- Email: [email protected]
A novel function for the Xslug gene: control of dorsal mesendoderm development by repressing BMP-4
Mechanisms of Development 97 (2000) 47±56
www.elsevier.com/locate/modo
A novel function for the Xslug gene: control of dorsal mesendoderm development by repressing BMP-4 R. Mayor*, N. Guerrero, R.M. Young, J.L. Gomez-Skarmeta, C. Cuellar Millennium Nucleus in Developmental Biology, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile Received 28 February 2000; received in revised form 19 June 2000; accepted 7 July 2000
Abstract The Snail family of genes comprise a group of transcription factors with characteristic zinc ®nger motifs. One of the members of this family is the Slug gene. Slug has been implicated in the development of neural crest in chick and Xenopus by antisense loss of function experiments. Here, we have generated functional derivatives of Xslug by constructing cDNAs that encode the Xslug protein fused with the transactivation domain of the virus-derived VP16 activator or with the repressor domain of the Drosophila Engrailed protein. Our results suggest that Xslug normally functions as a transcriptional repressor and that Xslug-VP16 behaves as a dominant negative of Xslug. In the present work, we con®rm and extend previous results that suggest that Xslug has an important function in neural crest development, by controlling its own transcription. In addition we have uncovered a new function for Xslug. We show that Xslug is expressed in the dorsal mesendoderm at the beginning of gastrulation, where is it able to upregulate the expression of dorsal genes. On the other hand when Xslug is expressed outside of the organizer it represses the expression of ventral genes. Our results indicate that this effect on mesodermal patterning depends on BMP activity, showing that Xslug can directly control the transcription of BMP-4. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Xslug; Xsnail; Organizer; Neural crest; BMP-4 pathway
1. Introduction Families of transcription factors present in restricted territories within embryos generate positional identity and control morphogenesis of different tissues. The Snail genes comprise a family of transcription factors with characteristic zinc ®nger motifs. In Drosophila, Snail is involved in the process of mesoderm formation (see Alberga et al., 1991; Leptin, 1991). This gene encodes a Zn-®nger transcription factor that represses the expression of neuroectodermal genes in the mesoderm (Leptin, 1991) and is required for the invagination of mesodermal cells during gastrulation (Ip et al., 1994). In Drosophila a second Snail-like gene, named Escargot (Whiteley et al., 1992), has been implicated in the maintenance of diploidy in imaginal cells (see Fuse et al., 1996). In tracheal cells, Escargot is required for normal expression of the cell adhesion molecule DE-cadherin (Tanaka et al., 1996) and the control of cell fusion during tracheal development (see Samakovlis et al., 1996). A third more distantly related Snail-like gene, Scratch, appears to be involved in the regu* Corresponding author. Tel.: 156-2-678-7351; fax: 156-2-271-2983. E-mail address: [email protected] (R. Mayor).
lation of neuronal development (Roark et al., 1995). Both Snail and Escargot proteins have similar DNA binding speci®city and have been proposed to act as repressors through interactions with the co-repressor C-terminal binding protein (CtBP) (Nibu et al., 1998). In vertebrates two members of the Snail family have been discovered in chick and Xenopus, named Snail and Slug. In Xenopus the expression of Xsnail begins at stage 9, prior to gastrulation, in the dorsal marginal zone, progresses laterally to the ventral side by stage 10 (Sargent and Bennett, 1990; Essex et al., 1993), and is subsequently found in the lateral plate mesoderm up until the tailbud stage. Xsnail expression can also be found in ectoderm and premigratory neural crest and their derivatives such as the branchial cartilages (Mayor et al., 1993). In contrast it has been reported that Xslug is expressed only in prospective neural crest and is not found in developing mesoderm before stage 17 (Mayor et al., 1995). The pattern of Snail and Slug expression in the chick is analogous to that seen in Xenopus, with the exception that Xsnail is found initially in prechordal mesoderm, but in the chick Snail is expressed preferentially in the right side lateral mesoderm, where it appears to play an important role in the development of asymmetry (Isaac et al., 1997). In the mouse, there is striking `swapping' in the
0925-4773/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(00)00412-3
48
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
pattern of Slug and Snail in many but not all expression domains (Sefton et al., 1998). In the zebra®sh, two Snaillike genes (Snail-1 and Snail-2) have been described (Hammerschmidt and Nusslein, 1993; Thisse et al., 1993, 1995); neither contains the 29 amino acid sequence motif characteristic of other vertebrate Slugs (see Sefton et al., 1998). In zebra®sh, Snail-1 is initially distributed widely up to the blastula stage. As in the case of Drosophila Snail, Snail-1 is associated with mesoderm formation. During gastrulation Snail-1 becomes restricted to the involuting cells of the germ ring and is absent from cells of the dorsal midline. Later in development, Snail-1 is found in the paraxial mesoderm, neural crest cells, and mesodermderived portions of the head (Thisse et al., 1993; Hammerschmidt and Nusslein, 1993). The Zebra®sh Snail-2 gene is expressed during gastrulation in cephalic mesendodermal cells, and is later restricted to the neural crest (Thisse et al., 1995). The function of Slug has been studied directly in the chicken and in the mouse. In the chick, embryos have been treated with antisense oligonucleotides directed against the Slug mRNA (Nieto et al., 1994). This treatment led to the inhibition of Slug expression and prevented neural crest migration. In contrast, mice homozygous for a null mutation in Slug are viable and display no obvious defects in neural crest formation, migration or development (Jiang et al., 1998), even though Slug is expressed in migratory neural crest cells (Jiang et al., 1998; Sefton et al., 1998). It has been suggested that the mouse may use the closely related gene Snail in place of Slug (Sefton et al., 1998), but this hypothesis has not been tested. The role of the Xslug gene in neural crest development in Xenopus has been recently analyzed using the injection of Xslug antisense RNA (Carl et al., 1999). The injection of Xslug antisense RNA led to reduced levels of both Xslug and Xsnail expression in the embryos, inhibition of neural crest migration and reduction of many neural crest derivatives. These results show that migration and differentiation of neural crest derivatives are dependent upon Xslug/Xsnail activity, but because Xslug antisense also reduced the level of Xsnail expression it is not possible to independently analyze the role of these two Zn-®nger genes by the antisense technique. In this study, we generated functional derivatives of Xslug by constructing cDNAs that encode the Xslug protein fused with the transactivation domain of the virus-derived VP16 protein (Friedman et al., 1988) or with the repressor domain of the Drosophila engrailed protein (Jaynes and O'Farrell, 1991). Our functional analysis using Xenopus embryos shows that Xslug-en produces the same effect as wild-type Xslug mRNA and the opposite effect to that of Xslug-VP16, suggesting that Xslug is likely to be a transcriptional repressor. In these experiments we con®rm the role of Xslug in neural crest development but, unexpectedly, we have uncovered a novel role for Xslug in mesodermal patterning. We show for the ®rst time that Xslug is expressed in the dorsal marginal zone at the early gastrula
stage and our results suggest that Xslug controls the development of dorsal mesoderm and represses the expression of ventral mesodermal genes. We propose a hierarchy of genes that control dorsal mesoderm development where Xslug functions as a repressor of BMP-4. 2. Results 2.1. Xslug is expressed in the Spemann organizer and the protein is localized in the cell nucleus It had been reported previously that the mesodermal expression of Xslug was ®rst detectable after stage 17 in the lateral plate (Mayor et al., 1995). In order to more carefully analyze the earliest mesodermal expression we performed more sensitive in situ hybridization experiments (see Mancilla and Mayor, 1996). We observed clear expression in the dorsal marginal zone from stage 10 onwards, above the dorsal blastopore lip (Fig. 1A). When these embryos were sectioned and analyzed for the in situ hybridization of chordin, cerberus and Xslug, the Xslug expression was not only in the prospective dorsal mesoderm but also in the deeper endodermal cells (Fig. 1B) which correspond to the cells that express Cerberus (Bouwmeester et al., 1996). At the neurula stage the expression is visible in the notochord, but it is weaker than the expression in the neural folds (Fig. 1C,D). The Slug gene encodes a putative transcription factor but to date there have been no reports that describe it to be localized in the cell nucleus. We constructed a cDNA that
Fig. 1. Xslug expression and cellular localization of Xslug protein. (A±D) In situ hybridization using a Xslug probe. (A) Vegetal view of a stage 10 embryo. d, dorsal; v, ventral; arrow, dorsal blastopore lip. Note the expression in the marginal zone above the blastopore lip. (B) Stage 10 embryo sectioned sagittally. An, animal pole; Vg, vegetal pole; D, dorsal; V, ventral; arrow, dorsal blastopore lip. Notice that Xslug is expressed in the dorsal marginal zone, including the super®cial and deep cells. (C) Dorsal view of a stage 15 embryo showing strong Xslug expression in the neural folds (arrowheads) and weaker expression in the notochord (n). (D) Paraf®n section of embryo shown in (C). Arrowheads, neural folds; n, notochord; s, somite. (E) Anti-myc immunostaining of a stage 11 embryo injected with Xslug-myc at the one cell stage. Note the nuclear localization of the Xslugmyc protein (arrows).
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
49
encodes the Xslug protein fused to a myc tag epitope, which can be recognized by immunostaining. The mRNA of this fused protein was injected at the one cell stage and at different stages the embryos were ®xed and the localization of the fused protein was analyzed. Nuclear localization of the construct was detected during all stages analyzed (Fig. 1E), showing for the ®rst time that Slug is a nuclear protein. 2.2. Xslug controls the expression of dorsal mesodermal genes We attempted to study Xslug function by injecting mRNAs that encode (a) the full length Xslug cDNA, (b) a fusion protein of Xslug and the repressor domain of the Drosophila engrailed (Xslug-en) (Jaynes and O'Farrell, 1991), and (c) the transactivation domain derived from VP16 (Xslug-VP16) (Friedman et al., 1988) (Fig. 2). The injection of Xslug or Xslug-en mRNAs at the one or two cell stage causes a partial and moderate secondary axis which is recognized as an expansion of the injected side (see Fig. 9C). However, injection of Xslug-VP16 at the two cell stage leads to a complete inhibition of head development, while the posterior regions of the embryo develop normally (Fig. 3). To explain this phenotype and taking into account that Xslug is expressed in the dorsal marginal zone, we decided to analyze if dorsal mesoderm genes, required for the normal development of the head, could be controlled by Xslug. The expression of chordin and cerberus (Fig. 4A,F) was analyzed in embryos injected with Xslug or the chimeric proteins. Injection of Xslug and Xslug-en mRNA produced a similar ectopic expression of chordin and cerberus in the ventral marginal zone (Fig. 4B,C,G,H), but the injection of Xslug-VP16 mRNA produced an inhibition of the dorsal markers at the site of injection (Fig. 4D,E,I,J). The similarity in the effect of Xslug and Xslugen and the opposite effect of Xslug-VP16 suggests that Xslug is likely to be a transcriptional suppressor of ventral meso-
Fig. 2. Schematic representation of wild-type and mutated Xslug molecules. The amino acids numbers are shown above each drawing of the predicted protein and the Zn-®ngers are indicated in blue. The top drawing represents wild-type Xslug protein. Repressing and activating forms of Xslug were constructed by replacing the N-terminal of Xslug with the repressor domain of the engrailed protein (red box) or the activation domain of VP16 (green box).
Fig. 3. Embryos were injected with 1 ng of Xslug-VP16 RNA at the two cell stage and analyzed at stage 35. Left column, uninjected embryos; right column, embryos injected with Xslug-VP16. Notice the absence of anterior development in the injected embryos (n 65, 42% of effect).
derm genes. In addition, the inhibition of these genes, particularly the inhibition of cerberus, could explain the absence of a head in embryos injected with Xslug-VP16. To determine the speci®city of the effects produced by these fused proteins on chordin and cerberus expression (Fig. 5A,F), a rescue experiment was performed. Embryos were injected with Xslug or Xslug-VP16 mRNA at the one cell stage, producing expansion (Fig. 5B,G) or inhibition (Fig. 5C,H) of the dorsal markers, respectively. When embryos were co-injected with Xslug-VP16 and Xslug mRNA a complete rescue on the expression of chordin and cerberus was observed (Fig. 5D,I). However, no rescue in the expression of these markers was observed when the embryos were co-injected with Xslug-VP16 and Xsnail mRNA (Fig. 5E,J), showing the speci®city of the XslugVP16 effect, which depends on Xslug and not on Xsnail activity. 2.3. Xslug is able to repress ventral mesodermal genes We have shown that Xslug is expressed in the prospective dorsal mesendoderm and is able to control the expression of genes expressed in that region. Our next question was whether Xslug was able to repress the expression of ventral mesodermal genes. Similar injections using the mRNAs encoding Xslug, Xslug-en and Xslug-VP16 were performed, and the expression of the ventral genes Xwnt-8 (Fig. 6A) and Xvent-1 (Fig. 6E) was analyzed. Injection of Xslug or Xslugen exhibited a similar effect: the expression of Xwnt-8 and Xvent-1 was decreased or abolished at the site of injection (Fig. 6B,C,F,G). This result suggests that the normal expression of Xslug at the dorsal marginal zone could have a role in repressing ventral mesodermal genes in that region. To analyze if Xslug was suf®cient to produce the dorsal repression of ventral genes we injected the Xslug-VP16 mRNA in the prospective dorsal mesoderm region of a four cell stage embryo and the expression of the ventral markers was
50
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
Fig. 4. Xslug controls the expression of dorsal genes. Embryos co-injected with the different constructs and b-gal mRNA (blue staining) were analyzed at stage 10 by in situ hybridization (purple staining) for the expression of chordin (A±E) and Cerberus (F±J). Vegetal view, dorsal is up. (A,F) Control embryos. (B,G) Embryos injected with 1 ng of Xslug mRNA in the ventral area of a two cell stage embryo. Arrow, site of injection. Notice the ectopic expression of the marker. (C,H) Embryo injected with 1 ng of Xslug-en RNA in the ventral area of a two cell stage embryo. Arrow, site of injection. Notice the ectopic expression of the marker. (D,I) Embryo injected with 1 ng of Xslug-VP16 mRNA intake in the dorsal side of one blastomere of a two cell stage embryo. Arrowhead, site of injection. Notice the inhibition in the expression of the marker. (E,J) Higher magni®cations of (D,I), respectively. Between 60 and 100 embryos were analyzed in each experiment. Between 42 and 82% of ectopic expression or inhibition was observed.
analyzed. No ectopic expression of Xwnt-8 or Xvent-1 was observed in the injected dorsal marginal zone (Fig. 6D,H), showing that although Xslug is able to repress the expression of ventral genes when expressed ventrally, its dorsal expression is not suf®cient for this suppressor activity. The Xsnail gene was also analyzed in a similar manner. This gene is expressed as a ring in the marginal zone at the beginning of gastrulation (Fig. 6I). The injection of Xslug or Xslug-en mRNA produced a strong inhibition of Xsnail (Fig. 6J,K), and the injection of Xslug-VP16 induced a very strong ectopic induction of Xsnail (Fig. 6L) that occasionally reached the animal pole (Fig. 6L, inset). This strong dependence of Xsnail on Xslug activity suggests that Xslug is a direct repressor of Xsnail in the dorsal mesoderm. 2.4. Xslug represses BMP-4 expression At the gastrula stage, the organizer secretes a variety of zygotic proteins that act as antagonists to various members
of the BMP and Wnt families of ligands, which are secreted by ventral and lateral mesodermal cells. BMPs and Wnts favor ventral development and the organizer antagonizes their activities. As Xslug is expressed in the organizer region and has the capacity to induce the expression of dorsal mesodermal markers we asked whether this activity depended on the inhibition of BMP or Wnt activity. The expression of chordin was analyzed under different conditions (Fig. 7A). Since Xslug mRNA injection produces an ectopic expansion of chordin (Fig. 7B) we analyzed whether this ectopic expression of a dorsal gene could be due to an inhibition of BMP-4 or Xwnt-8, two molecules that promote ventral mesoderm. We co-injected Xslug and BMP-4 mRNAs or Xslug and pCSKAXwnt-8, a cDNA that expresses Xwnt-8 after MBT transition (Christian and Moon, 1993), and we analyzed in which of theses conditions the effect of injecting Xslug could be inhibited. The injection of BMP-4 or pCSKAXwnt-8 leads to an inhibition in the expression of chordin (Fig. 7E,F).
Fig. 5. Speci®city of the effect of Xslug-VP16 on chordin expression. Embryos co-injected with the different constructs and b-gal mRNA (blue staining) were analyzed at stage 10 by in situ hybridization (purple staining) for the expression of chordin (A±E) and Cerberus (F±J). Vegetal views, dorsal is up. (A,F) Control embryos. (B,G) Embryos injected with 1 ng of Xslug mRNA in the ventral area of a two cell stage embryo. Arrow, site of injection. Note the ectopic expression of the marker (n 62, 45% of ectopic expression). (C,H) Embryo injected with 1 ng of Xslug-VP16 mRNA in the dorsal side of a one cell stage embryo. Arrowhead, site of injection. Note the inhibition in the expression of the marker (n 58, 85% of inhibition). (D,I) Embryo co-injected with 1 ng of Xslug-VP16 RNA and 1 ng of Xslug mRNA in the dorsal side of a one cell stage embryo (n 72, 10% of expression, 0% of inhibition). Note the rescue in the expression of the markers. (E,J) Embryo co-injected with 1 ng of Xslug-VP16 RNA and 1 ng of Xsnail mRNA in the dorsal side of a one cell stage embryo. Note that Xsnail is not able to rescue the inhibition of the dorsal markers by Xslug-VP16 (n 32, 89% of inhibition).
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
51
Fig. 6. Xslug inhibits the expression of ventral genes. Embryos co-injected with the different constructs and b-gal mRNA (blue staining) were analyzed at stage 10 by in situ hybridization (purple staining) for the expression of Xwnt-8 (A±D), Xvent-1 (E±H) and Xsnail (I±L). Vegetal views, dorsal is up. (A,E,I) Control embryos. (B,F,J) Embryos injected with 1 ng of Xslug mRNA in the ventral area of a one cell stage embryo. Arrowhead, site of injection. Note the inhibition in the expression of the gene. (C,G,K) Embryos injected with 1 ng of Xslug-en RNA in the ventral area of a one cell stage embryo. Arrowhead, site of injection. Note the inhibition in the expression of the gene. (D,H) Embryos injected with 1 ng of Xslug-VP16 RNA in the dorsal region of a one cell stage embryo. Arrowhead, site of injection. Note that no dorsal expression of the marker can be detected. (L) Embryo injected with 1 ng of Xslug-VP16 in one blastomere of a two cell stage embryo. Arrow, site of injection. Note the strong induction in the expression of Xsnail that reaches to the animal cap. Inset, animal view of the embryo. About 50 embryos were analyzed in each experiment. Between 40 and 70% of effect was observed in each experiment.
However, the co-injection of BMP-4 but not of pCSKAXwnt8 was able to block the effect of Xslug mRNA injection on chordin expression (Fig. 7C,D). This result suggests that the ability of Xslug to induce dorsal mesodermal genes in ventral regions is dependent on the inhibition of BMP but not on Xwnt-8 activity. To address more directly the possibility that Xslug was controlling BMP-4 transcription, we analyzed the expression of BMP-4 in control stage 11 embryos (Fig. 7E) or embryos injected with Xslug or Xslug-en (Fig. 7F). A clear inhibition of BMP-4 was observed in embryos injected with Xslug and Xslug-en. 2.5. Xslug controls the expression of neural crest markers in the ectoderm It had been previously reported that the injection of Xslug mRNA leads to an expansion of the neural crest marker Xtwist (LaBonne and Bronner-Fraser, 1998) and that injection of Xslug antisense RNA produces an inhibition in the development of neural crest derivatives (Carl et al., 1999). Using the constructs that we have developed for this study, we decided to analyze the effect of modifying Xslug activity on neural crest markers. Xslug injection leads to an expansion of the neural crest markers ADAM and Xtwist (data not shown), similar to the previously described results (LaBonne and Bronner-Fraser, 1998). Xslug-VP16 injected in one blastomere of a two cell stage (Fig. 8A) leads to a
complete inhibition of the expression of the ADAM and Xtwist neural crest markers on the injected side (Fig. 8B,C), similar to the effect produced by blocking Xslug activity by Xslug antisense RNA (Carl et al., 1999). However, as we have shown in this study Xslug has a role in patterning the mesoderm and as it is known that the induction of the neural crest depends on particular regions of the mesoderm (Marchant et al., 1998), it is possible that the neural crest phenotype seen in this case is an indirect consequence of affecting mesodermal patterning. To rule out this possibility we injected Xslug-VP16 at the one cell stage and at stage 11 the prospective neural crest of the injected embryos were grafted onto the prospective neural crest of a control embryo, where the mesoderm was intact (Fig. 8D). As a control, prospective neural crest taken from embryos injected only with b-gal mRNA was used. Control embryos exhibited normal expression of the neural crest marker Xtwist (Fig. 8E), but the embryo grafted with ectoderm containing Xslug-VP16 mRNA showed a complete inhibition of Xtwist in the graft. This result shows that, in addition to having a role in mesodermal patterning, Xslug controls the development of the neural crest. 2.6. Xslug expression is controlled by Pax3 and Xslug in the neural crest We next asked which gene may be upstream of Xslug in
52
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
Fig. 8. Effect of Xslug-VP16 on neural crest development. Embryos were injected with 1 ng of Xslug-VP16 in one blastomere of a two cell stage embryo (A) and the expression of ADAM (B) and Twist (C) was analyzed at stage 17. Arrow, injected side. Notice the inhibition of the neural crest marker in the injected side. (D) Embryos were injected at the one cell stage with 1 ng of Xslug-VP16 mRNA and b-gal mRNA, cultured until stage 11 and the prospective neural crest region was grafted into the equivalent region of a control embryo. The embryos were cultured until stage 17, and stained for b-gal (blue) and Xtwist expression (purple). (E) Control embryo where the grafts were taken from an embryo injected with b-gal. Notice the normal expression of Xtwist in the graft (arrow). (F) Embryo showing the absence of Xtwist expression in the grafts that contained XslugVP16 (arrow). About 20 embryos were analyzed in each experiment.
Fig. 7. Effect of Xslug overexpression can be blocked by Xslug BMP-4 but not by Xwnt-8. Embryos co-injected with the different constructs and b-gal mRNA (blue staining) were analyzed at stage 10 by in situ hybridization (purple staining) for the expression of chordin (A±F) or at stage 11 for the expression of BMP-4 (G,H). Vegetal views, dorsal is up. (A) Control. (B) Embryo injected with 1 ng of Xslug mRNA at the one cell stage. Arrow, site of injection. Notice the ectopic expression of chordin (n 40, 68% of ectopic expression). (C) Embryo co-injected with 1 ng of Xslug mRNA and 0.5 ng of BMP-4 mRNA. Arrow, site of injection. Notice that no ectopic expression of chordin can be detected (n 45, 10% of ectopic expression). (D) Embryo co-injected with 1 ng of Xslug mRNA and 0.4 ng of pCSKAXwnt-8 cDNA. Arrow, site of injection. Notice that an ectopic expression of chordin can be detected. In this embryo the dorsal side is to the left in order to better show the injected site (n 38, 58% of ectopic expression). (E) Embryo injected with 0.5 ng of BMP-4 mRNA at the one cell stage. Note the inhibition in the expression of chordin (n 28, 100% of inhibition). (F) Embryo injected with 0.2 ng of pCSKAXwnt-8 cDNA in one blastomere at the two cell stage. Arrowhead, site of injection showing the inhibition in the expression of chordin (n 29, 82% of inhibition). (G) Control stage 11 showing normal BMP-4 expression. (H) Embryo injected with 1 ng of Xslug-en mRNA showing inhibition of BMP-4 expression in the injected side (arrowhead, n 35, 48% of inhibition).
neural crest speci®cation. A possible candidate for this activity is the Pax3 gene product because in Xenopus it is
expressed earlier than Xslug and a mouse homozygous for a mutation in Pax3 displays obvious defects in neural crest (Bang et al., 1997; Conway et al., 1997). Pax3 mRNA was injected in one blastomere of a two cell stage embryo and the expression of different neural crest markers was analyzed. A small expansion in the ectodermal domains of Xsna, Xtwist and ADAM was observed (data not shown) but a strong expansion in the expression of Xslug expression was detected in early and advanced neurula on the side injected with Pax3 mRNA (Fig. 9A,B). Another candidate to regulate Xslug expression is Xslug itself. To analyze this auto-regulatory loop we injected embryos in one blastomere of a two cell stage with Xslug, Xslug-en, or Xslug-VP16. Xslug and Xslug-en injections produced an expansion in the expression of Xslug (Fig. 9C) while embryos injected with Xslug-VP16 showed an inhibition in the expression of Xslug (Fig. 9D). 3. Discussion The function of the Slug gene has been the subject of several studies published recently. All of them have focused on the possible function of Slug on neural crest development which is the most conspicuous site of Slug expression. In a study by Nieto et al. (1994), the role of Slug in neural crest
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
Fig. 9. Regulation of Xslug expression. (A,B) Embryos were injected at the two cell stage with 2 ng of Pax3 mRNA and b-gal mRNA. The embryos were ®xed at stage 14 (A) or stage 17 (B) and the b-gal staining and expression of Xslug were analyzed. Arrow, injected side. Notice the expansion in Xslug expression at the injected side (n 52, 54% of expansion). (C,D) Embryos were injected with 1 ng of Xslug-en mRNA or 1 ng of Xslug-VP16 mRNA at the two cell stage and Xslug expression was analyzed at stage 17. Notice the expansion (C) or inhibition (D) of Xslug expression in the injected side (arrow, above 50 embryos were analyzed in each experiment, above 60% of effect was observed).
behavior was analyzed using antisense oligonucleotides directed against two sites in the coding region. Treatment of embryos with these antisense oligonucleotides clearly induced a transient decrease in Slug RNA, an inhibition in neural tube closure and a blocking in neural crest migration. On the other hand, mice homozygous for a null mutation in Slug do not display obvious defects in neural crest (Jiang et al., 1998). This suggest that the expression of another neural crest specifying gene, possibly Snail, is suf®cient to support neural crest development. LaBonne and Bronner-Fraser (1998) found that over-expression of Xslug in Xenopus causes an increase in Xtwist expression and in melanophore numbers. Carl et al. (1999) have used an antisense RNA directed against the 3 0 UTR of the Xslug cDNA, which leads to a reduction in the level of Xslug and Xsnail expression in the prospective neural crest region in addition to an inhibition in neural crest migration. In this study, it was not possible to discriminate between the effect of Xslug or Xsnail downregulation on neural crest migration. In conclusion there is some contradictory data coming from different organisms, chick and Xenopus on one hand and mouse on the other hand, that do not allow the general conclusion that Xslug is a gene involved in neural crest development. In
53
addition the loss of function studies made so far in Xenopus and chicken are using antisense oligos or RNA, a technique that does not always produce clear results. By making use of chimeric proteins of Xslug and VP16 and engrailed, we show that Xslug-VP16 works as a dominant negative of Xslug. Our results support the sense and antisense RNAs injection results, since Xslug sense RNA produced an expansion, and Xslug-VP16 a reduction, on the expression of other neural crest markers. This is similar to the results described by LaBonne and Bronner-Fraser (1998) and Carl et al. (1999). However, in all these studies, it has not been shown that the effect observed on neural crest development is not an indirect consequence of affecting mesodermal development, which is known to be required for a proper induction of the prospective neural crest cells (Marchant et al., 1998). In this study we have analyzed this possibility by making grafts of prospective neural crest cells expressing Xslug-VP16 in the same region of an untreated embryo. Our results show that Xslug-VP16 is able to block the development of the neural crest directly by affecting Xslug function in the ectoderm. We also were able to study the relationship between Xslug and Pax3. Pax3 has been implicated in the development of neural crest cells in mouse (Conway et al., 1997) and in Xenopus it is expressed very early in an area that includes the neural crest cells (Bang et al., 1997). Our results show that Pax3 is upstream of Xslug in a cascade that probably controls neural crest development (Fig. 10B). In addition we have shown that Xslug is able to regulate its own transcription, probably by repressing the activity of a repressor of Xslug transcription. While this work was being prepared for publication, it was reported that expression of fused constructs with Xslug and activation or repression domains produced similar results on neural crest markers to those described in our
Fig. 10. Summary of interactions of genes with Xslug in mesoderm (A) and neural crest (B). Question marks indicate interactions not directly proven. See text for details.
54
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
report (LaBonne and Bronner-Fraser, 2000). It was concluded that Xslug functions as a transcriptional repressor and is required for the proper development of the neural crest, con®rming our observations. The most novel ®nding of our study is the early function that Xslug has in mesendoderm development. We show for the ®rst time that Xslug is expressed in the organizer region and is able to control the expression of organizer genes such as chordin, cerberus and other dorsal mesodermal genes such as goosecoide, pintallavis and Frzb (data not shown). That Xslug is required for the normal development of the embryo is shown by the injection of Xslug-VP16; these embryos lack a head. This phenotype is probably explained by the inhibition of the expression of some dorsal genes, particularly cerberus, which is known to be involved in head development. The lack of ectopic head structures and a proper secondary axis after Xslug injection could be explained by the fact that the level of cerberus induced ectopically by Xslug overexpression is not suf®cient to activate the cascade required by the organizer activity. When Xslug is expressed in ventral mesoderm it induces dorsal genes and represses ventral genes, BMP-4 being one of these ventral genes. However, when Xslug-VP16 RNA is injected in dorsal blastomeres, dorsal genes are inhibited while no expression of ventral mesodermal genes is observed. This can be explained by the requirement of BMP-4 for the expression of ventral markers (Xvent-1, Xwnt-8), and as BMP-4 is not expressed in the dorsal mesoderm no ventral marker can be induced there. Interestingly, our results show that Xslug is able to block BMP transcription, and as a consequence, Xvent-1, a gene downstream of BMP, is blocked by Xslug mRNA injection. In addition, the effect of Xslug-VP16 can be rescued only by co-injection of BMP mRNA but not with Xwnt-8. The inhibition of Xwnt-8 caused by Xslug overexpression could also be a consequence of the interference with BMP-4 transcription, since a dominant negative BMP-4 receptor suppresses Xwnt-8 expression (Schmidt et al., 1995; Hoppler and Moon, 1998). Taken together these results indicate that Xslug could be a repressor of BMP in dorsal mesoderm. Thus, when it is ectopically expressed in ventral mesoderm the downregulation of BMP leads to an upregulation of dorsal genes such as chordin or cerberus (Fig. 10A). The clear effect of Xslug and Xslug-VP16 on Xsnail expression suggests that there is cross-regulation in the expression of these two genes, as has been proposed by Carl et al. (1999). The expression of Xsnail begins at stage 9, prior to gastrulation, in the dorsal marginal zone and progresses laterally to the ventral side by stage 10 (Sargent and Bennett, 1990; Essex et al., 1993). Xslug expression is detected at stage 10 in the dorsal marginal zone and remains present during gastrulation in the notochord (this study). After stage 10, Xsnail disappears from the dorsal marginal zone and is never expressed in the notochord. Therefore, there is a complementary pattern in the expression of Xsnail and Xslug in the mesoderm. Our results
show that Xslug is able to repress Xsnail in the marginal zone. Based on this information, we propose the following model: (i) Xsnail is ®rst induced in the dorsal marginal zone during mesodermal induction (Sargent and Bennett, 1990), (ii) this ®rst expression of Xsnail activates Xslug expression at this region at the beginning of gastrulation, and (iii) Xslug expression feeds back negatively on Xsnail blocking its expression in the dorsal marginal zone (Fig. 10A). This regulatory loop maintains the complementary pattern of Xsnail and Xslug expression and restricts the expression of Xsnail to the ventral marginal zone. In the dorsal marginal zone only Xslug is expressed, where it controls the development of dorsal mesoderm as was discussed above. It is interesting to notice that while Xslug represses Xsnail in the dorsal marginal zone, a different situation is observed in the ectoderm, where both genes are expressed in the same cells (Mayor et al., 1995; Linker et al., 2000). This suggests that Xslug functions are context dependent, and that it works as a repressor of Xsnail or BMP-4 in mesoderm but not in ectoderm. 4. Materials and methods 4.1. Constructs and in vitro RNA synthesis An activating form of Xslug (Xslug-VP16) was created by replacing the N-terminal 100 aa of Xslug with the 81 aa activation domain of VP16 protein (VP16 AD). The repressing form of Xslug (Xslug-en) was created by replacing the N-terminal 110 aa of Xslug with the repressor domain of the engrailed protein (Jaynes and O'Farrell, 1991). The Xslugmyc construct was made by fusing the C-terminal of the Xslug cDNA to ®ve copies of the myc epitope. All the vectors were linearized and transcribed as described by Harland and Weintraub (1985) with GTP cap analog (New England Biolabs). SP6, T3 or T7 RNA polymerases were used. After DNAse treatment, RNA was extracted with phenol-chloroform and precipitated with ethanol. mRNA to be injected was resuspended in water. 4.2. Embryos, microinjections and grafts Xenopus embryos were obtained as described previously (Mayor et al., 1993) and staged according to Niewkoop and Faber (1967). Grafts were performed as previously described (Mancilla and Mayor, 1996; Mayor et al., 1997). Synthetic mRNA was injected in the one or two cell stage embryos in 8±12 nl volume as described (Mayor et al., 1993). 4.3. Whole-mount in situ hybridization and X-gal staining Antisense RNA probes were prepared from Xslug (Mayor et al., 1995), Chordin (Sasai et al., 1994), cerberus (Bouwmeester et al., 1996), Xwnt-8 (Christian et al., 1991; Smith and Harland, 1991), Xvent-1 (Gawantka et al., 1995), Xsnail
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
(Essex et al., 1993), ADAM (Alfandari et al., 1997), and Xtwist (Hopwood et al., 1989) cDNA using digoxigenin as the label. Specimens were prepared, hybridized and stained by the method of Harland (1991) with modi®cations (Mancilla and Mayor, 1996). X-gal staining was performed as described in Gomez-Skarmeta et al. (1998). 4.4. Histology and antibody staining Embryos were ®xed in 4% formaldehyde after wholemount in situ hybridization. After being embedded in paraf®n, embryos were sectioned. Whole-mount immunocytochemistry with an antibody against myc (BoehringerMannheim) was performed as described in Smith (1993). Acknowledgements We are very grateful to Dr B. Blumberg for the VP16 clone. We are grateful to Dr Miguel Allende for helpful comments on the manuscript, to E. DeRobertis, M. Sargent, R. Moon, J. Gurdon, D. Alfandari and R. Harland for cDNAs, and to F. Espinoza for technical assistance. This investigation was supported by grants from Fondecyt (1990570) and the Human Frontier Science Program (RG0042/98), and by the Millenium Program and Universidad de Chile. References Alberga, A., Boulay, J.L., Kempe, E., Dennefeld, C., Haenlin, M., 1991. The Snail gene required for mesoderm formation in Drosophila is expressed dynamically in derivatives of all three germ layers. Development 111, 983±992. Alfandari, D., Wolfsberg, T.G., White, J.M., DeSimone, D.W., 1997. ADAM 13: a novel ADAM expressed in somitic mesoderm and neural crest cells during Xenopus laevis development. Dev. Biol. 182, 314± 330. Bang, A.G., Papalopuliu, N., Kintner, C., Goulding, M.D., 1997. Expression of Pax-3 is initiated in the early neural plate by posteriorizing signals produced by the organizer and by posterior non-axial mesoderm. Development 124, 2075±2085. Bouwmeester, T., Kim, S., Sasai, Y., Lu, B., DeRobertis, E.M., 1996. Cerberus is a head-inducing secreted factor expressed in the anterior endoderm of Spemann's organizer. Nature 382, 595±601. Carl, T.F., Dufton, C., Hanken, J., Klymkowsky, M.W., 1999. Inhibition of neural crest migration in Xenopus using antisense slug RNA. Dev. Biol. 213, 101±115. Christian, J.L., Moon, R.T., 1993. Interactions between Xwnt-8 and Spemann organizer signaling pathways generate dorsoventral pattern in the embryonic mesoderm of Xenopus. Genes Dev. 7, 13±28. Christian, J.L., McMahon, J.A., McMahon, A.P., Moon, R.T., 1991. Molecular nature of Spemann organizer: the role of the Xenopus homeobox gene goosecoide. Cell 67, 1111±1120. Conway, S.J., Henderson, D.J., Copp, A.J., 1997. Pax3 is required for cardiac neural crest migration in the mouse: evidence from the splotch (Sp2H) mutant. Development 124, 505±514. Essex, L.J., Mayor, R., Sargent, M.G., 1993. Expression of Xenopus Snail in mesoderm and prospective neural fold ectoderm. Dev. Dyn. 198, 108±122. Friedman, A.D., Triezenberg, S.J., McKnight, S.L., 1988. Expression of a
55
truncated viral trans-activator selectively impedes lytic infection by its cognate virus. Nature 335, 452±454. Fuse, N., Hirose, S., Hayashi, S., 1996. Determination of wing cell fate by Escargot and Snail genes in Drosophila. Development 122, 1059±1067. Gawantka, V., Delius, H., Hirschfeld, K., Blumenstock, C., Niehrs, C., 1995. Antagonizing the Spemann organizer: role of the homeobox gene Xvent-1. EMBO J. 14, 6268±6279. Gomez-Skarmeta, J.L., Glavic, A., De laCalle-Mustienes, E., Modollel, J., Mayor, R., 1998. Xiro, a Xenopus homolog of the Drosophila Iroquois complex genes, controls development of neural plate. EMBO J. 17, 181±190. Hammerschmidt, M., Nusslein, V.C., 1993. The expression of a zebra®sh gene homologous to Drosophila Snail suggests a conserved function in invertebrates and vertebrate gastrulation. Development 119, 1107± 1118. Harland, R.M., 1991. In situ hybridization: an improved whole mount method for Xenopus embryos. Methods Cell Biol. 36, 685±695. Harland, R.M., Weintraub, H., 1985. Translation of mRNA injected into Xenopus oocytes is speci®cally inhibited by antisense RNA. J. Cell Biol. 101, 1094±1099. Hoppler, S., Moon, R.T., 1998. BMP2/4 and Wnt-8 cooperatively pattern the Xenopus mesoderm. Mech. Dev. 71, 119±129. Hopwood, N.D., Pluck, A., Gurdon, J.B., 1989. A Xenopus mRNA related to Drosophila twist is expressed in response to induction in the mesoderm and the neural crest. Cell 59, 893±903. Ip, Y.T., Maggert, K., Levine, M., 1994. Uncoupling gastrulation and mesoderm differentiation in the Drosophila embryo. EMBO J. 13, 5826± 5834. Isaac, A., Sargent, M.G., Cooke, J., 1997. Control of vertebrate left-right asymmetry by Snail-related zinc ®nger gene. Science 275, 1301±1304. Jaynes, J.B., O'Farrell, P.H., 1991. Active repression of transcription by the engrailed homeodomain protein. EMBO J. 10, 1427±1433. Jiang, R., Lan, Y., Norton, C.R., Sundberg, J.P., Gridley, T., 1998. The Slug gene is not essential for mesoderm or neural crest development in mice. Dev. Biol. 198, 277±285. LaBonne, C., Bronner-Fraser, M., 2000. Snail-related transcriptional repressors are required in Xenopus for both the induction of the neural crest and its subsequent migration. Dev. Biol. 221, 195±205. LaBonne, C., Bronner-Fraser, M., 1998. Neural crest induction in Xenopus: evidence for a two-signal model. Development 125, 2403±2414. Leptin, M., 1991. Twist and Snail as positive and negative regulators during Drosophila mesoderm development. Genes Dev. 5, 1568±1576. Linker, C., Bronner-Fraser, M., Mayor, R., 2000. Relationship between gene expression domains of Xsnail, Xslug and Xtwist and cell movement in the prospective neural crest of Xenopus. Dev. Biol. in press. Mancilla, A., Mayor, R., 1996. Neural crest formation in Xenopus laevis: mechanisms of Xslug induction. Dev. Biol. 177, 580±589. Marchant, L., Linker, C., Ruiz, R., Guerrero, N., Mayor, R., 1998. The inductive properties of the mesoderm suggest that the neural crest cells are speci®ed by a BMP gradient. Dev. Biol. 198, 319±329. Mayor, R., Essex, L.J., Bennet, M.F., Sargent, M.G., 1993. Distinct elements of the Xsna promoter are required for mesodermal and ectodermal expression. Development 119, 661±671. Mayor, R., Morgan, R., Sargent, M.G., 1995. Induction of the prospective neural crest of Xenopus. Development 121, 767±777. Mayor, R., Guerrero, N., Martinez, C., 1997. Role of FGF and noggin in neural crest induction. Dev. Biol. 189, 1±12. Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, S., Levine, M., 1998. DCtBP mediates transcriptional repression by Knirps, Kruppel and Snail in the Drosophila embryo. EMBO J. 17, 7009±7020. Nieto, M.A., Sargent, M.G., Wilkinson, D.G., Cooke, J., 1994. Control of cell behavior during vertebrate development by Slug, a zinc ®nger gene. Science 264, 835±839. Niewkoop, P.D., Faber, J., 1967. Normal table of Xenopus laevis (Daudin), North-Holland, Amsterdam. Roark, M., Sturtevant, M.A., Emery, J., Vaessin, H., Grell, E., Bier, E.,
56
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
1995. Scratch, a paneural gene encoding a zinc ®nger protein related to Snail, promotes neuronal development. Genes Dev. 9, 2384±2398. Samakovlis, C., Manning, G., Stenberg, P., Hacohen, N., Cantera, R., Krasnow, M.A., 1996. Genetic control of epithelial tube fusion during Drosophila tracheal development. Development 122, 3531±3536. Sargent, M.G., Bennett, M.F., 1990. Identi®cation in Xenopus of a structural homologue of the Drosophila gene Snail. Development 109, 967± 973. Sasai, Y., Lu, B., Steinbeisser, H., Geissert, D., Gont, L.K., DeRobertis, E.M., 1994. Xenopus chordin: a novel dorsalizing factor activated by organizer-speci®c homeobox genes. Cell 79, 779±790. Schmidt, J.E., Suzuki, A., Ueno, N., Kimelman, D., 1995. Localized BMP4 mediates dorsal/ventral patterning in the early Xenopus embryo. Dev. Biol. 169, 37±50. Sefton, M., Sanchez, S., Nieto, M.A., 1998. Conserved and divergent roles for members of the Snail family of transcription factors in the chick and mouse embryo. Development 125, 3111±3121. Smith, J.C., 1993. Purifying and assaying mesoderm inducing factors from vertebrate embryos. In: Hartley, D. (Ed.). Cellular Interactions in
Development ± A Practical Approach, Oxford University Press, Oxford, pp. 181±204. Smith, W.C., Harland, R.M., 1991. Injected Xwnt-8 RNA acts early in Xenopus embryos to promote formation of a vegetal dorsalizing center. Cell 67, 753±765. Tanaka, M.M., Uemura, T., Oda, H., Takeichi, M., Hayashi, S., 1996. Cadherin-mediated cell adhesion and cell motility in Drosophila trachea regulated by the transcription factor Escargot. Development 122, 3697±3705. Thisse, C., Thisse, B., Schilling, T.F., Postlethwait, J.H., 1993. Structure of the zebra®sh Snail1 gene and its expression in wild-type, spadetail and no tail mutant embryos. Development 119, 1203±1215. Thisse, C., Thisse, B., Postlethwait, J.H., 1995. Expression of Snail2, a second member of the zebra®sh Snail family, in cephalic mesoderm and presumptive neural crest of wild-type and spadetail mutant embryos. Dev. Biol. 172, 86±99. Whiteley, M., Noguchi, P.D., Sensabaugh, S.M., Odenwald, W.F., Kassis, J.A., 1992. The Drosophila gene Escargot encodes a zinc ®nger motif found in Snail-related genes. Mech. Dev. 36, 117±127.
www.elsevier.com/locate/modo
A novel function for the Xslug gene: control of dorsal mesendoderm development by repressing BMP-4 R. Mayor*, N. Guerrero, R.M. Young, J.L. Gomez-Skarmeta, C. Cuellar Millennium Nucleus in Developmental Biology, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile Received 28 February 2000; received in revised form 19 June 2000; accepted 7 July 2000
Abstract The Snail family of genes comprise a group of transcription factors with characteristic zinc ®nger motifs. One of the members of this family is the Slug gene. Slug has been implicated in the development of neural crest in chick and Xenopus by antisense loss of function experiments. Here, we have generated functional derivatives of Xslug by constructing cDNAs that encode the Xslug protein fused with the transactivation domain of the virus-derived VP16 activator or with the repressor domain of the Drosophila Engrailed protein. Our results suggest that Xslug normally functions as a transcriptional repressor and that Xslug-VP16 behaves as a dominant negative of Xslug. In the present work, we con®rm and extend previous results that suggest that Xslug has an important function in neural crest development, by controlling its own transcription. In addition we have uncovered a new function for Xslug. We show that Xslug is expressed in the dorsal mesendoderm at the beginning of gastrulation, where is it able to upregulate the expression of dorsal genes. On the other hand when Xslug is expressed outside of the organizer it represses the expression of ventral genes. Our results indicate that this effect on mesodermal patterning depends on BMP activity, showing that Xslug can directly control the transcription of BMP-4. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Xslug; Xsnail; Organizer; Neural crest; BMP-4 pathway
1. Introduction Families of transcription factors present in restricted territories within embryos generate positional identity and control morphogenesis of different tissues. The Snail genes comprise a family of transcription factors with characteristic zinc ®nger motifs. In Drosophila, Snail is involved in the process of mesoderm formation (see Alberga et al., 1991; Leptin, 1991). This gene encodes a Zn-®nger transcription factor that represses the expression of neuroectodermal genes in the mesoderm (Leptin, 1991) and is required for the invagination of mesodermal cells during gastrulation (Ip et al., 1994). In Drosophila a second Snail-like gene, named Escargot (Whiteley et al., 1992), has been implicated in the maintenance of diploidy in imaginal cells (see Fuse et al., 1996). In tracheal cells, Escargot is required for normal expression of the cell adhesion molecule DE-cadherin (Tanaka et al., 1996) and the control of cell fusion during tracheal development (see Samakovlis et al., 1996). A third more distantly related Snail-like gene, Scratch, appears to be involved in the regu* Corresponding author. Tel.: 156-2-678-7351; fax: 156-2-271-2983. E-mail address: [email protected] (R. Mayor).
lation of neuronal development (Roark et al., 1995). Both Snail and Escargot proteins have similar DNA binding speci®city and have been proposed to act as repressors through interactions with the co-repressor C-terminal binding protein (CtBP) (Nibu et al., 1998). In vertebrates two members of the Snail family have been discovered in chick and Xenopus, named Snail and Slug. In Xenopus the expression of Xsnail begins at stage 9, prior to gastrulation, in the dorsal marginal zone, progresses laterally to the ventral side by stage 10 (Sargent and Bennett, 1990; Essex et al., 1993), and is subsequently found in the lateral plate mesoderm up until the tailbud stage. Xsnail expression can also be found in ectoderm and premigratory neural crest and their derivatives such as the branchial cartilages (Mayor et al., 1993). In contrast it has been reported that Xslug is expressed only in prospective neural crest and is not found in developing mesoderm before stage 17 (Mayor et al., 1995). The pattern of Snail and Slug expression in the chick is analogous to that seen in Xenopus, with the exception that Xsnail is found initially in prechordal mesoderm, but in the chick Snail is expressed preferentially in the right side lateral mesoderm, where it appears to play an important role in the development of asymmetry (Isaac et al., 1997). In the mouse, there is striking `swapping' in the
0925-4773/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(00)00412-3
48
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
pattern of Slug and Snail in many but not all expression domains (Sefton et al., 1998). In the zebra®sh, two Snaillike genes (Snail-1 and Snail-2) have been described (Hammerschmidt and Nusslein, 1993; Thisse et al., 1993, 1995); neither contains the 29 amino acid sequence motif characteristic of other vertebrate Slugs (see Sefton et al., 1998). In zebra®sh, Snail-1 is initially distributed widely up to the blastula stage. As in the case of Drosophila Snail, Snail-1 is associated with mesoderm formation. During gastrulation Snail-1 becomes restricted to the involuting cells of the germ ring and is absent from cells of the dorsal midline. Later in development, Snail-1 is found in the paraxial mesoderm, neural crest cells, and mesodermderived portions of the head (Thisse et al., 1993; Hammerschmidt and Nusslein, 1993). The Zebra®sh Snail-2 gene is expressed during gastrulation in cephalic mesendodermal cells, and is later restricted to the neural crest (Thisse et al., 1995). The function of Slug has been studied directly in the chicken and in the mouse. In the chick, embryos have been treated with antisense oligonucleotides directed against the Slug mRNA (Nieto et al., 1994). This treatment led to the inhibition of Slug expression and prevented neural crest migration. In contrast, mice homozygous for a null mutation in Slug are viable and display no obvious defects in neural crest formation, migration or development (Jiang et al., 1998), even though Slug is expressed in migratory neural crest cells (Jiang et al., 1998; Sefton et al., 1998). It has been suggested that the mouse may use the closely related gene Snail in place of Slug (Sefton et al., 1998), but this hypothesis has not been tested. The role of the Xslug gene in neural crest development in Xenopus has been recently analyzed using the injection of Xslug antisense RNA (Carl et al., 1999). The injection of Xslug antisense RNA led to reduced levels of both Xslug and Xsnail expression in the embryos, inhibition of neural crest migration and reduction of many neural crest derivatives. These results show that migration and differentiation of neural crest derivatives are dependent upon Xslug/Xsnail activity, but because Xslug antisense also reduced the level of Xsnail expression it is not possible to independently analyze the role of these two Zn-®nger genes by the antisense technique. In this study, we generated functional derivatives of Xslug by constructing cDNAs that encode the Xslug protein fused with the transactivation domain of the virus-derived VP16 protein (Friedman et al., 1988) or with the repressor domain of the Drosophila engrailed protein (Jaynes and O'Farrell, 1991). Our functional analysis using Xenopus embryos shows that Xslug-en produces the same effect as wild-type Xslug mRNA and the opposite effect to that of Xslug-VP16, suggesting that Xslug is likely to be a transcriptional repressor. In these experiments we con®rm the role of Xslug in neural crest development but, unexpectedly, we have uncovered a novel role for Xslug in mesodermal patterning. We show for the ®rst time that Xslug is expressed in the dorsal marginal zone at the early gastrula
stage and our results suggest that Xslug controls the development of dorsal mesoderm and represses the expression of ventral mesodermal genes. We propose a hierarchy of genes that control dorsal mesoderm development where Xslug functions as a repressor of BMP-4. 2. Results 2.1. Xslug is expressed in the Spemann organizer and the protein is localized in the cell nucleus It had been reported previously that the mesodermal expression of Xslug was ®rst detectable after stage 17 in the lateral plate (Mayor et al., 1995). In order to more carefully analyze the earliest mesodermal expression we performed more sensitive in situ hybridization experiments (see Mancilla and Mayor, 1996). We observed clear expression in the dorsal marginal zone from stage 10 onwards, above the dorsal blastopore lip (Fig. 1A). When these embryos were sectioned and analyzed for the in situ hybridization of chordin, cerberus and Xslug, the Xslug expression was not only in the prospective dorsal mesoderm but also in the deeper endodermal cells (Fig. 1B) which correspond to the cells that express Cerberus (Bouwmeester et al., 1996). At the neurula stage the expression is visible in the notochord, but it is weaker than the expression in the neural folds (Fig. 1C,D). The Slug gene encodes a putative transcription factor but to date there have been no reports that describe it to be localized in the cell nucleus. We constructed a cDNA that
Fig. 1. Xslug expression and cellular localization of Xslug protein. (A±D) In situ hybridization using a Xslug probe. (A) Vegetal view of a stage 10 embryo. d, dorsal; v, ventral; arrow, dorsal blastopore lip. Note the expression in the marginal zone above the blastopore lip. (B) Stage 10 embryo sectioned sagittally. An, animal pole; Vg, vegetal pole; D, dorsal; V, ventral; arrow, dorsal blastopore lip. Notice that Xslug is expressed in the dorsal marginal zone, including the super®cial and deep cells. (C) Dorsal view of a stage 15 embryo showing strong Xslug expression in the neural folds (arrowheads) and weaker expression in the notochord (n). (D) Paraf®n section of embryo shown in (C). Arrowheads, neural folds; n, notochord; s, somite. (E) Anti-myc immunostaining of a stage 11 embryo injected with Xslug-myc at the one cell stage. Note the nuclear localization of the Xslugmyc protein (arrows).
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
49
encodes the Xslug protein fused to a myc tag epitope, which can be recognized by immunostaining. The mRNA of this fused protein was injected at the one cell stage and at different stages the embryos were ®xed and the localization of the fused protein was analyzed. Nuclear localization of the construct was detected during all stages analyzed (Fig. 1E), showing for the ®rst time that Slug is a nuclear protein. 2.2. Xslug controls the expression of dorsal mesodermal genes We attempted to study Xslug function by injecting mRNAs that encode (a) the full length Xslug cDNA, (b) a fusion protein of Xslug and the repressor domain of the Drosophila engrailed (Xslug-en) (Jaynes and O'Farrell, 1991), and (c) the transactivation domain derived from VP16 (Xslug-VP16) (Friedman et al., 1988) (Fig. 2). The injection of Xslug or Xslug-en mRNAs at the one or two cell stage causes a partial and moderate secondary axis which is recognized as an expansion of the injected side (see Fig. 9C). However, injection of Xslug-VP16 at the two cell stage leads to a complete inhibition of head development, while the posterior regions of the embryo develop normally (Fig. 3). To explain this phenotype and taking into account that Xslug is expressed in the dorsal marginal zone, we decided to analyze if dorsal mesoderm genes, required for the normal development of the head, could be controlled by Xslug. The expression of chordin and cerberus (Fig. 4A,F) was analyzed in embryos injected with Xslug or the chimeric proteins. Injection of Xslug and Xslug-en mRNA produced a similar ectopic expression of chordin and cerberus in the ventral marginal zone (Fig. 4B,C,G,H), but the injection of Xslug-VP16 mRNA produced an inhibition of the dorsal markers at the site of injection (Fig. 4D,E,I,J). The similarity in the effect of Xslug and Xslugen and the opposite effect of Xslug-VP16 suggests that Xslug is likely to be a transcriptional suppressor of ventral meso-
Fig. 2. Schematic representation of wild-type and mutated Xslug molecules. The amino acids numbers are shown above each drawing of the predicted protein and the Zn-®ngers are indicated in blue. The top drawing represents wild-type Xslug protein. Repressing and activating forms of Xslug were constructed by replacing the N-terminal of Xslug with the repressor domain of the engrailed protein (red box) or the activation domain of VP16 (green box).
Fig. 3. Embryos were injected with 1 ng of Xslug-VP16 RNA at the two cell stage and analyzed at stage 35. Left column, uninjected embryos; right column, embryos injected with Xslug-VP16. Notice the absence of anterior development in the injected embryos (n 65, 42% of effect).
derm genes. In addition, the inhibition of these genes, particularly the inhibition of cerberus, could explain the absence of a head in embryos injected with Xslug-VP16. To determine the speci®city of the effects produced by these fused proteins on chordin and cerberus expression (Fig. 5A,F), a rescue experiment was performed. Embryos were injected with Xslug or Xslug-VP16 mRNA at the one cell stage, producing expansion (Fig. 5B,G) or inhibition (Fig. 5C,H) of the dorsal markers, respectively. When embryos were co-injected with Xslug-VP16 and Xslug mRNA a complete rescue on the expression of chordin and cerberus was observed (Fig. 5D,I). However, no rescue in the expression of these markers was observed when the embryos were co-injected with Xslug-VP16 and Xsnail mRNA (Fig. 5E,J), showing the speci®city of the XslugVP16 effect, which depends on Xslug and not on Xsnail activity. 2.3. Xslug is able to repress ventral mesodermal genes We have shown that Xslug is expressed in the prospective dorsal mesendoderm and is able to control the expression of genes expressed in that region. Our next question was whether Xslug was able to repress the expression of ventral mesodermal genes. Similar injections using the mRNAs encoding Xslug, Xslug-en and Xslug-VP16 were performed, and the expression of the ventral genes Xwnt-8 (Fig. 6A) and Xvent-1 (Fig. 6E) was analyzed. Injection of Xslug or Xslugen exhibited a similar effect: the expression of Xwnt-8 and Xvent-1 was decreased or abolished at the site of injection (Fig. 6B,C,F,G). This result suggests that the normal expression of Xslug at the dorsal marginal zone could have a role in repressing ventral mesodermal genes in that region. To analyze if Xslug was suf®cient to produce the dorsal repression of ventral genes we injected the Xslug-VP16 mRNA in the prospective dorsal mesoderm region of a four cell stage embryo and the expression of the ventral markers was
50
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
Fig. 4. Xslug controls the expression of dorsal genes. Embryos co-injected with the different constructs and b-gal mRNA (blue staining) were analyzed at stage 10 by in situ hybridization (purple staining) for the expression of chordin (A±E) and Cerberus (F±J). Vegetal view, dorsal is up. (A,F) Control embryos. (B,G) Embryos injected with 1 ng of Xslug mRNA in the ventral area of a two cell stage embryo. Arrow, site of injection. Notice the ectopic expression of the marker. (C,H) Embryo injected with 1 ng of Xslug-en RNA in the ventral area of a two cell stage embryo. Arrow, site of injection. Notice the ectopic expression of the marker. (D,I) Embryo injected with 1 ng of Xslug-VP16 mRNA intake in the dorsal side of one blastomere of a two cell stage embryo. Arrowhead, site of injection. Notice the inhibition in the expression of the marker. (E,J) Higher magni®cations of (D,I), respectively. Between 60 and 100 embryos were analyzed in each experiment. Between 42 and 82% of ectopic expression or inhibition was observed.
analyzed. No ectopic expression of Xwnt-8 or Xvent-1 was observed in the injected dorsal marginal zone (Fig. 6D,H), showing that although Xslug is able to repress the expression of ventral genes when expressed ventrally, its dorsal expression is not suf®cient for this suppressor activity. The Xsnail gene was also analyzed in a similar manner. This gene is expressed as a ring in the marginal zone at the beginning of gastrulation (Fig. 6I). The injection of Xslug or Xslug-en mRNA produced a strong inhibition of Xsnail (Fig. 6J,K), and the injection of Xslug-VP16 induced a very strong ectopic induction of Xsnail (Fig. 6L) that occasionally reached the animal pole (Fig. 6L, inset). This strong dependence of Xsnail on Xslug activity suggests that Xslug is a direct repressor of Xsnail in the dorsal mesoderm. 2.4. Xslug represses BMP-4 expression At the gastrula stage, the organizer secretes a variety of zygotic proteins that act as antagonists to various members
of the BMP and Wnt families of ligands, which are secreted by ventral and lateral mesodermal cells. BMPs and Wnts favor ventral development and the organizer antagonizes their activities. As Xslug is expressed in the organizer region and has the capacity to induce the expression of dorsal mesodermal markers we asked whether this activity depended on the inhibition of BMP or Wnt activity. The expression of chordin was analyzed under different conditions (Fig. 7A). Since Xslug mRNA injection produces an ectopic expansion of chordin (Fig. 7B) we analyzed whether this ectopic expression of a dorsal gene could be due to an inhibition of BMP-4 or Xwnt-8, two molecules that promote ventral mesoderm. We co-injected Xslug and BMP-4 mRNAs or Xslug and pCSKAXwnt-8, a cDNA that expresses Xwnt-8 after MBT transition (Christian and Moon, 1993), and we analyzed in which of theses conditions the effect of injecting Xslug could be inhibited. The injection of BMP-4 or pCSKAXwnt-8 leads to an inhibition in the expression of chordin (Fig. 7E,F).
Fig. 5. Speci®city of the effect of Xslug-VP16 on chordin expression. Embryos co-injected with the different constructs and b-gal mRNA (blue staining) were analyzed at stage 10 by in situ hybridization (purple staining) for the expression of chordin (A±E) and Cerberus (F±J). Vegetal views, dorsal is up. (A,F) Control embryos. (B,G) Embryos injected with 1 ng of Xslug mRNA in the ventral area of a two cell stage embryo. Arrow, site of injection. Note the ectopic expression of the marker (n 62, 45% of ectopic expression). (C,H) Embryo injected with 1 ng of Xslug-VP16 mRNA in the dorsal side of a one cell stage embryo. Arrowhead, site of injection. Note the inhibition in the expression of the marker (n 58, 85% of inhibition). (D,I) Embryo co-injected with 1 ng of Xslug-VP16 RNA and 1 ng of Xslug mRNA in the dorsal side of a one cell stage embryo (n 72, 10% of expression, 0% of inhibition). Note the rescue in the expression of the markers. (E,J) Embryo co-injected with 1 ng of Xslug-VP16 RNA and 1 ng of Xsnail mRNA in the dorsal side of a one cell stage embryo. Note that Xsnail is not able to rescue the inhibition of the dorsal markers by Xslug-VP16 (n 32, 89% of inhibition).
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
51
Fig. 6. Xslug inhibits the expression of ventral genes. Embryos co-injected with the different constructs and b-gal mRNA (blue staining) were analyzed at stage 10 by in situ hybridization (purple staining) for the expression of Xwnt-8 (A±D), Xvent-1 (E±H) and Xsnail (I±L). Vegetal views, dorsal is up. (A,E,I) Control embryos. (B,F,J) Embryos injected with 1 ng of Xslug mRNA in the ventral area of a one cell stage embryo. Arrowhead, site of injection. Note the inhibition in the expression of the gene. (C,G,K) Embryos injected with 1 ng of Xslug-en RNA in the ventral area of a one cell stage embryo. Arrowhead, site of injection. Note the inhibition in the expression of the gene. (D,H) Embryos injected with 1 ng of Xslug-VP16 RNA in the dorsal region of a one cell stage embryo. Arrowhead, site of injection. Note that no dorsal expression of the marker can be detected. (L) Embryo injected with 1 ng of Xslug-VP16 in one blastomere of a two cell stage embryo. Arrow, site of injection. Note the strong induction in the expression of Xsnail that reaches to the animal cap. Inset, animal view of the embryo. About 50 embryos were analyzed in each experiment. Between 40 and 70% of effect was observed in each experiment.
However, the co-injection of BMP-4 but not of pCSKAXwnt8 was able to block the effect of Xslug mRNA injection on chordin expression (Fig. 7C,D). This result suggests that the ability of Xslug to induce dorsal mesodermal genes in ventral regions is dependent on the inhibition of BMP but not on Xwnt-8 activity. To address more directly the possibility that Xslug was controlling BMP-4 transcription, we analyzed the expression of BMP-4 in control stage 11 embryos (Fig. 7E) or embryos injected with Xslug or Xslug-en (Fig. 7F). A clear inhibition of BMP-4 was observed in embryos injected with Xslug and Xslug-en. 2.5. Xslug controls the expression of neural crest markers in the ectoderm It had been previously reported that the injection of Xslug mRNA leads to an expansion of the neural crest marker Xtwist (LaBonne and Bronner-Fraser, 1998) and that injection of Xslug antisense RNA produces an inhibition in the development of neural crest derivatives (Carl et al., 1999). Using the constructs that we have developed for this study, we decided to analyze the effect of modifying Xslug activity on neural crest markers. Xslug injection leads to an expansion of the neural crest markers ADAM and Xtwist (data not shown), similar to the previously described results (LaBonne and Bronner-Fraser, 1998). Xslug-VP16 injected in one blastomere of a two cell stage (Fig. 8A) leads to a
complete inhibition of the expression of the ADAM and Xtwist neural crest markers on the injected side (Fig. 8B,C), similar to the effect produced by blocking Xslug activity by Xslug antisense RNA (Carl et al., 1999). However, as we have shown in this study Xslug has a role in patterning the mesoderm and as it is known that the induction of the neural crest depends on particular regions of the mesoderm (Marchant et al., 1998), it is possible that the neural crest phenotype seen in this case is an indirect consequence of affecting mesodermal patterning. To rule out this possibility we injected Xslug-VP16 at the one cell stage and at stage 11 the prospective neural crest of the injected embryos were grafted onto the prospective neural crest of a control embryo, where the mesoderm was intact (Fig. 8D). As a control, prospective neural crest taken from embryos injected only with b-gal mRNA was used. Control embryos exhibited normal expression of the neural crest marker Xtwist (Fig. 8E), but the embryo grafted with ectoderm containing Xslug-VP16 mRNA showed a complete inhibition of Xtwist in the graft. This result shows that, in addition to having a role in mesodermal patterning, Xslug controls the development of the neural crest. 2.6. Xslug expression is controlled by Pax3 and Xslug in the neural crest We next asked which gene may be upstream of Xslug in
52
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
Fig. 8. Effect of Xslug-VP16 on neural crest development. Embryos were injected with 1 ng of Xslug-VP16 in one blastomere of a two cell stage embryo (A) and the expression of ADAM (B) and Twist (C) was analyzed at stage 17. Arrow, injected side. Notice the inhibition of the neural crest marker in the injected side. (D) Embryos were injected at the one cell stage with 1 ng of Xslug-VP16 mRNA and b-gal mRNA, cultured until stage 11 and the prospective neural crest region was grafted into the equivalent region of a control embryo. The embryos were cultured until stage 17, and stained for b-gal (blue) and Xtwist expression (purple). (E) Control embryo where the grafts were taken from an embryo injected with b-gal. Notice the normal expression of Xtwist in the graft (arrow). (F) Embryo showing the absence of Xtwist expression in the grafts that contained XslugVP16 (arrow). About 20 embryos were analyzed in each experiment.
Fig. 7. Effect of Xslug overexpression can be blocked by Xslug BMP-4 but not by Xwnt-8. Embryos co-injected with the different constructs and b-gal mRNA (blue staining) were analyzed at stage 10 by in situ hybridization (purple staining) for the expression of chordin (A±F) or at stage 11 for the expression of BMP-4 (G,H). Vegetal views, dorsal is up. (A) Control. (B) Embryo injected with 1 ng of Xslug mRNA at the one cell stage. Arrow, site of injection. Notice the ectopic expression of chordin (n 40, 68% of ectopic expression). (C) Embryo co-injected with 1 ng of Xslug mRNA and 0.5 ng of BMP-4 mRNA. Arrow, site of injection. Notice that no ectopic expression of chordin can be detected (n 45, 10% of ectopic expression). (D) Embryo co-injected with 1 ng of Xslug mRNA and 0.4 ng of pCSKAXwnt-8 cDNA. Arrow, site of injection. Notice that an ectopic expression of chordin can be detected. In this embryo the dorsal side is to the left in order to better show the injected site (n 38, 58% of ectopic expression). (E) Embryo injected with 0.5 ng of BMP-4 mRNA at the one cell stage. Note the inhibition in the expression of chordin (n 28, 100% of inhibition). (F) Embryo injected with 0.2 ng of pCSKAXwnt-8 cDNA in one blastomere at the two cell stage. Arrowhead, site of injection showing the inhibition in the expression of chordin (n 29, 82% of inhibition). (G) Control stage 11 showing normal BMP-4 expression. (H) Embryo injected with 1 ng of Xslug-en mRNA showing inhibition of BMP-4 expression in the injected side (arrowhead, n 35, 48% of inhibition).
neural crest speci®cation. A possible candidate for this activity is the Pax3 gene product because in Xenopus it is
expressed earlier than Xslug and a mouse homozygous for a mutation in Pax3 displays obvious defects in neural crest (Bang et al., 1997; Conway et al., 1997). Pax3 mRNA was injected in one blastomere of a two cell stage embryo and the expression of different neural crest markers was analyzed. A small expansion in the ectodermal domains of Xsna, Xtwist and ADAM was observed (data not shown) but a strong expansion in the expression of Xslug expression was detected in early and advanced neurula on the side injected with Pax3 mRNA (Fig. 9A,B). Another candidate to regulate Xslug expression is Xslug itself. To analyze this auto-regulatory loop we injected embryos in one blastomere of a two cell stage with Xslug, Xslug-en, or Xslug-VP16. Xslug and Xslug-en injections produced an expansion in the expression of Xslug (Fig. 9C) while embryos injected with Xslug-VP16 showed an inhibition in the expression of Xslug (Fig. 9D). 3. Discussion The function of the Slug gene has been the subject of several studies published recently. All of them have focused on the possible function of Slug on neural crest development which is the most conspicuous site of Slug expression. In a study by Nieto et al. (1994), the role of Slug in neural crest
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
Fig. 9. Regulation of Xslug expression. (A,B) Embryos were injected at the two cell stage with 2 ng of Pax3 mRNA and b-gal mRNA. The embryos were ®xed at stage 14 (A) or stage 17 (B) and the b-gal staining and expression of Xslug were analyzed. Arrow, injected side. Notice the expansion in Xslug expression at the injected side (n 52, 54% of expansion). (C,D) Embryos were injected with 1 ng of Xslug-en mRNA or 1 ng of Xslug-VP16 mRNA at the two cell stage and Xslug expression was analyzed at stage 17. Notice the expansion (C) or inhibition (D) of Xslug expression in the injected side (arrow, above 50 embryos were analyzed in each experiment, above 60% of effect was observed).
behavior was analyzed using antisense oligonucleotides directed against two sites in the coding region. Treatment of embryos with these antisense oligonucleotides clearly induced a transient decrease in Slug RNA, an inhibition in neural tube closure and a blocking in neural crest migration. On the other hand, mice homozygous for a null mutation in Slug do not display obvious defects in neural crest (Jiang et al., 1998). This suggest that the expression of another neural crest specifying gene, possibly Snail, is suf®cient to support neural crest development. LaBonne and Bronner-Fraser (1998) found that over-expression of Xslug in Xenopus causes an increase in Xtwist expression and in melanophore numbers. Carl et al. (1999) have used an antisense RNA directed against the 3 0 UTR of the Xslug cDNA, which leads to a reduction in the level of Xslug and Xsnail expression in the prospective neural crest region in addition to an inhibition in neural crest migration. In this study, it was not possible to discriminate between the effect of Xslug or Xsnail downregulation on neural crest migration. In conclusion there is some contradictory data coming from different organisms, chick and Xenopus on one hand and mouse on the other hand, that do not allow the general conclusion that Xslug is a gene involved in neural crest development. In
53
addition the loss of function studies made so far in Xenopus and chicken are using antisense oligos or RNA, a technique that does not always produce clear results. By making use of chimeric proteins of Xslug and VP16 and engrailed, we show that Xslug-VP16 works as a dominant negative of Xslug. Our results support the sense and antisense RNAs injection results, since Xslug sense RNA produced an expansion, and Xslug-VP16 a reduction, on the expression of other neural crest markers. This is similar to the results described by LaBonne and Bronner-Fraser (1998) and Carl et al. (1999). However, in all these studies, it has not been shown that the effect observed on neural crest development is not an indirect consequence of affecting mesodermal development, which is known to be required for a proper induction of the prospective neural crest cells (Marchant et al., 1998). In this study we have analyzed this possibility by making grafts of prospective neural crest cells expressing Xslug-VP16 in the same region of an untreated embryo. Our results show that Xslug-VP16 is able to block the development of the neural crest directly by affecting Xslug function in the ectoderm. We also were able to study the relationship between Xslug and Pax3. Pax3 has been implicated in the development of neural crest cells in mouse (Conway et al., 1997) and in Xenopus it is expressed very early in an area that includes the neural crest cells (Bang et al., 1997). Our results show that Pax3 is upstream of Xslug in a cascade that probably controls neural crest development (Fig. 10B). In addition we have shown that Xslug is able to regulate its own transcription, probably by repressing the activity of a repressor of Xslug transcription. While this work was being prepared for publication, it was reported that expression of fused constructs with Xslug and activation or repression domains produced similar results on neural crest markers to those described in our
Fig. 10. Summary of interactions of genes with Xslug in mesoderm (A) and neural crest (B). Question marks indicate interactions not directly proven. See text for details.
54
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
report (LaBonne and Bronner-Fraser, 2000). It was concluded that Xslug functions as a transcriptional repressor and is required for the proper development of the neural crest, con®rming our observations. The most novel ®nding of our study is the early function that Xslug has in mesendoderm development. We show for the ®rst time that Xslug is expressed in the organizer region and is able to control the expression of organizer genes such as chordin, cerberus and other dorsal mesodermal genes such as goosecoide, pintallavis and Frzb (data not shown). That Xslug is required for the normal development of the embryo is shown by the injection of Xslug-VP16; these embryos lack a head. This phenotype is probably explained by the inhibition of the expression of some dorsal genes, particularly cerberus, which is known to be involved in head development. The lack of ectopic head structures and a proper secondary axis after Xslug injection could be explained by the fact that the level of cerberus induced ectopically by Xslug overexpression is not suf®cient to activate the cascade required by the organizer activity. When Xslug is expressed in ventral mesoderm it induces dorsal genes and represses ventral genes, BMP-4 being one of these ventral genes. However, when Xslug-VP16 RNA is injected in dorsal blastomeres, dorsal genes are inhibited while no expression of ventral mesodermal genes is observed. This can be explained by the requirement of BMP-4 for the expression of ventral markers (Xvent-1, Xwnt-8), and as BMP-4 is not expressed in the dorsal mesoderm no ventral marker can be induced there. Interestingly, our results show that Xslug is able to block BMP transcription, and as a consequence, Xvent-1, a gene downstream of BMP, is blocked by Xslug mRNA injection. In addition, the effect of Xslug-VP16 can be rescued only by co-injection of BMP mRNA but not with Xwnt-8. The inhibition of Xwnt-8 caused by Xslug overexpression could also be a consequence of the interference with BMP-4 transcription, since a dominant negative BMP-4 receptor suppresses Xwnt-8 expression (Schmidt et al., 1995; Hoppler and Moon, 1998). Taken together these results indicate that Xslug could be a repressor of BMP in dorsal mesoderm. Thus, when it is ectopically expressed in ventral mesoderm the downregulation of BMP leads to an upregulation of dorsal genes such as chordin or cerberus (Fig. 10A). The clear effect of Xslug and Xslug-VP16 on Xsnail expression suggests that there is cross-regulation in the expression of these two genes, as has been proposed by Carl et al. (1999). The expression of Xsnail begins at stage 9, prior to gastrulation, in the dorsal marginal zone and progresses laterally to the ventral side by stage 10 (Sargent and Bennett, 1990; Essex et al., 1993). Xslug expression is detected at stage 10 in the dorsal marginal zone and remains present during gastrulation in the notochord (this study). After stage 10, Xsnail disappears from the dorsal marginal zone and is never expressed in the notochord. Therefore, there is a complementary pattern in the expression of Xsnail and Xslug in the mesoderm. Our results
show that Xslug is able to repress Xsnail in the marginal zone. Based on this information, we propose the following model: (i) Xsnail is ®rst induced in the dorsal marginal zone during mesodermal induction (Sargent and Bennett, 1990), (ii) this ®rst expression of Xsnail activates Xslug expression at this region at the beginning of gastrulation, and (iii) Xslug expression feeds back negatively on Xsnail blocking its expression in the dorsal marginal zone (Fig. 10A). This regulatory loop maintains the complementary pattern of Xsnail and Xslug expression and restricts the expression of Xsnail to the ventral marginal zone. In the dorsal marginal zone only Xslug is expressed, where it controls the development of dorsal mesoderm as was discussed above. It is interesting to notice that while Xslug represses Xsnail in the dorsal marginal zone, a different situation is observed in the ectoderm, where both genes are expressed in the same cells (Mayor et al., 1995; Linker et al., 2000). This suggests that Xslug functions are context dependent, and that it works as a repressor of Xsnail or BMP-4 in mesoderm but not in ectoderm. 4. Materials and methods 4.1. Constructs and in vitro RNA synthesis An activating form of Xslug (Xslug-VP16) was created by replacing the N-terminal 100 aa of Xslug with the 81 aa activation domain of VP16 protein (VP16 AD). The repressing form of Xslug (Xslug-en) was created by replacing the N-terminal 110 aa of Xslug with the repressor domain of the engrailed protein (Jaynes and O'Farrell, 1991). The Xslugmyc construct was made by fusing the C-terminal of the Xslug cDNA to ®ve copies of the myc epitope. All the vectors were linearized and transcribed as described by Harland and Weintraub (1985) with GTP cap analog (New England Biolabs). SP6, T3 or T7 RNA polymerases were used. After DNAse treatment, RNA was extracted with phenol-chloroform and precipitated with ethanol. mRNA to be injected was resuspended in water. 4.2. Embryos, microinjections and grafts Xenopus embryos were obtained as described previously (Mayor et al., 1993) and staged according to Niewkoop and Faber (1967). Grafts were performed as previously described (Mancilla and Mayor, 1996; Mayor et al., 1997). Synthetic mRNA was injected in the one or two cell stage embryos in 8±12 nl volume as described (Mayor et al., 1993). 4.3. Whole-mount in situ hybridization and X-gal staining Antisense RNA probes were prepared from Xslug (Mayor et al., 1995), Chordin (Sasai et al., 1994), cerberus (Bouwmeester et al., 1996), Xwnt-8 (Christian et al., 1991; Smith and Harland, 1991), Xvent-1 (Gawantka et al., 1995), Xsnail
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
(Essex et al., 1993), ADAM (Alfandari et al., 1997), and Xtwist (Hopwood et al., 1989) cDNA using digoxigenin as the label. Specimens were prepared, hybridized and stained by the method of Harland (1991) with modi®cations (Mancilla and Mayor, 1996). X-gal staining was performed as described in Gomez-Skarmeta et al. (1998). 4.4. Histology and antibody staining Embryos were ®xed in 4% formaldehyde after wholemount in situ hybridization. After being embedded in paraf®n, embryos were sectioned. Whole-mount immunocytochemistry with an antibody against myc (BoehringerMannheim) was performed as described in Smith (1993). Acknowledgements We are very grateful to Dr B. Blumberg for the VP16 clone. We are grateful to Dr Miguel Allende for helpful comments on the manuscript, to E. DeRobertis, M. Sargent, R. Moon, J. Gurdon, D. Alfandari and R. Harland for cDNAs, and to F. Espinoza for technical assistance. This investigation was supported by grants from Fondecyt (1990570) and the Human Frontier Science Program (RG0042/98), and by the Millenium Program and Universidad de Chile. References Alberga, A., Boulay, J.L., Kempe, E., Dennefeld, C., Haenlin, M., 1991. The Snail gene required for mesoderm formation in Drosophila is expressed dynamically in derivatives of all three germ layers. Development 111, 983±992. Alfandari, D., Wolfsberg, T.G., White, J.M., DeSimone, D.W., 1997. ADAM 13: a novel ADAM expressed in somitic mesoderm and neural crest cells during Xenopus laevis development. Dev. Biol. 182, 314± 330. Bang, A.G., Papalopuliu, N., Kintner, C., Goulding, M.D., 1997. Expression of Pax-3 is initiated in the early neural plate by posteriorizing signals produced by the organizer and by posterior non-axial mesoderm. Development 124, 2075±2085. Bouwmeester, T., Kim, S., Sasai, Y., Lu, B., DeRobertis, E.M., 1996. Cerberus is a head-inducing secreted factor expressed in the anterior endoderm of Spemann's organizer. Nature 382, 595±601. Carl, T.F., Dufton, C., Hanken, J., Klymkowsky, M.W., 1999. Inhibition of neural crest migration in Xenopus using antisense slug RNA. Dev. Biol. 213, 101±115. Christian, J.L., Moon, R.T., 1993. Interactions between Xwnt-8 and Spemann organizer signaling pathways generate dorsoventral pattern in the embryonic mesoderm of Xenopus. Genes Dev. 7, 13±28. Christian, J.L., McMahon, J.A., McMahon, A.P., Moon, R.T., 1991. Molecular nature of Spemann organizer: the role of the Xenopus homeobox gene goosecoide. Cell 67, 1111±1120. Conway, S.J., Henderson, D.J., Copp, A.J., 1997. Pax3 is required for cardiac neural crest migration in the mouse: evidence from the splotch (Sp2H) mutant. Development 124, 505±514. Essex, L.J., Mayor, R., Sargent, M.G., 1993. Expression of Xenopus Snail in mesoderm and prospective neural fold ectoderm. Dev. Dyn. 198, 108±122. Friedman, A.D., Triezenberg, S.J., McKnight, S.L., 1988. Expression of a
55
truncated viral trans-activator selectively impedes lytic infection by its cognate virus. Nature 335, 452±454. Fuse, N., Hirose, S., Hayashi, S., 1996. Determination of wing cell fate by Escargot and Snail genes in Drosophila. Development 122, 1059±1067. Gawantka, V., Delius, H., Hirschfeld, K., Blumenstock, C., Niehrs, C., 1995. Antagonizing the Spemann organizer: role of the homeobox gene Xvent-1. EMBO J. 14, 6268±6279. Gomez-Skarmeta, J.L., Glavic, A., De laCalle-Mustienes, E., Modollel, J., Mayor, R., 1998. Xiro, a Xenopus homolog of the Drosophila Iroquois complex genes, controls development of neural plate. EMBO J. 17, 181±190. Hammerschmidt, M., Nusslein, V.C., 1993. The expression of a zebra®sh gene homologous to Drosophila Snail suggests a conserved function in invertebrates and vertebrate gastrulation. Development 119, 1107± 1118. Harland, R.M., 1991. In situ hybridization: an improved whole mount method for Xenopus embryos. Methods Cell Biol. 36, 685±695. Harland, R.M., Weintraub, H., 1985. Translation of mRNA injected into Xenopus oocytes is speci®cally inhibited by antisense RNA. J. Cell Biol. 101, 1094±1099. Hoppler, S., Moon, R.T., 1998. BMP2/4 and Wnt-8 cooperatively pattern the Xenopus mesoderm. Mech. Dev. 71, 119±129. Hopwood, N.D., Pluck, A., Gurdon, J.B., 1989. A Xenopus mRNA related to Drosophila twist is expressed in response to induction in the mesoderm and the neural crest. Cell 59, 893±903. Ip, Y.T., Maggert, K., Levine, M., 1994. Uncoupling gastrulation and mesoderm differentiation in the Drosophila embryo. EMBO J. 13, 5826± 5834. Isaac, A., Sargent, M.G., Cooke, J., 1997. Control of vertebrate left-right asymmetry by Snail-related zinc ®nger gene. Science 275, 1301±1304. Jaynes, J.B., O'Farrell, P.H., 1991. Active repression of transcription by the engrailed homeodomain protein. EMBO J. 10, 1427±1433. Jiang, R., Lan, Y., Norton, C.R., Sundberg, J.P., Gridley, T., 1998. The Slug gene is not essential for mesoderm or neural crest development in mice. Dev. Biol. 198, 277±285. LaBonne, C., Bronner-Fraser, M., 2000. Snail-related transcriptional repressors are required in Xenopus for both the induction of the neural crest and its subsequent migration. Dev. Biol. 221, 195±205. LaBonne, C., Bronner-Fraser, M., 1998. Neural crest induction in Xenopus: evidence for a two-signal model. Development 125, 2403±2414. Leptin, M., 1991. Twist and Snail as positive and negative regulators during Drosophila mesoderm development. Genes Dev. 5, 1568±1576. Linker, C., Bronner-Fraser, M., Mayor, R., 2000. Relationship between gene expression domains of Xsnail, Xslug and Xtwist and cell movement in the prospective neural crest of Xenopus. Dev. Biol. in press. Mancilla, A., Mayor, R., 1996. Neural crest formation in Xenopus laevis: mechanisms of Xslug induction. Dev. Biol. 177, 580±589. Marchant, L., Linker, C., Ruiz, R., Guerrero, N., Mayor, R., 1998. The inductive properties of the mesoderm suggest that the neural crest cells are speci®ed by a BMP gradient. Dev. Biol. 198, 319±329. Mayor, R., Essex, L.J., Bennet, M.F., Sargent, M.G., 1993. Distinct elements of the Xsna promoter are required for mesodermal and ectodermal expression. Development 119, 661±671. Mayor, R., Morgan, R., Sargent, M.G., 1995. Induction of the prospective neural crest of Xenopus. Development 121, 767±777. Mayor, R., Guerrero, N., Martinez, C., 1997. Role of FGF and noggin in neural crest induction. Dev. Biol. 189, 1±12. Nibu, Y., Zhang, H., Bajor, E., Barolo, S., Small, S., Levine, M., 1998. DCtBP mediates transcriptional repression by Knirps, Kruppel and Snail in the Drosophila embryo. EMBO J. 17, 7009±7020. Nieto, M.A., Sargent, M.G., Wilkinson, D.G., Cooke, J., 1994. Control of cell behavior during vertebrate development by Slug, a zinc ®nger gene. Science 264, 835±839. Niewkoop, P.D., Faber, J., 1967. Normal table of Xenopus laevis (Daudin), North-Holland, Amsterdam. Roark, M., Sturtevant, M.A., Emery, J., Vaessin, H., Grell, E., Bier, E.,
56
R. Mayor et al. / Mechanisms of Development 97 (2000) 47±56
1995. Scratch, a paneural gene encoding a zinc ®nger protein related to Snail, promotes neuronal development. Genes Dev. 9, 2384±2398. Samakovlis, C., Manning, G., Stenberg, P., Hacohen, N., Cantera, R., Krasnow, M.A., 1996. Genetic control of epithelial tube fusion during Drosophila tracheal development. Development 122, 3531±3536. Sargent, M.G., Bennett, M.F., 1990. Identi®cation in Xenopus of a structural homologue of the Drosophila gene Snail. Development 109, 967± 973. Sasai, Y., Lu, B., Steinbeisser, H., Geissert, D., Gont, L.K., DeRobertis, E.M., 1994. Xenopus chordin: a novel dorsalizing factor activated by organizer-speci®c homeobox genes. Cell 79, 779±790. Schmidt, J.E., Suzuki, A., Ueno, N., Kimelman, D., 1995. Localized BMP4 mediates dorsal/ventral patterning in the early Xenopus embryo. Dev. Biol. 169, 37±50. Sefton, M., Sanchez, S., Nieto, M.A., 1998. Conserved and divergent roles for members of the Snail family of transcription factors in the chick and mouse embryo. Development 125, 3111±3121. Smith, J.C., 1993. Purifying and assaying mesoderm inducing factors from vertebrate embryos. In: Hartley, D. (Ed.). Cellular Interactions in
Development ± A Practical Approach, Oxford University Press, Oxford, pp. 181±204. Smith, W.C., Harland, R.M., 1991. Injected Xwnt-8 RNA acts early in Xenopus embryos to promote formation of a vegetal dorsalizing center. Cell 67, 753±765. Tanaka, M.M., Uemura, T., Oda, H., Takeichi, M., Hayashi, S., 1996. Cadherin-mediated cell adhesion and cell motility in Drosophila trachea regulated by the transcription factor Escargot. Development 122, 3697±3705. Thisse, C., Thisse, B., Schilling, T.F., Postlethwait, J.H., 1993. Structure of the zebra®sh Snail1 gene and its expression in wild-type, spadetail and no tail mutant embryos. Development 119, 1203±1215. Thisse, C., Thisse, B., Postlethwait, J.H., 1995. Expression of Snail2, a second member of the zebra®sh Snail family, in cephalic mesoderm and presumptive neural crest of wild-type and spadetail mutant embryos. Dev. Biol. 172, 86±99. Whiteley, M., Noguchi, P.D., Sensabaugh, S.M., Odenwald, W.F., Kassis, J.A., 1992. The Drosophila gene Escargot encodes a zinc ®nger motif found in Snail-related genes. Mech. Dev. 36, 117±127.