CDF-1-mediated repression of cell cycle genes targets a specific subset of transactivators
CDI-1-mediated repression of cell cycle genes targets a specific subset of transactivatorsJörk Zwicker+, Frances C. Lucibello, Valérie Jérôme, Sabine Brüsselbach and Rolf Müller*
Institut für Molekularbiologie und Tumorforschung (IMT), Philipps-Universität Marburg, Emil-Mannkopff-Strasse 2, D-35033 Marburg, Germany
Received September 8, 1997;Revised and Accepted October 19, 1997
ABSTRACT
The cdc25C, cyclin A and cdc2 genes are regulated during the cell cycle through two contiguous repressor binding sites, the CDE and CHR, located in the region of transcription initiation and interacting with a factor termed CDF-1. The target of this repression seems to be transcriptional activation of these promoters by transcription factors bound upstream. The majority of these factors falls into the class of glutamine-rich activators, suggesting that CDF-1-mediated repression might be activation domain specific. In the present study we have used chimeric promoter constructs to demonstrate that the cdc25C UAS, but not the core promoter, is crucial for repression. In addition, we show that only specific transcription factors and activation domains are responsive to CDE-CHR-mediated cell cycle regulation. These observations clearly indicate that CDF-1 interferes with activation of transcription by a specific subset of transactivators. The repressible activation domains belong to the same class of glutamine-rich activators, pointing to specific interactions of CDF-1 with components of the transcription machinery. In agreement with this conclusion we find that a simple inversion of the CDE-CHR module completely abrogates cell cycle-regulated repression.
We have chosen the human cdc25C gene (1 ) as a model system to investigate regulation of S/G2-specific transcription in mammalian cells (2 -5 ). To elucidate the mechanisms involved in cell cycle-regulated transcription of the cdc25C gene the human promoter was cloned and a comprehensive structure-function analysis was performed (3 ). Transient expression studies and in vivo footprinting studies led to the identification of two contiguous regulatory elements, termed the cell cycle-dependent element (CDE) and the cell cycle genes homology region (CHR) (3 ,4 ,6 ). These elements are located near the transcription initiation sites and play a key role in periodic transcription of the cdc25C gene. The CDE and CHR are bound by a transcriptional repressor in G0/G1 which is released in S/G2 (3 ,4 ). The CDE-CHR interacting factor has been identified and termed CDF-1 (see accompanying paper by Liu et al.). The CDE apparently does not interfere with basal transcription from the core promoter (3 ). Its function is dependent on a stretch of upstream sequences that is needed for transcriptional activation (UAS). This led to the hypothesis that CDF-1 may function by regulating the activity of upstream activators in a cell cycle-dependent fashion (3 ,5 ). This conclusion is supported by the observation that the proteins interacting with the cdc25C UAS do not only bind constitutively in vivo, but in a heterologous context also activate transcription in a way that is not significantly influenced by the cell cycle (5 ).
The major transactivator of the cdc25C UAS is the transcription factor CBF/NF-Y, which binds to three sites 5' of the CDE (5 ). A second important transactivator is Sp1 (or other members of the Sp family), which interacts with two sites further upstream (5 ). Interestingly, the major activation domains in both Sp1 (7 ) and NF-Y (8 ,9 ) are glutamine rich and both factors are therefore likely to contact a similar set of basal transcription factors, TAFs or other components of the preinitiation complex (10 ). It cannot, however, be ruled out at present that CDF-1 simply functions by preventing protein contacts through steric hindrance (11 ). It is therefore of obvious importance to consider the question as to whether this repression mechanism is restricted to a certain class(es) of activation domains (12 ) and thus activator-basal complex contacts (10 ). In support of such a hypothesis is the observation that the heterologous SV40 enhancer, which is bound by multiple transactivating factors belonging to different classes (13 ), is less efficiently repressed than the cdc25C UAS (3 ).
The relevance of this question is stressed by the fact that the promoters of various other cell cycle genes, such as cdc2 and cyclin A, are also regulated through CDE-CHR elements and harbor multiple Sp1 and NF-Y sites in the UASs (4 ,14 ). In addition, cell cycle genes repressed by transcription factor E2F also show a conspicuous preference for Sp1 and NF-Y sites 5' of the E2F sites (14 ). It is therefore likely that the molecular basis for E2F- and CDE-mediated negative regulation is very similar and that repression of glutamine-rich activators like Sp1 and NF-Y is a common mechanism of cell cycle-regulated transcription.
In the present study we have addressed the putative activator specificity of CDF-1-mediated repression in detail. To this end we have investigated how an exchange of the cdc25C UAS or the core promoter with heterologous sequences would affect CDE-CHR-mediated repression. In addition, we have analyzed the responsiveness of specific transcription factors and activation domains to CDE-CHR-mediated cell cycle regulation. The results of these analyses are compatible with the conclusion that CDF-1 interferes with activation of transcription by a specific subset of transactivators, which all belong to the class of glutamine-rich activators. We also find that a simple inversion of the CDE-CHR module completely abrogates cell cycle-regulated repression. Taken together, these observations point to specific interactions of CDF-1 with components of the transcription machinery.
NIH 3T3 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS), penicillin and streptomycin. Cells were transfected by the DEAE-dextran technique. For synchronization in G0 cells were maintained in serum-free medium for 2 days after transfection. Determination of luciferase activities were performed as published. Mel-ab cells (15 ) (provided by Prof. I.Hart, London, UK) were grown in DMEM plus 10% FCS, 200 nM TPA and 10 mM cholera toxin. C2C12 cells (16 ) (ATCC CRL-1772; obtained from Prof. H.H.Arnold, Braunschweig, Germany) were grown in DMEM plus 20% FCS. Synchronization of cells in G0/G1 was followed by FACS analysis as described (17 ).
The cdc25C promoter-luciferase constructs C290, C75, C33 and C20 as well the CDE mutants R1 and T7 (referred to as mCDE in this study) and the CHR mutant -6/-3 (referred to as mCHR in this study) have been described elsewhere (3 -5 ). Other constructs were generated by cloning of synthetic oligonucleotides with appropriate terminal overhangs or by PCR strategies as previously described (4 ,5 ,18 ). The following oligonucleotides were used (as five copies) for cloning of reporter constructs containing multiple transcription factor binding sites: NF-Y (Ea-Y, MHC class II promoter; 19 ), 5'-ATTTTTCTGATTGGTTAA;NF[kappa]-B; mouse [kappa] light chain enhancer (20 ), 5'-AGAGGGGACTTTCCGAGA; NF-1/CTF (high affinity binding site from adenovirus origin of replication; 21 ), 5'-TTTTGGCTACAAGCCAATA; Sp1, 5'-ATGGGGCGGAGA; (7 ); Gal4, CGGAGTACTGTCCTCCG (22 ). The oligonucleotides were sythesized with BamHI and BglII termini and cloned into the BamHI-digested cdc25C promoter construct C20 (3 ). The Gal4 expression vectors are based on pGal(1-147) (23 ) (GAYA plasmids, see Table 1 ) or pCG (22 ) (all other constructs). Chimeric promoters were generated by fusing the following enhancers or UASs to the cdc25C construct C20 (3 ): human troponin C promoter from -98 to +23 (24 ); human myf-4/myogenin promoter from -210 to +54 (25 ); human tyrosinase enhancer (-2.0 to -1.8 kb, EcoRI-NcoI fragment) (26 ); mouse TRP-1 promoter from -332 to -23 (27 ); SV40 promoter/enhancer from -281 to -45 (genomic region 273-36) (13 ); CMV enhancer (292-695 region from pcDNA3; Invitrogen). In the case of SV40 and CMV enhancers fusions were also made with the cdc25C constructs C33, C51 and C20/+30 (3 ). For inversion of the CDE-CHR in the cdc25C promoter the sequence 5'-GGGCTGGCGGAAGGTTTGAAT was changed to 5'-GTTCAAACCTTGCCT. Other constructs were cloned using PCR-generated promoter fragments as indicated in Results and the figures. All PCR-amplified fragments were verified by DNA sequencing using the dideoxynucleotide chain termination method using Sequenase 2.0 (US Biochemical) or Tth polymerase (Pharmacia).
We first addressed the question whether CDE-CHR-mediated repression might be dependent on a specific UAS. For this purpose the cdc25C UAS was exchanged with various heterologous enhancer sequences. Since it could not be excluded that the spacing between activators and repressor might be important, the heterologous sequences were fused with various cdc25C promoter fragments starting at -51, -33 or -20 and extending to +121. In addition, we used a fragment lacking most of the downstream sequence, extending to +30. All fragments were used in the wild-type form and, as controls, with mutated CDEs. The cdc25C promoter fragments were fused with both the cytomegalovirus (CMV) enhancer and the simian virus 40 (SV40) early enhancer/promoter regions and tested for cell cycle-regulated repression in transient luciferase assays in NIH 3T3 cells. As shown in Figure 1 , the chimeric CMV constructs did not show any significant cell cycle regulation, in contrast to the chimeric SV40 constructs. In the latter case cell cycle regulation was, however, only partially restored and ~3-fold below the induction value seen with the natural cdc25C promoter construct C290.
We next asked the question whether the core promoter might be of similar importance as the UAS. We use the term `core promoter' for the transcription initiation region of the cdc25C gene. This region, which shows basal promoter activity, extends from approximately +1 to +50 (see accompanying paper by Körner et al.). To address this issue we fused a cdc25C promoter fragment (-290 to +2) harboring the UAS and the CDE-CHR, but lacking the core promoter, to the core promoters of the human terminal deoxynucleotidyl transferase (TdT) (28 ) and SV40 early genes (13 ). As can be seen in Figure 2 , these chimeric promoters showed a similar cell cycle regulation as the wild-type promoter C290, indicating that CDE-CHR-mediated repression is not core promoter specific. In the subsequent experiments we therefore directed our attention to the observed activator specificity.
The observations described above suggest that CDE-CHR-mediated repression might work only in the context of specific activators. To address this question more directly we sought to investigate the function of the CDE-CHR module in the context of promoters which interact with only one transcription factor (family). To this end we constructed promoters where five binding sites for NF-Y, NF[kappa]-B, NF-I or Sp1 fused to the cdc25C minimal promoter construct C20 (-20 to +121). To be able to distinguish cell cycle effects exerted through the CDE-CHR from those due to other regulatory effects on the transactivators all constructs were generated with both a wild-type CDE and a mutated CDE. The constructs were tested in both growing and quiescent NIH 3T3 cells (Fig. 3 ). All four mutant constructs lacked any significant cell cycle regulation, as did the wild-type ones containing NF[kappa]-B or NF-I binding sites. In contrast, CDE-CHR-dependent cell cycle-regulated transcription was seen with NF-Y sites (3-fold) and, to a lesser extent, with Sp1 sites (2.2-fold). Since NF-Y and Sp1 are the transactivators of the cdc25C promoter (5 ), these observations may explain the observed requirement for a specific UAS. The fact that Sp1 is subject to cell cycle regulation via the CDE-CHR module may also offer an explanation for the observation that the SV40 promoter, which contains six Sp1 sites (13 ), was the only one which could replace the cdc25C UAS without totally obliterating cell cycle-regulated repression (Fig. 1 ).
In order to analyze the activator specificity of CDE-CHR-mediated repression in more detail we investigated the repressibility of specific transactivation domains fused to a fragment of Gal4 harboring the DNA binding domain (22 ,23 ). Plasmid vectors expressing fusion proteins of Gal4 and various activation domains (see Table 1 ) were co-transfected with a luciferase reporter containing five copies of a Gal4 binding site linked to the cdc25C minimal promoter C20 (-20 to +121). Luciferase activities were determined in both growing and quiescent NIH 3T3 cells with both wild-type and mutant CDE reporter plasmids (Fig. 4 ). CDE-CHR-mediated repression was seen with three different activation domains, the glutamine-rich domains of Oct-2 (Oct-N) (22 ), Sp1 (Sp1-Q1) (7 ) and different fragments of the Glu/Ser/Thr-rich activation domain of NF-Y (GAYA-5, -6 and -11) (8 ,29 ). The strongest cell cycle regulation was generally seen with GAYA-11, which contains the complete transactivation domain of the A subunit of human NF-Y (corresponding to the B subunit of rat CBF) (9 ,30 ). In contrast, no significant cell cycle-regulated repression was detected with other transactivation domains, i.e. ITF2 (31 ), VP-16 (22 ,32 ), Myc (33 ) and CTF (22 ,34 ). These domains are not glutamine rich and belong to the classes of Ser/Thr-rich, acidic or Pro-rich transcription factors (12 ). These data are in agreement with the results obtained in Figure 3 and confirm the conclusion that CDE-CHR-mediated repression is specific for a subset of transactivation domains. Significantly, the best repression was seen with NF-Y, which is the major transactivator of the cdc25C gene and also plays important roles in many other cell cycle genes. The fact that Sp1 (Fig. 3 ) as well as the Q1 domain of Sp1 (Fig. 4 ) are less efficiently repressed than NF-Y may, however, not reflect the physiological situation and could rather be due to the artificial experimental set-up, but this question is of minor importance with respect to the conclusions drawn from this study. The same applies to the fact that even the NF-Y-based constructs gave rise to a considerably lower cell cycle regulation than the natural cdc25C UAS. In addition, both Sp1 (7 ) and NF-Y (9 ) contain multiple activation domains, some of undefined nature, which may all contribute to cell cycle regulation in the context of the natural cdc25C promoter.
The specificity of repression reported above suggests that the CDE-CHR interacting factor establishes specific contacts with the transcription machinery. If so, the orientation of the CDE-CHR module should be important for its function. We therefore analyzed a series of constructs where the CDE-CHR module was inverted (Fig. 6 ). This inversion led to a complete abrogation of cell cycle-regulated repression (without affecting transcription levels in growing cells; not shown). The CDE-CHR inversion thus had a similar effect as mutation of the CDE, its inversion (and thus disruption of the repressor module) or mutation of both the CDE and CHR (Fig. 6 ). We have not formally shown that the inverted CDE-CHR still binds CDF-1, but previously published data strongly suggest that the nucleotides flanking the inverted sequence do not play a role in CDF-1 function (4 ). The finding that inversion of the CDE-CHR abrogates cell cycle regulation is therefore in line with the conclusion that CDE-CHR-mediated repression appears to involve stereospecific interactions with other transcription factors.
The data obtained in the present study suggest that CDE-CHR-mediated repression involves specific protein-protein interactions: (i) the cdc25C UAS could not be replaced by heterologous UASs or enhancers without partially or even completely impairing cell cycle-regulated repression; (ii) the glutamine-rich transactivation domains of NF-Y, Sp1 and Oct-2 were repressed in a cell cycle-dependent manner through the CDE-CHR module while several other activation domains did not show a comparable effect; (iii) inversion of the CDE-CHR sequence in the cdc25C promoter totally abrogated repression. It is therefore likely that the CDE-CHR binding factor CDF-1 or a protein associated with CDF-1 establishes contacts to a subset of activation domains, their associated co-activators (if any) or specific components of the basal machinery, such as basal transcription factors, TAFs or other cofactors that have been shown to be targeted by glutamine-rich transactivators. In this respect the cofactor PC-4 (35 ,36 ), hTAFII55 (37 ) and hTAFII130 (38 ), the human homolog of dTAFII110 (39 ) might be of particular interest. It has been shown that repression by the Drosophila protein Krüppel, which depends on interaction with TFIIE, is core promoter dependent and only functions in the context of a TATA box (40 ). The observation that repression by CDF-1 or associated factors is core promoter independent, at least for the two heterologous core promoters tested, does not rule out the possibility of a direct interaction with the basal machinery, since many of the components contained in the basal complexes formed on different core promoters are identical.
Several distinct mechanisms of transcriptional repression have been proposed (11 ), including a direct inhibition of general transcription factors (41 -43 ), a local change of chromatin structure near the promoter (44 -46 ) and inhibition of DNA binding by competition or steric hindrance (47 ,48 ). The mechanisms underlying transcriptional repression during the cell cycle are poorly understood. Repression of E2F-regulated genes depends on the recruitment of pRb, p107 and p130 to a promoter (49 ,50 ). These pocket proteins are able to repress transcription, at least in part, through a position-independent mechanism which apparently involves establishment of interactions with specific transcriptional activators and basal factors (49 ,51 ). It is possible that CDF-1 employs a similar mechanism of repression, even though it does not seem to be associated with pocket proteins of the pRb family (see accompanying paper by Liu et al.). In agreement with the latter observation, experiments performed with knockout mice that carry disrupted pRb, p107 and/or p130 genes show no change in expression of cdc25C (52 ).
The identification of CDF-1 together with the observations made in this study provide an important basis to investigate these questions in the future. Once CDF-1 is available in purified and/or recombinant form its interactions with the activation domains identified in the present study and with their interaction partners of the basal transcription machinery can be analyzed in detail in order to elucidate the molecular mechanisms underlying CDF-1-mediated repression of transcription during the early phases of the cell cycle.
We are grateful to M.Eilers, W.Herr, R.Mantovani, W.Schaffner, R.Tjian, W.Schaffner and T.Wirth for Gal4 constructs, to H.H.Arnold for the myf-4 promoter and C2C12 cells, to R.Vile and I.Hart for the tyrosinase and TRP-1 promoters and Mel-ab cells, to R.Treisman for NIH 3T3 cells and to Dr M.Krause for synthesis of oligonucleotides. This work was supported by grants from the DFG (SFB397 and Mu601/9-2) and the BMBF. J.Z. and V.J. were supported by fellowships from the Boehringer Ingelheim Fonds and EMBO respectively.
*To whom correspondence should be addressed. Tel: +49 6421 286236; Fax: +49 6421 288923; Email: mueller@imt.uni-marburg.de +Present address: Howard Hughes Medical Institute, Molecular and Cell Biology, 401 Barker Hall, University of California, Berkeley, CA 94720-3204, USA
The authors wish it to be known that, in their opinion, the first two authors should be considered as joint First Authors
H.-S. Martinsson-Ahlzen, V. Liberal, B. Grunenfelder, S. R. Chaves, C. H. Spruck, and S. I. Reed Cyclin-Dependent Kinase-Associated Proteins Cks1 and Cks2 Are Essential during Early Embryogenesis and for Cell Cycle Progression in Somatic Cells
Mol. Cell. Biol.,
September 15, 2008;
28(18):
5698 - 5709.
[Abstract][Full Text][PDF]
K. E. Knudsen, A. F. Fribourg, M. W. Strobeck, J.-M. Blanchard, and E. S. Knudsen Cyclin A Is a Functional Target of Retinoblastoma Tumor Suppressor Protein-mediated Cell Cycle Arrest
J. Biol. Chem.,
September 24, 1999;
274(39):
27632 - 27641.
[Abstract][Full Text][PDF]
K. Korner and R. Muller In Vivo Structure of the Cell Cycle-regulated Human cdc25C Promoter
J. Biol. Chem.,
June 16, 2000;
275(25):
18676 - 18681.
[Abstract][Full Text][PDF]
K. Korner, V. Jerome, T. Schmidt, and R. Muller Cell Cycle Regulation of the Murine cdc25B Promoter. ESSENTIAL ROLE FOR NUCLEAR FACTOR-Y AND A PROXIMAL REPRESSOR ELEMENT
J. Biol. Chem.,
March 23, 2001;
276(13):
9662 - 9669.
[Abstract][Full Text][PDF]
CiteULike 
Connotea 
Del.icio.us What's this?
