DEVELO

PMENT

3411RESEARCH ARTICLE

INTRODUCTIONPrimordial germ cells (PGCs) are the founding population of cells thatwill ultimately give rise to the mature gametes. Unlike organisms thathave a mosaically determined germline, PGCs in the mouse embryoare specified by an inductive mechanism that requires the presence ofseveral bone morphogenetic proteins (BMPs) emanating from thesurrounding somatic cells (Fujiwara et al., 2001; Lawson et al., 1999;Ying et al., 2000; Ying and Zhao, 2001). PGCs can first be detected inthe extra-embryonic mesoderm at 7.25 days post-coitus (dpc)(Ginsburg et al., 1990). By 8.5 dpc, PGCs enter the embryo proper andactively migrate through the hindgut endoderm, colonizing thedeveloping gonads between 10.5 and 11.5 dpc. During this time,PGCs proliferate from an initial population of 45 cells at 7.5 dpc to25,000 cells at 13.5 dpc when proliferation ceases (Tam and Snow,1981). Sexual differentiation of the germline begins at 13.5 dpc withfemale germ cells entering prophase I of meiosis. Within the malegonad, a signal thought to originate from the testis cords prevents entryinto meiosis and male PGCs enter a mitotic arrest by 14.5 dpc(McLaren, 1983). Prior to these changes, male and female PGCs aresexually indifferent, capable of following either the male or femalepathway (McLaren and Southee, 1997).

Shortly after PGCs enter the urogenital ridges, both male andfemale germ cells undergo a common set of changes independentof sexual differentiation. Changes in cell morphology and cell-

adhesion properties occur as the germ cells transition to a non-migratory state (De Felici et al., 1992; Donovan et al., 1986;Garcia-Castro et al., 1997). Male and female PGCs also ceaseproliferating, have decreased potential to form pluripotent stemcell lines (Matsui et al., 1992; McLaren, 1984; Resnick et al.,1992), and undergo a wave of apoptosis (Coucouvanis et al., 1993;Wang et al., 1998). These differentiation events are accompaniedby changes in gene expression as some germ cell marker genes,such as Tnap (Akp2 – Mouse Genome Informatics) and Zfp148, aredownregulated (Donovan et al., 1986; Hahnel et al., 1990;Takeuchi et al., 2003). Other genes, including Mvh (Ddx4 – MouseGenome Informatics), Scp3 (Sycp3 – Mouse Genome Informatics),Dazl, Mageb4 and Gcna1 are upregulated during this time (Cookeet al., 1996; Di Carlo et al., 2000; Fujiwara et al., 1994; Osterlundet al., 2000).

In addition to the differentiation events mentioned above, PGCsmediate two essential epigenetic processes. First, female PGCsreactivate their silenced X chromosome, thereby ensuring that eachoocyte carries an active X chromosome (Monk and McLaren, 1981;Tam et al., 1994). Interestingly, the ability to reactivate the inactiveX chromosome is not confined to female germ cells, as XXY malegerm cells also possess this reactivation capability (Mroz et al.,1999). Second, migratory germ cells carry parent-of-origin-specificimprinting marks and high levels of allele-specific methylation thatcontribute to monoallelic expression in migratory PGCs. Thesedifferentially methylated regions become hypomethylated as PGCscolonize the gonads, leading to a loss of imprinting and biallelicgene expression (Hajkova et al., 2002; Lee et al., 2002; Szabo et al.,2002). However, this wave of demethylation is not restricted toimprinted loci and genes of the X chromosome, as several non-imprinted genes and repetitive sequences also show decreasedmethylation at this time (Hajkova et al., 2002; Lane et al., 2003;Lees-Murdock et al., 2003).

DNA methylation is a primary mechanism for silencingpostmigratory primordial germ cell genes in both germ celland somatic cell lineagesDanielle M. Maatouk1, Lori D. Kellam1, Mellissa R. W. Mann2, Hong Lei3, En Li3, Marisa S. Bartolomei2 andJames L. Resnick1,*

DNA methylation is necessary for the silencing of endogenous retrotransposons and the maintenance of monoallelic geneexpression at imprinted loci and on the X chromosome. Dynamic changes in DNA methylation occur during the initial stages ofprimordial germ cell development; however, all consequences of this epigenetic reprogramming are not understood. DNAdemethylation in postmigratory primordial germ cells coincides with erasure of genomic imprints and reactivation of the inactive Xchromosome, as well as ongoing germ cell differentiation events. To investigate a possible role for DNA methylation changes ingerm cell differentiation, we have studied several marker genes that initiate expression at this time. Here, we show that thepostmigratory germ cell-specific genes Mvh, Dazl and Scp3 are demethylated in germ cells, but not in somatic cells. Premature lossof genomic methylation in Dnmt1 mutant embryos leads to early expression of these genes as well as GCNA1, a widely used germcell marker. In addition, GCNA1 is ectopically expressed by somatic cells in Dnmt1 mutants. These results provide in vivo evidencethat postmigratory germ cell-specific genes are silenced by DNA methylation in both premigratory germ cells and somatic cells. Thisis the first example of ectopic gene activation in Dnmt1 mutant mice and suggests that dynamic changes in DNA methylationregulate tissue-specific gene expression of a set of primordial germ cell-specific genes.

Key words: Mouse, Primordial germ cells

Development 133, 3411-3418 (2006) doi:10.1242/dev.02500

1Department of Molecular Genetics and Microbiology, PO Box 100266, University ofFlorida, Gainesville, FL 32610-0266, USA. 2Howard Hughes Medical Institute andDepartment of Cell and Developmental Biology, University of Pennsylvania School ofMedicine, Philadelphia, PA 19104, USA. 3Epigenetics Program, Models of DiseaseCenter, Novartis Institute for Biomedical Research, 250 Massachusetts Avenue,Cambridge, MA 02139, USA.

*Author for correspondence (e-mail: resnick@mgm.ufl.edu)

Accepted 19 June 2006

DEVELO

PMENT

3412

We have been investigating regulatory mechanisms underlyingpostmigratory germ cell differentiation. Several studies suggest thatcontinuing PGC development is regulated by a cell intrinsic programrather than by inductive signals from the gonads. PGCs located inectopic locations enter meiosis and initiate expression of thepostmigratory marker GCNA1 on schedule without exposure to theurogenital ridges (Wang et al., 1997). Embryonic stem cells havebeen shown to differentiate to form PGC-like cells that can go on toform cells resembling both oocytes and spermatocytes, furtherdemonstrating that PGC differentiation can occur independently ofthe gonadal environment (Geijsen et al., 2004; Hubner et al., 2003;Toyooka et al., 2003). Last, cessation of germ cell proliferation hasalso been suggested to be cell intrinsic (Ohkubo et al., 1996).

We previously tested the potential of premigratory germ cells todifferentiate in culture and reported that 8.5 dpc premigratory PGCsin feeder culture can differentiate to express GCNA1 on the correcttemporal schedule (Richards et al., 1999). Surprisingly, the rate ofdifferentiation in culture increased when PGCs were exposed to theDNA demethylating agent 5-azacytidine or the histone deacetylaseinhibitor trichostatin A (Maatouk and Resnick, 2003). This suggeststhat epigenetic mechanisms may contribute to the regulation of germcell differentiation.

Here, we further investigate the role of DNA methylation in theprocess of PGC differentiation. We present evidence that severalpostmigratory germ cell-specific genes are demethylated in germcells as they colonize the genital ridges and that DNA demethylationcontrols the temporal expression of these genes in vivo. In addition,we show that these postmigratory germ cell-specific genes areectopically expressed in DNA methyltransferase mutant embryos,suggesting that DNA methylation is a mechanism of silencing germcell-specific genes in somatic tissues. These results provide the firstin vivo evidence of tissue-specific embryonic gene regulationmediated by dynamic changes in DNA methylation.

MATERIALS AND METHODSMouse strainsPrimordial germ cells were purified from embryos obtained from timedmatings of B6C3F1 mice (Jackson Laboratories, Bar Harbor, ME). Noon ofthe day on which a mating plug was first visible was taken to be 0.5 dpc. For

RNA analysis, mice carrying either the Dnmt1n (Li et al., 1992) or Dnmt1c

allele (Lei et al., 1996) were maintained on the B6(CAST7) mixedbackground (Mann et al., 2003). Dnmt1c embryos used for immunostainingwere also maintained on a mixed background (129/SvJae � C57BL/6).

Primordial germ cell isolation and purificationGonads were collected from 10.5 dpc and 13.5 dpc embryos. At 13.5 dpc,embryos were sex segregated based on the presence of testis cords in themale gonad. PGCs were immunomagnetically purified using the TG-1antibody as described (Pesce and De Felici, 1995). Purified fractions weregreater than 85% (10.5 dpc) and 90% (13.5 dpc) germ cells, as judged byalkaline phosphatase staining. Immunodepleted fractions contained less than1% PGCs and primarily contained somatic cells from the gonad andmesonephros.

Bisulfite conversion and DNA sequencingGenomic DNA isolated from both purified and immunodepleted fractionswas subjected to bisulfite conversion as described (Clark et al., 1994).Bisulfite primers were designed against the converted DNA sequences andare listed in Table 1. PCR amplification was performed on 10% of onepurification (approximately one embryo equivalent) with HotStar Taq(Qiagen) using the following cycling conditions: 95°C for 15 minutesfollowed by 35 cycles of 95°C for 45 seconds, 53°C for 30 seconds and 72°Cfor 1.5 minutes. Bisulfite PCR amplifications were performed on twoindependent germ cell purifications to avoid inconsistencies that might arisefrom conducting PCR on small amounts of DNA. PCR products were gelpurified using Wizard DNA Clean-up System (Promega) and cloned usingthe pGEM-T Easy Vector System (Promega). Plasmid sequencing wascarried out using ABI Prism BigDye terminator (PerkinElmer) by the Centerfor Mammalian Genetics DNA Sequence Core.

RNA analysisRNA was isolated from 9.5 dpc embryos using the HighPure RNA TissueKit (Roche Molecular Biochemicals) with minor modifications of themanufacturer’s recommendations. Random-primed cDNA was preparedwith Superscript II as recommended (Invitrogen). For RT-PCR, forward andreverse primers were located in separate exons to exclude any bands thatmight arise from genomic DNA contamination. Primer sequences are listedin Table 1. RT-PCR was performed using Triplemaster Taq (Eppendorf) andPCR products were run on 2% agarose gels, Southern blotted andhybridized with �-[32PO4]-dCTP labeled probes. Probe fragments weregenerated by gel purification of the RT-PCR products obtained from testiscDNA.

RESEARCH ARTICLE Development 133 (17)

Table 1. PCR primers used in this studyPrimer name Primer sequence (5�-3�) Region analyzed

Mvh-F TTTGGCTCATATGATGCGGG 210 bpMvh-R ACACCCTTGTACTATCTGTCGAACTGAATGACC

Mage-b4-F ACGCGAGGTATCTCGGGC 180 bpMage-b4-R GGGCGTAAGTTGGCAACC

Dazl-F TTCTGCTCCACCTTCGAGGTT 335 bpDazl-R CTATCTTCTGCACATCCCAGTCATTA

Scp3-F CCAATCAGCAGAGAGCTTGG 450 bpScp3-R AGCTGTCGCTGTCCCCACAC

Hprt-F GCTGGTGAAAAGGACCTCT 250 bpHprt-R CACAGGACTAGAACACCTGC

B-Daz-F GGAAGAAAAAAACTAAGTCCTGATGGC 360 bp (–540 to –180)B-Daz-R AAACCCCCCCAATCCCTCAC

B-Mvh-F TGAATGAATATAATGGAATTGATGAGTT 248 bp (–378 to –77)B-Mvh-R AAAACAACAAATAACATCAAA

B-Scp3-F GAATGAGGATTTATGAGTAAAGATGGTT 381 bp (–241 to +141)B-Scp3-R CCCCCATCTCCTTAACCTCAA

B-Tnap-F GGAAGAAAAAAACTAAGTCCTGATGGC 128 bp (–213 to –85)B-Tnap-B AAACCCCCCCAATCCCTCAC

DEVELO

PMENT

Immunochemical methodsEmbryos were collected at 8.5 and 9.5 dpc and fixed overnight inmethanol:dimethyl sulfoxide (4:1) at 4°C. Endogenous peroxidase activitywas inactivated by a 2-hour incubation in methanol:dimethyl sulfoxide: 30%hydrogen peroxide (4:1:1) at room temperature. Embryos were stored in100% methanol at –20°C.

Antibody to GCNA1 was generously provided by George Enders. ForGCNA1 staining of paraffin sections, fixed embryos were dehydratedthrough an ethanol series, cleared in Citrisolv (Fisher) and embedded inparaffin. Cross-sections were cut at 7 �m. Immunostaining for GCNA1 inparaffin sections was performed using the 10D9G11 monoclonal antibodyas previously described (Richards et al., 1999). Whole-mount embryos wereimmunostained using the same procedure with several adjustments.Blocking was performed in PBSMT (PBS, 2% w:v nonfat dry milk, 0.1%Triton X-100) and a horseradish peroxidase (HRP)-conjugated mouse anti-rat IgM (Zymed) secondary antibody was used. Color development wasperformed using the Liquid DAB (3,3�-diaminobenzidine) Substrate Kitfollowing the manufacturer’s instructions (Zymed). Paraffin sections werestained for SSEA1 immediately after GCNA1 staining as previouslydescribed (Richards et al., 1999).

RESULTSPostmigratory germ cell-specific genes aredemethylated as PGCs colonize the urogenitalridgesWe have previously shown that initial expression of thepostmigratory germ cell marker GCNA1 is controlled by a cell-intrinsic timing mechanism. PGCs removed from 8.5 dpc embryosand plated in feeder culture initiate GCNA1 expression after a 2- to3-day delay, consistent with their in vivo temporal expression pattern(Richards et al., 1999). Furthermore, we found that the DNAdemethylating agent 5-azacytidine accelerates the rate and extent ofgerm cell differentiation in culture (Maatouk and Resnick, 2003).Because the expression of GCNA1 in cultured PGCs is sensitive tochanges in DNA methylation, we explored whether induction ofother postmigratory germ cell genes could be regulated by ademethylation event. As the gene encoding GCNA1 remainsunknown, we investigated the methylation status of several

postmigratory germ cell-specific genes that share a similarexpression pattern with GCNA1. Mvh, Scp3 and Dazl areexclusively expressed by both male and female germ cells as theyenter the urogenital ridges between 10.5-11.5 dpc (Cooke et al.,1996; Di Carlo et al., 2000; Fujiwara et al., 1994). Fig. 1 shows thatthe first exon of each of these genes is contained within a CpGisland.

To examine the methylation status of these postmigratory germcell genes, genomic DNA obtained from immunomagneticallypurified 10.5 and 13.5 dpc PGCs was subjected to bisulfite sequenceconversion. Approximately 20 CpG residues were analyzed for thepresence of methylation. For each gene, 10.5 dpc PGCs showed highlevels of methylation, but by 13.5 dpc all three genes showed asignificant loss of methylation with most clones being completelyunmethylated (Fig. 2). No significant differences in methylationwere observed between male and female PGCs, consistent with theidea that PGCs are indifferent prior to 12.5 dpc. These resultssuggest that DNA demethylation of the germline between 10.5 and12.5 dpc functions not only to reprogram imprinted loci andreactivate the X chromosome, but may contribute to additionalpostmigratory differentiation events.

Although most clones isolated from the PGC fraction weresignificantly hypomethylated at 13.5 dpc, some clones retained highlevels of methylation. Some of these clones may be explained by thepresence of contaminating somatic cells in the purified germ cellpreparations, as cells of the immunodepleted fraction remainedhighly methylated at 13.5 dpc (Fig. 2). However, because 13.5 dpcPGC preparations are routinely more than 90% pure, we favor theexplanation that some PGC genomes have not undergonedemethylation by 13.5 dpc. Consistent with this interpretation, somegerm cells initiate GCNA1 expression by 11.5 dpc, but many germcells do not express the marker until 14.5 dpc (Enders and May,1994).

Fig. 2 demonstrates that the germ cell-specific genes Mvh, Dazland Scp3 exhibit loss of methylation as they first become expressed.In contrast to this pattern of expression, several genes are negativelyregulated in PGCs as they differentiate into gonocytes. Tnap isexpressed as germ cells are initially allocated in the extra-embryonicmesoderm at 7.25 dpc, but expression is lost between 13.5 and 14.5dpc (Donovan et al., 1986; Ginsburg et al., 1990). We examined theTnap locus to explore any potential role of dynamic DNAmethylation changes on a gene that is negatively regulated ingonocytes. Bisulfite analysis of eight CpG dinucleotides upstreamof exon 1 shows that this region of Tnap is unmethylated at 10.5,13.5 and 14.5 dpc in both germ cells and somatic cells. This resultis consistent with the notion that DNA demethylation occurs ingenes that are positively regulated as germ cells transition intogonocytes, rather than a characteristic of all genes expressed in germcells.

Expression of GCNA1 is sensitive to DNAmethylation in primordial germ cellsPrimordial germ cells in culture express the postmigratory germ cellmarker GCNA1 on an accelerated schedule when exposed to theDNA demethylating agent 5-azacytidine (Maatouk and Resnick,2003). Although many genes respond to this agent in culture, Walshand Bestor (Walsh and Bestor, 1999) found that few genes aresubject to dynamic DNA methylation changes in vivo. To determinewhether GCNA1 expression is regulated by DNA methylation invivo, we next tested whether GCNA1 is prematurely expressed byPGCs in DNA methyltransferase 1 (Dnmt1) mutant embryos.Dnmt1 maintains methylation patterns during DNA replication and

3413RESEARCH ARTICLEDNA methylation silences gonocyte genes

Fig. 1. CpG island location in several germ cell-specific genes. Thegenomic structures from –2 kb to +2 kb relative to the transcriptionstart site are depicted for Dazl, Mvh, Scp3 and Tnap. For each gene,exon 1 (exon 1a for Tnap) is surrounded by a CpG island defined aspreviously described (Gardiner-Garden and Frommer, 1987). Blackboxes represent exons, gray bars represent CpG islands. The regionsamplified for bisulfite analysis are indicated by black bars below eachCpG island.

ATG

ATG

ATG

Dazl

Mvh

Scp3

Tnap

1 kb

DEVELO

PMENT

3414

the null Dnmt1c mutation results in a 98% loss of genomicmethylation (Lei et al., 1996). Dnmt1c/c mutant embryos were cross-sectioned and immunostained for the PGC marker SSEA1 and forthe postmigratory germ cell marker GCNA1. GCNA1 is normallynot detectable prior to 10.5 dpc (Enders and May, 1994). In 8.5 dpcDnmt1c/c embryos, PGCs located in the yolk sac endodermsimultaneously expressed GCNA1 and SSEA1, with GCNA1 beingdetected 2-3 days earlier than expected (Fig. 4B). PGCs in theposterior region of 9.5 dpc embryos also prematurely expressedGCNA1 (Fig. 4C). These results indicate that in vivo expression ofGCNA1 is temporally controlled by DNA methylation.

DNA methylation is a primary silencingmechanism for the postmigratory germ cellmarker GCNA1 in somatic cellsIn addition to demonstrating that DNA methylation regulatesGCNA1 expression in germ cells, Fig. 4 suggested that somesomatic cells also express this germ cell-specific marker when DNAmethylation is reduced by mutation of Dnmt1. To further investigatethe expression pattern of GCNA1 in Dnmt1 embryos, 9.5 dpc

Dnmt1n/n embryos were subjected to whole-mount immunostaining.This hypomorphic mutation leads to a 70% reduction in DNAmethylation, compared with the 98% reduction observed in Dnmt1c/c

embryos (Lei et al., 1996). Surprisingly, not only was GCNA1prematurely expressed in PGCs, but ectopic expression wasobserved in somatic cells scattered throughout the entire embryo(Fig. 5C,D). Because these cells do not express the PGC markersSSEA1 (Fig. 4) or OCT4 (Hattori et al., 2004), we suggest that theyare not PGCs that have migrated to aberrant locations.

As only a small number of cells ectopically expressed GCNA1, itseemed likely that the low levels of functional Dnmt1 enzymepresent in the hypomorphic Dnmt1n/n mutants might attenuatepromiscuous gene activation. To test this idea GCNA1immunostaining was also performed on 9.5 dpc Dnmt1c/c embryos.The more severe mutation consistently caused much higher levels ofectopic expression than observed in the Dnmt1n/n embryos (Fig.5F,G). As expected, wild-type and heterozygous embryos at thesestages exhibited no GCNA1 expression (Fig. 5A,E). Developmentof Dnmt1-deficient mutants is frequently retarded such that 9.5 dpcDnmt1c/c embryos more closely resemble wild-type 8.5 dpc

RESEARCH ARTICLE Development 133 (17)

A. Mvh

B. Scp3

C. Dazl

10.5 dpc germ cells

10.5 dpc somatic cells

13.5 dpc female germ cells 13.5 dpc male germ cells

13.5 dpc female somatic cells 13.5 dpc male somatic cells

10.5 dpc somatic cells

10.5 dpc germ cells

10.5 dpc germ cells

10.5 dpc somatic cells

13.5 dpc female germ cells

13.5 dpc female germ cells

13.5 dpc female somatic cells

13.5 dpc female somatic cells

13.5 dpc male germ cells

13.5 dpc male germ cells

13.5 dpc male somatic cells

13.5 dpc male somatic cells

Fig. 2. DNA methylation analysis ofpostmigratory germ cell-specificgenes. (A-C) Bisulfite sequence analysisof (A) Mvh, (B) Scp3 and (C) Dazl wasperformed on immunomagneticallypurified germ cell and depleted somaticcell fractions from 10.5 and 13.5 dpcembryos. Each line represents anindividually sequenced clone and circlesrepresent CpG residues. White circlesindicate unmethylated CpG sites; blackcircles represent methylated CpG sites.

DEVELO

PMENT

embryos. Fig. 5H demonstrates that GCNA1 expression is notreadily detected in a more closely stage-matched 8.5 dpc Dnmt1+/c

embryo. Together, these results indicate that the postmigratory germcell marker GCNA1 is ectopically expressed both temporally andspatially in embryos lacking a functional Dnmt1 enzyme.

Premature expression of postmigratoryprimordial germ cell genes in Dnmt1 mutantembryosSeveral postmigratory germ cell-specific genes are demethylated asthe germ cells colonize the developing gonads (Fig. 2). Additionally,GCNA1 is ectopically expressed under conditions of reducedmethylation. To determine if additional postmigratory germ cell-specific genes are prematurely expressed in Dnmt1 mutant embryos,RT-PCR analysis was performed to examine the expression profilesof Mvh, Scp3, Dazl and another PGC-specific gene that shares asimilar expression pattern, Mageb4 (Osterlund et al., 2000) (Fig. 6).As these genes are expressed only after PGCs enter the developinggonads, little or no expression was detected in wild-type andheterozygous 9.5 dpc embryos, as expected. However, embryoshomozygous for either the Dnmt1n or Dnmt1c mutation precociously

expressed each of the germ cell genes analyzed. In addition,expression seemed to be greater in the more severe Dnmt1c mutant.These results support the notion that postmigratory PGC geneexpression is dependent upon the genome-wide demethylation eventthat occurs during colonization of the gonads.

DISCUSSIONDNA methylation regulates germ cell-specificgene expressionRecent results from several laboratories demonstrate that between10.5 and 12.5 dpc the germ cell genome undergoes a wave ofgenomic demethylation that affects genes on the inactive Xchromosome, imprinted loci and some repetitive elements (Hajkovaet al., 2002; Lane et al., 2003; Lee et al., 2002; Lees-Murdock et al.,2003). Although most CpG islands are unmethylated regardless oftissue or expression status (Ioshikhes and Zhang, 2000; Rollins etal., 2006), we found that several germ cell-specific genes are highlymethylated at 10.5 dpc and are included in this wave of germ celldemethylation as they are first expressed (Fig. 2). These genesremain methylated and silent in somatic cells. Loss of methylation,as observed in Dnmt1 mutants, correlates with their premature

3415RESEARCH ARTICLEDNA methylation silences gonocyte genes

11

16

34

19

23

26

7

20

13

9

8

10.5 dpc germ cells 10.5 dpc somatic cells

13.5 dpc female germ cells 13.5 dpc female somatic cells

13.5 dpc male germ cells 13.5 dpc male somatic cells

14.5 dpc female germ cells 14.5 dpc female somatic cells

14.5 dpc male germ cells 14.5 dpc male somatic cells

Fig. 3. DNA methylation analysis of Tnap. Bisulfitesequence analysis of Tnap was performed onimmunomagnetically purified germ cell and somaticcell fractions from 10.5, 13.5 and 14.5 dpc embryos.Each line represents an individually sequenced clone.Numbers indicate the frequency of each observedclone. White circles indicate unmethylated CpG sites;black circles indicate methylated CpG sites.

Fig. 4. The postmigratory germ cell marker GCNA1 is precociously expressed in premigratory germ cells in Dnmt1-deficient embryos.Embryos were immunostained for SSEA1 with TG-1 (red) and for GCNA1 (black). (A) TG-1 and GCNA1 immunohistochemistry of a 12.5 dpcembryo, demonstrating that both markers specifically recognize germ cells. The inset shows that GCNA1 reactivity at this time is normally restrictedto the TG-1-positive population. Arrow and arrowhead indicate GCNA1-expressing and non-expressing germ cells, respectively. (B) 8.5 dpcDnmt1c/c yolk sac with a TG-1 and GCNA1-expressing cell. The inset shows a higher magnification view of the doubly stained cell. (C) NumerousTG-1- and GCNA1-positive cells in a 9.5 dpc Dnmt1c/c embryo.

DEVELO

PMENT

3416

expression in germ cells and ectopic expression in somatic cells.Together, these results strongly suggest that the naturally occurringdemethylation of these genes in germ cells is rate limiting for theirexpression, and that DNA methylation is necessary to maintainsilencing of these genes in somatic cells.

The rapid rate of demethylation and the presence of nuclearDnmt1 protein led Hajkova et al. to propose that germ celldemethylation results from an active mechanism, rather than thepassive process of replication without further methylation (Hajkovaet al., 2002). Our results are consistent with this proposal as weobserved individual PGC genomes having intermediate levels ofmethylation, probably representing PGCs in the process of beingactively demethylated (Fig. 2). Precocious expression of germ cell-specific genes, presumably owing to passive demethylation inDnmt1 mutants, suggests that the hypomethylated state is sufficientfor transcription and does not require the process of activedemethylation.

Several reports confirm a role for DNA methylation in generegulation in vivo. Monoallelically expressed genes, includingimprinted genes and X chromosome genes become biallelically

expressed, and IAP elements are transcribed in Dnmt1 mutants(Caspary et al., 1998; Li et al., 1993; Sado et al., 2000; Walsh et al.,1998). Bdnf expression is increased in neurons of mice bearing aconditionally deleted Dnmt1 gene (Martinowich et al., 2003).Although numerous genes are ectopically expressed in Dnmt1-deficient cells in culture, we are aware of only one previous reportof ectopic gene expression of a single copy gene in Dnmt1 mutantembryos. OCT4, a transcription factor present in premeiotic germcells, was found to be ectopically expressed in placentas of Dnmt1mutant embryos (Hattori et al., 2004). Here, we report extensiveectopic expression of germ line-specific genes resulting from loss ofDNA methylation.

Expression of postmigratory germ cell genes isattenuated in DNA-deficient mutantsIf DNA methylation is necessary to silence germ cell genes insomatic cells, why are only some cells positive for GCNA1 in theDnmt1-deficient embryos? Although Dnmt1c/c embryos lackdetectable Dnmt1 activity, Dnmt1n/n embryos produce low levels offunctional Dnmt1 enzyme and retain about 30% of genomicmethylation (Lei et al., 1996). The experiments reported here werenot performed under directly comparable conditions; however, Mvh,Dazl, Scp3 and Mageb4 all show greater expression in the Dnmt1c/c

compared with the Dnmt1n/n mutants relative to the Hprt control(Fig. 6). This was also observed for the expression of GCNA1 inDnmt1n/n compared with Dnmt1c/c embryos (Fig. 5). Repressivechromatin structure or compensation by de novo DNAmethyltransferases may maintain silencing in non-expressing cells.Alternatively, DNA demethylation in mutant embryos, which occursby a passive replication-dependent mechanism, may occur moreslowly in some cells as loss of Dnmt1 may decrease the rate of cellproliferation (Jackson-Grusby et al., 2001; Milutinovic et al., 2003).Slower cell cycles could lengthen the time it takes to passivelydemethylate, causing delayed gene activation. This may account forthe large number of cells observed in the Dnmt1c/c mutant that do notinitiate GCNA1 expression.

How could DNA methylation silence germ cell-specific genes inboth germ and somatic lineages? Several mechanisms, includingrestricted expression of positive acting transcription factors, stericinterference with transcription factor binding sites, attraction ofmethyl DNA binding proteins and DNA methylation induced

RESEARCH ARTICLE Development 133 (17)

Fig. 5. The postmigratory germ cell marker GCNA1 isectopically expressed in somatic cells of Dnmt1-deficient embryos. All embryos were immunostained forGCNA1 expression. (A) 9.5 dpc Dnmt1+/n embryo.(B) Enlarged view of the same embryo. (C) 9.5 dpcDnmt1n/n embryo (n=4/4). (D) Enlarged view of embryo inC, demonstrating scattered GCNA1 expression throughoutthe embryo. (E) 9.5 dpc Dnmt1+/c embryo. (F) 9.5 dpcDnmt1c/c embryo (n=5/5). (G) Enlarged view of embryo inF. (H) 8.5 dpc Dnmt+/c embryo. Micrographs in A,C,E,F areat the same magnification. Ectopic GCNA1 expression wasnot detected in any of 11 control littermate embryosstained in parallel (not shown).

Fig. 6. Postmigratory germ cell-specific genes are prematurelyexpressed in Dnmt1-deficient embryos. RT-PCR gene expressionanalysis for Dazl, Mvh, Mageb4 and Scp3, and was performed on 9.5dpc wild-type, heterozygous and mutant Dnmt1n and Dnmt1c embryos.Hprt amplification was used as a loading control.

DEVELO

PMENT

changes in histone modifications have been proposed (Jaenisch andBird, 2003). Interestingly, recent reports suggest that the repressivetranscription factor E2F6 is necessary to silence severalspermatogenic genes in somatic cells, and that promoters of thesegenes are hypomethylated in E2F6-deficient cells (Pohlers et al.,2005; Storre et al., 2005). We are currently investigating whetherE2F6 and DNA methylation share a common pathway to repressgerm cell-specific genes.

DNA methylation mediated regulation of germcell developmentDNA methylation has previously been proposed to regulate theexpression of tissue-specific genes; however, the lack of substantialin vivo evidence has narrowed the proposed role of methylation tosilencing of endogenous retrotransposons and maintainingmonoallelic gene expression within imprinted loci and on theinactive X chromosome in females (Jaenisch, 1997; Walsh andBestor, 1999). Our data provide strong evidence that methylationmay indeed control tissue-specific gene expression for a set of germcell-specific genes that are coordinately activated upon germ cellentry into the gonads.

Seki et al. (Seki et al., 2005) recently investigated genome-widechanges in chromatin modifications during primordial germ celldevelopment. Using antibodies to 5-methylcytosine, they observedthat PGCs at the base of the allantois at 8.0 dpc have similarmethylation levels as somatic cells; however, migrating PGCs in thehindgut displayed lower methylation levels. This first wave of germcell demethylation may signify the transition from a somatic cell fateto a more pluripotent state, as germ cells at this stage resemble cellsof the inner cell mass in their expression profiles and their ability togive rise to pluripotent stem cell lines (Donovan and de Miguel,2003; Matsui et al., 1992; Resnick et al., 1992).

Our data suggest that the second wave of demethylation, whichtemporally coincides with entry into the gonads, controls theexpression of several genes required for gametogenesis, as well ascontributing to imprint erasure, reactivation of the inactive Xchromosome and expression of IAP retrotransposons. Other aspectsof PGC differentiation may also be linked to DNA dimethylation;however, the lethality of Dnmt1 mutant embryos prior to 10.5 dpcprevents the examination postmigratory germ cell differentiationevents. Conditional deletion of Dnmt1 in PGCs might allow forfurther analysis of other changes that temporally overlap this waveof demethylation.

Cancer testis antigensEfforts to identify cancer-derived gene products as targets forimmunotherapy have revealed an association between genesnormally expressed only in germ cells, but ectopically activated intumors. Currently, 89 transcripts grouped into 44 families arerecognized as cancer testis (CT) antigens (Scanlan et al., 2004).Boon and colleagues (De Smet et al., 1996; De Smet et al., 2004)have demonstrated lower levels of promoter methylation in tumorsexpressing the MAGEA1 cancer testis antigen compared with non-expressing cells. Furthermore, MAGEA1 expression could beinduced in response to demethylating agents. This led to thesuggestion that the loss of DNA methylation that accompanies tumorprogression may be responsible for MAGE gene expression.

Koslowski et al. (Koslowski et al., 2004) reported that more thanhalf of CT genes are expressed in premeiotic germ cells and thatseveral could be induced in peripheral blood leukocytes by 5-azacytidine treatment. Similarly we found that several premeioticgerm cell-specific genes are expressed following loss of DNA

methylation, including Mageb4, the murine homolog of a human CTantigen. Our data provide direct in vivo evidence that premeioticgene expression is linked to hypomethylation, and provides a likelyexplanation for the frequent appearance of germ cell-specific genesin certain tumors.

Evolution of the germ cell lineageBoule and Vasa, the Drosophila homologs of Dazl and Mvh, wereoriginally identified as components of Drosophila germ plasm, andare highly conserved in germ cell development. While organismswith a mosaically determined germ line inherit these gene productsas maternal factors, Dazl and Mvh are expressed in postmigratorygerm cells in the mouse, 3-4 days after the germ line is specified.Interestingly, divergent mechanisms of germ cell specificationoperate within the amphibian class. The Xenopus germline ismosaically determined, while salamanders (Axolotl) specify germcells by an inductive mechanism similar to mammals (Johnson et al.,2003). As salamanders delay expression of axdazl and axvh untilgerm cells arrive at the gonad (Bachvarova et al., 2004), it would beinteresting to investigate the potential role of methylation in theexpression of axdazl and axvh. The observation that methylationwere to regulate expression of these genes in salamanders wouldsuggest that control of germ cell differentiation by DNA methylationmay be a widely conserved mechanism among species that useinductive signals to specify the germ cell lineage. Additionally, thiswould suggest that DNA methylation in the germ line initially aroseto regulate the timing of germ cell differentiation rather thanepigenetic processes such as genomic imprinting.

We gratefully acknowledge Kwon-Ho Hong, S. Paul Oh, Karen Johnstone,Chris Futtner, Andrew Johnson and George Enders for helpful advice andcomments. D.M.M. was supported by NIH Training Grant T32 CA09126. L.D.Kis a University of Florida Alumni Fellow. This work was supported by NIH GrantHD38429 to J.L.R.

ReferencesBachvarova, R. F., Masi, T., Drum, M., Parker, N., Mason, K., Patient, R. and

Johnson, A. D. (2004). Gene expression in the axolotl germ line: Axdazl, Axvh,Axoct-4, and Axkit. Dev. Dyn. 231, 871-880.

Caspary, T., Cleary, M. A., Baker, C. C., Guan, X. J. and Tilghman, S. M.(1998). Multiple mechanisms regulate imprinting of the mouse distalchromosome 7 gene cluster. Mol. Cell. Biol. 18, 3466-3474.

Clark, S. J., Harrison, J., Paul, C. L. and Frommer, M. (1994). High sensitivitymapping of methylated cytosines. Nucleic Acids Res. 22, 2990-2997.

Cooke, H. J., Lee, M., Kerr, S. and Ruggiu, M. (1996). A murine homologue ofthe human DAZ gene is autosomal and expressed only in male and femalegonads. Hum. Mol. Genet. 5, 513-516.

Coucouvanis, E. C., Sherwood, S. W., Carswell Crumpton, C., Spack, E. G.and Jones, P. P. (1993). Evidence that the mechanism of prenatal germ celldeath in the mouse is apoptosis. Exp. Cell Res. 209, 238-247.

De Felici, M., Dolci, S. and Pesce, M. (1992). Cellular and molecular aspects ofmouse primordial germ cell migration and proliferation in culture. Int. J. Dev.Biol. 36, 205-213.

De Smet, C., De Backer, O., Faraoni, I., Lurquin, C., Brasseur, F. and Boon, T.(1996). The activation of human gene MAGE-1 in tumor cells is correlated withgenome-wide demethylation. Proc. Natl. Acad. Sci. USA 93, 7149-7153.

De Smet, C., Loriot, A. and Boon, T. (2004). Promoter-dependent mechanismleading to selective hypomethylation within the 5� region of gene MAGE-A1 intumor cells. Mol. Cell. Biol. 24, 4781-4790.

Di Carlo, A., Travia, G. and De Felici, M. (2000). The meiotic specificsynaptonemal complex protein SCP3 is expressed by female and male primordialgerm cells of the mouse embryo. Int. J. Dev. Biol. 44, 241.

Donovan, P. J. and de Miguel, M. P. (2003). Turning germ cells into stem cells.Curr. Opin. Genet. Dev. 13, 463-471.

Donovan, P. J., Stott, D., Cairns, L. A., Heasman, J. and Wylie, C. C. (1986).Migratory and postmigratory mouse primordial germ cells behave differently inculture. Cell 44, 831-838.

Enders, G. C. and May, J. J. (1994). Developmentally regulated expression of amouse germ cell nuclear antigen examined from embryonic day 11 to adult inmale and female mice. Dev. Biol. 163, 331-340.

Fujiwara, T., Dunn, N. R. and Hogan, B. L. (2001). Bone morphogenetic protein

3417RESEARCH ARTICLEDNA methylation silences gonocyte genes

DEVELO

PMENT

3418

4 in the extraembryonic mesoderm is required for allantois development and thelocalization and survival of primordial germ cells in the mouse. Proc. Natl. Acad.Sci. USA 98, 13739-13744.

Fujiwara, Y., Komiya, T., Kawabata, H., Sato, M., Fujimoto, H., Furusawa, M.and Noce, T. (1994). Isolation of a DEAD-family protein gene that encodes amurine homolog of Drosophila vasa and its specific expression in germ celllineage. Proc. Natl. Acad. Sci. USA 91, 12258-12262.

Garcia-Castro, M. I., Anderson, R., Heasman, J. and Wylie, C. (1997).Interactions between germ cells and extracellular matrix glycoproteins duringmigration and gonad assembly in the mouse embryo. J. Cell Biol. 138, 471-480.

Gardiner-Garden, M. and Frommer, M. (1987). CpG islands in vertebrategenomes. J. Mol. Biol. 196, 261-282.

Geijsen, N., Horoschak, M., Kim, K., Gribnau, J., Eggan, K. and Daley, G. Q.(2004). Derivation of embryonic germ cells and male gametes from embryonicstem cells. Nature 427, 148-154.

Ginsburg, M., Snow, M. H. and McLaren, A. (1990). Primordial germ cells in themouse embryo during gastrulation. Development 110, 521-528.

Hahnel, A. C., Rappolee, D. A., Millan, J. L., Manes, T., Ziomek, C. A.,Theodosiou, N. G., Werb, Z., Pedersen, R. A. and Schultz, G. A. (1990). Twoalkaline phosphatase genes are expressed during early development in themouse embryo. Development 110, 555-564.

Hajkova, P., Erhardt, S., Lane, N., Haaf, T., El-Maarri, O., Reik, W., Walter, J.and Surani, M. (2002). Epigenetic reprogramming in mouse primordial germcells. Mech. Dev. 117, 15.

Hattori, N., Nishino, K., Ko, Y. G., Ohgane, J., Tanaka, S. and Shiota, K.(2004). Epigenetic control of mouse Oct-4 gene expression in embryonic stemcells and trophoblast stem cells. J. Biol. Chem. 279, 17063-17069.

Hubner, K., Fuhrmann, G., Christenson, L. K., Kehler, J., Reinbold, R., De LaFuente, R., Wood, J., Strauss, J. F., 3rd, Boiani, M. and Scholer, H. R.(2003). Derivation of oocytes from mouse embryonic stem cells. Science 300,1251-1256.

Ioshikhes, I. P. and Zhang, M. Q. (2000). Large-scale human promoter mappingusing CpG islands. Nat. Genet. 26, 61-63.

Jackson-Grusby, L., Beard, C., Possemato, R., Tudor, M., Fambrough, D.,Csankovszki, G., Dausman, J., Lee, P., Wilson, C., Lander, E. et al. (2001).Loss of genomic methylation causes p53-dependent apoptosis and epigeneticderegulation. Nat. Genet. 27, 31-39.

Jaenisch, R. (1997). DNA methylation and imprinting: why bother? Trends Genet.13, 323-329.

Jaenisch, R. and Bird, A. (2003). Epigenetic regulation of gene expression: howthe genome integrates intrinsic and environmental signals. Nat. Genet. 33, 245-254.

Johnson, A. D., Crother, B., White, M. E., Patient, R., Bachvarova, R. F.,Drum, M. and Masi, T. (2003). Regulative germ cell specification in axolotlembryos: a primitive trait conserved in the mammalian lineage. Philos. Trans. R.Soc. Lond. B Biol. Sci. 358, 1371-1379.

Koslowski, M., Bell, C., Seitz, G., Lehr, H. A., Roemer, K., Muntefering, H.,Huber, C., Sahin, U. and Tureci, O. (2004). Frequent nonrandom activation ofgerm-line genes in human cancer. Cancer Res. 64, 5988-5993.

Lane, N., Dean, W., Erhardt, S., Hajkova, P., Surani, A., Walter, J. and Reik,W. (2003). Resistance of IAPs to methylation reprogramming may provide amechanism for epigenetic inheritance in the mouse. Genesis 35, 88-93.

Lawson, K. A., Dunn, N. R., Roelen, B. A., Zeinstra, L. M., Davis, A. M.,Wright, C. V., Korving, J. P. and Hogan, B. L. (1999). Bmp4 is required for thegeneration of primordial germ cells in the mouse embryo. Genes Dev. 13, 424-436.

Lee, J., Inoue, K., Ono, R., Ogonuki, N., Kohda, T., Kaneko-Ishino, T., Ogura,A. and Ishino, F. (2002). Erasing genomic imprinting memory in mouse cloneembryos produced from day 11.5 primordial germ cells. Development 129,1807-1817.

Lees-Murdock, D. J., De Felici, M. and Walsh, C. P. (2003). Methylationdynamics of repetitive DNA elements in the mouse germ cell lineage. Genomics82, 230-237.

Lei, H., Oh, S. P., Okano, M., Juttermann, R., Goss, K. A., Jaenisch, R. and Li,E. (1996). De novo DNA cytosine methyltransferase activities in mouseembryonic stem cells. Development 122, 3195-3205.

Li, E., Bestor, T. and Jaenisch, R. (1992). Targeted mutation of the DNAmethylatrasferase gene results in embryonic lethality. Cell 69, 915-926.

Li, E., Beard, C. and Jaenisch, R. (1993). Role for DNA methylation in genomicimprinting. Nature 366, 362-365.

Maatouk, D. M. and Resnick, J. L. (2003). Continuing primordial germ celldifferentiation in the mouse embryo is a cell-intrinsic program sensitive to DNAmethylation. Dev. Biol. 258, 201-208.

Mann, M. R., Chung, Y. G., Nolen, L. D., Verona, R. I., Latham, K. E. andBartolomei, M. S. (2003). Disruption of imprinted gene methylation andexpression in cloned preimplantation stage mouse embryos. Biol. Reprod. 69,902-914.

Martinowich, K., Hattori, D., Wu, H., Fouse, S., He, F., Hu, Y., Fan, G. andSun, Y. E. (2003). DNA methylation-related chromatin remodeling in activity-dependent BDNF gene regulation. Science 302, 890-893.

Matsui, Y., Zsebo, K. and Hogan, B. L. (1992). Derivation of pluripotentialembryonic stem cells from murine primordial germ cells in culture. Cell 70, 841-847.

McLaren, A. (1983). Primordial germ cells in mice. Bibl. Anat. 24, 59-66.McLaren, A. (1984). Meiosis and differentiation of mouse germ cells. Symp. Soc.

Exp. Biol. 38, 7-23.McLaren, A. and Southee, D. (1997). Entry of mouse embryonic germ cells into

meiosis. Dev. Biol. 187, 107-113.Milutinovic, S., Zhuang, Q., Niveleau, A. and Szyf, M. (2003). Epigenomic

stress response. Knockdown of DNA methyltransferase 1 triggers an intra-S-phase arrest of DNA replication and induction of stress response genes. J. Biol.Chem. 278, 14985-14995.

Monk, M. and McLaren, A. (1981). X-chromosome activity in foetal germ cells ofthe mouse. J. Embryol. Exp. Morphol. 63, 75-84.

Mroz, K., Carrel, L. and Hunt, P. A. (1999). Germ cell development in the XXYmouse: Evidence that X chromosome reactivation is independent of sexualdifferentiation. Dev. Biol. 207, 229.

Ohkubo, Y., Shirayoshi, Y. and Nakatsuji, N. (1996). Autonomous regulation ofproliferation and growth arrest in mouse primordial germ cells studied by mixedand clonal cultures. Exp. Cell Res. 222, 291-297.

Osterlund, C., Tohonen, V., Forslund, K. O. and Nordqvist, K. (2000). Mage-b4, a novel melanoma tumor antigen (MAGE) gene specifically expressed duringgerm cell differentiation. Cancer Res. 60, 1054-1061.

Pesce, M. and De Felici, M. (1995). Purification of mouse primordial germ cells byMiniMACS magnetic separation system. Dev. Biol. 170, 722-725.

Pohlers, M., Truss, M., Frede, U., Scholz, A., Strehle, M., Kuban, R. J.,Hoffmann, B., Morkel, M., Birchmeier, C. and Hagemeier, C. (2005). A rolefor E2F6 in the restriction of male-germ-cell-specific gene expression. Curr. Biol.15, 1051-1057.

Resnick, J. L., Bixler, L. S., Cheng, L. and Donovan, P. J. (1992). Long-termproliferation of mouse primordial germ cells in culture. Nature 359, 550-551.

Richards, A. J., Enders, G. C. and Resnick, J. L. (1999). Differentiation of murinepremigratory primordial germ cells in culture. Biol. Reprod. 61, 1146-1151.

Rollins, R. A., Haghighi, F., Edwards, J. R., Das, R., Zhang, M. Q., Ju, J. andBestor, T. H. (2006). Large-scale structure of genomic methylation patterns.Genome Res. 16, 157-163.

Sado, T., Fenner, M. H., Tan, S. S., Tam, P., Shioda, T. and Li, E. (2000). Xinactivation in the mouse embryo deficient for Dnmt1: distinct effect ofhypomethylation on imprinted and random X inactivation. Dev. Biol. 225, 294-303.

Scanlan, M. J., Simpson, A. J. and Old, L. J. (2004). The cancer/testis genes:review, standardization, and commentary. Cancer Immun. 4, 1.

Seki, Y., Hayashi, K., Itoh, K., Mizugaki, M., Saitou, M. and Matsui, Y. (2005).Extensive and orderly reprogramming of genome-wide chromatin modificationsassociated with specification and early development of germ cells in mice. Dev.Biol. 278, 440-458.

Storre, J., Schafer, A., Reichert, N., Barbero, J. L., Hauser, S., Eilers, M. andGaubatz, S. (2005). Silencing of the meiotic genes SMC1beta and STAG3 insomatic cells by E2F6. J. Biol. Chem. 280, 41380-41386.

Szabo, P. E., Hubner, K., Scholer, H. and Mann, J. R. (2002). Allele-specificexpression of imprinted genes in mouse migratory primordial germ cells. Mech.Dev. 115, 157-160.

Takeuchi, A., Mishina, Y., Miyaishi, O., Kojima, E., Hasegawa, T. and Isobe,K. (2003). Heterozygosity with respect to Zfp148 causes complete loss of fetalgerm cells during mouse embryogenesis. Nat. Genet. 33, 172-176.

Tam, P. P. and Snow, M. H. (1981). Proliferation and migration of primordial germcells during compensatory growth in mouse embryos. J. Embryol. Exp. Morphol.64, 133-147.

Tam, P. P., Zhou, S. X. and Tan, S. S. (1994). X-chromosome activity of the mouseprimordial germ cells revealed by the expression of an X-linked lacZ transgene.Development 120, 2925-2932.

Toyooka, Y., Tsunekawa, N., Akasu, R. and Noce, T. (2003). Embryonic stemcells can form germ cells in vitro. Proc. Natl. Acad. Sci. USA 100, 11457-11462.

Walsh, C. P. and Bestor, T. H. (1999). Cytosine methylation and mammaliandevelopment. Genes Dev. 13, 26.

Walsh, C. P., Chaillet, J. R. and Bestor, T. H. (1998). Transcription of IAPendogenous retroviruses is constrained by cytosine methylation. Nat. Genet. 20,116-117.

Wang, D. H., Ikeda, Y., Parker, K. L. and Enders, G. C. (1997). Germ cell nuclearantigen (GCNA1) expression does not require a gonadal environment orsteroidogenic factor 1, Examination of GCNA1 in ectopic germ cells and in Ftz-F1 null mice. Mol. Reprod. Dev. 48, 154-158.

Wang, R. A., Nakane, P. K. and Koji, T. (1998). Autonomous cell death of mousemale germ cells during fetal and postnatal period. Biol. Reprod. 58, 1250.

Ying, Y. and Zhao, G. Q. (2001). Cooperation of endoderm-derived BMP2 andextraembryonic ectoderm- derived BMP4 in primordial germ cell generation inthe mouse. Dev. Biol. 232, 484-492.

Ying, Y., Liu, X. M., Marble, A., Lawson, K. A. and Zhao, G. Q. (2000).Requirement of Bmp8b for the generation of primordial germ cells in themouse. Mol. Endocrinol. 14, 1053-1063.

RESEARCH ARTICLE Development 133 (17)