PTK7 recruits dsh to regulate neural crest migration

Wnt signaling controls diverse developmental processes such as axis formation, anteroposterior patterning and the development of the neural crest(Moon et al., 1997; Raible and Ragland, 2005; Wodarz and Nusse, 1998). Precise regulation is necessary for many developmental processes and misregulation of several components of the canonical Wnt signaling pathway has been implicated in cancer formation(Polakis, 2007). Conversely,reduction or loss of function of Wnt signaling leads to general developmental defects or loss of organs, demonstrating the need for a tight regulation of the levels of Wnt signaling (Logan and Nusse, 2004).

The Wnt signaling pathway is highly conserved among all metazoans. Wnt ligands bind to frizzled (fz) transmembrane receptors leading to accumulation and nuclear localization of β-catenin, which serves as a transcriptional co-activator for TCF/Lef transcription factors(Logan and Nusse, 2004; Wodarz and Nusse, 1998). In addition to this so called `canonical' Wnt signaling pathway, fz receptors also activate alternative signaling pathways like the planar cell polarity(PCP) pathway, which defines the orientation of cells in the plane of an epithelium (Klein and Mlodzik,2005; Seifert and Mlodzik,2007). PCP signaling has been best characterized in Drosophila, where it determines for example the ommatidia organization in the eye and the bristle hair orientation in the wing(Axelrod and McNeill, 2002; Klein and Mlodzik, 2005). In vertebrates, PCP signaling is necessary for the orientation of the stereocilia bundles in the neurosensory epithelium of the inner ear, and dynamic convergent extension movements during gastrulation and neurulation(Wallingford et al., 2002; Wang and Nathans, 2007).

Dishevelled (dsh) is a key regulator of both the canonical as well as the PCP signaling pathway and regulates cell fate specification as well as cell movements such as convergent extension. Dsh consists of three major conserved domains, the DIX, PDZ and DEP domain, that have been implicated in different downstream signaling events (Boutros and Mlodzik, 1999; Wallingford and Habas, 2005). The DIX domain is used for canonical Wnt signaling,whereas the DEP domain is involved in PCP signaling(Boutros et al., 1998; Habas et al., 2003; Itoh et al., 2000; Rothbacher et al., 2000). By contrast, the PDZ domain is shared by both pathways. As downstream effectors of dsh are distinct for canonical and PCP signaling, it remains unclear how dsh selectively activates one or the other pathway. In the vertebrate PCP pathway, the signaling mechanisms affecting the subcellular localization and choice of effectors proteins of dsh are not well defined.

PTK7 (protein tyrosine kinase 7) is a regulator of PCP signaling that could modulate the dsh localization as well as the interaction with pathway-specific effector proteins. PTK7 regulates PCP in the inner ear hair cells and during neural tube closure in mice (Lu et al.,2004). In Xenopus, PTK7 is required for neural convergent extension (Lu et al., 2004). PTK7 is a transmembrane protein containing seven extracellular immunoglobulin domains and a kinase homology domain. Although the kinase domain lacks the DFG triplet necessary for catalytic activity, its overall structure is evolutionary conserved from Hydra to humans(Kroiher et al., 2001; Miller and Steele, 2000). The signaling mechanism of PTK7 has not been characterized, but considering the structure and function of PTK7, it could affect the localization of dsh as well as its downstream signaling.

Neural crest migration is a new system to analyze the mechanistic role of PTK7 in vertebrate PCP signaling. Neural crest cells are induced at the border region of the neural plate and migrate on defined routes throughout the embryo, where they give rise to a variety of derivatives ranging from neurons and glia cells of the peripheral nervous system, to cartilage and pigment cells. Although canonical Wnt signaling plays a role in the induction,delamination and differentiation of neural crest cells(Schmidt and Patel, 2005; Yanfeng et al., 2003), PCP signaling seems to be required for neural crest migration. De Calisto et al. have shown that a dsh mutant lacking PCP activity inhibits neural crest migration. Conversely, inhibition of neural crest migration by loss of Wnt11 function can be rescued by expressing a dsh mutant, which activates PCP signaling in neural crest cells (De Calisto et al., 2005). Furthermore, `core-PCP signals' such as Van Gogh/strabismus, prickle and daam are expressed in cranial neural crest cells(Bekman and Henrique, 2002; Darken et al., 2002; Goto and Keller, 2002; Nakaya et al., 2004),indicating that the PCP signaling cascade is active during neural crest migration.

Here, we identify a new function for PTK7 in neural crest migration. By analyzing the signaling mechanism of Xenopus PTK7 in vitro in animal cap explants and in vivo in migrating neural crest cells, we provide evidence that PTK7 localizes dsh to the plasma membrane and that this function is required for neural crest migration.

Myc-tagged PTK7 (PTK7-myc) was amplified by PCR (forward primers 5′CGGGATCCATGGGGCCGATTGTGCTC3′, reverse primers 5′CCATCGATACCCTTGTGTCTTGCTGCC3′), cut with BamHI and ClaI, and ligated into the respective restriction sites of pCS2-MT. HA-tagged PTK7 (PTK7-HA) was obtained from PTK7-myc by removing the myc-tag with ClaI and XbaI, and inserting an HA-tag generated using the following primers: forward primer,5′CGATATCCCTACGATGTTCCAGATTATGCATGATAA3′; reverse primer 5′CTAGTTGCGTAATCCGGTACATCGTAAGGGTAGT3′. The deletion of the predicted kinase domain of PTK7 (ΔkPTK7-myc) was introduced by PCR amplification of PTK7-myc using the following primers: forward,5′CTTGTCGCCAGAGCTGTGTC3′; reverse,5′TCTTCTGGCAGCAAGACACAAG3′. RNA transcribed from PTK7 constructs cloned in pCS2 is significantly more potent than RNA generated from the constructs published by Lu et al. (Lu et al., 2004); therefore lower concentrations were injected.

For neural crest-specific expression, a minimal slug promoter,a700BA (Vallin et al., 2001),was excised with SpeI/BamHI from a700BA-GFP(Vallin et al., 2001) and ligated into the SpeI/BamHI sites of BS-I-Sce-II KS(Pan et al., 2006) generating slug-BS-I-Sce-II KS. To express PTK7 under the control of the slug promoter, full-length PTK7 was excised using Sal/BspE1,blunted and ligated into the BamHI site of slug-BS-I-Sce-II KS generating slug-PTK7. For cloning of the kinase deletion mutant(slug-ΔkPTK7), full-length PTK7 was cut with SalI/Bsu361, blunted and ligated into the BamHI site of slug-BS-I-Sce-II KS. To express GFP in neural crest cells,the minimal slug promoter and the GFP-coding sequence were excised with SpeI/NotI from a700BA-GFP(Vallin et al., 2001) and inserted into the SpeI/NotI sites of BS-I-Sce-II KS. All constructs cloned in slug-BS-I-Sce-II KS were purified using QIAprep Spin Miniprep Kit (Qiagen) previous to injection in Xenopusembryos.

Xenopus injection experiments were performed as previously described (Borchers et al.,2001). The PTK7 MO sequences were published elsewhere(Lu et al., 2004). Here, a 1:1 mixture of MO2 and MO3 was used at total concentrations indicated in the text. As a toxicity control, identical concentrations of control MO (Gene Tools)were used. To evaluate neural crest migration independently of earlier neural tube closure defects, PTK7 MO injected embryos with severe neural tube closure phenotypes were excluded.

For RNA injection experiments, capped mRNA was synthesized using the mMESSAGE mMACHINE Kit (Ambion) according to the manufacturer's instructions. The following published plasmids were used to synthesize RNA for dsh-GFP(Yang-Snyder et al., 1996),dshΔDIX-GFP, dshΔPDZ-GFP, dshΔDEP-GFP(Miller et al., 1999), dsh-myc(Sokol, 1996), fz7-myc(Winklbauer et al., 2001) and lacZ (Smith and Harland,1991). For in situ hybridization, embryos were injected with RNA,plasmids or MO in one blastomere at the two-cell stage. GFP or lacZ RNA were used as lineage tracers as indicated in the text. To analyze protein localization in animal cap assays, embryos were injected in the animal pole at the one-cell stage.

Transplantation of neural crest cells was carried out as previously described (Borchers et al.,2000). Images were taken using a Leica MZFLIII microscope and a Leica DC500 camera (Leica Fire Cam 1.2.0 software for Macintosh).

β-Galactosidase staining and whole-mount in situ hybridization were performed as previously described (Borchers et al., 2002; Harland,1991). Antisense probes were synthesized from the following published plasmids using a DIG-RNA labeling kit (Roche): AP-2(Winning et al., 1991), engrailed (Brivanlou and Harland,1989), twist (Hopwood et al., 1989) and Sox10(Aoki et al., 2003). The PTK7 antisense probe was generated from full-length PTK7 (RZPD, IMAGp998I099552)linearized with SexA1.

For immunostaining on vibratome sections, embryos were fixed for 1 hour in MEMFA (3.7% formaldehyde, 0.1 M MOPS, 2 mM EGTA, 1 mM MgSO₄). Sectioning and immunostaining was performed as previously described(Borchers et al., 2000). Myc-tagged PTK7 was detected with anti-c-myc-Cy3 antibody (C6594, Sigma) and sections were imaged using an Axioplan 2 microscope (Zeiss) equipped with an AxioCam HRc digital camera.

Ectodermal explants (animal caps) were prepared as described by Wallingford and Harland (Wallingford and Harland,2001). At stage 10.5-11, animal caps were fixed in MEMFA and processed for immunostaining. Antibodies were used in the following dilutions:anti-c-myc-Cy3 (Sigma, C6594) 1:100, anti-HA (MMS-101P, Covance) 1:150,anti-GFP (AB290, Abcam) 1:1000, goat anti-mouse Alexa-fluor 488 nm (A11029,Invitrogen) 1:200, goat anti-mouse Cy5 (AB6563, Abcam) 1:200, and goat anti-rabbit IgG fluorescein isothiocyanate (FITC) (F7367, Sigma) 1:200. After immunostaining, protein localization was imaged by confocal microscopy using a LSM510 META confocal microscope with LSM510 META software (Carl Zeiss,MicroImaging).

For immunoprecipitation, 50 stage 11 embryos were homogenized in 500 μl NOP buffer (10 mM TrisHCl pH 7.8, 150 mM NaCl, 5% NP40) supplemented with protease and phosphatase inhibitors to a final concentration of 1 mM PMSF, 0.1 mM NaVO₄, 1 mM NaF and 1 mM β-glycerolphosphate. Inhibitor tablets (Roche) were added according to the manufacturer's instructions. Lysates were centrifuged at 16,000 g for 20 minutes at 4°C and the supernatant was pre-cleaned by a 30 minute incubation with Protein A sepharose (Amersham). After incubation with anti-myc antibodies (9E10, Sigma)for 1.5 hours at 4°C, protein A sepharose (15 μl of the bead volume)was added for 1 hour and protein complexes were precipitated by low-speed centrifugation (400 g). Sepharose beads were washed three times with NOP buffer, boiled in Laemmli sample buffer and analyzed by western blotting using anti-HA antibodies (Covance).

PTK7 is necessary for disparate processes regulated by PCP signaling such as neural tube closure and hair cell polarity in the inner ear(Lu et al., 2004); however,its precise signaling mechanism is unknown. Although the genetic interaction of PTK7 with Vangl2 (Lu et al.,2004) suggests that PTK7 is an integral part of the PCP signaling cascade, so far its intersection with it has not been defined. One possibility is that PTK7 signals via dsh, which like PTK7 regulates neural convergent extension (Wallingford and Harland,2001; Wang et al.,2006). A prerequisite for this would be that dsh co-localizes with PTK7 at the plasma membrane. To analyze whether this is the case, we employed the Xenopus animal cap assay, which has previously been used to demonstrate the interaction and co-localization of dsh with fz at the plasma membrane (Axelrod et al., 1998; Medina et al., 2000; Medina and Steinbeisser, 2000; Sheldahl et al., 1999; Yanagawa et al., 1995). Animal caps are ectodermal explants of blastula stage Xenopus embryos that allow the analysis of the cellular localization of overexpressed tagged proteins. For colocalization assays, Xenopus embryos were injected with RNA coding for the respective tagged constructs in the animal pole of a one-cell stage embryo. At blastula stage, the animal cap was dissected and cultured until gastrula stages, when cellular protein localization was determined using a laser scanning microscope. GFP-tagged dsh is predominantly localized in the cytoplasm of animal caps(Fig. 1A,G), but is translocated to the plasma membrane if myc-tagged fz7 is co-expressed(Fig. 1C,G). To examine whether PTK7 translocates dsh to the plasma membrane, myc-tagged PTK7 was co-expressed with GFP-tagged dsh. Like fz7, PTK7 can recruit dsh-GFP from the cytoplasm to the membrane (Fig. 1E,G). To further confirm the co-localization of PTK7 with dsh in a biochemical assay,we used glycerol gradient density centrifugation (see Fig. S1 in the supplementary material). This method allows the separation of proteins according to their molecular weight, and protein complex formation can be detected as a shift to higher molecular weight fractions. Indeed, the dsh protein tailed into higher density fractions in the presence of PTK7 (see Fig. S1 in the supplementary material), indicating complex formation of dsh with PTK7. Thus, the animal cap localization assay as well as the glycerol gradient density centrifugation indicate that PTK7 connects with the Wnt signaling pathway at the level of dsh.

To identify the molecular domains necessary for the interaction of PTK7 with dsh, deletion mutants were tested for co-localization in animal caps. As the intracellular domain of PTK7 contains a conserved tyrosine kinase motif,we analyzed whether its deletion abolishes translocation of dsh. Indeed, in contrast to overexpression of wild-type PTK7, this deletion mutant(ΔkPTK7) was not able to translocate dsh to the plasma membrane(Fig. 1F,G). Furthermore, this mutant also failed to shift dsh to higher molecular weight fractions in Xenopus lysates separated by glycerol gradient centrifugation (see Fig. S1C,D in the supplementary material). Taken together, this indicates that the kinase motif of PTK7 is required for dsh recruitment to the plasma membrane.

Dsh contains different functional domains that are involved in canonical and non-canonical Wnt signaling, respectively(Wallingford and Habas, 2005). To identify which of these domains are necessary for the PTK7-dependent dsh-translocation, we expressed GFP-tagged deletion mutants of the DIX, the PDZ and the DEP domain of dsh (Fig. 2A) in animal caps. In the absence of PTK7 expression, all mutant dsh proteins were mainly localized in the cytoplasm(Fig. 2B,D,F,H). However,co-expression of PTK7 transferred ΔDIX- as well as ΔDEP-dsh to the membrane (Fig. 2C,G,H). In the case of the ΔDIX-dsh-injected caps, a residual cytoplasmic staining remained that was not apparent in ΔDEP-dsh injected caps. By contrast,ΔPDZ-dsh was not translocated to the plasma membrane in the presence of PTK7 (Fig. 2E,H), indicating that the PDZ domain is required for function. In summary, these data show that the tyrosine kinase domain of PTK7 as well as the PDZ domain of dsh are necessary for the translocation of dsh to the plasma membrane.

The ability of PTK7 to control dsh localization suggests that the two proteins might interact. We tested for binding by co-expressing HA-tagged PTK7 with myc-tagged dsh and immunoprecipitating with either myc- or HA-antibodies. Independent of the antibodies used, we could not detect co-immunoprecipitation of PTK7 and dsh (data not shown), indicating that additional molecules are required for PTK7-mediated dsh localization. A likely candidate is fz7, which,like PTK7, is also able to recruit dsh to the plasma membrane. To test whether PTK7 forms a complex with dsh and fz7, we expressed HA-tagged PTK7 with either myc-tagged dsh or myc-tagged fz7, or a combination of the two (see Fig. S2 in the supplementary material). Although HA-tagged PTK7 was co-precipitated with myc-tagged fz7 in one out of three experiments, we detected only robust co-precipitation in combination with myc-tagged dsh and myc-tagged fz7 (see Fig. S2 in the supplementary material), indicating that PTK7 is part of a protein complex that includes fz7 and dsh.

If PTK7 forms a complex with fz7 and dsh, PTK7 could affect the fz7-mediated dsh recruitment. To test this, we analyzed animal caps, which endogenously express PTK7 (see Fig. S3 in the supplementary material), how loss of PTK7 function interferes with the fz7-mediated dsh recruitment. In the presence of control morpholino oligonucleotides (MOs), fz7 recruits GFP-tagged dsh to the plasma membrane (Fig. 3A,D). However, the PTK7 MO abolished the fz7-mediated dsh translocation (Fig. 3B,D). This could be rescued by co-injection of PTK7 RNA lacking the MO binding site (Fig. 3C,D).

As fz-mediated dsh membrane localization correlates with hyperphosphorylation of dsh (Rothbacher et al., 2000), we further examined whether PTK7 loss of function also affects the phosphorylation status of dsh. Animal caps expressing myc-tagged dsh alone or in combination with PTK7 show only a single band in western blots using anti-myc antibodies. However, an additional high molecular weight band representing hyperphosphorylated dsh is detected in lysates expressing dsh with fz7 (Fig. 3E). Interestingly, this fz7-mediated hyperphosphorylation of dsh is inhibited by the PTK7 MO, indicating that PTK7 is required for the fz7-mediated dsh hyperphosphorylation. Thus, these data support that PTK7 is part of a dsh-fz7 complex required for dsh localization and phosphorylation.

To further evaluate the in vivo relevance of the PTK7-mediated dsh membrane localization, we focused on Xenopus neural crest migration. Recently,PCP signaling has been implicated in the regulation of neural crest migration(De Calisto et al., 2005). As PTK7 is expressed in premigratory (Fig. 4A,B) as well as migratory cranial neural crest cells(Fig. 4C), we analyzed whether PTK7 also functions in neural crest development. Xenopus laevisembryos were injected with PTK7 MO and GFP RNA as a lineage tracer in one blastomere at the two-cell stage and neural crest migration was analyzed at neurula and tadpole stages using in situ hybridization for different neural crest markers. Starting at the neurula stage, injection of the PTK7 MO inhibited neural crest migration. Although twist-positive cells were induced, they failed to migrate in PTK7 MO-injected embryos(Fig. 4H), whereas induction and migration of twist-positive cells was normal in embryos injected with the control MO (Fig. 4F). At tadpole stages, a few migrating neural crest cells are found in the PTK7 MO-injected embryos; however, their number as well as their migration distance is dramatically reduced compared with the control(Fig. 4J,K). Similar migration defects are also seen with other neural crest markers such as AP-2(Fig. 4L,M) and Sox10 (data not shown), whereas the expression of the midbrain-hindbrain marker engrailed is not affected at these MO concentrations(Fig. 4I, right embryo). Thus,the PTK7 MO does not lead to morphological changes, but rather seems to specifically affect neural crest development.

To verify that the neural crest migration defect is caused by loss of PTK7 function, we performed rescue experiments. Embryos were injected with PTK7 MO in combination with increasing concentrations of wild-type PTK7 RNA lacking the MO-binding sites. Embryos were injected in one blastomere at the two-cell stage using lacZ RNA as a lineage tracer and neural crest migration was analyzed by twist in situ hybridization at tadpole stages. Fig. 3O summarizes a total of three injection experiments showing a concentration-dependent rescue effect of the PTK7 MO phenotype by wild-type PTK7 RNA. This demonstrates that the MO phenotype is specific to loss of PTK7 function.

In contrast to loss of PTK7 function, overexpression of PTK7 did not affect neural crest migration (Fig. 5). Different concentrations of myc-tagged and untagged PTK7 RNA were injected in one blastomere at the two-cell stage and neural crest migration was analyzed at neurula and tadpole stages. Neither neural crest induction nor neural crest migration were affected by overexpression of PTK7 (Fig. 5A-D), although the tagged PTK7 protein was detected by western blotting and immunostaining (Fig. 5E,F). Thus, PTK7 is required for neural crest migration; however,an excess of the protein seems not to disturb neural crest development. In summary, these data show that PTK7 is expressed and probably directly required in neural crest cells to enable their migration.

To address the function of PTK7 specifically in migrating neural crest cells, we used neural crest transplantation assays(Borchers et al., 2000). Embryos were co-injected with PTK7 MO and GFP RNA in one blastomere at the two-cell stage. At early neurula stages (14-16), the fluorescent neural crest was removed and transplanted into control embryos from which the local neural crest had been removed (Fig. 6A). To distinguish `non-migrating' grafts(Fig. 6D) from `migrating'grafts (Fig. 6B) the GFP fluorescence of the transplanted cells was monitored at different time points. Five hours after transplantation, grafts injected with the control MO showed streams of migrating neural crest cells(Fig. 6B,F). However, most grafts injected with the PTK7 MO did not migrate(Fig. 6D,F). Twelve hours after grafting, the number of migrating PTK7 MO grafts(Fig. 6E) increased, but was still significantly lower compared with control MO grafts(Fig. 6F). Furthermore, even if PTK7 MO grafts migrated, they showed fewer migrating cells and did not migrate as far to the ventral side as did control grafts(Fig. 6C,E). The PTK7 MO phenotype is rescued by co-injection of wild-type PTK7 RNA lacking the MO binding sites (Fig. 6F),although the rescue effect is not as pronounced as in whole embryo injection experiments, which is probably due to the more challenging experimental procedure. In summary, the transplantation assay shows that transplanted neural crest cells require PTK7 for migration.

To study the function of PTK7 in neural crest migration isolated from earlier developmental defects, we targeted the expression of PTK7 to neural crest cells. PTK7 constructs were expressed under the control of the neural crest-specific slug promoter, so that only the cells that are already specified to become neural crest cells are affected. Regions of the slug promoter that are sufficient to drive neural-crest-specific expression (Vallin et al.,2001) were used to express GFP(Fig. 7A,B), full-length PTK7(Fig. 7C,D) and a deletion mutant of the kinase motif, ΔkPTK7(Fig. 7E,F). ΔkPTK7 was used as a PTK7 antimorph, as it failed to recruit dsh to the plasma membrane in animal cap assays (Fig. 1F)and caused similar neural crest migration defects as the PTK7 MO in whole embryos (injection of 50 pg ΔkPTK7 RNA caused 44% neural crest migration defect, n=122). Plasmids were injected together with lacZ RNA as a lineage tracer in one blastomere at the two-cell stage and embryos were analyzed for twist expression at neurula and tadpole stages. As plasmid injections result in mosaic expression, the effects are generally not as strong as MO or RNA injections. Nevertheless, neural crest-specific expression of ΔkPTK7 significantly inhibited the migration of twist-expressing neural crest cells (Fig. 7E,F), whereas the expression of GFP or PTK7 rarely affected neural crest migration (Fig. 7A-D,G). Although we cannot completely rule out the possibility that PTK7 does also affect neural crest induction, these data demonstrate that PTK7 has an independent function in neural crest migration.

Having established that PTK7 is required for neural crest migration, we can use this system to analyze the in vivo relevance of the PTK7-mediated dsh membrane localization. By targeting the kinase deletion mutant of PTK7,ΔkPTK7, to neural crest cells, we have already demonstrated that failure of dsh localization (Fig. 1F,G)inhibits neural crest migration (Fig. 7F,P). Next, to characterize the genetic interaction of PTK7 with dsh in migrating neural crest cells, we targeted the expression of PTK7, as well as of different dsh mutants, to neural crest cells using the slug promoter system. First, we analyzed how dsh, ΔPDZ dsh andΔDEP dsh affect neural crest migration. We used ΔPDZ dsh, because this mutant abolishes the PTK7-mediated translocation, and ΔDEP dsh,because the DEP domain is required for PCP activity and affects neural crest migration (De Calisto et al.,2005; Wallingford and Habas,2005). Single injections of all dsh constructs and a single injection of slug-PTK7 produced only very mild defects(Fig. 7H,I,P). The same holds true for co-injection of slug-PTK7 with slug-dsh or slug-ΔPDZ (Fig. 7L-P). However, the combination of slug-PTK7 with slug-ΔDEP inhibited neural crest migration to an extent that is comparable with expression of slug-ΔkPTK7(Fig. 7J,K,P), which cannot recruit dsh to the plasma membrane. This suggests that, although ΔDEP dsh can be recruited by PTK7, its lack of PCP activity inhibits neural crest migration. Co-expression of wild-type dsh, which can be translocated but does not lack the DEP domain, does not affect neural crest migration. The same holds true for ΔPDZ, which is not translocated by PTK7. Thus, these data indicate that PTK7 is required to localize a functional DEP domain to the plasma membrane to enable neural crest migration.

PCP signaling in vertebrates regulates complex cell movements, but the molecular mechanisms that guide these processes are not well understood. Here,we show that PTK7, a regulator of vertebrate PCP signaling, is required for neural crest migration, and we provide in vitro as well as in vivo evidence that PTK7 regulates PCP signaling by recruiting dsh to the plasma membrane.

Neural crest cells migrate extensively on defined routes throughout the embryo to give rise to a range of derivatives. However, the signaling mechanisms guiding these organized cell movements are not well characterized. As overexpression of dsh mutants, which are defective in PCP signaling,inhibits neural crest migration in Xenopus, this indicates that PCP signaling plays a role (De Calisto et al.,2005). Although, most PCP effectors are expressed at the right time and location for a function in neural crest migration(Bekman and Henrique, 2002; Darken et al., 2002; Goto and Keller, 2002; Nakaya et al., 2004), their function in this process has so far not been analyzed. Here, we identify PTK7 as the first regulator of PCP with a function in neural crest migration. Loss of function of PTK7 inhibits neural crest migration, and transplantation assays show that neural crest cells directly require PTK7 for migration. Furthermore, this function is independent of the role of PTK7 in neural tube closure, as demonstrated by neural crest-specific expression.

The data presented here, suggests that PTK7 regulates neural crest migration by recruiting dsh to the plasma membrane. Dsh occupies a key position at the crossroad of all branches of the Wnt signaling cascade and several lines of evidence indicate that PTK7 channels dsh signaling towards PCP signaling. First, PTK7 affects processes regulated by PCP signaling such as convergent extension/neural tube closure and inner ear hair cell polarity in vertebrates (Lu et al.,2004). Second, PTK7 displays heteroallelic interaction with another regulator of PCP signaling, Vangl2 (Looptail mice)(Lu et al., 2004). Third, PTK7 does not activate β-catenin signaling, but does activate JNK signaling. Neither in TOPFLASH reporter assays (data not shown) nor in Xenopuslaevis double axis assays (0% in n=168) does PTK7 activate canonical Wnt signaling. However, PTK7 overexpression in Xenopus animal caps activates the phosphorylation of JNK (see Fig. S4 in the supplementary material), indicating PCP activity (Boutros et al., 1998; Habas et al.,2003; Li et al.,1999; Yamanaka et al.,2002). Thus, these data suggest that PTK7-mediated dsh recruitment activates PCP signaling.

In vivo evidence supports a model in which PTK7 recruits dsh to the plasma membrane, thereby regulating neural crest migration(Fig. 8A). This is experimentally supported by two observations. First, the PTK7 MO as well asΔkPTK7, which both abolish PTK7-dependent dsh-recruitment, inhibit neural crest migration (Fig. 8B). Second, overexpression of wild-type slug-PTK7 together with slug-ΔDEP dsh, which is deficient in PCP activity, but can be recruited to the membrane, inhibits neural crest migration to the same extent as slug-ΔkPTK7(Fig. 8C). Hence, if the DEP domain is missing or dsh cannot be localized to the plasma membrane, neural crest migration is inhibited (Fig. 8B,C). In contrast to RNA injection experiments, slug-driven expression of single dsh mutants did not affect neural crest migration. This could be caused by the late onset as well as the mosaic nature of the expression. Thus, overexpressed dsh proteins cannot effectively compete with endogenous dsh, unless ectopic PTK7 is simultaneously provided. Conversely,excess PTK7 does not affect neural crest migration unless it translocatesΔDEP dsh to the plasma membrane. Thus, PTK7 may function by anchoring dsh to the plasma membrane, which then regulates neural crest migration via its DEP domain probably by activating JNK signaling.

The mechanism by which PTK7 mediates dsh recruitment is unclear. Our data suggest that PTK7 is part of the fz-dsh complex that localizes dsh to the plasma membrane. First, PTK7 is co-precipitated with fz7/dsh in Xenopus lysates and HEK293 cells (data not shown). Second, PTK7 loss of function inhibits the fz7-mediated dsh recruitment and hyperphosphorylation. Alternatively, as the binding activity of PTK7/fz7/dsh is weak, PTK7 may contribute to the fz/dsh interaction at the signaling level possibly by promoting dsh phosphorylation. As the kinase domain of PTK7 lacks amino acid residues crucial for kinase activity(Kroiher et al., 2001; Miller and Steele, 2000), this scenario could involve additional PTK7-binding partners that mediate dsh phosphorylation. Independent of the molecular nature of the PTK7 interaction with fz/dsh - be it binding, signaling or both - it also remains to be seen whether PTK7 exclusively interacts with fz7 or whether other fz family members can replace it. The latter is supported by the finding that animal caps express other fz family members, including fz8 (see Fig. S3 in the supplementary material), which, like fz7, recruits dsh to the plasma membrane(Rothbacher et al., 2000). Furthermore, loss of function of fz7 does not affect the PTK7-mediated dsh localization in animal caps (data not shown), suggesting that either PTK7 can recruit dsh independently of fz7 or indeed that other fz family members can also interact with PTK7. As fz7 is expressed at the right time and place for a function in neural crest migration (De Calisto et al., 2005), it probably takes part in the in vivo PTK7-fz-dsh interaction. Future experiments will be needed to address the molecular composition, as well as the function, of the PTK7-fz-dsh interaction in neural crest migration.

Our findings supports the notion that PCP signaling is not limited to epithelia, but is also found in moving mesenchymal tissues such as migrating neural crest cells (Kuriyama and Mayor,2008). Neural crest cells migrate over long distances, but also stay in contact with each other (Teddy and Kulesa, 2004). Thus, PCP signaling may pattern migrating cells or orient cell protrusions. Indeed, PCP factors determine the polarization of neural crest cells by lamellipodia formation(De Calisto et al., 2005; Kuriyama and Mayor, 2008) and PTK7 could mediate this by activating dsh-JNK signaling at precise cellular locations. Additionally, as PTK7 contains extracellular immunoglobulin domains, which have been implicated in homophilic binding(Lu et al., 2004; Pulido et al., 1992), PTK7 may affect cell clustering or cell contact persistence. The latter was shown for fz7, which modulates local cell contact persistence to coordinate cell movements during zebrafish gastrulation(Witzel et al., 2006). Furthermore, PTK7 could also function as a receptor for signals that guide migrating neural crest cells. Indeed, functional analyses of the Drosophila PTK7 ortholog (otk, off-track) indicate that PTK7 is a co-receptor for semaphorin/plexins(Winberg et al., 2001), which are guidance cues for migrating neural crest cells in vertebrates(Gammill et al., 2007; Gammill et al., 2006; Yu and Moens, 2005). The next challenge will be to further dissect the signaling mechanism of vertebrate PTK7 and analyze whether it can also respond to extracellular signals to guide neural crest migration.

We thank R. Moon and H. Steinbeisser for providing plasmids, D. Frank, A. Hufton, H. Peradziryi and M. Podleschny for critical reading of the manuscript, and H. Steinbeisser and A. Wodarz for helpful critical discussion. Furthermore, we thank H. Steinbeisser for the fz7 MO and A. Wodarz for providing the laser scanning microscope and technical support, and I. Wunderlich for technical assistance. Finally, we thank T. Pieler for infrastructure and support. This work was supported by grants to A.B. by the Deutsche Forschungsgemeinschaft (DFG) and the DFG Center for Molecular Physiology of the Brain (CMPB). I.S. was supported by a stipend from the Christiane-Nüsslein-Volhard-Foundation.