Oncogene (1998) 17, 3463 ± 3469 ã 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Activation and repression of p21WAF1/CIP1 transcription by RB binding

Andrei L Gartel1, Eugene Goufman1, Sergei G Tevosian2, Heather Shih2, Amy S Yee2 and Angela L Tyner1

1Department of Molecular Genetics, University of Illinois at Chicago, 900 S. Ashland Ave., Chicago, Illinois 60607; 2Department of Biochemistry, Tufts University School of Medicine, 136 Harrison Ave., Boston, Massachusetts 02111, USA

The Cdk inhibitor p21WAF1/CIP1 is a negative regulator of As Caco-2 cells di€erentiate, expression of the the cell cycle, although its expression is induced by a initially increases, in parallel number of mitogens that promote cell proliferation. We with p21 (Gartel et al., 1996a). There are ®ve known have found that E2F1 and , transcription factors proteins which can heterodimerize with two DP that activate required for cell cycle progression, proteins to form complexes capable of transcriptionally are strong activators of the p21 promoter. In contrast, activating cellular genes encoding proteins required for HBP1 (HMG-box -1), a novel retinoblastoma cell cycle progression. Activity of these E2F proteins is protein-binding protein, can repress the p21 promoter negatively regulated by binding with the retinoblasto- and inhibit induction of p21 expression by E2F. Both ma protein (RB) or the RB related proteins p107 and E2Fs and HBP1 regulate p21 transcription through cis- p130 (for reviews, see Helin, 1998; Slansky and acting elements located between nucleotides 7119 to Farnham, 1996). Phosphorylation of RB by an active +16 of the p21 promoter and the DNA binding domains Cdk/cyclin complex results in release of E2F and of each of these proteins are required for activity. activation of genes required for cell cycle progression. Sequences between 7119 and 760 basepairs containing Disruption of the E2F1 in mice resulted in four Sp1 consensus elements and two noncanonical E2F increased proliferation of a number of cell types (Field binding sites are of major importance for E2F activation, et al., 1996; Yamasaki et al., 1996), highlighting a although E2F1 and E2F3 di€er in the extent of their possible role for E2F1 in the negative regulation of ability to activate expression when this segment is proliferation. These diverse studies underscore a deleted. The opposing e€ects of E2Fs and HBP1 on potential dual role for E2F1 in positive and negative p21 promoter activity suggest that interplay between control of proliferation. these factors may determine the level of p21 transcrip- Recently, a novel RB and p130 binding protein tion in vivo. termed HBP1 was cloned using a yeast two-hybrid screen (Tevosian et al., 1997). Like p21, HBP1, RB, Keywords: p21; E2F; HBP1; Sp1 and p130 are ubiquitously expressed, but their expression speci®cally increases during di€erentiation. HBP1 is a sequence-speci®c HMG box containing transcriptional repressor that can down-regulate Introduction transcription from the N- promoter. In cycling cells, the N-Myc promoter is activated by E2Fs p21WAF1/CIP1 (El-Deiry et al., 1993; Harper et al., 1993; (Mudryj et al., 1990). We have shown that E2F1- Jiang et al., 1994; Noda et al., 1994; Xiong et al., dependent activation can be counteracted by HBP1

1993), can bind and inhibit G1 cyclin/cyclin-dependent and have suggested that E2F1 and HBP1 may kinase (Cdk) complexes (Harper et al., 1993; Xiong et constitute positive and negative transcriptional signals al., 1993) leading to growth suppression (Lin et al., for the N-Myc promoter (Tevosian et al., 1997). 1996a). Exit from the cell cycle is a prerequisite for Additionally, HBP1 expression leads to cell cycle terminal di€erentiation, and p21 expression is induced arrest, a necessary pre-requisite for terminal differ- during terminal di€erentiation in a number of systems entiation. The expression pattern and functions suggest (Halevy et al., 1995; Jiang et al., 1994; Macleod et al., that HBP1 may play a role in general cell differentia- 1995; Parker et al., 1995; Steinman et al., 1994; for tion (Tevosian et al., 1997). reviews see Gartel et al., 1996b; Gartel and Tyner, A number of transcription factors, including 1998). In the adult small and large intestine, p21 (El-Deiry et al., 1993), AP2 (Zeng et al., 1997), expression is induced in epithelial cells as they migrate BRCA1 (Somasundaram et al., 1997), STATs (Chin et out of the crypts and begin to terminally di€erentiate al., 1996; Matsumura et al., 1997), C/EBP-a (Gartel et al., 1996a). p21 mRNA and protein levels (Timchenko et al., 1996), C/EBP-b (Chinery et al., also increase in Caco-2 cells, which were derived from 1997), Spl (Biggs et al., 1996; Datto et al., 1995; Li et a colon adenocarcinoma, during di€erentiation (Gartel al., 1998; Prowse et al., 1997) and Sp3 (Prowse et al., et al., 1996a). Caco-2 cells provide a useful in vitro 1997) have been shown to regulate transcription of the model system for studying the induction of p21 p21 gene. In this study, we have examined the ability transcription during di€erentiation. of E2F family members and HBP1 to regulate p21 transcription. Our results suggest that E2F1/E2F3 and HBP1 have opposite e€ects and may modulate Correspondence: AL Tyner positive and negative regulation of the p21 promo- Received 1 April 1998; revised 22 June 1998; accepted 22 June 1998 ter, respectively. Activation and repression of p21 transcription AL Gartel et al 3464 760, the reporter CAT102 was highly inducible by Results E2F1 and E2F3. The construct CATD7124/760 was 10-fold more inducible by E2F3 than by E2F1 (Figure Cis-acting elements located within 119 bp from the start 3b). This construct contains two additional Sp1 sites site of p21 transcription are required for basal and E2F located between 760 and 750 bp, which appear to be inducible transcription important for transactivation by E2F3, but not for To determine if E2F proteins can stimulate p21 E2F1. transcription, Caco-2 cells were transfected with di€erent E2F family expression constructs and The E2F1 DNA and Sp1 binding domains are both CAT2330 which contains 2330 bp of the p21 promoter required for high level induction of p21 promoter activity driving expression of a CAT reporter gene. Extracts were prepared and analysed for CAT activity. We To quantitatively estimate the degree of activation of found that both E2F1 and E2F3 are strong activators the minimal p21 promoter by E2F1 and E2F3, we of the p21 promoter, while did not have any cotransfected a series of di€erent concentrations of the e€ect on p21 transcription (Figure 1). E2F1 and E2F3 expression constructs with a constant To identify regions of the p21 promoter that are amount of the reporter plasmid CAT119. We found responsible for activation by E2F1 and E2F3, Caco-2 that E2F1 activates the minimal p21 promoter 15 ± 20- cells were transfected with p21 promoter-CAT con- fold (Figure 4a) and E2F3 activates this same construct structs containing varying lengths of promoter 10 ± 15 times (data not shown). The degree of sequence (Figure 2a) and the E2F1 and E2F3 activation is not signi®cantly dependent upon E2F1 expression constructs. All p21 constructs tested, and E2F3 concentrations. including the longest construct CAT2330 (72330 to To further investigate the mechanism of transactiva- +16 bp) and the shortest construct CAT119 (7119 to tion of the p21 promoter by E2F1, we used two E2F1 +16 bp), were activated by E2F1 (Figure 2b). E2F3 mutants: E132 with a point mutation in the DNA- was also capable of activating all p21 promoter binding domain (Shin et al., 1996), and d113 ± 120 constructs tested, and data with constructs having endpoints at 7595, 7296, 7133 and 7119 bp are shown in Figure 2b. From these experiments, we conclude that E2F1 and E2F3 responsive elements are located between positions 7119 and +16 bp of the p21 promoter. To further localize cis-acting elements within the p21 promoter that mediate activation by E2F1 and E2F3, we transfected Caco-2 cells with the p21 promoter- CAT reporter constructs CAT102 (7102 to +16 bp) and CATD7124/760 (7156 to +16 bp with a deletion of sequences between 7124 and 760 bp) and the E2F1 and E2F3 expression constructs. Basal transcription of the CAT102 and CATD7124/760 constructs was almost totally impaired, as can been seen when these two reporters are transfected with the empty expression vector (V) as a control (Figure 3b). This suggests that the two Sp1 sites located between 7102 and 7119 bp are important for regulating basal transcription. In contrast to the reporter CATD7124/

Figure 2 Cis-acting sequences between 7119 and +16 bp mediate induction of the p21 promoter by E2F1 and E2F3. (a) Di€erent promoter-deletion constructs of the human p21 promoter in the pJFCAT1 vector. (b) CAT assays were performed with extracts from Caco-2 cells transiently transfected Figure 1 E2F1 and E2F3 transactivate the p21 promoter in with the E2F1 expression construct or the vector control and p21 Caco-2 cells. CAT assays were performed with extracts from promoter-deletion constructs with endpoints extending from Caco-2 cells transfected with the empty vector or the E2F1, E2F3, 72230 to 7119 bp of the p21 gene. CAT assays were also and E2F4 expression constructs and the CAT 2330 p21 promoter performed with extracts from Caco-2 cells transiently transfected reporter construct. Transfections were performed in duplicate for with the E2F3 expression constructs and p21 promoter-deletion each experiment, and each experiment was repeated at least three constructs with endpoints extending from 7589 to 7119 bp. A times. A typical experiment is shown typical experiment is shown Activation and repression of p21 transcription AL Gartel et al 3465

Figure 3 Sequences between 7102 and 760 bp can mediate induction of the p21 promoter by E2F1 and E2F3. (a) p21 promoter deletion constructs with endpoints at 7102 bp (CAT102) and 7156 bp with a deletion of sequences between 7124 and 760 bp (CATD7124/760). (b) CAT assays with extracts from Caco-2 cells transfected with the E2F1 and E2F3 expression constructs and the CAT102 and CATD7124/760 Figure 4 E2F1 activates exogenous and endogenous p21 constructs. Sequences between 760 and +16 bp determine the transcription in Caco-2 cells and DNA-binding domain is di€erence between activation by E2F1 and E2F3. A typical CAT required for its activity. (a) Estimate of the degree of activation assay is shown. (c) Sequence of the p21 promoter with deletion of the p21 promoter by E2F1. CAT assays were performed with endpoints marked with arrows and consensus Sp1 binding sites in extracts from Caco-2 cells transfected with di€erent concentra- large bold font. Noncanonical E2F sites (a) and (b) as described tions of the E2F1 expression plasmid and the CAT119 construct. in Hiyama et al., 1998, are indicated by the smaller bold font The fold induction for each concentration of E2F1 is indicated at top with standard deviations shown in parentheses. (b) CAT assays were performed with extracts from Caco-2 cells transfected with di€erent E2F1 mutant expression constructs (E132 with a mutation in DNA-binding domain and d113-120 with a deletion containing a deletion in the Sp1-binding domain (Lin in the Sp1 binding domain) and the CAT119 reporter construct. et al., 1996b). We found that the DNA-binding mutant A typical CAT assay is shown. (c) RT ± PCR was performed with E132 lost its ability to transactivate the p21 promoter, total RNAs from Caco-2 cells mock transfected (7)or transfected with the E2F1 expression construct (+) using the and the Sp1-binding mutant was less active than wild human S14 ribosomal protein expression as an internal control. type E2F1 (Figure 4b). These data suggest that E2F1 The EtBr stained gel with the bands corresponding to the binds the p21 promoter and acts synergistically with di€erent transcripts is shown Sp1 to induce transcription. Basal transcription of the p21 gene in Caco-2 cells (Gartel et al., in preparation) and induction of the p21 promoter by E2F1 and E2F3 is mediated through the The novel RB family binding protein HBP1 represses region of the promoter that contains multiple Sp1 p21 transcription through cis-acting elements located binding sites. Since E2F1 DNA binding activity is between 7119 and +16 bp. required for activation, the E2F proteins may bind to the p21 promoter and directly activate transcription. Recently, Tevosian et al. (1997) demonstrated that the Hiyama and colleagues (Hiyama et al., 1998) recently novel RB and p130 binding protein HBP1 represses the demonstrated that recombinant E2F1 and DP can bind N-Myc promoter and represents a transcriptional to the sequence 5'-AAACGCGC-3' containing a partial signal that can directly oppose E2F1 activation of E2F binding site (underlined) found in the E2F(b) sites critical cell cycle genes (Tevosian et al., 1997). Because noted in Figure 3c. It is also possible that the E2F E2F1 strongly activated the p21 promoter, we asked if proteins may transactivate genes encoding other factors HBP1 could also repress expression of p21. The p21 responsible for p21 activation, such as Sp1. To address promoter was repressed by HBP1 expression. Intrigu- this possibility, we examined the levels of endogenous ingly, the p21 promoter sequences required for both Sp1 and p21 mRNA expressed in Caco-2 cells following E2F1 activation and HBP1 repression appear identical. transfection with the E2F1 expression construct and we To identify regions of the p21 promoter that are detected no increase in Sp1 mRNA levels, while p21 responsible for repression by HBP1, Caco-2 cells were levels increased approximately threefold (Figure 4c). transfected with constructs containing varying lengths Thus, transiently expressed E2F1 is capable of inducing of the p21 promoter (Figure 2a) and the wild type expression of the endogenous p21 gene, and does not HBP1 expression construct. All tested p21 promoter appear to transactivate the p21 promoter indirectly by constructs were repressed 2 ± 3 times by HBP1 (Figure activating the Sp1 gene. 5). This level of repression does not depend upon the Activation and repression of p21 transcription AL Gartel et al 3466

Figure 5 Cis-acting sequences between 7119 and +16 bp are responsible for p21 promoter repression by HBP1. CAT assays were performed with extracts from Caco-2 cells after transfection of the HBP1 expression construct with p21 promoter-reporter constructs with endpoints from 72230 to 7119 bp. A typical CAT assay is shown

eciency of transfection of the Caco-2 cells by the Figure 6 The HMG box is necessary, but not sucient for di€erent constructs. Thus, HBP1 represses the p21 repression of p21 promoter by HBP1. (a) Di€erent HBP1 expression constructs in the pEF-BOS vector. (b) CAT assay promoter through sequences between 7119 and with extracts from Caco-2 cells transiently transfected with E2F1, +16 bp of the p21 gene. di€erent HBP1 expression constructs and the CAT119 p21 promoter-reporter. A typical CAT assay is shown. (c) Quantita- tive estimate of fold repression of the p21 promoter by HBP1. HBP1 acts in opposition to E2F and its HMG box is CAT assays were performed with extracts from Caco-2 cells co- necessary, but not sucient for repression of the p21 transfected with 2 mg of E2F1 and di€erent concentrations of promoter HBP1 expression plasmids and the CAT119 construct. Standard deviations are shown in parentheses To test if HBP1 could act in opposition to E2F, we expressed di€erent HBP expression constructs (Figure 6a) with the minimal p21 promoter construct (CAT119) and the E2F1 expression construct. As Discussion shown in Figure 6b, wild type HBP1 repressed the p21 promoter in the presence of exogenous E2F1. Members of the RB family (RB, p107, p130) bind a To identify the domains of HBP1 required for variety of target proteins, including members of the repression, the activities of several mutant HBP1 E2F family (Slansky and Farnham, 1996) and HBP1 constructs (Tevosian et al., 1997; diagrammed in (Lavender et al., 1997; Tevosian et al., 1997). Release Figure 6a), were tested. The pmHMG construct of these target proteins is induced when RB family encodes full length HBP1 with a triple point mutation proteins are phosphorylated by active Cdk-cyclin resulting in disruption of its DNA binding activity, complexes. Ectopic expression of free E2F1 has been while the HMG box represents the DNA binding shown to result in the activation of a number of genes domain alone. HBP1 expressed from the pmL/IXC important for S-phase, and promotes cell cycle construct has point mutations in its LXCXE and progression (Johnson et al., 1993; Qin et al., 1994). IXCXE motifs, rendering it incapable of interacting An alternate role for E2F1 as an inhibitor of with RB. Repression of the p21 promoter by HBP1 proliferation has also been proposed (Sellers et al., required an intact HMG box DNA binding domain (pm 1995), as the RB-E2F1 complex itself can bind E2F cis- HMG) (Figure 6b). However, the HMG box alone acting elements in genes required for cell cycle (HMG box) did not have repressor activity. Mutation progression and act as a transcriptional repressor of the two RB-binding motifs (pm L/IXC) did not a€ect (Sellers et al., 1995; Weintraub et al., 1992, 1995). p21 repression, and suggested that repression of the p21 Disruption of E2F1 in mice results in hyperplasia and promoter was independent of RB in these cells (Figure tumor formation (Field et al., 1996; Yamasaki et al., 6b). Thus, HBP1 could repress E2F activated transcrip- 1996), further supporting a role for E2F as a negative tion of the p21 promoter and the HMG box of HBP1 is regulator of the cell cycle. E2F17/7 mice display visible necessary, but not sucient for this repression. enlargement of the thymus and lymph nodes within To quantitatively estimate the degree of HBP1 one month after birth, and increased thymocyte induced repression of the minimal p21 promoter, we proliferation in aging mice. These data imply that cotransfected increasing concentrations of HBP1 with E2F1 may perform a role in suppressing cell the reporter plasmid CAT119 and 2 mg of the E2F1 proliferation, perhaps via activation of p21. expression vector. We found that HBP1 may repress It has recently been suggested that induction of p21 the p21 promoter up to sevenfold, and the level of transcription by E2Fs may re¯ect a positive role for repression by HBP1 depends on the concentration of p21 in cell proliferation where it may be acting as a HBP1 (Figure 6c). Cdk/cyclin assembly factor (Hiyama et al., 1988). Activation and repression of p21 transcription AL Gartel et al 3467 Usually p21 acts as a negative regulator of cell-cycle Two E2F(b) sites (Hiyama et al., 1998) and four Sp1 progression (for reviews see Gartel et al., 1996b; Gartel sites are located between 7119 and 760 bp of the p21 and Tyner, 1998). However, induction of p21 gene (Figure 3c). As shown in Figure 3b, deletion of expression by mitogenic stimulation has been observed these sites greatly reduces E2F1 and E2F3 responsive- by several investigators (Mantel et al., 1996; Michieli et ness. Two Sp1 sites and one E2F(b) site that are al., 1994; Nourse et al., 1994). Thus far, the located between 761 bp and the start of transcription signi®cance of p21 induction following mitogenic do not appear to be necessary for activation of the p21 activation of the cell cycle is not well understood. promoter by E2F1, but they may be important for Proliferation of myeloid progenitor cells is compro- activation by E2F3. mised in mice lacking p21, suggesting a positive role Sp1 is a zinc-®nger DNA binding protein that has for p21 in the proliferation of these cells (Mantel et al., been shown to play an important role in the regulation 1996). Since E2Fs may induce cell cycle progression, it of p21 transcription. Sp1 mediates induction of the p21 is possible that induction of p21 by mitogenic stimuli is gene in response to phorbol ester and okadaic acid mediated through E2F. Experiments to test this (Biggs et al., 1996), butyrate (Nakano et al., 1997), hypothesis are currently in progress. lovastatin (Lee et al., 1998) and TGF-b (Li et al., We demonstrate that E2F1 and E2F3 can activate 1998), while the related factor Sp3 mediates induction expression of the Cdk inhibitor p21. E2F may induce of the p21 gene in response to calcium (Prowse et al., expression of Cdk inhibitors as a general means of 1997). It has been shown that Sp1 can bind to E2F1, controlling free E2F levels in the cell. In addition to , and E2F3, but not E2F4 and in vitro inducing p21 expression, ectopic E2F expression has also (Karlseder et al., 1996). Here we show that E2F1 and been shown to result in increased expression of a E2F3 can strongly transactivate the p21 promoter, and p16INK4A related transcript (Khleif et al., 1996). A it is possible that E2F1 and E2F3 induce p21 variety of data suggest a complex regulatory loop in transcription by interacting with Sp1 and directing which an E2F induces expression of Cdk inhibitors, transcription through distinct Sp1 binding sites. We which then inhibit Cdk activity and further activation of found that E2F1 DNA-binding activity is required for E2F target genes. p21 has been shown to inhibit E2F transactivation of the p21 promoter. In addition, activity and activation of E2F dependent promoters interaction of E2F1 with Sp1 can further increase including the DHFR, cdc2, E2F1 (Dimri et al., 1996) and levels of induction of the p21 promoter. Similarly, in cyclin A promoters (Spitkovsky et al., 1997). the context of the HSV TK promoter that is largely The novel RB and p130 binding protein HBP1 has dependent on the Sp1 sites, E2F1 activates transcrip- been shown to repress the N-Myc promoter and act in tion in the absence of canonical E2F sites through the opposition to E2F1 (Tevosian et al., 1997). Here, we Sp1 sites and requires the E2F1 N-terminal domain show that HBP1 acts in opposition to E2F and that supports Sp1 interaction (Lin et al., 1996b; Shin et represses basal and induced activity of the p21 al., 1996). Negative regulation of p21 transcription by promoter up to sevenfold. We found that the HBP1 HBP1 also requires the region of the promoter DNA binding domain is required, but not sucient for containing six Sp1 binding sites. The opposing repression of the p21 promoter. While the Rb binding activities of E2F1 and HBP1 suggest an intriguing motifs of HBP1 appear to play some role in the positive/negative transcriptional regulatory network repression of the N-Myc promoter, repression of the involving these RB binding proteins. p21 promoter in Caco-2 cells appears to be indepen- dent of RB binding. Overexpression of HBP1 may result in cellular transformation in certain cell lines (Lavender et al., 1997), which could be due in part to Materials and methods the negative regulation of growth suppressing genes such as p21. Plasmid constructs We have determined that both positive and negative The human p21WAF1/CIP1 promoter deletion-CAT reporter regulation of p21 transcription by E2F1/E2F3 and constructs in pJFCAT1 from CAT2330 to CAT296 (Figure HBP1 in Caco-2 cells requires only sequences extending 2a) were a gift from WS El-Deiry (Univ of Pennsylvania) to 7119 bp of the p21 gene. The sequence of the (El-Deiry et al., 1995). Plasmids CAT133, CAT119 (Figure human p21 promoter from 7180 to +20 bp is 2a), CAT102 and CATD7124/760 (Figure 3a) were a gift from J Biggs and A Kraft (Univ. of Colorado) (Biggs et al., diagrammed in Figure 3c. No sequence with a high 1996). These plasmids contain di€erent fragments of the degree of similarity to the known HBP1 binding site p21 promoter, with the longest fragment from 72330 to (TCAATGGG (Tevosian et al., 1997) was found in the +16 bp (CAT2330) and shortest from 7102 to +16 bp p21 promoter. In addition, no canonical E2F (CAT102). HBP expression constructs were created in the consensus sites (5'-TTTCGCGC-3') were found, pEF-BOS vector (Tevosian et al., 1997) (Figure 6a). E2F1 although a few partial sites (5'-CGCGC-3') were was cloned into pRc/CMV (Invitrogen). E2F1 and E2F3 noted. Recently Hiyama et al., also found that E2F1 were cloned into CMV expression vectors and were gifts could activate the p21 promoter, and these authors from J Lees (MIT) (Moberg et al., 1996). The E2F1 proposed that this occurs through binding of E2F1 to deletion mutants CMV-d113 ± 120 and E132 were a gift noncanonical E2F binding sites that they termed from J. Nevins (Duke Univ.) (Johnson et al., 1993). To monitor eciency of transfections, the lac Z expression E2F(a) and E2F(b) in the p21 promoter (Figure 3c). construct pCH110 (Pharmacia) was co-transfected. We found that deletion of the E2F(a) site (5'- CTCCGCGC-3', Hiyama et al., 1998) located between 7156 and 7149 bp of p21 promoter did not change Cell culture, transfections and CAT assays the degree of activation of the p21 promoter by E2F1 The Caco-2 human colon adenocarcinoma cell line was and E2F3 (Figure 2, compare CAT133 and CAT296). obtained from the American Type Culture Collection and Activation and repression of p21 transcription AL Gartel et al 3468 grown in DMEM/F12 with 20% fetal calf serum. Transient pair, and 1 min of extension/synthesis at 728C. For each transfections of undi€erentiated Caco-2 cells (2 days after combination of primers the kinetics of PCR ampli®cation plating) were performed using calcium phosphate precipi- was studied, the number of cycles corresponding to plateau tation with 2 ± 4 mg of reporter plasmid, 2 ± 15 mgof were determined, and PCR was performed at an expression vector plasmid and 3 mg of pCH110. Cells exponential range. were harvested 48 ± 72 h after transfection, cellular extracts were prepared and CAT assays were performed as described previously (Gorman et al., 1982). As a substrate Description Upstream primer Downstream primer Size for CAT assays we used 1-deoxy [14C]chloramphenicol (Amersham) instead of [14C]chloramphenicol. A CAT assay p21 ACTGTGATGC- ATGGTCTTCC- 233 bp using deoxychloramphenicol results in only one acetylated GCTAATGGC TCTGCTGTCC product, which makes it easier to quantify the results E2F1 TATGGTGGCA- TGCAGAGCAG- 220 bp GAGTCAGTGG ATGGTTATGG (Murray et al., 1991). The per cent of acetylation in each Sp1 CTTGCTACCTG- TGTGATGATA- 147 bp experiment was calculated using a Betascope 603 (Betagen). TCAACAGCG CCAAGCTGGC Variations in transfection eciency were corrected by S14 GGCAGACCGAG- CAGGTCCAGGG- 143 bp normalization for expression of b-galactosidase. All ATGAATCCTC GTCTTGGTCC transfections were repeated at least three times in duplicate. We analysed expression of the human S14 ribosomal protein as internal control (Foley et al., 1993). PCR products were separated on a 2% agarose gel and stained with RNA preparation and RT ± PCR ethidium bromide. Total RNA was prepared from Caco-2 cells by a single- step method using RNAzolTM B (TEL-TEST, INC). For synthesis of cDNA we used the SuperScriptTM Preamplifi- cation System (Gibco ± BRL). 2 mg of total cellular RNA Acknowledgements was annealed with 50 ng of random hexadeoxynucleotide We thank Wa®k El-Deiry, Joseph Biggs and Andrew Kraft primer at 708C for 10 min. cDNA was synthesized in 20 ml for p21 promoter-deletion constructs and Jacqueline Lees solution containing 20 mM Tris-HCl (pH 8.4), 50 mM KCl, and Joseph Nevins for E2F expression constructs. This 10 mM DTT, 500 mM dNTP and 200 units of SuperScript work was supported by NIH Grant DK48836 (ALT); II reverse transcriptase at 428C for 50 min, after which we award #97-01 from the American Cancer Society, Illinois added two units of RNAse H. 2.5 ml of cDNA product was Division (ALG); NIH Grant GM44634 (ASY); a core added to 30 ml of PCR reaction mixture that contained grant to (ASY) from the Digestive Disease Center at New

200 mM dNTP, 2.5 mM MgCl2, 75 pmol primers, 1.5 units England Medical Center (NIDDK, P30 DK-34928); and of AmpliTaq Polymerase (Gibco ± BRL). Each cycle of the NCI Oncology Research Faculty Development Pro- PCR included 45 s of denaturation at 948C, 1 min of gram (EG). ASY is an Established Investigator of the primer annealing at di€erent temperatures for each primer American Heart Association.

References

Biggs JR, Kudlow JE and Kraft AS. (1996). J. Biol. Chem., Gorman CM, Mo€at L and Howard B. (1982). Mol. Cell. 271, 901 ± 906. Biol., 2, 1044 ± 1051. Chin YE, Kitagawa M, Su W-CS, You Z-H, Iwamoto Y and Halevy O, Novitch BG, Spicer DB, Skapek SX, Rhee J, Fu X-Y. (1996). Science, 272, 719 ± 722. Hannon GJ, Beach D and Lassar AB. (1995). Science, 267, ChineryR,BrockmanJ,PeelerMO,ShyrY,BeauchampRD 1018 ± 1021. and Co€ey RJ. (1997). Nature Med., 3, 1233 ± 1241. Harper JW, Adami G, Wei N, Keyomarsi K and Elledge S. Datto MB, Yu Y and Wang X-F. (1995). J. Biol. Chem., 270, (1993). Cell, 75, 805 ± 816. 28623 ± 28628. Helin K. (1998). Curr.Opin.Genet.Dev.,8, 28 ± 35. DimriGP,NakanishiM,DesprezPY,SmithJRandCampisi Hiyama H, Iavarone A and Reeves SA. (1998). Oncogene, 16, J. (1996). Mol. Cell. Biol., 16, 2987 ± 2997. 1513 ± 1523. El-Deiry W, Tokino T, Velculescu V, Levy D, Parsons R, Jiang H, Lin J, Su Z, Collart F, Huberman E and Fisher P. Trent J, Lin D, Mercer W, Kinzler K and Vogelstein B. (1994). Oncogene, 9, 3397 ± 3406. (1993). Cell, 75, 817 ± 825. Johnson DG, Schwarz JK, Cress WD and Nevins JR. (1993). El-Deiry WS, Tokino T, Waldman T, Oliner JD, Velculscu Nature, 365, 349 ± 352. VE, Burrel M, Hill DE, Healy E, Rees JL, Hamilton SR, Karlseder J, Rotheneder H and Wintersberger E. (1996). Kinzler KW and Vogelstein B. (1995). Cancer Res., 55, Mol. Cell. Biol., 16, 1659 ± 1667. 2910 ± 2919. Khleif SN, DeGregori J, Yee CL, Otterson GA, Kaye FJ, Field SJ, Tsai FY, Kuo F, Zubiaga AM, Kaelin Jr W, Nevins JR and Howley PM. (1996). Proc. Natl. Acad. Sci. Livingston DM, Orkin SH and Greenberg ME. (1996). USA, 93, 4350 ± 4354. Cell, 85, 549 ± 561. Lavender P, Vandel L, Bannister AJ and Kouzarides T. Foley K, Leonard M and Engel J. (1993). Trends Genet., 9, (1997). Oncogene, 14, 2721 ± 2728. 380 ± 385. Lee S-J, Ha MJ, Lee J, Nguyen PM, Choi YH, Pirnia F, Gartel AL, Serfas MS, Gartel M, Goufman E, Wu GS, El- Kang W-K, Wang X-F, Kim S-J and Trepel JB. (1998). J. Deiry WS and Tyner AL. (1996a). Exp. Cell Res., 227, Biol. Chem., 273, 10618 ± 10623. 171 ± 181. Li J-M, Datto M, Shen X, Hu P, Yu Y and Wang X-F. Gartel AL, Serfas MS and Tyner AL. (1996b). Proc. Soc. (1998). Nucl. Acid Res., 26, 2449 ± 2456. Exp. Bio. Med., 213, 137 ± 148. Lin J, Reichner C, Wu X and Levine AJ. (1996a). Mol. Cell. Gartel AL and Tyner AL. (1998). In Progress in Molecular Biol., 16, 1786 ± 1793. and Subcellular Biology, Growth Inhibitors A Macieira- Lin SY, Black AR, Kostic D, Pajovic S, Hoover CN and Coelho, (Eds). Springer-Verlag: Heidelberg. vol. 22, in Azizkhan JC. (1996b). Mol. Cell. Biol., 16, 1668 ± 1675. press. Activation and repression of p21 transcription AL Gartel et al 3469 MacleodK,SherryN,HannonG,BeachD,TokinoT, Qin XQ, Livingston DM, Kaelin Jr W and Adams PD. Kinzler K, Vogelstein B and Jacks T. (1995). Genes & Dev., (1994). Proc. Natl. Acad. Sci. USA, 91, 10918 ± 10922. 9, 935 ± 944. Sellers WR, Rodgers JW and Kaelin Jr W. (1995). Proc. Natl. Mantel C, Luo Z, Can®eld J, Braun S, Deng C and Acad.Sci.USA,92, 11544 ± 11548. Broxmeyer HE. (1996). Blood, 88, 3710 ± 3719. Shin EK, Tevosian SG and Yee AS. (1996). J. Biol. Chem., Matsumura I, Ishikawa J, Nakajima K, Oritani K, 271, 12261 ± 12268. Tomiyama Y, Miyagawa J, Kato T, Miyazaki H, Slansky JE and Farnham PJ. (1996). Curr. Top. Microbiol. Matsuzawa Y and Kanakura Y. (1997). Mol. Cell. Biol., Immunol., 208, 1±30. 17, 2933 ± 2943. Somasundaram K, Zhang H, Zeng YX, Houvras Y, Peng Y, Michieli P, Chedid D, Lin D, Pierce J, Mercer W and Givol Zhang H, Wu GS, Licht JD, Weber BL and El-Deiry WS. D. (1994). Cancer Res., 54, 3391 ± 3395. (1997). Nature, 389, 187 ± 190. Moberg K, Starz MA and Lees JA. (1996). Mol. Cell. Biol., Spitkovsky D, Schulze A, Boye B and Jansen-Durr P. (1997). 16, 1436 ± 1449. Cell Growth Di€er., 8, 699 ± 710. Mudryj M, Hiebert SW and Nevins JR. (1990). EMBO J., 9, Steinman R, Ho€man B, Iro A, Guillouf C, Liebermann D 2179 ± 2184. and El-Houseini M. (1994). Oncogene, 9, 3389 ± 3396. Murray IA, Lewendon A, Williams JA, Cullis PM, Lashford Tevosian SG, Shih HH, Mendelson KG, Sheppard KA, AG and Shaw WV. (1991). Nucleic Acids Res., 19, 6648. Paulson KE and Yee AS. (1997). Genes Dev., 11, 383 ± 396. Nakano K, Mizuno T, Sowa Y, Orita T, Yoshino T, Timchenko NA, Wilde M, Nakanishi M, Smith JR and Okuyama Y, Fujita T, Ohtani-Fujita N, Matsukawa Y, Darlington GJ. (1996). Genes Dev., 10, 804 ± 815. Tokino T, Yamagishi H, Oka T, Nomura H and Sakai T. Weintraub SJ, Chow KN, Luo RX, Zhang SH, He S and (1997). J. Biol. Chem., 272, 22199 ± 22206. Dean DC. (1995). Nature, 375, 812 ± 815. Noda A, Ning Y, Venable SF, Pereira-Smith OM and Smith Weintraub SJ, Prater CA and Dean DC. (1992). Nature, 358, JR. (1994). Exp. Cell Res., 211, 90 ± 98. 259 ± 261. Nourse J, Firpo E, Flanagan WM, Coat S, Polyak K, Lee Xiong Y, Hannon G, Zhang H, Casso D, Kobayashi R and MH, Massague J, Crabtree GR and Roberts JM. (1994). Beach D. (1993). Nature, 366, 701 ± 704. Nature, 372, 570 ± 573. Yamasaki L, Jacks T, Bronson R, Goillot E, Harlow E and Parker S, Eichele G, Zhang P, Rawls A, Sands A, Bradley A, Dyson NJ. (1996). Cell, 85, 537 ± 548. Olson E, Harper J and Elledge S. (1995). Science, 267, Zeng YX, Somasundaram K and El-Deiry WS. (1997). Nat. 1024 ± 1027. Genet., 15, 78 ± 82. Prowse DM, Bolgan L, Molnar A and Dotto GP. (1997). J. Biol. Chem., 272, 1308 ± 1314.