pmc logo image
Logo of nihpaNIHPA bannerabout author manuscriptssubmit a manuscript
Journal List > NIHPA Author Manuscripts

Formats:
Oncogene. Author manuscript; available in PMC 2006 July 18.
Published in final edited form as:
PMCID: PMC1513044
NIHMSID: NIHMS10914
Copyright notice and Disclaimer
F-box protein Skp2: a novel transcriptional target of E2F
L Zhang and C Wang
Center for Molecular Biology of Oral Diseases, University of Illinois at Chicago, Chicago, IL, USA
Correspondence: Associate Professor C Wang, Center for Molecular Biology of Oral Diseases, University of Illinois at Chicago, 801 South Paulina Street, Room 530E, m/c 860, Chicago, IL 60612, USA. E-mail: chiayeng/at/uic.edu
Small right arrow pointing to: The publisher's final edited version of this article is available at Oncogene.
Small right arrow pointing to: See other articles in PMC that cite the published article.
Abstract
The F-box-containing protein Skp2 plays a critical role in coordinating the G1/S transition and progression through the S phase of the mammalian cell cycle. Skp2 is overexpressed in a broad spectrum of human cancers and the expression level correlates with tumor malignancy. However, the Skp2 gene is neither amplified nor rearranged in most human cancers and the underlying mechanism of Skp2 overexpression remains poorly understood. We show here that the Skp2 gene contains a functional E2F response element (hSRE2). Ectopic expression of E2F1 induces expression of the endogenous Skp2 gene in human fibroblast cells, whereas antisense-mediated knockdown of E2F1 in human tumor cell lines reduces expression of endogenous Skp2 gene. The hSRE2 element not only participates in activation of Skp2 promoter function during normal cell cycle progression into S phase, it is also required for the high-level Skp2 gene expression in many human tumor cell lines. These results reveal Skp2 as a novel target for E2F regulation that is disrupted in several human tumor cell lines.
Keywords: transcription, SCFskp2, E2F, cell cycle
Ubiquitin-directed protein degradation plays an important role in regulating a broad spectrum of biological processes, including transcription, signal transduction, cell cycle progression, apoptosis, cell growth and differentiation, and development. Alterations in ubiquitination (and deubiquitination) reactions are directly linked to the pathogenesis of many diseases (Ciechanover and Schwartz, 2004). Proteins that are targeted for degradation by this mechanism undergo series of ubiquitin-modification steps carried out by the successive activity of E1 (activating), E2 (conjugating), and E3 (ligating) ubiquitin enzymes (Hershko and Ciechanover, 1998; Pickart and Eddins, 2004). SCF complexes belong to a large family of multi-subunit E3 ubiquitin ligases that select and ubiqutinate specific proteins for targeted destruction by the 26S proteasome. SCF complexes are named for their core protein components Skp1, Cullin/cdc53, and Rbx1/Roc, and an F-box-containing protein that determines substrate specificity (Zheng et al., 2002). Different SCF complexes are distinguished by the types of F-box protein in association with the core proteins (Kipreos and Pagano, 2000; Willems et al., 2004).
Skp2, which was initially identified as a protein stably interacting with the cyclin A–cdk2 complex (Zhang et al., 1995; Bai et al., 1996), is the substrate-recognition subunit of the SCFskp2 E3 ligase complex. Several lines of evidence from both biochemical and genetic studies have shown that Skp2 is required for cell cycle progression at multiple stages, including G1/S transition, S phase progression, and S/G2 transition (Deshaies, 1999; Koepp et al., 1999; Nakayama et al., 2004). Skp2 knockout mice show cellular defects that include nuclear enlargement, centrosome duplication, and polyploidy (Nakayama et al., 2000). A principal substrate responsible for the activity of Skp2 in these cell cycle phases is p27kip1, an inhibitor of cdk2 and cdk1 activities that promote entry into DNA synthesis and into mitosis, respectively (Carrano et al., 1999; Sutterluty et al., 1999; Tsvetkov et al., 1999; Nakayama et al., 2000, 2004). Subsequently, several additional cell cycle regulatory proteins including Myc (Jin and Harper, 2003; Kim et al., 2003; von der Lehr et al., 2003; Amati, 2004), cyclin E (Yeh et al., 2001), p57kip2 (Kamura et al., 2003), p21WAP1 (Bornstein et al., 2003), p130 (Tedesco et al., 2002; Bhattacharya et al., 2003), Cdt1 (Li et al., 2003), hOrc1p (Mendez et al., 2002), and E2F1 (Marti et al., 1999) have also been reported as substrates of Skp2-mediated protein degradation.
Like many cell cycle regulators, the Skp2 gene is periodically expressed throughout cell cycle (Zhang et al., 1995; Wirbelauer et al., 2000). Skp2 expression is lowest in quiescent cells and in early-mid G1 of cycling cells. The level increases as cells approach the G1/S transition and stays high during S phase before it starts to decrease in G2 and M phases. Both mRNA and protein levels of Skp2 are similarly regulated throughout the cell cycle, although the level of protein expression is not completely dependent on the amount of mRNA present in cells (Wirbelauer et al., 2000; Bashir et al., 2004; Bashir and Pagano, 2004; Kurland and Tansey, 2004; Wei et al., 2004). This observation has led to the discovery that the overall levels of Skp2 protein expressed in cells are, in part, determined by changes in the rate of Skp2 degradation (stability) (Wirbelauer et al., 2000; Bashir et al., 2004; Bashir and Pagano, 2004; Kurland and Tansey, 2004; Wei et al., 2004). To date, very little is known about the transcription regulation of Skp2 mRNA. A recent study by Imaki et al. (2003) has demonstrated that a GA-binding protein is involved in regulating the promoter activity of the mouse Skp2 gene.
Owing to the central role of Skp2 in determining the degradation of a number of key cell cycle regulatory proteins, aberrant Skp2 expression is thought to play an active role in oncogenesis. Several additional lines of evidence support this notion. Skp2 cooperates with H-Ras in transforming primary rodent fibroblasts and in inducing tumor formation in nude mice (Gstaiger et al., 2001). Ectopic expression of Skp2 blocks cell differentiation (Dow et al., 2001; Koga et al., 2003) and induces anchorage-independent cell growth (Carrano and Pagano, 2001). Skp2 expression is greatly induced in a broad spectrum of human tumors and the levels of Skp2 mRNA and protein directly correlate with the grade of malignancy (Latres et al., 2001; Bloom and Pagano, 2003; Oliveira et al., 2003). Given these links, very little is known about the molecular mechanisms involved in the elevated Skp2 expression in malignant cells. In this report, we isolated and characterized a 4237-bp human Skp2 promoter sequence, and identified an E2F-responvie element, hSRE2, which plays a critical role in regulating the human Skp2 promoter activity in both normal and transformed cells.
Isolation and characterization of human Skp2 promoter
Skp2 mRNA and protein content is cell cycle regulated with highest expression levels at the G1/S transition and early S phase. There is evidence that this regulation occurs at the transcriptional and post-translational levels. Skp2 gene expression is constitutively elevated in many transformed human cells and primary tumors. It seems likely that aberrant regulation at one or both of these levels may contribute to the transformed phenotype.
In order to evaluate the mechanisms and factors that are involved in the transcriptional regulation of Skp2, we isolated 4237 bp of the human Skp2 gene as described under Materials and methods. The sequence of the 4237-bp PCR product has been submitted to GenBank (Accession Number DQ021501). In Figure 1aFigure 1, we show the proximal promoter sequence encompassing ~500 bp upstream of the ATG translational start site of the human Skp2 gene. Comparison of the 4237-bp human Skp2 promoter to that of mouse Skp2 promoter reported by Imaki et al. (2003) revealed that the two promoters shared strong sequence similarity (~67%) only within the ~500-bp proximal promoter regions (see Discussion). Using primer extension analysis, we localized the transcription start site to a region 186-bp upstream of the translational start site (Figure 1bFigure 1). The human Skp2 promoter did not contain any recognizable TATAA and CCAAT box sequences. Comparison to a transcription factor binding site database (confidence of 85) identified several potential regulatory elements including binding sites for transcription factors c-Rel/NF-kap, AP1, SRY, Sp1, GATA, and E2F. Also shown are the restriction sites used for constructing the CAT reporter vectors and the EMSA probes; these are indicated as N for NotI; S for SstII; A for ApaI, and P for PstI throughout this report. To avoid complications from possible translational effects, we removed the ATG translational start sequence at the 3′ end of the DNA fragment by using PstI ( + 212) restriction enzyme digestion before inserting the ~3.5-kb promoter fragment into pKSCAT reporter plasmid to generate p3CAT. In addition, we utilized both a 5′- and a 3′-series of deletion constructs to locate important regulatory regions (Figure 2aFigure 2). The promoter activity of these CAT constructs was evaluated following transient transfection into HeLa cells. As shown in Figure 2bFigure 2, the full-length sequence (p3CAT) possessed high level promoter activity. Progressive deletion of the ~3.5-kb of the 5′ flanking region down to the SstII site at −59 had little effect on promoter activity (pSPCAT). By contrast, the 3′- deletions results in substantial reduction in promoter activity when a specific region between + 65 to + 149 was deleted (cf pSACAT to pSNCAT). Consistent with the primer extension results, the region between + 78 to + 212 (del #19) did not have intrinsic promoter activity; however, this region contains the + 65 to + 149 segment that is critical for high level Skp2 promoter function.
Figure 1
Figure 1
Figure 1
(a) Comparative alignment of the human (top) and mouse (bottom) Skp2 promoter sequences (~500-bp) upstream of the ATG translation start site (in rectangle box) is presented. Transcription start sites ( + 1) are underlined in bold. Putative transcription (more ...)
Figure 2
Figure 2
Figure 2
(a) Schematic representation of 5′ and 3′ deletion constructs of the human Skp2 promoter-CAT reporter constructs. The coordinates of the Skp2 promoter present in each of the constructs are indicated in parenthesis. The restriction enzymes (more ...)
Increased Skp2 promoter activity in transformed cells correlates with elevated Skp2 mRNA levels
We screened a number of transformed human cell lines to identify those with increased Skp2 mRNA levels and Skp2 promoter activity compared to normal human diploid fibroblast (HF) cells. As shown in Figure 3aFigure 3, three of the four transformed cell lines (HeLa, MG63, U343) have increased Skp2 mRNA levels compared to the levels in normal fibroblast HF cells. Skp2 mRNA levels in the fourth transformed cell line, U87, were similar to the levels in HF cells. The southern analysis showed that the increased Skp2 content was not due to gene amplification, but more likely reflected increased transcription or mRNA stability. To distinguish these two possibilities, we tested whether the Skp2 promoter activity corresponded to the Skp2 mRNA levels in the transformed cell lines using transient transfection. We opted to use the pSPCAT promoter for these initial studies since its activity in HeLa cells is similar to the activity detected using the entire ~3.5-kb promoter. We also evaluated the pSNCAT promoter activity since this construct lacked the region critical for high-level promoter function and potentially represented the basal Skp2 promoter function. Finally, we included the SV40 early promoter (pSV40CAT) as a control for transfection efficiency or variability in transcription activity of the cell lines. Strikingly, the Skp2 promoter pSPCAT that contained the region + 65 to + 212 was expressed at high levels in all three transformed cell lines that had elevated Skp2 mRNA levels. By contrast, pSPCAT promoter activity was substantially lower in the U87 and HF cell lines where Skp2 mRNA levels were low (Figure 3bFigure 3). Collectively, these data support the notion that the increased Skp2 expression is associated with increased Skp2 promoter activity that leads to accumulation of Skp2 mRNA in these transformed cells.
Figure 3
Figure 3
Figure 3
(a) Top panel: Measurement of the levels of Skp2 mRNA in four human transformed (HeLa, MG63, U343, U87) and nontransformed (HF) cell lines by Northern blot. 36B4 was used to normalize sample loading. In total, 20 μg of total RNA was loaded per (more ...)
E2F positively regulates Skp2 promoter
The increased Skp2 promoter activity in transformed cells mapped to a minimum region from + 65 to + 212 (Figure 2bFigure 2). As noted in Figure 1bFigure 1, the region from + 65 to + 212 contained two of the three putative E2F recognition sites identified through computer analysis of transcription factor binding sites database; this raised the possibility that E2F might play a major role in regulating Skp2 gene transcription in transformed cells. This notion is supported by the findings of dysregulation of expression/or activity of E2F family members in many human tumors. To explore this regulatory link between E2F and Skp2 genes, we examined whether E2F controlled Skp2 gene expression and promoter activity. First, we ectopically expressed E2F1 using adenovirus infection of HF and U87 cells that normally have low levels of Skp2 and E2F1. As shown in Figure 4aFigure 4 (left panel), there was a substantial increase in Skp2 mRNA levels in E2F1-expressing cells compared to control cells.
Figure 4
Figure 4
Figure 4
(a) Ectopic E2F1 expression induces expression of the endogenous Skp2 gene. Human diploid fibroblast (HF) and human glioblastoma (U87) cell lines were infected with adenovirus containing either empty expression vector or E2F1 expression vector. The cells (more ...)
To investigate the mechanism for E2F activation of Skp2 gene expression, we tested whether the E2F1 expression vector could transactivate the Skp2 promoter. We used pSPCAT, pSACAT, and pSNCAT reporter constructs to identify which, if any, of the three putative E2F sites (referred to as human Skp2 regulatory element hSRE in this report) were responsive to E2F transactivation (Figure 4bFigure 4). The promoterless pKSCAT, as well as a Myogenin promoter CAT construct (MyoG-CAT), were used as negative controls since neither is known to be transactivated by E2F1. As shown in Figure 4bFigure 4, the pSPCAT construct that contained all three hSRE sites was transactivated by E2F1. Deletion of hSRE3 (pSACAT) did not change the transactivation ability; however, further deletion of the hSRE2 (pSNCAT) drastically reduced transactivation by E2F1. As expected, neither the MyoG promoter nor the promoterless CAT vector could be transactivated by E2F1. Together, these data pointed to hSRE2 as a key element for the E2F responsiveness of the Skp2 promoter and supported that the effect of ectopic E2F1 expression to induce endogenous Skp2 mRNA expression (Figure 4aFigure 4) occurred in part through transcriptional activation of the Skp2 gene.
E2F directly interacts within the human Skp2 promoter
We next asked whether the effect of E2F on the human Skp2 promoter was a direct one. To address this question, we first tested whether there was a direct interaction between E2F and Skp2 promoter in vivo using chromatin immunoprecipitation (ChIP) assay (Figure 5Figure 5). To this end, HeLa chromatin extract was immunoprecipitated with E2F1 or E2F4 antibody and analysed by PCR amplification using primers specific for the relevant Skp2 promoter region or for the myogenin promoter as a negative control. As expected, the myogenin promoter that is not regulated by E2F was not detected in the immunoprecipitated chromatin. However, the Skp2 promoter was present within this fraction, confirming that E2F family members regulate the Skp2 promoter in vivo.
Figure 5
Figure 5
Figure 5
Demonstration of direct interaction between E2F and human Skp2 promoter in vivo. ChIP assay was performed on HeLa cells as described in Materials and methods. Opposing arrows indicates locations of the PCR primer sets used to specifically detect the human (more ...)
Data presented in Figures 4Figure 4 and and55Figure 5 strongly indicated a direct association of E2F proteins to the hSRE containing proximal promoter region of the human Skp2 gene. To map which of the three hSRE sites were recognized by E2F, we utilized three DNA probes each containing a single hSRE: SN (hSRE1), NA (hSRE2), AP (hSRE3) in EMSA studies (Figure 6aFigure 6). Since E2F binds DNA as a heterodimer with E2F-related DP protein, we used purified GST-E2F1 and -DP1 as a protein source in the initial study. As a positive control for E2F binding, we utilized the well-characterized E2F site from adenovirus E2 promoter (referred to as E2 probe in this study). As shown in Figure 6bFigure 6, the E2 and NA probes readily formed nucleoprotein complex with GST-E2F1 and -DP proteins, whereas the SN or the AP probe had little or no binding. We used mutated E2 and NA probes to evaluate the sequence specificity of E2F/DP binding to the NA probe. Since E2F interacts with its DNA targets through the GCG core sequence, we replaced the GCG-rich nucleotides with TA to generate the mutant probes (see legend, Figures 6cFigure 6 and and7a).7aFigure 7). As shown in Figure 6cFigure 6, the adenovirus E2 site readily competed for binding of E2F/DP to the NA probe. However, mutation of the adenovirus E2F sequence (mE2) or of the Skp2 hSRE2 sequence (mNA) completely abolished the ability of these DNAs to compete for E2F/DP binding to NA. Likewise, the mNA DNA probe failed to form nucleoprotein complex with E2F/DP proteins. These results confirmed and extended the ChIP and the transactivation studies that identified hSRE2 as a critical cis-element for E2F activation of the Skp2 promoter.
Figure 6
Figure 6
Figure 6
(a) Schematic illustration of the DNA probes used to analyze DNA–protein interaction by EMSA. SN, NA, and AP probes contained single hSRE1, hSRE2 site, and hSRE3 site of the human Skp2 promoter, respectively. The mNA probe is identical to NA except (more ...)
Figure 7
Figure 7
Figure 7
(a) Mutational analysis of Skp2 promoter. Top panel: schematic illustration of the sequences and positions of the wild-type and mutant hSREs within the pSPCAT construct. The lower inset shows the specific mutation of each hSRE site (underlined). Bottom (more ...)
We also investigated the binding of nuclear extracts to the hSRE2 site and compared them to binding of the extracts to the adenovirus consensus E2F site. Binding to the latter has been extensively characterized and results in the formation of three distinct E2F-containing nucleoprotein complexes; in order of mobility, they are the S phase complex, the G0/G1 complex, and the free E2F heterodimers (Figure 6dFigure 6, left panel). We have confirmed the identity of each of these three complexes by supershift or blocking EMSA assay with antibodies specific for each complex as have been previously described (Devoto et al., 1992; Lees et al., 1992; Smith et al., 1996). The NA fragment containing only the hSRE2 site was able to compete for the formation of all three complexes, whereas the mNA, SA, and AP fragments were unable to compete for binding (Figure 6dFigure 6, left panel).
Direct binding of the nuclear extracts to hSRE2 probe resulted in the formation of two complexes (A and B) with different mobility than the complexes detected with the E2 probe (Figure 6dFigure 6, right panel). Of these two complexes, only complex B was competed by the adenovirus E2F site. The sequence specificity of complex A was confirmed by the ability of NA, but not the mutated hSRE2 site (mNA) to compete for formation of the complex. Inclusion of an antibody specific for DP1, a heterodimeric partner of most E2F members, effectively blocked formation of the complex B, but not complex A (Figure 6eFigure 6). To determine that E2F was present in complex A, we extracted proteins from complex A and analysed their content by Western blot using antibodies to E2F1 and E2F4. As shown in Figure 6eFigure 6 (inset), both E2F1 and E2F4 were indeed present in complex A (inset, lane1). As a negative control, we did not detect E2F proteins from lanes that contained only nuclear extract without DNA probe (inset, lane2). Using the same assay, we tested for the presence of additional proteins that are known to be present in E2F-containing nucleoprotein complexes. Although we could detect DP1 in complex A, we were not able to detect the presence of the other proteins (e.g., Sp1, p130, p107, pRB, cyclin A, and cdk2) that we examined (data not shown). Previous studies of non-consensus E2F-binding sites have also identified complexes that contain E2F and DP1 but lack other proteins usually associated with E2F complexes (Weinmann et al., 2001; Wells et al., 2002). We could not detect formation of complexes A or B using hSRE1 or hSRE3 probes with the same extracts (data not shown). Collectively, these DNA-binding analyses support a role for the hSRE2 in E2F interaction with the human Skp2 promoter, and the importance of the GCG core sequence of hSRE2 for recognition by E2F-containing complexes.
Role of hSRE2 in human Skp2 promoter function
While neither hSRE1- nor hSRE3-containing DNA fragment was capable of forming nucleoprotein complexes with purified E2F/DP1 or nuclear extracts by in vitro EMSA analysis, we could not exclude the possibility that either one could have a supporting role in regulating the Skp2 promoter. Thus, we created a series of pSPCAT constructs that contained mutations of the core GCG sequence of each hSRE or various combinations (Figure 7aFigure 7, top panel). To assay the effects of these mutations, E2F1 transactivation assays of the Skp2 promoter were conducted in HeLa cells. As shown in the bottom panel of Figure 7aFigure 7, mutation of hSRE1, hSRE3, or both had no effect on the transactivation of the Skp2 promoter by E2F1. By contrast, mutation of hSRE2 in any context abolished the transactivation of the Skp2 promoter in the pSPCAT constructs, for example, mpSPCAT-2, mpSPCAT-1/2, and mpSPCAT-2/3. These data unequivocally confirmed the essential role of hSRE2 element in mediating activation of the Skp2 promoter by E2F1 and ruled out an ancillary role for the other E2F sites.
Since E2F1 activity and Skp2 expression is increased during progression from G1 to S phase, we evaluated the role of the E2F response element hSRE2 in cell-cycle-dependent activation of the Skp2 promoter. NIH3T3 cell line stably transfected with pSPCAT constructs with either intact or a mutated hSRE2 element (mpSPCAT-2) were used for this study. Cells were synchronized to quiescence by serum deprivation followed by release into cell cycle by the addition of serum. At various times after addition of serum, RNA was isolated and the levels of CAT mRNA and endogenous Skp2 mRNA were measured (Figure 7bFigure 7). We observed an induction of the CAT mRNA levels from the wild-type construct in a time-dependent manner parallel to that of the endogenous Skp2 mRNA levels. By contrast, mutation of hSRE2 significantly impaired activation of the Skp2 promoter during the G1 to S phase transition, supporting a role for hSRE2 and E2F in cell-cycle-dependent activation of Skp2.
Next, we examined if hSRE2 was required for the increased Skp2 promoter activity observed in the Skp2 overexpressing transformed cell lines, HeLa, MG63, and U343 (Figure 3bFigure 3). Mutation of hSRE2 drastically reduced the relative levels of promoter activities in these transformed cell lines (Figure 7cFigure 7). By contrast, there was little effect of the mutation on promoter activity in normal human fibroblasts (HF). Finally, we found that there was a substantial increase in expression of several E2F family members (E2F1-3) in the HeLa, MG63, and U343 transformed cell lines compared to normal human HF cells (Figure 7dFigure 7, left panel). This was reflected in the amount of E2F complexes A and B that could be detected when nuclear extracts from the transformed cells were analysed for binding to hSRE2 site by EMSA (Figure 7dFigure 7, right panel). We also find increased levels of complexes A and B from additional human tumor cell lines that express high levels of E2F and Skp2, including U2OS (osteosarcoma), RD (embryonal rhabdomyosarcoma), and RH30 (alveolar rhabdomyosarcoma) (data not shown). Taken together, these data suggest a mechanistic basis linking elevated E2F expression to increased activation of the Skp2 promoter in many transformed cells.
Finally, we evaluated whether E2F activity is required for high-level Skp2 expression. In the first approach, we reduced endogenous E2F activity through ectopic expression of RB, a repressor of E2F activity, in C33A cells. These cells lack endogenous RB and have been previously utilized to study E2F activity (Ross et al., 2001). As shown in Figure 8aFigure 8, the E2F-responsive Skp2 promoter (pSACAT) had high-level activity in C33A cells. This activity was greatly reduced by RB expression. By contrast, the low-level activity of the Skp2 promoter lacking the hSRE2 (pSNCAT) was not affected by RB expression. In the second approach, we used previously described specific AS-ODN (Simile et al., 2004) to knockdown expression of E2F1 in HeLa cells. As shown in Figure 8bFigure 8, Skp2 mRNA and protein levels were substantially reduced when E2F1 levels were reduced by transfection of E2F1 AS-ODN. By contrast, transfection of S-ODN or unrelated control ODN had no effect on Skp2 mRNA or protein expression (Figure 8bFigure 8). Similar results were observed in C33A cell line (data not shown). Thus, endogenous E2F1 activity is required for the high-level Skp2 expression associated with human tumor cells as well as for high-level human Skp2 promoter activity.
Figure 8
Figure 8
Figure 8
(a) Inhibition of endogenous E2F-dependent transactivation of Skp2 promoter activity by RB. C33A cells were cotransfected with 2 μg of human Skp2 promoter CAT constructs (pSACAT and pSNCAT) with either empty pcDNA3 expression vector (15 μg) (more ...)
The SCFSkp2 E3 ligase complex is a critical regulator of cell cycle progression that targets several G1/S cell cycle regulators for degradation. Skp2, the substrate-binding subunit of the E3 ligase, determines the specificity of this complex and plays a key role in targeting proteins for 26S-proteasome-mediated degradation. Expression of Skp2 leads to the degradation of the major G1 cyclin kinase inhibitor p27Kip1 that controls G1/S transition by inhibiting cyclin A/E-associated cdk2 complexes. Given Skp2’s central role in promoting S phase entry, aberrant activation of Skp2 gene expression is detected in a large number of human cancers and the level of Skp2 activation is strongly correlated with tumor aggression and poor prognosis. Thus, identification of the regulatory mechanism(s) leading to increased Skp2 expression may offer new insight into the control of tumor cell proliferation.
Presently, the precise mechanisms involved in Skp2 upregulation in tumors remain poorly understood. Recent studies have reported that the entire chromosome 5 or a region of chromosome 5 containing the Skp2 gene is amplified in selective human tumors (Dowen et al., 2003; Yokoi et al., 2004; Coe et al., 2005). Thus, in some cases, amplification of the Skp2 gene results in higher expression of this gene in tumor cells. However, gene amplification is largely detected in tumors at metastatic stage, whereas an increased level of Skp2 expression can be readily detected in tumors even at early dysplastic stages. Thus, Skp2 gene amplification is likely to be associated with advanced stage of tumor progression while other control mechanisms not involving a gross genomic alteration are involved in the initial accumulation of Skp2 mRNA during oncogenesis. This implies a change in the rate of either transcription or mRNA degradation. Our preliminary analysis reveals no significant difference in the half-life of Skp2 mRNA between nontransformed and transformed human cell lines (Zhang and Wang, unpublished result). Therefore, one plausible mechanism is a dysregulated Skp2 promoter function.
In this report, we show for the first time that Skp2 promoter has increased activity in human tumor cell lines. Several lines of evidence indicates that E2F family members are directly involved in this aberrant activation: we have demonstrated that (1) there is a good correlation between elevated E2F levels and Skp2 mRNA levels in human tumor cell lines, (2) a direct activation of the human Skp2 promoter by E2F1 in transient transactivation assays suggests that the effect of increased E2F either in tumor cells or in ectopic expressing cells result from transcriptional activation, (3) E2F containing complexes bound to the Skp2 promoter can be detected in vivo using the ChIP assay, and direct binding of GST-E2F/DP heterodimers to the Skp2 promoter identifies hSRE2 site as critical for this binding, (4) a role of E2F activation of Skp2 promoter in transformed cells is indicated by our finding that mutation of the hSRE2 strongly reduces Skp2 promoter activity in tumor cells, (5) ectopic expression of E2F1 increases endogenous Skp2 mRNA content in normal human diploid fibroblasts, and (6) inhibition of E2F1 activity by RB expression or by E2F1 antisense oligonucleotides significantly reduces Skp2 mRNA levels in human tumor cell lines.
Functional studies using deletion and mutation strategies indicate that the hSRE2 site in the 5′ untranslated region of exon 1 is critical for the activation of the Skp2 promoter by E2F in both normal and transformed cells (Figure 7Figure 7). It is rare but not unprecedented for cell cycle genes to be transcriptionally regulated by factors acting downstream of the transcription start site (Kel et al., 2001). Studies of a large number of well-characterized E2F target genes suggest that, in general, E2F regulates cell-cycle-dependent gene expression by three modes of action: active induction, derepression, and active repression (Dyson, 1998; Koziczak et al., 2000; Muller et al., 2001). In the active induction model, a mutation of the E2F site prevents binding of active E2Fs (e.g., E2F1-3), thus leading to a loss of transcriptional activation by E2F. By contrast, a mutation of the E2F site in both derepression and active repression models results in early and/or constitutive activation of the promoter function. In our study, we showed that the mutation of the hSRE2 site in the Skp2 promoter did not increase promoter activity in normal cells and greatly reduced promoter activity in tumor cells. Thus, we believe that E2F most likely regulates Skp2 promoter function through an active induction mechanism where gene transcription is induced upon replacement of repressive E2Fs (e.g., E2F4-5) by active E2Fs (e.g., E2F1-3). This notion is supported by our finding in asynchronous cells of a direct interaction of both active (e.g., E2F1) and repressive (e.g. E2F4) E2Fs with the Skp2 promoter using in vitro EMSA and in vivo ChIP assays.
The role of E2F in regulating the transcription of genes that play a critical role in cell cycle progression (e.g., G1/S transition) and DNA replication is well established. Cell-cycle-dependent gene activation by E2F at the G1–S transition is principally determined through its association with the pocket proteins (pRB, p107, p130) (Helin et al., 1993; Qin et al., 1995; Frolov and Dyson, 2004). However, there is now increasing evidence to suggest that E2F may interact with other regulatory proteins in complexes that target promoters through noncanonical E2F DNA-binding sites (Kel et al., 2001; Muller et al., 2001; Weinmann et al., 2001). In this report, we have shown that a unique E2F-containing protein complex A binds to the hSRE2 site of the Skp2 promoter. Although this complex is not competed by the canonical adenoviral E2F-binding site, mutation of the hSRE2 central GCG core sequence, characteristic for E2F binding, disrupts binding to hSRE2. The behavior of complex A is similar to a previously described E2F-containing protein complex that interacts specifically with the b-myb promoter (Weinmann et al., 2001). In that study, the complex in question was effectively competed only by atypical E2F DNA-binding sequences identified from E2F ChIP assay, and does not appear to contain typical E2F-interacting proteins such as pRB or other pocket proteins. Other studies using ChIP assays have revealed that many E2F bound promoters do not contain elements closely matching to E2F consensus sequences (Weinmann et al., 2001; Wells et al., 2002). Thus, there is a growing evidence to support an expanded model of E2F action involving combinations of novel positive and negative regulatory E2F-interacting proteins other than pocket proteins bound to E2F-binding sites. Presumably, these are responsible for determining variable E2F target gene expression at different stages of the cell cycle, and have been hypothesized to also play a role in regulating non-cell-cycle-associated gene expression.
It is intriguing that the Skp2 promoter is activated by E2F1 since one of the targets of Skp2-mediated degradation during the late stages of S phase is E2F1 (Marti et al., 1999). Our results suggest a novel regulatory feedback loop between Skp2 and E2F in normal cell cycle progression, where E2F and Skp2 mutually control the expression of each other. The coupling of E2F1 as the positive regulator of Skp2 expression at G1/S boundary with Skp2 as the negative effector on E2F1 degradation in late S phase would serve as vital checkpoint control on cell cycle progression. Disruption of this regulatory loop may play a significant role in uncontrolled cell proliferation in human tumor cells. Indeed, many human tumors exhibit elevated expression of both E2F and Skp2 genes. Disruption of the E2F/RB regulatory pathway is known to be a major contributor to increased E2F action in many human tumors. Our finding that E2F directly controls the Skp2 promoter offers a novel mechanism for the upregulated Skp2 expression seen in tumor cells. Given the regulatory loop mentioned above, one might question why the elevated Skp2 expression in tumors does not lead to increased degradation of E2F and ultimately low E2F level in tumor cells. It is presently unknown whether the E2F proteins are more stable in tumor cells with elevated Skp2. Moreover, ubiquitination and degradation of E2F1 has been shown to also occur by Skp2-independent mechanism (Ohta and Xiong, 2001). Thus, the proposed regulatory loop does not need to be a closed or exclusive loop. Since tumor cells frequently exploit multiple mechanisms to evade normal regulatory processes that limit uncontrolled cell proliferation, it will not be surprising if a dysregulation of Skp2-dependent and/or -independent destruction of E2F does exist. Future studies will be needed to address these possibilities and determine the mechanisms involved.
Although the sequence of the proximal 500 nts of the mouse and human Skp2 promoters is highly conserved, it is notable that there are differences in nearly all putative transcription factor-binding sites identified in the TRANSFAC database (Figure 1aFigure 1). While this study of the human promoter points to the region downstream of the transcription start site as critical for high-level and cell cycle expression, a previous study of the mouse promoter points to the region upstream of the transcription start site. In the later study, a GABP response element is identified as the critical element for mouse promoter function (Imaki et al., 2003). We cannot reconcile this difference of their finding to ours that shows an E2F response element (hSRE2) as crucial for human promoter function. However, it is notable that the sequences for the GABP and E2F response elements, although present in similar positions of the mouse and human promoters, are not identical. Thus, it is possible that these factors may not be able to recognize and regulate the different promoters. This might suggest that different regulatory mechanisms exist for the cell cycle regulatory response in different species. If true, this might account for some aspects of the well-known failure of successful murine tumor therapies when translating into treatment for human tumors. Clearly, additional studies are required to address these intriguing findings. Finally, it should be noted that our results are in agreement with and validate the results of Vernell et al. (2003) who reported that Skp2 was a target for E2F activation in asynchronous cultures. It is not clear why other studies searching for E2F targets using ChIP-coupled microarray technology were not able to identify Skp2 (Ren et al., 2002; Weinmann et al., 2002). Clearly, these types of global strategies cannot be expected to identify all targets and the results are critically dependent on the experimental paradigm used to design the study as suggested by Vernell et al.
Reagents
We obtained DNA constructs containing CMV-E2F and CMV-RB from Dr Srilata Bagchi (University of Illinois at Chicago), human p27kip1 cDNA from Dr Kei-ichi Nakayama (Medical Institute of Bioregulation, Kyushu University, Fukuoka, Japan), and human p45Skp2 from Dr Hideyo Yasuda (Tokyo University of Pharmacy and Life Sciences, Japan). Adenoviruses containing CMV-expression vector and CMV-E2F vector were obtained from Dr Srilata Bagchi. Antibodies against E2F-1 (KH95) and E2F4 (C-108) for ChIP and against DP-1 (K-20) for EMSA were from Santa Cruz. The promoter for human Skp2 gene was obtained by PCR amplification of human diploid fibroblast genomic DNA using the following primers: (forward) 5′ ACCAAGAGCTGAGTTGGCGA 3′ and 5′ GAGAGAGACAGGGCAATCATACAC 3′ (reverse). The PCR reaction was carried out in the following steps: 1 one cycle at 94°C for 40 s→35 cycles of denaturing (94°C/40 s), annealing (55°C/1 min) and extension (72°C/4 min)→ 1 cycle 72°C at 5 min. The DNA sequence of two independent clones was compared to the NIH and Celera databases to verify the identity and accuracy of our PCR products.
Cell culture
All cell lines were maintained in Dulbecco’s modified Eagle’s high glucose medium supplemented with 200 U/ml of penicillin, 50 μg/ml of streptomycin, 1 mm glutamine, and 10% (v/v) bovine calf serum (NIH3T3 cells) or 10% fetal bovine serum (HeLa, MG63, U343, U87, human foreskin diploid fibroblast HF). NIH3T3 cells stably expressing CAT reporter constructs were established by CaPO4 transfection method followed by 3 weeks of 400 μg/ml G418 drug selection. Both clonal cell lines and pooled cells were prepared and analysed to show similar findings. For synchronized cell studies, the cells were arrested in G0 state by maintaining cells at 0.5% serum in methinoine-free media for 30 h. To induce re-entry into cell cycle, cells were rinsed with PBS and fed with 10% serum containing media. At various time points, cells were collected and poly A+-mRNA was prepared using RNA purification kit from Ambion under conditions recommended by the manufacturer. The RNA was analysed for Skp2, CAT, and 36B4 mRNA by Northern blot analysis.
Primer extension analysis
The primer extension analysis was performed under conditions recommended by the manufacturer (Promega). Reverse transcription was carried out in the presence of 10 μg of poly (A) + mRNA isolated from RH4 cells and 1 ng end-labeled Skp2 gene specific reverse primer (5′ TTTAAAATACGTG CATTAA 3′). In the parallel reaction, dideoxynucleotide-based sequencing analysis was carried out using the same reverse primer and human Skp2 promoter fragment as the DNA template. Products from the sequencing and extension reactions were analysed together by 6% denatured urea-polyacrylamide gel electrophoresis. The gel was then dried and exposed to X-ray film. Same results were obtained from primer extension analysis using RNA samples isolated from other human cell lines (e.g., HF, HeLa, U343, UU87).
CAT assay
For transient transfection, cells (HeLa, MG63, U343, U87, HF) were plated at a density of 1 × 106 cells per 100 mm tissue culture dish 24 h prior to transfection. Transient transfection was carried out with a total of 20 μg of DNA that includes 1–3 μg of β-galactosidase DNA (LacZ) driven by CMV promoter as an internal standard for monitoring transfection efficiency. Cells were exposed to DNA-CaPO4 precipitate for 17 h before they were transferred to growth medium for an additional 30–48 h. Cell lysates for LacZ and CAT assays were prepared as described previously (Ausubel, 1990; Cao and Wang, 2000). Deacetylase activity was inactivated by heating the lysates to 60°C for 7 min. (Mercola et al., 1985). A typical CAT assay reaction mixture consisted of 0.7 μg Acetyl-CoA, 0.2 μCi of [C14]chloramphenicol (1 Ci = 37 GBq), and cell lysates in a final volume of 200 μl. The amount of cell lysate used in each CAT reaction was standardized by the β-galactosidase activity. Routine CAT assays were carried out at 37°C for 1–3 h and terminated by extraction with 1 ml of ice-cold ethyl acetate. Overnight incubation was used when the transfection efficiency of a particular cell line was low. Quantitative analysis of CAT activity was carried out by phosphoimager analysis.
EMSA
EMSA was performed by preincubating bacterial expressed GST-fusion protein (50 ng) or nuclear extract (5–10 μg) with nonspecific salmon sperm DNA (0.05 μg/ml) in a binding buffer containing 20 mm HEPES pH 7.4, 0.5 mm EDTA, 10% glycerol, 5 mm DTT, 1 mm MgCl2, and 50 mM KCl for 5 min on ice. Routinely, 0.2 ng of a 32P-labeled DNA probe prepared by Klenow end labeling was added to the EMSA reaction mixture and allowed to form DNA–protein complexes during a 20-min incubation at room temperature. The reaction complex was subjected to electrophoresis on a 4% native polyacrylamide gel at 4°C in 0.25 × Tris-Boric acid buffers at 200 V until the bromophenol blue dye reached the bottom. Gel was dried and autoradiographed. E2F proteins were detected in complex A using EMSA and Western blot analysis. On a preparative EMSA gel, we ran the nuclear extract with and without the DNA probe in separate lanes. Following visualization of complex A by autoradiography, we cut out analogous portions of the two samples from the gel and isolated the proteins from each. The samples were then subjected to SDS-PAGE followed by Western blot analysis using antibodies specific for E2F1 or E2F4.
Northern and Southern hybridization
Total RNA was prepared from culture cells by Trizol extraction method according to the manufacturer’s recommendations (Invitrogen). Total RNA isolated from cells was denatured, electrophoresed on a formaldehyde–agarose gel, and blotted to nitrocellulose membrane. For normalization of sample loading, membrane was stripped and hybridized with nick-translated 36B4 DNA probe. Total genomic DNA was prepared as previously described (Blin and Stafford, 1976). A total of 20 μg of genomic DNA from different human cell lines was digested to completion with EcoRI. Digested DNA was size-fractionated by agarose gel electrophoresis and transferred to nitrocellulose membrane. The membrane was hybridized to nick-translated Skp2 DNA probe that detected a 3891-bp DNA fragment corresponding to nucleotide positions 69 278–73 169 of the human Skp2 gene (AC008942.6). For normalization, membrane was stripped and hybridized with nick-translated p27kip1 DNA probe that detected a 12.7 kb fragment corresponding to nucleotide positions 4561–17 281 of the human p27kip1 gene (AC008115.3).
Site-directed mutagenesis
We used the Quick change in vitro mutagenesis system to generate various mutant constructs under the conditions recommended by the supplier (Stratagene). All the constructs made using this method were verified by DNA sequence analysis. The sequences of the primers containing mutations (top strand shown) were: 5′ GGTATCTCGAAGATAGGTAAAGCTGC 3′ (mE2F-A); 5′ GTAGAGCCTTGCTATCGCAGTGG 3′ (mE2F-B); 5′ GCCGCCGCATACCAAAGCGGGAATC 3′ (mE2F-C).
ChIP
E2F-ChIP analysis was carried out using conditions previously described (Weinmann et al., 2001). In brief, a total of 1 × 108 HeLa cells were used for each immunoprecipitation reaction. Cells were crosslinked with 1% formaldehyde followed by chromatin isolation. Genomic DNA in the extract was sonicated to achieve an average of 500 bp in length and snapped frozen for storage until use. Extract containing sheared DNA was precleared with protein-G before specific antibody against E2F1 or E2F4 (1–3 μg/ml) were added to the extract. A fraction of the precleared extract served as total input control. Immunoprecipitated protein–DNA complexes were de-crosslinked and protein components removed by Proteinase K digestion. The released genomic DNA fragment was purified by PCR-purification kit under the condition recommended by the manufacturer (Qiagen).
PCR primers used for detecting hSkp2 promoter ChIP in HeLa cells were: 5′ GAGAGAGACAGGGCAATCATACAC 3′ (forward) and 5′ ATCACCAGAAGGCCTGCGGGCT 3′ (reverse). PCR primers used for detecting human Myogenin promoter were: 5′ TGCACAGGAGCCCGCCTGGGCCAG 3′ (forward) 5′ AACCTGAGCCCCCTCTAAGCTG 3′(reverse).
Antisense oligonucleotide (AS-ODN) and transfection
The AS-ODNs used to knockdown E2F1 and E2F4 were synthesized as s-oligonucleotides and purified by HPLC, as previously described (Simile et al., 2004). AS-ODN sequence for E2F1 was 5′ TAGATCCGATCCAGCTCAGTGACA 3′. Control ODNs contained the sense sequence complementary to AS-ODN and unrelated sequences 5′ CAGGTCTTTCATC TAGAACGATGCGGG 3′. HeLa cells were plated in six-well dish at 2 × 105 cells/well 1 day before transfection. Cells were transfected with 200 nm antisense, sense, or control ODN by lipofectamine method as recommended by the manufacturer (Invitrogen). Cells were harvested 48 h post-transfection to check the expression levels of the mRNA and protein.
Acknowledgments
We thank Drs Reed Graves, Peter Tontonoz, and Charles Stiles for their critical review of this manuscript. We also like to thank Ms Galina Grechkina for her technical support in EMSA analysis. The work is supported by grant from NIH (CA074907).
References
  • Amati B. Proc Natl Acad Sci USA. 2004;101:8843–8844. [PubMed]
  • Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA et al (eds). (1990). Current Protocols in Molecular Biology (Green and Siley-Interscience, New York), Vol. 1.
  • Bai C, Sen P, Hofmann K, Ma L, Goebl M, Harper JW, et al. Cell. 1996;86:263–274. [PubMed]
  • Bashir T, Dorrello NV, Amador V, Guardavaccaro D, Pagano M. Nature. 2004;428:190–193. [PubMed]
  • Bashir T, Pagano M. Cell Cycle. 2004;3:850–852. [PubMed]
  • Bhattacharya S, Garriga J, Calbo J, Yong T, Haines DS, Grana X. Oncogene. 2003;22:2443–2451. [PubMed]
  • Blin N, Stafford DW. Nucleic Acid Res. 1976;3:2303–2307. [PubMed]
  • Bloom J, Pagano M. Semin Cancer Biol. 2003;13:41–47. [PubMed]
  • Bornstein G, Bloom J, Sitry-Shevah D, Nakayama K, Pagano M, Hershko A. J Biol Chem. 2003;278:25752–25757. [PubMed]
  • Cao Y, Wang C. J Biol Chem. 2000;275:9854–9862. [PubMed]
  • Carrano AC, Eytan E, Hershko A, Pagano M. Nat Cell Biol. 1999;1:193–199. [PubMed]
  • Carrano AC, Pagano M. J Cell Biol. 2001;153:1381–1390. [PubMed]
  • Ciechanover A, Schwartz AL. Biochim Biophys Acta. 2004;1695:3–17. [PubMed]
  • Coe BP, Henderson LJ, Garnis C, Tsao MS, Gazdar AF, Minna J, et al. Genes Chromosomes Cancer. 2005;42:308–313. [PubMed]
  • Deshaies RJ. Annu Rev Cell Dev Biol. 1999;15:435–467. [PubMed]
  • Devoto SH, Mudryj M, Pines J, Hunter T, Nevins JR. Cell. 1992;68:167–176. [PubMed]
  • Dow R, Hendley J, Pirkmaier A, Musgrove EA, Germain D. J Biol Chem. 2001;276:45945–45951. [PubMed]
  • Dowen SE, Neutze DM, Pett MR, Cottage A, Stern P, Coleman N, et al. Oncogene. 2003;22:2531–2540. [PubMed]
  • Dyson N. Genes Dev. 1998;12:2245–2262. [PubMed]
  • Frolov MV, Dyson NJ. J Cell Sci. 2004;117:2173–2181. [PubMed]
  • Gstaiger M, Jordan R, Lim M, Catzavelos C, Mestan J, Slingerland J, et al. Proc Natl Acad Sci USA. 2001;98:5043–5048. [PubMed]
  • Helin K, Harlow E, Fattaey A. Mol Cell Biol. 1993;13:6501–6508. [PubMed]
  • Hershko A, Ciechanover A. Annu Rev Biochem. 1998;67:425–479. [PubMed]
  • Imaki H, Nakayama K, Delehouzee S, Handa H, Kitagawa M, Kamura T, et al. Cancer Res. 2003;63:4607–4613. [PubMed]
  • Jin J, Harper JW. Cancer Cell. 2003;3:517–518. [PubMed]
  • Kamura T, Hara T, Kotoshiba S, Yada M, Ishida N, Imaki H, et al. Proc Natl Acad Sci USA. 2003;100:10231–10236. [PubMed]
  • Kel AE, Kel-Margoulis OV, Farnham PJ, Bartley SM, Wingender E, Zhang MQ. J Mol Biol. 2001;309:99–120. [PubMed]
  • Kim SY, Herbst A, Tworkowski KA, Salghetti SE, Tansey WP. Mol Cell. 2003;11:1177–1188. [PubMed]
  • Kipreos ET, Pagano M. Genome Biol. 2000;1:REVIEWS3002. [PubMed]
  • Koepp DM, Harper JW, Elledge SJ. Cell. 1999;97:431–434. [PubMed]
  • Koga H, Harada M, Ohtsubo M, Shishido S, Kumemura H, Hanada S, et al. Hepatology. 2003;37:1086–1096. [PubMed]
  • Koziczak M, Krek W, Nagamine Y. Mol Cell Biol. 2000;20:2014–2022. [PubMed]
  • Kurland JF, Tansey WP. Cancer Cell. 2004;5:305–306. [PubMed]
  • Latres E, Chiarle R, Schulman BA, Pavletich NP, Pellicer A, Inghirami G, et al. Proc Natl Acad Sci USA. 2001;98:2515–2520. [PubMed]
  • Lees E, Faha B, Dulic V, Reed SI, Harlow E. Genes Dev. 1992;6:1874–1885. [PubMed]
  • Li X, Zhao Q, Liao R, Sun P, Wu X. J Biol Chem. 2003;278:30854–30858. [PubMed]
  • Marti A, Wirbelauer C, Scheffner M, Krek W. Nat Cell Biol. 1999;1:14–19. [PubMed]
  • Mendez J, Zou-Yang XH, Kim SY, Hidaka M, Tansey WP, Stillman B. Mol Cell. 2002;9:481–491. [PubMed]
  • Mercola M, Goverman J, Mirell C, Calame K. Science. 1985;227:266–270. [PubMed]
  • Muller H, Bracken AP, Vernell R, Moroni MC, Christians F, Grassilli E, et al. Genes Dev. 2001;15:267–285. [PubMed]
  • Nakayama K, Nagahama H, Minamishima YA, Matsumoto M, Nakamichi I, Kitagawa K, et al. EMBO J. 2000;19:2069–2081. [PubMed]
  • Nakayama K, Nagahama H, Minamishima YA, Miyake S, Ishida N, Hatakeyama S, et al. Dev Cell. 2004;6:661–672. [PubMed]
  • Ohta T, Xiong Y. Cancer Res. 2001;61:1347–1353. [PubMed]
  • Oliveira AM, Okuno SH, Nascimento AG, Lloyd RV. J Clin Oncol. 2003;21:722–727. [PubMed]
  • Pickart CM, Eddins MJ. Biochim Biophys Acta. 2004;1695:55–72. [PubMed]
  • Qin XQ, Livingston DM, Ewen M, Sellers WR, Arany Z, Kaelin WG., Jr Mol Cell Biol. 1995;15:742–755. [PubMed]
  • Ren B, Cam H, Takahashi Y, Volkert T, Terragni J, Young RA, et al. Genes Dev. 2002;16:245–256. [PubMed]
  • Ross JF, Naar A, Cam H, Gregory R, Dynlacht BD. Genes Dev. 2001;15:392–397. [PubMed]
  • Simile MM, De Miglio MR, Muroni MR, Frau M, Asara G, Serra S, et al. Carcinogenesis. 2004;25:333–341. [PubMed]
  • Smith EJ, Leone G, DeGregori J, Jakoi L, Nevins JR. Mol Cell Biol. 1996;16:6965–6976. [PubMed]
  • Sutterluty H, Chatelain E, Marti A, Wirbelauer C, Senften M, Muller U, et al. Nat Cell Biol. 1999;1:207–214. [PubMed]
  • Tedesco D, Lukas J, Reed SI. Genes Dev. 2002;16:2946–2957. [PubMed]
  • Tsvetkov LM, Yeh KH, Lee SJ, Sun H, Zhang H. Curr Biol. 1999;9:661–664. [PubMed]
  • Vernell R, Helin K, Muller H. J Biol Chem. 2003;278:46124–46137. [PubMed]
  • von der Lehr N, Johansson S, Wu S, Bahram F, Castell A, Cetinkaya C, et al. Mol Cell. 2003;11:1189–1200. [PubMed]
  • Wei W, Ayad NG, Wan Y, Zhang GJ, Kirschner MW, Kaelin WG., Jr Nature. 2004;428:194–198. [PubMed]
  • Weinmann AS, Bartley SM, Zhang T, Zhang MQ, Farnham PJ. Mol Cell Biol. 2001;21:6820–6832. [PubMed]
  • Weinmann AS, Yan PS, Oberley MJ, Huang TH, Farnham PJ. Genes Dev. 2002;16:235–244. [PubMed]
  • Wells J, Graveel CR, Bartley SM, Madore SJ, Farnham PJ. Proc Natl Acad Sci USA. 2002;99:3890–3895. [PubMed]
  • Willems AR, Schwab M, Tyers M. Biochim Biophys Acta. 2004;1695:133–170. [PubMed]
  • Wirbelauer C, Sutterluty H, Blondel M, Gstaiger M, Peter M, Reymond F, et al. EMBO J. 2000;19:5362–5375. [PubMed]
  • Yeh KH, Kondo T, Zheng J, Tsvetkov LM, Blair J, Zhang H. Biochem Biophys Res Commun. 2001;281:884–890. [PubMed]
  • Yokoi S, Yasui K, Mori M, Iizasa T, Fujisawa T, Inazawa J. Am J Pathol. 2004;165:175–180. [PubMed]
  • Zhang H, Kobayashi R, Galaktionov K, Beach D. Cell. 1995;82:915–925. [PubMed]
  • Zheng N, Schulman BA, Song L, Miller JJ, Jeffrey PD, Wang P, et al. Nature. 2002;416:703–709. [PubMed]

PubMed articles by these authors
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph
See more articles cited in this paragraph