XSIP1 is essential for early neural gene expression and neural differentiation by suppression of BMP signaling

Transcription

1 Developmental Biology 275 (2004) XSIP1 is essential for early neural gene expression and neural differentiation by suppression of BMP signaling Kazuhiro R. Nitta a, Kousuke Tanegashima b,1, Shuji Takahashi b, Makoto Asashima a,b,c, * a Department of Biological Sciences, Graduate School of Science, The University of Tokyo, Tokyo , Japan b Department of Life Sciences (Biology), Graduate School of Arts and Sciences, The University of Tokyo, Tokyo , Japan c ICORP, International Cooperative Research Project (JST), Japan Received for publication 1 March 2004, revised 7 August 2004, accepted 9 August 2004 Available online 2 September 2004 Abstract Neural differentiation is induced by inhibition of BMP signaling. Secreted inhibitors of BMP such as Chordin from the Spemann organizer contribute to the initial step of neural induction. Xenopus Smad-interacting protein-1 gene (XSIP1) is expressed in neuroectoderm from the early gastrula stage through to the neurula stage. XSIP1 is able to inhibit BMP signaling and overexpression of XSIP1 induces neural differentiation. To clarify the function of XSIP1 in neural differentiation, we performed a loss-of-function study of XSIP1. Knockdown of XSIP1 inhibited SoxD expression and neural differentiation. These results indicate that XSIP1 is essential for neural induction. Furthermore, loss-of-function experiments showed that SoxD is essential for XSIP1 transcription and for neural differentiation. However, inhibition of XSIP1 translation prevented neural differentiation induced by SoxD; thus, SoxD was not sufficient to mediate neural differentiation. Expression of XSIP1 was also required for inhibition of BMP signaling. Together, these results suggest that XSIP1 and SoxD interdependently function to maintain neural differentiation. D 2004 Elsevier Inc. All rights reserved. Keywords: XSIP1; Xenopus; Neural induction; def1/zfh family; SoxD; BMP; Zinc finger Introduction Classical experiments have shown that the dorsal lip region of the gastrula, called the bspemann organizerq, is a signaling source of neural induction and induces secondary neural tissue in host embryos when it is transplanted into the ventral side of embryos (Spemann and Mangold, 1924). In the last decade, molecular studies have revealed that these neural-inducing signals are secreted proteins such as Chordin (Sasai et al., 1995), Noggin (Lamb et al., 1993), * Corresponding author. Department of Life Sciences (Biology), The University of Tokyo, Komaba, Meguro, Tokyo , Japan. Fax: address: asashi@bio.c.u-tokyo.ac.jp (M. Asashima). 1 Current address: Laboratory of Molecular Genetics, National Institute of Child Health and Human Development, National Institute of Health, Bethesda, MD 20892, USA. Follistatin (Hemmati-Brivanlou et al., 1994), and Xnr3 (Haramoto et al., 2004; Smith et al., 1995; Yokota et al., 2003), all of which directly bind to bone morphogenetic protein (BMP). Attenuation of BMP signaling up-regulates the expression of early neural genes such as SoxD, Zic1-3, and XlPOU2, and induces neural tissue during embryogenesis (Fainsod et al., 1997; Matsuo-Takasaki et al., 1999; Mizuseki et al., 1998; Nakata et al., 1997, 1998; Piccolo et al., 1996; Sasai et al., 1995; Wilson and Hemmati- Brivanlou, 1995; Witta et al., 1995; Zimmerman et al., 1996). Binding of BMPs to their receptors induces phosphorylation of Smad1, Smad5, or Smad8, which are intracellular transducers of BMP signaling (Massagué and Wotton, 2000). The phosphorylated Smads form complexes with Smad4, then translocate to the nucleus and activate their target genes. Recently, several Smad-interacting proteins have been identified and shown to also regulate target genes /$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi: /j.ydbio

2 K.R. Nitta et al. / Developmental Biology 275 (2004) together with the Smad complex. One such vertebrate zinc finger protein, OAZ, is known to complex with BMPspecific Smads and mediate transcription (Hata et al., 2000). Another zinc finger protein, Drosophila Schnurri (Shn), directly interacts with Mad (Drosophila Smad homolog) and the resultant complex transduces Dpp signaling (Dai et al., 2000; Udagawa et al., 2000). Smad-interacting protein-1 (SIP1) was isolated in mouse by a yeast two-hybrid screening as a Smad1-interacting protein (Verschueren et al., 1999). SIP1 is a member of the def1/zfh family, which is comprised of transcriptional repressors containing a C2H2-type zinc finger. SIP1 homolog genes have been cloned in mouse, Xenopus, human, and chick (Cacheux et al., 2001; Eisaki et al., 2000; Tylzanowski et al., 2003; van Grunsven et al., 2000; Verschueren et al., 1999). In human, a SIP1 mutation leads to in Hirschsprung disease, also called aganglionic megacolon, which causes microcephaly, hypertelorism, convergent strabismus, and a wide nasal bridge (Cacheux et al., 2001; Wakamatsu et al., 2001; Yamada et al., 2001). SIP1-deficient mice also show defects in neural development (Higashi et al., 2002; Van de Putte et al., 2003). In Xenopus, XSIP1 was first isolated from animal caps treated with activin (Eisaki et al., 2000). XSIP1 is initially expressed in neuroectoderm at the gastrula stage and this expression is maintained in neural tissue throughout development. Overexpression of XSIP1 induces neural markers in animal caps and causes a hyperneuralized phenotype (Eisaki et al., 2000). Since a previous study showed that XSIP1 binds to XSmad1 as does mouse SIP1 (van Grunsven et al., 2000; Verschueren et al., 1999), it is likely that XSIP1 might inhibit BMP signaling to promote neural cell fate in the neuroectoderm. Furthermore E- cadherin, which is an epidermal marker, is directly repressed by XSIP1 (Comijn et al., 2001), suggesting that XSIP1 may not only inhibit the expression of BMPresponsive genes but also directly repress epidermal genes to maintain the neural cell fate. Another candidate for regulation by XSIP1 is Xenopus brachyury (Xbra), which has an important role in mesodermal formation. The Xbra promoter has CACCT-containing sequences to which the protein products of the def1/ ZFH family, including XSIP1, are known to bind. Mutation of this sequence in the Xbra promoter causes the expansion of reporter gene expression in the transgenic frog (Lerchner et al., 2000; Smith, 2001). Moreover, ectopic expression of XSIP1 in mesoderm causes down-regulation of Xbra expression (Papin et al., 2002). These results suggest that XSIP1 regulates the Xbra expression pattern. Recently, a similar result showed that SIP1 negatively regulates brachyury expression in chick (Sheng et al., 2003). Expression of SIP1 can be induced by Churchill, which is expressed in the neural region and down-regulates the mesodermal genes, brachyury and Tbx6. This suppression of mesodermal genes is dependent on SIP1, suggesting that the neural mesodermal boundary is established by SIP1 repressing mesodermal gene expression. Although it is thought that XSIP1 regulates the neural mesodermal boundary, it is still unknown how XSIP1 regulates neural induction in vivo. To examine the function of XSIP1 in the embryo, we used a loss-of-function approach employing antisense morpholino oligonucleotides. In this paper, we report that XSIP1 expression is induced by early neural genes and that this expression is essential for early neural induction. Materials and methods Embryos Xenopus laevis embryos were obtained by artificial fertilization and were cultured in 10% Steinberg s solution (SS) at 208C. They were staged according to Nieuwkoop and Faber (1956). Plasmid constructs The plasmid pcs2-rxsip1 was created by the polymerase chain reaction (PCR). Primers used are as follows: 5V-AAAGGATCCATGAAACAGGAGATCATGGCG- GATGG-3V 5V-TTCATCTTCCTCTTCGTCTCTTGGCA-3V The underline indicates a BamHI site and the following ATG is the start codon. G Y A and A Y G mutations are indicated in bold. PCR fragments were digested by BamHI and SphI and cloned into the equivalent sites of pcs2- XSIP1 (Eisaki et al., 2000). The plasmid pcs2-soxd was created as follows. SoxD inserts were digested from pbluescript(sk )-SoxD (Yabe et al., 2003) by BamHI and XhoI and cloned into the equivalent sites of pcs2+ vector. The plasmid pcs2-soxd-orf was created by PCR. The plasmid pcs2-zic3 was constructed as follows. Zic3 inserts were digested from pbluescript(sk )-Zic3 (Yabe et al., 2003) by EcoRI and cloned into the equivalent sites of pcs2+ vector. Myc-epitope-tagged constructs were made using the pcs2-mt vector. The pcs2-ha vector was constructed by inserting a sequence that codes for the HA-epitope dypydvpdya*t into the StuI/XhoI sites of the pcs2+ vector. HA-epitope-tagged constructs were made using the pcs2-ha vector. All tagged constructs were made with C- terminal tags. While pcs2-soxd-ha(utr+) encodes 5V UTR, pcs2-soxd-ha(utr ) does not encode this sequence. Morpholino-oligonucleotides The antisense oligonucleotides used were 25-mer morpholino-oligonucleotides (MO) (Gene Tools LLC) with the nucleotide sequence; 5V-CTTGCTTCATTGATAA-

3 260 K.R. Nitta et al. / Developmental Biology 275 (2004) GAGTGGGAT-3V for the XSIP1 MO and 5V-AGGA- GAAGCCTGGGTCCAACTGATC -3V for the SoxD MO. The control MO used was the standard control as supplied by Gene Tools LLC. Oligonucleotides were resuspended in filtered water and injected in maximal doses of 40 ng per embryo. Microinjection Microinjection was performed in 100% SS containing 5% Ficoll. mrna was synthesized using SP6 mmessage mmachine (Ambion) with templates from the following linearized plasmids: pcs2-xsip1 (Eisaki et al., 2000); pcs2-nls-lacz (Takahashi et al., 2000); pcs2-chordin (Sasai et al., 1994); pcs2-xsip1-mt, pcs2-rxsip1, pcs2- rxsip1-mt, pcs2-soxd, pcs2-soxd-ha (UTR ), pcs2- SoxD-HA (UTR+), and pcs2-zic3. Animal cap dissection Dissection of embryos and animal cap assays were performed in 100% SS. Caps were incubated in 100% SS containing 0.1% BSA until sampling. Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis Total RNA was extracted from Xenopus embryos using Isogen (Nippon Gene, Tokyo, Japan). First-strand cdna was synthesized from 1 Ag total RNA using an oligo-(dt) primer and SuperScriptk II RT (Invitrogen Corp., CA, USA). One-twentieth of the cdna was used as a template for RT-PCR. EF-1a was used as a positive control. Reverse transcriptase negative (RT ) reactions showed no evidence of genomic DNA contamination. Primer sequences were either as follows or obtained from the following sources: XSIP1, forward 5V-CACAGATCCGGACACAATTAG-3V and reverse 5V-GTGTACAACAGTGCGGGCAC-3V; N- CAM, Xenopus Molecular Marker Resource ( html); ms-actin, Stutz and Spohr, 1986; EF-1a, Krieg et al., 1989; BMP4, forward 5V-GCATGTACGGATAAGTC- GATC-3V and reverse, 5V-GATCTCAGACTCAACGG- CAC-3V; Msx1, Yamamoto et al., 2000; Vent1, Gawantka et al., 1995; Vent2, Onichtchouk et al., Histological analysis For histological analysis, embryos were fixed in Bouin s solution for 2 12 h, embedded in paraffin, sectioned at 10 Am, and stained in hematoxylin and eosin. Western blot analysis Myc-epitope-tagged proteins were detected using the 9E10 monoclonal antibody (Santa Cruz). HA-epitopetagged proteins were detected using the peroxidase-conjugated 3F10 monoclonal antibody (Roche). a-actin served as the loading control and was detected by the AC-40 monoclonal antibody (SIGMA). HRP-linked anti-mouse IgG (#7076, Cell Signaling Technology, Inc.) and peroxidase-conjugated goat anti-rabbit IgG (A6154, SIGMA) were used as the secondary antibodies. Lac-Z stain and whole-mount in situ hybridization Whole-mount in situ hybridization analysis was performed according to Harland (1991) using albino embryos of X. laevis. Albino embryos were co-injected with NLSbgal mrna and were pre-stained with Red-gal (Research Organics). Antisense RNA probes were synthesized using the following templates; pbluescript(sk )-XSIP1 (Eisaki et al., 2000); psp73-xbra (Smith et al., 1991); pbluescript(sk )-Chordin (Sasai et al., 1994); pbluescript(ks+)-n-tubulin (Richter et al., 1988); pbluescript(sk )-N-CAM (Takebayashi et al., 1997); pbluescript(sk )-SoxD; and pbluescript(sk )-Zic3 (Yabe et al., 2003). The signal was detected using BM purple (Roche). Whole-mount immunohistochemistry Wild-type embryos were bleached using 10% hydrogen peroxide in methanol. The NEU-1 monoclonal antibody has been described previously (Itoh and Kubota, 1989). The secondary antibody was goat anti-mouse IgG + IgM-AP conjugate (AMI0705, Biosource International). After incubation with the secondary antibody, we followed the same protocol as for in situ hybridization. Results XSIP1 regulates the expression of neural but not mesodermal genes Previous overexpression studies showed that XSIP1 regulates Xbra expression (Papin et al., 2002; Verschueren et al., 1999) and that XSIP1 mrna causes hyperneuralization in the injected area (Eisaki et al., 2000). However, it is still unknown which genes XSIP1 regulates in vivo. To address the function of XSIP1, we designed an MO against XSIP1 (XSIP1 MO) and confirmed its specificity. We found that the XSIP1 MO inhibited the translation of XSIP1-MT mrna but not rxsip1-mt mrna (Fig. 1A). The former encodes the wild-type XSIP1 protein with a myc-epitope sequence tag, while the latter encodes the same protein but without the sequence corresponding to the XSIP1 MO. In addition, RT-PCR analysis showed that injection of XSIP1 and rxsip1 mrna induced expression of the pan-neural marker, N-CAM but not the mesodermal marker, ms-actin in animal cap explants as previously shown (Eisaki et al.,

4 K.R. Nitta et al. / Developmental Biology 275 (2004) Fig. 1. XSIP1 MO inhibited neural development. (A) Western blot analysis of myc-tagged proteins. SDS-PAGE was performed with protein prepared from embryos injected into the blastomere of the two cell stage with 1 ng of XSIP1-MT mrna or rxsip1-mt mrna either with or without 40 ng of XSIP1 MO. a- Actin served as an internal control. (B) XSIP1 MO specifically inhibited neural differentiation induced by XSIP1 mrna. All embryos were injected with either 1 ng of XSIP1 or rxsip1 mrna alone or with 40 ng of XSIP1 MO. Animal caps were dissected at stage 9 and cultured. When sibling embryos reached stage 28, total RNA from 30 explants was extracted and some marker genes were examined by RT-PCR. (C E) Embryos injected with XSIP1 MO showed defects in neural and eye development. (C) Uninjected embryo. (D) Forty nanogram of XSIP1 MO was injected into the dorsal-animal pole of the eight-cell-stage embryo. (E) Another embryo similarly injected with 40 ng of XSIP1 MO and 50 pg of linearized pcs2-rxsip1. (F I) Transverse sections of control MO (F,G) and XSIP1 MO (H,I) injected embryos. Scale bars indicate 100 Am. 2000; Fig. 1B; lanes 2 and 4). Co-injection with the XSIP1 MO inhibited the induction of N-CAM by XSIP1 (Fig. 1B; lane 3) but not by rxsip1 (Fig. 1B; lane 5). These results suggest that the XSIP1 MO specifically inhibits the function of XSIP1. Next, we performed a knockdown experiment using XSIP1 MO to test the function of XSIP1 in early embryos. The XSIP1 MO was injected into the whole of two cellstage embryos or the dorsal-animal cells of eight cell-stage embryos. Injection of XSIP1 MO caused head and eye defects (Fig. 1D; 82%, n = 39), while no defects were observed in embryos injected with a control MO (Figs. 1F,G). The phenotypes of embryos injected at the eightcell stage were more severe than those injected at the twocell stage. These phenotypes were rescued by co-injection of pcs2-rxsip1 (Fig. 1E; 83%, n = 53). Histological sections showed that injection of XSIP1 MO caused disturbance of neural tube patterning and eye structure, while injection of the control MO had no effect (Figs. 1F I). Furthermore, in situ hybridization revealed a reduction in the expression of neural markers, N-tubulin (Fig. 2B; 97%, n = 30) and N-CAM (Fig. 2D; 78%, n = 18) in the Fig. 2. Injection of XSIP1 MO resulted in the down-regulation of terminal neural marker genes. Whole-mount in situ hybridization analysis of N- tubulin (A,B) and N-CAM (C,D) performed on stage 28 embryos. Embryos were injected with 40 ng of either control MO or XSIP1 MO into one dorsal-animal blastomere of the eight-cell stage, and co-injected with 250 pg of nuclear-localized lacz as a lineage tracer (red stained). (A,C) Control MO-injected embryo. Injected area (red) and marker expression (blue) overlapped. (B, D) XSIP1 MO-injected embryo. Neural markers, N-tubulin and N-CAM were not expressed in the injected area. Inset; Dorsal view of the same embryos.

5 262 K.R. Nitta et al. / Developmental Biology 275 (2004) cells derived from XSIP1 MO-injected blastomeres, while the expression of these markers was not affected in the cells derived from control MO-injected blastomeres (Figs. 2A,C). These results suggest that XSIP1 is required for neural tissue development. We further analyzed the expression of early mesodermal and neural genes in embryos injected with the XSIP1 MO into either the dorsal-marginal zone of the four-cell stage or the dorsalanimal cells of the eight-cell stage. Although overexpression studies suggested that XSIP1 regulates the expression of Xbra in the early gastrula, the XSIP1 MO did not affect the expression of Xbra (Fig. 3E; 98%, n = 49). The XSIP1 MO also did not affect the expression of chordin, which is expressed in dorsal mesoderm (100%, n = 41). In contrast, injection of the XSIP1 MO down-regulated expression of the early neural genes, SoxD (Fig. 3G; 100%, n = 46) and Zic3 (Fig. 3H; 81%, n = 26). These results suggest that XSIP1 does not regulate mesodermal gene expression, at least at the mid-gastrula stage, and that XSIP1 has an essential role in early neural gene expression. XSIP1 down-regulates BMP and BMP-inducible genes A recent study suggested that ZEB2, the human ortholog of SIP1, represses the transcription of BMP4 and Msx1 (Postigo et al., 2003). In Xenopus embryos, Msx1 is downstream of BMP4. Furthermore, the promoter activity of Vent2, another gene downstream of BMP4, was also repressed by ZEB2 in cultured cells (Postigo, 2003). We have already shown that overexpression of XSIP1 induces neural marker genes, a result that mimics the effect of BMP down-regulation. In addition, the XSIP1 MO downregulated SoxD and Zic3 expression, which is induced by BMP down-regulation. Therefore, we next examined whether XSIP1 also repressed BMP-dependent genes, in the same way as ZEB2. RT-PCR analysis showed that XSIP1 repressed BMP4, Msx1, Vent1, and Vent2 expression (Fig. 4). These results raised the possibility that XSIP1 represses BMP4- and BMP-dependent gene expression to induce neural gene expression. XSIP1 transcription is induced by early neural genes XSIP1 expression is induced by activin and Xbra (Papin et al., 2002). However, these genes are not expressed in neural tissue during embryogenesis. The expression of BMP antagonists is also reduced in neural tissue until the neurula stage (Sasai et al., 1994; Smith and Harland, 1992). In addition, after the neurula stage, expression of various BMP genes is observed around the neural tissue. Nevertheless, XSIP1 is specifically expressed in the neural tissue throughout development. It remains to be answered then by what mechanism XSIP1 expression is maintained. We therefore attempted to find candidates players involved in the regulation of XSIP1 expression. It is known that Chordin is important in early neural induction via its inhibition of BMP activity. RT-PCR analysis revealed that Chordin, like noggin, induced XSIP1 expression in animal caps (Fig. 5; Papin et al., 2002), suggesting that XSIP1 transcription responds to these neural-inducing signals. Several early neural genes, such as SoxD and Zic3, are induced following down-regulation of BMP signaling. These genes are expressed in a broad area of ectoderm in the early blastula. In the late blastula, this expression is restricted to the presumptive neuroectoderm (Figs. 6D F; Mizuseki et al., 1998; Nakata et al., 1997). Because XSIP1 exhibits a similar pattern of expression (Figs. 6A C; Eisaki et al., 2000; van Grunsven et al., 2000), we examined whether SoxD and Zic3 can induce the expression of XSIP1. Overexpression of these genes did induce XSIP1 expres- Fig. 3. Early neural genes were down-regulated by inhibition of XSIP1 mrna translation, while mesodermal genes were not influenced. Whole-mount in situ hybridization analysis of Xbra (A,E) and chordin (B,F) expression in stage 11 embryos, which had been injected with either control MO or XSIP1 MO into the dorsal marginal zone of the four-cell-stage embryo. An equivalent analysis of SoxD (C,G) and Zic3 (D,H) expression in stage 11 embryos, although one dorsalanimal blastomere of the eight-cell-stage embryo was injected in these embryos. SoxD and Zic3 expression were down-regulated in the XSIP1 MO-injected area (G,H; arrowhead), while they overlapped in the control MO-injected area (C,D). The dorsal view of all embryos is shown.

6 K.R. Nitta et al. / Developmental Biology 275 (2004) SoxD MO specifically inhibited SoxD translation. Injection of the SoxD MO into the dorsal-animal cells of eight cell-stage embryo resulted in embryos with neural tube and eye defects (Fig. 7C) and these defects were rescued by co-injection with pcs2-soxd-orf, which have no UTR sequence same as SoxD-HA(UTR ) so that SoxD MO does not bind to its transcripts (Fig. 7D; 84%, n = 75). Analysis of terminal neural differential markers by in situ hybridization revealed that major defects in neural development were observed in the regions of the embryo that had received the injected SoxD MO, as traced by coinjection of b-galactosidase mrna (Figs. 7F,H,J). These phenotypes were similar to that seen with the XSIP1 MOinjected embryos and SoxD dominant-negative experiments (Mizuseki et al., 1998). To address whether SoxD is required for XSIP1 expression in vivo, we carried out in situ hybridization analysis using embryos injected with SoxD MO. XSIP1 expression was down-regulated in SoxD MO-injected embryos from early through to late stages (Figs. 8D F), suggesting that SoxD plays an important role in the induction and maintenance of XSIP1 expression. XSIP1 is essential for SoxD-induced neural differentiation Fig. 4. XSIP1 down-regulated BMP and BMP-dependent genes. Animal caps were dissected from embryos injected with 1 ng XSIP1 mrna and cultured until sibling embryos reached stage 11. EF-1a is the loading control. Sibling control embryos served as a positive control (WE), and PCR without reverse transcriptase was performed (RT ) to verify the absence of genomic DNA. We revealed that XSIP1 was required for SoxD expression and was essential for proper neural development. We then analyzed whether XSIP1 was required for SoxD function. Misexpression of SoxD in the ventral marginal zone of four-cell-stage embryos resulted in formation of ectopic neural tissue (Figs. 9B,E), while the embryos co-injected with SoxD mrna and XSIP1 MO showed no ectopic neural tissue (Figs. 9C,F). This result suggests that the induction of XSIP1 by SoxD is a sion in animal caps (Fig. 5), raising the possibility that SoxD and Zic3 act to maintain XSIP1 expression during neural differentiation. An MO against SoxD inhibits neural differentiation and the expression of XSIP1 SoxD belongs to the Sox gene family, which encodes Sry-related transcription factors containing an HMG DNA-binding domain (Schepers et al., 2002). SoxD has a strong neural-inducing activity (Mizuseki et al., 1998). To determine the in vivo relationship between SoxD and XSIP1, we used an MO against SoxD (SoxD MO). To test the specificity of the SoxD MO, we constructed pcs2- SoxD-HA(UTR+) and pcs2-soxd-ha(utr ). Both encode an HA-tagged SoxD protein but the latter lacks the target sequence for the SoxD MO. The SoxD MO inhibited the translation of SoxD-HA(UTR+) mrna (Fig. 7A; lanes 2 and 3) but not of SoxD-HA(UTR ) mrna (Fig. 7A; lanes 4 and 5). This result suggested that the Fig. 5. The expression of XSIP1 in animal caps from embryos injected with mrna encoding neural inducers or early neural genes. Animal caps were dissected from embryos injected with 50 pg chordin, 500 pg SoxD, or 500 pg Zic3 mrna and cultured until sibling embryos reached stage 11. Sibling control embryos served as a positive control (WE), and PCR without reverse transcriptase (RT ) was performed to verify the absence of genomic DNA.

7 264 K.R. Nitta et al. / Developmental Biology 275 (2004) Fig. 6. Comparison of expression of XSIP1 and SoxD during embryogenesis. Whole-mount in situ hybridization of XSIP1 (A C) and SoxD (D F). (A,D) Dorsal view. (B,E) Anterior view. necessary step in the regulation of neural tissue development by SoxD. Discussion Neural induction and maintenance of neural fate with XSIP1 Previous studies have suggested that the expression of XSIP1 is highly correlated with that of early neural genes, such as SoxD, after the late gastrula stage (Eisaki et al., 2000; Mizuseki et al., 1998). In this study, we showed that misexpression of SoxD and Zic3 induced XSIP1 transcription and that knockdown of SoxD using a SoxD MO caused its reduction. This suggests that early neural genes control XSIP1 and that SoxD plays an essential role in XSIP1 expression. The mesodermal inducers, activin and Xbra, induce XSIP1 expression, and the same was found for the neural-inducing genes, noggin and tbr (Papin et al., 2002). In addition, XSIP1 was identified in our screen of a cdna library prepared from activin-treated animal caps (Eisaki et al., 2000). That induction of XSIP1 by the mesodermal inducers requires protein synthesis (Papin et al., 2002) might reflect neural induction from mesoderm. In fact, activin induces several kinds of neural inducers, such as Chordin, that are expressed in mesoderm (Sasai et al., 1994), and overexpression of Xbra induces the expression of noggin and follistatin (Strong et al., 2000). We showed that MOs against XSIP1 and SoxD inhibited each other s expression, suggesting that the expression of XSIP1 and SoxD are interdependent. However, the expres- Fig. 7. Injection of SoxD MO into Xenopus embryos resulted in a neural defect phenotype. (A) Western blot analysis of HA-tagged proteins. SDS-PAGE was performed with protein prepared from embryos injected with 1 ng SoxD-HA(UTR+) or SoxD-HA(UTR ) mrna either with or without 40 ng SoxD MO and cultured until stage 9. a-actin served as an internal control. (B D) Phenotype of embryos injected with 20 ng SoxD MO. (B) Uninjected embryo. (C) SoxD MO was injected into dorsal-animal blastomeres of eight-cell-stage embryos. (D) Linerized pcs2-soxd-orf (100 pg) was co-injected with SoxD MO into dorsal-animal blastomeres of eight-cell-stage embryos. (E J) Whole-mount in situ hybridization analysis of terminal differentiation markers N-tubulin (E H) and N-CAM (I,J). Embryos were injected with 40 ng SoxD MO into dorsal-animal blastomeres of eight-cell-stage embryos (F,H,J) and cultured until stage 15 (E,F) or 25 (G J). (E,F) Dorsal view.

8 K.R. Nitta et al. / Developmental Biology 275 (2004) Fig. 8. SoxD is required for XSIP1 expression. Whole-mount in situ hybridization analysis of XSIP1 expression in embryos cultured to stage 11 (A,D), 15 (B,E), and 25 (C,F). (A C) Uninjected embryos. (D F) Embryos injected with 40 ng of SoxD MO into dorsal-animal blastomeres of eight-cell-stage embryos. sion of SoxD is first activated in the entire ectoderm in the early gastrula (Mizuseki et al., 1998), whereas XSIP1 expression is restricted to the dorsal side (Papin et al., 2002). This fact suggests that the onset of XSIP1 and SoxD expression might be independent of each other. In fact, the overexpression of Chordin induced stronger expression of XSIP1 than either SoxD or Zic3 (data not shown), suggesting that inhibition of BMP signaling might be required for the full induction of XSIP1 expression. During gastrulation, expression of XSIP1 and SoxD becomes restricted to the neuroectoderm and their expression overlaps at the neurula stage (Eisaki et al., 2000; Mizuseki et al., 1998). In the late gastrula stage, the expression of SoxD is suppressed by BMP and the BMP-responsive genes, GATA1 and Msx1 (Mizuseki et al., 1998). We also showed here that XSIP1 suppressed the expression of BMP4 and BMPresponsive genes, suggesting that continuous suppression of BMP signaling by XSIP1 following by secreted BMP inhibitors is necessary for maintenance of SoxD expression and neural differentiation. In fact, suppression of BMP signaling is necessary well beyond the gastrula stages for neural development to proceed normally (Hartley et al., 2001). Therefore, XSIP1 might be the candidate BMP suppressor in neural tissue, although this hypothesis needs further investigation. Conserved function of XSIP1 in vertebrate embryogenesis It is known that mutation of the human SIP1 gene leads to Hirschsprung disease, which is associated with microcephaly, mental retardation, epilepsy, and characteristic facial features (Cacheux et al., 2001; Wakamatsu et al., 2001; Yamada et al., 2001). In the most severe cases, the patient shows severe brain atrophy on magnetic resonance imaging. XSIP1 MO-injected embryos showed similar and more severe defects including reduced head region, eye defects, and disturbed neural development. Because the patients with Hirschsprung disease have a frameshift mutation affecting only one allele, expression from the normal SIP1 allele may moderate the severe phenotype observed with the XSIP1 MO-injected embryos. These results suggest that the function of SIP1 might be highly conserved between frog and human and hence among vertebrates. The regulation of Xbra by genes of the def1/zfh family Three lines of evidences support that XSIP1 acts as a repressor of Xbra expression and suppresses it in the animal region. First, overexpression of both murine SIP1 and Xenopus XSIP1 caused down-regulation of expression of Fig. 9. XSIP1 is required for SoxD-mediated neural induction. Embryos were injected with 500 pg SoxD mrna alone (B,E) or co-injected with 40 ng XSIP1 MO (C,F) into the ventral marginal zone of four-cell-stage embryos and cultured until stage 32. (B,E) Ectopic neural tissue was observed on the ventral side (arrow head). (C,F) Ectopic neural tissue was not observed on the ventral side (arrow head).

9 266 K.R. Nitta et al. / Developmental Biology 275 (2004) brachyury (Papin et al., 2002; Verschueren et al., 1999). Second, deletion of the SIP1 consensus binding site from the Xbra promoter alters expression from a restricted pattern in the marginal zone to more general expression, spreading into the animal pole on the dorsal side (Lerchner et al., 2000; Remacle et al., 1999). Finally, expression of a fusion protein of XSIP1 and the VP16 activator domain (XSIP1- VP16) resulted in Xbra expression in animal caps (Papin et al., 2002). In contrast to these results, we showed that XSIP1 MO injection had no effect on Xbra expression, suggesting that XSIP1 is not sufficient to suppress Xbra expression in the Xenopus embryo. It is plausible that binding of XSIP1 or XSIP1-VP16 to the XSIP1 binding sites in the Xbra promoter could either up- or down-regulate Xbra expression depending on the context. Because this binding site is a consensus sequence for the def1/zfh family, the other members of this family could interact and affect the expression of Xbra. In fact, yef1 has been shown to bind this sequence in the Xbra promoter (Remacle et al., 1999). Our results raise the possibility that other members of the def1/zfh family might cooperate with XSIP1 to suppress Xbra expression in Xenopus embryos. Acknowledgments We thank Drs. J.C. Smith, Y. Sasai, I. Dawid, K. Takebayashi for their generous gifts of plasmids. This work was supported by Grants-in-Aid for Science Research from the Ministry of Education, Science, Sports and Culture of Japan and by ICORP (International Cooperative Research Project) of the Japan Science and Technology Agency (JST). S.T. and K.T. were supported by a research fellowship of the Japan Society for the Promotion Sciences for Young Scientists. References Cacheux, V., Dastot-Le Moal, F., N77ri7inen, H., Bondurand, N., Rintala, R., Boissier, B., Wilson, M., Mowat, D., Goossens, M., Loss-of-function mutations in SIP1 Smad interacting protein 1 result in a syndromic Hirschsprung disease. Hum. Mol. Genet. 10, Comijn, J., Berx, G., Vermassen, P., Verschueren, K., van Grunsven, L., Bruyneel, E., Mareel, M., Huylebroeck, D., van Roy, F., The twohanded E box binding zinc finger protein SIP1 downregulates E- cadherin and induces invasion. Mol. Cell 7, Dai, H., Hogan, C., Gopalakrishnan, B., Torres-Vazquez, J., Nguyen, M., Park, S., Raftery, L.A., Warrior, R., Arora, K., The zinc finger protein Schnurri acts as a Smad partner in mediating the transcriptional response to Decapentaplegic. Dev. Biol. 227, Eisaki, A., Kuroda, H., Fukui, A., Asashima, M., XSIP1, a member of two-handed zinc finger proteins, induced anterior neural markers in Xenopus laevis animal cap. Biochem. Biophys. Res. Commun. 271, Fainsod, A., Deissler, K., Yelin, R., Marom, K., Epstein, M., Pillemer, G., Steinbeisser, H., Blum, M., The dorsalizing and neural inducing gene follistatin is an antagonist of BMP-4. Mech. Dev. 63, Gawantka, V., Delius, H., Hirschfeld, K., Blumenstock, C., Niehrs, C., Antagonizing the Spemann organizer: role of the homeobox gene Xvent-1. EMBO J. 14, Haramoto, Y., Tanegashima, K., Onuma, Y., Takahashi, S., Sekizaki, H., Asashima, M., Xenopus tropicalis nodal-related gene 3 regulates BMP signaling: an essential role for the pro-region. Dev. Biol. 265, Harland, R.M., In situ hybridization; an improved whole-mount method for Xenopus embryo. Methods Cell Biol. 36, Hartley, K.O., Hardcastle, Z., Friday, R.V., Amaya, E., Papalopulu, N., Transgenic Xenopus embryos reveal that anterior neural development requires continued suppression of BMP signaling after gastrulation. Dev. Biol. 238, Hata, A., Seoane, J., Lagna, G., Montalvo, E., Hemmati-Brivanlou, A., Massagué, J., OAZ uses distinct DNA- and protein-binding zinc fingers in separate BMP-Smad and Olf signaling pathways. Cell 100, Hemmati-Brivanlou, A., Kelly, O.G., Melton, D.A., Follistatin, an antagonist of activin, is expressed in the Spemann organizer and displays direct neuralizing activity. Cell 77, Higashi, Y., Maruhashi, M., Nelles, L., Van de Putte, T., Verschueren, K., Miyoshi, T., Yoshimoto, A., Kondoh, H., Huylebroeck, D., Generation of the floxed allele of the SIP1 (Smad-interacting protein 1) gene for Cre-mediated conditional knockout in the mouse. Genesis 32, Itoh, K., Kubota, H.Y., Expression of neural antigens in normal Xenopus embryos and induced explants. Dev. Growth Differ. 31, Krieg, P.A., Varnum, S.M., Wormington, W.M., Melton, D.A., The mrna encoding elongation factor 1-alpha (EF-1 alpha) is a major transcript at the midblastula transition in Xenopus. Dev. Biol. 133, Lamb, T.M., Knecht, A.K., Smith, W.C., Stachel, S.E., Economides, A.N., Stahl, N., Yancopolous, G.D., Harland, R.M., Neural induction by the secreted polypeptide noggin. Science 262, Lerchner, W., Latinkic, B.V., Remacle, J.E., Huylebroeck, D., Smith, J.C., Region-specific activation of the Xenopus Brachyury promoter involves active repression in ectoderm and endoderm: a study using transgenic frog embryos. Development 127, Massagué, J., Wotton, D., Transcriptional control by the TGF-h/ Smad signaling system. EMBO J. 19, Matsuo-Takasaki, M., Lim, J.H., Sato, S.M., The POU domain gene, XlPOU 2 is an essential downstream determinant of neural induction. Mech. Dev. 89, Mizuseki, K., Kishi, M., Shiota, K., Nakanishi, S., Sasai, Y., SoxD: an essential mediator of induction of anterior neural tissues in Xenopus embryos. Neuron 21, Nakata, K., Nagai, T., Aruga, J., Mikoshiba, K., Xenopus Zic3, a primary regulator both in neural and neural crest development. Proc. Natl. Acad. Sci. U. S. A. 94, Nakata, K., Nagai, T., Aruga, J., Mikoshiba, K., Xenopus Zic family and its role in neural and neural crest development. Mech. Dev. 75, Nieuwkoop, P.D., Faber, J., Nomal Table of Xenopus laevis (Daudin). North-Holland, Amsterdam. Onichtchouk, D., Gawantka, V., Dosch, R., Delius, H., Hirschfeld, K., Blumenstock, C., Niehrs, C., The Xvent-2 homeobox gene is part of the BMP-4 signalling pathway controlling [correction of controling] dorsoventral patterning of Xenopus mesoderm. Development 122, (Erratum in: Development 124, precedi). Papin, C., van Grunsven, L.A., Verschueren, K., Huylebroeck, D., Smith, J.C., Dynamic regulation of Brachyury expression in the amphibian embryo by XSIP1. Mech. Dev. 111, Piccolo, S., Sasai, Y., Lu, B., De Robertis, E.M., Dorsoventral patterning in Xenopus: inhibition of ventral signals by direct binding of chordin to BMP-4. Cell 86,

10 K.R. Nitta et al. / Developmental Biology 275 (2004) Postigo, A.A., Opposing functions of ZEB proteins in the regulation of the TGFh/BMP signaling pathway. EMBO J. 22, Postigo, A.A., Depp, J.L., Taylor, J.J., Kroll, K.L., Regulation of Smad signaling through a differential recruitment of coactivators and corepressors by ZEB proteins. EMBO J. 22, Remacle, J.E., Kraft, H., Lerchner, W., Wuytens, G., Collart, C., Verschueren, K., Smith, J.C., Huylebroeck, D., New mode of DNA binding of multi-zinc finger transcription factors: yef1 family members bind with two hands to two target sites. EMBO J. 18, Richter, K., Grunz, H., Dawid, I.B., Gene expression in the embryonic nervous system of Xenopus laevis. Proc. Natl. Acad. Sci. U. S. A. 85, Sasai, Y., Lu, B., Steinbeisser, H., Geissert, D., Gont, L.K., De Robertis, E.M., Xenopus chordin: a novel dorsalizing factor activated by organizer-specific homeobox genes. Cell 79, Sasai, Y., Lu, B., Steinbeisser, H., De Robertis, E.M., Regulation of neural induction by the Chd and Bmp-4 antagonistic patterning signals in Xenopus. Nature 376, Schepers, G.E., Teasdale, R.D., Koopman, P., Twenty pairs of sox: extent, homology, and nomenclature of the mouse and human sox transcription factor gene families. Dev. Cell. 3, Sheng, G., dos Reis, M., Stern, C.D., Churchill, a zinc finger transcriptional activator, regulates the transition between gastrulation and neurulation. Cell 115, Smith, J.C., Making mesoderm upstream and downstream of Xbra. Int. J. Dev. Biol. 45, Smith, J.C., Price, B.M., Green, J.B., Weigel, D., Herrmann, B.G., Expression of a Xenopus homolog of Brachyury (T) is an immediate early response to mesoderm induction. Cell 67, Smith, W.C., Harland, R.M., Expression cloning of noggin, a new dorsalizing factor localized to the Spemann organizer in Xenopus embryos. Cell 70, Smith, W.C., McKendry, R., Ribisi Jr., S., Harland, R.M., A nodalrelated gene defines a physical and functional domain within the Spemann organizer. Cell 82, Spemann, H., Mangold, H., Über induktion von Embryonalagen durch Implantation Artfremder Organisatoren. Roux Arch. Entwickl. Mech. 100, Strong, C.F., Barnett, M.W., Hartman, D., Jones, E.A., Stott, D., Xbra3 induces mesoderm and neural tissue in Xenopus laevis. Dev. Biol. 222, Stutz, F., Spohr, G., Isolation and characterization of sarcomeric actin genes expressed in Xenopus laevis embryos. J. Mol. Biol. 187, Takahashi, S., Yokota, C., Takano, K., Tanegashima, K., Onuma, Y., Goto, J., Asashima, M., Two novel nodal-related genes initiate early inductive events in Xenopus Nieuwkoop center. Development 24, Takebayashi, K., Takahashi, S., Yokota, C., Tsuda, H., Nakanishi, S., Asashima, M., Kageyama, R., Conversion of ectoderm into a neural fate by ATH-3, a vertebrate basic helix-loop-helix gene homologous to Drosophila proneural gene atonal. EMBO J. 16, Tylzanowski, P., De Valck, D., Maes, V., Peeters, J., Luyten, F.P., Zfhx1a and Zfhx1b mrnas have non-overlapping expression domains during chick and mouse midgestation limb development. Gene Expression Patterns 3, Udagawa, Y., Hanai, J., Tada, K., Grieder, N.C., Momoeda, M., Taketani, Y., Affolter, M., Kawabata, M., Miyazono, K., Schnurri interacts with Mad in a Dpp-dependent manner. Genes Cells 5, van Grunsven, L.A., Papin, C., Avalosse, B., Opdecamp, K., Huylebroeck, D., Smith, J.C., Bellefroid, E.J., XSIP1, a Xenopus zinc finger/ homeodomain encoding gene highly expressed during early neural development. Mech. Dev. 94, Van de Putte, T., Maruhashi, M., Francis, A., Nelles, L., Kondoh, H., Huylebroeck, D., Higashi, Y., Mice lacking Zfhx1B, the gene that codes for Smad-interacting protein-1, reveal a role for multiple neural crest cell defects in the etiology of Hirschsprung disease mental retardation syndrome. Am. J. Hum. Genet. 72, Verschueren, K., Remacle, J.E., Collart, C., Kraft, H., Baker, B.S., Tylzanowski, P., Nelles, L., Wuytens, G., Su, M.T., Bodmer, R., Smith, J.C., Huylebroeck, D., SIP1, a novel zinc finger/homeodomain repressor, interacts with Smads proteins and binds to 5V-CACCT sequences in candidate target genes. J. Biol. Chem. 274, Wakamatsu, N., Yamada, Y., Yamada, K., Ono, T., Nomura, N., Taniguchi, H., Kitoh, H., Mutoh, N., Yamanaka, T., Mushiake, K., Kato, K., Sonta, S., Nagaya, M., Mutations in SIP1, encoding Smad interacting protein-1, cause a form of Hirschsprung disese. Nat. Genet. 27, Wilson, P.A., Hemmati-Brivanlou, A., Induction of epidermis and inhibition of neural fate by Bmp-4. Nature 376, Witta, S.E., Agarwal, V.R., Sato, S.M., XlPOU 2, a noggin-inducible gene, has direct neuralizing activity. Development 121, Yabe, S., Tanegashima, K., Haramoto, Y., Takahashi, S., Fujii, T., Kozuma, S., Taketani, Y., Asashima, M., FRL-1, a member of the EGF- CFC family, is essential for neural differentiation in Xenopus early development. Development 130, Yamada, K., Yamada, Y., Nomura, N., Miura, K., Wakako, R., Hayakawa, C., Matsumoto, A., Kumagai, T., Yoshimura, I., Miyazaki, S., Kato, K., Sonota, S., Ono, H., Yamanaka, T., Nagaya, M., Wakamatsu, N., Nonsense and Frameshift mutations in ZFHX1B, encoding Smadinteracting protein 1, cause a complex developmental disorder with a great variety of clinical features. Am. J. Hum. Genet. 69, Yamamoto, T.S., Takagi, C., Ueno, N., Requirement of Xmsx-1 in the BMP-triggered ventralization of Xenopus embryos. Mech. Dev. 91, Yokota, C., Kofron, M., Zuck, M., Houston, D.W., Isaacs, H., Asashima, M., Wylie, C.C., Heasman, J., A novel role for a nodal-related protein; Xnr3 regulates convergent extension movements via the FGF receptor. Development 130, Zimmerman, L.B., De Jesús-Escobar, J.M., Harland, R.M., The Spemann organizer signal noggin binds and inactivates bone morphogenetic protein 4. Cell 86,