Demethylase Activity Is Directed by Histone Acetylation*

Received for publication, May 1, 2001, and in revised form, August 2, 2001Published, JBC Papers in Press, August 27, 2001, DOI 10.1074/jbc.M103921200

Nadia Cervoni and Moshe Szyf‡

From the Department of Pharmacology and Therapeutics, McGill University, Montreal, PQ H3G 1Y6, Canada

Mammalian genomes are compartmentalized intodense inactive chromatin that is hypermethylated andactive open chromatin that is hypomethylated. It is gen-erally accepted that this bimodal pattern of methylationis established during development and is then faithfullyinherited through subsequent cell divisions by a main-tenance DNA methyltransferase (DNMT1). The patternof methylation is believed to direct local histone acety-lation states. In contrast to this well accepted consen-sus, we show here using a transient transfection modelthat an active demethylase is involved in shaping pat-terns of methylation in somatic cells. Demethylase ac-tivity is directed by the state of histone acetylation, andtherefore, the resulting methylation pattern is deter-mined by local histone acetylation states contrary to theaccepted model. Our data support a new model suggest-ing that the pattern of methylation is maintained by adynamic balance of methylation and demethylation ac-tivities and the local state of histone acetylation. Thisprovides a simple mechanism for explaining why activegenes are not methylated.

A hallmark of mammalian genomes is the compartmental-ization of the genome into dense inactive chromatin that ishypermethylated and active, open chromatin that is hypo-methylated (1). However, the mechanism responsible for estab-lishing this tight relationship remains unclear. The acceptedmodel is that a sequence of methylation and demethylationevents fashion the methylation pattern during development,but it is then faithfully inherited by a semiconservative DNAmethyltransferase, DNMT11 (2). Methylation of newly synthe-sized DNA is exclusively determined by the state of methyla-tion of the parental strand. The pattern of methylation istherefore believed to be fixed in somatic cells. Based on theassumption that DNA conserves its pattern of methylation insomatic cells, numerous experiments used transiently trans-fected methylated DNA to study the effects of DNA methyla-tion on gene expression. In many of these studies the state ofmethylation of these ectopically methylated genes followingtransfection was not determined assuming that the pattern of

methylation of the transfected gene did not change. MethylatedDNA is associated with methyl CpG binding proteins, such asMeCP2, which reside in a complex with histone deacetylaseactivity (3). The current model is therefore that the pattern ofmethylation dictates the state of histone acetylation and chro-matin configuration (4).

This attractive model explains the compartmentalization ofthe genome and its inheritance in somatic cells, but it can notexplain how genes are demethylated upon their activation. Analternative and opposite interpretation of the tight correlationbetween histone acetylation is that active chromatin causesdemethylation of associated sequences. Such a model can ex-plain why active genes are not methylated and how their un-methylated state is maintained through cell division.

This hypothesis that active chromatin can cause demethyla-tion is supported by previous data. Treatment of mammaliancells with general histone deacetylase inhibitors can causeglobal demethylation of human Epstein-Barr virus producercell lines’ genomes (5). Similarly in Neurospora the deacetylaseinhibitor TSA causes selective demethylation (6). Recent datasupport the claim that inhibition of histone deacetylation cancause selective demethylation of some genes such as the IgfIIreceptor (7) but not certain tumor suppressor genes (8).Whereas this data shows that activation of genes by histonedeacetylase inhibitors can lead toward loss of methylation, themechanism is unclear. This loss of methylation might either becaused by inhibition of the maintenance DNA methyltrans-ferase during replication, by site-specific proteins, by triggeringsite-specific repair activity, or by active site-specific or generaldemethylation.

In this report we use a transient transfection approach todirectly measure demethylation activity in human cells (HEK293) and show that the state of methylation of DNA is not fixedin somatic cells but is dynamically modulated to correlate withthe state of gene activity. This is accomplished by active de-methylase activity that is directed by histone acetylation.These data provide a simple mechanism for explaining howactive genes are demethylated and maintained in their un-methylated state.

MATERIALS AND METHODS

Cell Culture and CAT Assays—HEK 293 cells were plated at adensity of 8 � 104/well in a six-well tissue culture dish and transientlytransfected with 80 ng of plasmid DNA using the calcium phosphateprecipitation method as described previously (9). Transfections wererepeated a minimum of three times using different cultures of HEK 293cells. CAT assays were performed in triplicate as described previously(9).

Cell Culture and Flow Cytometry—HEK 293 cells were maintainedas a monolayer in Dulbecco’s modified Eagle’s medium (Life Technolo-gies, Inc.) containing 10% calf serum (Colorado Serum Co). To serumstarve the cells, confluent HEK 293 cells were cultured in a mediumcontaining 0.5% fetal calf serum for 72-h post-transfection. To deter-mine the percentage of cells at different stages of the cell cycle, cellswere stained with propidium iodide and the DNA content was meas-ured by flow cytometry.

In Vitro Methylation of Substrates—pMetCAT�, SV40CAT, pCMV-

* This work was supported by a grant from the Canadian Institute ofHealth Research (to M. S.). The costs of publication of this article weredefrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

‡ To whom correspondence should be addressed: Dept. of Pharmacol-ogy and Therapeutics, McGill University, 3655 Drummond St., Mon-treal PQ H3G 1Y6, Canada. Tel.: 514-398-7107; Fax: 514-398-6690;E-mail: mszyf@pharma.mcgill.ca.

1 The abbreviations used are: DNMT1, DNA methyltransferase; GFP,green fluorescence protein; EGFP, enhanced GFP; TSA, trichostatin A;dMTase, DNA demethylase; CAT, chloramphenicol acetyl transferase;MNase, micrococcal nuclease; CMV, cytomegalovirus; PCR, polymerasechain reaction; bp, base pair(s); kb, kilobase(s); CHIP, chromatinimmunoprecipitation.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 276, No. 44, Issue of November 2, pp. 40778–40787, 2001© 2001 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org40778

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

GFP, and GFP plasmids were methylated in vitro by incubating 10 �gof plasmid DNA with 20 units of SssI CpG DNA methyltransferase (10)(New England BioLabs Inc.) in a buffer recommended by the manufac-turer containing 160 �M S-adenosylmethionine, at 37 °C for 2 h. Afterrepeating this procedure three times, full protection from HpaII diges-tion was observed.

Bisulfite Mapping—Bisulfite mapping was performed as describedpreviously with minor modifications (11). 5 �g of sodium bisulfite-treated DNA samples was subjected to PCR amplification using thefirst set of primers described below. PCR products were used as tem-plates for subsequent PCR reactions utilizing nested primers. The PCRproducts of the second reaction were then subcloned using the Invitro-gen TA cloning kit (we followed the manufacturer’s protocol), and theclones were sequenced using the T7 sequencing kit (Amersham Phar-macia Biotech) (we followed the manufacturer’s protocol, procedure C). Theprimers used for the bacterial DNA chloramphenicol acetyltransferase(CAT) genomic region (GenBank 228 accession number U65077) were:CAT5�1, 5�-ttgtttaatgtatttataattacat-3�; CAT5�(nested), 5�-taaagaaaaataag-tataagtttta-3�; CAT3�1, 5�-ctcacccaaaaattaactaaaa-3�; CAT3�(nested), 5�-tt-taaaaaaataaaccaaattttca-3�. The primers used for the enhanced green fluo-rescence protein (pEGFP-1) (CLONTECH) (GenBankTM accession numberU55761) were: GFP5�1, 5�-gttattatggtgagtaaggg-3�; GFP5�(nested), 5�-ggggtggtgtttattttgg-3�; GFP3�1, 5�-tataactattataattatactcca-3�; GFP3�(nested), 5�-cttataccccaaaatattacc-3�.

Chromatin Immunoprecipitation Assay—CHIP assays (12) were per-formed by following the Upstate Biotechnology Chromatin Immunopre-cipitation (CHIP) assay kit protocol (catalog no. 17-295). HEK 293 cellswere transfected with 80 ng of in vitro methylated pMetCAT�,SV40CAT, and pCMV-GFP plasmids, using the calcium phosphatemethod (see above). A final concentration of 0.3 �M TSA was added ornot added to fresh medium 24 h after transfection. Formaldehyde wasadded to the culture media at a final concentration of 1%, 96-h post-transfection, and incubated at 37 °C for 10 min, and chromatin wasimmunoprecipitated using an anti-acetylated histone H3 antibody (Up-state Biotechnology) as recommended by the manufacturer. One-tenthof the lysate was kept to quantitate the amount of DNA present indifferent samples before immunoprecipitation. DNA purified from boththe immunoprecipitated and pre-immune (pre) samples was dilutedfrom 1:10 to 1:10,000 serially and was subjected to PCR amplificationusing the following primers for the CAT and GFP genes: CAT 5�,5�-cactggatataccaccgttga-3�; CAT 3�, 5�-aaaccctttagggaaataggc-3�; GFP5�, 5�-caagggcgaggagctgtt-3�; GFP 3�, 5�-cggccatgatatagacgttg-3�.

MNase Assay—Micrococcal nuclease (MNase) (Amersham Pharma-cia Biotech) digestions (0–60 units) were performed on purified nuclei(13) for 20 min at 20 °C in 50 mM Tris-HCl, pH 8.0, 0.05 mM CaCl2, and20% glycerol. The MNase reaction was terminated and DNA was ex-tracted using the DNeasy Tissue Kit (Qiagen).

Western Blot Analysis—Total cell extracts were prepared usingstandard protocols and resolved on a SDS-polyacrylamide gel electro-phoresis (7.5% for dMTase and 12.5% for GFP protein). After transfer-ring to nitrocellulose membrane and blocking the nonspecific bindingwith 5% milk, GFP protein was detected using rabbit polyclonal IgG(Santa Cruz, sc-8334) at 1:500 dilution, followed by peroxidase-conju-gated anti-rabbit IgG (Sigma) at 1:5000. Transfected dMTase proteinwas detected using Anti-Xpress mouse monoclonal IgG (Invitrogen,R910-25) at 1:5000 dilution, followed by peroxidase-conjugated anti-mouse IgG (Jackson ImmunoResearch) at 1:20,000 dilution and en-hanced chemiluminescence detection kit (Amersham PharmaciaBiotech).

RESULTS

Ectopically Methylated Sequence of DNA Is Actively De-methylated in Mammalian Cells When It Resides Downstreamfrom a Strong Promoter—If the methylation pattern of genes insomatic cells is static as predicted by the semiconservativemodel of methylation, an ectopically methylated gene will re-main methylated in somatic cells. If the methylation patternand its correlation with gene activity are a dynamic balance ofmethylation and demethylation, then the cell should recognizeectopically methylated DNA and fashion its pattern of methyl-ation in accordance with its state of activity. We therefore firsttested the hypothesis that somatic cells can recognize an ec-topically methylated gene and demethylate it when it residesdownstream from a strong promoter but not when it is founddownstream from an inactive or weak promoter. We chose to

compare the state of methylation of an identical reporter genesequence to exclude the possibility that any observed differ-ences in demethylation are a result of a sequence-specific prop-erty of the demethylated region. The pMetCAT� fusion con-struct expressing the bacterial (chloramphenicol acetyltransferase) CAT reporter gene under the direction of 2 kb ofthe DNA dnmt1 regulatory region was previously shown to behighly active in P19 cells (9) but is inactive in HEK 293 cells(Fig. 1A). In comparison, the CAT reporter gene, which iscontrolled by the SV40 promoter, is highly active in HEK 293cells (Fig. 1A). Both fusion constructs were fully methylated invitro with SssI methylase and transfected into HEK 293 cellsusing the calcium phosphate protocol (9). After 48 h, DNA fromthese cells was isolated and the methylation state of CpGsresiding in a 265-bp region within the CAT reporter gene wasdetermined by bisulfite mapping. 45% of CpGs became de-methylated within the CAT region under the control of theSV40 promoter, whereas almost no demethylation (3.5%) wasdetected within the CAT region downstream from the dnmt1regulatory region (Fig. 1B). This experiment demonstrates thatidentical sequences can be differentially demethylated in HEK293 cells depending on their state of expression.

We next determined whether this demethylation is unique tothe CAT reporter gene and whether it occurs exclusively whena reporter gene is found downstream from the SV40 promoter.Therefore, we performed the same assay using an in vitromethylated GFP (green fluorescence protein) reporter gene,

FIG. 1. An ectopically methylated reporter gene is demethyl-ated when an active promoter directs it. A, 5 �g of SV40CAT andpMetCAT� plasmids were transfected in triplicate into HEK293 cellsthen harvested after 48 h, and CAT activity was determined as de-scribed above. B, DNA was extracted from the SV40CAT and pMet-CAT� transfectants and treated with sodium bisulfite, and a regionwithin the CAT sequence (indicated in the physical map) was amplifiedby PCR, subcloned, and sequenced as previously described (27). Boththe outside and nested primers used for amplifying bisulfited DNA areindicated as hatched arrows. The result of this analysis is presented.Each line indicates one clone, filled circles are methylated, and emptycircles are unmethylated.

Histone Acetylation and Demethylation 40779

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

which is positioned under the control of a different strong viralpromoter from CMV (Fig. 2A). Sodium bisulfite mapping of aregion within the GFP gene, downstream from the CMV pro-moter, revealed that the GFP reporter became demethylatedalbeit to a lesser extent than SV40CAT (Fig. 2A, �TSA). Thisexperiment confirms that HEK 293 cells bear a demethylationactivity that demethylates ectopically methylated sequencesresiding downstream from strong promoters. Most or all theCpGs residing in the entire region of 250 bp of GFP are de-methylated in some plasmid molecules suggesting that dem-ethylation is regional and not site-specific.

Demethylation Is Directed by a Local State of Acetylation—Multiple mechanisms are responsible for modifying and acti-vating chromatin, including binding of activators, histoneacetylation and deacetylation, ATP-dependent remodeling andformation of a pre-initiation complex (14). To determinewhether regional demethylation requires histone acetylation,we tested whether the histone deacetylase inhibitor, trichosta-tin A (TSA) increases the demethylation of ectopically methyl-ated reporter genes, (Fig. 2, A and B). Bisulfite mapping of theGFP reporter in transfectants treated with 0.3 �M TSA for72 h, showed almost complete demethylation compared withlimited demethylation in the absence of TSA (Fig. 2 A). Arepresentative Southern blot of in vitro methylated GFP plas-mid transfected into HEK cells treated or untreated with TSAis shown in Fig. 2C. The histogram in Fig. 2D represents thequantification of demethylation of the GFP reporter gene asdetermined by Southern blot analyses. The degree of demethyl-ation was determined by quantifying the relative abundance ofthe fully digested HpaII fragment (529 bp) per lane, by imagedensitometry. The numbers and standard errors represent theaverage of demethylation from three independent experiments.A 4-fold increase in demethylation of transfected plasmid oc-curs in the presence of TSA. These results imply that deacety-lated histones protect ectopically methylated DNA from activedemethylation in HEK 293 cells.

TSA also induces a change in the state of expression of theGFP reporter plasmid, as expected. Fig. 2E shows HEK cells asexamined under fluorescent light with regular light overlaidwith the same field. Within the same size field, an average of18% of cells expressed GFP when treated with TSA, versus only0.13% of in the absence of TSA.

Histone Acetylation Can Trigger Demethylation Irrespectiveof the Presence of a Specific Promoter or Regulatory cis Ele-ments—Regional demethylation downstream from strong pro-moters may be a consequence of the interaction of specificfactors with distinct cis elements in the promoter region, or theformation of a pre-initiation complex that might be stimulatedby TSA inhibition of histone deacetylation (14). To determinewhether histone acetylation can trigger demethylation irre-spective of the presence of a specific promoter, we testedwhether a promoterless GFP construct becomes demethylatedin HEK cells. As illustrated in Fig. 2F, TSA stimulates de-methylation of a GFP reporter sequence residing in a promot-erless plasmid, as indicated by the almost complete digestion ofthe GFP sequence with the endonuclease HpaII (which cleavesthe sequence CCGG only when it is not methylated). The rel-ative abundance of fully demethylated HpaII fragments fromfour independent experiments was quantified using image den-sitometry and graphed in Fig. 2G. A 5-fold increase in de-methylation of transfected plasmid occurs in the presence ofTSA. This data supports the hypothesis that histone acetyla-tion and not necessarily the presence of a unique regulatoryregion directs the demethylation of a sequence of DNA in HEKcells.

TSA Increases Demethylation in a Dose- and Time-dependentManner—pCMV-GFP-transfected cells were treated with in-creasing concentrations of TSA (Fig. 3A) and subjected toSouthern blot analysis to investigate the dose response be-tween demethylation and inhibition of histone acetylation.HpaII sensitivity goes from undetectable with 0 �M TSA, tocomplete with 0.3 �M TSA, demonstrating that demethylationincreases with increasing TSA concentration. The extent ofdemethylation of GFP in the presence of TSA increases withincreasing time (Fig. 3B). HpaII sensitivity goes from undetect-able at 48 h post-transfection, to complete after 96 h. This timecourse also illustrates stability of the transfected plasmidthroughout the transfection experiment under these conditionsand suggests that demethylation of exogenously introducedmethylated plasmid occurs slowly, beginning only 72 hpost-transfection.

The Effect of TSA on Demethylation Is Independent ofChanges in Cell Cycle—Because the experimental paradigmused to detect demethylation requires the incubation of cellswith TSA for 72 h, we determined whether cell cycle arrest isnot a factor in changing the methylation pattern. In vitromethylated GFP-transfected HEK cells were replenished witheither Dulbecco’s modified Eagle’s medium containing either10% fetal calf serum in the presence or absence of TSA, or 0.5%fetal calf serum (fcs), and harvested 72 h post-transfection.Harvested cells were stained with propidium iodide and sub-jected to fluorescence-activated cell sorting analysis to deter-mine the cell cycle profile (Fig. 4A). Both cell populations thatwere either serum-starved (0.5% fetal calf serum) or treatedwith TSA displayed the same shift in cell cycle, from G1 towardG2 (Fig. 4A). The state of methylation of transfected GFPplasmid under the different conditions was analyzed by sodiumbisulfite mapping (Fig. 4B). Contrary to cells treated with TSA,which displayed 100% demethylation, no significant demethyl-ation occurred in transfected GFP plasmid in serum-starvedcells. Because the alterations in the cell cycle profiles of serum-starved and TSA-treated cells were similar, this experiment isconsistent with the conclusion that cell cycle changes caused byTSA per se are not responsible for the observed demethylation.

Demethylated Sequences Exhibit an Increased Associationwith Acetylated Histones in the Presence of TSA—To demon-strate that transiently transfected DNA is associated with ei-ther acetylated or deacetylated histones and that TSA can alterthe state of acetylation of histones bound to the GFP and CATreporter sequences, we performed a chromatin immunoprecipi-tation assay (CHIP) using anti acetyl-histone H3 antibody. Theresults presented in Fig. 5 show that there is an increasedassociation of the CAT gene sequence with acetylated histoneswhen they reside downstream from the strong SV40 promoter/enhancer (Fig. 5A, �TSA, �AB) relative to those residingdownstream from the pMet regulatory region (Fig. 5E, �TSA,�AB). TSA increases the abundance of both CAT and GFPsequences associated with acetylated histones (Fig. 5, A and C,�TSA, �AB). Interestingly, TSA treatment does not increaseacetylation of histones associated with pMetCAT� (Fig. 5E,�TSA, �AB), which is consistent with previous studies thatdemonstrate that not all sequences are acetylated followingTSA treatment (8). Perhaps, the context of factors interactingwith some regions renders them inaccessible to HATs. Consist-ent with our hypothesis, the pMetCAT� plasmid is not de-methylated even after TSA treatment (data not shown).

To provide a direct assessment of whether DNA associatedwith acetylated chromatin is demethylated, SV40CAT plasmidDNA isolated from anti-acetyl H3 histone immunoprecipitatedsamples, treated or not treated with TSA, was subjected tobisulfite mapping (Fig. 5B). All clones analyzed displayed al-

Histone Acetylation and Demethylation40780

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

FIG. 2. Histone deacetylase inhibitor TSA enhances demethylation. A, a physical map of the GFP region that was analyzed by bisulfitemapping is shown with the position of the CpGs indicated and numbered according to accession number U55761. In vitro methylated pCMV-GFPplasmid was transfected into HEK293 cells and treated with and without a final concentration of 0.3 �M TSA, 24 h post-transfection, and harvestedafter 96 h. DNA prepared from the transfectants was treated with bisulfite. The GFP region indicated in the map by dashed angled lines wasamplified, subcloned, and sequenced. Hatched arrows indicate the location of both the outside and nested primers used to amplify bisulfited DNA.Each line within the boxes represents an independent clone. A filled circle represents a methylated CG dinucleotide, and an empty circle representsa demethylated CG dinucleotide. The top box displays clones from control transfectants (�TSA), and the bottom box displays clones from cellstreated with a final concentration of 0.3 �M TSA (�TSA). B, representative sequencing gels of clones originating from cells treated and untreatedwith TSA. Arrows indicate the location of CG dinucleotides within the GFP gene. C and F, a fully methylated promoter or promoterless GFPplasmid, respectively, was transfected into HEK293 cells. The transfectants were either treated or not treated with a final concentration of 0.3 �M

TSA 24 h post-transfection and harvested after 96 h. 10 �g of isolated DNA was digested with 50 units of EcoRI followed by either digestion with20 units of MspI or HpaII restriction enzymes, fractionated on a 1.7% agarose gel, and was then subjected to Southern blot transfer andhybridization with a 32P-labeled GFP fragment (see map of probe in panel A, indicated by a dashed line flanked by restriction sites Cfr10I andAvaII). D and G, histograms represent the quantification of demethylation of the GFP reporter gene in the promoter (D) and promoterless (G)constructs as determined by Southern blot analyses. The degree of demethylation was determined by quantifying the relative abundance of thefully digested HpaII fragment (529 bp) per lane, by image densitometry. The numbers and standard errors represent the average of demethylationfrom three independent experiments for each the promoter and promoterless experiments. E, fluorescence microscopy of representative fields fromHEK cells transfected with GFP treated with or without TSA. The average percentage of expressing cells per field was calculated from between5 and 10 fields examined from three independent experiments.

Histone Acetylation and Demethylation 40781

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

most complete demethylation, in contrast to non-immunopre-cipitated DNA (Fig. 5B). The same degree of demethylation wasalso apparent for pCMVGFP plasmid DNA isolated from im-munoprecipitated samples; however, in this case DNA wasamplified only from cells treated with TSA (Fig. 5D). DNAcould not be immunoprecipitated when attempting to IP fromGFP-transfected cells, which were not treated with TSA, norfrom cells transfected with pMetCAT plasmid in the presenceor absence of TSA. This is consistent with previous studiesdemonstrating that methylated DNA is associated withdeacetylated histones as a result of recruitment of histonedeacetylases by methylated DNA binding proteins such asMeCP2 (15). The SV40 promoter seems to be less affected bymethylation as evident by the fact that it is partially immuno-precipitated by antibodies directed against acetylated histones.The fact that DNA that is associated with acetylated histones isdemethylated even in the absence of TSA treatment supportsthe hypothesis that demethylation following TSA treatmentrequires acetylation of the histones associated with the specificgene and is not a consequence of a general increase in de-methylase activity. This result can explain why TSA treatmentper se does not result in demethylation of all endogenous genes(8).

Transfected GFP Plasmid Assembles into a NucleosomeStructure—Experiments with transiently transfected DNAhave resulted in the abundance of information on the regula-tion of gene expression in eukaryotes as well as the effects ofDNA methylation and histone acetylation and deacetylation ongene expression. However, the issue of whether transientlytransfected DNA is assembled into nucleosomes has been con-troversial (16). Using plasmids transfected by calcium phos-phate precipitation, Reeves et al. (17) showed that nucleosomalladders are generated, which appeared indistinguishable fromthose of bulk cellular chromatin. On the other hand, Archerand colleagues (18) failed to observe nucleosome repeat pat-terns on transiently transfected murine mammary tumor viruspromoter. To address this question in our study we carried outmicrococcal nuclease (MNase) digestions of isolated nuclei from

cells transiently transfected with methylated GFP in the pres-ence of TSA (Fig. 6). Nuclei were subjected to digestion withincreasing amounts of MNase (0–600 units/ml) for 20 min at20 °C. Bulk DNA was isolated and fractionated on a 1.8%agarose gel (Fig. 6, first panel). Multiplicity of nucleosomalarrays are indicated (1 � mononucleosome, 2 � dinucleosome,etc.). The fractionated DNA was subjected to Southern blotanalysis and hybridization with a 32P-labeled GFP fragment(Fig. 6, second panel). Mononucleosomes and dinucleosomescan be detected but at lower doses of MNase than bulk DNA,consistent with previous findings showing that digestion oftranscriptionally active genes is more rapid than bulk DNA(19). Our data is consistent with the previously proposed modelthat nucleosomes are deposited onto non-replicating DNA intransfected cells; however, it is possible that the overall struc-ture may not be as organized as replicated cellular chromatin.Nevertheless, the transiently transfected DNA allows one todissect some of the basic principles of the interaction betweendemethylase and DNA assembled into nucleosomes.

Histone Acetylation Induces Replication-independent ActiveDemethylation of DNA—Recent data have demonstrated thatbinding of a transcriptional activator to its cognate site resultsin site-specific demethylation by a passive replication-depend-ent mechanism (20). Such a mechanism was proposed to ex-plain the demethylation of active genes during development

FIG. 3. Demethylation of transfected GFP is increased withincreasing TSA concentration and with time. A and B, in vitromethylated pCMV-GFP plasmid was transfected into HEK293 cells.24 h post-transfection the cells were treated with a final concentrationof 0.3 �M TSA (B) or with 0, 0.03, 0.1, or 0.3 �M TSA (A) and wereharvested after 96 h (A) or after 48, 72, or 96 h (B). A and B, 10 �g ofisolated DNA was digested with 50 units of EcoRI followed by eitherdigestion with 20 units of MspI or HpaII restriction enzymes. Alldigested DNA was fractionated on a 1.7% agarose gel and was thensubjected to Southern blot transfer and hybridization with a 32P-labeledGFP fragment (see map of probe in Fig. 2, panel A, indicated by adashed line flanked by restriction sites Cfr10I and AvaII).

FIG. 4. The effect of TSA on demethylation is independent ofcell cycle. A, HEK 293 cells were transfected with methylated GFPplasmid and grown for 72 h in the presence of 10% fetal calf serum inthe absence (Control) or presence of TSA (�TSA) or with 0.5% fetal calfserum (Serum Starved). Propidium iodide-stained cells were analyzedby fluorescence-activated cell sorting analysis and the percentage ofcells in G0/G1, S, and G2 for each condition is displayed by the histo-gram. B, DNA was harvested from each condition (Control, SerumStarved, and �TSA) subjected to bisulfite treatment followed by PCRamplification of the GFP gene, and cloning of the amplified fragment.Each line represents one clone. A filled circle represents a methylatedCG dinucleotide, and an empty circle represents a demethylated CGdinucleotide.

Histone Acetylation and Demethylation40782

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

(20). To test whether the demethylation detected after tran-sient transfection and treatment with TSA is occurring by asimilar mechanism, we digested DNA prepared from TSA-

treated HEK cells that were transiently transfected with invitro methylated pCMV-GFP, with the restriction enzymesDpn1 and XbaI. DpnI cleaves the sequence GATC only whenthe adenine is methylated on both strands (21). Because mam-malian cells do not bear a methylase that methylates the ade-nine in GATC, replication of the plasmid in HEK cells willrender it resistant to DpnI digestion. XbaI, on the other hand,does not cleave the sequence TCTAGAtc (which is present inpCMV-GFP) when the adenine is methylated (22). Replicationof the plasmid in mammalian cells will render it sensitive toXbaI digestion. However, the data presented in Fig. 7 showthat, although the transfected pCMV-GFP is fully cleaved withHpaII (Fig. 7A) indicating that it is fully demethylated follow-ing TSA treatment, it is completely digested with DpnI andresistant to XbaI (Fig. 7B). Control experiments shown in Fig.7C demonstrate that the plasmids are completely methylatedprior to their transfection into HEK cells, because they areresistant to HpaII cleavage. These data demonstrate that theplasmid did not replicate in HEK cells. Therefore, these datasupport the hypothesis that increasing histone acetylation in-duces replication-independent active demethylation of DNA.

Ectopic Expression of Mbd2b/Demethylase Stimulates De-methylation of Ectopically Methylated DNA—We recentlycloned and demonstrated that the methylated DNA bindingprotein MBD2b bears a demethylase (dMTase) activity (23).The availability of a cloned demethylase cDNA allowed us todetermine whether the rate-limiting step for demethylation isexclusively the state of acetylation of histones or whether in-creasing the level of demethylase in the cell could increasedemethylation. Methylated GFP was cotransfected with orwithout the Mbd2b/demethylase expression vector His-dMTase (23) into HEK 293 cells. We mapped the state ofmethylation of the transfected GFP gene by bisulfite mappingshown in Fig. 8A. The extent of demethylation within thehighly CG rich area of the GFP gene is 27% compared with 66%demethylation when dMTase is cotransfected with GFP. Fig.8B shows a representative Southern blot of transfected GFP inthe absence or presence of His-dMTase plasmid. The histo-grams in Fig. 8C represent the quantification of demethylationof the GFP reporter gene as determined by Southern blot anal-yses. The degree of demethylation was determined by quanti-fying the relative abundance of the fully digested HpaII frag-ment (529 bp) per lane, by image densitometry. The numbersand standard errors represent the average of demethylationfrom three independent experiments. These experiments dem-onstrate that demethylation of a given sequence is dependenton the abundance of demethylase in addition to the state ofhistone acetylation. Our data differ from another report thathas shown that Mbd2b/demethylase acts as a transcriptionalrepressor of a cotransfected methylated granulocyte/macro-phage-specific lysozyme gene enhancer and that it did notcause demethylation (24). A simple explanation to resolve thisinconsistency is that the interaction of MBD2b/demethylasewith methylated genes depends on the promoter context as wellas the accessibility of acetylated histones. Our unpublisheddata suggest that ectopic expression of His-dMTase does notdemethylate pMetCAT�, which is not associated with acety-lated histones even after TSA treatment (Fig. 5).

Because we have previously shown that His-dMTase de-methylates DNA indiscriminately in vitro (23), we also deter-mined whether demethylase could demethylate DNA in theabsence of a distinct regulatory sequence in the living cell. Asseen in Fig. 8, B and D, forced expression of His-dMTaseenhances the demethylation of pCMV-GFP and the promoter-less GFP (Fig. 8D) as indicated by the increase in the relativefraction of the plasmid that is digested with HpaII. The relative

FIG. 5. Bisulfite mapping and CHIP analysis of the associationof transfected plasmids with acetylated histones. A, C, and E,HEK 293 cells transfected with in vitro methylated plasmids(SV40CAT, pCMV-GFP, and pMetCAT�, respectively) were eithertreated or not treated with a final concentration of 0.3 �M TSA. The cellswere formaldehyde cross-linked after 96 h and subjected to a chromatinimmunoprecipitation assay using an antibody against acetylated his-tone 3 (AB�) or processed with no antibody (AB�) (see “Materials andMethods”). Pre denotes 10% of input DNA prior to immunoprecipita-tion. The CAT and GFP sequences were amplified, from purified DNA,by PCR using primers diluted to 50 �M, as indicated in Figs. 1B and 2A(respectively) as solid arrows, and as described under “Materials andMethods.” PCR was performed for each experiment on serial dilutions(from 1:10 to 1:10,000) of both immunoprecipitated and pre-immune(pre) DNA to ensure a linear relationship. A representative PCR isshown. PCR bands were quantified using image densitometry. �ABand �AB values were first normalized to pre (or input), before dividing�TSA values by �TSA values, for each. B and D, immunoprecipitatedDNA was subjected to bisulfite mapping analysis. Clones from DNA notimmunoprecipitated (NO IP CONTROL), immunoprecipitated withanit-acetylated H3 antibody from cells treated with TSA (Anti H3 IP; Band D) and not treated with TSA (Anti H3 IP �TSA; B only) arepresented.

Histone Acetylation and Demethylation 40783

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

abundance of fully demethylated HpaII fragments from fourindependent experiments was quantified using image densi-tometry and shown graphically in Fig. 8E. A Western blotanalysis using the Anti-Xpress antibody (which recognizes thehistidine-tagged protein) demonstrates expression of His-dMTase in transfected cells but not in untransfected HEKS(Fig. 8F), thereby confirming the expression of exogenouslyadded dMTase in transfected cells. This experiment shows thatincreasing the abundance of demethylase results in demethyl-ation of a gene in different contexts and that demethylase doesnot require specific regulatory sequences, nevertheless somepromoters can inhibit its activity as discussed above.

Ectopic Expression of Mbd2b/Demethylase Increases Expres-sion of Ectopically Methylated GFP—To determine whether theincreased demethylation is reflected in increased expression ofthe transfected pCMV-GFP plasmid and to demonstrate thatHis-dMTase does not suppress the activity of the pCMV-GFPpromoter, as has been shown for other promoters (25), cellextracts from HEK cells transfected with GFP plasmid andHis-dMTase or GFP alone, were subjected to Western blotanalysis using an antibody directed toward GFP protein. Cellextracts were loaded in triplicate, and equal loading was de-termined by Amido black staining (Fig. 9A). HEK cells cotrans-fected with both GFP and His-dMTase consistently expressedmore protein than GFP transfectants alone. In Fig. 9B, HEKcells expressing GFP, which were either cotransfected withGFP plasmid and His-dMTase or GFP alone, were counted 96 hpost-transfection by use of a hemacytometer with a fluores-cence microscope. The graph shows the results obtained fromtwo independent experiments. Whereas ectopic expression ofHis-dMTase increases the percentage of cells expressing GFP,it is clear that induction of expression by ectopic dMTase issignificantly less than by TSA treatment (compare with Fig.2E). These data suggest that demethylation is not sufficient toremove all the repression triggered initially by methylation.

DISCUSSION

The DNA methylation pattern is a fundamental constituentof the mechanisms regulating gene expression in vertebrates.It is therefore clear why it is important to understand howDNA methylation patterns are formed and how they cast theirintimate interrelationship with chromatin structure. The datapresented in this report support a simple but attractive modelthat explains the remarkable correlation between active genesand hypomethylation. Our data suggest that the state of de-methylation is determined by chromatin structure. The state of

methylation of a sequence is determined by a dynamic balanceof the abundance of demethylase activity in the cell and thestate of acetylation of histones.

We show that an identical sequence of DNA, which is ectopi-cally methylated, is actively demethylated in HEK 293 cellswhen it is associated with acetylated histones. Histone acety-lation is induced either by the SV40 enhancer that possiblyrecruits histone acetyl transferases to regulatory regions (Figs.1 and 5A) or by a pharmacological inhibitor of histone deacety-lation (Figs. 2, 5A, and 5C). DNA that is bound to acetylatedhistones also becomes demethylated (Fig. 5, B and D), thusproviding direct proof that DNA associated with acetylatedhistones is actively demethylated. The fact that SV40 DNAassociated with acetylated histones is completely demethylated,even in cells that were not treated with TSA, supports the hy-pothesis that active demethylation is a consequence of the asso-ciation with acetylated histones and not other side-effects of TSAtreatment. Augmenting demethylase activity by forced expres-sion of a cloned demethylase also stimulates demethylation (Fig.8) and gene expression (Fig. 9). Taken together, the data pre-sented in this paper fit a model proposing that the limiting stepin demethylation is the interaction between the demethylase andDNA. The probability of such an interaction is enhanced byeither histone acetylation or an increase in the abundance ofdemethylase. It is tempting to speculate that demethylase isinhibited from accessing the DNA by deacetylated histone tailsas has been shown to be the case with multiple transcriptionfactors (26).

Bisulfite mapping, which is followed by PCR amplification,

FIG. 6. Transfected GFP plasmid assembles as chromatin witha nucleosome structure. Genomic nuclei from HEK 293 cells treatedwith TSA were isolated and digested with 0, 5, 20, and 60 units ofMNase for 20 min at 20 °C. Isolated DNA was run on a 1.8% agarose gel(left lane). The right panel shows a Southern blot analysis and hybrid-ization with a 32P-labeled GFP fragment (see map of probe in Fig. 2,panel A, indicated by a dashed line flanked by restriction sites Cfr10Iand AvaII). Arrows point to mononucleotides, dinucleotides, etc. Thepositions of the size markers are indicated.

FIG. 7. Ectopically methylated plasmid DNA is actively de-methylated in the presence of TSA in a replication-independentmanner. A and B, in vitro methylated pCMV-GFP plasmid was trans-fected into HEK 293 cells. 24 h post-transfection the cells were treatedwith a final concentration of 0.3 �M TSA and harvested after 96 h. A, 10�g of isolated DNA was digested with 50 units of EcoRI followed byeither digestion with 20 units of MspI or HpaII restriction enzymes. B,5 �g of DNA was either undigested or digested with 40 units of EcoRI,40 units of DpnI, or 40 units of XbaI. C, 10 �g of plasmid (pCMV-GFP,GFP, pMetCAT�, and SV40CAT) was in vitro methylated with 20 unitsof SssI methylase (see “Materials and Methods”). 100 ng of DNA wasdigested with MspI or HpaII. A–C, all digested DNA was fractionatedon a 1.7% agarose gel and then subjected to Southern blot transfer andhybridization with a 32P-labeled GFP fragment (see map of probe inFig. 2, panel A, indicated by a dashed line flanked by restriction sitesCfr10I and AvaII).

Histone Acetylation and Demethylation40784

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

allows one to look at the demethylation events in a single DNAmolecule. A striking observation that emerges from this anal-ysis is that, with a few exceptions, active demethylation ofectopically methylated plasmids is regional and not site-spe-cific. Complete clusters of CGs are demethylated in some plas-mid molecules and none in others (Figs. 2 and 8). This patternof demethylation is inconsistent with distributive demethyla-tion but is consistent with a processive mechanism as haspreviously been suggested based on an in vitro analysis (27).

Regional and processive demethylation of CpG clusters is alsoconsistent with the picture that emerges from the mapping ofthe state of methylation of CpG-rich sequences in vivo. Forexample, all the CpG sequences in the CpG-rich first exon ofthe tumor suppressor p16 (28, 29) and the tumor suppressorp21 (30) are not methylated in most tissues, but the first exonis progressively completely methylated in some tumors that donot express p16 (28).

If histone acetylation is involved in regional demethylation of

FIG. 8. Forced expression of His-dMTase induces demethylation in a promoter independent manner. A, in vitro methylatedpCMV-GFP was transfected with or without the His-dMTase expression vector (23). DNA was harvested 96 h later and was subjected to bisulfitetreatment followed by PCR amplification of the GFP gene and cloning of the amplified fragment. Each line represents one clone. A filled circlerepresents a methylated CG dinucleotide, and an empty circle represents a demethylated CG dinucleotide. The top box contains clones frompCMV-GFP plasmid transfected alone (�dMTase), and the bottom box contains clones from pCMV-GFP plasmid cotransfected with His-dMTaseplasmid (�dMTase). B and D, fully methylated pCMV-GFP plasmid and promoterless GFP (respectively) were transfected into HEK 293 andharvested after 96 h. 10 �g of isolated DNA was digested with 50 units of EcoRI followed by either digestion with 20 units of MspI or HpaIIrestriction enzymes, fractionated on a 1.7% agarose gel, and was then subjected to Southern blot transfer and hybridization with a 32P-labeled GFP(see map of probe in Fig. 2, panel A, indicated by a dashed line flanked by restriction sites Cfr10I and AvaII). C and E, histograms represent thequantification of demethylation of the GFP reporter gene in the promoter (C) and promoterless (E) constructs as determined by Southern blotanalyses. The degree of demethylation was determined by quantifying the relative abundance of the fully digested HpaII fragment (529 bp) perlane, by image densitometry. The numbers and standard errors represent the average of demethylation from three to four independent experimentsfor each the promoter and promoterless constructs, respectively. F, total cell extracts from HEK cells untransfected, or transfected withHis-dMTase, were prepared using standard protocols and resolved on a 7.5% SDS-polyacrylamide gel. After transferring to nitrocellulosemembrane and blocking the nonspecific binding with 5% milk, transfected His-dMTase protein was detected using Anti-Xpress mouse monoclonalIgG (Invitrogen, R910-25) at 1:5000 dilution, followed by peroxidase-conjugated anti-mouse IgG (Jackson ImmunoResearch) at 1:20,000 dilution,and enhanced chemiluminescence detection with an ECL kit (Amersham Pharmacia Biotech).

Histone Acetylation and Demethylation 40785

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

CpG islands outside of a promoter region as suggested by ourdata, then histone acetylation should not be restricted only tothe promoter region per se. We show that the reporter genes weanalyzed are associated with acetylated histones. This obser-vation is in accordance with previous data showing that histoneacetylation is not restricted to nucleosomes within the pro-moter region but appears downstream of the transcription startsites (31), which can explain the demethylation of sequencesresiding a few hundred bases downstream from a promoter(Figs. 1 and 2). Recent data suggest that extended domains ofhistone acetylation can occur at large distances from matrixattachment regions (32).

Significant attention has been directed in the past towardsite-specific demethylation, particularly demethylation associ-ated with differentiation and activation of specific genes. Razinand Riggs (2) proposed that demethylation occurs in a site-specific and passive manner by binding of factors to specificsites during discrete time points in development that mask itfrom methylation during replication. This model has recentlygained experimental support (33). However, more recent datahave shown that passive demethylation is not simply depend-ent on the presence of DNA binding transcriptional activators(34). Yet still, other experiments have demonstrated site- andcell-specific active demethylation in transient transfection as-says (35), suggesting the presence of site- and cell-specificdemethylation machinery (36). The data presented here are notnecessarily in contradiction with these previous observations.It is possible that different mechanisms are responsible for thesite-specific demethylation that accompanies the activation ofgenes during development and the regional demethylationdemonstrated here.

An alternative reasonable hypothesis is that mechanismssimilar or identical to the demethylation activity described inthis report are also responsible for site-specific demethylation.Although regional demethylation might be triggered by a re-gional non-cell-selective histone acetylation, site-specific andcell type-specific demethylation might be directed by a specificrecruitment of histone acetyl transferases by certain activatorproteins. This model is consistent with a requirement for en-hancer or cis-acting sequences for site- and cell-specific de-methylation (35). For example, demethylation that is initiatedat the small pre-B cell stage of a single kappa light-chain allelerequires the presence in cis of both the intronic and 3� kappaenhancers (37). It is tempting to speculate that the activatorsthat interact with these cis elements trigger demethylation byrecruiting HATs and inducing histone acetylation.

The specific physiological function of the demethylase activ-ity detected by our transient transfection assay is unclear. Oneof the outstanding questions is: What is the physiological role ofdemethylase in somatic cells? Is it responsible for repair ofectopic methylation or does it play a dynamic role in maintain-ing the DNA methylation pattern and its correlation with chro-matin structure? It is also possible that the demethylase activ-ity studied here is an embryonic function that is aberrantlyactivated in some transformed cells. Nevertheless, the assaydescribed in this study reveals some of the basic principles ofthe demethylation reaction in living cells. These principlesmight explain a wide range of observations of demethylationevents in vivo as well as shed light on one of the fundamentalproperties of mammalian genomes, the correlation of gene ex-pression, chromatin structure, and the DNA methylationpattern.

The presence of demethylase that demethylates broad re-gions of ectopically methylated DNA in a somatic cell lineforces us to revisit our understanding of the DNA methylationpattern in somatic cells and the relationship between DNAmethylation and histone acetylation. The DNA methylationpattern at a given point like other biological signals might be asteady-state equilibrium of methylation and demethylation,whereas the direction of the reaction is determined by the localstate of histone acetylation. Recent data focused on how DNAmethylation changes histone acetylation; our data suggest thatit is possible that histone acetylation determines DNAmethylation.

Acknowledgment—We thank Johanne Theberge for her excellenttechnical assistance.

REFERENCES

1. Razin, A., and Cedar, H. (1977) Proc. Natl. Acad. Sci. U. S. A. 74, 2725–27282. Razin, A., and Riggs, A. D. (1980) Science 210, 604–6103. Jones, P. L., Veenstra, G. J., Wade, P. A., Vermaak, D., Kass, S. U., Landsberger,

N., Strouboulis, J., and Wolffe, A. P. (1998) Nat. Genet. 19, 187–1914. Eden, S., Hashimshony, T., Keshet, I., Cedar, H., and Thorne, A. W. (1998)

Nature 394, 8425. Szyf, M., Eliasson, L., Mann, V., Klein, G., and Razin, A. (1985) Proc. Natl.

Acad. Sci. U. S. A. 82, 8090–80946. Selker, E. U. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 9430–94357. Hu, J. F., Pham, J., Dey, I., Li, T., Vu, T. H., and Hoffman, A. R. (2000)

Endocrinology 141, 4428–44358. Cameron, E. E., Bachman, K. E., Myohanen, S., Herman, J. G., and Baylin,

S. B. (1999) Nat. Genet. 21, 103–1079. Rouleau, J., Tanigawa, G., and Szyf, M. (1992) J. Biol. Chem. 267, 7368–7377

10. Nur, I., Szyf, M., Razin, A., Glaser, G., Rottem, S., and Razin, S. (1985) JBacteriol 164, 19–24

11. Clark, S. J., Harrison, J., Paul, C. L., and Frommer, M. (1994) Nucleic AcidsRes. 22, 2990–2997

12. Crane-Robinson, C., Myers, F. A., Hebbes, T. R., Clayton, A. L., and Thorne,A. W. (1999) Methods Enzymol. 304, 533–547

13. Ramain, P., Bourouis, M., Dretzen, G., Richards, G., Sobkowiak, A., andBellard, M. (1986) Cell 45, 545–553

14. Workman, J. L., and Kingston, R. E. (1998) Annu. Rev. Biochem. 67, 545–57915. Nan, X., Ng, H. H., Johnson, C. A., Laherty, C. D., Turner, B. M., Eisenman,

R. N., and Bird, A. (1998) Nature 393, 386–389

FIG. 9. Expression of pCMV-GFP increases following ectopicexpression of His-dMTase. A, total cell extracts from HEK cellstransfected with pGFP-CMV plasmid and His-dMTase or pGFP-CMValone, were prepared using standard protocols and resolved on a 12.5%SDS-polyacrylamide gel in triplicates. After transferring to nitrocellu-lose membrane and blocking the nonspecific binding with 5% milk, GFPprotein was detected using rabbit polyclonal IgG (Santa Cruz Biotech-nology, sc-8334) at 1:500 dilution, followed by peroxidase-conjugatedanti-rabbit IgG (Sigma) at 1:5000, and enhanced chemiluminescencedetection with an ECL kit (Amersham Pharmacia Biotech). To showequal loading of the protein, the membrane was stained with Amidoblack. B, HEK cells were transfected with pGFP-CMV plasmid andHis-dMTase or pGFP-CMV alone. After 96 h, the number of cells ex-pressing GFP from each condition was counted using a hemacytometerwith a fluorescence microscope. The histogram shows the results fromtwo independent experiments.

Histone Acetylation and Demethylation40786

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

16. Smith, C. L., and Hager, G. L. (1997) J. Biol. Chem. 272, 27493–2749617. Reeves, R., Gorman, C. M., and Howard, B. (1985) Nucleic Acids Res. 13,

3599–361518. Archer, T. K., Lefebvre, P., Wolford, R. G., and Hager, G. L. (1992) Science 255,

1573–157619. Sun, Y. L., Xu, Y. Z., Bellard, M., and Chambon, P. (1986) EMBO J. 5, 293–30020. Hsieh, C. L. (1999) Mol. Cell. Biol. 19, 46–5621. McWhinney, C., and Leffak, M. (1990) Nucleic Acids Res. 18, 1233–124222. McClelland, M., Nelson, M., and Raschke, E. (1994) Nucleic Acids Res. 22,

3640–365923. Bhattacharya, S. K., Ramchandani, S., Cervoni, N., and Szyf, M. (1999) Nature

397, 579–58324. Boeke, J., Ammerpohl, O., Kegel, S., Moehren, U., and Renkawitz, R. (2000)

J. Biol. Chem. 275, 34963–3496725. Ng, H. H., Zhang, Y., Hendrich, B., Johnson, C. A., Turner, B. M., Erdjument-

Bromage, H., Tempst, P., Reinberg, D., and Bird, A. (1999) Nat. Genet. 23,58–61

26. Lee, D. Y., Hayes, J. J., Pruss, D., and Wolffe, A. P. (1993) Cell 72, 73–84

27. Cervoni, N., Bhattacharya, S., and Szyf, M. (1999) J. Biol. Chem. 274,8363–8366

28. Shapiro, G. I., Park, J. E., Edwards, C. D., Mao, L., Merlo, A., Sidransky, D.,Ewen, M. E., and Rollins, B. J. (1995) Cancer Res. 55, 6200–6209

29. Nguyen, T. T., Nguyen, C. T., Gonzales, F. A., Nichols, P. W., Yu, M. C., andJones, P. A. (2000) Prostate 43, 233–242

30. Milutinovic, S., Knox, J. D., and Szyf, M. (2000) J. Biol. Chem. 275, 6353–635931. Madisen, L., Krumm, A., Hebbes, T. R., and Groudine, M. (1998) Mol. Cell.

Biol. 18, 6281–629232. Fernandez, L. A., Winkler, M., and Grosschedl, R. (2001) Mol. Cell. Biol. 21,

196–20833. Lin, I. G., Tomzynski, T. J., Ou, Q., and Hsieh, C. L. (2000) Mol. Cell. Biol. 20,

2343–234934. Stunkel, W., Ait-Si-Ali, S., Jones, P. L., and Wolffe, A. P. (2001) J. Biol. Chem.

276, 20743–2074935. Paroush, Z., Keshet, I., Yisraeli, J., and Cedar, H. (1990) Cell 63, 1229–123736. Weiss, A., Keshet, I., Razin, A., and Cedar, H. (1996) Cell 86, 709–71837. Mostoslavsky, R., Singh, N., Kirillov, A., Pelanda, R., Cedar, H., Chess, A., and

Bergman, Y. (1998) Genes Dev. 12, 1801–1811

Histone Acetylation and Demethylation 40787

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from

Nadia Cervoni and Moshe SzyfDemethylase Activity Is Directed by Histone Acetylation

doi: 10.1074/jbc.M103921200 originally published online August 27, 20012001, 276:40778-40787.J. Biol. Chem. 

  10.1074/jbc.M103921200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/276/44/40778.full.html#ref-list-1

This article cites 37 references, 22 of which can be accessed free at

by guest on August 9, 2016

http://ww

w.jbc.org/

Dow

nloaded from