A genetic regulatory network for Xenopus mesendoderm formation

Developmental Biology 271 (2004) 467 – 478 www.elsevier.com/locate/ydbio

Genomes & Developmental Control

A genetic regulatory network for Xenopus mesendoderm formation Matthew Loose and Roger Patient * Institute of Genetics, University of Nottingham, Queen’s Medical Centre, Nottingham NG7 2UH, UK Received for publication 18 February 2004, revised 5 April 2004, accepted 19 April 2004 Available online 25 May 2004

Abstract We have constructed a genetic regulatory network (GRN) summarising the functional relationships between the transcription factors (TFs) and embryonic signals involved in Xenopus mesendoderm formation. It is supported by a relational database containing the experimental evidence and both are available in interactive form via the World Wide Web. This network highlights areas for further study and provides a framework for systematic interrogation of new data. Comparison with the equivalent network for the sea urchin identifies conserved features of the deuterostome ancestral pathway, including positive feedback loops, GATA factors, SoxB, Brachyury and a previously underemphasised role for h-catenin. In contrast, some features central to one species have not yet been found in the other, for example, Krox and Otx in sea urchin, and Mix and Nodal in Xenopus. Such differences may represent evolved features or may eventually be resolved. For example, in Xenopus, Nodal-related genes are positively regulated by h-catenin and at least one of them is repressed by Sox3, as is the uncharacterised early signal (ES) inducing endomesoderm in the sea urchin, suggesting that ES may be a Nodal-like TGF-h. Wider comparisons of such networks will inform our understanding of developmental evolution. D 2004 Elsevier Inc. All rights reserved. Keywords: Xenopus; Nodal; h-Catenin

Introduction The fates of cells are determined by their gene expression profiles. Key determinants of cell fate are therefore transcription factors (TFs). These proteins, often in response to extracellular signals, bind to DNA and regulate gene transcription. Amongst their target genes are other TFs and signals, and these cross-regulating, or epistatic, relationships serve to stabilise patterns of gene expression in the absence of the initiating signals. Thus, to a first level of approximation, extracellular signals, intracellular TF responses and the relationships between them provide insight into the molecular programming of specific subsets of cells during development. Amongst the earliest distinctions between cells during embryonic development is the formation of the primary germ layers, the endoderm, the ectoderm and the mesoderm. The endoderm and at least part of the mesoderm develop in close association, share inducing signals and may derive from a population of progenitors known as the mesendoderm or endomesoderm (Cameron and Davidson, 1991; Kimelman * Corresponding author. Fax: +44-115-970-9906. E-mail address: [email protected] (R. Patient). 0012-1606/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2004.04.014

and Griffin, 2000; Rodaway and Patient, 2001). To build into a relational format signalling and TF information on the formation of this bipotential progenitor in sea urchins, Bolouri and Davidson et al. (2002a,b) designed a computer programme called Netbuilder. This programme systematically represents the interactions between signals/TFs and their target genes, which initially code for other signals/TFs, allowing the construction of a genetic regulatory network (GRN). The data used to build such networks derive from expression profiles, perturbation experiments and the definition of cis-regulatory elements present in promoters and enhancers of the target genes. Substantial quantities of such information relevant to vertebrate mesendoderm formation are available for the amphibian, Xenopus laevis, so we have used it to construct the network outlined here. As for the sea urchin network, three criteria have been used for a given TF – target relationship to be considered direct: 1. The expression of the putative target is affected by perturbation of its activator or repressor (in the absence of cell loss). 2. The expression patterns of the target and its upstream activator/repressor must be consistent with their proposed relationship.

468

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

3. The promoter/enhancers of the target gene must contain binding sites for the activating/repressing transcription factor or the target must be shown to respond directly to the upstream factor, in the presence of protein synthesis inhibitors for example. Direct relationships are depicted by continuous lines ending in an arrowhead or crossbar, depending on whether the input is positive or negative, respectively (see key to Fig. 1). If criteria 1 and 2 are satisfied but criterion 3 is unknown, the interaction is described as ‘indirect’, which is depicted as a dashed line. Transcription factors that bind DNA via another transcription factor are considered direct if they are obligate members of a complex shown to bind DNA. The intercellular signalling molecules thought to be important in Xenopus mesendoderm formation are the Nodal-related family and Derriere (Hashimoto-Partyka et al., 2003; Jones et al., 1995; Joseph and Melton, 1997; Onuma et al., 2002; Rex et al., 2002; Smith et al., 1995; Sun et al., 1999; Takahashi et al., 2000; Yang et al., 2002; Yokota et al., 2003), and FGF (Schulte-Merker and Smith, 1995). We have included these signals and their known targets in the Xenopus mesendoderm network (XMN), relying on the same criteria for interaction, with the relevant binding sites being for the nuclear effectors of the signalling pathways, such as the SMADs and the forkhead transcription factor FAST/FoxH1 in the case of TGF-hs (Boggetti et al., 2000; Nomura and Li, 1998). To distinguish signalling molecules from transcription factors, their genes are marked with a circled ‘s’ (see Fig. 1) and chevrons are placed on their gene outputs to reflect their action via receptors. We have used Netbuilder to build the XMN and, importantly, have enhanced the utility of the resulting network by linking it to a database containing the data from which it was constructed. Both are available on the World Wide Web (Web Ref. 1). Currently, the database allows users to submit new data and criticise data already present within the network. For each connection made in the network, a summary of the known information about that link is provided as well as a link to the relevant published papers. This database is being developed to allow enhanced searching and querying of the data. Also available are interactive versions of the mesendoderm network diagrams. By selecting an individual gene, the user can see either those genes that have been identified as downstream targets, or those that activate or repress the selected gene. Clicking on the gene takes the user to the evidence for those links. The network will be continually refined as new data become available and networks for additional species are under construction.

The Xenopus mesendoderm network—overview Figs. 1 –4 show the current summary (version 1.0) of the XMN, but we recommend using the interactive mode of the network diagram on the internet (Web Ref. 1), particularly

as the interactions become more numerous , for example, in Figs. 3 and 4. An example of how targets for a particular regulator can be highlighted using this facility is given in Fig. 5 for the transcription factor VegT. For the sea urchin endomesoderm network, two ‘views’ were defined: the view from the genome and the view from the nucleus (Davidson et al., 2002a). The former contains all the known interactions between signals and TFs, whereas the latter contains only those active in a particular group of cells (or nuclei) at a particular time. We have chosen to represent the XMN as a series of views from the nucleus. Currently, the network is subdivided into four time periods, based on morphological and gene expression criteria, spanning the earliest interactions after fertilisation up to the end of gastrulation (stage 13). The first time point covers the period up to the midblastula transition (MBT, stage 7.5) and includes maternal effects and the zygotic expression of Xnr5 and Xnr6, which has been shown to occur before the MBT (Yang et al., 2002). Maternal RNAs and proteins are represented as genes without inputs, reflecting their synthesis before egg maturation and fertilisation. MBT is characterised by a surge of zygotic gene expression; however, a further burst is seen around stage 9. We have therefore chosen to treat these blastula stages as two time points: early (stages 7.5 – 9) and late (stages 9– 10.5). Gastrulation initiates at stage 10.5 and is over by stage 13 when the mesoderm and the endoderm are specified. Figs. 1 –3 consider the mesendoderm as a whole, whilst Fig. 4 (gastrulation, stages 10.5– 13) reflects the emergence of individual germ layers, with Brachyury and FGFs dominating the mesoderm and Mix family members with Sox17 dominating the endoderm. Within Figs. 1 – 3, gene products are placed on the network based on the time at which they are first expressed. To make comparison between different stages easier, these positions remain constant throughout the diagrams. Thus, a gene having an effect at a later stage of development will send an output ‘down the page’ to a target that is expressed later. This type of interaction corresponds to an initiation of expression. In some cases, initiation is distinct from maintenance of expression with the maintenance factors being expressed later in development than the targets. In these cases, genes send an output ‘up the page’. The current network is divided primarily by time, with only minimal spatial distinctions within the mesendoderm. Dorsally restricted genes have been placed on the right-hand side of the diagram, whilst genes placed on the left either have wider expression patterns throughout the mesendoderm or, as for Xwnt8, are excluded from dorsal. Where a gene is activated on one side of the embryo, but repressed on another, it is represented twice. For example, Xwnt8 is dorsally repressed by Goosecoid, but activated throughout the ventro-lateral regions of the embryo, possibly via VegT (Fig. 2). As gastrulation begins (stage 10.25), mesoderm and endoderm start to become distinguishable and the network has been redrawn to reflect this (Fig. 4). Spatial constraints, however, make it difficult to continue to represent time and

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

469

Fig. 1. Pre-MBT (Fertilisation to stage 7.5). Fast-1, Vg1, VegT, Sox3 and h-catenin are all expressed maternally, with Fast-1, Vg1 and VegT functioning as activators in the vegetal hemisphere, and Sox3 as a repressor in the animal hemisphere, restricting the domain of the mesendoderm. h-Catenin is localised on the dorsal side of the embryo and is considered to act with Xtcf3 as an activator. h-Catenin has been represented as nh-catenin, distinguishing the nuclearised h-catenin with transcriptional activity from the cytoplasmic pool. The only interactions so far characterised during this period are VegT and h-catenin activation of Xnr5 and 6, and Sox3 repression of Xnr5.

patterning within individual germ layers. As the data become available, we will further subdivide the network presented here to represent smaller subgroups of cells and time points to more accurately describe their states. It is not our intention to discuss every interaction contained within the XMN as all are extant in the literature and can be obtained by searching our website. We have chosen to report here those interactions that could currently be described as between major players in the network, together with significant observations made recently, summarised for each of the stages outlined. A comprehensive review of endoderm development has recently been published by Shivdasani (2002) and for a summary of events in mesoderm and endoderm induction, readers are referred to Smith and White (2003).

these targets, expression of Xnr5 and 6 has been demonstrated before the MBT, as early as the 256 cell stage, and their induction has been shown to be cell-autonomous (Takahashi et al., 2000; Yang et al., 2002). This expression of Xnr5 and 6 has been shown to depend on h-catenin and its DNA binding partner, XTcf3, as well as VegT (Rex et al., 2002; Xanthos et al., 2002; Yang et al., 2002). Thus, Nodal signalling appears to be activated before MBT, but only on the future dorsal side of the embryo where Xnr5 and 6 expression has been detected, coincident with nuclear hcatenin at this time. Very recently, Zhang et al. (2003) have shown that Xnr5 is negatively regulated by Sox3. Sox3 is expressed maternally in the animal hemisphere where it restricts mesendoderm induction (Penzel et al., 1997). Induction of these early Nodal-related genes may therefore also require downregulation of Sox3 activity or expression in the marginal zone.

Fertilisation to stage 7.5 ‘‘Pre-MBT’’ (Fig. 1) In X. laevis, the maternal T-box transcription factor VegT is an essential initiator of mesendoderm induction and has been discussed extensively in the literature (Clements and Woodland, 2003; Clements et al., 1999; Engleka et al., 2001; Kofron et al., 1999; Xanthos et al., 2001, 2002; Zhang et al., 1998). Maternal VegT RNA and protein are localised to the vegetal hemisphere of the egg and are thought to drive many zygotically expressed target genes including the TGF-h mesendoderm inducers, Xnr1, 2, 4, 5, 6 and Derriere, and the endoderm gene, Sox17 (Fig. 5). Of

Stages 7.5 –9 ‘‘early blastula’’ (Fig. 2) Entry into this phase is marked by the MBT. In addition to the earlier activation of Xnr5 and 6, Xnr1, 2, 4 and the TGF-h family member Derriere are now induced by VegT, and this establishes a high and continuous level of TGF-h signalling (Agius et al., 2000; Clements et al., 1999; White et al., 2002). Nuclear h-catenin spreads around the entire mesendodermal ring in the equator of the embryo (Schohl and Fagotto, 2003) and, now partnered by Lef-1 (Roel et

470

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

Fig. 2. Early blastula (stages 7.5 – 9). After MBT (dashed, horizontal line), the remaining Nodal genes are activated along with another TGF-h, Derriere. Xnr4 is dorsally restricted, whilst Xnr1 and 2 will ultimately have a wider expression pattern. Goosecoid and Siamois are rapidly induced and the earliest expression of Sox17, Brachyury, Pitx2 and Eomesodermin (Eomes) can be detected. The major activators within the network during this period are still the maternal signals, with h-catenin now partnered by Lef-1 (Roel et al., 2002), and the zygotic nodals, Xnr5 and Xnr6. Three T-box transcription factors (VegT, Brachyury and Eomes) are present at these stages. Whilst it is possible to functionally distinguish Brachyury and VegT targets, VegT and Eomes may share common targets (Conlon et al., 2001). The dorsal anterior mesendoderm is highlighted and Goosecoid is shown repressing Xwnt8 expression in these cells. It should be noted that although Vg1 is expressed at the right time and in the right place to be an inducer, all Vg1-dependent gene expression so far demonstrated results from overexpression of activated Vg1, and therefore may reflect the ability of Vg1 to mimic Nodal signalling as opposed to the in vivo activity of Vg1 itself.

al., 2002), can synergise with VegT to activate the other Xnrs, although it is not essential for their expression (Rex et al., 2002). Thus, the initial activation of Nodals is via VegT, either requiring or enhanced by h-catenin. Subsequently, the Nodals maintain their own expression by positive feedback, thereby locking the region into a mesendodermal fate. An antisense b-catenin morpholino oligonucleotide injected radially into the equatorial region of early cleavage embryos led to a downregulation of Brachyury as well as Goosecoid expression (Schohl and Fagotto, 2003). In contrast, Xwnt8 expression was unaffected by b-catenin morpholinos, but was inhibited by the dorsalising action of bcatenin overexpression (Schohl and Fagotto, 2003). To confirm the effect on Brachyury, a dominant negative XTcf3 construct was used to inhibit h-catenin activity on both sides of the embryo. As a consequence, the expression of Brachyury was inhibited on both sides of the embryo (Schohl and Fagotto, 2003). Thus, induction of the classical vertebrate mesoderm gene, Brachyury, appears to require bcatenin, possibly indirectly via Nodals. Another target of b-catenin is Siamois, which together with VegT and Nodal (Fast1) signalling, induces Goosecoid

(Carnac et al., 1996; Latinkic and Smith, 1999). Goosecoid is only expressed dorsally, being sensitive to the concentration/ timing of Nodal signalling and possibly also b-catenin via the activity of Siamois. Thus, high/early levels of b-catenin contribute to the regulation of Xwnt8 by activating its repressor, Goosecoid, in the dorsal-most mesendodermal cells. The coordination of VegT and Nodal inputs is illustrated by one of their target genes, Sox17, which is activated early in a VegT-dependent manner but requires TGF-h signalling for continued expression (Clements and Woodland, 2003). Supporting this model is the observation that a dominant negative form of Xnr2 or Derriere is able to block the rescue of Sox17 expression in VegT depleted embryos by VegT (Xanthos et al., 2001).

Stages 9.0 –10.25 ‘‘late blastula’’ (Fig. 3) The distinction between early blastula and late blastula is based on the observation that considerable gene expression initiates around stage 9 of development. Zygotic VegT expression begins around stage 9.5 yet all analysis of VegT suggests its major role is upstream of the Nodals and

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

471

Fig. 3. Late blastula (stages 9 – 10.25). At stage 9, a large number of genes are first activated. The relative timing for these genes is difficult to determine from the available literature. Maternal signals are still present, although presumably they are decreasing in level, and the task of driving the mesendoderm network forward is switching from VegT to the Nodal family.

effectively maternal. Other genes induced at this stage include the Mix and GATA families of TFs and signalling molecules such as FGFs, in addition to continued maintenance of the Nodals and Xwnt8. A reduction in phosphorylation of the FGF signal transduction component, MAP-K, has been observed as a consequence of b-catenin morpholino injection (Schohl and Fagotto, 2003). Embryonic FGF (eFGF) is a target for Brachyury and therefore presumably this reduction of activity is a consequence of reduced Brachyury expression. The FGF signalling pathway is important in the maintenance of Brachyury expression, providing yet another example of a positive feedback loop, utilising an extracellular signal to stabilise a particular programme of gene expression and establish a ‘‘community effect’’ in an emerging germ layer, in this case the future mesoderm (Schulte-Merker and Smith, 1995; Standley et al., 2001). The Mix family X. laevis has seven Mix family members, Mix.1, Mix.2, Mixer, Bix1/Mix.4, Bix2/Milk, Bix3 and Bix4 (Casey et al.,

1999; Ecochard et al., 1998; Henry and Melton, 1998; Rosa, 1989; Saka et al., 2000; Tada et al., 1998; Vize, 1996). They can be divided into two broad classes based on their upstream genes. Whilst all seven family members respond to VegT, only the Bix (Brachyury inducible homeobox) genes have been shown to respond to Brachyury, with some evidence that either Brachyury or VegT is required for activation (Saka et al., 2000; Tada et al., 1998). Mix.1, Mix.2, Mixer and Bix4 have been shown to respond to Nodal signals, indeed Mixer has been proposed to be a downstream effector of the Nodal signalling pathway (Kunwar et al., 2003; Trinh le et al., 2003; Xanthos et al., 2001). GATA factors Six GATA factors are found in all vertebrates (Patient and McGhee, 2002). They have classically been subdivided into two families, GATA 1– 3 being found in haematopoietic cells and ectodermal derivatives, whereas GATA 4– 6 are found in endodermal derivatives and the heart. The upstream regulation of these factors supports this distinction

472

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

Fig. 4. Gastrula (stages 10.25 – 13). Over this period, the mesendoderm is subdividing into mesoderm and endoderm. The network has been redrawn to reflect this subdivision with Brachyury dominating the mesoderm (red background) and Mix family members dominating the endoderm (blue background). GATA 4 – 6 are shown at the transition from endoderm to mesoderm because of their role in the specification of cardiac mesoderm as well as endoderm (Patient and McGhee, 2002). The anterior mesendoderm is distinguishable from the mesoderm and endoderm at this stage. The expression of Xnr3 and the repression of Brachyury by Sip1 in the ectoderm are also shown. By the onset of gastrulation, the maternal contributions to the network are no longer active and the expression of the early Nodal signals Xnr5 and Xnr6 is decreasing. Auto-regulation of the remaining Nodal family members, expressed in the endoderm but able to signal to adjacent tissues, continues to drive the network forward. Several genes expressed at this stage of development were activated by maternal factors, but the mechanism by which their expression is maintained is not known. They are therefore shown as genes with no inputs (e.g. Fgf3, Fgf8 and Xenf).

with GATA 4 –6 being regulated in a similar manner in response to VegT and the Xnrs (Xanthos et al., 2001).

Stages 10.25 –13 ‘‘gastrula’’ (Fig. 4) During this period, the mesoderm and the endoderm start to become distinct. This can be visualised for example by the increasing separation of expression domains of the future endodermal gene, Mix.1, and the future mesodermal gene, Brachyury (Lemaire et al., 1998). The mechanism of this separation is likely to include mutual repression. Thus, Mix.1 has been shown to downregulate the expression of Brachyury (Latinkic and Smith, 1999). Likewise, Bix2/milk is able to downregulate Brachyury, and a role for the Mix family in promoting endoderm as well as repressing mesoderm has been suggested (Ecochard et al., 1998). Mixer for example has been shown to upregulate GATA-5, which can induce endoderm in presumptive ectoderm (Mead et al., 1998; Weber et al., 2000). Sip1 represses Brachyury in the ectoderm where Sip1 is expressed with Xnr3 (Papin et al.,

2002). The remainder of the Xnrs and Derriere are now expressed within the endoderm and by the end of this period Xnr5 and Xnr6 expression has decreased (Takahashi et al., 2000). Finally, in the anterior mesendoderm, Brachyury, is repressed by Goosecoid.

Comparison with the sea urchin endomesoderm network Table 1 lists the genes important in the Xenopus or sea urchin mesendoderm networks, and their orthologues where they have been reported. Specification of the domain that gives rise to mesendoderm in the sea urchin is dependent on two inputs. One is an early signal (ES) from neighbouring cells, which is yet to be identified. The other input is the maternal nuclearisation of b-catenin in the prospective mesendoderm cells, a consequence of which is activation of Wnt8, which results in the further nuclearisation of bcatenin. This positive feedback loop is inferred in the sea urchin based on the observation that a negative Wnt8 ligand leads to the blocking of mesendoderm specification, but it is

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

473

Fig. 5. VegT highlight (fertilisation—stage 10.25). The network is shown up to stage 10.25 and the targets of maternal VegT have been highlighted in the manner achievable on the website by selecting the VegT gene (Web Ref. 1). In conjunction with h-catenin, VegT activates Xnr5 and Xnr6 before the MBT, establishing Nodal signalling in the dorsal – vegetal region of the embryo. As development proceeds past the MBT, the remaining members of the Nodal family (with the exception of Xnr3 which does not require VegT activity) are activated along with Derriere and Sox17a/h. The Nodal family members subsequently establish autoregulatory loops and activate many downstream targets shared with VegT. By stage 9, a range of genes including members of the Mix/Bix family, FGFs and GATA factors are activated by VegT. This activation may be a consequence of maternal or zygotic VegT activity. Maternal depletion of VegT results in failed expression of the majority of these genes, but this may be a consequence of failure to activate Nodal signalling. Upstream activators and repressors of individual genes can also be viewed on our website and the evidence behind their inclusion in the network can be viewed.

yet to be confirmed by the identification of h-catenin/TCF binding sites in the Wnt8 promoter (Davidson et al., 2002b). Whilst maternal h-catenin has not been shown to activate Xwnt8 in Xenopus, the canonical Wnt signalling pathway leading to nuclearisation of h-catenin is conserved and clearly active early in both species. The later zygotic expression of h-catenin is coincident with the expression of Wnt8, leaving open the possibility that the b-catenin/ Wnt8 loop observed in the sea urchin has been retained zygotically in the ventrolateral region of the Xenopus embryo.

After the b-catenin/Wnt8 loop is established in sea urchin embryos, a second loop involving GataE, Krox and Otx is activated and ultimately replaces the loop established by bcatenin. This loop has recently been shown to be active in the distantly related starfish, Asterina miniata, although its upstream activators and downstream targets are not conserved (Hinman et al., 2003). In the sea urchin, GataE contributes to the activity of six downstream genes, including the forkheads FoxA and FoxB, Bra, and GataC. GataC and GataE are initially expressed in the endomesoderm, but subsequently in the mesoderm and endoderm, respectively

474

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

Table 1 A comparison of transcription factors and embryonic signals involved in Xenopus and sea urchin mesendoderm specification (see Web Ref. 2) Gene class

Xenopus

Sea urchin

Comment

Catenin Wnt

h-catenin Xwnt8

Nh-TCF Wnt8

GATA Forkhead/Winged Helix Orthodonticle

Gata 1 – 3 Gata 4 – 6 Xfkh1 Otx2

GataC GataE FoxA FoxB Otx

Krox

no equivalent identified

Krox

T-Box

Brachyury

Bra

Sox factors

Sox3

SoxB1

Mix class homeobox Nodal

Sox17 Mix.1 Mix.2 Mixer Bix1/ Mix.4 Bix2/Milk Bix3 Bix4 Xnr 1 – 6

no equivalent identified no equivalent identified

T-Box

VegT

no equivalent identified

Goosecoid

Goosecoid

Gsc

Maternally expressed in both species. The role of the Wnt signalling pathway is conserved between the two species. Later mesodermal. Later endodermal. Downstream of GataE in sea urchin but Nodals in Xenopus. Sea urchin Otx and Xenopus Otx2 maternally expressed, but Otx2 induces anterior neurectoderm. Direct target of Otx and h-catenin in the sea urchin and activates Wnt8 and Bra. Regulated by GATA in the sea urchin and mouse. Mesodermal in Xenopus and mouse, endodermal in sea urchin. Maternal in both species. SoxB1 inhibits endomesoderm and Sox3 represses Xnr5. Xenopus endodermal marker. Many homeobox proteins identified in sea urchin genome but no Mix homologues. Essential for mesendoderm in Xenopus. Activated by h-catenin and VegT, and repressed by Sox3. ES, the unidentified early signal for sea urchin endomesoderm, activated by h-catenin and inhibited by SoxB1. Sea urchin TBr is related to VegT, but has no maternal component to its expression. Dependent on h-catenin in both but via Siamois in Xenopus.

no equivalent identified— may be ES

(Pancer et al., 1999). In sequence alignments, GataC is most closely related to GATA1, 2 and 3 in vertebrates, which have retained major roles in the mesodermal derivative, blood (Fig. 6 and Lowry and Atchley, 2000). Likewise, GataE is most closely related to GATA4, 5 and 6, which have retained major roles in endoderm development (Fig. 6). The downstream targets of GataE that are shared between the two species include other GATA factors and homeobox proteins including HNF1-b (Weber et al., 2000). In contrast, although Otx homologues have been identified in Xenopus, and Xotx2 is maternally expressed, they are considered inducers of anterior neurectoderm (Kuroda et al., 2000; Pannese et al., 1995). Likewise, the most similar gene to sea urchin Krox in Xenopus is Xblimp-1, which appears to repress anterior mesendoderm (de Souza et al., 1999). Thus further experiments are required, possibly involving other members of the Krox and Otx families, to determine the extent to which the entire loop is conserved in Xenopus. Brachyury is present in both networks suggesting an ancestral role. Some evidence supports regulation by GATA factors in mouse as in sea urchin (Yamaguchi et al., 1999), but GATA factors have not yet been shown to induce Brachyury in Xenopus (Shoichet et al., 2000; Weber et al., 2000). Furthermore, the later expression of Brachyury in sea urchin and vertebrates appears to differ: in sea urchin, Brachyury is an endodermal gene, whereas in Xenopus and mouse, it is mesodermal. When Brachyury homologues from a variety of species, including sea urchin, were overexpressed in Xenopus presumptive ectoderm, all induced only mesoderm, with the exception of ascidian and Drosophila Brachyury, which also induced endoderm (Marcel-

lini et al., 2003). The endoderm-inducing activity was traced to the absence of an N-terminal repressor of endodermal gene expression. Clearly, an explanation for the presence of this domain in sea urchin Brachyury requires further study. Brachyury is involved in controlling convergent extension of the mesoderm during gastrulation in Xenopus (Conlon and Smith, 1999), and in the sea urchin, the endoderm is formed by invagination, which is driven by convergent extension. Thus, Brachyury appears to have a conserved role in the cell movements associated with gastrulation, which explains its conserved expression around the blastopore (Technau and Scholz, 2003). A member of the SoxB family, SoxB1, in sea urchin contributes to the repression of mesendodermal gene expression before induction (Kenny et al., 1999; Web Ref. 2). This gene is expressed maternally but its expression is downregulated in the micromeres that contribute to the mesendoderm (Penzel et al., 1997). The Xenopus homologue of SoxB1, Sox3, has been shown to negatively regulate Xnr5, which contributes to the Nodal signalling essential for mesendoderm induction (Zhang et al., 2003). It will be interesting to determine if Sox3 represses other mesendodermal genes in Xenopus before MBT. This would indicate a conserved role for maternal SoxB proteins in preventing premature or ectopic mesendoderm induction. Thus, several but not all key components of mesendoderm induction in sea urchins appear to be conserved in Xenopus. Similarly, there are components of mesendoderm induction in Xenopus apparently not conserved in sea urchin. For example, Mix and Nodal-related proteins have

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

475

Fig. 6. Sequence relationships between sea urchin and vertebrate GATA factors. Trees constructed by comparing the C-terminal zinc fingers of the Xenopus laevis GATA family, GataC and GataE from Strongylocentrotus purpuratus, and Saccharomyces cerevisiae GAT-1. (a) S. cerevisiae GAT-1 is a single-fingered GATA factor and was included here to root the tree. GataE (in purple) groups with xGATA-4, 5a and 5b whereas GataC (in red) groups within xGATA-2 and 3. (b) To further explore this grouping, an un-rooted tree using both zinc fingers of the X. laevis GATA family and GataC and GataE was constructed and subjected to bootstrap analysis (Kumar et al., 2001). This analysis strongly supports the contention that GataC and GataE belong to the GATA 1 – 3 and 4 – 6 families, respectively. Furthermore, GataC groups most closely with GATA 2 and 3, whereas GataE groups closely with GATA 4 and 5.

not yet been identified in the sea urchin. In addition to mesendoderm induction, Nodals are strongly associated with dorsoventral patterning. The ascidian is a protochordate and thus thought to be the most primitive example of dorsoventral patterning. The recently published genome of the ascidian, Ciona intestinalis, contains an orthologue of Nodal (gene id grail.490.7.1) and Nodal has been identified in the ascidian, Halocynthia roretzi (Dehal et al., 2002; Morokuma et al., 2002). No examples of Nodal have yet been identified before the protochordates. It is therefore possible that the involvement of Nodal signalling in mesendoderm induction in vertebrates is linked with its role in dorsoventral patterning. However, it is not inconceivable

that the early signal (ES) passed from the micromeres to immediate ancestors of mesendodermal cells in sea urchin will turn out to be a TGF-h family member related to Nodal. If so, the Nodal signal recruited along with h-catenin to facilitate the patterning of the dorsoventral axis was already a player in general mesendoderm specification. Consistent with the idea that sea urchin has a Nodal activity, the Xenopus orthologue of sea urchin SoxB1, which represses mesendoderm formation, represses Nodal-related gene expression (Zhang et al., 2003). However, although Mixer has been proposed to be a downstream effector of the Nodal signalling pathway in vertebrates (Kunwar et al., 2003; Trinh le et al., 2003), a Mix orthologue has not yet been

476

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

identified in sea urchins or ascidians, suggesting a distinct evolutionary origin for this family. Likewise, no maternally expressed T-box transcription factor has been described in the sea urchin with an equivalent role to Xenopus VegT in mesendoderm induction. The sea urchin T-box transcription factor, Tbrain (HpTb), which is closely related to Xenopus VegT, is detected as maternal protein in the cytoplasm, although morpholino experiments suggest its zygotic product is not involved in endoderm formation (Fuchikami et al., 2002). Similar to the Nodal family, a VegT homologue does exist in the ascidian C. intestinalis, CiVegTR, and is maternally expressed and localised to the vegetal pole (Erives and Levine, 2000). In contrast, the zebrafish homologue of VegT, spadetail, is not expressed maternally, although the VegT paralogue, eomesodermin, has recently been shown to be maternally expressed and to regulate nodal targets apparently via nodal signalling (Bruce et al., 2003). Thus, it is currently unclear whether there is an ancient role for a maternal T-box factor in establishing the mesendoderm or if the Xenopus recruitment of VegT is a derived characteristic.

predicted elements is currently a bottle neck in the acquisition of data, but the identification of interactions and their presentation in a relational format should nevertheless be of significant predictive use. Actual experimental outcomes can be compared with those predicted from the model, and the model can be adapted to fit new or contradictory data. Such an iterative process should lead to greater precision in modelling.

Acknowledgments We would like to thank Aaron Zorn, Boni Afouda and Aldo Ciau-Uitz for adding data to the database. Also thanks to Boni Afouda, Martin Gering, Andrew Johnson and Maggie Walmsley for helpful discussion and comments on the manuscript, and to Chris Wade for assistance with the phylogenetic analysis. Finally, help from all members of the lab in the development of this network has been greatly appreciated.

Web references Concluding remarks We have created a regulatory network that represents the interactions between transcription factors and embryonic signals involved in the specification of the mesendoderm during Xenopus development. Comparison with the equivalent network for the sea urchin reveals remarkable similarities but also families of molecules not yet detected in one or other organism, suggesting areas for future study. For example, conservation considerations suggest that a Nodal-like molecule will be identified in sea urchins and that further exploration of early roles for members of the Krox and Otx families in Xenopus might be productive. Alternatively, some of the differences in molecular programming of this germ layer may eventually be informative with respect to the evolution of these organisms. The network presented here does not include regulation at the level of chromatin structure or posttranslational modifications of transcription factors. However, even chromatin modifications ultimately need to be targeted to specific DNA sequences via sequence-specific DNA binding proteins, which could be included amongst the ‘transcription factors’ for the purpose of building into the network their relationships with the other regulators listed here. We expect the network to be continually refined on the World Wide Web in the light of criticisms, corrections and new data (Web Ref. 1). As new high throughput methods for assaying gene function and expression are developed, the number of genes that can be accurately measured in one experiment will increase. Predictions from comparative genome sequence analysis of the cis-regulatory elements that control the activation of individual genes are also proceeding apace. Functional testing of

http://www.nottingham.ac.uk/genetics/networks http://sugp.caltech.edu/endomes

References Agius, E., Oelgeschlager, M., Wessely, O., Kemp, C., De Robertis, E.M., 2000. Endodermal Nodal-related signals and mesoderm induction in Xenopus. Development 127, 1173 – 1183. Boggetti, B., Argenton, F., Haffter, P., Bianchi, M.E., Cotelli, F., Beltrame, M., 2000. Cloning and expression pattern of a zebrafish homolog of forkhead activin signal transducer (FAST), a transcription factor mediating Nodal-related signals. Mech. Dev. 99, 187 – 190. Bruce, A.E., Howley, C., Zhou, Y., Vickers, S.L., Silver, L.M., King, M.L., Ho, R.K., 2003. The maternally expressed zebrafish T-box gene eomesodermin regulates organizer formation. Development 130, 5503 – 5517. Cameron, R.A., Davidson, E.H., 1991. Cell type specification during sea urchin development. Trends Genet. 7, 212 – 218. Carnac, G., Kodjabachian, L., Gurdon, J.B., Lemaire, P., 1996. The homeobox gene Siamois is a target of the Wnt dorsalisation pathway and triggers organiser activity in the absence of mesoderm. Development 122, 3055 – 3065. Casey, E.S., Tada, M., Fairclough, L., Wylie, C.C., Heasman, J., Smith, J.C., 1999. Bix4 is activated directly by VegT and mediates endoderm formation in Xenopus development. Development 126, 4193 – 4200. Clements, D., Woodland, H.R., 2003. VegT induces endoderm by a selflimiting mechanism and by changing the competence of cells to respond to TGF-beta signals. Dev. Biol. 258, 454 – 463. Clements, D., Friday, R.V., Woodland, H.R., 1999. Mode of action of VegT in mesoderm and endoderm formation. Development 126, 4903 – 4911. Conlon, F.L., Smith, J.C., 1999. Interference with brachyury function inhibits convergent extension, causes apoptosis, and reveals separate requirements in the FGF and activin signalling pathways. Dev. Biol. 213, 85 – 100. Conlon, F.L., Fairclough, L., Price, B.M., Casey, E.S., Smith, J.C.,

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478 2001. Determinants of T box protein specificity. Development 128, 3749 – 3758. Davidson, E.H., Rast, J.P., Oliveri, P., Ransick, A., Calestani, C., Yuh, C.H., Minokawa, T., Amore, G., Hinman, V., Arenas-Mena, C., Otim, O., Brown, C.T., Livi, C.B., Lee, P.Y., Revilla, R., Rust, A.G., Pan, Z., Schilstra, M.J., Clarke, P.J., Arnone, M.I., Rowen, L., Cameron, R.A., McClay, D.R., Hood, L., Bolouri, H., 2002a. A genomic regulatory network for development. Science 295, 1669 – 1678. Davidson, E.H., Rast, J.P., Oliveri, P., Ransick, A., Calestani, C., Yuh, C.H., Minokawa, T., Amore, G., Hinman, V., Arenas-Mena, C., Otim, O., Brown, C.T., Livi, C.B., Lee, P.Y., Revilla, R., Schilstra, M.J., Clarke, P.J., Rust, A.G., Pan, Z., Arnone, M.I., Rowen, L., Cameron, R.A., McClay, D.R., Hood, L., Bolouri, H., 2002b. A provisional regulatory gene network for specification of endomesoderm in the sea urchin embryo. Dev. Biol. 246, 162 – 190. Dehal, P., Satou, Y., Campbell, R.K., Chapman, J., Degnan, B., De Tomaso, A., Davidson, B., Di Gregorio, A., Gelpke, M., Goodstein, D.M., Harafuji, N., Hastings, K.E., Ho, I., Hotta, K., Huang, W., Kawashima, T., Lemaire, P., Martinez, D., Meinertzhagen, I.A., Necula, S., Nonaka, M., Putnam, N., Rash, S., Saiga, H., Satake, M., Terry, A., Yamada, L., Wang, H.G., Awazu, S., Azumi, K., Boore, J., Branno, M., Chin-Bow, S., DeSantis, R., Doyle, S., Francino, P., Keys, D.N., Haga, S., Hayashi, H., Hino, K., Imai, K.S., Inaba, K., Kano, S., Kobayashi, K., Kobayashi, M., Lee, B.I., Makabe, K.W., Manohar, C., Matassi, G., Medina, M., Mochizuki, Y., Mount, S., Morishita, T., Miura, S., Nakayama, A., Nishizaka, S., Nomoto, H., Ohta, F., Oishi, K., Rigoutsos, I., Sano, M., Sasaki, A., Sasakura, Y., Shoguchi, E., Shin-i, T., Spagnuolo, A., Stainier, D., Suzuki, M.M., Tassy, O., Takatori, N., Tokuoka, M., Yagi, K., Yoshizaki, F., Wada, S., Zhang, C., Hyatt, P.D., Larimer, F., Detter, C., Doggett, N., Glavina, T., Hawkins, T., Richardson, P., Lucas, S., Kohara, Y., Levine, M., Satoh, N., Rokhsar, D.S., 2002. The draft genome of Ciona intestinalis: insights into chordate and vertebrate origins. Science 298, 2157 – 2167. de Souza, F.S., Gawantka, V., Gomez, A.P., Delius, H., Ang, S.L., Niehrs, C., 1999. The zinc finger gene Xblimp1 controls anterior endomesodermal cell fate in Spemann’s organizer. EMBO J. 18, 6062 – 6072. Ecochard, V., Cayrol, C., Rey, S., Foulquier, F., Caillol, D., Lemaire, P., Duprat, A.M., 1998. A novel Xenopus mix-like gene milk involved in the control of the endomesodermal fates. Development 125, 2577 – 2585. Engleka, M.J., Craig, E.J., Kessler, D.S., 2001. VegT activation of Sox17 at the midblastula transition alters the response to nodal signals in the vegetal endoderm domain. Dev. Biol. 237, 159 – 172. Erives, A., Levine, M., 2000. Characterization of a maternal T-Box gene in Ciona intestinalis. Dev. Biol. 225, 169 – 178. Fuchikami, T., Mitsunaga-Nakatsubo, K., Amemiya, S., Hosomi, T., Watanabe, T., Kurokawa, D., Kataoka, M., Harada, Y., Satoh, N., Kusunoki, S., Takata, K., Shimotori, T., Yamamoto, T., Sakamoto, N., Shimada, H., Akasaka, K., 2002. T-brain homologue (HpTb) is involved in the archenteron induction signals of micromere descendant cells in the sea urchin embryo. Development 129, 5205 – 5216. Hashimoto-Partyka, M.K., Yuge, M., Cho, K.W., 2003. Nodal signaling in Xenopus gastrulae is cell-autonomous and patterned by beta-catenin. Dev. Biol. 253, 125 – 138. Henry, G.L., Melton, D.A., 1998. Mixer, a homeobox gene required for endoderm development. Science 281, 91 – 96. Hinman, V.F., Nguyen, A.T., Cameron, R.A., Davidson, E.H., 2003. Developmental gene regulatory network architecture across 500 million years of echinoderm evolution. Proc. Natl. Acad. Sci. U. S. A. 100, 13356 – 13361. Jones, C.M., Kuehn, M.R., Hogan, B.L., Smith, J.C., Wright, C.V., 1995. Nodal-related signals induce axial mesoderm and dorsalize mesoderm during gastrulation. Development 121, 3651 – 3662. Joseph, E.M., Melton, D.A., 1997. Xnr4: a Xenopus nodal-related gene expressed in the Spemann organizer. Dev. Biol. 184, 367 – 372. Kenny, A.P., Kozlowski, D., Oleksyn, D.W., Angerer, L.M., Angerer, R.C., 1999. SpSoxB1, a maternally encoded transcription factor asymmetri-

477

cally distributed among early sea urchin blastomeres. Development 126, 5473 – 5483. Kimelman, D., Griffin, K.J., 2000. Vertebrate mesendoderm induction and patterning. Curr. Opin. Genet. Dev. 10, 350 – 356. Kofron, M., Demel, T., Xanthos, J., Lohr, J., Sun, B., Sive, H., Osada, S., Wright, C., Wylie, C., Heasman, J., 1999. Mesoderm induction in Xenopus is a zygotic event regulated by maternal VegT via TGFbeta growth factors. Development 126, 5759 – 5770. Kumar, S., Tamura, K., Jakobsen, I.B., Nei, M., 2001. MEGA2: molecular evolutionary genetics analysis software. Bioinformatics 17, 1244 – 1245. Kunwar, P.S., Zimmerman, S., Bennett, J.T., Chen, Y., Whitman, M., Schier, A.F., 2003. Mixer/Bon and FoxH1/Sur have overlapping and divergent roles in Nodal signaling and mesendoderm induction. Development 130, 5589 – 5599. Kuroda, H., Hayata, T., Eisaki, A., Asashima, M., 2000. Cloning a novel developmental regulating gene, Xotx5: its potential role in anterior formation in Xenopus laevis. Dev. Growth Differ. 42, 87 – 93. Latinkic, B.V., Smith, J.C., 1999. Goosecoid and mix.1 repress Brachyury expression and are required for head formation in Xenopus. Development 126, 1769 – 1779. Lemaire, P., Darras, S., Caillol, D., Kodjabachian, L., 1998. A role for the vegetally expressed Xenopus gene Mix.1 in endoderm formation and in the restriction of mesoderm to the marginal zone. Development 125, 2371 – 2380. Lowry, J.A., Atchley, W.R., 2000. Molecular evolution of the GATA family of transcription factors: conservation within the DNA-binding domain. J. Mol. Evol. 50, 103 – 115. Marcellini, S., Technau, U., Smith, J.C., Lemaire, P., 2003. Evolution of Brachyury proteins: identification of a novel regulatory domain conserved within Bilateria. Dev. Biol. 260, 352 – 361. Mead, P.E., Zhou, Y., Lustig, K.D., Huber, T.L., Kirschner, M.W., Zon, L.I., 1998. Cloning of Mix-related homeodomain proteins using fast retrieval of gel shift activities, (FROGS), a technique for the isolation of DNA-binding proteins. Proc. Natl. Acad. Sci. U. S. A. 95, 11251 – 11256. Morokuma, J., Ueno, M., Kawanishi, H., Saiga, H., Nishida, H., 2002. HrNodal, the ascidian nodal-related gene, is expressed in the left side of the epidermis, and lies upstream of HrPitx. Dev. Genes Evol. 212, 439 – 446. Nomura, M., Li, E., 1998. Smad2 role in mesoderm formation, left-right patterning and craniofacial development. Nature 393, 786 – 790. Onuma, Y., Takahashi, S., Yokota, C., Asashima, M., 2002. Multiple nodalrelated genes act coordinately in Xenopus embryogenesis. Dev. Biol. 241, 94 – 105. Pancer, Z., Rast, J.P., Davidson, E.H., 1999. Origins of immunity: transcription factors and homologues of effector genes of the vertebrate immune system expressed in sea urchin coelomocytes. Immunogenetics 49, 773 – 786. Pannese, M., Polo, C., Andreazzoli, M., Vignali, R., Kablar, B., Barsacchi, G., Boncinelli, E., 1995. The Xenopus homologue of Otx2 is a maternal homeobox gene that demarcates and specifies anterior body regions. Development 121, 707 – 720. Papin, C., van Grunsven, L.A., Verschueren, K., Huylebroeck, D., Smith, J.C., 2002. Dynamic regulation of Brachyury expression in the amphibian embryo by XSIP1. Mech. Dev. 111, 37 – 46. Patient, R.K., McGhee, J.D., 2002. The GATA family (vertebrates and invertebrates). Curr. Opin. Genet. Dev. 12, 416 – 422. Penzel, R., Oschwald, R., Chen, Y., Tacke, L., Grunz, H., 1997. Characterization and early embryonic expression of a neural specific transcription factor xSOX3 in Xenopus laevis. Int. J. Dev. Biol. 41, 667 – 677. Rex, M., Hilton, E., Old, R., 2002. Multiple interactions between maternally-activated signalling pathways control Xenopus nodal-related genes. Int. J. Dev. Biol. 46, 217 – 226. Rodaway, A., Patient, R., 2001. Mesendoderm. An ancient germ layer? Cell 105, 169 – 172. Roel, G., Hamilton, F.S., Gent, Y., Bain, A.A., Destree, O., Hoppler, S.,

478

M. Loose, R. Patient / Developmental Biology 271 (2004) 467–478

2002. Lef-1 and Tcf-3 transcription factors mediate tissue-specific Wnt signaling during Xenopus development. Curr. Biol. 12, 1941 – 1945. Rosa, F.M., 1989. Mix.1, a homeobox mRNA inducible by mesoderm inducers, is expressed mostly in the presumptive endodermal cells of Xenopus embryos. Cell 57, 965 – 974. Saka, Y., Tada, M., Smith, J.C., 2000. A screen for targets of the Xenopus T-box gene Xbra. Mech. Dev. 93, 27 – 39. Schohl, A., Fagotto, F., 2003. A role for maternal beta-catenin in early mesoderm induction in Xenopus. EMBO J. 22, 3303 – 3313. Schulte-Merker, S., Smith, J.C., 1995. Mesoderm formation in response to Brachyury requires FGF signalling. Curr. Biol. 5, 62 – 67. Shivdasani, R.A., 2002. Molecular regulation of vertebrate early endoderm development. Dev. Biol. 249, 191 – 203. Shoichet, S.A., Malik, T.H., Rothman, J.H., Shivdasani, R.A., 2000. Action of the Caenorhabditis elegans GATA factor END-1 in Xenopus suggests that similar mechanisms initiate endoderm development in ecdysozoa and vertebrates. Proc. Natl. Acad. Sci. U. S. A. 97, 4076 – 4081. Smith, J.C., White, R., 2003. Patterning the Xenopus embryo. In: Tickle, C. (Ed.), Patterning in Vertebrate Development. Oxford Univ. Press, Oxford, UK, pp. 22 – 47. Smith, W.C., McKendry, R., Ribisi Jr., S., Harland, R.M., 1995. A nodalrelated gene defines a physical and functional domain within the Spemann organizer. Cell 82, 37 – 46. Standley, H.J., Zorn, A.M., Gurdon, J.B., 2001. eFGF and its mode of action in the community effect during Xenopus myogenesis. Development 128, 1347 – 1357. Sun, B.I., Bush, S.M., Collins-Racie, L.A., LaVallie, E.R., DiBlasio-Smith, E.A., Wolfman, N.M., McCoy, J.M., Sive, H.L., 1999. derriere: a TGFbeta family member required for posterior development in Xenopus. Development 126, 1467 – 1482. Tada, M., Casey, E.S., Fairclough, L., Smith, J.C., 1998. Bix1, a direct target of Xenopus T-box genes, causes formation of ventral mesoderm and endoderm. Development 125, 3997 – 4006. Takahashi, S., Yokota, C., Takano, K., Tanegashima, K., Onuma, Y., Goto, J., Asashima, M., 2000. Two novel nodal-related genes initiate early inductive events in Xenopus Nieuwkoop center. Development 127, 5319 – 5329.

Technau, U., Scholz, C.B., 2003. Origin and evolution of endoderm and mesoderm. Int. J. Dev. Biol. 47, 531 – 539. Trinh le, A., Meyer, D., Stainier, D.Y., 2003. The Mix family homeodomain gene bonnie and clyde functions with other components of the Nodal signaling pathway to regulate neural patterning in zebrafish. Development 130, 4989 – 4998. Vize, P.D., 1996. DNA sequences mediating the transcriptional response of the Mix.2 homeobox gene to mesoderm induction. Dev. Biol. 177, 226 – 231. Weber, H., Symes, C.E., Walmsley, M.E., Rodaway, A.R., Patient, R.K., 2000. A role for GATA5 in Xenopus endoderm specification. Development 127, 4345 – 4360. White, R.J., Sun, B.I., Sive, H.L., Smith, J.C., 2002. Direct and indirect regulation of derriere, a Xenopus mesoderm-inducing factor, by VegT. Development 129, 4867 – 4876. Xanthos, J.B., Kofron, M., Wylie, C., Heasman, J., 2001. Maternal VegT is the initiator of a molecular network specifying endoderm in Xenopus laevis. Development 128, 167 – 180. Xanthos, J.B., Kofron, M., Tao, Q., Schaible, K., Wylie, C., Heasman, J., 2002. The roles of three signaling pathways in the formation and function of the Spemann Organizer. Development 129, 4027 – 4043. Yamaguchi, H., Tanaka, K., Kitagawa, Y., Miki, K., 1999. A PEA3 site flanked by SP1, SP4, and GATA sites positively regulates the differentiation-dependent expression of Brachyury in embryonal carcinoma P19 cells. Biochem. Biophys. Res. Commun. 254, 542 – 547. Yang, J., Tan, C., Darken, R.S., Wilson, P.A., Klein, P.S., 2002. Betacatenin/Tcf-regulated transcription prior to the midblastula transition. Development 129, 5743 – 5752. Yokota, C., Kofron, M., Zuck, M., Houston, D.W., Isaacs, H., Asashima, M., Wylie, C.C., Heasman, J., 2003. A novel role for a nodal-related protein; Xnr3 regulates convergent extension movements via the FGF receptor. Development 130, 2199 – 2212. Zhang, J., Houston, D.W., King, M.L., Payne, C., Wylie, C., Heasman, J., 1998. The role of maternal VegT in establishing the primary germ layers in Xenopus embryos. Cell 94, 515 – 524. Zhang, C., Basta, T., Jensen, E.D., Klymkowsky, M.W., 2003. The {beta}catenin/VegT-regulated early zygotic gene Xnr5 is a direct target of SOX3 regulation. Development 130, 5609 – 5624.