The dorsal ectoderm of the limb bud is known to regulate anterior–posterior patterning as well as dorsal– ventral patterning during vertebrate limb morphogenesis. Wnt-7a, expressed in the dorsal ectoderm, encodes a key molecule implicated in these events. In the present study, chicken frizzled-10 (Fz-10) encoding a Wnt receptor was used to study mechanisms of Wnt-7a signaling during chick limb patterning, because its expression is restricted to the posterior-distal region of the dorsal limb bud. Fz-10 transcripts colocalize with Sonic hedgehog (Shh) in the dorsal side of stages 18–23 chick limb buds. It was demonstrated that Fz-10 interacts with Wnt-7a to induce synergistically the expression of Wnt-responsive genes, such as Siamois and Xnr3, in Xenopus animal cap assays. In the chick limb bud, Fz-10 expression is regulated by Shh and a signal from the dorsal ectoderm, presumably Wnt-7a, but not by signals from the apical ectodermal ridge. These results suggest that Fz-10 acts as a receptor for Wnt-7a and has a positive effect on Shh expression in the chick limb bud.
While much is known about tissue interactions that regulate patterning and outgrowth of the vertebrate limb bud, much remains unknown about the precise molecular mechanisms mediating these tissue interactions ( Tickle & Eichele 1994; Johnson & Tabin 1997). During vertebrate limb development, specialized structures termed signaling centers are established. Secreted factors are produced from the signaling centers and mediate tissue interactions essential for limb outgrowth and patterning ( Johnson & Tabin 1997). Wnt-7a, a member of the Wnt family of secretory proteins, is produced by the entire dorsal ectoderm and determines the dorsal–ventral patterning of the limb mesenchyme by inducing the transcription factor Lmx1 in the dorsal mesenchyme ( Parr & McMahon 1995; Riddle et al. 1995 ; Vogel et al. 1995 ; Chen et al. 1998 ). In addition, Wnt-7a is involved in the maintenance of Sonic hedgehog (Shh) expression ( Parr & McMahon 1995; Yang & Niswander 1995) and thus is involved in anterior–posterior patterning. Shh, expressed in the zone of polarizing activity (ZPA) located at the posterior-distal region of the limb mesenchyme, is essential for anterior–posterior patterning and maintaining limb outgrowth ( Riddle et al. 1993 ; Chiang et al. 1996 ; Hammerschmidt et al. 1997 ). Removal of Wnt-7a by gene targeting or surgical manipulation results in reduced Shh expression in addition to loss of Lmx1 expression ( Parr & McMahon 1995; Yang & Niswander 1995). Wnt-7a null mice exhibit truncations of the distal elements, which are also observed in Shh–/– mice ( Chiang et al. 1996 ). Implanting Wnt-7a-expressing cells to the posterior region of limb bud after dorsal ectoderm removal rescues Shh expression ( Yang & Niswander 1995). Although these reports show that Wnt-7a is required to maintain Shh expression in the ZPA, the mechanism by which Wnt-7a from the dorsal ectoderm influences the posterior Shh expression is not well understood.
The Wnt family is known to act through frizzled (Fz) receptors. The Fz proteins have an N-terminal extracellular cysteine-rich domain that possibly binds to Wnt proteins, a seven-pass transmembrane domain and a C-terminal cytoplasmic tail domain. To date, 10 members of the Fz family have been identified in vertebrates and several members are expressed during limb development ( Kengaku et al. 1997 ; Nohno et al. 1999 ; Kawakami et al. 2000). The interactions between Wnt members and Fz members have been examined in several experimental contexts, including in vitro binding assays, genetic studies and Xenopus development ( Bhanot et al. 1996 ; Yang-Snyder et al. 1996 ; He et al. 1997 ; Rocheleau et al. 1997 ; Liu et al. 1999 ). However, in order to understand the mechanisms of Wnt signaling during limb development, it is necessary to identify the Fz receptors involved in Wnt signaling during limb development.
Recently, we have identified chick Fz-10, which is expressed in the posterior-distal mesenchyme of the developing limb bud (Kawakami et al. 2000). The expression suggested a possible interaction between Fz-10 and Wnt-7a. In the present paper, we have analyzed the expression of Fz-10 to characterize its role in limb development. Several lines of evidence suggest that Fz-10 is involved in a positive interaction between Wnt-7a and Shh in the limb bud, by acting as a Wnt-7a receptor.
Materials and Methods
Cloning of chick Wnt-7a and Wnt-3a
When we performed reverse transcriptase–polymerase chain reaction (RT-PCR) on chick cDNA obtained from stages 18–21 embryos ( Tanda et al. 1995 ), we obtained Wnt-7a and Wnt-3a cDNA fragments. The entire coding sequence of chick Wnt-7a was obtained by rapid amplification of cDNA ends (RACE) from stage 21 limb bud cDNA according to the manufacturer’s instructions (GIBCO-BRL, Rockville, MD, USA). The entire coding sequence of chick Wnt-3a was obtained by RACE according to the previous method ( Nohno et al. 1997 ).
In situ hybridization
Chick embryos were staged according to Hamburger and Hamilton (1951). Whole-mount in situ hybridization was carried out as described previously ( Wilkinson 1992) using the following chick cDNA. An antisense RNA probe for Fz-10 has been reported (Kawakami et al. 2000). Shh was used as described previously ( Nohno et al. 1995 ; Wada et al. 1999 ). Chick Lmx1 was obtained by PCR based on the published sequence ( Riddle et al. 1995 ). Two-color in situ hybridization was performed according to the published method by using fluorescein isothiocyanate (FITC)-labeled Fz-10 probe and digoxigenin (DIG)-labeled Shh probe ( Hauptmann & Gerster 1994). Corresponding sense probes were used as controls and showed no significant hybridization signal under the same conditions as the antisense probes.
Injection of synthetic mRNA into Xenopus embryos
The entire coding sequences of chick Wnt-7a, chick Wnt-3a and Fz-10 were subcloned into pCS2(+) vector and the corresponding mRNA was synthesized using the SP6 mMessage Machine Kit (Ambion Inc., Austin, TX, USA). Microinjection of the RNA was performed as described ( Nishimatsu & Thomsen 1998). Animal caps were explanted at the early blastula stage (stage 8.5) and harvested at stage 11 for RT-PCR analysis. Polymerase chain reaction was carried out for 21 cycles to detect EF1α and 25 cycles to detect Siamois and Xnr3. Sequences of the PCR primers were as described ( Brannon & Kimelman 1996; McKendry et al. 1997 ; Nishimatsu & Thomsen 1998). We repeated the experiments three times to analyze interaction between Fz-10 and Wnts in Xenopus embryos and finally determined Siamois and Xnr3 expression quantitatively by using BAS-2000 imaging plate system (Fuji Photo Film Co., Tokyo, Japan).
To examine the role of Wnt-7a in Fz-10 expression, the dorsal ectoderm of the wing bud was removed at stages 20–21 without disturbing the underlying limb mesenchyme as described ( Yang & Niswander 1995). Because the embryos with dorsal ectoderm removed sometimes displayed defects in limb outgrowth caused by the toxic effects of Nile blue on the apical ectodermal ridge (AER), we treated six or more embryos at once and repeated the experiments at least three times. To misexpress Wnt-7a, chick embryonic fibroblasts were transfected with RCAS retrovirus vector bearing chick Wnt-7a and grafted to the posterior- distal region of the ventral side of stage 20 limb buds. Ectopic expression of Shh in the anterior margin of the limb bud was carried out as described using virus-resistant wild-type embryos as hosts at stages 20–21 ( Kawakami et al. 1996 ; Wada et al. 1999 ). Removing the posterior AER was done on stage 20–21 limb buds as previously reported using fine tungsten needles ( Nohno et al. 1997 ). The embryos were harvested at desired stages for gene expression analysis.
Cloning of chick Wnt-7a and Wnt-3a
To analyze the role of Fz-10 during chick limb development, first we isolated the entire coding sequence of chick Wnt-7a and Wnt-3a. These Wnts are possible ligands for Fz-10, because the Wnt-7a expression domain ( Dealy et al. 1993 ) and the Wnt-3a expression domain ( Kengaku et al. 1997 ) are in close proximity to that of Fz-10. Both Wnt-7a and Wnt-3a have 24 conserved cysteine residues that are characteristic of Wnt proteins ( Fig. 1). The deduced amino acid sequence of chick Wnt-7a has 92.8% identity to the mouse counterpart, while that of chick Wnt-3a has 81.8% identity.
Frizzled-10 transcripts colocalize with those of Shh in the limb bud
Our previous report has shown that Fz-10 expression in the developing limb mesenchyme is restricted to the posterior-dorsal region (Kawakami et al. 2000). The expression in the posterior-distal region has close proximity to that of Shh expression, which is maintained by Wnt-7a and fibroblast growth factors. Therefore, it has been suggested that Fz-10 acts in a Wnt signaling pathway that is involved in Shh expression. To examine this hypothesis, we compared the expression patterns of Fz-10 and Shh during limb development. Fz-10 is expressed weakly in the posterior region of the limb bud at stage 19 ( Fig. 2A) and the signal colocalizes with Shh only in the dorsal domain ( Fig. 2B). At stage 21, Fz-10 is expressed in the posterior-distal region of the limb bud and it is restricted to the dorsal region ( Fig. 2C,D). At this stage, it is clearly observed that the transcripts colocalize with Shh ( Fig. 2E) in the dorsal domain ( Fig. 2F). At stage 23, Fz-10 expression domain expands distally, but it still colocalizes with Shh in the posterior-distal region ( Fig. 2G).
Frizzled-10 interacts with Wnt-7a in the Xenopus embryo
Colocalization of Fz-10 and Shh transcripts in the posterior-dorsal limb suggested that Fz-10 acts as a receptor for Wnt-7a during chick limb development. In addition, the expression in the distal region also suggested that Fz-10 may act as a receptor for Wnt-3a. Therefore, we analyzed the possible interaction between Fz-10 and these two Wnts using Xenopus embryos. In the animal cap assay, Wnts have been shown to induce expression of Siamois and Xnr3 ( Brannon & Kimelman 1996; McKendry et al. 1997 ; Salic et al. 1997 ). Therefore, if Fz-10 can bind to Wnt-7a or Wnt-3a and transduce their signal, coexpression of both the ligand and the receptor would elevate both Siamois and Xnr3 expression in the animal cap assay. To test this possibility, we injected Wnt-7a, Wnt-3a and Fz-10 mRNA, alone or in combination, and analyzed Siamois and Xnr3 expression by RT-PCR. Co-injection of Wnt-7a RNA and Fz-10 RNA synergistically induced Siamois and Xnr3 expression ( Fig. 3A). The ratio of Siamois/EF1α in animal cap injected with both Wnt-7a and Fz-10 was 6.6-fold and 6.0-fold of single injection of Wnt-7a and Fz-10, respectively. The ratio of Xnr3/EF1α in animal cap injected with both Wnt-7a and Fz-10 was 4.0-fold and 5.6-fold of single injection of Wnt-7a and Fz-10, respectively ( Fig. 3B). In contrast, co-injection of Wnt-3a and Fz-10 did not induce Siamois or Xnr3 ( Fig. 3). These experiments were repeated three times with different doses and similar results were obtained. These results clearly demonstrate that Fz-10 can interact with Wnt-7a and transduce the signal to stimulate the responsive genes, as assayed in the Xenopus embryo.
Signal from the dorsal ectoderm is required for expression of Fz-10
In the chick limb bud, mesenchymal Fz-10 expression is possibly maintained by a signal emanating from the dorsal ectoderm as judged from its expression in the dorsal mesenchyme. To verify this possibility, we removed the dorsal ectoderm and examined the resulting Fz-10 expression. A slight reduction in Fz-10 expression was observed immediately after dorsal ectoderm removal ( Fig. 4A,B; n = 7 of 10), because Fz-10 is expressed in the dorsal ectoderm in addition to the posterior-dorsal mesenchyme. A reduction in Lmx1 expression was not observed (data not shown), confirming that our manipulation did not disturb the underlying tissue. At 12 h later, however, mesenchymal expression of Fz-10 was drastically decreased as compared with the 0 h control and with the contralateral limb bud of the same embryo ( Fig. 4B–D; n = 12 of 17). These results suggest that the signal from the dorsal ectoderm is implicated in the maintenance of mesenchymal Fz-10 expression.
The endogenous expression of Fz-10 in the dorsal region also suggested that signal from the dorsal ectoderm is involved in induction of Fz-10. The most likely candidate molecule for this is Wnt-7a. To test this, we implanted Wnt-7a-expressing cells to the posterior- distal region of the ventral side of the limb bud. We occasionally observed expression of Fz-10 on the ventral side ( Fig. 4E,F), which is not observed in the intact limb bud at this stage. However, expression was restricted to a small domain distal to the grafted cells. The control experiment, in which the limb bud was hybridized with Lmx1 probe after the grafting, indicated that grafted cells can induce genes downstream of Wnt-7a on the ventral side ( Fig. 4G). These experiments suggest that Wnt-7a is involved in induction of Fz-10, but another factor could be involved in this process.
Sonic hedgehog maintains Fz-10 expression in the posterior mesenchyme
To determine the role of Shh and the AER in the maintenance of Fz-10 expression in the posterior mesenchyme, we removed the posterior AER from the stage 20–21 limb bud. As reported previously, expression of Shh was downregulated shortly (6 h) after the manipulation ( Fig. 5A). We did not observe major change in the expression of Fz-10 at this time ( Fig. 5B). Fz-10 expression was downregulated 12 h after the AER removal ( Fig. 5D), when Shh expression was almost completely downregulated ( Fig. 5C). These results suggest that the downregulation of Fz-10 could occur due to loss of Shh expression, which precedes the downregulation of Fz-10 after AER removal. These experiments also indicate that the intact AER does not influence Fz-10 expression, because Fz-10 was not downregulated 6 h after the AER removal.
Implanting Shh-expressing cells induced Fz-10 expression in the anterior region of the limb ( Fig. 5E) and maintained it in the dorsal side (Kawakami et al. 2000). These results suggest that both Shh and the signal from the dorsal ectoderm, presumably Wnt-7a, cooperate to induce and maintain Fz-10 expression in the dorsal region of posterior-distal limb mesenchyme.
Expression pattern of Fz-10 suggests involvement in Shh expression
A number of Wnt genes are differentially expressed during chick limb development ( Dealy et al. 1993 ; Kengaku et al. 1997 ; Kawakami et al. 1999 ; Hartmann & Tabin 2000). One of these, Wnt-7a, is expressed in the entire dorsal ectoderm ( Dealy et al. 1993 ) and plays two different roles, in dorsal–ventral patterning of the limb mesenchyme ( Parr & McMahon 1995), and maintenance of Shh expression in the ZPA ( Parr & McMahon 1995; Yang & Niswander 1995). However, it is still unknown how Wnt-7a exhibits these different activities. One possible explanation would be that two different receptors for Wnt-7a are expressed and mediate the activities independently. The identification of the Fz family as Wnt receptors has begun to unravel this question. Several members of the Fz family, including Fz-1, Fz-2, Fz-7 ( Kengaku et al. 1997 ; Hartmann & Tabin 2000), Fz-4, Fz-6 and Fz-10 ( Nohno et al. 1999 ; Kawakami et al. 2000), are expressed during chick limb development. Among them, Fz-10 is expressed in the posterior-dorsal mesenchyme in close proximity to the dorsal ectoderm where Wnt-7a is expressed ( Dealy et al. 1993 ; Parr et al. 1993 ). To transduce the Wnt signal properly, the Fz genes should be expressed adjacent to the Wnt-producing cells, because Wnt protein can be trapped easily by the extracellular matrix and not diffuse over a long distance ( Cadigan & Nusse 1997). Therefore, the expression pattern of Fz-10 raises the possibility that Fz-10 functions as a receptor for Wnt-7a in the limb bud. If Fz-10 acts to maintain Shh expression in the posterior mesenchyme, its transcripts should colocalize with those of Shh. To examine this possibility, we compared the spatial and temporal expression patterns of Fz-10 and Shh. The two-color in situ hybridization revealed that Fz-10 and Shh transcripts colocalize in the dorsal mesenchyme ( Fig. 2B,F,G). These expression patterns suggest that Fz-10 is involved in the pathway in which Wnt-7a maintains Shh expression in the dorsal mesenchyme.
Frizzled-10 can transduce Wnt-7a signaling in the Xenopus embryo
At stages 23–26, Fz-10 expression extends to the distal and anterior regions of limb buds (Kawakami et al. 2000). This suggests that Fz-10 may interact not only with Wnt-7a, but also with Wnt-3a, which is expressed in the AER. To analyze the interaction between Fz-10 and these Wnts, we used Xenopus embryos. The Siamois gene and Xnr3 gene are Wnt-responsive marker genes in the Xenopus animal pole ( Brannon & Kimelman 1996; McKendry et al. 1997 ; Salic et al. 1997 ). The synergistic effect on the activation of Siamois and Xnr3 expression by Wnt-7a and Fz-10 indicated that Fz-10 can act as a receptor for chick Wnt-7a, but not for chick Wnt-3a. This also suggests a possible specificity between Fz and Wnts.
Regulation of Fz-10 expression by signaling from the dorsal ectoderm and Shh
If Fz-10 is involved in the interaction between Wnt-7a and Shh in the limb bud, it may be regulated by these factors. The expression pattern of Fz-10 after removal of the dorsal ectoderm indicated that the dorsal ectoderm is necessary for the expression of Fz-10, because it was downregulated after the operation ( Fig. 4A–D). Fz-10 was also downregulated after removal of the posterior AER, suggesting that it is maintained by Shh ( Fig. 5A–D). Removal of posterior AER resulted in the rapid decrease of Shh expression, as previously reported ( Laufer et al. 1994 ; Niswander et al. 1994 ), which preceded downregulation of Fz-10. From these observations, Fz-10 appears to be maintained by Shh. However, signals from the AER are not involved in maintaining Fz-10 expression in the posterior mesenchyme. Therefore, two signaling centers, the dorsal ectoderm and the ZPA, regulate Fz-10 expression in the posterior mesenchyme in the limb bud.
The endogenous posterior expression of Fz-10 ( Fig. 2) and its ectopic expression in the anterior mesenchyme after Shh treatment ( Fig. 5E) indicate that Shh can induce Fz-10 expression. The expression of Fz-10 after grafting Wnt-7a-expressing cells to the ventral side of the posterior-distal region of the limb is restricted to the region where Shh is expressed. Therefore, it is likely that Shh and Wnt-7a in combination induce Fz-10 expression. This is further supported by the temporal expression patterns in the normal limb bud. Wnt-7a expression starts at stage 15 ( Riddle et al. 1995 ), Shh expression starts at stage 17 ( Riddle et al. 1993 ) and Fz-10 expression in the limb mesenchyme starts at stage 18 (Kawakami et al. 2000). In conclusion, Shh and a signal from the dorsal ectoderm, presumably Wnt-7a, appear to cooperatively act on induction and maintenance of Fz-10 expression in the dorsal region of posterior-distal mesenchyme.
Role of Fz-10 in limb patterning
The present studies provide evidence that Fz-10, in response to Wnt-7a, could positively influence Shh expression. This idea is also supported by the fact that Fz-10 expression colocalizes to regions where Shh is expressed and involved in the epithelial–mesenchymal interactions, such as the branchial arches and feather buds (Kawakami et al. 2000). Shh is thought to have an essential role in epithelial–mesenchymal interaction ( Nohno et al. 1995 ; Wall & Hogan 1995; Helms et al. 1997 ). Fz-10 may be involved in these epithelial– mesenchymal interactions, because several Wnt members and Shh show overlapping expression patterns in the branchial arch, feather bud and surface ectoderm ( Dealy et al. 1993 ; Chuong et al. 1996 ; St-Jacques et al. 1998 ). However, Fz-10 is not necessarily involved in dorsal–ventral patterning, because it is not expressed in the entire dorsal mesenchyme. This suggests that another Fz family member is involved in Wnt-7a signaling during dorsal–ventral patterning of the limb mesenchyme.
We are grateful to C. Komaguchi and A. Shiga for technical assistance. We also appreciate the editorial assistance of J. K. Ng. This work was supported in part by Grants-in-Aid from the Japanese Ministry of Education, Science, Sports and Culture to T. N., Y. K. and N. W. and by Research Project Grants from Kawasaki Medical School to T. N.