Essential Role of the Transcription Factor Ets-2 inXenopus Early Development*

The fibroblast growth factor (FGF)/MAPK pathway plays an important role in early Xenopus developmental processes, including mesoderm patterning. The activation of the MAPK pathway leads to induction of Xenopus Brachyury (Xbra), which regulates the transcription of downstream mesoderm-specific genes in mesoderm patterning. However, the link between the FGF/MAPK pathway and the induction of Xbra has not been fully understood. Here we present evidence suggesting that Ets-2 is involved in the induction of Xbra and thus in the development of posterior mesoderm during early embryonic development. Overexpression of Ets-2 caused posteriorized embryos and led to the induction of mesoderm in ectodermal explants. Expression of a dominant-negative form of Ets-2 or injection of antisense morpholino oligonucleotides against Ets-2 inhibited the formation of the trunk and tail structures. Overexpression of Ets-2 resulted in the induction of Xbra, and expression of the dominant-negative Ets-2 inhibited FGF- or constitutively active MEK-induced Xbra expression. Moreover, overexpression of Ets-2 up-regulated the transcription from Xbra promoter reporter gene constructs. Ets-2 bound to the Xbra promoter region in vitro. These results taken together indicate thatXenopus Ets-2 plays an essential role in mesoderm patterning, lying between the FGF/MAPK pathway and the Xbra transcription.

During Xenopus early embryonic development, mesoderm arises from ectoderm by induction that requires signals from the vegetal hemisphere of the embryo (1,2). The FGF 1 /MAP kinase (MAPK) pathway has been shown to be involved in early mesodermal patterning (3)(4)(5)(6)(7)(8). One of the genes that are thought to be regulated directly by FGF via the MAPK signal transduction pathway is Xenopus Brachyury (Xbra). Brachyury is an important regulatory gene in early vertebrate development (9 -14). Loss of Brachyury function in mouse, zebrafish, and Xenopus embryos causes defects in posterior mesoderm and notochord differentiation (9,(15)(16)(17). However, the link between the FGF/MAPK pathway and the induction of Xbra expression has not been fully defined.
The ETS family of transcription factors, comprising more than 30 different members, has been found to play a crucial role in controlling transcription of a variety of genes involved in important cellular processes, such as proliferation and differentiation (19,20). They share a unique DNA binding domain, the ETS domain, which interacts specifically with GGA(A/T)based recognition sites (18). As targets of the Ras-MAPK signaling pathway, Ets transcription factors are phosphorylated by MAPK and function as critical nuclear integrators of ubiquitous signaling cascades. The ETS family is divided into subfamilies by sequence similarity based on the ETS domain or additional sequence motifs. Ets-2 is a member of the ETS subfamily, which consists of three members: Ets-1, Ets-2, and Drosophila Pointed (21,22). Studies with mammalian cultured cells have shown that Ets-2 is activated by MAPK-dependent phosphorylation of threonine 72 in an N-terminal regulatory domain (the Pointed domain) (21,23). Xenopus Ets-2 is maternally expressed in both the animal pole and the intermediate zone (24,25). On the basis of these results, we hypothesized that Ets-2 relays the FGF/MAPK signaling and induces mesoderm by inducing Xbra gene transcription. In this study, we have presented several lines of evidence indicating that Ets-2 plays an essential role in mesodermal patterning, lying between the FGF/MAPK pathway and the Xbra transcription.

EXPERIMENTAL PROCEDURES
Plasmid Construction-Xenopus Ets-2 cDNA was obtained by screening a ZAP II cDNA library made from stage 10.5 embryos. To construct CS2-En-N, the 888-bp fragment coding for amino acids 1-294 of Drosophila engrailed protein was inserted into the StuI site of CS2 ϩ . To generate Ets⌬N En-R (the repressor domain of Drosophila Engrailed), the DNA binding domain of Ets-2 (amino acid residues 310 -468) was subcloned into CS2-En-N. In vitro synthesis of capped mRNA was performed using mMESSAGE mMACHINE kit (Ambion) according to the manufacturer's instructions.
Embryo Manipulation and Animal Cap Assay-Embryos were in vitro fertilized, dejellied, and cultured in 0.1ϫ modified Barth's saline (1.5 mM HEPES, pH 7.4, 8.8 mM NaCl, 0.1 mM KCl, 0.24 mM NaHCO 3 , 0.082 mM MgSO 4 , 0.03 mM Ca(NO 3 ) 2 , and 0.041 mM CaCl 2 ). Embryos were staged according to Nieuwkoop and Faber (26). Embryos at the four-cell stage were injected with mRNA as described in the text and figure legends. Animal caps were dissected from the injected embryos at stages 8 -8.5 and cultured in 1ϫ Steinberg solution containing 0.1% bovine serum albumin to various stages for further analysis as described in the figure legends. RT-PCR experiments were performed according to standard protocol. The primer pairs used here for RT-PCR have been described elsewhere (27,28). For morpholino oligonucleotide injections, an Ets-2 antisense oligonucleotide with the sequence 5Ј-AGCTGAGGGAGGGTATGTCCTTCC-3Ј was obtained from Gene Tools, LLC. Oligonucleotides were resuspended in sterile, filtered water and injected into four-cell stage embryos with the indicated amounts (29). * This work was supported by grants from the Ministry of Education, Science and Culture of Japan (to E. N.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Luciferase Assay-Embryos were injected with 200 pg of Xbra-pOLUC (provided by Dr. K. W. Y. Cho) and 100 pg of pCMV-␤-galactosidase together with 1 ng of Ets-2 mRNA into animal poles at the two-cell stage. Dissected animal caps were assayed for luciferase and ␤-galactosidase activities at stage 11 (13,14). The ratio of luciferase to ␤-galactosidase activity provides a normalized measure of luciferase expression. In all cases, fold activation was calculated using the results from only Xbra pOLUC-and CMV-␤-galactosidase-injected embryos as background. Each experiment was performed three times to ensure reproducibility of results.
Gel Mobility Shift Assay-To obtain the recombinant GST-Ets-2 protein, the entire coding region of Ets-2 was subcloned into pGEX-6P. Production of GST-Ets-2 and GST protein was performed as described (30). A DNA sequence for oligonucleotides of the Ets binding site was 5Ј-CAGGTGTCAGTTCTTACTGGATGTAAGTTTATTGAAGGCA-3Ј. A gel mobility shift assay was carried out as described (31). For the competition assay, the Ets binding site was mutated using a sitedirected mutagenesis kit (QuikChange TM ; Stratagene. The Ets binding site was mutated using forward primer 5Ј-CAGGTGTCAGTTCTTACT-GGCCGTAAGTTTATTGAAGGC-3Ј and reverse primer 5Ј-TGCCT-TCAATAAAC TTACGGCCAGTAAGAACTGACACCTG-3Ј.

Ets-2 Is Required for Mesoderm Patterning in FGF Signaling-We first examined the effect of overexpression of wild type
Xenopus Ets-2 mRNA on early embryonic development. Ets-2 mRNA was injected into the dorsal or ventral marginal zones of four-cell stage Xenopus embryos. When compared with normal uninjected control embryos at tadpole stage 35, embryos with dorsal injections of Ets-2 showed anterior truncation and shortened body axis. The embryos injected with relatively high doses of Ets-2 mRNA (1-2 ng) lacked eyes and cement glands (Fig.  1A). About 70% of embryos injected with 2 ng of Ets-2 mRNA showed the severe anterior defects (n ϭ 34). A typical image is shown in Fig. 1A. About 40% of embryos injected with 1 ng of Ets-2 mRNA showed similar severe defects (n ϭ 45). A typical phenotype is shown (Fig. 1A, Ets-2 1 ng DMZ (dorsal marginal zone)). Moreover, injections with high doses of mRNA sometimes resulted in embryos with tail-like protrusions (data not shown). The ventral injections had little or no effect on the development of the embryos (Fig. 1A, Ets-2 1 ng VMZ (ventral marginal zone)). To investigate these embryos in more detail, we examined the wide range of molecular markers in Ets-2 mRNA-injected embryos. The results of the RT-PCR analysis carried out on stages 26 and 37 whole embryos injected with Ets-2 mRNA of 1 ng are shown (Fig. 1B). Sibling embryos were used as control. Overexpression of Ets-2 mRNA resulted in remarkable reduction of expression of a forebrain marker Otx2 and a cement gland marker XAG. Moreover, expression of a pan-neural marker neural cell adhesion molecule was also reduced by injection of Ets-2 mRNA. Thus, Ets-2 overexpression suppressed expression of anterior markers. Moreover, Ets-2 overexpression increased expression of the posterior markers such as Xbra and Xcad3 (Fig. 1B). A transverse section through the anterior of the posteriorized embryo showed no development of brain ventricles (data not shown). These data suggest that overexpression of Ets-2 causes anterior truncation and induction of posterior mesoderm. Next, to examine whether Ets-2 functions downstream of FGF signaling, we tested whether expression of Ets-2 could rescue the defects caused by inhibition of FGF signal in Xenopus embryos. Dominant-negative FGF receptor (XFD) mRNA with or without Ets-2 mRNA was injected into dorsal marginal zones of the four-cell stage. Overexpression of XFD caused a severe posterior defect (18 out of 18, Fig. 2). This defect is thought to result from inhibition of both gastrulation movement and posterior mesoderm formation (32). Co-injection of Ets-2 rescued this morphological defect, and the injected embryos developed almost normally (22 out of 47) (Fig. 2). This rescue of the XFD-induced phenotype by Ets-2 suggests that Ets-2 functions in mesoderm patterning downstream of FGFs.
To confirm the effects of Ets-2, we analyzed various markers. Animal caps were dissected from Ets-2 mRNA-injected embryos and cultured to stage 10.5 or 22. In stage 10.5 animal cap explants, Ets-2 induced the expression of the pan-mesodermal marker Xbra in a dose-dependent manner (Fig. 3A). The dorsal mesoderm marker Goosecoid and the ventral mesoderm marker Xwnt8 were not induced by Ets-2 overexpression (data not shown). In stage 22 animal caps, Ets-2 strongly induced the expression of the posterior markers Xcad3 and Xhox3 (Fig. 3B). The expression of Xlhbox6 (HoxB9), which is expressed in lateral mesoderm and spinal chord, was little induced by the injection of Ets-2 mRNA.
Inhibition of Ets-2 Causes Posterior Mesoderm Defects in Embryos-To examine whether Ets-2 is necessary for mesoderm patterning, we generated a dominant-negative form of Ets-2. The schematic diagram of the Ets-2-based construct used here was shown (Fig. 4A). Ets-2 possesses a highly conserved N-terminal regulatory domain (the Pointed domain) and a C-terminal motif that comprises the DNA binding domain (the ETS domain). Ets⌬N En-R was made consisting of the C-terminal region of Ets-2, which contains the DNA binding domain but lacks the activation function, fused to the transcription En-R. We examined the effect of overexpression of Ets⌬N En-R (Fig. 4B). Ets⌬N En-R mRNA was injected into dorsal marginal zones of four-cell stage embryos, and these embryos were cultured until stage 34. A typical image is shown in Fig. 4B. Ets⌬N En-R-injected embryos showed the severe posterior defect (47 out of 52). Body axis truncation and dorsal bending of embryos were also observed. These features of the injected embryos were similar to those of the XFD-injected embryos (Fig. 2). Embryos injected ventrally developed almost normally (Fig. 4B).
We then examined whether Ets⌬N En-R is able to block FGF-or constitutively active MAPKK (MAPKK SESE)-induced expression of Xbra (Fig. 4C). MAPKK (MEK) is a specific activator of classical MAPK (ERK MAPK) (33)(34)(35)(36). Although uninjected animal caps with FGF treatment expressed a high level of Xbra, Ets⌬N En-R-injected animal caps with FGF treatment did not induce Xbra expression. Although injection of active MAPKK alone induced Xbra expression strongly, coinjection of Ets⌬N En-R resulted in great reduction of Xbra expression. These results suggest that the FGF/MAPK pathway induces Xbra expression via Ets-2.
To test the specificity of the Ets⌬N En-R construct, Ets⌬N En-R mRNA was injected with wild type Ets-2 mRNA, and expression of Xbra was analyzed by RT-PCR with whole embryos (Fig. 4D). Ets⌬N En-R-injected whole embryos showed greatly reduced expression of Xbra when compared with uninjected whole embryos. The wild type Ets-2 co-injected embryos showed recovered expression of Xbra, indicating that wild type Ets-2 rescued the effect brought by Ets⌬N En-R.
Next, we tested the effect of expression of antisense morpholino oligonucleotides (MO) against Ets-2. We injected the Ets-2 MO into dorsal marginal zones in four-cell stage embryos. About 80% of embryos injected with the Ets-2 MO in dorsal sides showed a severe posterior defect (n ϭ 42). A typical phenotype is shown in Fig. 5. Expression of control MO in dorsal sides of embryos had no effect (Fig. 5). Embryos injected with the Ets-2 MO in ventral sides were almost normal (data not shown). To confirm that the Ets-2 MO-induced phenotype is specifically caused by blocking the Ets-2 function, we co-injected Ets-2 mRNA with Ets-2 MO. Co-injection of Ets-2 mRNA rescued about 60% of embryos from the defects in posterior structures induced by Ets-2 MO (n ϭ 28). A typical image is shown (Fig. 5, lower panel). Taken together, these results suggest that Ets-2 is necessary for patterning of posterior mesoderm.
Xbra Transcription Is Regulated by Ets-2-Our RT-PCR analysis showed that expression of Ets-2 is able to induce at the two-cell stage and were cultured until sibling embryos reached stage 11. On FGF treatment, animal caps were cultured in the medium including 50 ng/ml bFGF. Expression of Xbra was analyzed by RT-PCR. D, wild type Ets-2 rescued the reduction of Xbra expression caused by Ets⌬N En-R. Ets⌬N En-R mRNA (50 pg) was injected together with wild type Ets-2 mRNA (1 ng) into marginal zones of the two-cell embryo. Injected embryos were cultured until sibling embryos reached stage 11. Expression of Xbra was analyzed by RT-PCR with injected whole embryos. expression of Xbra. We supposed that Ets-2 regulates the Xbra transcription. To test the role of Ets-2 in regulation of Xbra transcription, we performed a reporter assay using the luciferase reporter plasmid containing the 1562-bp fragment of the Xbra promoter region (13,14). The Xbra reporter construct was co-injected with Ets-2 mRNA and the internal control ␤-galactosidase into animal poles of two-cell stage embryos. Animal caps were dissected at stage 8 and assayed for luciferase and ␤-galactosidase activities at stage 11. The ratio of luciferase to ␤-galactosidase activity provides a normalized measure of luciferase expression. Ets-2 activated luciferase expression about 4-fold relative to control (Fig. 6A). With FGF treatment, the luciferase activity by Ets-2 was slightly increased. On the contrary, Ets⌬N En-R repressed the luciferase activity markedly. Similarly, Ets⌬N En-R strongly inhibited the FGF-induced luciferase activity (Fig. 6A). These results are in good accordance with our hypothesis that Ets-2 regulates the transcription of Xbra. Then, we investigated whether Ets-2 protein binds to the promoter region of Xbra. A recent study has shown that a restricted upstream region of the Xbra promoter is necessary for its expression (37). We searched for the Ets binding sites within this region of Xbra promoter. We found two Ets binding motifs in the Xbra promoter and tested their ability to bind to Ets-2 in the gel mobility shift assay. The gel mobility shift assay was performed on Ϫ310 to Ϫ271 and Ϫ259 to Ϫ219 fragments of the Xbra promoter using the bacterially expressed glutathione S-transferase (GST)-Ets-2 fusion protein or GST alone. The Ϫ259/Ϫ219 fragment bound only very weakly to GST-Ets-2 (data not shown), whereas the Ϫ310/Ϫ271 region of the Xbra promoter bound to Ets-2 especially (Fig. 6C). Oligonucleotide probes corresponding to the wild type and the mutated Ϫ310/Ϫ271 region were designed for the gel mobility shift assay (Fig. 6B). The GST-Ets-2 fusion protein bound to the wild type oligonucleotide probe strongly, whereas GST alone did not bind to the probe (Fig. 6C). The binding was inhibited by preincubation with an excess of unlabeled oligonucleotide (Fig.  6C, competitor). However, the binding persisted with an excess of unlabeled mutant oligonucleotide (Fig. 6C, mutant competitor). Thus, Ets-2 is functional in forming a specific DNA-protein complex with a Xbra promoter region. DISCUSSION It is known that FGFs can induce mesoderm via the Ras/ MAPK pathway in early Xenopus development (3)(4)(5)(6)(7)(8). Previous studies have shown that members of at least six subfamilies of ETS proteins (ETS, YAN, ELG, PEA3, ERF, TCF) are nuclear targets of the Ras/MAPK pathway. Elk-1 is a member of the ternary complex factor subfamily, and it is well known that phosphorylation of Elk-1 by MAPK enhances activities of Elk-1. However, the FGF-induced Xbra expression was not reduced by overexpression of dominant-negative Elk-1 in animal cap assay (38). The other member of ETS family, ER81, which belongs to the PEA3 subfamily, was identified in Xenopus. Although XER81 has been reported to be a target of FGF signaling, XER81 alone did not induce Xbra expression in animal cap explants (39). Moreover, overexpression of XER81 did not change the expression pattern of Xbra transcript (40). Our results here have demonstrated that Ets-2 induces Xbra expression and plays an essential role in mesoderm patterning downstream of the FGF/MAPK pathway. Furthermore, overexpression of Ets-2 caused posteriorized embryos. Similar phenotypes were reported to be induced by expressing Xcad3 that is required for posterior development downstream of FGF signaling (41). Both Ets-2 and Xcad3 appear to posteriorize anterior neural tissue. How Ets-2 cooperates with Xcad3 and the Hox gene pathway in posterior development remains to be studied (42). FIG. 6. Ets-2 is essential for Xbra transcription. A, Ets-2-regulated expression of reporter gene construct containing a 1.5-kb fragment of the Xbra promoter region. The Xbra promoter construct was co-injected with Ets-2 mRNA or Ets⌬N En-R mRNA into animal poles of two-cell stage embryos. Animal caps were dissected at stage 8 and cultured with or without FGF (50 ng/ml). Cultured animal caps were assayed for luciferase activity at stage 11. B, the sequence of the wild type and the mutant Ets binding sites. C, Ets-2 bound to the Ϫ310/Ϫ271 region of the Xbra promoter. The gel mobility shift assay was performed using radiolabeled double-strand oligonucleotide probe of the Ets-binding site. After incubation of the radiolabeled probe with protein extracts, DNA-protein complex was analyzed by autoradiography following electrophoresis of binding reactions on 4% polyacrylamide gels. The upper arrow indicates the position of DNA-protein complex. The recombinant GST or GST-Ets-2 fusion protein was incubated with radiolabeled oligonucleotide probe containing the Ets-binding site. For competition assay, binding reactions were preincubated with a 200-fold molar excess of unlabeled oligonucleotide probe as competitor, and then binding reactions with GST-Ets-2 and labeled probe were incubated as described above. The same competition assay was performed with unlabeled mutant oligonucleotide.
In Xenopus mesoderm patterning, no transcription factor that would relay the FGF signal at the Xbra promoter has been identified (43). Our results suggest that Ets-2 regulates the transcription of Xbra downstream of the FGF/MAPK pathway. Mullick et al. (44) have recently shown that Ets-2 protein binds to a "weak" Ets-like site of the cytochrome P-450c27 promoter. They provide new insights on the role of putative weak consensus Ets sites in transcription activation, possibly through synergistic interaction with other gene-specific transcription activators (44). Because the putative weak Ets consensus sites are widely distributed on the Xbra promoter, it is possible that interactions with several Ets-like sites synergistically regulate Xbra transcription. Furthermore, other transcription factors, such as AP-1, potentially participate in transcriptional regulation of Xbra (45). It is likely that other transcription factors in coordination with Ets-2 regulate Xbra transcription. Elucidation of the precise relationship of Ets-2 with other transcription factors will be a necessary step toward understanding the regulation of Xbra.
Our results reported here strongly suggest that Ets-2 plays an essential role in mesoderm patterning. Previously, HpEts has been identified as a sea urchin homologue of Ets-2. Overexpression of HpEts in sea urchin embryos caused primary mesenchyme cells to extinguish cellular adhesion and to migrate (46,47). In Xenopus, it is likely that Ets-2 is also involved in cellular adhesion and migratory cell guidance. We are currently investigating other roles of Ets-2 in Xenopus developmental processes.