Xenopus brain factor-2 controls mesoderm, forebrain and neural

Xenopus brain factor-2controls mesoderm, forebrainand neural crest development

JoseLuis Gomez-Skarmetaa, Elisa de la Calle-Mustienesa,Juan Modolellb, Roberto Mayora,*

aLaboratorio de Biologıá del Desarrollo, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, ChilebCentro de Biologıá Molecular Severo Ochoa, Consejo Superior de Investigaciones Cientı´ficas and Universidad Autońoma de Madrid,

Cantoblanco, 28049 Madrid, Spain

Received 9 September 1998; revised version received 29 September 1998; accepted 30 September 1998

Abstract

The forkhead type Brain Factor 2 from mouse and chicken help pattern the forebrain, optic vesicle and kidney. We have isolated aXenopushomolog (Xbf2) and found that during gastrulation it is expressed in the dorsolateral mesoderm, where it helps specify this territoryby downregulatingBMP-4and its downstream genes. Indeed,Xbf2overexpression caused partial axis duplication. Interference with BMP-4signaling also occurs in isolated animal caps, sinceXbf2 induces neural tissue. Within the neurula forebrain,Xbf2and the relatedXbf1geneare expressed in the contiguous diencephalic and telencephalic territories, respectively, and each gene represses the other. Finally,Xbf2seems to participate in the control of neural crest migration. Our data suggest that XBF2 interferes with BMP-4 signaling, both in mesodermand ectoderm. 1999 Elsevier Science Ireland Ltd. All rights reserved

Keywords: Xbf2; Mesoderm; Neural induction; Neural crest;Xenopus

1. Introduction

Families of transcription factors present in restricted ter-ritories within embryos generate positional identity andcontrol morphogenesis of different tissues. The ‘wingedhelix’ or ‘forkhead’ family of transcription factors com-prise a large and growing group whose members aredefined by a common 110 amino acid DNA binding domain(reviewed in Kaufmann and Kno¨chel, 1996). The determi-nation of the crystal structure of the DNA-binding domainof one of its members (HNK-3g) has revealed that thisregion is a variant of the helix-turn-helix motif (Clark etal., 1993). A 20 amino acid region N-terminal to helix 3 hasbeen shown to be important for binding site specificity(Overdier et al., 1994).

Since its original discovery as a region of homologybetween the rat hepatocyte nuclear factor 3 (HNF-3; Laiet al., 1991) and theDrosophila forkheadgene (Weigeland Ja¨ckle, 1990), the forkhead domain has been found inmore than 60 proteins encoded by genes from species ran-ging from yeast to human (reviewed in Kaufmann andKnochel, 1996). Analysis of expression patterns as well asthe phenotypes observed under gain and loss of functionconditions have implicated these genes in pattern formationduring embryogenesis. Several members of the wingedhelix gene family are expressed in the central nervous sys-tem in different organisms (Ha¨cker et al., 1992; Tao andLai, 1992; Clevidence et al., 1993; Kaestner et al., 1993;Hatini et al., 1994; Murphy et al., 1994; Pierrou et al., 1994;Lef et al., 1996; Yuasa et al., 1996). The expression ofbrainfactor 1 (bf1) andbrain factor 2(bf2), two members of thisfamily, occupy adjacent domains within the developingmammalian and avian retina and forebrain (Tao and Lai,1992; Hatini et al., 1994). Loss of function experiments of

Mechanisms of Development 80 (1999) 15–27

0925-4773/99/$ - see front matter 1999 Elsevier Science Ireland Ltd. All rights reservedPII S0925-4773(98)00190-7

* Corresponding author. Tel.: +34-562-678-7271; fax: +34-562-271-2983.

bf1 and 2 indicate that they participate in forebrain andkidney development, respectively (Xuan et al., 1995; Hatiniet al., 1996).

A search forforkheadrelated genes inXenopusrevealedthe existence of a multigene family, all of whose membersshare the 110 amino acid conserved domain albeit at vary-ing degrees of similarity. They are calledXenopus forkheaddomainrelated genes orXFD (Knochel et al., 1992). EachXFD has a specific spatial pattern of expression and thesepatterns partially overlap with one another (Dirksen andJamrich, 1992; Kno¨chel et al., 1992; Ruiz i Altaba andJessell, 1992). These and other data obtained with severalorganisms suggest that members of this family participate incomplex networks required for pattern formation and tissuedifferentiation during early embryogenesis. The forkheadregion of XFD-6 and -9 share a high identity with that ofmouse and chick Brain Factor 2 (BF2). However, outside ofthis domain they diverge from each other and from BF2proteins, which suggests that they belong to a BF2 subfam-ily of forkhead, but are not true homologues of BF2. Inaddition, a partial cDNA fragment encoding aXenopushomologue ofbf1 has also been identified. This gene isexpressed in the prospective telencephalon (Papalopuluand Kintner, 1996). However, no functional analyses havebeen carried out.

In this paper we report the cloning of theXenopushomo-logue of the chicken and mousebrain factor 2. We call itXenopus brain factor 2(Xbf2). Xbf2 is expressed duringearly embryonic development in mesoderm, anterior neuralplate and neural crest cells. Overexpression experiments inembryos and in isolated tissues allowed us to discern pos-sible functions ofXbf2. Dorsolateral mesoderm expressionappears important for patterning of the mesoderm, as itrepresses ventral mesoderm genes and promotes theexpression of dorsolateral genes such asXmyoD. Anteriorneural plate expression ofXbf2, which repressesXbf1,could be necessary for patterning of the forebrain. Over-expression and grafting experiments suggest thatXbf2represses the migration of neural crest cells. All of thesefunctions may be mediated by an interference with BMP-4signaling.

2. Results

2.1. Molecular cloning of Xbf2

In a low stringency screen of aXenopusgastrula (stages10.5–11.5) cDNA library we identified a 1.8 kb cDNAclone which corresponded to a gene with strong expressionin the region of the neural folds of stage 17 embryos (seebelow). This cDNA encoded a protein that was a newmember of the winged helix or forkhead family of tran-scription factors. Moreover, the primary structure of thisprotein outside of the winged helix domain indicated that,most likely, it was theXenopushomologue of BF2 from

mouse and chicken (Hatini et al., 1994; Yuasa et al., 1996).Comparison between MBF2, CBF2 and XBF2 showed 97%identity and 99% similarity within the forkhead domainand 50% identity and 70% similarity outside this domain(Fig. 1).

2.2. Xbf2 gene is expressed in mesoderm, anterior neuralplate and prospective neural crest cells

The earliest detectable expression ofXbf2, as detected bywhole-mount in situ hybridization to developing embryos,occurs at gastrula stage 10 (Niewkoop and Faber, 1967) in adorso-lateral gradient at the marginal zone (Fig. 2A). Thisexpression continues in the prospective uninvoluted meso-derm (Fig. 2B), but it becomes undetectable during gastru-lation once the mesoderm has involuted. No expression isdetected in the ventral marginal zone. At neurula stage 12.5,expression occurs in the anterior neural plate and is loca-lized close to the dorsal midline, which corresponds to theprospective diencephalon and eye field (Fig. 2C, arrow). Atthis stage faint expression is also observed in the anteriorneural folds (Fig. 2C, arrowhead). As neurulation proceeds,Xbf2 expression becomes stronger in the anterior neuralplate and in the anterior neural folds (Fig. 2D,E). At stage17, the subdivision, within the cephalic neural crests, of theprospective mandibular, hyoid and branchial arches can bedistinguished (Fig. 2E, arrowhead). At a more advancedstage, the expression ofXbf2 becomes detectable in theoptic vesicle (Fig. 2F, arrow), while the gene is switchedoff in the prospective neural crest at the time these cellsbegin migration (Fig. 2F). At this stage, expression alsotakes place in the paraxial mesoderm (Fig. 2F, arrowhead).

2.3. Ectopic expression of Xbf2 induces a partial secondaryaxis and enlarges the neural plate

To analyze the function ofXbf2in Xenopusdevelopment,we injected 1- or 2-cell stage embryos with differentamounts ofXbf2 mRNA. Injection of 4 ng of mRNA in 1-cell stage embryos induced formation of a partial secondaryaxis and a significant reduction of the head (Fig. 3A). Whenthe same amount of mRNA was injected in one blastomereof a 2-cell stage embryo, a partial secondary axis was themost frequent outcome. Injection of 0.25 ng of mRNA inone cell of a 2-cell stage embryo produced a small expan-sion of the neural plate in approximately half of the embryos(not shown) and, with a low frequency (5%), a secondaryaxis. A reduction of head size was also observed in mostembryos when this amount of mRNA was injected at the 1-cell stage. The secondary axis produced under this conditionexpressed weak but detectable levels of neural crest markerssuch asXtwist(Hopwood et al., 1989b) andXslug(Mayor etal., 1995) (Fig. 3B,C, arrowheads). The expanded neuralplate was recognized by the expression of the neural platemarkerXsox2(Mizuseki et al., 1998) (Fig. 3D, arrowhead).High amounts of injected mRNA (4 ng) lead to an expanded

16 J.L. Gomez-Skarmeta / Mechanisms of Development 80 (1999) 15–27

ventral ectopic expression ofXsox2, which was apparentlyassociated with the duplicated axis present in most of theseembryos (Fig. 3E, arrow). The ectopic axis lacked the ante-rior-most tissues, such as eyes or cement gland (not shown).

In addition to neural markers,Xbf2 overexpression alsoaffected the paraxial mesodermal markerXMyoD (Hop-wood et al., 1989a), as its domain of expression wasexpanded (Fig. 3F, arrow).

Fig. 1. Comparison of the sequences of mouse, chicken andXenopusBF2 proteins (accession numbers Q61345, U47276, AJO11652). The forkhead domainis under and overlined. Note the high similarity between all BF2 proteins outside the forkhead domain.

17J.L. Gomez-Skarmeta / Mechanisms of Development 80 (1999) 15–27

2.4. Expression of Xbf2 in the ventral marginal zoneconfers on it neural inducer activity

The ectopic expression of neural plate markers and theinduction of a partial secondary axis inXbf2 injectedembryos might be due to a dorsalization of the ventral meso-

derm, which could endow it with neural inducer activity,and/or to a direct neuralization of the ectoderm. To inves-tigate the first possibility, we analyzed whether the expres-sion ofXbf2in ventral mesoderm was sufficient to allow thistissue to induce neural plate in naive ectoderm in vitro.Animal caps and dorsal and ventral marginal zones were

Fig. 2. Expression pattern ofXbf2 in Xenopusembryos at different developmental stages. (A–E) Dorsal views, anterior is up; (F) lateral view, anterior is up.(A) At stage 10Xbf2 is initially detected in the dorso-lateral prospective mesoderm. (B) Stage 11 embryo showingXbf2 expression in the uninvoluteddorsolateral prospective mesoderm. (C) Stage 12.5; at this stage expression appears in the anterior neural plate (arrow) and in the anterior neural folds(arrowhead). (D,E) At stages 14 and 17, respectively,Xbf2expression becomes stronger in the anterior neural plate (arrows) and in the anterior neural folds(arrowheads). At stage 17,Xbf2 is clearly expressed in the segmented premigratory cephalic neural crest (arrowhead). (F) Stage 20 embryo showingXbf2expression in the eye (arrow) and in the paraxial mesoderm (arrowhead). Note that at this stage the expression in the migratory neural crest has disappeared.


dissected at stage 10, were conjugated in different combina-tions and were cultured until the equivalent of stage 17,when expression of the neural plate markerXsox2was ana-lyzed. Animal caps cultured alone (Fig. 4A) or conjugatedwith ventral marginal zone (Fig. 4C) never expressed

Xsox2; but animal caps conjugated with ventral marginalzone taken from embryos injected with 4 ng ofXbf2mRNA, did expressXsox2. (Fig. 4D). This inducing activitywas similar, although not as strong, as the expressioninduced by dorsal marginal zone (Fig. 4B). As expected,

Fig. 3. Effect of Xbf2 overexpression on embryonic development. Embryos were injected at the 1-cell stage (A) or 2-cell stage (B–F) with differentconcentrations ofXbf2mRNA and 0.3 ng oflacZ mRNA as a lineage marker (blue dots). Embryos were cultured until stage 16 or 25, when the expressionof different genes was analyzed by in situ hybridization. Views are from the dorsal side, anterior is up. (A) Reduced head and absence of anterior structures,such as eyes or cement gland (arrow) and partial secondary axis (arrowhead) in an embryo injected with 4 ng ofXbf2mRNA (95%,n = 22). (B–D) Low levelof Xtwist (B) or Xslug (C) expression in the secondary axis of an embryo injected with 0.25 ng ofXbf2 mRNA. This duplicate axis is clearly detected byXsox2(D) expression (arrowheads point to the double axis that were found in about 5% of the injected embryos,n = 32, 43 and 54, respectively). (E) Ectopicexpression ofXsox2close to the ventral region of an embryo injected with 4 ng ofXbf2mRNA (arrow; 93%,n = 54). (F) Embryos injected with 4 ng ofXbf2mRNA also showed expansion of theXMyoD domain in the injected side (arrow; 75%,n = 40).


ventral marginal zones taken from injected embryos andcultured alone did not expressXsox2 (Fig. 4D, inset).These results indicate that expression ofXbf2 in the pro-spective ventral mesoderm changes its inducing capacity,possibly by transforming it to a more dorsal mesoderm. Asshown below,Xbf2can also directly neuralize the ectoderm.

2.5. Xbf2 inhibits ventral mesodermal genes

We examined whether the expression of several mesoder-mal genes was affected by the injection of 4 ng ofXbf2mRNA in the equatorial region of one blastomere of the2-cell stage embryo (arrowheads in Fig. 5 point to theinjected sites in all embryos). The expression of the dorsalmesodermal genesgsc(Cho et al., 1991) (Fig. 5A) andXlim1 (Taira et al., 1992) (Fig. 5B) was not affected. However,

XMyoDexpression, which is normally transcribed at a morelateral position, was strengthened in the injected side (Fig.5C) and clearly expanded at later stages (Fig. 3F). Ventraland lateral injections ofXbf2 mRNA did not affect theexpression ofXvent2, a slightly more ventral mesodermalgene (Onichtchouk et al., 1996) (Fig. 5D). However, dorsalinjections weakly inhibitedXvent2expression but only at itsmore dorsal domain (not shown). In contrast, the more ven-tral genesXwnt8(Christian et al., 1991; Smith and Harland,1991) andXvent1 (Gawantka et al., 1995) were stronglyinhibited in the injected side (Fig. 5E,F, respectively).SinceXvent1andXvent2are downstream genes of BMP-4(Gawantka et al., 1995; Onichtchouk, et al., 1996), we ana-lyzed the expression ofBMP-4 in the Xbf2 injectedembryos. Xbf2 overexpression apparently reduced theexpression ofBMP-4 (Fig. 5G) although not as strongly asthat of other ventral genes. Similarly,Xmsx1, another BMP-4 downstream gene (Maeda et al., 1997; Suzuki et al., 1997),was also inhibited at the injected site whenXbf2was over-expressed (Fig. 5 H). Taken together, these results suggestthatXbf2participates in the specification of the lateral pro-spective mesoderm (XMyoD domain), and that its overex-pression leads to a lateralization of the ventral mesoderm.

2.6. Direct neural induction by Xbf2

We tested whetherXbf2could neuralize ectodermal cellsdirectly, that is, in the absence of inducing signals emanat-ing from the mesoderm. Embryos were injected at the 1-cellstage with either 0.25 ng or 4 ng ofXbf2 mRNA. Animalcaps were dissected at stage 9 and cultured until the equiva-lent of stage 17, whenXsox2 expression was analyzed.Uninjected animal caps (Fig. 6A, inset) or caps injectedwith 0.25 ng ofXbf2 mRNA did not expressXsox2, whilecaps from embryos injected with 4 ng ofXbf2 mRNAstrongly expressed this neural plate marker (Fig. 6A).These caps did not express the mesodermal markerXmyoD (not shown). Since neural induction apparentlyrequires the suppression of BMP-4 signals in the ectoderm(reviewed in Hemmati-Brivanlou and Melton, 1997), weexamined whether expression ofXmsx1was affected inembryos injected with 4 ng ofXbf2 in one cell of the 2-cell stage. At stage 14, its expression was suppressed in theposterior ventral region, where it should be expressed verystrongly (not shown), and at the neurula stage, when thisgene is mainly expressed in the neural fold area, its expres-sion was also strongly reduced (Fig. 6B, arrowhead). Theinhibition of Xmsx1expression suggests that XBF2 inhibitsBMP signaling. As a consequence, ectodermal cells areneuralized.

2.7. Patterning the forebrain by mutual inhibition of Xbf1and Xbf2

In mouse and chicken,bf1andbf2genes are expressed inadjacent domains within the developing retina and forebrain

Fig. 4.Xbf2affects the neural inducer capacity of the mesoderm. Differenttissues were dissected from control orXbf2 injected stage 10 embryos,conjugated and cultured until the equivalent of stage 17, whenXsox2expression was analyzed. (A) Control animal caps showed noXsox2expression (n = 15). (B) All the conjugates of animal cap and dorsal mar-ginal zone strongly expressedXsox2(n = 14). (C) Animal caps conjugatedwith ventral marginal zone did not expressXsox2(n = 17). (D) Most of theconjugates of animal cap and ventral marginal zone taken from embryosinjected with 4 ng ofXbf2 mRNA expressedXsox2 (70%, n = 13),although at lower levels than conjugates of animal cap and dorsal marginalzone. In contrast, the ventral marginal zone alone taken from injectedembryos did not expressXsox2(inset;n = 10).


Fig. 5. Effect ofXbf2mRNA on mesodermal development. Embryos were injected in one cell at the 2-cell stage in the equatorial region with 4 ng ofXbf2mRNA and 0.3 ng oflacZmRNA and cultured until stage 10.5–11.5, when the expression of different mesodermal markers were analyzed. Arrowheads pointto the injected sites. Neithergsc (A) nor Xlim1 (B) expression were affected byXfb2 overexpression (100%,n = 18 and 13, respectively). (C)XMyoDexpression was strengthened in the lateral mesoderm in the injected side (60%,n = 34). (D) Ventral or lateralXbf2 injections did not affectXvent2expression(94%,n = 48). In contrast, overexpression ofXbf2strongly inhibited the expression ofXwnt8(E; 95%,n = 45) andXvent1(F; 92%n = 23) and weakly thatof BMP-4 (G; 80%n = 15) andXmsx1(H; 58%,n = 12) at the injection site.


(Tao and Lai, 1992; Hatini et al., 1994). We find that in theXenopusforebrain region these genes are also expressed inadjacent domains, that ofXbf1being more anterior than thatof Xbf2 (Fig. 7A). This complementary pattern of expres-sion suggests a mutually inhibitory regulation between thesetwo genes. To test this, we injected one blastomere of 2-cellstage embryos with low amounts ofXbf2mRNA (0.25 ng, toavoid secondary axis induction) and we examined theexpression ofXbf1 (Papalopulu and Kintner, 1996). Theinjected side showed a strong reduction or elimination ofXbf1 expression (Fig. 7C, arrowhead). Similarly, injectionof 0.5 ng ofXbf1mRNA inhibitedXbf2expression (Fig. 7B,arrowhead). This mutual inhibition could help patterningthe forebrain.

We also analyzed whether other forebrain-expressedgenes were affected byXbf2 overexpression.Xcpl1, agene expressed at the anterior-most region of the neuralfold (anterior toXbf1, Good et al., 1990), was also inhib-ited (Fig. 7D, arrowhead). However,Xotx2, a gene ex-pressed in a broad anterior area that includes the regionsof expression ofXbf1, Xbf2 and Xcpl1, was not affected(not shown). The inhibition ofXbf1 and Xcpl1 suggeststhat the overexpression ofXbf2 could transform the tele-ncephalic territory into a diencephalic region, while itmaintains an anterior character, as determined byXotx2expression.

2.8. Xbf2 inhibits neural crest migration

Xbf2 is expressed in the premigratory neural crest cells(Fig. 2C–E) and its expression is completely extinguishedonce these cells start to migrate. Thus, we examinedwhether Xbf2 extinction was required for migration ofneural crest cells by analyzing the effect ofXbf2 overex-pression on this process. Embryos injected in one blasto-mere at the 2-cell stage with 0.25 ng ofXbf2 mRNA wereanalyzed for the neural crest markerXtwistonce the neuralcrest cells began migrating (stage 25). At the uninjectedside, expression ofXtwist and migration of the crest cellswere normal (Fig. 8A, arrowhead). However, at the injectedside, levels ofXtwistexpression were normal but migrationdid not occur (Fig. 8B, arrow). To rule out the possibilitythat the overexpression ofXbf2was affecting the migratorypath of the crest cells, stage 10 ectoderm injected withlacZmRNA, as a lineage marker, with or withoutXbf2 mRNAwas grafted into the prospective neural crest region of nor-mal stage 11 embryos. Crest cells from the grafts migratednormally when they contained only the lineage marker (Fig.8C, arrowhead), but did not do so when they contained inaddition Xbf2 mRNA (Fig. 8D, arrow). Moreover,Xsox2expression was not detected in these grafts (Fig. 8D, inset)or in grafts cultured in isolation (not shown), indicating thatthe graft expressingXbf2 was not transformed into neuralplate. These results thus suggest that extinction ofXbf2expression is a requisite for the migration of the neuralcrest cells.

3. Discussion

3.1. Expression of Xbf2

Avian and mousebf1 andbf2 genes play important rolesin patterning the forebrain, retina and kidney (Xuan et al.,1995; Hatini et al., 1996). In the forebrain, they probablyspecify the positional identity of the telencephalic neuroe-pithelium and determine its proliferative potential duringbrain development (Hatini et al., 1994; Xuan et al., 1995).We now find that the expression of theXenopushomologueXbf2 in the anterior neural plate is compatible with thisfunction. However,Xbf2 is also expressed in two otherregions, which suggests additional functions. The expres-sion in premigratory neural crest cells is similar to theexpression of otherXenopus forkheadgenes (XFD-3, -6, -9, -10; Lef et al., 1996), while the early expression in pro-spective mesoderm is a novel one for abrain factor gene.Our data suggest that XBF2 acts as a transcriptional repres-sor. Indeed other forkhead proteins, such as Qin, Genesis,CWH-1, CWH-3 and CWH-3, behave as transcriptionalrepressors (Li et al., 1995; Sutton et al., 1996; Freyaldenho-ven et al., 1997).

3.2. Xbf2 participates in mesodermal development

The dorso-ventral patterning of the mesoderm is a com-plex process in which dorsal signals (e.g. noggin, chordin,Frzb-1) interact directly with ventral signals (BMP-4,Wnt8) (Holley et al., 1996; Piccolo et al., 1996; Zimmermanet al., 1996; Leyns et al., 1997; Wang et al., 1997; reviewedin Hemmati-Brivanlou and Melton, 1997 and Moon et al.,1997). As a result, dorso-ventral gradients of BMP-4 andXwnt8 are established. These locally activate transcriptionfactors which specify distinct mesodermal territories. Someof these factors are encoded by theXmsx1, Xvent1 andXvent2genes and have ventralizing activities (Gawantkaet al., 1995; Onichtchouk et al., 1996; Dosch et al., 1997;Maeda et al., 1997; Suzuki et al., 1997; Onichtchouk et al.,1998).

It has been proposed that the combination ofXvent1andXvent2 functions specifies different mesodermal cells(Onichtchouk et al., 1998). Thus, in ventral mesodermboth genes are expressed and this suppresses notochordand muscle differentiation. In dorsolateral mesoderm onlyXvent2is expressed and this inhibits notochord but allowsmuscle differentiation. The overexpression ofXbf2 inhibitsstronglyXvent1but very weaklyXvent2, and this only at itsdorsal-most domain of expression. As a consequence, theterritory that expressesXvent2alone is expanded and thisshould be compatible with muscle differentiation. Indeed,the domain of expression ofXMyoDis expanded in theXbf2injected embryos. In contrast,Xbf2overexpression does notinduce ectopic expression of dorsal genes such asgscandXlim1. A possible interpretation of these results is thatXbf2,which is expressed in the dorsal and dorso-lateral meso-


Fig. 6. Direct neuralization of the ectoderm byXbf2expression. Animal caps were dissected from stage 9 embryos and cultured until the equivalent of stage17, whenXsox2expression was analyzed. (A) Animal caps taken from embryos injected with 4 ng ofXbf2mRNA at the 1 cell stage.Xsox2is expressed inthese caps (58%,n = 12). Note also that these caps are not elongated as could be expected if dorsal mesoderm was induced. Inset: control animal capsshowing noXsox2expression (100%,n = 13). (B) Embryos were injected in one blastomere at the 2-cell stage with 4 ng ofXbf2mRNA and 0.3 ng oflacZmRNA. Xmsx1expression was analyzed at stage 16 embryos. Notice the absence ofXmsx1expression at the injected side (arrowhead; 83%,n = 18).

Fig. 7.Xbf2participates in patterning of the forebrain. All the embryos are shown from anterior side. The weaker expression ofXbf2 in neural crest cells isnot visible in A and B due to low staining development. (A) Double in situ hybridization of an stage 17 embryo. Blue (arrow) corresponds toXbf2expressionand purple (arrowhead) denotesXbf1expression. (B)Xbf2expression in an stage 17 embryo injected in one blastomere of a 2-cell stage embryo with 0.5 ngof Xbf1mRNA and 0.3 ng oflacZ mRNA. Notice the complete inhibition ofXbf2expression at the injected side (58%,n = 12). (C) Embryo injected in oneblastomere of a 2-cell stage embryo with 0.25 ng ofXbf2 mRNA and 0.3 nglacZ mRNA and cultured until stage 17.Xbf1 expression is suppressed in theinjected side (72%,n = 22). (D) Similar embryo as in C, but expression ofXcpl1 was analyzed. Notice the inhibition ofXcpl1 in the injected side (53%,n = 15).


derm, specifies the dorso-lateral mesoderm but is incapableof inducing a more dorsal fate by itself. The dorsalizingactivity of Xbf2 could then be due to an interference withBMP-4 and/orXwnt8. The observation that XBF2 inhibitsthe expression ofBMP-4 and of its downstream genes andinduces neuralization in caps suggests that XBF2 antago-nizes BMP-4 signaling. The inhibition ofXwnt8caused byXbf2 overexpression could then be due to this interferencewith BMP-4 signaling, since a dominant negative BMP-4receptor suppressesXwnt8expression (Schmidt et al., 1995;Hoppler and Moon, 1998).

In contrast to our observed inhibition ofXwnt8 togetherwith an expansion of theXMyoDdomain, it has been shownthat a dominant negative form of Xwnt8 inhibitsXMyoD(Hoppler et al., 1996). This apparent discrepancy may beexplained if this dominant negative has a broader than

expected range of activity and also antagonizes otherXwnt molecules (Hoppler et al., 1996) that may participatein XMyoD activation. In addition, XBF2 may activateXMyoD directly or via another gene required for its tran-scription.

3.3. Xbf2 and anterior neural plate development

In chicken and mouse,bf1 is expressed in the telencepha-lon and in the rostral half of the retina, adjacent to thedomains ofbf2expression within the diencephalon and pos-terior retina (Tao and Lai, 1992; Yuasa et al., 1996). Simi-larly, Xbf1 andXbf2 have adjacent domains of expressionwithin the forebrain. Our results suggest that, within theforebrain, eachXbf gene is required to specify the territorywhere it is expressed, as an expansion of the domain of one

Fig. 8. Overexpression ofXbf2blocks neural crest migration. (A,B) One blastomere of a 2-cell stage embryo was injected with 0.25 ng ofXbf2mRNA and0.3 ng oflacZ mRNA, was cultured until stage 25 and the expression ofXtwistwas analyzed. (A) Right side view of the embryo shown in (B). (A) View ofuninjected right side. Notice the normal expression ofXtwistand the migration of the neural crest cells (arrowhead). (B) View of injected left side. Notice thatthe cells still expressXtwistbut they do not migrate normally, most of them remaining in the dorsal side of the embryo (arrow; 58%,n = 17). (C,D) Grafts ofectodermal tissue, taken from stage 10 embryos previously injected with 0.3 ng oflacZ mRNA without (C) or with (D) 0.25 ng ofXbf2mRNA, were placedin the prospective neural crest region of a stage 11 embryo. The embryos were cultured until stage 28 andb-galactosidase was analyzed. (C) Control embryothat received a graft containinglacZ mRNA. Note in C the normal migration of the grafted cells (arrowhead; 87%,n = 15) and in D the permanence of thelabeled cells in the original site of the graft (arrow; 87%,n = 15). Inset: a similar, but more lateral (to better distinguishb-galactosidase staining fromXsox2expression) graft, does not showXsox2expression (100%,n = 12).


gene leads to the reduction of the domain of the other,probably by mutual repression. In addition,Xotx2 expres-sion is not affected inXbf2 injected embryos. This suggeststhat the anterior neural plate could be transformed into amore posterior territory, although still within theXotx2expression domain.

3.4. Xbf2 interferes with neural crest cell migration

Xbf2is expressed in the neural crest region up to the timecrest cells begin to migrate. Our results, obtained in over-expression and grafting experiments, suggest thatXbf2canrepress neural crest migration. As XBF2 exhibits a repres-sor activity in other tissues, it may also inhibit genes whichare necessary for neural crest cell migration. This couldexplain the switching off ofXbf2 before crest cells startto migrate. This migration requires an epithelial tomesenchymal transition of the cells, and although this pro-cess has been extensively studied and many molecules havebeen proposed to be involved in it, not much is knownabout the signal that triggers such change (see review inMayor et al., 1998). Thus, it is possible that XBF2 repressesa gene which is required for the initiation of the epithelial tomesenchymal transition that precedes neural crest migra-tion.

3.5. Xbf2 and BMP signaling

Ectopic neural development uponXbf2 overexpressioncan be due to the simultaneous dorso-lateralization of theventral mesoderm and the neuralization of ectoderm. Botheffects can be explained by XBF2 interference with BMP-4signaling. Note that the combined action of organizingmolecules (noggin, chordin and follistatin), which interferewith BMP-4 by direct binding, and XBF2, which repressestranscription ofBMP-4, may help establish the dorso-ven-tral gradient of BMP-4 activity necessary for proper devel-opment of the mesoderm.

The complementary domains of expression ofXbf1 andXbf2may also be related to BMP-4 signaling. In fact, it hasbeen found that the cement gland, an anterior marker, isinduced at slightly higher levels of BMP activity thanthose required for neural plate induction (Wilson et al.,1997). Thus, different levels of BMP-4 signaling may estab-lish the adjacent territories ofXbf1 and Xbf2 expression.Subsequently, their mutual repression and the interferenceof XBF2 with the BMP-4 pathway may help refine theirdomains of expression.

Finally, it is of interest that differentBMPsare expressedduring neural crest specification, migration and develop-ment (Liem et al., 1995; Schmidt et al., 1995). It is possiblethat XBF2 represses some BMPs and/or their downstreamfactors and that this repression is required for proper devel-opment and migration of the neural crest cells. Thus, it ispossible that XBF2 blocks BMP-4 signaling in all tissueswhere it is expressed and this interference is necessary for

mesoderm, anterior neural plate and neural crest develop-ment.

4. Materials and methods

4.1. Molecular cloning of the Xbf2 gene

Approximately 106 phages from aXenopusgastrula(stages 10.5–11.5) UniZap-XR cDNA library (Cho et al.,1991) were screened at low stringency with aDrosophilaabrupt probe, in hybridization buffer (40% formamide,4× SSC, 5× Denhartd’s, 0.1% SDS, 20 mM phosphate buffer(pH 6.8), 0.1 mg/ml salmon sperm DNA and 0.1% sodiumpyrophosphate) at 42°C. Filters were washed with 2× SSC,0.1% SDS at 42°C. Seven positives were obtained, werepurified and the corresponding plasmids were excised.Unexpectedly, one of the cDNAs encoded theXenopushomologue of the mouse and chickenbrain factor 2genes.

4.2. DNA sequencing

Sequencing was performed in an automatic DNA sequen-cer using ABI chemistry and T3, T7 and custom synthesizedoligonucleotides (ISOGEN Bioscience BV, Maarsesen, TheNetherlands) as primers. Consensus sequences wereassembled and analyzed with the University of WisconsinGCG software packages (Devereux et al., 1984).

4.3. Whole-mount in situ hybridization and X-Gal staining

Antisense RNA probes were prepared fromXbf-2cDNAandXslug(Mayor et al., 1995),gsc(Cho et al., 1991),Xlim1(Taira et al., 1992),Xsox2(kindly provided for Dr. R.M.Grainger), Otx-2 (Blitz and Cho, 1995; Pannese et al.,1995),Xtwist (Hopwood et al., 1989b),XMyoD (Hopwoodet al., 1989a),Xvent1 (Gawantka et al., 1995),Xvent2(Onichtchouk et al., 1996),Xcpl1 (Good et al., 1990),Xmsx1(Maeda et al., 1997; Suzuki et al., 1997),Xwnt8(Christian et al., 1991; Smith and Harland, 1991) andBMP-4 (Schmidt et al., 1995) cDNAs using digoxigeninor fluoresceine as labels. Specimens were prepared, hybri-dized and stained by the method of Harland (1991) withmodifications (Mancilla and Mayor, 1996). Double in situhybridizations were made as described in Go´mez-Skarmetaet al. (1998).

4.4. In vitro RNA synthesis and plasmid constructs

All the vectors were linearized and transcribed asdescribed by Harland and Weintraub (1985) with GTP capanalog (New England Biolabs). SP6, T3 or T7 RNA poly-merases were used. After DNase treatment, RNA wasextracted with phenol-chloroform and precipitated withethanol. mRNA to be injected was resuspended in water.


4.5. Embryos, explants, conjugates and microinjections

Xenopusembryos were obtained as described previously(Mayor et al., 1993) and staged according to Niewkoop andFaber (1967). Animal caps, marginal zone, conjugates andgrafts were performed as previously described (Mancillaand Mayor, 1996; Mayor et al., 1997). Synthetic mRNAwas injected at the 1- or 2-cell stage embryos in 8–12 nlvolume as described (Mayor et al., 1993).

Acknowledgements

We specially thank N. Papalopulu for theXbf1 cDNAprobe andXbf1 injected embryos prior to publication andM. Allende and C.R. Sharpe for helpful comments on themanuscript. We are grateful to A. Snape for theXenopuscDNA library; to R. Grainger, K.W.Y. Cho, R.T. Moon, M.Sargent, J.B. Gurdon, C. Niehrs, I. Dawid and H. Kung forcDNA probes; to E. Maduenõ and B. de Pablos for help insequencing; and F. Espinoza for technical assistance. Post-doctoral fellowships from Ministerio de Educacioń y Cul-tura to JLGS and Banco Bilbao Vizcaya visitingprofessorship at University of Cambridge to JM areacknowledged. This investigation was supported by grantsfrom FONDECYT (1960910, 7960009 and 3970017), ThirdWorld Academy of Sciences, Fundacioń Andes, Universi-dad de Chile, CSIC/CONICYT, Direccioń General deInvestigacion Cientıfica y Tecnica (PB93-0181), Comuni-dad Autonoma of Madrid, EC (contract CHRX-CT94-0692)and by an institutional grant from Fundacioń Ramon Arecesto the Centro de Biologıá Molecular Severo Ochoa.

References

Blitz, I.L., Cho, K.W.Y., 1995. Anterior neuroectodermis progressivelyinduced during gastrulation: the role of theXenopushomeobox geneorthodenticle. Development 121, 993–1004.

Cho, K.W.Y., Blumberg, B., Steinbeisser, H., de Robertis, E.M., 1991.Molecular nature of Spenmann’s organizer: the role of theXenopushomeobox genegoosecoid. Cell 67, 1111–1120.

Christian, J.L., McMahon, J.A., McMahon, A.P., Moon, R.T., 1991.Xwnt-8, a XenopusWnt-1/int-1-related gene responsive to mesoderm-inducinggrowth factors, may play a role in ventral mesodermal patterning duringembryogenesis. Development 111, 1045–1055.

Clark, K.L., Halay, E.D., Lai, E., Burley, S.K., 1993. Co-crystal structureof the HNF-3/fork head DNA-recognition motif resembles histone H5.Nature 364, 412–420.

Clevidence, D.E., Overdier, D.G., Tao, W., Qian, X., Pani, L., Lai, E.,Costa, R.H., 1993. Identification of nine tissue-specific transcriptionfactors of the hepatocyte nuclear factor 3/forkhead DNA-binding-domain family. Proc. Natl. Acad. Sci. USA 90, 3948–3952.

Devereux, J., Haeverli, P., Smithies, O., 1984. A comprehensive set ofsequence analysis programs for the VAX. Nucl. Acids Res. 12, 387–395.

Dirksen, M.L., Jamrich, M., 1992. A novel, activin-inducible, blastoporelip-specific gene ofXenopus laeviscontains a fork head DNA-bindingdomain. Genes Dev. 6, 599–608.

Dosch, R., Gawantka, V., Delius, H., Blumenstock, C., Niehrs, C., 1997.

Bmp-4 acts as a morphogen in dorsoventral mesoderm patterning inXenopus. Development 124, 2325–2334.

Freyaldenhoven, B.S., Freyaldenhoven, M.P., Iacovoni, J.S., Vogt, P.K.,1997. Avian winged helix proteins CWH-1, CWH-2 and CWH-3 represstranscription from Qin binding sites. Oncogene 15, 483–488.

Gawantka, V., Delius, H., Hirschfeld, K., Blumenstock, C., Niehrs, C.,1995. Antagonizing the Spemann organizer: role of the homeoboxgeneXvent-1. EMBO J. 14, 6268–6279.

Gomez-Skarmeta, J.L., Glavic, A., de la Calle-Mustienes, E., Modolell, J.,Mayor, R., 1998.Xiro, a Xenopushomolog of theDrosophila Iroquoiscomplex genes, controls development at the neural plate. EMBO J. 17,181–190.

Good, P.J., Richter, K., Dawid, I.B., 1990. A nervous system-specificisotype of the beta subunit of Na+,K(+)-ATPase expressed duringearly development ofXenopus laevis. Proc. Natl. Acad. Sci. USA 87,9088–9092.

Hacker, U., Grossniklaus, U., Gehring, W.J., Ja¨ckle, H., 1992. Develop-mentally regulatedDrosophila gene family encoding the fork headdomain. Proc. Natl. Acad. Sci. USA 89, 8754–8758.

Harland, R.M., 1991. In situ hybridization: an improved whole mountmethod forXenopusembryos. Methods Cell Biol. 36, 685–695.

Harland, R., Weintraub, H., 1985. Translation of mRNA injected intoXenopusoocytes is specifically inhibited by antisense RNA. J. CellBiol. 101, 1094–1099.

Hatini, V., Huh, S.O., Herzlinger, D., Soares, V.C., Lai, E., 1996. Essentialrole of stromal mesenchyme in kidney morphogenesis revealed by tar-geted disruption of Winged Helix transcription factor BF-2. Genes Dev.10, 1467–1478.

Hatini, V., Tao, W., Lai, E., 1994. Expression of winged helix genes, BF-1and BF-2, define adjacent domains within the developing forebrain andretina. J. Neurobiol. 25, 1293–1309.

Hemmati-Brivanlou, A., Melton, D.A., 1997. Vertebrate embryonic cellswill become nerve cells unless told otherwise. Cell 88, 13–17.

Holley, S.A., Neul, J.L., Attisano, L., Wrana, J.L., Sasai, Y., O’Connor,M.B., de Robertis, E.M., Ferguson, E.L., 1996. TheXenopusdorsalizingfactor noggin ventralizesDrosophilaembryos by preventing DPP fromactivating its receptor. Cell 86, 607–617.

Hoppler, S., Brown, J.D., Moon, R.T., 1996. Expression of a dominant-negative Wnt blocks induction ofMyoD in Xenopusembryos. GenesDev. 10, 2805–2817.

Hoppler, S., Moon, R.T., 1998. BMP-2/-4 and Wnt-8 cooperatively patternthe Xenopusmesoderm. Mech. Dev. 71, 119–129.

Hopwood, N.D., Pluck, A., Gurdon, J.B., 1989.MyoD expression in theforming somites is an early response to mesoderm induction inXenopusembryos. EMBO J. 8, 3409–3417.

Hopwood, N.D., Pluck, A., Gurdon, J.B., 1989. AXenopusmRNA relatedto Drosophila twistis expressed in response to induction in the meso-derm and the neural crest. Cell. 59, 893–903.

Kaestner, K.H., Lee, K.H., Schlondorff, J., Hiemisch, H., Monaghan, A.P.,Schutz, G., 1993. Six members of the mouse forkhead gene family aredevelopmentally regulated. Proc. Natl. Acad. Sci. USA 90, 7628–7631.

Kaufmann, E., Kno¨chel, W., 1996. Five years on the wings of fork head.Mech. Dev. 57, 3–20.

Knochel, S., Lef, J., Clement, J., Klocke, B., Hille, S., Ko¨ster, M.,Knochel, W., 1992. Activin A induced expression of a fork head relatedgene in posterior chordamesoderm (notochord) ofXenopus laevisembryos. Mech. Dev. 38, 157–165.

Lai, E., Prezioso, V.R., Tao, W.F., Chen, W.S., Darnell, J.E.J., 1991.Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammalsthat is homologous to theDrosophilahomeotic gene fork head. GenesDev. 5, 416–427.

Lef, J., Dege, P., Scheucher, M., Forsbach-Birk, V., Clement, J.H.,Knochel, W., 1996. A fork head related multigene family is transcribedin Xenopus laevisembryos. Int. J. Dev. Biol. 40, 245–253.

Leyns, L., Bouwmeester, T., Kim, S.H., Piccolo, S., de Robertis, E.M.,1997. Frzb-1 is a secreted antagonist of Wnt signaling expressed in theSpemann organizer. Cell 88, 747–756.


Li, J., Chang, H.W., Lai, E., Parker, E.J., Vogt, P.K., 1995. The oncogeneqin codes for a transcriptional repressor. Cancer Res. 55, 5540–5544.

Liem, K.F. Jr., Tremml, G., Roelink, H., Jessell, T.M., 1995. Dorsal dif-ferentiation of neural plate cells induced by BMP-mediated signals fromepidermal ectoderm. Cell 82, 969–979.

Maeda, R., Kobayashi, A., Sekine, R., Lin, J.J., Kung, H., Maeno, M.,1997. Xmsx-1modifies mesodermal tissue pattern along dorsoventralaxis in Xenopus laevisembryo. Development 124, 2553–2560.

Mancilla, A., Mayor, R., 1996. Neural crest Formation inXenopus laevis:mechanisms ofXslug induction. Dev. Biol. 177, 580–589.

Mayor, R., Essex, L.J., Bennett, M.F., Sargent, M.G., 1993. Distinct ele-ments of theXsnapromoter are required for mesodermal and ectodermalexpression. Development 119, 661–671.

Mayor, R., Guerrero, N., Martıńez, C., 1997. Role of FGF andnoggin inneural crest induction. Dev. Biol. 189, 1–12.

Mayor, R., Morgan, R., Sargent, M.G., 1995. Induction of the prospectiveneural crest inXenopus. Development 121, 767–777.

Mayor, R., Young, R., Vargas, A., 1998. Development of neural crests inXenopus. Curr. Top. Dev. Biol. 41, 85–114.

Mizuseki, K., Kishi, M., Matsui, M., Nakanishi, S., Sasai, Y., 1998.Xeno-pusZic-related-1 and Sox-2, two factors induced by chordin, have dis-tinct activities in the initiation of neural induction. Development 125,579–587.

Moon, R.T., Brown, J.D., Yang-Snyder, J.A., Miller, J.R., 1997. Structu-rally related receptors and antagonists compete for secreted Wnt ligands.Cell 88, 725–728.

Murphy, D.B., Wiese, S., Burfeind, P., Schmundt, D., Mattei, M.G.,Schulz-Schaeffer, W., Thies, U., 1994. Humanbrain factor 1, a newmember of the fork head gene family. Genomics 21, 551–557.

Niewkoop, P.D., Faber, J., 1967. Normal Table ofXenopus laevis(Dau-din). North-Holland, Amsterdam.

Onichtchouk, D., Gawantka, V., Dosch, R., Delius, H., Hirschfeld, K.,Blumenstock, C., Niehrs, C., 1996. TheXvent-2 homeobox gene ispart of the BMP-4 signalling pathway controlling dorsoventral pattern-ing of Xenopusmesoderm. Development 122, 3045–3053.

Onichtchouk, D., Glinka, A., Niehrs, C., 1998. Requirement forXvent-1and Xvent-2 gene function in dorsoventral patterning ofXenopusmesoderm. Development 125, 1447–1456.

Overdier, D.G., Porcella, A., Costa, R.H., 1994. The DNA-binding speci-ficity of the hepatocyte nuclear factor 3/forkhead domain is influencedby amino-acid residues adjacent to the recognition helix. Mol. Cell.Biol. 14, 2755–2766.

Pannese, M., Polo, C., Andreazzoli, M., Vignali, R., Kablar, B., Barsacchi,G., Boncinelli, E., 1995. TheXenopushomologue ofOtx-2 is a maternalhomeobox gene that demarcate and specifies anterior body regions.Development 121, 707–720.

Papalopulu, N., Kintner, C., 1996. A posteriorising factor, retinoic acid,reveals that anteroposterior patterning controls the timing of neuronal

differentiation in Xenopusneuroectoderm. Development 122, 3409–3418.

Piccolo, S., Sasai, Y., Lu, B., de Robertis, E.M., 1996. Dorsoventral pat-terning in Xenopus: inhibition of ventral signals by direct binding ofchordin to BMP-4. Cell 86, 589–598.

Pierrou, S., Hellqvist, M., Samuelsson, L., Enerba¨ck, S., Carlsson, P.,1994. Cloning and characterization of seven human forkhead proteins:binding site specificity and DNA bending. EMBO J. 13, 5002–5012.

Ruiz i Altaba, A., Jessell, T.M., 1992.Pintallavis, a gene expressed in theorganizer and midline cells of frog embryos: involvement in the devel-opment of the neural axis. Development 116, 81–93.

Schmidt, J.E., Suzuki, A., Ueno, N., Kimelman, D., 1995. LocalizedBMP-4 mediates dorsal/ventral patterning in the earlyXenopusembryo. Dev.Biol. 169, 37–50.

Smith, W.C., Harland, R.M., 1991. InjectedXwnt-8 RNA acts early inXenopus embryos to promote formation of a vegetal dorsalizingcenter. Cell 67, 753–765.

Sutton, J., Costa, R., Klug, M., Field, L., Xu, D., Largaespada, D.A.,Fletcher, C.F., Jenkins, N.A., Copeland, N.G., Klemsz, M., Hromas,R., 1996. Genesis, a winged helix transcriptional repressor with expres-sion restricted to embryonic stem cells. J. Biol. Chem. 271, 23126–23133.

Suzuki, A., Ueno, N., Hemmati-Brivanlou, A., 1997.Xenopus msx1med-iates epidermal induction and neural inhibition by BMP4. Development124, 3037–3044.

Taira, M., Jamrich, M., Good, P.J., Dawid, I.B., 1992. The LIM domain-containing homeo box geneXlim-1 is expressed specifically in the orga-nizer region ofXenopusgastrula embryos. Genes Dev. 6, 356–366.

Tao, W., Lai, E., 1992. Telencephalon-restricted expression of BF-1, a newmember of the HNF-3/fork head gene family, in the developing ratbrain. Neuron 8, 957–966.

Wang, S., Krinks, M., Lin, K., Luyten, F.P., Moos, M.J., 1997. Frzb, asecreted protein expressed in the Spemann organizer, binds and inhibitsWnt-8. Cell 88, 757–766.

Weigel, D., Ja¨ckle, H., 1990. The fork head domain: A novel DNA bindingmotif of eukaryotic transcription factors?. Cell 63, 455–456.

Wilson, P.A., Lagna, G., Sizuki, A., Hemmati-Brivanlou, A., 1997. Con-centration-dependent patterning of theXenopusectoderm by BMP4 andits signal transducer Smad1. Development 124, 3177–3184.

Xuan, S., Baptista, C.A., Balas, G., Tao, W., Soares, V.C., Lai, E., 1995.Winged helix transcription factor BF-1 is essential for the developmentof the cerebral hemispheres. Neuron 14, 1141–1152.

Yuasa, J., Hirano, S., Yamagata, M., Noda, M., 1996. Visual projectionmap specified by topographic expression of transcription factors in theretina. Nature 382, 632–635.

Zimmerman, L.B., de jesu´s-Escobar, J.M., Harland, R.M., 1996. The Spe-mann organizer signal noggin binds and inactivates bone morphogeneticprotein 4. Cell 86, 599–606.


Documents

Xenopus brain factor-2 controls mesoderm, forebrain and neural