Cell, Vol. 113, 631–642, May 30, 2003, Copyright !2003 by Cell Press

The Homeoprotein Nanog Is Required for Maintenanceof Pluripotency in Mouse Epiblast and ES Cells

avoid such ethical issues is to generate pluripotent cellsdirectly from somatic stem and other cells. The first steptoward this goal is to understand molecular mechanisms

Kaoru Mitsui,1 Yoshimi Tokuzawa,1 Hiroaki Itoh,1

Kohichi Segawa,1 Mirei Murakami,1

Kazutoshi Takahashi,1 Masayoshi Maruyama,1

underlying pluripotency.Mitsuyo Maeda,2 and Shinya Yamanaka1,*Leukemia inhibitory factor (LIF) has been utilized to1Laboratory of Animal Molecular Technology

maintain the symmetrical self-renewal of mouse ES cellsResearch and Education Center for Genetic(Smith et al., 1988; Williams et al., 1988). LIF causesInformationheterodimer formation of the LIF receptor and gp130,Nara Institute of Science and Technologywhich resulted in activation of the tyrosine kinase JakNara 630-0192(Ernst et al., 1996). The activated Jak phosphorylatesJapanseveral tyrosines of gp130, which serve as docking sites2 First Department of Anatomyfor proteins containing Src homology 2 (SH2) domains,Osaka City University Medical Schoolincluding the signal transducer and activator of tran-Osaka 545-8585scription (Stat) family of transcription factors (Ihle, 1996).Japan

Multiple lines of evidence indicated that Stat3 is thekey downstream transcription factor of the LIF/gp130pathway in ES cells. Forced expression of a dominant-Summarynegative Stat3 mutant caused differentiation of ES cellseven with the presence of LIF (Niwa et al., 1998). PointEmbryonic stem (ES) cells derived from the inner cellmutation of the tyrosine residue of gp130 responsiblemass (ICM) of blastocysts grow infinitely while main-for the Stat binding abrogated the ability of LIF to main-taining pluripotency. Leukemia inhibitory factor (LIF)tain self-renewal, while other tyrosines were dispens-can maintain self-renewal of mouse ES cells throughable (Ernst et al., 1999; Matsuda et al., 1999). ES cellsactivation of Stat3. However, LIF/Stat3 is dispensableexpressing a fusion molecule consisting of Stat3 andfor maintenance of ICM and human ES cells, sug-estrogen receptor could be maintained pluripotent bygesting that the pathway is not fundamental for pluri-the estrogen derivative tamoxifen (Matsuda et al., 1999).potency. In search of a critical factor(s) that underliesGene targeting experiments also demonstrated the im-pluripotency in both ICM and ES cells, we performed inportance of Stat3 for ES cell self-renewal (Raz et al.,silico differential display and identified several genes1999).specifically expressed in mouse ES cells and preimplan-

However, mutant embryos deficient in the LIF/gp130/tation embryos. We found that one of them, encodingStat3 pathway apparently form normal ICM; LIF-defi-the homeoprotein Nanog, was capable of maintainingcient mice develop normally (Stewart et al., 1992), whileES cell self-renewal independently of LIF/Stat3. na-LIF receptor-deficient mice showed perinatal lethalitynog-deficient ICM failed to generate epiblast and only(Li et al., 1995; Ware et al., 1995). Embryos deficient inproduced parietal endoderm-like cells. nanog-defi-gp130 progressively die between 12.5 days postcoitumcient ES cells lost pluripotency and differentiated into(dpc) and term (Yoshida et al., 1996). Stat3-deficientextraembryonic endoderm lineage. These data dem-embryos developed into the egg cylinder stage untilonstrate that Nanog is a critical factor underlying pluri-embryonic day (E)6.0 and showed a rapid degenerationpotency in both ICM and ES cells.between E6.5–E7.5 (Takeda et al., 1997). A role of LIF/gp130 in preimplantation embryos is evident only forIntroductionprolonged survival of blastocysts during diapause (Nich-ols et al., 2001). Furthermore, LIF is not required for self-

Embryonic stem (ES) cells derived from inner cell mass renewal of human ES cells (Reubinoff et al., 2000) andof mammalian blastocysts grow rapidly and infinitely several mouse ES cell lines (Berger and Sturm, 1997;while maintaining pluripotency, the ability to differenti- Dani et al., 1998). In addition, Stat3 is expressed in wideate into all types of cells (Evans and Kaufman, 1981; ranges of cell types and, in some cases, is essentialMartin, 1981). These properties of ES cells are main- for differentiation (Nakajima et al., 1996). These resultstained by symmetrical self-renewal, producing two iden- argue that LIF/gp130/Stat3 are not fundamental for pluri-tical stem cell daughters upon cell division (Burdon et potency and predict the existence of a novel pathway(s)al., 2002). These properties have raised the hope that that maintains pluripotency in both ICM and ES cells.ES cells could be used to treat a host of degenerative Several transcription factors have been shown essen-diseases such as Parkinson’s disease and diabetes tial for pluripotency in ICM, but none of them function(Smith, 1998). This hope was accelerated by the genera- independently of LIF/Stat3. Oct3/4 is a POU family tran-tion of pluripotent cells from human blastocysts (Thom- scription factor specifically expressed in ES cells, preim-son et al., 1998). However, human ES cells also raised plantation embryos, epiblast, and germ cells (Okamotosubstantial ethical issues since human embryos have et al., 1990; Scholer et al., 1990). Inactivation of oct3/4to be destroyed to generate ES cells (Colman and Burley, in embryos and ES cells resulted in loss of pluripotency2001; Lachmann, 2001; Mieth, 2000). One solution to and spontaneous differentiation into trophoblast lineage

(Niwa et al., 2000). However, ES cells expressing Oct3/4constitutively from an exogenous promoter still required*Correspondence: shinyay@gtc.aist-nara.ac.jp

Cell632

Table 1. Top 20 Genes with Highest Enrichment in ES Cells Identi-fied by Digital Differential Display

Frequency FrequencyUnigene ES/EC cell Others Symbol

Mm.139314 0.00295 0 ecat2Mm.10205 0.0028 0 utf1Mm.5090 0.00079 0 criptoMm.157658 0.00076 0 ecat1Mm.46472 0.00076 0 hnRNB-GMm.5180 0.00073 0 Nr0b1/dax-1Mm.6047 0.00067 0 ecat4/nanogMm.28369 0.00049 0 ecat3/fbx15Mm.17031 0.0004 0 oct3/4Mm.3396 0.00033 0 Zfp42/rex1Mm.45676 0.00033 0 ESTMm.249524 0.00033 0 ecat5/ERasMm.23310 0.0003 0 zfp296Mm.18154 0.00027 0 tcl1Mm.13433 0.00027 0 ecat7Mm.47904 0.00027 0 ecat8Mm.4213 0.00024 0 ecat9Mm.913 0.00024 0 brachyury (T )Mm.28388 0.00024 0 tex20Mm.158190 0.00021 0 ecat6

LIF for self-renewal (Niwa et al., 2000). Overexpressionof oct3/4 induced differentiation similar to that caused

Figure 1. Northern Blot Analyses of ecat Genesby stat3 inactivation, indicating crosstalk between theFive micrograms of total RNA was analyzed with mouse cDNAs oftwo transcription factors. Two other transcription fac-each ecat and H-ras as probes. Differentiation of ES cells was in-tors, Sox2 (Avilion et al., 2003) and FoxD3 (Hanna et al., duced with retinoic acid (3 ! 107 M) for 5 days.

2002), were recently shown essential for pluripotencyin embryos. Since both Sox2 and FoxD3 cooperatively

LIF-Independent Self-Renewal by Ecat4/Nanogfunction with Oct3/4 (Guo et al., 2002; Yuan et al., 1995),We expressed ecat, oct3/4, or sox2 constitutively in ES

it is likely that they also have crosstalk with Stat3. cells to explore their ability to render ES cells indepen-The purpose of this study is to identify the LIF/Stat3- dent of LIF for self-renewal. To this end, we utilized the

independent factor(s) that underlies pluripotency in both supertransfection system, in which more than 90% ofICM and ES cells. To this end, we conducted digital transfected cells expressed the gene of interest,differential display to identify genes expressed in ES allowing functional analyses of polyclonal populationcells as specifically as oct3/4. We explored functions of (Aubert et al., 2002; Niwa et al., 2002). As has beenthese genes by constitutive expression in ES cell and reported, forced expression of Oct3/4 resulted in differ-targeted gene disruption. Here, we report that one of entiation of ES cells (data not shown). Ecat9 and Sox2them, which we now call Nanog for Tir Na Nog (land of induced massive cell death. We were able to establishthe ever young), maintains pluripotency of both ICM and cells constitutively expressing the other eight Ecats.ES cells independently of LIF/Stat3. When cultured with LIF, all of them showed normal mor-

phology. When cultured without LIF, all but one differen-tiated normally as judged by flattened morphology (Fig-

Results ure 2A) and reduced oct3/4 expression (Figure 2B). Incontrast, cells constitutively expressing ecat4 did not

Identification of ecat by Digital Differential Display show such a morphological change even after pro-To identify candidates of the LIF/Stat3-independent fac- longed culture ("1 month) without LIF (Figure 2A). Ex-tor(s) essential for pluripotent cells, we performed digital pression of oct3/4 also remained normal (Figure 2B).differential display (http://www.ncbi.nlm.nih.gov/UniGene/ Because of this unique property, we renamed ecat intoinfo_ddd.shtml) to compare expressed sequence tag nanog from Tir Na Nog (land of the ever young).(EST) libraries from mouse ES cells and those from vari- These (CAG-nanog) cells expressed significantly moreous somatic tissues. A number of genes were found nanog transcripts (Figure 2B) and proteins (Figure 2C)overrepresented in ES cell-derived libraries (Table 1). than control cells. The nanog expression sustained evenAmong twenty genes with highest enrichment were after prolonged culture without LIF. Stat3 activation ki-oct3/4, utf1 (Okuda et al., 1998), and rex1 (Rogers et al., netics was unchanged (Figure 2D). Growth speed was1991), which had been experimentally identified as ES faster than control cells but significantly decreased bycell-specific markers. Northern blot analyses confirmed LIF removal (Figure 2E). These data demonstrated thatES cell-specific expression in nine uncharacterized genes constitutive expression of Nanog maintains the undiffer-(Figure 1). We designated these genes ecat, for ES cell entiated state of ES cells independently of the LIF/Stat3

pathway.associated transcripts.

Stat3-Independent Pluripotency Maintenance Factor633

Figure 2. Forced Expression of Nanog from a Constitutive Promoter

(A) Morphology of MG1.19 cells transfected with the parent plasmid (mock) or the Nanog expression vector (CAG-nanog). Cells were grownwith or without LIF for three passages after puromycin selection.(B) Expression levels of nanog, oct3/4, and nat1 in cells described in (A) were determined by RT-PCR.(C) Lysates were isolated from cells described in (A) and blotted with #-Nanog antiserum or #-Cdk4 antibody (Santa Cruz, sc-260-G). Lowerbands detected by #-Nanog antiserum are likely originated from proteolytic products.(D) LIF was removed for 60 min and was added back to culture medium. Cell lysates were collected at the indicated time points and blottedwith #-Nanog antiserum, #-Oct3/4 antibody (Santa Cruz, sc-5279), #-phospho-Stat3 (Tyr705) antibody (Cell Signaling Technology, #9131),and #-Stat3 antibody (Santa Cruz, sc-482).(E) Cells (10,000 cells per well of 24-well plates) were cultured with or without LIF for 5 days and cell numbers were counted. Shown areaverages and standard deviations of four independent experiments.

A Unique Homeoprotein Encoded by nanog region of 1077 nt containing a B2 repetitive element(data not shown). Homology search revealed that theThe nanog cDNA consisted of 2184 nucleotides (nt) and

contained a single open reading frame encoding poly- predicted Nanog protein contains a homeobox domain(Figure 3A). Phylogenic analysis showed that the Nanogpeptide of 305 amino acids. It has a long 3$ untranslated

Cell634

Figure 3. Sequence and Expression of Nanog

(A) Amino acid sequence of mouse Nanog iscompared with human hypothetical proteinFLJ12581 and mouse Nkx-2.3. Homeodo-mains are shown by closed boxes, in whichnumbers indicate amino acid identity with Na-nog. Also shown are Trp repeat, TN domain,and NK2-SD domain. Tyrosine shown isstrictly conserved among homeodomains ofthe NK-2 family, but not in Nanog orFLJ12581.(B) RT-PCR analysis showing expression pro-files of mouse nanog, oct3/4, and nat1. Lanes:1, undifferentiated MG1.19 ES cells; 2, differ-entiated MG1.19 ES cells; 3, undifferentiatedRF8 ES cells (RT minus control); 4, undifferen-tiated RF8 ES cells; 5, differentiated MG1.19ES cells; 6, brain; 7, heart; 8, kidney; 9, testis;10, spleen; 11, muscle; 12, lung; 13, stomach;14, ovary; 15, thymus; 16, liver; and 17, skin.(C) Lysate was collected from four ES celllines, maintained undifferentiated (indicatedby U) or treated with retinoic acid for 5 days(indicated by D). Western blot was performedwith #-Nanog antiserum or #-Cdk4 antibody.

homeobox is most similar to those of the Nk-2 gene et al., 1990) (Figure 3C). Retinoic acid treatment de-creased Nanog protein levels in all four lines. Recently,family (Harvey, 1996). However, the amino acid identity

is less than 50%. Furthermore, Nanog has valine at the Wang et al. (2003) reported the identification and expres-sion analysis of the same gene as nanog. Consistentposition corresponding to tyrosine strictly conserved

among the NK-2 family. The TN domain and the NK-2- with our data, they demonstrated specific expression inES cells and preimplantation embryos. DNA microarraysspecific domain, which are also conserved among the

NK-2 family, do not exist in Nanog. These data indicate analyses demonstrated that nanog is expressed in EScells, but not in neural or hematopoietic stem cells (Ra-that Nanog is distinct from the NK-2 family. Amino acid

identities of the Nanog homeodomain with other homeo- malho-Santos et al., 2002).Database analyses showed that the human hypotheti-proteins expressed in ES cells, namely Oct3/4, Pem (Fan

et al., 1999), and Ehox (Jackson et al., 2002), are less cal protein FLJ12581 is most similar to Nanog (Figure3A). Although overall amino acid identity is relativelythan 20% (data not shown). However, Nanog is identical

to a homeodomain protein recently described as being low (52%), its homeodomain is 85% identical to mouseNanog. FLJ12581 also contains Trp-rich repeats, inexpressed selectively in ES cells (Wang et al., 2003).

Nanog also contains a stretch of sequence in which Trp- which Typ-X-X-X is repeated eight times. In addition,FLJ12581 contains an Alu repetitive element in the 3$X-X-X is repeated ten times in tandem. The function(s)

of the Trp-rich repeats is not known. untranslated region. FLJ12581 is located on chromo-some 12, the syntenic region of mouse chromosome 6Transcripts of nanog were not detectable in twelve

somatic organs, including testis and ovary, by either where nanog is located. Blast search against the humandbEST identified EST clones corresponding to FLJ12581Northern blot (Figure 1) or RT-PCR (Figure 3B). Western

blot with #-Nanog rabbit antiserum, which we raised in libraries from NT2 human teratocarcinoma cells, germcell and testis tumors, marrow, and other tumors. Noagainst histidine-tagged Nanog protein, detected a spe-

cific band of expected size (35 kDa) in four independent EST clones was detected in libraries from normal so-matic tissues except those corresponding to the AluES cell lines: RF8 (Meiner et al., 1996), MG1.19 (Gass-

mann et al., 1995), J1 (Li et al., 1992), and CGR8 (Nichols element. Northern blot analysis did not detect FLJ12581

Stat3-Independent Pluripotency Maintenance Factor635

transcripts in 12 adult organs (data not shown). These (Palmieri et al., 1994). When cultured without feedercells, oct3/4 and rex1 were eliminated, while gata6 wasdata indicate that FLJ12581 is the human nanog or-

tholog. further induced (data not shown) consistent with theirparietal endoderm-like morphology (Figure 5A). Thesedata indicated that nanog disruption specified ES cellTargeted Disruption of the Mouse nanog Genefate exclusively into extraembryonic endoderm lineages.in ES Cells

Analysis with mouse genome databases demonstratedEstablishment of nanog-Deficient Micethat the nanog gene consisted of four exons. We gener-To study in vivo function(s) of Nanog, we injected %geoated two targeting vectors that replaced the homeodo-heterozygous ES cells into C57BL/6 blastocysts. Germ-main with either a fusion of %-galactosidase and theline transmission was obtained from three independentneomycin-resistance gene (%geo) (Mountford et al.,ES clones. Heterozygous mutant mice were normal in1994) or the hygromycin-resistant gene (hygr) (Figuregross appearance and fertile (data not shown). X-Gal4A). These vectors were introduced into RF8 ES cellsstaining of the preimplantation embryos detected %geoby electroporation, and drug-resistant colonies wereexpression in the inner cell mass of blastocysts and thescreened for homologous recombination. With the %geocentral portion of the morula, but not in 32-cell stage ortargeting vector, 27 out of 96 colonies were found posi-earlier embryos (Figure 6A).tive by PCR (data not shown). With the hygr vector, 8 out 27

were positive. Homologous recombination was confirmedwith Southern blot (Figure 4B) and PCR (Figure 4C). X-Gal Lack of Epiblast in nanog Null Embryosstaining of heterozygous ES cells confirmed that nanog Among 231 mice born from heterozygous intercrosses,is expressed in all undifferentiated ES cells but re- 80 were wild-type and 151 were heterozygous. Sincepressed by retinoic acid treatment (Figure 4E). no homozygous offspring was obtained, we analyzed

To obtain homozygous mutant ES cells, we intro- embryos of various stages. Among 36 decidual swellingsduced the hygr vector into heterozygous clones estab- obtained at 7.5 dpc, 14 contained wild-type embryoslished by the %geo vector. In 34 out of 83 hyg-resistant and 16 contained heterozygous mutants. Six deciduaeclones, homologous recombination of the hygr vector were empty. We therefore histologically examined E5.5was confirmed. Among these, 11 clones were found embryos. Of 33 embryos examined, 26 showed normalhomozygous for nanog mutation by Southern blot (Fig- morphology (Figure 6B). However, seven embryos ap-ure 4B) and PCR (Figure 4C). Northern blot (Figure 4D) peared to consist entirely of disorganized extraembry-and Western blot (data not shown) analyses confirmed onic tissues with no discernible epiblast or extraembry-the absence of expression in homozygous cells. In the onic ectoderm. The high incidence (7 out of 33) indicatesother 23 clones, the hygr vector replaced the %geo cas- that these abnormal embryos are nanog null. They aresette. We also introduced the %geo vector into heterozy- different from foxD3 null embryos that do have epiblastgous clones established by the hygr vector. In 114 out at this stage (Hanna et al., 2002) and Sox2 null embryosof 397 G418-resistant clones, homologous recombina- that lack epiblast but showed expansion of extraembry-tion of the %geo vector was confirmed. Among these, onic ectoderm (Avilion et al., 2003).six clones were found homozygous. Cells deficient in At 3.5 dpc, we found that 3 out of 15 blastocystsnanog obtained by both orders showed identical pheno- were homozygous and indistinguishable from normaltypes. embryos. When cultured on gelatin-coated plate, how-

ever, ICM of nanog null blastocysts failed to proliferate(Figure 6C). To further clarify the fate of mutant ICM, weLoss of Pluripotency and Extraembryonic Endodermremoved trophectoderm by immunosurgery. ICM defi-Differentiation in nanog Null Cellscient in nanog did not persist as undifferentiated massesCells deficient in nanog were distinct in morphologyin vitro but differentiated completely into parietal endo-from wild type and heterozygous ES cells. When main-derm-like cells within 4 days (Figure 6D). In contrast totained on STO feeder cells, they were larger and round-oct3/4 null ICM (Nichols et al., 1998), no trophoblastshaped (Figure 5A); when cultured without feeder cells,differentiation was observed. These data demonstratedthey further flattened and showed parietal endoderm-that Nanog is essential for maintenance of pluripotencylike morphology. Cells deficient in nanog grew signifi-of ICM at a stage after the initial requirement for Oct3/4.cantly slower that heterozygous ES cells (Figure 5B) but

could be serially passaged for at least 2 months (datanot shown). Identification of DNA Recognition Sequence

of NanogNorthern blot (Figure 5C) and RT-PCR (Figure 5D)analyses showed that nanog null cells expressed a num- We determined DNA recognition sequence of Nanog by

systematic evolution of ligands by exponential enrich-ber of endoderm transcription factors, parietal endodermmarkers, and visceral endoderm markers. In contrast, ment (SELEX). By selecting random olionucleotide with

the fusion protein consisting of maltose binding proteinmesoderm marker brachyury (t ), trophoblast markercdx2, primitive ectoderm marker fibroblast growth fac- (MBP) Nanog fusion proteins (see Experimental Proce-

dures for details), we identified a consensus sequencetor (fgf)-5, or the neuroectodermal marker islet-1 (isl1)were not induced. Pluripotent cell markers oct3/4 and (C/G)(G/A)(C/G)C(G/C)ATTAN(G/C) (Figure 7A). The high-

est conservation was observed with the tetramer ATTA,rex1 were greatly reduced though not completely elimi-nated when cultured on STO feeder cells. The persis- which is common for DNA recognition sequences of

many homeobox transcription factors. This consensustence of low oct3/4 expression may be related to itsproposed role in primitive endoderm specification sequence is different from that of the NK-2 family, T(C/T)

Cell636

Figure 4. Targeted Disruption of the Mouse nanog Gene

(A) Structures of the mouse nanog gene, targeting constructs, and correctly targeted loci. Positions of the probe and ScaI recognition sites(S) used for Southern blot were shown. Arrows indicate primers for PCR analyses.(B) Southern blot with the 3$ probe produced an 11 kb band from the wild type locus, a 15.4 kb band from the %geo-targeted locus, and a7.3 kb band from the hygr targeted locus.(C) Specific amplification by PCR with primers 6047-insS1.2 (CACCTACCACCATGCCAGGCTGAGAATGTC) and 6047-RACE1 (AGCTGGCATCGGTTCATCATGGTACA) produced a 3.2 kb band from the wild type locus, an 8.5 kb band from the %geo targeted locus, and a 4.2 kb bandfrom the hygr targeted locus.(D) Northern blot analysis confirmed the absence of nanog transcripts in knockout cells.(E) X-Gal staining of %geo/& ES cells that were maintained undifferentiated on STO cells or induced to differentiate by retinoic acid (RA).

AAGTG (Harvey, 1996), consistent with the amino acid Discussiondifferences between Nanog and NK-2. Gel-mobility shiftassay showed that mobility of consensus oligonucleo- Prior to this study, LIF/gp130/Stat3 was the only known

biochemical pathway that could maintain self-renewaltides was retarded by extracts of F9 embryonic carci-noma cells (Figure 7B) and RF8 ES cells (data not of mouse ES cells (Burdon et al., 2002). However, LIF/

gp130/Stat3 is dispensable for development of ICM andshown). The shift in mobility was abolished by excessamount of unlabeled consensus oligonucleotides, but self-renewal of human ES cells, suggesting that this

pathway is not fundamental for pluripotency (Nicholsnot by those with mutations in ATTA.

Stat3-Independent Pluripotency Maintenance Factor637

Figure 5. Analyses of nanog Null Cells

(A) Morphology of wild type RF8 cells grown on STO feeder cells (WT) and nanog null cells (KO) grown with or without STO feeder cells.(B) Three clones of heterozygous cells (&/') and homozygous cells ('/') were plated at 1 ! 104 cells per well of 24-well plates. Cell numberswere counted after 2, 4, and 6 days.(C) Total RNA was isolated from parent RF8 cells (WT), four clones with random integration of the %geo vector (&/&), four heterozygous clones(&/'), and four homozygous clones ('/'). Expression of marker genes was determined by Northern blot analyses.(D) Total RNA was isolated from wild-type RF8 cells (nanog, WT), nanog null cells (nanog, KO), ZHBTc4 cells (Niwa et al., 2000) culturedwithout tetracycline (oct3/4, WT), and ZHBTc4 cells cultured with tetracycline for 96 hr, which differentiated into trophoblast (oct3/4, KO).Expression of marker genes was determined by RT-PCR.

et al., 2001). In this study, we demonstrated that the resulted in loss of pluripotency in both ICM and ES cells.In the accompanying paper, Chambers et al. showedhomeoprotein Nanog can maintain self-renewal of ES

cells independently of the LIF/gp130/Stat3 pathway. that Nanog has an ability to maintain not only undifferen-tiated state, but also pluripotency independently of LIFFurthermore, we showed that deletion of the nanog gene

Cell638

Figure 6. Analyses of nanog Null Embryos

(A) X-Gal staining of preimplantation embryos obtained from intercross of heterozygous %geo mice.(B) Decidual swellings obtained from heterozygous intercross were removed from uterus, fixed with 10% buffered formalin, and embeddedin paraffin. Sections (5 (m each) were stained with hematoxylin and eosin.(C) Blastocysts obtained from heterozygous intercrosses were grown on gelatin-coated plate with ES cell medium. After 10 days, morphologywas recorded and genotype was determined by PCR. All homozygous blastocysts showed similar morphology shown here.(D) ICM isolated by immunosurgery was cultured in vitro for 4 days.

Stat3-Independent Pluripotency Maintenance Factor639

function phenotypes in ES cells are different betweenStat3 and Nanog. Inactivation of the Stat3 pathway byLIF depletion or a dominant-negative mutant resultedin differentiation into multiple cell lineages includingprimitive endoderm and mesoderm (Niwa et al., 1998).However, terminal differentiation into parietal and/or vis-ceral endoderm was not observed. Expression of gata4and gata6 was barely increased by those treatments(Fujikura et al., 2002). In contrast, nanog null cells spon-taneously differentiated into parietal and visceral endo-derm expressing high levels of gata4 and gata6. Third,LIF was still able to enhance growth of CAG-nanog cells.If Nanog were a downstream target of Stat3, these re-sults could not be explained.

How does Nanog maintain self-renewal of ES cells?One of the mechanisms may be transcriptional repres-sion of genes that promote differentiation, such as gata4and gata6. We found that both gata4 and gata6 wereupregulated in nanog null cells. Forced expression ofgata4 or gata6 in ES cells was sufficient to induce extra-embryonic endoderm differentiation (Fujikura et al.,2002). Expression patterns of marker genes were nearlyidentical between their gata6-overexpressing cells andour nanog null cells. Therefore, the upregulation of gata4and gata6 could be sufficient to explain spontaneousdifferentiation into extraembryonic endoderm in nanognull cells.Figure 7. DNA Recognition and Proposed Functions of Nanog

It has been shown that gata6 may be upstream of(A) Consensus binding sequence was determined by SELEX and

gata4, since inactivation of gata6 resulted in repressionshown with Sequence logo (Schneider and Stephens, 1990), whichof gata4 (Morrisey et al., 1998) while inactivation of gata4was generated by WebLogo (www.bio.cam.ac.uk/cgi-bin/seqlogo/increased gata6 expression (Kuo et al., 1997). We foundlogo.cgi). The height of each letter is proportional to its frequency,

and the letters are sorted so that the most common one is on top. that the gata6 enhancer region (Molkentin et al., 2000;The height of the entire stack is adjusted to signify the information Sun-Wada et al., 2000) contains the DNA recognitioncontent (measured in bits) of the sequence at that position. motif of Nanog determined by SELEX, suggesting that(B) 32P-labeled oligonucleotides containing the consensus sequence

Nanog might directly repress gata6, thereby inhibiting(wt: tgaggcccaccctCAGCCATTAaCctttttactttattct) or mutated se-ES cell differentiation. In addition, the Nanog consensusquence (mut: tgaggcccaccctCAGCCAGGAaCctttttactttattct) weresequence is found in the enhancer region of rex1 (Ben-incubated with nuclear extracts of F9 embryonic carcinoma cells,

separated by SDS-PAGE, and visualized by a phosphoimager (Toku- Shushan et al., 1998). Thus, Nanog may also regulatezawa et al., 2003). Specificity of the reaction was tested by addition ES cell-specific genes to maintain pluripotency. Furtherof excess amount of cold oligonucleotides. studies, such as reporter gene assays and gel-mobility(C) Proposed functions of Nanog, Oct3/4, and Stat3 in preimplanta- shift assays, are required to determine whether Nanogtion embryos (upper) and ES cells (lower).

indeed binds to the consensus sequence in the gata6or rex1 to regulate their expression.

Another important issue is the relationship between(Chambers et al., 2003 [Cell, this issue]). These data Nanog and Oct3/4. The loss-of-function phenotypes ofargue that Nanog is the missing fundamental factor that the two transcription factors are different. Disruptionunderlies pluripotency in both ICM and ES cells. of nanog resulted in cell populations that expressed

The ability to support LIF/Stat3-independent self- markers of both parietal and visceral endoderm andrenewal is unique to Nanog. We identified eight other could be serially passaged for several months. In con-ecat genes that are expressed in ES cells as specifically trast, oct3/4 null cells differentiated into trophectodermas nanog and oct3/4. We have shown that ecat3 encod- (Nichols et al., 1998; Niwa et al., 2000). Overexpressioning the F box protein Fbx15 is a novel target of Oct3/4 phenotypes are also different. Overexpression of nanogand Sox2 (Tokuzawa et al., 2003). We have also demon- led to LIF-independent self-renewal of ES cells, whilestrated that ecat5 encoding a constitutively active Ras oct3/4 overexpression resulted in differentiation intoprotein (ERas) plays an important role in tumor-like prop- multiple cell lineages, a similar phenotype caused byerty of ES cells (Takahashi et al., 2003). However, we Stat3 inactivation (Niwa et al., 2000). These results indi-could not achieve LIF/Stat3-independet self-renewal cate that the primary function of Oct3/4 is to preventwith Ecat3, Ecat5, other Ecats, Oct3/4, or Sox2. These trophectoderm differentiation of ICM and ES cells. Inresults highlighted the importance of Ecat4/Nanog in contrast, Nanog independently prevents differentiationmaintenance of pluripotency. into extraembryonic endoderm and actively maintains

Three lines of evidence indicate that Nanog functions pluripotency.independently of Stat3. First, phosphorylation kinetics of Based on the finding of our present work and previousStat3 is not elevated in CAG-nanog ES cells. It responded reports by others, we propose a model for the roles of

Nanog, Oct3/4, and Stat3 in the maintenance of pluripo-normally to LIF removal and stimulation. Second, loss-of-

Cell640

and introduced into RF8 ES cells by electroporation (Meiner et al.,tency (Figure 7C). During preimplantation development,1996). Genomic DNAs from G418- or hygromycin B-resistant colo-the first cell fate determination takes place at morulanies were screened for homologous recombination by PCR andstage, when cells remain pluripotent as ICM or differenti-Southern blot.

ate into trophectoderm. At this stage, nanog expressionis still low and Oct3/4 is the key determinant of cell fate.

Analyses of Marker Gene ExpressionCells that lost oct3/4 expression differentiate into troph-Total RNA was isolated with Trizol (Invitrogen). Northern blot analy-

ectoderm. ses were performed with formalin-agarose gel as previously de-The second cell fate determination subsequently scribed (Yamanaka et al., 2000, 1998). For RT-PCR, first strand cDNA

takes place at blastocyst stage when cells of ICM remain was synthesized with ReverTra Ace (Toyobo, Japan), and PCR wascarried out with ExTaq (Takara, Japan). Most PCR primers werepluripotent as epiblast or differentiate into primitive en-described by others (Fujikura et al., 2002). Other primers used aredoderm. Nanog is expressed at this stage and functions6047-S4 (AGGGTCTGCTACTGAGATGCTCTG) and 6047-AS5 (CAACas the crucial determinant of cell fate. Cells that expressCACTGGTTTTTCTGCCACCG) for nanog, Oct3/4-U474 (CTGAGGG

nanog remain pluripotent, while cells without nanog dif- CCAGGCAGGAGCACGAG) and Oct3/4- L935 (CTGTAGGGAGGGCferentiate into primitive endoderm. TTCGGGCACTT) for oct3/4, NAT1-U283 (ATTCTTCGTTGTCAAGC

ES cells derived from ICM require both Oct3/4 and CGCCAAAGTGGAG) and NAT1-L476 (AGTTGTTTGCTGCGGAGTTGTCATCTCGTC) for nat1, Cdx2-S775 (GGCGAAACCTGTGCGAGTGNanog to prevent differentiation into trophectoderm andGATGCGGAA) and Cdx2-AS1210 (GATTGCTGTGCCGCCGCCGCTprimitive endoderm, respectively. However, these twoTCAGACC) for cdx2, LimB.S354 (CTGTTCTGAAAGTGAATGTGGTGtranscription factors of normal expression level are notGCCCC) and LamB-AS945 (GTTTAATCGCCTTCTCTGCTGCAAC

sufficient for ES cell self-renewal. Additional factor(s), CTG) for lamininB1, and TM-S1071 (CCAGGCTCTTACTCCTGTA)such as Stat3 activated by LIF, are required to support and TM-AS1300 (TGGCACTGAAACTCGCAGTT) for thrombomo-prolonged maintenance of pluripotency. Increased ex- dulin.pression of Oct3/4 does not render ES cells independentof LIF, but instead induces primitive endoderm differen- Isolation of ICM by Immunosurgerytiation. This suggests that the function of Oct3/4 in ES Blastocysts were incubated for 1 min at room temperature in acidic

Tyrode’s solution (Sigma) to remove Zona Pellucida and washedcells is to specify cell fate into either pluripotent cellsthree times with M16 medium (Sigma). Zona-free embryos wereor primitive endoderm by blocking trophectoderm differ-incubated for 10 min at 37)C with #-mouse antiserum (diluted 20entiation. In contrast, our data demonstrated that ele-times in M16 medium), rinsed three times with M16, and incubatedvated expression of Nanog bypasses the LIF/Stat3 path- for 30 minutes at 37)C with guinea pig complement (ICN Biomedi-

way and maintains self-renewal of ES cells. Thus, Nanog cals, Inc., diluted 20 times in M16 medium). After washing three timesnot only blocks primitive endoderm differentiation, but with M16 medium, ICM were isolated from dead trophectoderm and

cultured in gelatin-coated plate with ES cell medium.also actively maintains pluripotency.In conclusion, our study revealed two important prop-

erties of the homeoprotein Nanog—the fundamental role SELEXThe coding region of the mouse nanog gene was introduced intoin pluripotency of both ICM and ES cells, and the abilitypIH1119 to produce a fusion protein consisting of MBP and Nanog.to maintain ES cell self-renewal without LIF. Oct3/4 andThe fusion was induced in E. coli BL21AI (Invitrogen) and was puri-Stat3, the two well-known players in pluripotency, havefied with amylose-beads (New England Biolabs). SELEX was per-only one of the two properties. Identification of criticalformed as follows. Double strand oligonucleotide was synthesized

regulatory genes such as nanog is a crucial step in in reaction mixture containing 1 (g SELEX-N20-Oligo (TAGGCATGTunderstanding early embryogenesis and in exploiting GGATCCGTCTGGCN20GCGTACCGGATCCACCTTCGAT), 1 (g

SELEX-N20RV primer (ATCGAAGGTGGATCCGGTACGC), 2.5 (lpluripotent cells for therapeutic goals.10! Klenow buffer, 4 (l 10 mM dNTP, and 14.7 (l distilled water.The mixture was denatured at 95)C for 5 min and annealed at 54)CExperimental Proceduresfor 5 min. Klenow fragment (0.225 U in 3.6 (l of 1! Klenow buffer)was added to the mixture, which was then incubated at 25)C forConstitutive Expression of the ecat Genes in ES Cells

The coding region of the ecat genes, oct3/4, or sox2 was introduced 40 min. The double-strand DNA was purified by phenol/chloroformextraction and ethanol precipitation and resuspended in 20 (l dis-with Gateway technology (Invitrogen) into pPyCAGIP vector that

allowed episomal expression in MG1.19 ES cells (Aubert et al., 2002; tilled water. DNA-protein binding was performed at 4)C for 30 minin reaction mixture consisting of 20 (l DNA, 30 (l MBP-Nanog-Niwa et al., 2002). Transfection into MG1.19 cells was performed with

LipofectAMINE 2000 (Invitrogen) according to the manufacturer’s bound beads, 20 (l of 5! binding buffer (100 mM HEPES-KOH [pH7.9], 1 mM EDTA, 1 M KCl, and 50% Glycerol), and 30 (l of water.protocol. Puromycin selection (2 (g/ml) was continued throughout

experiments. The beads were washed six times with 100 (l lysis buffer (20 mMTris-HCl [ph 7.4], 200 mM NaCl, and 1mM EDTA) supplementedwith 1 mM DTT and 0.2 mM PMSF. DNA was purified by phenol/Targeted Disruption of the Mouse nanog Gene

Two targeting vectors with promoter trap selection was designed chloroform extraction and ethanol precipitation and resuspendedin 20 (l distilled water. 5 (l of this was used for PCR amplificationto replace exon 2 of the mouse nanog gene, which contains the

homeodomain, with an IRES (internal ribosome entry site)-%geo cas- in reaction buffer containing 5 (l of 10! ExTaq buffer, 4 (l of 2.5mM dNTP, 1 (l of 30 (M SELEX-N20FW primer (TAGGCATGTGGsette (Mountford et al., 1994) or the IRES-hygr cassette. A 4 kb

fragment containing intron 1 and a 1.5 kb fragment containing exon ATCCGTCTGGC), 1 (l of 30 (M SELEX-N20RV primer, 0.25 (L ofExTaq polymerase, and 33.75 (L of water. PCR program consisted3, intron 3, and exon 4 were amplified from mouse genomic DNA

and used as the 5$ and 3$ homologous regions of the targeting of the initial denaturation at 95)C for 5 min, 20 cycles of 95)C for 1min, 54)C for 1.5 min, and 72)C for 1 min, and the final extensionvector. The 5$ arm was amplified by expand long template PCR

system (Roche) with primers S4 (AGGGTCTGCTACTGAGATGCTC at 72)C for 7 min. Amplified products were purified by Micro-biospin column (BioRad) and concentrated to 20 (l by ethanol precipi-TG) and AS4 (AGGCAGGTCTTCAGAGGAAGGGCG). The 3$ arm was

amplified with primes IntS3.2 (CGGGCTGTAGACCTGTCTGCATTC tation, of which half was applied to another selection round. Thisprocedure was repeated five times for enrichment. Amplified prod-TG) and RACE2 (GGTCCTTCTGTCTCATCCTCGAGAGT). Either the

%geo or hygr cassette was ligated between the two DNA fragments. ucts in the final reaction were ligated into pCR2.1 vector (Invitrogen).Thirty clones were randomly selected and sequenced.The resulting targeting vectors were linearized with SacII digestion

Stat3-Independent Pluripotency Maintenance Factor641

Acknowledgments to regulate endodermal-specific promoter expression. Proc. Natl.Acad. Sci. USA 99, 3663–3667.

We thank Drs. Ian Chambers, Jennifer Nichols, and Austin Smith Hanna, L.A., Foreman, R.K., Tarasenko, I.A., Kessler, D.S., and La-for sharing data prior to publication. We thank Tomoko Ichisaka and bosky, P.A. (2002). Requirement for Foxd3 in maintaining pluripotentYukiko Samitsu for blastocyst microinjection, Junko Iida and Chihiro cells of the early mouse embryo. Genes Dev. 16, 2650–2661.Takigawa for technical assistance, and Takashi Iwasaki for histologi-

Harvey, R.P. (1996). NK-2 homeobox genes and heart development.cal analyses. We also thank Dr. Hung Li for agreeing with the nomen-Dev. Biol. 178, 203–216.clature. We are grateful to Dr. Hitoshi Niwa for providing variousIhle, J.N. (1996). STATs: signal transducers and activators of tran-materials essential for this study. This work was supported in partscription. Cell 84, 331–334.by research grants from the Ministry of Education, Culture, Sports,

Science, and Technology of Japan; Uehara Memorial Foundation; Jackson, M., Baird, J.W., Cambray, N., Ansell, J.D., Forrester, L.M.,and Sumitomo Research Foundation (to S.Y.). and Graham, G.J. (2002). Cloning and characterization of Ehox, a

novel homeobox gene essential for embryonic stem cell differentia-Received: December 9, 2002 tion. J. Biol. Chem. 277, 38683–38692.Revised: April 29, 2003 Kuo, C.T., Morrisey, E.E., Anandappa, R., Sigrist, K., Lu, M.M., Par-Accepted: April 30, 2003 macek, M.S., Soudais, C., and Leiden, J.M. (1997). GATA4 transcrip-Published: May 29, 2003 tion factor is required for ventral morphogenesis and heart tube

formation. Genes Dev. 11, 1048–1060.References Lachmann, P. (2001). Stem cell research—why is it regarded as a

threat? An investigation of the economic and ethical argumentsAubert, J., Dunstan, H., Chambers, I., and Smith, A. (2002). Func- made against research with human embryonic stem cells. EMBOtional gene screening in embryonic stem cells implicates Wnt antag- Rep. 2, 165–168.onism in neural differentiation. Nat. Biotechnol. 20, 1240–1245.

Li, E., Bestor, T.H., and Jaenisch, R. (1992). Targeted mutation ofAvilion, A.A., Nicolis, S.K., Pevny, L.H., Perez, L., Vivian, N., and the DNA methyltransferase gene results in embryonic lethality. CellLovell-Badge, R. (2003). Multipotent cell lineages in early mouse

69, 915–926.development depend on SOX2 function. Genes Dev. 17, 126–140.

Li, M., Sendtner, M., and Smith, A. (1995). Essential function of LIFBen-Shushan, E., Thompson, J.R., Gudas, L.J., and Bergman, Y.

receptor in motor neurons. Nature 378, 724–727.(1998). Rex-1, a gene encoding a transcription factor expressed in

Martin, G.R. (1981). Isolation of a pluripotent cell line from earlythe early embryo, is regulated via Oct-3/4 and Oct-6 binding to anmouse embryos cultured in medium conditioned by teratocarcinomaoctamer site and a novel protein, Rox-1, binding to an adjacent site.stem cells. Proc. Natl. Acad. Sci. USA 78, 7634–7638.Mol. Cell. Biol. 18, 1866–1878.Matsuda, T., Nakamura, T., Nakao, K., Arai, T., Katsuki, M., Heike,Berger, C.N., and Sturm, K.S. (1997). Self renewal of embryonicT., and Yokota, T. (1999). STAT3 activation is sufficient to maintainstem cells in the absence of feeder cells and exogenous leukaemiaan undifferentiated state of mouse embryonic stem cells. EMBO J.inhibitory factor. Growth Factors 14, 145–159.18, 4261–4269.Burdon, T., Smith, A., and Savatier, P. (2002). Signalling, cell cycleMeiner, V.L., Cases, S., Myers, H.M., Sande, E.R., Bellosta, S.,and pluripotency in embryonic stem cells. Trends Cell Biol. 12,Schambelan, M., Pitas, R.E., McGuire, J., Herz, J., and Farese, R.V.,432–438.Jr. (1996). Disruption of the acyl-CoA:cholesterol acyltransferaseChambers, I., Colby, D., Robertson, M., Nichols, J., Lee, S., Tweedie,gene in mice: evidence suggesting multiple cholesterol esterificationS., and Smith, A. (2003). Functional expression cloning of Nanog, aenzymes in mammals. Proc. Natl. Acad. Sci. USA 93, 14041–14046.pluripotency sustaining factor in embryonic stem cells. Cell 113, thisMieth, D. (2000). Going to the roots of the stem cell debate. Theissue.ethical problems of using embryos for research. EMBO Rep. 1, 4–6.Colman, A., and Burley, J.C. (2001). A legal and ethical tightrope.Molkentin, J.D., Antos, C., Mercer, B., Taigen, T., Miano, J.M., andScience, ethics and legislation of stem cell research. EMBO Rep.Olson, E.N. (2000). Direct activation of a GATA6 cardiac enhancer2, 2–5.by Nkx2.5: evidence for a reinforcing regulatory network of Nkx2.5Dani, C., Chambers, I., Johnstone, S., Robertson, M., Ebrahimi, B.,and GATA transcription factors in the developing heart. Dev. Biol.Saito, M., Taga, T., Li, M., Burdon, T., Nichols, J., and Smith, A.217, 301–309.(1998). Paracrine induction of stem cell renewal by LIF-deficient

cells: a new ES cell regulatory pathway. Dev. Biol. 203, 149–162. Morrisey, E.E., Tang, Z., Sigrist, K., Lu, M.M., Jiang, F., Ip, H.S., andParmacek, M.S. (1998). GATA6 regulates HNF4 and is required forErnst, M., Oates, A., and Dunn, A.R. (1996). Gp130-mediated signaldifferentiation of visceral endoderm in the mouse embryo. Genestransduction in embryonic stem cells involves activation of Jak andDev. 12, 3579–3590.Ras/mitogen-activated protein kinase pathways. J. Biol. Chem.

271, 30136–30143. Mountford, P., Zevnik, B., Duwel, A., Nichols, J., Li, M., Dani, C.,Robertson, M., Chambers, I., and Smith, A. (1994). Dicistronic tar-Ernst, M., Novak, U., Nicholson, S.E., Layton, J.E., and Dunn, A.R.geting constructs: reporters and modifiers of mammalian gene ex-(1999). The carboxyl-terminal domains of gp130-related cytokinepression. Proc. Natl. Acad. Sci. USA 91, 4303–4307.receptors are necessary for suppressing embryonic stem cell differ-

entiation. Involvement of STAT3. J. Biol. Chem. 274, 9729–9737. Nakajima, K., Yamanaka, Y., Nakae, K., Kojima, H., Ichiba, M., Kiuchi,N., Kitaoka, T., Fukada, T., Hibi, M., and Hirano, T. (1996). A centralEvans, M.J., and Kaufman, M.H. (1981). Establishment in culture ofrole for Stat3 in IL-6-induced regulation of growth and differentiationpluripotential cells from mouse embryos. Nature 292, 154–156.in M1 leukemia cells. EMBO J. 15, 3651–3658.Fan, Y., Melhem, M.F., and Chaillet, J.R. (1999). Forced expressionNichols, J., Chambers, I., Taga, T., and Smith, A. (2001). Physiologi-of the homeobox-containing gene Pem blocks differentiation of em-cal rationale for responsiveness of mouse embryonic stem cells tobryonic stem cells. Dev. Biol. 210, 481–496.gp130 cytokines. Development 128, 2333–2339.Fujikura, J., Yamato, E., Yonemura, S., Hosoda, K., Masui, S., Nakao,Nichols, J., Evans, E.P., and Smith, A.G. (1990). Establishment ofK., Miyazaki Ji, J., and Niwa, H. (2002). Differentiation of embryonicgerm-line-competent embryonic stem (ES) cells using differentiationstem cells is induced by GATA factors. Genes Dev. 16, 784–789.inhibiting activity. Development 110, 1341–1348.Gassmann, M., Donoho, G., and Berg, P. (1995). Maintenance of anNichols, J., Zevnik, B., Anastassiadis, K., Niwa, H., Klewe-Nebenius,extrachromosomal plasmid vector in mouse embryonic stem cells.D., Chambers, I., Scholer, H., and Smith, A. (1998). Formation ofProc. Natl. Acad. Sci. USA 92, 1292–1296.pluripotent stem cells in the mammalian embryo depends on theGuo, Y., Costa, R., Ramsey, H., Starnes, T., Vance, G., Robertson,POU transcription factor Oct4. Cell 95, 379–391.K., Kelley, M., Reinbold, R., Scholer, H., and Hromas, R. (2002). The

embryonic stem cell transcription factors Oct-4 and FoxD3 interact Niwa, H., Burdon, T., Chambers, I., and Smith, A. (1998). Self-renewal

Cell642

of pluripotent embryonic stem cells is mediated via activation of target of Oct3/4 but is dispensable for embryonic stem cell self-renewal and mouse development. Mol. Cell. Biol. 23, 2699–2708.STAT3. Genes Dev. 12, 2048–2060.

Wang, S.H., Tsai, M.S., Chiang, M.F., and Li, H. (2003). A novel NK-Niwa, H., Miyazaki, J., and Smith, A.G. (2000). Quantitative expres-type homeobox gene, ENK (early embryo specific NK), preferentiallysion of Oct-3/4 defines differentiation, dedifferentiation or self-expressed in embryonic stem cells. Gene Expr. Patterns 3, 99–103.renewal of ES cells. Nat. Genet. 24, 372–376.Ware, C.B., Horowitz, M.C., Renshaw, B.R., Hunt, J.S., Liggitt, D.,Niwa, H., Masui, S., Chambers, I., Smith, A.G., and Miyazaki, J.Koblar, S.A., Gliniak, B.C., McKenna, H.J., Papayannopoulou, T.,(2002). Phenotypic complementation establishes requirements forThoma, B., et al. (1995). Targeted disruption of the low-affinity leuke-specific POU domain and generic transactivation function of Oct-mia inhibitory factor receptor gene causes placental, skeletal, neural3/4 in embryonic stem cells. Mol. Cell. Biol. 22, 1526–1536.and metabolic defects and results in perinatal death. DevelopmentOkamoto, K., Okazawa, H., Okuda, A., Sakai, M., Muramatsu, M.,121, 1283–1299.and Hamada, H. (1990). A novel octamer binding transcription factorWilliams, R.L., Hilton, D.J., Pease, S., Willson, T.A., Stewart, C.L.,is differentially expressed in mouse embryonic cells. Cell 60,Gearing, D.P., Wagner, E.F., Metcalf, D., Nicola, N.A., and Gough,461–472.N.M. (1988). Myeloid leukaemia inhibitory factor maintains the devel-

Okuda, A., Fukushima, A., Nishimoto, M., Orimo, A., Yamagishi, T.,opmental potential of embryonic stem cells. Nature 336, 684–687.

Nabeshima, Y., Kuro-o, M., Nabeshima, Y., Boon, K., Keaveney, M.,Yamanaka, S., Zhang, X.Y., Miura, K., Kim, S., and Iwao, H. (1998).et al. (1998). UTF1, a novel transcriptional coactivator expressed inThe human gene encoding the lectin-type oxidized LDL receptorpluripotent embryonic stem cells and extra-embryonic cells. EMBO(OLR1) is a novel member of the natural killer gene complex with aJ. 17, 2019-2032.unique expression profile. Genomics 54, 191–199.

Palmieri, S.L., Peter, W., Hess, H., and Scholer, H.R. (1994). Oct-4Yamanaka, S., Zhang, X.Y., Maeda, M., Miura, K., Wang, S., Farese,transcription factor is differentially expressed in the mouse embryoR.V., Jr., Iwao, H., and Innerarity, T.L. (2000). Essential role of NAT1/during establishment of the first two extraembryonic cell lineagesp97/DAP5 in embryonic differentiation and the retinoic acid path-involved in implantation. Dev. Biol. 166, 259–267.way. EMBO J. 19, 5533–5541.

Ramalho-Santos, M., Yoon, S., Matsuzaki, Y., Mulligan, R.C., andYoshida, K., Taga, T., Saito, M., Suematsu, S., Kumanogoh, A., Ta-Melton, D.A. (2002). “Stemness”: transcriptional profiling of embry-naka, T., Fujiwara, H., Hirata, M., Yamagami, T., Nakahata, T., et al.onic and adult stem cells. Science 298, 597–600.(1996). Targeted disruption of gp130, a common signal transducer

Raz, R., Lee, C.K., Cannizzaro, L.A., d’Eustachio, P., and Levy, D.E. for the interleukin 6 family of cytokines, leads to myocardial and(1999). Essential role of STAT3 for embryonic stem cell pluripotency. hematological disorders. Proc. Natl. Acad. Sci. USA 93, 407–411.Proc. Natl. Acad. Sci. USA 96, 2846–2851.

Yuan, H., Corbi, N., Basilico, C., and Dailey, L. (1995). Develop-Reubinoff, B.E., Pera, M.F., Fong, C.Y., Trounson, A., and Bongso, A. mental-specific activity of the FGF-4 enhancer requires the syner-(2000). Embryonic stem cell lines from human blastocysts: somatic gistic action of Sox2 and Oct-3. Genes Dev. 9, 2635–2645.differentiation in vitro. Nat. Biotechnol. 18, 399–404.

Rogers, M.B., Hosler, B.A., and Gudas, L.J. (1991). Specific expres- Accession Numberssion of a retinoic acid-regulated, zinc-finger gene, Rex-1, in preim-plantation embryos, trophoblast and spermatocytes. Development The GenBank accession number for the mouse nanog cDNA se-113, 815–824. quence reported in this paper is AB093574.Schneider, T.D., and Stephens, R.M. (1990). Sequence logos: a newway to display consensus sequences. Nucleic Acids Res. 18, 6097–6100.

Scholer, H.R., Dressler, G.R., Balling, R., Rohdewohld, H., and Gruss,P. (1990). Oct-4: a germline-specific transcription factor mappingto the mouse t-complex. EMBO J. 9, 2185–2195.

Smith, A. (1998). Cell therapy: in search of pluripotency. Curr. Biol.8, R802–R804.

Smith, A.G., Heath, J.K., Donaldson, D.D., Wong, G.G., Moreau, J.,Stahl, M., and Rogers, D. (1988). Inhibition of pluripotential embry-onic stem cell differentiation by purified polypeptides. Nature 336,688–690.

Stewart, C.L., Kaspar, P., Brunet, L.J., Bhatt, H., Gadi, I., Kontgen,F., and Abbondanzo, S.J. (1992). Blastocyst implantation dependson maternal expression of leukaemia inhibitory factor. Nature 359,76–79.

Sun-Wada, G.H., Manabe, S., Yoshimizu, T., Yamaguchi, C., Oka,T., Wada, Y., and Futai, M. (2000). Upstream regions directing heart-specific expression of the GATA6 gene during mouse early develop-ment. J. Biochem. (Tokyo) 127, 703–709.

Takahashi, K., Mitsui, K., and Yamanaka, S. (2003). Role of ERas intumor-like properties in mouse embryonic stem cells. Nature 423,541–545.

Takeda, K., Noguchi, K., Shi, W., Tanaka, T., Matsumoto, M., Yo-shida, N., Kishimoto, T., and Akira, S. (1997). Targeted disruption ofthe mouse Stat3 gene leads to early embryonic lethality. Proc. Natl.Acad. Sci. USA 94, 3801–3804.

Thomson, J.A., Itskovitz-Eldor, J., Shapiro, S.S., Waknitz, M.A.,Swiergiel, J.J., Marshall, V.S., and Jones, J.M. (1998). Embryonicstem cell lines derived from human blastocysts. Science 282, 1145–1147.

Tokuzawa, Y., Kaiho, E., Maruyama, M., Takahashi, K., Mitsui, K.,Maeda, M., Niwa, H., and Yamanaka, S. (2003). Fbx15 is a novel