Background : The Wnt signal transduction pathway regulates various aspects of embryonal development and has been implicated in promoting cancer. Signalling by Wnts leads to the stabilization of cytosolic β-catenin, which then associates with TCF transcription factors to regulate expression of Wnt-target genes. The Wnt pathway is further subject to cross-regulation at various levels by other components.
Results: Recent evidence suggests that a specific MAP kinase pathway involving the MAP kinase kinase kinase TAK1 and the MAP kinase NLK counteract Wnt signalling. In particular, it has been shown that TAK1 activates NLK, which phosphorylates TCFs bound to β-catenin. This phosphorylation down-regulates the DNA-binding activity of a TCF-4/β-catenin complex, and blocks activation of their target genes. To investigate the role of NLK in Xenopus development, we isolated xNLK, a Xenopus homologue of NLK. Our findings indicate that xNLK is expressed in neural tissues and induces the anterior-neural marker gene, Otx-2. Moreover, xSox11, which is induced by the expression of Chordin, co-operates with xNLK to induce neural development. These molecules also interact in mammalian cells, and expression of a mutant of xNLK lacking kinase activity was found to suppress the induction of neural marker gene expression by xSox11.
Conclusions : Our findings indicate that xNLK may play a role in neural development together with xSox11 during early Xenopus embryogenesis.
Wnt stimulation of the Frizzled receptors and activation of the Wnt signalling pathway leads to the stabilization of β-catenin via the cytoplasmic protein Dishevelled, which antagonizes the destabilizing effect of glycogen synthetase kinase 3 (GSK3) and Axin (Cadigan & Nusse 1997; Sokol 1999). The maternal β-catenin, complexed with the TCF family of transcriptional regulators, activates several dorsal region-specific target genes, such as Siamois and Twin, and is among the earliest steps leading to induction of the organizer (Cadigan & Nusse 1997; Harland & Gerhart 1997; Moon et al. 1997).
Recent evidence suggests that a specific MAP kinase pathway involving the MAP kinase kinase kinase TAK1 (TGF-β activated kinase 1) and the MAP kinase NLK (Nemo like kinase) antagonize Wnt signalling (Behrens 2000). In particular, homologues of TAK1 and NLK in Caenorhabditis elegans, termed MOM-4 and LIT-1 (Rocheleau et al. 1999; Shin et al. 1999), negatively regulate cell fate normally promoted by Wnt signalling in early embryos (Meneghini et al. 1999). Moreover, TAK1 activates NLK, which in turn phosphorylates TCFs bound to β-catenin, thereby down-regulating the DNA-binding activity of a TCF-4/β-catenin complex and blocking interaction with target gene promoters. In Xenopus embryos, injection of NLK mRNA can block secondary axis formation induced by ectopic Wnt signalling (Ishitani et al. 1999). Thus, the TAK1-NLK pathway negatively regulates the Wnt signalling pathway.
The Sox gene family encodes Sry-related transcription factors that contain an HMG DNA-binding domain (Pevny & Lovell-Badge 1997). A recent study indicates that Sox proteins interact with β-catenin and inhibit transcriptional activation by TCF-β-catenin complexes, which results in ventralization of Xenopus embryos (Zorn et al. 1999). These results show that Sox proteins modulate Wnt signalling pathways. In addition, some Sox proteins, such as the neurally expressed Sox2 and SoxD, are required for neural differentiation of early Xenopus ectoderm (Kishi et al. 2000; Mizuseki et al. 1998b).
In order to analyse the function of NLK during Xenopus development, we isolated xNLK, a Xenopus homologueue of NLK. We first examined its potential role in axis formation in Xenopus embryos. We also found that xNLK was expressed in neural tissues and induced the anterior-neural marker gene, Otx-2. Moreover, in mammalian cells xNLK could interact with xSox11, and these molecules co-operated to induce neural development. Our findings thus indicate that xNLK may play a role together with xSox11 in neural development during Xenopus early embryogenesis.
Isolation of the Xenopus NLK cDNA and analysis of its expression patterns in Xenopus embryos
To analyse the function of NLK in embryonic development, cDNAs encoding Xenopus NLK (xNLK) were isolated from an oocyte cDNA library under low stringency conditions using the mouse NLK (mNLK) cDNA (Brott et al. 1998) as probe. The longest xNLK cDNA recovered was approximately 3.5 kb. Sequence analysis of the longest cDNA for xNLK revealed a single open reading frame. The nucleotide sequences were predicted to encode a protein of 447 amino acids with a molecular mass of 50 kDa for xNLK (Fig. 1A). The amino acid sequence comparison between xNLK and mNLK or Drosophila Nemo (Choi & Benzer 1994), shown in Fig. 1B, reveals that it is 66% identical to mNLK or 68% to Nemo.
RT-PCR analysis of xNLK mRNA obtained during early embryogenesis revealed that its transcript is stored maternally and expressed throughout development (Fig. 2A). During early gastrula stages, when neural induction first takes place, xNLK is expressed widely in the ectoderm, as determined by whole mount in situ hybridization (Fig. 2B). At the end of neurulation, whole mount in situ hybridization of embryos showed that xNLK transcripts become localized to the nervous system, and are restricted to the central nervous system, eye and head neural crest cell populations by the early tadpole stages.
Injection of xNLK mRNA inhibits Wnt-induced axis formation in Xenopus embryos
It is well established that the Wnt pathway plays a crucial role in the development of the Xenopus embryonic axis, and injection of mRNA encoding various components of the Wnt pathway into Xenopus embryos are known to induce the formation of an ectopic axis (Cadigan & Nusse 1997; Miller & Moon 1996). Previously, we have shown that injection of mNLK mRNA can block secondary axis formation induced by ectopic Wnt signalling (Ishitani et al. 1999). To test whether xNLK also exerts its effects via the Wnt pathway, we examined the effects of xNLK on axis duplication induced by β-catenin or Siamois. Injection of β-catenin into the vegetal ventral region of early cleavage-stage embryos leads to axis duplication (Funayama et al. 1995). Siamois is a homeobox gene whose expression is specifically activated by β-catenin-TCF and which mediates the effects of the Wnt pathway on axis formation (Brannon et al. 1997; Lemaire et al. 1995). Co-injection of xNLK strongly suppressed β-catenin-induced axis duplication and in most cases resulted in the normal development of the embryo (Fig. 3A). Axis duplication induced by injection of Siamois mRNA could not be blocked by xNLK. Moreover, the expression of xNLK inhibited the expression of the Xnr-3 and Siamois genes (Carnac et al. 1996; Smith et al. 1995), which are direct target genes for Wnt signalling, while the expression of the kinase negative type xNLK (xNLK-KN) did not (Fig. 3B). Thus, injection of xNLK mRNA can block secondary dorsal axis formation in Xenopus embryos, apparently by interfering with signalling through the Wnt pathway at a point downstream of β-catenin, and upstream of Siamois. Taken together, these results suggest that xNLK also functions as a negative regulator of the Wnt pathway in Xenopus development.
xNLK can promote neural differentiation in animal caps
We next used RT-PCR to investigate the roles of xNLK in neural induction. When 200 pg of xNLK mRNA was injected into the animal pole of two-cell embryos, expression of the pan-neural marker N-CAM was observed in the animal caps (Kintner & Melton 1987) (Fig. 4), but not in the control caps. Moreover, xNLK mRNA injection increased expression of the anterior-neural marker, Otx-2 (Lamb et al. 1993), whereas the midbrain marker En-2 (Hemmati-Brivanlou et al. 1990), hindbrain marker Krox20 (Bradley et al. 1993) and spinal cord marker HoxB9 (Wright et al. 1990) were not induced. These results show that xNLK is sufficient to induce anterior neural differentiation in the animal cap explants.
xSox11 and xNLK can promote neural differentiation
Sox proteins, such as Sox2, Sox11 and SoxD, are known regulators of neural differentiation (Hiraoka et al. 1997; Kishi et al. 2000; Mizuseki et al. 1998b). Sox proteins have also been shown to bind to β-catenin and inhibit its TCF-mediated signalling activity (Zorn et al. 1999). Moreover, we previously reported that mNLK binds to TCF and inhibits Wnt signalling (Ishitani et al. 1999). Therefore, we next examined by RT-PCR the possibility that Sox proteins may work together with xNLK to induce neural development. In animal cap assays, injection of Xenopus Sox11 (xSox11) induced the anterior-neural marker genes Otx-2 and En-2 (Fig. 5A), and the pan-neural marker N-CAM. Moreover, RT-PCR analysis showed that a combination of xSox11 and xNLK further enhanced expression of these neural markers. On the other hand, xNLK did not affect the neural marker gene induction by either Sox2 or SoxD (data not shown). In addition to their role in inducing neural development, Sox2 and SoxD are known to be induced by the Chordin gene (Mizuseki et al. 1998a,b). Therefore, we tested whether Chordin also induces the expression of xSox11. We found that the xSox11 transcript, but not the xNLK transcript, was induced by expression of Chordin (Fig. 5B). Furthermore, a kinase-negative mutant of xNLK (xNLK-KN) inhibited the induction of anterior neural marker genes by xSox11 or Chordin (Fig. 5A,C). Whole mount in situ hybridization of embryos showed that xSox11 transcripts become localized to the nervous system at the end of neurulation, and are restricted to the central nervous system, eye and head neural crest cell populations by the early tadpole stages (Fig. 2C). These data suggest that both xSox11 and xNLK are expressed in the same dorsal site and co-operate in the neural development.
xSox11 binds xNLK
To determine how xNLK enhances xSox11’s neural inducing activity, we asked if xNLK and xSox11 could physically interact, analogous to the interaction of mammalian NLK with TCF, an HMG protein. To confirm that the interaction between xNLK and xSox11 occurs in vivo, co-immunoprecipitation experiments were performed with Flag-tagged xNLK and T7-tagged xSox11. Two hundred and ninety-three cells were transfected T7-xSox11 with or without Flag-xNLK. xNLK was precipitated from the resulting cell lysates with anti-Flag antibodies, and the immunoprecipitates were then subjected to anti-T7 epitope Western blotting. xSox11 co-precipitated with Flag-xNLK (Fig. 6). We conclude that the neural induction by xNLK may be due largely to the physical interaction between xNLK and xSox11.
In this work, we isolated xNLK, a Xenopus homologue of mouse NLK and Drosophila Nemo, and demonstrated that it plays an essential role in neural differentiation during Xenopus development. Our recent studies in mammalian cells indicated that NLK binds to and phosphorylates the HMG-domain protein TCF-4, thereby down-regulating the DNA-binding activity of a TCF-4/β-catenin complex (Ishitani et al. 1999). Our results here suggest that xNLK also inhibits the Wnt pathway downstream of β-catenin and upstream of the Wnt signal target gene, Siamois, indicating that xNLK may function as a negative regulator of Wnt signalling in the determination of dorso-ventral polarity, by the same system as mouse NLK.
Our results also demonstrate that xNLK may play a role in neural development in co-operation with xSox11, which is a transcription factor containing a DNA-binding HMG box (Hiraoka et al. 1997). The expression of the Sox genes has been detected in the central nervous system during development in mouse, chick and zebrafish, and they are known to play diverse roles in vertebrate differentiation and development (Collignon et al. 1996; de Martino et al. 2000; Kamachi et al. 1995; Uwanogho et al. 1995). Since it has been shown that mNLK interacts with HMG-domain proteins, such as a TCF/Lef-1, we expected that xNLK could directly interact with Sox proteins containing HMG domains and regulate their transcriptional activities to promote anterior neural differentiation. Indeed, xNLK can be co-immunoprecipitated with xSox11, providing evidence that xNLK and xSox11 might form a complex. In addition, we show that the kinase-negative form of xNLK (xNLK-KN) inhibits the expression of anterior-neural marker genes induced by xSox11 in an animal cap assay. These results raise the possibility that phosphorylation and interaction with xNLK could be important for the transactivation of xSox11, similar to what is observed with the mammalian TCFs, and that these complexes may act as positive regulators for anterior neural differentiation.
A number of models could explain how xSox11 co-operates with xNLK to promote neural induction. Our data suggest that xNLK is the responsible kinase, as it binds to and increases the phosphorylation of Sox11. The simplest possibility is that phosphorylated xSox11 competes with TCF for binding to xNLK. In cells with higher levels of phosphorylated xSox11 protein, xNLK would be sequestered and rendered unavailable for interaction with TCF. Thus, the relative levels of TCF and xSox11 proteins would determine how a cell responds to a Wnt signal. Alternatively, it is possible that xSox11 bind to a complex of xNLK and TCF. In this model, xSox11 would repress TCF/xNLK activity as part of a ternary or higher order complex. Clarification of this model must await further investigation.
In Xenopus, neural inducers such as Noggin, Chordin and Follistatin are secreted by the Spemann organizer and can promote neuralization of animal cap ectoderm by inactivating BMP4 (Hemmati-Brivanlou & Melton 1997; Sasai & De Robertis 1997). It is known that the transcription of the Sox genes family is regulated by Chordin/BMP signals (Mizuseki et al. 1998b). We have shown here that Chordin induces the transcription of xSox11 but not xNLK. Moreover, the expression of anterior-neural marker genes induced by the injection of Chordin mRNA in animal caps is inhibited by co-injection of an mRNA encoding a kinase-inactive mutant of xNLK. Our results indicate that the xSox11 transcript can be induced by Chordin, and that xSox11 forms a complex with xNLK to induce neural differentiation. Furthermore, co-injection of xNLK and xSox11 mRNAs enhance the induction of anterior-neural marker genes, suggesting that a complex formed of xNLK and xSox11 plays a role in the neural differentiation induced by Chordin.
Previous reports have shown that NLK is activated by TAK1 in the presence of its activator TAB1, and regulates the Wnt signalling pathway (Ishitani et al. 1999). More recent studies have indicated that NLK/Nemo/LIT-1 plays a role in two Wnt-dependent processes in Drosophila and in C. elegans, controlling embryonic epidermal patterning and planar cell polarity, respectively (Rocheleau et al. 1999; Verheyen et al. 2001). Moreover, Nemo was involved in various signalling pathways including Notch, Wingless and Dpp pathways (Verheyen et al. 2001), and was shown to be required for development in various tissues. In this work, we demonstrate that xNLK may regulate dorso-ventral patterning and neural development. It is therefore possible that NLK plays an important role in tissue modelling during development. However, it remains unclear as to what targets are regulated by the xNLK-xSox11 complex and what upstream factors may mediate this signalling. The relationship between NLK and Sox/TCF molecules awaits a more detailed analysis, as does their precise roles in Wnt signalling.
cDNA cloning and construct
A Xenopus oocyte cDNA library in the Lambda Zap vector was screened using random-primed 32P-labelled mouse NLK cDNA as probe. About 1.5 × 106 plaques were hybridized at 60 °C on nylon filters, and subsequently washed in 1 × SSC, 0.1% SDS, at 60 °C. Positive clones were initially isolated and subcloned into the pBluescript vector. Full-length cDNAs were identified and sequenced on both strands by primer walking. A kinase-negative (KN) mutant allele of xNLK was generated by site-directed mutagenesis using the polymerase chain reaction (PCR). In xNLK-KN, the critical lysine at 89 in the ATP-binding site was mutated to arginine.
Whole-mount in situ hybridization
Whole-mount in situ hybridization was carried out essentially as described (Harland 1991). Digoxygenin-labelled riboprobes were prepared from a pBluescript SK(-) plasmid containing the entire xNLK and xSox11 cDNAs. Transcripts were detected using the BM-purple substrate (Boehringer).
Animal cap assay
The xNLK, xNLK-KN and xSox11 cDNAs were cloned into the CS2+ vector (Rupp et al. 1994). Capped mRNA was synthesized from linearized vectors using the mMessage Machine kit (Ambion). mRNAs were then injected into the animal poles of early stage embryos. The amounts of injected in vitro synthesized mRNAs and sites of injection are described in the text. Animal cap explants were removed with hair knives at late blastula stages and allowed to grow until control sibling embryos reached either the gastrula or neural stages. Total RNA was then extracted and analysed with RT-PCR. RT-PCR assays and primer sequences were as published (Hemmati-Brivanlou & Melton 1994; Sasai et al. 1995; Wilson & Hemmati-Brivanlou 1995). The sequences of the primer pairs for xNLK were follows; forward primer: 5′-TTTACCCCAAGCTCTCAGC-3′, reverse primer: 5′-GAGGAAACCTTGATGTGG-3′.
Antibodies, immunoprecipitation and Western blot
The monoclonal antibody against the T7 epitope was purchased from Novagen. The anti-Flag monoclonal antibody was purchased from Eastman Kodak Co. Two hundred and ninety-three cells were transiently transfected with the indicated constructs by the calcium phosphate method. Forty-eight hours after transfection, cells were lysed in lysis buffer (50 mm Tris-HCl, pH 7.5, 150 mm NaCl, 5 mm EDTA, 0.5% NP-40, 50 mm NaF, 1 mm dithiothreitol, 1 mm sodium orthovanadate, 1 mm (p-amidinophenyl) methanesulphonyl fluoride-HCl, 10 µg/mL aprotinin, 10 µg/mL pepstatin, and 20 µg/mL leupeptin). For co-immunoprecipitation, lysates were incubated with 10 µL of protein A-Sepharose beads to preclear and then incubated with the appropriate antibodies coupled to protein A-Sepharose beads at 4 °C for 1 h. Immunoprecipitates were washed four times with 200 µL of lysis buffer and analysed by Western blot analysis with the appropriate antibodies.
We thank Marc Lamphier for critical reading of the manuscript. We are grateful to Yoriko Oda for her excellent assistance. This work was supported by Grants-in-Aid for scientific research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan, the Research Grant for Longevity Sciences from the Ministry of Health, Labour and Welfare of Japan, Mitsui Life Social Welfare Foundation and Yamanouchi Foundation for Research on Metabolic Disorders.