DNA Methylation Is Required for Silencing of Ant4 , an Adenine Nucleotide Translocase Selectively Expressed in Mouse Embryonic Stem Cells and Germ Cells

Original Article

Stem Cells 2005;23:1314–1323 www.StemCells.com

DNA Methylation Is Required for Silencing of Ant4, an Adenine

Nucleotide Translocase Selectively Expressed in Mouse Embryonic

Stem Cells and Germ Cells

Nemanja RodiĆ,a Masahiro Oka,a Takashi Hamazaki,a Matthew R. Murawski,a Marda Jorgensen,b Danielle M. Maatouk,c James L. Resnick,c En Li,d Naohiro Teradaa,b

aDepartment of Pathology, bShands Cancer Center, cDepartment of Molecular Genetics and Microbiology,

University of Florida College of Medicine, Gainesville, Florida, USA; dCardiovascular Research Center,

Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts, USA

Key Words. Embryonic stem cell • DNA methylation • Differentiation • Germ cellAdenine nucleotide translocase • Gene repression

Correspondence: Naohiro Terada, M.D., Ph.D., Department of Pathology, University of Florida College of Medicine, 1600 SW Archer Road, Gainesville, Florida 32610, USA. Telephone: 352-392-2696; Fax: 352-392-6249; e-mail: [email protected] Received March 16, 2005; accepted for publication May 10, 2005; first published online in Stem Cells EXPRESS July 28, 2005. ©AlphaMed Press 1066-5099/2005/$12.00/0 doi: 10.1634/stemcells.2005-0119

AbstractThe capacity for cellular differentiation is governed not only by the repertoire of available transcription factors but by the accessibility of cis-regulatory elements. Studying changes in epigenetic modifications during stem cell differentiation will help us understand how cells maintain or lose differentiation potential. We investigated changes in DNA methylation dur-ing the transition of pluripotent embryonic stem cells (ESCs) into differentiated cell types. Using a methylation-sensitive restriction fingerprinting method, we identified a novel ade-nine nucleotide (ADP/ATP) translocase gene, Ant4, that was selectively hypomethylated and expressed in undifferentiated mouse ESCs. In contrast to other pluripotent stem cell–spe-cific genes such as Oct-4 and Nanog, the Ant4 gene was readily derepressed in differentiated cells after 5-aza-2'-deoxycyti-

dine treatment. Moreover, expression of de novo DNA meth-yltransferases Dnmt3a and Dnmt3b was essential for repres-sion and DNA methylation of the Ant4 gene during ESC differ-entiation. Although the deduced amino acid sequence of Ant4 is highly homologous to the previously identified Ant isoforms, the expression of Ant4 was uniquely restricted to developing gametes in adult mice, and its promoter hypomethylation was observed only in testis. Additionally, Ant4 was expressed in primordial germ cells. These data indicate that Ant4 is a plu-ripotent stem cell– and germ cell–specific isoform of adenine nucleotide translocase in mouse and that DNA methylation plays a primary role in its transcriptional silencing in somatic cells. Stem Cells 2005;23:1314–1323

IntroductionDNA methylation, or the addition of a methyl group to the 5'-

position of cytosine within the CpG dinucleotide, is a heri-

table modification that contributes to gene silencing. Most

CpG sites in mammalian cells are methylated in a nonrandom

fashion. For instance, repetitive and parasitic elements tend to

be hypermethylated, whereas CpG island–associated promot-

ers are usually hypomethylated [1]. The complex process of

DNA methylation has been proven essential for normal devel-

opment [2], X-chromosome inactivation [3], imprinting [4],

This material is protected by U.S. Copyright law. Unauthorized reproduction is prohibited.

For reprints contact: [email protected]

Rodić, Oka, Hamazaki et al. 1315

and the suppression of parasitic DNA sequences [5]. Although

DNA methylation is a modification that promotes genomic

integrity and ensures proper temporal and spatial gene expres-

sion during development, it also associates with malignancy

when aberrantly controlled. Both hypermethylation and

hypomethylation have been attributed to cancer development

[6, 7]. DNA methylation is proposed to prevent the binding

of transcription factors and recruit repressor complexes that

induce the formation of inactive chromatin complexes [8].

For example, it was recently shown that methylation of a CpG

dinucleotide within the glial fibrillary acidic protein promoter

prevents STAT3 binding [9]. In another instance, DNA meth-

ylation–mediated gene silencing is dependent on the pres-

ence of methyl-CpG binding protein, MeCP2, which forms

a complex with histone deacetylases and a repressor protein,

mSin3A, to repress transcription in a methylation-dependent

manner [10]. It is still unclear in most cases, however, whether

DNA methylation is a causal event in gene silencing or, rather,

a consequence of gene silencing.

It is becoming increasingly clear that epigenetic modifica-

tions play a critical role in the regulation of gene expression in

many cellular processes [8, 11]. Studying changes in epigenetic

modifications during stem cell differentiation will help us under-

stand how cells maintain or lose differentiation potential. In the

present study, we attempted to identify differentially methylated

loci that are hypomethylated in undifferentiated embryonic stem

cells (ESCs) and become hypermethylated after differentiation.

Murine ESCs are originally derived from the inner cell mass of

a developing blastocyst and have the ability to differentiate into

all cell types of an adult animal [12]. Pluripotency of ESCs can be

maintained in vitro when the cells are cultured in a serum-con-

taining medium supplemented with leukemia inhibitory factor

(LIF) [13]. When LIF is removed from the medium, the ESCs

begin to differentiate in vitro into all three embryonic germ lay-

ers. This in vitro ESC differentiation system serves as an excel-

lent model to study the regulation of gene expression required for

stem cell self-renewal and pluripotency [14–16]. Recent studies

on molecules involved in epigenetic modifications have revealed

a unique expression pattern of DNA methyltransferases [17], his-

tone deacetylases [18], and methyl-binding proteins [19] in ESCs.

ESCs also have a differential genome-wide DNA methylation

pattern compared with their descendant differentiated cells [20,

21]. However, exact genomic loci of such differentially methyl-

ated regions remain unknown.

Using methylation-sensitive restriction fingerprinting

(MSRF), we identified a novel gene encoding an adenine nucle-

otide (ADP/ATP) translocase homologue that is specifically

expressed in undifferentiated ESCs and germ cells. Further-

more, we show that DNA methylation, but not the availability

of transcription factors, is the dominant factor restricting the

gene’s expression.

Materials and Methods

In Vitro ESC DifferentiationMouse ESC lines R1, J1, Dnmt3a-null, Dnmt3b-null, and

Dnmt3a-Dnmt3b double-null ESCs [22] were maintained on gel-

atin-coated tissue culture dishes in Dulbecco’s modified Eagle’s

medium optimized for ESCs (Invitrogen, Carlsbad, CA, http://

www.invitrogen.com) containing 1,000 U/ml recombinant mouse

LIF (ESGRO; Chemicon, Temecula, CA, http://www.chemicon.

com), 10% knockout serum replacement (Invitrogen), 1% fetal

calf serum (Atlanta Biologicals, Lawrenceville, GA, http://www.

atlantabio.com), 20 mM HEPES (Invitrogen), 300 μM monothio-

glycerol (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.

com), 100 U/ml penicillin, 100 μg/ml streptomycin (Invitrogen),

and 2 mM L-glutamine (Invitrogen). In vitro ESC differentiation

was induced in an Iscove’s modified Dulbecco’s medium (Invit-

rogen) containing 20% fetal calf serum, 2 mM L-glutamine, 100

U/ml penicillin, 100 μg/ml streptomycin, and 300 μM monothio-

glycerol. The embryoid body (EB) formation was induced in a

hanging drop containing approximately 2,000 cells in differentia-

tion medium for 2 days as described previously [23].

MSRFMSRF was performed according to the original method estab-

lished by Huang et al. [24]. Briefly, ESCs and EBs were harvested

by gentle cell scraping followed by a 5-minute centrifugation in a

bench-top centrifuge. Genomic DNA was extracted using Wizard

Genomic DNA Purification Kit (Promega, Madison, WI, http://

www.promega.com). Extracted DNA was digested with a methyla-

tion-insensitive restriction enzyme (MseI) either by itself or in com-

bination with a methylation-sensitive restriction enzyme, BstUI, at

10 U per 1 μg of DNA for each enzyme. Digestion with BstUI was

performed for 2 hours at 60°C, followed by overnight incubation

with MseI at 37°C. The polymerase chain reaction (PCR) was per-

formed in 20-μl reaction mixtures containing 2 μl of digested DNA

(50–100 ng), 0.4 μM of primers, 1.25 U of HotMasterTaq DNA

polymerase (Eppendorf, Hamburg, Germany), 200 μM deoxy-

nucleotide triphosphates, 1 mCi/μl [α-32P] dCTP (3,000 Ci/mmol,

Amersham, Piscataway, NJ), and 2 μl of ×10 reaction buffer. The

primers used in this study were Bs-1 (5'-AGCGGCCGCG), Bs-2

(5'-GCCCCCGCGA), Bs-3 (5'-CGGGGCGCGA), and Bs-4 (5'-

ACCCCACCCG). The PCR reaction consisted of an initial dena-

turing step for 5 minutes at 94°C and 30 cycles of the following: 2

minutes at 94°C, 1 minute at 40°C, and 2 minutes at 72°C. A final

step at 72°C was 8 minutes. After PCR amplification, 5 μl of each

sample was mixed with 1 μl of ×6 loading dye solution. Each sample

was then separated in a 4.5% nondenaturing polyacrylamide gel.

All reactions were run as duplicates to account for loading error.

Wet gels were laid on 3M-filter paper, wrapped with plastic wrap,

and exposed to Kodak X-OMAT LS film (Kodak, New Haven, CT,

http://www.kodak.com) for 24–48 hours at −80°C.

1316 DNA Methylation–Dependent Repression of Ant4

Cloning and SequencingDNA segments of interest were excised from polyacrylamide

gels using a sterile scalpel. DNA was eluted by incubation of

gel fragments in 50 μl of sterile deionized water for 10 minutes

at 100°C. Eluted DNA was reamplified by the identical prim-

ers and PCR conditions as used in the MSRF method. Ream-

plified DNA fragments were excised from the gel, ligated into

a TA-cloning vector using pCRII-Topo Cloning Kit (Invit-

rogen), and sequenced. To determine the identity of resulting

DNA sequences, searches were performed against the follow-

ing mouse genomic databases: http://www.ensambl.org, http://

genome.ucsc.edu, and http://www.ncbi.nlm.nih.gov/genome/

seq/MmBlast.html. All of the recombinant DNA experiments

here and below were performed under the National Institutes of

Health guidelines.

Bisulfite Sequencing and Combined Bisulfite Restriction AnalysisDNA was extracted from ESCs, EBs, and adult tissues using

the DNA Wizard Genomic DNA Purification Kit (Promega).

Mouse testes were decapsulated, and seminiferous tubules

were collected for analysis. A bisulfite reaction was performed

using EZ DNA Methylation Kit (Zymo Research, Orange,

CA, http://www.zymoresearch.com). Up to 2 μg of genomic

DNA was used for conversion with the bisulfite reagent.

Approximately 80 ng of bisulfite-converted DNA was used as

a template for each PCR analysis. Primers used for Ant4 com-

bined bisulfite restriction analysis (COBRA) and for bisulfite

sequencing were 5'-TTGTTGTGTATTGATTGAGTATG

and 5'-ACACTAAAAAAAACTAAAAAACC (40 cycles).

Primers used for bisulfite sequencing were 5'-AAGGTGTGT-

GTTATTATTGTTTGATGT and 5'-TACCCCTCATCTAT-

CATATTCCCTA; 5'-TTGAGGAGGAGTTAATAAGTT-

TAGGG and 5'-CCAAAAAACACACTCTAACCAAATAC;

5'-GTAGGTAAATTAATTGTGGATTAAATAGTA and

5'-TACACAACAACTTTTACAAAAAAAC; 5'-AGTTGTTGT-

GTATTGATTGAGTATG and 5'-TCTTTAAAAACTACTTCT-

TAAAAAATTC; and 5'-TTTTTAAGAAGTAGTTTTTA-

AAGAAGG and 5'-AAACAATCCAACATACCCTTATAAC.

PCR fragments were cloned into pCRII-TOPO cloning vector

(Invitrogen), and individual clones were sequenced.

Reverse Transcription–Polymerase Chain ReactionTotal RNA was isolated from ESCs and EBs using RNAque-

ous kit (Ambion, Austin, TX, http://www.ambion.com). Two

micrograms of RNA was used as a template for reverse tran-

scription (RT) reactions using SuperScript Synthesis for First

Strand Synthesis Kit (Invitrogen). Sequences of forward and

reverse primer pairs were as follows: Ant4 (5'-TGGAGCAA-

CATCCTTGTGTG, 5'-AGAAATGGGGTTTCCTTTGG),

Oct-4 (5'-TGGAGACTTTGCAGCCTGAG, 5'-TGAATGCAT-

GGGAGAGCCCA), Nanog (5'-AGGGTCTGCTACTGAGA-

TGCTCTG, 5'-CAACCACTGGTTTTTCTGCCACCG), Ttr

(5'-CTCACCACAGATGAGAAG, 5'-CCGTGAGTCTCT-

CAATTC), Fgf-5 (5'-AAAGTCAATGGCTCCCACGAA, 5'-

CTTCAGTCTGTACTTCACTGG), β-actin (5'-TTCCTTCTT-

GGGTATGGAAT, 5'-GAGCAATGATCTTGATCTTC),

Gapdh (5'-CCCTTCATTGACCTCACTACATGG, 5'-CCT-

GCTTCACCACCTTCTTGATGTC), and Hprt (5'-GCTGGT-

GAAAAGGACCTCT, 5'-CACAGG ACTAGAACACCTGC).

Northern BlottingNorthern blotting was performed using multiple premade blots

(Clontech, Palo Alto, CA, http://www.clontech.com). Briefly,

the membrane was preincubated with 5 ml of ExpressHyb

Hybridization Solution at 68°C for 1 hour. Seventy-five nano-

grams of the full-length Ant4 or β-actin cDNA fragment was

radio-labeled with [α-32P] dCTP nucleotides. Ant1 and Ant2

probes were PCR-amplified using the following primers: Ant1

(5'-GGCGCTACTTTGCTGGTAAC, 5'-GCAATCTTCCTC-

CAGCAGTC), Ant2 (5'-CAGCTGGATGATTGCACAGT,

5'-CAAGCCCAGAGAATCTGTCC). Unincorporated nucleo-

tides were removed using ProbeQuant G-50 Micro Column

(Amersham Biosciences, Piscataway, NJ, http://www.amer-

sham.com). Heat-denatured and briefly chilled probe (~20

ng/ml) was added to the hybridization solution. The membrane

was incubated with gentle shaking for 24 hours at 68°C and

washed twice for half an hour per wash in ×2 standard saline

citrate (SSC)/0.05% SDS at room temperature and twice in

×0.1 SSC/0.1% SDS at 50°C. The membrane was wrapped with

plastic wrap and exposed to X-OMAT Kodak film at −80°C for

18–24 hours.

ImmunohistochemistryRabbit polyclonal antibodies were raised against mouse Ant4

peptide (1-MSNESSKKQSSKKALFD-17) and purif ied

through an affinity column using the same peptide (Sigma

Genosys, The Woodlands, TX, http://www.sigma-genosys.

com/index.asp). Mouse ovary, testis, and liver were harvested

from 3-month-old BALB/c mice and immediately frozen into

optimal cutting temperature blocks (Tissue-Tek, Torrance, CA,

http://www.sakura.com). Frozen sections were cut at 5 μm,

air-dried for 2 hours, and then fixed in acetone for 10 minutes

before being rehydrated and blocked for endogenous peroxidase

activity. After a serum-blocking step, affinity-purified rabbit

polyclonal anti-Ant4 antibodies were applied at 2 μg/ml and

incubated overnight at 4°C. Staining was achieved using the

EnVision+ HRP kit (DakoCytomation, Glostrup, Denmark,

http://www.dakocytomation.com) following the manufactur-

er’s directions. Positive signal was detected with DAB+ (Dako-

Cytomation), and slides were counterstained using Light Green

SF Yellowish (Sigma-Aldrich).

Rodić, Oka, Hamazaki et al. 1317

Primordial Germ Cell PreparationPrimordial germ cells (PGCs) were obtained from timed matings

of B6C3F1 mice purchased from Jackson Laboratory (Bar Har-

bor, ME, http://www.jax.org). Noon of the day on which a mating

plug was first observed was taken to be 0.5 days post coitus (dpc).

PGCs were immunomagnetically purified using anti-SSEA1

antibody from isolated urogenital ridges as described [25]. Geni-

tal ridges at 12.5 dpc were sex-separated by visual inspection

for testis cords. The immunodepleted fraction refers to cells not

retained on the magnetic column.

Results

Identification of a CpG-Rich DNA Sequence That Undergoes De Novo DNA Methylation During ESC DifferentiationWe investigated changes in DNA methylation during the transi-

tion of ESCs into differentiated cell types using an MSRF method.

Mouse ESCs were differentiated in a culture medium without LIF

using a hanging drop method. DNA was extracted from undif-

ferentiated ESCs (day 0) and differentiated EBs (days 5 and

10). The DNA was then digested by a combination of methyla-

tion-sensitive (BstUI) and insensitive (MseI) restriction enzymes

and amplified by PCR using CpG-rich 10-mer primers (Fig. 1A).

Using a combination of four different CpG-rich primers, we iden-

tified a total of eight bands that exhibited differential methylation

patterns during ESC differentiation (data not shown). The bands

were excised from the gels, and their DNA sequences were deter-

mined. Database searches revealed that one of these methylated

fragments (methylated fragment 1, MF1 in Figs. 1B, 1C) mapped

to the exon 1/intron 1 boundary of a previously uncharacterized

gene (RIKEN cDNA 1700034J06). The other seven DNA frag-

ments were localized outside of any CpG islands and were derived

from various genomic regions. Because the MF1 DNA fragment

mapped to a genomic region that partially overlapped with a CpG

island, we decided to further analyze the associated gene.

The associated gene on mouse chromosome 13 did not

contain a classic TATA box sequence at its 5'-flanking region.

Because TATA-less genes are often associated with multiple

transcriptional initiation sites, we performed a 5'RACE (rapid

amplification of cDNA ends) reaction using cDNA isolated from

undifferentiated ESCs. 5'RACE identified five different tran-

scriptional initiation sites, four of which were previously reported

in Expressed Sequence Tag (EST) databases (data not shown).

Because there are multiple transcriptional initiation sites, nucleo-

tide positions are marked relative to the translation initiation site

(+1) throughout the manuscript. The major transcript initiated at

−60 bp, whereas the longest transcript initiated 304 bp upstream

of the translation initiation site.

The MF1-associated gene was predicted to encode a 320-amino-

acid protein and shared high amino acid sequence homology with

the other mouse adenine nucleotide translocase (Ant) proteins pre-

viously identified (70.1% and 69.1% overall amino acid identity to

Ant1 and Ant2 isoform, respectively) (Fig. 2A). The MF1-associated

Figure 1. Identification of differentially methylated CpG-rich frag-

ment in ES cells and EBs. (A): MSRF. Genomic DNA was prepared

from undifferentiated ES cells and differentiating EBs (days 5 and

10). DNA was digested either by MseI alone or MseI and BstUI and

subjected to PCR amplification using CpG-rich 10-mer primers in the

presence of radiolabeled dCTP. Amplified DNA fragments were sep-

arated in 4.5% polyacrylamide gels. In MseI/BstUI double-digested

samples, the intensity of a band indicated by arrowhead (MF1)

increased after ES cell differentiation. (B): Nucleotide sequence of

methylated fragment 1 (MF1). The DNA band was cloned into a PCR

cloning vector and sequenced. Bold letters indicate primers used,

whereas BstUI sites are underlined. MF1 corresponds to the exon

1/intron 1 boundary region of a previously uncharacterized gene on

mouse chromosome 13; black solid boxes represent exons predicted

by Expressed Sequence Tag and Ensemble mouse genome databases.

The translation initiation site is marked as +1. The arrow represents the

major transcription initiation site (−60 bp). Abbreviations: EB, embry-

oid body; ESC, embryonic stem cell; MseI, methylation-insensitive

restriction enzyme; MSRF, methylation-sensitive restriction finger-

printing; PCR, polymerase chain reaction.

1318 DNA Methylation–Dependent Repression of Ant4

gene also contained three tandem repeats of a domain of approxi-

mately 100 residues, each domain containing two transmembrane

regions, a characteristic shared by all members of the Ant family

[26]. Because another isoform of Ant, ANT3, has been reported in

human [27], we tentatively named the present gene adenine nucleo-

tide translocase 4 (Ant4). All known mammalian Ant isoforms

(unlike plant adenine nucleotide translocases) lack an N-terminal

mitochondrial localization sequence, yet they localize to mitochon-

dria [28]. In agreement with this observation, Ant4 also does not

contain a classic mitochondrial localization sequence. However, it

did localize to mitochondria when N-terminal FLAG-tagged Ant4

was expressed in NIH3T3 fibroblast cells (data not shown).

The human orthologue of ANT4 is located on chromosome

4q28.1, and its deduced amino acid sequence shares 85.9% over-

all amino acid identity with mouse Ant4 (Fig. 2B). In addition,

the genomic architecture with six exons is conserved between

mouse and human (data not shown). It is notable that the human

genome contains at least seven ANT pseudogenes on the X chro-

mosome; however, in contrast to ANT4, such pseudogenes do

not reveal conserved exon/intron architecture or translatable

coding regions [29].

Ant4 Promoter Locus Undergoes De Novo DNA Methylation After ESC DifferentiationTo determine DNA methylation patterns across the Ant4 pro-

moter locus during ESC differentiation, genomic DNAs from

ESCs or day-10 EBs were treated with bisulfite and subjected

to sequence analysis. We analyzed the Ant4 promoter region

that encompasses a total of 47 CpG dinucleotides, from 516

bp upstream of the translation initiation site to the 3'-end of

exon 1. Sequencing of individual bisulfite-converted genomic

DNAs revealed that the Ant4 promoter and associated CpG

island region were mostly unmethylated in undifferentiated

ESCs (Fig. 3). In contrast, EBs at day 10 showed significant

hypermethylation around the Ant4 promoter region, indicating

that after ESC differentiation, the Ant4 promoter undergoes de

novo DNA methylation. The further upstream region (>1 kb) of

the Ant4 promoter outside of the CpG island revealed hyper-

methylation regardless of the differentiation status of ESCs

(Fig. 3). Ant4 promoter methylation was confirmed by the

quantitative method of COBRA [30]. The frequency of methyl-

ation at a specific CpG site overlapping an HhaI restriction site

(−169 bp) was evaluated by the COBRA assay. In agreement

with the bisulfite sequencing data, we detected low levels of

DNA methylation in ESCs and high methylation levels (~75%)

in EBs at day 10.

Ant4 mRNA Expression Is Downregulated After ESC DifferentiationWe then examined Ant4 mRNA expression levels in ESCs and

EBs using semiquantitative RT-PCR. Ant4 mRNA was easily

detectable in undifferentiated ESCs, whereas the relative abun-

dance of Ant4 mRNA decreased after ESC differentiation (Fig.

4A). β-Actin expression was relatively unchanged during ESC dif-

ferentiation. We then used the DNA demethylating agent 5-aza-2'-

deoxycytidine (5-aza-dC) to study the effect of demethylation on

transcription of the Ant4 gene. 5-aza-dC induces DNA demeth-

ylation in cells by depleting DNA methyltransferases through its

covalent and irreversible binding to the enzymes. Notably, the

Ant4 gene was readily derepressed by the addition of 5-aza-dC

to differentiated EBs (day 10) (Fig. 4B). In contrast, expression

of Oct-4 and Nanog, transcription factors selectively expressed

in pluripotent ESCs and primordial germ cells, was not affected.

Ant4 (but not Oct-4 or Nanog) was similarly derepressed by addi-

tion of 5-aza-dC in NIH3T3 cells and an immortalized endoderm

precursor cell line (OBAT, unpublished data).

DNA Methylation of Ant4 Gene During ESC Differentiation Requires Dnmt3DNA methyltransferase 3a and 3b (Dnmt3a and Dnmt3b) has

been proposed to play a primary role in de novo methylation dur-

ing murine embryonic development [31]. To investigate the role of

the Dnmt3 in the establishment of DNA methylation of the Ant4

Figure 2. Ant4 encodes a novel isoform of adenine nucleotide trans-

locase. (A): Deduced amino acid sequence of the mouse Ant4 gene

is aligned with previously identified mouse Ant proteins (Ant1 and

Ant2). (B): Deduced amino acid sequence of the mouse Ant4 gene is

aligned with ANT4 human orthologue.

Rodić, Oka, Hamazaki et al. 1319

promoter during ESC differentiation, we used ESCs homozy-

gously deleted for both the Dnmt3a and Dnmt3b genes. Dnmt3a-

Dnmt3b double-null ESCs (earlier passage cell lines) were sub-

jected to the same differentiation protocol as parental (wild-type)

J1 ESCs. Of interest, Ant4 expression was not suppressed but

rather increased in differentiated double-null cells (Fig. 5A). In

contrast, the pluripotent ESC marker Oct-4 was downregulated,

whereas the primitive ectoderm marker fibroblast growth factor,

Fgf-5, and endoderm marker transthyretin, Ttr, were upregulated

after ESC differentiation. These results indicate that Dnmt3a-

Dnmt3b double-null ESCs proceed with a normal differentiation

pattern. COBRA assays confirmed that there was an absence of

DNA methylation at the Ant4 promoter region (−169 bp) in the

double knockout ESCs (Fig. 5B). These results indicate that func-

tional Dnmt3 is required for repression of the Ant4 gene during

ESC differentiation.

In addition, we examined which member of the Dnmt3

gene family, Dnmt3a or Dnmt3b, played a predominant role in

Ant4 methylation using ESCs deleting either gene. As shown in

Figure 5C, downregulation of Ant4 expression was incomplete

when Dnmt3a-null ESCs or Dnmt3b-null ESCs were differenti-

ated. Furthermore, both Dnmt3a-null ESCs and Dnmt3b-null

ESCs demonstrated virtually no DNA methylation on the CpG

site assayed by COBRA. These data indicated that both Dnmt3a

and Dnmt3b are required for DNA methylation and repression

of Ant4.

Figure 4. Ant4 is repressed during ES cell differentiation. (A): Total

RNA was extracted from undifferentiated ES cells and differentiat-

ing EBs at days 5 and 10. RNA expression levels of Ant4 and β-actin

were evaluated by semiquantitative RT-PCR analysis using various

cycles of DNA amplification (20 to 30 cycles). (B): Ant4 derepres-

sion by 5-aza-dC. EBs were treated with various concentrations

(0–10 μM) of 5-aza-dC for 48 hours from day 8 and harvested at day

10. RNA expression levels of Ant4, Oct-4, Nanog, and β-actin genes

were examined by RT-PCR. Abbreviations: 5-aza-dC, 5-aza-2'-deox-

ycytidine; EB, embryoid body; ESC, embryonic stem cell; RT-PCR,

reverse transcription–polymerase chain reaction.

Figure 3. Ant4 promoter region undergoes DNA methylation dur-

ing ES cell differentiation. Summary of Ant4 methylation levels in

ES cells and EBs (day 10) is illustrated. For each primer pair, up to

14 clones were sequenced after bisulfite treatment of either ES cells

or EBs (day 10) to determine the rate of DNA methylation. Closed

circles represent methylated CpG, whereas open circles represent

unmethylated CpG. The y-axis of the graph represents percent meth-

ylation for each CpG residue. Diagram at the bottom represents rela-

tive position of CpG residues within the 5'-end regulatory region of

Ant4 gene. Nucleotide positions are marked relative to the translation

initiation site (+1). Arrow represents a major transcription start site

(−60 bp), black rectangle represents the first exon, and vertical lines

mark the location of individual CpG residues. Abbreviations: EB,

embryoid body; ESC, embryonic stem cell.

1320 DNA Methylation–Dependent Repression of Ant4

Hypomethylation of the Ant4 Promoter and Ant4 Expression Are Restricted to Testicular Germ CellsTo determine if Ant4 is specifically expressed in ESCs, we ini-

tially used a Basic Local Alignment Search Tool (BLAST) search

against EST databases. We used full-length Ant4 cDNA as a query

sequence and found a total of eight cDNA clones (with scores of

>200 bits). All clones identified were from testis. We then inves-

tigated expression patterns of Ant4 in adult mouse organs using

Northern blot analysis. Ant4 mRNA was found specifically in tes-

tes, at the predicted approximately 1.6-kb transcript size, but was

Figure 5. Dnmt3 is required for Ant4 repression during ESC dif-

ferentiation. (A): Expression of the Ant4 gene in WT ES cells and

Dnmt3a-Dnmt3b double-null ESCs. WT J1 ES cells and Dnmt3a-

Dnmt3b double-null ESCs (Dnmt3a−/−, Dnmt3b−/−) were differ-

entiated as described above. RNA expression was evaluated using

RT-PCR as described above. Oct-4, Ttr, and Fgf-5 are markers for

undifferentiated ESCs, early visceral endoderm, and early ecto-

derm, respectively. (B): Ant4 promoter DNA methylation deter-

mined by COBRA assay. DNA was extracted from WT ESCs and

Dnmt3a-Dnmt3b-null ESCs (Dnmt3a−/−, Dnmt3b−/−) and treated

with bisulfite. The Ant4 promoter region was amplified and sub-

jected to overnight digestion with HhaI restriction enzyme, which

cuts GCGC sites at −169 bp. DNA methylation of the CpG protects

the site from bisulfite conversion; thus, the polymerase chain reac-

tion fragments are digested by HhaI only when the template genomic

DNA is methylated at the site. The digested DNA samples were sepa-

rated in 4.5% polyacrylamide gels and visualized using a SyBr-green

dye. (C): Expression of the Ant4 gene and Ant4 promoter methyla-

tion in Dnmt3a-null ESCs and Dnmt3b-null ESCs. Dnmt3a-null

ESCs (Dnmt3a−/−) and Dnmt3b-null ESCs (Dnmt3b−/−) were differ-

entiated, and Ant4 expression and DNA methylation of the Ant4 pro-

moter were examined as described above. Abbreviations: COBRA,

combined bisulfite restriction analysis; EB, embryoid body; ESC,

embryonic stem cell; RT-PCR, reverse transcription–polymerase

chain reaction; WT, wild-type.

Figure 6. Ant4 expression and promoter DNA methylation in adult

organs. (A): Northern blot analysis of Ant4 mRNA expression in vari-

ous organs from adult mice (8 weeks old). The blot was hybridized to

specific cDNA probes for Ant4, Ant1, Ant2, and β-actin. (B): Immu-

nohistochemical analysis of Ant4 expression in testis, ovary, and liver.

Mouse ovary, testis, and liver were harvested from 3-month-old mice,

and frozen sections were stained with affinity-purified rabbit poly-

clonal anti-Ant4 antibodies and visualized using horse radish peroxi-

dase (brown). Slides were counterstained using Light Green SF Yel-

lowish. (C): Bisulfite analysis of the Ant4 promoter in various organs

from adult mice (8 weeks old). DNA samples were extracted from the

indicated mice organs and subjected to bisulfite conversion. Seven to

eight individual clones were sequenced for each sample. Nucleotide

positions are marked relative to the translation initiation site. Arrow

represents a major transcription start site (−60 bp). (D): Ant4 mRNA

is expressed in purified primordial germ cells. Primordial germ cells

were obtained from E11.5 and 12.5-dpc genital ridges and purified

using anti-SSEA1 magnetic beads. The immunodepleted fraction

(dep) refers to cells not retained on the magnetic column. Individual

samples were subjected to reverse transcription–polymerase chain

reaction analysis.

Rodić, Oka, Hamazaki et al. 1321

undetectable in any other organs examined in Figure 6A. There

was no detectable Ant4 mRNA expression in stomach, small

intestine, skeletal muscle, ovary, thymus, uterus, or placenta (data

not shown). In contrast, the previously identified Ant isoforms,

Ant1 and Ant2, were expressed in many non-germ cell organs at

various levels but were low or undetectable in testes.

We then examined which cells within testis express the Ant4

protein using specific antibodies raised against mouse Ant4 (Fig.

6B). Results obtained from immunohistochemical analyses con-

firmed the presence of Ant4 protein within testicular tissue. The

strong cytoplasmic staining of spermatagonia, spermatocytes,

and spermatids within the seminiferous tubules, coupled with the

lack of signal in interstitial, capillary, and capsular cells, suggests

testicular germ cell–specific expression of Ant4. Of interest, it

seems that mature sperm is absent for Ant4 expression. Although

Northern blot analyses using whole ovaries did not detect any

Ant4 gene expression (described above), immunohistochemical

analyses revealed that Ant4 protein is selectively expressed in the

cytoplasm of oocytes (Fig. 6B, middle panels). In contrast, stain-

ing performed on liver sections was negative. These data indicate

that Ant4 is specifically expressed in developing gametes in testis

and ovary. Similar data were obtained using in situ RNA hybrid-

ization with riboprobes against Ant4 transcript (data not shown).

Further, we determined DNA methylation levels across the

Ant4 promoter region in adult mice organs. Only testis showed

hypomethylation of the Ant4 promoter, whereas other tissues

examined were hypermethylated (Fig. 6C). Additionally, Ant4

was expressed in purified primordial germ cells, obtained from

E11.5 and 12.5-dpc genital ridges when they were purified using

anti-SSEA1 magnetic beads (Fig. 6D). The data indicate that Ant4

is expressed in premeiotic fetal germ cells as well.

DiscussionWe have identified a novel isoform of the adenine nucleotide trans-

locase genes, Ant4, by screening differential DNA methylation

patterns in ESCs and EBs. Selective expression status of Ant4 in

undifferentiated ESCs, primordial germ cells, and developing

gametes in adult testis and ovary indicates that Ant4 is a pluripotent

cell-specific isoform of the adenine nucleotide translocase family

in mouse. This contrasts with the previously identified isoforms

of mouse Ant, Ant1 and Ant2, which are predominately expressed

in somatic cells. BLAST analysis revealed the human orthologue

(ANT4), which shares a high homology in deduced amino acid

sequence with mouse Ant4. Moreover, the mouse Ant4 and human

ANT4 genes have a conserved genomic architecture. While we

were preparing this manuscript, Dolce et al. [32] reported this

human orthologue as AAC4 (ADP/ATP carrier protein 4), which

they identified by screening human genome databases for homol-

ogy to AAC1 (ANT1). Their study demonstrated that AAC4

actively exchanges ADP for ATP by an electrogenic antiport mech-

anism. Although AAC4 expression was highest in testis among

human tissues, as we showed here with mouse Ant4, their real-time

PCR analysis indicated that AAC4 was expressed in other somatic

organs, including liver and brain. This is in contrast to the pattern

of Ant4 expression in mice demonstrated by Northern blotting

and immunohistochemical analyses in the present study, which is

highly restricted to developing gametes among adult tissues.

Ant exchanges mitochondrial ATP for cytosolic ADP and

therefore plays a pivotal role in cellular metabolism in eukaryotic

cells. It is intriguing that germ cell–specific isoforms also exist

in other proteins involved in mitochondrial energy metabolism.

Developing spermatocytes are known to express testis-specific

isoforms of pyruvate dehydrogenase complex E1α subunit,

Pdha-2 [33], testis-specific cytochrome c, Cyt cT, [34], and sub-

unit VIb of cytochrome c oxidase [35]. Interestingly, Cyt cT–null

mice exhibit early testicular atrophy and apoptosis [34]. Ant

has been implicated in the mitochondrial permeability transi-

tion, which is a common feature of apoptosis. It is possible that

this germ cell–specific isoform of Ant, probably interacting with

other germ cell–specific mitochondrial membrane proteins, could

also be critical for germ cell development and maintenance. Fur-

ther, a recent study indicates that an antiapoptotic protein, Bcl2,

supports ESC survival and maintenance in harsh culture condi-

tions such as serum deprivation [36], implying that antiapoptotic

mechanisms of the mitochondrial membrane play a role in ESC

self-renewal. We are now generating Ant4-null ESCs and knock-

out mice, which will elucidate the in vivo function of the protein

in spermatogenesis and ESC maintenance.

The present data indicate that de novo DNA methylation,

mediated by Dnmt3a and Dnmt3b, plays a pivotal role in gene

repression of Ant4 during ESC differentiation. In adult organs, as

well as during in vitro ESC differentiation, Ant4 expression levels

inversely correlated with the DNA methylation status of the gene.

Further, Ant4 was readily derepressed by addition of the demethyl-

ating agent 5-aza-dC. In contrast, Oct-4 and Nanog, genes specifi-

cally expressed in pluripotent stem cells and primordial germ cells,

were not derepressed by 5-aza-dC. This implies that transcription

factors involved in Ant4 gene expression are present and active in

differentiated EBs and that DNA methylation may play a primary

role in the suppression of the Ant4 gene in somatic cells. We con-

firmed derepression of Ant4, but not Oct-4 or Nanog, by addition of

5-aza-dC into two other somatic cell lines. Further, this hypothesis

was supported by the fact that functional deletion of Dnmt3a and

Dnmt3b led to failure of gene suppression of Ant4, but not Oct-4,

during ESC differentiation. Taken together, these data suggest that

DNA methylation is required for the transcriptional repression of

the Ant4 gene. The present data also indicate that both members

of the Dnmt3 gene family are required for complete repression of

the Ant4 gene. At the particular CpG site that we investigated by

COBRA, DNA methylation was hardly detected in single-knock-

out EBs within 10 days of differentiation. Dnmt3a and Dnmt3b

may induce DNA methylation cooperatively at some CpG sites.

1322 DNA Methylation–Dependent Repression of Ant4

Undifferentiated ESCs are known to be transcriptionally

permissive for expression of some germ line cell–specific genes

[37]. This might be due to a conserved mode of transcriptional

regulation between ESCs and germ cells or simply reflect the

fact that ESCs are derived from largely unmethylated preim-

plantation embryos (i.e., blastocyst). Recent reports demon-

strated that DNA methylation is involved in the transcriptional

regulation of several germ cell–specific genes, including pgk2

[38], pdha-2 [39], the MAGE gene family [40], and Tact1/Actl7b

[41]. Because Ant4 is expressed exclusively in the testes of the

adult mouse, the regulation of Ant4 transcription may be closely

related to that of other germ cell–specific genes. Of interest, the

promoter region of Ant4 contains the Coup-Tf binding element

at −586 bp, which has been implicated in selective silencing of

the c-Mos proto-oncogene in somatic cells [42, 43]. In addition,

the Ant4 promoter has multiple elements that potentially bind to

Sox/Sry factors, which have also been implicated in early cell

fate specification in development and germ cell–specific tran-

scription [44, 45]. By elucidating the role of DNA methylation

in gene regulation of pluripotent cells, we may uncover common

gene regulatory pathways underlying the transition between plu-

ripotent stem/germ cells and somatic cells.

AcknowledgmentsThe authors thank Drs. Thomas Yang, Paul Oh, Jorg Bungert,

Keith Robertson, and Michael Rutenberg for helpful discussions

and critical reading of the manuscript. This work was supported

in part by National Institutes of Health grants DK59699 and

RR17001 (to N.T.).

DisclosuresN.T. owns stock in RegenMed, Inc.

References

1 Robertson AK, Geiman TM, Sankpal UT et al. Effects of chromatin

structure on the enzymatic and DNA binding functions of DNA methyl-

transferases DNMT1 and Dnmt3a in vitro. Biochem Biophys Res Com-

mun 2004;322:110–118.

2 Li E, Bestor TH, Jaenisch R. Targeted mutation of the DNA methyltrans-

ferase gene results in embryonic lethality. Cell 1992;69:915–926.

3 Panning B, Jaenisch R. RNA and the epigenetic regulation of X chromo-

some inactivation. Cell 1998;93:305–308.

4 Li E, Beard C, Jaenisch R. Role for DNA methylation in genomic imprint-

ing. Nature 1993;366:362–365.

5 Walsh CP, Chaillet JR, Bestor TH. Transcription of IAP endogenous ret-

roviruses is constrained by cytosine methylation. Nat Genet 1998;20:116–

117.

6 Jones PA, Laird PW. Cancer epigenetics comes of age. Nat Genet

1999;21:163–167.

7 Baylin SB, Esteller M, Rountree MR et al. Aberrant patterns of DNA

methylation, chromatin formation and gene expression in cancer. Hum

Mol Genet 2001;10:687–692.

8 Bird AP, Wolffe AP. Methylation-induced repression: belts, braces, and

chromatin. Cell 1999;99:451–454.

9 Takizawa T, Nakashima K, Namihira M et al. DNA methylation is a

critical cell-intrinsic determinant of astrocyte differentiation in the fetal

brain. Dev Cell 2001;1:749–758.

10 Nan X, Ng HH, Johnson CA et al. Transcriptional repression by the

methyl-CpG-binding protein MeCP2 involves a histone deacetylase com-

plex. Nature 1998;393:386–389.

11 Grewal SI, Moazed D. Heterochromatin and epigenetic control of gene

expression. Science 2003;301:798–802.

12 Evans MJ, Kaufman MH. Establishment in culture of pluripotential cells

from mouse embryos. Nature 1981;292:154–156.

13 Smith AG, Heath JK, Donaldson DD et al. Inhibition of pluripotential

embryonic stem cell differentiation by purified polypeptides. Nature

1988;336:688–690.

14 Chambers I, Colby D, Robertson M et al. Functional expression cloning

of Nanog, a pluripotency sustaining factor in embryonic stem cells. Cell

2003;113:643–655.

15 Mitsui K, Tokuzawa Y, Itoh H et al. The homeoprotein Nanog is required

for maintenance of pluripotency in mouse epiblast and ESCs. Cell

2003;113:631–642.

16 Niwa H, Miyazaki J, Smith AG. Quantitative expression of Oct-3/4

defines differentiation, dedifferentiation or self-renewal of ESCs. Nat

Genet 2000;24:372–376.

17 Okano M, Xie S, Li E. Cloning and characterization of a family of

novel mammalian DNA (cytosine-5) methyltransferases. Nat Genet

1998;19:219–220.

18 Lee JH, Hart SRL, Skalnik DG. Histone deacetylase activity is required

for embryonic stem cell differentiation. Genesis 2004;38:32–38.

19 Huntriss J, Hinkins M, Oliver B et al. Expression of mRNAs for DNA

methyltransferases and methyl-CpG-binding proteins in the human

female germ line, preimplantation embryos, and embryonic stem cells.

Mol Reprod Dev 2004;67:323–336.

20 Kremenskoy M, Kremenska Y, Ohgane J et al. Genome-wide analysis of

DNA methylation status of CpG islands in embryoid bodies, teratomas,

and fetuses. Biochem Biophys Res Commun 2003;311:884–890.

21 Shiota K, Kogo Y, Ohgane J et al. Epigenetic marks by DNA methylation

specific to stem, germ and somatic cells in mice. Genes Cells 2002;7:961–

969.

22 Chen T, Ueda Y, Dodge JE et al. Establishment and maintenance of

genomic methylation patterns in mouse embryonic stem cells by Dnmt3a

and Dnmt3b. Mol Cell Biol 2003;23:5594–5605.

23 Hamazaki T, Iiboshi Y, Oka M et al. Hepatic maturation in differentiating

embryonic stem cells in vitro. FEBS Lett 2001;497:15–19.

24 Huang TH, Laux DE, Hamlin BC et al. Identification of DNA methylation

markers for human breast carcinomas using the methylation-sensitive

restriction fingerprinting technique. Cancer Res 1997;57:1030–1034.

25 Pesce M, De Felici M. Purification of mouse primordial germ cells by

MiniMACS magnetic separation system. Dev Biol 1995;170:722–725.

26 Belzacq AS, Vieira HL, Kroemer G et al. The adenine nucleotide translo-

cator in apoptosis. Biochimie 2002;84:167–176.

Rodić, Oka, Hamazaki et al. 1323

27 Slim R, Levilliers J, Ludecke HJ et al. A human pseudoautosomal gene

encodes the ANT3 ADP/ATP translocase and escapes X-inactivation.

Genomics 1993;16:26–33.

28 Mozo T, Fischer K, Flugge UI et al. The N-terminal extension of the ADP/

ATP translocator is not involved in targeting to plant mitochondria in vivo.

Plant J 1995;7:1015–1020.

29 Chen ST, Chang CD, Huebner K et al. A human ADP/ATP translocase

gene has seven pseudogenes and localizes to chromosome X. Somat Cell

Mol Genet 1990;16:143–149.

30 Xiong Z, Laird PW. COBRA: a sensitive and quantitative DNA methyla-

tion assay. Nucleic Acids Res 1997;25:2532–2534.

31 Okano M, Bell DW, Haber DA et al. DNA methyltransferases Dnmt3a and

Dnmt3b are essential for de novo methylation and mammalian develop-

ment. Cell 1999;99:247–257.

32 Dolce V, Scarcia P, Iacopetta D et al. A fourth ADP/ATP carrier isoform in

man: identification, bacterial expression, functional characterization and

tissue distribution. FEBS Lett 2005;579:633–637.

33 Brown RM, Dahl HH, Brown GK. Pyruvate dehydrogenase E1 alpha sub-

unit genes in the mouse: mapping and comparison with human homologs.

Somat Cell Mol Genet 1990;16:487–492.

34 Narisawa S, Hecht NB, Goldberg E et al. Testis-specific cytochrome c-

null mice produce functional sperm but undergo early testicular atrophy.

Mol Cell Biol 2002;22:5554–5562.

35 Huttemann M, Jaradat S, Grossman LI. Cytochrome c oxidase of mam-

mals contains a testes-specific isoform of subunit VIb–the counterpart to

testes-specific cytochrome c? Mol Reprod Dev 2003;66:8–16.

36 Yamane T, Dylla SJ, Muijtjens M et al. Enforced Bcl-2 expression over

rides serum and feeder cell requirements for mouse embryonic stem cell

self-renewal. Proc Natl Acad Sci U S A 2005;102:3312–3317.

37 Geijsen N, Horoschak M, Kim K et al. Derivation of embryonic germ cells

and male gametes from embryonic stem cells. Nature 2004;427:148–154.

38 Zhang LP, Stroud JC, Walter CA et al. A gene-specific promoter in trans-

genic mice directs testis-specific demethylation prior to transcriptional

activation In vivo. Biol Reprod 1998;59:284–292.

39 Iannello RC, Gould JA, Young JC et al. Methylation-dependent silenc-

ing of the testis-specific Pdha-2 basal promoter occurs through selective

targeting of an activating transcription factor/cAMP-responsive element-

binding site. J Biol Chem 2000;275:19603–19608.

40 De Smet C, Lurquin C, Lethe B et al. DNA methylation is the primary

silencing mechanism for a set of germ line- and tumor-specific genes with

a CpG-rich promoter. Mol Cell Biol 1999;19:7327–7335.

41 Hisano M, Ohta H, Nishimune Y et al. Methylation of CpG dinucleotides

in the open reading frame of a testicular germ cell-specific intronless gene,

Tact1/Actl7b, represses its expression in somatic cells. Nucleic Acids Res

2003;31:4797–4804.

42 Xu W, Cooper GM. Identification of a candidate c-mos repressor

that restricts transcription of germ cell-specific genes. Mol Cell Biol

1995;15:5369–5375.

43 Lin HB, Jurk M, Gulick T et al. Identification of COUP-TF as a tran-

scriptional repressor of the c-mos proto-oncogene. J Biol Chem

1999;274:36796–36800.

44 Tchenio T, Casella JF, Heidmann T. Members of the SRY family regulate

the human LINE retrotransposons. Nucleic Acids Res 2000;28:411–415.

45 Clarkson MJ, Harley VR. Sex with two SOX on: SRY and SOX9 in testis

development. Trends Endocrinol Metab 2002;13:106–111.