• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of fasebjClick here to go to the full-featured web site.Click here to learn how to submit an article to The FASEB Journal.Never miss an issue. Click here to subscribe to The FASEB Journal s full content.Click here for the latest news from The FASEB Journal.Find out what s being published every day. Click here to subscribe to RSS feeds.Click here to visit The FASEB Journal's full-featured site.
FASEB J. Jan 2013; 27(1): 232–242.
PMCID: PMC3528318

Methylation of RASSF1A gene promoter is regulated by p53 and DAXX


Inactivation of the tumor suppressor Ras-association domain family 1 isoform A (RASSF1A) due to epigenetic silencing occurs in a variety of human cancers, and still largely unknown are the regulators and mechanisms underlying RASSF1A gene promoter methylation. Herein, we report that this methylation is regulated by p53 and death-associated protein 6 (DAXX) in acute lymphoblastic leukemia (ALL). We found that p53 bound to the RASSF1A promoter, recruiting DAXX as well as DNA methyltransferase 1 (DNMT1) for DNA methylation, which subsequently resulted in inactivation of RASSF1A in wild-type p53 ALL cells. Although the presence of p53 was required for the recruitment of DAXX and DNMT1 to the RASSF1A promoter, fluctuation in p53 protein levels did not affect the rates of RASSF1A methylation. Conversely, methylation of RASSF1A promoter was critically controlled by DAXX, as the enforced overexpression of DAXX led to enhanced RASSF1A promoter methylation, whereas inhibition of DAXX reduced RASSF1A methylation. Interestingly, we found that the p53/DAXX-mediated RASSF1A methylation regulated murine double minute 2 (MDM2) protein stability in ALL. Our results reveal a novel function for p53 in the methylation of RASSF1A promoter by its interaction with DAXX. Discovery of this mechanism provides new insight into the interactions among the tumor-related factors p53, RASSF1A, DAXX, and MDM2 in cancer pathogenesis.—Zhang, H., He, J., Li, J., Tian, D., Gu, L., Zhou, M. Methylation of RASSF1A gene promoter is regulated by p53 and DAXX.

Keywords: acute lymphoid leukemia, epigenetic gene silencing, protein-DNA binding

Ras-association domain family 1 isoform A (RASSF1A) is a tumor suppressor that is reported to have multiple functions. Through interaction with different cellular proteins, it regulates several key biological processes, including cell-cycle progression and apoptosis (1, 2). For example, RASSF1A can associate with murine double minute 2 (MDM2) and death-associated protein 6 (DAXX) in the nucleus, thereby disrupting formation of the MDM2-DAXX-HAUSP complex, resulting in the promotion of MDM2 self-ubiquitination and, thus, prevention of p53 degradation (3). In fact, RASSF1A is one of the most frequently inactivated proteins in human cancers, and inactivation of RASSF1A is implicated in cancer development and promotion (4). Unlike tumor suppressor p53 inactivation that occurs mainly by gene mutation, the most common reason for loss of RASSF1A function is transcriptional silencing through promoter hypermethylation (5). Loss or reduced RASSF1A expression due to epigenetic silencing occurs in a variety of human cancers (6). Still largely unknown are the regulators and mechanisms underlying RASSF1A promoter methylation in individual types of cancer.

The p53 tumor suppressor gene plays an important role in inducing apoptosis in response to various types of cellular stresses. P53 functions as a transcription factor that can either positively or negatively regulate transcription of a particular gene promoter (7). Previous studies have shown that p53 induces apoptosis by activating apoptosis-promoting genes such as Bax, DR5, and Fas (8,10), or by repressing apoptosis-inhibiting genes such as MDR, Bcl-2, and survivin (11,13). A study using microarrays representing >33,000 individual human genes aimed at identifying differentially expressed genes in response to p53 has found that a total of 1501 genes (4.4%) responded to p53, and ~80% of these were repressed by p53 (14). In activating gene expression, p53 functions via DNA sequence-specific binding (15). Transcriptional repression mediated by p53 is complex and may occur via multiple mechanisms. For example, p53 has been shown to suppress survivin expression by both direct DNA binding and interaction with DNMT1 to induce DNA methylation (16).

DAXX is a H3.3-specific histone chaperone, which is initially identified as an apoptosis-regulating protein that links the death receptor Fas to the c-Jun NH2-terminal kinase JNK (17). Predominantly localized in the nucleus, DAXX binds to sumo-modified PML-oncogenic domains forming the nuclear body (18). DAXX is involved in many key cellular processes, such as cell proliferation, differentiation, pro- and antiapoptosis, and transcription regulation (19). For example, DAXX interacts with MDM2 and HAUSP to form a tertiary complex, which inhibits the self-ubiquitination of MDM2, maintaining MDM2 ligase activity toward p53 (20). As a transcriptional regulator, DAXX can associate with many transcription factors and participates in regulation, either activating or repressive, of many target genes (21,23). It is reported that DAXX interacts with the NF-κB member RelB, repressing the RelB target promoter via DNA methyltransferase 1 (DNMT1) recruitment and DNA hypermethylation (24).

Based on the previous observations for the interaction and functional role of p53 and DAXX in mediating gene expression, we investigated whether p53 and DAXX have roles in the regulation of RASSF1A epigenetic silencing. We found that wild-type (WT)-p53 indeed participates in RASSF1A promoter methylation. We also revealed that DAXX plays a critical role in p53-dependent methylation of RASSF1A promoter. Herein, we delineate the mechanism by which p53 interacts with DAXX to induce gene promoter methylation of RASSF1A, which ultimately leads to MDM2 stabilization and promotion of cancer growth.


Cell lines

The 15 acute lymphoblastic leukemia (ALL) cell lines used in this study have known p53 status and MDM2 expression, as characterized in a prior publication (25). Nine of these cell lines (EU-1 to EU-9) were established at Emory University, whereas Sup-B7, Sup-B13, Sup-B15, and UOC-B3 were obtained from Stephen D. Smith (Department of Pediatrics, University of Illinois College of Medicine at Peoria, Peoria, IL, USA), and the Reh and KT cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA).

Plasmids and transfection

The expression plasmids for both WT-p53 (pC53-SN3) and different types of mutant p53 (pC53-143, pC53-175, pC53-248, and pC53-273) were provided by B. Vogelstein (Johns Hopkins University, Baltimore, MD, USA). The DAXX expression plasmid was generated by cloning the DAXX cDNA into the pCMV-HA vector (Clontech, Mountain View, CA, USA). To generate the RASSF1A full promoter and various deletion constructs, PCR primer pairs were determined from the corresponding sites in the RASSF1A promoter sequence. The Quick Change site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) was used to mutate the C and G residues within the RASSF1A 10-bp motif that strongly interfered with its binding to p53. Next, constructs including the different deleted or mutated fragments were ligated to the pGL3 basic vector (Promega, Madison, WI, USA). DNA sequencing was then performed to confirm that the sequence of the PCR products were correct when compared to the sequence of the RASSF1A gene promoter found by search of the human genome online.

Gene transfections were performed in EU-4 cells by electroporation at 290 V, 950 μF, using a Gene Pulser II System (Bio-Rad, Hercules, CA, USA). For the gene reporter assay, cells were transiently cotransfected with the RASSF1A promoter-luciferase constructs, as described above, plus p53 (WT and mutant) expression plasmids. The Renilla luciferase (pRL)-CMV vector (Promega) was also cotransfected, as an internal control. The transfected cells were resuspended in 10 ml of RPMI containing 10% FBS and incubated 24–36 h, and then cell extracts were prepared with 1 × lysis buffer. Supernatant aliquots of 20 μl were mixed first with 100 μl of Luciferase Assay Reagent II (Promega), to measure the firefly luciferase activity, and then Stop & Glo Reagent (Promega) was added to the same sample to measure Renilla luciferase activity. These luciferase activities were analyzed via Microplate Instrumentation (BioTek, Winooski, VT, USA).

siRNA and transfection

The siRNAs used to target DAXX and p53 were purchased from Dharmacon (Chicago, IL, USA). A scrambled siRNA was provided as a control. Cells were transfected with 50-200 nM of chemically synthesized siRNA using Lipofectamine reagent (Invitrogen, Carlsbad, CA, USA). Briefly, the cells were seeded at 1 × 106/well in 12-well plates and cultured overnight. Then siRNA solution was mixed with Lipofectamine reagent in Opti-MEM medium (Invitrogen) for 20 min, following the manufacturer's protocol, prior to addition to the plated cells in a total volume of 0.5 ml. After 4 h incubation, 1 ml of normal medium was added to the cell culture. All treatments were performed in triplicate.

Bisulfite modification and methylation analysis

Methylation of the RASSF1A promoter was determined by bisulfite modification, methylation-specific PCR (MSP), and genomic sequencing. For bisulfite modification, 1 μg of genomic DNA was first denatured and then treated with sodium bisulfite, using the CpGenome DNA modification kit (Chemicon International, Temecula, CA, USA) to convert all of the unmethylated cytosines to uracils, whereas the methylated cytosines remained unchanged. The modified DNA was purified and subjected to MSP or sequencing.

MSP was performed as described previously (26), using the methylation-specific primers 5′-GTGTTAACGCGTTGCGTATC-3′ and 5′-AACCCCGCGAACTAAAAACGA-3′ and the unmethylation-specific primers 5′-TTTGGTTGGAGTGTGTTAATGTG-3′ and 5′-CAAACCCCACAAACTAAAAACAA-3′, plus 100 ng of bisulfite-modified DNA. PCR was conducted using the GeneAmp DNA Amplification Kit and AmpliTaq Gold polymerase (Applied Biosystems, Foster City, CA, USA). The optimized thermal profile included initial denaturing at 95°C for 12 min, followed by 35 or 55 cycles at 95°C for 45 s, at 60°C for 45 s, then 72°C for 1 min, and a final extension at 72°C for 10 min. PCR products were analyzed by agarose gel electrophoresis and ethidium bromide staining.

The quantity of RASSF1A promoter that was methylated was determined by a bisulfite genomic sequencing protocol (5). Briefly 100 ng bisulfite-modified DNA was amplified in a reaction buffer using 2 primers, 5′-GTTTTGGTAGTTTAATGAGTTTAGGTTTTTT-3′ and 5′-ACCCTCTTCCTCTAACACAATAAAACTAACC-3′. PCR was performed at 95°C for 1 min, 55°C for 1 min, and 74°C for 2 min in 30 cycles. Next, seminested PCR was performed under similar conditions using 1/50 of the amplified product and the primers 5′-GTTTTGGTAGTTTAATGAGTTTAGGTTTTTT-3′ and 5′-CCCCACAATCCCTACACCCAAAT-3′. The PCR products were purified and then ligated into the pCR2.1 vector (Invitrogen). Six to 10 clones were sequenced in order to obtain average methylation levels.

In vitro methylation analysis

The effects of p53 and DAXX on DNMT1-mediated RASSF1A promoter methylation were measured by in vitro methylation assay. Briefly, cell lysates were prepared from EU-1 cells and then digested with DNase to remove any contaminated endogenous DNA. Cell lysates were immunoprecipitated with antibodies against p53, DAXX, and DNMT1, respectively. Next, the immunocomplexes were incubated with RASSF1A promoter DNA that was a PCR product of the RASSF1A promoter plasmid. After centrifugation and washing, the DNA was eluted, purified, and subjected to MSP assay, as described above. In addition, the in vitro methylation assay was performed by directly using recombinant p53, DAXX, and DNMT1 proteins and the purified PCR products of the RASSF1A promoter DNA, to measure the effect of p53 and DAXX on DNMT1-mediated methylation of RASSF1A.

Chromatin immunoprecipitation (ChIP) assay and electrophoretic mobility shift assay (EMSA)

The ChIP assay was performed to analyze the in vivo binding of p53 to the RASSF1A promoter. Cells were treated with formaldehyde to cross-link the protein and DNA. Then 1 × 106 cells were washed and resuspended in lysis buffer (1% SDS, 10 mM EDTA, and 50 mM Tris-HCI, pH 8.1) with 1 mM PMSF, 1 μg/ml aprotinin, and 1 μg/ml pepstatin A. After sonication, the lysates were cleared by centrifugation and incubated in suspension with antip53 or control antibodies. Immunoprecipitated complexes were collected by protein A/G plus agarose and washed. The precipitated DNA was recovered by performing phenol/chloroform extraction and ethanol precipitation, and then it was subjected to PCR using primers that specifically amplify the RASSF1A core promoter. PCR products were analyzed by agarose gel electrophoresis and ethidium bromide staining.

The binding capacity of p53 to the RASSF1A promoter, in vitro, was analyzed by EMSA. Extracted nuclear proteins or recombinant proteins (rhp53 from Alpha Diagnostic International, San Antonio, TX, USA; rhDNMT1 from BPS Bioscience, San Diego, CA, USA) were incubated in a binding buffer (10 mM Tri-HCl, 50 mM NaCl, 0.5 mM EDTA, 10% glycerol, 1 mM DTT, and 7.5 mM MgCl) plus 0.1 μg of poly (dI-dC) carrier and a 32P-labeled probe spanning −417 to −390 of the RASSF1A core promoter that contains the second p53-binding element. A mutated probe with changes in key sequence of the p53-binding site served as control. Anti-p53, anti-DAXX, and anti-DNMT1 (and rabbit IgG control) antibodies were used to supershift the specific complexes of interest by pretreating the extract for 1 h at 4°C. These samples were electrophoresed on a 5% polyacrylamide gel, then dried, and an X-ray image was developed with an intensifier screen at −70°C.

Expression of glutathione S-transferase (GST)-tagged proteins

The expression and purification of GST-fused DAXX protein were performed as described previously (27). Briefly, after transfection of the GST-fused DAXX plasmid into BL21 Escherichia coli, the bacteria were incubated in Lennox broth medium, and then incubated with 0.1 mM isopropyl-β-d-thiogalactoside (IPTG) for 2 h and harvested at the end of that period. To purify GST-fused DAXX protein, the induced cells were lysed by sonication, after which the complex was isolated with glutathione sepharose beads (GE Healthcare, Milwaukee, WI, USA). The purity and correct expression of the GST-fused DAXX protein were confirmed by gel electrophoresis with Coomassie G250 staining, plus a Western blot assay using antibodies against GST or DAXX.

Immunoprecipitation and Western blot assays

All cells were lysed in a buffer composed of 50 mM Tris (pH 7.6), 150 mM NaCl, 1%Nonidet P-40, 10 mM sodium phosphate, 10 mM NaF, 1 mM sodium orthovanadate, 2 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 10 μg/ml pepstatin. After centrifugation, the clarified cell lysate was separated from the pellet of cell debris and incubated with 15 μl Protein G plus/Protein A-agarose (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and 1 μg of antibodies, overnight, at 4°C. For the Western blot, the resulting cell lysates or immunoprecipitates were resolved by SDS-PAGE, transferred to a nitrocellulose filter, and probed with specific antibodies. Protein expression was visualized with a chemiluminescent detection system (Pierce, Rockford, IL USA).

Pulse-chase assay

Protein turnover was tested for by a standard protein-synthesis-inhibitor cycloheximide (CHX) assay. Briefly, cells were treated with 50 μg/ml CHX for different times before lysis in the presence or absence of treatment, and then MDM2 and p53 expression were detected by Western blot analysis.


RT-PCR was performed with an Access RT-PCR Kit (Promega), according to the manufacturer's protocols. The primers used for measuring RASSF1A mRNA expression were 5′-GGCGTCGTGCGCAAAGGCC-3′ and 5′-ATGAAGCCTGTGTAAGAACCGTCCTTG-3′.

Apoptosis assays

An annexin-V assay (Oncogene, San Diego, CA, USA) was used to quantitate apoptotic cells. Briefly, cells with or without a given treatment were washed once with PBS and then stained with FITC-annexin-V for 30 min, according to the manufacturer's instructions. Stained cells were detected using the FACScan (Becton Dickinson, Franklin Lakes, NJ, USA) and analyzed using WinList software (Verity Software House, Topsham, ME, USA).


RASSF1A promoter methylation is associated with p53 status in ALL cell lines

We tested the possible association between RASSF1A expression and p53 status as well as DAXX expression in 15 ALL cell lines by Western blot. We found that the expression of RASSF1A was closely associated with p53 status but not DAXX expression. As shown in Fig. 1A, totally lost or reduced RASSF1A expression occurred in the WT-p53 ALL cells, whereas the p53-deficient lines expressed normal levels of RASSF1A. In contrast, the expression of DAXX as both high and low levels was detected in cells either with normal or reduced RASSF1A expression.

Figure 1.
Association of RASSF1A methylation with p53 status in ALL. A) Western blot results showed the expression RASSF1A and DAXX in 15 ALL cell lines with differing p53 status, as compared with the levels in normal peripheral blood lymphocytes (NPBL). B) Top ...

Because the expression of RASSF1A is mainly regulated by its promoter's methylation, we performed MSP to evaluate RASSF1A promoter methylation in these 15 ALL cell lines. Corresponding with the lost or reduced RASSF1A expression, we found that complete or partial methylation of RASSF1A promoter occurred only in the WT-p53 cell lines that were studied (Fig. 1B, top panel). All p53-deficient lines, including those with a mutation or a null phenotype, did not have RASSF1A promoter methylation, although a low percentage (1–5%) of CpG site methylations could be detected in the p53-mutated ALL lines (Supplemental Fig. S1). Similarly corresponding to the methylation status, the expression of RASSF1A mRNA was lost or reduced in WT-p53 cells, but not in p53-deficient cells (Fig. 1B, bottom panel). Figure 1C shows examples of detailed bisulfate sequencing results for RASSF1A promoter CpG site methylation in the 15 cell lines.

Identification of RASSF1A core promoter region and variations in promoter activity due to p53

Because RASSF1A promoter methylation occurs in WT-p53, but not in p53-deficient ALL cells, we evaluated whether p53 participates in the methylation of RASSF1A promoter. The transcription factor activity of WT-p53 either induces or represses gene expression by binding to the gene's promoter. First, we analyzed the RASSF1A gene promoter and then investigated whether its activity was regulated due to direct binding of p53. We found that the DNA sequence of the RASSF1A promoter contains 3 potential p53 response consensus sites, located between nucleotides −444 to −423, −414 to −392, and −337 to −307, which are all upstream of the known 16 CpG sites (ref. 5 and Fig. 2A). We found these putative p53 response elements associated with a canonical p53 consensus sequence that consists of 2 copies of a decamer motif, 5′RRRCWWGYYY-3′, separated by 0 to 13 bp of random nucleotides. Within the motif, r = G or A, W = T or A, Y = C or T (15).

Figure 2.
Analysis of RASSF1A promoter. A) Nucleotide sequence of RASSF1A promoter that contains 3 putative p53-binding sites (underlined). The 16 CpG-islands (green) are all located downstream of the p53-binding sites. The cDNA start site (red background) and ...

To find the core promoter region of the RASSF1A gene, we generated a series of deleted constructs of the RASSF1A promoter (Fig. 2B) and performed transfections prior to luciferase activity assays. Our results demonstrated that the core promoter region contained 150 bases and resided between −449 and −299. As shown in Fig. 2C, the construct pLuc-479 (−449 to +30) expressed the maximum luciferase activity, similar to that of the full-length promoter pLuc-659, while the construct pLuc-329 (−299 to +30) showed significantly reduced promoter activity when transfected into p53-null EU-4 cells. To further confirm that the sequence between −449 and −299 is critical for RASSF1A promoter activity, we cloned that 150-base sequence into the pGL3-basic vector, where it had luciferase activity comparable to that of the larger (−449 to +30) fragment.

Interestingly, all 3 p53 response elements resided within the core promoter region. Therefore, we tested for the effect of WT-p53 on RASSF1A promoter activity. We cotransfected the various RASSF1A promoter-luciferase constructs and a plasmid containing WT-p53 into EU-4 cells. As presented in Fig. 2C, we saw that although WT-p53 produced a significant decrease in RASSF1A-promoter activity (relative to the vehicle control plasmid), the mutant p53 expression plasmid did not inhibit RASSF1A promoter activity (Supplemental Fig S2). However, luciferase activity of the construct pLuc-150 (−449 to −299) with deletion of the downstream CpG sites was not inhibited by WT-p53, and furthermore, mutations of the p53 response elements, particularly the second and third p53 response elements, in the construct pLuc-479 significantly abrogated the effect of WT-p53 on repression of RASSF1A promoter activity. These suggested that repression of RASSF1A promoter activity by WT-p53 depends on the presence of the CpG islands and binding of p53 to the core promoter region.

The RASSF1A promoter is bound by p53, which also physically interacts with DAXX and DNMT1

We performed experiments to confirm that p53 is able to bind to the RASSF1A promoter. We performed the ChIP assay in WT-p53 EU-1 cells. After cross-linking and immunoprecipitation, PCR (using a pair of primers for the fragment that contains the 150-bp RASSF1A core promoter) showed a DNA band in the complex immunoprecipitated with anti-p53 antibody (Fig. 3A), but not in the sample immunoprecipitated with anti-DAXX, DNMT1 and MDM2 antibodies (Supplemental Fig. S3A).

Figure 3.
Binding of p53 to the RASSF1A promoter and the DAXX and DNMT1 proteins. A) Agarose gel electrophoresis shows the PCR results from each ChIP assay in EU-1 cells. Primer pair A, for amplifying the 150 bp RASSF1A core promoter; primer pair B, for amplifying ...

To further test whether p53 actually binds to the identified RASSF1A core promoter in cells, EMSA was carried out in EU-1 cells. The binding reactions were performed using WT or mutated oligonucleotides corresponding to the RASSF1A core promoter. As shown in Fig. 3B, one DNA-protein complex that was specific for the WT probe, which effectively competed with WT cold oligonucleotide. The DNA-protein complex was not observed using mutant probe. Further supershift assays, performed by incubation with antibodies, demonstrated that a large DNA-protein band was detected in the reaction with anti-p53 antibody, but was not detected in a sample treated with control antibody (IgG) or antibodies against DAXX and DNMT1 (Supplemental Fig. S3B), suggesting that p53, but not DAXX and DNMT1, binds directly to the RASSF1A promoter. In addition, we performed UV cross-linking and gel electrophoresis with a 32P-labeled RASSF1A promoter probe and recombinant human p53, DAXX and DNMT1 proteins. In fact, our results showed that the p53 but not DAXX and DNMT1 strongly and specifically bound to the RASSF1A promoter in vitro, without any other cellular elements present (Fig. 3C).

Previous studies demonstrated that p53, DAXX and DNMT1 interact with each other (16, 24, 28). By immunoprecipitation and Western blot assay, we confirmed that p53, DAXX, and DNMT1 did indeed physically bind with each other in EU-1 ALL cells (Fig. 3D).

p53 and DAXX can induce methylation of RASSF1A promoter

The results shown in Fig. 2C indicated that WT-p53 represses the activity of a RASSF1A promoter that is intact, i.e., the promoter contains CpG sites that are located downstream of the core promoter region. We investigated whether p53 represses RASSF1A promoter activity through regulation of RASSF1A promoter methylation. We performed MSP analysis to evaluate the effects of p53 on methylation of RASSF1A promoter in 4 p53-null cell lines EU-4, EU-5, EU-8, and EU-9 (without RASSF1A methylation). Transfection of p53 into these cells, either alone or in combination with DAXX, induced RASSF1A promoter methylation (Fig. 4A and Supplemental Fig. S4). Because p53, DAXX, and DNMT1 in WT-p53 EU-1 cells physically bind with each other, which may contribute to the induction of RASSF1A promoter methylation in these cells, we tested for the ability of the p53-DAXX-DNMT1 complex to induce this methylation in vitro. The p53-DAXX-DNMT1 complex was immunoprecipitated either by p53, DAXX, or DNMT1 and then incubated with recombinant RASSF1A promoter DNA. The complex induced RASSF1A promoter DNA methylation (Fig. 4B). In addition, we found that the recombinant p53 and DAXX protein together increased DNMT1 activity that methylated RASSF1A in vitro (Fig. 4C).

Figure 4.
Induction of RASSF1A methylation by p53 and DAXX. A) EU-4 cells were transfected with p53 and DAXX alone, or their combination. RASSF1A methylation in the transfected cells was evaluated by MSP analysis. B) Cell lysates from EU-1 were coimmunoprecipitated ...

We further investigated the regulation of RASSF1A promoter methylation by p53 and DAXX in dose-dependent and quantitative assays. After bisulfite sequencing analysis for methylation of RASSF1A promoter in EU-4 cells that were transfected with different doses of p53 and DAXX plasmids, we saw that an increase in p53 expression level did not enhance promoter methylation of RASSF1A, whereas an increase in DAXX, in the presence of p53, stimulated this methylation. This enhancement of RASSF1A promoter methylation by DAXX and p53 was experimentally diminished by the methylation inhibitor 5-aza-CdR (Fig. 4D, top panel). As is also shown in Fig. 4D (bottom panel), Western blots demonstrated there were corresponding changes of RASSF1A protein expression following the transfection of p53 and DAXX into EU-4 cells.

We also performed a gene transfection and reporter assay to evaluate the effect of p53 and DAXX, in different doses, on RASSF1A promoter activity. Consistent with results of RASSF1A promoter methylation, an increase in the amount of p53 did not enhance the repression of RASSF1A promoter activity; in contrast, increasing DAXX in the presence of p53 significantly enhanced the repression of RASSF1A promoter activity. Transfection of DAXX alone, without p53, did not (Supplemental Fig. S5A). This RASSF1A promoter methylation was not regulated by changes in levels of p53 was further confirmed by treatment of EU-3 cells (WT-p53 with partial RASSF1A methylation) with doxorubicin, a chemotherapeutic drug that induces WT-p53 expression due to DNA damage, which failed to enhance promoter methylation of RASSF1A (Fig. 4E and Supplemental Fig. S5B).

Inhibition of DAXX reactivates RASSF1A

The data shown in Fig. 1B indicate that WT-p53 ALL cell lines have RASSF1A promoter methylation, either completely or partially. All 3 RASSF1A promoter partially methylated WT-p53 cell lines had reduced DAXX expression. Therefore, we evaluated whether inhibition of DAXX expression by siRNA in a completely methylated RASSF1A cell line was able to induce demethylation and reactivation of RASSF1A. MSP results showed that treatment with DAXX-targeting siRNA induced demethylation of RASSF1A promoter in EU-1 cells, which have 100% CpG island methylation (Fig. 5A). Consistent with the demethylation of RASSF1A by DAXX siRNA, we found that the expression of RASSF1A was indeed induced (Fig. 5B). Quantitative sequencing analysis demonstrated that the induction of demethylation of RASSF1A promoter by DAXX siRNA was dose-dependent studied in 2 cell lines (Fig. 5C). In contrast, demethylation of RASSF1A promoter and endogenous RASSF1A expression did not occur in the same cells treated with p53 siRNA (Fig. 5B, C).

Figure 5.
Demethylation of RASSF1A by inhibition of DAXX. A, B) MSP analysis of RASSF1A unmethylation (A) and Western blot for protein expression (B) in EU-1 cells that were treated with DAXX and p53 siRNA or control siRNA. C) Levels of RASSF1A methylation in EU-1 ...

We recently reported that berberine, a natural product derived from a plant used in Chinese herbal medicine, is able to inhibit DAXX (29). Herein, we tested to see whether berberine induces RASSF1A promoter demethylation due to inhibition of DAXX. MSP analysis showed that berberine treatment of EU-1 cells did induce demethylation of RASSF1A promoter (Fig. 5D), resulting in reactivation of RASSF1A (Fig. 5E). Reactivation of RASSF1A by berberine was most likely a consequence of inhibition of DAXX, as no similar changes of RASSF1A were detected in doxorubicin-treated EU-1 cells, nor was DAXX inhibited by doxorubicin treatment (Fig. 5D, E).

Inhibition of RASSF1A gene promoter methylation down-regulates MDM2, induces p53, and sensitizes ALL cells to apoptosis

Previous studies demonstrated that RASSF1A plays an important role in disrupting the DAXX-HAUSP-MDM2 interactions, resulting in MDM2 self-degradation and p53 activation (3). We have indeed found that ALL cells with RASSF1A promoter methylation (inactivation) commonly overexpressed MDM2 (Supplemental Fig. S6). To investigate whether the overexpressed MDM2 can be reduced by inhibition of RASSF1A promoter methylation, we treated EU-1 cells (having RASSF1A methylation) with berberine to reactivate RASSF1A. As shown in Fig. 6A, berberine treatment induced RASSF1A expression, which then bound to MDM2 and the DAXX protein, resulting in dissociation of MDM2 from DAXX and HAUSP. Results from pulse-chase analyses showed that reactivation of RASSF1A by berberine increased the turnover of MDM2. Thus, while the half-life of MDM2 in untreated cells was larger than 90 min, berberine treatment reduced the MDM2 half-life to less than 30 min. In contrast, the half-life of p53 in untreated cells was less than 30 min, whereas the half-life of p53 in berberine-treated cells was remarkably increased (Fig. 6B), resulting in activation of p53, as expression of the p53 transcriptional targets p21 and PUMA was increased in berberine-treated cells (Fig. 6C).

Figure 6.
Regulation of MDM2 and p53 as well as sensitivity to apoptosis by inhibition of DAXX-mediated RASSF1A demethylation. A) EU-1 cells were treated either with or without berberine, and cell lysates were IP with antibodies, as indicated. Protein expression ...

As we reported previously (27), EU-1 cells having high-level MDM2 expression are resistant to doxorubicin. In this study we tested to see whether this sensitivity could be altered by combining doxorubicin treatment with berberine or DAXX siRNA, to affect demethylation of RASSF1A promoter. As seen in Fig. 6D, a flow-cytometry apoptosis assay revealed there was a notably increased percentage of apoptosis of EU-1 cells treated with combined doxorubicin and DAXX siRNA, as compared to those receiving doxorubicin plus a control siRNA. Furthermore, we sought to evaluate whether berberine provided a possible synergistic effect in doxorubicin-induced apoptosis by treating EU-1 cells with both drugs, which when given alone induce 20–50% apoptosis after a 24 h treatment. When the same doses were given to EU-1 cells in combination, we found that >80% of cells underwent apoptosis after 24 h (Fig. 6E), which indeed suggested that berberine had synergy or a chemosensitization effect to doxorubicin-treated ALL cells having the WT-p53 and a MDM2-overexpressing phenotype.


The most commonly inactivated tumor suppressor genes in human cancer are p53 and RASSF1A. Inactivation of p53 occurs either through genetic mutations (30) or via degradation by its inhibitor MDM2 (31), whereas loss of RASSF1A function is most frequently via epigenetic silencing (1). A well-studied transcription factor that either positively or negatively regulates its target genes, p53 induces cell-cycle arrest and apoptosis. RASSF1A can also induce cell-cycle arrest and apoptosis; however, the mechanisms behind its activity are not completely understood. Emerging evidence suggests that RASSF1A acts as a scaffolding protein to assemble and modulate multiple effector protein complexes (1). Whether the tumor suppressor p53 and RASSF1A are either linked or mutually regulated in their contribution to cancer control or tumorigenesis, respectively, has never been investigated, with the exception of a previous report that RASSF1A can modulate the MDM2-DAXX-HAUSP complex to indirectly induce p53 activation (3).

In the present study, we reveal that p53 negatively regulated RASSF1A, by directly binding to the RASSF1A gene promoter to effectuate its methylation and silencing. Silencing of RASSF1A may subsequently result in release of the disruption of the MDM2-DAXX-HAUSP complex, as activated RASSF1A disrupts the MDM2-DAXX-HAUSP interactions. Because the MDM2-DAXX-HAUSP interactions are necessary to maintain MDM2 stabilization and p53 degradation (20), the p53-mediated RASSF1A inactivation leads to MDM2 stabilization and p53 degradation. Therefore, we propose a p53-RASSF1A-MDM2 negative feedback loop exists in WT-p53 cancer cells, in which p53 inhibits RASSF1A by inducing its promoter methylation, RASSF1A inhibits MDM2 by modulation of MDM2 self-ubiquitination, and MDM2 inhibits p53 by targeted ubiquitination (Fig. 7). Due to this negative feedback regulation, RASSF1A became methylated, MDM2 became overexpressed, and p53 became inhibited, which all results in cell proliferation, tumor promotion, and possible development of chemoresistance in these WT-p53 cancers. In fact, previous studies have reported that WT-p53 ALL cells with MDM2 overexpression were doxorubicin resistant (32).

Figure 7.
Model depicting the p53-RASSF1A-MDM2 negative feedback loop and the central role of DAXX in regulation of this p53-RASSF1A-MDM2 feedback loop.

In our cell model, the methylation of RASSF1A promoter was not detected in ALL cells that have p53 mutations or lost WT-p53 expression. Because RASSF1A is normally expressed in these p53-deficient cancer cells, it induces MDM2 protein instability, which could explain why almost all p53-deficient ALL cell lines express no or very low levels of MDM2 (ref. 25 and Supplemental Fig. S5). In fact, our study results demonstrated that the half-life of MDM2 in these p53-deficient ALL cells was much shorter than in WT-p53 ALL cells (Supplemental Fig. S6). The rapid turnover of MDM2 in p53-deficient cancer cells is indeed due to expression of RASSF1A, because re-expression of WT-p53 in the p53-null cells induced methylation and silencing of RASSF1A gene expression, which resulted in stabilization and increased expression of MDM2. In addition to the induction of MDM2 self-ubiquitination and degradation, RASSF1A has been reported to induce cell-cycle arrest and senescence in human cancer cells by p53-independent regulation of p21/WAF1 (33). Further studies demonstrated that RASSF1A can induce cell-cycle arrest by regulating the cyclin D1 and JNK pathways (34, 35), plus it promotes apoptosis by interacting with MST1 and CNK1 (36, 37). RASSF1A-mediated cell-cycle arrest and apoptosis are likely an important supplement to the loss of p53 function, in p53-deficient cancer cells, inhibiting tumor cell growth and disease progression. As observed clinically, some p53-mutated cancer patients have had a favorable outcome in response to anticancer treatment (38, 39).

An interesting finding of the present study is the critical role of DAXX in the regulation of RASSF1A promoter methylation. We discovered that DAXX represses the p53-targeting promoter RASSF1A in a similar way as it represses the RelB target promoter (24), through DNA methylation. Importantly, DAXX plays a critical role in the regulation of RASSF1A promoter methylation, as fluctuations of DAXX, but not p53, levels do significantly affect the degree of methylation and demethylation of RASSF1A promoter. Also, DAXX-mediated RASSF1A promoter methylation absolutely required the presence of WT-p53. As was observed in p53-null ALL cells, there was no RASSF1A promoter methylation detected, regardless of whether the cells in question were expressing or not expressing DAXX. In contrast, the degree of RASSF1A promoter methylation was found to be closely associated with DAXX expression levels in WT-p53 ALL cells. All 4 of the cell lines with completely methylated RASSF1A expressed high levels of DAXX, whereas all 3 of the lines where RASSF1A was partially methylated, expressed low levels of DAXX. This was confirmed by experiments in which enforced DAXX expression induced methylation of RASSF1A promoter, whereas the inhibition of DAXX by siRNA induced RASSF1A promoter demethylation in a dose-dependent manner.

Taken together with previous findings that DAXX regulates MDM2 and p53, we proposed a model that could account for the central role that DAXX plays in the regulation of the p53-RASSF1A-MDM2 negative feedback loop, contributing to tumorigenesis of WT-p53 cancer cells. As is illustrated in Fig. 7, we found that DAXX interacts with p53 in the presence of DNMT1 to induce RASSF1A promoter methylation and subsequent inactivation. Because DAXX interacts with the RASSF1A protein in the presence of HAUSP, it induces MDM2 degradation. The inactivation of RASSF1A would result in stabilization of MDM2 after DAXX interacts with MDM2 and HAUSP, which would subsequently induce p53 ubiquitination and degradation.

According to this regulation model, one can envision what would happen when DAXX is removed from cancer cells. In fact, previous studies have shown that loss of DAXX results in extensive apoptosis in early development (40), and silencing of DAXX expression by RNAi in cancer cells ends up sensitizing them to multiple apoptosis pathways (41, 42). More importantly, a recent study demonstrated that pancreatic cancer patients whose tumor cells lost DAXX expression due to gene mutation had a prolonged survival relative to those patients whose tumor cells lacked DAXX gene mutation (43). In addition, our recent studies have shown that inhibition of DAXX by berberine, induces strong apoptosis of WT-p53/MDM2-overexpressing cancer cells (29). Observations from our present study further indicated that inhibition of DAXX by either siRNA or berberine treatment of WT-p53 ALL cells sensitized these cells to the chemotherapeutic drug doxorubicin. We believe that disruption of the p53-RASSF1A-MDM2 regulation loop by removal of DAXX, which led to RASSF1A reactivation, MDM2 degradation, and beneficial p53 tumor suppressor activation, is a major mechanism for the induction of apoptosis of an important subset of leukemia cells, and perhaps other cancers as well.

Supplementary Material

Supplemental Data:


The authors acknowledge support from U.S. National Institutes of Health/National Cancer Institute grants R01 CA123490 and R01 CA143107 (M.Z.), the Leukemia and Lymphoma Society 6033-08 (M.Z.), and CURE Childhood Cancer (M.Z. and L.G.).

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

acute lymphoblastic leukemia
chromatin immunoprecipitation
death-associated protein 6
DNA methyltransferase 1
electrophoretic mobility shift assay
glutathione S-transferase
murine double minute 2
methylation-specific PCR
Ras-association domain family 1 isoform A


1. Donninger H., Vos M. D., Clark G. J. (2007) The RASSF1A tumor suppressor. J. Cell Sci. 120, 3163–3172 [PubMed]
2. Van der Weyden L., Adams D. J. (2007) The Ras-association domain family (RASSF) members and their role in human tumourigenesis. Biochim. Biophys. Acta 1776, 58–85 [PMC free article] [PubMed]
3. Song M. S., Song S. J., Kim S. Y., Oh H. J., Lim D. S. (2008) The tumour suppressor RASSF1A promotes MDM2 self-ubiquitination by disrupting the MDM2-DAXX-HAUSP complex. EMBO J. 27, 1863–1874 [PMC free article] [PubMed]
4. Richter A. M., Pfeifer G. P., Dammann R. H. (2009) The RASSF proteins in cancer: from epigenetic silencing to functional characterization. Biochim. Biophys. Acta 1796, 114–128 [PubMed]
5. Dammann R., Li C., Yoon J. H., Chin P. L., Bates S., Pfeifer G. P. (2000) Epigenetic inactivation of a RAS association domain family protein from the lung tumour suppressor locus 3p21.3. Nat. Genet. 25, 315–319 [PubMed]
6. Pfeifer G. P., Yoon J. H., Liu L., Tommasi S., Wilczynski S. P., Dammann R. (2002) Methylation of the RASSF1A gene in human cancers. Biol. Chem. 383, 907–914 [PubMed]
7. Levine A. J. (1997) p53, the cellular gatekeeper for growth and division. Cell 88, 323–331 [PubMed]
8. Miyashita T., Reed J. C. (1995) Tumor suppressor p53 is a direct transcriptional activator of the human bax gene. Cell 80, 293–299 [PubMed]
9. Takimoto R., El-Deiry W. S. (2000) Wild-type p53 transactivates the KILLER/DR5 gene through an intronic sequence-specific DNA-binding site. Oncogene 19, 1735–1743 [PubMed]
10. Owen-Schaub L. B., Zhang W., Cusack J. C., Angelo L. S., Santee S. M., Fujiwara T., Roth J. A., Deisseroth A. B., Zhang W. W., Kruzel E. (1995) Wild-type human p53 and a temperature-sensitive mutant induce Fas/APO-1 expression. Mol. Cell. Biol. 15, 3032–3040 [PMC free article] [PubMed]
11. Zastawny R. L., Salvino R., Chen J., Benchimol S., Ling V. (1993) The core promoter region of the P-glycoprotein gene is sufficient to confer differential responsiveness to wild-type and mutant p53. Oncogene 8, 1529–1535 [PubMed]
12. Miyashita T., Harigai M., Hanada M., Reed J. C. (1994) Identification of a p53-dependent negative response element in the bcl-2 gene. Cancer Res. 54, 3131–3135 [PubMed]
13. Hoffman W. H., Biade S., Zilfou J. T., Chen J., Murphy M. (2002) Transcriptional repression of the anti-apoptotic survivin gene by wild type p53. J. Biol. Chem. 277, 3247–3257 [PubMed]
14. Mirza A., Wu Q., Wang L., McClanahan T., Bishop W. R., Gheyas F., Ding W., Hutchins B., Hockenberry T., Kirschmeier P., Greene J. R., Liu S. (2003) Global transcriptional program of p53 target genes during the process of apoptosis and cell cycle progression. Oncogene 22, 3645–3654 [PubMed]
15. Kern S. E., Kinzler K. W., Bruskin A., Jarosz D., Friedman P., Prives C., Vogelstein B. (1991) Identification of p53 as a sequence-specific DNA-binding protein. Science 252, 1708–1711 [PubMed]
16. Esteve P. O., Chin H. G., Pradhan S. (2005) Human maintenance DNA (cytosine-5)-methyltransferase and p53 modulate expression of p53-repressed promoters. Proc. Natl. Acad. Sci. U. S. A. 102, 1000–1005 [PMC free article] [PubMed]
17. Yang X., Khosravi-Far R., Chang H. Y., Baltimore D. (1997) Daxx, a novel Fas-binding protein that activates JNK and apoptosis. Cell 89, 1067–1076 [PMC free article] [PubMed]
18. Torii S., Egan D. A., Evans R. A., Reed J. C. (1999) Human Daxx regulates Fas-induced apoptosis from nuclear PML oncogenic domains (PODs). EMBO J. 18, 6037–6049 [PMC free article] [PubMed]
19. Chang H. Y., Nishitoh H., Yang X., Ichijo H., Baltimore D. (1998) Activation of apoptosis signal-regulating kinase 1 (ASK1) by the adapter protein Daxx. Science 281, 1860–1863 [PubMed]
20. Tang J., Qu L. K., Zhang J., Wang W., Michaelson J. S., Degenhardt Y. Y., El-Deiry W. S., Yang X. (2006) Critical role for Daxx in regulating Mdm2. Nat. Cell Biol. 8, 855–862 [PubMed]
21. Huang Y. S., Shih H. M. (2009) Daxx positively modulates beta-catenin/TCF4-mediated transcriptional potential. Biochem. Biophys. Res. Commun. 386, 762–768 [PubMed]
22. Croxton R., Puto L. A., de Belle I., Thomas M., Torii S., Hanaii F., Cuddy M., Reed J. C. (2006) Daxx represses expression of a subset of antiapoptotic genes regulated by nuclear factor-kappaB. Cancer Res. 66, 9026–9035 [PubMed]
23. Wethkamp N., Klempnauer K. H. (2009) Daxx is a transcriptional repressor of CCAAT/enhancer-binding protein beta. J. Biol. Chem. 284, 28783–28794 [PMC free article] [PubMed]
24. Puto L. A., Reed J. C. (2008) Daxx represses RelB target promoters via DNA methyltransferase recruitment and DNA hypermethylation. Genes Dev. 22, 998–1010 [PMC free article] [PubMed]
25. Zhou M., Yeager A. M., Smith S. D., Findley H. W. (1995) Overexpression of the MDM2 gene by childhood acute lymphoblastic leukemia cells expressing the wild-type p53 gene. Blood 85, 1608–1614 [PubMed]
26. Herman J. G., Graff J. R., Myohanen S., Nelkin B. D., Baylin S. B. (1996) Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc. Natl. Acad. Sci. U. S. A. 93, 9821–9826 [PMC free article] [PubMed]
27. Gu L., Zhu N., Zhang H., Durden D. L., Feng Y., Zhou M. (2009) Regulation of XIAP translation and induction by MDM2 following irradiation. Cancer Cell 15, 363–375 [PMC free article] [PubMed]
28. Gostissa M., Morelli M., Mantovani F., Guida E., Piazza S., Collavin L., Brancolini C., Schneider C., Del Sal G. (2004) The transcriptional repressor hDaxx potentiates p53-dependent apoptosis. J. Biol. Chem. 279, 48013–48023 [PubMed]
29. Zhang X., Gu L., Li J., Shah N., He J., Yang L., Hu Q., Zhou M. (2010) Degradation of MDM2 by the interaction between berberine and DAXX leads to potent apoptosis in MDM2-overexpressing cancer cells. Cancer Res. 70, 9895–9904 [PMC free article] [PubMed]
30. Hollstein M., Sidransky D., Vogelstein B., Harris C. C. (1991) p53 mutations in human cancers. Science 253, 49–53 [PubMed]
31. Haupt Y., Maya R., Kazaz A., Oren M. (1997) Mdm2 promotes the rapid degradation of p53. Nature 387, 296–299 [PubMed]
32. Zhou M., Gu L., Abshire T. C., Homans A., Billett A. L., Yeager A. M., Findley H. W. (2000) Incidence and prognostic significance of MDM2 oncoprotein overexpression in relapsed childhood acute lymphoblastic leukemia. Leukemia 14, 61–67 [PubMed]
33. Thaler S., Hahnel P. S., Schad A., Dammann R., Schuler M. (2009) RASSF1A mediates p21Cip1/Waf1-dependent cell cycle arrest and senescence through modulation of the Raf-MEK-ERK pathway and inhibition of Akt. Cancer Res. 69, 1748–1757 [PubMed]
34. Shivakumar L., Minna J., Sakamaki T., Pestell R., White M. A. (2002) The RASSF1A tumor suppressor blocks cell cycle progression and inhibits cyclin D1 accumulation. Mol. Cell. Biol. 22, 4309–4318 [PMC free article] [PubMed]
35. Whang Y. M., Kim Y. H., Kim J. S., Yoo Y. D. (2005) RASSF1A suppresses the c-Jun-NH2-kinase pathway and inhibits cell cycle progression. Cancer Res. 65, 3682–3690 [PubMed]
36. Oh H. J., Lee K. K., Song S. J., Jin M. S., Song M. S., Lee J. H., Im C. R., Lee J. O., Yonehara S., Lim D. S. (2006) Role of the tumor suppressor RASSF1A in Mst1-mediated apoptosis. Cancer Res. 66, 2562–2569 [PubMed]
37. Rabizadeh S., Xavier R. J., Ishiguro K., Bernabeortiz J., Lopez-Ilasaca M., Khokhlatchev A., Mollahan P., Pfeifer G. P., Avruch J., Seed B. (2004) The scaffold protein CNK1 interacts with the tumor suppressor RASSF1A and augments RASSF1A-induced cell death. J. Biol. Chem. 279, 29247–29254 [PubMed]
38. Preudhomme C., Dervite I., Wattel E., Vanrumbeke M., Flactif M., Lai J. L., Hecquet B., Coppin M. C., Nelken B., Gosselin B. (1995) Clinical significance of p53 mutations in newly diagnosed Burkitt's lymphoma and acute lymphoblastic leukemia: a report of 48 cases. J. Clin. Oncol. 13, 812–820 [PubMed]
39. King T. C., Akerley W., Fan A. C., Moore T., Mangray S., Hsiu Chen M., Safran H. (2000) p53 mutations do not predict response to paclitaxel in metastatic nonsmall cell lung carcinoma. Cancer 89, 769–773 [PubMed]
40. Michaelson J. S., Bader D., Kuo F., Kozak C., Leder P. (1999) Loss of Daxx, a promiscuously interacting protein, results in extensive apoptosis in early mouse development. Genes Dev. 13, 1918–1923 [PMC free article] [PubMed]
41. Michaelson J. S., Leder P. (2003) RNAi reveals anti-apoptotic and transcriptionally repressive activities of DAXX. J. Cell Sci. 116, 345–352 [PubMed]
42. Chen L. Y., Chen J. D. (2003) Daxx silencing sensitizes cells to multiple apoptotic pathways. Mol. Cell. Biol. 23, 7108–7121 [PMC free article] [PubMed]
43. Jiao Y., Shi C., Edil B. H., de Wilde R. F., Klimstra D. S., Maitra A., Schulick R. D., Tang L. H., Wolfgang C. L., Choti M. A., Velculescu V. E., Diaz L. A., Jr., Vogelstein B., Kinzler K. W., Hruban R. H., Papadopoulos N. (2011) DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors. Science 331, 1199–1203 [PMC free article] [PubMed]

Articles from The FASEB Journal are provided here courtesy of The Federation of American Societies for Experimental Biology
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...