Oct4 Maintains the Pluripotency of Human Embryonic Stem Cells by Inactivating p53 Through Sirt1-Mediated Deacetylation

Authors

  • Zhen-Ning Zhang,

    1. Section of Molecular Biology, Division of Biological Sciences, University of California, La Jolla, California, USA
    Search for more papers by this author
  • Sun-Ku Chung,

    1. Section of Molecular Biology, Division of Biological Sciences, University of California, La Jolla, California, USA
    Search for more papers by this author
  • Zheng Xu,

    1. Section of Molecular Biology, Division of Biological Sciences, University of California, La Jolla, California, USA
    Search for more papers by this author
  • Yang Xu

    Corresponding author
    1. Section of Molecular Biology, Division of Biological Sciences, University of California, La Jolla, California, USA
    • Correspondence: Yang Xu, Ph.D., Section of Molecular Biology, Division of Biological Sciences, University of California, San Diego, 9500 Gilman Drive, La Jolla, California 92093, USA. Telephone: 858-822-1084; Fax: 858-534-0053; e-mail: yangxu@ucsd.edu

    Search for more papers by this author

Abstract

Oct4 is critical to maintain the pluripotency of human embryonic stem cells (hESCs); however, the underlying mechanism remains to be fully understood. Here, we report that silencing of Oct4 in hESCs leads to the activation of tumor suppressor p53, inducing the differentiation of hESCs since acute disruption of p53 in p53 conditional knockout (p53CKO) hESCs prevents the differentiation of hESCs after Oct4 depletion. We further discovered that the silencing of Oct4 significantly reduces the expression of Sirt1, a deacetylase known to inhibit p53 activity and the differentiation of ESCs, leading to increased acetylation of p53 at lysine 120 and 164. The importance of Sirt1 in mediating Oct4-dependent pluripotency is revealed by the finding that the ectopic expression of Sirt1 in Oct4-silenced hESCs prevents p53 activation and hESC differentiation. In addition, using knock-in approach, we revealed that the acetylation of p53 at lysine 120 and 164 is required for both stabilization and activity of p53 in hESCs. In summary, our findings reveal a novel role of Oct4 in maintaining the pluripotency of hESCs by suppressing pathways that induce differentiation. Considering that p53 suppresses pluripotency after DNA damage response in ESCs, our findings further underscore the stringent mechanism to coordinate DNA damage response pathways and pluripotency pathways in order to maintain the pluripotency and genomic stability of hESCs. Stem Cells 2014;32:157–165

Introduction

Human embryonic stem cells (hESCs) are capable of unlimited self-renewal and retain the pluripotency to differentiate into all cell types in the body. The core transcription factors required to maintain the pluripotency of ESCs include Oct4, Sox2, and Nanog [1]. In addition, the importance of Oct4 in pluripotency is further underscored by the findings that Oct4 is critical for reprogramming somatic cells into the pluripotent state [2]. Higher or lower than normal levels of Oct4 induce the differentiation of ESCs into primitive endoderm/mesoderm or trophectoderm, respectively, indicating that Oct4 is involved in cell fate decision during the early stages of embryonic development [3, 4]. Oct4 and Sox2 form heterodimers through the POU and Sox domains to bind to the promoters or enhancers of their target genes to activate or suppress their expression [5]. ChIP-Seq analyses have identified the global DNA-binding sites of Oct4 in the genome of hESCs [6, 7]. However, it remains to be fully understood which target genes of Oct4 are important for mediating its critical roles in maintaining pluripotency.

The tumor suppressor p53 maintains genomic stability by directly regulating the expression of genes involved in cell cycle arrest (p21, 14-3-3σ), apoptosis (Puma, Noxa), senescence (PAI-1), and cellular differentiation [8]. Recent studies have identified the important roles of p53 in suppressing pluripotency and promoting differentiation. p53 suppresses the expression of the pluripotency factor Nanog upon DNA damage and induces the differentiation of DNA damaged ESCs to ensure genetic stability [9]. In addition, both the protein level and activity of p53 are induced during ESC differentiation [9, 10], and p53 activates the expression of several miRNAs, which have been shown to induce the differentiation of ESCs [10]. Therefore, p53 functions as a critical coordinator of DNA damage response and pluripotency in ESCs. However, it remains unclear how pluripotency pathways regulate the p53 activity for ESCs to remain in the pluripotent state.

The stability and activity of p53 are negatively regulated by various proteins, including its E3 ligase Mdm2 [8]. The interaction between p53 and Mdm2 or other E3 ligases leads to the ubiquitination and degradation of p53, and also suppresses p53 activity [8]. The post-translational modifications of p53 such as phosphorylation and acetylation are important in regulating p53 stability and activity in response to various stresses [11]. Previous studies have shown that human p53 can be acetylated at multiple lysine residue by Tip60 and CBP, including K120, K164, K320, and the C-terminal K370/372/373/381/382/386 [12-16]. In addition, p53 is phosphorylated and acetylated during the differentiation of ESCs, leading to the activation of p53 in this process [9, 10]. In this context, the acetylation and activity of p53 are negatively regulated by Sirt1, a NAD+-dependent Class III histone deacetylase that plays important roles in a wide variety of physiological processes including stress response, metabolism, apoptosis, and aging [17-19]. In ESCs, Sirt1 attenuated Reactive oxygen species-induced apoptosis and Nanog expression by deacetylating p53 and blocking its nuclear translocation [20], and is required for the genomic stability of human ESCs [21]. While there has been a clear correlation between the acetylation and activity of p53 and Sirt1 during the differentiation of ESCs [10], the physiological importance of Sirt1-mediated deacetylation of p53 in the differentiation of hESCs remains unclear.

Recent biochemical and cell line transfection studies have indicated the involvement of the acetylation at K120 and K164 in activating p53-dependent responses to DNA damage [14-16]. While the mutation of the K120 and K164 leads to a partial defect in p53 activities in these cell line studies, the mutation of all eight acetylation sites (K120, 164, and the six C-terminal lysine residues) abolishes the p53-dependent transcription without any impact on p53 stability and its DNA-binding activity, suggesting that these acetylation events play synergistic roles in activating p53-dependent transcription [16]. Mouse knock-in studies have shown that acetylation at K120/164 is dispensable for p53 stabilization after DNA damage but is required for p53 activity [22]. However, considering the apparent difference in p53 regulation and function between mouse and human as well as between somatic and pluripotent stem cells, the physiological roles of these acetylation events in human pluripotent stem cells remain to be established.

Materials and Methods

Generation of p53CDK hESCs

To facilitate the generation of p53 conditional knockout hESCs, we took advantage of the p53+/− hESCs we previously generated [23]. p53+/− hESCs, in which exons 2–6 of one p53 allele were replaced by CAG-Neo selection cassette, are genetically stable and can undergo normal differentiation [23]. The BAC vector for conditional knockout was generated by flanking exons 2–4 of the p53 gene with two LoxP sites. The translational initiating ATG of the p53 gene is encoded by exon 2. Therefore, deletion of exons 2–4 of the p53 gene will eliminate p53 expression. The FRT-flanked CAG-Neo-Ires-Puro selection cassette was inserted into the intron 4. The BAC conditional knockout vector was electroporated into the p53+/− hESCs that were selected with puromycin. The homologous recombination between the targeting vector and the remaining WT allele in p53+/− hESCs generated the p53 conditional knockout hESCs. The homologous recombination event was screened by PCR with primers P1–3. The homologous recombination between the targeting vector and the remaining WT allele led to the loss of amplification by this PCR reaction. The homologous recombinants were confirmed by Southern blotting with NheI digestion and hybridization to a probe spanning exon 11, giving rise to a 9.6 kb germline band, a 10.1 kb knockout band, and a 14 kb conditional knockout band. The selection marker cassette inserted into the conditional knockout allele suppresses p53 expression, leading to a p53-null phenotype. By excising FRT-flanked selection marker from the genome with transient expression of FLP in hESCs, WT p53 is expressed from the conditional knockout allele, leading to p53 conditional knockout (p53CDK) hESCs. The presence of the two LoxP sites flanking the exons 2–4 in the conditional knockout allele was confirmed by DNA sequencing. Expression of Cre in p53CDK hESCs or their derivatives will lead to LoxP/Cre-mediated deletion of exons 2–4 of the p53 gene from the conditional knockout allele, leading to a p53-null genotype.

Generation of K2R p53 Acetylation Site Knock-in hESCs

The BAC clone containing human p53 gene (RP11–199F11) was purchased from Invitrogen (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). The targeting vector was generated through recombineering in the Escherichia coli strain SW102 as previously described [23]. To obtain p53+/− hESCs, Neomycin marker was initially used for selection [23]. To target the WT allele in p53+/− hESCs, the knock-in vector was constructed by inserting the FRP-flanked selection cassette (CAG-Neo-IRES-Puro-polyA) into the intron four of the p53 gene. To eliminate the selection cassette from the knock-in allele, a plasmid encoding the FLP gene was transiently transfected into the knock-in hESCs. The full-length p53 cDNA derived from the knock-in hESCs was sequenced to confirm that only the K2R mutation but no other mutation was introduced into knock-in allele.

Cell Culture for hESCs

The hESC lines were cultured on feeder layer in knockout Dulbecco's modified Eagle's medium supplemented with 10% knockout serum replacement, 10% plasmanate, 1% PenStrep, 1% Glutamine, 1% nonessential amino acids, 10 ng/ml bFGF (basic fibroblast growth factor), and 55 µM β-mercaptoethanol, as described [23]. For Oct4-depletion-induced differentiation, hESCs were cultured on matrigel-coated plates in the mTeSR1 medium with 5× supplement (05850, Stem Cell Technologies, BC, Canada, http://www.stemcell.com). To induce the differentiation of hESCs by retinoic acid (RA), hESCs were cultured in the differentiation medium (hESC medium without FGF) containing 1 µM RA for 3 days. All human ESC work has been approved by UCSD institutional ESCRO and IRB.

Oct4 Knockdown

The knockdown of Oct4 in hESCs was performed as previously described [24]. The hESCs were cultured under the feeder-free condition using hESC-qualified Matrix (354277, BD Biosciences, Biosciences, San Jose, CA, http://www.bdbiosciences.com) and mTeSR1 defined medium (05850, Stem Cell Technologies). The knockdown constructs were transfected into hESCs with electroporation. A Rho-associated kinase inhibitor Y27632 (04-0012, Stemgent, Inc., Cambridge, MA, https://www.stemgent.com) was added to the media for 24 hours after transfection. Puromycin (0.3 mg/ml) was added to the media 24 hours after transfection to enrich the shRNA-transfected cells.

Western Blotting

Protein extracts from transfected hESCs were resolved on 6%–10% SDS-PAGE gels and transferred to nitrocellulose membrane, which was probed with a monoclonal antibody against Oct3/4 (sc-5279; Santa Cruz, Dallas, Texas, http://www.scbt.com), monoclonal antibody against p53 (pAb1801; Santa Cruz), polyclonal antibody against Sirt1 (Ab13749; Abcam, Cambridge, MA, http://www.abcam.com), p53 acetylation-specific antibodies against acetylated K120 and K164 (gifts from Dr. Wei Gu) or against acetylated K382 (2525s, Cell Signaling, Danvers, MA, http://www.cellsignal.com), polyclonal p53 phosphorylation-specific antibody against phosphorylated Ser15 (9284l, Cell Signaling), monoclonal antibodies against Mdm2 (Ab-2; Oncogene Research, Boston, MA, http://www.apoptosis.com), β-actin (Santa Cruz Biotechnology), or tubulin (clone B-5-1-2; Sigma, St. Louis, MO, http://www.sigmaaldrich.com). Membrane was subsequently probed with a horseradish peroxidase-conjugated secondary antibody and developed with ECL detection kit (Amersham Biosciences, Pittsburgh, PA, http://www.gelifesciences.com). For coimmunoprecipitation analysis, 1–2 mg of whole cell protein extracts was immunoprecipitated with anti-p53 antibody (FL393; Santa Cruz Biotechnology). The amount of p53 and Mdm2 in the immunoprecipitate was analyzed by Western blotting using monoclonal antibody against p53 (pAb1801; Santa Cruz), monoclonal antibody against Mdm2 (SMP-14; Santa Cruz Biotechnology).

Quantitative Real-Time PCR

Real-time PCR was performed as previously described [25]. Briefly, total RNA was purified from hESCs with RNeasy Mini kit (Qiagen, 74106). Total RNA (1 µg) was reversely transcribed into cDNA and analyzed by quantitative real-time PCR with FastStart Universal SYBR Green Master (ROX) (Roche, 04 913 850 001). The sequences of the primers used were listed in Supporting Information Table S1. The average threshold (Ct) was determined for each gene and normalized to GAPDH mRNA levels.

Flow Cytometric Analysis

hESCs were stained with either phycoerythrin (PE)-conjugated anti-SSEA3 antibody (560879, BD Pharmingen), anti-TRA-1–60 antibody (560193, BD Pharmingen), or PE-conjugated isotype-matched normal antibody (553943, BD Pharmingen) as negative control, and analyzed by a BD LSR-II using FACS Diva software (Becton Dickinson) as we previously described [23].

Immunostaining

hESCs were cultured on Matrigel-coated coverslips in mTESR1 medium and fixed with 4% paraformaldehyde in Phosphate-Buffered Saline (PBS) for 15 minutes at room temperature (RT). After being washed three times with PBS, cells were incubated in 0.3% Triton X-100 in TBS (TBST) for 10 minutes at RT. After blocked with blocking buffer (2% Fetal Bovine Serum in TBST) for 20 minutes at RT, the cells were incubated with anti-Oct4, anti-p53, or anti-Sirt1 antibody overnight at 4°C, followed by being washed and incubated with the secondary antibody (Alexa-Fluor goat anti-rabbit 488 for Oct4 and Alexa-Fluor goat anti-mouse 568 for p53 or Alexa-Fluor goat anti-mouse 488 for Oct4 and Alexa-Fluor goat anti-rabbit 568 for Sirt1) for 45 minutes in the dark at RT. Coverslips were mounted on glass slides using Vectashield mounting medium for fluorescence. The cells were examined and photomicrographed using an Olympus confocal microscope.

Analysis of p53 Stability

For in vivo p53 degradation assays, hESCs were incubated with cycloheximide (100 µg/ml). Cells were harvested at 20, 40, and 80 minutes later, subjected to Western blotting using anti-p53 antibody (pAb1801; Santa Cruz Biotechnology). The intensity of bands was measured by densitometry.

Chromatin Immunoprecipitation

Chromatin immunoprecipitation (ChIP) assay was performed essentially according to the Pierce Agarose ChIP kit (cat# 26156). Briefly, human ESCs were incubated in culture media containing 1% formaldehyde with gentle shaking for 15 minutes at RT, and crosslinking was stopped by addition of 2.5 M glycine to a final concentration of 0.125 M glycine. After two washes with cold PBS, cells were harvested and their genomic DNA was digested with 2.5 U Micrococcal Nuclease (ChIP Grade) to an average of 500 bp after membrane extraction. Aliquots of nuclear extracted supernatants were used for immunoprecipitation by 2 µg Oct4 antibodies (Cat# 5677, Cell Signaling). After extensive washing with wash buffer 1–3 provided in the kit for 5 minutes each time, the proteins were eluted from the beads by 150 µl elution buffer. The DNA samples were recovered by ChIP prep Module after reversal of crosslinking. The purified DNA was then analyzed by quantitative real-time PCR using Applied Biosystems StepOnePlus.

Results

To investigate the roles of Oct4 in maintaining the pluripotent state of hESCs, we developed an episomal vector to efficiently silence the expression of Oct4 in hESCs (Fig. 1A) [24]. As expected, the silencing of Oct4 in hESCs leads to their rapid differentiation as indicated by the downregulation of hESC-specific surface marker SSEA3 and TRA-1–60 (Fig. 1B). Consistent with this notion, the expression levels of the pluripotency genes such as Nanog, FoxD3, Rex1, Cripto, and Tert were reduced by Oct4 silencing, while the expression of lineage-specific genes such as endodermal marker GATA6 and Osteopontin, mesodermal marker Brachyury, ectodermal marker Cdx2, and left-right determination factor 1 (Lefty) were greatly increased (Fig. 1C). Interestingly, the silencing of Oct4 in hESCs led to increased p53 protein levels and p53-dependent transcription (Fig. 1D), but slightly decreased p53 mRNA levels, suggesting increased stability of p53 protein in Oct4-depleted hESCs (Fig. 1E). In support of this finding, immunofluorescence analysis indicated that the protein levels of p53 were higher in Oct4-silenced hESCs when compared with mock-treated hESCs (Supporting Information Fig. S1A). Because p53 suppresses the pluripotency of ESCs as well as induced pluripotency [9, 26-31], and previous findings have shown that p53 activation is important for RA-induced differentiation of hESCs [10], the activation of p53 might play an important role in driving the differentiation of hESCs after Oct4 depletion.

Figure 1.

Oct4 maintains the pluripotency of human embryonic stem cells (hESCs) by suppressing p53. (A): Silencing of Oct4 in hESCs with an episomal vector expressing either nonspecific RNAi (USi) or Oct4-specific RNAi (Oct4i). The episomal vector contains a puromycin resistance gene, allowing the selection of the transfectants with puromycin. The protein levels of Oct4 in hESCs at different time points after transfection were determined by Western blotting with tubulin as the loading control. (B): Silencing of Oct4 in hESCs leads to spontaneous differentiation. Flow cytometric analysis of the surface expression of SSEA3 and TRA-1–60 7 days after the transfection with USi (red) or Oct4i vector (blue), or untransfected control (orange). (C): Quantitative real-time PCR analysis of the mRNA levels of pluripotency genes and lineage-specific genes 7 days after transfection of the knockdown vector. Data are means ± sdv from three independent experiments. (D): Silencing of Oct4 in hESCs stabilizes and activates p53. Left panel, the protein levels of p53 and Oct4 in Oct4-silenced hESCs 7 days after transfection. Right panel, qPCR analysis of the expression levels of p53 target genes p21 and Puma 7 days after transfection. Mean data ± sdv from three independent experiments are shown. (E): Silencing of Oct4 in hESCs has little impact on p53 mRNA levels. qPCR analysis of the mRNA levels of p53 7 days after transfection. Mean data ± sdv from three independent experiments are shown. Abbreviations: BRA, brachyury; MEF, mouse embryonic fibroblast; OPN, osteopontin.

To test this hypothesis, we examined the differentiation of Oct4-depleted hESCs when p53 was disrupted in the same hESCs. p53 is critical for maintaining the genomic stability of hESCs, and p53−/− hESCs accumulate extensive genomic instability after several passages [23]. Therefore, p53−/− hESCs are not suitable for this purpose. To address this issue, we developed a p53 conditional knockout (p53CKO) hESCs, in which the exons 2–6 of one p53 allele were replaced with CAG-Neor and the exons 2–4 of the other allele flanked with two LoxP sites (Supporting Information Fig. S2). By expressing the Cre enzyme from the same episomal vector used to silence Oct4, we were able to simultaneously silence the expression of p53 and Oct4 in the same hESCs (Fig. 2A). By examining the surface expression of hESC-specific markers as well as the expression of the pluripotency genes and lineage-specific genes, p53-deficiency effectively prevented the differentiation of Oct4-depleted hESCs (Fig. 2B–2D). In this context, the expression of Oct4 target pluripotency genes such as Nanog, FoxD3, and Rex1 remained reduced in the Oct4- and p53-depleted hESCs. In contrast, the expression of other pluripotency markers such as Cripto and Tert was rescued by the p53 deficiency in hESCs after Oct4 depletion, consistent with the inhibition of hESC differentiation and the expression of lineage-specific genes (Fig. 2D). These findings indicate a novel role of Oct4 in maintaining the pluripotency of hESCs by suppressing the activity of p53.

Figure 2.

Disruption of p53 prevents the differentiation of Oct4-depleted human embryonic stem cells (hESCs). (A): Acute disruption of p53 in Oct4-depleted p53CKO hESCs via LoxP/Cre-mediated deletion. Left panel, the protein levels of p53 and Oct4 in p53CKO hESCs 7 days after the transfection of the episomal vector expressing Cre and Oct4i. Right panel, qPCR analysis of the mRNA levels of p53 target genes p21 and Puma 7 days after transfection. (B): Acute disruption of p53 in Oct4-silenced p53CKO hESCs prevents their differentiation. Flow cytometric analysis of SSEA3 and TRA-1–60 expression in p53CKO hESCs 7 days after transfection with USi (red), Oct4i (blue), or CRE-Oct4i plasmid (orange). (C): Phase-contrast images of hESCs cultured on matrigel 7 days after transfection with episomal vector expressing USi, Oct4i, or CRE-Oct4i. Scale bar = 100 µm. (D): Quantitative real-time PCR analysis of the mRNA levels of the pluripotency genes and lineage-specific genes 7 days after transfection of the episomal vector expressing USi, Oct4i, or Cre-Oct4i. Data are means ± sdv from three independent experiments. Abbreviations: BRA, brachyury; OPN, osteopontin.

The p53 gene is not a direct transcriptional target of Oct4 and the p53 mRNA levels are slightly decreased in Oct4-silenced hESCs, suggesting that the increased protein levels of p53 in Oct4-depleted hESCs are primarily through the post-translational mechanisms (Fig. 1E). By examining the published ChIP-seq data of the global binding sites of Oct4, we found that Oct4 binds to the promoter region of Sirt1, a deacetylase known to inactivate p53 through deacetylation (Fig. 3A). Using quantitative ChIP analysis, we confirmed that Oct4 was bound to its consensus-binding site within the promoter region of Sirt1 (Fig. 3B). In addition, both Sirt1 mRNA levels (Fig. 3C) and protein levels (Fig. 3D; Supporting Information Fig. S1B) were significantly decreased in the Oct4-silenced hESCs. Together, these data support the notion that Oct4 directly activates the transcription of Sirt1. We speculated that the reduced Sirt1 expression in Oct4-silenced hESCs could account for the p53 activation by reducing the p53 acetylation. In support of this notion, the acetylation of p53 at K120, K164, and K382, also the targets of the deacetylase Sirt1 [18], was increased in Oct4-silenced hESCs, while phosphorylation of p53 at Ser15 was not affected (Fig. 3D). Furthermore, when the expression levels of Sirt1 were restored to normal levels in the Oct4-silenced hESCs by the ectopic expression of Sirt1, this inhibited the p53 activation in Oct4-silenced hESCs and the differentiation of Oct4-silenced hESCs (Fig. 3E–3G).

Figure 3.

Oct4 inactivates p53 by inducing the expression of Sirt1. (A): The consensus DNA-binding site of Oct4 in the promoter region of Sirt1. (B): Chromatin immunoprecipitation analysis confirms that Oct4 binds to the promoter region of Sirt1. Mean value from three independent experiments is presented with sdv. (C): Silence of Oct4 in human embryonic stem cells (hESCs) leads to the reduction of the mRNA levels of Sirt1. Mean value from three independent experiments is presented with sdv. (D): Ectopic expression of Sirt1 in Oct4-silenced hESCs prevents the acetylation and activation of p53. Left panel, the ectopic expression of Sirt1 in Oct4-silenced hESCs suppresses p53 acetylation and activation. Right panel, qPCR analysis of the mRNA levels of the p53 target genes p21 and Puma 7 days after transfection. Data are means ± sdv from three independent experiments. (E): Ectopic expression of Sirt1 in Oct4-silenced hESCs prevents their differentiation. Flow cytometric analysis of SSEA3 and TRA-1–60 expression in hESCs 7 days after transfection with USi (red), Oct4i (blue), or Oct4i/Sirt1 expressing plasmid (orange). (F): Phase-contrast images of hESC colonies cultured on matrigel 7 days after transfection with USi, Oct4i, or Oct4i/Sirt1 expressing episomal vector. Scale bar = 100 µm. (G): Quantitative real-time PCR analysis of the mRNA levels of the pluripotency genes and lineage-specific genes 7 days after transfection of the episomal vector expressing USi, Oct4i, or Oct4i/Sirt1. Data are means ± sdv from three independent experiments. Abbreviations: BRA, brachyury; OPN, osteopontin.

To further determine the importance of the acetylation events of p53 at K120 and 164 in activating p53 in hESCs, we introduced two knock-in mutations (K120/164R) into the WT allele of the p53+/− hESCs, generating the K2R knock-in hESCs (Supporting Information Fig. S3), and the loss of K120/164 acetylation was verified by Western blotting (Fig. 4A). The p53 protein levels in K2R hESCs were lower than those in p53+/− hESCs and could not be increased by DNA damage, indicating that the two acetylation events are important for p53 stabilization in hESCs (Fig. 4A). To elucidate the basis of the reduced p53 protein levels in K2R cells, we analyzed the half-life of the p53 protein in p53+/− and K2R hESCs. Our findings indicated that the half-life of p53 in p53+/− hESCs is significantly longer than that in K2R hESCs, indicating that p53 is unstable in K2R (Fig. 4B). Consistent with this finding, the interaction between p53 and its E3 ligase Mdm2 was significantly increased in K2R hESCs, indicating that the increased interaction between p53 and Mdm2 in K2R cells led to p53 destabilization (Fig. 4C). In further support of this conclusion, the p53 protein levels in K2R hESCs could be increased by the treatment with Nutlin-3, which specifically disrupts the interaction between p53 and Mdm2 (Fig. 4D). However, the p53-dependent gene expression is essentially abolished in K2R hESCs even after the Nutlin-3 treatment and DNA damage (Fig. 4E). Therefore, the acetylation events at K120 and K164 are required for p53 stabilization and activity in hESCs.

Figure 4.

The stability and activity of p53 in K2R human embryonic stem cells (hESCs). (A): The acetylation and stabilization of p53 in p53+/− and K2R hESCs after doxorubicin treatment. The acetylated p53, total p53, and actin are indicated. (B): The protein levels of p53 in p53+/− and K2R hESCs at different time points after treatment with CHX that inhibits protein synthesis. (C): The interaction between p53 and Mdm2 in p53+/− and K2R hESCs was analyzed by coimmunoprecipitation. p53−/− hESCs were used as a negative control. (D): The stabilization of p53 in p53+/− and K2R hESCs by treating with Nutlin-3 for 24 hours. The concentrations of Nutlin-3 used are indicated on the top. (E): p53-dependent gene expression in p53+/−, K2R, and p53−/− hESCs after doxorubicin and/or Nutlin-3 treatment. The mRNA levels of p53 target genes were determined by qPCR and standardized by the mRNA levels of GAPDH. Mean value ± sdv from three independent experiments are shown. Abbreviation: CHX, cycloheximide.

To study the functional importance of Sirt1-mediated deacetylation of p53 in maintaining Oct4-dependent pluripotency, we examined the p53 stability and activity in Oct4-silenced K2R hESCs. Oct4 depletion greatly increased p53 protein levels and p53-dependent transcription in p53+/− hESCs, but not in K2R hESCs, indicating that these two acetylation events are important to stabilize p53 in Oct4-depleted hESCs (Fig. 5A). In addition, the differentiation of K2R hESCs after Oct4 depletion, as indicated by the downregulation of hESC-specific surface marker and pluripotency genes as well as the increased expression of various lineage-specific genes, was also abolished (Fig. 5B, 5C).

Figure 5.

Silencing of Oct4 in K2R human embryonic stem cells (hESCs) fails to activate p53 and induce differentiation. (A): Silencing of Oct4 stabilizes and activates p53 in p53+/− hESCs but not in K2R hESCs. Left panel, the protein levels of p53 and Oct4 in Oct4-silenced p53+/− hESCs and K2R hESCs 7 days after transfection. Right panel, qPCR analysis of the mRNA levels of p53 target genes p21 and Puma in p53+/− hESCs and K2R hESCs 7 days after transfection of the knockdown vector. Mean value ± sdv from three independent experiments are shown. (B): Silencing of Oct4 induces the differentiation of p53+/− hESCs but not K2R hESCs. Flow cytometric analysis of the surface expression of SSEA3 or TRA-1–60 on p53+/− hESCs and K2R hESCs 7 days after transfection. (C): Quantitative real-time PCR analysis of the expression levels of the pluripotency genes and lineage-specific genes in p53+/− hESCs and K2R hESCs 7 days after transfection. Data are mean value ± sdv from three independent experiments. (D): Schematic diagram of the Oct4-Sirt1-p53 pathway in maintaining the pluripotency of hESCs. Abbreviations: BRA, brachyury; OPN, osteopontin.

To further examine the roles of the acetylation of p53 at K120 and K164 in the differentiation of hESCs, we examined the RA-induced differentiation of p53+/− and K2R hESCs. Consistent with the published data [10], the treatment of hESCs with RA leads to p53 activation and differentiation of hESCs as indicated by the reduced expression of pluripotency genes and the increased expression of various lineage-specific genes (Supporting Information Fig. S4). In addition, Sirt1 protein levels were significantly decreased while p53 acetylation at K120 and K164 was greatly increased under RA stimulation (Supporting Information Fig. S4). In response to RA treatment, p53 protein levels and p53-dependent transcription were increased only in p53+/− hESCs but not in K2R hESCs, indicating that the acetylation is important for activating p53 during RA-induced differentiation (Supporting Information Fig. S5A). In addition, RA-induced differentiation of K2R hESCs was inhibited as indicated by the expression of hESC-specific surface marker, pluripotency genes, and various lineage-specific genes (Supporting Information Fig. S5B, S5C). Therefore, these two acetylation events are required for activating p53 during the differentiation of hESCs in response to various stimuli.

Discussion

It has been well accepted that Oct4 plays important roles in maintaining the pluripotency of ESCs by regulating the expression of a network of pluripotency genes such as Nanog and Sox2 [1]. Here, we have identified a new mechanism for Oct4 to maintain pluripotency of hESCs by inducing the expression of Sirt1, a known inhibitor of p53 activity. In this context, by establishing the acetylation site knock-in hESC lines, our study provides the physiological importance of acetylation at K120 and K164 in stabilizing and activating p53 in hESCs. We further demonstrate that these acetylation events represent a common mechanism to activate p53 in response to distinct differentiation stimuli such as Oct4 depletion or RA treatment. This conclusion is important because it is the first time to demonstrate that post-translational modification is indispensable for both p53 stabilization and transcription activity in hESCs. In the mouse knock-in studies, the corresponding K120/164R mutations have no impact on the p53 stability but abolish p53-dependent transcription [22]. This discrepancy might be due to the differential regulation of p53 between mouse and human cells or between somatic cells and pluripotent stem cells. Together with the findings that p53 can induce the differentiation of ESCs by suppressing the expression of the other pluripotency gene Nanog [9], these data support a novel reciprocal feedback regulatory mechanism in hESCs to coordinate the DNA damage responses and pluripotency (Fig. 5D).

When compared with somatic cells, pluripotent stem cells have lower levels of oxidative phosphorylation and primarily rely on glycolysis as the source of energy [32]. Recent gene profiling of hESCs and Oct4-depleted hESCs also suggests that Oct4-depletion leads to increase in expression of genes involved in oxidative stresses and decrease in genes involved in glycolysis [24, 33]. Since p53 is known to suppress glycolysis [34], our findings that Oct4 inactivates p53 in hESCs provide a mechanism how Oct4 promotes anaerobic metabolism in hESCs. Since pluripotency is dependent on anaerobic metabolism and excessive oxidative stress is detrimental to ESCs [32], the key roles of Oct4 in maintaining pluripotency could depend on its regulation of metabolism in hESCs by inactivating p53.

It has been reported that, when ectopically expressed in the adult tissues, Oct4 exhibits oncogenic potential [35]. In addition, Oct4 is frequently overexpressed in many types of human cancers and its expression correlated with the poor prognosis of the cancer patients [36, 37]. Our findings that Oct4 suppresses the activity of p53, which is the most crucial tumor suppressor, could account for the oncogenic potential of Oct4 in tumorigenesis. Oct4 is also one of the key reprograming factors required for reprogramming somatic nucleus into pluripotent state [2]. Considering that p53 is a potent inhibitor of induced pluripotency [26], this novel functional interaction between Oct4 and p53 could also account for the irreplaceable roles of Oct4 in induced pluripotency.

Conclusion

While the involvement of Sirt1 and p53 in the differentiation of hESCs has been well-characterized [9, 10], our studies have produced several novel findings important to understand the mechanism safeguarding the pluripotency of ESCs. First, the mechanism underlying Oct4-dependent pluripotency remains to be fully understood. Our findings reveal a novel role of Oct4 in maintaining the pluripotency of hESCs by suppressing the pathways required to induce the differentiation of hESCs. This novel Oct4-Sirt1-p53 axis further underscores the stringent reciprocal regulation between pluripotency pathways and the DNA damage pathways important to maintain pluripotency and genomic stability of hESCs. Second, while previous studies have produced a correlation between Sirt1 and p53 in promoting the differentiation of hESCs [10], our studies provide the first physiological evidence to prove that the acetylation of p53 at K120/K164, which is inhibited by Sirt1, is required to activate p53 response to induce hESC differentiation. This conclusion shed light on the mechanism underlying the functional interaction between Sirt1 and p53 in hESCs. Third, while recent knock-in studies have shown that acetylation of p53 at K120/164 in mice is required for p53 activity but is dispensable for p53 stabilization [16, 22], our findings that these acetylation events are required for both p53 stabilization and activity in hESCs further highlights the species-specific (mouse vs. human) roles of post-translational modification of p53.

Acknowledgments

We thank Drs. Sarah Kinnings and Wei Wang for help with the bioinformatics analysis and Dr. Wei Gu for providing p53 acetylation-specific antibodies. This work was supported by grants from NIH (CA094254) and California Institute for Regenerative Medicine (RC1–0148) to Y.X.

Author Contributions

Z.N.Z. and S-K.C.: conception and design, collection and/or assembly of data, data analysis and interpretation, and manuscript writing; Z.X.: collection and/or assembly of data and data analysis and interpretation; Y.X.: conception and design, financial support, administrative support, data analysis and interpretation, and manuscript writing. Z.-N.Z. and S.-K.C. contribute equally to this work.

Disclosure of Potential Conflicts of Interest

The authors indicate no conflict of interest.

Ancillary