blimp1 is a critical determinant of the germ cell lineage in mice

8/9/2019 Blimp1 is a critical determinant of the germ cell lineage in mice

1/7

Blimp1 is a critical determinant of thegerm cell lineage in miceYasuhide Ohinata 1*, Bernhard Payer 2*, Donal OCarroll 3*, Katia Ancelin 2 , Yukiko Ono 1 , Mitsue Sano 1 ,Sheila C. Barton 2, Tetyana Obukhanych 4 , Michel Nussenzweig 4 , Alexander Tarakhovsky 3 , Mitinori Saitou 1,5,6

& M. Azim Surani 2

Germ cell fate in mice is induced in pluripotent epiblast cells in response to signals from extraembryonic tissues. Thespecication of approximately 40 founder primordial germ cells and their segregation from somatic neighbours areimportant events in early development. We have proposed that a critical event during this specication includesrepression of a somatic programme that is adopted by neighbouring cells. Here we show that Blimp1 (also known asPrdm1), a known transcriptional repressor, has a critical role in the foundation of the mouse germ cell lineage, as itsdisruption causes a block early in the process of primordial germ cell formation. Blimp1-decient mutant embryos form atight cluster of about 20 primordial germ cell-like cells, which fail to show the characteristic migration, proliferation andconsistent repression of homeobox genes that normally accompany specication of primordial germ cells. Furthermore,our genetic lineage-tracing experiments indicate that the Blimp1-positive cells originating from the proximal posteriorepiblast cells are indeed the lineage-restricted primordial germ cell precursors.

Primordial germ cells (PGCs) are the source of totipotency, aunique state generated by this lineage. Germ cells have many exceptional properties, including the potential for extensive epi-genetic reprogramming of the genome 1,2 . A detailed understandingof germ cell properties, together with the mechanisms of germ cellspecication and their segregation from somatic neighbours, could

greatly advance our knowledge of epigenetic mechanisms thatregulate cellular differentiation, genome reprogramming and stemcell biology.

Previous studies have shown that germ cell fate in mice is imposedupon pluripotent epiblast cells, which are also the source of allsomatic cells36 . The competence to form PGCs in mice is not aninherited property, but is induced in the proximal epiblast cells by signals from extraembryonic tissues 58 . We have previously shownthat germ cell competence results in the upregulation of fragilis(also known as interferon-induced transmembrane protein 3 of Itm3) in the proximal epiblast at embryonic day (E)6.5 (ref. 9).Subsequently, approximately 40 cells acquire PGC fate at aboutE7.5, as detected by expression of stella (also known as develop-mental pluripotency-associated 3 or Dppa3), the earliest knownmarker of founder PGCs 9,10 . A key event leading to PGC speci-cation also involves repression of homeobox ( Hox ) genes9 .Similar repression of the somatic programme is commonly observed in the emerging germ cell lineage in a number of organisms, but the underlying molecular mechanisms involveddiffer markedly 1114 . Here we describe the critical role of Blimp1, aknown transcriptional repressor 1517 , during PGC specication inmice. We also show that Blimp1-positive cells constitute lineage-restricted PGC precursors much earlier in development thanpreviously thought 5.

Blimp1 expression marks the origin of PGCsTo identify genes with a key role during specication and segregationof PGCs from their somatic neighbours, we performed differentialscreen and expression analyses for candidates, including histonemethyltransferases, using single-cell complementary DNAs fromE7.5 founder PGCs and neighbouring somatic cells 9. Among thecandidates, B-lymphocyte-induced maturation protein-1 ( Blimp1)showed remarkably specic expression in PGCs but not in somaticcells (Supplementary Fig. 1 and data not shown). Blimp1 encodes atranscriptional repressor with a SET domain and Kru ppel-type zinc-ngers. It drives terminal differentiation of B cells into immuno-globulin-secreting plasma cells 15,16 by repressing the mature B-cellgene expression programme 17 . Blimp1 is widely expressed duringdevelopment, including in migrating germ cells 18 . We thereforedecided to investigate this gene for its role in PGC specication.

We rst used in situ hybridization to examine the expression of Blimp1 in late streak stage embryos (E7.25) and observed a strongsignal in the posterior-proximal extraembryonic mesoderm, wherefounder PGCs reside (Fig. 1a), as well as in the visceral endoderm 18 .Compared with fragilis19 , Blimp1 expression appears to be morerestricted and punctate. Furthermore, we detected 2025 Blimp1-positivecells,and only 28 stella-positivecellsat this stage. Single-cellcDNA analysis conrmed that Blimp1 expression precedes that of stella (see Supplementary Fig. 1a). However, at the early bud stage,stella expression was fairly consistently detected in the Blimp1-positive cells. These ndings led us to reason that PGC specicationprogresses in the Blimp1-positive cells that go on to express stella.

Blimp1-positive cells in the early epiblastAlthough we initially detected Blimp1 expression in the visceral

ARTICLES

1Laboratory for Mammalian Germ Cell Biology, Center for Developmental Biology, RIKEN Kobe Institute, 2-2-3 Minatojima-minamimachi, Chuo-ku, Kobe, Hyogo 650-0047Japan. 2 Wellcome Trust/Cancer Research UK Gurdon Institute of Cancer and Developmental Biology, University of Cambridge, Tennis Court Road, Cambridge, CB2 1QN, UK.3 The Laboratory for Lymphocyte Signaling and 4 The Laboratory of Molecular Immunology, Howard Hughes Medical Institute, The Rockefeller University, 1230 York Avenue, NewYork, New York 10021, USA. 5 Precursory Research for Embryonic Science and Technology, Japan Science and Technology Agency, 4-1-8 Honcho, Kawaguchi, Saitama 332-0012,Japan. 6 Laboratory of Molecular Cell Biology and Development, Graduate School of Biostudies, Kyoto University, Oiwake-cho, Kitashirakawa, Sakyo-ku, Kyoto 606-8502, Japan.*These authors contributed equally to this work.

Vol 436 |14 July 2005 |doi:10.1038/nature03813

207

!""# Nature Publishing Group


2/7

endoderm (an extraembryonic tissue layer) at E5.5, the earliestexpression in the embryo proper was seen at E6.25 (an unexpectedly early stage) in the most proximal layer of the epiblast (Fig. 1b).Notably, these Blimp1-positive cells were restricted to only the futureposterior side of the shorter embryonic axis (Fig. 1c, d), which hasrecently been shown to form the primitive streak after dynamic

morphological rearrangement 20,21 . This highly restricted location isconsistent with the expression of other posterior markers 22 inBlimp1-positive cells at the mid-streak stage, such as nodal( Nodal ), cripto ( Cfc1), Fgf8, brachyury (T ) and Evx1 (refs 2022and M.S., unpublished observations). Although fragilis is alsoexpressed in the most proximal epiblast in the pre/early streak embryos 9,19 , it is detected in the entire rim of the proximal epiblast(both anterior and posterior), and in cells that were 35 layers

deep (see Supplementary Fig. 1b). In contrast, the single layer of Blimp1-positive epiblast cells (Fig. 1c, d) is probably in direct contactwith the extra-embryonic ectoderm. At the early streak stage, a few cells in the nascent mesodermal region also started to show Blimp1expression (Fig. 1d), and in slightly later embryos undergoing

Figure 1 | Expression of Blimp1 in gastrulating embryos. a, fragilis, Blimp1and stella expression at the late streak stage. Anterior is to the left. Redarrowheads indicate the location of founder PGCs and the black arrowheadindicates Blimp1 expression in the visceral endoderm. Scale bar, 100 mm.b , Blimp1 expression between pre-streak and early bud stage. Top, lateral view; bottom, posterior view. Scale bar, 200 mm. PS,pre-streak; 0S,no streak;ES, early streak; MS, mid-streak; LS, late streak; 0B, no bud; EB, early bud.c, d , Consecutive 8- mm sectionsof the 0S( c)andES( d) embryosshownin b .The left position of the long embryonic axis corresponds to the futureanterior (see Fig.2k). Blimp1 is detected in thetop layer of proximal epiblastcells at one side of the embryo (red arrowheads). Expression in the visceralendoderm (black arrowheads) is primarily on the opposite side. Scale bar ind , 100 mm.

Figure 2 | Origin and allocation of Blimp1mEGFP cells. ad , Lateral viewsof pre/no streak ( a , b) and early streak ( c, d) embryos stained withphalloidine (red in b , d). Expression of Blimp1-mEGFP is in the mostproximal epiblast cells in contact with extra-embryonic ectoderm (whitearrowheads in a and c). e j, Transverse views of pre/no-streak ( e , f), early streak ( g, h) and mid-streak ( i, j) embryos with GFP expression (yellow) inthe epiblast (arrowheads) and visceral endoderm (arrows). DAPI (nuclear)staining is shown in cyan. Scale bars, 50 mm. k, Rotation of the embryonicaxis and distribution of Blimp1-positive cells during gastrulation. Blimp1-positive cells (red), Blimp1-negative epiblast cells (green), nascentmesoderm (yellow) and visceral endoderm (blue) are shown. A, anterior;P, posterior.

ARTICLES NATURE |Vol 436 |14 July 2005

208



3/7

extensive gastrulation movement, the signal was mostly observed inthe mesodermal region of the primitive streak (data not shown).

To gain further insight into the origin of the Blimp1-positive cells,we used two independent transgenic mouse lines generated using amodied 230-kilobase (kb) bacterial articial chromosome (BAC)expressing membrane-targeted enhanced green uorescence protein(mEGFP) under the control of Blimp1 regulatory elements (seeSupplementary Fig. 2). The expression pattern of the reporter

transgenes was indistinguishable from in situ hybridization analysis(see Supplementary Fig. 2bd). Serial confocal images of transverse

sections of pre/no-streak stage embryos conrmed the presence of uorescent cells at one side of the most proximal layer of the epiblast,which is in direct contact with the extra-embryonic ectoderm(Fig. 2a, b, e, f, k). During subsequent development, the number of Blimp1-positive cells apparently increased (Fig. 2c, d, g, h, k), and atthe mid-streak stage, the uorescent cells were detected as a cluster atthe posterior end of the embryo (Fig. 2ik). These cells developed acytoplasmic spot in the presumptive Golgi apparatus, which is a

characteristic associated with PGCs23

(Supplementary Fig. 2d anddata not shown).We next determined the numbers of Blimp1-positive cells at each

stage of development. In pre/no-streak embryos (E6.25), we detectedabout six uorescent cells (Fig. 3a), which increased to 16 positivecells at the early streak stage (E6.5), including a few within thenascent mesoderm (Figs 2g and 3a). The relatively abrupt increase intheir numbers (Fig. 3a) might be due to a lag period necessary forgenerating detectable levels of EGFP. These cells form a tight clusterof approximately 2028 cells at the mid- to late-streak stage (E7.25,Fig. 2ik), from which tissue non-specic alkaline phosphatase(TNAP)- and stella-positive cells may eventually emerge. In early

Figure3 | Blimp1-positive cellsbecome TNAP- and stella-positivePGCs. a,Blimp1mEGFP cell numbers at different embryonic stages using twotransgenic mouse lines (with either8 copies(blue) or 1 copy (red) of Blimp1-mEGFP ). bd , Base of theallantois regionof early budstageembryos stainedfor TNAP ( b , red) and Blimp1mEGFP ( c, green). A merged image is shownin d. Scale bar, 50 mm. Blimp1mEGFP cells overlap with TNAP-positivePGCs, with very few exceptions (arrow in c and d). eg, Posterior view of embryos at early head fold stage stained for stella ( e , red) and Blimp1mEGFP ( f, green). A merged image (with DAPIstaining in blue) is shown ing. Scale bar, 100 mm. Arrowheads denote visceral endoderm. h , Almost allstella-positive cells are also Blimp1mEGFP positive.

Figure 4 | Blimp1-positive cells are lineage-restricted PGCs. a, Diagramshowing the genetic cross for lineage tracing experiments. bd, Late budstage embryo (E7.8) stained with anti-GFP ( b , green) and anti-stella ( c, red)antibodies. Lateral confocal sections showing the base of the allantois. Amerged image (with DAPI staining in blue) is shown in d . Scale bar, 100 mm.All of the GFP-positive cells are stella-positive, except in the visceralendoderm (arrowheads). e , PGC cell counts (stella-positive) and theproportion of GFP-positive PGCs. f, Cell fates observed in this experiment.Blimp1-negative cells (grey) become somatic cells. Cells expressing Blimp1and with successful excision of the oxed Stop (green) become stella- andGFP-positive PGCs (yellow). Cells that are Blimp1-positive but Cre-negative(pink) later become stella-positive but remain GFP-negative PGCs (red).

NATURE |Vol 436 |14 July 2005 ARTICLES

209



4/7

bud stage embryos (E7.5) that begin to form the allantois, weobserved a cluster of approximately 40 Blimp1-positive cells withthe cytoplasmic spots characteristic of PGCs (Fig. 3a and Sup-plementary Fig. 2hj). We did not detect any Blimp1-mEGFP cellsin the allantoic mesoderm (the closest somatic relative of PGCs), the yolk sac mesoderm or blood islands. We also note an absence of Blimp1-mEGFP-positive cells in mice mutant for bone morpho-genetic protein 4 ( Bmp4; data not shown), which is consistent with a

lack of PGCs in these mutants7,24

. However, wedo notcurrentlyknow whether there is a relationship between Bmp4 and Blimp1 expressionin the epiblast.

Blimp1-positive cells are lineage-restricted PGCsNext, we examined whether the Blimp1-positive cells are lineage-restricted to give rise to only PGCs. First, we found that all but a few Blimp1-positive cells were co-stained with TNAP (Fig. 3bd), a

classical marker of founder PGCs, at the early bud stage 23 and alsoat the mid bud stage (see Supplementary Fig. 3). Furthermore, at theearly head fold stage, we found an almost complete overlap betweenBlimp1-mEGFP-positive cells and cells that were co-stained with ananti-stella antibody (Fig. 3eg), which is a denitive marker of newly established founder PGCs 9,10,25 . Counts of PGCs in individualembryos showed that 94100% of cells showed expression of bothmarkers, suggesting that the Blimp1-mEGFP-positive cells in the

mesoderm are indeed PGCs (Fig. 3h).To rmly establish that the Blimp1-positive cells are lineage-restricted PGCs, we performed a genetic lineage tracing experiment.We created Blimp1-Cre-BAC transgenic mice, which we crossedwith mice containing a ROSA26 EGFP f reporter 26 . In doing so,Blimp1-Cre-expressing cells and their clonal descendents wouldbecome permanently and genetically marked by the onset of GFPexpression following Cre-mediated deletion of the Stop sequence in

Figure 5 | Loss of PGCs in Blimp1 decient embryos. a, TNAP-positivefounder-PGCs at E7.5 (early bud to late head fold stage) in wild-type (WT;n 4), Blimp1 heterozygote ( /2 ; n 8) and null ( 2 /2 ; n 10)embryos. Single asterisk, P 0.016 for WT versus Blimp1 /

2; double

asterisk, P 0.002 for WT versus Blimp12 / 2 embryos. b, Linear regression

analysis of log PGC number versus somite number at E8.5 for WT (bluesolid line, y 0.1134 x 1.224; n 6) and Blimp1 /

2(green dashed line,

y 0.0763 x 1.098; n 16) embryos. cf, Wild-type ( c) and Blimp12 / 2

(d) embryos at E8.5 (23 somites, lateral view, anterior to the left) and

corresponding TNAPstaining ( e , f, posterior view) revealmigratingPGCs in wild-type embryos ( e , white arrows) but only a few clustered cells inBlimp1

2 / 2 mutant embryos ( f, black arrow). Al, allantois. Scale bars,200 mm. g, h , PGCs in embryos generated from control and Blimp1

2 / 2

embryonic stem cells. g, stella-positive PGCs in E7.5 control (early bud tolate head fold stage; n 11) and Blimp1

2 / 2 embryos ( n 10). Asterisks,P 0.0001. h , Comparison of PGCs detected by TNAP staining at E8.5E9.5in control embryos ( n 14) versus Blimp1

2 / 2 embryos ( n 22). See alsoSupplementary Fig. 5.


210



5/7

front of the ROSA26 EGFP reporter (Fig. 4a, f ). The resultingembryos from this cross were collected between the late bud (E7.8)and 3-somite stage, and double-stained with anti-GFP and anti-stellaantibodies (Fig. 4bd). We found that 5576% of the PGCs (deter-mined by stella expression) were also GFP-positive (Fig. 4e). Thereason why not all PGCs were GFP-positive is because of incompleteBlimp1-Cre-mediated deletion, which is also reported to occur withother Cre-lines 26,27 . However, all of the GFP-positive cells (with the

exception of one) were also stella-positive (Fig. 4e), suggesting thatBlimp1-positive cells constitute germline-restricted progenitors of PGCs. Our combined results show that Blimp1-positive cells give riseto PGCs, which unequivocally indicates their early lineagerestriction.

Mechanism and role of Blimp1 in PGC specicationTo explore whether Blimp1 is indeed necessary for PGC specication,wecreateda null allelein which exon5 ofthe Blimp1 gene wasankedby loxP sites andsubsequently deleted (see SupplementaryFig. 4). Allembryos of different genotypes generated from heterozygous inter-crosses were apparently normal until at least E8.5 (Fig. 5cf and datanot shown) and were present in an appropriate mendelian ratio.

Loss of founder PGCs in Blimp1 mutantsTo examine the fate of PGCs in these Blimp1 mutant embryos, westained them for TNAP 23 and counted the clearly identiable PGCscontaining a cytoplasmic spot. At the early bud and early head foldstages of late E7.5 embryos, we detected approximately 25 TNAP-positive PGCs in wild-type embryos. These numbers were reduced toabout 17.5 cells in heterozygotes ( Blimp1 /

2), and a maximum of

ve PGCs were counted in individual Blimp1 null (Blimp12 / 2 )

embryos (Fig. 5a). The relative differences in the number of PGCsin control and mutant embryos are statistically signicant, but thetotal number of PGCs is an underestimate because at these earlierstages, many of the PGCs still form a tight cluster inwhich individualcells are difcult to count (see below, including Figs 5g and 6a, forfurther analysis).

At E8.5 (Fig. 5b), the relative numbers of PGCs in homozygousmutants were also very low when compared with control mice. Only a few PGC-like cells were detected near the base of the allantois, andvirtually none seemed to be migrating appropriately. Analysis of thenumbers of PGCs shows no signicant differences in the slopes of theregression lines when comparing wild-type and heterozygousembryos ( P 0.581) (ref. 7), suggesting that the dose-dependenteffect of Blimp1 is primarily on the genesis of the founder PGCpopulation and not on their subsequent survival or proliferation. Arecent study of Blimp1 mutant embryos conrms this observation 28 .Furthermore, we did not detect any abnormalities of the allantois inE8.5 embryos (Fig. 5e, f), suggesting that the Blimp1-positivecellsaredestined for the germ cell fate and are set apart from other epiblastcells that give rise to the allantois and other somatic tissues.

To conrm that the effect of the Blimp1 mutation on germ cellformation is direct and not a consequence of its expression in thevisceral endoderm (which is important for early embryonic pattern-ing 22 ), we performed tetraploid rescue experiments in whichBlimp1

2 /2 embryonic stem cells were injected into wild-type hosttetraploid blastocysts. The tetraploid host cells contribute almostexclusively to extraembryonic tissues, including the visceral endo-derm, while the injected ES cells only contribute to the embryoproper 29 . We determined the number of PGCs in E7.5 embryos using

an anti-stella antibody 10

, and detected an average of 35 PGCs incontrol embryos and 6.5 stella-positive cells in Blimp12 / 2 embryos(P 0.0001) (Fig. 5g). The numbers of TNAP-positive cells atE8.5E9.5 (Fig. 5h and Supplementary Fig. 5) were also signicantly lower in mutants compared with control wild-type mice. The few TNAP-positive PGC-like cells that we detected near the base of theallantois did not seem to be migrating appropriately in the mutants.Taken together, these results demonstrate that the PGC-intrinsicactivity of Blimp1 is essential for the formation of PGCs, and cannotbe substituted for by the expression of Blimp1 within the visceralendoderm.

Aberrant PGC-like cells in Blimp1 decient miceWe investigated the phenotypic characteristics of PGC-like cells inmutant embryos at the earlier E7.5 stages (early bud to late bud)

using TNAP staining and confocal microscopy (Fig. 6a). At the early budstage, all theembryos showeda TNAP-positivecluster at thebaseof the allantois, but only in the wild-type embryos were a few cellsmigrating out of the cluster. Thesephenotypic differences weremuchmore prominent at the late bud stage, when most of the TNAP-positive cells containing cytoplasmic spots had started migrating incontrol embryos. In contrast, the cluster in mutant embryosremained compact and tight, with only a rare migrating cell(Fig. 6a). Furthermore, the number of TNAP-positive cells in controlembryos increased from approximately 30 to 58 between the early bud and late bud stages, but their numbers in mutants rarely exceeded 20 at similar stages of development (data not shown); atight cluster of cells is visible even at E8.5 (see Fig. 5f). It thereforeseems that in mutant embryos, the number of cells in the clustermight not increase from the early bud stage onwards, and perhaps

Figure 6 | Aberrant phenotype of Blimp1 decient cells. a, Posterior viewsshowing uorescent TNAP staining of Blimp1 wild-type ( / ),heterozygote ( /2 ) and homozygote ( 2 /2 ) embryos. Arrowheads point toPGCs migrating out from the cluster. Scale bar, 50 mm. b, Single-cell cDNAanalysis reveals aberrant gene expression in Blimp1

2 / 2 PGC-like cells atE7.5 (early bud to late bud stage). PGCs were identied by expression of stella and Blimp1 transcripts. The mutant Blimp1 transcript (denoted by anasterisk) served as a marker only and does not giverise to functional protein(see Supplementary Fig. 4e).


211



6/7

even from the mid-streak stage onwards. Although the number of TNAP-positive cells in heterozygote embryos was similar to those inthe null embryos, these cells were more like those in control embryosin that they also seemed to be dispersed and migrating (Fig. 6a).These ndings indicate that in the absence of Blimp1, TNAP-positivePGC-like cells areformed, but they areaberrantby theearly budstageand they do not proliferate or migrate properly.

Blimp1 null cells fail to repress Hox genesAmong theTNAP-positive cells in mutant embryos, we detected very few stella-positive cells (Fig. 5g); these cells probably represent thegreatest extent to whichmutant cells areable to develop any PGC-likecharacteristics. Using embryos generated from embryonic stem cellsin tetraploid hosts, we used single-cell cDNA analysis to comparemutant stella-positive PGC-like cells with cells from controlembryos 9 (Fig. 6b). In control embryos at mid-bud to late budstages, we identied 7/60 cells that were Blimp1- and stella-positive(Fig. 6b). Most of these showed characteristic repression of bothHoxa1 (5/7) and Hoxb1 (6/7). However, we also detected four cellsthat were Blimp1-positive and stella-negative, and which showedexpression of Hoxa1 and Hoxb1 (data not shown). These cells mightbe at an earlier stage of germ cell development and have yet to inducerepression of Hox genes, which reinforces the notion that repressionof Hox genes and expression of stella are temporally linked events inPGCs9. Almost all somatic cells that were negative for Blimp1 alsoshowed expression of these Hox genes (not shown), which isconsistent with the observation that Hoxb1 expression starts in themesodermal cells at about the mid-to-late-streak stage 30 .

From mutant embryos, we detected 5/160 of the relatively rarestella-positive cells expressing the Blimp1 mutant transcript (seeSupplementary Fig. 4e) at the early bud to late bud stage. Notably,4/5 of these cells showed expression of Hoxa1 and 3/5 showed Hoxb1expression (Fig. 6b), indicating that the repression of Hox genes hadnot occurred consistently in these cells. In contrast, somatic cellsfrom both the control and mutant embryos showed Hox expressionandwere therefore unaffected at least in this respect. Furthermore, weobserved that although PGCs from control embryos consistently

expressed Sox2 (7/7) and Nanos3 (6/7; Fig. 6b and M.S. and M.A.S.,unpublished observations), 2/5 mutant cells did not show thisexpression pattern. Therefore, stella-positive mutant cells seem tobe perturbed in several aspects compared to wild-type PGCs. They show inconsistent gene expression patterns, fail to proliferate (Figs 5and 6a), and may eventually undergo apoptosis or adopt a somaticcell fate.

DiscussionOur study shows that lineage-restricted PGCs are Blimp1-positivecells that are rst detected in the proximal epiblast and accumulatesubsequently. The mechanisms underlying the initiation of Blimp1expression in epiblast cells and the accretion of additional Blimp1-positive cells from E6.5 onwards are unknown. These cells, whichform a cluster of approximately 20 cells at the mid-streak stage, could

generate the founder PGCs in a single round of cell division by theearly bud stage. In contrast, a previous study on PGC specicationusing clonal analysis showed that at the E6.5 early streak stage,descendants of single marked cells that contribute to PGCs alwaysinclude oneor twodistinct somatic lineages, most predominantly theallantoic mesoderm 5. Here, we did not detect a contribution of Blimp1-positive cells to the allantois or to any other mesodermalcells. It is possible that theclonal analysis in ref. 5 failedto mark oneof the few Blimp1-positive PGC lineage-restricted cells at E6.5 (Fig. 3a).Assuming this to be the case, after one or more divisions of themarked cells, some of them might subsequently proceed towards thePGC fate after expression of Blimp1, while the other descendentscontribute to the somatic lineage. Further experiments are needed totrace the fate of the initial Blimp1-positive cells before the nal PGCspecication event.

Our study shows that mutation of the Blimp1 gene affects thelineage-restricted PGCprecursors at a stage that is earlier thanthe early bud stage, because unlike the cluster in control embryos (where thecells proliferate and subsequently migrate out as PGCs), almost all thecells in mutant embryos fail to proliferate or exit from the cluster. Thefew stella-positive cells in mutants are also aberrant as they show inconsistent repression of Hox genes, which accompanies thenal stages of PGC specication in normal embryos. The precise

molecular mechanism of Blimp1 in PGC specication remains to beelucidated.

METHODSEmbryo isolationand single-cell cDNA preparationwere performed as describedpreviously 9,31 . For comparisons between mutant and control embryos (Fig. 6),mutant embryos were generated from Blimp1

2 / 2 embryonic stem cells, andloxP/loxP embryonic stem cell-derived or normal wild-type embryos were usedas controls for the isolation of single cells.

In situ hybridization was performed as described previously 32 . Antibody stainings on dispersed E7.5 germ cell regions and on whole embryos werecarried out essentially as described previously 25,33 . TNAP staining of PGCs wasperformed as described previously 7,34 .

PGC numbers at E7.5 were compared using the nonparametric Mann-Whitney U -test and linear regression lines were compared using analysis of covariance (ANCOVA) with GraphPad Prism software .

More detailed methods, concerning particularly the generation of transgenicand knockout mice described in this paper, can be found in the Supplementary Information. The Blimp1 mutant mice, Blimp-1-mEGFP mice and Blimp1-Cremice are available from A.T., M.S. and M.N., respectively.

Received 17 March; accepted 10 May 2005.Published online 5 June 2005.

1. Surani, M. A. Reprogramming of genome function through epigeneticinheritance. Nature 414, 122-128 (2001).

2. Li, E. Chromatin modication and epigenetic reprogramming in mammaliandevelopment. Nature Rev. Genet. 3, 662-673 (2002).

3. Gardner, R. L. & Rossant, J. Investigation of the fate of 4-5 day post-coitummouse inner cell mass cells by blastocyst injection. J. Embryol. Exp. Morphol. 52,141-152 (1979).

4. Lawson, K. A., Meneses, J. J. & Pedersen, R. A. Clonal analysis of epiblast fateduring germ layer formation in the mouse embryo. Development 113, 891-911(1991).

5. Lawson, K. A. & Hage, W. J. Clonal analysis of the origin of primordial germcells in the mouse. Ciba Found. Symp. 182, 68-84 (1994).

6. McLaren, A. Primordial germ cells in the mouse. Dev. Biol. 262, 1-15 (2003).7. Lawson, K. A. et al. Bmp4 is required for the generation of primordial germ

cells in the mouse embryo. Genes Dev. 13, 424-436 (1999).8. Gardner, R. L., Lyon, M. F., Evans, E. P. & Burtenshaw, M. D. Clonal analysis of

X-chromosome inactivation and the origin of the germ line in the mouseembryo. J. Embryol. Exp. Morphol. 88, 349-363 (1985).

9. Saitou, M., Barton, S. C. & Surani, M. A. A molecular programme for thespecication of germ cell fate in mice. Nature 418, 293-300 (2002).

10. Sato, M. et al. Identication of PGC7, a new gene expressed specically inpreimplantation embryos and germ cells. Mech. Dev. 113, 91-94 (2002).

11. Surani, M. A. et al. Mechanism of mouse germ cell specication: a geneticprogram regulating epigenetic reprogramming. Cold Spring Harbor Symp. Quant.Biol. 69, 1-10 (2004).

12. Seydoux, G. & Strome, S. Launching the germline in Caenorhabditis elegans :regulation of gene expression in early germ cells. Development 126, 3275-3283(1999).

13. Blackwell, T. K. Germ cells: nding programs of mass repression. Curr. Biol. 14,R229-R230 (2004).

14. Santos, A. C. & Lehmann, R. Germ cell specication and migration in Drosophilaand beyond. Curr. Biol. 14, R578-R589 (2004).

15. Turner, C. A. Jr, Mack, D. H. & Davis, M. M. Blimp-1, a novel zinc nger-containing protein that can drive the maturation of B lymphocytes intoimmunoglobulin-secreting cells. Cell 77, 297-306 (1994).

16. Shapiro-Shelef, M. et al. Blimp-1 is required for the formation ofimmunoglobulin secreting plasma cells and pre-plasma memory B cells.Immunity 19, 607-620 (2003).

17. Shaffer, A. L. et al. Blimp-1 orchestrates plasma cell differentiation byextinguishing the mature B cell gene expression program. Immunity 17, 51-62(2002).

18. Chang, D. H., Cattoretti, G. & Calame, K. L. The dynamic expression pattern ofB lymphocyte induced maturation protein-1 (Blimp-1) during mouse embryonicdevelopment. Mech. Dev. 117, 305-309 (2002).

19. Lange, U. C., Saitou, M., Western, P. S., Barton, S. C. & Surani, M. A. TheFragilis interferon-inducible gene family of transmembrane proteins is


212



7/7

associated with germ cell specication in mice. BMC Dev. Biol. 3, doi:10.1186/1471-213X-3-1 (2003).

20. Perea-Gomez, A. et al. Initiation of gastrulation in the mouse embryo ispreceded by an apparent shift in the orientation of the anterior-posterior axis.Curr. Biol. 14, 197-207 (2004).

21. Mesnard, D., Filipe, M., Belo, J. A. & Zernicka-Goetz, M. The anterior-posterioraxis emerges respecting the morphology of the mouse embryo that changesand aligns with the uterus before gastrulation. Curr. Biol. 14, 184-196 (2004).

22. Rossant, J. & Tam, P. P. Emerging asymmetry and embryonic patterning inearly mouse development. Dev. Cell 7, 155-164 (2004).

23. Ginsburg, M., Snow, M. H. & McLaren, A. Primordial germ cells in the mouseembryo during gastrulation. Development 110, 521-528 (1990).24. de Sousa Lopes, S. M. et al. BMP signalling mediated by ALK2 in the visceral

endoderm is necessary for the generation of primordial germ cells in themouse embryo. Genes Dev. 18, 1838-1849 (2004).

25. Payer, B. et al. stella is a maternal effect gene required for normal earlydevelopment in mice. Curr. Biol. 13, 2110-2117 (2003).

26. Mao, X., Fujiwara, Y., Chapdelaine, A., Yang, H. & Orkin, S. H. Activation ofEGFP expression by Cre-mediated excision in a new ROSA26 reporter mousestrain. Blood 97, 324-326 (2001).

27. Lomeli, H., Ramos-Mejia, V., Gertsenstein, M., Lobe, C. G. & Nagy, A. Targetedinsertion of Cre recombinase into the TNAP gene: excision in primordial germcells. Genesis 26, 116-117 (2000).

28. Vincent, S. D. et al. The zinc nger transcriptional repressor Blimp1/Prdm1 isdispensable for early axis formation but is required for specication ofprimordial germ cells in the mouse. Development 132, 1315-1325 (2005).

29. Nagy, A. Manipulating the Mouse Embryo: a Laboratory Manual (Cold SpringHarbor Lab. Press, Cold Spring Harbor, 2003).

30. Forlani, S., Lawson, K. A. & Deschamps, J. Acquisition of Hox codes duringgastrulation and axial elongation in the mouse embryo. Development 130,3807-3819 (2003).

31. Saito, H., Kubota, M., Roberts, R. W., Chi, Q. & Matsunami, H. RTP familymembers induce functional expression of mammalian odorant receptors. Cell119, 679-691 (2004).

32. Wilkinson, D. G. & Nieto, M. A. Detection of messenger RNA by in situhybridization to tissue sections and whole mounts. Methods Enzymol. 225,361-373 (1993).

33. Seki, Y. et al. Extensive and orderly reprogramming of genome-wide chromatinmodications associated with specication and early development of germcells in mice. Dev. Biol. 278, 440-458 (2005).

34. Cox, W. G. & Singer, V. L. A high-resolution, uorescence-based method for

localization of endogenous alkaline phosphatase activity. J. Histochem.Cytochem. 47, 1443-1456 (1999).

Supplementary Information is linked to the online version of the paper atwww.nature.com/nature.

Acknowledgements We thank S. Chuva de Sousa Lopes, K. Nakao, H. Miyachiand R. Nakayama for technical help, and T. Nakano for anti-stella/PGC7antibody. B.P. was supported by a Wellcome Trust PhD studentship. M.A.S. isfunded by the BBSRC, the Wellcome Trust and the EU Epigenome Programme.M.S. is supported by the Ministry of Education, Culture, Sports, Science andTechnology, and a PRESTO grant by the JST. D.OC. acknowledges the supportof the Irvington Institute for Immunological Research and is their NationalGenetics Foundation Fellow. Thanks to K. Lawson and A. McLaren fordiscussions and critical comments.

Author Information Reprints and permissions information is available atnpg.nature.com/reprintsandpermissions. The authors declare no competing

nancial interests. Correspondence and requests for materials should beaddressed to M.A.S. ([email protected]) or M.S.([email protected]).


213

blimp1 is a critical determinant of the germ cell lineage in mice

Documents