Repression of transcription of the p27Kip1 cyclin-dependent kinase ...

ã

Oncogene (2001) 20, 1688 ± 1702 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Repression of transcription of the p27Kip1 cyclin-dependent kinase inhibitor gene by c-Myc William Yang1,3, Jian Shen1,3, Min Wu1, Marcello Arsura1, Mark FitzGerald1, Zalman Suldan2, Dong W Kim1, Claudia S Hofmann1, Stefania Pianetti1, RaphaeÈlle Romieu-Mourez1, Leonard P Freedman2 and Gail E Sonenshein*,1 1

Department of Biochemistry, Boston University Medical School, Boston, Maryland, MA 02118, USA; 2Cell Biology Program, Memorial Sloan-Kettering Cancer Center, New York, NY 10021, USA

Upon engagement of the B Cell Receptor (BCR) of WEHI 231 immature B cells, a drop in c-Myc expression is followed by activation of the cyclindependent kinase inhibitor (CKI) p27Kip1, which induces growth arrest and apoptosis. Here, we report inverse patterns of p27 and c-Myc protein expression follow BCR engagement. We present evidence demonstrating, for the ®rst time, that the p27Kip1 gene is a target of transcriptional repression by c-Myc. Speci®cally, the changes in p27 protein levels correlated with changes in p27 mRNA levels, and gene transcription. Induction of p27 promoter activity followed BCR engagement of WEHI 231 cells, and this induction could be repressed upon co-transfection of a c-Myc expression vector. Inhibition of the TATA-less p27 promoter by c-Myc was also observed in Jurkat T cells, vascular smooth muscle, and Hs578T breast cancer cells, extending the observation beyond immune cells. Consistent with a putative Inr element CCAGACC (where +1 is underlined) at the start site of transcription in the p27 promoter, deletion of Myc homology box II reduced the extent of repression. Furthermore, enhanced repression was observed upon transfection of the c-Myc `superrepressor', with mutation of Phe115 to Leu. The sequences mediating transcriptional activity and c-Myc repression were mapped to bp 720 to +20 of the p27 gene. Finally, binding of Max was shown to facilitate cMyc binding and repression of p27 promoter activity. Overall, these studies identify the p27 CKI gene as a new target whereby c-Myc can control cell proliferation, survival and neoplastic transformation. Oncogene (2001) 20, 1688 ± 1702. Keywords: c-Myc; Max; p27Kip1; WEHI 231 cells; smooth muscle cells; breast cancer

*Correspondence: GE Sonenshein 3 The ®rst two authors contributed equally to this work. Received 1 November 2000; revised 4 January 2001; accepted 9 January 2001

Introduction The c-Myc protein has been found to function in both positive and negative regulation of transactivation (rev. in Arsura and Sonenshein, 2001; Classen and Hann, 1999; Dang, 1999; Facchini and Penn, 1998; Grandori and Eisenman, 1997; Lemaitre et al., 1996). Heterodimers of c-Myc and its partner Max stimulate transcription via a consensus E box sequence (Grandori and Eisenman, 1997; Lemaitre et al., 1996). In addition, c-Myc plays a direct role in repression of transcription mediated by an initiator (Inr) element (Classen and Hann, 1999; Dang, 1999; Facchini and Penn, 1998). An Inr element, which facilitates formation of the initiation complex, has the consensus sequence: CCAGACC (where +1 is underlined) (Lo and Smale, 1996). Using cell-free transcription analyses, Roy and coworkers (1993) demonstrated that cMyc can repress Inr-mediated transactivation. In vivo inhibition of Inr-mediated transcription has been shown by several laboratories (rev. in Classen and Hann, 1999; Dang, 1999). Repression appears to involve AA 106 ± 143, including the Myc homology box II (MBII) (Li et al., 1994), encompassed within AA 129 ± 143 (Classen and Hann, 1999). Moreover, mutation of phenylalanine to leucine at position 115 of c-Myc, which has been found in multiple lymphomaderived c-myc genes (Hoang et al., 1995), increases transcriptional repression by c-Myc (Lee et al., 1996). In contrast, the ®rst 100 amino acids of c-Myc, which encompasses the transactivation domain, can be deleted without loss of its ability to repress (Xiao et al., 1998). The WEHI 231 immature B lymphoma cells express surface IgM, which serves as B Cell Receptor (BCR). As with normal immature B cells (Norvell et al., 1995), BCR engagement upon anti-IgM treatment causes Rb hypophosphorylation between 8 ± 12 h (Maheswaran et al., 1991), growth arrest in the late G1/S phase transition by 18 h (Warner et al., 1992), followed by apoptosis at 24 ± 48 h (Benhamou et al., 1990; Wu et al., 1996a). Thus, these cells have been commonly used as models for self-induced tolerance via clonal deletion (Boyd and Schrader, 1981). Work from our laboratory and others has implicated a drop in c-Myc in the

c-Myc represses p27 gene transcription W Yang et al

growth arrest and induction of apoptosis of WEHI 231 cells. A dramatic drop in c-myc mRNA and protein levels occurs by 6 ± 8 h following BCR engagement (Levine et al., 1986; Maheswaran et al., 1991; McCormack et al., 1984). Ectopic expression of cMyc promotes survival and continued proliferation of WEHI 231 cells (Wu et al., 1996a; Yi et al., 1996), and direct inhibition of c-myc expression with an antisense vector is sucient to induce apoptosis of these cells (Wu et al., 1999). A similar role for c-myc in promoting cell survival has been demonstrated in an increasing number of cell types under various treatment conditions, including pre-B, erthyroleukemia, immature T cells, leukemia, lymphoid, melanoma and breast cancer cells (Besancon et al., 1998; KiharaNegishi et al., 1998; Wang et al., 1999; rev. in Thompson, 1998). Ezhevsky et al. (1996) showed induction of the levels of p27Kip1 cyclin-dependent kinase (CDK) inhibitor (CKI) promotes anti-IgM-induced cell cycle arrest of WEHI 231 cells. The p27 CKI was originally characterized as a protein able to interrupt CDK2 kinase activity in TGF-b1-treated cell, and in a yeast two hybrid system as being able to interact with cyclin D and CDK4 (Polyak et al., 1994; Toyoshima and Hunter, 1994). The CKI p27 is now known to bind to and inhibit complexes formed by cyclin E-CDK2, cyclin A-CDK2, and cyclin D-CDK4, playing essential roles in transition through the G1 phase, in particular the restriction point (Sherr and Roberts, 1999). More recently, we demonstrated that anti-IgM treatment of WEHI 231 cells also results in p53-mediated induction of the CKI p21WAF1/CIP1 (Wu et al., 1998). Interestingly, we found that inhibition of p21 expression provided WEHI 231 cells partial protection from apoptosis, while inhibition of both p21 and p27 provided additional protection (Wu et al., 1998). Thus, we tested directly the role of p27 in apoptosis of WEHI 231 cells, and found that induction of p27 leads to death of WEHI 231 cells (Wu et al., 1999). Interestingly, we observed that apoptosis induced speci®cally upon inhibition of c-myc expression could be ablated by inhibition of p27 expression, and that a drop in cMyc levels was sucient to induce endogenous p27 levels (Wu et al., 1999). Thus, here we have explored the regulation of p27 expression and the possible role of c-Myc in this control. We report that in anti-IgMtreated WEHI 231 cells, p27 protein and mRNA and gene transcription levels increased following the drop in c-Myc expression with the overall patterns of p27 and c-Myc expression varying inversely. Transfection analysis con®rmed the presence of an Inr element at the start site of the TATA-less p27 promoter DNA sequence (Kwon et al., 1996), and the ability of c-Myc to repress p27 promoter activity in multiple cell types. Direct binding of c-Myc to the p27 promoter and functional repression of its activity was facilitated by Max. These ®ndings represent the ®rst report identifying the p27Kip1 CDK inhibitor gene as a target of repression by c-Myc, and suggest a novel mechanism involving Max.

1689

Results p27 and c-Myc protein expression vary inversely following anti-IgM treatment of WEHI 231 cells To measure the eects of anti-IgM treatment on p27 expression, proteins were extracted from WEHI 231 cells in exponential growth and following treatment for 2, 4, 6, 8, 10 or 24 h and subjected to immunoblot analysis. The level of p27 protein, which appeared to drop slightly at 2 h, increased above basal values by 8 h, and continued to increase further by 24 h (Figure 1a). As judged by densitometry, a 2.5- or 3-fold induction was seen at 24 h relative to the value at the 0 or 2 h time points, respectively. Thus, the induction of p27 precedes the appearance of hypophosphorylated Rb protein, which begins to be detected between 8 and 12 h (Maheswaran et al., 1991), consistent with its role in growth arrest of these cells (Ezhevsky et al., 1996). Interestingly, the changes in p27 expression appeared to vary inversely with the previously observed changes in c-Myc expression (Maheswaran et al., 1991). To con®rm this inverse correlation, the immunoblot was probed for c-Myc protein levels (Figure 1a). Indeed, an initial increase in c-Myc protein levels was followed by a dramatic decrease such that the levels dropped below control values between 6 to 8 h, consistent with our previous ®ndings. Of note, the decrease in c-Myc expression is followed immediately by an increase in p27 levels. Recently several groups have reported changes in RNA levels that account for altered p27 protein expression (Kamesaki et al., 1998; Kim et al., 1998 and see below). To determine whether a pre-translational site of control was involved in mediating the changes in p27 protein levels, mRNA expression was monitored. RNA was isolated at 0, 1, 2, 4, 6, 8, or 10 h following anti-IgM treatment and subjected to Northern blot analysis for p27 transcripts (Figure 1b). Changes in RNA levels commensurate with the ¯uctuations in protein expression were noted. Speci®cally an early drop in the amount of p27 mRNA was followed by an increase such that by 4 to 6 h, transcript levels were above baseline values, paralleling and preceding the increase in protein expression, as expected. A 2.3- or 4.3-fold induction was seen at 10 h relative to the value at the 0 or 2 h time point, respectively. Consistent with our previous ®ndings (Levine et al., 1986), an increase in c-myc mRNA levels seen between 1 and 2 h that followed by a dramatic decline such that between 4 to 6 h levels dropped below control values (Figure 1b). Ethidium bromide staining of the gel con®rmed equal loading and RNA integrity. Overall, these studies indicate that changes in mRNA levels can account for the initial decrease and later increase in p27 expression. To evaluate whether the changes in p27 mRNA levels were due to altered transcription, nuclei were isolated from WEHI 231 cells in exponential growth or following anti-IgM treatment for 1.5 or 5.5 h and subjected to nuclear run-on analysis (Figure 1c). An Oncogene


1690

Figure 1 Levels of p27 protein and mRNA and gene transcription correlate inversely with c-Myc expression during anti-IgMinduced apoptosis of WEHI 231 cells. (a) p27 protein. Total cellular proteins were isolated from exponentially growing WEHI 231 cells treated with anti-IgM for the indicated times (in h). Samples (50 mg) were subjected to immunoblotting for either p27 or c-Myc expression. Detection was achieved using a horseradish peroxidase conjugated secondary antibody and chemiluminescence, as previously described (Wu et al., 1996a). (b) RNA levels. Cytoplasmic RNA was isolated from control or anti-IgM treated WEHI 231 cells, and samples (20 mg) subjected to Northern blotting for p27 (pGlu-p27H, gift from Dr Y Xiong) or for c-myc (Stanton et al., 1983). Bottom panel shows ethidium bromide staining of the 28S and 18S rRNA on the Northern blot, which indicates equal loading and RNA integrity. (c) Gene transcription. Nuclei were isolated from WEHI 231 cells in exponential growth or following treatment with anti-IgM for 1.5 and 5.5 h and subjected to nuclear run-on analysis. Radiolabeled transcripts (2 ml at 56106/ml) were used as probe with the DNAs (10 mg/slot) for the p27, b-actin and p21 genes and pUC19, as a control for non-speci®c hybridization. (d). Quantitation of gene transcription analysis. Data from the nuclear run-on were subjected to densitometry twice and the results presented normalized for b-actin+the standard deviation (s.d.)

early drop in the rate of p27 gene transcription was noted by 1.5 h post anti-IgM treatment. This decline appeared selective in that transcription of a second member of the CKI family, p21, increased as a result of anti-IgM treatment. Transcription of the actin gene, while somewhat low, appeared relatively unaected. By 5.5 h of anti-IgM treatment, the level of p27 gene transcription had increased. As seen in Figure 1d, densitometry indicated p27 gene transcription normalized for b-actin declined initially from 4.6+0.08 vs 0.7+0.03 after 1.5 h and then increased to 9.4+2.0. In contrast, the rate of p21 gene transcription increased following anti-IgM treatment, consistent with the p53mediated increase in p21 mRNA levels noted previously (Wu et al., 1998). The normalized value increased from 14.2+0.2 in exponentially growing cells to 35.3+0.6 at 1.5 h, and then dropped to 22.2+0.9 at 5.5 h. Thus for p27 expression, the drop and return in mRNA levels appear to relate to an initial decrease Oncogene

followed by an increase in the rate of transcription of this gene. BCR-mediated induction of p27 promoter activity in WEHI 231 cells is ablated by c-Myc We next sought to determine whether anti-IgM treatment of WEHI 231 cells causes an increase in p27 promoter activity comparable to the increase in mRNA and gene transcription levels. The pl34 ± 11 p27 promoter construct, which contains 72002 to +154 bp of the p27 promoter linked to the pGL-2 basic luciferase reporter vector, was used in transfection analysis. To avoid the initial burst of c-Myc expression, cells were pre-treated for 4 h with antiIgM. Thus, cultures of WEHI 231 cells in exponential growth or following the 4 h pre-treatment with antiIgM were transfected, in duplicate, with the pl34 ± 11 vector in the presence of pSV40-b-gal to normalize for


transfection eciency. Cells were harvested after an additional 6 h of incubation and analysed (Figure 2a). Anti-IgM treatment caused a 3.9-fold increase in p27 promoter activity compared to untreated cells. BCR engagement did not alter the transfection eciency over this initial 6 ± 8 h period (data not shown); in three additional experiments an average 3.7-fold increase was noted upon normalization to protein levels. These ®ndings are consistent with the changes in p27 mRNA and gene transcription levels. Analysis of the sequence of the TATA-less p27 promoter DNA sequence (Kwon et al., 1996) revealed a putative Inr element at the start site: CCAGACC (where +1 is underlined) (Lo and Smale, 1996). Since c-Myc has been shown to repress transcription of genes containing Inr elements (rev. in Classen and Hann, 1999; Dang, 1999), we assessed the ability of c-Myc expression to repress the induction of p27 promoter activity following anti-IgM treatment (Figure 2b). Cells were either left untreated or treated with anti-IgM for 6 h prior to transfection. Two vectors were employed that express wild type c-Myc2 protein, pRc-CMV-Myc

(Hann et al., 1994), and pSV-Hu-myc-2 (Land et al., 1986); both of these vectors have been shown to activate transcription of an EMS-driven reporter construct (Arsura et al., 1995; Xiao et al., 1998). Cells were harvested after 13.5 h. A 3.5-fold increase was seen upon anti-IgM induction, consistent with the results presented above, and with our previous work showing induction an Adenoviral Inr-driven construct in WEHI 231 cells following BCR engagement (FitzGerald et al., 1999). Co-transfection of either cMyc expression vector almost completely abrogated the normal increase in p27 promoter activity of the pl34 ± 11 construct in WEHI 231 cells treated with anti-IgM (Figure 2b). Upon ectopic c-Myc expression, the activity of the p27 promoter remained essentially comparable to the uninduced values. Thus, expression of c-Myc eectively represses the anti-IgM-mediated induction of p27 promoter activity. Taken together, the above ®ndings suggest that the drop in c-Myc during receptor-mediated apoptosis led to de-repression of the p27 promoter, permitting activation of transcription of this gene.

1691

Figure 2 Induction of the p27 promoter following anti-IgM treatment of WEHI 231 cells is ablated by ectopic c-Myc expression. (a) The p27 promoter is induced following anti-IgM treatment. Exponentially growing WEHI 231 cells were left untreated or pretreated with anti-IgM for 4 h. Both cell populations were electroporated, in duplicate, with 15 mg pl34 ± 11 in the presence of 5 ug SV40b-gal with 10 mg empty vector DNA (total of 30 mg). Cells were harvested after an additional 6 h and extracts normalized for equal amounts of b-gal activity, were assayed for luciferase activity. Averages values are presented+s.d. (b) c-Myc represses the induction of p27 promoter. Exponentially growing WEHI 231 cells were left untreated (no treatment) or pretreated with anti-IgM for 6 h. Cells were electroporated, in triplicate, with 30 mg pl34 ± 11 in the absence (no Myc) or the presence of either 5 mg pCMV-cMyc or pSV-Hu-myc-2 expression vector. Anti-IgM was added to the pretreated cells immediately following transfection. Cells were harvested 7.5 h after transfection (13.5 h anti-IgM) and extracts, normalized for equal amounts of protein, were assayed for luciferase activity; average values are presented+s.d. Oncogene


1692

Wild type but not MBII mutant c-Myc represses the p27 promoter in Jurkat T cells We next sought to determine whether c-Myc repression of p27 promoter activity involved the Myc homology box II (MBII) domain, as has been seen for Inrmediated repression of other genes by c-Myc (Li et al., 1994; Classen and Hann, 1999). To avoid problems associated with the early transient induction of c-Myc following anti-IgM treatment, we performed the transient co-transfection analysis in Jurkat T cells, where higher levels of Inr-mediated transcription and basal levels of p27 promoter activity were observed in preliminary analyses (Manzano-Winkler et al., 1996, and Figure 3a). We ®rst con®rmed the ability of wild type c-Myc to repress the p27 promoter using increasing concentrations of the pRc-CMV-Myc expression vector. Co-expression of c-Myc repressed the activity of the pl34 ± 11 p27 promoter construct in a dose-dependent fashion when co-transfected into Jurkat T cells (Figure 3a). An approximate fourfold decrease in activity was noted in four separate experiments. Similar data were obtained with the pM21 expression vector, which expresses human cMyc (data not shown). We then compared the MLV-Myc, which expresses wild type c-Myc, with the MLV-Myc dlMBII vector, which expresses a mutant c-Myc protein with deletion of 18 amino acids within the MBII (kindly provided by E Zi). While these paired MLV-promoter driven vectors yield lower levels of c-Myc proteins than the one above driven by the CMV LTR, they have been successfully used to demonstrate the role of the MBII in repression (Li et al., 1994). In co-transfection analysis in Jurkat T cells, MLV-Myc signi®cantly reduced the activity of the pl34 ± 11 p27 promoter construct (Figure 3b; P50.01); although, to a lesser extent than seen above, as expected. No signi®cant repression of the pl34 ± 11 p27 promoter activity construct was seen upon co-transfection of MLV-Myc dlMBII vector (Figure 3b; P40.05). These vectors yield equivalent levels of c-Myc protein in multiple cell types (Li et al., 1994), including NIH3T3 cells (data not shown). Thus, c-Myc potently represses the p27 promoter in Jurkat T cells, and this inhibition appears to involve the Myc homology box II domain of c-Myc, consistent with Inr-mediated repression. Leu115 mutant `super-repressor' c-Myc potently inhibits p27 promoter activity Mutation of phenylalanine to leucine at position 115 of c-Myc, which is naturally occurring in various Burkitt's lymphomas (Discussion), was shown to increase transcriptional repression by c-Myc while not aecting activation mediated via an E-box (Lee et al., 1996). To compare the eect of this mutation on c-Myc-mediated inhibition of the p27 promoter, Jurkat cells were cotransfected with low doses of vectors pMLV-Myc or pMLV-MycB2, expressing either wild type c-Myc or cMycLeu115, respectively (Figure 3c). With wild type cOncogene

Myc expression, a low level of repression was seen with the low doses employed. In contrast, c-MycLeu115 caused a dramatic drop in p27 promoter activity (Figure 3c). Equal levels of c-Myc protein expression by these vector preparations were con®rmed following transfection into NIH3T3 cells (data not shown). Thus, the Phe115 to Leu115 mutation enhances the ability of c-Myc to repress the activity of the p27 promoter. c-Myc represses p27 promoter activity in non-immune cells We next asked whether repression of the p27 promoter by c-Myc occurs in other cell types. Expression of p27 mRNA and protein has been linked inversely with growth of vascular smooth muscle cells (SMCs) both in vivo and in vitro (Chen et al., 1997; Gallo et al., 1999; Tanner et al., 1998). Thus, we next assessed the potential role of c-Myc in repression of the p27 promoter activity in primary cultures of aortic SMCs using transient transfection analysis. Cells were plated at approximately 40 ± 50% con¯uence, incubated overnight and then transfected with the pl34 ± 11 p27 promoter construct in the absence or presence of an increasing dose of pRc-CMV-Myc expression vector (Figure 4a). Activity of the p27 promoter was repressed by c-Myc in a dose-dependent fashion. An approximate 83% inhibition of promoter activity was noted with co-transfection of 3 mg pRc-CMV-Myc vector DNA. Decreased levels of p27 have been found to correlate with poor prognosis and increased grade of many cancers, including breast cancer (Catzavelos et al., 1997; Kim et al., 1998; Porter et al., 1997). Conversely, increased c-Myc expression has been associated with increased tumor grade (Berns et al., 1992; Borg et al., 1992). Thus, we next assessed the role of c-Myc in p27 promoter activity in breast cancer cells, using the human breast cancer line Hs578T in transient transfection analysis. Cells in exponential growth were transfected with the pl34 ± 11 p27 promoter construct in the absence or presence of 2.5 or 5 mg of the pRcCMV-Myc expression vector. As can be seen in Figure 4b, the p27 promoter activity was decreased approximately ®vefold at the higher dose of c-myc expression vector. Thus, c-Myc expression decreases the activity of the p27 promoter in multiple non-immune cells, including SMCs and breast cancer cells. The p27 +1 region has an Inr element that is repressed by c-Myc To begin to localize the sequences mediating transcription initiation and repression, a smaller p27 promoter construct, DXhoI (745 to +154 bp), with only 45-bp of sequence upstream of the transcription start site, was similarly analysed. Co-transfection into Jurkat T cells with a c-Myc expression vector repressed an average of sixfold in four experiments (data not shown). To further de®ne these sequences, the region between 720 and +20 bp of the p27 gene was fused to the


1693

Figure 3 Wild type and `super-repressor' but not MBII deleted c-Myc represses the p27 promoter activity in Jurkat T cells. (a) cMyc repression is dose-dependent. Exponentially growing Jurkat T cells were electroporated, in duplicate, with 10 mg pl34 ± 11 in the absence (0 mg) or presence of 1, 2.5, or 5 mg of pRc-CMV-Myc c-Myc expression vector and enough pBluescript DNA to make a ®nal amount of 40 mg. Extracts, prepared after 24 h, were normalized for amount of protein and assayed for luciferase activity. Results are presented as the average+s.d. (b) Mutation of MBII ablates repression of the p27 promoter. Exponentially growing Jurkat T cells were electroporated, in duplicate, with 10 mg pl34 ± 11 in the absence (no Myc) or presence of 5 mg either MLV-Myc wild type or MLV-Myc d1MBII mutant c-Myc expression vector, and enough pBluescript DNA to make a ®nal amount of 40 mg. Extracts, containing equal amounts of protein, were assayed for luciferase activity, and the results presented relative to the p27 vector alone. No Myc to MLV-Myc wild type: P50.01; MLV-Myc wild type to MLV-Myc dlMBII: P50.05; No Myc to MLVMyc dlMBII: P40.05. (c) Mutation of Phe115 to Leu in c-Myc enhances repression of the p27 promoter. Exponentially growing Jurkat T cells were cotransfected, in triplicate, using the GenePorter transfection reagent with 1 mg pl34 ± 11 p27 promoter reporter construct, and 0 (7), 0.1, 0.5 and 2.5 mg of either pMLV-Myc or pMLV-MycB2 vectors, expressing wild type and mutated c-Myc protein, respectively, 0.5 mg SV40-b-gal and enough pMLV parental vector to make a ®nal amount of 4 mg. Cells were harvested after 48 h, and extracts, normalized for bgal activity, were assayed for luciferase activity. Results are presented as the average+s.d.

promoter-less pXP2 luciferase construct to generate pXP720/+20. The parental and pXP720/+20 vectors were used in transient transfection analyses into Jurkat T cells. As expected, the parental vector exhibited very low activity, whereas pXP720/+20

displayed a signi®cantly higher level of luciferase activity (Figure 5a). The activity of pXP720/+20 was approximately one-half that observed with the DXhoI p27 promoter construct (data not shown). Furthermore, co-transfection with the CMV-driven cOncogene


1694

Figure 4 c-Myc represses p27 promoter activity in non-lymphoid cells. (a) Aortic SMCs. SMCs were plated at 36105/ 60 mm2 dish (approximately 40 ± 50% con¯uence). After overnight incubation, cells were transfected by lipofection, in duplicate, with 2 mg pl34 ± 11 reporter in the absence (0 mg) or presence of 0.1, 0.5, 1, or 3 mg of pRc-CMV-Myc c-Myc expression vector, 0.5 mg of SV40-b-gal, and enough pCMV parental vector to make a ®nal amount of 5 mg. Cells were harvested after 48 h, and extracts, normalized to bgal activity, were processed. Results are presented as the average+s.d. (b) Hs578T breast cancer cells. Hs578T cells in exponential growth were transfected by calcium phosphate, in duplicate, with 15 mg pl34 ± 11 in the absence (0 mg) or presence of 2.5 or 5 mg of pRc-CMV-Myc c-Myc expression vector, 2.5 mg of MoECAT and enough pUC19 DNA to make a ®nal amount of 22.5 mg. Extracts, normalized to CAT activity, were processed and results are presented as above

Figure 5 The p27 promoter contains an active Inr. (a) p27 promoter region. Exponentially growing Jurkat T cells were electroporated, in duplicate, with 15 mg of either pXP2 or pXP720/+20 in the presence (+), or absence (7) of 8 mg of pRc-CMVMyc. Samples contained a total of 25 mg of DNA, normalized with pBluescript plasmid DNA. Extracts, harvested 24 h after electroporation, were assayed for luciferase activity. Results were normalized for amounts of protein in the extracts, and are presented as the average+s.d. (b) Basal activity of the MUTpXP720/+20 construct is greatly diminished. Exponentially growing Jurkat T cells were transfected, in duplicate, with 10 mg of the parental pXP2 vector (stippled box) or 10 mg of wild type pXP720/ +20 (®lled boxes) or MUTpXP720/+20 (open boxes) in the presence of 0, 2.5 or 5 mg pRc-CMV-myc (Myc), as indicated. Samples contained a total of 30 mg of DNA, normalized with pBluescript DNA. Extracts, harvested 18 h after electroporation, were assayed for luciferase activity. Results were normalized for amounts of protein in the extracts, and are presented as the average+s.d. Oncogene


Myc expression vector reduced the activity of pXP720/+20 by approximately 70%. When similar transfection analysis was performed with a construct containing a 4-bp mutation at the 3' end of the Inr element, termed MUTpXP720/+20, basal activity was greatly diminished (Figure 5b). The level of activity of the MUTpXP720/+20 construct was decreased by approximately 90% compared to the wild type pXP720/+20. This activity was somewhat higher than that observed with the parental pXP2 luciferase DNA. This low basal activity was unaected upon cotransfection with the pRc-CMV-Myc expression vector, although the signi®cance of this observation is unclear given the extremely large reduction resulting from the mutation. As expected, the activity of the wild type pXP720/+20 was reduced in a dose-dependent fashion. Thus, the region immediately surrounding the transcriptional start site of the p27 gene functions as an initiator element, promoting transcription in a fashion that can be repressed by c-Myc. c-Myc can interact with the p27 Inr element through association with Max To begin to characterize the initiation complex, we next performed EMSA with nuclear extracts from WEHI 231 cells using an oligonucleotide containing 720 to +20 bp sequences as probe. The EMSA yielded three major shifted complexes (Figure 6a, lane 1, upper panel). A 100-fold molar excess of unlabeled wild type oligonucleotide eciently competed for binding, whereas addition of an oligonucleotide with the mutated sequences used in the MUTpXP720/+20 vector, described above, failed to compete (Figure 6a, lower panel). Surprisingly, the formation of the two slower migrating complexes was diminished upon addition of anity puri®ed c-Myc antibody (Figure 6a, lane 2). As expected, the c-Myc antibody preparation had no eect on the band pattern obtained with an irrelevant probe for NF-kB (data not shown), indicating the speci®city of its eects on binding to the p27 +1 region. Further, we tested for the presence of another b/HLH/LZ factor, USF, which has been found to bind to Inr elements. Addition of a speci®c antibody against USF had no eect on the band pattern with the p27 +1 oligonucleotide (Figure 6a), although it had a dramatic eect on binding to an E-box (data not shown). These data suggest the presence of c-Myc in complexes that bind the sequences within the 720 to +20 bp start site of p27 promoter sequences. To con®rm the ability of c-Myc to bind to the +1 oligonucleotide, bacterially expressed proteins were used in EMSA. No complexes were detected upon addition of up to 300 ng of c-Myc-Glutathione Stransferase (GST) fusion protein alone to the binding reaction (Figure 6b). Since Max has been identi®ed as the major binding partner of c-Myc, we next assessed the eects of Max protein in gel shift assays. Addition of 75 ng of GST-Max yielded a single shifted complex (Figure 6b). The ability of Max to facilitate association of c-Myc with the Inr element was monitored next. The

amount of Max in the reactions was held constant at 75 ng, while the amount of Myc fusion protein was increased from 75 to 300 ng, such that the molar ratio of c-Myc and Max varied from 1 : 2 to 2 : 1. When cMyc and Max preparations were mixed and allowed to bind the +1 oligonucleotide, the formation of a new complex with slower mobility was observed in addition to the band present with Max alone (Figure 6b). The intensity of the new band increased dramatically as the amount of c-Myc in the reaction was increased. Thus, formation of this complex occurred in a c-Myc dosedependent fashion. Competition analyses were performed to con®rm the speci®city of complex formation. A 50-fold molar excess of unlabeled wild type oligonucleotide eciently competed both bands, whereas addition of the mutant oligonucleotide, described above, failed to compete (Figure 6b). As expected, addition of an anti-c-Myc antibody preparation reduced the intensity of the upper band (Figure 6c). Addition of an anti-Max antibody cleared the lower band seen with Max alone resulting in a huge supershifted band, making it dicult to assess its eects on the upper complex. Overall, the upper complex formed contains c-Myc and likely also has a Max component since it only formed in the presence of Max protein. Because the lower band was present with Max alone and cleared with an anti-Max antibody, it most likely represents Max homodimer binding.

1695

The Max variant dMax cannot bind to the Inr region nor facilitate repression by c-Myc The naturally occurring splice variant of Max, termed dMax, is missing the basic region and part of the HLH region but retains the LZ (Arsura et al., 1995). The dMax protein is unable to bind to an E-box, but is able to interact in vivo and in vitro with c-Myc. To further explore the role of Max in binding of c-Myc to the p27 start site, similar EMSA was performed using dMax protein. No binding was detected following addition of 75 ng of GST-dMax, whereas, control GST-Max protein bound as above (Figure 7). Furthermore, dMax was unable to facilitate binding of c-Myc to this sequence. SDS ± PAGE and in vitro pull down experiments con®rmed the presence of intact GSTdMax protein, as well as its ability to interact with cMyc and other Max binding partners (see FitzGerald et al., 1999). Next, we assessed mutations of c-Myc that aect binding to Max, i.e., the HLH and LZ domains. Deletion of either the HLH or LZ regions essentially eliminated the ability of c-Myc to bind with Max to the p27 start site oligonucleotide (Figure 7). These results indicate that association of c-Myc with Max is required for c-Myc to bind to the Inr element. We next wished to investigate the functional role of Max in repression by c-Myc. The rat PC12 pheochromocytoma tumor cell line, which has been shown to express a defective Max protein, unable to heterodimerize with c-Myc, and possibly an altered c-Myc protein (Hopewell and Zi, 1995, and data not shown), was employed. PC12 cells were transfected with 10 mg Oncogene


1696

Figure 6 c-Myc protein can bind the p27 start site oligonucleotide. (a) c-Myc present in WEHI 231 nuclear extracts participates in complex formation on the p27 +1 region. Upper panel. Nuclear protein extracts (5 mg) from exponentially growing WEHI 231 cells were incubated with 32P-labeled, double-stranded wild type p27 +1 oligonucleotide probe and separated on a 5% polyacrylamide, 5% glycerol-containing gel as described in Materials and methods. Where indicated, anti-c-Myc or anti-USF antibody preparations (1 ml) were added prior to the addition of labeled probe. The positions of the three complexes detected are as indicated. Lower panel. Nuclear protein extracts (5 mg) from exponentially growing WEHI 231 cells were incubated with 32P-labeled, double-stranded wild type p27 +1 oligonucleotide probe in the absence or presence of 100-fold molar excess wild type or mutant p27 +1 oligonucleotide, and complexes resolved by EMSA. (b) Bacterially synthesized c-Myc protein binds the p27 +1 region with Max. GST-c-Myc or GST-Max alone, or various combinations of the two were incubated with 32P-labeled p27 +1 probe and separated on a 4.5% polyacrylamide, 5%-glycerol-containing gel as described in Materials and methods. The amounts of proteins (in ng) added are given above the lanes. Where indicated, 50-fold molar excess unlabeled wild type or mutant oligonucleotide competitor DNA was added prior to the addition of labeled probe. (c) Bacterially synthesized GST-c-Myc (300 ng) and Max (75 ng) proteins were used in gel shift analysis. Where indicated 1 ml of either anti-c-Myc or anti-Max antibody preparation was added

pl34 ± 11 DNA, 0.5 mg pRc-CMV-Myc DNA and 5 or 10 mg of either pMT2T-Max or pMT2T-dMax DNA expressing wt Max or dMax, respectively. Parental pMT2T vector DNA was added to maintain a constant ®nal amount of pMT2T-based DNA of 10 mg. A dosedependent drop in p27 promoter activity was observed upon co-transfection with 5 or 10 mg wt Max expression vector (Figure 8). In contrast, dMax was unable to facilitate repression of the p27 promoter. These ®ndings suggest that expression of wild type Max facilitates repression of the p27 promoter activity by cMyc, and extend the repression to PC12 pheochromocytoma tumor cells. Taken together with the binding studies, described above, these results indicate that binding to the promoter region is required for Max to facilitate repression of the p27Kip1 gene by c-Myc. Discussion Here, we demonstrate the p27 gene is a target for repression by c-Myc in WEHI 231 immature B cells, Jurkat T, Hs578T breast cancer, vascular smooth Oncogene

muscle and PC12 cells. Furthermore, Max appears to facilitate the repression via promoting binding of cMyc to the p27 promoter region. The TATA-less p27 promoter has a functional Inr element. Consistent with previous ®ndings, repression of p27 promoter activity required AA 122 to 140, encompassing the MBII domain, and was enhanced by mutation of Phe115 to Leu (Li et al., 1994; Lee et al., 1996). Treatment of WEHI 231 cells with anti-IgM was previously shown to result in a decline in the constitutive levels of both NF-kB/Rel and A-Myb that leads directly to the drop in c-Myc level (Lee et al., 1995; Arsura et al., 2000). Furthermore, inhibition of c-Myc leads to induction of p27 levels (Wu et al., 1999). Here, we show that c-Myc expression prevents induction of p27 gene expression. Together these ®ndings indicate that the drop in c-Myc induced by BCR engagement allows for the derepression, and thereby elevated expression of p27. This increase in p27 expression leads to G1/S phase growth arrest (Ezhevsky et al., 1996) and apoptosis (Wu et al., 1999). Consistent with the evidence for cMyc repression of p27 gene transcription, many cells exhibit an inverse pattern in expression of c-Myc and


1697

Figure 7 The alternatively spliced dMax variant protein is unable to facilitate binding to the p27 +1 oligonucleotide by cMyc. GST-c-Myc (300 ng) or GST-DLZ-c-Myc (300 ng), GSTDHLH-c-Myc (300 ng) and GST-Max (75 ng) or GST-dMax (75 ng) were added alone, or in the indicated combinations in the binding reactions with 32P-labeled p27 +1 probe. The complexes were separated on a 4.5% polyacrylamide, 5%-glycerol-containing gel as described in the legend to Figure 6

p27 during growth induction, dierentiation and transformation (see below). Furthermore, Sedivy and coworkers have recently found that expression of both p27 mRNA and protein levels are elevated in rat1 cells null for c-myc (Mateyak et al., 1999), consistent with loss of c-Myc-mediated repression, as noted previously for the GADD45 gene (Mateyak et al., 1997). Overall these ®ndings identify repression of CKI p27Kip1 gene transcription as a mechanism for c-Myc-mediated control of cell proliferation and survival. Here we show that c-Myc can repress p27 gene transcription. It has also been shown that p27 function can be inhibited by sequestration, which prevents its association with the cyclin D-CDK complex; interestingly c-Myc has been implicated in this pathway, as well (Vlach et al., 1996; Bouchard et al., 1999; PerezRoger et al., 1999). Interestingly in WEHI 231 cells, Ezhevsky et al. (1996) have shown that p27 is associated with cyclin A and not with the D type cyclins. Overall, the regulation of p27 appears quite complex. In the initial studies, post-translational sites of regulation were observed (Polyak et al., 1994; Toyoshima and Hunter, 1994). For example, protein stability was identi®ed as a major site of control. Accumulation of p27 protein in density-mediated growth arrest of HeLa cells and ®broblasts was found to occur through changes in protein half-life (Hengst and Reed, 1996). In certain colorectal carcinomas, low p27 protein levels were shown to result from increased proteolytic activity, due to enhanced proteasome-

Figure 8 Max, but not dMax, facilitates c-Myc-mediated repression of the p27 promoter in PC12 cells. PC12 cells were transfected, in duplicate, with 10 mg pl34 ± 11, 0.5 mg pRc-CMVMyc (pCMV-Myc) and 0, 5, or 10 mg of either the Max or dMax expression vectors, pMT2T-Max and pMT2T-dMax respectively, plus 2 mg SV40-bgal expression vector, added to normalize for transfection eciency. Parental pMT2T vector DNA was added to adjust the total amount of pMT2T-based DNA to 10 mg. After overnight incubation, extracts were prepared, and luciferase activity was measured in samples, normalized for b-gal. Values are given as the average+s.d.

mediated degradation (Loda et al., 1997; Steiner et al., 1996). More recent studies have indicated p27 expression is also controlled at pre-translational levels. Kamesaki et al., (1998) observed increased p27 mRNA and protein levels during TGF-b1-mediated growth arrest of CH 31 and WEHI 231 immature B cells. Furthermore, during contact inhibition of Balbc/3T3 cells, p27 mRNA and protein levels were elevated as cell density increased, which corresponded with a decrease in CDK2 activity (Yanagisawa et al., 1999). Treatment of human A375 melanoma cells with IL-6 led to growth arrest and an increased level of p27 mRNA, and a corresponding increase in p27 protein (Kortylewski et al., 1999). Ectopic expression of the Von Hippel-Lindau (VHL) tumor suppressor gene resulted in G1 cell cycle arrest and growth inhibition of renal and breast cancer cell lines due to induction in levels of the p27 protein, mRNA, and gene transcription (Kim et al., 1998). In addition, as discussed above, rat1 cells null for c-myc displayed increased p27 mRNA levels (Mateyak et al., 1999). Overall, the altered p27 expression that accompanies changes in growth state appear to be regulated at multiple sites. Our studies demonstrate control of transcription of the Inr-driven promoter of the p27 gene via repression mediated by c-Myc. Our ®ndings provide one of the ®rst pieces of evidence that Max can play a role in repression by cOncogene


1698

Oncogene

Myc. Binding of bacterially expressed c-Myc to the +1 region required Max, which bound to the 720 to +20 bp oligonucleotide even in the absence of c-Myc. Similar binding data were obtained using the T cell receptor Vb Inr element as probe (data not shown). The variant dMax, which can interact with c-Myc (Arsura et al., 1995), was found unable to bind to the Inr element of the p27 promoter, or to facilitate c-Myc binding or repression. These ®ndings are consistent with our previous data showing that the induction of an Adenoviral Inr- construct in WEHI 231 cells was not inhibited upon co-transfection of a dMax expression vector (FitzGerald et al., 1999). Importantly, dMax could not enhance repression of the p27 promoter by c-Myc in PC12 cells, which contain a mutant Max protein (Hopewell and Zi, 1995), whereas the wt Max protein very eectively enhanced repression. Furthermore LZ and HLH mutants of cMyc, unable to interact with Max, also failed to bind. Interestingly, Max has been found essential for transformation by Myc (Amati et al., 1993). Our work suggests a potential role for Max in repression of the p27 gene by c-Myc. Multiple pathways of repression by Myc have been elucidated. The work of Roy and coworkers (1983) ®rst implicated Myc in repression of TFII-I-mediated Inr-driven transcription. Facchini et al. (1997) noted an obligatory role for Myc-Max heterodimerization in autoregulation of c-myc gene transcription from its P2 promoter. In another interesting twist, Peukert et al. (1997) observed that positive transactivation of the cyclin D1 promoter by Miz-1 can be ablated via Miz-1 interaction with c-Myc. Importantly, c-Myc has been found to mediate repression of a growing list of genes involved in control of cell cycle or growth arrest, such as GADD34, GADD45, GADD153, gas1, and p21 (Lee and Dang, 1997; Lee et al., 1997; Amundson et al., 1998; Warner et al., 1999; Classen and Hann, 2000), and dierentiation (see below). The mechanism(s) of cMyc-mediated repression of these genes, including the potential role of Max, remains to be explored. BCR engagement of WEHI 231 cells initiates a complex signal cascade resulting in apoptosis. Recently, we demonstrated that the induction of p27 activity plays a direct role in activation of apoptosis in WEHI 231 cells (Wu et al., 1999). Ectopic overexpression of p27 induced apoptosis, and induction of apoptosis upon microinjection of an antisense c-myc expression vector was ablated by inhibition of p27 expression. The results presented here argue that the drop in c-Myc leads to de-repression of the p27 promoter. Interestingly, the link between BCR engagement and increased p27 levels is mediated, in part, by a drop in NF-kB/Rel activity (Wu et al., 1996b), indicating another mechanism for the observed eects of the Rel family of transcription factors on cell proliferation and apoptosis. In addition to the NF-kB/ c-Myc/p27 pathway, we have recently demonstrated BCR engagement also leads to increased expression of p53 which induces p21 CKI mRNA by 2 h posttreatment (Wu et al., 1998). Consistently, the nuclear

run-on demonstrated an early increase in p21 gene transcription. The p53/p21 pathway, which accounts for approximately 45 ± 50% of the death of WEHI 231 cells upon anti-IgM treatment, appears to function independently of the c-Myc and p27 pathway. A dominant negative p53 failed to ablate either the decrease in c-Myc or increase in p27 following antiIgM treatment (Wu et al., 1999). Engagement of the CD40 receptor protects WEHI 231 cells from BCRinduced cell death. Interestingly, we have shown that treatment with CD40 ligand results in an early and sustained induction of NF-kB which prevents the drop in c-Myc (Schauer et al., 1996, 1998) and the induction of p27 (Wu et al., 1999). The induction of p27 has been shown to lead to accumulation of repressor complexes of the E2F family (Shiyanov et al., 1997), raising a question as to whether changes in expression and/or function of the E2F family of transcription factors are downstream targets of these CKIs. A Phe to Leu substitution mutation within c-Myc at amino acid 115 (MycB2 mutant) was found in several Burkitt's lymphomas (Clark et al., 1994; Hoang et al., 1995; Yano et al., 1993). This mutated c-Myc was shown to repress an Inr-containing adenoviral promoter construct in transfection analysis more potently than wild type c-Myc, whereas, its transactivation activity via an E-box element remained the same (Lee et al., 1996). When compared with wild type c-Myc, this `super-repressing' c-Myc displayed enhanced ability in a Ras co-transformation in primary embryonic cells and less contact inhibition in rat1a cells (Lee et al., 1996). In our study, the MycB2 c-Myc showed more potent repression of p27 promoter, suggesting that suppression of p27 might be one of the c-Myc pathways that leads to neoplastic transformation, and implicating this mechanism in the in vivo growth selection advantage in cells containing such a mutation. Interestingly, a growing body of literature has demonstrated decreased levels of p27 protein in various tumor types (Catzavelos et al., 1997; Porter et al., 1997). The decreased level of p27 has been characterized as an independent factor of poor prognosis in various human cancers such as breast, colon, and prostate adenocarcinoma (Lioyd et al., 1999; Tsihlias et al., 1999; Tsuchiya et al., 1999). Although some tumors only showed changes in p27 protein levels, suggesting post-translational control, others displayed changes in the levels of both p27 mRNA and protein (Kim et al., 1998; Lee et al., 1999). Recent studies have indicated that c-Myc levels are elevated in as many as one seventh of human cancer deaths in the United States (Dang, 1999). The elevated level of c-Myc has also been identi®ed as a poor prognostic marker (Berns et al., 1992; Borg et al., 1992). Taken together, these ®ndings suggest that c-Myc-mediated repression of p27 gene transcription is, in part, responsible for the inverse correlation of these two factors implicated in tumor development. A transient or permanent drop in expression of cMyc has been shown to play a pivotal role during terminal dierentiation. For example, constitutive


ectopic c-myc expression was shown to prevent terminal dierentiation of mouse erythroleukemia (MEL) cells (Dmitrovsky et al., 1986), while selective reduction in c-myc function was found to promote dierentiation of these cells directly (Prochownik et al., 1988). Interestingly, many genes reported to be repressed by c-Myc through an Inr-mediated mechanism are expressed in a dierentiation-speci®c pattern, e.g., albumin and C/EBPa in hepatocytes (Li et al., 1994; Runge et al., 1997), l5 and terminal deoxynucleotidyl transferase in pre-B cells (Mai and Martensson, 1995), and mim1 and lysozyme in myelomonocytes (Mink et al., 1996) (rev. in Classen and Hann, 1999; Dang, 1999). The CKI p27 has been found to play an important role in control of dierentiation. Thus, it is tempting to speculate that a cell-type speci®c dierentiation inducing signal in concert with the drop in cMyc, plays two essential roles in dierentiation: allowing for de-repression of genes that encode proteins that arrest growth (e.g., p27) and dierentiation-speci®c markers. In this model, repression by cMyc plays a central role in the observed function of this oncoprotein as an essential `gatekeeper' of dierentiation. Work is in progress on this question. Interestingly overexpression of c-Myc has recently been shown by Iritani and Eisenman (1999) to yield a larger cell size. Our experiments suggest a possible mechanism for this observation. The p27 protein is involved in control of the restriction point where cells are blocked in the G1 phase preventing entry into the S phase of the cell cycle (Sherr and Roberts, 1999). For example, induction of the p27 CKI in anti-IgM-treated WEHI 231 cells has been implicated in G1/S phase growth arrest (Ezhevsky et al., 1996). Thus, repression of the p27 gene by c-Myc likely facilitates transition from G1 into S phase, which would lead to an overall increase in cell size. Consistent with this hypothesis are older observations demonstrating overexpression of c-myc in growing cells leads to reduced growth factor requirements and a shortening of the time to pass through G1 (rev. in Lemaitre et al., 1996; Spencer and Groudine, 1991). Materials and methods Cell culture and transient transfection conditions WEHI 231 cells were maintained and treated with 1 : 1000 dilution of anti-m heavy chain antibody as previously described (Schauer et al., 1996). Cells were transfected via electroporation either in exponential growth or following anti-IgM treatment for either 4 or 6 h, as described previously (Wu et al., 1996a). Anti-IgM treatment had no detectable eect on transfection eciency (data not shown). Following incubation at 378C for the indicated times, cells were harvested and the resulting extracts were normalized either for b-gal reporter activity (upon co-transfection with SV40 b-gal) (Bellas et al., 1995) or for total protein (Bio-Rad protein quantitation kit), which all yielded similar results. Jurkat T cells were maintained in Dulbecco's modi®ed Eagle's medium supplemented with 10% fetal bovine serum, glutamine, non-essential amino acids, penicillin, and streptomycin and transfected either by electroporation, essentially as

described above, or using the GenePorter transfection reagent (Gene Therapy Systems, San Diego, CA, USA), as indicated. The Hs578T breast tumor cell line, which was established from a patient with in®ltrating ductal carcinoma and does not express estrogen receptors (Hackett et al., 1977), was maintained and transfected using a modi®ed calcium phosphate method, as described previously (Sovak et al., 1997). Aortic SMCs were derived from primary tissue explant of aortae of female calves, and cultures maintained and transfected using lipofectamine reagent (Life Technologies) as we have described previously (Kypreos et al., 1999). All SMCs were within passage 4 from primary explant. PC12 (pheochromocytoma) cells, obtained from ATCC (American Type Culture Collection, Rockville MD, USA), were cultured in DMEM/F12 medium containing 10% heat inactivated horse serum, 5% heat inactivated FBS, glucose, glutamine, non-essential amino acids, penicillin, and streptomycin. Cells were co-transfected via electroporation, incubated overnight at 378C, harvested and analysed as above.

1699

Protein and RNA blot analysis For analysis of levels of p27 and c-Myc proteins, whole cell extracts were prepared as described previously (Wu et al., 1996a). Brie¯y, cells were rinsed two times in ice cold PBS and lysed in RIPA buer (10 mM Tris-HCl, pH 7.5 150 mM NaCl, 1% NP-40, 0.1% SDS, 1% sodium sarcosyl, 0.2 mM phenylmethylsulfonyl ¯uoride (PMSF), 10 mg/ml leupeptin, 1 mM dithiothreitol) with repeated shearing through needles down to 25G. Extracts were spun in a microcentrifuge to remove insoluble material before quantitation and electrophoresis. Immunoblotting was performed as described previously (Schauer et al., 1996) using p27 antibody (Transduction Laboratories, Lexington KY, USA), or anity puri®ed c-Myc speci®c antibody (50 ± 23) (kindly provided by S Hann, Vanderbilt University, Memphis, TN, USA). Cytoplasmic RNA was isolated and analysed as previously described (Schauer et al., 1996) using either the HindIII ± BamHI fragment of the pGlu-p27H vector (generously provided by Y Xiong, University of North Carolina, Chapel Hill, NC, USA), or c-myc54 cDNA vector (Stanton et al., 1983), as probe. Nuclear run-on analysis Nuclei isolation and transcription analysis was performed as described by Greenberg and Zi (1984), essentially as we have described previously (Levine et al., 1986). Nitrocellulose ®lters were prepared with 10 mg of DNA per slot and hybridization mixtures contained 2.56106 cpm/ml per ®lter lane, as described previously (Levine et al., 1986). Doublestranded DNA was used for the following probes: p27, pOPRSVI-27 vector DNA (Wu et al., 1999); rat b-actin, (Bond and Farmer, 1983); p21, pOPRSVI-p21 (Wu et al., 1998); pUC19 DNA. Data were subjected to densitometry twice and the results presented normalized for b-actin+s.d. Plasmids and promoter constructs The p27 promoter construct pl34 ± 11 contains 2.2 kb of the upstream region of the p27 gene fused to the ®re¯y luciferase gene. Promoter deletion constructs were made by digesting pl34 ± 11 with XhoI (at 745 from the start site of transcription, respectively), and then religating the remaining promoter-reporter construct. Promoter activity of these constructs termed pl34 ± 11 (72002 to +154 bp) and DXhoI (745 to +154 bp) was con®rmed by transfection of these Oncogene


1700

constructs into HeLa cells (data not shown). The wild type (wt) c-Myc expression vectors used include: pRc-CMV-Myc, which is under the control of the CMV promoter (Hann et al., 1994), and pSV-Hu-myc2, which is driven by an SV40 enhancer-promoter (Land et al., 1986). Both vectors express c-Myc2 protein, the predominant form which is initiated at the AUG codon start site in exon 2. The human c-Myc homology box II (MB2) mutant expression construct MLVMyc dlMBII, encoding a protein with a deletion of amino acids 122 to 140, and its wt counterpart are both driven by the MLV LTR (Li et al., 1994). These were generously provided by E Zi (New York University Medical Center, New York, NY, USA). The pMLV-MycB2 vector, which expresses a human c-Myc with a mutation of phenylalanine at position 115 to leucine, driven by the MLV LTR (Lee et al., 1996). The pMLV-MycB2 and parental c-Myc expression vector pMLV-Myc were kindly provided by C Dang (The Johns Hopkins University School of Medicine, Baltimore MD, USA). To generate the pXP 720/+20 construct, the oligonucleotide 5'-CGCGGATCCCTGCGCCTCTCTTCCCCAGACCTGCGCGCTACTGCGGATCCC-3' and its complement were annealed, digested with BamHI, and ligated into the BglII site of the promoterless luciferase vector pXP2. The +1 site is underlined, and the Inr element indicated in bold letters. To generate the MUTpXP720/+20 construct, the oligonucleotide 5'-CGCCCATCCCTGCGCCTCTCTTCCCCAGACGGGGGGGCTACTGCGGATCCC-3' and its complement were similarly annealed, digested and introduced into the pXP2 vector, where the mutated bases are indicated in italic font. The dMax, MycDLZ and MycDHLH-GST fusion protein constructs were as described (FitzGerald et al., 1999). GST-c-Myc, and -Max constructs, bearing full length murine c-Myc and Max sequences, respectively, were generously provided by S Hann. Preparation of the GSTdMax construct was described previously (FitzGerald et al., 1999). The preparation of the pMT2T-Max and pMT2TdMax constructs was described previously (Arsura et al., 1995). EMSA The sequence of the wild type p27 oligonucleotide probe is as follows: p27+1, 5'-CCTGCGCCTCTCTTCCCCAGACCT-

GCGCGCTACTGCGG-3', where the Inr consensus sequence is in bold font. Double-stranded oligonucleotides were endlabeled by ®ll-in synthesis using Klenow polymerase and 32PdCTP and 32P-dGTP (NEN Life Science). Nuclear protein extracts were prepared according to a modi®cation of the method of Dignam et al., (1983). Modi®cations included replacement of HEPES with Tris, pH 7.9 in all buers, and substitution of KCl for NaCl in the extraction buer (Buer C). GST fusion proteins were induced in E. coli (protease de®cient strain BL-21) transformed with the appropriate plasmids and isolated as described (Frangioni and Neel, 1993). Puri®cation was con®rmed by gel electrophoresis. Binding assays using either nuclear extracts (5 mg of protein extract per reaction) or puri®ed proteins (amounts speci®ed in ®gure legends) were performed according to ManzanoWinkler et al. (1996). Speci®cally, the indicated protein amounts were pre-incubated for 10 min at 308C in reactions containing 20 mM Tris (pH 7.9), 0.2 mM EDTA, 5 mM dithiothreitol, 0.5 mM PMSF, 10% glycerol, and 0.1 mg/ml bovine serum albumin. Then, 1.5 ng of labeled probe and 100 ng of poly (dA:dT) were added, and the reaction incubated for an additional 20 min at 308C. (While similar bands were detected with dI:dC as non-speci®c competitor, their intensities were diminished.) Where indicated, antibody preparations or unlabeled competitors were included in the pre-incubation. Electrophoresis was performed in either 5% (for assays using nuclear extracts) or 4.5% (for assays using puri®ed proteins) polyacrylamide gels containing 5% glycerol and 0.5X TBE (45 mM Tris-Borate, 1 mM EDTA).

Acknowledgments We thank Drs C Dang, S Desiderio, S Hann, A Ko, Y Xiong, and E Zi for generously providing cloned DNAs or antibody preparations. This work was supported by National Institutes of Health grants CA 36355 (GE Sonenshein), HL13262 (GE Sonenshein), and CA 82742 (GE Sonenshein) and training grants HL07429 (W Yang) and CA64070 (MJ FitzGerald), and Department of Army grant DAMD17-98-1-8034 (DW Kim) and Cure for Lymphoma Foundation (M Arsura).

References Amati B, Brooks MW, Levy N, Littlewood TD, Evan GI and Land H. (1993). Cell, 72, 233 ± 245. Amundson SA, Zhan Q, Penn LZ and Fornace AJ. (1998). Oncogene, 17, 2149 ± 2154. Arsura M, Hofmann CS, Golay J, Introna M and Sonenshein GE. (2000). Blood, 96, 1013 ± 1020. Arsura M, Deshpande A, Hann SR and Sonenshein GE. (1995). Mol. Cell. Biol., 15, 6702 ± 6709. Arsura M and Sonenshein G.E. (2001). Transcription Factors: Normal and Malignant Development of Blood Cells. Ravid K and Licht J. (eds). J. Wiley & Sons. pp521 ± 549. Bellas RE, Lee JS and Sonenshein GE. (1995). J. Clin. Invest., 96, 2521 ± 2527. Benhamou L, Cazenave P-A and Sarthou P. (1990). Eur. J. Immunol., 20, 1405 ± 1407. Berns E, van Putten KJW, van Staveren I, Portengen H and Foekens J. (1992). Cancer Res., 52, 1107 ± 1113. Besancon F, At® A, Gespach C, Cayre YE and Bourgeade MF. (1998). Proc. Natl. Acad. Sci. USA, 95, 8081 ± 8086. Oncogene

Bond J and Farmer SR. (1983). Mol. Cell. Biol., 3, 1333 ± 1342. Borg A, Baldetorp B, Ferno M, Olsson H and Sigurdsson H. (1992). Eur. J. Cancer, 51, 687 ± 691. Bouchard C, Thieke K, Maier A, Sarich R, Hanley-Hyde J, Ansorge W, Reed S, Sicinski P, Bartek J and Eilers M. (1999). EMBO J., 18, 5321 ± 5333. Boyd A and Schrader J. (1981). J. Immunol., 126, 2466 ± 2469. Catzavelos C, Bhattacharya N, Ung YC, Wilson JA, Roncari L, Sandhu C, Shaw P, Yeger H, Morava-Protzner I, Kapusta L, Franssen E, Pritchard KI and Slingerland JM. (1997). Nature Med., 3, 227 ± 230. Chen D, Krasinski K, Sylvester A, Chen J, Nisen PD and Andres V. (1997). J. Clin. Invest., 99, 2334 ± 2341. Clark HM, Yano T, Otsuki T, Jae ES, Shibata D and Raeld M. (1994). Cancer Res., 54, 3383 ± 3386. Classen GF and Hann SR. (1999). Oncogene, 18, 2925 ± 2933. Classen GF and Hann SR. (2000). Proc. Natl. Acad. Sci. USA, 97, 9498 ± 9503.


Dang CV. (1999). Mol. Cell. Biol., 19, 1 ± 11. Dignam JD, Martin PL, Shastry BS and Roeder RG. (1983). Methods Enzymol., 101, 582 ± 598. Dmitrovsky E, Kuehl WM, Hollis GF, Kirsch IR, Bender TP and Segal S. (1986). Nature, 322, 748 ± 750. Ezhevsky S, Toyoshima H, Hunter T and Scott D. (1996). Mol. Biol. Cell, 7, 553 ± 564. Facchini LM and Penn LZ. (1998). FASEB J., 12, 633 ± 651. Facchini LM, Chen S, Marhin WW, Lear JN and Penn LZ. (1997). Mol. Cell Biol., 17, 100 ± 114. FitzGerald MJ, Arsura M, Bellas RE, Yang W, Wu M, Chin L, Mann KK, DePinho RA and Sonenshein GE. (1999). Oncogene, 18, 2489 ± 2498. Frangioni JV and Neel BG. (1993). Anal. Biochem., 210, 179 ± 187. Gallo R, Padurean A, Jayaraman T, Marx S, Roque M, Adelman S, Chesebro J, Fallon J, Fuster V, Marks A and Badimon JJ. (1999). Circulation, 99, 2164 ± 2170. Grandori C and Eisenman RN. (1997). Trends Biochem. Sci., 22, 177 ± 181. Greenberg ME and Zi EB. (1984). Nature, 311, 433 ± 438. Hackett AJ, Smith HS, Springer EL, Owens RB, NelsonRees WA, Riggs JL and Gardner MB. (1977). J. Nat. Cancer Inst., 58, 1795 ± 1806. Hann SR, Dixit V, Sears RC and Sealy L. (1994). Genes Dev., 8, 2441 ± 2452. Hengst L and Reed SI. (1996). Science, 271, 1861 ± 1864. Hoang A, Lutterbach B, Lewis BC, Yano T, Chou T, Barrett J, Raeld M, Hann SR and Dang CV. (1995). Mol. Cell. Biol., 15, 4031 ± 4042. Hopewell R and Zi EB. (1995). Mol. Cell Biol., 15, 3470 ± 3478. Iritani BM and Eisenman RN. (1999). Proc. Natl. Acad. Sci. USA, 96, 13180 ± 13185. Kamesaki H, Nishizawa K, Michaud GY, Cossman J and Kiyono T. (1998). J. Immunol., 160, 770 ± 777. Kihara-Newgishi F, Yamada T, Kubota Y, Kondah N, Yamamoto H, Abe M, Shirai T, Hashimoto Y and Oikawa T. (1998). Int. J. Cancer, 18, 523 ± 530. Kim M, Katayose Y, Li Q, Rakkar AN, Li Z, Hwang SG, Katayose D, Trepel J, Cowan KH and Seth P. (1998). Biochem. Biophys. Res. Commun., 253, 672 ± 677. Kortylewski M, Heinrich PC, Mackiewicz A, Schniertshauer U, Klingmuller U, Nakajima K, Hirano T, Horn F and Behrmann I. (1999). Oncogene, 18, 3742 ± 3753. Kwon T, Nagel J, Bucholz M and Nordin A. (1996). Gene, 180, 113 ± 120. Kypreos KE, Marhamati DJ and Sonenshein GE. (1999). Matrix Biology, 18, 275 ± 285. Land H, Chen AC, Morgenstern JP, Parada LF and Weinberg RA. (1986). Mol. Cell Biol., 6, 1917 ± 1925. Lee CC, Ichihara T, Yamamoto S, Wanibuchi H, Sugimura K, Wada S, Kishimoto T and Fukushima S. (1999). Carcinogenesis., 20, 1697 ± 1708. Lee LA and Dang CV. (1997). Curr. Topics Micro. Immunol., 224, 131 ± 135. Lee H, Arsura M, Wu M, Duyao M, Buckler AJ and Sonenshein GE. (1995). J. Exp. Med., 181, 1169 ± 1177. Lee LA, Dolde C, Barrett J, Wu CS and Dang CV. (1996). J. Clin. Investig., 97, 1687 ± 1695. Lee TC, Li L, Philipson L and Zi EB. (1997). Proc. Natl Acad Sci USA, 94, 12886 ± 12891. Lemaitre J-M, Buckle RS and Mechali M. (1996). Adv. Cancer Res., 70, 95 ± 144. Levine RA, McCormack JE, Buckler AJ and Sonenshein GE. (1986). Mol. Cell Biol., 6, 4112 ± 4116.

Li LH, Nerlov C, Prendergast G, MacGregor D and Zi EB. (1994). EMBO J., 17, 4070 ± 4079. Lioyd RV, Erickson LA, Lin L, Kulig E, QianX, Cheville JC and Scheithauer BW. (1999). Am J. Path., 154, 313 ± 323. Lo K and Smale ST. (1996). Gene, 182, 13 ± 22. Loda M, Cukor B, Tam SW, Lavin P, Fiorentino M, Dratta GF, Jessup JM and Pagano M. (1997). Nature Med., 3, 231 ± 234. Maheswaran S, McCormack JE and Sonenshein GE. (1991). Oncogene, 6, 1965 ± 1971. Mai S and Martensson IL. (1995). Nucleic Acids Res., 23, 1 ± 9. Manzano-Winkler R, Novina CD and Roy AL. (1996). J. Biol. Chem., 271, 12076 ± 12081. Mateyak MK, Obaya AJ, Adachi S and Sedivy JM. (1997). Cell Growth Dier., 8, 1039 ± 1048. Mateyak MK, Obaya AJ and Sedivy JM. (1999). Mol. Cell. Biol., 19, 4672 ± 4683. McCormack JE, Pepe VH, Kent RB, Dean M, MarshakRothstein A and Sonenshein GE. (1984). Proc. Natl. Acad. Sci. USA, 81, 5546 ± 5550. Mink S, Mutschler B, Weiskirchen R, Bister K and Klempnauer KH. (1996). Proc. Natl. Acad. Sci. USA, 93, 6635 ± 6640. Norvell A, Mandik L and Monroe JG. (1995). J. Immunol., 154, 4403 ± 4413. Perez-Roger I, Kim SH, Griths B, Sewing A and Land H. (1999). EMBO J., 18, 5310 ± 5320. Peukert K, Staller P, Schneider A, Carmichael G, Hanel F and Eilers M. (1997). EMBO J., 16, 5672 ± 5686. Polyak K, Kato JY, Solomon MJ, Sherr CJ, Massague J, Roberts JM and Ko A. (1994). Genes Dev., 8, 9 ± 22. Porter PL, Malone KE, Heagerty PJ, Alexander GM, Gatti LA, Firpo EJ, Daling JR and Roberts JM. (1997). Nature Med., 3, 222 ± 225. Prochownik EV, Kukowska J and Rodgers C. (1988). Mol. Cell. Biol., 8, 3683 ± 3695. Roy AL, Garruthers C, Gutjahr T and Roeder RG. (1993). Nature, 365, 359 ± 361. Runge D, Runge DM, Bowen WC, Locker J and Michalopoulos GK. (1997). J. Biol. Chem., 378, 873 ± 881. Schauer SL, Wang Z, Sonenshein GE and Rothstein TL. (1996). J. Immunol., 157, 81 ± 86. Schauer SL, Bellas RE and Sonenshein GE. (1998). J. Immunol., 160, 4398 ± 4405. Sherr C and Roberts JM. (1999). Genes Dev., 13, 1501 ± 1512. Shiyanov P, Hayes S, Chen N, Pestov D, Lau L and Raychaudhuri P. (1997). Mol. Biol. Cell., 8, 1815 ± 1827. Sovak MA, Bellas RE, Kim DW, Zanieski GJ, Rogers AE, Traish AM and Sonenshein GE. (1997). J. Clin. Invest., 100, 2952 ± 2960. Spencer CA and Groudine M. (1991). Adv. Cancer Res., 56, 1 ± 48. Stanton L, Watt R and Marcu K. (1983). Nature, 303, 401 ± 406. Steiner P, Rudolph B, Muller D and Eilers M. (1996). Prog. Cell Cycle Res., 2, 73 ± 82. Tanner FC, Yang ZY, Duckers E, Gordon D, Nabel GJ, Nabel EG and Thompson EB. (1998). Circ. Res., 82, 396 ± 403. Thompson EB. (1998). Annu. Rev. Physiol., 60, 575 ± 600. Toyoshima H and Hunter T. (1994). Cell, 78, 67 ± 74. Tsihlias J, Kapusta L and Slingerland J. (1999). Annu. Rev. Med., 50, 401 ± 423. Tsuchiya A, Zhang GJ and Kanno M. (1999). J. Surg. Oncol., 70, 230 ± 234.

1701

Oncogene


1702

Oncogene

Vlach J, Hennecke S, Alevizopoulos K, Conti D and Amati B. (1996). EMBO J., 15, 6595 ± 6604. Xiao Q, Claasen G, Shi J, Adachi S, Sedivy J and Hann S. (1998). Genes Dev., 12, 3803 ± 3808. Wang W, Wykrzykowska J, Johnson T, Sen R and Sen J. (1999). J. Immunol., 162, 314 ± 322. Warner BJ, Blain SW, Seoane J and Massague J. (1999). Mol. Cell. Biol., 19, 5913 ± 5922. Warner GL, Ludlow J, Nelson DO, Gaur A and Scott D. (1992). Cell Growth Dier., 3, 175 ± 181. Wu M, Arsura M, Bellas RE, FitzGerald MJ, Lee H, Schauer SL, Sherr DH and Sonenshein GE. (1996a). Mol. Cell. Biol., 16, 5015 ± 5025.

Wu M, Bellas RE, Shen J and Sonenshein GE. (1998). J. Exp. Med., 187, 1671 ± 1679. Wu M, Bellas RE, Shen J, Yang W and Sonenshein GE. (1999). J. Immunol., 163, 6530 ± 6535. Wu M, Lee H, Bellas RE, Schauer SL, Arsura M, Katz D, FitzGerald MJ, Rothstein TL, Sherr DH and Sonenshein GE. (1996b). EMBO J., 15, 4682 ± 4690. Yanagisawa K, Kosaka A, Iwahana H, Nakanishi M and Tominaga S. (1999). J. Biochem. (Tokyo), 125, 36 ± 40. Yano T, Sander C, Clark H, Dolezal M, Jae E and Raeld M. (1993). Oncogene., 8, 2741 ± 2748. Yi AK, Hornbeck P, Lafrenz DE and Krieg AM. (1996). J. Immunol., 157, 4918 ± 4925.