Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Mol Cell. Author manuscript; available in PMC 2013 Apr 13.
Published in final edited form as:
PMCID: PMC3327774

Distinct regulatory mechanisms and functions of p53-activated and p53-repressed DNA damage response genes in embryonic stem cells


p53 is critical in regulating the differentiation of ES and induced pluripotent stem (iPS) cells. Here, we report a whole-genome study of p53-mediated DNA damage signaling in mouse ES cells. Systems analyses reveal that binding of p53 at promoter region significantly correlates with gene activation but not with repression. Unexpectedly, we identify a regulatory mode for p53-mediated repression through interfering with distal enhancer activity. Importantly, many ES cell-enriched core transcription factors are p53-repressed genes. Further analyses demonstrate that p53-repressed genes are functionally associated with ES/iPS cell status while p53-activated genes are linked to differentiation. p53-activated genes and -repressed genes also display distinguishable features of expression levels and epigenetic markers. Upon DNA damage, p53 regulates the self-renewal and pluripotency of ES cells. Together, these results support a model that, in response to DNA damage, p53 affects the status of ES cells through activating differentiation-associated genes and repressing ES cell-enriched genes.

Keywords: embryonic stem cells, p53, genomics, epigenetics, transcription


The tumor suppressor p53 is a sequence-specific transcription factor. It regulates cell cycle arrest, apoptosis, senescence and stem cells differentiation by activating and repressing a battery of its downstream targets (Vousden and Prives, 2009). It is an interesting question as to how p53 “decides” when to activate and when to repress. At a single-gene level, multiple models of p53-mediated repression have been proposed, which include the recruitment of co-repressors, competition of binding sites with other transcription factors, and interference with the activity of other factors (Laptenko and Prives, 2006). In somatic cells, both p53-activated genes and p53-repressed genes are critical for mediating p53’s function (Vousden and Prives, 2009). As the guardian of genome, p53 is also believed to play key roles in maintaining the genomic stability of ES cells by affecting the differentiation and/or apoptosis of ES cells (Li and Huang, 2010; Lin et al., 2005). Transcriptional repression of the Nanog gene by p53 upon DNA damage could partially explain the roles of p53 in ES cells (Lin et al., 2005).

In addition to regulating the differentiation of ES cells, p53 also plays an inhibitory role in generating induced pluripotent stem (iPS) cells. Blocking p53-mediated DNA damage signaling dramatically increases the reprogramming efficiency (Kawamura et al., 2009; Marion et al., 2009; Neveu et al., 2010; Takahashi and Yamanaka, 2006). Using a systems biology approach, a recent report has linked the aberrant reprogramming and p53 transcriptional gene network in ES cells to tumorigenesis (Mizuno et al., 2010). However, little is known about the transcriptional targets of p53 in mES cells. The mechanisms underlying p53-mediated differentiation and reprogramming regulations have not been fully appreciated. The relationship between p53-mediated differentiation/reprogramming regulation and tumorigenesis is also elusive (Zhang and Huang, 2010). Therefore, a genome-wide picture of p53 signaling in ES cells will greatly facilitate our understanding of the biological function of p53 in ES and iPS cells.

Here, we use ChIP-seq (chromatin immunoprecipitation followed by deep sequencing) combined with gene expression microarray to profile a whole-genome p53 signaling in mES cells. The main objects of this study are to identify factors that distinguish between p53-activated genes and p53-repressed genes, and to explore the functions of these two groups of genes in controlling ES cell differentiation and iPS cell generation at a genome-wide level. Our results show that the mechanisms used by p53 to regulate the activated genes and the repressed genes are drastically different. In addition, p53-activated genes and p53-repressed genes are two functionally separable transcriptional units during ES cell differentiation and somatic cell reprogramming. We also discover that the interference with the enhancer activity by the distal binding of p53 is a mechanism underlying the transcriptional repression of some p53 targets. Our results depict a global view of p53 signaling in mES cells and provide a molecular basis for understanding its roles in regulating the differentiation and reprogramming.


Genome-wide profiling of p53 chromatin binding

To explore the roles of p53 in mES cells in a genome-wide manner, we set out to identify p53 binding sites using ChIP-seq. mES cells were either untreated or treated with adriamycin, a DNA damage agent widely used to activate p53 (Huang et al., 2006; Huang et al., 2007; Lee et al., 2010). In addition to using a pan-p53 antibody that recognizes total p53, we also profiled the binding sites of a well known post-translational modification (PTM) of p53, Serine 18 phosphorylation (S18P, S15P in human). S18P is generally thought to be involved in the activation of p53 after DNA damage (Bode and Dong, 2004; Toledo and Wahl, 2006) and our original goal was to use this PTM as an indicator for p53 activation.

Using peak finding algorithm (Zhang et al., 2008), we identified 7749 p53 peaks from untreated cells (p53_Ctr), 53475 p53 peaks from cells treated with adriamycin (p53_Adr), 3758 S18P peaks from untreated cells (S18P_Ctr), and 30327 S18P peaks from adriamycin treated cells (S18P_Adr), all with high stringency (Figure 1A, Table S1 and Supplementary Methods).

Figure 1
Genomic profiling of p53 and S18P

One interesting observation is that p53 binds to chromatin without extrinsic DNA damage stress, suggesting that p53 is poised for activation on a significant portion (14.4%) of its binding sites before DNA damage (Figure 1A and 1B and Figure S1). On a genome-wide scale, p53 is hypo-phosphorylated at S18 before DNA damage, implying that p53 is generally less activated before DNA damage than after DNA damage (Figure S1B). At a single-peak level, 3717 p53 peaks have detectable S18P signal even in the absence of extrinsic DNA damage stress, suggesting that p53 might be activated at these sites by some intrinsic stresses, such as replicative stress.

We then performed genome-wide comparative analyses for p53 and S18P peaks. DNA damage increases the occupancy of total p53 and, to a larger extent, the S18P signal (Figure 1C, S1B and S1E), reflecting the multiple layers of regulation during p53 activation. An unresolved question is whether all p53 binding locations contain phosphorylated p53 at S18. We found that S18P signal was present at 28810 (53%) p53 peaks, while the remaining 24637 peaks had no S18P signal (Figure 1D). The 24637 peaks without S18P signal could result from the passive diffusion of p53 to chromatin after stabilization, or the dephosphorylation of p53 by some phosphatases at these regions after binding. Interestingly, the p53 peaks with S18P signal (n=28810) have higher occupancy than those without S18P signal (n=24637) (Figure 1D, lower panel), suggesting that S18P may contribute to the chromatin binding of p53.

The repression of many ES-specific core transcription factors is p53-dependent

We then integrated p53 binding data with gene expression microarray data to identify p53 direct target genes (Figure S2A, S2B, S2C and Supplementary Methods). Using this approach, 3697 genes were determined to be p53 direct targets, which have gene expression change in a p53-dependent manner and at least one p53 peak associated with the gene (Table S2). Among these genes, 2070 (56%) are activated while 1627 (44%) are repressed. Compared to other transcription factors, the number of p53 targets is similar to that of c-Myc (3542 genes) in mES cells, but much higher than that of Nanog (1284 genes) or Oct4 (783 genes) (Kim et al., 2008). Several Wnt ligands and canonical p53 direct targets were identified as p53-activated genes, as previously shown (Figure 2A) (Lee et al., 2010). Most notably, p53 represses many genes encoding key transcription regulators in mES cells, such as the Oct4, Nanog, Sox2, Zic3, Jmjd1c, Esrrb, Tcfcp2l1, Utf1, n-Myc, c-Myc and Prdm14 genes (Figure 2A) (Chen et al., 2008; Young, 2011). The p53-dependent repression of these genes was validated using realtime PCR and immunoblotting (Figure 2B and 2C). A dosage-dependent effect of adriamycin treatment on the repression of these genes was observed (Figure 2D and 2E). To rule out the possibility that the repression of these genes is limited to adriamycin, we treated the cells with aphidicolin, a DNA polymerase inhibitor that induces replicative and DNA damage stress responses (Sheaff et al., 1991). Similar p53-dependent repression was observed with aphidicolin treatment (Figure S2D), suggesting that the repression of these genes is not specific to adriamycin. Together, our results indicate that the repression of these ES-specific transcription factors is dependent on p53.

Figure 2
p53 represses many core transcription factors in mES cells

The decrease of the mRNA levels of some of these genes occurs as early as 2 hours in a p53-dependent manner, further suggesting that the repression is direct. In addition, the decrease of the mRNA levels of these p53-repressed genes precedes apoptosis since adriamycin-induced apoptosis becomes obvious after 8-hour treatment (Lee et al., 2010). To test the potential relationship between the repression of these genes and the cell cycle arrest caused by adriamycin or aphidicolin, we performed cell cycle analysis with p53+/+ and p53−/− mES cells. Adriamycin treatment leads to G2/M arrest, while aphidicolin causes G1/S arrest in a p53-independent manner (Figure S2E and S2F). Thus, the p53-dependent repression cannot simply be explained by the cell cycle arrest induced by these treatments. Together, our genome-wide study reveals a widespread p53-mediated transcriptional repression of ES cell-specific transcription factors.

p53-activated genes and p53-repressed genes have distinct p53 and S18P binding patterns

Because little is known about which factor(s) affects p53-mediated transcriptional outcome, i.e., activation or repression, we wanted to explore the potential relationship between p53 binding and the transcriptional outcome. We hypothesized that p53-activated genes and p53-repressed genes might have different p53 and/or S18P binding patterns, correlating with their activation and repression outcome. As an initial test, p53-activated genes and p53-repressed genes were grouped into two separate sets, and a set of 3697 randomly selected genes was used as a control to test whether a result was random. We defined three non-overlapping regions for each gene, a promoter-proximal region, a gene body region, and a distal region (Figure S2B and Supplementary Methods). For each region, p53 or S18P binding was determined and the percentage of genes containing at least one p53 or S18P peak was calculated (Figure 3A and 3B). Statistical analyses at each region were also performed (Table S3).

Figure 3
Promoter binding of p53 correlates with activation

In cells without extrinsic stress, p53 sits on a large fraction of its target genes: 51.6% of p53-activated genes and 26.2% of p53 repressed genes contain at least one p53 binding site (Figure 3B, left, p53Ctr sub-section). These p53 binding sites are either at the promoter-proximal region, gene body region, or distal region. In the promoter region, after DNA damage, the percentage of p53-activated genes with p53 binding greatly increases to 60–70%, while the percentage of the repressed genes with p53 binding increases only to 31.5%, which is slightly higher than that for the random genes (23.3%) (Figure 3B, p53Adr sub-section). At the gene body region or the distal region, we did not observe dramatic difference of p53 binding between p53-activated genes and p53-repressed genes. These results strongly suggest that the promoter-proximal region binding of p53 can differentiate p53-activated genes from p53-repressed genes. In addition, p53 peaks from the repressed genes are less likely to be phosphorylated at S18 than those from the activated genes (Figure S3A). 32% p53 peaks associated with the repressed genes have S18P signal, which is in drastic contrast to 73% for p53-activated genes and 55% for the random genes (Figure S3A). Thus, p53-activated genes and p53-repressed genes have their unique features of p53 and S18P binding patterns.

We then compared the occupancy of p53 and S18P peaks associated with the activated genes, the repressed genes and the random genes (Figure S3B). In any case, peaks associated with p53-repressed genes have significantly less fold enrichment than those in the activated and random groups. Using peaks associated with the p53-activated genes, we successfully generated a consensus motif, which is similar to the previously described p53 motif (Figure S3C) (Smeenk et al., 2008). However, we could not consistently derive a consensus motif using peaks linking to p53-repressed genes, suggesting that the peaks associated with the repressed genes may represent indirect chromatin recruitment of p53 probably by other transcription factor(s), or that the response elements of p53 in these peaks are significantly different from the consensus motif. This notion is also supported by the observation that stronger p53 binding sites are more likely to contain the consensus p53 motif than those weaker ones (Figure S3D and S3E). Therefore, at the peak level, p53-activated genes and p53-repressed genes also display distinct p53 and S18P binding patterns.

p53 binding in the promoter region is correlated with transcriptional activation

To gain insights into the role of p53 binding in transcription activation, we inspected the binding strength of p53 and S18P in the promoter-proximal and gene body regions on average and on individual p53 target (Figure 3C, 3D and 3E). Both p53 and S18P peaks spread around the TSS of p53 activated genes (Figure 3C and 3E). For p53-activated genes, the binding strength of p53 and S18P decreases along with the decrease of fold induction of gene expression (Figure 3C, walking average). However, for p53-repressed genes, the linkage of p53 and S18P binding with the gene expression change is not obvious. To further support this conclusion in a statistically manner, p53-activated genes and p53-repressed genes were divided into three quantiles according to their expression induction. The binding strength of p53 for each quantile and the differences between quantiles were calculated (Figure 3D). For the activated genes, as the expression induction decreases, the binding strength of p53 and S18P also attenuates (p<2.7e-5). However, for p53-repressed genes, p53 binding strength is unchanged (p>0.11) as expression induction alters, suggesting that, for the repressed genes, p53 binding at the promoter region is uncoupled with expression induction. There is a strong association between p53 binding in the promoter-proximal region and gene activation (p=3.2e-13, Table S4). We did not observe a correlation between p53 binding at the gene body or distal region with change of expression (Table S4). In the promoter region, DNA damage has a notable effect on p53 and S18P binding for p53-activated genes but not for p53-repressed genes (Figure 3E). In the gene body region, however, for both p53-activated genes and p53-repressed genes, DNA damage has a minimal effect on the binding of p53 and S18P (Figure 3E). These results suggest that the change of p53 and S18P binding strength in promoter is associated with gene activation.

Interference with distal enhancers is a mode for p53-mediated repression

We attempted to search for the region that correlates with p53-mediated repression. As p53 transits from an activator to a repressor, the binding density of p53 in the promoter region gradually decreases (Figure 4A, left). In the gene body region, the percentage of genes with p53 binding also decreases along with the decrease of fold induction of gene expression, albeit to a lesser extent than that in the promoter-proximal region (Figure 4A, middle). On the contrary, at the distal region, the binding of p53, on average, increases from the activated genes to the repressed genes (Figure 4A, right). Since most of the models for p53-mediated repression focused on the promoter-proximal regions (Laptenko and Prives, 2006), we were intrigued by the observation that p53-repressed genes on average tend to have more p53 binding at the distal region than at the promoter-proximal region. In addition, results from the association study show that the binding of p53 in the distal region significantly correlates with transcriptional repression of p53-repressed genes (p=0.003, Table S4). Furthermore, several p53-repressed core transcriptional factors, such as Nanog, Sox2, Oct4 and Sall4, have p53 binding sites at the distal regions (Figure 4B). Together, our analyses suggest that there might be a mechanism underlying p53-mediated repression in the distal region.

Figure 4
p53-repressed genes have a higher tendency to have distal p53 binding than the activated genes

We then aimed to explore the putative mechanism that links distal binding of p53 to repression. Enhancers were found at the distal regions (both upstream and downstream of a gene), as well as at the gene body regions (Blackwood and Kadonaga, 1998). A recent report showed that distal enhancer regions are critical for regulating the expression of many core transcription factors in ES cells (Kagey et al., 2010). Using a recently published dataset that maps the distal enhancers in mES cells (Creyghton et al., 2010), we found that p53-repressed genes are more likely to associate with distal enhancer than p53-activated genes (p<2.2e-16), with more than 62% p53-repressed genes and only 45% p53-activated genes having at least one distal enhancer (Figure 4C). Therefore, it is possible that the binding of p53 at the distal region might interfere with the activity of the enhancers of p53-repressed genes.

To test this hypothesis, we used a well-established luciferase-based enhancer test (Figure 5A) (Chen et al., 2008; Heintzman et al., 2007). The promoter of the Nanog gene drove the luciferase activity more than 27 fold (Figure 5B). The Nanog enhancer further increased the promoter-driven luciferase activity by about 10-fold, indicating that the assay can detect enhancer activity. We found that p53 did not alter the luciferase activity from a vector with only the Nanog promoter; however, it greatly decreased the luciferase activity from the vector carrying both the promoter and enhancer. Importantly, the R175H mutant did not decrease the luciferase activity despite the fact that it was expressed at a similar level in cells compared to the wild type. This result suggests that an intact DNA binding domain of p53 is required for this activity. It is worth noting that the Nanog enhancer is at the distal region upstream of the TSS (Figure 5B).

Figure 5
Interference with the distal enhancer activity by p53 binding is one of the mechanisms underlying p53-mediated repression

To test whether the interference of the enhancer activity depends on the relative position of the enhancer to the TSS of a gene, we performed similar experiments using a promoter-enhancer pair from another ES cell-enriched gene, Utf1, because the Utf1 enhancer is at the downstream distal region (Figure 5B). Similar to Nanog, the interference with the Utf1 enhancer activity by p53 was observed as well. Because ChIP-seq also detected another p53 binding site in the distal region of the Nanog gene (Figure 5C, dr, distal region), we tested whether this p53 binding site has enhancer activity and confers p53-mediated repression (Figure 5C). The result showed that this region had no enhancer activity (p>0.1) and was not involved in p53-mediated repression, further supporting the conclusion that p53-mediated repression of the Nanog gene relies on the interference of the enhancer activity. Therefore, enhancer interference at the distal region represents one of the mechanisms for p53-mediated repression.

Since a significant portion of p53-activated genes also contains p53 binding sites in their distal regions (Figure 3B), we decided to test the effect of these distal p53 binding sites in the context of transcriptional activation (Figure 5D). There are two possible outcomes: the distal p53 binding sites of the activated genes either antagonize or augment the activating effect of the promoters of these genes. Expectedly, p53 activated the expression of luciferase through the promoters of the Mdm2 and Btg2 genes. The distal p53 binding in the Mdm2 and Btg2 genes did not decrease the activating activity by the promoters. On the contrary, they slightly augment the activation driven by the promoters (Figure 5D). Therefore, the distal binding of p53 by itself does not necessarily lead to transcription repression. The ultimate transcriptional outcome may depend on the architecture of promoter and enhancer of a given gene.

Because no canonical consensus p53 response element was detected in the distal p53 binding sites associated with the repressed genes, the repressed gene-related DNA sequences of the distal p53 binding sites could be fundamentally different from those in the activated genes. Investigation of the sequences of the Nanog and Utf1 enhancers revealed a non-canonical p53 binding site in each enhancer. Both putative binding sites have a 15-nucleotide spacer between the two half sites (Figure 5E). Interestingly, earlier studies have suggested that the length of the spacer dictates the transcriptional outcome (activation versus repression) for p53-repressed genes, such as the Survivin and Mdr1 genes (Hoffman et al., 2002; Johnson et al., 2001). Therefore, we deleted the spacer of the putative p53 binding site in the enhancer of the Nanog or Utf1 gene, and tested whether the deletion can reverse p53-mediated repression (Figure 5F). The result shows that the deletion of the spacer does not change the repression mediated by p53, and the deletion does not significantly alter the activity of the Nanog and Utf1 enhancers either. Thus, these results suggest that the enhancer-mediated repressive mechanism uncovered in this study is fundamentally different from that involving the promoter regions of the Survivin and Mdr1 genes.

p53-activated genes and p53-repressed genes have distinct functions

We next turned our attention to the relationship between p53-regulated genes in mES cells and the genes associated with the differentiation of ES cells and the generation of iPS cells. Given that p53 regulates the transcription of many genes in mES cells, we used the Gene Set Enrichment Analysis (GSEA) to address this issue at a gene-set level (Subramanian et al., 2005). p53-activated genes are significantly enriched in the differentiation-associated genes, while p53-repressed genes are highly associated with ES cell-associated genes (Figure 6A, upper panels). This result indicates that p53 has dual functions in the regulation of ES cell status. On the one hand, it activates the differentiation-associated genes, and on the other hand, it represses genes that are required for maintaining ES cell status. We performed a similar test to investigate the roles of p53-activated genes and p53-repressed genes in somatic cell reprogramming. p53-activated genes are enriched in genes that are highly expressed in mouse embryonic fibroblast cells, and p53-repressed genes are significantly associated with genes expressed in iPS cells (Figure 6A, lower panels). Together, results from these analyses suggest that p53-activated genes and p53-repressed genes are functionally separable, and that they represent two sets of genes with opposite effects in the processes of ES cell differentiation and iPS cell generation. Of note, our results do not necessarily indicate that differentiation cues elicit p53-mediated DNA damage signal and vice versa. Instead, it suggests that p53 and differentiation cues merge at the same subset of downstream targets (Table S5). Thus, mES cells may use the same set of genes to cope with both differentiation signals and stress signals. To support the conclusion that p53-activated genes and p53-repressed genes have different functions, we assessed the relationship of p53-activated genes and p53-repressed genes with an ES cell-module consisting of 111 genes that are highly correlated with ES cell status (Kim et al., 2010). The overlap of p53-activated genes with the ES cell-module is minimal (p=0.92), while the overlap of p53-repressed genes with the module is highly significant (p=8.2e-36), with more than half of the genes in the module in the repressed genes (Figure 4B). This result further indicates that p53-repressed genes functionally link to ES cells status.

Figure 6
p53-activated genes are linked to ES cell differentiation and p53-repressed genes are associated with ES cell status

We then aimed to further understand the potential functions of p53-activated genes. The activation of some p53 induced genes, such as Cdkn1a, Phf2, Fosl1, Foxo3a, Sdc4, Btg2 and Ddit4, was validated by the realtime PCR analysis (Figure S4A). Similar to the repressed genes (Figure 2B and S2D), the activation of these genes also responds to aphidicolin treatment in a p53-dependent manner (Figure S4B). Having validated the activation of these candidate p53-activated genes, we conducted pathway analysis, as previously shown (Lee et al., 2010). There are 553 p53-activated genes that are overlapping with differentiation associated genes (Table S5). Among these genes, 13 pathways are enriched (p<0.05). Interestingly, extracellular matrix-receptor interaction, p53 signaling, MAP kinase signaling, and TGF-beta signaling pathways have been widely implicated in both differentiation/development and cancer. Uniquely enriched in the 1074 non-overlapping p53-activated genes are 25 pathways, many of which are cancer-related, for example, glioma pathway, basal cell carcinoma pathway, and pancreatic cancer pathway. These results shed light on the roles of p53 in the regulation of ES cell differentiation as well as in tumor suppression.

If p53-activated genes are linked to differentiation, their expression should be low in mES cells. In the same vein, the expression of p53-repressed genes should be high since they are enriched in ES cell-associated genes (Figure 6B). Gene expression microarray data supports this hypothesis and shows that the expression levels of p53-activated genes are about 4-fold less than those of p53-repressed genes (Figure 6C). To gain insights into the differential expression of p53-activated genes and p53-repressed genes, we interrogated three major histone modifications in ES cells, histone H3 lysine 4 trimethylation (H3K4me3), lysine 27 trimethylation (H3K27me3), and lysine 79 di-methylation (H3K79me2). H3K4me3 generally links to transcription initiation and H3K79me2 is normally associated with transcription elongation (Rahl et al., 2010). High levels of H3K27me3 correlate with low gene expression (Mikkelsen et al., 2007). On p53-activated gene, we observed low levels of H3K4me3 and H3K79me2 but high levels of H3K27me3 (Figure 6D), consistent with the observation that these genes are expressed at low levels in the absence of extrinsic stress (Figure 6C). For p53-repressed genes, H3K4me3 and H3K79me2 levels are high while H3K27me3 levels are low. On average, RNA polymerase II (Pol II) recruitment is higher on p53-repressed genes than on p53-activated genes. Therefore, without extrinsic stress, p53-activated genes are epigenetically suppressed while p53-activated genes are actively transcribed. Upon DNA damage, p53 induces the expression of differentiation-associated genes and simultaneously represses ES cell status-linked repressed genes.

DNA damage affects the self-renewal and pluripotency of mES cells

Earlier work by others and our current study suggest that p53 is critical for regulating the self-renewal and pluripotency of mES cells in response to DNA damage stress. Using the alkaline phosphatase assay, we tested the effect of DNA damage stress on the self-renewal of mES cells as previously described (Lee et al., 2010; Loh et al., 2006; Pease et al., 1990). We observed that adriamycin treatment increased the number of differentiated colonies in p53+/+ mES cells, suggesting that DNA damage stress decreases the self-renewal of p53+/+ mES cells (Figure 7A). However, DNA damage failed to increase the number of differentiated colonies in p53−/− mES cells, indicating that the pro-differentiation effect of DNA damage stress is p53-dependent.

Figure 7
p53 regulates mES cell differentiation and pluripotency

We also tested whether DNA damage affects the pluripotency of mES cells using the embryoid body (EB) formation assay (Desbaillets et al., 2000). mES cell-derived EBs consist of three embryonic germ layers: endoderm, mesoderm and ectoderm. p53+/+ and p53−/− mES cells either untreated or treated with adriamycin were allowed to form EBs for 0, 2, 4 and 6 days. Realtime PCR was used to measure the levels of markers for all three germ layers (Gata4 and Sox17 for endoderm; Flk1 and alpha-actin for mesoderm; Th and Tubb3 for ectoderm) (Figure 7B). Adriamycin treatment dramatically increased the levels of markers for the three germ layers in day2 EBs derived from p53+/+ mES cells, suggesting that DNA damage also up-regulates the differentiation-associated genes at this stage of EB formation (Figure 7C). However, this up-regulation was transient, and the expression of most the markers decreased in day4 and day6 EBs, to a level lower than that in EBs from undamaged cells. In contrast, the expression of all the markers gradually increased in EBs derived from undamaged mES cells. Importantly, we did not observe similar results in EBs derived from p53−/− mES cells (Figure S5A). Together, our results suggest that DNA damage affects the pluripotency of mES cells in a p53-dependent manner.


Using genome-wide approaches, we depict a comprehensive picture of p53-mediated transcriptional network in mES cells and provide a molecular basis for understanding p53’s roles in regulating ES cells differentiation upon DNA damage (Figure 7D). In response to DNA damage, p53 represses many ES cell- or iPS cell-associated genes, such as Oct4, Sox2, Nanog, Sall4, Esrrb, Sall4, Utf1, Prdm14, n-Myc and c-Myc. Interestingly, none of these transcription factors had been identified previously using a ChIP-chip platform, which only covers the promoter regions of genes (Lee et al., 2010). We here demonstrate that this is partially because p53-repressed genes tend to have less p53 binding in the promoter region. Therefore, promoter-focused arrays naturally favor the detection of p53-activated genes over p53-repressed genes (Figure S5B). Moreover, peaks associated with p53-repressed genes are generally more difficult to detect because they generally have weaker binding intensity than those associated with the activated genes (Figure S3B). ChIP-seq with enough sequencing depth is able to detect these low-intensity peaks. Thus, our whole-genome p53 binding studies provide a powerful tool to systematically probe the regulatory action of p53.

Post-translational modifications (PTMs) of p53

Upon stresses, p53 is subject to hundreds of PTMs, which mainly include phosphorylation, acetylation, ubiquitination, and methylation (Bode and Dong, 2004; Kruse and Gu, 2009). Although it is debatable about the biological function of PTMs, the general view is that they fine-tune the function of p53 independently or cooperatively (Toledo and Wahl, 2006). We found that p53 binds to chromatin without extrinsic stress and that the S18P signal of these p53 peaks, on average, is low (Figure 1 and S1). It is currently unknown whether chromatin-bound p53 without extrinsic stress bears phosphorylation at other sites or is actively regulated by other PTMs. In the past several years, work by others and us have proposed an attractive model that methylation keeps p53 inactive before stresses. After stresses, p53 is demethylated and switches to active PTMs, such as phosphorylation and acetylation (Berger, 2010; Huang et al., 2006; Huang et al., 2007; Shi et al., 2007). One limitation of this model is that it is derived from results on several candidate p53 targets. Whether it holds on a genome-wide scale is unclear. Armed with powerful computational biology approaches as exemplified in this study, we shall be able to decode the complex regulation of p53 by PTMs in the future.

Distal binding of p53

The mechanisms underlying the transcriptional repression by p53 are far less understood than those underlying activation (Laptenko and Prives, 2006). Multiple models have been proposed for p53-mediated transcriptional repression in somatic cells (Laptenko and Prives, 2006; Murphy et al., 1999), and most of these mechanisms focus on the promoter region. Guided by clues obtained from genome-wide analyses, we demonstrate that the interference with the distal enhancer activity is one of the repressive mechanisms (Figure 5). It is worth noting that our enhancer interference model may not explain all the repression events mediated by p53, and it is expected that p53 uses many different mechanisms to conduct transcriptional repression. Indeed, 38% p53-repressed genes do not have a distal enhancer. Promoter-based models proposed by others could also play roles in p53-mediated repression in mES cells. Nevertheless, our results significantly expand the controlling domains of p53 in mES cell genome to the distal region.

Why p53 “chooses” to repress some genes at the distal regions in mES cells? Given that enhancers normally are associated with cell-type-specific gene transcription, and in human cells more than half of the enhancers locate at the distal regions (Creyghton et al., 2010), it is attractive to speculate that this observation may reflect a ES cell-specific role of p53 in regulating the status of ES cells. Indeed, p53-repressed genes are significantly overlapping with ES cell-specific genes (Figure 6A and 6B).

p53 in reprogramming

p53-mediated DNA damage response is one of the major barriers for reprogramming (Kawamura et al., 2009; Marion et al., 2009). It has been suggested that p53 may inhibit reprogramming at an early stage by activating senescence genes, such as Cdkn1a (Kawamura et al., 2009; Marion et al., 2009). Here, we discovered that p53 represses the transcription of many reprogramming factors, such as Oct4, Nanog, Sox2, n-Myc, c-Myc and Esrrb (Blelloch et al., 2007; Feng et al., 2009; Takahashi and Yamanaka, 2006; Yu et al., 2007). Because full reprogramming requires a sustained endogenous expression of these reprogramming factors (Takahashi and Yamanaka, 2006; Yu et al., 2007), our results raise a provocative possibility that p53 might also have a role during late stages of reprogramming.

It is not a practical approach to increase the reprogramming efficiency by simply removing p53 because this will lead to genome instability. Therefore, understanding the role of each or a subset of p53 target genes in reprogramming is crucial for the development of more efficient ways to generate iPS cells while maintaining their genome stability. For example, we found that p53 up-regulates TGF-beta pathway in ES cells (Table S5). Inhibition of TGF-beta signaling has recently been shown to increase reprogramming efficiency (Ichida et al., 2009). It will be interesting to test whether the modulation of other common pathways also can increase reprogramming efficiency. Thus, our whole-genome profile of p53 signaling in mES cells has set the stage for exploring this issue in the future.


ChIP-seq and peak calling

ChIP assay was performed as previously described (Huang et al., 2007; Lee et al., 2010). For the next generation sequencing, 10–20 ng DNA were used to make the ChIP-seq library as per instructions from Illumina’s ChIP-Seq Sample Prep kit. The ChIP-seq library was loaded on Illumina’s Cluster Station and sequenced with Illumina Genome Analyzer IIx. All these steps were performed at the sequencing facility at the Center for Cancer Research. Sequenced reads were aligned to the mouse genome (mm9) and reads with more than two aligned positions were removed. The remaining tags were further filtered by quality score and redundancy. Only non-redundant reads that have passed the quality score were kept for downstream analysis. ChIP-seq tags (treatment) and whole genome sequencing tags (control) were then analyzed using MACS (Zhang et al., 2008) to identify the peaks. For details, see Supplementary Methods.

Other methods

Full methods are provided in the Supplementary Methods.

Supplementary Material






We thank Glenn Merlino, Tom Misteli, Nan Roche and Li Guo for critically reading the manuscript and their comments. J.H.’s laboratory was funded by the intramural research program and partially by the Office of Science and Technology Partnerships at the Center for Cancer Research (CCR), the National Cancer Institute (NCI) at the National Institutes of Health (NIH), and partially by the NCI Director’s Innovation Award (to J.S. and J.H.). The computational analyses utilized the high-performance computational capabilities of the Helix Systems at NIH (http://helix.nih.gov).


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


ChIP-seq and gene expression microarray data were combined and deposited into the GEO database with a super-series number, GSE26360.


M.L., Y.H., and J.H. performed the ChIP and ChIP-seq assays and enhancer test. X. W. did gene expression microarray and partially analyzed the data. W.D. carried out MEF and mES cell-related work. J. S. performed statistical analyses and J.H. did computational analyses. M.L. carried out the rest of the experiments. M.L. and J.H. conceived the experiment design and wrote the manuscript.


The authors declared no conflict of interest.


  • Berger SL. Keeping p53 in check: a high-stakes balancing act. Cell. 2010;142:17–19. [PMC free article] [PubMed]
  • Blackwood EM, Kadonaga JT. Going the distance: a current view of enhancer action. Science. 1998;281:60–63. [PubMed]
  • Blelloch R, Venere M, Yen J, Ramalho-Santos M. Generation of induced pluripotent stem cells in the absence of drug selection. Cell Stem Cell. 2007;1:245–247. [PMC free article] [PubMed]
  • Bode AM, Dong Z. Post-translational modification of p53 in tumorigenesis. Nat Rev Cancer. 2004;4:793–805. [PubMed]
  • Chen X, Xu H, Yuan P, Fang F, Huss M, Vega VB, Wong E, Orlov YL, Zhang W, Jiang J, et al. Integration of external signaling pathways with the core transcriptional network in embryonic stem cells. Cell. 2008;133:1106–1117. [PubMed]
  • Creyghton MP, Cheng AW, Welstead GG, Kooistra T, Carey BW, Steine EJ, Hanna J, Lodato MA, Frampton GM, Sharp PA, et al. Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proc Natl Acad Sci U S A. 2010;107:21931–21936. [PMC free article] [PubMed]
  • Desbaillets I, Ziegler U, Groscurth P, Gassmann M. Embryoid bodies: an in vitro model of mouse embryogenesis. Exp Physiol. 2000;85:645–651. [PubMed]
  • Feng B, Jiang J, Kraus P, Ng JH, Heng JC, Chan YS, Yaw LP, Zhang W, Loh YH, Han J, et al. Reprogramming of fibroblasts into induced pluripotent stem cells with orphan nuclear receptor Esrrb. Nat Cell Biol. 2009;11:197–203. [PubMed]
  • Heintzman ND, Stuart RK, Hon G, Fu Y, Ching CW, Hawkins RD, Barrera LO, Van Calcar S, Qu C, Ching KA, et al. Distinct and predictive chromatin signatures of transcriptional promoters and enhancers in the human genome. Nat Genet. 2007;39:311–318. [PubMed]
  • Hoffman WH, Biade S, Zilfou JT, Chen J, Murphy M. Transcriptional repression of the anti-apoptotic survivin gene by wild type p53. J Biol Chem. 2002;277:3247–3257. [PubMed]
  • Huang J, Perez-Burgos L, Placek BJ, Sengupta R, Richter M, Dorsey JA, Kubicek S, Opravil S, Jenuwein T, Berger SL. Repression of p53 activity by Smyd2-mediated methylation. Nature. 2006;444:629–632. [PubMed]
  • Huang J, Sengupta R, Espejo AB, Lee MG, Dorsey JA, Richter M, Opravil S, Shiekhattar R, Bedford MT, Jenuwein T, et al. p53 is regulated by the lysine demethylase LSD1. Nature. 2007;449:105–108. [PubMed]
  • Ichida JK, Blanchard J, Lam K, Son EY, Chung JE, Egli D, Loh KM, Carter AC, Di Giorgio FP, Koszka K, et al. A small-molecule inhibitor of tgf-Beta signaling replaces sox2 in reprogramming by inducing nanog. Cell Stem Cell. 2009;5:491–503. [PMC free article] [PubMed]
  • Johnson RA, Ince TA, Scotto KW. Transcriptional repression by p53 through direct binding to a novel DNA element. J Biol Chem. 2001;276:27716–27720. [PubMed]
  • Kagey MH, Newman JJ, Bilodeau S, Zhan Y, Orlando DA, van Berkum NL, Ebmeier CC, Goossens J, Rahl PB, Levine SS, et al. Mediator and cohesin connect gene expression and chromatin architecture. Nature. 2010;467:430–435. [PMC free article] [PubMed]
  • Kawamura T, Suzuki J, Wang YV, Menendez S, Morera LB, Raya A, Wahl GM, Belmonte JC. Linking the p53 tumour suppressor pathway to somatic cell reprogramming. Nature. 2009;460:1140–1144. [PMC free article] [PubMed]
  • Kim J, Chu J, Shen X, Wang J, Orkin SH. An extended transcriptional network for pluripotency of embryonic stem cells. Cell. 2008;132:1049–1061. [PMC free article] [PubMed]
  • Kim J, Woo AJ, Chu J, Snow JW, Fujiwara Y, Kim CG, Cantor AB, Orkin SH. A Myc network accounts for similarities between embryonic stem and cancer cell transcription programs. Cell. 2010;143:313–324. [PMC free article] [PubMed]
  • Kruse JP, Gu W. Modes of p53 regulation. Cell. 2009;137:609–622. [PMC free article] [PubMed]
  • Laptenko O, Prives C. Transcriptional regulation by p53: one protein, many possibilities. Cell Death Differ. 2006;13:951–961. [PubMed]
  • Lee KH, Li M, Michalowski AM, Zhang X, Liao H, Chen L, Xu Y, Wu X, Huang J. A genomewide study identifies the Wnt signaling pathway as a major target of p53 in murine embryonic stem cells. Proc Natl Acad Sci U S A. 2010;107:69–74. [PMC free article] [PubMed]
  • Li M, Huang J. A new puzzling role of p53 in mouse embryonic stem cells. Cell Cycle. 2010;9:1669–1670. [PMC free article] [PubMed]
  • Lin T, Chao C, Saito S, Mazur SJ, Murphy ME, Appella E, Xu Y. p53 induces differentiation of mouse embryonic stem cells by suppressing Nanog expression. Nat Cell Biol. 2005;7:165–171. [PubMed]
  • Loh YH, Wu Q, Chew JL, Vega VB, Zhang W, Chen X, Bourque G, George J, Leong B, Liu J, et al. The Oct4 and Nanog transcription network regulates pluripotency in mouse embryonic stem cells. Nat Genet. 2006;38:431–440. [PubMed]
  • Marion RM, Strati K, Li H, Murga M, Blanco R, Ortega S, Fernandez-Capetillo O, Serrano M, Blasco MA. A p53-mediated DNA damage response limits reprogramming to ensure iPS cell genomic integrity. Nature. 2009;460:1149–1153. [PMC free article] [PubMed]
  • Mikkelsen TS, Ku M, Jaffe DB, Issac B, Lieberman E, Giannoukos G, Alvarez P, Brockman W, Kim TK, Koche RP, et al. Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature. 2007;448:553–560. [PMC free article] [PubMed]
  • Mizuno H, Spike BT, Wahl GM, Levine AJ. Inactivation of p53 in breast cancers correlates with stem cell transcriptional signatures. Proc Natl Acad Sci U S A. 2010;107:22745–22750. [PMC free article] [PubMed]
  • Murphy M, Ahn J, Walker KK, Hoffman WH, Evans RM, Levine AJ, George DL. Transcriptional repression by wild-type p53 utilizes histone deacetylases, mediated by interaction with mSin3a. Genes Dev. 1999;13:2490–2501. [PMC free article] [PubMed]
  • Neveu P, Kye MJ, Qi S, Buchholz DE, Clegg DO, Sahin M, Park IH, Kim KS, Daley GQ, Kornblum HI, et al. MicroRNA profiling reveals two distinct p53-related human pluripotent stem cell states. Cell Stem Cell. 2010;7:671–681. [PubMed]
  • Pease S, Braghetta P, Gearing D, Grail D, Williams RL. Isolation of embryonic stem (ES) cells in media supplemented with recombinant leukemia inhibitory factor (LIF) Dev Biol. 1990;141:344–352. [PubMed]
  • Rahl PB, Lin CY, Seila AC, Flynn RA, McCuine S, Burge CB, Sharp PA, Young RA. c-Myc regulates transcriptional pause release. Cell. 2010;141:432–445. [PMC free article] [PubMed]
  • Sheaff R, Ilsley D, Kuchta R. Mechanism of DNA polymerase alpha inhibition by aphidicolin. Biochemistry. 1991;30:8590–8597. [PubMed]
  • Shi X, Kachirskaia I, Yamaguchi H, West LE, Wen H, Wang EW, Dutta S, Appella E, Gozani O. Modulation of p53 function by SET8-mediated methylation at lysine 382. Mol Cell. 2007;27:636–646. [PMC free article] [PubMed]
  • Smeenk L, van Heeringen SJ, Koeppel M, van Driel MA, Bartels SJ, Akkers RC, Denissov S, Stunnenberg HG, Lohrum M. Characterization of genome-wide p53-binding sites upon stress response. Nucleic Acids Res. 2008;36:3639–3654. [PMC free article] [PubMed]
  • Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A, Pomeroy SL, Golub TR, Lander ES, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–15550. [PMC free article] [PubMed]
  • Takahashi K, Yamanaka S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell. 2006;126:663–676. [PubMed]
  • Toledo F, Wahl GM. Regulating the p53 pathway: in vitro hypotheses, in vivo veritas. Nat Rev Cancer. 2006;6:909–923. [PubMed]
  • Vousden KH, Prives C. Blinded by the Light: The Growing Complexity of p53. Cell. 2009;137:413–431. [PubMed]
  • Young RA. Control of the embryonic stem cell state. Cell. 2011;144:940–954. [PMC free article] [PubMed]
  • Yu J, Vodyanik MA, Smuga-Otto K, Antosiewicz-Bourget J, Frane JL, Tian S, Nie J, Jonsdottir GA, Ruotti V, Stewart R, et al. Induced pluripotent stem cell lines derived from human somatic cells. Science. 2007;318:1917–1920. [PubMed]
  • Zhang X, Huang J. Integrative genome-wide approaches in embryonic stem cell research. Integr Biol (Camb) 2010;2:510–516. [PMC free article] [PubMed]
  • Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nusbaum C, Myers RM, Brown M, Li W, et al. Model-based analysis of ChIP-Seq (MACS) Genome Biol. 2008;9:R137. [PMC free article] [PubMed]
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • BioProject
    BioProject links
  • Gene
    Gene records that cite the current articles. Citations in Gene are added manually by NCBI or imported from outside public resources.
  • GEO DataSets
    GEO DataSets
    Gene expression and molecular abundance data reported in the current articles that are also included in the curated Gene Expression Omnibus (GEO) DataSets.
  • GEO Profiles
    GEO Profiles
    Gene Expression Omnibus (GEO) Profiles of molecular abundance data. The current articles are references on the Gene record associated with the GEO profile.
  • HomoloGene
    HomoloGene clusters of homologous genes and sequences that cite the current articles. These are references on the Gene and sequence records in the HomoloGene entry.
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...