8549, Japan; 2Section of Oral Surgery, Graduate School, Tokyo Medical and Dental University, 1-5-45, Yushima, Bunkyo-ku, Tokyo 113-8549, Japan

The activity of E2F transcription factors plays a crucial 1991)and thymidine kinase (TK)(Dou et al., 1992)). role in mammalian cell-cycle progression and is controlled The six members of the E2F family (E2F1-6)that have by physical association with the pocket proteins (pRb and been identified thus far bind to DNA by heterodimeriz- its related p107 and p130). The E2F1 promoter, which ing with DP family members, DP1 and DP2, and the contains two overlapping E2F-binding sites, is activated at activity of E2F1-5 is regulated by physical interactions the G1/S transition in an E2F-dependent manner. with the pocket proteins (pRb and the pRb-related Mutational experiments have shown that the distal E2F- proteins p107 and p130). The E2F members can be binding site on the E2F1 promoter is required for divided into three distinct subgroups based on their transcriptional repression in the G0 phase, whereas the function and ability to bind pocket proteins. The proximal E2F-binding site contributes to transcriptional members of the first subgroup, E2F1, E2F2 and E2F3, activation at the G1/S boundary. Consistent with these are potent transcriptional activators and bind almost results, chromatin immunoprecipitation assays have re- exclusively to pRb, whereas the members of the second vealed that the E2F4/p130 repressor complex specifically subgroup, E2F4 and E2F5, contribute to ‘active binds to the distal E2F-binding site, whereas E2F1and repression’ of E2F-responsive genes and bind preferen- E2F3 activators preferentially bind to the proximal E2F- tially to p107 and p130 (Dyson, 1998; Harbour and binding site. The assays also showed that the specific Dean, 2000; Trimarchi and Lees, 2002). E2F6, the sole binding of E2F4/p130 complex to the distal site was member of the third subgroup, acts as a transcriptional dramatically impaired by a mutation introduced into the repressor in a pocket protein-independent manner, contiguous repression site (cell Cycle gene Homology because it lacks the pocket protein-binding domain. Region; CHR). Taken together, these findings indicate The precise characteristics of E2F6, however, remain that the two E2F-binding sites play distinct roles in the unclear (Cartwright et al., 1998; Gaubatz et al., 1998; regulation of E2F1 transcription by interacting with Ogawa et al., 2002). different sets of E2F members and cooperating with the There are a number of findings demonstrating the contiguous repressor element. diverse transcriptional regulation of E2F-responsive Oncogene (2003) 22, 7632–7641. doi:10.1038/sj.onc.1206840 genes. First, the expression patterns of E2F-responsive genes during the cell cycle are not simple. The promoter Keywords: E2F1; cell cycle; transcription; E2F; core- activities of E2F1 and B-myb are upregulated at the pressor G1/S boundary (Lam and Watson, 1993; Hsiao et al., 1994; Johnson et al., 1994; Neuman et al., 1994), whereas those of cyclin A and cdc2 are upregulated in the late S-G2 phase (Dalton, 1992; Schulze et al., 1995; Tommasi and Pfeifer, 1995; Liu et al., 1998)and that of Introduction Rb is almost unchanged throughout the cell cycle (Shan et al., 1994). Second, mutational studies have shown The E2F family of transcriptional factors plays a crucial that mutations introduced into E2F-binding sites in the role in the mammalian cell cycle by controlling the B-myb and cdc6 promoters cause derepression in G0 expression of genes that mediate cell-cycle progression (Lam and Watson, 1993; Bennett et al., 1996; Hateboer (e. g., cyclin E (Ohtani et al., 1995; Botz et al., 1996)and et al., 1998; Ohtani et al., 1998), whereas those in the cdc2 (Dalton, 1992; Tommasi and Pfeifer, 1995)) and dihydrofolate reductase (DHFR)and TK promoters DNA synthesis (e.g., DNA polymerase a (Pearson et al., result in the absence of transactivation during cell-cycle progression (Lin et al., 1996; Karlseder et al., 1996; Rotheneder et al., 1999). Third, studies using mouse *Correspondence: M-A Ikeda, Section of Molecular Embryology, embryonic fibroblast (MEFs)deficient in pRb or p107/ Graduate School, Tokyo Medical and Dental University, 1-5-45, p130 have demonstrated that the transcriptional activ- Yushima, Bunkyo-ku, Tokyo 113-8549, Japan; E-mail: [email protected] ities of the B-myb, cdc2, E2F1, and cyclin A2 promoters Received 17 February 2003; revised 30 May 2003; accepted 2 June 2003 are derepressed in MEFs lacking p107/p130, but not in Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7633 pRb-deficient MEFs. On the other hand, derepression of the presence and copy number of these integrated the cyclin E and p107 promoters occurs in MEFs plasmids by Southern blot analyses (Figure 1c). Lucifer- lacking pRb alone, suggesting that the expression of ase and b-galactosidase DNA fragments were hybri- these E2F-responsive genes is regulated by different sets dized to genomic DNA isolated from cell lines of E2F-pocket protein complexes (Hurford et al., 1997). transfected with plasmids, but not to genomic DNA It is currently uncertain how individual E2F members isolated from the untransfected REF52 cell line. The recognize a particular E2F-binding site in each stage of quantitation of integrated plasmids normalized by the the cell cycle. One possibility is that the DNA binding intensity of endogenous GAPDH (data not shown) specificity of E2F members is influenced by other indicated that these cells harbored on average 16.770.8 transcriptional regulatory factors that bind to sites copies of the reporter plasmids (Figure 1c, top)and contiguous to the E2F-binding sites (e.g., Sp1 sites in the 3.170.3 copies of pCMV b-gal (Figure 1c, bottom). TK and DHFR promoters (Lin et al., 1996; Karlseder Furthermore, PCR analysis using pairs of primers et al., 1996; Rotheneder et al., 1999), cell-cycle gene designed to detect the exogenous E2F1 promoter homology region (CHR)/downstream repressive site (Figure 1e)indicated that the transfected E2F1 reporter (DRS)in the B- myb promoter (Catchpole et al., 2002), is integrated into chromatin, along with cotransfected and the YY1-binding site in the cdc6 promoter (Schlisio expression vectors driven by the CMV promoter et al., 2002)). These factors have been deduced to (Figure 1d). stabilize the interactions of specific free E2F members or E2F-pocket protein complexes with DNA and act Transcriptional activity of the wild-type and various synergistically in the activation or repression of tran- mutant E2F1 promoters during the cell cycle scription. To elucidate the role of E2F-binding sites in regulat- To examine the relative contribution of the E2F-binding ing the expression of E2F-responsive genes, in this sites to cell-cycle-dependent E2F1 transcription, the study, we analysed the transcriptional regulation of the activity of the integrated E2F1 reporters was assayed E2F1 promoter, which contains two overlapping E2F- during the cell cycle. First, we investigated the cell-cycle binding sites and provide a model for the specific roles profile of these stably transfected cell lines and of the E2F-binding sites in transcriptional regulation of confirmed that the REF52 cells that had integrated the E2F1 promoter. E2F1-Luc WT were efficiently induced into quiescence by serum starvation and synchronously re-entered the cell cycle after serum restimulation (Figure 2a). Con- et al Results sistent with previous studies (Johnson ., 1994; Smith et al., 1996; Lindeman et al., 1997; Leone et al., 1998), Establishment of REF52 cell lines with stably integrated Western blotting analysis of REF52 cells with stably E2F1 promoter reporter plasmids integrated E2F1-Luc WT showed that E2F4 and p130, components of the G0 repressor complex, were localized To investigate the role of E2F-binding sites in control- to the nucleus in quiescent cells. In late G1 (16 h), p130 ling the expression of E2F-responsive genes, we analysed was hyperphosphorylated, the majority of E2F4 was the transcriptional activity of the E2F1 promoter, which localized to the cytoplasm and the expression of E2F1 includes two overlapping E2F-binding sites and is and E2F3 was induced (Figure 2b). These results activated at the G1/S transition in an E2F-dependent indicated that these cell lines retained orderly protein manner, as a model of E2F-responsive genes. Mutations expression and cell-cycle progression similar to those of were introduced into both or either of these sites in the parental REF52 cell line. For the cell-cycle- reporter plasmids carrying the luciferase gene under the dependent transcription assays, REF52 cells that had human E2F1 promoter containing the region 242 bp integrated reporter plasmids were serum starved and upstream of the transcription initiation site that retains harvested every 4 h after serum restimulation (Figure 2c). elements required for cell-cycle-regulated E2F1 tran- Consistent with previous studies, the activity of wild- scription (Figure 1a,b)(Johnson et al. 1994). Hence- type E2F1 promoter was low in G0 (0 h)and increased forth, we refer to the distal side (position À31 to À20)of approximately sevenfold from late G1 to S phase (20 h). E2F-binding sites as ‘E2F site A’ and to the proximal The activity of the promoter mutated into both E2F side (position À14 to À3)as ‘E2F site B’. To examine sites A and B was fivefold greater than that of the wild- functional correlation between the role of the E2F- type promoter in G0, and only slightly elevated binding sites in controlling the E2F1 promoter activity throughout the cell cycle (Smith et al., 1996). To define and the interaction of E2F members with these sites, we their roles more precisely, we investigated the activity of first generated REF52 cell lines that integrate these the E2F1 promoter mutated into each of them. As reporter plasmids into chromatin. REF52 cells were shown in Figure 2c, the mutation in E2F site A transfected with these reporters along with a neomycin abolished transcriptional repression of the E2F1 pro- resistance expression vector and pCMV b-gal as an moter in G0, whereas the mutation in site B had no internal control, and selected with Geneticin. To significant effect on transcriptional repression, but minimize particular positional effects on integrated dramatically impaired G1/S transcriptional upregula- plasmids in chromatin, we combined a large number tion. We therefore concluded that these two E2F- of Geneticin-resistant colonies and examined them for binding sites play distinct roles in controlling cell-

Oncogene Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7634 cycle-dependent E2F1 transcription: the E2F site A Analysis of E2F binding to the E2F1 promoter in vitro represses transcriptional activity of the E2F1 promoter It is conceivable that the distinct roles of these two sites in G0, whereas the E2F site B confers upregulation on in controlling E2F1 transcription are attributable to the transcription at the G1/S boundary. different binding specificities of individual E2F members to particular sites in each stage of the cell cycle. We therefore carried out electrophoretic mobility shift assays (EMSAs)by incubating radio-labeled oligonucleotides containing E2F site A (Mut E2F site B)or B (Mut E2F site A)with nuclear extracts from serum- starved REF52 cells (Figure 3a). Controversially, the results of the assays indicated that E2F/p130 G0 complexes designated by antibody supershift assays with anti-p130 antibody and competition assays using excess amounts of unlabeled oligonucleotides can bind to the E2F site B, but binding to the E2F site A was hardly detected. To make these binding clear, probes were incubated with whole-cell extract from U2OS cells overexpressing E2F4 product (Figure 3b), which is the predominant E2F in E2F/p130 complexes (Vairo et al., 1995; Ikeda et al., 1996). Similar to the E2F/p130 complexes in Figure 3a, E2F4 showed overwhelming binding preference for the E2F site B. Next, we performed EMSAs to investigate the binding preference of the E2F1 product, which plays a positive role in the transcriptional control of E2F- responsive genes. The same as E2F4, all forms of E2F1 derived from the transfection of an E2F1 expression vector, that is, free E2F1 and the E2F1/endogenous Rb family complexes confirmed by antibody supershift assays as shown in Figure 3c left, also bound much more to the E2F site B (Figure 3c, right). Taken together, these findings indicated that E2F1 as well as E2F4 preferentially bind to the E2F site B in vitro.

Different promoter occupancy of the integrated mutant E2F1 promoters The conﬂicting results shown in Figure 3a and b prompted us to investigate E2F binding under physio-

3––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Figure 1 Establishment of REF52 cell lines with stably integrated E2F1 promoter reporter plasmids. (a)Schematic representation of the wild-type and mutants of E2F1 promoter reporter plasmids. Closed and open boxes show the consensus sequences of E2F-binding sites and CHR, respectively. The mutations are represented by cross symbols. Bent arrows indicate the transcription initiation site. (b)The various mutations introduced into E2F-binding sites and CHR are underlined. (c)Southern blotting analyses of integrated plasmids. Genomic DNA isolated from the indicated cell lines was probed with luciferase (top)or b-galactosidase (bottom). Densitometric values of the luciferase and b-galactosidase genes were normalized to those of endogenous GAPDH. Relative copy numbers of stably integrated genes in each cell line were estimated based on the values of control DNA plasmids (10 ng of E2F1-Luc was set equal to 2 and 7 pg of pCMV b-gal was set equal to 1). (d)Detection of integrated reporter gene by PCR. Genomic DNA (100 ng)isolated from untransfected and E2F1-Luc WT transfected REF52 cells was ampliﬁed by using E2F1 promoter-luciferase (top), CMV promoter (middle), or GAPDH (bottom)primers. For positive control, 250 fg of E2F1-Luc WT (top) or pCMV b-gal DNA (middle)was used. ( e)The positions of E2F1 promoter-luciferase primers are schematically shown by small arrows

Oncogene Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7635

Figure 2 Transcriptional activity of the wild-type and various mutant E2F1 promoters during the cell cycle. (a)LSC analysis of REF52 cell line with stably integrated E2F1-Luc WT. Cells were induced into quiescence by serum starvation (0 h), restimulated with serum, and harvested at the times indicated. Cell number is plotted against DNA content determined by propidium iodide intensity. (b) Distribution of E2Fs and p130 during the cell cycle. Expression and subcellular localization of E2Fs and p130 were analysed by Western blotting of nuclear (N)and cytoplasmic (C)cell extracts in G0 (0 h)and late G1 (16 h)with the antibodies indicated. ( c) Relative transcriptional activity of the wild-type and mutant E2F1 promoters. REF52 cell lines that had stably integrated the indicated E2F1-Luc reporters were serum starved and restimulated with serum. Cells were harvested at the times indicated and assayed for luciferase activity normalized to b-galactosidase activity. Data are shown as fold activation. The mean value of wild-type E2F1 promoter activity at 0 h was set equal to 1. Bars show the standard errors of three independent assays

logical conditions. In recent studies, chromatin immu- plasmids. Consistent with previous studies demonstrat- noprecipitation assays (ChIPs)have been used to ing endogenous E2F1 promoter occupancy in T98G reveal in vivo promoter occupancy in many mammalian cells (Takahashi et al., 2000), we detected robust genes, and thus far this approach has provided faithful binding of E2F4 and p130 to the integrated wild-type understanding of the E2F binding to promoters E2F1 promoter in G0 and of E2F1 and E2F3 in late G1 in vivo (Wells et al., 2000; Takahashi et al., 2000). (Figure 4a, top). The specificity of our experiments We therefore performed ChIPs with REF52 cell was confirmed, since we did not detect any binding to lines that had integrated E2F1 reporter plasmids the cointegrated CMV promoter, which does not using the pairs of primers specific to the exogenous show the response to E2Fs (Figure 4a, bottom). ChIPs

Oncogene Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7636 with cells that had integrated various mutant whereas E2F1 and E2F3 intimately bound to the reporters revealed that E2F4 and p130 exclusively E2F site B (Mut E2F site A), and slightly to the E2F bound to the E2F site A (Mut E2F site B)in G0, site A in late G1 (Figure 4b,c). From these results, we concluded that transcriptional repression of the E2F site A in G0 is mediated by the recruitment of E2F4/ p130 repressor complex to the site A, and transcriptional upregulation of the E2F site B in late G1 is conferred by the interaction of E2F1 and E2F3 activators with the site B. Mutations in both E2F sites A and B precluded E2F binding to the E2F1 promoter in G0 and late G1, indicating the E2F-binding sites-speciﬁc recruitment of these proteins to the E2F1 promoter (Figure 4d).

CHR contributes to transcriptional repression of E2F1 The question then arose as to how the E2F4/p130 complex recognizes the E2F site A specifically in G0, in spite of the great binding preference of the E2F/p130 complex to the E2F site B in vitro. One plausible answer currently proposed is that the binding specificity of E2F members is dependent on other factors binding to sites located contiguous to E2F-binding sites. We therefore focused on the CHR/DRS in the B-myb promoter, which acts as a corepressor element by stabilizing the interaction of E2F/p130 repressor complex with the E2F-binding site (Catchpole et al., 2002). The CHR-like sequence is located downstream of the E2F site A and the relative spacing between the E2F-binding site and this region is identical to that in the B-myb promoter (Figure 5a)(Zwicker and Muller, 1997). To investigate the transcriptional role of the putative CHR in transcriptional regulation of the E2F1 promoter, transient reporter assays were performed using reporters that were also mutated into the putative CHR as shown in Figure 1a (Figure 5b). In agreement with previous observations with the B-myb promoter, mutation in the putative CHR (Mut CHR)led to a fivefold elevation of the E2F1 promoter activity in G0, and it is noteworthy that the presence of the mutation in the E2F site A, in addition to the mutation in the putative CHR (Mut E2F site A and CHR), had no additional effect on the promoter activity in G0, suggesting that mutation in the putative CHR abrogates the E2F site A-mediated transcriptional repression. We therefore carried out ChIPs to elucidate the correlation between the CHR and E2F site A in transcriptional repression of E2F1 (Figure 4e). As shown in Figure 4c, E2F4/p130 strongly

3––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Figure 3 Analysis of E2F binding to the E2F1 promoter in vitro. Gel mobility shift assays were performed with 2 mg of nuclear cell extracts from serum-starved REF52 cells (a)or whole-cell extracts from U2OS cells transfected with either pCMV E2F4 and pCMV-HA-DP1 (b)or pCMV E2F1 and pCMV-HA-DP1 (c)(0.5 mg each). Extracts were incubated with probes in the absence (À)or presence of the antibodies indicated and a 100-fold excess of unlabeled wild-type or mutant type (Mut E2F sites A and B)oligonucleotides. The probes used are indicated at the bottom of lanes (a and c left)or represented as þ (b and c right). Whole-cell extracts from U2OS cells transfected with 1.0 mg of a control CMV vector were used in the asterisk lane (C left)

Oncogene Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7637

Figure 5 CHR contributes to transcriptional repression of E2F1. (a)Sequence comparison of E2F-binding sites and CHRs between the E2F1 and B-myb promoters. E2F-binding sites and CHRs are highlighted by bold characters. (b)Transcriptional activity of the E2F1 promoters mutated into E2F site A and/or the CHR during the cell cycle. REF52 cells were transiently transfected with 0.9 mg of E2F1-Luc reporters along with 0.1 mg of pCMV b-gal as an internal control. Cells were induced into quiescence by serum starvation, restimulated with serum, and then harvested at the times indicated. Luciferase activity is shown normalized to b- galactosidase activity. Experiments were repeated three times in duplicate and independently. The mean value of wild-type E2F1 promoter activity at 0 h was set equal to 1. Bars show the standard errors of three independent assays

Discussion Figure 4 Different promoter occupancy of the integrated mutant E2F1 promoters. Chromatin immunoprecipitations were per- In this study we analysed the transcriptional activity of formed on REF52 cells that had stably integrated the E2F1-Luc the E2F1 promoter and found that two E2F-binding reporters in G0 (0 h)and late G1 (16 h)with the antibodies sites play distinct roles in controlling E2F1 transcription indicated. The immunoprecipitates obtained were amplified by using the following pairs of primers: E2F1 promoter-luciferase (Figure 2c). Similar to the E2F1 promoter, a dual role of (a top, b, c, d and e)and CMV promoter (a bottom).Rabbit IgG E2F members in positive and negative control of was used for mock immunoprecipitation transcription has previously been documented in several promoters containing two E2F-binding sites (e.g., Htf9- a/RanBP1 (Di Fiore et al., 1999), cdc6 (Schlisio et al., 2002)), yet the in vivo occupancy of their particular E2F- bound to the E2F site A (Mut E2F site B), but these binding sites was not demonstrated. We carried out interactions were significantly impaired by mutation in ChIPs with cells containing exogenously integrated the CHR (Mut E2F site B and CHR). Taken together, reporter genes, and this approach provided powerful these findings demonstrate that the CHR is essential to information on in vivo protein–DNA interaction with- the recruitment of E2F/p130 repressor complexes to the out improper protein expression and cell-cycle progres- E2F site A and acts as a corepressor element for sion (Figure 4). Our findings point to different in vivo transcriptional repression of E2F1. occupancies of the particular E2F-binding sites in G0

Oncogene Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7638 binding to the CHR and to determine the effect of the mutation in the CHR on E2F–DNA interaction by in vitro binding assay (data not shown). The CHR was found in cyclin A, cdc2,andcdc25C promoters, which include a ‘cell-cycle-dependent element’ (CDE)and are upregulated at the S/G2 transition, suggesting that CDE-CHR elements contribute to S/G2 phase specific gene expressio(Zwicker et al., 1995). The cdc6 and cdc25A promoters, which contain two E2F-binding sites, also include the CHR, and their gene expression is induced at the G1/S transition (Ohtani et al., 1998; Hateboer et al., 1998; Vigo et al., 1999). Although the role of the CHR in E2F-dependent transcriptional regulation of the cdc6 and cdc25A promoters has not yet been elucidated, it is noteworthy that transcriptional repression of these promoters in G0 is mediated by one of two E2F-binding sites and that the relative spacing Figure 6 Model for in vivo occupancy of the E2F1 promoter between the repressive E2F-binding site and the CHR is during the cell cycle. In quiescent cells, the interaction of E2F4/ identical to the spacing in the E2F1 and B-myb p130 with the E2F site A, which appears to be stabilized by a CHR promoters (Hateboer et al., 1998; Vigo et al., 1999; binding factor, mediates transcriptional repression. In contrast, E2F1 and 3 preferentially bind to the E2F site B and lead to Iavarone and Massague, 1999; Wu et al., 2000; Schlisio transcriptional activation in late G1 et al., 2002). Based on these observations and our study, we propose that the diverse transcriptional regulation of E2F-responsive genes would be caused by interaction of specific E2F and Rb family members with other and late G1; the E2F4/p130 repressor complex exclu- transcriptional regulatory factors. This is strongly sively binds to the E2F site A and is responsible for the supported by undetectable in vivo binding of E2F4 to transcriptional repression in G0, whereas E2F1 and the E2F site B, in spite of its high binding affinity to this E2F3 activators preferentially bind to the E2F site B site in vitro. and contribute to the upregulation on transcription at We observed functional cooperation between the E2F the G1/S transition (Figure 6). It has been previously site A and the CHR in the E2F4/p130-mediated shown that induction of ectopic E2F1 overproduction is transcriptional repression in G0, and yet transcriptional sufficient to drive quiescent cells into S and the derepression caused by the mutation in the CHR was deregulated expression of E2F1 in normal cells results found throughout the cell cycle (Figure 5b). As shown in in apoptosis (Dyson, 1998; Harbour and Dean, 2000). Figure 2b, the bulk of p130 is hyperphosphorylated and These observations suggest that E2F1 is a critical cell- E2F4 is localized in the cytoplasm in late G1, proposing cycle regulator in the timing of the G1/S transition and the possibility that the CHR element acts as a the appropriate expression of E2F1 is important for cell transcriptional repressor in addition to in an E2F4/ survival. Taken together, we propose that the combina- p130-dependent manner. tion of positive and negative control through the two Transcriptional activity of the E2F1 promoter in late E2F-binding sites enables to confer the tight regulation G1 is increased approximately twofold by the mutation of E2F1 transcription. in the E2F site A (Figure 2c), suggesting that the E2F We have shown that E2F1 and E2F3 preferentially site A contributes to transcriptional repression in late bind to the E2F site B in late G1 (Figure 4b,c). This may G1 as well as G0. Previous studies have demonstrated reflect the results of the in vitro binding assay, that is, that E2F/p130 repressor complexes are primarily found that the binding affinity of E2F1 and E2F3 is dependent in G0, while E2F/pRb complexes, composed of pRb and on the DNA sequences of the E2F-binding sites. E2F1-3, are most evident in late G1 (Ikeda et al., 1996; Previous reports have described the physical interaction Dyson, 1998). As shown in Figure 4c and e, we detected between Sp1 and E2F1-3 and demonstrated their definite binding of E2F1 and E2F3 to the E2F site A cooperative DNA binding and transcriptional regula- in late G1. Thus, to investigate whether the E2F site tion of the TK promoter (Karlseder et al., 1996; A-mediated transcriptional repression in late G1 is Rotheneder et al., 1999). Since the E2F1 promoter attributable to the E2F/pRb pathway, we carried out contains several sites for other transcriptional factors, ChIPs with anti-pRb antibody (sc-49, Santa Cruz such as Sp1 sites, an alternative possibility is that the Biotechnology). Although we could not rule out the DNA binding of E2F1 and E2F3 is influenced by other possibility that this antibody is incapable of being used adjacent factors. in ChIPs, no association between pRb and the E2F1 Our results provide evidence that the CHR element promoter was detected in late G1 (data not shown), contributes to stabilizing the interaction of the E2F4/ which is supported by previous experiments analysing in p130 repressor complex with the E2F site A and is vivo binding to endogenous E2F1 promoter (Takahashi essential for the transcriptional repression of E2F1 in et al., 2000). While it is possible that the elevation of the G0. Nevertheless, we were unable to identify the factor E2F1 promoter activity caused by the mutation in E2F

Oncogene Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7639 site A is attributable to the exclusion of a functional E2F1-Luc reporter plasmids along with pCMV b-gal, as an repressor complex other than the pocket protein internal control, and pcDNA3-neo at a ratio of 10 : 3 : 1 and complexes, an alternative explanation is that the selected with 800 mg/ml of G418 for 2 weeks. The resulting promoter activity is affected by the acetylated state of G418-resistant colonies were pooled and expanded under nucleosomal histones that is closely related to the local selection. Southern blotting analyses were performed by using structure of chromatin DNA. A recent study has standard procedures. EcoRV–XbaI digested genomic DNA demonstrated that p130 and p107 are responsible for (5 mg)from the stable transfectants and the indicated amount the recruitment of a chromatin-modifying factor, of E2F1-Luc plasmid DNA were separated and transferred to histone deacetylase (HDAC1), to the E2F1 promoter, a positively charged nylon membrane (Roche Diagnostics). and loss of HDAC1 recruitment is well correlated with For b-galactosidase gene, DraI–EcoRV digested genomic elevation of histone acetylation and activation of the DNA (6 mg)and the indicated amount of pCMV b-gal were promoter (Rayman et al., 2002). Thus, evidence separated and transferred. The blots were probed with the presented here supports the notion that the exclusion EcoRV–XbaI fragment of the luciferase gene and DraI–EcoRV of p130 by the mutation in the E2F site A prevents the fragment of the b-galactosidase gene, respectively, and the recruitment of HDAC1 to the E2F1 promoter, causing bands were quantitated on a BIO imaging analyzer 2500 (Fujifilm). For sample loading controls, the blots were the subsequent elevation of histone acetylation which reprobed with a GAPDH cDNA. The densitometric values appears to facilitate the accessibility of transcriptional of the luciferase and b-galactosidase genes in each cell line were regulatory proteins and/or components of the basal normalized to those of endogenous GAPDH. transcriptional machinery to the promoter in G0. In this regard, an enhancement of the site A mutated promoter Laser scanning cytometry activity in late G1 might be due to acceleration of the formation of a functional transcriptional complex on the REF52 cells stably transfected with E2F1-Luc WT were plated promoter upon serum stimulation. It will be of interest on a coverslip and serum starved. Cells were harvested at the indicated times after serum stimulation. Cells were fixed with to make clear the roles of E2F site A and CHR element 75% ethanol and stained with buffer containing propidium in control of the transcriptional activity of the E2F1 iodide (50 mg/ml)and RNaseA (200 mg/ml). The cell-cycle promoter in late G1. profile was analysed with a Laser Scanning Cytometer (LSC101 Olympus).

Materials and methods Western blotting analysis Western blotting analysis was performed by using standard Plasmids procedures. Equal amounts (50 mg)of nuclear and cytoplasmic The human E2F1 promoter reporter plasmid (E2F1-Luc extracts prepared from REF52 cells that had been stably (–242)) WT was provided by Dr K Ohtani (Johnson et al., transfected with E2F1-Luc WT were separated and transferred 1994). Mutations were introduced into both or either of two to a polyvinylidene diﬂuoride membrane (Pall Gelman putative E2F recognition sites and/or the CHR on the E2F1 Laboratory). The blot was probed with primary antibodies promoter region of the E2F1-Luc (–242)WT by site-directed directed against E2F1 (sc-193), E2F3 (sc-878), E2F4 (sc-1082), mutagenesis using a U.S.E. Mutagenesis Kit (Pharmacia p130 (sc-317), or actin (sc-7210) (Santa Cruz Biotechnology). Biotech). Wild-type and mutants of the E2F1 reporter are schematically illustrated in Figure 1a. EMSAs EMSAs were performed as described previously (Ikeda et al., Cell culture 1996). Equal amounts of the extracts (2 mg)were incubated REF52 and U2OS were grown in Dulbecco’s modiﬁed Eagle’s with 0.4 ng of radio-labeled probes in the absence or presence medium (DMEM)containing 10% fetal calf serum (FCS).To of unlabeled oligonucleotide competitors (40 ng)and 100 ng of induce cells into quiescence, REF52 cells were washed three antibodies (E2F1, sc-193; E2F4, sc-1082; Rb, sc-50; p107, sc- times with PBS and then placed in DMEM containing 0.1% 318; p130, sc-317). The following double-stranded oligonu- FCS for 48–72 h. The cells were stimulated by replacement cleotides were used as probes and cold competitors: with DMEM containing 20% FCS to allow cell-cycle re-entry. E2F1 WT (50-CGTGGCTCTTTCGCGGCAAAAAGGAT TTGGCGCGTAAAAGT-30); E2F1 Mut E2F site A (50-CGT Transfection and luciferase assay GGCTCTTTCGATGCAAAAAGGATTTGGCGCGTAAAA GT-30); E2F1 Mut E2F site B (50-CGTGGCTCTTTCGC REF52 cells plated in 12-well plates (4 Â 104 cells/well)were GGCAAAAAGGATTTGGATCGTAAAAGT-30); E2F1 Mut transfected with E2F1-Luc reporter plasmids along with b- E2F site A and B (50-CGTGGCTCTTTCGATGCAAAAAG- galactosidase expression plasmid pCMV b-gal, as an internal GATTTGGATCGTAAAAGT-30). control, by using FuGENE6 transfection reagent (Roche Diagnostics). Luciferase activity was normalized to b-galacto- CHIPs sidase activity. All assays were performed at least three times in duplicate and independently. ChIPs were performed as described previously (Takahashi et al., 2000). Each cell line of REF52 cells cultured on f150 mm dishes (2 Â 107 cells)was crosslinked with formalde- Generation of stably transfected cell lines and southern blotting hyde solution and lysed. Nuclei were sonicated and separated analyses by CsCl step-gradient centrifugation. Crosslinked chromatins To generate REF52 cell lines stably transfected with E2F1-Luc were collected and precleared with ProteinA/Sepharose 4B reporter genes, cells were transfected with either one of the beads (Amersham Biosciences)previously blocked with

Oncogene Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7640 sonicated salmon sperm DNA (100 mg/ml)and BSA (100 mg/ was performed using the following pairs of primers: E2F1 ml). Precleared chromatin was incubated with 2 mg of each promoter-luciferase gene (50-AGGAACCGCCGCCGTTG antibody (E2F1, sc-193; E2F3, sc-878; E2F4, sc-1082; p130, sc- TTCCCGT-30 and 50-GGATAGAATGGCGCCGGGCCTTT 317)at 4 1C overnight. Normal rabbit IgG was used for the C-30); CMV promoter (50-TTCCCATAGTAACGCCAA- mock immunoprecipitation. Immunocomplexes were recov- TAGGGA-30 and 50-GCATCACCATGGTAATAGCGAT- ered with 30 ml of ProteinA/Sepharose 4B beads and washed GA-30); GAPDH (50-ACCACAGTCCATGCCATCAC-30 eight times with RIPA buffer (0.5 m LiCl, 50 mm HEPES (pH and 50-TCCACCACCCTGTTGCTGTA-30). PCR products 7.5), 1 mm EDTA, 1% NP40, 0.7% sodium deoxycholate) were separated by electrophoresis through an 8% polyacryla- containing proteinase inhibitors (complete mini, Roche Diag- mide gel and visualized by using a BIO imaging analyzer 2500 nostics)and twice with TE buffer. To reverse the crosslinks, (Fujifilm). the pellets were incubated in 100 ml of TE buffer containing 0.5% SDS, 10 mg RNaseA, and 10 mg Proteinase K at 551C for 3 h and then at 651C for 16 h. The resulting DNA was purified by phenol–chloroform and ethanol precipitation. Pellets were Acknowledgements This work was supported in part by the Grant-in-Aid for dissolved in 50 mlH2O and subjected to semiquantitative PCR. For input control, 1/200 volume of chromatin was amplified. General Scientific Research and Cancer Research from the Each experiment was repeated two to three times, and Ministry of Education, Science, Sports, and Culture of Japan representative data are shown. and Japan Society for the Promotion of Science and in part by ‘Ground Research for Space Utilization’ promoted by NASDA and Japan Space Forum. PCR reaction We thank Dr K Ohtani for pBS-GAPDH and for critically PCR amplifications were carried out in 10 ml with 0.25 U of reading the manuscript, members of SME for support and Taq DNA polymerase (Sigma), 0.5 mCi of [a-32P]dCTP, useful discussions, and Solachuddin JA. Ichwan for excellent 10 pmol of each primer set, and 4% of DMSO. Each reaction technical assistance.

References

Bennett JD, Farlie PG and Watson RJ. (1996). Oncogene, 13, Lin SY, Black AR, Kostic D, Pajovic S, Hoover CN 1073–1082. and Azizkhan JC. (1996). Mol. Cell. Biol., 16, Botz J, Zerfass-Thome K, Spitkovsky D, Delius H, Vogt B, 1668–1675. Eilers M, Hatzigeorgiou A and Jansen-Durr P. (1996). Mol. Lindeman GJ, Gaubatz S, Livingston DM and Cell. Biol., 16, 3401–3409. Ginsberg D. (1997). Proc. Natl. Acad. Sci. USA, 94, Cartwright P, Muller H, Wagener C, Holm K and Helin K. 5095–5100. (1998). Oncogene, 17, 611–623. Liu N, Lucibello FC, Engeland K and Muller R. (1998). Catchpole S, Tavner F, Le Cam L, Sardet C and Watson RJ. Oncogene, 16, 2957–2963. (2002). J. Biol. Chem., 277, 39015–39024. Neuman E, Flemington EK, Sellers WR and Kaelin WG. Dalton S. (1992). EMBO J., 11, 1797–1804. (1994). Mol. Cell. Biol., 14, 6607–6615. Di Fiore B, Guarguaglini G, Palena A, Kerkhoven RM, Ogawa H, Ishiguro K, Gaubatz S, Livingston DM and Bernards R and Lavia P. (1999). J. Biol. Chem., 274, Nakatani Y. (2002). Science, 296, 1132–1136. 10339–10348. Ohtani K, DeGregori J and Nevins JR. (1995). Proc. Natl. Dou QP, Markell PJ and Pardee AB. (1992). Proc. Natl. Acad. Acad. Sci. USA, 92, 12146–12150. Sci. USA, 89, 3256–3260. Ohtani K, Tsujimoto A, Ikeda MA and Nakamura M. (1998). Dyson N. (1998). Genes Dev., 12, 2245–2262. Oncogene, 17, 1777–1785. Gaubatz S, Wood JG and Livingston DM. (1998). Proc. Natl. Pearson BE, Nasheuer HP and Wang TS. (1991). Mol. Cell. Acad. Sci. USA, 95, 9190–9195. Biol., 11, 2081–2095. Harbour JW and Dean DC. (2000). Genes Dev., 14, Rayman JB, Takahashi Y, Indjeian VB, Dannenberg JH, 2393–2409. Catchpole S, Watson RJ, te Riele H and Dynlacht BD. Hateboer G, Wobst A, Petersen BO, Le Cam L, Vigo E, (2002). Genes Dev., 16, 933–947. Sardet C and Helin K. (1998). Mol. Cell. Biol., 18, Rotheneder H, Geymayer S and Haidweger E. (1999). J. Mol. 6679–6697. Biol., 293, 1005–1015. Hsiao KM, McMahon SL and Farnham PJ. (1994). Genes Schlisio S, Halperin T, Vidal M and Nevins JR. (2002). EMBO Dev., 8, 1526–1537. J., 21, 5775–5786. Hurford RK, Cobrinik D, Lee MH and Dyson N. (1997). Schulze A, Zerfass K, Spitkovsky D, Middendorp S, Berges J, Genes Dev., 11, 1447–1463. Helin K, Jansen-Durr P and Henglein B. (1995). Proc. Natl. Iavarone A and Massague J. (1999). Mol. Cell. Biol., 19, 916– Acad. Sci. USA, 92, 11264–11268. 922. Shan B, Chang CY, Jones D and Lee WH. (1994). Mol. Cell. Ikeda MA, Jakoi L and Nevins JR. (1996). Proc. Natl. Acad. Biol., 14, 299–309. Sci. USA, 93, 3215–3220. Smith EJ, Leone G, DeGregori J, Jakoi L and Nevins JR. Johnson DG, Ohtani K and Nevins JR. (1994). Genes Dev., 8, (1996). Mol. Cell. Biol., 16, 6965–6976. 1514–1525. Takahashi Y, Rayman JB and Dynlacht BD. (2000). Genes Karlseder J, Rotheneder H and Wintersberger E. (1996). Mol. Dev., 14, 804–816. Cell. Biol., 16, 1659–1667. Tommasi S and Pfeifer GP. (1995). Mol. Cell. Biol., 15, Lam EW and Watson RJ. (1993). EMBO J., 12, 2705–2713. 6901–6913. Leone G, DeGregori J, Van Z, Jakoi L, Ishida S, Williams RS Trimarchi JM and Lees JA. (2002). Nat. Rev. Mol. Cell Biol., and Nevins JR. (1998). Genes Dev., 12, 2120–2130. 3, 11–20.

Oncogene Distinct roles of E2F-binding sites in the E2F1 promoter K Araki et al 7641 Vairo G, Livingston DM and Ginsberg D. (1995). Genes Dev., Wu L, Goodwin EC, Naeger LK, Vigo E, Galaktionov K, 9, 869–881. Helin K and DiMaio D. (2000). Mol. Cell. Biol., 20, Vigo E, Muller H, Prosperini E, Hateboer G, Cartwright P, 7059–7067. Moroni MC and Helin K. (1999). Mol. Cell. Biol., 19, Zwicker J, Lucibello FC, Wolfraim LA, Gross C, 6379–6395. Truss M, Engeland K and Muller R. (1995). EMBO J., 14, Wells J, Boyd KE, Fry CJ, Bartley SM and Farnham PJ. 4514–4522. (2000). Mol. Cell. Biol., 20, 5797–5807. Zwicker J and Muller R. (1997). Trends. Genet., 13, 3–6.

Oncogene