Cell cycle-regulated expression of mammalian CDC6 is dependent on E2F

Abstract

The E2F transcription factors are essential regulators of cell growth in multicellular organisms, controlling the expression of a number of genes whose products are involved in DNA replication and cell proliferation. In Saccharomyces cerevisiae, the MBF and SBF transcription complexes have functions similar to those of E2F proteins in higher eukaryotes, by regulating the timed expression of genes implicated in cell cycle progression and DNA synthesis. The CDC6 gene is a target for MBF and SBF-regulated transcription. S. cerevisiae Cdc6p induces the formation of the prereplication complex and is essential for initiation of DNA replication. Interestingly, the Cdc6p homolog in Schizosaccharomyces pombe, Cdc18p, is regulated by DSC1, the S. pombe homolog of MBF. By cloning the promoter for the human homolog of Cdc6p and Cdc18p, we demonstrate here that the cell cycle-regulated transcription of this gene is dependent on E2F. In vivo footprinting data demonstrate that the identified E2F sites are occupied in resting cells and in exponentially growing cells, suggesting that E2F is responsible for downregulating the promoter in early phases of the cell cycle and the subsequent upregulation when cells enter S phase. Our data also demonstrate that the human CDC6 protein (hCDC6) is essential and limiting for DNA synthesis, since microinjection of an anti-CDC6 rabbit antiserum blocks DNA synthesis and CDC6 cooperates with cyclin E to induce entry into S phase in cotransfection experiments. Furthermore, E2F is sufficient to induce expression of the endogenous CDC6 gene even in the absence of de novo protein synthesis. In conclusion, our results provide a direct link between regulated progression through G1 controlled by the pRB pathway and the expression of proteins essential for the initiation of DNA replication.

FIG. 1

Expression of endogenous human, rat, and mouse CDC6 mRNA. (A) Northern blot containing total RNA from asynchronously growing human tumor cells U937 (histiocytic lymphoma), HT230 (colon adenocarcinoma), IMR32 (neuroblastoma), ML1 (premyeloid leukemia), HL60 (promyeloid leukemia), NGP (neuroblastoma), SiHa (cervix carcinoma), and T98G (glioblastoma). The blot was probed with a partial human CDC6 cDNA probe. (B) Northern blot containing total RNA from the human fibroblast cell lines Detroit 551 (embryonic skin), IMR90 (fetal lung), Hs27 (newborn foreskin), and HEL 299 (embryonal lung), the rat cell line Rat-1, and the mouse cell line NIH 3T3. Hybridization was done with a human CDC6 cDNA probe (lanes 1 to 5) and with a mouse CDC6 cDNA probe (lanes 6 and 7). Blots were subsequently probed with a partial rat GAPDH probe, as depicted (lower blots).

FIG. 2

Cell cycle-regulated expression of CDC6 mRNA. (A) Swiss 3T3 cells were starved in low-serum-containing medium, subsequently stimulated with high serum, and used for FACS analysis. (B) Total RNA was isolated from Swiss 3T3 cells treated similarly and used for Northern blotting. The blot was probed first for mouse CDC6 and E2F-1 mRNAs (upper blots). Subsequently the blot was probed for GAPDH expression (lower blot).

FIG. 3

Expression of CDC6 during a proliferating cell cycle. Asynchronously growing NIH 3T3 cells were fractionated by elutriation. Ten fractions are shown. (A and B) Total RNA from these elutriated cells was used for RT-PCR with sets of specific primers which amplified the cDNA obtained from mouse CDC6 mRNA (A) and mouse GAPDH mRNA (B). (C) The cell cycle distribution of cells from 10 fractions was determined by FACS.

FIG. 4

Sequence and schematic representation of the human CDC6 promoter. (A) Nucleotide sequence of the 1,759-bp CDC6 promoter. The previously reported 355-bp fragment is underlined (79). Three putative E2F DNA-binding sites are boxed. A putative CHR region (bp −30 to −26) and the putative INR region (bp −16 to −11) are shown in lowercase. Restriction sites used for cloning are italic, and the transcription start site (bp +1) is bold. (B) Schematic representation of transcription factor-binding sites in the large 1,759-bp human CDC6 promoter. The transcription start site is depicted with an arrow. A CHR site, a CCAAT box, and a putative Sp1 element (Sp1/?) which are found to be protected in in vivo footprint assays, are located next to the two 3′ E2F sites close to the start site. Positions of consensus binding sites for AP-2, C/EBP, Ets-1, and NF-κB are also given. Putative recognition sites upstream of bp −325 and downstream of the start site are not shown. INR, initiator region.

FIG. 4

Sequence and schematic representation of the human CDC6 promoter. (A) Nucleotide sequence of the 1,759-bp CDC6 promoter. The previously reported 355-bp fragment is underlined (79). Three putative E2F DNA-binding sites are boxed. A putative CHR region (bp −30 to −26) and the putative INR region (bp −16 to −11) are shown in lowercase. Restriction sites used for cloning are italic, and the transcription start site (bp +1) is bold. (B) Schematic representation of transcription factor-binding sites in the large 1,759-bp human CDC6 promoter. The transcription start site is depicted with an arrow. A CHR site, a CCAAT box, and a putative Sp1 element (Sp1/?) which are found to be protected in in vivo footprint assays, are located next to the two 3′ E2F sites close to the start site. Positions of consensus binding sites for AP-2, C/EBP, Ets-1, and NF-κB are also given. Putative recognition sites upstream of bp −325 and downstream of the start site are not shown. INR, initiator region.

FIG. 5

Luciferase activity mediated by different CDC6 promoter constructs. (A) Schematic representation of the large human CDC6 promoter. The transcription start site (+1) is given by an arrow. Three putative E2F sites are depicted by black bars. The two E2F sites in the 355-bp fragment (shaded) are numbered 1 and 2. The endogenous restriction sites BamHI and NaeI were used to construct p[−570,+98] and p[−266,+98]. Mutations introduced in the two most downstream E2F sites to obtain pGL3-DM, -SM1, and -SM2 are shown with big black crosses, and their sequences are given below the constructs. (B) Responses of p[−130,+225], pGL3-SM1, GL3-SM2, pGL3-DM, and p[−1534,+225] to E2F-1 and DP-1 cotransfection in U2-OS cells. The activity from the p[−130,+225] construct is upregulated approximately 13-fold by coexpression of E2F-1 and DP-1, while the mutant constructs pGL3-DM, -SM1, and -SM2 do respond significantly less to E2F-1 and DP-1. The larger p[−1534,+225] construct is upregulated approximately twofold. The p[−130,+225] construct without E2F-1 and/or DP-1 cotransfection is depicted as 100 adjusted luciferase counts after measuring β-Gal activity. All other values are given relative to this with standard deviations from the means.

FIG. 5

Luciferase activity mediated by different CDC6 promoter constructs. (A) Schematic representation of the large human CDC6 promoter. The transcription start site (+1) is given by an arrow. Three putative E2F sites are depicted by black bars. The two E2F sites in the 355-bp fragment (shaded) are numbered 1 and 2. The endogenous restriction sites BamHI and NaeI were used to construct p[−570,+98] and p[−266,+98]. Mutations introduced in the two most downstream E2F sites to obtain pGL3-DM, -SM1, and -SM2 are shown with big black crosses, and their sequences are given below the constructs. (B) Responses of p[−130,+225], pGL3-SM1, GL3-SM2, pGL3-DM, and p[−1534,+225] to E2F-1 and DP-1 cotransfection in U2-OS cells. The activity from the p[−130,+225] construct is upregulated approximately 13-fold by coexpression of E2F-1 and DP-1, while the mutant constructs pGL3-DM, -SM1, and -SM2 do respond significantly less to E2F-1 and DP-1. The larger p[−1534,+225] construct is upregulated approximately twofold. The p[−130,+225] construct without E2F-1 and/or DP-1 cotransfection is depicted as 100 adjusted luciferase counts after measuring β-Gal activity. All other values are given relative to this with standard deviations from the means.

FIG. 6

Cell cycle-regulated expression of the CDC6 promoter is dependent on E2F DNA-binding sites. NIH 3T3 cells were transiently transfected with p[−130,+225], pGL3-DM, p[−570,+98], and p[−266,+98]. Forty-eight hours posttransfection, cells were serum starved for 24 h and subsequently stimulated with fresh medium containing 10% BCS. Lysates were made after the depicted time spans. Asynchronous (A) samples were valued as 100 adjusted luciferase counts, while the others were calculated in comparison to this. Transfection efficiencies were determined by pCMV-βgal cotransfection. Samples were obtained in duplicate, and the presented data are representative for at least three independent experiments.

FIG. 7

EMSA using labeled wild-type (wt) and double mutant (DM) probes. CDC6 promoter probes were obtained by PCR on p[−130,+225] and pGL3-DM and used in combination with nuclear extract from MRC5 human fibroblasts. Specific E2F complexes that interact with the wt probe are depicted with a bracket to the right of the blot. Non-E2F containing protein-DNA complexes are given with asterisks to the left. The cold wt probe (in lane 3) and the cold mutant (derived from DM) were added in 50-fold excess over the quantity of radiolabeled fragment. Antibodies directed against DP-1, E2F-4, and pRB (lanes 5 to 9) shift different subsets of E2F-containing complexes, while the antibodies against p107 and p130 (lanes 10 to 12) do not. M1 (lane 13) is a monoclonal antibody raised against adenovirus E1A. FTβ (lane 14) is a polyclonal serum raised against farnesyl transferase. M1 and FTβ served as negative (neg.) controls. -, nothing added; PC, polyclonal antibody.

FIG. 8

In vivo footprinting analysis of the transcription start site region of the hCDC6 promoter. LMPCRs were performed with primer S1, S2, or [minus (−) strand] or AS1, AS2, or AS3 [plus (+) strand] on genomic DNA templates obtained from serum-starved (G₀) or exponentionally growing (expo.) MCF7 cells treated in vivo with the guanosine methylating agent, DMS. Similar LMPCRs were performed with DMS-methylated naked DNA (vitro lanes). Protected residues and hyperreactive residues detected between in vitro- and in vivo-methylated DNAs are indicated as circles and arrowheads, respectively. Weak (white circles) and strong (black circles) in vivo protection is indicated. The transcription start site is indicated with an arrowhead (+1) to the left of the blots. Amplified DNA ladders that are visible correspond to guanines of the hCDC6 promoter. (A) Positive-sense strand. (B) Negative-sense strand. (C) Summary of DNA-protein contacts observed by in vivo footprinting on both strands of the hCDC6 promoter upstream of the transcription start site (black arrow, +1). Putative consensus binding sites are indicated as open boxes. A protein-bound element with a sequence similar to that of an Sp1 consensus site is depicted as Sp1/?. A protected site around the putative initiator region (INR) (bp −16 to −10) is indicated with a question mark.

FIG. 8

In vivo footprinting analysis of the transcription start site region of the hCDC6 promoter. LMPCRs were performed with primer S1, S2, or [minus (−) strand] or AS1, AS2, or AS3 [plus (+) strand] on genomic DNA templates obtained from serum-starved (G₀) or exponentionally growing (expo.) MCF7 cells treated in vivo with the guanosine methylating agent, DMS. Similar LMPCRs were performed with DMS-methylated naked DNA (vitro lanes). Protected residues and hyperreactive residues detected between in vitro- and in vivo-methylated DNAs are indicated as circles and arrowheads, respectively. Weak (white circles) and strong (black circles) in vivo protection is indicated. The transcription start site is indicated with an arrowhead (+1) to the left of the blots. Amplified DNA ladders that are visible correspond to guanines of the hCDC6 promoter. (A) Positive-sense strand. (B) Negative-sense strand. (C) Summary of DNA-protein contacts observed by in vivo footprinting on both strands of the hCDC6 promoter upstream of the transcription start site (black arrow, +1). Putative consensus binding sites are indicated as open boxes. A protein-bound element with a sequence similar to that of an Sp1 consensus site is depicted as Sp1/?. A protected site around the putative initiator region (INR) (bp −16 to −10) is indicated with a question mark.

FIG. 9

Cell cycle-regulated occupation of protein-binding sites in the human CDC6 promoter. Human MCF7 cells were starved in medium lacking serum and isoleucine and subsequently stimulated in normal medium with a high level of serum. (A) FACS analysis of synchronized MCF7 cells after starvation (0 h) and stimulation (12, 16, and 20 h). To block cells in G₁/S, cells were also treated with hydroxyurea (HU). (B) Northern blot with total RNA from MCF7 cells that were treated as described above for panel A and then probed for human CDC6 mRNA expression (upper blot). Equal loads were ensured by ethidium bromide (EtBr) staining (lower blot). Lane A contains RNA isolated from asynchronously growing cells. (C) In vivo footprinting analysis [plus (+) strand] on genomic DNA templates from MCF7 cells that were treated as described above panels for A and B. Vitro lane contains LMPCR-treated DMS-methylated naked DNA. HU lane contains samples from hydroxyurea-treated cells. Weak (white circles) and strong (black circles) in vivo protection is indicated. Hyperactive residues, compared to in vitro-methylated DNA, are indicated with arrowheads. Sp1 refers to the putative Sp1 site upstream of the two E2F (E2F/1 and E2F/2) sites. The protected site around the putative initiator region is indicated with a question mark.

FIG. 9

Cell cycle-regulated occupation of protein-binding sites in the human CDC6 promoter. Human MCF7 cells were starved in medium lacking serum and isoleucine and subsequently stimulated in normal medium with a high level of serum. (A) FACS analysis of synchronized MCF7 cells after starvation (0 h) and stimulation (12, 16, and 20 h). To block cells in G₁/S, cells were also treated with hydroxyurea (HU). (B) Northern blot with total RNA from MCF7 cells that were treated as described above for panel A and then probed for human CDC6 mRNA expression (upper blot). Equal loads were ensured by ethidium bromide (EtBr) staining (lower blot). Lane A contains RNA isolated from asynchronously growing cells. (C) In vivo footprinting analysis [plus (+) strand] on genomic DNA templates from MCF7 cells that were treated as described above panels for A and B. Vitro lane contains LMPCR-treated DMS-methylated naked DNA. HU lane contains samples from hydroxyurea-treated cells. Weak (white circles) and strong (black circles) in vivo protection is indicated. Hyperactive residues, compared to in vitro-methylated DNA, are indicated with arrowheads. Sp1 refers to the putative Sp1 site upstream of the two E2F (E2F/1 and E2F/2) sites. The protected site around the putative initiator region is indicated with a question mark.

FIG. 9

Cell cycle-regulated occupation of protein-binding sites in the human CDC6 promoter. Human MCF7 cells were starved in medium lacking serum and isoleucine and subsequently stimulated in normal medium with a high level of serum. (A) FACS analysis of synchronized MCF7 cells after starvation (0 h) and stimulation (12, 16, and 20 h). To block cells in G₁/S, cells were also treated with hydroxyurea (HU). (B) Northern blot with total RNA from MCF7 cells that were treated as described above for panel A and then probed for human CDC6 mRNA expression (upper blot). Equal loads were ensured by ethidium bromide (EtBr) staining (lower blot). Lane A contains RNA isolated from asynchronously growing cells. (C) In vivo footprinting analysis [plus (+) strand] on genomic DNA templates from MCF7 cells that were treated as described above panels for A and B. Vitro lane contains LMPCR-treated DMS-methylated naked DNA. HU lane contains samples from hydroxyurea-treated cells. Weak (white circles) and strong (black circles) in vivo protection is indicated. Hyperactive residues, compared to in vitro-methylated DNA, are indicated with arrowheads. Sp1 refers to the putative Sp1 site upstream of the two E2F (E2F/1 and E2F/2) sites. The protected site around the putative initiator region is indicated with a question mark.

FIG. 10

Direct regulation of CDC6 expression by E2F-1 transcription factor. (A) Northern blot containing total RNA purified from Rat-1 cells stably expressing a fusion between full-length wild-type E2F-1 and the ligand-binding domain of the ER (ER-E2F-1). E2F-1 activity was induced by the addition of OHT in the absence (lanes 3 to 5) or presence of CHX (lanes 7 to 9). The effect of CHX addition alone (lanes 11 to 13) was also monitored. The Northern blot was probed with a partial mouse CDC6 probe and subsequently hybridized with a partial rat GAPDH probe. Lane A contains the asynchronous cell sample. (B) RT-PCR on RNA obtained from the cells described above for panel A. Cells were stimulated for up to 16 h with OHT or BCS or for 4 h with OHT in the absence or presence of CHX (C). Induction of CDC6 mRNA expression was measured by a linear range radioactive RT-PCR, and the rat endogenous GAPDH mRNA served as a control (lower blot). (D) RT-PCR on total RNA purified from Rat-1 cells stably expressing a fusion between a full-length DNA-binding mutant (E132) of E2F-1 and the ligand-binding domain of ER after the addition of CHX, OHT plus CHX, or OHT alone for 4 h. The CDC6 signal was measured by an RT-PCR, and GAPDH mRNA served as a control (lower blot).

FIG. 10

Direct regulation of CDC6 expression by E2F-1 transcription factor. (A) Northern blot containing total RNA purified from Rat-1 cells stably expressing a fusion between full-length wild-type E2F-1 and the ligand-binding domain of the ER (ER-E2F-1). E2F-1 activity was induced by the addition of OHT in the absence (lanes 3 to 5) or presence of CHX (lanes 7 to 9). The effect of CHX addition alone (lanes 11 to 13) was also monitored. The Northern blot was probed with a partial mouse CDC6 probe and subsequently hybridized with a partial rat GAPDH probe. Lane A contains the asynchronous cell sample. (B) RT-PCR on RNA obtained from the cells described above for panel A. Cells were stimulated for up to 16 h with OHT or BCS or for 4 h with OHT in the absence or presence of CHX (C). Induction of CDC6 mRNA expression was measured by a linear range radioactive RT-PCR, and the rat endogenous GAPDH mRNA served as a control (lower blot). (D) RT-PCR on total RNA purified from Rat-1 cells stably expressing a fusion between a full-length DNA-binding mutant (E132) of E2F-1 and the ligand-binding domain of ER after the addition of CHX, OHT plus CHX, or OHT alone for 4 h. The CDC6 signal was measured by an RT-PCR, and GAPDH mRNA served as a control (lower blot).

FIG. 11

Block of DNA synthesis by microinjection of specific antibodies directed against human CDC6 protein. (A) Human T98G cells were starved for 48 h in low-serum-containing medium and subsequently stimulated with medium with high serum plus BrdU. After 16 h, cells were injected with either an affinity-purified specific anti-hCDC6 rabbit antiserum (L20) or with affinity-purified rabbit IgG. After 2 h, cells were fixed and treated with DAPI (blue; upper panels) and with antibodies for BrdU (red; middle panels) and with antirabbit antibodies to score for the injected antisera (green; lower panels). Arrowheads indicate injected cells that are not incorporating BrdU in the case of L20 or that are synthesizing DNA in the case of rabbit IgG. (B) Noninjected cells were scored for BrdU incorporation after 0, 14, 16, and 18 h following serum stimulation and compared to the total number of cells. Injected cells that were stained green because of their injected antisera were scored for BrdU incorporation and compared to the total number of injected cells. The diagram is the average of three independent experiments in which approximately 200 cells per coverslip were injected. FCS, fetal calf serum; BCS, bovine calf serum; rIgG, rabbit IgG.

FIG. 11

Block of DNA synthesis by microinjection of specific antibodies directed against human CDC6 protein. (A) Human T98G cells were starved for 48 h in low-serum-containing medium and subsequently stimulated with medium with high serum plus BrdU. After 16 h, cells were injected with either an affinity-purified specific anti-hCDC6 rabbit antiserum (L20) or with affinity-purified rabbit IgG. After 2 h, cells were fixed and treated with DAPI (blue; upper panels) and with antibodies for BrdU (red; middle panels) and with antirabbit antibodies to score for the injected antisera (green; lower panels). Arrowheads indicate injected cells that are not incorporating BrdU in the case of L20 or that are synthesizing DNA in the case of rabbit IgG. (B) Noninjected cells were scored for BrdU incorporation after 0, 14, 16, and 18 h following serum stimulation and compared to the total number of cells. Injected cells that were stained green because of their injected antisera were scored for BrdU incorporation and compared to the total number of injected cells. The diagram is the average of three independent experiments in which approximately 200 cells per coverslip were injected. FCS, fetal calf serum; BCS, bovine calf serum; rIgG, rabbit IgG.