site-specific methylation of the promoter alters deoxyribonucleic acid-protein interactions and...

602

BIOLOGY OF REPRODUCTION 64, 602–610 (2001)

Site-Specific Methylation of the Promoter Alters DeoxyribonucleicAcid-Protein Interactions and Prevents Follicle-Stimulating Hormone ReceptorGene Transcription1

Michael D. Griswold2 and Jeong-Seon Kim

School of Molecular Biosciences, Center for Reproductive Biology, Washington State University,Pullman, Washington 99164-4660

ABSTRACT

In the male gonad, the FSH receptor (FSHR) gene is ex-pressed only in Sertoli cells. To date, the mechanism(s) respon-sible for Sertoli cell-specific expression of the FSHR gene areunknown. In this study, DNA methylation at specific sites in thepromoter are shown to lead to changes in the DNA-protein in-teractions at those sites and, subsequently, to transcriptional re-pression of the gene. The extent of methylation of cytosine res-idues within the core promoter region of genomic DNA isolatedfrom cells/tissues that expressed, or did not express, the FSHRgene was analyzed by the sodium bisulfite conversion technique.All seven cytosine residues in CpG dinucleotides within the corepromoter region were found to be unmethylated in primary cul-tured rat Sertoli cells that were actively expressing FSHR mRNA.In contrast, in tissues not expressing FSHR the same region ofthe gene was methylated at each of the CpG dinucleotides ex-amined. In addition, DNA-protein interactions in three primaryregulatory regions of the promoter were examined by electro-phoretic mobility shift assays (EMSA) with synthetic oligonucle-otides containing selectively methylated cytosine residues.Methylation of a CpG sequence within a consensus E box ele-ment (CACGTG, 2124/2119) decreased the binding affinity ofUSF1/2 transcription factors for this element. Methylation of theCpG sequence in the Inr region (CCGG, 285/282) allowed theformation of an additional DNA-protein complex. Methylationat both cytosine residues in the E2F element (mCGmCG) gener-ated a new methylcytosine-specific DNA-protein complex. Thecore FSHR promoter region of a mouse Sertoli cell line (MSC-1) that does not express FSHR was shown to be methylated atfour CpG dinucleotides. The demethylation of these four sitesby treatment of the MSC-1 cells with 5-aza-29-deoxycytidine (5-azaCdR) activated the transcription of the FSHR gene. Taken to-gether, these results suggest that cytosine methylation is a majorfactor in the repression of the expression of the FSHR gene.

FSH, gene regulation, hormone action, Sertoli cells

INTRODUCTION

The biosynthetic properties and characteristic structureof Sertoli cells define their fundamental role as nurse cellsfor spermatogenesis [1, 2]. Follicle-stimulating hormonestimulates the expression of genes in Sertoli cells that pro-mote the initiation and maintenance of spermatogenesis [3,4]. The FSH signal is mediated by the FSH receptor

1Supported by NICHD grant no. HD 10808.2Correspondence: Michael D. Griswold, School of Molecular Biosciences,630 Fulmer Hall, Washington State University, Pullman, WA 99164-4660.FAX: 509 335 9688; e-mail: [email protected]

Received: 7 August 2000.First decision: 6 September 2000.Accepted: 22 September 2000.Q 2001 by the Society for the Study of Reproduction, Inc.ISSN: 0006-3363. http://www.biolreprod.org

(FSHR) whose expression is highly restricted to Sertolicells in the testis and granulosa cells in the ovary. Overall,in the male rodent, FSH appears to be essential for quan-titatively normal reproductive function, although it is notrequired for fertility [5].

Cell-specific gene expression is generally considered tobe regulated by the interaction of basal and cell-specifictranscription factors with cis-acting promoter elements [6].In the TATA-less rat FSHR promoter, three major regula-tory regions have been identified. First, the initiation oftranscription in the FSHR gene takes place within Inr ele-ments (GCAGATC, 2100/294; ACAGTGT, 281/275)that encompass two major transcriptional initiation sites atposition 298 and 280 base pairs (bp) [7, 8]. Second, an Ebox sequence (CACGTG, 2124/2119) within the FSHRpromoter has been identified as an essential positive regu-latory element that binds with the ubiquitous transcriptionfactor USF1/2. Mutation of the E box element decreasedreporter gene activity in transfected cells by more than 50%[8, 9]. Third, a consensus E2F (TTTCGCG, 245/239) el-ement of the FSHR promoter was reported as a possibleregulatory element for transcription activity of the FSHRgene [9]. These previous studies failed to identify a cis-acting element that could account for the cell specificity ofthe rat FSHR gene expression, and the proximal promoterwas shown to be active in initiating transcription of reportergenes when transfected into a variety of cells in vitro [7–10].

Some mammalian genes exhibit an inverse correlationbetween the extent of DNA methylation and gene activity[11–14]. Methylation of DNA is a heritable biological sig-nal and occurs primarily at the fifth position of a cytosineresidue within CpG dinucleotide motifs on both strands ofDNA [15, 16]. Previously, using methylation-sensitive re-striction endonucleases, a cytosine base at nucleotide 284was shown to be methylated in the promoter of a mouseSertoli cell line (MSC-1). Expression of the FSHR gene inthis cell line is silenced [10]. The core rat FSHR promoterhas seven potential methylation sites, including position284, but the methylation pattern of the entire core promoterhas not been examined.

In this study, the extent of and consequence of methyl-ation of all cytosine residues in the FSHR core promoterwas determined. A sodium bisulfite conversion techniquewas used to analyze the in vivo methylation pattern of thecore promoter region in tissues where FSHR expressionwas silenced and in Sertoli cells that were actively tran-scribing the FSHR gene. The targeted methylation of cis-acting elements in the FSHR promoter region altered theinteractions of the promoter with nuclear proteins from theSertoli cells. In addition, the extent of demethylation of thepromoter region in MSC-1 cells was shown to correlatedirectly with the level of endogenous FSHR gene tran-

603CYTOSINE METHYLATION, FSH RECEPTOR GENE EXPRESSION

scripts. The results show that DNA methylation correlatesdirectly with the cell-specific silencing of the FSHR gene.

MATERIALS AND METHODS

Cell Cultures of Primary Sertoli Cells and MSC-1Cell Line

Primary Sertoli cells were cultured from 20-day-oldSprague-Dawley male rats as previously described [17].Cells were plated in 150-mm culture dishes with Hams F-12 medium (Gibco BRL Life Technologies, Inc., Grand Is-land, NY) and maintained at 328C in a humidified 5% CO2and 95% air atmosphere. The media was changed on thethird day of culture. Prior to extraction of nuclear proteins,cells were treated with 10% bovine calf serum with F-12medium for 24 h. Cells were harvested on the fifth day forgenomic DNA isolation and nuclear protein extraction. Themouse Sertoli cell line, MSC-1, was grown in Dulbeccosmodified Eagles medium (DMEM) supplemented with 10%bovine calf serum at 378C in the 5% CO2 and 95% airconditions.

Sodium Bisulfite Conversion of Genomic DNA

The genomic sequencing was based upon the bisulfitemodification and subsequent polymerase chain reaction(PCR) amplification of single-stranded DNA originally de-scribed by Frommer [18]. Genomic DNA was isolated fromcultured rat Sertoli cells, tissues from 20-day-old rats(brain, kidney, liver, and spleen), as well as MSC-1 cells[19]. The genomic DNA was digested with the restrictionenzyme BamHI, and then 2 mg of DNA was denatured with0.3 N NaOH in a volume of 25 ml for 20 min at 378C. Thedenatured DNA was treated with sodium bisulfite by ad-dition of 250 ml of a freshly prepared solution of 4.0 Msodium bisulfite (NaHSO3) and 10 mM hydroquinone at pH5.0. The reaction was overlayered with mineral oil and in-cubated at 558C for 8–12 h. The bisulfite-treated DNA wasrecovered using the DNA Clean-Up System (Promega,Madison, WI), according to the manufacturer’s instructions.The DNA was desulfonated by treatment with 0.3 N NaOHfor 20 min at 378C, and the reaction was neutralized bytreatment with 3 M ammonium acetate (CH3COONH4, pH7.0). Finally, the DNA was precipitated with ethanol andresuspended in 100 ml of TE buffer (10 mM Tris-HCl, 0.5mM EDTA, pH 7.5). The bisulfite-converted DNA (50 ng)was used for PCR amplification.

Polymerase Chain Reaction Amplificationof the Promoter Region

The sense strand of the region of interest of the FSHRpromoter was amplified from bisulfite-treated DNA by PCRusing strand-specific external primers. This reaction product(5 ml) was used directly in a second PCR reaction with in-ternal primers. These two external (primer set A) and internal(primer set B) primer sets were designed with regard to CpGsites because cytosine residues in any given site would bemodified after bisulfite conversion reaction (indicated in un-derlines). Thus, the primer pairs can recognize the bisulfite-modified sense strand and cannot anneal to the nonmodifiedDNA strand. The following PCR primers were used for bi-sulfite genomic DNA sequencing of the 2278- to 146-bpregion of the rat FSHR core promoter: primer set A, PA1:59-GGAGAAGATAGTAGTGATTAGTAGGGAT-39(2278/2250), PA2: 59-ATACTCAAATAAAAACCACA-CAACTAL-39 (1171/1196); primer set B, PB1: 59-

TGGGGGTTAAGGAATAAAAAATATAGGTT-39 (2278/2250), PB2: 59-ATCCCATACCCAAAAATACCAACAAA-39 (121/146).

The first round of PCR was performed in a volume of50 ml containing 50 ng bisulfite-treated DNA, 10 mM ofprimer set A designed to read the upper strand of the re-gion, 200 mM dNTPs, 3 mM MgCl2, and 0.5 units of TaqDNA polymerase (Gibco BRL Life Technologies). Ampli-fication was performed under the following conditions: onecycle of 948C for 2 min; five cycles of 948C for 1 min,488C for 2 min, 728C for 2 min; 30 cycles of 948C for 0.5min, 488C for 2 min, 728C for 1.5 min; one cycle of 728Cfor 6 min. The first PCR product (5 ml) was used in asecond round PCR reaction that contained 10 mM of primerset B. The PCR reaction was performed under the sameconditions as described above except an annealing temper-ature of 538C was used. Following PCR amplification withthe gene-specific primer, only 5-methylcytosine residues areamplified as cytosine, whereas all uracil residues, originallycytosine in the template, are amplified as thymine. To vi-sualize the genomic DNA sequences, at least five separatebisulfite-reacted genomic DNA preparations were used forsequencing.

Electrophoretic Mobility Shift Assay

The ability of protein to form complexes with syntheticoligonucleotides was detected by electrophoretic mobilityshift assay (EMSA), according to the method of Kerr [20].Nuclear extracts were prepared from Sertoli cells and othertissues as follows [21]: Cultured cells were washed withPBS, collected with a scraper, and precipitated by centri-fugation at 1500 3 g for 5 min. The cells were resuspendedin 400 ml of ice-cold hypotonic buffer (10 mM Hepes, pH7.9, 1.5 mM MgCl2, 10 mM KCl, 0.2 mM PMSF, 0.5 mMdithiothreitol [DTT]), allowed to swell for 15 min, and pre-cipitated by centrifugation at 1500 3 g for 5 min. The cellswere then resuspended in high salt buffer (20 mM Hepes,pH 7.9, 25% glycerol, 1.5 mM MgCl2, 420 mM NaCl, 0.2mM EDTA, 0.2 mM PMSF, 0.5 mM DTT) and incubatedon ice for 20 min. After centrifugation, the supernatant wasfrozen in aliquots at 2708C. The protein concentration inthe nuclear extracts was determined using the micro-bicin-choninic acid protein assay (Pierce Chemical Co., Rock-ford, IL). The probes used in the EMSA were double-stranded synthetic oligonucleotides corresponding to posi-tions of the FSHR promoter region. Several oligonucleo-tides, containing an E box, an Inr, or an E2F motif weresynthesized using a DNA synthesizer (PE Biosystems, Fos-ter City, CA). For synthesis of methylated oligonucleotides,5-methylcytidine phosphoramidite replaced cytosine phos-phoramidite at the CpG site of the sequences and annealedto double strands. The following oligonucleotide sequenceswere used as complementary strand probes: E box (21-mer,2133/2113), E box-WT: TTGGTGGGTCACGTGATC-TTG, E box-Me: TTGGTGGGTCAMGTGATCTTG; Inr(28-mer, 2105/278), Inr-WT: TCCAAGCAGATCTCT-CTTATCCGGACAG, Inr-Me: TCCAAGCAGATCTCT-CTTATCMGGACAG, Inr-Mut: TCCAAGCAGATCTCT-CTATCtaGGCAG; E2F (23-mer, 254/232), E2F-WT:TGTGGAAGTTTTCGCGCTGATGC, E2F-MeA: TGT-GGAAGTTTTMGCGCTGATGC, E2F-MeB: TGTGGA-AGTTTTCGMGCTGATGC, E2F-MeC: TGTGGAAGT-TTTMGMGCTGATGC. Italicized sequences represent aconsensus cis-acting element. Among sequences, M indi-cates 5-methylcytosine. Mutated nucleotides other than

604 GRISWOLD AND KIM

methylation are shown as lowercase letters. The WT, Me,and Mut represent wild-type, methylated, and mutatedprobes, respectively. Both single-stranded oligonucleotideswere end-labeled using 1 unit of T4 polynucleotide kinaseand 12.5 pmole of [g-32P]ATP at 378C for 30 min, and thelabeled complementary strands were then annealed. Thedouble-stranded oligonucleotides were purified using a 20%native polyacrylamide gel.

The 32P-labeled double-stranded oligonucleotides, eithermethylated or unmethylated, corresponding to a specific re-gion of FSHR gene promoter were incubated with the nu-clear extracts as described below. The binding reaction wasperformed by incubation of 2- to 4-mg nuclear extracts withapproximately 4 pmole (50 000 cpm) of end-labeled DNAat 48C for 30 min in the presence of 2 mg of poly(dI-dC)(dI-dC) or poly(dA-dT)(dA-dT) in a final volume of 15 ml,containing 12.5 mM Hepes, pH 7.9, 5% glycerol, 25 mMKCl, 10 mM MgCl2, and 0.5 mM DTT. The DNA-proteincomplex was separated from the unbound probe on a 5%native polyacrylamide gel (29:1 acrylamide:bis-acrylam-ide) in 0.53 TBE (22.5 mM Tris-borate, pH 8.2, 0.5 mMEDTA). Gels were dried and exposed to autoradiographicx-ray film. In competition experiments, molar excesses ofunlabeled oligonucleotides were incubated with bindingbuffer prior to addition of the labeled probe.

Treatment of MSC-1 Cell Line with 5-azaCdR

The MSC-1 cells were cultured in 100-mm plates withDMEM and 10% bovine calf serum (BCS) to 60% conflu-ence at a density of 1 3 106 cells per plate. The MSC-1cells were treated with 0.5 mM, 1.0 mM, or 2 mM 5-aza-29-deoxycytidine (5-azaCdR, Sigma) for 24 h [22]. The cul-tured medium was changed 24 h after treatment and at in-tervals of 3 days. The cells were allowed to grow in DMEMwith 10% BCS for 9 days. On the ninth day followingtreatment, the cells were treated again with 5-azaCdR usingthe same conditions. After incubation for 24 h, the mediawere replaced and the cells incubated for another 3 days.Genomic DNA and total RNA were isolated from controlcells and 12-day treated cells as described [19, 23].

Analysis of DNA Demethylation by 5-azaCdRin the MSC-1 Cells

Demethylation of DNA by 5-azaCdR was correlated withFSHR expression by analysis of the methylation pattern inthe core promoter with bisulfite genomic sequencing as pre-viously described [18]. After treatment of genomic DNAwith sodium bisulfite, the sense strand of the region of FSHRpromoter was amplified using mouse sequence-specific prim-ers. Both external (set A) and internal (set B) primer sets forgenomic sequencing of the mouse FSHR core promoter wereused. The following PCR primers were used for bisulfitesequencing of the 2273- to 146-bp region of the core pro-moter of mouse FSHR gene: primer set A, mPA1: 59-ATTTTGATATTATTGAGAAGAGAGTAGTG-39 (2425/2397), mPA2: 59-ATACTCAAATAAAAAACACAAAAC-TA-39 (2171/2204); primer set B, mPB1: 59-ATAGGTTTT-GAAGGATAAGATAGGTGTTT-39 (2273/2245), mPB2:59-ATCCCAAACCCAAAAATACCAACAAA-39 (121/146).

The first PCR reaction was performed under followingconditions: one cycle of 948C for 2 min; five cycles of 948Cfor 1 min, 578C for 2 min, 728C for 2 min; 30 cycles of948C for 0.5 min, 588C for 2 min, 728C for 1.5 min; onecycle of 728C for 6 min. The second PCR reaction was

performed under the same conditions as described aboveexcept at an annealing temperature of 588C. The final am-plified PCR product was 330-bp of the mouse FSHR corepromoter.

Isolation of RNA and Reverse Transcription-PCR

Total RNA was isolated from 5-azaCdR-treated MSC-1cells and nontreated total testis as described by Kingston[23]. After the 12-day treatment with 5-azaCdR, cells werelysed with 3.5 ml of 4 M guanidinium solution containing350 ml of b-mercaptoethanol. The cell lysate was drawnthrough a syringe with a 20-gauge needle to shear genomicDNA. The lysate was then layered on top of 1.5 ml of 5.7M CsCl in 5-ml ultracentrifuge tubes. After centrifugationat 35 000 rpm for 16 h, the supernatant was decanted, andthe RNA pellet was washed twice in 70% ethanol preparedwith diethyl pyrocarbonate-treated water. Total RNA fromthe whole testes of normal adult mice was isolated for pos-itive control of the reverse transcription (RT)-PCR reaction.The total RNA of the testis was extracted with Trizol re-agent as recommended by the supplier (Gibco BRL LifeTechnologies). For cDNA synthesis, RT reactions were per-formed using total RNA (3 mg), 5 pmole of sequence-spe-cific primer nA2, 5 pmole of dNTPs, and 1 unit of reversetranscriptase Superscript II (Gibco BRL Life Technologies)in a 20-ml reaction. The reaction mixture was incubated 1h at 428C, and the reaction was terminated by incubationfor 15 min at 758C. Negative control RT reactions wereperformed under the same conditions without total RNA.The cDNA samples were then amplified with primers (nB1and nB2) specific for the mouse FSHR promoter. The PCRwas performed in a volume of 50 ml containing 2 ml of thecDNA from the RT reaction, 10 mM of primer set nB, 200mM dNTPs, 2 mM MgCl2, and 0.5 unit of Taq DNA poly-merase (Gibco BRL Life Technologies). Amplification wasperformed under following conditions: one cycle of 948Cfor 3 min; five cycles of 948C for 1 min, 488C for 2 min,728C for 2 min, 25 cycles of 948C for 0.5 min, 508C for 2min, 728C for 1.5 min; one cycle of 728C for 5 min. AfterPCR, the amplified cDNAs were electrophoresed on 0.8%agarose gel. The RT-PCR was also done with genomicDNA before treatment of the cells as a negative control andin the total mouse testis as a positive control. Primer nPA2was used for the RT reaction, and primer sets nB used forPCR were designed to amplify 290 bp of the mouse FSHRpromoter. Primer sequences are as follows: nPA2: 59-GAC-GATACTCACAGTTCAATG-39 (1143/1163); primer setnB, nPB1: 59-TTGAAGGATAAGACAGGTGCT-39(2266/2246), nPB2: 59-AAGAATGCCAGCAAGGAGA-39 (116/132).

RESULTS

Analysis of Methylation Sites Within the Core Promoterof the FSHR Gene in Cells and Tissues

The pattern of cytosine methylation within the core pro-moter of the FSHR promoter region was analyzed using thedifferences in bisulfite reactivity for cytosine and 5-meth-ylcytosine residues. A comparison was made between ge-nomic DNAs from Sertoli cells that express the FSHR geneand various tissues that do not express the FSHR gene. The320-bp region analyzed by bisulfite genomic sequencingencompassed seven CpG dinucleotides in the rat FSHR pro-moter that are potential methylation sites (221 bp shown inFig. 1), and four CpG dinucleotides in the mouse promoter


FIG. 1. Nucleotide sequence alignment of the FSHR core promoter re-gion. Numbers indicate base positions with respect to the translation ini-tiation site. Arrows indicate major transcription initiation sites of themouse and rat core FSHR promoter, 298, and 280, identified by primerextension analysis [7]. Lines indicate the potential CpG methylation sites.The consensus sequences for major regulatory elements, E box, E2F, andthe Inr regions are indicated in bold type. The translation initiation siteATG is indicated with (11).

FIG. 3. Summary of analysis of methylation sites within the core pro-moter of the FSHR gene in cells and tissues. The DNA methylation statusfrom genomic DNA from primary Sertoli cells, various tissues, and MSC-1 cells. For an analysis of the rat FSHR promoter, the region betweenposition 2278 and 146, relative to the translation start site was selected.Triangles indicate the potential methylation site. Symbol (1) indicates thesites that 5-methylcytosine residues are present at CpG dinucleotides inall of at least three experiments. Symbol (/) indicates that cytosine residuesare present at CpG dinucleotides in all of at least three experiments.

FIG. 2. Representative results of sodium bisulfite genomic sequencing. Sequence analysis of a rat promoter in a region encompassing the core promoterregion including E box, Inr, and E2F elements following sodium bisulfite treatment. In vivo methylation of CpG dinucleotides of the FSHR promoterwas determined by using the bisulfite conversion reaction, following PCR amplification on genomic DNA from Sertoli cells and brain tissue. Resultsshown are representative of results from three experiments with each tissue.

(230 bp shown in Fig. 1). The CpG dinucleotides in the ratFSHR promoter were found at nucleotides 2155, 2122,2110, 284, 242, 240, and 220. Some of the methylatedcytosine bases were located within major regulatory regionsof the promoter such as the E box (CACGTG, nucleotide[nt] 2122), the Inr region (-CCGG-, nt 284), as well asthe consensus sequence for transcription factor E2F(TTTCGCG, nt 242 and 240) (Fig. 1) [7–10]. The analysisof the sequence showed that all cytosine residues that arepotential methylation sites in the region of interest werefully converted to thymine in the FSHR promoter of ratSertoli cells. Representative sequencing results for the threemajor regulatory regions are shown in Figure 2. Likewise,all cytosine residues that were adjacent to adenine, thymine,and cytosine residues in the original template were con-verted to thymine, indicating that the bisulfite conversionreaction was complete. These data showed that every CpGdinucleotide present in the core promoter region of the

FSHR gene was unmethylated in the rat Sertoli cells (Figs.2 and 3). On the contrary, in the promoter of brain tissuein which the gene was not expressed, cytosine residues inCpG dinucelotides were not converted to thymine (Fig. 2).We extended this analysis to a variety of different rat tissuessuch as kidney, liver, and spleen and found identical resultsin that all seven CpG dinucleotides were methylated in allgenomic DNAs examined (results summarized in Fig. 3).

The same analysis was done on DNA from the mouseMSC-1 cells (summarized in Fig. 3). MSC-1 is a well-char-acterized Sertoli cell line that does not express the FSHRmRNA [24, 25]. The methylation pattern of the DNA fromthe MSC-1 cells was consistent with that found in the non-


FIG. 4. Characterization and effect of DNA methylation on E box byEMSA. Nuclear extracts derived from Sertoli cells (3 mg) were incubatedwith labeled unmethylated or methylated E box oligonucleotides and sep-arated on a polyacrylamide gel. As a probe, a synthetic oligonucleotidefor the E box element (2133/2113) was used that contained either wild-type (-CACGTG-) or methylated cytosine (-CAMGTG-) within the corerecognition site. Probe (50 000 cpm) was used with 1, 2.5, 5, 7.5, or 10mg of proteins from Sertoli cell nuclear extracts.

FIG. 5. Characterization and effect of DNA methylation on the Inr re-gion by EMSA. An EMSA with the Inr element (2105/278), containingwild-type (-TATCCGG-), methylated sequences (-TATCMGG-), or mutatedsequence (-CAAGCGG-) within the core recognition site. Specific DNA-protein complexes are indicated by multiple retarded bands within theputative transcription initiation region. Protein bound to the methylatedelement has similar affinity to the wild-type element with the additionalbinding activity (complex M).

expressing rat tissues. Four CpG methylation sites arefound at nucleotides 2161, 2121, 284, and 243 in themouse FSHR promoter. Three of these sites are found inthe major regulatory elements of the mouse promoter in-cluding the consensus E box (CACGTG, nt 2121), the Inrregion (-CCGG-, nt 284), and the E2F (TTTCGCT, nt243).

Effects of Methylation on the Regulatory Regionof the FSHR Promoter

The presence of the CpG dinucleotides in the regulatoryregions of the FSHR core promoter suggested a possibleinteraction between cytosine methylation and the bindingof transcription factors. The EMSAs with synthetic oligo-nucleotides that contained either unmethylated or methyl-ated cytosine base(s) in the CpG sequence were used toinvestigate whether site-specific DNA methylation of thecis-acting regulatory region could directly affect the bind-ing affinity of the transcription complexes.

It was shown in previous studies that a consensus E boxelement in the FSHR promoter is a functional binding sitefor the USF1/2 heterodimers, and high affinity binding tothis site is required for activation of transcription [8, 9].The EMSA binding reactions were done with selectivelymethylated or unmethylated oligonucleotides containing theE box sequence and nuclear extracts from cultured rat Ser-toli cells. Major retarded bands, previously shown to con-sist largely of complexes of DNA and USF heterodimerswere formed on the gel with the unmethylated E box probe(E box-WT, 21mer) that contained a consensus -CACGTG-sequence (Fig. 4, lane 1). The EMSA results with the meth-ylated probe E box-Me, where the CpG dinucleotide wasreplaced with 5-methyl CpG (-CAMGTG-), were qualita-tively the same as with the unmethylated probe (E box-

WT), but the bands were formed with a reduced affinity(lane 2). Furthermore, the level of binding of the protein tothe methylated E box element was compared by EMSAwhere probe was incubated with increasing amounts of nu-clear extracts (Fig. 4). The differential binding affinity inthe complex was estimated by densitometric analysis to beat least 3.5-fold.

The region (-CCGG-, 285/282) between the two Inrelements also contained differentially methylated cytosineresidues. The Inr element is known to provide basal tran-scriptional initiation by binding of a minimal set of proteinswhen a TATA box is not present in the promoter [26–28].Protein binding activity was determined using EMSA withan oligonucleotide probe corresponding to the region 39 ofthe first Inr element and including the CpG dinucleotide.Multiple DNA-protein complexes were generated by theunmethylated Inr probe (Inr-WT, -TATCCGG-) and rat Ser-toli cell nuclear extracts (lane 1; Fig. 5). An EMSA withthe methylated probe Inr-Me (-TATCMGG-), where theCpG dinucleotide was replaced by 5-methyl CpG, revealedan additional faster migrating complex (complex M; lane2; Fig. 5). When a mutated probe Inr-Mut (-TATCTAG-)where CpG was replaced by TpA was used in the EMSA,all of the faster migrating bands were selectively reducedwithout generating a new complex (Fig. 5, lane 3).

From the results of bisulfite sequencing analysis, twocytosine residues of the E2F element were found to be


FIG. 6. Characterization of and effect of DNA methylation on the E2Felement. Binding of Sertoli cell nuclear proteins to the E2F element wasexamined by EMSA. The probes were E2F-WT (-TTTCGCG-) in A and E2F-MeC (-TTTMGMG-) in B. The competing oligonucleotides at 100-fold mo-lar excess were: E2F-WT in lanes 2 and 8; E2F-MeC in lanes 3 and 7;E2F-MeA (-TTTMGCG-) in lanes 4 and 9; E2F-MeB (-TTTCGMG-) in lanes5 and 10.

methylated (see Figs. 2 and 3). Sequence analysis indicatedthat a putative regulatory E2F sequence (TTTCGCG), lo-cated about 40 bp upstream of the translational initiationsite, was located in this transcribed region of the gene. Twodistinctive DNA-protein complexes (I and II) were formedin an EMSA analysis using an unmethylated probe, indi-cating that putative binding proteins for the E2F elementwere present in nuclear extracts from Sertoli cells (Fig. 6).An E2F-1-specific antibody (Santa Cruz) was used inEMSA experiments, but no supershifted complex was de-tected (data not shown). When either a methylated probeE2F-MeA (-TTTMGCG-) or E2F-MeB (-TTTCGMG-) wasused to examine relative effects of the methylation of eachsite on protein binding, it was shown that the methylationat one of the CpG sites resulted in a weaker formation ofboth complexes, especially complex I (not shown). A cor-responding mutation of the E2F sequences E2F-MutA (-TTTATCG-) had a modest reduction in binding activity thatwas similar to the effect of methylation of either CpG di-nucleotides within the consensus E2F sequence. WhenE2F-MeC that contained two methylated cytosine residues(-mCGmCG-) was used as the probe, a new complex (Fig.6, complex III, lane 6) was formed with a similar, althoughnot identical, mobility to that complex I (Fig. 6, lane 1).

Competition analysis was used to further identify theDNA-protein complexes shown with probes E2F-WT andE2F-MeC, respectively. A 100-fold molar excess of unla-beled oligonucleotide E2F-WT significantly competed withtwo complexes formed with the probe E2F-WT (self-com-petition) (Fig. 6, lane 2). The intensity of the doublet com-plexes generated with the E2F-WT probe was also clearlyreduced by a molar excess of the unlabeled oligonucleo-tides containing a methylated cytosine residue in either thefirst or the second site (Fig. 6, lanes 4 and 5). However, anexcess of the unlabeled oligonucleotide E2F-MeC contain-ing mCGmCG sequences failed to compete with the for-mation of the doublet complexes at a 100-fold molar excess(Fig. 6, lane 3). Thus, it was clear that the doublet com-plexes were specific DNA-protein interactions with thewild-type E2F binding site. Conversely, the complexformed with the E2F-MeC probe could compete only weak-ly with an excess of unlabeled E2F-MeA or E2F-MeB ol-igonucleotides (Fig. 6, lanes 9 and 10). An excess of un-labeled oligonucleotide E2F-WT did not compete with thecomplex III formed by the probe containing two methyl-cytosine sequences (Fig. 6, lane 8), indicating that the com-plex III resulted from interaction of proteins specificallywith the mCGmCG sequence.

Analysis of the Methylation Pattern of the FSHR CorePromoter after Treatment of the MSC-1 Cell Linewith 5-azaCdR

Bisulfite-DNA sequence analysis of the MSC-1 cells al-lowed the correlation of cytosine methylation with the en-dogenous FSHR gene inactivation during the formation ofa tumorigenic cell line. Treatment of the MSC-1 cells with5-azaCdR led to demethylation of the promoter and allowedthe direct correlation between demethylation and activationof transcription. The effective conditions for demethylationof genomic DNA from the MSC-1 cells was determinedafter a 12-day treatment with varying concentrations of 5-azaCdR. The results of the bisulfite sequencing of DNAfrom cells treated with the lowest concentrations of 5-azaCdR (0.5–1 mM) showed four cytosine residues at CpGdinucleotides in the core promoter region were partially

converted to thymine. Representative sequencing results forthe same three regions shown in Figure 2 for the rat arepresented (see color plate, Fig. 7). Even in cells treated with0.5 mM of 5-azaCdR, DNA conversion was detected mostoften in the E box motif, indicating that demethylation oc-curred more frequently at this site. When cells were cul-tured with 2 mM 5-azaCdR, all cytosine residues in thesequence were converted to thymine, indicating that the 5-azaCdR-treated gene changed all methylated cytosine res-idues into the demethylated state in these conditions (Fig.7). Although the final sequencing products were 330 bp ofthe mouse FSHR core promoter, a cytosine residue at nt2161 was not clearly identifiable under our conditions.

The expression level of the FSHR gene in the 5-azaCdR-treated cells was assayed by RT-PCR analysis before andafter treatment of MSC-1 cells with 5-azaCdR for 12 days.With PCR primers specific for the mouse FSHR cDNA se-quences, a 290-bp region of the mRNA was amplified andanalyzed by agarose gel electrophoresis. As shown in Fig-ure 8, the treatment of MSC-1 cells with 5-azaCdR causedthe FSHR gene expression to increase in a dose-dependentmanner (lanes 1–7). The FSHR mRNA was only detectableafter treatment of the MSC-1 cells with 1 and 2 mM 5-azaCdR and in the whole testis control.

DISCUSSION

Cell-Specific Methylation in the Core FSHR Promoter

Cell-specific expression of a gene can be regulated bythe cell-specific transcription factors interacting with thepromoter of the gene [6, 29]. The region of the FSHR genebetween 2383 and 11 bp, relative to the translational ini-tiation site, is essential for maximum promoter activity andhas been considered the core promoter [8]. This region ofthe core promoter contains two Inr-mediated transcriptionalstart sites, an E box, and an E2F consensus sequence. Even


FIG. 7. Genomic sequencing profile of bisulfite-converted DNA from MSC-1 cells treated with 5-azaCdR. Effect of 5-azaCdR on demethylation of theFSHR promoter of MSC-1 cells was examined by bisulfite genomic sequencing. The MSC-1 cells were cultured in the presence (0.5, 1.0, and 2.0 mM)or absence of 5-azaCdR for 12 days. Genomic DNA was isolated from the cells and subjected to bisulfite conversion reaction. Data are shown in thecolor plate, Figure 7. Note that the sequence of the mouse FSHR in these regions differs somewhat from the rat sequence. With no treatment all of theCpG sites were methylated, as the C residues in these positions remained unconverted by bisulfite (data not shown). Following treatment with 1 mMor less of 5-azaCdR, one CpG site was partially demethylated as indicated by the TpG sequence within the E box element. When the 2-mM concentrationof 5-azaCdR was used, all CpG sites were completely demethylated.

FIG. 8. Comparison of the FSHR gene expression of MSC-1 cells treatedwith 5-azaCdR. Total RNA was isolated from untreated cells or treatedcells with 5-azaCdR (1.0 and 2.0 mM) for 12 days. The RNA from mousetestis was used as a positive control (lanes 8 and 9). Expression of theFSHR mRNA was detected as a 300-bp PCR product by RT-PCR (lanes 5,7, and 9). The 2 and 1 symbols refer to the absence (2) or presence (1)of reverse transcriptase in the cDNA synthesis reaction. Relative molec-ular size of DNA fragments shown with a 100-bp ladder marker (GibcoBRL Life Technologies) (lane 10).

though some aspects of the mechanism of regulation of theFSHR gene have been investigated, there are no reports ofcell-specific DNA-protein complexes in the FSHR promot-er [7–10]. In addition, reporter gene assays driven by therat FSHR promoter have shown that this region could pro-mote the promiscuous transcription of the gene in severalcell lines in which the expression of the endogenous FSHRgene was originally repressed [10]. Lastly, a transgenic lineof mice was reported to have testis-specific expressionwhen a reporter gene was driven by 5 kb of the promotersequence [10].

In this study, we have shown that seven specific CpG

sites within the FSHR core promoter are methylated in cellsthat do not express the gene. In addition, the extent of de-methylation of this gene correlates directly with the levelof the FSHR expression in the tumor-derived Sertoli cellline MSC-1. These results indicate that DNA methylationis one of the major factors regulating Sertoli cell-specificexpression of the FSHR gene.

In mammals, DNA methylation patterns appear to playa critical role in terms of gene regulation in differentiatingand differentiated cells [11, 30, 31]. There is considerableevidence that a variety of cell-specific genes have distinc-tive DNA methylation patterns in expressing and nonex-pressing cells and undergo demethylation at the stage ofgene activation [11, 13, 31, 32].

DNA Methylation Can Inhibit TranscriptionFactor Binding

Several studies have shown that the inhibitory effects ofmethylation in a promoter can be the result of altered bind-ing of transcription factor(s) to the methylated binding mo-tif [33–35]. Recently, it has been shown that the silencingof the testis-specific Pdha-2 gene occurs via the selectivemethylation of a CpG dinucleotide within the ATF/CREbinding site [36]. Generally, under conditions where thereare low levels of CpG dinucleotides, cytosine methylationcan function as a mutation within the factor-binding site toprevent the binding of transcription factors [35, 37, 38].Selectively methylated site-specific elements in the pro-moters correspond to the inactive state of transcription in avariety of genes [35, 37–40]. A possible mechanism is thata single CpG site within the cis-acting regulatory elementsis sufficient to block the binding of the trans-acting factorsto the DNA [33, 35, 41–43]. A single cytosine methylationcould alter the local environment in the major groove ofDNA that serves as the site for the binding of transcriptionfactors [35, 44].

Within the 320-bp region of the rat FSHR core promoter,four of the seven potential 5-methylcytosine residues arelocated within known protein-DNA binding sites. Themethyl group(s) within cis-acting regulatory elements may


interrupt the binding of trans-acting factors to their targetsites on the promoter. It is reasonable to consider this in-hibition mechanism because the promoter region of theFSHR gene exhibits rather low G1C content (40%) and noextensive CpG-rich sequences. DNA methylation of the Ebox element (CACGTG, 2124/2119) in the FSHR pro-moter region inhibited the binding of nuclear extracts fromSertoli cells by 3.5-fold. In other studies, methylation ofthe E box motif has been shown to significantly decreasethe binding affinity of c-Myc/c-Myn [43, 45].

A methylated CpG dinucleotide within the Inr region ofthe FSHR gene in the nonexpressing tissues enhanced thebinding affinity of additional proteins that were specific forthe methylated sequence. Because the Inr element is a cen-tral site for the formation of the transcription initiation com-plex in the TATA-less promoter, the methylated DNA se-quence might recruit binding of a methyl-CpG-specific pro-tein and subsequently lead to the repression of the promoteractivity. This additional complex clearly showed the pres-ence of proteins within the nuclear extracts from Sertolicells with high affinity for the methylated CpG sequence.However, detection of the known methyl-binding proteinsin supershift experiments using two different polyclonal an-tibodies against MeCP2 provided by Dr. A. Bird was un-successful (data not shown). Optimal binding conditions forthe detection of the MeCP2 protein complex with the linearDNA sequence requires the DNA to be in the form of nu-cleosomes [46, 47]. The nature of the methylation-specificcomplexes in the Inr region is under investigation.

The E2F consensus sequence TTTCGCG, containing twoCpG dinucleotides (245/239) is located in the FSHR pro-moter. Kovesdi et al. have previously shown that the bindingof transcription factors to the E2F element was sensitive tomethylation at the second cytosine [48]. Previous studieshave shown that the methylcytosine binding protein, MeCP2,will bind at the methylated target sequence mCGmCG in theE2F element [49]. Our results showed that methylation ateither cytosine residue in this element had the same effectas a mutation, i.e., the affinity of protein binding to this sitewas decreased. Recently, the positive functional role of theE2F motif in the FSHR promoter has been characterizedwith transient transfection assays that show that the tran-scription factor E2F-1 binds to this consensus sequence butnot to the methylated sequence [50]. When the methylatedprobe containing -mCGmCG- dinucleotides was used inEMSA, a new DNA-protein complex was detected.

Recent evidence suggests that cytosine methylation canalso exert inhibitory effects through altering chromatinstructure and not by inhibition of the transcription machin-ery directly [14, 30, 51]. When CpG islands are located inthe promoter region, DNA methylation of this region canbe accompanied by changes in the chromatin structure forinactivation of a gene [52, 53]. Distinctive chromatin struc-ture on methylated DNA can be explained by displacementof histone H1 with MeCP2 [54]. The MeCP2 has recentlybeen shown to exist in a multiple repression complex inthe promoter region, providing strong evidence for the al-ternative inhibitory mechanism by DNA methylation [55–57]. The repressive chromatin structure results from the re-cruitment of histone deacetylases (HDACs) by MeCP2 innucleosomal DNA [47, 55].

Demethylation of MSC-1 Cell and Expressionof FSHR Gene

The mouse Sertoli cell line (MSC-1), derived from atesticular tumor, has an inactive FSHR promoter [25]. The

inactive state of transcription in MSC-1 cells correlates withthe cytosine methylation in the core promoter of the FSHRgene. The drug, 5-azaCdR is known to block DNA meth-ylation in newly replicated DNA molecules. It has beenused to activate silent genes in cells when the expressionof that gene is controlled by DNA methylation [58–60].Likewise, treatment of MSC-1 cells with 5-azaCdR resultedin the demethylation of the mouse FSHR promoter and re-activated the transcription of the FSHR gene.

Altogether, the data presented in this study support thenotion that the transcription of the rat or mouse FSHR geneis regulated by methylation or demethylation of CpG di-nucleotides in the proximal promoter. Although CpG di-nucleotides are not abundant in the rat or mouse FSHRpromoter, DNA methylation at very specific sites within thepromoter correlated with gene inactivation. Methylation ofDNA within the FSHR promoter may decrease the bindingof required transcription factors and/or may recruit methyl-specific binding proteins. The methyl-specific binding pro-teins could also directly interfere with transcription factorbinding or could recruit transcriptional silencers such asHDACs or negative cis-acting elements, [61]. Proteins thatspecifically recognize methylated DNA may also play in-dependent or interrelated roles in modulating gene activity.The maintenance of tissue specificity in gene expressionprobably requires multiple mechanisms, and our studiessupport the idea that DNA methylation plays a major rolein this process in the regulation of the rat or mouse FSHRgene. It is noteworthy that the proximal 231 bp of the pro-moter for the human FSHR gene, while 80% homologousto the promoter for the rat or mouse gene in this regionlacks any of the seven specific CpG sites and must be reg-ulated by some other mechanism [62].

REFERENCES1. Russell LD. Form, dimensions, and cytology of mammalian Sertoli

cells. In: Russell LD, Griswold MD (eds.), The Sertoli Cell. Clear-water, FL: Cache Revier Press; 1993: 1–38.

2. Griswold MD. Protein secretions of Sertoli cells. Int Rev Cytol 1988;110:133–156.

3. Zirkin BR, Awoniyi C, Griswold MD, Russell LD, Sharpe R. Is FSHrequired for adult spermatogenesis? J Androl 1994; 15:273–276.

4. Griswold MD. Action of FSH on mammalian Sertoli cells. In: RussellLD, Griswold MD (eds.), The Sertoli Cell. Clearwater, FL: CacheRevier Press; 1993: 493–508.

5. Kumar TR, Wang Y, Lu N, Matzuk MM. Follicle stimulating hormoneis required for ovarian follicle maturation but not male fertility. NatureGenet 1997; 15:201–204.

6. Roeder RG. The role of general initiation factors in transcription byRNA polymerase II. Trends Biochem Sci 1996; 21:327–334.

7. Heckert LL, Daley IJ, Griswold MD. Structural organization of thefollicle-stimulating hormone receptor gene. Mol Endocrinol 1992; 6:70–80.

8. Goetz TL, Lloyd TL, Griswold MD. Role of E box and initiator regionin the expression of the rat follicle-stimulating hormone receptor. JBiol Chem 1996; 271:33317–33324.

9. Heckert LL, Daggett MA, Chen J. Multiple promoter elements con-tribute to activity of the follicle-stimulating hormone receptor (FSHR)gene in testicular Sertoli cells. Mol Endocrinol 1998; 12:1499–1512.

10. Linder CC, Heckert LL, Goetz TL, Griswold MD. Follicle-stimulatinghormone receptor gene promoter activity. Endocrine 1994; 2:957–966.

11. Eden S, Cedar H. Role of DNA methylation in the regulation of tran-scription. Curr Opin Genet Dev 1994; 4:255–259.

12. Bird AP. Gene number, noise reduction and biological complexity.Trends Genet 1995; 11:94–100.

13. Beard C, Li E, Jaenisch R. Loss of methylation activates Xist in so-matic but not in embryonic cells. Cell 1995; 65:229–239.

14. Siegfried Z, Cedar H. DNA methylation: molecular lock. Curr Biol1997; 7:305–307.

15. Gruenbaum Y, Cedar H, Razin A. Substrate and sequence specificityof a eukaryotic DNA methylase. Nature 1982; 295:620–622.


16. Kafri T, Gao X, Razin A. Mechanistic aspects of genome-wide de-methylation in the preimplantation mouse embryo. Proc Natl AcadSci U S A 1993; 90:10558–10562.

17. Karl AF, Griswold MD. Sertoli cells of the testis: preparation of cul-ture and effect of retinoids. Methods Enzymol 1990; 190:71–75.

18. Frommer M, McDonald LE, Millar DS, Collis CM, Watt F, Grigg GW,Molly PL, Paul CL. A genomic sequencing protocol that yields apositive display of 5-methylcytosine residues in individual DNAstrands. Proc Natl Acad Sci U S A 1992; 89:1827–1831.

19. Blin N, Stafford DW. A general method for isolation of high molecularweight DNA from eukaryotes. Nucleic Acids Res 1976; 3:2303–2308.

20. Kerr LD. Electrophoretic mobility shift assay. Methods Enzymol1995; 254:619–632.

21. Dignam JD, Lebovitz RM, Roeder RG. Accurate transcription initia-tion by RNA polymerase II in a soluble extract from isolated mam-malian nuclei. Nucleic Acids Res 1983; 11:1475–1489.

22. Taylor SM, Jones PA. Mechanisms of action of eukaryotic DNAmethyltransferase: use of 5-azacytosine containing DNA. J Mol Biol1982; 162:679–692.

23. Kingston RE. Preparation and analysis of RNA. In: Ausubel FMEA(ed.), Current Protocols in Molecular Biology. New York: John Wileyand Sons, Inc.; 1997: 403.

24. Perchon JJ, Behringer RR, Cate RL, Harwood KA, Idzerda RL, Brins-ter R, Palmiter RD. Directed expression of an oncogene to Sertolicells in transgenic mice using inhibiting substance regulatory se-quence. Mol Endocrinol 1992; 6:1403–1411.

25. McGuinness MP, Linder CC, Morales CR, Heckert LL, Pikus J, Gris-wold MD. Relationship of a mouse Sertoli cell line (MSC-1) to normalSertoli cells. Biol Reprod 1994; 51:116–124.

26. Carcamo J, Buckbinder L, Reinberg D. The initiator directs the as-sembly of a transcription factor IID-dependent transcription complex.Proc Natl Acad Sci U S A 1991; 88:8052–8056.

27. Weis L, Reinberg D. Accurate positioning of RNA polymerase II ona natural TATA-less promoter is independent of TATA-binding-pro-tein-associated factors and initiator-binding proteins. Mol Cell Biol1997; 17:2973–2984.

28. Smale ST, Schmit MC, Berk AJ, Baltimore D. Transcriptional activationby Sp1 as directed through TATA or initiator: specific requirement formammalian transcription factor IID. Proc Natl Acad Sci U S A 1990;87:4509–4513.

29. Wasylyk B. Enhancers and transcription factors in the control of geneexpression. Biochim Biophys Acta 1988; 951:17–35.

30. Tate PH, Bird A. Effect of DNA methylation on DNA-binding pro-teins and gene expression. Curr Opin Genet Dev 1993; 3:226–231.

31. El-Deiry WS, Nelkin BD, Celano P. High expression of the DNAmethyltransferase gene characterizes human neoplastic cells and pro-gression stages of colon cancer. Proc Natl Acad Sci U S A 1991; 88:3470–3474.

32. Nickel J, Short ML, Schmitz A, Eggert M, Renkawitz R. Methylationof the mouse M-lysozyme downstream enhancer inhibits heterotetra-meric GABP binding. Nucleic Acids Res 1995; 23:4785–4792.

33. Comb M, Goodman HM. CpG methylation inhibits proenkephalingene expression and binding of the transcription factor AP-2. NucleicAcids Res 1990; 18:3975–3982.

34. Robertson KD, Hayward SD, Ling PD, Samid D, Ambinder RF. Tran-scriptional activation of the Epstein-Barr virus latency C promoterafter 5-azacytidine treatment: evidence that demethylation at a singleCpG site is crucial. Mol Cell Biol 1995; 15:6150–6159.

35. Watt F, Molloy PL. Cytosine methylation prevents binding to DNAof a HeLa cell transcription factor required for optimal expression ofthe adenovirus major late promoter. Genes Dev 1988; 2:1136–1143.

36. Iannello RC, Gould JD, Young JC, Giudice A, Medcalf R, Kola I.Methylation-dependent silencing of the testis-specific Pdha-2 basalpromoter occurs through selective targeting of an activating transcrip-tion factor/cAMP-responsive element-binding site. J Biol Chem 2000;275:19603–19608.

37. Iannello RC, Young J, Sumarsono S, Tymms MJ, Dahl HM, Gould J,Hedger M, Kola I. Regulation of Pdha-2 expression is mediated byproximal promoter sequences and CpG methylation. Mol Cell Biol1997; 17:612–619.

38. Okuse K, Matsuoka I, Kurihara K. Tissue-specific methylation occursin the essential promoter element of the tyrosine hydroxylase gene.Mol Brain Res 1997; 46:197–207.

39. Singal R, Ferris R, Little JA, Wang SZ, Ginder DG. Methylation ofthe minimal promoter of an embryonic globin gene silences transcrip-tion in primary erythroid cells. Proc Natl Acad Sci U S A 1997; 94:13724–13729.

40. Ngo V, Gourdji D, Laverriere J-N. Site-specific methylation of the ratprolacting and growth hormone promoters correlates with gene ex-pression. Mol Cell Biol 1996; 16:3245–3254.

41. Kovesdi I, Reichel R, Nevins JR. Identification of a cellular transcrip-tion factor involved in E1A trans-activation. Cell 1986; 45:219–228.

42. Iguchi-Ariga SM, Schaffner W. CpG methylation of the cAMP-re-sponsive enhancer/promoter sequence TGACGTCA abolishes specificfactor binding as well as transcriptional activation. Genes Dev 1989;3:612–619.

43. Prendergast GC, Lawe D, Ziff EB. Association of Myn, the murinehomolog of Max, with c-myc stimulates methylation-sensitive DNAbinding and ras cotransformation. Cell 1991; 65:395–407.

44. Fox KK. Effect of HhaI methylation on DNA local structure. BiochemJ 1986; 234:213–216.

45. Prendergast GC, Ziff RB. Methylation-sensitive sequence-specificDNA binding by the c-myc basic region. Science 1991; 251:186–189.

46. Nan X, Campoy J, Bird AP. MeCP2 is a transcriptional repressor withabundant binding sites in genomic chromatin. Cell 1997; 88:471–481.

47. Nan X, Ng H-H, Jonson CA, Laherty CD, Turner BM, Eisenman RN,Bird A. Transcriptional repression by the methyl-CpG-binding proteinMeCP2 involves a histone deacetylase complex. Nature 1998; 393:386–389.

48. Kovesdi I, Reichel R, Nevins JR. Role of an adenovirus E2 promoterbinding factor in E1A-mediated coordinate gene control. Proc NatlAcad Sci U S A 1987; 84:2180–2184.

49. Di Fiore B, Palena A, Felsani A, Palitti F, Caruso M, Lavia P. Cytosinemethylation transforms an E2F site in the retinoblastoma gene pro-moter into a binding site for the general repressor methylcytosine-binding protein 2 (MeCP2). Nucleic Acids Res 1999; 27:2852–2859.

50. Kim J-S, Griswold MD. E2F and GATA-1 are required in the tran-scriptional regulation of the follicle-stimulating hormone receptor(FSHR) promoter. 2000; (in press).

51. Richardson R, Yung R. Role of DNA methylation in the regulationofcell function. J Lab Clin Med 1999; 134:333–340.

52. Ng HH, Bird A. DNA methylation and chromatin structure. Curr OpinGenet Dev 1999; 9:158–163.

53. Pfeifer GP, Tanguay RL, Steigerwald SD, Riggs AD. In vivo footprintand methylation analysis by PCR-aided genomic sequencing: com-parison of active and inactive X chromosomal DNA at the CpG islandand promoter of human PGK-1. Genes Dev 1990; 4:1277–1287.

54. Kass SU, Pruss D, Wolffe AP. How does DNA methylation represstranscription? Trends Genet 1997; 13:444–449.

55. Jones PL, Veenstra GJ, Wade PA, Vermaak D, Kass SU, LandsbergerN, Strouboulis J, Wolffe AP. Methylated DNA and MeCP2 recruithistone deacetylase to repress transcription. Nat Genet 1998; 19:187–190.

56. Chandler SP, Guschin D, Landsberger N, Wolffe AP. The methyl-CpGbinding transcriptional repressor MeCP2 stably associates with nucle-osomal DNA. Biochemistry 1999; 38:7008–7018.

57. Coffee B, Zhang F, Warren ST, Reines D. Acetylated histones areassociated with FMR1 in normal but not fragile X-syndrome cells.Nat Genet 1999; 22:98–101.

58. Ottavianon YL, Issa J-P, Parl FF, Smith HS, Baylin SB, Davidson NE.Methylation of the estrogen receptor gene CpG island marks loss ofestrogen receptor expression in human breast cancer cells. Cancer Res1994; 54:2552–2555.

59. Bovenzi V, Le NL, Cote S, Sinnett D, Momparler LF, Momparler RL.DNA methylation of retinoic acid receptor beta in breast cancer andpossible therapeutic role of 5-aza-29-deoxcytidine. Anticancer Drugs1999; 10:471–476.

60. Bender CM, Pao MM, Jones PA. Inhibition of DNA methylation by5-aza-29-deoxycytidine suppresses the growth of human tumor celllines. Cancer Res 1998; 58:95–101.

61. Mizuno K, Goto M, Masamune Y, Nakanishi Y. A silencer-like ciselement for the testis-specific phosphoglycerate-kinase-2-encodinggene. Gene 1992; 119:293–297.

62. Gromoll J, Dankbar B, Gudermann T. Characterization of the 59 flank-ing region of the human follicle-stimulating hormone receptor gene.Mol Cell Endocrinol 1994; 102:93–102.

site-specific methylation of the promoter alters deoxyribonucleic acid-protein interactions and...

Documents