E2f1, E2f2, and E2f3 control E2F target expression and cellular proliferation via a p53-dependent negative feedback loop

Abstract

E2F-mediated control of gene expression is believed to have an essential role in the control of cellular proliferation. Using a conditional gene-targeting approach, we show that the targeted disruption of the entire E2F activator subclass composed of E2f1, E2f2, and E2f3 in mouse embryonic fibroblasts leads to the activation of p53 and the induction of p53 target genes, including p21(CIP1). Consequently, cyclin-dependent kinase activity and retinoblastoma (Rb) phosphorylation are dramatically inhibited, leading to Rb/E2F-mediated repression of E2F target gene expression and a severe block in cellular proliferation. Inactivation of p53 in E2f1-, E2f2-, and E2f3-deficient cells, either by spontaneous mutation or by conditional gene ablation, prevented the induction of p21(CIP1) and many other p53 target genes. As a result, cyclin-dependent kinase activity, Rb phosphorylation, and E2F target gene expression were restored to nearly normal levels, rendering cells responsive to normal growth signals. These findings suggest that a critical function of the E2F1, E2F2, and E2F3 activators is in the control of a p53-dependent axis that indirectly regulates E2F-mediated transcriptional repression and cellular proliferation.

FIG. 1.

Loss of E2f1-3 leads to the activation of p53 target genes in primary cells. (A) Primary E2F1^−/− E2F2^−/− E2F3^f/f (123^f/f) cells were infected with control (−) or cre (+) retroviruses. Genomic DNA was then extracted from control- and cre-treated cells and used for PCR to determine the extent of E2f3 deletion. (B) BrdU incorporation of the primary 123^f/f cell line. Primary 123^f/f cells were treated as described for panel A and used for a BrdU incorporation assay. Cells were brought to quiescence by serum starvation and stimulated to proliferate by the addition of serum. Cells were harvested for BrdU incorporation at 0 and 18 h after serum stimulation. At least 500 cells were counted at each time point. (C) Real-time PCR analysis of E2F target genes. Total RNA was harvested from control-treated (−) or cre-treated (+) 123^f/f cells and used to produce cDNA. Real-time PCR analysis was done to determine the relative levels of the indicated E2F target genes. Results are shown as the induction (n-fold) of gene expression, where levels for control-treated samples were standardized to equal 1. (D) Real-time PCR analysis of p53 target genes in the 123^f/f cell line. Primary 123^f/f cells were treated as described for panel A, and cells were harvested under proliferating conditions. Total RNA was extracted and used to produce cDNA that was then used to look at the relative levels of the indicated p53 target genes. Results are shown as the induction (n-fold) of gene expression in the E2f1^−/− E2f2^−/− E2f3^−/− cells (+) compared to that in the control-treated E2f1^−/− E2f2^−/− E2f3^f/f cells (−). (E) p19^ARF mRNA levels are elevated in TKO cells. RNA from cells treated as described for panel C was used for real-time analysis. Shown are levels of p19^ARF in control-treated or cre-treated 123^f/f cells. (F) Increased levels of p19^ARF protein in TKO cells. Cellular lysates from 123^f/f cells infected with either control-expressing or cre-expressing retroviruses were used for a Western blot probed with the p19^ARF, p53, and tubulin antibodies.

FIG. 2.

Loss of E2f1-3 leads to p53 phosphorylation on serine 15. (A) Real-time PCR analysis of p53 target genes in the 123^f/f cell line. Established 123^f/f MEFs were infected with either control-expressing or cre-expressing (+) retrovirus and selected with puromycin. The cells were harvested under proliferating conditions. Total RNA was extracted and used to produce cDNA that was then used to look at relative levels of the indicated p53 target genes. Results are shown as induction (n-fold) of gene expression in the E2f1^−/− E2f2^−/− E2f3^−/− cells compared to that in the control-treated E2f1^−/− E2f2^−/− E2f3^f/f cells. (B) p21 levels are increased in TKO cells. The 123^f/fcell line was infected with either control-expressing or cre-expressing retroviruses and selected with puromycin. Protein was then extracted from these cells, and equal amounts of lysates from control- and cre-treated 123^f/f cells were used for Western blot analysis with antibodies against E2F3 (top panel), p21^CIP1 (middle panel), and tubulin as a loading control (bottom panel). (C) p19^ARF mRNA levels in TKO cells. RNA from cells treated as described for panels A and B was used for real-time PCR analysis. Shown here are levels of p19^ARF in control-treated or cre-treated 123^f/f cells. (D) Phosphorylated p53 protein levels are increased in TKO cells. Lysates from cells treated as described for panels A and B were used for Western blot analysis. For the Western blot, antibodies against total p53, p53 (ser¹⁵) and p19^ARF were used along with tubulin as a loading control. (E) Acetylation of histone H4. Cell lysates from experiments similar to those described above were harvested and used for chromatin immunoprecipitation assays with antibodies against acetylated histone H4. Real-time PCR was then performed using primers around either the p53-binding sites of the p21 and bax promoters or the E2F sites in the cyclin A2, dhfr, cdc6, and cyclin E1 promoters. Results are shown as inductions (n-fold) of histone H4 acetylation in control versus TKO cells from one representative experiment.

FIG. 3.

Analysis of proliferating TKO colonies. (A) Colony formation assay of TKO cells. Cells treated as described in the legend for Fig. 2A were counted and plated on 100-mm plates. After 2 weeks, the colonies were fixed, stained, and counted. The mean and standard deviation of colony counts from triplicate plates are shown. Colonies from the cre-infected cells have been corrected for deletion of E2f3 using the method described in panel B. (B) To determine the efficiency of E2f3 deletion, cells treated with the cre retrovirus were also plated into 96-well plates. DNA was extracted from wells containing single colonies and analyzed for deletion of E2f3 by PCR genotyping. A total of 32 colonies for each cell line was analyzed. Shown are results for seven representative colonies. (C) Colonies obtained from 123^f/f cell line 4 from an experiment similar to the one described above were expanded and subjected to various assays. E2f3 PCR genotyping of genomic DNA extracted from the 123^f/f cell line 4 colonies is shown. PCR on genomic DNA from the original population of cells is shown in lanes 2 and 3. Lanes 4 to 15 represent 12 different colonies. Asterisks identify TKO colonies. Colonies without asterisks represent cells that were infected with the cre retrovirus and survived selection but failed to delete E2f3. (D) Western blot analysis of colonies for several cell cycle regulators. Protein lysates were made from the original cell populations and the 12 colonies. After electrophoresis of equal amounts of protein, the ensuing Western blots were probed with antibodies against p21^CIP1 (top panel), p53 (middle panel), and cdk4 (bottom panel) proteins. (E) p53 promoter activity in colonies of 123^f/f cell line 4. Cells were transfected with either a wild-type p21^CIP1 promoter construct (pGL3.1-mp21) or a mutant construct deleted for the p53-binding sites (pGL3.1-mp21Δp53). The graph demonstrates the relative activity obtained from the wild-type construct versus that from the mutated construct for each colony. 123^f/f1, 123^f/f cell line 1; 123^f/f4, 123^f/f cell line 4.

FIG. 4.

Genetic deletion of p53 rescues the cellular proliferation defect of E2f1^−/− E2f2^−/− E2f3^−/− cells. The 123^f/f and p53^f/f 123^f/f established cell lines (passage number, >20) were infected first with control- or E2F3a-expressing retroviruses and then with either control- or cre-expressing retroviruses, selected for hygromycin and puromycin resistance, and then used in the following assays. (A) E2f3 (top panel) and p53 (bottom two panels) PCR genotyping of the population of cells from the experiment described above showing efficient deletion of both E2f3 and p53. The floxed allele (E2f3^f/f) produces a 184-bp PCR fragment, and the knockout E2f3 allele produces a 416-bp fragment. The p53 conditional and knockout alleles produce 584-bp and 612-bp fragments, respectively. (B) BrdU incorporation of established 123^f/f and p53^f/f 123^f/f cell lines. Proliferating cells were incubated with BrdU for 12 h and then fixed and stained. A total of 500 DAPI-stained nuclei from each cell line was counted, and the percent positive for BrdU incorporation is shown. (C) Cells from the experiment described above were also tested for their long-term growth potentials by colony formation assay. Values shown have been corrected for deletion of E2f3 and p53 by colony PCR. (D) E2f3 (top panel) and p53 (bottom two panels) PCR genotyping on genomic DNA from the cre-infected 123^f/f and p53^f/f 123^f/f colonies. E2f3 was not deleted in colonies arising from the 123^f/f cell line but was deleted in all the colonies arising from the cre-infected p53^f/f 123^f/f population of cells. (E) BrdU incorporation of the primary p53^f/f 123^f/f cell line. Primary p53^f/f 123^f/f cells were infected with control-expressing (−) or cre-expressing (+) retroviruses. Cells were synchronized by serum starvation, stimulated by the addition of serum, and harvested at the indicated time points. At least 500 cells were counted at each time point, and the percent positive for BrdU incorporation is shown.

FIG. 5.

Loss of p53 in TKO cells prevents the activation of p53-regulated genes and restores kinase activity. (A) Real-time PCR analysis of p53 target genes in TKO and QKO cells. The 123^f/f and p53^f/f 123^f/f cell lines were infected with either control-expressing (−) or cre-expressing (+) retroviruses and selected with puromycin. Total RNA was extracted from proliferating cells and then used to produce cDNA to determine relative levels of the indicated p53 target genes. Results are shown as induction (n-fold) of gene expression in the cre-treated versus the control-treated cells from one representative experiment. (B) Quadruple knockout cells have normal kinase activity. Relative activities of cyclin E-dependent kinase in protein lysates derived from TKO and QKO cells were determined by harvesting cells after 72 h of serum starvation (0 hrs) or 18 h of serum stimulation (18 hrs) and using histone 1 as a substrate. The blot shown in the inlay was scanned and quantified, and relative activity is shown. (C) Rb phosphorylation status of cells used in the experiment described for panel B as determined by Western blot analysis. (D and E) Cell lysates from TKO and QKO cells treated with either control (−) or cre (+) retrovirus were harvested and then used for chromatin immunoprecipitation assays with antibodies against E2F4 and p130. Real-time PCR was then performed using primers around the E2F sites in the cyclin A2, bmyb, and gapdh promoters. Results are shown as recruitment (n-fold) in cre-treated cells relative to that in the cells infected with control virus from one representative experiment.

FIG. 6.

E2F target genes are expressed in QKO cells. (A) cDNA prepared for the experiment shown in Fig. 5 was also used to determine the expression levels of E2F target genes in proliferating TKO versus QKO cells. In contrast to the results shown below, this graph demonstrates repression (n-fold) of gene expression in the cre-treated cells relative to that in the cells infected with control virus. (B) Induction of S phase and real-time PCR analysis of E2F target genes in TKO and QKO cells. Cells were synchronized by serum starvation, stimulated by the addition of serum, and harvested at the indicated time points. Total RNA was extracted from the cells and subsequently used to produce cDNA and determine the levels of the indicated E2F target genes. In order to determine the timing of S-phase entry in these cells relative to that of gene expression, the incorporation of 5-bromodeoxyuridine was also analyzed. BrdU was added to the plates at the same time as the serum, and harvesting took place at the zero time point as well as at the 8-hour time point. A total of 500 DAPI-stained nuclei from each cell line was counted, and the percent positive for BrdU incorporation is shown. Percent BrdU incorporation is indicated by the dashed lines, and solid lines represent induction (n-fold) of target genes. Symbols: □, cells treated with control virus; •, cre-treated cells.

FIG. 7.

Histone H4 acetylation status of E2F and p53 target promoters. (A) Cell lysates from TKO cells treated with either control (−) or cre (+) retrovirus were harvested after 2 days (left panel) and 4 days (right panel) of selection and then used for chromatin immunoprecipitation assays with antibodies against acetylated histone H4. Real-time PCR was then performed using primers around either the E2F sites in the cyclin A2, dhfr, and cdc6 promoters or the p53-binding sites of the p21 and bax promoters. Results are shown as induction (n-fold) of histone H4 acetylation in control versus TKO cells from one representative experiment. (B) In addition to cell lysates from cells in the experiment described above, total RNA was extracted and used to produce cDNA. Real-time PCR analysis was then used to look at the relative levels of the indicated E2F and p53 target genes. Results are shown as induction (n-fold) of gene expression. (C) Repeat of the experiment described above with the addition of QKO cells treated with control or cre retrovirus.