Portal de Periódicos da CAPES

The zinc finger gene Xblimp1 controls anterior endomesodermal cell fate in Spemann's organizer

1999; Springer Nature; Volume: 18; Issue: 21 Linguagem: Inglês

10.1093/emboj/18.21.6062

1460-2075

Flávio S. J. de Souza,

Genomics and Chromatin Dynamics

Article1 November 1999 The zinc finger gene Xblimp1 controls anterior endomesodermal cell fate in Spemann's organizer Flávio S.J. de Souza Flávio S.J. de Souza Division of Molecular Embryology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Search for more papers by this author Volker Gawantka Volker Gawantka Division of Molecular Embryology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Search for more papers by this author Aitana Perea Gómez Aitana Perea Gómez Institut de Génétique et de Biologie Moleculaire et Cellulaire, CNRS/INSERM/Université Louis Pasteur/Collège de France, BP163, 67404 Illkirch cedex, CU de Strasbourg, France Search for more papers by this author Hajo Delius Hajo Delius Division of Applied Tumour Virology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Search for more papers by this author Siew-Lan Ang Siew-Lan Ang Institut de Génétique et de Biologie Moleculaire et Cellulaire, CNRS/INSERM/Université Louis Pasteur/Collège de France, BP163, 67404 Illkirch cedex, CU de Strasbourg, France Search for more papers by this author Christof Niehrs Christof Niehrs Division of Molecular Embryology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Search for more papers by this author Flávio S.J. de Souza Flávio S.J. de Souza Division of Molecular Embryology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Search for more papers by this author Volker Gawantka Volker Gawantka Division of Molecular Embryology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Search for more papers by this author Aitana Perea Gómez Aitana Perea Gómez Institut de Génétique et de Biologie Moleculaire et Cellulaire, CNRS/INSERM/Université Louis Pasteur/Collège de France, BP163, 67404 Illkirch cedex, CU de Strasbourg, France Search for more papers by this author Hajo Delius Hajo Delius Division of Applied Tumour Virology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Search for more papers by this author Siew-Lan Ang Siew-Lan Ang Institut de Génétique et de Biologie Moleculaire et Cellulaire, CNRS/INSERM/Université Louis Pasteur/Collège de France, BP163, 67404 Illkirch cedex, CU de Strasbourg, France Search for more papers by this author Christof Niehrs Christof Niehrs Division of Molecular Embryology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany Search for more papers by this author Author Information Flávio S.J. de Souza1, Volker Gawantka1, Aitana Perea Gómez2, Hajo Delius3, Siew-Lan Ang2 and Christof Niehrs1 1Division of Molecular Embryology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany 2Institut de Génétique et de Biologie Moleculaire et Cellulaire, CNRS/INSERM/Université Louis Pasteur/Collège de France, BP163, 67404 Illkirch cedex, CU de Strasbourg, France 3Division of Applied Tumour Virology, Deutsches Krebsforschungszentrum, Im Neuenheimer Feld 280, D-69120 Heidelberg, Germany The EMBO Journal (1999)18:6062-6072https://doi.org/10.1093/emboj/18.21.6062 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The anterior endomesoderm of the early Xenopus gastrula is a part of Spemann's organizer and is important for head induction. Here we describe Xblimp1, which encodes a zinc finger transcriptional repressor expressed in the anterior endomesoderm. Xblimp1 represses trunk mesoderm and induces anterior endomesoderm in a cooperative manner with the pan-endodermal gene Mix.1. Furthermore, Xblimp1 can cooperate with the BMP inhibitor chordin to induce ectopic heads, while a dominant-negative Xblimp1 inhibits head formation. The head inducer cerberus is positively regulated by Xblimp1 and is able to rescue microcephalic embryos caused by dominant-negative Xblimp1. Our results indicate that Xblimp1 is required for anterior endomesodermal cell fate and head induction. Introduction The study of early endoderm development has been a relatively neglected topic until very recently, when the availability of early endodermal markers allowed the identification of genes important for endoderm formation in Xenopus embryos. Transforming growth factor-β (TGF-β) signalling by Vg1 and nodal-related proteins is strongly implicated in early endoderm formation (Henry et al., 1996; Joseph and Melton, 1998; Zorn et al., 1999). Transcription factors involved in endoderm specification include members of the HMG-box (Hudson et al., 1997), T-box (Zhang et al., 1998) and paired-homeobox (Ecochard et al., 1998; Henry and Melton, 1998; Lemaire et al., 1998) families. These genes are expressed throughout the whole embryonic endoderm and are implicated in pan-endodermal development. The expression of many organizer genes reveals that the anterior endoderm is already regionalized in the early gastrula. At this stage, the anterior endodermal cells are mixed with mesoderm precursors and are therefore referred to as anterior endomesoderm. These cells are the first to migrate under the blastocoel roof during gastrulation, and give rise to liver, foregut and prechordal endomesoderm (Pasteels, 1949; Nieuwkoop and Florschütz, 1950; Keller, 1991; Bouwmeester et al., 1996). The activity of one gene expressed in the anterior endomesoderm, cerberus, has given strong molecular support to the idea that this region is crucial in the process of head induction (reviewed in Slack and Tannahill, 1992; Gilbert and Saxen, 1993; Bouwmeester and Leyns, 1997; Niehrs, 1999). Cerberus is a secreted factor able to induce ectopic heads including forebrain, eye, cement gland and heart in Xenopus (Bouwmeester et al., 1996). Independent evidence for the importance of endoderm in forebrain induction comes from studies in mouse, where ablation of anterior visceral endodermal cells (Thomas and Beddington, 1996) as well as inactivation of genes such as nodal (Varlet et al., 1997) and otx2 (Rhinn et al., 1998) in visceral endoderm cause rostral truncations of the neural tube (for a review, see Beddington and Robertson, 1999). The anterior portion of the visceral endoderm (AVE) lies beneath the future forebrain at the beginning of gastrulation and expresses markers similar to those found in the anterior endomesoderm of Xenopus, including cerberus-like (Belo et al., 1997; Biben et al., 1998; Shawlot et al., 1998), the homeobox gene Hex (Newman et al., 1997; Thomas et al., 1998) and the murine homologue of dkk1 (Glinka et al., 1998; Pearce et al., 1999), suggesting that the AVE and the anterior endomesoderm of Xenopus are homologous structures. In addition to cerberus, frzb (Leyns et al., 1997; Wang et al., 1997) and dkk1 (Glinka et al., 1998) are expressed in Xenopus anterior endomesoderm. These genes encode secreted proteins that inhibit signalling by Wnt family members, and when co-expressed with BMP inhibitors such as Chordin (Sasai et al., 1994) and Noggin (Smith and Harland, 1992) can induce a head (Glinka et al., 1997). According to the two-inhibitor model (Glinka et al., 1997), the head organizer, consisting of anterior endomesoderm, secretes BMP and Wnt inhibitors, while the trunk organizer, corresponding to the more posterior chordamesoderm, secretes predominantly BMP inhibitors. It was found recently that the head inducer cerberus can bind to and inactivate BMP, Wnt and nodal-related molecules, suggesting that the inhibition of nodal signalling is also important for head induction (Piccolo et al., 1999). While this presents a framework for the signals involved in Spemann organizer function, little is known about transcription factors specifying the anterior endomesoderm. Here, we present Xblimp1, a gene expressed in the anterior endomesoderm of the Xenopus organizer. Xblimp1 encodes a zinc finger transcriptional repressor which is the Xenopus homologue of human PRDI-BF1 (Keller and Maniatis, 1991) and mouse Blimp-1 (Turner et al., 1994). In mammals, PRDI-BF1 represses transcription from the β-interferon promoter after viral infection (Keller and Maniatis, 1991) and Blimp-1 promotes survival and maturation of B-cell precursors in the immune system (Turner et al., 1994; Messika et al., 1998) by a mechanism that involves the direct repression of the c-myc promoter (Lin et al., 1997). Here we show that Xblimp1 is expressed in the anterior endomesoderm and its overexpression represses trunk mesoderm and induces anterior endomesoderm. Studies with a dominant-negative Xblimp1 suggest that the gene is required for cerberus expression and head formation. In addition, in co-injections with the BMP inhibitor chordin, Xblimp1 can induce ectopic heads. The results suggest an important role for Xblimp1 in the anterior endomesoderm, controlling cerberus expression and head formation. Results Cloning of Xblimp1 We are carrying out a large-scale in situ hybridization screen to identify genes differentially expressed during early Xenopus development. During this screen, a 1 kb cDNA, 23E9.1, was found to be expressed in the anterior endomesodermal cells (Gawantka et al., 1998) and was chosen for further characterization. A 3.4 kb cDNA clone was recovered subsequently, which contained a complete open reading frame (ORF) coding for a putative protein of 780 amino acids. The deduced amino acid sequence is most similar to the zinc finger protein known as PRDI-BF1 in humans (Keller and Maniatis, 1991) and Blimp-1 in the mouse (Turner et al., 1994), and we named the gene Xblimp1. As Figure 1 shows, Xblimp1 and mouse Blimp-1 proteins have nearly the same size and share an overall identity of 71%. The identity is distributed along the whole sequence, but is particularly high (93%) in the five zinc finger motifs found near the C-terminus of the proteins. Along with the mammalian sequences, an ORF in Caenorhabditis elegans (Wilson et al., 1994) encodes a putative protein containing five zinc finger motifs which has 35% overall identity and 42% similarity to Xblimp1, indicating that the PRDI-BF1/Blimp-1 factor is of ancient origin. Figure 1.Alignment of the deduced amino acid sequences of Xblimp1 (top) and mouse Blimp-1 (bottom; Turner et al., 1994). Boxes show identical residues. The five zinc finger motifs near the C-termini are underlined in black. Underlined in grey is the domain mediating transcriptional repression by co-repressors of the Groucho family (repressor domain; Ren et al., 1999). The identity between both proteins is 71% overall and 93% in the zinc finger motifs. Download figure Download PowerPoint We conclude that Xblimp1 is a Xenopus homologue of the mammalian zinc finger gene PRDI-BF1/Blimp-1. Xblimp1 is expressed in anterior endomesoderm and prechordal plate By RT–PCR analysis, Xblimp1 transcription starts after mid-blastula transition (stage 8.5), and transcripts can still be found at least until tadpole stages (not shown). Whole-mount in situ hybridization of sagittal cuts of early gastrulae show that Xblimp1 transcripts are present in anterior endomesodermal cells, the region primarily fated to become liver and prechordal plate (Pasteels, 1949; Nieuwkoop and Florschütz, 1950; Keller, 1991), in a pattern similar to that of cerberus (stage 10.5, Figure 2A and B) (Bouwmeester et al., 1996), but in a narrower arc. In addition to deep cells, Xblimp1 is also weakly expressed in the ectoderm. By the end of gastrulation (stage 13, Figure 2D), Xblimp1 is no longer expressed in the anterior-most endodermal cells, but maintained in the prechordal plate, in both the mesodermal and endodermal layers. Transcripts are also seen in both neural plate and future epidermis in three longitudinal stripes of expression (Figure 2C, asterisks). Expression is also found in 2–4 stripes near the dorsal midline of the embryo (Figure 2C), and transversal cuts indicate that the expression is in paraxial mesoderm (not shown). By tadpole stages (stage 30, Figure 2E), Xblimp1 is expressed in a variety of tissues: the presumptive pituitary gland, the eyes, the dorsal part of the otic vesicles, the pharyngeal arches (especially the third and fourth), the fin and two stripes in the forming somites near the tip of the tail, which correspond to future somites and can be regarded as a continuation of the expression in the paraxial mesoderm seen at stage 13. Figure 2.Expression of Blimp1 in Xenopus and mouse. Whole-mount in situ hybridizations are shown in all panels. (A and B) Sagittal cuts of Xenopus gastrulae (stage 10.5) showing the expression of Xblimp1 and cerberus. The dorsal lip is indicated by an arrowhead. Note that both genes are expressed similarly in deep anterior endomesodermal cells (ant endomes). (C) Dorsal view of a Xenopus neurula (stage 13) showing Xblimp1 expression in stripes in the neural plate and epidermis (asterisks) and blocks of paraxial mesoderm (p mes). Anterior is to the left. (D) Expression of Xblimp1 in the prechordal plate seen in a sagittally cut neurula. Anterior is to the left. (E) Expression of Xblimp1 in a whole-mount tailbud embryo (stage 30). Transcripts are detected in the future pituitary gland (pit), eye, otic vesicle, pharyngeal arches, fin and forming somites. (F) Lateral view of a 6.5 d.p.c. mouse embryo; anterior is to the left. Blimp-1 is weakly expressed in the anterior (arrow) and posterior visceral endoderm. Expression is also seen in presumptive definitive endoderm (arrowhead). (G) Lateral view of a 7.0 d.p.c. mouse embryo; anterior is to the left. Blimp-1 expression becomes more restricted to the anterior side of the embryo (arrow). The black line shows the plane of the section shown in (H). (H) Transversal cut of the embryo shown in (G); anterior is to the bottom. Blimp-1 is expressed in the prechordal plate and in more lateral anterior definitive endoderm. (I) Head-fold-stage embryo in anterior view. Blimp-1 transcripts are found in the foregut and midgut. The black line indicates the plane of the section shown in (J). (J) Transversal cut of the embryo in (I), showing Blimp-1 expression in the foregut (arrow) and in the endoderm overlying the posterior-most part of the primitive streak (arrowhead). Download figure Download PowerPoint Since the expression of the mammalian homologues of Xblimp1 has not been studied during development, we performed whole-mount in situ hybridizations on mouse embryos. Transcripts can first be detected soon after gastrulation starts. As shown in Figure 2F, mouse Blimp-1 is weakly expressed at 6.5 days post-coitum (d.p.c.) in the anterior and posterior visceral endoderm and in presumptive definitive endoderm cells. At 7.0 d.p.c. (Figure 2G and H), transcripts are found in the anterior region of the embryo in the prechordal plate as well as in the anterior definitive endoderm. At head fold stage, Blimp-1 expression is seen in the foregut and adjacent mesoderm (Figure 2I and J). At the 2–3 somite stage, expression is detectable in the anterior neural ridge. Later in development (8.5 d.p.c. onwards), expression is found in the otic placode, ventral telencephalon, branchial arches and somites (not shown). A full description of early Blimp-1 expression will be published elsewhere (A.P.Gómez and S.-L.Ang, in preparation). In conclusion, Xblimp1 and Blimp-1 are expressed in equivalent regions in the early gastrula, namely anterior endomesoderm and prechordal plate in Xenopus, and AVE and prechordal plate in the mouse. Xblimp1 inhibits mesoderm development The expression of Xblimp1 during early gastrulation in the anterior endomesoderm suggested that the gene could play a role in the dorso-anterior region of the embryo, particularly in the endodermal layer. To gain insight into the role of Xblimp1 during gastrula stages, we carried out a series of overexpression experiments by microinjecting synthetic Xblimp1 RNA into early Xenopus embryos. Increasing doses of Xblimp1 RNA cause progressive reduction of the antero-posterior axis of the embryos, with retardation of gastrulation and spina bifida defects appearing with higher doses of injected RNA (Figure 3B). These effects suggest that the formation of the main axial structures of the embryo, such as somites and notochord, is impaired by Xblimp1 overexpression. Figure 3.Xblimp1 inhibits mesoderm. (A) Scheme of the marginal zone experiment. Embryos are injected with RNA into all blastomeres at the 4-cell stage, and dorsal (D) and ventral (V) marginal zone explants are cut at the gastrula stage. Explants are cultured until stage 11, total RNA is extracted and analysed by RT–PCR. (B) The indicated amounts of Xblimp1 RNA were injected into all blastomeres of 4-cell stage embryos. Note that the axis is reduced with increasing amounts of RNA injected. (C) Embryos were either uninjected (control) or injected with Xblimp1 RNA (100 pg/blastomere), and dorsal (D) and ventral (V) marginal zone explants were analysed at stage 11 by RT–PCR for the expression of the indicated genes. Note that the trunk mesoderm markers Xbra, chordin and myf5 are inhibited by Xblimp1, while the anterior markers cerberus (Cer), goosecoid (Gsc) and otx2 are induced. The lower panel shows the same experiment performed with mouse Blimp-1 (100 pg/blastomere) at stage 13. H4, histone 4; RT-, uninjected whole embryo control sample without reverse transcription. Download figure Download PowerPoint To find out which changes in gene expression can be caused by Xblimp1 during gastrulation, dorsal and ventral marginal zone explants (DMZ and VMZ) injected with Xblimp1 RNA were analysed by RT–PCR (Figure 3C). The trunk mesoderm markers Xbra (pan-mesodermal, Smith et al., 1991), chordin (prospective notochord, Sasai et al., 1994) and myf5 (muscle, Hopwood et al., 1991; Dosch et al., 1997) are repressed by Xblimp1. The dorsal homeobox gene Xlim1 (Taira et al., 1994) is also repressed by Xblimp1. In the VMZ, the dorso-anterior markers goosecoid (Cho et al., 1991), otx2 (Blitz and Cho, 1995; Pannese et al., 1995) and cerberus (Bouwmeester et al., 1996), which are co-expressed with Xblimp1 in the anterior endomesoderm, are weakly induced. Like the Xenopus gene, mouse Blimp-1 injection represses axial structures (not shown), induces cerberus and represses Xbra in marginal zone explants (Figure 3C, lower panel), suggesting that both genes have similar properties. Since Xblimp1 represses mesodermal markers, it is likely that the reduction of the axis seen in injected embryos is caused by a deficiency in the development of axial mesodermal structures such as notochord and somitic mesoderm. We conclude that Blimp genes are able to repress mesoderm. Xblimp1 induces endoderm and cooperates with Mix.1 in inducing cerberus Since mesoderm inhibitory effects similar to that of Xblimp1 are typically observed for endoderm-promoting genes, such as Mix.1 and Milk (Rosa, 1989; Ecochard et al., 1998; Lemaire et al., 1998), we tested whether Xblimp1 could induce endoderm in animal cap explants. RT–PCR analyses (Figure 4B) show that Xblimp1 can induce pan-endodermal markers such as endodermin (Sasai et al., 1996), Xsox17α (Hudson et al., 1997) and the new pan-endodermal marker 26D10.1 (Gawantka et al., 1998) in ectoderm. Xblimp1 could also cause some neuralization, as observed by neural cell adhesion molecule (NCAM) induction (Kintner and Melton, 1987), which occurred in the absence of mesoderm, since no Xbra or muscle actin (Mohun et al., 1984) was induced (Figure 4B). In some experiments, the blood marker α-globin (Banville and Williams, 1985) was induced (not shown). Finally, Xblimp1 did not induce cerberus in animal caps, unlike in VMZs (see Figure 3C). Figure 4.Xblimp1 induces anterior endomesoderm. (A) Diagram of animal cap experiments. Embryos were microinjected with RNA into four animal blastomeres at the 8-cell stage, animal caps were cut at late blastula stages and explants were cultivated up to stages 26 and 10.5 (B) or 11 (C) and total RNA was extracted and analysed by RT–PCR. (B) Animal caps were either uninjected (Co) or injected with increasing concentrations of Xblimp1 RNA (50, 100 and 200 pg/blastomere). At stage 26, explants injected with Xblimp1 express the endodermal markers Xsox17α, endodermin (Edd) and 26D10.1, as well as the neural marker NCAM. At stage 10.5, neither Xbra nor cerberus is induced by Xblimp1. (C) Xblimp1 (Xbl, 100 pg/blastomere) and Mix.1 (M, 12.5 pg/blastomere) RNAs were either injected alone or in combination (Xbl/M). Note the induction of cerberus expression (Cer) by the combination. M-Actin, muscle actin; H4, histone 4; RT+, RT−, uninjected whole embryo control sample with or without reverse transcription, respectively. Download figure Download PowerPoint The fact that Xblimp1 induces cerberus in VMZs, but not in animal caps (Figure 4B), suggested that a competence factor is missing in ectoderm. Recently, it was found that Mix.1 serves as a competence factor to siamois in the induction of cerberus (Lemaire et al., 1995, 1998). Figure 4C shows that Mix.1 also serves as a competence factor to Xblimp1, with both genes cooperating to induce high levels of cerberus, while neither Xblimp1 nor Mix.1 can induce considerable cerberus expression when injected alone. Among other genes expressed in the anterior endomesoderm, goosecoid and XHex (Newman et al., 1997) are also induced by the combination, while the head inducers dkk1 and frzb (Leyns et al., 1997; Wang et al., 1997; Glinka et al., 1998) are not induced in a cooperative manner. The competence effect was specific to Mix.1, because the co-expression of Xblimp1 and another endoderm-promoting transcription factor, Xsox17α (Hudson et al., 1997), was not able to induce anterior endomesodermal markers cooperatively (not shown). In conclusion, Xblimp1 is an endodermal inducer in ectodermal explants, and it can cooperate specifically with the pan-endodermal gene Mix.1 in inducing anterior endomesoderm. Xblimp1 is required for cerberus expression and head formation The mammalian zinc finger PRDI-BF1/Blimp-1 works as a transcriptional repressor of the promoters of the β-interferon and c-myc genes (Keller and Maniatis, 1991; Lin et al., 1997), and it was shown recently that this repression is mediated by members of the Groucho family of transcriptional co-repressors, which can bind to a repressor domain in the middle of human PRDI-BF1 protein (Figure 1, Ren et al., 1999). In view of that, we constructed a dominant-negative version of Xblimp1 (XbVP, Figure 6A) by fusing its zinc finger domain with the transcriptional activator domain of the viral protein VP16 (Friedman et al., 1988), an approach employed successfully for other transcriptional repressors (Ferreiro et al., 1998; Onichtchouk et al., 1998; Latinkic and Smith, 1999). As a control, we fused the zinc finger domain of Xblimp1 with the transcriptional repressor domain of Drosophila Even-skipped (Han and Manley, 1993). This construct (Xbeve, Figure 5A) should repress transcription of its target genes, like wild-type Xblimp1. Figure 5.Xblimp1 acts as a transcriptional repressor during early development. (A) Scheme of wild-type Xblimp1 protein and the Xbeve construct, in which the N-terminal domain of Xblimp1 is replaced by the repressor domain of Even-skipped (eve). Zinc fingers are shown in dark grey. (B) Embryo injected into all blastomeres at the 4-cell stage with Xbeve RNA (250 pg/blastomere) has axial defects. The phenotype is similar to that of Xblimp1 RNA injection (Figure 3B). (C) RT–PCR of dorsal (D) and ventral (V) marginal zones at stage 10.5 injected with Xblimp1 (125 pg/blastomere) and Xbeve (250 pg/blastomere) RNAs. Note that both Xblimp1 and Xbeve can induce cerberus and repress Xbra and chordin. (D) RT–PCR of animal cap explants at stage 11 injected with Mix.1 (12.5 pg/blastomere), Xbeve (beve, 60 pg/blastomere) or XbVP (bVP, 30 pg/blastomere) RNAs alone or in combination. Note the cerberus induction by the co-injection of Mix.1 and Xbeve (M/beve), but not by Mix.1 and XbVP (M/bVP). Co, uninjected control; H4, histone 4; RT+, RT−, uninjected whole embryo control sample with or without reverse transcription, respectively. Download figure Download PowerPoint Figure 6.Requirement for Xblimp1 in head formation. (A) Scheme of wild-type Xblimp1 protein and the dominant-negative construct XbVP, in which the N-terminus of Xblimp1 was substituted by the activator domain of VP16 (VP). The five zinc fingers are in dark grey. (B) Embryos were either uninjected (control) or injected into two dorsal blastomeres at the 4-cell stage with 125 pg of XbVP, or co-injected with 125 pg of XbVP and either 50 pg of Xblimp1 RNA or 50 pg of cerberus DNA. Note that both wild-type Xblimp1 and cerberus can rescue the microcephalic phenotype caused by XbVP. (C) RT–PCR analysis of stage 10.5 dorsal (D) or ventral (V) marginal zones explanted from embryos that were either uninjected (control) or injected with the indicated RNAs into two dorsal blastomeres at the 4-cell stage. XbVP injection (125 pg/embryo) inhibits cerberus expression, and this effect can be rescued by co-injection with Xblimp1 (100–200 pg/embryo). H4, histone 4; RT−, uninjected whole embryo control sample without reverse transcription. Download figure Download PowerPoint The injection of RNA for the repressor construct Xbeve causes a phenotype similar to that of wild-type Xblimp1, with reduction of the dorso-anterior axis and gastrulation defects (Figure 5B). Xbeve also induced cerberus and repressed Xbra and chordin in marginal zones, as did wild-type Xblimp1 (Figure 5C). Importantly, Xbeve was able to cooperate with Mix.1 in inducing cerberus-like wild-type Xblimp1, while XbVP was not (Figure 5D). We conclude that, like its mammalian homologues, Xblimp1 acts as a transcriptional repressor during early Xenopus development. RNA injection of the activator construct XbVP into the dorsal side of embryos led to a microcephalic phenotype, with embryos lacking cement gland and eyes or having cyclopic eyes, indicating disturbances in the induction and/or patterning of the head (Figure 6B). The phenotype is specific, since it can be rescued by the co-injection of XbVP with wild-type Xblimp1 RNA (Figure 6B; Table I, I-1), but not by the co-injection with RNA for another zinc finger gene, GATA2 (Table I, I-3; Zon et al., 1991). Table 1. XbVP causes a microcephalic phenotype that can be rescued by Xblimp1 RNA and cerberus DNA RNA or DNA injected (pg/embryo) Two eyes (%) Cyclopia (%) No eyes (%) No. of embryos I-1 XbVP (100–130) 5.2 34.8 60 90 XbVP + Xblimp1 (25) 33.1 60.6 6.4 37 XbVP + Xblimp1 (50) 65.7 17.1 17.4 75 XbVP + Xblimp1 (100) 59.1 10.9 30 27 I-2 XbVP (100–200) 18 59.1 22.5 48 XbVP + Xblimp (50–100) 76.2 18.8 5 60 I-3 XbVP (100) 13.1 47.2 29 39 XbVP + GATA2 (62) 13.6 56.6 29.8 37 XbVP + GATA2 (125) 23.4 55.6 21.1 38 XbVP + GATA2 (250) 13.5 32.5 54.1 37 I-4 XbVP (120) 17.2 33.5 49.4 58 XbVP + DNA Cer (25) 31.9 33.7 34.5 42 XbVP + DNA Cer (50) 57.9 12.8 29.2 65 XbVP + DNA Cer (100) 38.4 45 29.1 43 Four-cell (I-1, I-3 and I-4) or 8-cell (I-2) stage embryos were microinjected into two dorsal or dorso-vegetal blastomeres with XbVP RNA alone or in combination with Xblimp1 (I-1 and I-2) or GATA2 (I-3) RNAs or with plasmid DNA for cerberus (I-4). Only embryos without gastrulation defects (spina bifida) were scored. The microcephaly caused by XbVP may be due to a direct interference with the prospective neuroectoderm, since Xblimp1 is also weakly expressed in the ectoderm, or it could be due to indirect interference with head-inducing anterior endomesoderm. To distinguish between these possibilities, XbVP RNA was injected into dorso-vegetal blastomeres of 8-cell-stage embryos, so that the prospective neuroectoderm was not targeted by the injection. Embryos injected in this manner also have head defects and can again be rescued by co-injections with wild-type Xblimp1 RNA (Table I, I-2), suggesting that the phenotype induced by XbVP is caused by an interference with Xblimp1 function in the anterior endomesoderm. The anterior endomesoderm secretes the head inducers Cerberus, Frzb and Dkk1, and the microcephaly caused by XbVP could indicate that it inhibits the expression of one or more of these factors. To test this, RT–PCR analyses were carried out with gastrula dorsal explants injected with either XbVP alone or combinations of XbVP and Xblimp1 RNAs. XbVP injection reduces the expression of cerberus, but not that of the head inducers frzb and dkk1 (Figure 6C). Importantly, the co-injection of wild-type Xblimp1 RNA rescues cerberus expression repressed by XbVP (Figure 6C). These results suggest a genetic interaction between Xblimp1 and cerberus in the anterior endomesoderm. Since cerberus was the only gene among the known head inducers whose expression was affected by XbVP, we tried to rescue the microcephaly by co-injecting XbVP RNA with cerberus. Cerberus protein is a nodal inhibitor, and RNA injections interfere precociously with nodal signalling, inhibiting endomesoderm and axis formation (Bouwmeester et al., 1996; Piccolo et al., 1999). To circumvent this, we overexpressed cerberus from plasmid DNA after mid-blastula transition. The co-injection of XbVP RNA and cerberus DNA could rescue head formation, increasing the number of embryos with normal heads up to 60% (Figure 6B; Table I, I-4). The results indicate that Xblimp1 works as a transcriptional repressor during early Xenopus development, a